' .. \ .. VOLUME III SPECIAL INVESTIGATIONS AND EXPERIMENTAL DATA ·~it!..). 'I" .. 1'~ "" .. "\ CONTENTS PAGB INTRODUCTION 8 19 PART I SPECIAL INVESTIGA nONS CHAPTER 1. THE BAHR EL JEBEL FLOOD-PLAIN BETWEEN JUBA AND BOR I. OBJECTIVES OF SURVEY 823 2. ACCOUNT OF SURVEY 823 METHODS. 824 RESULlS 824 3. JUBA TO BOR: GENERAL 825 4. THE ALIAB VALLEY 825 T OPOGRAPHY 825 HYDROLOGY 828 SOILS 83 1 VEGETATION 832 LAND UTILIZATION 834 FISHERIES ... ... ... -.. 835 EFFECTS OF THE EQUATORIAL NILE PROJECT .. 838 5. THE MONGALLA- GEMMEIZA TOICH 839 TOPOGRAPHY 839 HYDROLOGY 840 SOILS 842 VEGETATION 843 LAND UTILIZATION 844 FISHERIES . 844 EFFECTS OF THE EQUATORIAL NILE PROJECT 845 6. CONCLUSIONS 846 CHAPTER 2. AN ANALYSIS OF THE WHITE NILE FLOOD BETWEEN MALAKAL AND RENK by J. W . WRIGHT. M.A., F.R.I.C.S. INTRODUCTION ... .. . 849 PRESENT CONDITIONS ON THE WIIlll! NILB .. . 849 EFFECTS OF PROJECTED CONTROL WORKS 850 EsTIMA 1l!S OF THE FLOOD-PLAIN AREA 850 SECTION I . THE THEORY OF THE IDEALIZED TROUGH 851 IDI!ALIZATION OF A SINGLE CROSS-SECTION ... ... ... 851 APPUCAT ION TO SEVERAL CROSS-SECTIONS AND DERIVATION OF A MEAN .. 852 THE IDEALIZED TROUGH ... .. . 853 CiIARAC'IERlSTICS OF THE IDEALIZED TROUGH 854 OlII1JNE OF THE METHOD OF ANALYSIS USED 855 INFLow ...... .. . .. . 856 COMMEN1S ON THE ANALYSIS IN THE FIRST INTERIM REpORT 856 THE INDMDUAL Ill!MS OF THE WATER ACCOUNT ... 857 SECTION II. CUMULATIVE OBSERVED LOSS 858 PERIOD COVERED BY THE OBSERVATIONS 858 MBTHOD OF CoMPUTATION ... 858 MBTHODS OF MEAstJREMENT OF THE DISCHARGES . . . 859 ). . PROBABLE ERRORS OF OBSeRVED lossES (AssUMING No SVS1l!MATIC ERRORS) 859 '!HE POSSmlLlTY AND EFFECTS OF SYS1l!MATIC ERRORS 859 , PAG~ ( SECTION m. TROl}GH VOLUME 861 PREVIOUS EsnMA TES 861 METHOD OF CALCULATION 862 " COMPUTATION OF MEAN G AUGES 863 · PROBAB LE ERRORS 864" TRIBUTARY KHORS 864 SECTION IV. RAINFALL . 867 SOURCES OF ERROR 867 DATA AND METHOD OF CALCULATION 868 SECTION V. EVAPORATION 868 RISING STAGE 868 FALLING STAGE 869 METHOD OF CALCU LA llON 870 D,SCUSS,ON OF RESULTS .. 870 METHOD OF EsTIMATION IN THE ANALYSES .. . 871 SECTION VI. ABSORPTION 871 PREVIOUS EsTIMATES 871 METHOD OF CALCULATION 872 SELECTION OF YEARS OF No INFLOW 872 C ALCULATION Of A VERAGE DEPTH 873 CONDITIONS DURING THE fuSI NG STAGE 873 SECTION VII. ANALYSES OF INDIVIDUAL YEARS 875 INTRODUCTION ... .. . ... .. . 875 COMPLETE ANALYSIS OF ONE YEAR (1946-47) .. . 877 SUMMARY ANALYSES OF FIFTEEN FLOODS (1928-46) 885 SECTION VIII. I NFLOW AND THE EVIDENCE FOR IT 902 MEASURED DISCHARGEs OF TRIBUTARIES 902 OBSER VEO INFLOW 902 1929-30 FLOOD 903 1938-39 FLOOD 903 1942-43 FLOOD 904 1946-47 FLOOD 904 SECTION I X. DRY SEASON I NFLOW AND POSSIBLE SOURCES 906 ERRORS IN THE OBSERVED OR COMPUTED LOSSFS 906 BARO SPILL-METHOO OF CALCULATION 906 POSSIBLE ROUTES 907 APPENDIX. THE EFFECTS OF AN ERROR OF ONE METRE IN THE LEVELS OF CROSS-SECTION No. 4 ON THE TABLES OF FLOOD-PLAIN AND SURFACE AREAS AND OF VOLUMES ON THE WHITE NILE 908 CHAPTER 3. AN ANALYSIS OF THE SOBAT FLOOD by J. W. WRlGHT, M.A., F.R.l.C.S. INTRODUCTION 913 MAIN FEATURES OF THE SOBAT 913 SUMMARY OF RESULTS ... 914 H,STORY OF THE PRESENT INVESTIGATION 915 FIRST ANALYSIS OF THE SOBAT FLOOD 915 THE CoNCEPT OF THE INCISED TROUGH 915 RECONSIDERATION OF ABSORPTION DEPTH 915 SECTION I. THE SHAPE OF THE SOBAT VALLEY 916 THE LOWER SOBAT 916 THE UPPER SOBAT 917 (i) THE I NCISBD TROUGH ... 917 (a) LOSSES ON THE LOWER SOBAT 918 (b) GAINS FROM AND LOSSES TO TRIBUTARIES ... 918 (c) NET EVAPORATION Loss ON THE WHOLE SOBAT 918 (d) ABsORPTION DEPTH ON THE WHOLE SOBAT 918 METHOD OF DERIVING MeANS ... 919 CALCULATION OP AREA AND VOLUME TABLES 919 COMPARISON WITH AIR SURVBY 919 (ii) THE SURROUNDING PLAIN 919 DERIVATION OF FORMULAE... ... 920 CALCULATION OP MeAN PLAIN SLOPE 921 DISCUSSION OP REsuLTS 921 1935-36 FLOOD ... ... .. . ... ... .. . ...... 922 1946-47 FLOOD .. . ... .. . ... ... ... ...... 922 CALCULATION OF SURFACE AREA AND VOLUME TABLES FOR HIGH FLOODS 922 ~ PAG" SECfION I!. ANALYSIS OF THE SOBAT FLOOD 93 1 COMPUTBD LOSSES 93 1 TROUGH VOLUMB ." 93 1 RAINFALL AND EVAPORATION ~3 1 AllSORPTION 932 OllSERVED LOSSES 932 MAIN RIVER DISCHAR GES 932 TRIBUTARY DISCHARGES 932 DISCUSSION OF RESU LTS.. 933 Low FLOODS 933 HIGH FLOODS. .. 933 CONCLUSIONS .. . 934 FUTURE CONTROL OF THE SOUI\T 934 SECTION m. THE REGIMES OF THE TRIBUTARIES 962 (i) THE TwALOR 962 (ii) THE WAKAU 963 GENERAL 963 MAIN DEDUCTIONS 963 BEHAVIOUR DURING 1946 flOOD 964 (iii) OTHER TRIB UTARIES .. . 964 PROPORTIONS OF HIGH LOSSES AND GA INS CAR RI ED BY NORTH.ERN TR JlJUTAR IES.. 964 ARGUMENTS AGA INST SUPPOSED NORTHERN SPilL AN D SOUTHERN I NFLOW 965 ARGUMENTS AGAINST SUPPOSED LARGE EXCHANGES WITH MA CHAR MARSHES 965 CONCLUSIONS 966 CHAPTER 4. THE MACHAR MARSHES I. I NTRODUCTION .... 97 1 2. HYDROLOGY ·972 W ATER SUPPLIES 972 SPiLL FROM THE BARD 972 THE EASTERN TORRENTS 973 RAINFALL 974 SPILL FROM THE SOBAT WEST OF THE B ARo-PIBOR CONFLUENCE 975 LoSSES FROM THE SWAMP 975 DRAlNAGE TO THE WHITE NILE 975 THE KHOR ADAR 975 THE KHOR WOL 977 DEDUCTIONS FROM THE WHITE N ILE A NA LYSIS 977 DRAINAGE TO THE SOBAT 978 THE MACHAR MARSHES .. 978 EVAPORATION AND THE GENERAL W ATER B ALANCE 978 MOVEMENT OF WATER THROUGH THE S\VAMPS 979 AIR SURVEY 979 GROUND RECONNAlSSANCE 979 OllSERVED DISCHARGE OF THE WAKAU 979 OBSERVED DISCHARGES OF WHITE NILE KHORS 1946-47 980 3. OTHER ASPECTS OF THE AREA . .. 980 GENERAL INFORMATION ... 980 DlNKA CoUNTRY ... 982 R OUTES fNTO AND ACROSS THE M ACHAR 983 SoUTHERN ROUTES 983 NORTHERN ROUTES ... 984 PART II EXPERIMENTAL DATA CHAPTER 5. AGRICULTURAL EXPERIMENTS 1. INTRODUCfION 987 OBJECTS OF THE EXPERIMENTAL WORK 987 ORGA"NIZATION AND DIFFICULTIES 987 J PAGE ,.., 2 . SUMMARY OF PREVIOUS EXPERIMENTAL WORK IN THE FLOOD REGION 987 CEREALS 987 SORGHUM 987 MAIZE 988 . . BULRUSH MILLET 988 OIL CROPS 988 SESAME . !188 SUNFLOWER 988 LEGUMINOUS CROPS 988 GROUNDNUTS •.. 988 OTHER LEGUMES 989 FIBRE CROPS 989 COTTON .. 989 OTHER FIBRE CRoPS .. . 990 3. MALAKAL EXPERIMENTAL FARM 990 ENVIRONMENT 990 SOILS 990 CLIMATE 990 UTlUZATION OF HIGH LA ND 991 EXPERIMENTS IN IMPROVEMENT OF HUSBANDRY 991 MANURIAL AND SP AC ING EXPERIMENTS. , 99 1 EXPERIMENTS IN D ATES OF PLA NTING AN D WEEDI NG OF RAIN-GROWN COTTON 993 MIXED CROPPING E X PERJMENTS 994 EXPERIMENTS IN nrE CONTROL OF STRIGA HERMO NTHICA 994 EXPERJMENTS IN CROP INTRODUCTION 995 CEREALS 995 SORGHUM ... 995 M AIZE 996 BULRUSH MILLET 996 FINGER MILLET 996 MISCELLANEOUS CEREALS 997 OIL CRops 997 SESAME 997 SUNFLOWER 997 HYPTIS SPICIGERA 997 SAPPLOWER 998 CASTOR • . . 998 LEGUMINOUS CROPS 998 GROUNDNUTS 998 SOYA BEAN 1000 COWPEAS . . . . .. 1000 MISCELLANEOUS LEGUMINOUS CRoPS 1000 FIBRE CROPS 1001 COTTON 1001 JUTE 1003 DECCAN HEMP ... 1003 FORAGE CROPS 1003 UTILIZATION OF INTERMEDIATE LAND 1005 DRAINAGE EXPERIMENTS . . . 1005 UTILIZATION Of K ILLfFER RIDGED LAND EXPERIMENTS . .. . . . 1005 EXPERIMENTS ON THE INTRODUCTION Of NEW CROPS AND VAR IETIES ON KILLlFBR RIDGES 1006 CEREALS 1006 O IL CROPS 1007 LEGUMTNOUS CROPS 1007 FIBRE CRops 1008 FORAGE CROPS ... ... •. . . .. . .. 1008 EXPERIMENTS ON SUBSIDIARY IRRIGATION OF CRops 1009 RICH ••. . . . 1009 CoARSE FIBRES 1009 JUTS 1012 DECCAN HEMP ... 1012 SUNN HEMP .. . ... .. . . . . 1012 UTILIZATION OP THE RIYBRAlN FLOOD-PLAIN 1012 RICH .. . 1012 CoARSE FIBRES 1013 JUTB ... 1013 DECCAN HEMP 1013 SUNN HEMP lOB D ISEASES AND PEsTS 101.3 PAGE 4. OTHER EXPERIMENTAL FARMS 1015 ARrBLDEK .. . 1015 CEREALS 101 6 OIL CRoPS 1017 LEGUMINOUS CRops 1017 FIBRE CRops 1017 FANGAK 1017 KODOK 101 8 CHAPTER 6. GRASSLAND EXPERIMENTS I. INTRODUCTION.. 1021 2. NATURAL PASTURES... 1021 THE EASTERN PLAIN 1021 NATURE OF THE INVESTIGATION 1022 AlR RECONNAISSANCE... . .. 1023 THE COLLECTION AND ANALYSIS OF GRASS SAMPLES 1023 GRASS SAMPLING UNITS: SUMMARY 1024 THE PENGKO TEST AREA 1029 CONCLUSIONS 1033 THE RlYER FLOOO-PLAIN 1033 NATURE OF THE INVESTIGATION 1033 THE COLLECTION AND ANALYSIS OF GRASS SAMPlES 1034 GRAZING EXPERIMENTS 1038 SHALLOw-FLOODED PASTURE 1038 DEEP-FLOODED PASTURE ... 1040 THE SroCK-CARRYING CAPACITIES OF RIVERAIN PASTURES 1041 SHALLOW-FLOODED PASTURE 1041 DEEP-FLOODED PASTURE 1042 3. MISCELLANEOUS GRASSLANDS 1042 THE FLOOD REGION 1043 GONIO GRAZING TRIAL 1043 KODOK GRAZING TRIAL 1043 NAGDIAR GRAZING TRIAL 1044 CONCLUSIONS ... 1044 THE SEMI-ARID REGION ... 1045 JEBEL MEGaNlS EXPERIMENT 1046 FIRST SEASON (1950) 1046 GRASS SAMPLES FOR ANALYSIS 1046 HAY SAMPLES ... 1046 SECOND SEASON (1951) 1046 GRASS SAMPLES FOR ANALYSIS 1046 HAY SAMPlES 1047 INTERPRETATION OF RESULTS AND ANALYSES 1047 SABA-AsUDA EXPERIMENT 1052 4. IRRIGATED PASTURE. . . 1053 WATER-DuTY: RATE AND FREQUENCY 1053 EXPERIMENT A .. 1053 FIRST SEASON (1950-51) 1054 SECOND SEASON (1951-52) 1054 EXPERIMENT B ... 1056 FIRST SEASON (1950-51) ... 1056 SECOND SEASON (1951-52) 1056 MANAGEMENT ... .. . 1059 FIRST SBASON (1950-51) .. . 1059 IRRIGATION 1059 INTENSITI' OF STOCKINO 1060 • CONDITION OF CAT ILE 1060 HERBAOE PRODUCTION ... 1062 SECOND SEASON (1951-52) 1063 IRJUOATION 1063 INTENsITY OF STocKlNO. . . 1064 CONDITION OF CATTLE .. . 1064 HeRBAOE PRODUCTION .. . 1066 CoNCLUSION 1067 J PAGB --. GRASS AND LEGUME TRIALS . . . 1067 INDIGENOUS GRASSES AND LEGUMES 1067 PHRAGMITES COMMUNIS 1067 · HYPARRHENIA RUFA .. . 1067 SETARIA INCRASSATA .. . 1067 · [SCHAEMUM BRACHYATHERUM 1068 SETARIA SP. (PROBABLY S. LYNSElI) .. 1068 PANICUM REPENS 1069 EeH/HoeHLOA PYRAMJDALlS . . . 1069 ECHJNOCHLOA STAGN/NA 1069 VOSSIA CUSP/DATA 1069 ECHINOCHLOA COLONA 1069 SPOROBOLUS PYRAMIDALIS 1070 VET/YERIA NIGR/TANA 1070 CYNODON DACTYLON ... 1070 HYPARRHENIA SP. (H. PSEUDOCYMBARIA) . .. 1070 PAN/CUM SPP. ... . .. 1070 INTRODUCED GRASSI'S AND LEGUMES 1070 GRASSI'S 1071 CHLORIS GAYANA 1071 ERAGROSTIS SP. . .. 1011 TEFP GRASSES 107 1 BOTHRIOCHLOA S. .. ·1071 BRACHIA RIA BRIZANTHA 1071 PAN/CUM SPP . . .. 1071 PENNISETUM SP• . 1071 PASPALUM SPP. 107I SETARIA SPLENDIDA 1072 LEGUMES 1072 TRIFOLI UM SUBTERRANEUM 1072 TRIFOLIUM JOHNS TO N! 1072 MED/CAGO SATIVA (LUCERNE, ALFAlfA) 1072 MEDICA GO LUPLULINA 1072 ALYSICARPUS SPP. (ALICE CLoVERS) 1072 MELILOrlS SPP. (SWEET CLoVERS) 1072 LESPEDEZA SPP. 1072 STYLOSANTHES SPP. 1072 DESMODIUM S. .. 1073 GLYCINE JA VANICA 1073 INDIGOFERA SPP. 1073 CENTROSEMA SPP. 1073 CROTOLARIA INT£RMEDIA 1073 PHASEOLUS L A THYRO/DES 1073 PHASEOLUS SEMI·£RECTUS (KOROOPAN PEA) 1073 PHASEOLUS TRILOBUS .. 1073 VIGNA SPP. 1074 DOL/CHOS SPP. 1074 LUP/N US LUTENS 1074 DESMANTHUS DEPRESSUS 1074 CAJANUS BICOLOR 1074 PUERARIA PHASEOLOJDES 1074 FODDER SHRUBS 1074 5. PALATABILITY TRIALS 1074 MBTHOD 1074 RESULTS 1074 LIST OF TABLES NUMBBR TITLE PAGP. 375 ALIAR VALLEY: HYDROLOGICAL DETAILS 829 376 ALLAR VALLI!Y : TOPOGRAPHICAL DETAILS- B AN K 829 377 AUAB VALLBY: TOPOGRAPHICAL DETAILS- TOICH 830 378 DURATION OF IO-DAY MEAN DISCHARGES AT MONGALLA 1946-50 . 830 379 ALIAB V ALLEY (NORTH-EAST OF RIVER ALlAB): AREAS FLooOEO RELATEO TO M ON- GALLA DISCHARGES ANO A VBRAGB DURATION 830 380 ALIAR VALLEY CROSS-SECTION No. 13: RESULTS OF MECHANICA L AND C HEMICA L ANALYSIS OF SOIL SAMPLES 833 381 ALLAR VALLEY: PRESeNT DISTRIBUTION OF VEGETATION SPECIES 834 382 ALIAB VALLEY: DURATION OF FLOODING OF VEGETATION SPECIES 834 383 TOTAL FISH CATCHES T ABULATED BY SPECIES 836 384 FISH CATCHES T ABULATED BY FISHING SITES 837 385 FIsH CATCHES TABULATED BY REACHES 838 386 ALLAR VALLEY: FUTURE DISTRIBUTION OF VEGETATION SPECIES 838 387 ALLAR VALLEY (NORTH-EAST OF RIVER ALlAB) : PRESENT AND FUTURE DISTRIBUTION OF VEGETATION SPECIES 838 388 MONGALLA-GEMMEIZA TOICH ,' HYDROLOGICAL DETAILS 841 389 MONGALLA-GBMMElZA TOICH,' T OPOGRAPHICA L DETAI LS- BANK 841 390 MONGALLA-GBMMEIZA TOICH ,' TOPOGRAPHICAL DETAILS- ToICH 84 2 391 MONGALLA-GEMMEIZA TOICH,' AREAS FLOODED RELATED TO MONGALLA DISCHARGES AND AVERAGE DURATION 842 392 MONGALLA-GEMMEIZA TOICH ,' PRESENT DISTRIBUTION OF VEGETATION SPECIES ... 843 393 MONGALLA-GEMMEIZA TOICH,' DURATION OF FLOODING OF VEGETATION SPECIES 844 394 MONGALLA-GEMMEIZA TOICH,' FUTURE DISTRIBUTION OF VEGETATION SPECIES 845 395 MONGALLA-GEMMElZ~ TOICH,' PRESENT AND fuTURE DISTRIBUTION OF V EGETATION ' SPECIES .. . .. . 846 396 i'REuMINARY SUMMARY OF FLOODS ON THE WHITE NILE BETWEEN M ALAKAL AND RENK. 1928-47 861 397 SURFACE AREAS BETWEEN MALAKAL AND MELUT AN D BETWEEN MELUT AND RENK ~ FOR VARIOUS MEAN GAUGE V ALUES 865 398 TROUGH VOLUMES ABOVE MEAN GAUGE 9·80 (i) MALAKAL TO MELUT 865 (il) MeLUT TO RENK 866 399 PROBABLE ERRORS OF TROUGH VOLUMES A.BOVE MEAN GAUGe 9 ·80 : 867 (I) MALAKAL TO MELUT 867 (il) MELUT TO RENK 867 400 AVERAGE EVAPORATION PROM AN OpeN WATER SURFACE 869 401 CALCULATION OF EVAPORATION DEPTH 871 402 (i) MALAKAL GAUGE AND APPARENT ABsORPTION IN SOME YEARS .. . 874 (il) CALCULATION OF AVERAGE AoSORYTION DEPTH 875 403 CoMPLETE ANALYSIS OF ONE YEAR (1946-47): (i) GAUGe-READINGS AND TROUGH VOLUMES 879 (il) RAINFALL AND EVAPORATION 881 • (iii) CuMULATIVE OBSERVED AND COMPUTED LOSSES . .. 883 404 SUMMARY ANALYSIS OF THE YEAR 1928- 29 886 405 SUMMARY ANALYSiS OF THE YEAR 1929- 30 887 406 SUMMARY ANALYSIS OF THE YEAR 193~31 888 407 SUMMARY ANALYSIS OF THE YeAR 1931- 32 889 408 SUMMARY ANALYSIS OF THE YEAR 1932-33 890 409 SUMMARY ANALYSIS OF THE YEAR 1936- 37 89 1 410 SUMMARY ANALYSIS OF THE YEAR 1937-38 892 411 SUMMARY ANALYSIS OF THE YEAR 1938-39 893 412 SUMMARY ANALYSIS OF THE YEAR 1939-40 894 413 SUMMARY ANALYSIS OF THE YEAR 1940-41 895 414 SUMMARY ANALYSIS OF THE YEAR 1941-42 896 415 SUMMARY ANALYSIS OF THE YEAR 1942-43 897 416 SUMMARY ANALYSIS OF THE YEAR 1943-44 898 411 SUMMARY ANALYSIS OF THE YEAR 1944-45 899 418 SUMMARY ANALYSIS OF THE YEAR 1945-46 900 419 SUMMARY OF SlXTBBN FLOODS BETWEEN MALAKAL AND RENK DURING THE YEARS 1928 TO 1946 901 4~ DEDUCED INFLow INTO THE Wm.TE N1LI! BETWEEN MALAKAL AND RENK IN THE MoNTHS WHW iT .BXCEEDED 40 MILLIONS BETWEEN 1928 AND 1947.. . 905 r-.'UMBER TITLE PAGE · 421 BARO SPILL AND DRY Sl!A50N L>,FLOw ... 907 422 THE EfFECTS ON FLOOD AND TOTAL SURFACE WIDTHS BETWEEN MALAlCAI. AND MaUT OF AN E RROR OP ONtO METRE IN THE REDUCED LEVELS OF CROSS-SECTION NO.4... 909 423 loNGITUDINA L PROFILES OF THE RIVER SOBAT 922 ' 424 MEASURED CRoss-SEenONS ON THE LOWER SOBAT 923 425 S URFACE AJfPARISON OF CO>fPUTED AND OBSERVED LoSSES 942 445 SUMMA RY A NA LYSIS Of THE YEAR 1938- 39: (i) GAUGES AN D RAI NFALL D ATA 943 (ii) OBSERVED DISCHARGES 943 (iii) COMPARISON OF CO'fPUTED AND OBSERVED losses 944 446 SUMMARY ANALYSIS OF THE YEAR 1939-40: (i) GAUGES AND RAlNFALL DATA 945 (ii) OBSERVED DISCHARGES 945 (iii) COMPARISON OF COMPUTED AND OBSERVED LossES 946 447 SUMMARY A NA LYSIS OF THE YEAR 1941 - 42: (i) GAUGES AND RAINFALL DATA 947 (ii) OBSERVED DISCHARGES 947 (iii) COMPARISON OF CoMPUTED AND OBSERVED losses 948 448 SUMMARY A NALYSIS OF THE YEAR 1942-43: (i) GAUGES AND RAJNFALL D ATA 949 (ii) OBSERVED DISCHARGES 949 (iii) COMPARISON OF COMPUTED AND OBSERVED LoSSES 950 449 SUMMARY ANALYSIS OF THE YEAR 194>-44 : (i) GAUGES AND RAINFALL DATA 951 (ii) OBSERVED DISCHARGES 951 (iii) COMPARISON OF COMPUTED AND OBSERVED LoSSES 9 52 NUMBER TITLe PAGE 450 SUMMARY ANALYSIS OF THE YEAR 1944-·45: (i) GAUGES AND R AINFALL D ATA 953 (ii) OBSERVED DISCHARGES ... 953 (iii) COMPARISON OF COMPUTeD AND OBSERVED LossES 954 451 SUMMARY ANALYSIS OF THE Y EA R 1945--46 : (i) G AUGES AND RAI NFALL D ATA 955 (ii) OBSERVED DISCHARGES 955 (iii) COMPARISON OF COMPUTeD AND O BSERVED LOSSES 956 452 SUMMARY ANALYSIS OF THE YEAR 1946-47: (i) GAUGES AND RAINFALL DATA 957 (ii) OBSERVED DISCHARGES 957 (iii) CoMPARISON OF COMPUTeD AND OBSE RVED LOSSES 958 453 SUMMARY ANALYSIS OF THE YEAR 1947- 48 : (i) GAUGES AND RAINFALL DATA 959 (ii) OBSERVED DISCH ARGES 959 (iii) COMPARISON OF COMPUTED AND OBSERVED LOSSES 960 454 COMPARISON OF COMPUTED AND OBSERVED LOSSES: AVERAGE FOR FI VE Low FLOODS 961 4~5 COMPARISON OF COMPUTED AND OBSERVED l OSSES: AVERAGES FOR SEVEN HIGH FLOODS (1946-47 EXCLUDED) 96 1 456 COMPARISON OF COMPUTED AN D OBSERVED LOSSES: AVERAGE OF TWELVE FLOODS (1946-47 exCLUDED) 962 457 PREDICTED LossES ON THE SOBAT DURING THE TIMELY SEASON FOR VAR IOUS D IS· CHARGES AT SOBAT HEAD 962 458 COMPARISON OF THE DISCHARGES IN TH E PIBO' AND TWALOR 966 459 KHOR WAKAU DISCHARGES AND NASIR G AUGE BEFORE THE REVERSAL OF FLOW .. 967 460 THE RBVERSAL OF FLOW IN THE KHOR W AKAU 967 461 DISCHARGES rN THE KHOR WAKAU AND GAUGE·READINGS AT NASIR DUR ING AND AFTER THE PEAK OF THE FLOOD 967 462 LoSSES AND GAINS AT THE PEAK OF HIGH FLOODS 968 463 LARGE RECORDED DISCHARGES IN THE SOUTHERN TRIBUTARIES 968 464 SPILL ON THE BARO 973 465 THE EASTERN TORRENTS-NORTHERN AREA: EsTIMATE OF RUN·OFF 974 466 RAINFALL ON THE MACHAR MARSHES 974 467 KHOR ADAR: HYDRAULIC DETAILS OF CHANNEL 976 468 KHOR ADAR: DETAILS OF FLOOD· PLAIN 977 469 SUMMARY OF RESULTS OF SORGHUM V ARIETY TRIALS IN UPPER NILE PROVI NCE, 1933-49 988 470 GROUNDNUT YIELDS AT KODOK , 1935-45 988 471 GROUNDNUT YIELDS AT SELIM BANGA, 1935- 36 989 472 EFFECT OF PLOUGHING ON YIELDS OF COTTON ... 989 473 EFFECT OF SPACING ON YIELDS OF COTTON 989 474 SOILS OF THE EAST B ANK FARM 990 475 SOILS OF THE WEST B ANK FARM . 990 476 RAINFALL 195(}"'52 991 477 EFFECT OF FERTILIZERS AND SPACING ON YlELDS OF SORGHUM, 195(}...5 1 991 478 EFFECT OF FERTILIZERS AND SPACING ON YIELDS OF SORGHUM, 1951-52 992 479 EFFECT OF FERTILIZERS ON YIELDS OF JUCE, 1952-53. 992 480 EFFECT OF CATTLE·PEGGING ON YIELDS OF SORGHUM, 195(}"'51 993 481 EFFECT OF CATTLE·PEGGING AND GRASS BURNING ON YIELDS OF SORGHUM AGONO, 1952-53 993 '482 EFFECT OF D ATE OF PLANTING ON YlELDS OF COTTON, 1952-53 993 483 EFFECT OF DATE OF WEEDING ON YIELDS OF COTTON, 1952-53 994 484 EFFECT OF INTeRFLANTING SORGHUM WITH lEGUMES, 1951-52 994 485 EFFECT OF SPRAYING AND HAND-PULLING OF STRIGA HERMONTHICA ON YIELDS OF SORGRUM, 1952-53 994 486 EFFECTOFSPRAYING AND H AND-PuLLING OF STRIGA HERMONTHICAoNYIELDSOF MAIZE, 1952-3 995 487 YIELDS OF SORGHUM VARIETIES, 195(}"'53 995 488 YIELDS OF MAIze VARIETIES, 1951-53 996 489 YIELDS OF BULRUSH MILLET VARlETIES, 195(}"'53 996 490 YIELDS OF FINGER MILLET VARlETIES, 195(}"'53 996 491 YIELDS OF MISCELLANEOUS CEReALS, 195(}"'53 997 492 YIELDS OF SESAME VARIETIES, 1951-53 997 493 YIELDS OF SUNFLOWER V ARIIlTIES, 195(}"'53 997 494 YIELDS OF HYPTIS V AR/ETlES, 1951-53 ... 998 495 YIELDS OF SAFFLOWER VAIUBTIES (LATE SOWN), 1950-52 998 496 YmLDs 9" CASTOR V ARlETlBS, 195(}"'52 ... 998 NUMBER TmE PAGE 497 YIELDS OP G~OUNDNUT VAlU£TlES, 1950-53 999 498 YIBLDS OF SOYA BEAN VARIETrES, 1950-53 1000 499 YI£LDS OF COWPEA VARIETIES, 1950-53 1000 500 YIELDS OF MISCELLANEOUS LEGUMINOUS CROPS, 1950-53 1001. 501 YIELDS OF COTTON VARIETIES (YIELDS OF SEEO COTTON), 1950-53 ld02 502 YIELDS OF SWEET SORGHUM AND SUDAN GRASS, 1950-52 1003 503 Y IELDS OF LEGUMINOUS FORAGE CROPS, 1950-53 1"004 504 Y,ELDS OF FORAGE MIXTURES: 1951-52 ... 1004 505 YIELDS OF STOVERS, 1950-53 1004 506 REsULTS OF CHEMICAL ANALYSIS OF FODDER CROPS AND M IXTURES 1005 507 Y IELDS OF SORGHUM AND M AIZE ON DrFFERENT TyPES OF KILLIFER RIDGES, 1951- 52 1005 508 EFFECT OF I NTERPLANTING ON CROPS GROWN ON KILLI FER RIDGES, 1951- 52 1006 509 YIELDS OF SORGHUM VARIETIES ON KrLLlFER RIDGES, 195 1- 53 1006 510 Y IELDS OF CEREALS ON KILLIFER RIDGES, 1951- 53 1007 51 I YIELDS OF OrL CROPS ON KILLIFER RIDGES, 1951- 52 . 1007 512 YI£LDS OF L EGUM INOUS CROPS ON KlLLlFER R IDGES, 1951- 53 1007 513 Y , ELDS OF COTTON ON KILLIFER RIDGES. 1951- 53 1008 514 YIELDS OF FORAGE CROPS ON KILLIFER RIDGES, 1951-52 1008 515 YIELDS OP FORAGE MIXTURES ON K ILLIFER RIDGES, 1951-52 .. 1009 516 YIELDS OF STOVERS OF LOCAL SORGHUM ANO OF GROUNONUTS, 1951 -53 '1009 517 IRRIGATION ANO T OTAL QUANTITY OF WATER SUPPLIED TO Rr CE IN GROWING SEASON, 1951- 53 . . . 1010 518 YIELDS OF RICE IN THE IRRIGATION TREATMENT TRIAL, 1952-53 1010 519 YIELDS OF RICE VAR IETIES, 1951-53 1011 520 YIELDS OF RICE VARJETrES ON TOICH. 1951- 52 .. . 1012 52 1 YI£LDS OF RICE V ARiETIES ON TOICH, 1952- 53 . . . 1013 522 CROP DrSEASES AT MALAKAL EXPER"IENTAL F ARM, 1950-52 (IOENTIFIED BY THE PLANT PATHOLOG IST, MINISTRY OF AGRICULTURE RESEARCH DI VISION) 1014 523 CROP P ESTS AT M ALAKAL EXPERIMENTAL FARM, 1950-53 • 1015 524 REsULTS OF ANALYSIS OF SOIL SAMPLES FROM ARTELBEK 1016 525 YIELDS OF SORGHUM AT ARIELBEK, 1951- 52 1016 526 YIELDS OF OTHER CEREALS AT ARJELBEK, 1951- 52 1017 527 Y,ELDS OF OIL C ROPS AT ARrElBEK, 1951 - 52 1017 528 YIELDS OF LEGUMINOUS CROPS AT ARIELBEK , 1951-52 .. . 1017 529 YIELDS OF OBSERVATION P LOTS AT FANGAK, 1951- 52 .. . 101 8 530 YIELDS OF OBSERVATION PLOTS AT KODOK, 1952- 53 1018 531 HYPARRHENIA RUFA REGROWTH: ANALYSIS 1025 532 E ASTERN PLAIN SAMPLING UNlTS: AIR-DRIED WE IGHTS, 1951 1026 533 EASTERN PLAIN SAMPLING UNITS : ABsOLUTE D RY WEIGHTS, 1951 1026 534 E ASTERN PLAIN SAMPLI NG U NlTS: DRY MATTER PER FEDDAN, 1951 1027 535 EASTERN PLAIN SAMPLING UNITS: CRUDE PROTEIN PER FEDDAN, 195 1 1027 536 EASTERN PLAIN SAMPLI NG UNITS: AIR-DRIED WEIGHTS, 1952 1028 537 HYPARRHENIA RUFA REGROWTH. SUMMARY OF A NALYSES, 1951 1028 538 ANDROPOGON REGROWTR, 1950-51 1029 539 REGROWTH FROM M IXED P ASTU RE AT MWOT TOT, 1950-5 1 1029 540 PBNGKO TEST AREA: INTENSITY OF STOCKING, 195 1 1031 541 PBNGKO TEST AREA: QUANTITY OF REGROWTH, 1951 103 1 542 PENGKO T EST HERO: M ILK YIELDS, 1951-52 1032 543 PBNGKO TEST HERD: HERD RECORD 1032 544 SHALLOW-FLOODED GRASS: ANALYSIS 1031 545 D EEP-FLOODED GRASS: ANALYSIS 1036 546 ECHINOCHLO A PYRAMIDAL/S : ANALYSIS, DIGESTIBILITY COEFFIClENTS, AND DIGESTIBLB NUTRIENTS OF MA TORE GRASS 1037 547 SHALLOW-FLOODED PASTLrRE : HERBAGE YIELDS 1037 548 DEEP-FLOODED PASTURE: HERBAGE YIELDS 1037 549 SHALLOW-FLOODED PASTURE: SUMMARY OF STOCKING I NTENSITIES, L IVE WEIGHT GAINS, AND AVAILABLE P ASTURE 1039 550 SHALLOW-FLOODED PASTURE: L IVESTOCK WEIGHTS, 1953 1039 551 DEEP-FLOODED PASTURE: SUMMARY OF STOCKING INTENSITIES AND LIVE WElOHT GAINS 1040 552 DEEP-FLOODED PASTURE: LrvESTOCK WEIGHTS, 1953 1041 553 GONIO GRAZING TRrAL : LrVESTOCK WEIGHTS, 1952 1045 554 ANALYSIS OF SOME MATURE SUDAN GRASSES 1045 555 JEEEL MEGElNIS: ANALYSIS OF GRASS SAMPLES, 1950 .. . 1048 556 JEEEL MEGEINIS: ANALYSIS OP GRASS SAMPLES, 1951 .. . 1049 557 JEBEL MEOElNls: ANALYSIS OF ARlSTIDA MUTABILIS HAy, 1950 Ui50 NUMBER TITLE PAGB 558 PBRCJlNTAGE OF ARISTIDA MUTABLIS H AY CONSUMED OY SHEEP 1050 559 JEBEL MEGEINlS: ANALYSIS OF H AY SAM I'LES, 195 1 1051 560 SABA-AsUDA WBLL-CE!NTRE: Nut-mER OF ANIMALS RECORDED WATE.RING, DRY SEASON 1950 1052 561 S ADA-AsUDA WELL-CENTRE: AVERAGE NUM"ER OF ANIM,\LS WATERING D AILY 1053 562 EFFECT OF DIFFERENT WATER ING TREATM.N1~ ON GRASS YIELDS. 1950- 51 1054 563 EFFECT OF DIFFERENT WATERING TREATMENTI ON GRASS Y IELDS. 195 1- 52 1055 564 SETARIA INCRASSATA , ANALYSIS DF GRASS SAMPLES. 1951- 52 1055 565 EFFECT OF DIFFEReNT WATERING TREATMENTS ON GRASS YIELDS, 1950- 51 : ECHINOCHLOA PYRAMIDALIS 1056 566 EFFECT OF DIFFERENT WATER ING TREATMENTS ON GRASS YmLDs. 1951- 52: ECHINOCHWA PYRAMIDALIS 1056 567 ECHINOCHWA PYRAMIDALIS , ANALYSIS OF GRASS SAMPLES, 195 1-52 1057 568 ANALYSIS OF HAY 1057 569 ANALYSIS OF HAY UseD IN FEEDI NG TRIAL 1057 570 ANALYSIS OF SILAGE.. . 1058 571 GRASS YIELDS FROM EXPERIMENT A 1058 572 GRASS YIELDS FROM EXPERIMENT B 1059 573 IRRIGATION: FIRST SEASON. 1950-5 1 1059 574 INTENSITY OF STOCKING, 1950-51 ... 1060 575 INTENSITY OF STOCKING PER PADDOCK PER ROTATION, 1950-51 1061 576 LIVESTOCK WEIGHTI, 1950-51 1062 577 MONTHLY MILK YIELDS, 1950-51 1062 578 HERBAGE PRODUCED PER ROTATION. 1950-51 1063 579 HERBAGE PRODUCED PER PADDOCK PER SEASON, 1950-51 1063 580 IRRIGATION: SECOND SEASON, 1951- 52 1064 581 GRASS Y,ELDS FROM P ADDOCKS, 1951 1064 582 ANALYSIS OF HAY ... 1064 583 LIVESTOCK WEIGHTS, '1951-52 1065 584 HERBAGE PRODUCED PER ROTATION, 1951- 52 1066 585 HERBAGE PRODUCED PER PADDOCK PER SEASON. 1951- 52 1066 586 ANALYSIS OF GRASS FROM PADDOCKS, 1951- 52 1066 587 HYPARRHENIA RUFA, Y , ELDS AT DIFFERENT GROWTH STAGES 1067 588 SETARIA INCRASSA TA, Y,ELDS AT DIFFERE NT GROWTH STAGES 1068 589 !SCHAEMUM BRACHYATHERUM, YIELDS AT DIFFERE NT GROWTH STAGES 1068 590 SETARIA SP. (S. LYNSEIJ): YIELDS AT DIFFERENT GROWTH STAGES . 1068 591 PA NICUM R£PENS, YIELDS AT DIFFERENT GROWTH STAGES 1069 592 ECHINOCHLOA PYRAMIDALIS , YieLDS AT DIFFERENT GROWTH STAGES 1069 593 CYNODON DACTYLON , YIELDS AT DIFFERe NT GROWT H STAGES 1070 594 SUMMARY OF RESULTI OF PALATABILITY TRIALS 1075 ILLUSTRATIONS PLATE (between pages 842 and 843) IX I & 2. T AKlNG SOlL SAMPLES (ALIAB VALLEY). 3. DeEP POOL WITH NILe CABeAGe (PISTIA STRATOIDES) . X I. RIVER FLOOO-PLAIN UNDER WATeR (NeAR BOR). 2 . INLET ON EAST BANK OF BAHR EL JEBeL NBAR BOR. 3. A CLOSe VIEW OF THE EDGE OF THE FLOOD-PLAIN WITH FLOOO SEASON GROWTH OF ECHINOCHLOA STAGNINA RECENTLY EXPOSED. 4. EXPOSED RIVER FLOOD-PLAIN NI!AR JONGLEI WITH COARSE GROWTH OF ECHINOCHLOA PYRAMIDALIS. 5. A CLOSE VIEW OF REGROWTH OP ECHINOCHLOA PRYAMIDALIS AFTeR BURNING. 6. HYPARRHENIA RUFA , COARSE RAINS SEASON GROWTH BEING FIRBD TO ENCOURAGE REGROWTH. PENGKO GRAZING TRI ALS (DECEMBER 1951). 7. HYPARRHENJA RUFA. EASTERN PLAIN: SAMPLING UNIT No. VI (MAY 195 1). 8. ROBUST REGROWTH OF HYPARRHENIA RUFA AFTER BURNING Of COARSE GROWTH. EASTERN PLAIN: SAMPLI NG UNIT No. IV (MAY 1951). 9. NILOTIC cow GRAZING ON DEEP-FLOODED PASTURE IN FADDOI POOL. CENTRAL NUER DISTRICT (DECEMBER 1951). 10. DURA (AGONO) ON ' KILLIFER' RIDGES: NAGDlAR ROAD NBAR MALAKAL (DECEMBER 1951). II. DURA (AGONO) ON • KILLIFER' RIDGES (LEFT) AND ON THE FLAT (RIGHT). 12. THE EFFECTS OF FLOODING. DURA SOWN ON THE FLAT. \3. THE EFFECTS OF DROUGHT. FETBRITA SOWN IN OCTOBER 1951 . TWO MONTHS BEFORE THE PHOTO WAS TAKEN. BOR DISTRICT (DECEMBER 1951), 14. CASTOR SOWN 7.6.51. BOR DISTRICT (DECEMBER 1951). 15. PLOUGHING OXEN (NILOTIC). BOR DISTRICT (DECEMBER {951). INTRODUCfION This volume of our report contains the results of special surveys. investigations, and experi- ments upon which many of our conclusions concerning the eftects of the Equatorial Nile Project and our recommendations for remedial measures are based. Chapter I describes survey work carried out in the reach of the Bahr el Jebel between Juba and Bor. This survey is of particular importance since our estimates of the vegetation species on the flood-plain of that reach are based upon the detailed in vestigation of sample areas. It has also tbrown much light on the problem of the factors which govern the distribution of vegetation on toich land and swamp in other reaches of the river. Chapters 2, 3, and 4 need particular mention. They cover respectively the White Ni le between Malakal and Renk, the Sobat, and the Machar Marshes which form the greater part of the area lying between those rivers. Chapters 2 and 3 have been wholly, and the hydrological section of Chapter 4 partly, written by Mr. J. W. Wright of the Sudan Survey Department. Mr. Wright became interested in the Equatorial Nile Project when, in 1947, he started the compilation of maps of this area from the American ai r photographs. Our grateful thanks are due to him for his substantial contributions to our investigation and his great interest in our work. It had from the first been one of the Team's main problems to relate the areas flooded by the White Nile each year to the height of tbe river. This was stressed in the First Interim Report, a rough estimate of the average total area flooded annually being deduced indirectly from the observed discharges at Malakal and Renk. Our predecessors also asked for air photographs to be taken at various stages of a flood in order to obtain this relationship directly. Mr. Wright showed that in this reach ai.r photographs were unlikely to give an accurate result, mainly because of the height mthe grass, and also that tbey were unnecessary since a sufficiently accurate result could be obtained from the cross-sections which had been measured some years ago by the Egyptian Irrigation Department. His results were included as Appendix IV in the Team's Third In/erim Report in 1949, and they have been used by us in estimating the effects on grazing along the White Nile of various proposals for control of the river upstream of Malakal. Meanwhile, from this relationship between the rise above mean low level and the average width, Mr. Wright had been able to calculate tables of surface area and trough volume in terms of tbe gauge-readings between Malakal and Renk. From tbese and records of rainfall and evaporation he was able to make an analysis of fifteen floods on the White Nile, and to estimate during the course of eacb the volumes of water stored temporarily in the trough, gained by rainfall, and lost by evaporation and by absorption into the flood-plain. Comparing the totals of these wi tb the difference between the observed discbarges into tbis reach at Malakal and out of it at Renk, he was able to deduce the amounts of water contributed by the small tributary watercourses entering from both sides along this reach. While it seems likely tbat his estimates for this inflow may be in error for the years when it was small, in three out of the four years where his deduced flow exceeded a milliard there was ample evidence, in local administrativerecords, of beavy discbarges in these tributary khors. For the fourth year no records could be found . These results form the chapter entitled ' An Analysis of the White Nile Flood between Malakal and Renk', which was received by the Team in 1949. Mr. Wright then ,turned to the Sobat, and attempted to apply the same method of analysis but with two important differences. The first was tbat there were only four complete cross- sections of the Sobat available and these were all on the lower third of its course, whicb is affected by the backwater of the White Nile and wbere tbe average rise and fall and the width in flood are very much less than they are in the upper two-thirds. Tbe second difference was that the tributary watercourses of this river bad apparently been investigated more thoroughly than those of the White Nile, so that in most years all important tributary discharges bad been recorded. It therefore seemed likely that in most years the differences between tbe observed discharges into the Sobat at Sobat Head and those out of it at Hillet Doleib (8 km. above its mouth) were almost entirely due to the cbange in trough volume, gains by rainfall, and losses by evaporation and absorption, once all the observed tributary discbarges were allowed for. Mr. Wright then assumed that on the upper Sobat the general sbape of the valley bore a family resemblance to that of the White Nile, and deduced tbe actual average cross-section from the observed discharges at its head and tail by a process of successive approximation, until he was able to produce a curve of average cumulative differences between tbe two sets of dis- cilarges which agreed closely with that observed. He was then able to calculate from this average 819 cross-section tables of surface area and trougb volume in terms of the gauge-readings along the Sobat, and so to analyse each of thirteen floods exactly as on tbe White Nile. He also studied the bebaviour of the tributary watercourses-particularly tbe Wakau and tbe Twalor-in more detail tban had been done before, and was able to deduce from tbem tbe probable behaviour of the water in tbat part of tbe Machar Marshes lying immediately north of the Sobat. These results form the chapter entitled' An Analysis of tbe Sobat Flood " received by us in June 1953. Cbapter 4, wbich concerns the Machar Marshes, was originally written by us before ·this analysis of the Sobat flood was available. As tbe deductions in tbe analysis differed in some respects from ours, Chapter 4 required a certain amount of re-writing, and tbis was undertaken on our behalf by Mr. Wright after the main outlines had been agreed between us. The result is that all three parts form a coherent whole, in which our field investigations have been com- bined with Mr. Wright's theoretical analyses, to produce what we hope is a reliable picture of conditions in the arca. It should be emphasized, however, that wbile Cbapter 4 and certain conclusions, given in Volume I, wbicb are based on Cbapter 2 and 3 of this volume, represent our considered views as a Team, tbese two cbapters tbemselves are Mr. Wright's individual contributions, and his conclusions must remain to some extent a matter of opinion until tbeir very theoretical basis is supported by further observations. Chapters 5 and 6 of tbis volume contain records of data obtained by experimental metbods. It sbould be stressed again that the time available for experimental work was very small. Qualified staff was not available until comparatively late in the Team's tbree-year programme, and all members of the Team had many other duties to perform which carried them far and wide over tbe vast area involved. In most cases the experiments can best be described as , sbort-term trials '. The results have had to be accepted with considerable reserve and tbere are naturally many omissi<;>Ds. 820 PART I. SPECIAL INVESTIGATIONS 8~1 ·CHAPTER l. THE BAHR EL JEBEL FLOOD-PLAIN BETWEEN JUBA AND BOR 1. OBJECTIVES OF SURVEY Tbe objectives of tbe survey may be defined as follows: (I) Investigation of the topography and hydrology of the Bahr el Jebel Rood-plain between Juba and Bor, with concentration on two specimen a reas. the Aliab Valley between Tombe and Bor and the toich on the east bank between Mongalla and Gemmeiza, in order to correlate levels and areas of flooding wit h Mongalla discharges. (2) Collection of information on the ecology of the area- in particu lar the relationship between flooding, soils, and vegetation. (3) Collection of basic information on the economy of the area-agriculture, animal hus- bandry, and fishing- and its relat ionship to the ecological regime. (4) Assessment of the effects of the Equatorial Nile Project in this reach of the river . . (5) Investigation of the possibility of carrying high discharges through this reach by bank- ing of the Bahr el Jebel or by canalization of the River Aliab. (6) Investigation of the possibility of using the Aliab Valley or other parts of the flood-plain as a remedy for losses under the Equatorial Nile Project by producing artificial grazin g or irrigated crop production schemes. 2. ACCOUNT OF SURVEY The importance of a survey of the Aliab Valley was first stressed by the Team in 1947 (see Second Interim Report, pp. 8, '7, lOS), when it was proposed to use the area to produce a rtificial grazing, grain, and sugar, assuming that the River Aliab would be canalized and the left bank of the Bahr el Jebel strengthened to eliminate spil l. In 1948 tbe Team recommended in their Third Interim Report that a detailed survey should be made of this area, and later in that year it was decided that conditions, though not favourable, were good enough to start work immedi- ately. The plan was that cross-sections should be surveyed as the basis of a contour map of the Aliab Valley; it would then be possible to make the preliminary design for an irrigation scheme. Survey work began on 10th December 1948, and ended on 8th May 1949. The average strength of the party was seventy, including chainmen and unskilled labour to clear the lines. Extremely high discharges and river levels- December to March average 70·6 mi d-resulted in very swampy conditions and made levelling on the toiel. impracticable until very late in the dry season. The preliminary work of clearing a road and survey line from Tombe gauge to Yalakot (see Fig. H I) and levelling to Khor Gwir was begun, but the extreme density of the forest and undergrowth on the western side of the valley made progress slow; average clearance on the line was about 300 m. a day, with a labour force reduced by sickness to 50. By the end of the season the longitudinal section had been cleared and surveyed as far as Yalakot, and two and a half cross-sections of the toich were completed (see Progress Report, 1948-49, p. 7). A programme of soil sampling was started. . No work was done in this area during the 1949-50 season as the Team was concentrating its activities as far as .p ossible in the Northern Zone, but the survey was continued between 20th January and 10th May 1951 under favourable conditions (December to March average discharge 42·3 mi d). At the southern end of the valley the survey line already completed was used as a framework; in the north, where communications were easier, the sections were based directly on existing E.I.D. bench-marks. Twelve and a balf cross-sections of the vaUey between Tombe and Bor-several of them over 10 km. in length-making a total of 15, and two cross- sections of Khor Gwir, were completed. The opportunity was taken to relate the cross-sections to soil types and vegetation. A comprehensive soil survey was made, involving the collection of 900 samples from 150 soil pits, and a detailed survey of vegetation was also carried out. At the same time a survey of cattle-camps was made, their positions being noted, together with cattle population and movements and tbe way in whicb pasture was utilized. The distrihution of fish was also investigated. These cross-sections were plotted; from these the alluvial formation of the banks of the Babr el Jebel became evident. Since the sections also indicated a fall in bank level relative to water-level from south to north, it was realized that a longitudinal section of the bank crest, including alI 'spill-<:hannels, was necessary to determine the amount of spilling into the Aliab 823 Valley and thus to complete the survey. This was done during the 1951-52 field season, when the survey was extended farther south to cover the whole Bahr el Jebel flood-plain between Juba and Bor. As it was possible to start work early in the dry season at the southern end of the reach, where conditions were fai rly dry. the survey began on 24th November 1951 and was ' completed by 7th May 1952. Twelve cross-sections of the flood-plain were surveyed from the high ground on the east to that on the west, while four AJiab sections were extended as far as the high ground on tbe east, to complete sixteen cross-sections at approximately 10 km. intervals from Juba to Bor. In addition to the Aliab Valley one other area , the Mongalla-Gemmeiza (oich, was surveyed in detail. Eight additional cross-sections of the latter were interposed to give twelve evenly·spaced sections, while longitudinal bank sections from Tombe to Bor and from Mongalla to Gemmeiza were completed. Three sections were extended to the east of the Mongalla-Gemmeiza road. All these sections were based directly on the line of E.I.D. bench-marks down the east bank. METHODS The cross-sections were set out by compass and ranging-rod and later positioned on air photographs or air survey maps. Nilotic labourers in parties of 20 cleared the lines under the supervision of a senior chainman. Progress varied according to the conditions; it averaged. a few hundred metres a day in thick forest, one kilometre in medium bush or thick grass, and two kilometres in good conditions on the (oich. Where tbe survey lines passed through water, taping was combined with tbe soundings of channels and swamps; pegs were driven to water- level for the data of the soundings to be determined by tbe levelling party. It was often necessary to use small boats to cross deep channels, and on occasion three boats were used for taping, one for each end of the tape, and a third for the sounding party. Soundings of the main channels were taken 'at first by means of a sounding cable and later by tacheometry; the intervals at which sou ndings were taken were measured by reading a staff in the boat from an instrument on the bank . Surface current speed; were taken by float and stop-watch. At first stakes were driven into the mud to support the instrument and steves when levelling in swamp, but it was found that the necessary accuracy-that of control levelling-could be achieved by judicious placing of the instrument and by levelling in one direction, reading both metric and arbitrary scales. At change-points the staff was placed either on specially constructed large base-plates or on a large peg. In the Aliab Valley the level was often carried across lake or swamp by driving stakes to water-level at both sides of the obstacle. A survey of vegetation was made along each line in conjunction with the taping; the propor- tion of ground covered by each species was noted, together with the position of changes. Specimens were collected, thei r vernacular names being noted, and were later identified. Soil sample pits were dug to a depth of six reet at kilometre or half-kilometre intervals along the sections, samples being collected at foot intervals, and described, together with the site, vegeta- tion, and moisture content. All cattle-camps and permanent villages were visited, and the tribe and sect ion were noted, together with animal population and movements, watering places and grazing areas, and also areas of cultivation. Two fishermen visited all the khors and lakes with casting-nets; their catch was tabulated, the species, weight, and size being noted, together with the time spent in fishing. RESULTS As far as possible, the results of the investigation have been reduced to diagrammatic form. These are dealt with in detail below, and comprise: JUBA-BoR REACH Figure No. 16 cross-sections of the flood-plain ... .., A 2-9 3 cross-sections of the forest to the east of the river A I ()"I 2 Key plan AI ALIAS VALLEY J 5 cross-sections of the valley H4-9 2 cross-sections of KhoT Gwir H6 Longitudinal bank section . HI()"II Map-contours and flooding . . . HI Map-soil texture, surface organic horizon, aDd water·table H2 Map-vegetation, pasture utilization, aDd fishing H3 MONOALLA-GEMMElZA TOICH 12 cross-sections of the valley H 14-t6 LOngitudinal bank section ... H 17- t8 Map--c.ontours and flooding . .. . .. . .. HI2 Map- vegetation, pasture utilization, and fishing H 13 824 3. JUBA TO BOR: GENERAL Sixteen cross-sections of the flood-plain at approximately 10 kl11. ihtervals between Juba . and Bar, together with a map o r the a rea, give a genera l picture or this reach. The flood -plain . defined by high banks on either side which mark the li mit of the forest. increases steadily in Width from 4·5 km. at Juba to 9 km. at Ba r. The Bahr el Jebel, confined in parts to one channel and in other parts divided into as many as three channels. swings from one side of the flood-plain to the other, dividing the a rea up into segments bounded by the river and the forest or into islands. Between Juba and Bar there are eight main segments of flood-pl a in , three on each bank of the river and two large islands. The two islands, between Gemmeiza a nd Tombc, a rc heavily flooded; this is possibly due to the banking-up above Tombe or the water which used to flow down the River Aliab, since the latter is now blocked by .rudd. Apart from these islands, flooding is least near Juba and greatest in the north near Bar. The cross-sections, which include nonnaJ high and low water-levels-based o n maximum and minimum 10-day normals, 92 mi d and 58 m /d at Mongalla-show how the water-level rises from so ulh to north in relation to the levels of the bank and flood-plain. Fig. H 21, which is deri ved from the cross-sections, shows this most clearly. This rise is reflected in increased flooding and a transition in vegetation fio/1l the Phragmifes dominated foich near Juba to the papyrus swamp opposite Bar. But this transition is not uniform. Each segment of the flood-plain is a hydrological unit, and, as Fig. H 21 also shows, within each unit there is a similar transition fro m south to north in relative height of bank and therefore in flood ing and in vegetation. In order to elucidate this double transition , within and between uni ts of the fl ood-plain , two specimen areas were surveyed in detail. These areas, the Aliab Valley and the MongalIa-Gemmeiza loic", will be dealt with in turn below. 4. THE ALTAB VALLEY TOPOGRAPHY The topographical description of the Aliab Valley is based on the survey work carried out, together with general observations in the field. The survey work (see Fig. H I) consists of fifteen cross-sections from the high ground on the west of the valley to tbe Bahr el Jebel, from Tombe to near Bar. At the southern end. where they are based on a longitudinal survey through the forest, the sections are at 2 km . intervals and vary in length from I km. to 6 km ., the average being 4 km. Farther north, where they are based on E.1.0. bench-marks, the sections are at approximately 4 km. intervals and the average length is 9 km. Two cross- sections of Khor Gwir, one near Lake Dijir and one 10 km. upstream , and a longitudina l section aJong the left bank of the Bahr el Jebel from Tombe to opposite Bar, 58 km. long, complete the work. These sechons have been positioned on a 1/ 50,000 map of the Aliab Valley, which was traced from E.I.D. air survey maps based on photographs taken during the years 1930-3t. A contour map has been prepared from this work. Because of the broken alluviaJ nature of the ground there are considerable variations in height, but the mean ground level has been taken in determining the intercepts of contours along the sections. Between the cross-sections the contours have been interpolated according to the topography, and tbey would therefore be more correctly described as form lines; the large number of small khors and depressions would make ;t impracticable to produce a contour map on this scale, however much survey work was done. The map does, however, give a general picture of the topography, brings o ut its inverted nature-the rivers being above the plains they traverse-and thus complements the cross-sections. The names of as many channels as possible have been included on the map, but there is not space for all. These, together with general and detailed topographicaJ informa- tion, are available in traverse notes and the descriptions of soil sample sites. The Aliab VaJley is the name-derived from the Aliab Dinka who utilize the western side during the dry season-given to an arbitrary portion of the flOOd-plain bounded on the west by forested high ground and on the east by the Tombe channel (from Tombe where it leaves the forest), by the west channel, and finally by the main channel, of the Bahr el Jebel. Thus, although the flood-plain widens steadily from south to north from an over-all width of 8 km. at Tombe to 9 km. at Bar, the Aliab Valley varies considerably in width, as the map shows. The total Tength of the valley from Tombe to Lake Papiu is about 90 km., and its total area is approximately 480 sq. km. The area surveyed is 250 sq. km., or about half the whole vaUey. The western limit of the flood-plain is a comparatively steep bank, whose height remains fairly constant at about 4 m. The higher ground west of this bank is thickly forested; the bank itself forms a well-defined boundary between forest and open flood-plain. Between Tombe and YaJakot, where the forest is extremely thick, several small dra inage-channels form re-entrants 825 into the forest and bank. Just north of Panabang the bank is broken by Khor Gwir, a seasonal stream fed by local rai.nfall, wh ich spreads out from a well-defined channel into a grass-choked depression feeding Lake Dijir. N(lrth of Khor Gwir the high ground is set back to the west, and the forest spreads to the east of it and merges gradually into tbe flat, open toic". At ' M inkaman the forest gives way to thick bush, but the bank once again marks the boundary. between higher ground. covered with thick bush, and loic". . The Aliab Valley has as its eastern boundary the left bank of tbe Bahr el Jebel whicb,. by deposit ion of si lt , has been built up to a height of 2 m. above the adjacent flood-plain . This process is conti nuing; in many places it was found that levees were fo rming on the river side of former banks, and in such cases the higher of the two banks was surveyed. The longitudinal sect ion of the bank from Tombe to Bor (Figs. H 10- 1 I), showing the crest level of tbe bank every 50 m. and the invert level of every spill-channel , gives a clear picture o f tbe bank. As its origin would lead one to expect, the crest o f the bank is extremely even in heigbt, variations from a mea n line drawn through the sect ion being of the order of on ly 15 em. The bank is, however, broken by many spill-channels, la rge and small ; between Tombe and Bor there arc 22 channels whose depUls are greater than 1m., 67 between I m. and 50 cm. , and 28 1 less tha n 50 cm. They vary in width from I m. a t bank level to 28 m. in the case of Nyin Akujai near Bor, the average width being 2 m. The la rger channels are dammed a t their mouths by the Dinka during the dry season to a llow the loich to dry out for grazing and to enable the ca ttle to reach the grazing a reas by a route along the Bahr el Jebel. These dams are usua lly washed "way as the river rises. The Dinka also dig small channels where the bank is low in order to trap fi sh. bu t most channels appea r to be started by hippos forcing their way on to the bank of the river. Th is is a point which must be borne in mind when banking of the rivcr to prevent spi ll is cO.nsidered. T he heads of a ll cha nnels have the same appearance, a funnel openi ng towa rds the ri ver. the channel being narrowest a t its highest point or invert, then sloping away dow n to the loic" . Once such a track has been started catt le use it as a wall' ring place. and when the r iver rises spill-water find s its way through the gap in the bank ami erodes " cha nnel down the slope from Ihe ri ver to the loid,. 'Many of the smaller channels are choke<.l wit h P"rag lllil eO' ('Oll1l11l1l1iO', so that the channel grows or closes up according to whether scour from thc river compctes successfu lly or no t with grass growth and sil t. The map. on wh ich a ll the large r chan nels are marked , shows how they have formed where flow fro m the river is grea test, e ithcr on the o utside o f a bend or where there is a drainage-cbannel parallel to Ihe river and close to the bank. This means that a la rge proportion of the spill- wa ter is kd <.Iirectly in to dra inage-channels a nd is ca rried north in a defi ned channel. Com parison of the present topography with air photographs and maps dat ing from 1931 , and also with ground surveys made a t that time, shows that tbe la rge channels have changed little in the interval. Alt hough the southern head of Khor Ker at Amathhom is now so silted up that its co urse is unrecognizable on the ground and the bank of the Tombe channel is unbroken, the course of Khor Ker still shows up clearly from the ai r, so it may have been blocked before the map was made. Two changes in the course of the river are noticeable; the two loops in the river near B.M. A 18 and B.M. A 22 have been cut off and remain as ox-bow lagoons. When the bank crest was su rveyed , the water-level in the Bahr el Jebel was taken every ki lomet re. The river level was almost constant over the period of the survey, but these levels have been reduced to a single Tombe ga uge-reading to give one river profile between Tombe and Bar; the slope is found to be constant in each cha nnel, but it cbanges at ebanneljunctions. A correlat ion between Mongalla discharges and Tombe, Malek, and Bor gauge-readings has been deduced from the monthly means of 1938 to 1942 (see Fig. H 20); for eacb steady Mongalla discharge between 50 mi d and 120 mi d , at 10 mi d intervals, a river profile has been drawn and su perimposed on the bank section (see Figs. H 10 and H II). This combined section demonstrates clearly that the longitudinal slope of the bank is greater than U,e slope of the river water-level at a ll discharges. Therefore for each steady discharge above a certain level there exists a point at which the bank and water-level coincide. As their angle of incidence is very small it is necessary, in order to find this point accurately, to re-plot the longitudinal section on a greatly increased vertical scale (see Fig. H 21 for a "educed copy of the working d rawing). At each kilometre the mean level of Lbe bank over the kilometre bas been plotted, together with a lower point, such that 100 m. of the bank'is below it. This latter point, about 10 cm. below the mean level, is taken to be the effective bank level, and the points at which the water profile for eacb Mongalla discharge cross this curve have been deduced and are marked on the river bank on the contour and flooding map (Fig. H I). As the discharge increases, so the point of intersection moves farther upstream. North of this point the bank is under water aod widespread spilling occurs. Because of the height of the alluvial 826 bank above the flood-plain this water spreads laterally from the river, hindered only by the resistance of the vegetation, until an obstacle is reached. South of the p.oint where tbe bank is inundated, spill occurs through chan nels and by seepage through the alluvial bank. As already . stated, most of this spill is led directly into channels parallel to the river and Hows northwards, largely confined to these channels, until an obstacle in the form of a lateral bank is reached or until the channel debouches into widespread flooding; this spill-water helps to maintain the flooding level across the valley. It has therefore been assu med that this combinat ion of spill over the bank and through channels will maintain backwater flood ing at the level of the point of intersection, and for this reason the contours have been followed back from these points to give the areas flooded at each discharge (see Fig. HI). This picture of flooding spreading from north to so uth is supported by field observation. The width of the alluvial bank increases from south to north, indicating that spill is greatest in the north. Where the bank is low in relation to water-level south of Bor, channels, dry at the time, were seen leading away from the river but onl y sta rti ng about 30 m. from the bank, here unbroken; in the same area in May 1952 the bank was submerged and water was pouring over it. In 1951 Lake Barnyieu, fed largely by Nyin Akujai, spread back to C.S. 14 and C.S. IS, while the water-levels on C.S. 13 were approxi mately the same as the lake level. The picture fits in well with the discharges at which spi lling is known to occur so uth of Bor and Malek. As·Fig. H 20 shows, spill ing increases rapidly south of Bor at a Mongalla discharge of 65 mI d, and south of Malek at between 75 and 95 mId. The valley itself, the depression contained between the Bahr el Jebel and the afforested high ground, is too extensive and varied to describe in detail . The cross-sections (Figs. H 4-9) should be studied, but field observations may be summarized as follows. On all lines the vaUey was found to change in character at Kbor Ker ; not only was the ground itself flat or rolling to the west and more broken to tbe east. but the vegetation (see Fig. H 3), the deptb of flooding of soil pits (see Fig. H 2), and the nature of drainage-channels were different. Varia- tion was not continuous from south to north. Between Khor Ker and tbe Bahr el Jebel, C.S. 2, C.S. 3, and C.S. 4 got steadily wetter. C.S. 5 was comparatively dry, but C.S. 6 and C.S. 8 grew wetter ; C.S. 9 was again dry, but north of it the lines became more and more flooded until C.S. IS, crossing Lake Barnyieu and a papyrus swamp, was reached. As the cross-sections, together with the contou r map, show, the area is transeeted by a network of spill and drainage-channels whose general direction is pa rallel to the Bahr el Jebel. AU these spill-channels have formed their own alluvial banks; the River Aliab or Khor Ker, the largest channel, wind ing down the centre of the valley, has banks little lower than the Bahr el Jebel itself and forms a barrier to lateral flooding. Although the river is covered with vegetation and partly blocked throughout its course, it carries away to the north most of the water which tops its bank. The original head of the river at Amathhom has silted up, and the river is now fed by a bead near B.M. A 22. Its course from this head to Yalakot provides a transverse obstacle to drainage from south to north and isolates the two basins to the south. Thus the banks of the River Aliab effectively divide the area surveyed into four basins, each with its own sources of flooding and drainage systems. The largest basin, north-east of the River A1iab, receives spill from several large channels near Mathiang and south of Malek ; this spill drains north, concentrated in Nyin Koriom and the system of lakes down the centre of the basin, until backwater flooding is reached. The latter governs flooding in this area, which varies from high ground on the banks of the Khor Ker, seldom inundated, to an area of permanent flooding in the north. The map shows a separate system to the west in the smaller basin contained between the River Aliab and the forest from Yalakot to Minkaman. Protected normally by the high banks of the River A1iab from river flooding, this area is fed by Khor Gwir and by run-off collected by Khor Nyankojii south of Lake Dijir. A drainage system runs north from Lake Dijir and finally joins Kbor Ker where it comes close to the forest at Abou. This basin, being free of silt-laden spill-water, is characterized by a smaller variation in flooding, flat or gently rolling ground, and open (sudd-free) channels without alluvial banks. The two small basins south of Yalakot, divided by Kbor Ker between Amathhom and Gukthon, each receive spill through several large channels and are inundated by spill over the bank at discharges between 90 and 100 mI d. However, there is a difference between them. Whereas iii the eastern basin Khor Bargeik and Khor Kolnyang drain into a cul-de-sac, pro- ducing the lake marked on the map and an area of semi-permanent swamp only subject to evaporation, the western basin is partly drained by Khor Lalop, which flows into Khor Ker at Yalakot ; to the east of Khor Laiop, Khor Aduteier fans out into the formation shown on the map, as was confirmed during a reconnaissance flight. Thus the western basin dries out first as the river drops; once the eastern basin is fuU, spill from the river flows straight into a 827 flooded area and silt is deposited near the river. For this reason the eastern basin is about I m. lower than the western and has a high and narrow alluvial bank. This partition into four basins explains all the phenomena noticed in the field and mentioned earlier. Thus, given the topographical conditions described above, there is for each steady Mongalla discharge an equilibrium position of flooding. Because the flooding of the valley is determined by backwater and because the river bank falls relatively to the water surface, a change in discharge causes a greater change in the level of flooding in the valley than in the level ofthe river. The areas normally flooded and uncovered seasonally are dealt with quantitatively in the next section. HYDROLOGY We now turn to the hydrological aspects of the Aliab Valley, which must be introduced with a brief summary of the relevant topographical features. The total area surveyed amounts to 250 sq. km. Approximately 156 sq. km. are flooded by the Bahr el Jebel north and east of the River Aliab; 66 sq. km. west of the Aliab receive flooding from Khor Gwir; and 27 sq. km. south of Khor Ker form separate basins flooded by the Bahr el Jebel. The River Aliab, or Khor Ker, takes a winding course down the centre of the valley, which it effectively divides, by its alluvial ridge, into eastern and western portions. The old course of the Aliab from the southern apex of the valley is blocked with silt; the river is now fed by a new offtake near Gukthon , and a new channel, the Khor Ker, runs from south-east to north- west across the valley to Yalakot. Since this too has alluvial banks it also divides the valley into small areas to the south and larger areas to the north. The level of the floor of the valley is on the average nearly 2 m. below the crest level of the left bank of the river, varying from 0·9 m. in the s0uth to 2·6 m. in the north of the area surveyed. This bank is not more than 200-300 m. wide at the southern end but increases in width north- wards, so that from Cross-Section 9, downstream of the juncti~n of the eastern and western channels of the river, it averages I km. in width. The reader should refer to Tables 375, 376, and 377 for· further details. The following average slopes from south to north have been calculated: Average slope of the bank crest 10·8 em/ krrt Average water slope at the mean Mengalia discharge of 75 m/ d 9·75 cm j km. Average loic}' slope 15-2 em/ Iem. As already explained (see Vol. I, p. 15), water reaches the Aliab Valley in five different ways: (I) From the BahT eJ Jebel by spill over the crest of its left bank and in large and small channels. (2) From the Bahr el Jebel by backwater from the northern end of the valley. (3) From Khor Gwir, a watercourse of local origin, and from the River Gel or Tapari in the area outside this survey. (4) From the underground water-table. (5) From direct rainfall. Spill through large channels, of which there are 22 between Tombe and Bor, begins at the lowest discharges of 50 mi d or 60 mi d at Mong::llla. The depth of water in Nyin Akujai (Unyam Koji) just south of Bor is 4·3 m. when the Monga lla discharge is 50 m/ d. The southern part of the va lley between Tombe and Khor Ker is flooded through spill-channels which draw water at the low discharge of 50 mi d. The water spilled is ponded by the ridge formation of Khor Ker, and the areas bounded by these features are therefore semi-pennanent swamp. There are 67 smaller channels between 1·0 m. and 0·5 m. in depth and 281 of less than 0·5 m. between Tombe and Bor. There are· also low points on the bank 15 em.. below tbe crest levels given in Table 376. Jt can therefore be safely assumed that when the water-level reaches the crest levels shown in Table 376 spill is passing over the cres( in appreciable quantites. Spill over the bank crest begins at a Mongalla discharge of70 mId at the northernmost end, and the whole bank is submerged when the discharge reaches 120 mi d at Mongalla. It will also be noted that the largest spill-channel is the Nyin Akujai, at the northern end of the area surveyed, which feeds the backwater lake . The River Aliab forms a bank effectively barring the ingress of Bahr el Jebel spill to the western portion of the valley. which is therefore only watered by rainfall and run-off from Khor Gwir (except possibly at the very highest levels). Field observations showed a marked difference both in gr.ass types and in the underground water-table in soil sample pits east and west of the Aliab channel. The main underground water-table extends from the Bahr el Jebel to a distance of about 1·5 km. within its alluvial left bank. There are also subsidiary water-tables created by surface Hooding. and by flooding in spill-channels. The soil is by no means homogeneous, and contains banks of silty sand deposited by old and recent minor channels in the valley. Rainfall over the area averages 903 mm. per year at Bor. 828 We have calcula ted the a reas of the valley which are inundated at Mongalla discharge above 50 mid. a nd have plotted the duration of each d ischarge fo r t h~ period 1946- 50. wilh the figures for 1950 incl uded twice. according to an empi rica l rule (see Vol. r. p. 156) decided . upon by the Team (Table 379. p. 830. and Fig. H 22). The area to the west of the River Aliab has not been included in these areas because there is evidence 10 show thai it was no t Aooded by spill from the Bahr el Jebel in the period considered. and the two areas south of Khor Ker are also not included. The analysis of grass species shows tha t Aow in la rge spill -channels is cOMned to those channels as they run down the va lley and to a narrow strip on each side of them. The factor which controls widespread Hooding is the backwater from the north . TABLE 375 ALIAB VALLEY HYDROLOGICAL DETA ILS So:tion W ,HER·LEVELS AT M ONOALlA D ISCHARGES IN /II 0 _~_um_be_r _• __~ __ I --;;- -1- 10'--1' 80----1-- 90 --1-' 100 -~~J ~_' I i~_ __ I i 42J.98t ') I 424·50 424-81 r:;-25~0-'1 4;5-';5 425-40 r4 25 ·50 ! 425060(') 2 I ) . J) ~ 3-91(3) 4·27 4·50 I 4·65 4·79 , 4·88 3 I 4-96 3-30(') , 3-65(') 4-01 4-24 4·39 4·53 4·62 4'10(') 4 HOt'X') ! 3'42 I ) ,79 4-01 I 4w-16 4·29 4-38 4·45 5 2-66 6 2·38 I Wl:ll:j: In~ ~.~~ iI ~~; 4· 12 4·20(') H4 3·90 7 2-09(') I 2'58 2-95(') 3' 18 3·)2 3-45 ) -52 3-58 8 1-89(') I 2-)9 I 2'15t') 2098 3·12 3-24 ) '2'J 3·35 9 1·19(') I 2·19 2·55 2·11 H I 3-03 3·10 3' 15 10 1·29(')(')1 1·78 1·16 2·36 2·50 2·60 2·67 2-10 :i b~1~:l\,) ! 1·56 1·92 2·1J 2-26 2·36 2-42(1) 2-46 1·24 1·61 \ ·8 1 1·94 2-03(') I 1-08 2' 11 13 420'42\') ' 0·90 1·24 1·42 1·54 1·62(» 1·61 1·69 14 41 9-99(1)(') 1 0·41 0·17 0·92 1·03 1' 10(') 1·14 1·11 15 419'60(')(') 1 420·06 420·3) 420·46 420·56(') 420·63 420·66 420·68 ~:~ ~: ::~:~:l :g:~~ r,:;~~e~~~r~l:rtt:~~ill~~!~~~. l'} Walcr·le\_ p.:.m L 1_ _ 1 I ! 0·07 - 4 7-35 8'30 I 8'60 350 - 0·36 2 j 0·06 0·01 5 8·05 I 8·55 8·70 260 i - 0·25 i 7 6 3 0·02 - 2 5-80 5·75 6·30 510 9·4 0·64 2 4 (}Ol I - I 2 I 6·00 6·60 ! 6'60 700 I 10·4 0·97 I I 5 (}OS - 2 HO 4·50 6·10 1,060 I 9·8 1'38 7 6 0·10 0'01 5 6·20 6·60 6·20 760 8·8 0·89 10 7 0-(J8 - 3 5·10 I 6·20 6·70 910 I 11 ·1 1·34 8 8 0·11 - 4 5·70 6·20 I 6-65 880 8·6 1·01 19 9 0'05 - 3 5090 I 6'55 7·00 820 9·0 0·99 -10 (}05 - 3 4-45 6·60 I 6'50 260 8·5 0·26 I 17 11 0·49 0·40 8 4·95 5·10 6·45 1,150 I g·S 1·32 IS 12 0·08 - 4 5,)0 5·20 4·50 1,460 9·9 J'92 24 I) 0·10 - I 3 6·50 I 6·85 6·45 1,200 10·7 1·69 23 14 0'13 - 3 HS 5-25 - 2,320 10·8 3·]4 9 15 0'14 0·05 4 5·95 6·90 7-30 810 8·8 I 0·96 10 16 0·10 - 3 6·80 HO 7-30 840 8·9 1·00 12 17 0·21 0 ·12 3 4·75 4080 4·50 1,390 9·9 1·84 5 18 0-fT/ - 3 4·70 6·00 7·00 1,160 10·2 1·58 7 19 0·09 - 4 H O 6'05 5·90 860 8·8 1·01 10 20 0·05 - 7 6·40 6·80 6·75 650 10·5 0·90 7 21 0·03 - 2 6·30 6·65 6·60 640 8'7 0·75 6 833 TABLE 381 ALIAB VALLEY PRESENT DISTRIBUTION OF VEGETATION SPECIES AJuAs MEAsURED fROM NORTH Area covered Species in association Area covered From To ! with other spp. , Pen::en~ - - sq. ,' sq. km. sQ. kIn. km. SQ. Jon. NORTH-EAsT OF THE RIvER AL1AB Open Water 0·0 19'5 19·5 19·5 IH Cypuus papyrus 19·5 33-2 (45'8) 13-7 (26-3) 11-2 7·2 VOSS;Q cuspidalo 31-1 (19-5) SO-O 48·9 (60'5) 20·1 12·9 £Chinocllloa slogl/ina 49·5 104-9 55-4 29-8 19·1 £Chinochloa pyramidalis 31 ' 1 155-8 124-7 71·5 45-9 PhrQgmi/~s communis 145·3 155-8 10-5 3-7 2-4 ----- Total area 155-8 100'0 WEST OP TIlE Rlv ER A.uAB Ethinochloa slogn;na I 21·0 11-1 16·7 £ChiMchioo pyramidalis - 65-4 47·1 70·8 Oryta barfhi; ,-, .. . 20'1 8'0 12·0 Ph'Qgmil~s communis .. . I I 0 ·9 0 ·3 0-5 L .. _J ------------Total area 66-5 100-0 Two B ASlNS SoUTH Of Y ALAKOT Vossia cuspidolo ! 8'3 3-9 14-4 Echinochloa stag nina .,' l 9-2 6·2 23-0 r,,~~n;~~/:sa :::r,::~~~~lis : I 24·9 16-4 60·7 I 1·4 0-5 1·9 i I I - - --------- I 1 Total area I 27-0 100·0 T OTAL AREA SURVEYED 1 Open Waler ... i 19'5 19-5 7-8 Cyptrus papyrus 13·7 (26'3) 11-2 4·5 Yossio cuspido/Q 57-2 (68-8) 24·0 9·6 Echinochloa slognillo 85-6 47·1 18-9 &1I;,lOC"'OO pyramidalis 215'0 135-0 54·2 OrYUlbar,hii 20-1 8-0 3-2 Phrogmites communis 12-8 4-5 1,8 ------- Total area 249-3 lao-a Brackets refer to pllrli:al COVtt. TABLE 382 ALlAB VALLEY DURATION OF FLOODING OF VEGETATION SPECIES (From Fig. H 22) Species Minimum Maximum Optimum Days Days Days Open Water ... 360 360 360 Cyperla papYrllS 281 360 329 VassiQ cuspidatQ 198 300 260 Echillochloa sragnina 178 240 204 Echinochloa pyramidalis 25 287 126 Phragmius communis 25 58 44 LAND UTILIZATION Except for a very small amount of maize grown by the Mandari during the dry season on the river bank near Tombe, the riverain flood-plain in the area covered by the survey is used exclusively for dry season grazing. As the flood-plain on the right bank of the Bam !'l 834 Jebel is extremely limited in extent between the latitudes of Tombe and Bor, the Bor Gok Dinka as well as the AJinb Dinka are dependent on the Aliab Valley past~rc (see Figs. E 5, E 6) . Although the Bahr el Jebel forms the boundary between Upper Nile and Babr el Ghazal 'Provinces, the Bor Dinka cross the river to dry season cattle-camps in the AUab Va lley; the River Aliab fom1s a rough boundary between Bor Dinka and Aliab Dinka camps, the form er being to the east of its channel and the latter to the west, with a few camps o n the east bank . . The period of utilization varies from year to year, depending on the rains and the fall and rise of the river. Before the cat tle are moved on to the loic" , a period o f three weeks is allowed after flooding has receded for the ground to dry out and for the grass to be sufficiently dry fo r burning. Thus the time at which cattle-camps are first occupied may vary fro m early November to January and also varies according to the camp, but the end of November or beginning of December is considered normal. The tin1e of the move away from the loic" is more constant, and occurs generally at the end of April or the beginning of May. The positions of all the cattle-camps in the area have been marked on the vegetation map (see Fig. H 3); altered conditions cause only minor variations in camp sites, which are limited to the highest ground, although the direction of paslure utilizatio n may change each year. It will be noted that all camps are situated where Echilloch/oa pyramidalis and P"ragmiles communis indicate relatively high ground, generally on the sloping banks of the River AUab or the Bahr el Jebel, where they are a lso within easy reach of water. Each cattle-camp was visited, and the cattle population was estin1ated by counting the number of cattle groups and estin1ating the average number in each group by counting a few typical ones. The figures obtained can be found on the map. The numbers of sheep and goats were also estimated, though with less accuracy. The total animal popUlation belonging to each tribe over the area covered by the survey is tabulated below: Tribe Cattle Sheep Goals Ar\!a (sq. km.) Aliab Dinka l 4,'19O 2,870 l,370 Bor Gok Dinka 22,860 3,920 2,080 ---- _. ---_._-------.. _-_. Total 37,650 6,790 3,450 250 (including 20 open water) The directions of grazing movements from each camp at the time of the survey-April 1951, towards the end of the dry season-have been indicated on the map by arrows. These show the concentration of cattle around the north-eastern basin, where flooding is governed by backwater and where tbe range of flooding and the areas successively exposed as the discbarge decreases from 90 mi d to 50 mi d are greates!. Earlier in the dry season the grazing west of the River Aliab is used, but by April the lower areas, which are indicated by the presence of Vossia cuspidata and Echinochloa stagnina and which dry out last, are being grazed. The southern basins where flooding is produced by spill support relatively smaller populations, possibly because flooding is more constant and all parts of the area tend to dry out simultaneously. To complete the picture some mention should be made of the game found in the area, although to the Dinka of this area it is of little economic importance except that it competes for pasture. Between December and April large herds of game migrate from the west and graze the western fringe of the flood-plain . Very many bulfalo, estimated at about 3,000, and over 1,000 tiang are found between Tombe and Fad unyiel, concentrated between Lake Dijir and Minkaman where the River Allab is set back to the east and leaves a wide belt accessible. Also to be seen in smaller numbers are elephant, giraffe, lion, waterbuck, reedbuck, bushbuck, Mrs. Gray's lechwe, and warthog. FISHERIES An account of the fish resources of the Jonglei Area and their importance as an item of diet is given elsewhere (Vol. 1, p. 387 and p. 245) . The statistics collected in the Aliab Valley give a measure of these resources in a small area. During the period of the survey-13th March to 24th April 1951-the fish were concen- trated in the Bahr el Jebel and a few khors. Two Shilluk fishermen, using small fine-mesh Casting-nets, fished the Tombe channel, some spill-channels and, with the help of a Dinka guide, all the inland lagoons and drainage-channels in the area. The position of each fishing site is shown on the map by a letter. The species, weight, and size of each fish caught were reco,ded, and also the time .spent fishing in each place. 835 These results have been tabulated in two ways, according to species and according to fishing sites (see Tables 383 and 384, pp. 836-7). The first table gives the totals caught in two years' fishing, the first (1950-51) in the Aliab Valley, and the second (1951-52) between Juba and Bor, only a small proportion being in the Aliab. This table also gives an indication of the relative frequency of each species over the area-Tilapia, Heterolis, and Clarias being by far the most common. It also shows the weight of fish generally caught in casting-nets and the large yield which can be obtained by this method. The total catch for 33 days' fishing by two men in 1950-51 was 662 kg., or 10 kg. of fish per man per day. The second table shows for each fishing site, arranged approximately from south to north from Juba to Bor, total catches and the percentage of each species in the catch. Positions G to X are in the Aliab Valley. As the fishermen often fished a long way from camp, the total catch tended to be what they could carry back , so kg. per man-hour of fishing is the only objective basis for comparison. In spite of weather cbanges yields were very consistent at a given place on different days. Even so, it is not an absolute measure of distribution, as some areas are more suitable than others for the casting-net and bottom fish such as Clarias are less likely to be caught by this method. However, the table, wi th the map, shows that catches were largest in tbe Bahr el Jebel, and especiaUy at Khor Lalop bead. Away from the river, they were very large in a pool near Yalakot in the nearly dry bed of the River Aliab, and fairly constant elsewhere over the whole vaUey, even in the isolated Khor Gwir system. The proportion of Tilapia caught was every- where large; tbat of Clarias greatest near tbe Bahr el Jebel, and tbat of Heterotis largest away from the river. PolypTerus were caught, in small quantities, only in inland kbors. At present, exploitation of these resources is limited in extent and methods. Monythany Dinka (see Vol. I, p. 384), who rely almost entirely on fishing , have permanent villages down the west side of the Aliab Valley at Akot, Ahou, Minkaman, Panabang, and Yalakot. All these vi llages are close to open water, in Lake Barnyieu, Khor Tetang, Lake Dijir, and Khor Ker at Yalakot, which they fish from canoes using harpoon ana line. Monythany from the east bank also fish Lakes Majir a nd Guol. The majority of the population, however, rely on random spearing and occasional mass spearing in nearly dry khors or behind traps; when the river rises Clarias especially are caught on their way to the /oich. The amount caught by two men shows that although suitable khors are probably too scattered a nd inaccessible for commercial exploitation on any scale, improvement of present fi shing methods wou ld greatly increase local catches of fish, though there might be some danger of overfishing. TABLE 383 TOTAL FISH CATCHES TABULATED BY SPECtES (1950- 19S I) (1951-1952) Species Number Weight Average i Number Weigh_t J _Av erage (kg.) (kg.) I __t •. (kg) .~~!. Tilapio sp . .. . 443 423 ' 1 0'960 413 430·6 1·040 H~/U()lis nifOficus ... 84 136·7 1-630 33 96-5 2-920 Clar;as sp. 53 73-4 1·390 3 4·1 1·370 Cllhorinus sp. 25 6·2 0 ·250 39 48·8 1'250 Polypterul sp. 55 15-0 0 ·270 7 4-1 0'S9O Hydrocyon linea/US 5 0-8 0-160 31 IH 0·430 SinOtMmJiJ spp. 5 0·9 0-180 37 12·5 H40 Labto sp. _,. 16 12-3 0·770 A./utes spp. 10 1·2 0·120 44 8-4 0'190 LAtes niloticus 6 8-4 1·400 Auchenglanis sp. 0' 1 0-100 12 6·0 O·SOO Distichodus sp. . .. 1·5 0-170 S 2-0 0-2S1) Gymnarclius ni/o/icu.f ::: I' 1·0 0-500 Mormyrus spp. .. . . .. 0·9 0-130 EuJropius niloticus or ScM/be sp -- I O'S 0 ·120 0'1 0·100 Barrus bayiJd __ _ __ 0·2 0-200 0·5 0·500 Clarotes /aliceps __ _ __ i 0'1 0·100 0·2 0·200 A.aiNu sp. . .. 0·1 0·100 '1'- --- --. ----~:-~ - - '~ -I 706 661·7 i 652 647·7 I 33 days' fishing by two fisbermen 44 days' .6shin£ by two fishenneo (117 hours' nshing by two fishermen) (157-5 hours' fishing by two fisbennen) A vcrage catcb/ d.y -100 1<&. Average catcb/ day - N 1<&. Average catcb/bour := 2-8 kg_ Averaae catcb/bour -= 2-1 kg_ 836 TABLE 384 FISH CATCHES TABULATED BY FISHING SITES FISHING SITE ! PERCENTAGE OF C ATCH ACCORDrNO TO SPECIES Total Hours Ra~ -L-eu-er-.-I----'--J-)es-cri.-Pb-.o-n ----- I N"D:;~ or I Period 1- TiI~~i" TH:~r:;: 1-~:r:-I-c/;:;:~G;.,:- 1 .-1W eight (2 men) Ici/br· Others y. fi eld observations ; C.S. 9, e.S. 10, and e.S. 12 were surveyed at about this discharge and spill-water was pouring over the bank at the time. At higher discharges, bank inundation and backwater flooding spread farther south, as shown on the map, until tbe whole area is flooded at a discharge of 120 m/ d. The loich itself is, as shown by the cross-sections, extremely broken and irregular. It is divided into a number of longitudinal basins by the alluvial ridges of watercourses, now silted up, which run parallel to the river. These ridges are almost as high as the bank of the Bahr 839 c1 Jebel , except in the north where they tend to disappear as the flat bed of Lake Buri is reached. Thus at the southern end of the valley spill from the Bahr el Jebel through channels and seepage is li mited to a fairly narrow belt beside the river, while run-off is confined to a smail area near . the forest. This is illustrated by C.S. 2, surveyed after the river had just risen. The water- level near the river was rising while the small drainage-channel near the forest was drying up, and the greater part of the section was dry, unaffected by spill or run-off. This is also illus- trated by the vegetation, as wil l be shown later. However, as these longitudina l channels do not form any separate basins like the Khor Gwir basin in the Aliab Valley, it has bccn assumed that once th e bank crest is completely inu ndated backwater flooding will spread laterally as far as the forest, and the whole loich has been treated as a unit in the next section in assessing areas flooded. HYDROLOGY The Mongalla-Gemmeiza [oiell is 48 km. long and varies from 2·5 to 8·0 km. in width. Its total a rea is 161 sq. km. The va lley is bounded by high grou nd on the east, from 2 to 6 m. above the average lOi('h level in the first marked rise from it. On the west the area is bounded by channels of the Bahr el Jebel whose right bank is alluvial and whose level is on the aveFage 1·67 m. above the average [vicll level. The [Dicit is also divided into a number of longitudinal sections by the a ll uvial ridges of a number of modern and old watercourses which traverse it from so uth to north. Over most of the area these subsidiary ridges are roughly as high as thc ma in river bank , but in the north thcy tend to disappear. The following arc the significant slopes from south to north: Average water slope at Mongalla discharge 75 mi d t7·8 em/ km. Average b;.\ nk slope t9 em/km. Average to;ch slope 26 em/ km. The following sections refer to the five dilTerent ways in which the flood-plains in the Southern Zonc rcceivc water. Details are given in Tables 388 to 390 (pp. 841-2), to which the reader shou ld refer. ( I ) Spi ll ovcr the bank crest begins at Mongall~ discharges or rrom 120 m/ d:::lt the southern end to 90 mi d at the nor thern end. Spill into thc va llcy begins in the largest khors and spill< hannels (over 1·0 m. decp), of wh ich there ure 14, at disch.uges vary ing from 70 mi d to 50 mi d, according to location. T here arc in addition 214 spi ll-cha nnels less than 0·5 m. decp, and 52 between 0·5 m. and 1·0 m. in depth. There arc also points on the b~lnks 0' 15 m. on the average lower than the levels given. Thus when wa ter-level rcaches the b~ nk level given here it can be safely assumed Ihat water is spilling over . the crest of Ihe bank in appreciable quantities. Owing to the ridges which divide the toich into strips longitudina lly, even when the river spills over the crest of the bank water cannot spread vcry far eastwa rd s. perhaps from I to 1·5 km., since the ridGes form banks effectively barring the way. Only a narrow sirip borllcring lhe river is affected dircctly by river spill. W aLer pouring into the vn lley in decp khors is retained in them until the khors disappear in the lake a t the northern end. (2) Except ror these watercourses, noodi nG from the river in the vaUey is almost entirely due to backwater from the northern end. when the bank is topped, and through the large channel just south of Gernmcil.',I. (3) Another important source of nooding which affects the easlcrn edge of the vaHey is run-off from tbe va lley side, and Ihul brought in by some small watercourses of local origin. (4) Si nee there is :'1 larger percentage of coarse sa nd in the soil here than in that of the ALiab Valley. the en-cct of underground scep:'lge both from the river and from the watercourses running through the va lley must be borne in mind. The width of Ihe alluvial bank east of the river vades from 200 to SOO m. (5) The averaGe annual r:1infaJl at Mongalla is 929 mm. On the basis of the a bove points, we have calculated areas of the valley inundated by tbe river a t different discharges a t Monga Ua. Up to a discharge of 70 mi d only a small part of the fl ood-plain is inundated. At discharges from 70 mi d to 110 mi d the valley becomes submerged at a fairly steady rate, and at 120 mi d becomes almost fully inundated. The relationship between pasture and flooding in the MongaUa-Gemmeiza [Did, is one of the most important features. On tbe hydrological side we have calculated the areas inundated related to MongaUa discharges a nd the duration of flooding which occurred in the 5-year period end ing in 1950. As already stated, the Team has decided that the reasonable empirical rule to be applied for the duration of flooding of pasture examined should be the duration in the five previous years, i.e. from 1946-50, weighted in favour of the final year by including that year's flooding twice (see Vol. I. p. 156). We have calculated the areas in- undated on the assumption that flooding starts at the northern end and moves southward.s. , < 840 When considering vegetation , due allowance must be made fo r run-ofT in small local khors, spill in channels, the underground water-tables which are mentioned in the above description , .. and the average annua! rainfall of 929 mm. at Mongalla. TABLE 388 MONGALLA-GEMMEIZA TO fCH HYDROLOGICAL DETAILS I WATER-LEVELS CoRRESPONDING TO MONGALLA DlSCHARGES Cross-Section : millions of c·ubic mel rcs per day Number :_ .• ________ , ____ .• ___ . _ __._ __. ______ ~ __ _ 1_ _so_ _I ~_I __7_ 0_: __8_0 _ :_ ~ __ :_~ __ I_ _ ~ I~ 120 I 437-77 438·03 438.39<'1 : 438·68(' ) : 4)8 ·95 4)9,20 439·49 4)9-68(', 6·84 7·18 7-49<' , . 7·79 8·0) 8-28 8'50 8·69<" : 5-89 6'24(', 6·53 6·82(') 7·08 7·)2 I 7·53(" 7·7, '" !. :~(') ! :~~('W' !:~ ~:~~ ::: ::! I :~:'" :~" 3'59 3-95 4·24", 4·52 4·76 4·99 5· ' 8", 5·)2 2-4()(" 2·77 3·05 3·33 3·58 3·19<', 3-91 4·10 1·73 2·09<') 2·38 2-66 2-89 3·09<') 3-21 ) ·)9 0'97'~ I ,)) 1·60 1·88 2· 11 2-)0<') 2-46 2'58 10 O·2S('} 0·60 0·88 1·15 1·38", 1·56 1·10 1·80 II 429'50 429-851' ) 0" 2 0·39 0'62(') 0-18 0·90 1-00 12 8·85t') 9·19 429·45 429·70 429'94,'1 430·10 430·20 430·30 e' l Bahr cl Jebel rilM bank crest kvcl. (') IDvcl1lenlorlal'lclpill-cblnnds. (') A~'c:t1IlC: l{)fe}, level. Water slopes :n MoncalLa MQn Oiscbarrc or 15 mi d : MOf\&.Jla tokm.1J 19-7cm/ km. km.13 lokm. 34 18·. . km.l4tokm. .s7-" 1S-7 TABLE 389 MONGALLA-GEMMEIZA TOICH TOPOGRAPffiCAL DETAILS-BANK ~oss..sECT10N D istance Invert Average on Bank R.L. of B.nIt Fall Bank Slope Level Deep Told Level Number Bank km. Bank cm/lcm. Khors R.L. km_ 1·7 439·10 438·30 438·10 6·1 4'4 8·62 1·08 24·5 1·50 10·6 4·1 7·52 1·10 24-4 1·10 6·10 6·00 1503 4·1 6·10 0·82 11·5 5-30 HO 5·2S 20·0 4·1 5-68 1·02 21 ·1 4·00 23-3 3-3 5·06 0·62 18·8 4·00 3·00 30·0 6·7 3-65 1'41 21·0 2-31 2·00 33-8 3-8 3·00 0·65 11·1 1·90 1·00 ( I 38·8 5·0 2' 15 0·85 11·0 1·50 430·35 430·50 10 1 -'43,5 4·7 )028 0·81 18·5 429·85 429·50 \1 48-4 4-9 430·50 0·78 15-9 9·15 8·80 12 I 52-7 4·3 429·90 0·60 14·0 8'60 425·15 427-50 -----.- T01= 1 51·0 9·80 19-2 .. 1 • • / .' --r 841 ;' TABLE 390 MONGALLA-GEMMEIZA TOICH TOPOGRAPHICAL DETAIlS-TOICH CRoss-SECTION ! TOICH LEvELS .\ Average Average Height \If Lowes; - Toicll Toich Bank , ---- Straight line - Average - -.- I Fall Slope I above A veroge Number Distance R.L. R.L. cm/ km. ; Tojrh km. 1 ---. ,--- --~----'-- -- -' - -,,-- ._--.. _1 ---- I 438-50 436·85 1·20 2 3-8 7-50 6-00 1-0 26·3 1·12 3 4·1 6·00 4·90 1·5 36·6 1·52 4 H 5·25 3-80 0·75 19·2 1·45 5 4·1 4-00 2-20 1·25 30·5 1-68 6 H 3·00 1-00 1-0 31-2 2·06 7 3-8 2·00 430·60 1-0 26·3 1-65 8 H 1·00 429-60 1·0 27-8 2·00 9 4-3 430-50 8-75 0·5 11 ·6 1-65 10 4-1 429·50 1·0 24-4 1·78 11 3-7 8·80 i~ ; 0-7 18·9 1 1-70 12 3-8 427'50 4-27·25 -I-- -1-3 -II ·_~ _L._2'4O ,...: Total and 11 '-0-0 ---t Averages ! 42·4 1 25·9 I 1·68 1- 8 I 26'5 7·50 28·3 8-12 . 15-9 3·50 22-0 TABLE 391 MONGALLA-GEMMEIZA TOICH AREAS FLOODED RELATED TO MONGALLA DISCHARGES AND AVERAGE DURATION MongaJJa Discharge Area Inundated Average Duration mi d sq. km. 1946-50 plus 1950 days 120 161 ·1 (Total) 7 110 157·7 20 100 127·3 37 90 67-9 71 80 38·2 108 70 17·7 173 60 11 ·5 227 50 8'4 293 40 5·0 347 30 2'5 360 SOILS During 1951-52 the soil survey was extended to cover the reach from Juba to Tombe, 58 soil sample pits being dug on cross-sections A to L, as marked on the diagrams (Figs. A 2- 7 and H 14-16). Of these, 19 pits were between MongaUa and Gemmeiza, as shown on the map (Fig. H 12); only 15 were actually on the toic/z, so the survey was far less detailed than in the Aliab Valley. Without analysis, the descriptions of the samples on the toich do no more than indicate a greater proportion of sand than in the Aliab Valley, though the soils may still be classified as clayey; a number of holes contained stratified layers of sand and clay. In the drainage- channel carrying run-off north beside the forest, clay seems to predominate. To the east of the Mongalla-Gemme;za road on the afforested high ground, 31 pits were dug at I Ian. intervals along C.S. X, C.S. Y, and C.S. Z (Figs. A IG--12) and examined. 124 samples were collected from 22 pits. These samples have been sent for analysis';,n~connection with the proposed sugar plantation scheme, and results are expected shortly but not in time for inclusion in this report. The holes dug on C.S. X, 4 km. north of MongaUa, and the traverse notes show that ironstone soils extend from 3·150 to 1·850 km., and east of 0·970 km. on the section; sand predominates elsewhere. C.S. Y, east of B.M. 64, indicates a predominance of clay from the road to the rise 3 km. to the east, and a predominance of sand on slightly higher ground_ 1>!:yo,n~:. 842 and 2. Taking soil samples (A liab Val- ley). - 3. Deep pool with Ni le Cabbage (Pislia slralo;des). .' 10 14 (, PLATE\!.;.I ,,) \,. ,) ') 1. River fl ood-pla in under water (ncaf Bor), Nc;\ rl y a ll the swnmp grass is £('/1I110Cll/Oll .\'I(I~nil/a, Grass Sampling Uni t No. I was sited by the nca r sick of the huge pool (Decembe r 1951). ~ 2. Inlet on east bank of Ba hr eI Jebel nca r Bor. Flood-water in backgro und has receded and the \. swamp grass has a lread y been grazed flat ( 195 1). 3. ) close view of the edge of the Aood-plain with !load season growth o f £ch;l1uddu{/ .,·WKII;II{/ recently exposed. It is a lrea dy parti y gra zed and trampled. All this thick growth is edible (December 195 1). 4. Exposed ri ver fl ood-plain nea r Jo ngki with coarse g rowth or EchillodJ/oa pyramidalis. This is burnt and the subseq uent regrowth grazed (November 1951). 5.\.. A d ose view of regrowth of £chilloch/oa pyramidalis after burning. Note the very poor ground ·\:over. A 2:} ~ fi lm box indicates the height of the regrowth . 6. Hrparrlienia n~fa. Coa rse rains season growth being fired to en.cou rage regrowth . Pengko grazing trials (December 1951). 7. H.I-'parrhenio rufo. The co~ rse rains seaso n growth has been burn t, but no regrowth h HS yet occurred. Eastern Pla in : Sampling Unit No. VI (May 1951). 8. Robust regrowth of H),parrlien;a ntIa afte r bu rning of coarse growth. Eastern Plain: Sampling Unit No. rv (May 195 1). 9. Nilo tic cow grazing on deep-flooded pastu re in Faddoi Pool, Central Nuer Distri ct (December 195 1). 10. Dura (Agona) on . Killifer ' ridges: Nagdiar road nea r Malakal (December 1951). 1 J. Dura (AgOI/O) on' Killifer ' ridges (le ft ) and o n the fl at (right). Note the uneven spacing and indifferent growth of the latte r (December 195 1). 12. The effects of fl ooding. Dura sown on the fla t. Note the tall, flouri shing stand of d ura (cf. No. II ) in the background . 13. The effects· of drought. Feterita sown in October 1951, two months before the photo was taken. Germination, survival, and growth were very poor. Bor Dis trict (December 195 1). ""'" 14. Castor sown 7.6.51 Bar Distri ct (December 195 1). 15. Ploughing oxen (N ilotic), Bar Dist ri ct (December 1951) . ./ "v .1 C.S. Z, east of B.M. 72, shows a predo minance of clay thro ughout. If the resulls of analysis confirm this, the three sections illustra te a tra nsitio n from south to north in soil type as in \ topography. Yl'?ETATION ,~A vegetation map (Fig. H 13) of the Mongalla- Gemmeiza loiel, has been compiled from tra" erses along the cross-sections, notcs taken during the survey of the bank sectio n, and fro m air photographs. This map, like that of the Aliab Valley, illustrates the importance o r hyd ro- logical conditions, mainly depth and duration of fl ooding, among the factors governi ng distribution of vegetation on the fl ood-plain. Comparing this map with the con tour and flooding map, it will be noted that open wa ter and Cyperlls papyrlls occu r althe northern end of the valley, permanently flooded through the head of Lake Buri . A band of C)'perus pap)'rus and Vossia clispidata is found between 40 and 48 km. on the bank sect io n, where the bank is relatively low and therefore inundated at 80 mi d, and where permanent spilling occur. through channels. This spill flows north into Lake Buri on the eastern side of C.S. I I, where a combina- tion of flooding, run-off, and spill produces a band of Echinochloa slag"i"a, mixed with Vossia cuspidata near the lake. '-. Large spill-channels at 30, 35, 3::' and 40 km. on the bank section lead into areas with little dra inage to the north ; this spill combines with seasonal inundation of the bank to produce near the river a band of Echinochloa slag"i"a, Vossia clIspidala, and some Cyperl/S papyrl/s. South of Terakeka, where backwater flooding occurs for a very short peri od, spill and seepage produce a belt, parallel to the river, of a little Echi"ochloa slag"i"" and Vossia clispidala interspersed with EchiJlochloa pyramidalis and Phragmiles commltnis. The large channel nea r the river, which carries this spill north, is blocked with Ecltillocft/oa slagni"a and Vossia cuspidata. Near the forest, where run-off from the high ground collects in depressio ns, Ecltinochloa stagnina is found. South of the 80 mi d flooding line, in the centre of the vaUey where longitudinal banks eliminate the effects of spill or run-off, Ecltillocft/oa p)'I'amidalis and Phragmites cOl11mW1is are predominant. These a re also found o n the bank of the Bah r el Jebel. The areas in square kilometres over which each associat ion is found, and the areas covered by each species, have been tabulated; these bave also been given as percentages of the total area surveyed (see Table 392 below). Tbese distributions have been plotted on a graph (Fig. G 23) showing areas flooded a t given Mongalla discbarges, and for given durations in days. The grasses have been plotted in steps according to associations, those found in the north of the valley in the a rea of greatest flooding being plotted on tbe left of the graph, and the others in order from north to south. ExceptioDs to this are those areas where some Ecltillocft/oa slaglliltQ and Vossia (,ltspidala, due to spill with little flooding, are found interspersed with Phragmites COllllllllllis so uth of Tcrakeka. Tbese bave been plotted separately from otber areas of the same species, as have the a reas of Ed.inochloa stagniJlQ caused by rUD-off in the south. This grapb shows the optimum durations of flooding for each vegetative species tabulated below (see Table 393). TABLE 392 MONGALLA-GEMMEIZA TOICH PRESENT DISTRIBUTION OF VEGETATtON SPECtES I AREA MEASURED f ROM j ! ___ .. NOR~ __ Areai;v~red l ATea Total Covered A rea Pcrexi"lta3c assoclatlon 1 1_ . from_ .__ to . wi~p~~her , ____________ ~: km: 1 sq.km. I sq.km. ! sq. km. ! sq. km. 1 Open Water .. . II 2·3 2·3 2-3 2-3 t ·4 Cypuus papyrus . .. 2-3 10·9 8-6 4-8 4-8 3-0 Vossla cusp/data. ", "' 1 ) ,6 13-5 9·9 3-0 3-9 2-4 £CJrinochioa Iol.l}jnina ... 4-0 36·6 32·6 19-6 32·8 20-4 Eclmwchloa pyramIdalis ... 20·6 155-3 134·7 84· 7 84-7 52-6 Pltrogmiles communisI .. , ::: , 56·7 ISS,) 98 ·6 J2·6 32·6 20·2 Vossia cuspidQIQ spill and I' 56·7 65·4 8·7 0·9 l Echinochloa slagnina seepage 57·4 78·6 21,2 Echfnochloa SIDgllina (run-oH) .. . I 155·3 . 16H 5-8 7-4 • - I - ___5_ '8_ __ ~_-= __ -= I Total 161·t 161-t I JOO'O 843 { TABLE 393 MONGALLA-GEMMElZA TOICH DURATION OF FLOODING OF VEGETATION SPECIES (From Fig. H 23) ----------------------------~----~~~------------------------__f,~ Species Minimum Days Maximum Days Optimum Days ___- L_ ____ _ _ L_ _ .... _. . '._ .. ._ Open Water .. 360 360 360 Cypults papyrus 256 360 320 Vossia cuspidotQ 216 360 256 Echinochloa sfOgnina 110 360 180 Echinochloa pyramidalis 10 160 98 Phragmius communis 10 82 42 LAND UTILIZATION The Mongalla-Gemmeiza ,oich is used by Mandari from both east and west of the Bahr el JebeL On the east bank the Mandari extend from north of Mongalla to about 20 km. north of Gemmeiza, and their permanent villages are grouped along the ridge near the Mongalla-Gemmeiza road. Where the Bahr el Jebel adjoins the forest south of Terakeka, the loich on the east of the river is used by Mandari from the west bank. On the east bank, crop husbandry plays a more important part in the economy than among the Dinka farther north, and it is possible to use the loich for this purpose as it is less frequently inundated than that in the Aliab Valley. Besides the millet and groundnuts grown in clearings in the forest, maize and some tobacco are grown on the banks of the Bahr el Jebel wherever these are easily accessible. Cultivations are to be seen in a narrow strip extending up to 50 m. from the river between the following points on tbe bank section-from 0 to 3·2 km., easily reached from the east bank, from 14 to 26 km., accessible from the west bank, and from 53·6 to 56·2 km., near Gemmeiza. The total area on this one bank is estimated at 100 feddans. In April 1952 cultivations were prepared for planting in the south, and crops were well established near Gemmeiza. The positions of villages and cattle-camps have been marked on the vegetation map (Fig. H 13), together with figures of cattle population. Permanent villages and their associated cattle-camps are joined by a broken line, while approximate grazing areas are marked by arrows. It will be seen that all cattle-camps are on the banks of the Bahr el Jebel. At the extreme south of the loich and north of Terakeka there is movement from permanent villages to cattle-camps, possibly because a number of drainage-channels separate the high ground from the grazing produced by spill near the river. For 8 km. south of Terakeka the grazing near the river is used by cattle-camps from the west bank, while on the east side the cattle are not moved away from the permanent villages, presumably because adequate grazing is accessible nearby. Cattle-camps are normally occupied from December or January to the middle of April, but, as permanent villages are near the loich, some use is made of the verge of the loich throughout the year. The total number of cattle supported by the Mongalla-Gemmeiza foich (on the east bank) is given below, with estimated totals of sheep and goats: Tribe Cattle Sheep and Goats Area (sq. km.) Mandan (E. Bank) . 4,010 Mandari (W. Bank) .. 1,220 Total 5,230 10,430 161 FISHERIES A fishery survey, as carried out in the A1iab Valley, was continued during 1951-52 to the south as far as J uba. On the east bank between M ongalla and Gemmeiza only one stretch of open water suitable for tbe casting-net was found-Lake Buri, south of Gemmeiza-so it will be convenient to deal with the whole reach from Juba to Gemmeiza. The resul~ of the survey have been tabulated according to species and according to fishing sites in Tables 383 and 384 (pp. 836-7). The first, giving the total 1951- 52 catcb-hy species, includes some fishing in the Tombe channel, but shows that, as in the Aliab, Tilapia and HelerOlis were most frequently caught; Clarias were seldom found, while Cifharinus were more common than farther north . . Fishing sites A to E are south of Gemmeiza; sites D and E are marked on the map (Fig. H 13). In spite of extensive search and enquiry, no other sites were found on the east bank. As the second table shows, an attempt was made to use a casting-net in various channels near JUb~, 844 but failed because of the high banks, the depth of water, and the strength of the current. In Khor Nyaraya, near Gondokoro, catches were small. The la ke so uth of Mongalla and also ;Lakes Buri and Moni were fished more successfully; over the whole reach, in fact, catches increased from south to north. I . There is at present a commercial fishing ca mp on Lake Buri which sells dried fi sh to the south. Apart from this there is no permanent open water and there are few permanent khors. On the east bank at least, fishing seems to playa relatively unimporta nt pa rt in the economy; there is no specialist fishing section like the Monythany among the Dinka, and the Manda ri rely solely on somewhat haphazard methods of spearing fish, especially during the ra ins. Near Lake Buri, however, where one area of seasonally fl ooded loich is named' The Lake of Fish " fishing is carried on throughout the year. South of MongalJa the Bari fish with hooks in the Bahr el Jebel, and also fish in pools during the rains. On the west bank , however, Lake M oni is one of several large lagoons where the Fisheries Section have shown that commercial fi shing has greater possibili ties (Vol. I, p. 387). EFFECTS OF THE EQUATORIAL NILE PROJECT ' .. In order to assess the effects of the Project on the Mongalla- Gemmeiza loich the area has had to be divided into three main sections : the portion fl ooded at a Mongalla discharge of 57 mi d ; that exposed at 57 mi d but flooded at 90 mi d ; and that not fl ooded at 90 mid. The second and third portions have each been sub-divided into three types of area: those affected at present respectively by spill and seepage from the river, by run-off from the forest, and by neither. The hydrological effects of the Project in the area may then be tabulated as follows: Number From I To Area Effects ... _--_._------- - 0'0 10·5 10·,5 sq. km. 360 days' Hooding 10·5 33-5 23·0 .. 192 .. ftooding plus 168 seepage and rainfall Jl·S 39·6 6·1 192 .. flooding plus 168 run-off and rainfall 39·6 67-9 28-) 192 .. flooding plus 168 rainfall 67-9 87-6 19·7 192 .. seepage plus 180 rainfall 87-6 93-4 5-9 180 .. run-off and rainfall 93-4 161-1 67-7 180 rainfall By comparison with areas where similar conditions are found at present, and with the present vegetation in the area, the effects on vegetation under the Project have been estimated and are shown in Table 394. These effects have also been plotted in diagram form (see Fig. H 25). For comparison, the present and future estimated distribution of species are given in Table 395. TABLE 394 MONGALLA-GI!MME1ZA TOIcn FUTURE DISTRIBUTION OF VEGETATION SPECIES Number Area Open Cyp~l'us Vossio n;:tf~a I ~tf~Q II Phragm/f,e HYl!orr' l1 sq.km. I w_.t_e_r +_P_QP_Y_ru_' +_'"_,'P_;d_Q_'Q+_" _QK_n_;n_Q+ pY_'_Qm_'_'da_"," I_communls hema sp. --~--+I -b-:-:?-i--~ · ,lj f: :1 :: ~- :: -It D~': 4 28-3 - - 8-S 14·2 5·6 - - season 5 19·7 - - - 4·9 8·9 5·9 6 H - - - 5·8 - - 7 67'7 - - - - 20·3 33·9 ,i-s T_ o_t_a1_ _'_ "-jI_ I_6_' -'_ -t_ _7_' 3_-+-_'4_-7_i-_17_'8_i-_3_3_'2_+-_3_4:~ ' 1--3-9'-8 - .. 13·5 - - - Percentaaes I 100-0 4·5 9·1 11·0 20-6 21-6 I 24·8 8·4 845 TABLE 395 MONGALLA-GEMMElZA TOICH PRESENT AND FUTURE DISTRmUTION OF VEGETATION SPECIES Species ~ESEI'\'T ESTlMAT,W DISTR£Bl1TION F11 ___tJ_TUR _ _£ _E_sT1_M_AT, £_O _D_ I_""_ "D_tJ_TI_O_N,_ I s~. kIn. : % sq. kIn. ! % ~f,;r:~~;yrus :.· ··!I··· -- ~) : 839 1---21: ~4- -'1--- 7·) 4·5 14·7 9· 1 Voss;o ruspidQIQ 17·8 11 ·0 Ffc7,~~~~7.~:: ~~!a~:,~~afis ~~ :~ ;g:: I )3·2 20·6 )4 ·8 21·6 Phrogmilt!s cO/wHlmls ~ 32·6 20·2 I )9·8 24·8 In:ned ia~: ~P_. 1),5 _(H_J'parrhnliO) "'_t I 8·4 I ----j-,- T OTAL "' 1 16 1·1 100·0 I 161·1 100·0 6. CONCLUSIONS The conclusions reached may best be summarized with reference to the objectives of the survey, defined earlier. It has been shown how each unit of the flood-plain must be treated in isolation, and how a complete survey of the unit area, including the river ba!)k, is required for the determination of flooding. The way in 'which flooding is mainly due to backwater has been demonstrated, and how it is caused by the fall in bank level relative to river level over the whole reach and along each section of the reach (see Fig. H 21). It has been pointed out that at the northern end of each lOich tbere is a channel through which rainfall or spill from higher up drains into the river and through which flooding occurs as the river rises. This has been noted in the Sudd region farther north , and the hypothesis has been put forward that the Bahr el Jebel formed its bank when it flowed at a steeper gradient, and that the lagoons in this region were formed by spill-water forcing an exit at the northern end of each reach (see The Nile Basin, Vol. I, p. 78). In the Juba-Bor reach analysis of the correlation between Rejaf, Juba, Mongalla, and Tombe gauges (see Vol. II, p. 534) has confirmed that the river water profile is tilting about a point between Juba and Mongalla because of the effects of sil ting or scouring. This explains the present inclination of water profile to bank profile and the inclination of both to the foich profile, brought out most clearly in Fig. H 21 , and thus makes clear the way in which flooding is greatest at the northern end of the reach, and also at the northern end of each unit of flood- plain. This process, if it has been continuous in the past, may explain the discrepancy between early descriptions of the large herds of cattle between Juba and Mongalla (see F. Werne, The White Nile , tr. C. W. O'Reilly, London, 1849 ; Vol. II, pp. 71 , 95) and the present shortage of good toich grazing in this reach. It was claimed by a guide, who pointed out the erosion of the old post at Gondokoro, that there used to be far more fl ooding in this area than at present. The relation between soil texture and surface organic horizon and the hydrological regime has been described, as has the relation between flooding and vegetation. The graphs showing areas flooded in the Aliab Valley and the Mongalla-Gemmeiza toich (see Figs. H 22 and H 23), when compared, illustrate the differences between the hydrological regimes in the two areas- widespread flooding occurring in the southern area at higher discharges and therefore for shorter periods than in the northern area. The distribution of vegetation in the two areas may be compared by referring to the two vegetation maps (Figs. H 3 and H 13), Tables 381 and 392 (pp. 834, 843) showing areas of species, and the graphic representation of these tables on Figs. H 22 and H 23. Though other factors, including soil and slope, have to be considered, the dominant factor governing distribu- tion of species has been shown to be depth and duration of flooding. Compariso'l of Tables 382 and 393 (pp. 834, 844) shows that the optimum duration of flooding for each species is found to be the same in the two areas. But the two areas have different hydrological regimes, and therefore, although species occur in the same order from north to south in both areas, the difference in duration of flooding produces a different proportion of vegetation species-more Phragmites communis and less open water, Cyperus papyrus, and Vossia cuspidata in the southern than in the northern area. . '< 846 The dependence of tbe economy of the two areas on the hyd rological regime and the resulting vegetation has been described. The adjustment of the economy to tbe differences between the two areas has also been brought out- the greater degree of reliance on an imal . husbandry in the Aliab, and on crop husbandry in the Mongalla- Gemmeiza a rea, and the greater importance of fishing in the Aliab Valley. The effects of the Equatorial Nile Project in areas and duration of fl ooding and in vegela- tion changes have been described and shown in diagrammatic form. This has formed the basis for the estimate of losses in the Southern Zone. Basic infonnation for the design of banking of the Bahr el Jebel or canali za tion of lhe River Aliab in this reach, and also for the design of irri gation schemes in the Aliab Valley and Mongalla-Gemmeiza /oich, has been obtained ; the actual design of lhese schemes is descri bed elsewhere (see Vol. II, p. 664) . To sum up, tbe intensive survey of all aspects of a small area has yielded basic informati on applicable not only in the area itself but also in similar areas elsewhere. 11 is therefore of exceptional significance in an investigation of the Equatorial Nile Project and its effects in the Sudan and demonstrates the value of intensive sample surveys. 847 CHAPTER 2. AN ANALYSIS OF THE WHITE NILE FLOOD BETWEEN MALAKAL AND RENK by J. W. Wright, M.A. , F.R.f.e.S. INTRODU CTION This analysis of the White Nile Flood. together with a preliminary note on Ihe Sobal . was written in the summer of 1949. as the direct sequel 10 the determination of Aood-plain areas along the White Nile which was described in Appendi x IV of the Third il1lerim Report of the Jonglei Investigation Team. That appendix showed how lhe surface a reas of the ri ver at different stages could be calculated quite simply from its height above normal low level. These surface areas were then used to ca lculate the trough volume. the gains by rainfall, the losses by evaporation. and the losses by absorpli on during a number of floods on the White Nile. The results were compared with the differences between the ob erved discharges at Malakal and Renk . and this comparison gave figures for the amounts of water which had flowed into the river between those two points through tributary khors. Thus the shape of lhe river trough was used to calculate the main elements of its fl ood. In lhe preliminary note on the Sobat originally attached to this analysis an oUlline was given of how ra ther more complete observations of discharges on that river and on its tributaries might be used to derive from the elements of ils flood cycle the shape of ilS trough. The whole of this thesis was produced in a limited ' edition ' of nine typed copies for sludy by members of the Jonglei Investigation Team and certain other people who took an interesl in the subject. Since the production of that ' edilion ' several events have occurred which have led to modifications of the original thesis, which is now presented in a slightly different form. First of all the investigation of the Sobal fl ood has been carried oul in much greater detail and lhe results have made the preliminary note no longer worth publishing. It is therefore omitted from this paper and a full account is given later in this volume (Chapter 3). Secondly the calculations of the White Nile analysis have been checked by irrigation engineers of the Jonglei Investigalion Team, with resulting minor corrections to the values given in the first' edition '. None of these alterations is significan t but lhey have been incor- porated. Thirdly an error was found in the drawing of one of the cross-sections supplied by the Egyptian Irrigation Department on which the tables of surface areas. and lhus practically all the other calculations. were based. The effects of lhis error are in mosl cases less than the probable errors of the tables themselves, and to take full acco unt of it would necessitate com- pletely re-computing them. This bas therefore not been done. but a note is included in the present edition indicating the effects of this correction on the tables of both Appendix IV of the Third Interim Report and of this analysis. Fourthly I have had the opportunity of dis- cussing this paper with Dr. H. E. Hurst of the Directorate of Nile Control. and sometime Director of the Egyptian Physical Department which was responsible for most of the data on whicb it is based. His main criticism was that I had not made sufficient allowance for the possibility of systematic errors in the observed discharges. and 1 have therefore added a note on this at the end of Section II which deals with the observed discharges in detail. Finally. the last section---{)n future conditions-has been omitted as the subject has been dealt with in Volume 1I, Chapter I\, This brief note replaces the original summary of the analysis which formed the first para- graph of the introduction. Some other parts of this now seem redundant and have been omitted, but apart from this the original text of the paper has been retained with only minor alteralions. PRESENT CONDITIONS ON THE WHJTE NILE At the time this analysis was made there was available no detailed description of the White Nile betwe~n Malakal and Reok. In Volume I of The Nile Basin there is a brief account. comprising three and a half pages of text and illustrated by fourteen photographs. In spite of its brevity this included an outline of the main features of the river valley. drew attention to the unexpectedly small losses observed between Malakal and Khartoum. and suggested that there might be a considerable contribution from the many khors which join the river between these two places, particularly in the southern part of the reach. As will appear. this estimate. which I am able to confirm. was lost sight of by later investigators. Volume V of the same 849 work described the Lake Plateau and the Bahr el Jebel, and included a chapter on the White Nile between Lake No and Sobat mouth; it ended with the hope that one more volume would complete the detailed account of the whole Nile Basin and its hydrology, and this was produced in 1950 as Volume VIII. Various proposals are made in Volume VII (' The Future Conserva- tion of the Nile '); if these are carried out there will be profound changes in the regime of th~ White Nile north of Malaka!. It therefore seems worth some trouble to analyse in more detail than has been done hitherto the hydrology of this reach, for only thus can the effects of the proposed changes he forecast witb confidence. At present the flow in this part of the Nile comes essentially from two sources, one almost constant, the other with a considerable annual fluctuation. These two sources are the main channel below the mouth of the Zeraf, and the River Sobat (see Map I). The first has a discharge which remains practically constant throughout tbe year, the mean value being about 39 million cubic metres a day. (This unit of discbarge will be referred to hereafter simply as , millions a day' or ' m/ d '. It may be helpful in following the arguments below to think of a million as a square kilometre covered to a depth of one metre. The corresponding British unit is an acre-foot, which is self-explanatory ; there are approximately 830 acre-feet in a million cubic metres.) The mean Sobat discharge, on the other hand, varies from about 7 millions a day in the dry season to 67 at the height of the flood. Thus at Malakal on the average the discharge of the White Nile, being the sum of these two, varies from 46 to 106 millions a day, the minimum being in March or April and the maximum in October or November. EFFECTS OF PROJECTED CONTROL WORKS. As is well known the Egyptian proposa ls include dams at Lakes Albert and Victoria and a canal through the Sudd from Jonglei, in latitude 6° 50', to a point on the White Nile just upstream of the mouth or the Sobat. By this means it will be possible to vary the flow at this point considerably so as to counterbalance tbe varying discbarge of tbe Sobat and cause a much more constant flow downstream of the junction in any ('ne year. The discharge of a river is closely connected with the level which it occupies in its bed, and the effect of running the White Nile at a constant discharge would be to keep it at a correspondingly constant level. Under present conditions there is a variation in its level at Malakal which averages just over two metres each year. The river is low from February to April, rises slowly for six months till October or November when it holds its maximum level for about a month, and then falls fairly rapidly so that by February it has again almost returned to its lowest leve!. Because its valley is very flat this small change in level is enough to flood and dry out a considerable area of land alongside the main channel, in what is here called the flood-plain. On this land, which is uncovered as tbe river falls in the driest part of the year, the pastoral tribes who live in this region depend for the only grazing which is then available to them. As explained elsewhere in this report, there may be grazing inJand at this time but it cannot be used for lack of drinking-water or because it is unpalatable. In considering the Egyptian proposals for Nile control the Sudan Government has been greatly concerned at the possibility of losing this valuable natural grazing. To investigate the problem has therefore always been considered one of the most important tasks of the Jonglei Investigation Team. The first stage in this investigation was clearly to estimate the area of land flooded in a normal year under present conditions so that some idea could be obtained of how much was going to be lost. This could not be done by measurements from existing maps in the stretch south of Jebe1ein, because the best available maps were only on 1/ 250,000 scale and did not show contours near the river. In the First Inlerim Report of the Jonglei Investigation Team, published in 1946, an extremely ingenious attempt was made to estimate the area of the flooded land between Malakal and Renk by a consideration of the discharges which are measured at these two points. I shall describe this analysis in some detail because the present paper is based to a large extent on it. I have gone into more detail and used certain data about the river valley which were not fully appreciated at that time, and as a result I have been able to turn the analysis the other way round and work from the shape of the valley to the water account instead of using this to estimate some of the cbaracteristics of the flood-plain. Nevertheless lowe to this analysis most of the fundamental idea of relating observed discharges to theoretical ones and so being able to analyse the water account. ESTIMATES OF THE FLOOD-PLAIN AREA The outlines of the analysis in the Firs/Interim Report are briefly as follows. Study of the observed discharges at Malakal and Renk shows that when the river is rising those at Malakal are the greater, and when it is falling tbey are the smaller. This was thought to be due to water' spilling' over the banks of the main channel on to the flood-plain alongside when the 'r ~50 river rises, and being returned when it falls. At the end of the Aood there is a deficit beca use some of the water which has been' spilt ' has been lost by eva poralion from its surface, and some was also lost when the water first covered the ground and was absorbed into it. If the total depth of water per unit area lost by these two agencies is estimated, the tota l a rea Ilooded can also be calculated; it is only necessary to divide the total deficit by the tota l average depth of water lost. By this method an estimate of 500 sq. km. was obtained for the maxim um area flooded in an average year between Malakal and Renk, corresponding to a rise on the Malakal gauge of 2·26 m. , which was the average for the years 1912--42. As wi ll appea r below, tllis estimate was to prove very close to that obtained by an entirely diRerent method , in spite of the fact, as will also appear below in due course, that most of the data a nd facto rs on which the estimate were based were incorrect. It was reali zed at the time that s.ince the total depths estimated for evaporation and absorption were necessarily very rough, the estimated area must be regarded as only an approximate fi gure, so that it could not be regarded as a final solution to the problem. It was suggested therefore at this time, and proposed again more strongly early in 1948, that several series of air photographs should be taken at different stages of the flood in order to obtain a more reliable result , and one which could be expressed in terms of different levels of the river and so be used for a detailed estimate of future losses due to alt~rations in these levels. This proposal was fully discussed by me in Appendix IV of the Third Interim Report, which was published early in 1948. A modified version of this was also publ.ished in the Geographical Journal (Vol. CXIV, pp. 173- 90). I showed that the proposed air survey might well prove disappointing, chiefly because of the masking effect of the long grass on the Hood- plain; and I was also able to show that it was not really necessary. This was because the series of existing cross-sections which had been surveyed by the Egyptian Irrigation Department be- tween Malakal and Iebelein could be used to give for any level of the river an average value of the flood-plain width on which considerable reliance could be placed. From internal evidence it appeared that the probable errors of these mean widths were in most cases below 15 per cent. of their value; and this was confirmed independently from air photographs taken at the height of the 1944 flood , whose levels were of course known . These photographs covered a stretch of river nearly 60 km. long south of Reok, and also nearly the whole of the reach between Malakal and Melut; and the differences in both these stretches between the ' observed' and predicted flood-plain areas were within the probable errors of the latter. In order to make these estimates I made use of a concept, which may be original, in the shape of an idealized bank profile. This and the development of an idealized trough from it are described in the next section. By this means the maltimum area flooded in an average year between Malakal and Renk was estimated to be 470 sq. km.----{)nly 6 per cent. less than that estimated in the analysis already described. The immediate and practical part of the problem was in this way apparently solved, and it was clear that tbe effect of any proposed changes in the river regime could be turned into estimates of tbe flooded areas which would be lost, with a fairly small margin of error. The close agreement with the previous estimate, which was obtained by such very different metbods, was very gratifying, especially in view of the scanty data by which it in particular bad been obtained. Nevertheless this agreement in results was in fact largely fortuitous, because the earlier analysis was based on faulty grounds and did not present a true picture of the mechanism of the White Nile flood. To what extent the correct understanding of this is of practical importance at this juncture must be a matter of opinion; but since I believe that a true picture can now be presented it has seemed worth doing so, although it entails analysing tbe available records in considerable detail. Before I proceed to that part of the paper T feel it is wonb outlining the theoretiC'll grounds on whicb the analysis is based, and tbis is therefore done in the next section. I bave been unable so far to trace any record of a similar approach to the subject by any previous student of it, and it seems therefore that my ideas may to a certain extent be original. Moreover they certainly differ considerably from the current conceptions of the hydrology of the White Nile and the Sobat held by the irrigation engineers of Egypt and the Sudan. Tbere is therefore some excuse for describing tbem in detail, even though it means repeating some of the ideas expressed in the Third fllIerim Report in a slightly different form. SECTION 1. THE THEORY OF THE IDEALIZED TROUGH lDEALIZATION OF A SINGLE CROSS-SEcrION As was clearly shown in Volume I of The Nile Basin by description and by the use of aerial photographs, tbe Sobat and the White Nile valleys have similar though rather unusual forma- tions. On the White Nile this formation is confirmed by tbe measured cross-sections already -mention~, but unfortunately the Sobat valley above Abwong has never been surveyed on the 851 ground and there are therefore no complete cross-sections of it. The typical cross-section of the White Nile valley is of the form shown in Fig. K 14 (i), the vertical scale being exaggerated some 200 times. (For some actual cross-sections reference may be made to Diagram 5 of Appendix IV of the Third Inlerim Reporl and to Fig. 2 of the paper in the Geographical Journal.) It will be seen to consist of a main channel with a number of subsidiary shallower channels or lagoons running parallel to it; these only communicate with it by cross-channels at infrequent intervals. The total width of the water surface at any level h above the low water-level is given by the sum of the widths of the separate channels, i.e., aa' + bb' + ce' + dd' . The idea which I developed in the Third Inlerim Reporl for estimating the average widths of the flood-plain was that of a mean idealized bank profile. This time I want to make use of a mean idealized cross-section and so of an idealized trough. If the distances aa', bb', etc. are measured on the true cross-section for various values of h from zero up to the highest level, we can construct an idealized version of the cross-section in which the width at the corresponding height h above low level is always equal to the sum of these separate widths. This is one property of the idealized cross-section, the other being that both banks have the same profile so that the cross- section is symmetrical. This idea lized version is shown on the right of the diagram, and it will be clear that zz' = aa' + bb' + ce' + dd', for any value of h. APPLICATION TO SEVERAL CROSS-SECTIONS AND DERIVATION OF A MEAN When this idealization is applied to a number of cross-sections on the White Nile it is found that their idealized versions have a remarkably close resemblance to each other, in spite of the dissimilarity of the cross-section s themselves. This is only apparent if the widths are compared at the same heig ht above normal low level. This will of course be at a different height above sea level at each cross-section, owing to the longitudinal slope of the river bed. This is the first step in which 1 differ from current idea$, for the widths of-the White N ile between Melut and Jebel Aulia have been worked out by the Egyptian Irrigation Department from their cross- sections, but it was all done in terms of reduced levels, or in other words in terms of heights ahove sea level. The differences between our methods are discussed more fully in Section III (p. 862), but this important difference may be noted here. When a number of idealized cross-sections have been obtained in this way it is then possible to construct a mean idealized cross-section, by simply taking an arithmetical mean of the wid ths of the river valley in the different idealized cross-sections at the same height ahove low water level, and doing this for all the different heights from zero up to the highest flood level. This mean idealized cross-section has two properties of immediate practical value, and a third property which wi ll be used in considering the Sobat. The firs t property is that it is simple a nd symmetrical and therefore its width at any height above low river level (or below it) can be expressed in terms of that height. The second property is rather unexpected: it is that the meall cross-section obtained in this way from several cross-sections distributed over a long stretch of river gives a more accurate value for the mean width at any level in a comparatively short stretch than one obta ined from only one or two cross-sections. This is true even when these lie actually in the short stretch, so long as this is not less than 20 km. long. This second property of the long-term mean is due to the fact that there is a definite short-term variation of the total width of the cross-section of the valley at any given level above low river in terms of distance along the river. This short-term variation is almost periodic in character with a period of between 10 and 20 km. and an amplitude which is sometimes nearly as big as the average width. Over only a few kilometres the actual total width at any height may vary from about half to over twice the average value. This variation is brought out in Fig. J I , which was compiled from measurements made during the writing of the Supplement to Appendix IV of the Third Inlerim Reporl . It shows the flood-plain width between Sobat mouth and Melut as taken otl'the new 1/ 100,000 map, which was compiled from photographs taken at the height of the 1944-45 flood, when the maximum extent of the flood could easily be seen by a change in the tone of tbe ground covered by it. It will be clear from this diagram that the cbances of one or two cross-sections in a short stretch being located at points where the actual width is close to the average are .small. On Ihe other band, in spite of this local variation, it will a lso be seen that the long-term variation is small, and the mean of the whole reach from Sobat mouth to Melut gives a representative idea of the means of any shorter stretches, so long as these are not less than 20 km. long, and if allowance is made for the gradually decreasing rise of the flood downstream of Malakal. The practical consequences of these two characteristics of the mean-its short-term variation and its long-term uniformity-are that in estimating the average width at any level for a short 852 stretch, it is better to use a mean derived from nil the cross-sections in the reach ra ther than the one or two which happen to fall in or at the ends of the short strotch . . The proofs of th is were given in Appendix IV of the Third IlIlerim Reporl and in the pa per in th" Geographical JoumaL already cited, whero the means fo r several short stretches were deri ved both from ono or two cross-sections in them and also from the long-term mean, and compared with mean widths (or with areas, which is the same thing) obtained either from large-scale contoured maps or from lbe water surface at a known level plotted from an air photograph. On the average tbe long- term mean gave a value twice as close to the true mean as the mean obtained from only two cross-sections. It will be clear that this long-term mean must have some limits, and in fact it was found that lbe mean did change significantly if too 10llg a stretch of river was co nsidered. 1n fact throc distinct means were found between Malakal and Jebel Aulia, with the boundaries between them approximately at Melut and Jebelein respectively. But witllin each of these stretches the mean idealized cross-section appeared to be comparatively uiliform. By subtracting from each measured width the width of the low-level channel (or channels) at the same point, it was possible to get the shape of the idealized bank profile in terms of the height above low level. This disclosed the third property of the mean idealized cross-section, which is that its bank profiles, which are by definition symmetrical, are very approximately parabolas. This seemed to be true for all three reaches, and it was possible to express the di!ferences between Utem simply by varying the constant of the parabola. In considering the White Nile, analyt ical treatment of the ballk profiles in this way is quite unnecessa ry because they, and the mean idealized cross-sections of the dilferent reaches, are obtained by direct arithmetical mea n.ing of the widths measured on the ditIerent individual cross-sections. But in considering the upper Sobat, on which no cross-sections have been measured, it is necessary to cOllSider the shape of lbe bank pr06.le analytically and to derive a simple formula for it ; it is for this reasou that I have mentioned this aspect of it here. THE IDEALIZED TROUGH Once the concept of an idealized cross-section is understood tbere should be no difficulty in proceeding to the next step, which is to conceive of an idealized trough, since this merely represents a stretch of river in which the idealized cross-section is constant. Here again, however, I di!fer from current ideas since, in their published works at least, I think it is fair to say that the e!fect of studying the Slidd extensively has bred in those who have worked on the Nile an unnecessarily complex idea of conditions on the White Nile and on the Sobat. In the Sudd lbere is no doubt that the river valley- if one can caU it tbat- is in two distinct parts: the low-stage channel, and the flood-plain over which this' spills ' when the river bas risen beyond a certain point and has overtopped its banks. In fact in the Slidd this plain is observed actually to slope away from tbe central channel and not towards it. Because the White Nile flood-plain is also flat and wide, and to a large extellt is separated from the main channel by banks which rise above the main level of the plain, something of the same ide-a has persisted. Allbough in lbe actual working of the Jebel Aulia Reservoir, so far as I understand it, the whole valley seems to be thought of as one, on the Sobat certainly there is still a tendency to regard the valley as being in two parts. I believe that this is wrong, and that it is possible for a great many purposes, though not, as 1 sball explain, for aU, to treat the river vaUey as one unified feature, and to think of its idealized form as a single trough of the shape I have described, with sides which slope comparatively steeply near the edges of the low-water channel but fiatten out rapidly as they get farther away from it. If this concept is being used the idea of' spill' must be abandoned. True ' spill' is surely non-returnable water, and lbe difference between the ideas outlined here and those given in previous published works may be illustrated by a simple analogy. If a plate full of soup is tilted, some of the soup will lap over on to the flatter rim of the plate. But this soup is not spilt, for as soon as the plate is levelled it will return to tbe deeper central part. OnJy if the plato is tilted so far that the soup actually pours over the outer rim wi ll it be truly spilt, in the senso of being unable to return. It may be argued that some of the water on the flood-plain, because Qf absorptipn and evaporation losses, will not return to the main river trough when the level of lbe river falls, and this is of course true. But I see no reason for distinguishing between this loss from the flood-plain and the loss which is taking place all the time from the surface of the main channel. As will be seen below, in Section V (p. 869), when the river is falling during the dry months this last source of loss may be at least as large as the loss from the outer parts of lbe trough, i.e., from the flood-plain. Thus anolber respect in which I differ from previous published work on lbe hydrology of the White Nile is that I prefer to think of the 853 water as being held a t all times in the trough of the river, and not as being ' spilt' from it into a separate feature-the flood-plain-alongside. It will be seen below to what extent this concept can be justified, both by the ease with which it produces results and by the accuracy of them. CHARACTERISTICS OF THE IDEALIZED TROUGH It will be clear tbat the idealized trough is on ly equivalent to tbe actual river valley in some respects; in particular it cannot be applied when conditions of flow through the vaIley are being considered, since the irregular nature of the actual flood-plain and the grass which grows on it will prevent any but very small discharges over it. An estimate made by means of the idealized trough of the proportions of the discharge distributed between the main channel and the flood-plain would give a very misleading idea of the real conditions. Moreover, even when conditions of flow into and out of a section of river are being considered, as they are throughout this paper, the idealized trough must be used with care. When the river is rising or falling there will obviously be some sort of a lag between the water-level in the side-channels and the level recorded on the gauges, which are in the main channel. But this lag, as I shall show, is not very great except in the early stages of the flood, and it does not seriously invalidate the conception of the whole river va lley as a single entity. On the other hand the idealized trough has several advantages in the way it makes possible a detailed analysis of the flood. Before considering these 1 should like to note one more thing abo ut it. It is definitely an ideal, and like most ideals, it is seldom if ever realized in actual fact. The only case I know of where an actual bank profile approximated to the parabolic shape of the ideal was not a natural one, but an incidental result of some experiments recently carried out in the United States. This experiment, described as ' A Laboratory Study of the Meandering of Alluvial Rivers " was carried o ui during tbe war by the Mississippi River Commission and the results are described by J. F. Friedkin in a paper with tbis title whi~b tbe Government Geologist kindly let me see. It was conducted with a scale model consisting of a gently sloping sand-table over whicb a regularly varying supply of water was poured, simulating the conditions of an annual flood . This water was started in a straight uniform channel and the main purpose of the experiments was to study how different conditions of flow and slope afrected the meanders which tended to form. With tllese results we are not concerned here, but it was noticeable tbat when the experimenters tried to control the meanders by banking up one side of tbe channel witb a hard substance to prevent further erosion of it, the other side did tend to take up a profile similar to tbat exhibited in tbe mean idealized cross-sections of the White Nile. In this case of course the characteristics of the rela tionship between rise above low level and width of surface were achieved by only one bank, the other being vi rtually vertical; and not by a symmetrical cross-section which has o nly been used in this paper because it is easier to imagine. The bank profile created in the experiments was naturally very much steeper than that observed on the White N ile, but this is expected to occur witb models. Nevertheless, though under ideal conditions in the laboratory this profile was created it does not seem to be found in nature, where conditions a re less uniform and where the forces concerned have been working for a longer time. It must be realized tberefore that the near parabolic mean bank profile is only an idealized concept which bas the ad vantages of being much simpler and yet equivalent in some respects to the many and varying channels of the actual river. Tbe fundamental advantage of this concept as compared with the older one of a main channel flanked by a fairly fl at but irregular flood-plain is that the total width of the river at any level can be estimated easily. From this the surface area of any known length of river can be calcula ted, and by a process of integrat ion the volume of the trough above low level can also be calculated for any given level of the river. Since the surface area is known, it is possible to estimate the loss by evaporation and the gain by rainfall once the depths of water abstracted or added by tbese have been assumed or recorded. Since the surface area of the newly flooded land is also known it is possi ble to estimate the loss by absorption once the deptb of this has been assumed. But tbere is more to it than this. If the discbarges at the two ends of a trough are measured regularly throughout several floods, these depths of evapora- tion and absorption can be calculated by comparing the observed and tbeoretical differences between the end discbarges at different stages of each flood. Finally, I believe that by applying tbese estimated depths to another river flowing in similar conditions it is possible to turn the whole computation inside out and use the measured discharges on such a river to work backwards to its mean idealized cross-section, and so obtain an idea of the sbape of its valley witbout the expense of an extensive ground survey. This has been attempted in tho analysis of the Sobat flood (pp. 913-70). 854 OUTLINE OF THE METHOD OF ANALYS IS USED After these preliminary general remarks, the practical details of the theory may be de- veloped. In order to wo rk out the theoretica l losses for comparison with the observed (jifferences between the discharges at the two ends of a section of river, it is necessary first of all to assume depths for the processes of evapo ration and absorption of which the laller in particular is not at present very accurately determined . By a fortunate chance, on the Whito Nile these two sources of loss do not have their maximum effects at the same time; in fact each one's maximum coincides with the other's minimum. When the ri ver rises it wets the ground and this absorbs water until it is saturated. This happens during the rainy season and so evaporation is small ; in fact it is for a month or two more than counterbalanced by the amount of rain falli ng on the surface of the river. On the other hand, by the time the river has begun to fall the ground which it covers may be assumed with some confidence to have absorbed all it can take; rainfall has ceased, and evaporat ion is rapid ly approaching its maximum value, which it maintains almost until the ri ver has returned to its lowest level. Once the river has reached its highest or lowest level and remained there more or less stat ionary for about a month, it may be assumed that the delays in fi ll ing or emptying the minor channels and lagoons alongside the main channel wi ll have been taken up, and thai the volume contained in the actual river valley will therefore approximate very closely to that held in the idea}jzed trough. During the periods between these two stages this will not be so, and since the amount of delay and the differences between the idealized and actual trough volumes at any time must be uncertain, the calculation of depths for absorption and evaporation must be done for the whole of the rising and falling stages respectively. While the river is rising the discharges at the upper end of a given stretch will obviously be appreciably larger than those at the lower end. Normally these discharges are given on the White Nile at Malakal and Renk in the form of ten-day means, the last period of each month being sometimes eleven days or (in February) eight days. If the mean discharge for each period is multiplied by the number of days in it, the total discharge for that period is obtained. By summing cumulatively the difl'erences between these totals at the two ends of the reach from the time when the river begins to rise until it begins to faU , the cumulative sum of the • losses' during this stage is obtained. I have called these' losses' though one of the items in the account-rainfall-is actually a gain, and another-the trough volume-is only a temporary loss because it is a form of storage. But, if this convention is followed, the equation below is immediately derived for the end of the rising stage of the flood: CumulatIve Sum of Differences of Observed 11 Maxunum Trough Volume minus Trough Volume at Start of Rise Discharges plus Total Absorption Loss (upper rnmus lower) - plus Total Evaporation Loss or CUMULATIVE OBSERVED LOSS minus Ramfall Gam When the river is falling, as already explained, rainfall and absorption have virtually ceased, and the discharges at the lower end of the reach will be greater than those at the upper end because of the water being returned to the main channel from the side lagoons; in other words from the emptying of the trough. When the river has returned to its minimum level and remained there for a month, the following equation can therefore be derived: Cumulative Sum of Differences of Observed} {MaXimUm Trough Volume minus Trough Volume Discharges at End of Fall (lower minus upper) = minus or CUMULATIVE OBSERVED GAIN Totat Evaporation Loss during this stage These equations have already been applied- in sum-in the analyses of the White Nile and Sobat floods carried out in the First Interim Report by H. A. W. Morrice and in Sobat Hydraulics by A. D. Butcher. It is only when the total volume of the river tTOugh- not merely that of the main channel as used by Butcher-is estimated that it becomes possible to separate out the two stages of rise and fall and have some chance of estimating the size of the individual terms in the separate equations. Even when they have been separated out by doing this, it is still impossible to get reliableand consistent results in all years by considering each stage as a whole, and it is of no value therefore to take a mean of several years. The reason for this is that there is yet another term in the equations for which allowance has to be made, and it is one which cannot be estimated directly at all. This term is what I have called inflow, and its effects m.us t now be discussed . ' 855 INFLOW Inflow may be defined as water fl owing into the river from a definitely outside source, that is to say it is quite distinct from water which may have been' spilt' from the river originally and which I consider, as explained above, to be still held in the trough. As was pointed out in Volume I of The Nile Basill. there is definitely inflow into tllis part of the Nile in some years, although there are practica lly speak ing no measurements of its amount. Probably because there were no reliable data about it it has tended to be forgotten in more recent work, such as the Interim Reports of the Jonglei Investigation Team, and to be regarded as negligible ; but in fact, as I sha ll show, with a value in some years of over a rnilliard, it is far from being so. Obviously, unless it can be allowed for in some way, it will completely invalidate the equations given above. I t can on ly be allowed for by dist inguishing the years in which it occurred, or in which it was present fo r part of the flood. In my opinion the only way in which this can be done is by making a preliminary study of each flood separately in great detail, comparing the observed and computed or theoretical losses all the way through the flood , and not just at the ends of the rising and fa lling stages in the way ou tlined above. It is this detai led a nalysis which makes the present paper so bu lky, especia lly as regards the separate fl oods detailed in Tables 404-18 (pp. 885.9(0), and it has meant a great deal of work; but I do not see how this could have been avo ided or the resulls acbieved in any o ther shorter way. In order to get this close and con- tinuous comparison of the observed and computed losses throughout the length of each flood it was necessary to calcu late tile values of each of the five terms of the equations-cumulative observed loss, trough volume, ra infall. evapo ration, and absorption- for each of the ten-day periods of each Oood. These formed the natural unit of time to use because it is for each of them that the mea n discharges and the ga uge- readings are given. It is clear that the equations will not hold true during these stages in the way that they will at their ends, when the lags have been taken up ; but nevert heless 1 believe that it is possible by comparing the differences between observed and com puted loss in one yea r wi th those in another to detect when inflow has occurred and to estimate its amount. COMMENTS ON THE ANALYSIS IN THE FIRST INTERIM REPORT Before I realized the necessity for individual treatment of each year in this way I tried to work from means, usi ng the cumulat ive discharge diAerences and gauge-readings given in Table 16 of the First Interim Report , from which the first estimate of the flooded areas was calculated in the way T have a lready described in the introduction. It was very difficul t to get any correlation between these which made sense, a nd for a long time I could not discover why, and I began to think that tile whole theory of an idealized trough must have some fundamental fault. Later, however, I discovered why the gauge-readings and observed discharge clifferences did not agree, the reaso n being that this table had been constructed from means which were not strictly comparable. Discharges have been measured regularly at Malakal since 1912, but measurements started at Renk only in 1928. The means used in Table 16 of the First IllIerim Report were taken from the third supplement to Volume IV of The Nile Basin ; those at Malakal being the averages for tbe years 1912-42, whereas those at Renk were of course averages for the years 1928-42 only. During this period the means had changed by about 2 per cent, which is less than the average error of a single discharge, and would in most hydro- logical work be regarded as unimportant. But in this particular investigation the very quantities being studied- the differences between the two sets of discharges-are themselves not much larger than this, and so a systematic cha nge even of this small size in the means completely upsets the data. As is pointed out below, although the error of a single discharge may amount to a few per cent, the systematic errors in them are very much less than this, and over several years amount to only a fract ion of one per cent. Thus the results obtained by comparing these means were actually quite erroneous, and it was only a series of accidents which made the resulting estimate of tile maximum flooded area turn ou t to be very near that obtained by more direct means in the Third Interim Report. In this earlier analysis an average value of 540 millions was found for the tota! loss during the year (February to February); but if the average clischarges for 1928-42 are compared it will be found that the true average loss per year is very mucb less, in fact it is about 200 millions. (See Table 419, p. 901.) This diJrerence is more tban accounted for by an average value of 700 millions for inflow; and the true average losses by evaporation and absorption, less rainfall, I estimate to be nearly a milliard, as may also be seen from this table. The assumed depths on which the earlier analysis was based are adrnitted to have been very approximate, and these differences in the items of the water account from tbe figu res obtained by a later more detailed estimate were only to be expected. The curious fact is that by a series of coincidences this 856 earlier analysis should have arrived at practically the same value for the average maximum flood- plain area as I obtained last year by direct arithmetical averaging from the cross-sections. It can be seen from the above that the average value of the inflow makes it an item of some importance, but whereas the other items can be correlated with the height of the flood , and so .their average values can be estimated from the average height, this is not true of inflow. So far it seems to have very little correlation with any obvious feature of the flood, and it is for this reason that each year has to be studied separately, sinco the average value of the inflow is no indication of the amount to be expected in a year like 1944-45, which in other respects may be regarded as average. THE INDTVlDUAL ITEMS OF THE WATER ACCOUNT In the next five sections each of the five measurable items of the water account is discussed in detail, and the way in which their values are calculated is described. In Section VII (p. 875) the analyses of sixteen years are given, the last one, for 194~7, being given ill complete detail in Table 403 (pp. 877-83) and the others in summary form in Tables 404-18 (pp. 885-900). In Table 403 the steps of the calculation are all shown and fully explained. In Tables 404-18 ooIy the essential data and the different items of the account are shown. The intermediate steps can be worked out by anyone who wishes to do so, using the data provided and the auxiliary tables of surface areas and volumes (Tables 397 and 398, pp. 865-6) and following the procedure described and illustrated in Table 403 . The five items of the water account are placed ill the order of directness of tbeir derivation, which is not necessarily the order of accuracy. Thus the observed losses come first because they are obtained directly from the observed discharges by making the necessary aflowance for the number of days in each ten-day period. Next come the trough volume and the rainfall gain, which are both obtained by direct calculation from measurements. The trough volumes are obtained from the direct means of the measured cross-sections, and so are the surface areas; and the depth of rainfall is deduced directly from the average of tbe avai.lable records. It is true that there is an element of uncertainty in this last factor, due to the variable nature of the rainfall and the small number of stations; nevertheless the calculation is a direct one and requires no estimated factor. This is not true of evaporation, since the relation between instrumental and actual evaporation is still uncertain, and I bave therefore thought it worth while to make an independent estimate of this factor. As it turns out the value which I get is very near that which has been assumed previously and I do not therefore make use of it. For absorption, on the otber hand, practically no reliable previous estimates seem to have been made, and I therefore have first to make an estimate of its depth before the losses which it causes can be calculated at all. Of these five items this is therefore clearly the one which is obtained most indirectly, and for that reason it is described last and shown last before the total computed losses in the individual analyses. In calculating trough volumes and surface areas I assume all the time tbat tbe water surface across any given cross-section is horizontal. Obviously this is not true, for when the river is rising it will be higher in the main channel than at the sides, and when it is falling the level in the main channel will be lower. Nevertheless, until observations have been made of the amount of this difference of level-as they have in the Sudd-tbere is little point in trying to guess what it is. I feel that this source of uncertainty is better omitted from all the items which can other- wise be calculated directly, and for trough volume, rainfall, and evaporation the surface areas and volumes used are these theoretical ones. The effects of this uncertain difference between the theoretical and actual shapes of the transverse water profile have therefore been included in an item which I call apparent or mean absorption. This is explained more fully in Section VI (p. 872). I should like to make it quite clear, though, that I consider that I have allowed for the difference between the longitudinal profiles in the rising and falling stages, by breaking the stretch hetween Malakal and Reok into two parts separated by Melut, and obtaining for each of these at all times a mean gauge-reading derived from the gauges at its ends. There seems at times to be some confusion of thought about the effects of tbe reseFVoir. It does not ever cause backward flow at this point and I cannot see that it can invalidate measured discharges at Reok. What it does do of course is to lessen the water slope and so slow down the rate of discharge; and it also invalidates the normal gauge/ discharge relationship. It is because of this last effect in particular that I am unwilling to analyse years in which the mean discharges at Reok or Melut are based ooIy on gauge/ discharge curves supported by few-if any- actual discharges. But when the analyses are based on measured discharges compared with measured levels of tbe water surface I am sure that the backwater effect is fully catered for, even when, as in later years, it is considerable. This may be checked by studying Tables 404-18 and Table 419 carefully. 857 . Tbe last item of tbe water account is inflow, and this is entirely incalculable except as a residual after all the others have been derived. I therefore feel tbat it should be left until after the analyses, since it contributes nothing to tbeir calculation and is in fact obtained from tbem. There are virtually no measurements of this item-since the few recorded discharges of tributaries are not complete-but there is a certain amount of supporting evidence of a des- criptive kind, which I have summarized after the analyses in Section VIII (p. 902). We may now proceed to di scuss the individual items in detail , beginning with the observed losses. SECTION II. CUMULATIVE OBSERVED LOSS PERIOD COVERED BY THE OBSERVATIONS The cumulative observed loss in a given period is the algebraic sum of the differences between the di scharges at Malakal and Renk during the period . Discharges at Malakal have been measured regularly since 19 12, but at Renk they only began in 1928 and there have been two quite long gaps since then. The first was from June 1930 to November 1932, and the second from March 1933 to June 1936. Discbarges were tben measured regularly at Renk until August 1947, when the discharge site was moved to Melut, wbere it has remained ever since. Thus for the noods of 193 1-32. 1933-34, 1934-35, and 1935-36 there were no discharge measu rements at all a t Renk; and fo r the fl oods of 1930-3 1, 1932-33, and 1936--37 the measure- ments were not complete, though in the last flood only those of the first month and a half are missing. In the Supplements of Vol ume IV of The Nile Basin mean discharges are shown for the whole period from 1928 to 1942 . In the years when no discharges at all were measured the ten-day meanS were obta ined from a general gauge/ discharge curve ; in the years when only some measurement. were made the means were obtained from a gauge/ discharge curve fitted to these measurements. Means obtained from a general curve, as in the first case, and especially over such a long gap as the second mentioned above, seem to me too unreliable for the delicate balance required in estimating the losses, and I have therefore not thought it worth while to analyse the fl oods between 1933 and 1936. 1 have analysed the fl oods of 1930-3 1 and 1931-32, though here also only a few measure- ments were made, hecause it was a shorter gap, and a lso because the floods were smaller so that the errors in the estimated discharges are probably less. I have not, however, used the figures obtained from these years in estimating the evaporation and abso rption depths. Movement of the discharge site to Meilit in 1947 a lters a ll the items of the analysis and tbis makes com- parison of the last two floods with the earlier ones impossible. Moreover discharge measu re- ments a t Meilit were far from regular and there were gaps of several months during this period. I have therefo re not included here any analysis of the last two floods (1947-48 and 1948-49), though I have made a preliminary analysis of the last one which is quoted below. A proper analysis of these two floods might be worth doing when the mean discharges at Melut have been published by the Di rectorate of Nile Control. Thus the discharges at the lower end of the reach are the limiting factor in deciding what years should be analysed. and I have confined myself to those between 1928 and 1933, and between 1936 and 1947. giving a total of sixteen years to be analysed in the reach between Malakal and Renk. In each analysis I have worked not by calendar years but by floods, each flood being taken from when the river started to rise to when it stopped falling, which means normally from the beginning of May one year to the end of the following April. This practice makes it much easier to eliminate from any consideration of the whole flood the errors in the estimated trough volume, which is nearly a t a maximum at the end of the calendar year. It a lso eliminates from the final totals of each flood any uncertainty about the delays in filling or emptying the side-channels and basins, which make ftxact correpondence between ideal and actual conditions impossible, especially when the river is changing its level as fast as it often is at the end of the calendar year. METHOD OF COMPUTATION The total discharge for each ten-day period in the following analyses has been computed hy multiplying the mean for the period by the number of days in it. The means were either taken from the Supplements to Volume IV of The Nile Basin or were kindly supplied by the Egyptian Irrigation Department in Khartoum, to whom my grateful thanks are due. The differences between these ten-day totals for MalakaI and Renk were derived and then summed cumulatively from the end of the ten-day period chosen as the start of each flood. The results 858 are the cumulative observed losses. The final totals at the end of each flood are given in Table 396 (po 86 1). In this table are also given the initial and maximum Malakal gauge- readings for each flood . the ten-day periods in which the rise began and ended and after which the fall began; and also the averages of the rainfall totals at Malakal. Melut. a nd Renk. This table shows therefore the data which were ava ilable when this analysis was started , and as such represents a preliminary summary of the floods. METHODS OF MEASUREMENT OF THE DlSCHARGES Details of the methods of measuring the discharges are given in the introduction to Volume II of The Nile Basin , which also includes a discussion of their accuracy. It may be noted here that the probable error of an individual measurement is now assessed at about 3 per cent. ,Il) . and that the systematic error is thought to be very much less. Every precaution is taken to prevent systematic error: the current meters are used only for a limited number of measure- ments before being replaced, the original meters being sent to Egypt for checking. At Malakal and Renk tie banks have been built to constrict the river so that all the water has to pass through the cross-section used for the discharge measurements. So far this has not been done at Melut. It might be of interest to cut a channel there at right angles to the river across the flood-plain and take measurements across it at the height of the flood , since at present there are no figures for the proportion of the total flow which passes over this part of the river bed. There is little doubt that the proportion is small but it might be of value to know what it is. PROBABLE ERRORS OF OBSERVED LOSSES (ASSUMING NO SYSTEMATIC ERRORS) The frequency of measurement and the probable errors of the observed losses may now be considered. At Malakal the discharges are measured fairly regularly every five days, or twice in each ten-day period. At, Renk the frequency was more variable, but it was never less than this, except for the occasional month which can be ignored. On the other band over quite long periodS the discharges at Reok were measured nearly everyotherdaY,and theassump- tion of the same frequency as at Malakal will not therefore give an over-optimistic idea of the accuracy of the means at Renk. The mean of two measurements each having a probable error of 3 per cent. may be taken as having a p.e. of 2 per cent. The probable erro r of the difference between two such means, provided they 'are approximately equal, may be taken as roughly equal to 3 per cent. of their mean. Thus the probable error of the observed loss during a single ten-day period may be taken as 3 per cent. oftbe mean of the total discharges at Malakal and Reok during that period. It should be noted that it is a function of the size of these dis- charges and not of the difference between them. At low river these discharges average about 40 mId, or 400 during a ten-day period; at high river they average about 1,000 mj d during a ten-day period. Thus the average probable errors of the observed losses in each ten-day period at these stages are 12 and 30 millions respec- tively, giving for both the rising and falling stages an approximate mean of 20 millions per ten-day period. The rise lasts as a rule for seven months (210 days) and the fall for about five months (150 days), so that the probable errors of the cumulative observed discharges during these periods (from which absorption and evaporation depths will be calculated) are roughly 90 and 75 millions respectively, assuming that there is no systematic effect. The probable error of the final cumulative observed loss in an average flood is about 120 millions, if the same assumption is made. Clearly this assumption will not hold in years when no discharges were measured, and no estimate of the probable errors of the observed discharges in these years can be made. It is almost certain tluit they will be greater than in years of regular discharge measurements; but how much greater I should not like to say. THE POSSIBILITY AND EFFECfS OF SYSTEMATIC ERRORS Mter this paper was completed in draft it was suggested to me that insufficient account had been taken of the possibility of systematic errors in the observed discharges. The following note has therefore been added and in it I try to assess these and outline what effects they would have on the analysis. I have preferred to deal with this question as a whole here rather than in parts under the various sections concerned. It may be found clearer, however, when these have been read, and so reference to this note is made in them where it is relevant. There were said to be two principal ways in which systematic errors could arise. First the softness of the bottom could cause a systematic difference in the soundings taken by different engineers and so affect the area of the measured cross-section. Secondly the factor used in 859 converting the velocity measured at half depth to the mean velocity of the corresponding part of the cross-section might be wrong. This factor is the mean of many years' observations and in the long run is probably very accurate; but changes in the conditions of the river bed in individual years may cause deviations from its mean value during any particular flood. Opinions as to the possible size of these errors differ, but they may in theory amount to one or two per cent. with corresponding effects on the accuracy of the discharges.(2) There is, however, one piece of evidence which indicates that they may not be as large as this. I have outlined on p. 855 above and consider in more detail in Section V (p. 869) how in the falling stage the average evaporation depth can be obtained by comparing the difference between the observed discharges and the volume of water released by the emptying of the trough. Varying values for this depth are obtained in different years, but it is noteworthy that none of them exceed by very much the average evaporation depth obtained from instrumental records and adopted by the Egyptian Irrigation Department. The absence of unusually largo values for the evaporation depth calculated by this means is to me an indication of the compara- tive accuracy of the observed discharges when averaged over this falling stage, unless there have been errors which in all cases have been counterbalanced by inflow. The effects of such errors on my results may now be considered. They are twofold. In the first place if larger errors are allowed for in the observed discharges then the valueS' of evaporation and absorption depth obtained in Sections V and VI, by the methods indicated on pages 870 and 872, may be allowed a wider range. Since all the highest values are already included, this entails adding only smaller values and so decreasing the means obtained from them all. This is of little consequence in the case of evaporation because I do not use the means obtained in this way but rely on estimates observed from instrumental records ; but in the case of absorption it would entail a decrease in the assumed depth. This is supported by other evidence and it seems likely in fact that the value of 80 cm. which I have obtained is actually too large and that 50 cm. may be a more reasonable figure. The second effect of allowing for systematic errors in the observed discharges is to increase the probable errors of the deduced inflow. It will be observed that in some years, notably 1932-33, 1936-37, 1939-40, and 1943-44, this has a very steady value, as may be seen from the curves of Figs. J 5, 6, 7, 8, 9, which in these years have a steady rather than a sudden separation. It may well be that in these years this systematic decrease in the observed losses is not due to inflow counterbalancing the true losses by evaporation and absorption, but simply to systematic errors in the observed discharges causing these at Renk to appear relatively larger than they reallyare. It has also been argued that there should be some correlation between the height of the White Nile and inflow into it, and that the poorness of this correlation with my figures for deduced inflow renders them very much open to question. I do not agree with this view, for the following reasons. In the first place there is unequivocal evidence of inflow in 1942-3 on the west bank (see Section VlII, p. 904) and yet in this year the maximum gauge-reading at Malakal was 12·38-only 12 cm. above the average. Secondly the arguments for this correlation seem to me much weaker than those in favour of correlation with the height of the Baro and Sobat which are explained in Section IX (p. 906) ; and in this case a fair degree of correlation does exist as is shown there. The maximum height of the White Nile in each flood is due to two factors: the height at which it begins the cycle, and the volume of water added to it by the Sobat. In other words a considerable part of its maximum height is due to the flow from the Great Lakes for a year or more beforehand and there is no reason why conditions there should be connected with inflow below Malakal, over six hundred miles farther north. A third reason which inclines me to stand by my figures for deduced inflow, in spite of the possibility of their being invalidated by systematic discharge errors, is the correspondence between my maximum figures and the supporting evidence obtained from the few records available. I think it is quite possible that my totals for deduced inflow in each year may be wrong by 100% in years when it is small, and by 30 or 50% in years when it is large. Neverthe- less it is worth noting that in three out of the four floods when I deduce maximum totals of inflow there is unequivocal evidence (given in Section VIII, p. 902) that it was larger than usual; for the fourth year- 1929-30-no evidence either way is available since practically no records can be found . Moreover in the two months of 1947 when my deduced totals are 'very much larger than in any other month during the whole period considered, there is again plenty of evidence that inflow in the Paloich area was without precedent during the previous twenty years. Thus, though the possihility of systematic errors in the discharges cannot be excluded, there is a fair amount of evidence that they are not large enough to invalidate seriously the results which I have obtained by ignoring them. . 860 TABLE 396 PRELIMINARY SUMM ARY O F F LOODS ON TH E WHITE NILE BETWEEN MALAKA L AN D REN K 1928-41 Year in which IC u~i~~~ive : Initial I Max imum I First P.:riod I Last Period I Last Period I Average Flood began I Observed I Malakal M;'I lnka l Rise I Rise or Pe.1k T~la l : Loss ~ Gauge Gauge , I Raln r,, 11 I i _ ..!! ----19-28----'---6;;---i---;1; ----;;3~ T ~~~I ) T~:'~ ;'3) -' -D-ec-.-(l;---;~-- :::: - ~~: /' :::: :~ ~: : ::: :~: _Ii ::.~ . :~; ~::. :~: ::~ 1931 (281) 9·66 12-26 May ( I) Nov. ( I) Nov. (3) 690 1932 845 I 9·98 12-92 May ( I ) Dec. ( I ) Dec. (3) 513 1933 (20) 10·29 12·84 May (I) 0.:1. ( I) Nov. (I ) 645 1934 (- 190) 10·11 12·56 Ap'. ( I) Nov. (I) , Dec. (2) 566 1935 ( - 490) 10' 19 12·36 Ma y ( I) OCI. (3) Dec. ( I) 559 1936 551 10-12 12·1 7 M.y ( I) ()cl. (l) Nov. (3) 614 1931 - I I I 9·87 12-44 M ay ( I) Oc •. (3) N ov. (2) 622 1938 - 201 10-09 12·63 Ap,. (2) Nov. ( I ) Ja n. (I) 601 1939 100 10·39 12·26 M.), ( I ) Oct. (3) Nov. (3) 692 1940 85 10·03 11 ·94 May (2) Ocl. ( I) Nov. (2) 520 1941 - 140 9·11 12· 19 May (I ) Nov. (3) Dec. (2) 6ll 1942 - 298 9·83 12-38 Ap'. (3) Oct. (1) Nov. (2) 604 1943 10·04 12·17 Ap'. ( l ) Nov. ( I) Nov. (3) .5 14 1944 471 10·00 11-24 Ap,- (3) Oct. (2) Nov. (2) 64 1 1945 805 9-44 12·37 Ap,- (3) Oc• . (3) Nov. (l) 865 1946 - 52 9·52 12·92 Ap'. (3) Nov. (2) Jan . (2) 868 Nons: LosK! in aull ions, saufCS i.e nlettes, fa!nr. U in millimetreJ. Obsc""cd I05XS ill brack-tis are derived (rom , cocnl pUle/di$cha~c curves onl)' Do t Renk. Fluctuations or under) em. lit tho peak or the flood h..:Ivo been i&norc4 In calculal ina ill duration. SECTION III. TROUGH VOLUME As the river rises it fills the trough, and the trough extends from the bottom of the river to the highest flood limits without a break. This is the burden or my thesis, and the change of trough volume as the river rises and faUs is the biggest factor in causing the differences which are observed hetween the ctiscbarges at Malakal and Renk during these stages. In calculating the theoretical differences this item in the account has two advantages: it can be calculated entirely from existing measurements without any assumptions having to be made or factors to be estimated, so long as no attempt is made to allow for lag; and it is cancelled out at the end of each flood when the river bas returned virtually to the level it occupied at the beginning. This means that an estimate or the total loss during anyone flood is not affected by any errors in the estimated trough volume. PREVIOUS ESTIMATES I have incticated in the introduction that I believe my approach to this question to be rather different from that used previously, and it may therefore be worth outlining what this previous approach was, and what I understand to be the current methods of working out the trough volume of this part of the river. It may be noted that previous estimates have not been con- IilCmed with the river above Melut (so far as I know) because they have been made in connection 861 with the Jebel Au li a Reservoir whose effects are not felt above Melut. In Appendix IV of the Tilird J"leri", Reporll gave an acco unt of the tables, prepared originally by Topouzada Bey and revised by Dr. Amin, wh ich a re used in the wo rk ing of the reservoir, and which the Egyptian Irri gation Dera rtment in Khartoum kindly let me see. The essence of this account will be repea ted hcre for convenience. In these tables the river between Melut and Jebel Autia is divided into 22 stretches of between 20 and 30 km. long, with a cross-section in the middle of each. The first part of the tables gives the widt hs of the wa ter surface at each of these cross- sections correspondi ng to different reduced levels at it. That is to say the widths are given in terms of heights above sea level and not above the surface of the river; and so the different cross-sections are related to a series of horizontal planes instead of to a series of planes parallel to the low-level slope of the ri ver. This system of reference makes impossible any comparison between successive cross-sections and so has prevented in the past any realization of the com- parati ve unifor mity of the cross-scction or assessment of the probable errors of its mean width at va ri ous levels of the river. Valtles for the surface areas and volumes of the different stretches were compiled in these Egyptian lables by assuming Ihat Ihe width of the surface at each cross-section was the mean for the width of th e stretch in which it lay. As I have shown this gives on the average a less accurate va lue for each stretch than one derived from a large number of cross-sections taken over a much longcr reach of thc ri ver. so long as this has reasonably uniform cha racteristics. Clearly, when such a long reach is being co nsidered , the to tal surface area or volu me will come out the same whcther it is computed by summin g the val ues obtained for the separate stretches of which itis composed. or whet her it is done by taking a mean of the widths at the cross-sections and then multiplying this hy the lengt h of the reach. But it wi ll also be clear that when a large number of volumes co rrespondi ng to many differen t states of the river are required the computations wi ll be shortened considerably by taking the mean out first. In estimating the "o tume of a long reach the first step must be to ass ume or predict a longi- tudinal profile for the water surface along it. and th is is expressed .;n the first instance as a series of gaugc-readi ngs at vari ous po ints along the reach. From these, in the Egyptian tables, the red uced level of the water surface at each of the cross-sections is interpolated, and the corresponding widt h or the section in which it lies is then deduced as described above. In my method , descrihed below. the tables of vo lumes have already been prepared in terms of mean ga uge-readi ngs for each of the st retches into which the gauges divide the river. Thus the volume of the whole reach ("an be taken ou t direct ly once these ga uge-readings have been assumed. and thc volume of any shorter stretch ca n be obta ined by reduction in the pro- portion of its length to that of the whole reach. In other words the whole reach between successive gauges is rega rded as un.ifonn and a ny shorter length required is chopped off like a sausage. T he calcul at ion of the trough volumes in any given conditions, even over so short a reach as null between Melu t a nd Renk , must be a fairly lengthy calculation by the previous method; but it ca n be taken out frolll Table 398 (i), p. 866, in a few seconds once the gauge- read.ings at Melut a nd Renk have bee n. assumed. METHOD OF CALCULATION T he data for calculating the trough volumes which I have used were Ule 22 cross-sections surveyed between Ma laka l and Jebelein by the Egyptian Irrigation Department between 1928 and 1943. (These shou ld not be co nfused with the 22 sections mentioned above into which the ri ver between Melu t and Jebel Au lia was divided quite arbitrarily in the Egyptian tables.) Following the principles outlined above r made use of separate means for the idealized bank profi les between Malakal and Melut and between Melut and Renk . Moreover in calculating the la ller r did not omit from the mea n the cross-sections below Renk becau\e, as I have shown, there is no significa nt change in the cross-section at that point, and the more cross-sections used the better, even if they lie outside the stretch for which the mean is required . To obtain the total width of the water surface at various gauges it is necessary to add to the fl ood-plain widths the mean width of the low-level channel. At mean gauges of 10·0 this is 400 m. between Malakal and Melut and 560 m. between Melut and Jebelein. I obtained the mean widths in this way because it made it possible to use the tables already compiled and pubtished in the Third Inlerim Reporl ; the means of the low-level widths were obtained from the individual values given in Table 24 and the mean flood-plain widths are taken directly from Table 26, with rises turned into gauge-readings by adding 10·0 to them. The lengths of tile two reaches Malakal to Melut and Melut to Renk are 141 and 182 km. respectively; if the total mean widths are multiplied by these the corresponding total surface areas are immediately obtained. It is thus possible to construct a table showing surfa~ 862 areas in terms of mean gauge-readings, and this has been done in Table 397 (p. 865). This table is used for calculating the volumes of wa ter gai ned by rainfa ll and lost by evaporation. as is described in more detai l in the next two sect ions. But it ca n also be used fo r ca lculati ng the volume of water held in tbe trough at any particular gauge-read ing. For the present pur- 'poses we are concerned. not with the tota l volume of the trough. but with the extra volume caused by a rise above mean low level. If this is taken as correspondi ng to a mea n ga uge-reading of 9'80, the trough volume held a t a ny particular height above it can be obtained by a process of integration. This is done in Table 398 (p. 865) which was calcu lated by add ing successively the increments of volume corresponding to each decimetre rise. each increment being obtained by mUltiplying the me.1n surface area (taken from Table 397) during the rise by 0·1. Between the values for the separate decimetres direct interpolat ion could bc used since the relationship between gauge and volume may be rega rd ed as linear over such a short range. COMPUTATION OF MEAN GAUGES (see Table 403 (i)) i In order to make use of Table 398 it is necessary first of all to obtain the mean gauge- readings for each of the two reaches for each of the ten-day periods. The mea ns for these a t Malakal. Melut, and Renk were either obtained directly from Volume 1Il of The Nile Basill and its Supplements or, for the years after 1942. from figures kindly supplied by the Egyptian Irrigation Department in Khartoum. The values at Kodok were not used because the profile be- tween Malakal and Melut appeared to be substantially straight at a ll times ; the val ues at Meteimer were also omitted because they were only available for part of the time. The mean gauge-readings for the two reaches were therefore obtained by a direct mean of those at Malakal and Melut, and at Melut and Renk respectively. In actual fact there is not quite the same correspondence between the rise at Renk and its gauge-reading as there is at Malakal and Melut, since the average low-level reading at Renk ga uge is usually between one and two decimetres above those at the other two places. But the amount is small and uncertain, and to allow for it would have complicated the oalculations, so I have not done so. There is, however, one more step which has to be taken in o rder to achieve d irect corres- pondence between the theoretical and the observed losses. This is to convert from the mean gauge-readings as derived above, which correspond to the middle of the ten-day period, to the mean gauge-readings corresponding to the end of the period, the moment for which each cumulative observed loss is given. When the river is ri sing or falling steeply this difference may amount to several decimetres over the whole length of the reach and cannot therefore be ignored. It may be argued that exact correspondence cannot exist betwe¢n the theoretical and actual trough volumes, and this is of course true owing to the complex nature of the actual trough ; nevertheless there is little point in confusing the analysis by introducing a large varying difference into the theoretical volume when a little extra computation will remove it. It is necessary therefore first of all to mean the individual gauge-readings, and tben to take out the mean of each ten-day period with that of the period immediately following it. When this has been done we are in a position to derive the theoretical ' loss' to the trough volume at the end of each ten-day period. We take first of all the mean gauge-reading at the end of the last ten-day period before the river begins to rise. This is our datum for the year, and we get from Table 398 the corresponding volumes held above (or below) tbe mean gauge-reading of 9·80 in each of the two reaches. This value is recorded, but of course the actual loss to the trough at this moment is nil. The trough volumes for all subsequent ten-day periods in this year are taken from Table 398 and corrected by subtraction of these datum values, which are of course given their correct signs. Thus we obtain for the end of each ten-day period the theoretical value of the.trough volume held above the river surface at tbe start of the year, and this largest item in the computed losses is in direct correspondence with the observed losses. The value of 9·80 was chosen for the datum of the table because it gave the smallest initial value (either positive or negative) on tbe whole for the twenty years under consideration here and so made easier the correction of tbe volumes obtained from tbe table in order to convert them to • losses '. When this invest igation was first started I calculated the cumulative loss to the trough volume by working out separately tbe cbange of volume for each ten-day period, and then summing fuese increments cumulatively. The advantages of tabulation over this method are obvious: the increments are cbosen so as to be easily calculable and, once the basic table has been checked, the individual trough volumes taken from it are independent of each other. When they are obtained by cumulative summing of calculated increments any error in one of these or in the addition will affect all the cumulative sums which follow it. The years 1928-29, 1929-30, and 1930-31 had been worked out and checked by calculation of individual increments 863 before I had fully realized the advantages of tabulation ; I have not worked them out again and small differences may therefore be found in the trough volumes which would be obtained for these years if they were taken from Table 398. PROBABLE ERRORS Since the trough volume, even when it is taken from the table, is obtained by summing a succession of increments, its theoretical probable error at any level is obtained easily once we know the probable errors of the increments. Tltis is done quite simply. Each increment may be thought of as a disc whose thickness equals the change of mean gauge-reading to which it corresponds. Its length is the length of the reach and its width the mean width of the reach corresponding to the height above low level at which it lies. Only this last factor has any appreciable error and we may therefore derive the probable error of the volume of the increment directly from it. In Table 26 of the Third Interim Report are shown the probable errors of the mean widths at various heights above low river level expressed as percentages of the widths of the flood-plain. Except at very low levels the va riations in the width of the low-level channel are a small proportion of the flood-plain variations and will not seriously affect the probable errors of the means; in fact the main effect of adding the low-level width is to reduce the percentage error. We may therefore for practical purposes take these percentages as giving ·the corresponding percentage errors of the volumes of the increments. In Table 399 (p. 867) the calculation from these and from Table 398 of the probable errors of the trough volumes for each of the two reaches is shown, for each even decimetre of the gauge-readings up to the maximum values recorded during the years under consideration in this paper. It will be seen that these probable errors are not large; in fact they are under ten per cent. for the upper reach and under five per cent. for the lower one, except at very low gauges. In a normal year, when .the mean gauges in the two reaches attain maxima of 12·2 and 12·0 respectively the probable error of the whole trough vol ume between Malakal and Renk will be: V14' + 8' = 16. • For a high year such as 1946, when the mean gauges reached 12·8 and 12·6 respectively, the probable error of the total was: V2O' + 13' = 24. If therefore we assume that the probable errors of the totals in average and high years are 25 and 40 millions respectively it should not be an underestimate, even allowing for any systematic tendency there may be in the errors of the mean at different levels. TRIBUTARY KHORS It will be convenient at this point to deal with the subject of storage in the beds of the tributary khors, and also of their flood-pl ain areas. There are about a dozen of these khors on each side of the river between Malakal and Renk. From a study of the air photographs it is clear that few of them have a width, evell when full, of more than two or three hundred metres. The opportunity of a surveyors visit to Kodok in 1952 was taken to obtain a cross-section of the Khor Nyadwai, which enters there and is one of the biggest tributaries on the west bank in tills reach. As will appear below it can on occasion provide a substantial portion of the infl ow in some years. Nevertheless its capacity for storage is very small. The cross-section is triangular, with both banks sloping evenly upwards at an angle of about one in forty from a very narrow bed. The bed is at a level corresponding to Kodok gauge 11 ·4 so that no water would enter the khor until the gauges at Malakal and Melut reached this value. In an average year (maximum Malakal gauge 12'3) the khor wo uld have under a metre of water in it even at the mouth. Its surface width there would be about 80 m. and its cross-sectional area about 40 sq. m. It is reported locally that in an average year the water reaches inland to a distance of about 10 km., so the stored volume would be only about a trurd of a million or less. It will be clear therefore that even twenty of these khors are very unlikely to hold much more than 10 millions in an average year. Tills may be compared with the average maximum trough volume (as shown in Table419, p. 901) of about 700 millions, of wlllch it is only therefore just over one per cent. The Khor Adar, which goes back very much farther and obviously has a flatter slope-though not a much bigger cross-section-may hold a few millioruo when full ; but even if it is included it will be clear that the storage in the tributary khors and the flooded areas wlllch they provide for grazing must be almost negligible when compared with the same features on the main river along the whole stretch. The grazing 'areas on the tributary khors may assume a slightly greater degree of importance than is warranted by their size because of their relative accessibili ty; but even so it does not seem that they can be regarded as a serious proportion of the whole. 864 TABLE 397 SURFACE AREAS BETWEEN MALAKAL AND MELUT AND BETWEEN MBLUT AND RENK, FOR VARIOUS MEAN GAUGE VALUES (i) MALAKAL TO MELUT (ii) MELUT TO RENK I I M ALAKAL-M.L\JT !-CoRREcrEl.> V ALUes AFTER ALLOW1NO fOR Mean Gauge ~ Mean Widlh I,S ur~:a~rea I E~:~_~d:~N4 iI ' Mean Gauge Mean Widt h i Total ~ Width ..l Mea IS u,faco A,,,,, ---~-:g--'- ~~ -1~l1 I! ~~ ~i--'----~f--II-- "i~g -i--l~ 10-0 400 56 410 58 10-0 560 102 10-1 410 , 58 430 61 10-1 580 105 10-2 420 59 440 62 10-2 605 110 10-3 460 65 480 68 10-3 670 122 10-4 500 70 530 75 10-4 730 133 10-5 520 73 550 78 10-5 785 143 10-6 540 76 570 80 10-6 840 153 10-7 570 80 600 85 10-7 880 160 10-8 600 84 6)0 89 10-8 920 167 10-9 630 89 670 94 10-9 970 177 11-0 660 93 700 99 11 -0 1,020 185 11-1 700 99 740 104 II-I 1,090 198 11-2 730 103 770 109 11 -2 1, 160 211 11-3 770 108 810 11 4 11 -3 1_250 227 11-4 810 114 850 120 11 -4 1_340 243 11 -5 920 130 980 138 11 ·5 1,470 267 11-6 1,030 145 1,100 155 11 -6 1,600 291 11·7 1, 170 165 1_250 176 11 -7 1,705 310 11-8 1,)00 185 1,380 195 11 -8 1,810 329 11-9 1,480 210 1,570 221 11 ·9 2,020 367 12-0 1,650 235 1,750 247 12-0 2,230 405 12-1 1,750 247 1,850 261 12-1 2,440 443 12-2 1,850 261 1,950 275 12-2 2_650 480 12-3 2,000 282 2,110 298 12-) 2_860 520 12-4 2,150 303 2,270 320 IN 3,070 560 12-5 2,280 322 2,390 337 12-5 3,210 584 12-6 2,400 338 2_500 352 12-6 3,350 608 12-7 2,520 355 2,660 375 12·7 3,460 628 12-8 2,650 374 2,830 399 NOTU: ~cse~l~i;~ :?:::I~~_.z~ic~~~~~t~d~~: f~on:l~ t~kcz!7':~':~~:ind~~~~~!lC~~i~~~~~)(:~~~~i~~ f:::t~!r tis. two ruches as 141 and. 182 Idlotneucs. Gaulcs and widllu in metra, ar~.u in $Quare kilomclICS. TABLE 398 TROUGH VOLUMES ABOVE MEAN GAUGE 9-80 (i) MALAKAL TO MELUT Mean 0-0) Gauee 0-00 : 002 : 0-03 I, 0-04 i 0-05 i 0-06 I 007 I 008 I Mean 0-09 I Gauge 9-8 0 1 I 2 2 i , 3 3 4 4 5 9-8 9-9 5 10-0 II j 6 6 7 7 8 8 9 9 , 10 9-9 11 12 12 13 i 13 14 14 15 I 16 10-0 10-1 17 18 18 19 19 , 20 20 21 22 , 2) 10-1 102 23 10-3 29 I 24 24 25 26 26 27 : 27 : 28 29 10-2 30 31 32 32 33 34 35 35 ! 35 10-) 10-4 36 36 37 38 38 39 40 40 41 42 10·4 10-5 43 44 44 45 46 47 47 , 48 49 49 10-5 10-6 50 51 52 52 53 54 55 56 56 57 10-6 10·7 58 59 60 60 61 , 62 62 , 63 64 65 10-7 10-8 66 67 68 68 69 I 70 71 72 73 74 10-8 10-9 75 76 77 79 79 80 81 , 82 83 83 10-9 11-0 84 85 86 87 88 89 90 91 92 93 11-0 11-1 94 95 96 97 98 99 100 ; 101 102 i 103 11·1 11·2 104 105 106 107 108 110 111 112 , 113 114 11·2 11 -3 115 116 117 118 11 9 121 , 122 : 123 , 124 125 11-3 11 -4 126 127 128 129 130 132 : 133 134 135 137 11 -4 11 -5 138 139 141 142 144 145 146 148 I 149 151 11-5 11 -6 152 154 155 157 158 160 , 162 163 ! 165 166 11-6 11 -7 168 170 172 173 175 177 I 179 ,: 181 182 184 11-7 11-8 186 188 190 192 194 196 198 200 : 202 204 11 -8 11-9 206 208 210 212 215 217 219 222 ; 224 226 11-9 12-0 228 230 233 235 238 240 243 245 247 250 12-0 12·1 ' 252 254 257 259 262 264 267 269 : 272 274 12-1 12-2 277 280 283 285 288 291 293 296 I 299 301 12-2 12-3 304 307 310 313 316 31 8 321 324 327 330 12-3 12-4 333 336 339 342 345 349 352 355 ; 358 361 12-4 12-5 364 367 371 374 377 380 384 387 391 394 12-5 12-6 397 400 404 : 407 411 414 418 421 425 428 12-6 12-7 431 434 438 441 445 449 i 452 456 I 459 463 12-7 . 12-8 467 471 476 480 485 489 494 499 I 503 I 508 12-8 Non. OauJI;:!I IQ metres aad volumes 10 mlIliolll . - 865 TABLE 398 (continued) CORRECTED VALUES FOR EACH DECIMETRE ALLOWING FOR THE ERROR IN CROSS-SECTION No_ 4 Mean Gauge _J Trough Volume I Mean Gauge Trough. Volume Mean Gauge Trough Volume --.;----- --- 9-8 9-9 10-0 12 11 -0 89 12-0 240 10-1 18 II -I 99 12-1 265 10-2 24 11 -2 110 12-2 292 10-3 30 11-3 121 12-3 321 10-4 37 11-4 133 I N 352 10-5 45 11 -5 146 12-5 385 10-6 53 11 -6 161 12-6 419 10-7 6 1 11-7 177 12-7 455 10-8 70 11 -8 196 I H 494 10-9 79 11 -9 217 12-9 534 - TROUGH VOLUMES ABOVE MEAN GAUGE 9-80 (11) MELUT TO RENK I Mean Gauge 0-00 0-01 0-02 0-03 0-04 0-05 0-06 0-07 0-08 I 0-09 ; Mean I Gauge ------ ----;I --- --i-l--i----:-T- 9-8 4 i 5 6 ; 7 8 i 9 ! 9-8 9-9 10 II 12 13 14 ; 15 16 ! 17 18 i 19 ; 9-9 ,I 10-0 20 21 22 23 24 25 i 26 27 28 ,I ! 29 10-0 10-1 30 31 32 33 34 35 I 37 38 39 40 10-1 10·2 4 1 42 43 44 45 46 48 49 50 51 10-2 10-3 53 54 55 57 58 59 60 6 1 63 I 64 10-3 10-4 66 67 69 70 72 , 73 74 76 11 I 18 10-4 10-5 80 81 83 84 86 88 89 91 92 94 10-5 10-6 95 96 98 99 101 103 104 106 101 109 I 10-6 10-7 III 11 3 11 4 11 6 11 8 119 12 1 123 124 I 126 10-7 10-8 127 129 130 132 133 135 137 138 140 142 10-8 10-9 144 146 148 149 151 153 155 156 158 160 10-9 11 -0 162 164 166 170 110 ; 171 I ! 173 175 177 179 11 -0 II -I I 181 183 185 187 189 I 19 1 I 193 : 195 ; 197 199 II-I 11-2 ; 201 203 205 208 210 i 212 I 214 217 219 22 1 11 -2 11 -3 223 225 228 230 232 ! 235 237 I 240 242 245 11 -3 11 -4 247 249 252 255 257 260 I I 262 I 265 268 271 IH ! ! 11-5 I 273 276 279 281 284 i 287 290 I 292 , 295 298 11-5 11-6 301 304 307 310 313 316 319 322 325 328 11-6 11-7 : 331 334 338 341 344 347 350 ! 354 357 360 11-1 I i'8 363 366 370 373 377 380 384 387 391 394 11 -8 11 -9 398 402 406 409 413 i 417 421 i 424 428 432 11 -9 ; I 12-0 436 j 440 444 I I 449 453 i 457 461 466 410 474 12-0 12-1 418 , 482 487 491 496 ! 501 505 510 ! 514 519 12-1 12-2 524 529 , 534 539 544 I 549 554 559 564 569 12-2 12-3 574 I 579 585 590 12-4 628 634 639 645 , I 601 606 612 , 617 622 12-3 565961 657 662 668 i 614 680 12-4 I 12-5 685 i 691 697 703 709 115 721 727 i 733 739 12-5 12-6 745 i 751 757 I 763 769 I 775 781 I 787 793 199 12-6 I Non: Gau;es in metres aD.d volumes 10 mllIiODJ. 866 TABLE 399 PROBABLE ERRORS OF TROUGH VOLUMES ABOVE MEAN GAUGE 9'80 (;) MAlAKAL TO MELUT Mean Gauge I Percentage I Cumulative IP robable Error Inmmen! Probable Error · Square of Error Probable Error Sum of of Sqllares Trough Volume - - ---l----------I- -----"_.- ... ~ . _--_. _._--, --" (I) (2) (3) (4) (5) (6) (7) 9·8 0 0 0 0 0 10'0 II 25 3 9 9 10·2 12 25 3 9 18 10'4 13 25 3 9 27 10·6 14 21 3 9 36 10·8 16 17 I 3 9 45 II'O 18 17 3 9 54 1 11 ·2 20 17 3 9 63 8 IH 22 16 4 16 79 9 11·6 ! 26 14 4 16 95 10 11 ·8 34 13 4 16 III II 12·0 I 42 14 6 36 147 12 12-2 49 14 7 49 196 14 IN S6 13 7 49 245 16 IH I 64 13 8 64 309 18 IH I 70 13 9 81 I 390 20 (H) MElUT TO RENK 9-8 0 0 0 0 I 0 0 10·0 20 II 2 4 4 I 2 10·2 21 II 2 4 8 I 3 10·, 25 12 3 9 11 , ! , 10·6 29 9 3 9 26 5 10·8 32 1 2 30 5 11 ·0 35 5 2 4 34 6 11 ·2 39 3 I , ,I 3S 6 II" 46 3 I I I 36 6 11·6 54 2 40 6 11·8 62 4 2 4 44 7 12·0 73 S 4 16 I 60 8 12·2 88 5 4 I 16 76 9 IN 104 6 6 36 112 10 12-6 111 6 1 I 49 I 161 13 Naru: The mcrcmcnls m Col. (2) a~ liken ftom Tablll 398. !he pcr.::;cnL1l.II~ 10 Col. (3) (rom Table 26 of the Thfrd Inurun R(poT/. Col. (4) I' the product of these lWo divided by 100 and i.$ thus the p.c. of eacb increment. 10 Col. (5) the!.: arc squared Ilod in Col. (6) the sqU.:U'.:l IlR summed cumwatively. Col. (1) ,bows the square rooCl of each cumulative sum lind therefore Il:iv~ the IHob:lblc cn or o f the volume held III the put.c-ruding at the bclinllio& of the line. The figures in Coli. (2). (-4), Ind (1) are in millions. The pro~ble erron. (or the smleh between MaLakal . nd Melut would be rkcr~aJ~d by the eorr«tion to Cross-So.::tion No. 4 ~inee thjs no"," 'pproximales very closely to the mean, whereu previously it WllS uceptioDally (ar (rom it. SECTION IV. RAINFALL SOURCES OF ERROR In a country where the rain falls gently and uniformly over a wide area the average depth of rain which has fallen on any area in a given period can be estimated with fair accuracy from a few rain-gauges evenly scattered over it. Unfortunately tropical rain is seldom gentle and nearly always localized; moreover in the Southern Sudan the rain-gauges are still too far apart to give a truly representative picture. This is shown by the fact that in the same year two adjacent stations may record totals which are well above and well below their respective averages. Nevertheless, in the present analysis I have placed the gains due to rainfall next after the change of trough volume because no estimate of a factor can or need be made. The relationship between the mean of the various rain-gauge records and the average depth over the whole area may be uncertain and variable; but there are no grounds for assuming any other r.elationship than equality between them. The gains due to rain falling on the surface of the river are therefore calculated very easily once the rainfall stations to be used have been selected and when the average surface area during the period concerned is known. This is derived in the manner already described. In this case, in calculating the rainfall gain for each ten-day period, we u.se of course the mean surface area for the period, and not the value at its end, and we therefore enter Table 397 (p. 865) with the mean gauge-readings corresponding t.o the iniddle of each period . ' 867 Apart from the major source of error described above, about which nothing can be done as far as past years are concerned, there are two further uncertainties which must be considered, thougb neither of tbem is serious. Tbe first is the unknown time-lag between the theoretical and actual surface areas. I bave.neglected this altogether, as I did tbe lag between tbe theo- retical and actual trough volumes, and as I do below the lag between theoretical and actual absorption. As is explained below, all these time-lags are combined in the apparent absorption which is derived empirically and only compared incidentally with its theoretical value. The second uncertainty is the amount of run-off from that part of the trough or valley not covered by the theoretical water surface. This does not seem to be a large item since there is no noticeable diflercnce between the years when rainfall along the river has been high and those when it has been low. A possible exception to this is the year 1928, when unusually heavy rain in April coincided witll an anomalous value of the apparent absorption; this is discussed in more detail in the next section but one, which deals with that subject. Run-offfrom areas outside the trough, entering by well-defined khors, is another matter altogether; as will appear later this is probably one of the most important sources of inflow and is therefore described in a specia l section which follows the analyses of the individual years. In general therefore the gains due to rainfall are obtained by direct multiplication of the mean of the readings at the selected stations during each ten-day period by the mean surface area during that period. Thus, though the results may be subject to error, they are obtained by direct calculation entirely from measured quantities. DATA AND METHOD OF CALCULATION The latest avai lable maps of Sudan rainfall show that conditions are not the same inside and outside the river valley, and I have therefore made use only of records obtained from gauges sited on the river. From Malakal to Renk., during the twenty years considered here, continuous records have only been kept at Malakal, Kodok, Melu t, and Renk. Althougb it reduces sti ll further the number of stations from wbich the averap.es are derived, I have not used the records at Kodok . Conditions change fairly rapidly northwards from Malakal, and there is no corresponding station between Melut and Renk, so that I felt that to include tbe Kodok records in tbe mean would weight it unfairly. The total effect of tbe rainfall is not a large item in the analysis and it did not seem to warrant the complications of a weighted mean. It will be observed that in an average year the mean tota l surface area during the rainy season is about 400 sq. km.; with an average total fall of a li llIe over 600 millimetres this gives a total gain of abo ut 250 millions. Errors in the surface area will not be appreciable compared with the uncertainty in the amo unt of the rai nfall, and if this is assumed to be 20 per cent. we arrive at a probable error in the total of the oruer of 50 millions. For tbe years before 1937 the monthly totals at these three places are given in Volume VI of The Nile Basill, and I simply divided these by three to get the ten-day totals. For the later years, since monthly totals were not readily available, I went back to the original daily records which the Government Meteorologist was kind enough to lend me, and from which I derived the ten-day totals directly myself. Since the monthly totals were available for the earlier years this ex tra refinement did not seem to justify the extra labour involved for them. Even the heaviest falls occurring when the ri ver surface was near its maximum (as for example in September 1939) ca used only a minor effect on the curve of total computed loss. SECTION V. EVAPORATION RISING STAGE Evaporation differs from rainfall in the very important fact that whereas it is possible to measure directly the amount of rain falling on a small known area, and to deduce from the result directly, if not accurately, the rainfall over a large area, there is as yet no really reliable method of relating instrumental observations of evaporation to the total over a large area of open water. Tbere is no need here to discuss the various experiments which have been made and the various estimates given for tbe daily evaporation from an open water surface in these regions at different times of the year. I had hoped that this present investigation. might add materially to the discussion of this problem, but it may as well be admitted at once that this has not happened , though r have obtained rough figures which agree with previous estimates for tbe conditions during the dry season surprisingly well. Tbese are discussed below; at the moment I am concerned with the effect of evaporation during the rising stage of the river. Here we are on fairly sure ground, since by the time the surface area has become large enough for evaporation losses to become appreciable, evaporation itself has become so small that its 868 effects are almost negligible; and this happy state of things persists nearly to the peak of the flood- or even beyond in low years. This is fortunate because it would be impossible from any consideration of the observed losses during this stage to make a ny distinction betwccn those due to evaporation and those due to absorption. It is necessary therefore to rely entirely on previous estimates for evaporation during tbis stage; but on the other hand any errors in these estimates will not have a large effect on tbe losses. In Volume I of The Nile Basin (pp. 57-61) the subject of evaporat ion is discussed thoroughly and a value of 1-2 mm. per day is estimated for this part of the river in the months June to October inclusive. I understand that the modem pract ice in the Egypt ian Irrigation Depart- ment is to halve tbe readings of the Piche tube evaporimeters which are situated a t va rious meteorological sta tions in the Jebel Aulia Reservoir; and I therefore give below the values obtained by halving the monthly means of these read ings at Mulakal and Renk during the years when records were taken: TABLE 400 AVERAGE EVAPORATION FROM AN OPEN WATER SURFACE (piche Tu be Readings(t) x O'S) I May June July I Malalcal (191.1-47) I H ] 2.5 41··6 1 1.3 Renk (1938-48) 8·1 6·0 1 2·7 z1..5s 1 2.0 I H II H8· 5 I· H H H 8·5 1 89-·68 118··61 I 116··01 ··· I~4-2 z.s i 20 N I -;~I~~~ 8~ l-9-.2-+-9-.8-1~ Mean ... (') Supplied by tho Sudan GovernmeM Mt'tcorololisl . Ev.pomtioa in millimeue.s pn day. The average for the faUing stage (December to April inclusive) may be noted as g.g 111m. per day. Taking aU tbese factors into consideration there seems to be little chance of serious error if a value of 2 mm. is accepted for the montbs of August and September, 2-4 mm. for July and October, and 5-8 mm. for May and November; with some allowance made for years which are abnormally dry or wet in the early or late stages of the season. All in aU the total loss by evaporation during this stage, which amounts on the average to a little over 100 millions, may be estimated within 20 or 30 millions in an average yea r. The net gain or loss by evapora tion minus rainfall during this stage may then be reckoned as being accurate to about 50 or 60 millions. It is worth noting before we leave the subject tbat the totaL net loss or gain from these factors up to the peak of the flood is very small and varies li ttle from one year to another. FALLING STAGE During the rising stage therefore the net loss of evaporation minus rainfall can be estimated with a fair degree of accuracy and is a fair.ly small quantity; it is these facts wbich make it possible to estimate the absorption factor in years when there is apparently no inflow, as will be described in the next section. But as tbe river falls tbe situation is reversed: absorption may then be assumed with some confidence to have ceased, and therefore all the losses can be ascribed without serious error to evaporation. By making allowance fo r tbe volume of water reLeased by the emptying of the trough it is possibLe to deduce from the observed differences between the discharges at the two ends how much has been lost from its surface. Unfortunately, however, it is in this part of the year that infiow most commonly occurs, and there are apparently very few years in which it is at all certain that there has been none. Thus, as will be seen below, only three determinations of the evaporation factor can be made from the data at present available. This is not enough to give one much confidence in the results (though they agree with each other and with previous estimates quite well), nor is it enough to determine whether tho whole surface of the river should be treated as uniform or divided into two parts : the open channel and the flood-plain. This requires a word of explanation. Surface conditions in these two areas are different after the peak of the flood, because on the flood-plain the grass has grown up enough to protect the water ·surface when it begins to fall, and so to produce conditions for evaporation very different from those on the main channel where the surface remains completely open the whole time. During the rising stage this difference probably does not occur, if the water rises faster than the grass can grow ; and in any case evaporation at that time is so small that this difference will have little effect on the total. But in the faUing stage, when evaporation from an open water surface is of the order of 10 = .p er day during one or two months, the difference between 869 the evaporation from these two areas may be considerable. It has for example been estimated that evaporation from sudd is of the order of 4 mm. per day, and if this sort of value were assumed for the flood-plain it would affect the estimates for the total area con- siderably. If enough records were available from years of no inflow, condition equations could be derived from which the most probable values of the evaporation over both areas could be obtained. As it is at present, with only three years of no inflow in which discharges were fully recorded at Renk, we can only obtain an average value for the whole surface during the falling stage. METHO D OF CALCULATION In Table 401 (p. 871) the calculation of this average value is shown, using the only six years in which discharges were measured at Renk during this stage, and in which the losses during it have anything like reasonable values. Eacb of the six years has been allotted a column in the table, and the different lines, whicb are numbered, sbow the data and the successive stages in the calculation of tbe average evaporation depth. Line (I) gives the cumulative observed loss allhe peak of each fl ood, just before tbe river began to fall . Tbis level has usually been held for at least a month and all the delays in filling the side basins may be assumed to have been taken up, so that the theoretical and actual trough volumes should be equal. Line (2) shows the cumulative observed loss under similar conditions when the river has fallen to its low level and been more or less stationary for a month or so. The figures in both these lines are taken from the second parts of tbe relevant years of Table 404-18 (pp. 885-900). Line (3) gives the difference between them or the total amo unt of water, less that lost by evaporation, which has apparently been released by the decrease in volume of tbe trough. It is also the excess of the discbarges at Renk over those at Malakal during this falling stage. This of course is assuming that there has been no inflow; the way in which 1 endeavour to eliminate this is sbown below. Lines (4) and (5) give the tbeoretical volumes of the lrough /l.t the same stages of tbe river, i.e. at the peak and at tbe end of the falling s:age. Tbey also are taken from tbe second parts of the relevant years in Tables 404-18. The difference between tbem is given in line (6), and it will be clear that this represents the whole of the theoretical amount of water released by the emptying of tbe trough. The diJference between this and line (3) must therefore be the amount lost by evaporation, assuming that there has been no inflow. This diJference is given in line (7) . We tbus bave in this line the total observed loss due to evaporation, and in order to derive fro m this the average depth of evaporation in a given period of time it must be divided by the product of tbe average area of the water surface and the time during whicb the loss has taken place. This product is most easily derived by adding together tbe surface areas corresponding to each of the ten-day periods between the peak of the flood and its end, between which points botb the tbeoretical and the observed losses obtained above have been taken. These cumulative totals of surface area are taken from tbe last column oftbe first parts of Tables 404-18. lfthe values in line (7)-the deduced loss by evaporation in millions-are multiplied by a bundred and tben divided by this cumulative surface area tbe result wi ll be the average evaporation in centimetres per ten-day period, or in millimetres per day. DISCUSSION OF RESULTS It is clear at once that only three years-I929, 1937, and 1946-give results which are at all consistent with eacb o ther and witb the average of half the Piche tube readings quoted above. This was 8·8 mm., which may be compared with the mean of7·8 rom. obtained from these three years. It will be seen tbat the general figures given by previous estimates are confirmed if inflow is assumed in small quantities in the other three years, as well as in large amounts in the remaining years excluded fro m tbe table. It may also be argued that the slightly low value which r have obtained is due to the presence of grass over part of the surface area, as already explained above. But are we justified in assuming that the calculation in the other three years sbown in Table 399 (p. 867) is invalidated by inflow, and that the differences in the results are not merely due to errors in the otber items of the account? This question can only be answered by a consideration of these errors in the year 1944-45 where the average depth comes closest to the mean without being accepted for it. (1 have rejected the value obtained from 1932-33 because for two out of the four months no discbarges were measured at Renk.) In order to bring the evaporation depth in 1945 to the mean we must add over 250 millions to the value of the deduced loss by evaporation given in line (7), for there cannot be a serious error in the cumulative total of the surface areas. This deduced loss is obtained by the difference between the cumulative observed loss during the falling stage and the maximum trough volume minus the trough volume at the end of the falling stage. It can be shown from the arguments 870 given on pages 859 and 864 that tbe theoretical probable error in the trough volume was less than a quarter of the probable error in Ute observed loss, and that this scarcely exceeded 100 millions during this period."1 If therefore lhere was no inflow during this stage we should have to assume an unlikely combination of errors to account for the fact that the deri ved evaporation depth is so far from the mean. 1 see no reason for doing this when it seems a t least possible that inflow has occurred, and I therefore feel justified in rejecting the value obtai ned in this year. This rejection, based entirely on a consideration of the interna l errors of the account, is confirmed by lhe fact that the three concordant results agree also with the val ues which are used in the Jebel Aulia Reservoir as a result of many years' experience. METHOD OF ESTIMATION IN THE ANALYSES I have treated the preceding investigation purely as a confirmation of the previous estimates and of my theories, and for calculating the evaporation losses during the falling stage in all the years treated in Tables 403 and 404-18 I have used values based on the average of half the Piche tube readings as shown on page 869. These have been rounded off and some allowance has been made in abnormal years; but I have not scrupled to use 10 instead of 9 or II in order to ease the calculation of the losses. The effect on the total computed loss is negligible com- pared with the effects of the inevitable uncertainties in the values of the other items which go to make it up. TABLE 401 CALCULATION OF EVAPORATION DEPTH 1928-29 I I 932- lJ I 1936-37 I 1937- 38 I 1944-45 I 1945-46 Observed Loss at Flood Peak (1) 948 1.67 1 ; 725 58 1 , 1,045 I 1.362 Final Observed Loss (2) 625 I 845 55 1 I - III i 472 I 805 Difference .. . ... (3) 323 I 826 174 I 692 I 573 I 557 I Maximum Trough Volume (4) 667 1,062 571 I 746 672 834 Fmal Trough Volume ... (5) - 2 I 52 - 23 I 27 I -59 I 4 Difference, equals wal'er re- I I I turned (6) 669 I 1,0 10 , 594 I 719 I 731 i 830 I , Deduced Loss by Evapo ra tion (7) I 346 184 420 27 I 158 I 273 Cumulative Total of Surface I1 Area ... ... ... (8) 3,993 5,279 4, 143 5,513 5,376 5,60 1 Average Evaporation in mill i- metres per day.. . ... (9) I 8·7 3·5 10 I 0·5 i 2·9 i I 4·9 Nons . AJJ ftptaarc l.Dmillionsexcept thoselllllna(a)a.nd(9)whlCh ut:I.OJqUMt:kiJomdru and mllbmelfes rupec;:lIvely. lirm fi~2r 'i1:s(U) ~~~ m::: I~ee: f;:..mth~}o~~ ~~u,:.n o~~~1wI~ ~;ctP/~~f1a~~Jr~r81. y~~~ i{6i~b:~) ~~8{S).Lir=~:{~ is (6) m.inu (l). ~n (:?~'I~~m:: ~&':'d~!id!sm~"(~rIY tht: totals ill the b It co lumn oflht: fint parts ofthc relcv3nt run in Tabid ,,().I....I 8. Lino (9) SECTION VI. ABSORPTION PREVIOUS ESTIMATES As the river rises it covers land which has, apart from rain, been dry since the previous flood and which will therefore absorb a certain amount of water. I have not seen any previous published measurements or estimates of the average depth of absorption, apart from the estimate of 150 em. given on page 31 of the First Interim Report, which is frankly described as " little more than a hopeful guess". This estimate was based on conditions in the Geziea, where a fortnightly watering of 10 cm. is made for a period of eight months. In my opinion this figure is too large for the conditions here, for two reasons. In the first place a series of shallow temporary floods will probably lead to more absorption than a single prolonged one, during which the ground will have no chance to dry out or to pass moisture up to the air through the plants rooted in it. In the second place the author of that section of the report has confirmed my suspicion that in thinking of these repeated shallow floods as being easily absorbed he forgot that they were also being dried up by evaporation, and he made no allowance for this, although over such a long period it forms a considerable proportion of the total loss. I under- stand that in the working of tbe Jehel Aulia Reservoir a very much smaller absorption depth- of. t'h e order of 25 to 50 crn.-is used; but in this case I believe no allowance has been made for 871 inflow, which may quite possibly occur even north of Renk. Certainly one of the known and proven sources of inflow comes into the river only just above Renk, as will be shown in Section VIII (p. 902). In the Gash, wbere conditions are perhaps not too dissimilar, the resident engineer Mr. H. Bell has informed me that tbe allowance for loss by both evaporation and absorption is 80 to 100 em. This covers a period of about 30 days, in July or August, when the average evapora- tion may be taken as 5 mm. per day, or 15 cm. in a ll, and the average rainfall in a month is about 10 em. Thus tbe assumed absorption deptb there may be taken as between 75 and 95 cm., and it will be seen that although I learnt of this figure after my calculations for its value on the White Nile were complete, there is a surprisingly close agreement between the two results. METHOD OF CALCULATION In the following attempt to estimate absorption on the White Nile both the factors of evaporation and rainfall have been allowed for, and I have a lso done the best I can to eliminate the uncerta inties due to inflow. The analysis is in three parts: first an a ttempt is made to distinguish the years in which during the rising stage inflow was either negligible or small enough to be allowed for ; secondly from these years a mean value for the average absorption depth is deri ved so that the total absorption loss at the peak of the flood can be estimated in all YeaIS; and finally a rough and ready method is evolved for interpolating the apparent absorption losses throughout the rising stage in all years, so as to complete the running totals of the com- puted loss for comparison with those of the observed loss throughout this stage. It must be clear that a lthough we may derive a value for the total absorption loss, and so for the average depth, when the river has risen to its peak and been steady there for about a month, this depth cannot be used directly to estimate the absorption and total losses during the intermediate stages of the rise. The reason for this is that the actual absorption loss will be offset by an apparent gain due to the delays in filling the side basins and channels; delays which make the true trough volume at any stage and a lso the tota l and flood-plain surface areas less than their theoretical values. r have therefore included a ll these apparent gains, of whicb by far the largest must be the lag in the trough volume, in a term which I have called the' apparent' absorption when I am calculating it , and the ' mean' absorption when I am using a mean derived from the apparent values. 'True' absorption can only be shown at and after the end of the flood peak, when all its lags have been taken up. In the individual ten-day periods of the risi ng stage, or ra ther at their ends, the apparent absorption can be obtained from the following equation, wh ich is only that of page 855 of Section I expressed in a different form. The equation is : 'Cumulative Observed Loss minus Apparent Absorption (= True Absorption Theoretical Trough Volume minus ' gains' due to de lays in tilling the minus side-chaonels) Theoretical Evaporation Loss plus Theoretical Rainfall Gain In using this eq uation it is clearly only worth while to work from years in whicb discharges at Renk have been actually measured and a re not derived on ly from a gauge/ discharge curve. The year 1930 can therefore not be used. In some of the remaining years, as will be seen from the second parts of Table 404-18, the apparent absorption as derived above is never large and is even negative; I have assumed that in these years the equations are invalidated by inflow, which is deliberately excl uded from them. In yet other years it has a negative value to start witb, or in the early stages, and in others again after a satisfactory start it suddenly begins to decrease, and becomes negative just when, by theory, it ought to be attaining a comparatively large value. In both these I assume inflow during part of the time, and in the first case I attempt to make an aUowance for it. In only two years of recorded discharges at Renk-1944 and 1945 -does the apparent absorption behave as theory predicts: starting slowly but steadily increasing at an increasing rate, and then tailing ofT to a maximum value which is held at the peak of the flood . SELECTION OF YEARS OF NO fNFLOW After these general remarks we may proceed to the first step of selecting the years from which a mean val ue of the total average depth of absorption can be derived. In Table 402 (i), p. 874, and Fig. J 2 apparent absorption as derived above is shown for the only six years in which it has a reasonable value. It is compared with Malakal gauge since this can be roughly correlated with the area of the flood-plain, as was shown in Appendix IV of the Third Interim Report. The considerable differences in different years did not seem to make it worth while 872 to compare actual computed flo od-plain areas and apparent abso rption directly, though this should ideally have been done. In Fig. J 2 the different characteristics of the different years can perhaps be distinguished more easi ly than in Table 402 (i), and reference should be made to it during the discussion which foll ows. It will be observed that the curves fall into three different groups: (8) 1928 and 1936, in which absorption is apparently negative for the firs t metre or so of the rise , but then increases rapidly so that its final value is quite large. (b) 1944 and 1945, in which it increases steadily from a very small value to one which is quite large. (e) 1940 and 1946, in which it starts steadily as in group (b), but suddenly begins to decrease before thc river reaches its highest level, and finishes with a very small or even negative va luc. Tn both the tab le and the diagram] have stopped showing its va lues when it becomes negative. There are, I think, some grounds for saying that one factor in the apparent absorption should be fairly constant from year to year, assliming a fairly constant rate of rise, and that is the total time-lag between theoretical and observed losses due to tbe delay in filling the side basins. The large differences whicb appear in apparent absorption in the ea rly stage between different years can therefore be ascribed wi th some reason to inflow, assuming a small amount carried on from the previous year. Certainly negative apparent absorption seems to be found foll;\lrp. GaLlgl! : Ab:.;urp. G:t u~c ; Abs('Irp. G.luge A bsurp . G auge Absorp. Gauge Absorp. -- -_. - - -- - I - - 9'89 , IU·12 !(- 3) 10 ·15 -- 3 10'02 - 18 9'44 5 9·5) 0 1O'(l'} 2! 10·2.\ ( 54) 10·38 - 19 10·5 t 4 9·50 7 9·70 2 10,)4 " 10'52 j(-' 8) 10·68 - 26 10-75 5 10-17 IJ 9·87 IJ \U' ~ 5 ·- 25 1ll·7(, • \ 74) 10·86 - 9 10 ·89 35 10 ·50 16 10·16 32 10·7U - IO'9\) I 47 10·90 + 5 11 ·10 46 10·63 29 to'48 27 10 ·S.1 I 2 1 11 · 17 ·_· I \) 11 ·09 35 11 ·22 76 10 '83 55 10'77 36 10·97 25 11-3 1 ! I 10 11 ·30 59 11·)4 117 10 ·99 84 10·97 68 11 ·12 '" 11 ·4) I 35 11·46 62 II-4S 15 t II · t4 126 tHI 100 11 ·26 19 tl ' 5~ I 5) 11 ·55 56 11 -64 180 11 '27 162 11 '23 152 11'40 11 11 '()5 I 62 11'69 65 11·70 204 11 ·42 196 11·56 20) 11 ·53 47 11 ·75 I 62 11 ·8 1 49 11·77 226 11 ·60 205 11 ·8) 172 11 ·70 57 I I ·g) 75 11 ·85 19 11·86 246 11 ·)2 236 12·37 140 11 ·'10 53 11 ·9 1 I 8) 11·89 2 11 ·9') 290 11 ·79 24 1 IH t 80 11 ·94 32 11 .1)7 % 11 '9~ I 12 12·03 306 1) ,96 227 12-74 IJ 12·0 1 27 12-04 I 109 - -- 12·10 320 12' 1) 26 1 IH I - I 12·09 36 12-08 115 - .- 12' 19 345 12·25 299 - - 12· 15 54 12·12 IJ I _. . -- 12-22 398 12-34 378 - - 12·2 1 67 12' 15 136 - .-- IN 4 4 1) 12-37 449 - - 12·28 103 12· 17 140 - - - 12-22 40 1 12·37 5 t8 - - 12'30 I h l 11·17 146 12-36 568 12-30 216 - - - - - - -- - - - - 12·35 592 - - 12-30 !73 -.-- -- - - - - 12-34 606 - -12·29 3 12 --- - - - - 12-32 607 -12-28 -37 1 - j - - - - - - - - 874 TABLE 402 (col/tinued) (i i) CA LCULATION OF AVERAGE ABSORPTION DEPTH _____________ ._ __ .~~ L 1944 1945 ._.1936 Apparent nbsorption at peak of flood ... . .. 1 37 1 11 146 607 Co~~tio~r n:~'im"~ nes"i". va.I.~e of ~pparc~:1 25 83 18 Corf7Cted total absorption, ... .. . ... . .. I 396 229 4 19 607 M aXlmum area of Hood-plain ... ... .. . .. I 484 I 419 507 601 ---f----·- - Average depth of wa ter absorbed in centimetres "' 1 82 55 83 11 0 1 I NOT'I! : AlltiaurcsArtin,nilUonJuceplln thc lllulinc, SECTION VII. ANALYSES OF INDIVIDUAL YEARS INTRODUCTION In each of the preceding sections tbe methods of calculation and estimation for each item have been explained . In Table 403 (pp . 877-83) tbe working out of one year, 194~7, is shown in full detail, with a page of explanation opposite each of the three parts which go to make up the table. In Tables 404-18 (pp. 885-900) summary analyses of the remaining fifteen years for which adequate discharge records exist are shown. These summaries are in two parts only: tbe first shows the rainfall data and gauge-readings from which the analysis was made, and the second shows the different items and totals of the computed loss. In this part are also shown the observed discharges at Malakal and Renk and the • observed' losses derived from them. In the last two columns (sometimes one is omitted) are shown the apparent absorption and the in1Iow, which are both deduced by comparing the observed and computed losses. Table 419 (p. 901) summarizes all these items for tbe different fl oods, giving their maximum or final values. It will be noticed that even the values of these last items bave been left to the last figure and not rounded off, though I have tried to eliminate minor differences between the observed and computed totals which are too small to be certainly called in1Iow, and this is not sbown till its cumulative value exceed~ 100 millions. I thought that it might be easier for anyone who took the trouble to follow the construction of these tables if the figures were left just as they were obtained by calculation, though of course, particularly in a high flood, the values of the total losses at the end of the flood may have uncertainties of well over a hundred millions. 875 TABLE 403. COMPLETE ANALYSIS OF ONE YEAR (1946-47) (i) GAUGE-READINGS AND TROUGH VOLUMES (ii) RAINFALL AND EVAPORATION (iii) CUMULATIVE OBSERVED AND COMPUTED LOSSES 877 TABLE 4lll (i). GAUGE-READINGS AND TROUGH VOLUMES ExPLANATORY NOTES This part of the table shows the gauge-readings and the trough volumes obtained from them. (See SectiOD lll). i.e., tfTI)is+ (U)idC~~ f~)~ t~~ ~~~~f~~:!a~~!i~I&j~ O)~~he~~~ ~~~~~iat~~o\f~~nt;it 1n~~ ':a~~ %:~~: ~a%~~: ~~et~en~e~au~i~Cs~i~g,~~r ~:~g~~:~i~g~tIR~~kf~r ~~=dl~u:,~ :~et~:n:d2y 0i,e~~t~n;~~ ~[.i~)~! !~: direct mean of this with the corresponding gauge-reading a t Melut, i.c. H(S) + (2)). Col. (7) is the mean of each value in Col. (6) with the one immediately fo llowing it in the same column; thus it is the mean gauge-reading for the Mclul- Renk reach corresponding 10 the end of the tcn-day period at the beginning of the line. Cols. (8) and (9) arc the trough volumes for each of the two reaches corresponding to the mean gauge-readings in Cols. (4) and (7) respectively; they are taken rrom Table 398, pans (i) and (ii) respectively. The top line, where the figures are in brackets, represents the volumes held above mean gauge-readings or 9·80 corresponding to the mean gauge- readi ngs at the beginnings or Cols. (4) and (7). In this case these are below 9·80 and so the trough volumes are negative. The remaining values in these two columns are obtained by entering Table 398 with the corresponding mean puge- readings in Cols. (4) and (7), and then adding to the results the figures in the brack.ets (i.e., 15 in Col. (8) and 12 in Col. (9». Col. (10) is the sum or Cols. (8) and (9). It is represemed by the dotted li ne in Fig. J 3. 878 TABLE 403 COMPLETE ANALYSIS OF ONE YEAR (1946-41) (i) GAUGE-READINGS AND TROUGH VOLUMES TROUGH VOLUME.' Ten-day Period - Malakal l Mel,ut Direc::t '! IA1 cnd Gauge I Direct T":; ~,;;i M;iakal l Mc\ut-~ I I Mean 10 Days Mean 10 Days - McJut Renk 0 a ( I) (2) (3) (4) (5) (6) (7) (8) (9) ( IO) (- IS) April _ .. (- 12) (-21) 9·52 9·52 9·52 1 9'52 9·78 9·65 9·64 0 0 0 May ... ... 9'53 9·49 9·51 9'58 9·77 9·63 9·68 3 3 6 9·70 9-61 9·66 9·74 9'83 9·72 9·78 II 10 21 9·87 9·76 9·82 9·94 9·92 9 ·84 9'93 22 25 47 June ... 10·16 9·97 10·06 10·22 10·06 10·02 10·16 39 49 88 10·48 10·30 10·39 10·52 10·30 10·30 I I 10·4 1 59 79 138 10'77 10·56 10·66 10·77 10'49 10·52 10·62 I 79 110 189 I July .. . . i 10·97 I 10·78 10·88 10·95 10·67 10·72 10·80 94 139 233 I 11·1\ 10·94 11·02 1\ ·08 10·83 10·88 10·93 107 161 268 1\ ·23 11 ·05 11 ·14 11 ·29 10·92 10·98 11 ·09 I 129 191 320 Auaust .. . 11 ·56 11 ·31 11 '44 11·57 11·09 11 ·20 I 11 ·33 I 163 242 405 11·83 11·58 11 ·70 11 ·95 11-35 11 ·46 11 ·65 232 328 560 12·37 12·02 12-20 12·34 11 ·67 11 ·84 11 '99 331 444 175 September 12-61 12-33 12·47 12·54 11 ·96 12·14 12·22 392 546 i 938 I 12-74 11;50 12'62 12·66 12 ' 12 12·31 12·38 433 629 1,062 i 12·81 12·60 12·70 12·71 12-28 12·44 12-48 449 686 1,135 October ... ! "' 1 12-82 12·63 12·72 12·72 12·40 12·52 12·55 453 727 1,180 I 12·81 12·64 12·72 12·74 12·51 12·58 12·60 460 757 1,217 12·84 12·69 12·76 12-78 12·55 12·62 12·63 474 775 1,249 November 12·87 12·71 12·79 12·80 12·58 12·64 12'65 482 I 787 1,269 "I1 2-90 12·73 12·82 12·82 12·60 12·66 12·66 491 I 793 1,284 12·90 12·74 12-82 12-83 12·56 12-65 12-66 495 793 1,288 December 12-92 12-75 12·84 12·84 12·56 12·66 12·64 500 781 1,28 1 "I 12-92 12-74 12·83 12·83 12·52 12·63 12·62 495 I 769 1,264 12·92 12'74 12·83 12·82 12-50 12-62 12·63 49 1 775 1,266 , January " . I I 12·92 12·74 12·83 12·82 12·52 12·63 12·64 12·90 12·74 12·82 12·80 12·53 12·64 12·64 I 491 I 781 1,272 482 j 781 I 1,273 12·84 12·74 12·79 12·74 12·53 12·64 12 '63 t I 460 I 775 1,235 February . 1 12'66 12-70 12·68 12·50 12·53 12·62 12·55 379 727 1,106 12' 12 12-49 12-30 12·00 12-48 12-48 12·32 243 597 840 I 11·39 12·00 11 ·70 11 ·41 12-32 12·16 11·92 142 418 560 March I I ". 10-86 11 ·37 11 ·12 10·88 11 ·98 11·68 11 '44 88 269 357 1 10'46 10·81 10·64 10'SO 11 ·58 11 ·19 , 11·28 58 231 289 I 10·23 10'48 10·36 10·30 11 ·26 11 ·37 11 ·01 34 176 210 April ··· 1 1M3 10·31 10·22 10·23 10'96 10·64 I 10·59 40 106 146 I 10·23 10·26 10·24 10·28 10·82 10·54 10'52 43 95 138 10 32 10·32 10·32 10·31 10·62 10·50 10·47 45 88 III May ... ... 1 .... 10·30 10·28 10·30 10·30 10·61 10·45 I 10·45 I 44 86 130 Nons : Gauccs att in metm ud vollolITlC!l in millions. For ClIplanation of the lable _ the opposite pace. 879 TABLE 403 (iil. RAINFALL AND EVAPORATION ExPLANATORY NOTES This part of the table shows the basic rainfall data and assumed evaporation depths, and tbe calculation therefrom of the net evaporation loss. (See Section IV.) Cols. (1), (2), and (3) show the recorded rainfall for each ten-day period at Malakal, Melut, and Renk respectively, in millimetres. Col. (4) shows the arithmetic mean, or i(I) + (2) + (3)). Col. (5) is the assumed evaporation in milUmetres per day. and Col. (6) is the net evaporation or (5) - f/ IO x (4). Col. (4) has to be divided by ten to bring it to millimetres per ten-day period. Col. (7) and Col. (8) are the surface areas corresponding to the middles of tbe ten-day periods, and are taken from Table 397 using the mean gauges in Cols. (3) and (6) of part (i) of this table. Col. (9) is the sum of these, i.e. (7) + (8). The flood-plain area at any time required for estimating the absorption is derived by subtracting the value in this column from the value at the bead of it, which represents the surface area of the low-level channeJ. 10 this case it is 130. Col. (10) is the net evaporation loss for each ten-day period in millions, and is obtained by multiplying the value in Col. (6) by that in Col. (9) and also by the number of days in the period (ten, eleven or eight) aDd then dividiog by 1,000 to bring the result to millions of cubic metres, i.e., (10) _@..x (9) ;~. I~~) 880 TABLE 403 (continlled) COMPLETE ANALYSIS OF ONE YEAR (1946-411 (iii RAINFALL AND EVA PORATION RAINFALL l EVAPORAnON I AREAS I Net Ten-Day Period ~~~~: '~e: i-Gro~ --I ~I ~a'ak::l' l Melut To tal IEVLtlr:nra-• Evap. Evap. M clul Renk ass . -" '--"-r.c: --- .. - . - -- millimetrcs __ : m~~mde~;es square ki lometres mil lions ----------+---~---,----~I- ,-- (I) (2) (ll (4) (5) (6) (11 (81 (9) (10) I May ... ... 3 24 9 50 80 1)0 o o I I o 52 85 1J1 23 55 26 54 99 15) I June ·'1 91 13 19 I 8) - 3 51 103 160 - 5 31 95 31 I 54 - 1 10 122 192 - 2 41 51 34 o 18 145 223 0 July 41 42 6: i 51 - 3 88 161 249 - 8 8 JJ 42 28 - 1 94 115 269 - 3 'i 119 59 134 104 - 8 101 18) 284 - 25 1 AUlust ... I 142 41 98 94 -1 120 211 3)1 - 24 , 10) 91 16 12 -5 165 251 422 - 21 64 16 65 68 2 I - 5 261 345 , 606 - 33 1 I September ... ... ,! 58 51 I 41 52 2 i - 3 3\1 458 115 - 2) , 42 38 I 41 40 2 ! - 2 341 524 ! 865 - 11 I '8 83 3) 2 - 1 )55 ! 510 925 - 9 : October .. . ... i 44 7 I 35 29 I 3 0 )59 589 948 0 , 31 19 101 50 3 -2 359 603 962 - 19 , 13 9 0 i 8 i 3 +2 367 613 980 +22 I ! November ... ... II 0 I 4 I 5 4 3 312 616 988 30 i -- I - 1 - 4 4 378 620 998 - - ! - i 40 - I 5 5 318 618 996 50 i I December .. . 1 I ''' 1 - - 1 - I - I 6 6 382 620 1,002 60 - - - - 7 7 380 615 995 I 70 , ! - - I - - I 8 8 380 613 993 I 88 I 1 January .. . ... - - 1 - - 8 8 378 618 996 80 , - - I - - 9 9 378 618 I 996 90 - - - - 10 10 312 618 990 109 February .. . - - - I - I 10 10 344 613 957 96 .1 - - - I - 10 10 269 584 85) 85 - - - - 10 10 131 477 608 49 March ... .. . - - - - 10 10 98 306 404 40 , - - - - 10 10 76 21 1 287 29 - - - - 10 10 67 177 244 26 April .. . .. . ... 27 6 0 11 8 I 7 60 157 217 15 0 0 0 0 8 8 63 147 210 17 1 7 11 0 6 8 7 66 144 210 14 May (I) ... I ''' 1 1 0 I 0 o I 6 I 6 64 139 203 12 I Nons: Tbc pcrioch muked • Ire not len days. For explanation of thiJ tlble see the opposite ~. 881 TABLE 403 (iii). CUMULATIVE OBSERVED AND COMPUTED LOSSES ExPLANATORY NOTES This part of the table shows the observed discharges, the observed losses deduced from them. the computed losses obtained from the previous parts of the table, and the calculation of apparent absorption and deduced inflow whicb are obtained from comparison of the observed and computed losses. (See Sections II and V.) Cois. (1) and (2) show the mean observed discharges at Malakal and Renk respectively multiplied by tbe number of days in the period. Col. (3) is the difference between them. or (1) - (2). Col. (4) is the cumulative observed loss, or the cumulative swn of the differences in Col. (3). It is represented by a pecked line in Fig. J 3 and in Figs. J 4, S, 6, 7, 8 and 9. Col. (5) is the trough volume held above the mean gauge at the start of the flood. and merely repeats the last column of the first part of this table. Col. (6) is the cumulative oet evaporation and is obtained by summing cumulatively the values given in the last column of the second part of this table. It is represented by the dash and dot line in Fig. J 3. Col. (7) is the mean absorption . Its maximum value is obtained by subtracting from the maximum area shown in Col. (9) of the second part of this table the area shown at the head of the column, and multiplying the result by 0·8. Intermediate va lues up to the maximum (of 700 millions reached at the end of November) are obtained by a form of interpolation as described in Section IV, It is represented by the line of crosses in Fig, J 3, Col. (8) shows the computed' total of the cumulative losses, or (5) + (6) + (7), It is represented by the continuous black line in Fig. J 3 and in all the P:lrts of Figs, J 4-9, Col. (9) shows the apparent absorption. or (4) - (5) - (6). It is not shown when it becomes negative. Col. (10) shows the deduced inflow or (8) - (4) . It is shown only when it exceeds 100 millions. 882 TABLE 403 (contilluetf) COMPLETE ANALYSIS OF ONE YEAR (1 946-47) (iu) CUMULATIVE OBSERVED AND COMPUTED LOSSES M::~::k ' O{~' ~~' Trough ' CUN~~e, Deduced Volume Evapor. ' MAbeasn. ' cTumotavl. , ' AApbpsa.r , Innow ------- ------;.;--(-1-)--1+--(:)- -.- ~~4) (5) r (6) I (7) (8) (9) (10) ! May .. . 354 342 12 I 12 6 6 I 12 o 39Q 365 25 37 21 14 35 2 477 435 42 79 47 19 66 13 uno ... 496 441 55 134 88 14 o 102 J2 565 522 43 177 138 12 10 160 27 636 576 60 237 189 12 30 231 36 uIy .. , 69Q 622 68 305 2JJ 4 SO 35 1 68 729 665 64 369 268 I 80 349 100 • 837 758 79 448 320 24 I 120 416 152 1- AUi:USt 848 736 112 560 405 160 517 203 924 821 103 663 560 i::: :: I 691 112 • 1,190 1,040 ISO 813 1 775 102 220500 I 923 140 110 September 1,130 1,050 80 1- I' ; - '25 300 1,113 80 220 1,160 1,120 40 :~; ! I ,~~ 1- 142 360 1,280 13 347 1,210 1,}6O 50 983 1,135 420 1,404 421 - 15 ' :' October 1,220 1,180 40 1.1 80 - 151 480 1,509 486 1,200 1,200 o 1,023 I' 1,023 1,217 -170 540 1,587 564 • 1,360 1,330 30 1,053 1,249 -148 59Q 1,691 638 I November 1,240 1,220 20 1,073 1,269 - 1' 8 I 630 1,781 708 1,240 1,2 10 30 1,103 1,284 - 78 I 670 1,876 773 1,250 1, 19Q 60 1,163 1,288 - 28 i 700 1,970 807 I December I,2SO 1, 19Q 60 1,223 1,281 + 32 I 700 2,013 79Q 1,250 1,180 70 1,293 1,264 102 700 2,066 773 • 1,380 1,270 110 1,403 1,266 19Q 700 2, 156 753 I January 1,240 1,170 70 I 1,473 , 1,272 270 700 2,242 769 1,200 1,160 40 1,513 ' 1,273 360 700 2,333 820 • 1.310 1,270 40 1,553 1,235 469 700 2,404 851 February 1,000 1,160 - 160 1,393 1,106 565 700 2,37 1 978 767 1,130 -363 1,030 840 650 700 2,19Q 1, 160 462 838 -376 654 560 699 700 1,959 1,305 March 517 838 -321 333 357 739 700 1,796 1,463 448 641 -193 140 289 768 I 700 1,757 1,617 480 599 -119 21 210 794 j 700 1,704 1,683 April .. , 439 496 - 57 - 36 146 809 700 1,655 1,691 5J2 484 + 28 - 8 138 826 700 1,664 1,672 484 SOl - 16 - 24 III 840 I 700 1,673 1,697 May (I) 482 510 - 28 - 52 i 130 852 i ! 700 1,682 1,744 Nons: Tbe periods marked· art lIot ten days.. FOI U;p laDatiOD of the table _ the opposite pa,.. All fitW'tS &tc ill millions. 883 TABLES 404-18 SUMMARY ANALYSIS OF FIFTEEN FLOODS ON THE WHITE NILE BETWEEN MALAKAL AND RENK IN THE YEARS 1928 TO 1946 885 TABLE 404 SUMMARY ANALYSIS OF THE YEAR 1928-29 (i) THE DATA Ten-Day 1___ ~G~_ __ . - ~._1 .--_ _ RAJNF~___ ACt~~ed I Total Surface Period Malakal 1 Melu t 1 Renk ! Malakal ; Melut ! Rcok Evaporation A.ea ~;;;; --;--~-.~~--. --~-~~- -I -;-.~~-: 200 I 52 -.-L 46 ~ I 154 : 10,09 10-00 10{)6 I 6 162 May ". I 10·34 10' 18 I 10'16 i I 6 170 _I ~g:~~ ~g:~j I ~g:!j 201 I 34 83 43 ; 218 June .. . 1 ~g:~~ ~&:~i i ~g:~~ i 234 li S 76 78 i 249 263 July ....\ I ~: :~~ :~:~! !~:~ : I ~ 282 : : :~ :n~ j :~ :~ ! 298 195 , 98 I 66 i 325 August 11 -10 11 '49 11 ·18 i 2 378 :: :;~ ll :~~ 1 n::: 232 95 208 ~ 448 499 September I 12·0 1 I 11 -83 i 11·57 2 525 g:~ 1 : l:r, :: :~r 156 81 63 ~ 553 569 October ",I g:i~ 586 I g:&A I :: : ~; 94 95 32 i 618 12 '30 12·] J 11 -7 1 2 632 November .. ! 12·30 12-11 1) ·7) 2 636 12'30 12·11 11 ·73 - I - - .. 636 :;:i: !;::: :: :~! , 6J6 I' ~ 638 12 ·23 12· 10 11 ·7) I - - - 8 630 11 ·94 11 ·96 11 -69 9 559 January n:;g n :~~ n :~6 - I _ _ :g 409 287 10'63 10·64 10·14 J' 10 238 Februa ry 10-43 I J O'42 \ 10·49 I 10 209 10-35 10 '31 I 10·38 !\' - - - 10 193 10·27 10-23 10-32 )0 182 March 10-18 I 10-)5 10·15 10 169 10·10 10,09 I 10,18 I - I! - - 10 166 10'01 ! 10·0) 10' 11 10 159 April (1) 9·92 I 9-94 10'07 : 10 154 NoTU : Penods muked .. arc nOlten days. Oluies arc m melru., raint:&U totals uc aD millimc:tres per month, c\'aporatlon IS III millimctrcs per day, and the surface area is ic s.quare kilometres . SUMMARY ANALYSIS OF THE YEAR 1928-29 Oi) CUMULATIVE OBSERVED AND COMPUTED LOSSES ! DISCHAROES 1 Cumulati\'e 1 COMPUTED LosSEs I D EDUao Ten-Day R -- _I Observed '- - -- -,- -.---- ~- --Period 1 Loss I Trllu&h I Net I Mean 1 I Apparent 1 Malakal Renk Volume Evaporation{ Absorption Total i Absorption Innow , - -- April 444 421 23 ! 16 5 0 21 i 2 - 479 473 29 41 10 i 0 5 1 I -22 I - May ". , 537 501 I 65 I 72 15 0 I 87 -22 - , 588 I 554 99 105 I 19 , 10 134 -25 -692 647 144 IJ2 21 , 20 ! 173 - 9 -June 658 601 201 161 19 30 210 +21 - 689 ! 656 234 !, 192 I17. 50 ! 259 25 I -721 698 257 ,, 225 60 299 18 -July , 758 727 288 261 8 80 349 19 i - , 802 I 759 331 298 2 110 ! 410 31 i -926 869 388 345 i - 4 1 130 471 47 - August 886 829 445 1 408 - 20 160 ! 548 57 103 , 929 887 487 472 I - 38 i 190 624 53 137 1,0H 1.032 490 516 - 58 230 ! 688 32 198 Septembc 964 I 95' sao 547 - 7. 260 733 27 : 233 997 975 521 I 577 - 91 i 290 1,020 991 551 605 - 108 320 ! 776 36 ; 254 8 17 54 266 October 1.040 1,010 58 1 6J3 - 11 9 I 340 854 67 I 273 1,070 1,010 , 641 : 657 - 119 350 i 888 103 267 1.200 1.120 72 1 667 - 107 360 920 161 I 199 NovCTD~r 1,090 ,I 1.020 791 I 667 - 94 I 370 943 218 152 1,090 1,010 871 667 - 69 I 380 I 978 273 107 1,060 983 948 667 I - 31 390 1,026 312 - Decembe 1,060 ! 968 1,040 662 + 7 390 I i 1.0.s9 371 -, 1,020 980 1,080 609 57 390 1,056 414 -969 1,039 1,010 474 : 113 390 977 423 - January 701 , 611 \ 868 1 843 306 I 153 390 : 849 - - 703 75 1 : 184 I 181 390 755 - -612 656 707 11 4 205 390 709 - - February. I 5J8 i 534 , 691 78 226 390 ! 694 - -502 522 ,, 671 I 62 I 245 390 390 . I 400 661 I 49 i 263 390 , 697 -- -702 - March ... 470 486 645 36 I 280 390 700 - -458 463 ! 640 23 287 390 700 - -486 479 647 I II 313 390 714 - -April (1) 429 i 451 I 625 - 2 I 328 390 I 716 - - NOTES . PcnodJ marked aro Dot len clan. .AU fi,u.res arc in millions. Apparent absorption is not shOWII .ncr th. peak of the Good, and deduced inflow II ollly shown ... ben i~ c:umuIa.IJ~ value CJ[~ 100 millions. TroU,Jh VOIWDeS ue cakubled and not taken froro Table 398. 886 TABLE 405 SUMMARY ANALYSIS OF THE YEAR 1929- 30 (i) THE DATA Ten-Day _ GAUOfS 1 RArNfALL I --_., .- I AO~~;;d ! Tol u l Period 1Mal:l:'~ I--I-~eIU: -I' Renk -- ~alllk·o. l -I--Mclut · Rcnk Evaporation l S'f;:~ April 9-88 10·19 I :g; ---:~-r-;-- ~:: -l--:: ---- :---- is -- - i2~4S,~ ----' Moy 10·60 10·70 10·92 10·12 10-58 June 11·12 10-90 10-72 ! 4 26 1 11 ·28 11 -06 10-85 180 98 2.S 4 283 11-38 11 -)7 10-95 " 300 July 11 ·51 11 -30 11 -()4 J )2 1 11·62 11 -43 11 -16 224 IS] 222 ) 150 11 ·80 11 -64 11 -) 1 3 341 AUJUst 11 ·89 11 ·16 11044 2 480 11 ·94 11-80 II -53 290 50 130 2 506 12-01 11-84 II -57 2 ns Septem~r ... 12·10 11-95 11 -65 2 566 12'16 12-01 11 ·71 76 124 52 2 598 12· 18 12-04 11 ·73 2 607 October 12·26 IHO 11 -74 2 6JJ 12·33 12'14 11 ·75 106 62 63 3 650 12·34 12·17 11 ·77 4 665 Noyember ... 12·36 tH 8 11 -76 5 667 12·34 12·18 11 ·76 .)4 5 - 6 665 12·33 12-16 tt -75 7 659 December ... 12·33 12-16 11 ·73 8 652 12·32 12-15 11 -1) - - - 9 I 652 12·25 January ... 11·96 :~ :~i !! :~~ 1 :~ I ~~~ 11·43 lI-s8 11 -47 - - - I 10 I1 402 10·98 11-0 1 11 -06 I', 10 283 February .. . 10·14 10'70 10·72 10 24) 10'49 10-47 10'53 - - - 10 2 16 10·]2 Marcb 10·30 : ~:i~ ' ~~:;~ ,i I' :~ 1 : ~~ 10·25 10-19 10-24 - - - 10 ' 171 10·09 10-06 10-17 I 10 I 164 April(l) .. , 9·91 9-94 10-09 8 . 158 NOTD. Pe:nocls marked· Ire 001 len dll)". Gau,a are an melrU, raanr.U lo uts Irc If! It\lllime~J per month, e ..... por.lJon IJ In mllhmetm p.:r day, and. lhe St1ffaoe aru is lIIlquaR ItilometJQ. SUMMARY ANALYSIS OF THE YEAR 1929-30 (ii) CUMULATIVE OBSERVED AND COMPUTED LOSSES I Ten-Day __D _"_CH,A_'_O_" _ _ I C6~~~~de ___ DEDUCfD - ,-_C_O_M_'U__T £,D-- LOS3£S Period MnJakal I Renk Lo" I Trou&h : Net -Mean 1, TO-:;--! Apparent ~ Volume ,EvaporatiOn Absorption i Absorpl ion April (3) 504 464 40 42 6 o 48 May ... ,96 >49 87 92 10 o 102 619 6 11 9S 124 14 10 148 742 690 147 164 19 20 203 June 728 670 20S 20S 22 30 260 771 710 266 244 2l 40 309 N 800 743 m 276 28 60 364 July .. . 838 778 383 308 IS 80 403 E 862 800 44' 364 I 100 46S 984 943 486 436 -16 120 S40 V Aucust 913 902 497 477 -30 140 S87 922 9S? 462 S02 -4l 170 627 E 16' 1,037 1,045 414 >41 -6 1 200 680 216 R Septembe 976 981 449 S86 - 67 230 749 300 998 1,000 447 610 - 73 260 797 3S0 1,010 1.010 447 633 - 79 300 854 p 407 October 1.050 1,020 477 6S9 - 8S 330 904 427 1,070 1,040 ~j I 678 - 8S 360 9S3 o 446 1, 190 1.160 691 - 78 390 1,003 466 Novembe 1,090 1.030 69 1 - 4l 400 1.046 S 449 1,040 S9~:g;g S97 7 Ii 686 - S 410 1.09 1 494 1,020 S87 67l +4 1 410 1,126 I S39 Deeembe 1,020 1,020 S87 I 6" 93 410 1. 174 T S87 1,010 1,010 6" 1S2 410 1.217 630 1,050 1,090 >S8477 S97 21S 410 1, 222 I 67S January 8>4 966 410 699 689 908 423S I 438 271 1. 134 16 284 311 410 V 789 639 818 37 183 339 410 I.~~ 9l> February 146 617 I 11 8 363 410 89 1 E 93S S03 S66 -10474 76 38S 410 871 978 377 433 - 163 l2 404 410 866 1,029 March .. 466 S07 1, 418 499 :~~ II 422 410 87l 079 3 1 439 410 880 1,125 477 l26 4-294 1 32 I' 44l 410 867 1,161 April (I) 416 461 -339 4 467 410 881 1,220 887 TABLE 406 SUMMARY ANALYSIS OF THE YEAR 19l0-31 (i) THE DATA Ten-Day ~ ---------------------I R.A.lNPALL GAUGU . AGssruomsse d I Total Period SurfaQC Malaka l I Melut I Reok Malakal I Mclut I Renk Evaporation Aroo April 9·97 9·94 10-09 - - 10-()6 9-92 10·04 35 - - 7 1S8 10·08 10·00 10-10 7 161 May 10·19 10, )0 I 10-13 6 ! 164 10,12 10-) I 10-18 62 3S 2 1 6 165 10-13 10·10 10' 14 6 164 June 10-)4 10·24 10'21 4 178 10-69 10-5) I 10-4 1 99 60 71 4 216 10·92 10·75 I 10·58 4 I 243 July i 1I {)5 10-90 I 10·72 3 260 11 ' 16 JJ{)() 10·81 92 99 86 3 27> 11 ' ]1 11 ·13 I 10·90 3 291 AUJUSl .. ..I 11 -40 11 ·24 I 11-01 2 I 309 11'50 11 ·)5 11·12 170 161 186 2 333 II -56 11 '42 11-19 2 355 September ... 11 ·64 11 ·51 I 11·71 11 ,5 7 11 '75 11 ·62 I 11·24 2 382 11·28 70 44 91 2 401 11 ,)0 2 I 419 OClober .\ 11 ·79 11 ·66 I I ')] . 2 436 1) -82 11'68 IJ-3S 76 54 30 2 447 11·82 11'69 11 '34 2 449 November H' I 11 ·84 11·70 11 ,)6 2 453 II 'S3 11 ·70 11·36 20 ! - - 4 ! 451 11 ·72 ) 1·64 11 -]4 5 426 December . 11·40 11-42 11 ·25 6 348 11.o J 11{)6 IH)] - - - 7 288 10·73 10·77 10·78 8 247 Jaouary ... . 10-44 10 ·47 10·54 9 215 10·24 10·26 tO'lS - - - 10 184 10·13 10 '14 10·2 1 10 167 February 10·00 10·02 10·13 10 160 9·92 9·95 10-05 - - - 10 158 9-86 ·9·88 10 '00 10 155 March 9 '84 9-85 i 9·96 10 I 153 9·74 9·79 9·71 9'74 1 9·9} - - - 10 lS I I 9·90 : 10 150 April _. . 9 '7 1 9'75 9·92 10 150 , 9 ·12 9'74 9'91 3 10 - 9 150 9·71 9·74 9·9 1 8 ISO May (I) 9-66 9·70 I 9 ·88 (18) I (5) (15) 7 I ISO NC7TD. Ptnods marked are no t tcn d.y,. aauacs are In meltes.. t1oln(aU totals .to In mllllmetteS pcr month, but (Ot May (I) 19)1 one tlurd o ( the 10111 is ,h'Cn ( in bladets). E\'4pot1olioo 11 in millimct.te.s pcr day, and tbe surface area is in 5qUaR ltiIometrcs. SUMMARY ANALYSIS OF THE YEAR 19lO-l1 (ii) CUMULATIVE OBSERVED AND COMPUTED LOSSES Ten·Day DISCHAROES I Cumulillive I COMPUTED loSSES Period -Milik~ I-[ -~enk-- Observed --T;:Qugh---I --N-;t- - - Mean-- -----Loss Volume Evaporation f Absorption Total April 455 :rsl I 4 14 0 18 467 -I 16 27 0 43 May :1 506 484t 21 22 39 0 61 485 492t 14 2 1 47 0 68 " 540 I S33t 21 30 52 0 82 June .. . i S40 SJ3 48 66 56 0 122 I 617 S73 92 118 58 10 186 : 672 631 133 161 S8 20 239 July ... 712 677 168 192 58 40 290 750 I 708 210 222 S8 60 340 ' i 883 I 811 282 255 58 80 393 Aueusl I 829 : 772 339 288 46 100 434 852 805 386 317 33 120 470 ' j 950 913 423 342 18 140 500 Seplemb« \ 878 848 453 368 18 160 546 890 8S8 485 385 18 180 583 899 868 516 401 18 200 6 19 OClobcr ... 907 878 545 416 18 210 644 915 883 577 422 18 220 660 1,007 969 615 424 18 230 672 November "1 9 19 886 648 424 27 240 691 904 874 678 4 12 45 240 697 1 845 809 714 357 67 240 664 December 734 772 676 268 88 240 596 647 698 625 183 108 240 531 663 683 604 114 130 240 484 January . 530 562 572 62 149 240 45 1 497 SI7 552 33 167 240 440 m 529 546 14 186 240 440 February 439 459 526 - I 203 240 442 422 43S SJ3 - 12 220 240 448 340 334 519 -20 234 240 454 March . 410 408 521 - 27 249 240 462 402 397 526 -34 264 240 470 438 423 54 1 -36 280 240 484 April ... 398 394 545 - 36 295 240 499 399 391 553 -37 309 240 512 398 391 560 -39 321 240 522 May (I) . 391 380 S7I -41 330 240 529 Nons : The tiods m.rked • a Dot len da The dlsc;ha cs lD&lked !he 888 TABLE 407 SUMMARY ANALYSIS OF THE YEAR 1931-32 (i) THE DATA Ten-Day Period April (3) 9·74 9·90 May 9·70 9·88 9· 70 9·8S 9·80 9·92 June 10'14 10,00 10'OJ 10·37 10·23 10·2 1 IO'6~ 10'43 10· ]2 July 10·99 10'74 10'52 11-19 10·97 10·74 11 ·)5 11'15 10·89 August 11 ·48 11 ·30 \I -O! 11 ·62 11-44 11 , \2 127 143 120 11 -72 11 '58 11 ·25 September .. 11 ·85 11'70 11 ·)6 2 455 12·00 11-86 11 -45 277 146 104 2 12·05 11 -94 11'56 2 '5"" October 12' 12 12·01 11 ·6 1 2 574 I H9 12-06 11 ·64 92 72 78 2 596 12·22 12·10 11 ·67 2 612 November .. . 12·28 12'13 11 ·70 4 632 12·27 12· \ 5 11 ·72 15 6 645 12-26 12'16 11 ·73 8 645 December .. 12'23 12·14 11 -73 8 640 12·09 12,08 11 ·72 > 607 11·65 11-85 11 ·62 V S9) Janua.ry 11·01 11'45 11 ·25 10 340 10·58 10·84 10,82 IV 250 10·38 10-41 10'49 10 208 February 10·21 10·27 )0,)3 10 185 10·19 10·19 10'25 10 171 10'11 10·11 10-20 10 166 March 10,04 1082 385 391 85 - 40 44. 280 688 - 60) 895 TABLE 414 SUMMARY ANALYSIS OF THE YEAR 1941-42 (i) THE DATA GAUG~ RAINFALL Tell-Day Assumed Total Period Gross Surface Malakal Mclut Reok Malakal Melut Renk Evaporation Area - ---,---- May ... 9·77 I ; 9·71 9·93 9 '84 9·74 9·92 68 73 32 5 151 10·25 I 10-04 10·09 65 98 57 4 162 June 10·61 10·36 10·)3 39 38 4 198 10·90 10·66 10·51 66 19 6 4 234 11014 10·89 10·69 57 18 16 4 260 luly I 11 ·27 11-04 10·83 10 78 15 3 281 11 ·40 II 'IS 10·95 64 22 26 3 300 "j 11 '51 11 ·30 11·09 48 28 130 3 325 August 11 ·61 11 ·40 11 '10 60 50 7 2 357 11 ·7] J ) -49 11 ')4 60 25 36 2 395 11 ·84 11·62 11 '48 32 100 79 2 450 September .. .. 1 11 ·91 11 '72 II -51 49 35 3 2 487 J ) ,95 11 ·76 11 ·65 2 48 42 2 504 11 ·98 11 ·81 11·74 50 13 22 2 535 October 12·02 11 ·85 11 ·77 J3 II 2 553 12-04 11 ·88 11 ·11 8 9 3 562 11·07 11 ·90 11·71 14 14 4 57S November ... IH2 11 '93 11 '78 22 3 5 589 12'14 11·95 1t '81 6 599 IH7 11 ·97 11 -82 7 610 December .. 12·19 1) ·99 11 ·84 8 621 IH9 12-00 11 ·86 9 626 12'13 J ),99 11·88 10 624 January 11 ·84 1) ·85 11 ·85 10 543 11·25 11 ·43 11·68 10 392 10·73 10'85 11 ']2 10 280 February .. -:I 10'44 10-47 11 ,01 10 235 10·27 10'26 10·79 10 208 10'11 . 10·18 10·6) 10 192 March 10'17 JO.J I 10'46 10 178 10'20 10'10 10·31 10 169 10')0 10'19 10')0 10 176 April 10'10 10,07 10'14 10 166 9·91 9 ·88 10·09 10 IS7 9·83 9·77 9·98 8 154 lions : Prriocb marked· are n01 tCII cbys. GaUl" arC' in metres., rain(all to tals in miltimctm pcr leood. " period, cnpontioD in milllmctres per day, and the surface area in tqUl rc kilometres. SUMMARY ANALYSIS OF THE YEAR 1941-42 (ii) CUMULATIVE OBSERVED AND COMPUTED LOSSES DISCHARGI!S CumulatiVe1! --- '- COMPUTED LossES I DEDUCED Ten-Day Period , Malakal I --- Observed Renk LoS$ I Trougb I Net ! Mean i Total Apparent I I ft 1 Volume Evaporationl Absorptio n I ___- I-A_ b_'_O.,p:-_tio_n-;-_n _o_w_ May .. .• 404 22 -13 550 JUDe ... 588 I m I l~~ li~ I ~ i r- l~ 65 + 5 1- 132 5 655 199 5 702 658 I 228 230 0 30 260 - 2 July ,_, 731 713 246 268 0 SO 318 -22 762 760 248 306 - 3 70 373 126 870 871 247 344 - 11 90 417 170 Aucw( 816 809 254 383 - 24 lJO 469 21S 845 828 271 431 - 32 130 529 258 960 935 296 483 - 56 I SO 577 281 September 893 867 322 516 - 61 170 625 )03 904 885 341 S46 - 66 190 670 329 91 4 904 351 573 - 71 210 712 36 1 October 923 912 362 587 - 11 240 756 394 928 912 378 59S - 60 270 805 427 1,030 1,003 40S 608 - 41 290 857 452 November 943 915 433 622 - 17 310 916 412 947 921 459 637 + 19 330 986 527 950 923 486 649 61 350 1,060 574 Deccmbc 053 928 Sl1 658 III 360 1. 129 618 952 933 530 660 167 370 1,191 667 1,003 1.021 !l06 6J!i 23!i 380 1,230 724 January 802 918 390 !l82 289 380 1,2!11 861 650 812 228 318 328 380 1,026 798 597 137 88 199 3!19 380 938 850 493 !I!I 1 30 133 383 380 896 866 February I 467 507 - 10 100 404 380 884 894 366 389 - 33 80 .19 380 879 912 March ... • 453 467 - 47 66 437 380 883 9)0 460 4!1 1 - 38 67 .54 380 901 939 525 523 - 36 63 47. 380 917 953 Apru .. '1 427 417 - 86 39 491 380 920 1,006 398 433 - 121 16 501 380 903 1,024 389 408 - 140 10 519 380 909 1,049 Nons . '!be penocls masked are not ten days. All fI,ura are in milliollS. The appare.at absorptJoo IS not sltowo whe.o It hu a nepu. ... YlI~ with a hip puae at Malilil. Deduced i.oJ'Iow b mowu oDJ,y wtKn Its cumulative vaI~ o:cceeds 100 mllUoDs. , . 896 TABLE 415 SUMMARY ANALYSIS OF TH E YEAR 1942-43 (il THE DATA G ,WOIlS R AI",'ALL AG"ruomssC d~l.S TuorrA,,C1C Malab.l I Melut Renk M:lIa';:;4 'r Me;~t--r_ _Ren k --+--1------;,- 1I E ~a~=at ':" I _~a April (3) ... 9 'S3 9·77 May 9 ·92 9·81 9 ·98 I 1 ! 5 I S4 10·09 9095 10·56 10·31 JUDO 10-81 10·61 !H! lu 10 I 5 158 I I ~ )1 5 , 192 3 4 22. 10·91 10-74 10'59 J2 \7 10 4 24 1 11 ·16 10'92 10·74 2 19 14 4 266 July 11 ,)0 1I-()& 10·90 39 78 66 2 281 11 -43 11 ·22 11'01 1) 22 40 2 310 11 '56 11 ,)4 SO 2 m AUl\lst 11 ·71 11 -50 1111 '·2192 4)00 I' 2561 15 2 388 11·84 11 ·62 11-43 44 25 25 2 443 12·00 11 -76 11'54 112 100 41 2 506 September ", 12-20 11 ·94 11·66 38 3S 114 2 m 12-26 12·04 11 ·8 1 9 I 48 9 2 629 12·29 12·09 !: ::~ s~ I~ 3 2 661 October 12,)7 12'14 1 4 2 6.3 12-38 )2' 18 11 ·96 10 I 4 3 109 12-)7 12·)7 11 -91 I I 4 107 November .. 12,)5 12·16 11 ·98 5 704 12')5 12' 14 11·98 6 696 12-33 12·14 11·99 1 696 December ... 12-29 12·12 11 ·99 8 688 12·18 12·07 11·97 9 663 11·78 11-85 11 '90 10 541 January ... 11 ·11 11')1 11 ·6' 10 365 10-69 10·71 1I ·J2 10 210 10'48 10-46 11·12 10 238 February ... 10'41 10'38 II{)S 10 m 10'32 10-29 10·89 10 211 10·2) 10-19 10·71 10 191 March 10'16 10·11 10,52 10 182 10· 11 10,04 10·33 10 161 10,06 9 '98 10·20 10 161 April 10'10 9·99 10-14 10 161 10·08 9·99 10'14 161 10-04 9·95 10,12 •8 IS. SUMMARY ANALYSIS OF THE YEAR 1942-43 (iI) CUMULATIVE OBSERVED AND COMPUTED LOSSES DlSCIiAROr.s II I CoMPl1TED L OSSES Te o-Da y Cumulative I uced Period Observed Malakal Renk Lou --Tr-ou-gh- -I,- - Net I Mean .I: I Jnftow Evaporation Abso rption Total I Volume I May ... j 411 410 1 12 0 I 0 12 - · 451 446 6 50 2 0 52 -613 562 51 109 0 0 I 109 I -J uno ... 604 58 1 1 151 9 1 10 110 - 639 605 114 181 14 I 20 221 101 6.6 658 152 232 22 40 294 142 July ... 142 121 113 214 13 60 )41 114 · 110 758 185 311 1 .0 408 m 810 863 192 359 0 120 .79 281 Au&ust ... 823 828 181 414 -- 4 160 560 J1) · IS9 861 185 411 - 8 180 643 458 1,005 916 214 552 4 1 210 121 501 September •.. 915 921 268 621 -- 64 240 804 536 988 950 306 680 - 64 280 i 896 590 996 962 340 111 64 320 973 I 63J October 1,020 911 38. 141 - SO 360 1,057 i 668 · 1,020 ' 97 1 438 155 - 36 )90 1,109 , 671 1,11 0 1,071 411 149 - , 4 10 1,154 611 November ... 1,000 915 S02 142 + 30 420 1, 192 690 I ,~ 915 >21 139 12 430 1,241 114 916 543 134 121 430 1.285 142 December '" 91. 916 543 110 116 430 1,316 m 918 949 512 611 236 430 1,283 111 831 1.001 348 45. 29. 430 1,182 834 January ... 600 193 155 285 332 430 1,047 8.2 520 638 31 18. 359 4)0 918 941 523 584 - 24 151 )85 430 966 990 February ... .. 460 515 - 19 128 408 430 966 1,000S 442 498 -J3S 101 430 4)0 96 1 1,096 342 380 -113 16 446 430 952 1, 125 March ... 4" 460 -206 '6 464 430 9SO 1.156 · 434 417 -249 J9 48 1 430 950 1,199 483 506 - 272 32 491 4)0 .5. 1,231 'April ... 454 446 - 264 30 513 430 913 1,23 7 434 445 -215 28 526 430 984 1,259 426 449 - 298 " 5J9 430 996 1.294 897 TABLE 416 SUMMARY ANALYSIS OF THE YEAR 1943-44 i) THE DATA I GAUOU I lWNPALL I Assumed ToW Tp~;i~~y !-~~I':;-T --;.;-,;:.~ .-;. .; - Mal.k.1 I Swfaco M,lu! R,nk Eva~~tioD Area :::~, ~l-~;-l- ;;;'i-- ---I1 ---+----+----+-- -May 10,10 9-97 '\1 10' 12 12 160 10'20 JO-04 10-)4 32 4 '! 163 10'2S 10' 14 10·23 2 18 20 168 June 10-)3 10-)5 10-22 3 I 3 :g;; \I 176 ... I 19~1 :g~~ I~l 1 16 17 201 \I 16 232 July ... I 111~ :g~~ :g~l s1 I 8 3 249 27 16 269 AUJust ....\ :: : ~~ :~! :~!~ ;~ i 22 107 294 : : 93 9 322 U:;; n:~ n:ll n 62 96 378 . 1 38 23 ! 443 September .. 11 ·90 11·70 II -54 13 48 12 480 11 -96 11 -76 11 ·65 SO 87 lS SIO 12-04 11 ·84 11-75 23 33 2 549 October 12-08 11 ·89 11 ·80 16 i 51 7 S74 12'10 11-91 11 ·82 10 ",·.'I 32 27 S87 1l.J3 11-93 11 ·86 2 8 606 November 12' 16 11'96 11 ·90 S 622 I 12')6 11 '91 11 -92 6 626 12'17 11·97 11-94 7 633 December .. . 1 12, )4 11 -97 J ) ,96 8 632 11 '94 11 ·89 11 ·95 9 S90 January .... 1 :~::~ :: :b~ :: :!~ 10 444 10 32S I 10·60 10· 76 11 ·44 10 277 • 10-41 )0-43 11-3.5 10 247 February ....... 1 10-)7 10- 10 237 10·28 10' 2lS7 1111·-2098 10 222 10-17 10-11 10-84 10 199 March 10-09 10·04 10-63 10 \8l .; Ig:~; ;:n :g:~ 10 16s 10 IS8 April 10 ' )4 9-98 10-15 10 161 10'17 10-08 10-23 9 166 10-00 9 '94 10-)4 8 IS9 SUMMARY ANALYSIS OF THE YEAR 1943-44 (ii) CUMULATIVE OBSERVED AND COMPUTED LOSSES Ten-Day ~ DIlCHAIt.CES I Cumulative ! Co/rofPU'T'ED LoSSES ~ Deduced Period !--~k;-T~=-- O~~ed 11:-T-'-OU-g-h '7[- Net I Mean I Total Inflow Volume I Evaporation I Absorption May 461 4S6 S 492 463 34 ! 6 10 0 I' 16 I '1 S58 553 39 JUDe S24 499 64 li !~ H 1 ~ !: 'I: S82 S37 109 90 39 20 149 624 S93 I I 40 I3I 39 30 20 0 - July .. . 6S8 621 177 170 i 44 so 264 - 708 I 660 ; 22S 21S 44 I 70 i 329 104 . ! 841 AUKUSt 796 I 778 288 263 7lS ! ".i 329 I 316 i I 38 I 90 I 391 103 32 110 4S8 129 844 804 369 383 9 \ \30 S22 \S3 9S0 i 933 ! 386 424 I 0 ISO i 51. 188 September .. I 890 864 ! 412 474 0 180 654 242 i 916 890 438 SI7 -20 I 210 I 701 269 942 i 907 I 473 553 -20 240 773 300 October ... ! 9SO ! 9 16 S07 S70 I -26 270 I 814 307 9S0 92' m 588 -20 300 " 1,055 I 1,011 S71 607 + 6 320 I 868 33S 933 362 November .. i 968 I 924 968 924 I 61S I 620 37 340 997 382 6S9 626 I 15 I 360 1,061 402 : 968 924 I 703 628 119 370 1,117 414 D ecember ... 933 924 I 712 603 169 380 1,152 440 I 844 890 666 S04 222 380 1, 106 440 156 909 SI3 lS3 270 380 1,003 490 January 510 721 I 362 I 238 302 I 380 920 "8 , S06 6\3 lSS I7l 330 380 883 628 S27 S83 199 \36 351 380 873 674 February 472 SI7 IS4 III 381 380 874 720 4S4 484 124 80 403 380 863 739 ' I 394 428 90 S8 421 380 8S9 769 Marc:h 'J 430 449 71 24 439 380 843 772 427 446 S2 - 2 455 380 833 781 463 474 41 - 2 471 380 849 808 April . ! 448 437 S2 + 1\ 487 380 878 826 4" 470 33 - 7 S02 380 889 8S6 I 426 I 460 - 1 - 6 I SID 380 884 883 Nons. The J)U1ocb tIlUked are 001 ten days.. All fiaures u. Ul mi.I.1iom. Tho appat'CDt ablOrptIoD 11 11001 abown beca ...... ,t it Devor poutivc. Deduud iaBow is sbowu only when itJ cumulatiYe value exceeds 100 milliona. 898 TABLE 417 SUMMARY ANALYSIS OF THE YEAR 1944-45 (i) THE DATA GAUOES RAINFALL Ten-Day Tota l Period Ma\akal I Melut I Rcnk Mnlakal I Surratt Melu l R• .m k Are" -~-~r-yl1-(3-l-+-:~~ :::~ -I:~:~ : 1 !I .1 11 ··-1 --,-1 - " T-- r ill 10'51 10-22 10·23 IO'7S 10·53 10·49 32 : 1une 10-89 10-66 10 '58 39 1IH1·2O2 1100··9874 '1 11006890 1 1142 T 1 3! I g~ July 11-)4 11-09 1088 6 32 ! 72 i 281 11'48 11 '22 1100 116 11·64 11·39 II IS 4 ;1 ! ~~ U; August 11'70 11 '49 11)0 67 14 39 ,2 ) 85 11·77 11·56 I 1139 22 87 27 2 4 J9 IJ.86 tl ·64 11 '47 98 6 38 2 451 September _. 11·99 11·75 11 ·56 28 112 n 2 ~o:s 12·03 11·84 1l ·70 47 22 20 2 544 12 '10 11'90 11'79 52 )1 5 2 579 October 12·19 11·98 Ii'SS 14 27 17 2 620 12'22 ]2·03 11'91 )0 9 1 53 2 644 12·24 12·07 11 '95 5 2 665 November .. n '22 12·06 11 ·98 665 12'18 12·04 12'00 662 12-14 12-01 11 ·99 650 December , _ 12·02 11 ·94 11 '96 9 616 , 11 ·72 11'76 11 '9 1 10 517 IHO 11-39 11 '78 10 .\97 January ". ! 10·98 IJ 'O) 11·61 10 32 ) 10·68 10-74 11 '49 10 2S0 "'.'I1 10-43 10-46 11-3] 10 248 February 10·30 10·)2 11 '25 10 229 10 '23 10·22 11·04 10 214 10·17 10'14 10·87 10 202 March 10,08 10,05 10·63 10 18) o! 9-96 9'94 10'4] 10 164 9·83 9'80 10-17 10 155 April 9·71 9·68 9·97 10 149 9·63 9·61 9·91 9 146 9·51 9-51 9·84 8 14 \ SUMMARY ANALYSIS OF THE YEAR 1944-45 (til CUMULATIVE OBSERVED AND COMPUTED LOSSES I DISCHARCES II Cumulative .1' _._ -,--_CoM_YV_T'ED, -L o_ssE_S - ,-___I_ __D E,D U-C_EO Tp~;PodY i __ I : Malakal I R nk ~ Ob(~~ed ! Trough 1 Net I' Mean 'I Apparent 1 e , Volume Evaporation Absorption Total Absorption Inflow ~-.I- ~~1-1 672 *i -11· ;H -T-~!I-!-!f i ;! ! ;~! T-~:! ' June ". j m I 611 294 189 49 40 278 56 712 .53 1 353 220 57 60 337 76 July ... 742 679 416 254 45 80 379 117 - 179 711 418 301 .2.6 100 427 151 I -903 831 544 345 19 120 I 48' 180 I -~u~t 840 193 59 ) 383 4 140 m 204 -858 809 640 4 18 1 _ 160 I 574 226 - 969 912 697 463 1 -12 180 631 246 - Septembe 924 , 848 713 515 -32 2\0 .93 290 -933 881 825 556 ·-31 240 159 306 - 959 901 m 600 1 -43 270 827 ; 320 - October 9., "4 938 636 I -43 300 893 345 -994 942 990 661 -69 330 922 398 - : 1,102 1,055 - 1,031 619 -55 350 974 4Il -Nov"". . l 916 968 ! 1,045 672 -28 380 1,024 401 - 950 976 1,019 663 +18 400 1,08 1 - : - 924 950 993 63. 70 400 j 1,104 - III D=mb~ 881 907 961 567 I 126 400 1,093 - \26 183 813 817 449 178 400 1,027 , - 150 .I 138 861 148 332 222 400 9,. - 206 Januuy 601 616 614 247 I 1$4 400 901 i - 221 531 624 581 184 282 400 866 - 219 532 582 531 121 309 400 836 - 299 FOb,ua'y·1 463 491 503 109 332 400 841 - l38 45J .68 488 83 I 3S3 400 836 - 348 .157 369 416 58 369 400 821 - 15I 431 448 465 30 381 400 8\1 - J52 ~ j 41\ 4\6 460 3 403 400 806 - 346 418 426 452 -21 419 400 798 - 346 A,priI .. , 366 362 456 -37 434 400 791 - 341 "9 "2 463 - 49 448 400 199 - 33. 346 331 472 -59 I 459 400 800 - l28 Nons. n.~ marked are nOl ten days. All ficurcs are III millioQS. Tho apparenl absorptioo 1J shown only up to the pc.ak of tho nood aDd tile deduced iDflow il shotI'D oDI,. wbcnlu eumulative value olreeeds tOO milliolU. 899 TABLE 418 SUMMARY ANALYSIS OF THE YEAR 1945-46 (0 THE DATA Ten.Day !-_ ._ GAtJ~"";- _ ___. _ RAINFALL I Assumed Total Oro" Surfaco Period ! Malakal - Melut - RenJc Malakal I Mclut Renk Evaporation 1 : I Area 1 I April (3) ... 9·5 1 9·51 9·97 - - - - - May 9·44 i 9·48 9·78 1 7II. 20 - 7 144 · 9-50 , 9-47 9·76 69 12 10·17 9-90 10·03 1 ! 6 17 -June 10·50 10·3 1 10-34 i 73 7. 104 I 6 144 5 157 194 10-63 10-45 10·44 12 30 126 •• 211 10-83 10·62 10·54 10 4 8 4 232 July 10-99 I 10-77 I 10·66 54 - 13 3 249 · 11-14 I 10-93 I 10·80 8 41 160 3 268 11 -27 11 -08 10·9 1 41 57 - 3 287 August 1) ·42 I 11-23 ! 11 -02 90 I 214 78 2 309 11·60 : 11 '41 11 -20 ! 36 17 106 2 357 11 -72 11 ·53 11 -)9 54 75 145 2 407 S~ptember .. 11 -79 ; 11 ·6 1 II -52 35 I JS 85 11 -96 11'74 t 11 ·64 33 18 2 2 447 500 12- 13 11 ·9 1 11 -80 lIS 53 I - 2 589 October ... 12'25 i I 12')4 I 12-06 11 -92 6 56 21 2 654 12-12 I 12-00 116 2 38 2 696 12-37 12 '18 12-05 16 - - 2 729 November _. 12·37 ; 12·2 1 12-09 J8 I - 5 3 741 12-)6 12-22 I 12' 10 ! - - - 4 745 12-35 i 12-21 12-11 - - - 5 743 December . 12·34 i 12-20 12-12 - - - 6 741 · 12'32 12-19 I 12-12 I - - - 7 738 12'28 12' 16 12 ·13 - - - 9 723 January 12 '12 12·09 12-11 I - - - 10 690 · 11 -63 1 11 -82 i 12-04 I - - - 10 547 10-94 11 -20 11 -77 I - - - 10 359 February 10-49 1 10 ·62 11'49 - -- - 10 268 · 10'27 10·)2 11 -26 - 1 - 10 231 10-15 i 10-18 ! 11-18 I - - - I 10 217 March 10{)2 JO·O 4 10-92 - - - 1 10 19 8 .1 9-87 10-6) 10 170 9-73 )0-36 10 153 April .. . [ 9-61 10-09 8 150 9-56 9·90 "2 8 146 9-52 9-78 49 2S 7 I3S May(l) 9 '49 9-77 3 24 1 6 I3S Noru : PCTiods tNlrkcd • a~ DOl ten day,_ Gauses are in llIelreS, rainfall totllis in mil.limctrts per tcn-day period, evaporation in rnilllmelres per day. and the ,wf.lt.e Olea is i.e square k.ilometres, SUMMARY ANALYSIS OF THE YEAR 1945-46 (ii) CUMULATIVE OBSERVED AND COMPUTED LOSSES D ISCHARGES 1 C 1 ti I CoMPtJTl!D Lossu TPe~i~day .,----.---•• -! ~~~~,ede 1----·-- ,·-,--- ---II A'b~~::,ti~~ I MaiakaJ ! R~nk , Loss Trough j Nct _ I Mea n, I Total . ' , ~ Volume ! Eva pora l,on : AbsorplJon ; ~.--3J-4-·--3-25- --, 9 l - 4 8 '--0--1 --4--' --5 - ;~~ Ui 11~ + ~~ I~ g I 15~ Ii Junc 559 .5 19 156 13) 7 10 150 16 589 550 195 163 ) 20 186 29 658 589 264 199 10 ~O 2)9 55 J uly 689 62 1 332 236 12 50 298 84 733 662 40] 284 - 7 70 347 126 8]4 770 467 312 - 7 90 395 162 August 784 735 516 360 - 40 120 440 196 .', 8] 1 778 569 418 - 54 150 514 205 I 950 904 615 468 - 89 J90 569 236 Septcmber .- 890 847 658 519 - 102 210 627 241 933 864 727 602 - 102 250 750 227 985 890 822 686 - 125 290 851 261 October 1.020 924 918 751 -132 330 949 299 1.054 950 1,022 797 - 15) 370 1,014 378 November·; : :~~ I ,~:: ~:~i! !;~ =: ;~ :i~ !:~~~ ~: ,. I 1,054 976 1,)01 834 - 101 450 1,183 568 1,037 976 1.)62 834 - 64 470 1,240 592 Decembcr 1,028 976 1,414 828 - 20 480 1,288 606 .j 1,020 976 1,458 819 oJ. 32 480 1,33 1 6()7 ,.3 1 .~ 1~~ m ~ m I~a ~ January ,., : 901 950 1,444 682 172 480 1,334 746 881 1,309 506 227 480 1.2 1) 641 893 1.057 3 16 267 480 1,063 Februa", . ..'1 499 642 914 2 19 294 480 993 .1 461 517 858 174 317 480 971 352 )92 818 146 335 480 961 March .. j 420 45 1 787 99 355 480 934 394 417 764 64 372 480 916 408 424 148 38 387 480 905 April · 1 351 344 761 19 399 480 898 352 )34 779 6 41 I 480 897 351 )37 793 0 418 480 898 May (-I) 354 342 426 480 910 900 TABLE 419 SUMMARY OF SIXTEEN FLOODS BETWEEN MALAKAL AND RENK DURING THE YEARS 1928 TO 1946 First Year Period Period I Tota l F ' "AL COMPUTED Loss" Ic umul'- 1 Cumula- of Flood Aood Flood Began Ended t.:'al3kal Tr~u-;-I N.~- -I M.;an -j-.c-on; U'''J Ob~;~Cd I D.C~~:~cd DIs..::harge Volu me EVI:lPOr.i- Ab~orp- T~:J I " LOSS ', InHow I (Max.) l.Ion tlon 1928 April (2) i April (1) 27,900 I ::~ ::: 1 ::- :~~ - i' ~ ::; -- I ~~O 1929 April (3) April (1) 28,268 1930 April (2) : May (1) 25,045 424 330 240 529 57 1 Nil j 1931 May (1) i May (I) 26,108 645 , 413 I 400 863 (287) (576) 1932 May (I) ! May (2) I 32,748 1,062 I 835 620 1,507 845 662 1933 1934 NO T ANALYSED 1935 1936 May (1) 1 April (3) ! 25, 163 57J 384 340 701 55 1 150 1937 May (I) April (2) ! 26,261 746 483 440 950 - II I 1,061 1938 April (2) April (I) 29,373 859 718 S20 1,27 1 - 20 1 1,472 1939 May (1) May (1) 27,600 586 448 360 747 i 100 647 1940 May (3) April (3) 21,866 500 448 1941 May (2) April (3) 25,250 660 519 1942 May (1) April (3) 25,860 755 539 ::: i :: I =~ :: , :: 1943 May (1) April (3) 24,746 628 510 380 I 884 I I 863 1944 May (1) April (3) 25,220 679 459 400 ' 800 472 328 1945 May (1) May (I) 26'447 836 426 480 h910 I 805 I ~~- ~ May (I) May (1) ~ 1,288 ~,~ I~I~_~ Sums:. - - , 430,592 1I,599 ! 8, 159 I 6,770 , 15,034 I 3,199 i 11 ,691 1 MEANs : ... May (I.) April (3) 1~I~ I--5-10--~:-~l~ :-;;-f.l.) I I I NOTU : All 6a:\ltII e""pl !bose in lbe first three wlull1fIS are in aW.li.OCJ. In 1911 tho ~ ill the I.ast two columns IlI"C iII brac:kc:u bccau.se they are IIppro)fimate owina to bd. or measured di5(hultcs u Rcnk iDthis)'Ut. No deduced in1low is 'hown when ill tiJal value wu leu tb.a.D 100 miWotu or. in 1945, 1t\.1.n it:!: probable error. 901 SECTION VIII. INFLOW AND THE EVIDENCE FOR IT In Table 420 (p. 905) are shown the values attained by deduced inflow in individual months during the floods analysed in this paper. These are monthly totals, and they have been derived from Tables 403 and 40~ 18 by comparison of the cumulative totals of deduced inflow at the end of each month. The arbitrary lower limit of 40 m/ d has been taken so that minor fluctua- tions or errors in the discharges may be eliminated as far as possible. Even with this limit there are some months-e.g. October 1936-when it may be doubted whether inflow has really occurred. In a few months there was a negative result with a value over 40 millions; these have been indicated in the table though the amounts are not shown. As will be seen one at least of these is confirmed by independent testimony, since when the river is really high it seems that the storage power of the Adar increases considerably and there may be a fair amount offlow back into it. In Table 420 averages have been derived for each month, and these repeat the tendency, which is well marked in the years of biggest inflow, towards a double peak in its values, the first peak occurring in August or September, and the second in January or February. The first peak is easily explained, for it corresponds to the peak of the rains; and there can be little doubt that it is due to the inflow brought by the steeper khors from a comparatively short distance away. The second peak is less easy to account for, and I discuss in the next section the most probable sources of supply at this time, particularly that of spill from the Baro travelling through the Machar Marshes. Before proceeding to discuss the inflow any further it will be as well to consider the evidence which, in my opinion, confirms beyond all doubt that inflow does exist, if the previous analysis by which it was deduced has not been sufficiently convincing. MEASURED DISCHARGES OF TRIBUTARIES The evidence for inflow takes two forms: actual measurements of the discharge of tributary khors, and observations of their flow and its effects which have been recorded in the province or district monthly diaries. Neither of these sources is anything like comprehensive, and reliable confirmation of deduced inflow goes only so far as agreement with the values deduced above in exceptional years. It must be remembered that the tributary khors, as was explained above in Section III (p. 864), take in a little water every year and pass it out again when the river drops. Thus they have water in them every year and it is only when this is higher than normal and flow- ing much faster tban normal that it is likely to be observed or commented on specially, because it has caused damage. Actual measured discharges are very few. In 1930 a few measurements were made at the mouth of the Khor Adar, which joins the White Nile opposite a point upstream of Melut. Between September 21st and December 10th in that year the maximum flow into the Nile which was recorded in the Adar was under half a million a day; during some of the period of measurement the flow was the other way. Lately the discharges of this khor have been under constant observation, and they have been measured when they were at all large. They are described in more detail in the next section, but it may be noted here that the maximum ever recorded was 3 millions a day, in October 1947, and that in 1948 the maximum was under half a million a day, in August. In general the flow in this khor seems to coincide with the first peak of inflow mentioned above, and not to amount as a rule to more than a million a day. The only other measurement of tributary discharge which I have been able to trace was in the Khor Doleib, which enters the White Nile just above Renk; it is a very significant one. The Egyptian Irrigation Department recorded of this khor in 1938:w .. The kbor began to rise on August 15th, reached its maximum from 29th September to 4th October, and then feU away fast. On September 30th a discharge of 108 m'/ sec. (9 '3 millions a day) was recorded." Comparison with Table 420 will show that in this year the deduced inflow began in July and reached its maximum in September and October, with an average daily inflow of nearly 10 m/ d. As will be seen below, the Khor Doleib was not the only source of inflow at this time, so the correlation is quite good. OBSERVED INFLOW This completes the very meagre records of actual measured inflow, and we are left with the much more unsatisfactory evidence of administrative records ; unsatisfactory because they must of necessity be purely qualitative and even accidental, and not even reliably objective. Administrative officers are changed frequently and each new man has a different standard ; he must therefore rely at first entirely on local reports for the comparative size of what he sees or hears about. The available evidence is contained mainly in province and district diaries from 902 which the extracts given below are quoted . A set of the Upper Nile Province diaries is held in the Civil Secretary's Office, and I went through these myself. But they only go hack as far as 1937. and Mr. D. F. Ferguson of the Jonglei Investigation Tea m was kind enough to sea rch through the Upper Nile archives, in which he found some va luab le ev idence from district diaries. Unfortunately so far no trace has- been discovered of the records of 1929- 30, which seems to have been an interest ing year. since a large amount of inflow coincided with a low White Nile-or rather a not exceptionally high one-and also with a record high Baro. An aUempt was made by the province staff to locate these records but without success. From a study of Tables 419 and 420 and Fig. J 10 it will be clear that four fl oods stand out as pre-eminent as regards inflow: 1929-30. 1938- 39, 1942-43, and 1946-47. From the pre- liminary analysis which I have made of the last flood - I 948-49- between Malakal and Melut it appears that in this also there was an unusual amount of inflow, with maxima in August and September, and again in December. There was a fair amount of inflow also in 1937- 38 and in 1941-42, amounting to just over a miUiard, but I propose to confine myself here to a study . of the four floods mentioned above, and to discuss them individually below. 1929-30 FLOOD Tbe deduced inflow in this flood totalled over 1,200 millions, witb two maxima, in August and September 1929 and in January 1930. Up to the time of writing no evidence of inflow has been obtained, as described above, but it is worth mentionin g that this year coincided with a record high Baro, and that the amount of water spilt from this into the Machar Marshes, as described fully below in the next section, was over six milliards, a milliard more than that recorded even in 1946. Tbus the records of this year may be of considerable value in deciding the difficult question of what proportion of the second peak of inflow comes from this source. Since the White Nile was not particularly higb in this year-the maximum Malakal gauge was 12'36-there would be more chance of distinguishing between flow from these marshes and general Nile flooding than tliere was in 1946. 1938-39 FLOOD In this tlood the deduced inflow was nearly 1,500 millions with maxima in September and December 1938. The measurement of the Khor Doleib discharge mentioned above is confirmed by the following extracts, which also reveal that the khors on the west bank flo wed unusually hard that year. The first extract is from the Renk District diary, and the others from the province diary. NOVEMBER 1938 . .. An embankment was erected at Khor Mario as well as at Khor Doleib, but un- fortunately as the Khor Doleib tlood was unusual this year the embankment was washed away." NOVEMBER 1938 (Province Diary) . .. Of all the causeways between Kaka and Malakal there remain only those at Abaraoie, Thuor, and Famad. Three years' work has gone in as many months." DECEMBER 1938 . .. The opening of the motor roads bas been considerably delayed by the very bigh Nile and tlooded rivers tlowing from inland to the Nile. The most serious hold-up was the breaking of the causeway just north of Kodok which was cut in two places to let the tlood-water escape which was coming from inland. These breaches are 5 and 15 metres wide. On 30th November, in spite of the high Nile, the water was still flowing hard from inland through these breaches. Usually the road is open as far as Melut in mid-November. .. At the beginning of October the Khor Doleib was still tlowing ; it dried in the beginning of November and came down in spate in mid-November." So big was this flow in the west bank khors that a special journey was made by the A.D.C. Kodok and Mr. H. Ben (at that time a member of the Egyptian Irrigation Department) to check on a Shilluk legend tbat this water came from the White Nile above Malakal, and that it was in fact a spill from near Tonga. They found that none of the tributary khors near Tonga went back any great distance and also that all the streams between here and the stretch 903 north of Malakal were flowing southwards, and not northwards as this theory would require. They concluded that the water must come from the Nuba Mountains, though it must be re- marked that neither on the existing maps nor on the recent American air photographs can any such connection be traced. 1942-43 FLOOD In this year deduced inflow reached a total of nearly 1,300 millions, with the usual two maxima, one in August 1942, and the other in January and February 1943. The province monthly diary for October 1942 recorded: " NORTHERN DISTRlcr, SHILLUK SUD·DISTRIcr, W. BANK " Heavy rains in the hinterland brought the watercourses down in a similar spate to that of 1938, when all the ramped culverts were destroyed. The brick bridges built after that have apparently been unable to cope with the volume of water and the earthworks at Akuinkual, Famad, Fadit, and Thuro have been washed away. " In October the spate subsided and repairs to bridged watercourses could be made sufficiently to allow of the passage of cars. The water in each place cut a channel for itself through the ramp of an average of20 metres." There is no mention in the diaries of an unusual inflow in the follo wing January and February, either from this source or from any other. 1946-47 FLOOD In this flood the deduced inflow reached the highest total and also-in February 1947- the highest monthly total during the period under review. The yearly total was over 1,700 millions and there were two well·marked peaks, with an actual negative value between them, in September 1946 and February and March 1947. The confi rmatory evidence is considerable, and it has already been mentioned in the Second Jnterim Report (pp. 93-94). The following extracts from the Nortbern District monthly diaries, collected by Mr. Ferguson, do something to expand that account. OcrODER 1946 "The timbers of the P.W. D. bridge over the Khor Adar have apparently been swept into the Nile. " JANUARY 1947 " Lorries can now reach Paloich from Melut and Gelhak from Renk, but the khors to the south and east of Paloich rose during tbe month with overflows from the river and Adar." This sounds more like simple flooding from the high river than true inllow, especially as it follows the negative inflow wbich may be seen from Table 420 to have occurred in December 1946. FEBRUARY 1947 "The six or seven khors between Paloich and Leweng were found by the Deputy. Governor wherl he walked this stretch to contain twice as much water as in the previous month . The Khor Lool (7 Lui) which in December was:f mile wide, became some 6 miles wide by the middle of February and was still so at the end of the month." Melut gauge had by then dropped to 12·0 from a maximum of 12·74 which it held throughout December and January; this effect must therefore have been due to inflow, and it confirms the record total of over 450 millions for this month shown in Table 420. MARCH 1947 " The khors in the Paloich area are still rising at the end of the month. From where the water is coming is a mystery." " The three miles of road which was under water from flooding by the Khor Wol on the south was dry by the end of the month, but water to a depth of two feet was still flowing over the bridge and there was a mile of water to the north of it. " (Continuation of same diary.) 904 ApRIL 1947 .. The people compare this year to the great floods in the time of the Turkish Govern- ment (1878) and affirm that these present conditions will prevail throughout next year." "The Khor Wol bridge appeared above water for the first time since Christmas." .. Work was begun on the ramps and bridges. many of which have been washed away completely, on the Tonga-Kodok road." As already explained I have not included analyses of the floods of 1947-48 and 1948-49 in this paper because records were not made of the discbarges at Renk aner the middle of 1947 ; but it may be worth noting here that the summer of 1947 saw a continuation of these fl oods on the east side of the river, and it was not till early 1948 that the water there really began to dry out. From these extracts it is clear that the last three exceptional floods-1938-39. 1942-43, and I 946-47-have tbe inflow which I deduced in their analyses well supported by independent testimony; and it is to be hoped that similar evidence will be found in due course for the first of these exceptional floods-in 1929- 30. For the remaining floods when deduced inflow was under a milliard tbe evidence is unsatisfactory and fragmentary ; and I do not propose to quote from it. It is clear that only exceptional inflow, particularly when it was in the form of a violent spate wbich caused damage, is likely to merit a special reference in the monthly diaries . Many of tbe tributary khors are bridged, and there is no reason why these should not take a fair amount of flow. This may be shown by quoting the Khor Nyadwai, near Kodok, of which the cross-section bas already been described on page 864. In an average year tbis will have at maximum level a cross-section of about 40 square metres, and consequently if flowing at an average rate of half a metre per second would have a discharge of20 cubic metres per second or nearly two millions a day. It is dangerous to generalize in this way, but this is sufficient to show that a few khors flowing in this manner could easily account for the monthly totals of inflow shown for 1941 in Table 420, and it is reasonable to suppose thaI flow of this sort of size might not cause damage or be remarkable enough to merit special mention in a district diary. Thus I do not think it is at all difficult to believe that inflow of the order of half a milliard or so could take place during the rainy season without causing remark ; what cannot be denied is that the losses on this reach, as pointed out in Volume I of The Nile Basin , are very much less on the average than they should be by the most conservative estimates. It is less easy to feel sure that the deduced inflow shown in Table 420 for the early months of several years, when from all accounts the tributary khors are normally dry, has really occurred except in the obviously unusual years such as 1947. I have therefore devoted the next section entirely to a discussion of this problem, and have attempted to show that at any rate it does not seem to be due to overflow from the Machar Marshes caused by Baro spill, although this was gi ven as the most likely explanation in the Third Imerim Report , in the passage quoted above. (But see Chapter 4, pp. 976-7.) TABLE 420t DEDUCED INFLOW INTO THE WHITE NILE .BETWEEN MALAKAL AND RENK IN THE MONTHS WHEN IT EXCEEDED 40 MILLIONS BETWEEN 1928 AND 1947 ___Fl_ O_od_ _. LM_aY~_J_un_e .l July Aug. I Sept. I Oct. I NOV. , Dec. Jan. Feb. !MarCh ! A pril 28-29 29-30 I I ~ l In I I~~ 1-6~ I -;~ I - I 1 136 220 1J4 I 133 I 98 30-31 Insufficient data but probably small or nil 31-32 32-33 I = I = I = I 101 -1 ~: I ~: i ~: i 33-36 Insufficient Data 3&-37 I 53 , 37- 38 120 182 168 48 ' 216 51 47 40 67 38-39 I 57 212 273 241 109 21 6 169 63 53 39-40 46 92 ! 48 142 57 46 71 75 51 40-41 54 147 41 53 73 84 56 68 41-42 138 III 80 91 122 150 126 62 41 96 42-43 51 90 145 220 126 44 65 92 156 135 106 63 43-44 43 85 112 62 52 , 76 184 95 75 44-45 1741 ; 95 93 52 45-46- 46-47 142 311 217 169 ! - v~ 98 454 378 I MEAN ... ···1 13 39 93 97 59 64 1 72 89 82 66 31 905 SECTION IX. DRY SEASON INFLOW AND POSSIBLE SOURCES The second peak of inflow shown in many years in Table 420, coming usually in December or January, averages some two millions a day and in more than half the floods analysed appears to continue until May. This wo uld appear to mean that as often as not one or more of the tributary khors between Malakal and Renk must be flowing at a time when, according to local information , they are normally dry-except of course in an unusual year such as 1947. It may be worth mentioning that the time when this phenomenon seems to have been most frequent was during the years 1939-43, when the adminis trative staff was reduced owing to the wa r, and this may account partly for the lack of records of its having been observed . But the questions of whether this delayed inflow does occur, and if so where it comes from, are of some imponance. In the first place it is the onJy logical consequence of the preceding analyses which appears to run counter to the general evidence; and in the second this inflow is deduced as occurring in the timely season when it is of considerable practical interest. The inflow which occurs during the peaks both of the rains and the flood is unlikely to be denied and is unim- portant ; inflow coming in the dry months of January to April is neither. There are three possible causes for this deduced inflow : errors in the computed or observed losses; spill from the Baro or Sobat which has travelled thro ugh the marshes east of the Nile; o r rain falling on these or other areas at a distance from the river and taking several months to reach it. Little can really be said abo ut this last owing to the lack of rain-gauges in the a reas concerned. The first of these possibl.ities may now be considered. ERRORS IN THE OBSERVED OR COMPUTED LOSSES (see also the last part of Section II , p. 859) At this time of the year absorption may be assumed to have ceased and rainfall has certainly done so; the only items from wh ich the deduced inflow is derived are therefore the observed losses, trough vo lume, and evaporation. In the sections dealing with these I have made esti- mates of the probable eirors which they have in an average year: in the first two they can be derived fairly directly, a lthough this does not necessa rily mean that they are correct. In Section V (p. 870) I show that they would appear to be too small to account for the low evapora- tion depths reached in two o ut of the six years considered in Table 401 (p. 871), and I therefore postulate inflow in those years. These were borderline cases, but the same arguments may be applied with more force to the years which were excluded from that table because they wo uld have given actual negative values to the evaporation depth . For example in order to eliminate the deduced inflow in the early part of 1943, the three quantities from which it is derived would have to have a total error of 460 millions, whereas their probable errors are 130, 20, and about 40 millions, giving for the probable error of the deduced inflow a value of less than a third of this. I feel therefore that unless there is something seriously wrong with the observed discharges the only possible explanation is inflow. It is noticeable that it does behave much as one would expect if it were real and due to either of the two second causes given above: that is to say in general it follows years in which inflow during the rains has been above average; and it shows a stead ier rate than this inflow, whieh would be expected if it had come from a long distance, probably through swamps. Only in 1947 does it show the characteristics of a spate, and 1947 was an exceptio nal yea r. It is arguable that if it is due to errors and not to real causes it would be much more va riable and be unlikely to exhibit such a systematic tendency, not only in each year but in the several years in which it occurs. BARO SPILL- METHOD OF CALCULATIONU) The hydrology of the Baro has been discussed by A. D. Butcher in Sabat Hydraulics and by Dr. Mohammed Amin in ' River Baro Losses ' (unpublished), and also in the Second Interim Report (pp. 90-94). For the present purposes I have adopted the conclusion reached in the last named that when the monthly discharge at Gambela exceeds 1,500 millions the excess is spilt into the Machar Marshes, most of it being carried by the Khor Machar which leaves the Baro 57 km. upstream of its junction wi th the Sobat. This is true spill in the sense that it does not return to the Baro and apparently only a small proportion-if any-reaches the White Nile. A basic difficulty in assessing the contribution made by this spill to the White Nile is that it is difficult to distinguish its effects from those of rainfall in this eastern area. At first sight correlation between Baro spill and late inflow seems good, but it must be remembered that a large amount of spill on the Baro is due to a high flood and this is due in turn to heavy rains in its basin; these are likely to coincide with heavy rain farther north in the basins of the Yabus and Daga which are the likely alternative sources, or even with heavy rain actually on the marshes themselves. It is noticeable that even the Blue Nile flood shows a certain degree of correlation with this late inflow, though there is obviously no direct connection and the correspondence must be due to the chain of circumstances outlined above. 906 Because of this basic difficulty in distinguishing the sources of this inflow, a difficulty which is made almost insuperable by the lack of rainfall and discharge records in the areas concerned, I do not intend here to devote much space to discussing possible correlations, but will merely summarize the various factors and leave further discussion to others who are or will be better acquainted with the area. In Table 421 (below) are shown the va lues of Baro spill--<:alculated as described above-in the last twenty years, and r have also repea ted for conveniepce the totals of inflow for the following dry season so that such correlation as there is may be seen at a glance. I think that the years 1931, 1937, 1943, and 1944 a re indications that the Baro spill does not provide all of this inflow since in these years a low Baro fl ood was followed by an appreciable amount of dry season inflow, whereas the high floods of 1928 and 1945 were not fonowed by inflow. POSSIBLE ROUTES Here again local knowledge and some investigation of past records and memories may reveal evidence which at present is almost non-existent; and it would be premature at this stage to devote much space to a discussion of the routes by which this late inflow may have arrived. We may note only that the Egyptian engineers have never recorded a discharge of the Adar in these dry months, even in 1947, though I have been informed by Mr. Calder (then A.D.C. Renk) that in March of that year he saw this khor flowing strongly. In the preceding section it may be noted that the Khor Wol in 1947 showed signs of having flowed very strongly, and the bridge across it did not appear above water until April. No discharges of this khor have been measured, but it does not enter the Nile directly and may therefore easily have escaped observation since its waters would come into the river through the Khor Awilwil. This is one of the channels parallel to the main river and it would be emptying at this time naturally; inflow by this route would therefore lead only to an accentuation of this tendency. Neverthe- less there might be some chance of obtaining confirmation from the missionaries who were at Rom on the banks of this khor. Once again records from the early months of 1930 would be of value since the inflow in that year was exceptionally large but the White Nile had not been anything like as high as in 1946. On the whole it seems unlikely that this dry season inflow can come from the west side of the river since even in 1946 the khors on that side seem to have finished their flow before tbe end of the flood peak. But this is as far as one can go at the present, and I suggest that if the question is considered of sufficient importance, a tborough search must be made through all province and district records; and that the missionaries who live on tbe east side of tbe river should also be consulted. TABLE 421' BARO SPILL AND DRY SEASON INFLOW or Totallnftow in Year Number of Total Gambela Difference following Jan.- Flood Months SpiU Discharge I MoNnuthms ~Xr oIf,SO O I equals Spill April inclusive 1928 5 12,140 7,500 4,640 nil 1929 5 1l,630 7,500 6,1l0 545 1930 3 5,730 4,500 1,230 prob. nil 1931 4 8,960 6,000 2,960 266 1932 4 9,890 6,000 3,890 (321) 1933 4 9,550 6,000 3,550 , 1 no data 1934 4- 10,430 6,000 4.430 available 1935 5 11 ,900 7,500 4,400 1936 4 I 6,000 3,420 nil 1937 4 ~:~ I 6,000 1,990 Ii 167 1938 4 10,800 6,000 4,800 298 1939 5 11,600 1,500 4,100 253 1940 No records oWIng to the war 281 1941 11 ,860 1,500 4,360 )25 1942 8,590 4,500 4,090 460 1943 8,490 6,000 2,490 393 1944 6,190 4,500 2,290 122 1945 10,090 6,000 4,090 nil 1946 11,140 6,000 5,140 944 1941 9,960 6,000 3,960 not estimated 1948 10,130 6,000 4,730 (200) Non:s: l:: ~1!~.~~ li~:..~bcl :a. ~(er only to lin month. in which . IOllll of 1,500 millions was exceeded. All fl&\l,rd CX<;epl in lhe fin! Fi~ ue taken from TM Nil, Btuln, VoL IV and iu Supplements, or (rom unpublUhed rccordJ kindly supplied by the EJYPtian Itris:atlon ~!niniD~:;~:=is not based on continuou. d..ischarae fCICOr45, • RefczTed touT.ble II in Fie: . J 10. 907 APPENDIX THE EFFECTS OF AN ERROR OF ONE METRE IN THE LEVELS OF CRO~S-SECTION NO.4 ON THE TABLES OF FLOOD-PLAIN AND SURFACE AREAS AND OF VOLUMES ON THE WHITE NILE It has been established that the drawing of Cross-Section No.4, 80 km. downstream of Malakal, showed values for the levels which were one metre too great. Thus in effect the datum used for measuring widths on this cross-section was one metre too low, and in consequence the widths are all much too small. The effects may be divided into those concerning the tables of Appendix IV of the Third Interim Report of the Jonglei Investigation Team, and those concerning the tables in this analysis of the White Nile flood. EFFECTS ON FLOOD-PLAIN WIDTHS AND AREAS The old and new val ues, the resulting differences in the widths obtained at Cross-Section No. 4, and the effects of these on the mean flood-plain widths for the reach Malakal to Melut, are shown in Table 422. It will be observed that the average flood-plain widths for this reach, and so the corresponding flood-plain areas at almost all rises. should be increased by between 4 and 10 per cent., with an average of about 6 per cent. Since for most stages of the river the flood-plain area of this reach is only about one-third of the total from Malakal to Jebelein, the effect of the error on these totals (given in Table 28 of the appendix quoted) is only about 2 per cent., and this is only about a third of the estimated probable error. It may be noted, however, that the difference between predicted and ' observed' flood-plain areas between Malaka! and Melut. given in Table 33 of the supplement to that appendix, is now very much less, in fact on ly half its previous value. For the mean rise found there the new flood-plain width would be 1,520 m. instead of 1,430 m .. and tilis reduces the difference between mean predicted and mean observed flood-plain widths from 170 to 80 m., and from I I to under 6 per cent. EFFECTS ON SURFACE AREAS AND TROUGH VOLUMES The effects on total widths, and so on surface areas and volumes used in the White Nile flood analysis, are comparatively small if the whole reach from Malakal to Renk is being considered . From the last line of Table 422 it is clear that the effect of this error on the total mean widths between Malakal and Melut is very nearly constant for all rises of the river, the average being about 6 per cent. This wil l therefore be the proportional effect on the trough volumes of this reach also, both areas and vo lumes needing to be increased from the values previously given. But these quantities are only about two-fifths of the totals between Malakal and Renk, and the effect of this error on these totals is therefore only about 2 per cent., which is less than or equal to their estimated probable errors. Since this error has therefore comparatively minor effects on these tables of volumes and surface areas, and so on the analyses themselves, and since it was not discovered until after they had been finally computed and typed, the analyses have not been corrected; but it wiU be possible from this note to carry out at any time such corrections as may be required. The corrected versions of the area and volume tables have been given with the original versions. 908 TABLE 422 THE EFFBCTS ON F LOOD AND TOTAL SURFACE WIDTHS BETWEEN MALAKAL AND MELUT OF AN ERROR OF ONE METRE IN THE REDUCED LEVELS OF CROSS-SECTION No.4 Rise above Datum 0·0 I 0·2 I 0'4 I 0 ·6 I 0·8 1·0 I 1·2 I 1·4 I 1·6 I 1·8 I 2·0 2·2 I 2-4 2-6 I H I 3·0 I : (I) NeW Total Width at Section NO.4 ... ... 400 450 500 540 600 680 .790 850 1,080 1,220 1,460 1,620 1,800 1,920 I 2,620 I 2,800 (2) New FloodwPJain Width at Section No, 4 ... ... 0 50 I I 100 140 200 280 390 450 680 820 I 1,060 1,220 1,400 1,520 2,220 l 2,400 (3) Old Flood-Plaio Width at Section No.4 ... ... 0 0 I 10 30 60 120 170 220 260 320 400 510 570 800 940 1,180 (4) Difference ... .. . 0 I 50 90 110 140 160 220 i 230 420 500 i 660 710 830 720 1,280 , 1,220 (5) 1/8th of Difference ... 0 10 10 10 20 20 30 30 50 . 60 I 80 90 100 I 90 160 : 150 I (6) Old Mean Flood-PlaiD Width (Malaka! to Melut) ... 0 20 100 140 200 i I 260 330 410 630 I 900 1,250 I 1,450 1,750 2,000 2,250 2,550 i I (7) l18th Difference as per.c..e ntage of Old Mean Flood- I Plain Width ... ... ... 0 50 10 7 10 I 8 I I 9 7 8 7 6 ,I 6 i I 6 ! 4 I I 7 6 (8) Old Total Width at Section No.4 ... ... I 280 280 290 310 340 400 450 500 540 600 680 790 I 850 1,080 i 1,220 1,460 (9) Difference from New Total Width at Section No.4 120 170 210 230 260 280 I 340 350 540 620 780 I 830 950 840 I 1,400 1,340 (10) 11 8th Difference of Total Widths at Section No, 4 .. 10 I 20 30 I 30 30 40 40 I 40 70 80 I 100 100 120 100 i 180 170 (II) Old Mean Total Width ... ... I 400 420 500 540 600 660 I 730 810 1,030 1,300 i 1,650 1,850 2,150 2,400 2,650 2,950 (12) 1/ 8th Difference as Percentage of Old Mean Total ... ... I I I ! I I I Width ." 2 5 6 I 6 5 6 , 5 5 7 6 I 6 5 I 6 ,I 4 I 7 6 I : i I Nons : AU 118ucet 10 metre, clleept in lbe ICvetl.th and Ilul lines wbkh Ire ptrcentll,es. UDe ( I) is obUli.ned by measurements (rom lhe rorTectly plolled Cl'oas-scctioo: Line( J) from Tlblc 2<4 of Appcndu IV of lhe Third IliUM Rrpo,,; Unc 4 iJ (rom Table 26 of tM $Ilme p"b1icaUon. NOTES AND REFERENCES C') Verbal communication by Mr. E. S. WaUer. In the Introduction to Volume II of Th~ Nile Busin the error ofa sinale discharge was estimated to be S%. (I) Erron of this order were deduced from the ana lyses of fi ve floods below averalilc on the Sobal (see Chapler J , p. 933); but in above averacc Hoods systematic errors up to 5% were found. A compari50n of dist:har,cs above and below Khartoum also shows that errors up to S% rnosy occur, according to a private communication from Dr. H. E. Hurst. (') But see the last part of Section 11 (pp. 859--60). (') See also the lut part of Section II (pp. 859-®), CO) Private communication. But see Volume VIII of The Nile Busill, p. 77. (-) A more detailed account of this and of possible routes is ,iven later (see Chapler -I) . 911 CHAPTER 3. AN ANALYSrS OF THE SOBAT FLOOD by J. W. Wright. , M .A., F.R.I.C.S. INTRODUCTION This paper describes a detailed analysis of the flood cycle of the River Sobat, derived alm ost entirely from the gauge-readings and discharges observed by the Egyptian Irrigation Depart- ment in the Sudan. From these records, and those of rainfall and evaporati on kept by the Sudan Meteorological Service, an attempt has been made to ana lyse the fl ood and to estimate the average cross-section of the river valley by first assuming that it has a general form similar to that of the White Nile, then predicting the diJl'erences which should occur during the flood cycle between the discharges at the head and the mouth of the river, and finally comparing these with the differences actually observed. This cycle of operations has been gone through more than once, so that the final results represent a series of successive approximations. This may appear to invalidate to some extent the extremely close resemblance which was finally obtained between the average predicted and observed differences in the discbarges at the two ends of the river. Nevertheless it is hoped that in the main a more accurate picture has been drawn of tbe behaviour of the river in flood and of the shape of its trough than was avai lable hitberto; and that the results may be of value in estimating the effects of any control schemes wbich may be projected either for the Sobat itself, its tributaries, or the Machar Marshes to the north . The work was all done as a spare-time occupation and was made possible by a grant from the Leverhulme Trustees in 1950, followed by a second grant in 1952 when a reconsid eration of the basic assumptions neces,sitated recomputation of most of the tab les. Acknowledgment is also due to the Egyptian Irrigation Department in the Sudan, and the Sudan Meteorological Service for information supplied and for facilities to study original records, and to the Sudan Railways Accounts Department for calculating Tables 454-6 (pp. 96 1-2). MAIN FEATURES OF THE SOBAT For detailed accounts of the topography and hydrology of the Sobat and its tributaries, the Baro and Pibor, reference should be made to the relevant sections of The Nile Basin ; only a brief outline will be given here. The Sobat is the principal tributary of the White Nile, and is responsible for the greater part of the fluctuations in level and discbarge as far down as Jebelein. It contributes about 13 milliards annually, or just under half the total White Nile discharge. But whereas the White Nile discharge upstream of Sobat mouth varies only by a few per cent. throughout the year, the Sobat discharge varies from under 10 mi d in February and March to over 60 mi d during October, when it forms more than half the White Nile discharge below the junction. The Sobat is formed by the junction of two streams, the Baro and the Pibor. The first rises and ruDS almost entirely in Ethiopia, and the second, though it runs in the Sudan, gets most of its water also from Etltiopia. From this junction, which is called Sobat Head, to its mouth tbe Sobat is 348 km. long. The water surface has a mean slope of between three and four centimetres per kilometre, but tbe lower third is rather flatter than this, particularly during the eDd of the flood, owing to the backwater effect of the White Nile. Apart from the Baro tbe Sobat has no tributaries which run during the wbole year, and only two of the khors which join it have their discharges recorded every year. On the average these contribute about half a milliard, or less than a twentieth of tbe total annual discharge; but in very high years such as 1946 the amount may be very much larger. There are a number of bends in the course of the river, aDd one of its characteristic features is the presence on the inside of these of marshy basins which bave formed in the abandoned channels, and which are filled with water when the river is iD flood. These therefore form a definite part of the river trough, and, though tbey are too overgrown to pass any appreciable portion of the discharge, tbe water which they absorb' and store temporarily during the flood causes an appreciable lag in the passage of the flood down the river. . Apart from uncontoured air survey maps, a line of levels along tbe river, and four cross- sections in the lowest 100 km.ofits course, there are no data about the shape of the valley. In this attempt to analyse the flood, therefore, use has had to be made of the concept of an average or idealized trough, with a form such that its filling and emptying during the passage of the flood cause differences in the discharges at the two ends which are, on the average, close to tbose 913 actually observed. In each year these differenccs are very simila r to those observed on the White Ni le between Malakal and Renk, where cross-sections at intervals of 20 km. make possible a d irect estimate of the shape of the tro ugh. Simi lar differences would probably be found OD the Baro betwecn GHmbcla and Sobat Head were they no t masked by the large amounts which that river spills over its banks. Il is very noticeable from the air how the marshy basins on the Sobat, o n the ins ides of its mea nders, a re continued in apparen tly exactly the same form up the BarD. SUMMARY OF RES ULTS The firs t section of this paper is co ncerned wi th the shapes of the lower and upper Sobat troughs. The second comprises an analysis o f the Oood cycle, leading to comparisons between predicted and observed losses between the discharge sites at Sobat Head , a nd an estimate of the losses if the river were controlled . In th e third section the regimes of the tributa ries a re discussed and the relationship of the Sobat with the Machar Marshes is investigated . The results may be bricOy summarized as [ollows: Sf.in -ION I. TIII~ S IIA !'I' 01' Till! SOII"T VALU!Y (i) Four cro)'s-sccliolls of Ihe lower Sohal (from !-linct Oolcib up to Abwong) show that its trough has a forlll such tha l lhc ;.average widt h o f its sur f'ucc remains virtua lly unc htlngcd for the fil'611wo metres of ri~ uhovc InC' lIl luw love l. Thereuner the uverage width increases ns tho squurc of the ex tra rise (i) U!'il1l(, th ...· se forms I'm the tmush. nnd this va ilic fo r the absorp tio n dep th, nnd a llowing for losses and gain!' hy evapHril l i{l1l ami minfa ll. the va lue .. of cUlllulativo obscrvcdl.lnd cumulutive predic ted luss hc twcen Sohat 11e:'ld :lIId I li llc t I.>uleib avcrngec..l fmlll twelve flooc..l s arc fou nt.l to agree with in I ';;', or the c.urrcspondillJ,! Sohnt IleH d cUlll lll nlive toln l of d ischurgo throughout the ris ing stage o f Ihe.: II ltut!, fro m May 10 Novcmher inclus ive. Agreemen t in the fa ll ing s tage o f the nood is equa lly l,(oml rlu' li ve nooLls beluw :Ivcragc heigh t (i.e. cont:J illClI withi n the incised trough) , bu t no t quite ~I) I;UI,X I ror ~cven noods above lIverugc, wherc ubscrvetl losses by the ent.l of February uro less lhan thos\,) prcl li1.:1ed . J}l'ohahly hecn use o f umccordcd inllow from tributaries. (ii) This dose ugrcc lllc lIl hetwee n ohs\,)rvcd ;I nd pred icted losses m CH IlS tha t HI)lIrt from the measured csdw lIges or w:l lcr Ih rolll,(h the Twalor a nd Wukilu (whkh arc regu lar ly observed), the Sobat flood is in nearl y :tli yc: u'S virtlHtlly sc lr-.... ontu ined. T he di scharges obscrvctf a t Sobn t l-Ieau p lus the waler :Hlded 10 the IIl'l lial nnllJ ~ lII"f~lCe by n linr;)lI . less tlte loss frolll Ihis by evtlporntion and tho loss by IIhsllrpl io ll il110 the I1 l1m l ~p l llill. whe n correded for the disl,,' hnrgcs o r these two tributa ries, Qccou nt 0 11 the uvel':ll{c ell tirely ill IIIW yellrs. t\lul a l ll1o~1 cnl irely in high o nes. COl' the discharges u t Hillel ()uli.lih throughout the Ilood <:yd o. (iii) If thc Sobat were kepi ~Il any cons tulll dischurge o f' between 10 und 50 mi d the losses during Ule timel y SColson (J.UIUUI')' to Ju nc) would be o j' the orde r of 2 to 31}'o. !:i1!rnON Ill. T ill! R I:OIMI~" OF T II U Tl( lUlITMm.s (I) The T walm is l'i how lI 10 act os :t s rill -ch:lIlI1d from lhe Pibor in ye'lrs when Ihis is high. (ii) The Wlika n is shown In UI,,' I us a s imple spill -dml1ncl from the Soba l while this rises from 7·5 to 10 ·4 0 11 Nas il' l:ttl ugc , t:lki l1 8 ou t un Ilvcragc of 150 m illi un.'; duri ng this stage. Aner this the flow in the Wnkuu reverses. 11111..1 it I'clu rns I1 t leas l un clj uu l nmount and on the average con tribu tes u further 250 l11illiull)) to the Soha t l..Iuring nn'" nner tho pcak of its nood. T his behavio ur is shown to be most prohably due to a n ini lin l spi ll thro ugh the Wuku ll into the Tierbor sys tem, wh ich runs north of the Soh:lt unll roughly lJ:tl':l lld to it , this being lutcr fillet! by sp ill from the no rth ballk of the Baro to such II heigh t thnt il is able to ovornow lhrough the Wukuu b:.lck into the Sobot oven a t the height of li s nood. (iii) Rcusolls .11'0 given for Iho helief llmt QI)3 rt fro m drainage from the sou thern plains in oxccptiona l yeurs no no or Iho o ther tributuries takes fro m or contribu tes to the Sobat uoy appreciable amount. 1t is shoW II th:11 Iho lu rge losses tint! gains which occur at the peak of each high nood arc unUke.ly to be due 10 s l) ill thro lls h the nort hern tributaries fo ll owcu ciUler by sudden henvy innow from UlO ~out h. 01' by II re turn now frOI1l the Tic.rOOr sys tem similllr to Ihol whieh takes p lace through the Wukuu. Tho muin rcusons for this beliof aro the large s ize ond I'npidity or Lhesc losses and gains, the inv~' ritlb lc closo coincitluncc in time between tJ10 change fro m one to tho other with the chonae rrolll II ri sing to a fa iling ri ver, nnd the hick or obsurvod discharges in these lributnry khors of the ordor of s izo requ ired. It is shown thut tJlcse losses and gains can most simply be explained by the ussulllpLion thut they uro due to the noodina nod subsequent dra inage or u wide nrcn of pln.in above nnd immediately nunking tho pcrmnnont swnmps of 010 Sobot flood-plain. 914 HISTORY OF THE PRESENT INVESTIGATION It has already been mentioned in the account of the White Nile Flood (Chapter 2, pp. 902- 5 that heavy inflow on that river was deduced in four years and that in the three of these for which records were available there was ample confirmation in the adm inistrat ive reco rds of the area. This confirmation of the method of analysis led to the present paper, which attempts to apply it to the River Sobat, though in a somewhat different form. On the White Nile the form of the trough could be obtained from measured cross-sections, and the inflow of the tributaries had to be deduced from a com parison of the observed and co mputed losses. On the Sobat there are very few measured cross-sections, but on the other hand the fl ow in the tributaries appears to be measured fairl y comprehensively, so that, except in very high years, a more reliable comparison can be made between the observed discharges at the two ends of the river than on the White Nile. Basically, therefore, the method used in this analysis was to start with the assumption that the total increase in the width of the river above low level was equal to a constant times the square of its average rise above that level, and then to calculate this constant from a comparison of the end discharges and the value of the rise in differen t years. This assumption made it possible to express the surface area, and hence the volume of the trough and the loss by absorption, in terms of the rise and the constant. By equating in several floods the losses computed from these to the losses derived from the differences between the discbarges at the two ends of the river during the rising stage, an approximate value was obtained for the average constant of tbe whole river. Tbis was the first step, taken before this paper was actually begun ; and the close (tho ugh fortuitous) agreement between vailies of the constant derived from a number of floods was an encouragement to investigate the matter more closely, and to outline progress to date in a letter to Natu,e (I) . FIRST ANALYSIS OF THE SOBAT FLOOD At this stage the author's attention was drawn to four cross-seclions which had been measured of the lower Sobat. From these the average cross-section as far up as Abwong could be calculated directly as on th'e White Nile, leaving only the average cross-section of lhe upper two-thirds of the river to be derived indirectly. Thirteen Hoods were studied and twelve (excluding 1946--47) Hoods were found in which the records of main and tributary discharges were sufficiently complete for their differences to be used in estimating the shape of the average cross-section of the upper Sobat. From these a mean value for the constant already referred to was obtained, and used to compute tables of surface area and trough volume in terms of average rise above mean low level. Similar tables had been prepared for the lower Sobat from the measured cross-sections. Using tbese tables, as on the White Nile, it was possible to prepare detailed figures of the cumulative losses or gains due to evaporation, rainfall , change of trough volume, and absorption along the wbole Sobat in th irleen Hoods. The total of these (total computed cumulative loss) for the end of each ten-day period was compared with that observed, and in the mean of the twelve floods wbose records were reasonably complete the differences were not very great. THE CONCEPT OF THE INCISED TROUGH This was expected to complete the analysis ; but the individual comparisons for each flood now available showed that they divided sharply into two groups. One group consisted of five , low' floods where the maximum rise was below the average maximum of thirty years, and the other of seven ' high ' floods where the maximum rise was above the average. In tbe first group the curve of observed loss had very much the predicted shape. In the high floods, on the other hand, once the water had risen above the average maximum flood level, the observed losses (except in 1946) increased very much faster than could be accounted for by the assumed average cross-section; and after· the peak of the flood there were correspondingly large gains. The most obvious, and quite reasonable, explanation of this was that the assumed shapeofthe trough only extended up to the level of the average Hood, and that beyond this the land surface stretched out in a flat or very gently sloping plain, over which the water spread and from which it returned, giving the large and rapid losses and subsequent gains deduced from the observed discbarges. On the White Nile the trough is deep enough to contain all floods up to the highest recorded; but on the upper Sobat it seems that the trough has been incised by, and only just contains, the average flood. RECONSIDERATION OF ABSORPTION DEPTH this modification of the assumed shape of the trough necessitated complete recomputation of the basic tables of area and volume on the upper Sobat, and thereafter of tbe computed losses on the whole Sobat. The opportunity was also taken to review another of the basic assump- tions-that regarding absorption. On the White Nile flood-plain an average absorption depth 915 of 80 cm. had been deduced and used throughout the analysis, but there was some indication that this was probably an over-estimate. Nevertheless, for lack of any definite evidence the same value was originally assumed for the Sobat. The necessity of complete recomputation seemed to justify making a fresh estimate of absorption in the incised trough, using the five low floo ds. With approximate values for the maximum flooded area, and for evaporation losses and rainfall gains, these gave a mean value of only 30 em., and this was therefore adopted. Investigations showed that no other value between 0 and 80 em. appreciably affected the differences from the mean in the calculations of the slope of the surrounding plain, and so the same value was assumed for this. This last stage of the analysis, which had been started in 1949, was carried out at Ed Darner in the Northern Sudan during 1952-53 without the earlier resul ts having been discussed at all with the Jonglei Investigation Team, who were working in the Southern Sudan at Malakal. They had been studying theMachar Marshes and had available some information gained from ground and aerial reconnaissances of which I had had no knowledge. From these they had drawn independently their conclusions about conditions in and alongside the Sobat which did not differ very much from those which I had drawn. Such differences as existed were resolved during a visit to Malakal in June 1953 and the result is a general picture of the Sobat regime which seems likely to be as near the truth as can be hoped for until further information is available. It differs in some respects from the account given in Volume VIII of The Nile Bas;1! by Dr. Hurst, but considering that virtually no reference was made to this either by the Jonglei Investigation Team or myself until our meeting in 1953 it is surprising how closely the results of the three independent investigations agree. SECTION I. THE SHAPE OF THE SOBAT VALLEY On the Sobat the regular discharge sites are at its head and at Hillet Doleib, 8 km. from its mouth; cross-sections are only available for the lower Sobat, as far up as Abwong. Above this there are no contoured maps or complete cross-sections at all, so that an indirect method of estimating the trough shape had to be used. It had been found on the White Nile, and also on the lower Sobat and for a part of the River Pibor which is a tributary of the Sobat, that the curve of the mean idealized bank profile was approximately parabolic; and it therefore seemed reasonable to assume a similar form for that of the upper Sobat. From this assumption, as will be shown below, the shape of the idealized trough can be calculated, using the rise from low to flood level, the difference between the discharges at the two ends of the trough, and measured or assumed depths for rainfall, evaporation, and absorption. The method of deriving an idealized trough has already been outlined (see p. 853) and the calculations may now be considered in more detail. In the first place it was clearly advisable to divide the Sobat into two reaches, both because of the existence of cross-sections which did not extend very far up the river, and also because the backwater effects of the White Nile were likewise limited. Fortunately the limit in both cases was approximately at Abwong; and as this is a gauge site, it formed the obvious place to put the boundary between what for convenience may be called the upper and lower reaches of the Sobat. THE LOWER SaBAT On the lower Sobat four complete cross-sections have been measured('); they are shown in Figs. K I and K 2. In order to idealize them it was necessary first of all to establish at each the datum of normal low river level. This was done by plotting the longitudinal profile of the river surface from the mean gauge-readings converted to reduced levels as shown in Table 423 (p. 922). The profile is shown in Fig. K 3, and the positions of the cross-sections have been plotted on it, enabling their datums to be read off directly. On each cross-section the total width at each half metre of rise above the datum was then measured and the results are shown in Table 424. This table also gives, in the second line for each cross-section, the corresponding flood-plain widths, which are obtained by subtracting from each total width that of the low-level surface. The means were then abstracted, and it will be seen that they approximate closely to the formula: Width of Flood-Plain - 210 (rise minus two metres)' or in other words the mean idealized bank profile of this reach is nearly a parabola which starts at two metres above normal low level. For calculation of the surface areas and volumes in the lower Sobat (Tables 425 and 426, pp. 923-4) this formula was used, assuming the length of the trough to be 116 km. This method was more convenient than using the actual mean profile, and introduced no appreciable inaccuracy into the resuits when these were combined with the same quantities on the upper Sobat, where the trough is not only twice as long but is filled to a higher level. 916 THE UPPER SOBA T (I) THE INCISED TROUGH For calculating the trougb volumes and surface areas on the upper Sabat Ule shnpe of the incised trough has first to be found . At first it was assumed to have a form simila r to that of the lower Sabat, with idealized bank profiles of parabolic form sta rting at some poin t above mean low level. This assumption had to be ruled out at an eady stage beca llse it failed co m- pletely to account for the considerable losses which invariably occurred each year during the early part of the rise, and which could only be accounted for by assuming that the incised trougb began to widen as soon as its surface was above the mean low level. Thus the average width of the water surface was assumed to be related to the average rise above mean low level by the following formula : W - w. + 2. K .h' ( I) where W - average tOlal wid th W . ... average width of (he low-level chE L OWER SOBAT (Mean of Hillet Doleib and Abwong - hJ 3·32 3-38 3-)8 3-24 3-24 MEAN RJsB UPPER SoBAT (Mean of Abwong, Nasir. and Sobat Head = h~) ... 4-73 4·77 4084 4-66 4-60 TABLE 428 CORRECTIONS FOR ACTUAL STARTING GAUGES (DIFFERENCES OF THESE FROM MEAN LOW LEVEL DATUM) FOR FIVE LOW FLOODS 1936 1939 1941 1943 1944 Ten-Day Period before Losses Begin I Apr. 1 Apr. 3 Apr. 3 Apr. 3 I Apr. 3 Start Gauge at Hillel Doleib ... .. . ! 11 ·02 11 ·50 10·78 11 ·11 11 ·06 Difference from Datum ... -0·03 + 0·45 -0·27 + 0·06 ; + 0·01 : Starting Gauge at Abwong 10·27 10·89 10·12 10·)) 10·28 Difference from Datum -0·04 + 0·58 - 0·19 ! + 0·02 - 0·03 Starting Gauge at Nasir 5·12 HO 5·08 5·34 5-4) Difference from Datum -0·01 + 0·67 -0·05 + 0·21 + 0·30 Start Gauge at Sobat Head 4082 5·50 4·78 5·04 5·14 Difference from Datum + 0·02 + 0·70 -0·02 + 0·24 + 0·)4 MEAN Dlffl!Jf.ENCE LoWER SOBAT (Start minus Datum = d1) • • • -0·04 +0·52 -0·23 +0·04 - 0·01 MEAN DIFFERENCE UPPER. SoBAT (Start minus Datum = dJ ... -0·01 +0·65 -0·09 +0· 16 + 0·20 925 TABLE 429 COMPUTATION OF LOSSES ON THE LOWER SOBAT IN FIVE LOW FLOODS I 1936 1939 1941 1943 1944 Mean Rise = h" rrom-;~;I~ ~-2;----------1;--3-'3-2- ;--3-'3-8-;--3-'3-8- +--3- '-24-+--3'-24- (h, - 2)' .. . II 1·74 1·90 1·90 1·54 1·54 (h, - 2)' 2-30 2-63 2-63 1·91 1·91 h, - d, ... .. . . . II 3'36 2-86 H I 3-20 3·25 (d , rrom Table 428) Loss to Main Channel = 12(h1 - dJ .. I 40 34 43 38 38 Loss by Absorption to Flood-Plain ;:a 6·9 (hi - 2)1 .. , ! 12 13 13 to 10 Loss to Volume above Flood-Plain ,",", 7·63(hl - 2)' ... I 18 20 20 IS JS TOTAL ... : M ~ U ~ ~ I Non: L.oSK$ o.re in millions or cubic IlXtrcs, all other I'IgurM! . AI in mella.. TABLE 430 (i) DETERMINATION OF AVERAGE ABSORPTION DEPTH FROM FIVE LOW FLOODS : Corrected ~ AMPPIlJ(lI~XlOlum·'r ante I Final Approximate ' Change in Final Deduced Absorption Yea_r. __ _l __ Ol:l.served Evaporation Trough Absorption Flood-Plain Depth ._~OSS I Loss , Volume Loss s:.~. em. '''--1- - -. -- -- .-------1--- - - -;-- --+--- 1936-)7 77 + 10 + 12 55 338 16 1939-40 493 + 2 - 15 506 320 158 1941-42 29 1-85 + 17 -73 350 - 29 1943-44 - 177 + 5 - 8 - 174 320 -54 1944-45 221 + 5 - 5 221 309 72 Means 128 21 0 107 327 32-6 Mean absorption depth from ra tio or sums in the last two columns - 30·6 an. - -- - - -- - ._-- - - -------_._--- (ii) DIFFERENCES OF DEDUCED ABSORPTION DEPTHS FROM 30 CM. EXPRESSED AS PERCENTAGE ERRORS OF SOBAT HEAD DISCHARGE T~~?~;: I-'-{~~~~ -1 ._S~~~::d I Difference Year as Percenlagc 1936-37 - 0-4 1939-40 194 1-42 ---:~f -'-r -~1~i - 1 i~:~ + 3-4 - 1·6 1943-44 1944-45 92 I -2·3 + 129 11 ,800 +H Mean (regard less of sign) 1·8 TABLE 431 COMPUTATION OF DIRECTLY CALCULABLE LOSSES ON THE UPPER AND LOWER SOBAT FOR FIVE LOW FLOODS _____ .L 1941 1943 1944 1936 ___1 _9_39_ .1--_ _ + __- -+ _ _ _ Mean Rise on Upper Sobat -= hi 4·73 4'77 4-84 4-66 4·60 h' 22·4 22·8 23-4 21·7 21 ·2 h: 105-8 IOS-5 1Il-4 101 ·2 97·3 Using d" from Table 428, hi - d, ... .. . 4·74 4·12 4·93 4·50 4-40 Tola] Loss on the Lower Sobat, less Evaporaljon (from Table 429) ...... ......... 70 67 76 63 63 Approximate Net Evo.poration Loss for the Whole Sobat . - 85 - 107 - 32 -92 - lOS Loss to Permanent Channel of Upper Sobat = W. L(h, - dJ 154 133 160 146 143 Sum of Calculable Losses 139 93 204 117 98 926 TABLE 432 COMPUTATION OF THE PARABOLIC CONSTANT OF THE UPPER SOBAT INCISED TROUGH FROM F IVE LOW FLOODS (EQUATION 7) ._!_ 1939 1941 . 1936 .. 1 194] 1 1944 Observed Loss to Peak of Flood 6]9 966 792 Observed In60w of Tributaries - 33 - 44 - S6 ~m I-J~- Corrected Observed Loss (Q. - Q.) 606 922 736 ] ij] 621 Sum of Calculable Losses tR - E - L.W. h) 139 93 204 11 7 98 Net Observed Loss = DIVIDEND 467 829 532 266 S23 pU oSl t .L. h 3 -_ Flood-Plain Loss _ K 16-3 16·7 17-5 15-0 ~ ~) 2L.D.h' = Absorption Loss 3-1 ] -2 3-3 H r I K Sum - Coefficient of K = DIVISOR . 19-4 19-9 20·8 1709 QUOTIENT - K .. 24-1 41 ·7 25-6 14·3 29·] Mean K (from Ratio of Sums) 27-1 27 ' 1 27- 1 27 ·1 27 ·1 Difference from Mean .. -]-0 + 14-6 - 1-5 - 12-8 +2-1 Difference x Divisor '"' - 58 + 291 - 22 - 238 + 38 Total Discha.rae at Sobat Head ... 8,780 8,950 8,480 8,340 8,490 Difference 35 Percentage of Discharge - 0-7 + ] -] - 0 ,] - 2'1 + 0'4 TABLE 433 SURFACE AREAS OF THE INCISED TROUGH OF THE UPPER SOBAT, FOR MEA N GAUGE-READINGS FROM 6-6 TO 11 ·6 This table was calculated from lhe foUowing fomlula : Surface Area _ W.l + 2K.Lh'. where W. _ mean width at low level -= 140 metres L - length of reach = 231 km . K - Parabolic Constant - 27· , from Table 432 b - mean rise along the reach above the mean of the low-level gauge-readings al Abwong, Nasir, and Sobat Head, whjch is 6·7. Mean Rise Flood- Tolal Mean Rise Flood- TOlal Gauge Squared Plain Surface Plain Surface AIea AIea Gauge Squared Area Are. . ~ , 6-6 0-00 0-0 i 12 9-1 I 5-76 72-1 104 6-7 0-00 0-0 32 9-2 I 6-2S 78 '2 110 6-8 0·01 0- 1 12 9-] 6 -76 84'1 116 6-9 0·04 0·5 ! 12 H 7-29 9 1·2 124 7-0 0-09 I-I 33 9·5 H4 98-2 III 7-1 0-16 2·0 34 9·6 8·41 lOB 138 7-2 I 0·25 3-1 3S 9-7 9-00 lIN 145 7-3 0·36 4-5 37 H HI 120-3 153 H 0-49 6-1 38 9-9 10-24 128-2 160 7-S I 0-64 8-0 40 10-0 10-89 136·3 169 H 0-81 10-1 42 10-1 II-56 144-7 177 H 1-00 12·5 45 10-2 INS 153-3 186 H HI IS-I 47 10·3 12-96 162-2 ; 194 7-9 1-44 18-0 50 10-4 IH9 171·3 204 8-0 1'69 I 2H 53 10·S 14-44 180·8 21l I 8-1 1-96 24-5 I I 57 10-6 IH I 190·4 , 223 8-2 2-25 28-1 60 10-7 16-00 200-5 233 8-3 2-56 )200 64 10-8 16-81 210-5 24] N 2-89 36·1 68 10·9 17-64 220·9 253 8-5 3-24 40-5 I 73 11-0 18-49 231-5 264 I 8-6 HI 4H I 77 IH 19-36 242-4 27S 8-7 4-00 50-I 82 11 -2 20·25 253-3 286 8-8 HI 5H 88 11 -3 21-16 264·4 2~7 8-9 4-84 j 60-6 93 11·4 22-09 276-S 309 9-0 H9 66·2 98 11·5 23·04 288-4 321 - - 7' I - 11'6 I 24-01 lOO'5 333 Nous: ~eso:::: =';l:o" U:~~~:90.km. (rom Hillol Doleib, tbo equivalent sun.ec .re.u a. .. obtained by muhiptyiDa ... alua cMriwd (rom this table by 190/ D I, wbkh - 0-11]. 927 TABLE 434 VOLUME OF THE INCISED TROUGH OF THE UPPER SOBAT ABOVE MEAN LOW LEVEL (FROM MEAN GAUGE 6·7 TO 11-6) Mean 'l ' ' I' M Gauge ~ o~oo _! ~~.~04 ! 0·06~ I-G_ae_:8"_e-+_0'_00_!---0_'_02--1_0_'0_4+-_0_'06 0·08 8' 1 57 58 59 61 62 6·7 o I I 2 3 8·2 63 1 64 65 67 68 6-8 3 4 5 5 6 8'3 69 70 72 73 74 6·9 7 8 9 9 10 8-4 76 77 79 80 81 7·0 10 II 12 12 13 8-5 83 84 86 87 89 7·1 13 14 15 15 16 8-6 90 92 93 95 96 N 17 18 19 19 20 8·7 98 100 102 103 105 7-3 21 22 23 23 24 g.g 106 108 110 I III 113 7·4 25 26 27 27 28 8·9 115 117 119 ' 121 123 7·5 29 30 31 31 32 9·0 125 127 129 III 133 7·6 33 34 35 35 36 9·1 135 137 139 142 144 7·7 37 38 39 40 41 9·2 146 148 150 153 155 H 42 43 44 45 46 9'3 157 159 162 164 167 7·9 47 48 49 50 51 9·4 169 171 174 176 179 8-0 52 53 54 55 56 9·5 182 , 185 188 1 190 1 193 ---- 10:00 -'0'01-;0 :0;-'-0:03' 004 I 0·05 1. 0·06 I 0·07 ! 0'08 I 0·09 I - --~- '-- ~;:' -;;7-r-~ -i--2~---2~1---~- ; ~~ G;-~~--- 9·7 210 2 11 213 214 216 217 219 220 222 i 223 9·8 225 226 228 229 23 1 232 234 236 237 239 9·9 24 1 242 244 246 247 249 250 252 254 256 10·0 257 259 262 262 264 266 268 269 27 1 273 10·1 274 276 277 279 281 283 i 284 286 288 290 10·2 292 294 296 298 300 301 303 305 307 309 10·3 31 1 313 315 317 319 321 323 325 327 329 10·4 lJ l 333 lJ5 337 339 342 344 346 348 350 10'5 1 352 354 356 358 360 363 365 367 370 372 10·6 374 376 378 380 382 385 387 389 392 394 10-7 397 399 402 404 406 409 411 414 416 418 10-8 421 423 426 428 431 433 436 438 441 443 10·9 446 448 451 453 456 458 461 463 466 469 I 11 ·0 472 475 478 480 483 486 488 490 493 495 1 11 ·1 498 50 1 504 507 509 511 514 517 520 523 11 ·2 526 529 532 535 537 540 543 546 549 552 11 -3 555 558 561 564 567 570 573 576 579 I 582 11 '4 585 588 592 595 598 601 605 608 6 11 614 11 ·5 617 620 62 3 ! 626 629 632 ' 636 639 642 646 11 ·6 , 650 NOTU : Volumes li re in millions orcubi<: metres. Wben {he di s.::hullc sil t is al Nuir, 190 km. (rom Hillel Do lcib, the eouivalc:nt trou¥b "oh,loxs arc obtained by rnullipl),il'la va lues derived (rom this table by 190/ 2) 1, or 0-82). TABLE 435 CHECK CALCU LATION OF THE VO LUME OF THE INCISED TROUGH OF THE UPPER SODAT This table was calculated from the following Connula: Trough Volume .. W .. L.h + 2. K 3L. h~ where W" - wid lh of the main channel =: 140 metres L .::; lenglh of (he reach :II 23 1 km. K ___ the parabolic constant """ 27 ·1 from Table 432 h c< mean rise along the re.'\ch above the mean or lhc low-level ga uge-readings al A,b wong. Nasir. and S,o bat Head, which is 6·7. Mean Li Gauge ___ ___ L_ __ i I W• . L.h i·K.L.h1 Trough h' _L __ 1 Volume ------- --.--1--- -- 7·2 0-5 0·1 16·2 0·5 16'7 7·7 1·0 1·0 32-4 4·2 36·6 8·2 ) ,5 3-4 48-6 14·1 62·7 8-7 2·0 8·0 64·8 33-4 98·2 9·2 2-5 15·6 81-0 65·2 146'2 9·7 3-0 27·0 97·2 11 2·7 . 209'9 10-2 3-5 42·9 11 3·4 178·9 292·3 10·7 4·0 64'0 129·6 267· 1 396'7 11 ·2 4·5 91·1 145-8 380·2 526·0 11-6 4·9 117·6 158·8 490·7 649·5 Non: The last lluc. COIUmDS arc in llI.ilUoaa 0( cubiC metreI. 928 TABLE 436 CALCULATION OF THE SLOPE OF THE PLAIN TOWARDS THE UPPER SOBAT TROUGH i ITI!M 1934-35 1935-36 1937-38 1938-39 1945-46 1946-47 1947-48 ._-I- ----- (1) Period before Flood overtops Trough Sep. (3) Scp. (2) Sep. (3) Scp. (3) Sep. (2) Aug. (3) Scp. (I) (2) Period of Highest Observed Loss ,-, Nov. (2) Nov. (2) Oct. (2) Nov. (2) Nov. (3) Nov. (2) Nov. (2) (3) Mean Gauge at end of this Period .. . 11·71 11·58 11'70 11·96 11 ·69 12041 11 ·88 (4) Maximum Mean Gauge ", ... ... ,,' ," 11·86 11 ·75 11 ·70 11·98 11·87 12046 12·02 (5) Mean of the two, minus 11·60 = h' (centimetres)... . .. 18 6 10 37 18 84 35 (6) Lower Sobat Losses by Absorption and Trough Volume. 118 91 91 118 105 245 116 (7) Evaporation on Whole Sobat, up to top of Trough ... - 78 -61 - 122 - 114 - 85 -77 -83 (8) Sum of Preliminary Losses (millions of m')... ... . 40 • 30 - 31 4 20 167 33 (9) Net Evapora tion Depth from top of Trough to Peak (d) ... ". .. . 5 6 - I 6 10 - I 5 10) Swn of Height, Absorption and Evaporation Depths (h ' + d + D) H I 53 42 39 73 58 110 70 111) S~(h; + D + d) (millions of m.') '" ... ... ... . .. 177 140 130 243 197 367 233 12) Vt - D.W• . L ... '" ... ... ... ... . .. . .. 640 640 640 640 640 640 640 13) Sum of (8), (II), and (12) '" .. . ... ... ... .. . ·· · 1 857 810 739 887 857 1.1 74 906 14) Mnimum Observed Loss (Corrected for Tributaries) ' " .. . . ,I 1,677 1,796 519 1,399 1,286 741 2,174 1}~ Ditreh~~2cg~4~(~~t~~~ .~esidu.~~ Lo~.~ DI~'.I.DENO... 820 986 -220 512 429 -433 1.268 -= ::: 0·83 0·72 0·69 1·03 0·88 1·40 1{)() (17) L(h ' + 2D + d) (L - 231) . i 192 166 159 238 203 323 23 1 (18) L.h'(h' + 2D + d) = DIVISOR ... 34·5 8·0 15·9 88· 1 36·5 27 1 80·8 (19) (IS) divided by (18) = QUOTIENT = P 23-8 123-2 - 13-8 5·8 11 ·8 - 1·6 15·7 1 Mc:.n Ql.lOtient (e:u;:ludin, 1935 and 1946) .. )051/ 284 '] _ 100S. TABLE 437 DIFFERENCES FROM THE MEA N PLAIN SLOPE EXPRESSED AS DISCHARGE ERRORS AT SOBAT HEAD I lTEM 1934-35 1 1935-36 1937- 38 1938-39 1942-43 1945-46 ! I 194<>-47 ,1 1947-48 --------------_. ------,- _. .... - Difference of Individual Values of p from the Mean + 13·0 + 112'4 -24-6 - 5'0 - 2· \ + I.() - 12·4 + 4·9 Differences multip lied by Individua l Divisors + 448 +889 - 391 - 440 - 60 + 36 - 3,360 +396 Total Discharge at Sobat Head during the period" 10,354 11 .2 12 7,911 10.079 8,989 10,672 11 ,598 10.694 Difference as percentage of Discharge at Sohat Head + 4·3 +8'0 -4-9 - H - 0'7 + 0-4 - 29'0 + 3·7 Total Sobat Head Discharge for May to October inclusive ... 8,866 9,986 8,6 13 8,733 8,989 8,724 9,978 9,405 Maximum Mean Gauge (mean of Abwong, Nasir, and Sobat Head) " J 1·86 11 ·75 11 ·70 11 ·98 11 ·76 11 ·87 1H6 12-02 TABLE 438 SURFACE AREAS ON THE UPPER SOBAT IN HIGH FLOODS This table is calculated from the formula : Surface Area = 5t + 2.L.p.h', where SI - the surface area of the incised, trough, -= 333 km.' from Table 433 L - length of the reach =- 231 km. p -= one thousandth of the slope of the plain on each side of the trough = It h' - the rise above the top of the trough, wruch corresponds to mean gauge 11·60. ) Mean Gauge 0-00 0·01 I 0-02 0'03 0-04 I 0·05 .L 0·06 I 0·07 ._- _. -- L~L~'09 _ r - ' .- 11 ·6 333 384 435 485 536 587 I 638 689 739 790 11 ·7 84 1 892 ! 943 993 1,010 1,090 1,150 1,200 1,250 1,300 JJ ·8 1,350 1,400 I 1,450 1,500 1,550 1,600 i 1,650 1,700 1,750 1,8 10 11 ·9 1,860 1,9 10 1,960 2.0 10 2,060 2, 110 2, 160 2,2 10 2,260 2,3 10 12·0 2,360 2,4 10 2,460 2,510 2,560 2.610 2,670 2,720 2,770 2.820 12-1 2.870 2.920 2,970 3,020 3,070 3,130 3, 180 3.230 3,280 3.330 12-2 3,380 3,430 3.480 3,530 3.580 3.640 3,690 3,740 3,790 3,840 12·3 3,890 3,940 3,990 4.040 4.090 4,150 4,200 4,2~0 4,300 4,350 12-4 ! 4,400 4.450 4,500 4,550 4,600 4,650 4,700 4,750 TABLE 439 CALCU LATlON OF . TROUGH' VOLUMES ON THE UPPER SOBAT FOR EACH DECIMETRE ABOVE AVERAGE FLOOD LEVEL This table is calcu lated from the formu la: • Trough ' Volume z: VI + SI .h' + L.p.h" . where Yl =-- volume of the incised trough = 650 millions SI ..,. surface area of the incised trough _ 333 km' L =- the length of the reach - 23 1 "m. p ..:a one thousandth or the slope or the pluin on each side or the trough c:: 11 h' = the rise above the top or the incised trough, which corresponds 10 the mean gauge 11·60. Mean h' h" 3J3h' 2541h" • Trough ' G ;lUge (em .) Volume 11 ·60 0'0 0-0 0 0 650 11 ·70 0-1 0·01 33 25 708 11 ·80 0·2 0-04 67 102 819 11 ·90 0'3 0·09 100 229 979 12·00 0'4 0-1 6 133 406 1,189 12· 10 0·5 0·25 167 635 1,452 12·20 0-6 0'36 200 915 1,765 12·30 0·7 0'49 233 1,245 2,128 12-40 0-8 0·64 267 1,626 2,543 12-50 0-9 0-8 1 300 2,058 3,008 Volumes io millionl 01 m', TABLE 440 , TROUGH ' VOLUMES ON THE UPPER SOBAT IN HIGH FLOODS ~·u·;e . __O_' o_o I 0·01 1_ ~:0~ 1~~3_ _1 _~~1~_1_0·~. 1 0·07 I 0'08 I 0-09 --'-1-'6 I 650 l ~-;';- 659 I I ~~~ i ~~ 678 1 684 I 690 I 696 702 11 ·7 708 716 755 1 766 778 792 805 JJ'8 819 833 ~!? i I ,~ I I,~ 895 I 912 929 I 945 962 11 ·9 II 979 1,000 l :~~ I 1,080 I 1,100 1,130 1,150 1,170 ]2,0 1 1,190 1,220 1,270 1,300 1,320 I 1,340 1,370 1,400 1,430 12· 1 1,450 1,480 1,5 10 1,540 I 1,570 1,600 I t :~~g 1,670 1,700 1,730 12-2 1,760 1,800 1,840 1,870 i 1,900 1,940 2,010 2,050 2,080 g:~ ! U~ U~ 2,210 2,260 ! 2,300 2.340 2,380 2,420 2,460 2,500 2,640 2,~0 I 2,?!.0 ; 2,770 2,810 2,860 2,910 2,960 12·5 I 3,010 I - 930 SECTION 11. ANALYSIS OF THE SOBAT FLOOD COMPUTED LOSSES Once the tables of surface area and trough volume have been computed, calculating the individual items of loss in dillerent years is a compa rati vely si mple, though tedious, operation. The method used foHows al most exactly that evolved in lhe ana lysis of the White Nile nood, of which a detailed account has already been given (pr. 86 1-75). The computed losses com- prise the following: (i) Increase of trough volume afler the s t ~1I1 or the cycle. (ii) Net loss by evapor:ltion minus r~in fa ll . (iii) Absorption in the newly wetted ground. TROUGH VOLUME For this the mean gauge-read ings for each ten-day period at Hillet Dolcib, AblVong. Nasir, and Sobat Head were used"); they are given in (i) of Tables 44 1- 53 (pp. 935--60). The mean gauge for the reach was fi rst obta ined for the ten-day peri od and then this was meaned with the mean of the ten-day period immediately followin g. This was necessa ry in order to obtain the increase in trough volume at the end. and not the middle, of each ten-day period, since this was the instant to which all other items of loss were referred . The mea n of the first two gauges was used for the trough volume of the lower Soba t, by entering Table 426 (p. 924) ; and the mean of the las t three for the upper Sobat, by entering Tables 434 and 440 (pp. 928 and 930). It is the losses due to change of trough volume wltich a re required , and so in each flood the trough volume at the start of "the cycle has first to be ca lculated. This usually small quan tity is then deducted from the actual quantities obta ined direct ly from the tables. It will be noticed that in some years the trough volumes in February are negative. This is because tbe water surface had by then fallen below its level at the beginning of the cycle, and th is level was the datum from which changes of trough volume were measured . This is shown in 1937-38 (Table 444, p. 941). RAINFALL AND EVAPORATION The volumes of water added to or taken from the Sobat between its bead and Hillet Doleib by rainfall and evaporation during each fl ood were calculated by estimating the depths of these quantities during each ten-day period, and then multiplying Ihe difference by the average surface area of the river during the same period. No allowance was made for rain fa lling o n the flood-plain before it was covered by the river. Rainfall depths were obtained by taking the mean of the readings at Abwong, Nasir, and Doleib Hill up to 1939, and of the first two tbereafter when observations at Doleib Hill had been stopped . The Sudan Government Meteorologist kindly lent the original records and the ten-day sums were made from these. Evaporation depths were obtained by hal ving the average Piche tube readings at Malakal for each month, and then interpolating between the monthly averages for tbe first and third periods of eacb month. The results were rounded off to the nearest centimetre per ten-day period; they are shown in tbe first parts of Tables 441-53 (pp. 935-60) for each year. The differences between these and the corresponding mean rainfall (of which the data are also shown there) were then mUltiplied by the average surface area, which is a lso shown in the first part of the tables. These areas were obtained for tbe lower Sobat by entering Table 425 (p. 923) with the means of the gauge-readings at Hillet Doleib and Abwong. and for the upper Sobat by entering Tables 433 (p. 927) or 438 (p.930) wi th the means of the gauge-readings at Abwong, Nasir, and Sobat Head. In this calculation the direct means for each ten-day period were used, since they give the averages for the period. It will be seen from Table 456 (p. 962) that the average net gain by rainfall minus evapora- tion during the rising stage of the Sobat flood is just over one hundred nti llions, as compared with an almost negligible quantity on the White Nile during its rising stage. There are two reasons for this. Firstly the Sobat basin is actually wetter, havi ng more rainfall and less evaporation than that of the-Wltite Nile between Malakal and Renk ; and secondly, because its flood peak is earlier, its maximum surface area coincides rather with the immediate aftennath of the rains than with the drier conditions which come later. 931 ADSORPTION As on the White Nile, the losses by absorption are the least reliable of all, since they are subject to no meas urement and the average depth on which they are based has been derived indirectly. The mean absorption depth of 30 cm. was obtained as a/ready described on page 918 and Table 430 of Section I. For each flood this was then multiplied by the maximum area inundated, i.e. the maximum surface area less the surface area at the beginning. The result was then taken as having occurred approximately one month after the occurrence of this maximum inundation ; and it was assumed that no further absorption occurred after that until the fo llowing year. The cumulative value of the loss during the early par t of each cycle was interpolated from zero up to this maximum va lue roughly in accordance with the increase in the flooded area, with a lag behind it of about one month . The results cannot be claimed as accurate, but they follow what may be fairl y presumed to be the characteristics of this form of loss . OBSERVED LOSSES The data and the deriva ti on of the observed losses in each year are shown in the second parts of Tables 441 - 53. The mean daily discharges for each ' ten-day period " as published in Volume IVand its supplements of The Nile Basin, or as supplied by the Egyptian Irrigation Department, were mul tiplied by the number of days in each period . The difference between discha rges at Sobat Head and Hillet Doleib during each period is then the uncorrected loss during that period , and these were summed cumulatively to give the uncorrected cumulative loss at the end of each peri od. The observed discharges of the tributaries were obtained in the same way, and also summed cumulatively. The difference is then the cumulative sum for each ten-day period of the corrected observed loss, and this is shown in the last column of the second parts of Tables 441 - 53, and transferred to the third parts of the ta bles for comparison with the compu ted losses. The accuracy of these losses depends very much on the accuracy of the discharges, because they are obtained from the differences between them, and these di fferences are normally only about one- tenth of the discharges. Only during the three or four periods in November or December when the river is fa lling fast and the basins are returning their stored water to the main channel a rc these differences compara ble in size with the discharges. In consequence of tlus it is particula rly important to estimate the accuracy of measurement of the discharges as ca refull y as possible, and to take account of any movements of the discharge sites which may affect them. MAIN RIVER D ISCHARGES On the main Sobat the discharges have been measured regularly since 1905 Dear its mouth at Hillet Doleib ; and since 1929 they have been measured just below the junction of the Pibor and the Baro at Sobat Head . The latter is not easily reached except during the Sobat fl ood; during tJ1C low stages, therefore, the discharge si te is moved to Nasir. On the average this site is used fro m Jan uary to April inclusive, but the exact dates are shown in the second parts of Tables 44 1- 53. Because the river is fairly low at this time, this change of discharge site does not greatly affect the calculations of net evaporation ; but it has been allowed for in computing the va lues of trough volume and surface area which are shown. An important effect of this change of site is that when the discharges are measured at Nasir no allowance has to be made fo r those of the Wakau, since the mouth of this lies between Nasir and Sobat Head. It is fortunate that this is so because their measurement usually ceases soon after the transfer of the main discharge measurements to Nasir, although the khor has not by any means always ceased flowing at that time. TRIBUTARY DISCJ, 5-46 I: 5-82 '"10-99 11 11-69 - - - 100 4] 5-63 6 -00 IHI I 11-67 - - _ 100 47 5-55 5-91 11-17 11 -72 - - - 100 46 MarCh(l)· , 5-10 i 5-40 i 11 -00 11-48 - - - - - Non: The periods muked • are not ten days_ SUMMARY ANALYSIS OF THE YEAR 1935- 36 (;;) OBSER YED DISCHARGES I DUCHARG" Un- I Om",n IN'COW Ic umUladv' l eo"""ed Ten-Day _ _ ._- correcfc.d . ---- -- Inflow Cumulative Period Sobat DHioll1e~ltb I Di1fereDce CumLuolsas tive Nyandin, ! Twalor ! Wakau Total Loss Head May '" lOOt 68 32 32 - 1 3 - 4 28 · 209t 164 45 77 - 4 - 7 -IS 62 ,07 274 33 110 - 7 - , - 31 7' June 398 309 89 199 47l 376 " 2'8 I - 7 -- ,9 - 47 152 - 6 - 62 236 "7 4" 102 400 -- ,6 - 14 - 82 lI8 July '6' 4'4 111 S11 -23 -110 401 · 59J 484 109 620 I - 3 -32 - 145 47l 673 "9 114 734 - 2 -37 - 184 "0 AUlJ\Ist 627 98 832 · 636 '5"4 ' " 923 I -27 - 211 621 + °2 -20 - 229 694 71S 61S 100 1,023 6 -11 -234 789 September 667 ,S,7,4 9J 1,116 11 + 4 -219 897 616 " 1,201 'd 18 - 186 1,015 683 612 - 11 1,272 'E I 2"7 26 -133 1,139 October 681 628 1,331 68 32 -33 1.298 · 68' 643 "46 1,377 ~ 81 3S 128 31 1,408 103 41 , + 83 1,460 7l' 227 1,635 November 64' 661 - 16 1,]92 0 c 100 36 '63 1,755 '81 664 83 1,]09 " 29 4,,8,7 1,796 '0' 661 - 156 1,15] 78 26 1,744 Dccembe 422 6S! - 229 · 382 5611,' - 237 68334t - lSI '" 53 42 686 1,611 7 32 64 182 1.469 gn '06 6 - 188 1,294 January 223 - 64 442 - - 188 1,230 · 1'3 - 10 432 - - 188 1,220 . 12It 13S - 14 418 - - 188 1,206 February 114t III + 3 421 - - 188 1,209 · ,11 ~nt + 24 44' - - 188 1,233 122 - 12 - 433 - - 788 1,221 March ___ 71t '0 13 420 - - 788 1,208 SOt 67 ' - 8 .4.1 ,2 - -- 788 'Ot " - 1 - 788 t:j~ NOTES: Periods muked • I" Ilot ten clays_ Diseharaes marked t were obSCfVCli It Nulr, whlcb is below the moutb of the Wakau_ 937 TABLE 442 (continued) SUMMARY ANALYSIS OF THE YEAR 1935-36 (HIGH FLOOD) (iii) COMPARISON OF COMPUTED AND OBSERVED LOSSES Ten-Day L .- .- -- C~M~~~ __ -- -- --i ~rrected DHTereoco __ .": i~ _J _ i~f~~~_~~~;o~~ll0n ! AbsorplJon I T~lO_b_t:_o~_c_d-+_c_o.m_PU_'_cd-_ob_'C_"'_c_d May I.1I 20 o 20 28 - 8 lS 2 11 62 + IS 138 2 140 19 61 June · 1 218 3 o 221 152 69 289 - I 10 298 236 62 ... l 351 - 3 20 31' 318 56 Jul)' 420 - I 30 449 '01 48 .j 412 - 6 '0 S08 415 23 516 - 12 SO SS4 550 4 August ··· 1 562 31 60 S91 621 - 30 576 - 31 10 61S 694 - 19 September " 61S - 45 80 6S0 189 - 139 i 6S8 - 4S 90 103 891 - 194 693 60 100 133 1,01S 136 - 11 120 185 1,139 nood October ... 181 -108 ISO 829 1,298 over- 81' - 108 896 1.460 :-5:6E4}190 tors 81S - 71 2'0 98' 1,63.5 - 651 inclSC(\ November 169 - 27 300 1,042 1,7.55 - 113 troullh 103 - 1 3S0 1.046 1,796 -7.50 633 + 11 3S0 1,000 1,744 - 744 December 513 39 3S0 902 1.61 1 -709 322 SO 3S0 122 1,469 -7.7 102 61 3S0 SIl 1,294 -761 January 'S 61 3S0 462 1,230 -168 26 12 lSO 448 1,220 -772 I. 11 3S0 446 1,206 -160 February ... 20 82 lSO 4S2 1,209 -757 23 81 lSO 460 1,233 -711 " 92 3S0 451 1,22 1 -164 NOTT-S : The period' Marked • D.n: no t ten days, For the period, m;r.rked t 1M upper di$o:harae .itt w,u III N.uir, 190 km. from Hillel Doleib. and no t lit Sohal Head. The figuru (or net cVlporluion, .b.orption, lind cOfT(Cu:d o bserved Iou I tO cumulatiye. 938 TABLE 443 SUM M ARY ANALYS IS OF TH E YEAR 1936- 37 Ii) GAUGES AND RAINFALL DATA I RAI NFAl.L GAUOES I ! A~~:;:d i T olal T~~;I~dY I';--SO-b-al-.,----'----Hi·~~; --I.-~~$~r - -..~ b-\:o:g~L~ D~lr:b Ar~ ll Head Nasir Abwonc Dolcib miUimc lrcs - jEVapO-ra-li-o ni! sq. km. ~-; ~~ --I 4'95 10·13 11 ·01 4-99 5-28 10'41 11-<)2 10 60 5-39 5-68 10 '70 11 ' \4 ,. 4 2l '0 •' •9 MIlY 5·36 I 5·67 10·71 11' 18 12 7 40 48 ~~l 6')6 11 ' 11 11·.\0 '"8 14 2. '0 60 Ii ,·14 11" 4 11·62 12 28 6 40 87 June ... ; 7-51 8·00 12·26 11·92 3 1 10 9 40 i ' ·84 8·45 1.2·65 12·22 26 t4 48 30 I S. 8·08 8·69 12·90 12·40 ,6 84 20 'tJ"2 July 8-)1 8·98 13-12 n ·$) '0 .3 34 20 194 8-60 9·20 1) ')0 12·65 "16 4' " 67 20 2IS 8·83 9-4) 13·48 12·79 12 60 20 238 AUl\lst 8-98 9-61 13·66 12·9) I. '3 62 10 258 9'10 9·75 1) ·78 13·0) 63 '0 41 10 274 9-22 9-90 13 '92 n ·1l '3 IH '3 10 29) 9-29 \0·02 14·0) n ·n 6. 44 10 308 9·]4 '6 10-18 [-"IS 1) ·)0 t29 26 24 10 326 9-40 10-32 14-26 1) ·) 1 40 I. 90 10 H2 October 9-4. 10-)9 14 '36 13·4) 34 10 354 9 ·47 10-42 14·43 1) ·48 6 I' 20 362 9 -51 10'46 14·4' 13·50 2 1 29 30 369 No\'emberj 9-45 10-46 14·51 13 ·51 3 40 37. 8-93 10·27 14·" 13·52 '0 348 8·24 9·12 \4 '38 1) ·5'J 60 289 ' -51 g'81 13·94 1) ·)3 60 209 6'62 "60 11·88 12·9) " 70 "3 6-06 6-56 11 ·80 12·30 80 70 Janua.ry $-03 6·01 11-15 11·81 .0 54 5·36 5·68 10'91 11·64 .0 49 5' 13 5'4) 10·73 11 ·44 100 46 February . '-04 5·33 • 10·59 11·)5 100 46 4,-·0901 S·22 10·50 11 ·28 100 5'29 10'55 11·25 100 "46 Non : Thc periods marktd • arc nOI Icn days. SUMMARY A NA LYS IS OF THE YEAR 1936-37 (i l) OBSERVED DISCHARGES DISCHARGU • Uncorrected I DeSERVED I N"FLOW Ten-Day [ Cu mulative ~ Corrected Period I! Diffe rence J CumLoul,',tive ! Innow Cumulative ,: Sobat Head Hillet Doleib : Twalor Wakau Total Loss - , I : , - J April 71 46 2' 2' 1 - , - , 2' 104 84 20 45 - ! - - I 45 May --- 104 i 96 i 8 ! 53 - - - 53 · 168 I 131 I 37 .0 I - - ! - 90 213 237 36 ! 126 I - i - - 126 June 337 289 I '8 114 -- I 1 - 174 406 3.6 60 I 234 !1 ( -17) I - 234 448 J77 71 30' - (- 17) 288 July --- 47' · 499 I , 411 i .. 369 - 1 3 1 338 453 46 i 41 ' - 1 J (-42) - 45 m Sl8 45 460 - ,( t- ) 370 ") .01 Auaust .'I. ,' '19 I 36 I 496 -· 510 541 29 I '2' -- !1 ( -'I) ~- 97~ 420 3 432 614 31 SS6 - 110 446 September -- '96 I S7J 23 S79 -3 - 13 - 126 4SJ 601 586 I '9' +1 0 - 125 469 60' I '96 I "9 603 4 + 11 - 110 4.3 October 612 - 603 9 I 61l 6 ! 16 - 88 '24 6 18 608 ,i 10 i 622 1 19 - 62 510 689 672 17 63' 8 2 1 - 33 606 November ••• · 601 612 - II 628 I , 23 - 10 618 496 6JJ -111 ., +35 546 391 "2 -201 '3I1t0 ,I 74 109 419 December --. 314 'IS - 201 109 6' 174 283 · 238 3'1 - 113 4 33 207 203 178t 231 - '3 -51 I - 207 1' 0 1anuary l~l 149 - 34 - 91 ~ 207 · 11' -24 I , - II' Ii -- - " 6 207 92 97 - 13 -128 e 207 - 79 February .... ~f " - , - 133 i - , 207 74 · 66 - 2 -135 g I - 207 SSt '0 + -130 - 72 207 71 March .-- m '9 2 -128 48 - - 134 ! - 207 79 · 6 207 36t 3" + I -Ill I I -- 73 I 207 14 939 TABLE 443 (continued) SUMMARY ANALYSIS OF THE YEAR 1936-37 (LOW FLOOD) (i ii) COMPARlSON OF COMPUTED AND OBSERVED LOSSES CoMPUTED I Corrected I Differenoc -- , I Observed C ~~?~~~ ' Eva-~~-tio-n 1; --A-bs-or-p-tio-n- '- TOTAL 1 Loss omputcd- Observcd ~ ·-r ---1-' --;-- - - - -IS-- ----2S--- ----: --- .. April 26 29 4S - 16 May 41 45 53 - 8 76 81 90 - 9 134 141 126 +15 June 200 209 174 3S 147 2.56 234 22 m ill ill 4 July 334 - 2 10 342 338 .. 381 - 6 10 385 370 IS 423 -10 20 4)) 40 1 32 August 464 - 20 20 464 420 44 :1 497 -31 30 496 ·U2 74 $32 -52 40 520 446 7. September 56) - 68 50 54) 4H 90 600 - 85 60 575 469 106 628 -99 70 599 493 106 OClober . 65) -99 80 632 524 108 664 - 96 90 658 570 88 672 - 93 90 669 606 63 November 635 - 78 100 657 618 39 53S - 60 )00 !in 546 29 392 -42 110 460 419 41 December 227 -31 110 306 283 23 IJ6 -23 110 203 203 0 'j 55 -17 110 148 150 - 2 Janu ary "J 34 -13 110 131 116 +15 26 - 9 110 121 92 35 18 - 4 110 124 79 45 February ... j 14 . + 1 110 125 74 51 j 13 S 11 0 128 72 56 'j 12 10 110 132 77 55 NOTES : Tho periods marked .. arc not ten days. For the peliods marked t Ihc upper d..Ltcharic site "I'll at Nasir, 190 Jun. f/om Hw.:t Dokib, .ad lIot al Sobat Head. The fieuru ror nCI cyaporolion, absorption, and corrected observed loss are cumulative:. 940 I J TABLE 444 SUMMARY ANALYS IS OF TH E YEAR 1937- 38 (i) GAUGES AND RAINFALL DATA i GAUGES L - - I~~~.f"''.L~_. . ___l AS5umcd I Tolal Ten-Day _______- ;-_ Dolcib ' Gross Surr3("'C Period Sobat : . I Nasir Abwone Hill lE.va po rolion ArCIl s4 . km. __- ;I_ _HC_ '_d Nasir A~:onl .-LI'~~:~~ I,- -_.- .- _n~~lIi~e~~ .-=-~'-'-~.' April (3) 4-87 5-16 10'30 10-90 May 5·23 $·5) 10 -47 10 ·98 48 12 I 40 40 5·95 6 ·26 11 ,0) 11 -26 30 34 SO 40 SO 6·7) 7- 13 11 '62 11 -60 52 10 44 8. Iune 7-20 7·85 12 ·23 11·9 7 8 5 85 40 li S '7·-2.n8. 7·96 12 -4) 12 -10 125 98 48 30 12) 8· 12 12-5] 12-24 66 39 16 20 Il2 July 7·80 8-48 12-74 12,)5 64 116 64 20 IS. .1 8'2 1 8·85 l HlO 12 -5 1 7) 21 46 20 18) 8-77 9-]2 13-)) 12 -68 40 45 125 22 22. I AUlUst 9- 10 9 ·68 13 ·69 1) -16 57 43 90 iO 269 .' 9-33 9-96 13-91 13 -24 2] 22 40 10 )02 9-45 10-15 14- 12 13-38 31 70 )9 11 318 September; 9-51 10 '3 1 14 ·)2 1) -47 2 83 IS 10 lSI 9-57 10-44 14-43 13-5] 53 65 S6 10 )69 9·6 1 10-54 14-58 13-61 65 117 76 10 )88 October 9·63 10-60 14·74 13 ·7\ 22 14 8 10 70) 9·65 10·64 14'79 13 ·75 14 52 14 20 858 9·62 10·64 14·83 13 ·80 6 33 91l Novtmber~ 9'49 10'59 14·82 13 ·78 40 S55 9·26 10 '48 14·78 13 '74 50 390 8·93 10·26 14-69 13 ·70 60 360 December' 8·18 9 ·65 14·50 13 ·63 60 295 7·31 8·64 13·92 13 ·44 70 227 6·63 7'33 12·73 11·94 88 106 January 5·81 6·26 11 ·60 12-30 80 53 5·44 5'77 11 '02 11 ·85 90 .3 5·22 5'54 10'77 11 '6 1 110 41 February , 5·02 5·) I 10'60 11 '49 100 39 4'80 5·09 ' 10,)9 11 ·36 100 38 4·76 5·05 10,)0 11·26 80 38 Non : The pcriodJ; Olilfhd • are 1\01 len d.1I)'5 . SUMMARY ANALYS IS OF THE YEAR 1937-38 ( ii) OBSERVED DISC HARGES D ISCHARGES I Uncorrected I Ten-Day --0-8.S-ER-VE;D -IN-fL-OW- -'I Cumulative 1 Corrected Period iI I DifferCDce CU mulative I InHow i Cumulative Sobal Head Hillet DOleib f I Loss Twalor \Vakau I Total I Loss -M-a-y----, , -,.I~--8-9-t-4I----6-2--~---2-7--+----27---' 27 I i~; I n~ ~~ ~: I -~ .. , - 5 -- ! ~~ unc , 318 I 285 ! 33 II I - 2 - 5 - Il 11 8 328 304 i 24 ISS -2 - 3 I - 18 I 137 344 325 19 17. 0 0 I - 18 156 July .. , 389 3.7 .2 216 - 3 0 I - 21 19' · 448 392 56 272 i 1 -. - I - 26 ! 2.6 564 '79 I 85 1 I 3,7 I 0 10 i - 36 ; 321 AUJUSt .. , 553 511 , '2 399 0 -22 - 58 3.1 · 584 5" ,9 458 i 0 - 26 - 84 )7' 664 612 52 510 0 -2) - 107 40) 1 I September . 621 . 579 .2 I m 0 - 12 - 119 .33 63. '98 38 590 0 + 3 - 116 .7. 640 629 17 607 0 I ' -101 500 October .. , 651 .64 - .I.l 59' ! 0 21 - 80 .73 51' · m - 20 57. 0 - 55 '19 702 748 - 528 0 "27 I -28 500 November ... '96 672 I - 7. 472 0 20 8 464 542 058 I -1 16 330 , 18 + I' 351 4" 641 -159 177 I' 23 53 230 December ___ 404 607 - 203 - 26 15 60 128 102 · 321 5" I -194 -220 Il 71 212 8 2.8 302 -114 -"4 12 35 "9 -75 January .. , 1l8! I8J -45 -379 - - "9 -120 · 9. 128 - 32 -411 I - - 259 - 152 89 II) - 2. -43" - - "9 -176 February .. , 08t 8. - 18 -453 I - - - 194 · 52t 468. ' I -- 16 - 469 I - - "9 259 -210 39j 7 -476 - - "9 -217 941 TABLE 444 (continued) SUMMARY ANALYSIS OF THE YEAR 1937- 38 (HlGH FLOOD) (liil COMPARISON OF COMPUTED AND OBSERVED LOSSES Ten-Day ;- ______ •_ __CO MJ'UTE~ -- __1 _·· ____1 '6b~~:1 I Diiference Period : ~~~~~~ Eva:a~~tion 1 Absorption I TOTAL ! Loss I Computed-Observed May Ii IH-T . ---,-- - - .. ''' ,If '- H -1- -~2~ June IS3 4 157 JlS 39 J68 - 4 164 137 27 199 - 7 192 IS6 36 July 246 - 16 o 230 19S 3S 319 -21 10 288 246 42 4(16 -32 20 394 321 73 August .. . 1 486 - 46 30 410 341 69 549 - SI 40 S38 37. 164 " 600 - 64 '0 S86 403 183 September ! 642 - 71 60 631 433 198 677 90 70 6S7 474 183 718 -122 80 676 S06 270 October . 7S9 -122 90 727 SI4 77S -131 120 764 SI9 2I3} Dood 7S1 - 104 180 817 SOO ~t~ ~:~f:r Novembe r 70S - 82 220 843 464 379 trough 648 - 62 260 846 3SJ 49S S40 - 4(1 260 760 230 >30 December 403 - 22 260 641 102 539 221 - 6 260 47S 8 483 89 + 3 260 3S2 - 7> 427 January ". 3S 8 260 303 - 120 4" 19 13 260 291 - 1.52 443 10 18 260 288 - 176 464 February .. - 2 23 260 281 -194 47S - 4 27 260 283 - 210 493 - 4 31 260 287 - 217 So. NOTU: AU lones I rc in millions or cubic metres. The periods marked • are not (CIl days. For the periods marked t Lbc upper di5chal'Jc site wu at Nuir, 190 km. rro m Hillet Doleib, and 110t 11.1 Sobat Head. The 6aures for nct C: V:IIpon.tion, abl.Orptlon, :lnd corTecleci observed 10" are cumulative. 942 TABLE 445 SUMMARY ANALYSIS OF THE YEAR 1938- )9 (il GAUGES AND RA lNFALL DATA v .woo RAINfALL Teo-Day - f Doleib : AG~~;;cl ~ Toll' Period Sobat AbwODE I--~illc~ ' I Nasir J Ab\Yol1~~_ Hill Eva pOl1llJo._n. ._ sqM. ""m . Head Nasir : Dolcib I millin'lOlrcs . l IO':~--I-I .7--1 --1----1-·- April (J) !H) May '·74 10-67 11 ·2) I 29 • 16 40 42 S-S0 10-89 11'36 12 48 40 43 6-12 I l {)() 11·41 14 16 44 47 June , .. 1·12 t HiS 11 -67 H n 40 72 1 ·9] \2·21 12{)3 11 4 8•6 48 JO 11 6 8·44 12·60 12·28 $' 6 8 20 m July 8-90 11·94 12-5 1 II 9$ 8 20 19$ 9 ' 18 1],20 12-67 23 41 22 20 206 9-4$ 13·42 12 '84 $4 80 Il7 22 2)7 Au,ust I 9-66 J) '6) 1) -0 \ 27 27 $7 10 260 9 ·19 13 ·19 1HZ 1.0.J 9J 4J 10 279 9 ·97 1l ·94 13 ·26 6$ $6 97 II 303 Seplembt 10-19 14- NOVOmbe'j 9-45 10-55 14-55 13-60 J 10 $ .0 179 9-22 10-47 14-55 13-60 $0 "5 8-77 10-20 14-50 13 -58 .0 JH December; 7-96 9-47 14·29 13-52 60 20$ 7-13 8-18 13-55 13-27 70 I6J 6-12 6-64 12-03 12-60 SS 6J 5-16 6-17 11-36 12-04 80 $0 Januuy :1 5-48 5-83 11 -02 11 ·76 90 4$ 5-24 5-55 10-74 11'55 110 42 February 5'16 "46 10-55 11 -43 10 41 5-19 5·49 ,10-62 11 ·41 10 41 10-57 11 -]1 9 ., Much (1;1 5-14 >-44 $-06 5-)5 10-'0 11 -34 Non: The periocb m;uhd .. 3.n: Dot ten days_ SUMMARY ANALYSIS OF THE YEAR 1939-40 (ii) OBSERVED DISCHARGES , D"CHAROU I I Ul1correc~cd j DeseRVE D INfl OW Cumulative ! Corrected Ten-Day I- - - "--- Difference i CumulatIve 1 Inllow I Curnu] ",ti. .. \: Period . Sobat Head Hillel DO]eib ! I Loss I Twalor Wa).::au Tota l Loss 2$ ::~j ·r I T - f: 27$9 I 79 lJ. IJ6 , 19 1 - 3 - 8 - 11 180 416 JJ9 77 ; 2. 8 -J - 8 -22 , 246 4" 374 84 Jll -. - 7 - 33 319 July . 487 408 79 4J1 -4 - 10 - 47 384 m 440 73 , 104 -2 - 1$ -64 440 190 $0. 84 l88 -2 - 2J - 89 ! 499 Auaust 1$0 484 66 654 - -20 - 109 I $4$ , 1$' $01 1$ 709 - - 19 -128 I $80 618 168 '0 i 7>9 - - 20 - 148 I .11 September __ $72 lJ8 J4 i 79J - - lJ - 161 632 >92 m 37 830 - + 2 -1S9 I .71 61J . ,71 42 872 - I 13 -146 726 October ... 622 m 47 919 +J 21 -122 797 , .26 '94 j 12 i 9ll 11 24 -87 864 689 674 ! 1$ , 966 14 2. -44 922 November .. 614 .6219 , I 961 13 27 4 7 9>7 m ! -74 887 11 28 +3$ 922 460 61J -153 73. 9 38 82 816 December ... 346 $.2 , -216 $18 10 81 173 .91 27J 437 -164 I "4 12 63 248 602 116t 2$7 - 81 27J 0 (I) 248 m January 128t 138 - 10 I 263 - - 248 $11 lOOt 106 - m - - 248 $0$ SSt 90 - •2 I 2$' - - 248 'OJ February ". 73t 77 - 4 2S1 - - 248 '99 , 1St 77 - 2 ; 249 - - 248 .97 64t I 68 - 4 ! 245 - - 248 49' 945 TABLE 446 (continued) SUMMARY ANALYSIS OF THE YEAR 1939-40 (LOW FLOOD) (iii) COMPARISON OF COMPUTED AND OBSERVED LOSSES COMPUTED Ten-Day Corrected Period Trough Net Observed IC ompDutiCedf"-enco Absorption T OTAL Loss Qbscrved -- --Vo-lu-me- Evaporation --- -- May .. . ; 23 24 2$ - I 1 76 77 7. - 2 ' [ 123 124 136 - 12 June - i 164 168 180 - 12 217 224 246 - 22 I 26$ 272 319 - 47 July --- I 311 3 - 10 32. 384 -60 , 1$7 7 10 360 440 - 80 399 -27 20 392 499 -107 AUi\lst - , 430 - 3$ 20 41$ $4$ -130 457 - 41 30 446 $80 - 134 487 - 53 40 474 611 - 137 September 521 - 72 50 499 632 -133 568 - 86 60 542 671 - 129 604 - 108 70 $66 726 - 160 October . - , ; 627 -11 5 80 592 797 -205 641 -III 90 620 864 - 244 650 - 100 90 640 922 -282 November 636 - 89 100 647 9$7 - 310 576 ·· 71 100 60$ 922 - 311 459 - 53 110 516 816 - 300 December 2~4 - 40 110 35' 691 -331 II I -32 110 191 602 -4Jl 't- n - 26 110 116 m -40$ January __ t l 9 - 22 110 97 - 5 11 - 414 1. - 3 18 110 89 505 - 416 '\1 -10 - 13 110 87 $03 -416 February Hi: - 12 8 110 90 499 -409 -12 - 3 110 95 ; 497 -402 't, -15 + 2 110 97 493 -396 I NOTES : The ~riod. roarhd • are nOI len days. For the period! marked t lbe upper dischatalt silt was &1 Nuir, 190 kIl:! . (rom Hillel D oJeib, and DOt _t Sobat Head. The tiaures (or net evaporation, absorption, lad corrected ob$crved lou :ICe cumulative. 946 TABLE 447 SUMMARY ANALYSIS OF THE YEAR 194 1- 42 (i) GAUGES AND RAINFALL DATA I RAIN~'ALL GAUOES Tp:ri~~)' !- 'SObal Do lcib Total --j Nasir Abwong Hill l!"l Hill ANIl Nasir Abwon~ I sq. km_ Head Doleib millimetre! Apri1(3) ;-~-· ·I- 5-03 10 -02 10·77 May ... j 4·83 5-12 10-21 10·79 , I 40 44 5-65 to·4 1 10·89 190 40 46 ,6.'"4 2 I 6·76 11'36 11-)7 '9 107 25 44 71 Jun, 7·30 7'78 12'0 2 11 -77 64 40 112 7'96 8·5) 12-55 12-12 1" 0 103 JO ISS 8'37 8·96 12-90 12-39 39 4 20 186 July 8·69 9-25 13 · 15 12 -55 27 8J 20 III . ~ ::~i 9-51 13 '34 12-68 21 16 20 236 9·71 13 ·5 1 12-82 16 21 251 August ! 9·\7 9'84 1) ·67 12-94 79 "II 10 272 9'23 9-96 1)·80 13 -08 ,.. 9-31 10·08 13-99 13 -20 "J2 3 10 119 II J09 September 9-35 10-18 14·06 13-29 10 )20 9·36 10·23 I'HI 13 -34 I 61 10 J25 I 9'41 10·30 i4' IS 13 -37 2. 2 10 JJ6 October I 9-47 10-38 14-13 13-41 41 10 346 9 ·:5) la-SO 14,]0 13-4) . 1 12 40 9'57 10-58 14-39 13-48 6 , 20 J57 JJ 369 November; 9-59 10'63 14-47 13 -50 42 .0 m 9'59 10·67 14 -57 I )-53 '0 9-42 10·64 14·61 13 -SS J" 60 38. December: 9·02 10·47 14 ·62 I) -58 60 36J 8-49 ]0-08 14 '53 I) -57 70 J2J . ; 7-51 9-16 14-17 1)-47 88 2" January I 6'98 7-69 13-0] \J -07 80 107 ; 6-00 6-47 , 11·72 12-34 90 57 ·i 5-76 5-93 11 '06 11 -80 110 46 February 5-30 5-63 10·17 II -5 1 100 41 5·22 :5'53 10-58 11 -32 100 40 5'18 5-49 10·5 8 11 -26 80 39 March (1) , 5·32 10·64 11 -14 I Non: The periocb muke4 • are nOI teo daYI_ SUMMARY ANALYSIS OF THE YEAR 1941-42 (ii) OBSERVED D ISCHARGES I; ---DIS-C--H..ARCESTen-Day - -- - -- I IU ncorrected I OBSERVEO INFLOW I Cumulative I Corrected Difference Cumulative I Inflow Cumu lative Period Sobat Head ; Hillel Doleib Loss 1··T-wal-or -;- -W-akau -, Total Loss ~-1- iIl .--.~~--.,-- 11 I- If, ---I-' - J~ .... '- I III ill I H I ill ' -_31 -- : - : m4>8 384 74 347 12 - 21 I 326 July ---. 487 411 76 m 436 77 m 73 I 423 -4 1 -19 - 44 379 '00 -4 - 26 -74 426 "4 m -3 -32 I -109 ! 464 Aua:ust -- -. '43 492 SI 624 -2 -29 - 140 ; 484 "2 S21 31 OS> 0 -26 ! -166 I 489 618 592 26 681 +2 -23 I -187 494 September -- 569 >47 22 I 703 +3 -17 I - 201 '02 572 S53 , 19 I 722 4 -13 -210 m S?9 558 21 I 143 ,4 - 8 - 114 "9 October ,,4 '.7 17 760 - I -210 "0 '97 576 21 I 781 9 + 15 -186 '9' 667 6" 12 793 12 28 -146 647 November. :·1 611 607 4 I 797 13 30 -103 694 6ts 620 5 792 14 II - '6 736 I '62 630 - 68 724 14 December 466 - 169 m I II - 9 7tS 6" 12 36 + 39 '94 384 621 -238 317 10 S? 106 423 ·1 327 60' - 278 39 13 88 207 246 January -j un 382 - 1S5 - 116 -- -- 207 91 202 I - '0 -166 207 41 1I6t I2S - 9 - 175 - - 207 32 February 80t 84 ,1 - 4 -179 - - 207 28 m 73, 0 -179 --·1 - - 207 28 56 I + I -178 - - 207 29 947 TABLE 447 (continued) SUMMARY ANALYSIS OF THE YEAR 1941-42 (LOW FLOOD) (iii) COMPARISON OF COMPlITED AND OBSERVED LOSSES J CoMPUTED Ten~Day Corrected Difference Period I I Observed J~r~:e Eva~~~tion I Computed-Obscrved Absorption TOTAl. Loss ~---"·-I·· III ---=J- -1--~i--~i-1 21 -10 64 -26 123 -31 JUDe: ... I i~ =g g U~ I 196 -24 266 -21 m323 -17 0 306 I 326 -20 July . i =~: Ig ~gi 379 -27 'j 426 -18 463 -26 10 441 464 -17 August ... W ~ Will / 484 - , 529 -40 20 509 489 +20 · 1 S59 -62 30 527 494 33 September --. I ~~~ =~ ~ ~~ 502 58 I 512 64 :~; =~ ~ ~: 529 79 October . ...• 11 1 550 88 666 - 14 80 672 595 77 ... I 687 -67 80 100 647 '3 November 704 -59 90 73.5 694 51 707 -39 90 158 736 22 ' 410 - 16 100 754 715 39 December ... • 590 + 6 100 696 594 102 / 444 29 100 573 423 1' 0 January ... ttl1 ~~ ~: : : ~~ 246 160 91 167 ·t 48 65 100 213 41 172 32 70 100 202 32 170 February .. .. .t l H 100 199 28 171 t 80 100 198 28 170 ·t 85 100 202 29 173 NC1T'£S: The periods marked· arc no! itn day •. Fo r the periods marked t the upper ditcharaa silt wu at N:lsir, 190 Icm. (rom Hillel Ooleib, and not at Sobal HC:IId . The Ii~res ror ntt evaporalion. abJOrptkln. and cor-m;ted obse:n 'ed lou arc cumulative. 948 TABLE 448 SUMMARY ANALYSIS OF THE YEAR 1942-43 (i) GAUGES AND RAINFALL DATA Ten-Day 1------ RAI NrA l. L Assumed . Total GAUOES " '-j ~'!--~\::g-- Gross ! Evaporation . Area Period Sobat Head I Nasir I Abwons I sq. km. H illel Doloib l-- '" millime tres 1- 10 ';S :- - _. ,--_. ~~".~l 1:~~-!,- ;:~:'-I'- :~:~~ 'I 10·97 I 81•2 • 16 .0 38 5·94 6-24 10·86 11 -16 23 40 . ; 7·04 1 7-49 11-88 ! 11 ·72 I SO n 44 1•0•0 lun. ..,! ~:~: . ~:~! :~ :~6 11 -99 49 3 40 \l· 17 18 70 lO "14S' 8· 16 Ij 8'S I 12'83 12040 I 105 44 20 III July 8,42 I 9-09 13·09 12·56 5 1 36 20 199 ::~~ ~:~ g:~~ 12·71 71 26 20 221 12'87 85 91 22 252 August 9-09 9·83 13·73 13·04 70 70 10 m 9·23 10-02 13'89 13 ' 19 8t 70 10 29' 9-38 10-21 14-11 13·37 23 95 II 326 September .. 9·51 10-40 14·)2 13 ·57 41 46 10 J58 ~ :~~ i :g:~ ]!:!: 13·63 7 • 10 372 13·67 l3 10 387 October 9·68 1 10-74 14-63 I) ·7S 6S J7 10 .02 9·7 1 10·79 14·74 13 ·78 IS 20 ::~! i :~:~~ :: :~~ 13 ·77 • JJ '82427 November · : 13 ·75 .0 78S ~:~: ! :g:~~ l::U 13·74- SO 393 13 ·72 60 365 December. 8-10 9·79 14·54 1) ·66 60 301 7·)6 8·85 14·0 1 13·49 70 211 6-)4 7-10 12·63 12·94 88 . 0 Janua.ry 5·70 6·09 11 ·39 12 '18 SO 5-38 $·70 10·90 11 ·74 9'00 '2 $-25 5-56 10·67 11 '53 110 40 February I 5· 14 5-44 10'52 11 ·48 100 19 4·97 5-26 10'44 11 ·38 100 38 4·75 5-04 10·21 \1 ·28 80 38 March (1) ... I 4-65 4-9'" 10-08 11 ·21 Nons : The periods m.ark~ • arc nOI len days. No rainfall obsuvations were made :101 Oolt.ib HUt durin, thU flood . SUMMARY ANALYSIS OF THE YEAR 1942-43 (LOW FLOOD) (ii) OBSERVED DISCHARGES Ten- Day I1_ _. .~ ~~~~~-.---. __ I cor~e~-ted I - .?~S'~ RVfD jNr~I ___.; c umUlalive l Corrected Inflow Cumulative Period Sobat Hill~t I Difference Cumulative Nyandini' Wakau Twalor ! Tota l Loss :: -rf11r :i i ft: ~- - c r ' ofT]: 1uly... 480 403 II~ 77 338 l I - 7 - 3 -28 I 310 508 432 76 414 3 - 16 - 2 - 43 ,. 371 • 591 507 I 84 498 4 -25 -2 -66 432 Auaust 551 492 59 557 2 - 22 -I - 81 470 563 517 46 603 7 -15 -I - 96 507 • 647 603 44 647 21 - 6 + 1 -80 561 September 625 581 44 691 °8 + 7 4 -61 i 630 642 589 53 144 ° 18 9 -34 I 710 651 596 55 799 25 15 + 6 805 October 664 627 37 836 0 31 19 56 892 669 653 16 852 0 )4 21 III 963 736 730 6 858 0 37 24 172 I' 1,030 November 601 665 - 64 794 ° 32 22 226 1,020 494 666 -172 622 0 28 20 274 896 December ;:~ ::; =~;~ ~~ I ~ :~ ~: I :~~ ! :~ 265 527 -262 - 158 - 95 23 530 · 372 • 188t 353 -165 -323 1 - (29) 4 534 I 211 January 1251 158 - 33 -356 - - - 534 178 • ';;+ :~ :;: ~ =~~! = = = ~j: 11 ~~~ February Ut ~; = ~ =~ ~ = = = ~j: g6 34t 41 .. 7 -371 - - - 534 163 949 TABLE 448 (continued) SUMMARY ANALYSIS OF THE YEAR 1942-43 (HlGH FLOOD) (iii) COMPARISON OF COMPUTED AND OBSERVED LOSSES CoMPUTED Teo-Day I'_ ___~ ____- -;_ ___ --;-__ Corrected Difference _ periO_O __i -__VT_or _lO_uu;.:_;.~ ~-Ev-apN-oer-tat-ion~ -A-bso-rp-tion+ _-T-OTA-l ~--O-bL'-oes-Ns ~-~-c-om-p-.t-~---O-b~-N-~-- May n 17 I 18 6 12 38 o 38 42 " 122 I 123 92 + 31• June 172 2 174 117 S1 232 I 233 173 60 289 - 8 281 242 39 July 339 12 10 337 310 27 393 - 19 20 394 371 13 4S1 - 36 30 44S 432 13 August 494 - S3 40 481 410 II S48 - 74 SO S24 S01 17 608 - 91 60 m S61 10 September 6S6 - 102 70 624 630 6 684 - 102 80 662 110 726 - 106 90 110 80S October . 791 - 131 100 760 892 83S - 123 =I~} ~~~_ 170 882 963 814 - 98 220 936 1,030 =: ~~ il~~ November 729 - 61 230 892 1,020 -128 trough 661 - 41 240 8S4 896 - 42 SSI - 2S 240 766 704 + 62 December 389 - 7 240 622 SI6 106 196 + 8 240 444 372 72 ' j' 66 16 240 322 211 III January ... ··.f 32 21 240 294 178 116 j" 20 2S 240 28S 176 109 ' f\ 13' 29 240 282 180 102 February ... .. .t 6 33 240 279 111 102 f' I 37 240 278 170 108 'fi -7 41 240 274 163 III Nons : TM periods marked • are nOI len cb)'J. For the periods marked t tho upper dischilra:e site was It Nasir, 190 kin. (rom Hillel Doltib. and not at Soblll HIt,d. The fi,ures of net e~po"'lion. absorption, I.Dd eorrec:led ol»crved lou ate cumulative. 95~ TABLE 449 SUMMARY ANALYS IS OF THE YEAR 194)-44 (i) GAUGES AND RAINFALL DATA - __ O"U(;ES RAINFALL Assumed , Ten-Day ._. I Gros~ I Total _._-~. Period Evaporation : s.q~ r:~. ___L Sobo.t Head ~ Nasir Abwong -·-·-r·--·_·-I ..... L. April (3) .. i 5·26 10·21 11 ·08 5·42 10-45 11 -14 .s-84 10-79 11 ·27 S5 77 5·75 10·72 11 ·3 1 7JJ0 I 40 45 '0 >0 44 49 ::: ::-1 6·60 11 ·19 11 ·4J ) 40 66 7·12 11 ·69 11 ·75 16 7 )0 82 7·64 12·06 12·0 1 19 61 20 106 July 8·22 12·40 12 ·2 1 2 62 20 \36 8-74 12·74 12·42 17 4 20 169 9-0' 1),05 12·64 73 99 Z2 194 August :.1 9·28 13 -)0 12·80 118 )' 10 218 9·55 13·60 13·07 68 113 10 ,.. 9·74 1] ·74 13· \ 7 28 52 II 27 1 September .. !I 9·87 1J ·83 13 -24 2' l5 10 287 9'96 1 3'~4 13·)2 68 91 10 )00 10·08 1 14-03 1) ,42 10 )IS October 10-19 14'/1 1) '46 "17 ") 10 J29 10·28 14-18 ]3,50 31 97 20 ))J 10')9 14·29 13 ·5 1 II )) 352 November .. 10'46 14,)7 1),54 12 40 )6 1 10'41 14-41 IJ'56 10,03 '0 350 14·40 13·55 60 312 December .. 9·26 14 ·16 \ 3-50 60 246 7'96 1) ·3 1 13 ·2 1 70 145 6·75 12-08 12·56 .. J8 January ... ! 6·05 1) ·24 11 '96 SO ,. 5·6] 10-82 11·66 90 48 5'51 10·57 J \'47 110 46 February 5·48 10·64 11·42 100 45 5·28 10·44 11·)4 100 44 '-06 10·20 11·11 90 44 March (I) ... ! 4·98 10,06 I I -II NOlU: Tb< periods marked· Btl: nOI teo days. No rainfall observ.lioru \Yea made al Doleib Hill durina tbi, 1\004. SUMMARY ANA LYSIS OF THE YEAR 1943_ (ii) OBSERVED DISCHARGES I, DISCHARGES ! Diffcren~ ~:~~~~: ! !C umulative ! Corrttt~ Ten-Day -- - ; , Inflow ' Cumulative Period ; Sobat Head ' Hillet Doleib! Lou I Wakau ; Total ! Loss i . __, _ _____. . Twa.lor I I. May 74 5 , , 109 I) 18 18 J)2 - 9 9 9 Jun. :::,1 !!!! ' 150 +58 67 67 246 234 12 79 79 ! 314 296 18 97 July 428 349 79 116 -, 97 - 7 12 164 469 414 '46 '09 " 231 - 6 - 8 - 26 20' 37 268 -, -12 - 4) AUJUSt ",1 532 492 40 )08 - ) - 16 - 6) '" 580 ,45 35 34) - I -25 - 8. 254 672 621 'I 394 o -)) -122 '" 272 September .. ! 632 '82 '0 444 o -2' - l SI 29) ! 644 '9' " 489 o - 28 - 179 )10 658 623 " ,5,2.4 o - 24 -20] 321 October ... 1 668 638 )0 17 -220 674 648 26 ,.0 - 10 -230 350 'I 74..8 719 '" 29 609 + 4 -226 383 November •.• i ,68 4 688 4 60' +) 22 -20 1 404 698 - 110 49' 12 6) -126 )69 .58 696 -238 257 12 J8 -36 221 n=mb"",1 )68 646 - 278 - 21 12 + 51 + 30 291 4.0 -199 -220 19 6"3 133 -87 222t 298 - 76 -296 133 -163 January .. , ' 128/ 148 - 20 - 316 133 - 18] m 100 7 -313 133 - 190 lcib Hill (lurinr Ih i" fl oo-J. SUMMARY ANALYS IS OF THE YEAR 1944-45 ( i i) OBSERVED DISCHARGES D ISCHAJWES II u ncorrectedl OBSU,Vr.O INFLOW Ten·Day , .. Cumulative Corrected Inflow Cumu lative Period : S b t Hillet !: Cumulalive : ! ~~~_L_..?~:ib Difference I Loss Nyandini 1 Twa lor ! Wakau Total Loss May ... ! 104 6) --r---- 41 41 I ' ;; ; I :: . 41 96 ;i: : ~ 118 JUDO -, - 14 157 II -} -- ,2 - 21 211 430 371 .59 297 0 -2 - 2S 269 July 444 394 50 347 I + I -2 - 7 - 36 3 11 ,I 479 420 59 406 I -2 -12 - 49 357 547 493 54 460 0 -3 - 18 - 70 ]90 A UlUst _I SIS 460 5S SIS -} -23 - 95 423 544 472 72 .590 - } -2' +124 466 6n 535 78 668 - 2 -32 - 156 512 scptemberi ' 67 , Sl2 " 72l i I I -28 - IS4 I 570 '" 52 m 0 0 - 26 - 210 "SS.S 57. 536 . 18 81} 0 0 - II -22 1 '.2 October i '" 558 Il 826 0 0 + , -2 16 610 I '72 '68 830 0 0 7 - 209 621 632 634 - •2 828 0 0 10 - 199 I 62. Novcmbe:j '4" I S74 -57 771 0 0 I . - 180 " 1 " '64 - 111 ! 660 0 0 34 - 146 SI . 3. 6 ,.6 -150 SlO 0 0 - 8. 421 December! 120 482 - 162 348 - - 6"1 - 18 320 ,I 2'2" 35 1 - .6 2S2 13 - IS 237 3t I 2SS j -32 220 -- -- 0 - 15 I 20' January I 162t 178 I - 16 204 - - - - 15 I 189 . J07t .4 + Il 217 - - - - IS 202 p~ruuy:1 SSt 7l 1,2 229 - - - - IS 214 61t ,.; 23' -- --- -- - 15 219 Sit 4. 2 236 -15 221 46t I 46 0 j 236 - - - - 15 221 1 I NO"Rll: The periods marked· :are not ten daya. The observation$ marked, were tilken at Nasir, downstream tM mouth or th. Wakau, 953 TABLE 450 (continued) SUMMARY ANALYSIS OF THE YEAR 1944-45 (LOW FLOOD) (u ;) COMPARISON OF COMPUTED AND OBSERVED LOSSES Ten-Day 1 ' CoMPUTED -- _ ----I '6'b:'= 1 O;ffeds marked t lilt upper di5Charae lile W:IS ,I Nub. 190 kin. rrom Killel Doleib, and not at Sobat Head. The f\aures for ntt evaporation, absorption, and corret lecl obser"\'ed loss are cumulative. i ,.I, _.,;J 954 TABLE 451 SUMMARY ANALYSIS OF THE YEAR 1945--46 (;) GAUGES AND RA INFALL DATA GAl)O F.S R AINfA LL ! Assumed I Ten-OilY I - I - I . - " Gross Total Penod ! Nasir ; Abwong l Evaporation Area Sobat Head Nasir Abwong ! }-hllet DOk,b ! ' .;-;-___1 - .. '-'-~ sq. km. z:.~I(3) :~~ -- ~~!-'~ · : i~:--I--'-5-' m'II":"'~ I. - 4-0 --'----37- 5·48 5'78 10,]5 10·51 94 43 40 40 6·98 7·38 11 -76 11·31 38 JO 44 80 JUDe 6·44 7·29 1\ ·95 I I ·M S6 16 40 74 7·33 7'89 12·24 11 ·81 13 9S I' 30 100 t 7·65 8·28 12 ·53 12·03 )7 14 20 122 July "'1' 87··9391 88··6946 1\22··9779 1122-,2)92 5222 9189 I 2200 117950 8·74 9-28 1)'23 12 '57 54 99 22 219 AUlwt .. 9-G4 9-60 13 '47 12-74 114 9 1 JO 249 . ~:l~ 19:~5 :~ :~~ :~ :r~ i~ :1 ! :? ~~ September 9-56 10-37 14-07 1) ·20 62 47 ! 10 ]]) .. ! ~:;6 :g:~~ :: :~; :~ :~~ ;! ~~ ! : 8 ~l~ October .. ! 9·83 10-82 14-74 \3 ·65 36 28 10 1,4\4 9 ·80 10·86 14·84 1] .7) 42 4 20 J,569 • I 9-79 10·86 14-98 13 ·77 2 6) 33 1,774 November .. 9·79 10·84 15·01 13·77 1 40 1,838 9·74 10·81 15·02 13 '76 3 50 1,718 I 9·57 10·14 1S'02 1J ' 74 60 1,328 December ... : ~:~~ :8:~~ :~:~ n:~i ~ j~! .. j 1·79 9·10 14·70 1J'65 88 199 January ... .. • 1' 16··1185 87·'8514 1123··2976 1113'·4825 9800 18654 5·66 1·04 11 ·32 12·0 1 110 55 February 5·30 5·62 10·84 11 ·54 100 42 .!t ~:~; i:~ :8:!; :: : ~~ I~ j~ March (I) ... 4·18 5·06 10·20 11 ·04 Nons : The ~riods marked .. are not ten days. No n.info.ll obun'atioru were made at Dvleib Hill dluinll: this flood . SUMMARY ANALYSIS OF THE YEAR 1945--4. (;;) OBSERVED DISCHARGES . I I DLSCHAACES . Cumulative I OBSERVED INFI,.OW \' Cumulative Corrected ToPne-rDIoady I' Difference I Uocorrected'----r-- -~·.. ._-- Inflow Cumulative Sobat Hillet , Loss ! Nyandina: I Twalor Wakau Total Loss Head Doleib , , .---- M ay 32t 34 - 1 I - 2 - I - - I - - 2 '8t 55 + 42 +42 - - - - +42 297t 212 85 125 - - - - 125 uno 251t 252 5 I 130 - I I I 5 ! 6 136 338t 284 54 184 - II - 4 - 6 I - 15 '6' 373t 332 41 225 - 7 -4 I - .4 - 30 ' .5 uly ... 416 36. 47 212 0 -3 -- 4 -37 235 · 413 3.7 76 348 - I -4 - 51 343 56. 470 99 447 -I -5 I -19 - 76 371 A u&us' 548 460 88 5J5 - 2 -4 -27 -109 426 · 583 513 70 605 0 - 2 -37 I - 148 457 668 5.6 n 677 0 0 I - 35 - 183 4'4 ptember 619 563 56 733 + I +5 -II - 188 545 657 610 47 780 6 15 I -16 - 183 ,., 675 637 38 i 818 10 .27, I + 28 - 118 644 October 683 665 18 836 6 I 1 33 - 34 802 · 686 684 2 838 0 78 34 + 78 . 16 752 763 - II 827 0 110 37 225 1,052 Novembe 671 701 - 25 802 0 100 34 35. 1,161 665 70S - 40 762 0 100 34 493 1,255 606 704 - '8 664 0 I .8 31 622 1,286 Decembe 506 703 - 191 467 0 84 28 734 - · 408 6.8 -290 177 0 59 40 I - -357 126 -369 -192 0 46 80 ! -- -January {!H 522 -217 -469 -- ! -- -- -· 292 - 141 -615 - - - -- -I06t 161 - 55 - 611 I -February 94 - 13 -694 - · U! 68 -- 8 - 702 - I -- -- -- -- 39 45 6 - 708 - I - I - - - Nons: The pcriociJ matkcd .. are DOt ten days. The dLSchups marked t .ueauumcd to have been taken at N:asir, downstream Ibc mouth o(lhe W.kau. 955 TABLE 451 (continued) SUMMARY ANALYSIS OF THE YEAR 1945-46 (HIGH FLOOD) (jH) COMPARlSON OF COMPUTED AND OBSERVED LOSSES -- -- - COMPUTED 1 Ten-Day - ---- - -- -.- - --1 Corrected Period Trough Observed IC ompDuticffde-"Qn. . No' Absorption TOTAL I loss bserved Volume Evaporation '----T-- May --- t 9 1 0 10 I -2 12 t l 48 0 0 48 42 6 ' t 99 2 0 101 12> -24 June .. .t i 111 2 0 119 136 -\1 t ' 160 - 1 10 169 169 0 t ! 191 - 3 10 204 195 + 9 July 284 - 10 20 294 235 59 J39 - 14 20 345 343 2 404 - 22 30 412 311 41 August 418 - 41 40 411 426 45 540 - 59 50 53 1 451 14 591 - 61 60 590 494 96 September 655 - 82 10 643 545 98 730 - 93 80 717 597 840 - 122 100 818 644 17240 ) October . 929 - I SO 160 939 802 137 Ilood 982 - I SO 280 1.1 12 916 196 over- 1,036 - I SO 400 1,286 1,052 tops November 1,021 - 76 470 1,414 1,161 22>334 J incised 939 + 10 510 1.459 1.255 204 trollah 794 90 540 1.424 1.286 t38 December 694 117 540 1.)51 1,201 ISO 562 145 540 1.247 1.000 2., 378 173 540 1,09 1 751 334 January . -- -t l 172 \89 540 90 1 480 42\ t' 89 199 540 828 333 495 't 47 207 540 794 218 516 February .. .. ,t ; 26 2\2 540 778 2S5 523 t 20 217 540 777 247 530 't t2 222 540 774 241 533 I No:Tl!J ; The pcriod$ fIUI rktd • are nOI len days. Fur the periocb marked t fhe upper diKhOltIJe ~ itc iSlinunlcd 10 bave been a l Nasir, 190 km. (rom Hillel Do lcib, lind Dol a l Sobal Head. The !laure, (or nel evaporation, IIb$Orplio ll . alld corre;;led 17 48. 3"3 0-6 567 - 78 S' S4J 503 40 0 -1 597 - 93 64 568 m 36 0-5 October .. . 620 - 96 72 598 563 33 0-5 6J7 - 97 82 622 600 22 0-) 659 -92 8' 641 637 4 0-0 November 632 - 80 9. 646 653 - 7 0-1 572 -62 9. 604 611 - 13 0 -1 .65 -.3 100 S22 518 + • 0-0 December •.. 319 - 28 100 39 1 383 + 8 0 -1 186 - 17 100 269 276 - 1 0 -1 92 - 8 100 184 192 - 8 0 -1 January ... 41 - 2 100 IJ9 145 - 6 0-1 21 + 2 100 123 130 - 1 0 -1 II + 7 100 118 128 - 10 0 -1 February . 6 12 100 118 128 -10 0-1 -:1 3 17 100 120 128 - 8 0-1 I 21 100 112 129 - 7 0 -1 NOTU ; The periods IDIIrkcd • 1f'C cot teo days. TABLE 455 COMPARlSON OF COMPUTED AND OBSERVED LOSSES AVERAGES FOR SEVEN HlGH FLOODS (1946-47 EXCLUDED) Teo-Day Period May lune lu1y AUlJWt September October ... November December ... l ..u ary February ... 961 TABLE 456 COMPARlSON OF COMPUTED AND OBSERVED LOSSES AVERAGE OF 1WELVE FLOODS (1946-47 EXCLUDED) CO ____- ,-_ _ _ ."-.P, .O._ T.O_ _ ----,___ _ 1I Corrected I Dilfcrcnce PAeprpcreonxtilm,el toor Ten-Day Period Trouah I Net. I . I I ObL!~ed :g~~~~ Cumulative .V_ _ol_um__e_ ___ ,E, _v apo~ _A bsorp_uo_n '_ T_O_TA_L _ I_ ____ JI _ ___ .. __~~:_:aa:_ a~ l Moy ... ! JUDe I~! i ~ i ~ I :~! i j! I ~ ~ I Q-4 ~~ - 1 1·0 I ! m 1:: : II 0·8 0·4 July 309 - 7 J I 313 309 1 4 0·2 364 - II 17 ]70 372 - 2 0,) 418 - 22 27 423 429 - 6 0·2 August 468 - 34 33 467 472 - .5 0-1 ~M ~ :~ 0'0 46051 ng ~1~ g .1 0·1 Septem.ber 604 - 72 592 60S 11 0-2 648 - 83 70 635 664 - 29 O·S 704 - 97 82 689 735 - 46 0·6 October 763 - liS 110 758 818 - 60 0-8 81S - 118 154 8.51 903 - .n 0·6 ' 1 838 - )01 215 952 995 43 O'S November 813 - 61 268 1,020 1.056 - 36 0·4 739 - 18 300 1,021 1,05) .. ]2 0 ,) 613 + 9 32 1 943 977 - 34 0-) December .. 484 43 321 848 840 8 0·1 339 62 321 722 695 27 0'3 206 79 321 606 541 65 0·6 January 106 89 321 516 417 99 0·9 54 97 321 472 )62 110 Hl 25 103 321 449 287 162 I ,) February . ) I 109 321 441 214 167 ) , 3 6· 113 321 440 269 171 I,) 2 117 321 440 264 176 1-3 NOTH : The periods m3tbd • :1ft not ten d.ys. The: a\'C'raRC' !lood over-tops the incUed troup from .bout the beainnin, or Oc:tober untilthc: middle or November. TABLE 457 PREDICTED LOSSES ON THE SOBAT DURING THE TIMELY SEASON FOR VARIOUS DISCHARGES AT SOBAT HEAD AT SOBAT HEAD I CORRESPOND LNG SURFACE AREAS (km' ) --- - - ----- -- MEAN GAUGES T(molil.l1i oLnoss)s 1 PeLr~o nstsa"a e Daily Discharge Total 1- -U~~r~": --I Upper j Lower of Discharge Discharge -'-- - --.-- 10 i 1,800 7·) I H'O I )4 12 -I' 46 2-6 20 I 3.600 8-0 11 ·7 S3 12 6S 1·8 30 S,4OO 8·9 I 12·3 I 93 12 lOS 2-0 40 i 7,200 9·8 12·9 IS3 r Il 166 2,) SO 9.000 10·6 I 13·5 I 223 I 27 I 2S0 I 2-8 I I SECTION III. THE REGIMES OF THE TRIBUTARIES Quite apart from their effect on the main Sohat flood these two trihutaries of the Sohat present some features of interest which are worth describing in detail. The original account has been modified to include the comments of the Jonglei Investigation Team, which are based on field observations and other additional data. (I) THE TWALOR This enters the Sobat from the south just below Nasir. Unfortunately its upper reaches are not shown on the latest 1/ 100,000 maps; these were compiled from aerial photographs and this area was covered by cloud when the photographs were taken. . There is, however, little doubt that the Twalor is connected to the Pibor downstream of the mouth of the Gila. Butcher suspected this in 1939 and suggested that the Twalor took water froll). thel'iliOi in years 962 when this was high. The observations made since 1934 of the discharges in both rivers are shown in Table 458 (p. 966), and they confirm this suggestion. From this table it can be seen that when the combined Pibor and Gila discharges are less than two milliards for the months August to December inclusive, the corresponding total discharge of the Twalor is normally under 100 millions. But when, as in 1934 and 1946, the total Pibor discharge in this period was over three milliards, in spite of the fact that it caused the Gila to flow backwards, then the corresponding recorded Twalor totals were a milliard and two milliards respectively. In the second case its discharge was only recorded till the end of December 1946 although it was then still flowing hard ; and its actual total may have been considerably greater. It is probable that this khor contributed to the excess discharges at Hillet Doleib in January 1947 which have already been mentioned. (II) THE WAKAU GENERAL The hydrology of the Wakau is more complex tban that of the Twalor, and it presents some very interesting features for which only a tentative explanation can at present be given. Their main interest is the light which the bebaviour of this khor appears to tbrow on conditions in the Macbar Marshes, for which virtually no direct evidence is available. One of the striking and unusual features about the discharges of this khor is the fact that it both takes from and, later, discharges into the Sobat a considerable quantity of water. The amount taken averages about 150 millions, whereas no other tributary takes more that 40 millions as a rule during the rising stage. In Table 459 (p. 967) are shown the monthly negative discharges of the Wakau in June, July, and August in different years, during the rising stage of the Sobat. This table also shows the mean Nasir gauge-readings for the same periods, and tbe relation- ship between the two is shown also in Fig. K13. From this it is clear that while the Sobat is rising the Wakau begins to act as a simple spill-channel as soon as Nasir gauge reads about 7·5 ; and that its gauge/ discharge curve is a fairly smooth and constant one. By the end of August, however, though the Sobat continues to rise, this relationship no longer holds good: the flow into the Wakau diminishes and eventually cbanges to a flow out of it into the Sobat some time between the beginning of September and the end of October. Details of this reversal are given in Table 460 (p. 967). It is noticeable from this that although the date of this reversal of flow varies from year to year, the level of tbe Sobat at which it occurs does not vary very much, since it occurs as a rule within a few centimetres of Nasir gauge-reading 10·4. It bears no relationship to the date of tbe maximum reading at Nasir gauge. Up to this point the total discharge from the Sobat into tbe Wakau normally amounts to between 150 and 200 millions, as shown in Table 460. From then unti l the end of the year there is normally an uninterrupted discharge from the khor into the Sobat, as shown in Table 461 (p. 967). This discharge is fairly steady during the' peak' of the flood (which may be defined as the part over 10·4 on Nasir gauge), but as the river drops below this gauge-reading it increases to a brief maximum, and indications are that it ceases by the time Nasir gauge is down to about 7·0. Because the discharge measurements nearly always stop at the end of December, whatever the state of the river, the total amount of the Wakau discharge after the peak of the flood is only recorded in low years; but it appears likely that it amounts on the average to between 150 and 200 millions, or to much the same as the total discharge out of the river into the kbor during the rise. During the peak of the flood the total discharge varies considerably from one year to another. On the whole it tends to be large when the river is high, and small when it is low; but this variation is due rather to the varying length of the peak than to any great variation in the rate of discharge during it. The table shows that in 1946, which was a record bigh year on the Sobat, the inflow from the Wakau during the peak was not particularly large; this fact will be referred to again below. MAIN DBDUCTIONS From these three tables certain definite conclusions were originally drawn, and they can now be amplified from data gathered by the Jong1ei Investigation Team. In the first place it is clear that in order to cause this considerable initial outflow from the Sobat there must be a large area of ground, or length of channel, to tbe north of Nasir where the water surface lies below that of the Sobat during the first part of its rise. Since some 150 millions is spilt into this area annually via the Wakau while the Sobat rises three metres, the area can scarcely be less than 50 sq. kID. In fact, some of the water probably spreads out much more than this, so that the area flooded from the Wakau may easily exceed a hundred square kilometres. 963 10 Secondly, since there is a reversal of flow and the Wakau begins to discharge into the Sobat as soon as Nasir gauge reads 10·4, it follows that at this stage the water-level in part of this area north of Nasir must have risen by at least three metres. The zero of Nasir gauge is at R.L. 386-0, so the levels of this water surface when flow in the Wakau begins and when it is reversed must be below 393·5 and above 396·4 respectively. Moreover, since flow into the river normally continues throughout the peak of the flood, this water surface must continue to rise in sympathy with that in the Sobat and so at the top of the flood be at a level equal to or above 386·0 plus the maximum reading on the Nasir gauge for the year. Thus during the rising stage of a high flood there may be a rise of about four metres in the level of the water surface over part of the area north east of Nasir. Thirdly, since the discharge of the Wakau appears to cease when Nasir gauge has fallen below about 7'0, by this time the outside water surface must have fallen to below that of the Sobat. From these conclusions it is definite that there must be to the north of Nasir a separate water channel in wbich the water-level fluctuates more or less in sympathy with that of the main Sobat channel, but througb a greater range. This much follows logically from the observations given in Tables 458 and 459. From the data collected by the Jonglei Investigation Team, whicb is included in the description on tbe Machar Marshes (see p. 979) tbere can be little doubt that this channel is that of the Machar-Tierbor system, which runs parallel to the Sobat about 30 km. away on the north side, from tbe Khor Machar right through to the Khor Wol ; and that its fluctuations are caused by the passage down it of part of the spill from the north bank of the Baro. BEHAVIOUR DURING 1946 FLOOD A striking confirmation of this theory of tbe behaviour of the water in tbe Khor Tierbor can be obtained from a study of tbe 1946 records. In that year, though the Baro was high, the Pibor was also very higb and its flood was prolonged after tbat of the Baro had passed, so that the Sobat was kept unusually high until after the end of December. As sbown in Table 461 , in the last ten days of December Nasir gauge averaged 10'55, instead of the normal 7·99. This high Sobat was accompanied by second reversal of flow in tbe Wakau, whicb took 10 millions out of the river during tbe last 20 days of December. Unfortunately the observations then stopped, so that there is no record of tbe discharge in January 1947; but it is clear that by tbe end of December 1946 the level north-east of Nasir had dropped below that in the Sobat. This is what would be expected if it were due to Baro spill, so long as the Sobat was still being kept high hy the discharge of the Pibor. Moreover, as already mentioned, the total inflow from the Wakau during the peak of the flood was not above average in this year in spite of the very higb total of Baro spill. There are therefore good grounds for supposing that part at least of the spill from the Baro follows the western branch of the Khor Machar and travels down the Khor Tierbor parallel to the Sobat. It seems likely that all the spill west of Adura Tail will follow this route and, as may be seen from Table 464 (p. 973) in the description of the Machar Marshes, this would amount annually to at least a milliard on the average. Some 400 millions of this return via the Wakau to the Sobat during and after the peak of its flood . The remainder may be supposed to follow westwards the 150-200 millions which the Wakau had spilt out of the Sobat during the early stage of its rise, so that in all rather less than a milliard may be supposed to flow annually into the Tierbor marsh system west of Nasir from channels starting east of it . This quantity should be quite sufficient to flood and maintain these marshes without any substantial additions from spill downstream of Nasir. (Ill) OTHER TRJBUTARJES PROPORTIONS OF HIGH LoSSES AND GAINS CAluu.Eo BY NORTIfERN TRmUTARIES There are a number of north bank tributaries do",nstream of Nasir, but no cross-sections or discharges of them appear to have been measured. It has been suggested that these also have regimes similar to that of the Wakau, and the local inhabitants believe that the western part of the Tierbor system is flooded by Sobat spill passing through these tributaries. In the author's opinion any exchanges between the Tierbor and Sobat systems through these western tributaries are inconsiderable compared with the quantities of water which each receives from other sources-in the Sobat swamps and flanking plain from its main channel, and in the Tierbor system from Baro spill and from the Wakau. It has been shown in this analysis that in high floods each of these quantities may exceed a milliard, and it has boen considered that less than a hundred millions is exchanged between them. Although hydrologically this is a relatively small proportion, it is of course sufficient to fill up all the channels connecting the two systems, 964 and so account for the local view that the Tierbor system is supplied directly fro m the Sobat, since they may in fact be connected during the height of the Hood. This question of whether these tributaries are merely fi lled and emptied by (he Sobat, or do pass appreciable quantit ies of water between it and the Tierbor system, is of some importance in the analysis of Sobat hydrology. Hitherto in this analysis it has been assumed that the Jarge losses and subsequent ga ins observed in Roods above average height are due almost entirely to the inundation and subsequent drainage of areas above and immediately fl anking the per- manent swamps which horder its main channel and which form its main flo od-plain. There are, however, two other possible mechanisms by which they co uld be caused , and these must be considered separately and in some detail to show why they have been rejected. These supposed mechanisms may be summarized as follows : (a) The losses in the rising stage are caused by true spill , through channels in the north bank, over a plaln sloping away rrom the river. The subsequent gains would then be !1 EAST B ANK FARM Oank Mal"kul Month Aerodrome Far,n A\'(:ruge 1950 _._-----_ 1950 1951 1952 19$2 R:lillfall .. January "' j February ... \ no' no, Marclt ... 5 April 62 recorded 51 recorded 29 May .. . 136 31 87 86 June 70 72 45 127 July ... "I 152 198 229 100 16 1 1(,9 August 189 220 203 8 1 HI I ~ I September .,. 95 136 262 102 12 1 119 October ::: ! 69 37 62 72 20 71 November ... I 6 9 December . . .. 1 I ~ - Total • .. i 774 359 538 8 17 UTILIZATION OF HI GH LAN D EXPERIMENTS IN IMPROVEMENT OF HUSBANDRY M ANURlAL AND SPACING E XPERIMENTS The aim of these experiments was not to investigate the possibilities of the application of artificial fertilizers to local crops such as sorghum, but to find out the plant nutrient status of local soi ls and whether this or soil moisture was the main limiting factor. In the 1950-51 season the effect of the application of gypsum (rate 3 tons p.f.) and nitrogen (ammonium sulphate, 100 kg.p.f.) on sorghum crops planted at different spacings was tested. Nitrogen depressed yields of grain significantly, while differences due to manure on yields of straw were not significant. The effect of spacing on the yield of grain was not significant, while the yield of straw increased significantly with the close spacings. TABLE 477 EFFECf OF FERTILIZERS AN D SPACING ON YIELDS OF SORGHUM, 1950-5 1 ! Y IELD IN KG. , PLOT(') Main effect or fertilizers I--arain Slrnw Control .. . 4·7 49'2 Ammonium nitrate . . H 48·9 Gypsum 5·0 51·1 Significant difference (p - 0-05) 1·2 NS I Y IELD IN KO./ Plo -n ') Main .effect or ~pacing I- --G~ . J-~Iraw - . - --.--- - - I ----;- 4·5 i 64·9 40cm. x 4Ocm. "' 1 4-S I lO-3 ~~: ~ I~g~-: .._. ___ '" 4-4 )4,1 OO5 '·· 1 NS 10·4 S_is_ni'f_i c_an_' _d_iffi_e"n,_C_C lP_-__' _ l_ _ Coefficient of variability y. ", · .. 1 24-8 18-6 (') sa. o( plou _ Ij 401h (odd. n. In the 1951- 52 season a huger trial with a sorghum crop was carried out. The crop was p1!'llted at spacings similar to those of the previous year and manured with nitrogen (ammonium sulphate at 150 Ib.p.f.), phosphorus (super phosphate at 100 Ib.p.f.), and potash (muriate of 991 potash at 100 lb.p.f.), as well as dung (5 tons p.f.). The experiment showed a significant difference between spacing treatments, in both grain and straw, but no significant effect with manuring and no significant interaction between spacing and manuring. TABLE 478 EFFECf OF FERTILlZERS AND SPACING ON YIELDS OF SORGHUM, 1951- 52 YIELD IN KO ./ PLOrl ' ) Main effect of fertilizers Grain Straw Control ... '" ... i 7046 34-9 Ammonium suJphllte 7-97 34-9 Superphosphate 9-44 36·2 Mu ria te of potash ... i 8·49 35·7 Fannyard manure .. i 9-44 33-9 Significant difference (p =. 0-05) . . . i NS NS Main effect of spacing I Grain I Straw 40 em. x 40 em. 6-88 "1--- ;~~- -- 80 em. x 80 em. 11 ·27 120cm. x 120 em. _J 7-53 21·3 Significant difference (p - 0-05) I 9'1 Coefficient of va riabilhy (manures) 21-4 Coefficient of variability (spacing) " 14·9 In tbe 1952-53 season a manurial trial was again arranged with sorghum as an indicator crop, but owing to drought and very bad infestation witb stem-borer, the yields on all plots were practically nil. In the same season a manurial trial, involving unreplicated observation plots (4/ 5th feddan), was carried out with irrigated rice .(' ) Nitrogen and potash alone depressed the yields, while phosphorus alone slightly increased them. The application of aU three combined more than doubled the yields. Rates of application were very heavy and the appearance of the crop on the phosphorus fertilized plot suggested that application at this rate interfered with the absorption of iron or magnesium. The nitrogen manured plot and the plot receiving complete manure suffered a heavy attack of He/mintltosporium oryzae. TABLE 479 EFFECf OF FERTILIZERS ON YfELDS OF RICE, 1952- 53 Treatment Rate of application Ib.p.r. Yield in kg./ plot I Control ... '" ~T -- -;~-- -- II 457 Ammonium su lphate 367 Superphosphate (triple) 529 Potassium sulphate : I 252 Complete 1,00> 1,060 The results of these experiments suggest that a heavy application of nitrogenous manure might actually depress the yields, while the application of other plant nutrients without irriga- tion is unlikely to lead to yield increase, as moisture is probably the main limiting factor. On the other hand, when moisture ceases to be a limiting factor, as when irrigation water is applied, phosphorus is probably the main limiting factor among the plant nutrients, while additional response can be achieved by the application of aU three main elements. In the 1952-53 season a smaU experiment on the application of minor elements was carried out. The only significant response was due to the application of zinc (zinc sulphate at 20 lb. p.r.). Response to manganese, boron, copper and iron was not significant. 992 In the 195~51 season a small experiment on observation plots only was carried out to enquire into the value of the native practice of pegging cattle at night on the fields to be sown. The manured plots received the manure of 400 cattle-nights each. Local II/ali type sorghum was used as an indicator crop. The yields of both grain and straw were increased on the manured plots. TABLE 480 EFFECT OFCATTLE·PEGG INGON YIELDS OF SORG HUM . I 9s()"5 I YLELO IN KG. P.F. Treatment Grain Control 301 Manured ... 345 In the 1952-53 season a similar but replicated experiment was carried out, involving pegging of cattle and burning of grass (path: ShilluJc practice). The results, given in Table 481, showed no statistically significant differences due to these forms of manuring. The light rainfall and high stem-borer infestation during this season should be noted. TABLE 48J EFFECT OF CATTLE·PEGGING AND GRASS BURNI NG ON YIELDS OF SORGHUM AGONO, 1952-53 Treatment Control .. . I 100 2,228 Grass-burning 89 2,131 Call1e·pegging .". !I 55 2,158 Significant difference I (p - 0'05) i NS NS EXPERJ.MENTS IN D ATES OF PLANTING AND WEEDING OF RAIN-GROWN COTTON Cotton is the most important cash crop in the area, but local standards of cultivation are generally low. It has been noticed by the officers responsible for the supervision of its cultiva- tion that late planting and late weeding are often responsible for bad yields. In the 1952-53 season a replicated experiment was laid down to ascertain the effect of date of planting on yield of cotton, to confirm these observations. TABLE 482 EFFECT OF DATE OF PLANTING ON YIELDS OF COTTON, t952-53 Variety Yield in I Yield in kg. p.r. kanla" (100) p.r. ___~~ _P_~ _~ ,-____+ __1_ ~_ ___1 _ H Sianificant difference (p = O-OS) ... 82 i 1·8 As may be seen from the above table, the late planting of cotton gave a highly significant reduction of yield, though the difference between the second and third date of planting was only 26 days. The importance of early weeding was indicated by the observation plots trial carried out in the s= season. The early weeded plots gave yields equal to those weeded continuously. 993 .1 1 TABLE 483 EFFECT OF DATE OF WEEDING ON YIELDS OF COTTON, 1952-53 Variety sr 84 SP 84 SP 84 SP 84 SP 84 SP 84 SP 84 MIXED CROPPING EXPERIMENTS Mixed cropping is the general nat ive practice. In the 1951-52 season an observation plot trial was laid down to compare the yields of pure stands with those of mixed stands. These trials indicated that sorghum benefits from interplanting with legumes such as cowpeas, but not with groundnuts, the harvesting of which takes place earlier than that of sorghum, causing disturbance to the soil round the sorghum roots at a critical moment. The legumes su lTered from interplanting with sorghum. Infestation with S/riga hermon/hiea seemed to be reduced in mixed stands. TABLE 484 EFFECT OF INTERPlANTING SORGHUM WITH LEGUMES, 195 1- j2 Crop Inlcrplanted "ilh Yield in kg.p.L Sorghum Pure stand 490 SorGhum Groundnuis 324 Sorghum Cowpcas 704 Groundnuts Pure stand 860 GrOUI.ldnU1S Sorghum 300 Cowpeas Pure sland 280 Cowpeas ."..' 1 Sorghum .". '.1 130 EXPERIMENTS IN TH E CONTROL OF STRIGA HERMONTHICA Slriga hermol1lhi{'{/ is definitely one of the most important adverse factors in local agricullure; for example, in the 1951 -52 season the weight of grai n in 100 plants of sorghum affected with this parasite was 2·4 kg., while the weight of grain of 100 plants not affected was 5· 3 kg. , i.e. Siriga herllloll/hiea caused a 53% reduction in yield. In the 1952-53 season experiments on the possibility of controlling this parasite by spraying with 2-4D (Sodium dichloro-phenoxy-acetate) were laid down. Unfortunately drought and very serious infestation with stem-borer reduced the yields of sorghum, which was used as an indicator crop, to practically nothing. However, the yield of straw was significantly higher on both the sprayed plots and plots from which Siriga hermon/hiea was hand-pulled. TABLE 485 EFFECT OF SPRAYING AND HAND·PULUNG OF STR/GA HERMONTHICA ON YIELDS OF SORGHUM, 1952-53 Yield of Slraw Variety t Treatment I kg.p.f. Agono .. 1 C2-o4nDtr aotl t I.b... 512 Agono I p.f. I' 782 Agono . Hand·pulling 960 S-;g-nifi-ca-nt -d-ifI-i. -re-nc-e--'(p-..,. 0'05) ... ... I 266 994 An experiment on the control of Slriga hern/(J1/lhica by spraying and hand-pull ing in maize was carried out in the same season. The plot was only lightly infested, and increases in yield were not significant. Pre-germination spray ing at the rate of I Ib.p.r. and 2t Ib.p.f. of 2-40 was done. TABLE 486 EFFECT OF SPRAYING AND HAND-PULLING OF STRIGA HERMONTHICA ON YIELDS OF MAIZE. 1952- 53 Variety Treatment Yiel~g~~.~rain l ocal Control 416 Local 2-40 at I lb. p.r. .. . 436 Local 2-4D at 2t Ib.p.f .. .. 516 Local Haod·pulling 534 Significant difference (p - 0'05) NS EXPERIMENTS IN CROP INTRODUCTION For both agronomic and economic reasons, tbe extension of the range of crops and va rieties grown in this area is desirable. A series of experiments was carried out to that end, and the following is a summary of the results. All new crops and varieties can be classified, on the basis of these results, as • probable' , • possible', • improbable', • impossible '. CEREALs SORGHUM Twenty-two varieties of sorghum, local and introduced, were tried, with the following results. It should be noted that tbe local mixed agono and luali types, as weU as a selection thereof, bave given generally better and less variable yields tban any of the introduced varieties. TABLE 487 YIELDS OF SORGHUM VARIETIES, 1950-53 YfILD Variety II Number of I Number of 1950--51 ! i9SI-si -'! . Years Trials kg.p.L kg.p. L I ; I LocAL V AIUETIES --4-:- r--4O-8- -I Agono type (mixed) ~~ -----I. ------I Luali type (mixed) ... (143) 290 Adwel (Agono type) 840 72 456 AlaI (AgODO type) .. . 550 284 417 Awet (Agone type) 940 ChaUa (Agono type) .. - I 290 7176 417883 I' Boweng {Asono type} ". . ! IlO ___~ ~~~ ~~g (Luali ty~~_'_" _1 400 i I~ m 48)--1,- - )()8- -;- 3io-'j - ----:-- Average ----i-- , . ---:- -- -'- - --'-- --- NOR;taViifrus "-.--- ~~!-- (ll2) 170 ! 98 110 Feterita Managil ... ; 20S I 70 137 ~=:~!: ~k~ ::: ! 3~ . ~~ I ~~ Wad Alcer , 350 90 no ~~.J:abaI : :: (ml igg I l~ m I Milo Dwarf " ::: I' 325 117 221 : ~~~:. ... (F) Hg .L IH m I '---A- v-eraB;: ._- ..• ! , 264 : 81 I 171 I "T' ~:~bin-e-----I ~~.---:--1---:-1-'---:--- :---'-'-1--- _ OWloos_nin_.B._O ___ -- . 150 90 ' 120 I 2 I I 120 50 i 85 __ .-'_ 2 i I .- ~. ,:--.-- 120 65 .. ---- 92 , ---j- Average Over-all average. m I 324 88 -2JST-----,--- FiJUra In bntekeu are. of' rl.toon . harvest oaly. 995 12 MAIZe Seven varieties of maize were tried; some introduced varieties were rather better than local varieties. On the whole yields are not inferior to those of sorghum, provided the maize can be protected from pests. Nearly aU the 195~51 crop was lost for this reason. TABLE 488 YlELDS OF MAIZE VARIETIES, 1951-S3 Variety Number of years ECR/ 741 450 190 320 ECR/ 744 4S0 180 31S ECR/ 49O 340 18S 262 ECR/ 49O (ex Shend;) S60 S60 ECR/ 320 Peruvian Yellow 760 80 420 ECR/ 321 Mexican June 420 lOS 262 Local (Control) 490 416 4S3 ._ - -- --- _L- Average 496 193 370 BULRUSH MILLET Three varieties of bulrush millet were tried. Yields were generally superior to those of sorghum, partly because the awned varieties are bird-resistant. TABLE 489 YIELDS OF BULRUSH MILLET VARIETIES, 19S~S3 I ! YIELD IN KG .P.F. ! Variety 1------I- ---.-----.------. --I Number of : 19S~SI year> 19SI-S2 I 1952- 53 I A verage 1 ECR/ 344 ... - --, -;;--- - 7;--' --~44 --- [- -;;; -j- ECR/ 20S A Yellow awned ... : 418 236 327 ECR/ 20S B B~C~WDe:'.." __" _' ~. _____ j _ _ ~S~_J 164 307 Average 546 lSI 390 FINGER MILLET Nine varieties of finger millet were tried. Some of them gave yields at least as good as those of sorghum, although MaJakal must be considered the northern limit for tbis crop when grown without irrigation. The introduction of this crop is worth considering because of its superior storage qualities. TABLE 490 YIELDS OF FINGER MILLET 'vARIETIES, 19~53 Variety _ ___ Y_I_E_LD_ iN_ K_G_.P.F_._ ___ __ . _ I Number of ! 19SI-S2 -i- -I 952-S3 : A vcragc yean ECR/ 377 I 1~ l~ -- - m- --------544 ECR/ 380 363 ECR/ 401 123 ECR!707 103 310 40 151 ECR/ 929 ••• ! 220 114 167 AG 21 ••• 1 110 56 g3 AO 22 105 S2 78 AG 23 220 93 IS6 AO 24 ___ __ . 140 i_ __8 _3 ___ -'~I_ ______ Average -'-----::T--28-3--:---~ : 88 172 996 MISCELLANEOUS CERSALS Yellow manna (Seraria iralica) and pearl millet (PelllliserulI'I glaucum) are worth considering. It should be noted that Seraria, mainly Seraria pallidifusca, is a naturally dominant species on this type of land in the Flood Region. Buckwheat was tried twice without success, Job's tears flowered but failed to produce grain, and crown millet failed . TABLE 491 YIELDS OF MISCELLANEOUS CEREALS, 1950-11 YLELD IN KO. P.F. Variety Numberof 1950-51 1951-52' -,- 1952-53 \ Average years -'j-- . I- - --- YeUow Manna. ECR/ II OO (S~/ar;a /fol/CO) ... 160 100 130 1 Pearl Millet, ECRjI209 (Penniselum flouellm) 180 24 102 Crown Millet F Buckwheat .. . . .. F Job's Tean (CDix' lachryma jobl) F OlL CRopsW SESAME Two introduced varieties have given better yields than local varieties. In the 1952-53 season, when rainfall was similar to that of the Semi-Arid Region, yields were much better than in the 1951-52 season of average rainfall. This indicates that in years of nonnal rainfall sesame is not a promising crop for the heavy soils of the Flood Region . TABLE 492 YIELDS OF SESAME VARIETIES, 195 1-53 Y IELD [N KG.P.f. Variety iI 1. Number or Oi l Oil 1951-5'2 J9S2-SJ A verage years X kg.p.f. --~----~------~--- ECR/ 95 I ... ! 71 294 182 I ECR/ 99 I Local . i i~ l~ :~ I -!--------~ - -- ._ __ m lJ~. A verace 39 196 117 56 SUNFLOWER Ten varieties of sunflower were tried in Malakal, and this crop, easy to grow, can be classified as promising. TABLE 493 YJELDS OF SUNFLOWER VAR IETIES, 1950-53 1I YtaD IN KG .P.F. Variety -lI Number of I Oil Oil 195(}..St 1-~9;;-5;- -i 1952-53 I-Average year> X kg.p.r. , I Dwarf Mancburian 338 83 305 242 30·6 74 ~~~~key S46 75 120 247 29-6 7J 398 flO 395 308 27-2 84 ECR/ 616 .. . 520 310 415 30'4 126 ECR/ 618 .. . 400 225 312 28·7 90 ECR/ 623 .. . 300 225 262 32·0 84 ECR/821 ... 200 260 . 230 30·8 71 ECR/ 834 .. . 440 290 365 30·5 111 ECR/ 841 ... :.:..: I 400 270 335 28·8 96 ECR/ 1059 660 88 374 26'3 98 ------------ ------_. Averaae .. . 427 321 249 309 91 HYPTIS SPICIGERA. This crop, popular in the Ironstone Region, was tried in Malakal but gave disappointing' yields. It should be classified as improbable. 997 TABLE 494 YIELDS OF HYPTIS VARIETIES, 1951-53 YIELD IN KG.P.F. Variety J Oil ' 1951-52 ! 1952-5i I Average kg.p.r. ECR/ 1188 . . . ECR/ 1I89 .. . ::I I~~ - -'--:~ .-- J~ 24·0 II 24·4 20 EB 2 \7'0 20 EB 4 ... I 45 _30_ _ ____ 3_7 24·0 9 Average 65 75 70 15 SAFFLOWER This crop was tried with two different dates of sowing. The early sowing resulted in almost complete fai lure. As can be seen from Table 495, the yields of this crop, even when it is sown late (September), are disappointing. It might, however, be useful as a catch crop grown after rice or as a post-flood cash crop on intermediate land. TABLE 495 YIELDS OF SAFFLOWER VARIETIES (LATE SOWN) 1950-52 YIELD IN KG .P.F. I Variety . _____. _ ---,-- ___ 1 Number of Oil Oil I 1950-51 I 1951- 52 1 Average I years % kg. p.r. ~_~_:e_~~!/r_:_i£_~_'J J : -r- ~ 1 31·6 27 : 23·0 , Average 77 45 18 CASTOR This crop can be classified as improbable. In addition to giving very low yields, it does not survive the dry season. However, trials with other varieties should be continued . TABLE 496 YIELDS OF CASTOR VARlETIES, 1950-52 YtELD 11'1 XC .P, F. r- 1 Variety Number of 1--19 50-51 1951-52 - Avera~;-- -' years ._-- .. _-- -------- - - ':'- ---- -----Ii- -- ---- ECR/278 .. ' j 35 35 35 LEGUMINOUS CROPS GROUNDNUTS Experiments with groundnuts make it clear that though this crop is very laborious to harvest in these heavy soils, provided drainage is reasonable it yields very well and must be regarded as one of the most promising crops. Experiments also suggest that to obtain the best yields close spacing is essential, especially for the bunch type varieties. Response to' ridging requires further investigation. The crop probably benebts from ridging in years of heavier, rather than in years of average, rainfall. On the other hand if the field is ridged this crop suffers at harvest time from increased damage by birds. 998 TABLE 497 YIELDS OF GROUNDNUT VARIETIES. 1950-S3 On. CoNTENT 1950-51 I 19SI-S2 Variety _ O~_8~,Y~EI LOO~H~~~I~FOO H~;:211 -S~n ti~; ,-o;rid;e'-I ~ Over-.II 1 ~W:;fri~JI ~:~lS On ft., I On fta' II O n ,idgcs j Oil Oil SO ~. 70:n- 70 ~m. Average 30~. 30 ~m . 3O:n' 30 ~m. 30 ~m . Average Average 'I X I kg.p.r. '-SO 1em. 1 -30 e1m. -30 e-m. '--- 80 COl. 60 em. 40 em. 20 em. 80 em. --:- --~~~- -~~-- - ~~ - - 200 -r --=- i--:------._----- -- _ . - LoeaI 368 47-9 176 American erect :: I ::: I .~:: I ::: S96 572 504 728 265 I S33 614 S24 47 ·1 247 Barberton .,_ 900 1.123 1.011 600 748 712 904 340 Ii 661 761 630 SO-7 319 Medani erecl 980 640 688 668 160 560 656 488 5(}3 245 KCA I : 1: I 72 1 580 39 1 50-S 197 ___I_ __ ____ - ----;- I , --_. - Average 3S5 738 971 - 584 677 241 I ~~ --l ~ T I 480 49-3 237 SOYA BEAN The yields of this crop, though comparable to yields of local cowpeas, were not high enough to classify it as more than an improbable cash crop. It might, however, be considered as a promising crop for local consumption. TABLE 498 YIELDS OF SOYA BEAN VARIETIES, 1950-53 YIELDS IN KO. P.F. Variety , Number of , Oil Oil 1950-51 _1,951-52 1 _~~:2-S3 .l~ra&e I years % kg.p.r. I ---I - -'-' .. ECR/ 124 _ 100 90 ! 95 I 11-5 II i ECR/ 262 _. . 180 F 90 13-4 12 ECR/ 299 110 80 95 16-3 15 ECR/759 10 10 17-5 ECR/ 973 _ 243 90 400 244 18-7 46 ECR/ 998 . 140 300 220 16-4 36 ECR/ 999 _ 180 250 215 16-8 36 ECR/ IOOI I 30 30 18'6 Average -- - 105 187 125 20 CoWPEAS Six introduced varieties were tried. None of them was outstandingly better than local varieties. TABLE 499 YIELDS OF COWPEA VARIETIES, 1950-53 YIELD IN KG.P.F. Variety Number of 1950-51 1951- 52 1952- 53 Average years --- - -- Local 122 260 80 154 ECR_n 7 69 180 90 113 ECR/ 78 .. . 150 210 20 127 ECR/ 169 -- - I 189 140 165 I ECR/ 939 ! 150 200 175 ECR/ 94O 135 170 152 I BA 13 ... --. ! 60 60 60 I- Average ... 114 169 109 135 MISCELLANEOUS LEGUMINOUS CROPS Of the introductions, velvet beans, green gram, black gram and tepary beap must be considered the most promising. Cluster bean and babun (Vigna vexillata) gave yields at least as good as cowpeas, and must therefore be classified as probable. Chick peas and pigeon peas (the latter do not normally survive the dry season) should be considered as possible. The value of the former as a catch crop after rice, possibly without additional irrigation, should be IlIken into account. Haricot beans and tick beans were failures. 1000 TABLE 500 YIELDS OF MISCELLANEOUS LEGUMINOUS CROPS, \9lO-Sl 195.o:5~~~1_ ---Y-IE-L-D Crop TIN K-G.-P-. ~·. --- ---- ---- Number or years __1-. 1951 - 52 J9S2- S3 I ~vc rage -- I Bambara groundnuts 285 12 148 I Tepary bean 349 140 260 249 ! Green gram ECR/ 6S 171 2 10 220 200 Black grnm 350 230 290 Ouster bean 281 45 163 Babun ,,- 250 100 175 Haricot bean F Tick bean F Chick pea 80 80 Pigeon pea -- I 18 65 12 Velvet bean 281 520 400 FIBRE CRoPS COTTON Seven varieties of cotton , both long staple Egyptian type and short staple American type, were tried, with and without irrigation, and the results are given in Table 501. It should be noted that these varieties were grown in a badly drained field, and therefore suffered consider- ably from waterlogging. In the 1951-52 season the irrigated plots received approximately 3,000 rn3. of water per ' feddan between 21st November 1951 and 5th March 1952. In the 1952-53 season approxi- mately 2,500 m3. of irrigation water was given, but irrigation was stopped at the end of December. One of the main objects of these trials was to find out the blackarm resistance of these varieties under Flood Region conditions. In the 1951- 52 season of average rainfaU, only the varieties with B2 and B3 gene proved resistant (BAR 73, BAR 78, BAR 12/ 3, BAR 14/ 25 and BAR 4/ 16) while the other varieties, including SP 84, BAR XL, BAR XA and BAR NT 96, were aU heavily infested. In the dry 1952- 53 season blackarm infestation was absent. 1001 TABLE 501 YIELDS OF COlTON VARIETIES (YIELDS OF SEED COTTON)_ 1950-53 SP 84 347 116 23 1 5- 1 2 482 482 10-7 BAR NT 96 230 30 130 2-9 128 562 345 7-7 BAR 73 106 106 2-4 BAR 78 196 109 152 176 518 347 7-7 I BAR 12/3 186 186 4-1 --- -------j- - -I A_;_.Ve_rag_e _______ I_ _34_ 7_ ___ :~ _____~ J_~7~___ ~-~ 149 521 293 6-5 ____ --I--- ---i- ._-_._. Average I kantan (31S) p_f. -- ----------- -- EoYPTlAN VARJEIDS I Domains sakel 162 162 I-I BAR XL 35 35 0-3 108 I 108 0-8 BAR XA_ 298 298 2-1 BAR 14/25 17 17 0-1 166 I 654 410 2-9 BAR 4/ 16 39 F 19 0-1 -______1 _- :. ___ -'--__719 5-1 ------1- _ --''-____- '-_ ----- Avera", 37 23 0-2 183 686 339 2-4 JUrE Jute was tried in the 1950- 51 and 195 1- 52 seasons on the high land. In the 1950-51 season variety ECR/ 576 yielded only 11 8 kg.p.r, of clean fibre, and in Ihe 195 1- 52 season the three varieties tried (ECR/ 584, ECR/585 and ECR/ 612) did nol grow ta ll enough to be worth harvesting for fibre. The yields were better on intermed iate land with subsidiary irrigati on (see p. 1012), D ECCAN HEMP Tn the 1950-51 season Deccan hemp was a complete fai lure mainly owing to heavy infesta- tion with flea beetle. In the 1951-52 season the va riety ECR/ 985 yielded 210 kg.p.f, clean fibre. FORAGE CRoPs Though there is no immediate need for forage crops in this part of the Jonglei Area, the need may arise when, as the result of the Equatorial Nile Project, the area of riverain grazing is considerably reduced, A number of forage crops were therefore tried in the 1950-5 1 and 1951- 52 seasons, The crops were cut at the stage most suitable for fodder (generally the seed fo rmation stage), Tbey were weighed after cutting and after they had dried out in the fi eld . Losses of fine tissue (leaves, etc.) in handling the dry crop are unavoidable, and therefore the difference between fresh and dry weight represents loss not only of moisture but also of part of the plant tissue, In the 1951-52 season it was found impossible to ascertain tbe dry weight of some fodder crops as, owing to late rains and high humidity, they rotted before they dried out. Sudan grass and sweet sorghum, the two most promising fodder cereals, gave yields as shown in Table 502. TABLE 502 YIELDS OF SWEET SORGHUM AND SUDAN GRASS, 1950-52 YIELD IN RO.F. F. I I Y,UD IN KO.P.F. I 'NumberOf 1951-52 AveRAoe kO.P.F. 1950-51 Number0rj Variety ! cuts _____ cuts I fresh . dry i I fresh dry fresh dry +---'-, ._---!--- -,'--- ---.I ------ Swttr ! SORGHUM Swaziliutd I 4,905 2,860 5,342 2,690 5, 123 2.775 Island 4,420 2,843 4.990 2,01 9 I 4,705 2,431 Red Amber ' 1 I 4,180 1,710 4,180 1,710 Columbus ... 4,400 1,800 4,440 1,800 Minnesota 4,780 2,320 4,780 2,320 Medani 5,200 2,800 5,200 2,800 Evelyns J 5,600 2,780 I 5,600 2,780 Fadasi 5,260 2,880 5,260 2,880 Abu Hegtein 3,640 1,816 3,640 1,816 Sudaniensis 1,100 400 1.100 400 VcrticutiBorum 2,620 1,040 2,620 1,040 Avence ·· 1 4,662 2.851 4,283 2,023 4,241 2,068 SUDAN GRASS ECR/ IOn 3,000 1,930 3,220 1,320 3,110 1,625 I!CR/ I201 i 2,740 1,410 2,740 1,410 Yellow I 3,490 1,420 3,490 1.420 Red ::: I 4.480 1,880 4,480 1,880 Brown L_ _ 4,860 2,340 ,I- -- 4,860 2,340 I Avera,e .. . , 3,000 1,930 I i-3,758 1,674 I 3,736 1,735 Table 503 lists the yields of leguminous crops harvested for forage, i.e. at tbe seed-forming stage. It shows that some-legumes can compete with sweet sorghum and Sudan grass for soiling, but on drying out lose not only moisture but a good deal of the finer plant tissue, which in fact is the most nutritious part of the crop. 1003 TABLE 503 YIELDS OF LEGUMINOUS FORAGE CROPS, 1950-53 YU!LD I N XO.P.f . I y~~~ ~N I Y~~~.:.N No. of Crop 1950-51 _~~1-52 _ _. !~2-53 Average Obser- vations fresh dry 1 fresh fresh dry fresh - - - - Cowpeas ... - - 3,540 I 3,079 I 1,295 3,310 2 Soya bean I 3,680 1,060 - I - - 3,680 1 Tepary bean ... 1.450 644 181 1,336 360 991 I 3 Grceo gram 3,148 1,258 1,350 2,000 940 4,166 3 Cluster bean 608 283 640 - I PhjlJipisara - 624 2 1.644 191 - - I - 1,644 1 Babun - - 1.960 1,400 360 1,680 2 Sword bean - - 3.920 - ! - 3,920 1 Velvet bean -- - 3,920 -- - 3.920 1 Dolichos lablab i - 3.840 - i - 3,840 1 In the 1951-52 season an experiment was laid down to investigate the possibilities of growing sweet sorghum and ordinary local sorghum together with leguminous crops as a forage mixture. The sweet sorghum mixture was cut when the sorghum was forming seed, while the ordinary sorghum mixture was cut after the sorghum grain had been harvested separately. It is interesting to note that yields of sorghum grain were generally above average. TABLE 504 YIELDS OF FORAGE MIXTURES, 1951-52 YlaD IN KO .P.P. I Crop Grain fresh Fi ?er dry -~_ No. of CU .. Sweet sorghum and Phillipisa ra ... Sweet sorghum and Do lichos lablab . I I~ :~~g i "U~'r'-" --­ --. -.----·-----,1--·-.. --~I :~-·-i~;o Average . i ',- I '- -;:;60 --1'---- --- Sorghum Agono and Cowpeas 285 2.060 Sorghum Agono and Green gr3m 645 2.820 1.860 Sorghum Agono and Cluster bean 390 3.480 Sorihum Agona and Velvet bean ll5 3.960 22.,232100 I' Sorghum Agono and Phillipisa ra ." 480 7.050 4,980 Sorghum Agono and Dolichos lablab 540 3, 120 2.880 I Ave rage: . 446 3.148 2.595 Besides specially grown fodder crops, stovers of numerous crops have a value as animal fodder. In Table 505 we give the yields of stovers, averaged over three years and over all experiments of the main crops tried. TABLE 505 YIELDS OF STOVERS, 1950-53 Average dry weight I Number Crop I of stover of kg.p.f. Observations - -_ ... -_.- '-.. ----1- Sorghum Agono 1,905 13 Sorghum Luali . 1,586 4 Maize 1.506 2 Oroundnuts 175 20 Cowpeas .. 384 15 Samples of tho main forage crops and stovers were sent flY.' analysis to the Faculty of Agriculture, U.C.K. Table 506 gives the average analytical results. Unfortunately no digestibility trials could be performed and therefore these figures are of only limited value. 1004 TABLE 506 RESULTS OF CHEMICAL ANALYSIS OF FODDER CROPS AND MI XTURES , No or I I ' Crop or Mixturo I Sa~ples Mo~tLlre Crud~ Ether Crude ! N·Free . Averaged % Protean E.'lot;iSl. Ministry of AJl"iculture RcseJtch Divis ion. 4. OTHER EXPERIMENTAL FARMS ARIELBEK Experiments at ArielOOk were started in order that a rough assessment might be made of the yields and problems under conditions prevailing in the southern part of the Flood Region and in the neighbourhood of the Eastern Plain, which was thought to offer special grazing possibilities. They were discontinued after the first season, as it was found that the personnel available was insufficient. The results of analysis of soil samples from Arielbek are as follows: lOIS TABLE 524 RESULTS OF ANALYSIS OF SOIL SAMPLES FROM ARIELBEK MECHANICAL A NALYSIS Depth StOCe5 and gravel I Coarse sand Fine sand Sill Clay ___fl. _ _ 1_' __ .~ __ X X X X ()"I 21 35 42 1-2 24 34 34 2-3 24 31 39 3-4 17 27 46 4-5 IS 31 41 5-6 15 14 28 39 CHal1CAL A NALYSJS Depth pH SailS , CaCO, Sodium Nitrogen Carbon ft , --., X X value p.p.m. --' . X ()..I 8'32 0,027 0·4) 277 0,254 1- 2 9' 13 0·048 0 '48 13 217 0,203 2-3 9,32 0'069 0·95 17 284 0·240 3-4 9,)6 0'090 1'50 16 200 0,167 4-5 9-36 0,081 2·40 17 2 10 0,138 5-6 9'26 ' 0·090 2' )0 15 175 0,097 The rain-gauge at Arielbek was not installed until August 1951. Records for that year are as follow: September J08 mm, October 92 mm. November 42 rom. December Nil Previous rainfall in this year was definitely below average. Crop yields were poor, largely for this reason. CEREALS Yields of the cereal varieties sown were as follows : TABLE 525 YIELDS OF SORGHUM AT ARtELBEK. 1951-51 Variety ---1i S-ow-n on 29 .5.51 I Sown on 27.10.5 1 :h'[(,'-- 57--- -1--- 105 Akur<') 55 100 Challat·, 10 I ~g ~~~:~a ~~Ttb~~ ::: ! 11~ ! J3 S Feterita Maluk .. " I 45 Feteri[a Suki . . SO -, "',-" Average ... 1 53 I 76 I (') Dink. "atlety. (") Shliluk variety. As can be seen from this table. the yields of the late sowrf (angul) crop were better than those of the early sown (rap jal1g) crop. The yields of other cereals were also very poor. as can be seen from Table 526. 1016 1 'j TABLE 526 YIELDS OF OTHER CEREALS AT ARIELBEK, 1951-52 ____c ::~ ____ J_ Varitty Yield in kg.p.f. Maize . .. . Local 35 Bulrush millet .. . :j ECR /J44 65 Fiogermillct ECR; 117 60 Rice ,_, .- I Congo 065 35 OIL CRoPS The yields of oil crops were also low, as shown in the following table, though the castor yield was surprisin gly good. TABLE 527 YIELDS OF OIL CROPS AT ARIELBEK, 1951-52 Crop . Va riery I Yield in kg.p.r. ;c::------..- .: I ~ ---I--'-;-'-- Sunflower "'1 Dwarf Manchurian 190 Striped Grey I' 210 E.'( Kodok 105 A\rcrage . - 168 Safflower ::: 'I' Debeira ECR/ 601 Ii 70 Castor . ... ECR/ 278. 240 LEGUMINOUS CROPS Yields of leguminous crops were as follows: TABLE 528 YIELDS OF LEGUMINOUS CROPS AT ARIELBEK , 1951-52 Crop Variety Yield in kg. p.r. Groundnuts local 202 American erect 200 Average 201 Soya bean ... . .. ECR/ 971 150 Bambara groundout . 100 Cowpea Local 25 Tepary bean 80 Cluster bean 45 Green ,ram ECR/ 65 60 Haricot bean . ..-. I Local 20 FmRE CROPS The yield of colton was unexpectedly good, the variety SP 84 yielding 310 kg.p.f. of seed cotton. Considering tbe poor rainfall, the yield of Deccan bemp was also surprisingly good: ECR/985. 150 kg.p.f. clean fibre. ECR/986. 170 kg.p.f. clean fibre. Yields of experiments at Arielbek were poor not only because of bad rainfall distribution but also because tbe inevitable lack of supervision resulted in standards of cultivation little above those of the local inhabitants. FANGAK During tbe 1951-52 season the Team carried out, in collaboration with tbe District Com- missioner, Central Nuer District, a series of trials on observation plots at Fangak. 1017 Rainfall at Fangak during the 1951-52 season was well below average: Rainfall, 1951-52 578 mm. Average 1,1 I I nun. Yields were as follows: TABLE 529 YIELDS OF OBSERVATION PLOTS AT FANGAK, 195 1-52 Crop Variety Yield in kg. p.r. -.----- ClREALS Sorghum Dair 410 Bilwich 450 Agono 433 Feterita 1931 225 Lafon 105 Olosingo 122 Average "' 1 291 Bulrush miUet I ECR / 344 195 Finger millet ECR/ 377 90 H I OIL CROPS ! Sunflower .. ' I. D warf Manchurian l 60 Striped G rey . 480 Ex Kodak 280 Average , 273 LEGUMINOUS CROPS , Groundnuis l.oc:ll 230 So)'a bean ECR/ 973 120 Cowpeas . :I Local 200 . Tepary bean .H 180 Cluster bean 140 G reen gram i 180 The standard of cultivat ion of these plots can be described as moderate. KODOK In the 1952-53 season the Team supervised a few observation plots at Kodak , the experi- ments being carried out by the local Agricultural Officer. Rainfall during the season was ' above average: 1952-53 814mm. Average 663 mro . Pests, in particular birds, caused considerable damage to some crops. TABLE 530 YIE LDS OF OBSERVATION PLOTS AT KODOK, 1952-53 , Crop Variety : Yield in kg.p.r. -_._----- , --_.- --, ._- CEREALS Sorghum . ..• J Agono 17 Luali 176 felerita 56 Average ... .1 83 Bulrush millet i 1\3 Finger millet ... F Rice (rain-grown) 65 OIL CROPS I Sesame . 1 F Sunflower !D WS~~G~~;jan 106 262 , Ex Kodak 308 Average 225 LEGUMINOUS CROPS i Groundouts Local 170 Soya ~3n ... 20 Cowpeas '" Loca l 239 Tepary bean ' 1 Green gram .. , 10 76 I Low yields were partly due to low standards of cultivation. lOl8 NOTES AND REFERENCES (I) Fereoson. R , • Note on Experiments in Tonj Aweil Area 1942-16', Sudan Government Reporl. Unpubl ished . (") Molt. T. R. O. (.) Moir. T. R. 0., ' Malak.' Agricultural Station Reports, 1952-53. Report on Rice Cuhivalion Experiments ', Sudan Govern· ment Report . Unpublished . , (I) ~6~rJit:~;~e:i~~~t~~~~~~~fi~~: ~~~c~~~~n~tb~Y cl~:m9c~lc~~::i~~i~.C~:d~~~ld5~~tP~!::s~~fl; 9:el -~~t~r~:d lb; mechanical edraction. (') For further details see: Moir. T. R. G., ' Report OD Experiments in Rice Produ..:tion 195 1 ', and' Malakul Agricuhu ral Station Reports 1952-53 ' (includ ia& : • Report on Rice Cultivation Experiments '; • Waler Requirements of Ri~ Crops UDder Pump Irric-tion ': • Fibre Crops '). 1019 CHAPTER 6. GRASSLAND EXPERIMENTS. I. INTRODUCTION The most obvious, practicable, and least expensive alternative to ri verain pastures no longer available under the Equatorial Nile Project is to make use of the vast areas of grassland to be found inland away from the rivers. The princi pal limiting factor appea rs at first sight to be the absence of adequate water supplies during the dry months of the year. Our in vestiga· tion of domestic water supplies indicated that there would be little or no difficu lty in providing suitably distributed water-points, in the form of either surfa ce storage or deep-bo re wells. It remained debatable, however, whether these grasslands were sufficien tly pa latab le and nutri- tive for pasture, and if they were, whether there was adequate grazing to support. for the period required, the numbers of animals displaced from the river-front. Trials and experiments were obviously necessary to determine these very important points, as wel1 as to find ou t whether inland grasslands could be improved by any of the vari ous methods suggested in Chapter 6, Volume II. The only other alternative-apart from alternative livelihood in the form of crop husbandry or mixed husbandry-is the provision of irrigated pastures, either gravity or pump irrigation schemes, which a rc notoriously expensive to instal1 and run. Irrigated pasture, however, might be the only possible remedy whatever the economic implicat ions, and trials and experiments had to be carried out in the establishmen t, water·duties, herbage production, and value of irrigated pasture grasses, both indigenous and exotic. It must be noted that the Jonglei Investigation Team was specifical1y instructed by the Jonglei Committee to carry out only such short-term trial s as might give indicalions of the answers to these problems. More precise answers could on ly be expected from long-term , trials for which there was clearly insufficient time. It must also be appreciated that the post of Pasture Research Officer on the Team was only fil1ed at the end of 1950, five years after the start of the investigation and two years after the present Team had begun its programme of work. Moreover, the Veterinary and Pasture Section of the Team was engaged in more general survey work carrying its two members to all parts of the Jonglei Area, and it was not possible to give all experimental work the time it most certainly deserved. Local Sudanese sta ft· had to be trained, and it was quite impossible to give constant or adequate supervision to their work on trials which extended from Kosti to Bor, a distance of about 570 miles by road. The experiments and trials recorded in tbis chapter are therefore in many cases of a very preliminary and superficial nature and the results must be confi rmed or otherwise by a more elaborate programme of experimental work. They provide nonetheless some idea of the herbage production, nutriti ve value, palatability, and stock-carrying capacities of the main types of pasture found in the Jonglei Area, information which was completely lacki ng at the ou tset of our investigation. 2. NATURAL PASTURES THE EASTERN PLAIN Tbe Eastern Plain lies between the high land along the east side of the Bahr el Jebel and the high land along tbe west banks oftbe Khor Veveno and Pibor River, between latitudes 6° N. and 8° N. approximately. At the outset of our investigation existing maps of the area showed no features except a few irregular water channels along the southern and northern limits, the Khors Geni and Tuni to the east, and the Khor Amwom to the west. The possibilities of this vast plain as an alternative to riverain grazing in tbe dry season were first given prominence by Harrison (T.I.R., Appendix Ill) after he had made a hurried tour round the perimeter in 1948. This was in tbe middle of the dry season when most of the coarse grass had been burnt off and the extent of regrowth was impressive. Local information suggested that tbe whole plain was .. a continuous thick cover of Hyparrhenia nt/a witb a sprinkling of Andropogon gayanus and a trace of both Panicum spp. and Sporobolus pyramidalis " and tbat the regrowth in the centre of the plain was as good as, if not better than, that seen from the motor road. The need for a thorough investigation of the plain in general, and of the quality, quantity, and reliability of Hyparrhenia regrowth in particular, was ful1y appreciated, but the personnel was not available. Eventually work started in April 1950 under the Veterinary Inspector, advised by the P·asture Research Officer, S.V.S. (Harrison), and with the co-operation of the Dean, Faculty of Agriculture (Boyns). The coUection of samples of grass regrowth over such a large area, in which the few roads are only open in the dry season and in which the grass 1021 grows from widely spaced tussocks about 20 cm. high (like large molehills). presented many new problems. Some of them were only partly solved, and inaccuracies in the data collected had to be accepted and allowance made. The main objects of the investigation were to determine: (i) The distribution of dry season regrowth after the coarse rains season growth of grass had been burnt. (ii) The quality and quantity of dry season regrowth at all stages after burning. (iii) The reliability of such regrowth in different areas and at different seasons ; and hence (iv) The general stock-carry ing capacity of the plain in the dry season, January to May. NATURE OF THE INVESTIGATION To give an indication of the quantity and quality of regrov"th produced, samples of re- growth had to be collected, weighed, and analysed over a number of dry seasons. A series of sampling units was erected around the perimeter of the plain, accessible from the motor road passing through Ayod, Kongor, Bar, Pibor, Akobo, Waat, and back to Ayod, a distance of 490 km. Since the raised road alters the natural surface drainage and hence is likely to affect grass regrowth, each sampling unit had to be divided into two blocks, Block I being on the side of the road with greater flooding and Block II on the side of lesser flooding (see Fig. L 3). Each block was a square measuring 8 m. by 8 m., protected by a wire-netting fence, and was sub- divided into four quarters ca]Jed Plots A, B, C, and D (each 4 m. x 4 m.). The intention was that each plot within a block should be subject to a different treatment, namely: A to be cut every two weeks. B four weeks. C ,. eight weeks . D .. twelve weeks and at the end of the experimental season. By taking samples at different intervals and at different stages of growth it was hoped to get an indication of the best system of pasture management. Sampling was to consist of: recording < the average length of grass regrowth ; cutting the regrowth, with scissors, to ground level as though it had been heavily grazed ; weighin g the collected sample as soon as it was cut; storing it in a special collecting basket made of very fine mesh wire-netting in order to dry it; weighing the sample when dry ; retaining the sa mple in a cloth bag for analysis. The ana lysis of several hundred individual samples was not considered practicable or necessary. and it was decided to bulk all samples having simil u treatment and have them analysed on a dry matter basis. This was the intended plan. but it could not be followed exactly for vario us p'ractical reason~. First, by the time the motor road had opened (late December) some of the areas where units were to be sited were already burnt and the regrowth partly grazed, while other areas were still too wet to burn. This meant that plots baving similar treatments would not all be comparable because of different dates of burning. Secondly, it was soon found that two weeks was too short a period to permit of sufficient regrowth for sampling purposes. In addition, delays due to transport troubles aF\d to other necessary work made it impossible to work to a fixed time- table; but taking the date of burning as the sta rting point, plots generally were cut as follows: A after 5, 8, J 2, and 16 weeks. B after 8 aod 16 weeks. C after 12 and 16 weeks. Dafter 16 weeks. Because of the difficulty of bulking like samples, all samples from the 1950--51 series were analysed separately. To relate the information gathered from the sampling units to actual visible stock-carrying capacity an area of sixty feddans of apparently typical plain was fenced in April 1950 near Pengko in Bar District. (We say' apparently' because subsequent investigation revealed that this area actually lay within a wide drainage depression .) This area was divided into two halves, one being treated as a single paddock and the other sub-divided into three paddocks each often feddans. A herd of cattle was bought and a cement-faced brick trough erected beside the well at Pengko village for watering the herd during the dry season. The cattle grazed in and around the village during the rains, and in February 1951, three weeks after the coarse rains growth of grass had been burnt, the test herd moved into the fenced area. The small paddocks, 1,2 and 3, were grazed in rotation-ten days in each paddock; the large single paddock, No.4, was gr• . zed continu- ously throughout the dry season. Tbe information gained is recorded in full in the succeeding pages. It was intended to trek on foot across the plain in one or more places to confirm or refute the information of the local inhabitants that the whole plain was one uniform · stand of Hyparrhenia rllfa and that the regrowth probably improved towards the centre of the plain. 1 So,, , ",' i, ""' ,im,l, Do,"" '"' i";:'"'' 'f«< ili, "'~, w'," ili= ;, ""'~" " ,J .,J A";nlriin2-w'iter in pools, Ule C?lHSC gruss is St) high Ihal. v i ~ jhiliLy is nil. 'I here arc no paLhs or ?f game nmmals. . In thu dry SCII SO I~ , when I he f.!r:l SS hils heen burnt , vi:-. ibi lily gOIng muc~ less trYlilg. but WU h':'I'-P(u lil s an,' Iuw ;1fltl di ll ic.:u lt to locale. For th~~t~~~,~~:~u~ll;d~;e~:rl~la~ek~lng was abnn?oncnl vitic casonably reliable alternative (0 lost ri vera in I""ture. Having described the nature of the survey, we turn (0 the detai ls. To get a general picture of the extent or the Easlern Plain anulhe "egree or regrow th over 1 whole area at the end of the dry season, an air recon nai.sance was ca rried out on 16th and ~th March 1951. This consisted of two Rigbls, and a sketch map showing Ihe degree or ~8towth seen on these two days has been prepared (see Fig. E 13). ';.' From the actual route reports of the fl.ighls some general observalions can be made. Il be recognized, however, tbat interpretat,on or gro und cond ,tions, grass species, types or owth, and other such details is not easy from an alhlode of 1,500 reet, at which the recon- nee was mainly conducted. The Eastern Plain is not a great, homogeneous area producing green regrowth throughout dry season. Approximately 50% of the area bounded by Duk Fadiat-Pengko-Pi bor-Akobo was not burnt or had oot produced visible regrowth; 30% had prod uced regrowth had dried to a red-purple colour ; 15% had been burnt and had produced a faint green ; aod 5% had been burnt and shU produced strong, green regrowtb at the time or Without exception tbe areas of strong, green regrowth were con.fined to depressions a general south-north bearin.g (or rather so uth to north -west and so ulh to north-east). _," u~'o'o cattle or game were seen the fact was recorded, but as botb flights were completed by .:Mttlid-m" rnu· 12 (10.44 and 11.39 hrs.) many herds of cattle must still have been ill their vi llages, the Nilohc practice to release the cattle for grazing rather late in tbe morning. How- preseoce of cattle coin.cided with the presence of opell water-except near to Duk where it is known that there are wells. On several occasions herds of game (liang) and seen grazing on land which showed no green regrowth . Sometimes lbe land ap- as if there were no regrowth, although the animals were defioitely grazing there; places herds were grazing where there appeared to be regrowth but dried and purple It is possible therefore that the impression of the quali ty of regrowlb obtained by reconnaissance is unduly pessimistic. Tbe main cattle concentrations were 011 tbe lower and Lolilla, the lower and middle reaches of KIlors Oeni and TUlli, the upper KIlor and the upper Khor Fullus. 'fhe areas where regrowth appeared strong and green but where ca ttle were not seen were : (i) The western margin of the plain between Kongor and Pengko. Only two isolated pools were seen in this vicinity. so it may be assumed tbat lack of drinking-water accounted for the absence of cattle. (iy The 20-25 miles of plain eastwards from Fengko. Though cattle were Dot observed, it is known that cattle do graze in this area as long as there is adequate drinking-water. The sketch map suggests that this area consists of irregular depressions which in aU probability become continuous northwards as tbe upper KhoT Gcoi. It would be reasonable to suppose that the plain's regrowth between Pengko and the middle Geni was as useful in the middle as at either end. QU) The upper KhoT Alar, north of Chieth Bridge . Although there were six herds of cattle seen in ~hc near vicinity of Chieth Bridge. the area of green regrowth appeared to be capable of accommodatmg a much greater number. There was not much open water and this may have been a limiting factor. areas where regrowth was faint green or bad dried to a purple colour were: The plain east of the line Duk Fadiat to Kongor. Herds of cattle were seen at the northern end of .this area where there were pools of open water. Farther south no catt le were seeD and only one line of dried-out pools was observed. The plain east and west of the upper Khor Nyanding, part icuJarly to the east. This is a known dry season haunt of tiang. AND ANALYSIS OF GRASS SAMPLES 1950 and February 195 1 eleven sampling units were set up on the Eastern Plain. The units were not placed at random and tlle reason for their . briefly thu~ : 'rbis is near Ayod, about the northern limit of the nre.1 ~ throughout the dry season .per9.! waterlog from wells in the upper Khor Atllr (vide Appendix Ill, T.I.R.) . 1023 S.U. II and S.U. III. Placcd at oncHhird a nd two-thirds dis ta nce from Duk Faiwil to Konsor. Bot h an! dev i ~cd to reduce lhe numbe r of unalyses. but owing 10 Ih..: IIlll·erl:lil.t y or it... applil.:;JliOI1 ;111 salll pi..: :-- rrolll the fi rs t SCilson ( 1950- 51) were ana lysed iIHli viduall y. /\ seco lld series "I' ,a "'ples w", co llected (1951 - 52) lIl1d weights recorded, but thefl' W' IS nl' ilhcr time nor need for furth er ; ln~tl yses. I:ull records have not heen included here hut the inro rmation is summa rized in Tables 531 to ~.19 he low. ( ; RASS SAMI' I.I N(; UNIt'S : SU MMAR Y I. l,' r")I)) 7(, ~;JllJpk pints at devcll widdy separ:tleu plaec>; around thc Eastern Plain samples or regrowth Wl' rc clIl at varyi ng intervals durillg the ury season and fresh and dry weights rCl"~)rdell . 2. Only , a"'ples I'llr the seaso l1 1'150--5 1 were subjected to chemical ana lysis. The best samples gave :111 :l1l:1lys ifi, 011 dry mailer basis, simila r to good English meadow hay. :1. The grea test qua ntity o f crude pro tein pcr plot nearly a lways ca me rrom first cuts al'll.: r burning. or th e L:O: lrSL: grow!ll . Thai is to say that. 0 11 the nverage. n plot still green in Mard. arter tw" previ"us cut s would 0 111 )' prod uce a rracliol1 (60"/,, ) or that rrom a p lot of silllil:II' size which had not prev iolls ly been cut. 4. Over Ihe whole seasou. plol s Il pruduced the greatest quanLity or crude protein as well ,, :-, the g.realcs t dried weight. Ir this trend is confirmed, it is a useful guide in deciding a system or managc l11enl. (Plo ts 13, though the ir coarse grass w(ts burnt olr in December-January, werc len unt oul'hed until March : a second cui was laken in May.) 5. Over the whole season and t"king totals ror a ll plots, the average air-dried weight of regrnwth pcr fcddan W!l S 123 kg. and 106 kg. in 1950-5 1 and 1951- 52 respectively. The mean is 11 4 kg. per redd.l1. 6. On the avcrage, one rcddan produced 9·3 kg. crude protein ill U,e season 1950-51. 7. Assuming th"t 10 kg. of dry regrowth wi ll provide a maintenance-plus ration for one Animal Unit I'or one day. the aren required ror one /\nil11al Unit for U,e five-month period January to May inclusive wo uld be 12·3 reddans. Ir only tbe areas or robust regrowth in the shallow depressions nrc considered, 7·3 feddans per Animal Unit is enougb.. 1024 TABLE 531 frYPARRHEN1A RUFA REGROWTH ANALYSIS (On Dry Maller Basis) (Selected Samples from 16 sq . rn. Plots) Samp~ CuI X x x No. Ash SiO. Ca p -,I "-'-_.'-I, - - 51 / 80 23·04 51·2) 0·11 13-93 i 8-4 1 0·40 ---I 51 /95 53·79 11 ·89 7-54 0·43 0·22 0 v.... .. 51/ 110 51/ 125 TABLE 532 EASTERN PLAIN SAMPLING UN ITS AIR-DRIED WEIGHTS 1951 (In grammes) Un; • ._1 Plot I No. of Block 1 PBlo ' 1 NoA Cuts eu. uo f 1 PClo . I Neou.l So f Total ---I : ~;;--'--T~;-- -- 752 438 T '--T 2,285 - I: I 157 206 73 90 526 II 537 635 223 465 1,860 II II 424 525 483 400 1,832 III 567 524 406 290 1,787 IV 637 1,066 1,203 1,029 3,935 IV II 442 1 888 465 975 2,770 V I 956 1,160 919 3,965 V II 897310 I 1,247 1,181 1,308 4,607 VI I 195 149 409 455 1,208 VI II 224 240 202 237 903 VII 506 320 203 264 1,293 VII II 314 198 253 303 1,068 Vlll 72 279 63 330 744 IX III 147 323 1 I 255 836 IX II 103 281 139 657 X I 318354 I 359 J75 519 I 1,638 X II 384 578 496 427 . ____,_ . _--.1.._~~ ___. _. __ ._ _1 ---,~- TOlal for 18 feddans J 7,642 1 --T~ I 8,843 33,799 (I) ~rtr~~:n~~~ ~~ic~~::e~ea~nJ\~:~~~(a~::~Jh!e'f,~~w~ri(e~~~~. 1 was arealest where Ihe road incrcalCd the dC11ft o r Roodinl. The llVeralC (2) Tbe ,fe;ate$t quantil), o r rclfowth was produced (rom the IwO units s.itua ted wilhin 15 mild (rom Pen&ko, i.e . to the call (sec Fia. Ell). (:t) Area or nch plo l 16 ~. m. WeiShl taken 10 neueSI lramme. TABLE 533 EASTERN PLAIN SAMPLING UN ITS ABSOLUTE DRY WEIGHTS, 1951 (In grammes) Unit Block Plol No. of Plot I No. of I' Plot No. of I Piol No. of A C UI S B i Cuts C Culs D Cuts Total I I 3 ~!: 'I I I' 543 414 I 2,195 II 715220 I 71 I 85 I 506 I 517 4 608 215 2 438 I 1,778 II 405 4 504 465 I 377 I 1,751 I .\44 I 3 504 391 2 273 I 1,7 12 I 604 4 1,012 1,137 2 968 I 3,721 II 352 3 846 444 2 939 I 2,581 I 884 4 912 1,115 2 869 I 3,780 II 8JJ 4 1,198 1,137 2 1,258 1 4,426 VI I 189 I 143 I 428 I 1,l3l VI II 217 I 23 1 ~;~ I 225 I 868 VII I 490 I 307 197 I 252 I 1,246 VII II 304 I 191 1 243 I 288 I 1,026 VIIl I 68 I 269 I 61 1 312 I 710 IX I 107 2 143 312 I 241 I S03 IX II 128 2 99 27 1 2 131 I 629 X I 368 4 344 361 2 492 I , 1,565 X II 369 3 555 473 2 403 I 1,800 I 8,002 I 8,393 I 32,228 Area or eaclI plot 16 Iq . til. Wejpt ... ken to oeaJ'W lralDlDL 1026 TABLE 534 EASTERN PLAIN SAMPLING UN ITS DRY MATTER PER FEDDAN, 195 1 (In kilogrnmmcs) UD_i_' _~ __ Block _. _ _ _ .p~o,~ J_ P1o1 B PIOlD Total 189 136 T 143 109 577 II 40 52 19 22 III II 116 160 56 11 5 467 II II 106 132 I 122 99 459 III 143 132 ! 103 72 450 IV 159 266 298 254 977 IV n 92 222 117 246 677 V 232 239 293 228 992 V 11 2 19 )14 298 330 1,16 1 VI 50 )8 97 11 2 297 VI 1I 57 61 51 59 228 VII 129 81 52 66 328 VII II 80 50 64 76 270 VIII 18 71 16 82 187 IX 28 38 82 63 211 IX II 34 26 71 34 165 X 97 90 95 129 411 X II 97 146 124 106 473 Total for 18 feddans 1,906 2,254 2, 101 2,202 8,463 Absolute dry matter based on 16 sq . m. plots. TABLE 535 EASTERN PLAIN SAMPLING UNITS CR UDE PROTEIN PER FEDDAN, 1951 (In k ilogrammes) Unit I Block I PlatA I Plot B Plot C I Plot 0 I , Total I I I 17-8 I 11 ·2 10·2 6·4 45-6 I II i 4·3 3-8 I I,) 1·2 10-6 II I I 14·6 12-8 4-8 8·4 40·6 II II ,i 12,) i, 12·5 10·1 l-4 I 38·3 I III I 12·7 I 11·9 8· 1 2,) 35·0 I V I 18·0 22·5 I 24·6 16·6 81·7 IV II j 10·3 , 18·6 11·0 19·5 I I 59'4 V I 203 19·1 19·8 6·2 1 65-4 V II 24·2 28·0 24·0 20·0 96·2 VI I ,I 4·7 ),1 I , 7·2 5·9 I 20'9 VI 11 5·5 5-2 ),7 l-4 : 17-8 VII I 9·6 5-4 H 2·9 21·0 VII II 6·4 ),2 ) ·8 4·0 11-4 VIIl I 1·5 I 5·6 1·5 4·6 13·2 IX I 3·0 J-6 5,) 3·5 1504 IX II ) ' 1 i I 2·2 5-4 1·9 12-6 X I 10·2 8·5 7-8 9·2 ) 5-7 X II i 107 i 1l-4 I 104 6·9 41'4 Total for 18 feddans .. . ... 189-2 I 190·6 162· 1 126,) 668·2 .Based on 16 sq. m. plots . . 1027 TABLE 536 EASTERN PLAIN SAMPLING UNITS AIR-DRIED WEIGHTS, 1952 On grammes) Unit Block PIOl No. of Plot No. of I Plol No. of I Total A Cuts B Cuts 0 Cuts for Unit 11 I! I 506 562 I I ,~, II II 224 260 III 158 199 IV 643 372 tE I I 625 2,E: IV I! 219 712 555 356 1,842 V 756 837 1,215 ! 1,329 4,137 V 11 1,497 1, 165 VI I 980 1,032 ;I l '~:: i 1 ,:~ ~:~~ VI 11 445 433 636 , 602 2,116 VI! 203 301 234 1 332 1,070 VI! IJ 542 472 vm I 961 1,092 IX 162 I I,:: ~~! ~:ill IX 1I 70 49 X 76 X IJ _-+_ _+ --_---:I~ ___I_ ._1_65_,_ _ . ~ Total for 18 plots 7,442 7,486 7,427 ! I 6,891 I T 29,246 I t (I) Ma:x.imum ITJfowth (rom ploU CUI Illhe bcJinnin.c of MaKb and apin in May (Plol B). (2) tn the uven c:ompuablc units or ' .... ·0 blocks the rt'P'O .... 'th was &lUtest when lbe road inaeasecl tbe de,rce or floodina. The averille dilfereoco is calculll.ted I' 8·7 q. (air-dried wei,lu) per reddan. (3) The pealcsi qua.ntity o f rt'JI'owth .... -as produced rrom the three uniu between Ptn,ako and the eastero Bot-PiOOr boundary (IV, V, Ind VI). (.) Area of each "lo t 16 '"1:. m. TABLE 537 HYPARRHEN1A RUFA REGROWTH SUMMARY OF ANALYSES, 1951 I NITIAL CllTS AFTIR B URNING ! - Average 1 Quantity No. of Avy~ge AV~~ge I Air-Dried Weight Crude Protein Samples I Crude Protein Crude Fibre %r ~~t '7; :~t I : .~;-----;~~~---,- .- ;5:-----r;; --- February 14 March 18 8·14 24-48 269·1 I 21·90 April 15 6·73 24·26 396·3 26'67 May 16 j 5·04 25-48 410·0 I 20·66 SUBSEQUENT CuTS ! Average Quantity No. of Period of I Average 1 Average ; Air-Dried Weight CUi No. Crude Protein Samples Regrowth % % , 1 in Weeks I , per Plot per Plot Crude Protein Crude Fibre ; in gm. ingm. , , 2 4 l 10·14 24·28 130·8 13·26 3 4 9·83 26·91 156·2 15-35 I 8 I 5 9·87 26·17 64·2 6·34 5 8 8·24 24·98 317·7 26·17 i 4 9 9·32 26·65 127·1 11·83 1 i 3 3 6 9·97 24·52 81 ·8 8'15 6 I 7 10'11 26·01 I 152·0 15-36 I ,I I 10·28 24·71 102-8 I. 10·56 Area of each plot 16 sq. m. 1028 TABLE 538 ANDROPOGON REGROWTH. 1950-51 Percentage Average Period of I l crcenlagc Present I:-ength Regrowth Crude PrOlein mem. in weeks in Dry Malter Over 40 9-070 (shorl ) Over 40 10 II 10·344 (medium) Over 40 19-4 II 9·101 (long) 10 11-6 13 8·447 (long) TABLE 539 REGROWTH FROM MIXED PASTURE AT MWOT TOT. 1950-5 1 Period of Length of Percentage Sample No. Regrowth Regrowth Cut No. Crude Protein Dried Weight in in em. in Dry Mauer in gm. weeks 65 a 19 11 ·07 42 ·5 65 b 10·)2 179·0 65 c 8'87 507-4 65 d 7-77 620·0 65 • 6·76 114'0 ( I) Samples NN. 6.3 d. band t Ire aU (rom Plot A: Sample No. 65 d is (rom Plot B; Sample No. 65 ~ it from Plot C. As , result or urly and re,uJu CUItinj: lbe percental' COIde protein of PIOl A samples Is bi~cr. evUl in S()~)7;;/a:~ J Re~;:~~t~i~~ed..s 158 200 12.12.52 2 30 - - whole 627·8 I 480 556 I 58 3598 4492 ! 1197 , 867 I cut at Melut I M ature growth TABLE 545 DEEP-FLOODED G RASS ANALYSIS Jonglei I Dale o f Area cut IA verage wei8hi~ O{g~amPle Fraction \ we~ of MoY:ture _. ___ . __ 9 10· .58 6·83 .. ' matur e flood-seaso n growth 118 / 69 4.5.51 16 0·50 0·16 whole 99-2 3·60 8·08 1·38 3 1·46 45 ·46 13·62 9·06 .. 34 days' regrowth after 'I ~m_ 132 / 175 28.2.52 29·25 I part 517-2 6-28 4·34 1·23 33 ·72 43·34 17·37 13·58 Mi)C.ed sample of mature E. stog"ina, E. pyramidalis and OrYlA 133 / 176 17.4.52 2·00 1·25 part 57 1·7 6·63 13·63 I 2·41 28·15 40·52 15·29 9·10 ; E. stagnina: appro)C. . 60 days' re- \ I growth, part sample 134 / 177 4.652 3,00 0·50 whole 6·28 478·0 I 8·22 1·33 33·82 I 43·61 13·02 8·32 I £. stag"ina: 4~u~~' regrowth after 135 / 178 5.3.52 16 150 13·25 2·30 .\ part 1,050'0 6·63 6·28 I 1·55 31·89 48'39 11 ·89 8·49 I' .. mature Hood season growth 137 / 180 5.5.52 16 170 JO·oo 1·25 part 55 1·6 7-45 13·8 1 ii' 2·49 , 31·3 1 40'53 11 ·86 6·46 I .. 30 days' regrowth after 138 / 181 5.1.52 16 100 3,00 0·50 whole 550·0 7-28 cutting 8·12 I -II I, 31·33 48 '92 JO-52 5·99 .. regrowth after grazing 139 / 182 5.2.52 16 30 11 ·25 1·50 whole 1,350-2 I 6·45 , 9·87 1'18 34· 15 43-07 11·74 5·22 I .. 31 days' regrowth after I cutting 140 / 183 6.4.52 16 60 6·00 2'50 part 863-8 I 6·65 I 11 ·72 I 1·66 27·% 45·56 13,10 6·83 .. 60 days' regrowth after 141 I cutting / 184 5.5.52 16 50 3·75 1·00 part 457·5 6·80 16·76 2·02 30·40 36'39 14·43 7, 11 I .. 2~~j~:' regrowth after 142 / 185 9. 12.5 1 4 60 5·9 0'75 pan 740·0 I 6-48 I 7-07 1·37 32·83 I 41-89 16·84 13-74 .I .. 4 SQ. m. out to ground i I :=~~~g r~:~~: cut 143 / 186 4.2.52 16 50 11 '00 2'50 whole 2,585-0 6-88 , 11 ·42 2·04 30·96 40-93 14-66 9· 13 \ .. unit had been under I I water; regrowth robust 144 / 187 4.3.52 16 6 1·25 1·00 whole 639·9 6-95 I 9·88 ) ,83 28·7) 47 ·88 11 ·69 7·36 " poor regrowth from area I c.."Ut for 142 145 / 188 4.4.52 16 17 1·00 0·50 whole 349'1 7·65 10·71 2·04 29,56 45·19 12·5 1 8·45 " 31 days' regrowth, not 146 /18q 4.5.52 strong 16 21 4·75 1,00 part 478-3 7·83 13·39 2·09 29·98 41·85 12-69 7·32 " 30 days' regrowth, strong 155 / 197 10.4.52 16 2·75 0·75 whole 761·6 6·43 7·06 1·44 30·01 43-53 17-95 13·57 49 days' regrowth after cutting 157 / 199 30.3.52 16 3·75 3·75 part 1,648'3 4·73 4·1 8 1·24 31·80 48-47 14-31 10·60 mature growth partly I consumed by grazing animaJs. TABLE 546 ECHINOCHLOA P YRAM IDALIS ANALYSIS, D1GESTIJl ILlTY COEFFICIENTS, AND DIGESTIBLE NUTRI ENTS OF MATURE GRASS (One Sample: longlci Rer. No . 44: Shambat Ref. D .T. No. 37) Mois- IO rganic I Crude I Ether I Crude : N.Fr~ II Silica I tu re 1 Matter I Protein Extract . I Fibre ; E~~:I,;- ~ Ash ; SiO: S_E. JT _D_N_ _ ____ 1 1 • ;--_. - _.-- Composition of Dry 'I Maller --- 1 4- 38 87-07 3-30 1-13 30-96 51-69 12-92 Digestibility Coefficients! 60-30 21-02 43-52 63 -)2 61-37 A'f~er ~ctu~c o~: i swned by sbeep . 1 66-30 24-23 48-75 68-57 67-73 52-SO 0-69 0-49 19-60 31 -72 Digestible Nutrients "' \ Above, in parts of fodder actually con- sumed by sheep . . . , TABLE 547 SHALLOW. FLOODED PASTURE HERBAG E YIELDS (All weights in kg.) 1st CUts Total Average Total UoitNo. Year after Subsequent Cuts Fresh Percentage Dry Burning Weight Moisture Weight -- VI 1951 0-50 I-50 2-00 70 0-6 VI 1952 2-60 0-75 3-35 70 1-00 VII 1951 I 0-50 1-15 2-25 10 0-68 VII 1952 0-25 2-25 2-50 10 0-75 Vill 1952 I 0-25 2-75 3-00 70 0-90 1 I -I ----.. Total from 80 sq. m. 4-10 9-00 13-10 70 3-93 ! Total from 1 feddan I (4,200 sq- m_) --- I 688 206 TABLE 548 DEEP·FLOODED PASTURE HERBAGE YIELDS (All weights in kg.) 1st Cuts Unit No. Year after Subsequent Total Average Total Cuts Fresh Perceotage Dry Burning Weight Moisture Weieht ._--- 1951 14-00 3)-25 47-25 80 9-45 1952 3-50 10-00 IJ-SO 80 2-70 I(a) 1952 6-00 5-25 11 -25 80 2-25 n 1951 5-30 8-00 13-30 80 2-66 II 1952 IJ-25 10-25 23 -50 80 4-70 ill 1952 9-15 2-00 11 -75 80 2-35 IV 1951 62-40 0-50 62·90 80 12-58 IV 1952 9-50 10-00 19-50 80 3-90 V 195 1 28-50 0-25 28-75 80 5-75 V 1952 23-60 18-25 41-85 80 8-37 Total for 160 sq. m. 273 -55 I -- -80 54-11 Total from 1 feddan I (4,200 sq_ m _) __ . I 7, 181 i 1,436 1037 GRAZING EXPERIMENTS The direct method of ascertaining the stock-carrying capacity of a particular type of pasture is to select a known area typical of that pasture and have it grazed to capacity, recording the number of Animal Unit days grazing provided throughout the season. To be reliable the experiment should continue for a number of years so that an average figure may be obtained. The two experiments which are described below were of too short duration to yield figures of ' accuracy, but they give a useful indication of the carrying capacities of the pastures investigated. SHALLOW-FLOODED PASTURE Ea rly in 1950, as soon as the flood-waters from the 1949 high river had receded and the coa rse growth of grass had been burnt, an area of 30 feddans of typical shallow-flooded pasture was fenced on Awarajok Island (nea r Malakal) and divided into four paddocks ; Paddocks I, 3, and 4 were each 5 feddans and Paddock 2 measured 15 feddans. Normally most of Awarajok Island is under water from July until the following January or February, and in 1950, owing to pressure of other duties, the burni ng of the grass, erection of the fences, and transference to the island of the test herd of callie were not completed until the beginning of May. By that time the regrowth of grass was so vigoro us that a total of 90 head of cattle could not graze down more than 25 of the 30 feddans. The attempt to mow the regrowth by machine (an Allen scythe) was not successful owing to the uneven surface created by the carpet of charred stems. In 195 1 the river level fell sufficiently by mid-February to allow work to commence. A 5-metre strip of grass was cut by ha nd on either side of all fences, to protect the posts, and the whole area was fi red on 12th February. Since the grass was not perfectly dry, combustion was not complete alld a thin layer of scorched and twisted stalks remained. The new growth appeared and the herd was transferred to the island camp on 20th March ; but on 22nd March a fire from the other end of the island spread to and fired the layer of scorched residue within th e fenced a rea. In this disaster some of the fence and all of the pasture in Paddocks I and 2 were destroyed. As a result of this only Paddocks 3 and 4 were available for grazing until 20th May. and even then the regrowth was weak and the ground cover poor. Thus in 1951 a herd of 30 cattle could not be maintained throughout the dry season on an area 5 feddans larger than that which had supported 90 head, though for a shorter period. the previous year. These details a re recorded to illustrate the signifi cant relationship between time of burning and vigour of subsequent regrowth and the need fo r control of burning if the fullest use is to be made of riverain pasture. In 1952 the test area was burnt 01T on 5th March (the delay was due to the necessity of protecting the posts by fire-lines) and the test herd entered the field on 5th April. Between then alld 29 th June 3,094 Animal Unit days grazing was produced by this 30-feddan field over a period of 85 days, i.e. 60 Animal Units per day or 2 Animal Units per feddan. But the records show that there was a general loss of body weight of 2·6 kg. per head, although there was plenty of good pasture towards the end of the grazing period. The field was obviously overstocked ; there should have been an increase in weight over this seaso n (cf. 1953). In 1953 the experiment was co ntinued by the Sudan Veterinary Service. A herd of 15 young bullocks entered the experimental fi eld on 25th March and grazed there until 29th June ( I,S80 Animal Unit days) during which time there was a live weight gain of 17·7 kg. per head. It shou ld be added tha t the regrowth was assisted by good early rains. The conclusion is that the carrying capacity of shallow-flooded pasture at Malakal is approximately 0·5 Anima l Units per feddan (119 A.U. per sq. km.) over the season when shallow-fl ooded regrowth is available, i.e. March to June inclusive. One fact of note is the rate of growth of Echillochloa pyramidalis after rainfall and before flooding by ri ver spill. Permanent grass sampling plots were protected within each paddock of the grazing field on Awarajok ]sland; the grass in the plots in Paddock I was cut down to ground level on the day the cattle were removed from Paddock I to the next paddock in the rotation. This was repeated in each paddock. Before cutting, the average height of the plants was recorded and the figu res from one plot are given below: --D-A:T-E- O-F TSAM-PL;fN:O -_I I A vcrage Daily Inlerval Average Height in days in em. Growth in COl • .. _---._--. ---_.. ..- 28; I 5~~--! 38 2·24 5/ 6 ! 25/ 6 20 5084 1038 Raillrall at Malakal up to 4/6/ 53 was 199 mm. and between 4/6/ 53 and 24/ 6/53 a further 80 nun. rainfall was recorded, a total of 279 mm. up until 24th June. Fresh and dry weights of all grass cu t from these plots were also recorded; the figures are swnmarized at the bottom of Table 549 below. The high yield of regrowth in 1953 was due to the relatively heavy early rains. TABLE 549 SHALLOW-FLOODED PASTURE SUMMARY OF STOCKING INTENSITIES. LIVE WEIGHT GA INS. AND AVA ILABLE PASTURE Intensity of Stocking in Animal Unit days Paddock I (5 fOOdans) 625 170 510 233 Paddock 2 (15 fOOdan,) NR 652 2.548 900 Paddock 3 (5 feddans) 7&0 470 900 225 Paddock 4 (5 fOOdan,) NG 405 1. 136 522 --,------ ·Total for 30 fOOdans 1,405+ 1.697 5.094 1.880 - ---_._ - - - ------ -------!-- -- - ----_ .. .. ---.-------- Live weight gain or Joss per head in kg... N R i NR -2·8 + 17 ·7 I Fresh weight of grass in kg. per feddan NR 2.683 3.763 7.875 Dry weight of grass in kg. per feddan NR 787 904 2.012 NR _ ' not recorded' NG - 'nolarlltcd' TABLE 550 SHALLOW-FLOODED PASTURE LIVESTOCK WEIGHTS. 1953 : _____ .• _~E~~_ I~ Ko. ______ Bullock No. Beginning of E.nd of Experiment Experiment 21 275 302 24 364 362 25 324 34 1 26 355 370 29 343 345 33 339 350 34 316 333 43 325 347 45 281 302 54 321 340 46 290 326 48 307 337 49 286 306 55 249 57 304 ------- . Total 4.679 4.945 - - ------ ----------- Total gain durink experiment 266 Average gain during c}(periment 17-7 1039 DEEP-FLOODED _ PASTURE An area of 30 feddans of typica l deep-flooded pasture was fenced on the Nile flood-plain (east bank) immedia tely downstream of Ma laka l. Echinochloa slagnina was dominant and tbere was a fair percentage of Oryza sp. on the ridges. The fence was erected in May 1951 so that th e coarse, fl ood season growth of grass would not be disturbed by local livestock in the following dry (low river) season. The fenced area was divided into tbree 100feddan paddocks ' a nd these were grazed in rolation. It will be appreciated that the great bulk of flood season growth permitted either a high stock ing ra te for a short period foll owed by a long period a t a low stocking rate on regrowth on ly, or a moderate stocking rate over a prolonged period. We wished to kecp the sa me number in th e experimental fie ld throughout tbe low river season. In 1952 the herd consisted of 50 adult bullocks with an average live weigh t of370 kg. They entered the protected area on 4th Februa ry and grazed the mature growth of Paddock 1 unLii 28 th February, by which time little remained but a carpet of trampled stems. Paddock 2 was then grazed until 21st Ma rch, and then Paddock 3 unti l 15th Apri l. At the end of this first cyclc there was an average live weight increase of 0·3 kg. per beast ; i.e. at that stocking rate thc pasture was providing lillie more than a mai ntenance ra tion. By this time the regrowth in Paddock I was quite strong a nd that in Paddock 2 was noticea ble. During the remai nder of the dry season the size of the herd was not altered (except that two casua lties were removed a nd not replaced). The second cycle, which began on 16th April, was complcted by 25th May, approximately half the du ration of the firs t cycle, and there was a n average loss of live wcigh t of 3· 3 kg. per head. The third grazing cycle was completed in 19 days, because of the sbortage of regrowth, a nd there was a further loss of li ve weight of 0·5 kg. per head. In the remaining five days before the end of the ex peri men t on 18th June 1952 there was yet another loss of 2 kg. per head. Thus, at a stocking ra te of 1·6 Animal Units per feddan, there was a gross loss of live weight of the herd amounting to 266 kg., o r 5·5 kg. per head. Under 1952 conditions the stocking rale was much too high , but the ra in fall was considera bly below average and this affected the amo unt of regrowth , especially in Ule laller half of the test period. In 1953 Ule test herd was reduced to 30 head, out of which there were two casualties (not replaced) during the season ; the average li ve weight was 223 kg. These animals were a completely new batch and were either ad ult or growing bullocks. The first grazing cycle began on 2nd February and ended on 23rd March, a period of seven weeks (cf. 10 weeks in 1952) and, there was an average li ve weigh t ga in of 29'5 kg. per head. The second grazing cycle was longer than intended, 24 th March to 13th May (7 wccks) a nd there was a live weight loss of 1·9 kg. per head. But in the third cycle ( 14th May to 25th J une- 6 weeks) a gain of 8·4 kg. per bead was recorded. T hus, a t a stock ing ra te of I Anima l Unit per feddan (actually 0·93 A.U.) there was a live weigh t gain of 36·0 kg. per head per dry season, o r 0·25 kg. per head per day. In contrast to 1952, the ea rl y rains of 1953 werc rela ti vely heavy and the regrowth at the end of the experi- mental period benelited . TABL E 55 1 DEEJ'·FLOOOED I'ASTURE SUMMA RY OF STOCK I NG I NTENSIT IES ANI) LI VE WEIG HT GA1NS Anima l Unit Days I- 1952 r -j9~;-- 11allJock I (10 fcddllns) T 2,016 I):ah.h,x':\. 2 ( .. I 2,)52 I 1';uJlltXk 3 ( .. I 2,064 r= To!;!! for 30 fcddans I 6,432 I 4,032 Live.weight gain or loss per head in kg .. - 5·5 I + 45 1040 TABLE 552 DEEP-FLOODED PASTURE LIVESTOCK WE IGHTS, 1953 (All weights in kg.) On Purchase I Beg;nning of End of Bullock No. 26/1/53 Ex~m~3cnt 23/3/53 13/5/53 Experiment 25/6/53 -----}l ~- ·-I -----.-.. --_._--' 307 338 331 340 271 266 297 291 296 273 285 308 316 319 286 279 295 234 24 1 307 309 340 3)7 338 303 2~ )00 309 323 225 223 257 266 272 282 287 335 )3) 333 205 196 226 226 300' 10 254 259 285 283 287 11 255 252 287 283 292 12 187 184 210 225 229 13 267 274 304 293 294 14 174 167 192 193 210 15 151 155 184 175 182 16 210 204 234 243 17 194 183 285 29 1 297- 18 184 186 212 2 11 228 19 187 185 214 260 274 20 175 180 215 265 267 21 226 219 265 261 263 22 216 214 240 222 248 23 192 184 185 159 171 24 152 151 200 188 229 25 197 179 212 216 23 1 I 26 271 264 292 I 278 261 I 27 223 218 237 I 232 235 28 272 265 297 I 288 I 297 Total ··. 1 5,8«<) 5,760 6,498 I~~-- -I- 6,660 T~~rygam -af-t~-.:+I------r-----~--7-3-8--_+i-_-.--~-9---r_-_-___~-_ __ _ Average gain after 1- I entry ... 29· 5 i 2H THE STOCK-CARRYING CAPACITIES OF RIVERAIN PASTURES SHALWw-FLOODED PASTURE From the experience of the grazing experiment on Awarajok Island we have stated (see p. 1038) that shallow-flooded pasture has a stock-carrying capacity of 0·5 Animal Units per fed- dan from March to June at MaJakal (119 A.U. per sq. km.). From the 1953 data we extract the following figures : 1041 __________ ~tOJlstMarch !_lst April to 29tb June Animal Unil Days per 30 [eddans I 105 I 1,775 Animal UnH Days per sq. km. ... 833 I 14,082 Anim31 Units per sq. km. . .. i 27,8 (1 month) i 156-5 (3 months) This ignores any contribution from the coarse, flood season growth before burning; while its value as fodder is low it is not entirely worthless. The only digestibility trial on shallow-flooded grasses concerned a sample of Echinochloa pyramidalis cut in October, dried, and stacked as hay (Shambat No. D .T. 37). This had a crude protein content of 3·30% on a dry matter basis, but the digestible crude protein was 0·69% . Further, the sample was cut while green and alive; had it been left to mature, wilt, and' die ' in the ground, the percentage of crude fibre would have increased and that of crude protein decreased. For example, mature Setaria incrassafa was found to contain 1·62% crude protein and 40'55% crude fibre (dry matter basis) but digestible crude protein was zero (Shambat No. D.T. 39). We assume that one feddan of shallow-flooded pasture will produce flood season growth of 1,000 kg. dry matter having 0·4% digestible crude protein, and we have calculated the daily requirement (maintenance plus low production) of one Animal Unit to be 0 ·25 kg. protein equivalent per day; the stock-carrying capacity can be calculated at 16 Animal Unit days per feddan. But at least half of the shallow-fl ooded pasture is burnt as soon as it is exposed, so that the carrying capacity is reduced to 8 Animal Unit days per feddan or 63·5 Animal Unit months per sq. km. When this is added to the 27 ·8 A.U. months' regrowth (up to 31st March, above) the total carrying capacity up to 31st March is seen to be 91·3 Animal Unit months per sq. km. Since there can be· no claim to accuracy, we have taken the stock-carrying capacity of one square kilometre of shallow-flooded pasture at Malakal to be 100 Animal Unit months until 31st March and 160 Animal Units per lTtonth thereafter. DEEP-FLOODED P ASTURE It will be remembered that tbe field work of the Team ended in June 1952; on tbe informa- tion collected up to tbat date, we calculate the stock-carrying capacity of deep-flooded pasture , to be 400 Animal Units per sq. km. (1 '7 A.U. per feddan) for the first two months when the bulky flood season growth is available, and thereafter 240 Animal Units per sq . km. (1 A.U. per feddall) on regrowth alone. This can be expressed as approximately 190 Animal Unit days per feddan. Results obtained by the Veterinary Service in 1953 confirm our assessment, but suggest that it may be Slightly high. However, only a continuation of experiments over a period of several years can be expected to produce more accurate figures. If we turn to the figures of herbage yields for confirmation we are at a great disadvantage in not knowing the digestible coefficients of deep-flooded grasses. One feddan of deep-flooded pasture can produce a t least 155 kg. of crude protein (see p. 1034). One sample of Echinochloa slagnina hay, having a crude protein content of 6·04% was shown by BoynsCl) to have a diges- tible crude protein content of 2·23%. Since the ratio crude protein: digestible crude protein would be closer in the case of regrowth, we can take this figure with safety and apply it to our known yield, 155 kg., which figure is also incomplete. The yield of digestible crude protein per feddan would then be 57·35 kg. We have calculated (see p. 571) that the daily requirement of one Animal Unit for maintenance and low production is 0·25 kg. digestible crude protein. Thus one feddan should support 229 Animal Unit days or 1·5 Animal Units throughout the 150 days' season on riverain pasture (357 A.U. per sq. kIn.). This is reasonable confirmation, but the fact that these figures are based on very meagre data, and are therefore accepted with caution, is too evident to require emphasis. 3. MISCELLANEOUS GRASSLANDS In many parts of the Jonglei Area grasslands inland are not grazed in the dry season for two main reasons . Eitber the swamp pastures at tbe main rivers, which are fresh ·and green in the dry season, are sufficient for local needs and the dry inland grasses are not required, or the inland pastures are inaccessible owing to lack of water supplies. However, it was necessary, to establish the actual grazing value of these inland grasses in· the dry season in order to determine whether inland pastures could be opened up as alternatives to riverain pastures lost under the Project. 1042 THE FLOOD REGION In the Flood Region there is no lack of coarse grass remaining at the end of the rainy season but this is either unpalatable or is of low nutritive value. Such is the local belief, and to prove it true or false a series of simple experiments was arranged. Three small areas, each of fOllr feddans, were fenced in, their perimeters protected by fire-lines, and the contained grasses grazed by a few test animals whose live weight gains and losses were recorded. Two of the fields, at . Nagdiar and Kodok , were so far from Malakal that the experimental animals were only weighed twice, before moving to the test field at the beginning of the dry season, and on returning to Malakal at the end of the dry season-or when all the grass within the test fie ld had been con- sumed. The third field, at Gonio, was close to Malakal and tile experimental anima ls were weighed at weekly intervals. Records over two seasons are therefo re more complete for the Gonio field and so this experiment is described in fu ll , the others more briefly. The chemical analysis of some of these mature grasses is tabulated at the end of this section. GONIO GRAZING TRIAL The four-feddan field, situated about 300 yards from the right bank of the Ni le just upstream of Malakal, was fenced in December 1951 and protected by fi re-lines. Solaria incrassata and HyparrhelJia spp. were the principal grasses ; Sorghum purpureo-sericeum and Andropogon gayallus were less evident and there was a large number of unidentified annual grasses. By the time the experiment began (February 1952) all these grasses, particularly the four named, had become coarse and brittle and much of the seed had been shaken from the heads. The four cattle selected for the trial were Nos. 21 and 22, immature 4-year old Nilotic bullocks, No. 28, an adult Nilotic bullock, and No. 29, an immature Arab bull. At night tbey were kept tied at the stockyard at the Malakal Experimental Farm- because of the danger of attack from wild animals- and they walked the short distance to and from the Gonio field each morning and evening. Drinking-water was provided at the fie ld. The trial began on 1st February 1952. Whilst the seed-heads of the grasses were sti ll 1. ' available all four beasts actually gained weight, but as the heads disappeared and as leaf became , scarce live weight decreased regularly until the end of May. By mid-April all the Sorghum, H),parrhenili, and the various annuals had been eaten, and only the stiff stems and rough leaves of Setaria and A/1dropogon remajned. In May the early showers of rain encouraged germina- tion and growth of the new season's grass and from June onwards steady gains in weight were recorded. As is seen in the following record of live weights, the losses during the months January-May were 29 kg., 47 kg., 71 kg., and 42 kg. respectively. No. 28 suffered the greatest loss, perhaps because it received an injury to an eye which interfered with its sight and later with its ability to graze. There is no doubt that the loss of weight recorded was a good indication of the low nutritive value of the grass, though it has to be admitted that, just before the new season's grass appeared, grazing was lacking in quantity as well as quality. To show that the poor quality of the grassland was confined to the dry season when the plants were mature or dead, the trial was continued with the same animals throughout the rainy season of 1952. Heavy rajn showers frequen tly cause such surface flooding that this heavy clay land cannot be grazed for days at a time, but in 1952 the weather was such that continuous grazing was possible. With the growth of fresh grass all the test animals gained weight steadily. Bullock No. 28 became increasingly blind and this, combined with its age, prevented it making as good a recovery as the other three, wruch gained weight at an average rate of 0·5 kg. per head per day during the four months June to September. They not only gained the weight lost in the dry season but added a further 27 kg. per head on the average. However, in so doing they consumed most of the grass and by mid-October grazing was already scarce. At the end of October three of the four beasts were removed in the hope that the grass remaining would be sufficient for the remaining animal throughout the dry season. But by this time all growth had ceased and the remaining bullock gradually lost weight until removed in December 1952. The conclusions to be drawn fro m this brief experiment are that this natural high land, or rugh intermediate land, pasture has a stock-carrying capacity of one Animal Vnit per feddan during the rainy season but that, if grazed to this extent, no pasture remajns for the dry season. Conversely, if not grazed during the rajny season this grassland may be stocked in the dry season at the rate of one A.V. per feddan and the livestock will survive, but will certajnly not thrive. It is noteworthy that, given no al ternative, Nilotic cattle can graze this coarse, mature grass, but it is very unlikely that it would allow them to maintain their condition even if an unlimited .quantity were available. Live weight gajns and losses are recorded in Table 553. KODOK GRAZING TRIAL The four-feddan field u~ed for this experiment was about 400 yards in from the left bank of the Nile two miles upstream of Kodok (Shilluk District) and was chosen as being repre- sentalive of much of the land within a few miles of the river between Kodok and Renk. 1043 16 Sorghum purpureo-sericeum CAr. anis) was dominant and there was a fair admixture of Brachiaria Oblusiflora and Selaria incrassata in addition to various· annual grasses. Because of difficulties over fencing and transport, this trial did not start until 7th April 1952, and because the dry season was nearly at an end six beasts were used. The site of the trial was 45 miles from Malakal, and, there being no weighing machine available at Kodok, cattle were weighed at Malakal on 5th April, embarked on a steamer and taken to Kodok, and entered the field on 7th April. Grazing was reported to be in short supply by 24th April at which time the livestock were' still in good condition, but the animals remained within the field until 11 th May by which time all had lost condition. Until they could be brought back to Malakal (5th June) they were grazed on land similar to that within the field, but the new growth brought on by the early rains helped them partly to regain their lost condition. When weighed on 5tb June they were found to have lost an average of II kg. per head . This field lVas not grazed during the rai.ny season because of the impossibility of supervision, but on 19th November one bullock, No. 23, was admitted to the field and remained there until 19th May 1953. At first it maintained condition well, but towards the end ofthe dry season the grass became scarce until finally the only grazing remaining consisted of the fresh shoots of new growth. The animal lost 59 kg. (from 314 kg. to 255 kg.) during the trial (the weight on 20th May was actually 234 kg. but this was immediately after a hungry, and perhaps a thirsty, journey from Kodok; 255 kg. was the weight recorded two days later) . Thus four feddans of this high land, or high intermediate land , could not support one A.V. throughout the dry season. We must add that the 1952 rains had been relatively light and that the growth of grass was much less luxuriant than in previous years. A further piece of informa- tion is that a second control bullock, which was weighed on the same days as No. 23 and which accompanied the latter to and from Kodok, lost 21 kg. weight although it was grazing natural pasture along with the Government herd ; for this loss we can offer no explanation except to repeat that in this particular year grass was not as abundant as is usua\. NAGDJAR GRAZING TRIAL About two miles west of the Sobat, alongside the Malakal-Nagdiar road, this four-feddan field was on land subject to heavier flooding than the two previous fields. The principal grasses were Selaria il/craSSala and Andropogon gayanus, with small quantities of Hyparrhenia, Sorghum, Panicum, Sporobolus, and Echillochloa species. As with the Kodok experiment, there was some delay in starting, and six bullocks were therefore used in the first season. These were weighed , at Malakal on 6th March 1952 and taken by steamer to Nagdiar, at which village they were to be tethered nightly. Grazing of the test field began on 7th March 1952. Because of the distance from the river, water could not be supplied at the field and the animals were only able to drink at early morning, before moving to the field, and in the evelling on their return to camp. They all appeared to be affected by this long, thirsty day in tbe sun and a grass roof was built to provide shade; but the Selaria incrassala, of which the roof was made, was evidently more attractive than the local grass and was soon consumed. By the end of April all six beasts were looking thin althougb there was plenty of grass. One animal died on 15th Mayas a result of a digestive disorder of which constipation was the main symptom. Bone meal was fed to two of the bullocks but without any effect. On 19th June the animals were brought back to Malakal and weighed; there was an average loss of 4·4 kg. per head, although two animals-not those receiving bone meal-actually showed a gain. It must be remembered that the new season's growth of grass had already appeared, as at Gonia and Kodak, and was by this time quite appreciable at Nagdiar. This trial, like the Kodok one, was repeated in the following dry season commencing in November 1952, when the grass, though mature, was not yet dry and dead. Only two bullocks were used this time, both grazing within the field from 26th November 1952 until 23rd May 1953, by which time the new season's growth of grass was about 20 cm. high. Nevertheless both animals lost weight, No. 31 from 313 kg. to 245 kg., and No. 42 from 283 kg. to 226 kg.-an average loss of 62 kg. per head. There was no shortage of grass at any time during this second dry season, although latterly it was coarse, and both animals appeared to keep their condition well, particularly No. 31 which lost the more weight. CONCLUSIONS From the information gained from these three short grazing experiments it appears that mature high land, or high intermediate land, grasses of the types described cannot satisfactorily support cattle throughout the dry season, and that to supply suffiCient bulk alone at least four feddans would be required for each Animal Unit. 1044 I ,j" ~. ~ '. TABLE 553 ~. ,~. . GONIO GRAZING TRIAL '. LIVESTOCK WEIGHTS. 1952 t, Date Bullock Bullock Bullock Bullock No. 21 No. 22 No. 28 No. 29 1 February 225 258 318 268 1 229 260 325 269 14 .. , , 221 254 319 265 21 ... 224 250 316 257 28 :. : 225 254 320' 265 7 Ma~~h ~ 222 258 313 256 14 ' ! 222 254 313 252 21 , 221 256 311 259 28 , .. 2 19 253 309 251 7 Ap~'i1 219 248 290 245 14 : ! 209 232 285 231 2 1 209 232 283 235 29 ... 206 229 211 235 6 May 210 226 269 235 13 205 221 259 226 20 198 22. 250 228 21 196 222 2lO 228 I 3 June. 0 •• ! 198 221 2lO 230 10 ......... 200 229 247 232 18 ' 211 231 248 246 21 218 234 260 256 6 J.iiy ::: 219 240 255 255 23 235 256 263 280 S A~~~'t 234 244 251 257 12 249 250 262 292 19 251 259 211 299 26 252 255 213 304 2 September 248 260 278 305 9 246 263 283 310 16 246 261 270 303 23 243 263 282 )03 30 254 273 214 304 7 October 248 261 264 298 14 245 268 264 301 TABLE 554 ANALYSIS OF SOME MATURE SUDAN GRASSES (ON A DRY MATTER BASIS) I, M I oi,tu'" I' C",d~ I Fa. II SCoalrubbole- '~ Fibr. Ash I Remarks Protem hydrate ' I Hyp"',h.niaspp. .. . . .. 1 4'0 ' "3'1 1--1-'5-'-'- 4-6-,4----37-. 3---,--1~ .:- !I CUt N=~alakal. SorghumpUl'pureo-serfceum j 4·0 2·6 1·3 47 ·8 37 ·0 t 1·0 Cut Nov. near Kodok Grazing Trial. PenniStlUm romosum.. . I 3·0 )·7 1·8 41·0 37·0 10,0 Cut Nov . near Malakal . 1 Setaria spp. _. . 1 5{) 7{) 2·0 4S'2 29'1 j 16-0 I Mainly S. incrasSOfO hay (At' I' cut at flowering StqCl. Scluwv/tldio Iracilu 2{) 1-1 51 ·6 36'4 8·9 Cut Nov. in Kosti Ois- IUnIn juella) . lrict, common near . Saba-Asuda weil-cmlre. THE SEMI-ARID REGION At the outset of the investigation it was known that some grasslands in the Semi-Arid Region provided good grazing all the year round, but the stock-carrying capacities of these pastures were unknown_ The only unused grasslands of this Region were in the south, border- . ing on the Transitional Belt between the Semi-Arid and Flood Regions, and grazing potentialities were not known_ Since time was short and trained staff not available, simple experiments had to be designed to give an indication of the nutritive value or these grasses and their stock- . carrying capacities. Such .experiments were conducted in two parts of Kosti District, one near Jebel Megeillis, and one around the Saba-Asuda well-centre, about 30 km. south-west of KostL 1045 16· JEBEL MEGEINIS EXPERIMENT The grassland in the vicinity of Jehel Megeinis is predominantly hariq grassland, hut there is a small proportion of the annual short grass type which is more characteristic of the northern part of the Semi-Arid Region. With the ohject of assessing the feeding and grazing values of these pastures, clippings of the more prominent grasses were taken throughout the period July- November 1950 and in July-October 1951. The samples were then submitted for chemical · analysis. In additi on samples from different hays made in both years were also submitted for analysis, and it was hoped that larger samples could be fed to sheep so that some indication of the palatability of the herbage could be obtained. As far as was possible grass samples of the selected grasses were taken from pure stands. (These were Aristida mutabilis (Ar. dambalab) which is common in the Semi-Arid Region on the short grass plain, Sorghum purpureo-ser;ceul/l (Ar. allis) , Hyparrhellia sp. , probably H . pseudocymbaria (Ar. anzora), and Cymbopogon lIervatus (Ar. nal) which are the common grasses of the hariq grasslands.) Although this was the intention it will be appreciated that it was not always possible to find completely pure stands and tbe analyses should be considered as those of mixtures dominated by the grass species indicated. FIRST SEASON (1950) In the first year many imperfections in the sampling technique became apparent, and the experiment must be considered incomplete in many respects. The main reason for this was that tbe sampling area was far away from Malakal, with the result that no supervision was possible during the rainy season when the samples were being collected, GRASS SAMPLES FOR ANALYSIS As far as was possible different areas, carrying pure stands of the above four grasses, were selected monthly between the months of July and November inclusive and enough herbage for analysis (roughly one kilogram me) was cut from each selected site. Prior to cutting, the heigbt of the grass was recorded. This was done by taking the average of the heights of twenty plants. Immediately after cutting the berbage was weighed, allowed to dry, then re-weighed. Unfortunately, since the areas from wruch the samples were taken were irregular in sbape, the size of the plots (given in Table 555 at the end of this section) are only approximate. No replicates of the samples were taken. HAY SAMPLES Hay was made only from Aristida mutabilis (Ar. dambalab). Two cuts were taken. The first was on 24th August prior to the flowering of the grass, tbe cut herbage taking about four days to dry ; the second cut was on 10th October, i.e. well after flowering when the herbage was fairly coarse. Unfortunately no record of the size of the area from which the grass was cut was made and no weigbts were recorded. Samples were submitted for analysis and a digestibility trial was conducted at the Faculty of Agriculture, Shambat. Tbe results of the analysis and trial are given in Tables 557 and 558. SECOND SEASON (1951) In the first season samples sufficient for chemical analysis were collected but could not be related to areas; in tbe second season trus deficiency was made good and all samples were duplicated. For each of the four grass species two sampling units, each 16 sq. m. , were selected and fenced. Each unit was quartered and each quarter was cut once during tbe season, the first in July, the second in August, tbe third in September, and tbe fourth in October. The heights of the grasses prior to cutting were recorded and the weights of the cut herbage, both fresh and dry, were noted. Hay was also made from additional areas, as in the first season, but greater care was taken to ensure that the area from which the grass was cut was measured and that yields were recorded. GRASS SAMPLES FOR ANALYSIS The heights and dry weights of the samples and the areas from which they were cut are recorded in Table 556, together with the result of the cbemical analysis. It will be noted that the· cuts "for each species have been duplicated . It will be further noted that the chemical analysis is in two parts, the second part being on a soil-free basis. This was found necessary because . when the samples arrived at the Faculty of Agriculture, Sham bat, tlaey were found to be heavily contaminated with soil. The analysis was conducted in the usual manner and then the per- centage of soil in each sample analysed was calculated. The figures expressing each constituent 1046 as a percentage of the dry matter on a soi l-free basis were then calculated and are given in the second part of the table. TI,ese are the figures which should be used in any comparative study, since they give as accurately as can be assessed the true chemical composition for each of the grasses at the different growth stages . .H AY SAMPLES For each of the four main grass species two additional units were fen ced, the first of 100 sq. m. and the second of 25 sq. m. The larger units were cut for hay in late August and the smaller in late November. It was intended that all eight hay sa mples should be submitted to chemical analysis and be fed to sheep to determine digestibility but, partly owing to the light rains, the quantity was insufficient fo r feeding trials. The results of the chemical analyses are given in Table 559. INTERPRETATION OF RESULTS AND ANALYSES The clip samples of aU four species indicate a general trend in change of chemical com- position during the season in both years. In the 1950 results the rapidity with which the nutritives value declined between July and August and thereafter is striking, the most significant change taking place with Sorghum purpureo-sericeum which had a surprisingly high protein content in the July cut. The decline in nutriti ve value was not so rapid in the 195 1 samples ; in fact there was a gradual falling olf. In comparing the analyses for the two years it should be noted that the growth in 1951 was at least one month behind that of 1950 ; rainfall was prohahly the factor most responsible. The results of both years' analyses indicated that the feeding val ue of aU four grasses was high in the early rains period (July 1950, August 1951). This was particularly so in the case of Sorghum purpureo-sericeum which had the highest crude protein content of all the first samples in both years. The results of the analyses of the second samples differed considerably; those of the 1950 samples indicated low feeding values, whereas those of 195 1 were only slightly inferior to the first cuts and indicated a higher feeding value than the grasses cut at the compar- able period in 1950. By the time the third cut was taken samples for all four grasses and for both years showed extremely low protein and soluble ash and high crude fibre content. Even allowing for high digestibility of the crude fibre it is not unreasonable to assume that such fodders fed alone would not supply maintenance requirements. The results of the ten hay samples analysed indicate that hay made from grasses after flowering (say in October or November) is 'of poor feeding value and would not supply maintenance requirements. On the other hand hay made from grasses cut prior to flowering or in August should provide reasonably good fodder, with the exception of that made from Aristida mutabilis. 1047 TABLE 555 JEBEL MEGEINIS ANALYSIS OF GRASS SAMPLES, 1950 (on dry matter basis) Size of I, Average Grass Date of Area CU, Height Crude N-Free Silica. Phosphorus culling sq. m. orOrass Cru Ash Protedien .I' E EXllhraecrt . I Fibre EXlrdct. Si0 p 2 I I I in em. , I I Aristlda mUJabilis (Ar. Dam- 21. 7.50 1·5 I 18 1,021'5 II 138·88 1 13'59 I 9'52 2·85 30-42 40·75 16-46 10'30 0·53 0 ·13 bolob.) 21. 8.50 2·0 49 H5 1·98 34·83 43·70 14-04 9 · 11 0-36 0·07 : 1,135,0 296·66 26· 14 \ 21. 9.50 2·0 71 ' 1,8 16'0 44H5 I 24·52 3-21 1·52 36-53 47·68 11 ·06 8-46 0·31 0 ·11 21.10.50 50 1·56 1-00 38·13 45-2 1 14·10 12·63 0·28 0·04 21.11.50 I 908·0 683-68 75-30 'I 50 908'0 876·82 96·57 1·46 0·99 37·75 48 ·77 11·03 9·63 0 ·23 0 ·14 20 21.11.50 I 1·36 0 -8 1 38· 16 48-46 11·21 9·65 0·22 0·16 \ Sorghum purpllr~o-seriullm 17. 7.50 I 6-0 30 I 1,IlS I 133-43 1 11 ·76 20'36 2·90 22-32 35 '56 18·86 3·86 0·99 0'28 (Ar. Anis.) 17. 8.50 I' ] ·0 100 25·33 5·90 1·48 34·29 45·96 12·36 3-2 1 0·52 0·17 17. 9'50 0 ·5 150 : :~;~ 1 ;~~:;~ 1 41-61 2·47 0 ·94 38·90 49-38 8·31 4·65 0'39 0 ·16 17.10.50 200 2,~ 79~41 3~7 1·27 0·75 42·17 48 ·53 7 ·28 4·75 0'30 0·08 17.10.50 Ii 1·54 0 ·78 41 ·40 47 ·34 8·94 4·97 0·45 0·08 i 22.11.50 I \ 200 I 1·45 0·60 41·05 47·33 9,57 5· 17 0-39 0·06 20 908 i 90 1'38 99-27 22.11.50 1,37 1·28 41 ,15 47·59 8·6] 5·01 0 '39 0 ·06 Hypo'rrhe"ia psclldocym- 2 1. 7~-1; 0 20 1,135 134'04 11-81 8·90 3·35 28·54 44·27 14,94 9·64 0 ,44 0·26 haria {Ar. Anzoro) 2 1. 8.50 1·5 61 1.362 289'44 21'25 4·86 2-41 32·42 49'04 11 ·27 6·64 0'39 0· 13 21. 9.50 0·5 150 1.816 527, 18 29-03 3·2 1 2·04 37-45 49·33 7·97 5·03 0-32 0·13 21.10.50 II - 150 908 598·30 65·89 2·38 1·9 1 37·77 50'19 7·75 5·72 0·34 0·15 21.11.50 2·0 150 908 895·61 I 98·64 1·82 1·17 40·84 48·76 7-41 5·72 0·27 0' 12 21.11.50 r 1·45 0-88 39·96 47-61 10·10 5-50 0·38 0'18 Cymbopogoll neno/us (Ar. 21. 7.50 0·5 15 1,135 71·25 I 6.28 9·96 2·71 24·20 49·76 13·37 5'10 0 ·90 0-37 No/) 21. 8.50 0·5 46 1.362 185'18 I 13·60 6·45 3·35 29·66 49·90 10 '64 5-81 0·58 0·27 21. 9.50 ] ·0 100 1,81 6 537,93 29·62 3·18 2-00 )4-92 50·16 9'74 6·18 0 ·33 0·19 • 21.10.50 150 \ 908 792'73 I 87-30 4-00 2·53 32·94 51 ,82 8·71 5'59 0·37 0·16 21.11.50 I I ],27 0·5 100 908 900'74 99'20 1·69 36·58 52-22 8'24 5·80 0·31 0 ,11 21.11.50 1·37 1·34 39-42 50'35 7·52 5·01 0-32 0'10 1 TABLE 556 JEBEL MEGEINIS ANALYSIS OF GRASS SAMPLES, 1951 (on dry matter basis) Date Size of Average Grass of Area Height ~rt Mo~ture :I Crude ' Ether I Crude 'I N-Free I Total I S;I;ca I SZi ;n I Crude I Ether Crude N-Free ! Ash Cutting Cut of Grass Sample in Protein Extract. Fibre Extract. Ash SiO, Sample Protein I Extract. Fibre Extract' l I .sq.m. m em, in gm. Sample I ' Ar;'Ma mutabiUs (Ar. Dam- I 25. 7.51 4 1 7 91-'8---'1i-3-'-so-+I--3-'8-5-t-0- '5- 7-'i-1-0'-36-+1- 1-7-67 1 67-54 I 56·82 I 62-90 I'~ ~~~8--II,- ; '~4- ' 2792 ~;-I~- bolab: I~ ~: n: : 16 m~ : l~ 1:~ ? ~~ ~.~~ i~2~ ~:~ nn i ~n2 I~~ l~~ ~2~~ :m , 12j: , ~~: ~:~ : : j& 2130 I ~ii ::~~ n~ ~~1~ 4253 l: ~ : ~~~ :~~ i ~ :~g iif 1 ~l~~ I ~:~~ : ~~ ; iU~: : 1 19 , ~m lI ~~l , ~5~ r~~ iYN ~.~ i : ~~~ : ~~i 'I : ~~t II ni g~ m~ I ml m 1 25.10.51+ _4 _; -_50- -._ 703_·5 + _3-3_8 _: __2-49 1·74 32-81 46·92 I 16·04 14·2 1 13·10 2-87 2'()() 37'76 ,' 53-99 3-38 ___ ____- +' _ _ --'-__. ..1-._ _' --_-'-! __. ..1-._ _ ._ ._ __ _________- -''--_--'-_ _ _ ~nl : I : 1\ ~ ~:~} Insufficient Material for A nalysis ~. SO[r:~':tn&rpureo-serjceum ~;: ~:~ : ! I !g tn:~ ~:~~ , g:;~ i ~:f~ ! ~i:~ 43-25 19·02 ~ :~:~ ! 14·77 2-49 , 24-92 , 48·)6 9-46 44-26 17·02 14-69 I' 2-35 25-09 48·15 9·72 25. 9.5 1 4 15 214-9 5·75 I %6 i 2-44 125-46 i 402 IH2 I 9.()() : 10·28 2-60 27-09 , 47·58 12-45 ~L~l: : ~ i~~~ ~:'i'i :~ I li~ ~gg I 48·84 14'55 51·17 13·71 , ~.~ 2-28 28·23 , 52-26 8·56 1 1·89 30·67 i 54·30 8·43 25.10.51 4 30 322.0 n5 ! 4·14 , 1'82 31·17 49·90 12·97 7'53 ml 1·92 )2-91 1 i 52-69 8· 11 -.------.----~--r___I----T----r--~- I ------'------ - -._--'--- Hyparrhenfa pseudocymbarJa (Ar. Anzora) ~~: i:~ ! ~ i:~ } Insufficient Material for Analysis ~~~l : : g 37-24 24·76 8·37 3·08 24-47 ! 40·64 17·88 1m 1~g : : n~ I ~~~ I ~~i} 40·06 19-69 6-43 25 ·87 1 42-8 1 14'18 ~l ~l : : i~ }~6 ~.~ II ~~i i ~;~ ~g~ 50·64 12-85 4·90 21-·8481 I' 28·24 8·36 I 46·52 11'61 9·2 1 HI 26·73 I liP. : 9·26 25.10.5 1 4 75 203-8 4·35 4·55 1 41 33 23 51·06 9·75 3·40 1·46 34·40 52-86 6·57 25.10.51 4 80 316'7 4·03 4·74 I 141 i 33 11 50·65 10-09 3-83 1·47 ! 34·4) 52-67 6·50 ! Cymbopogon nervalllS (Ar. 25. 7.51 5 1·0 Nal) 25 7.51 4 0·5 I} Insufficient Material for Analysis 25. 8.5 1 5 29·0 10·73 1].08 4·18 21 '5 1 I 21 ·40 13-38 12·13 25. 8.5 1 I 14-89 4·76 24-4S 45,)) 10·54 5 16' 1 5·28 11 ·13 H5 18·6) !r:~~ ! 24·01 14-52 1)-46 25 . 9.51 40 261 ·2 I 49-62 13-29 7·10 I 4·79 , 1~~~ 5-37 , 48·13 12·19 1 6·33 8·92 4-48 2H8 4-98 i ~! ::~ 52·12 8-66 25. 9.5 1 15 167·5 5<63 8,)9 2· 12 22·78 48·77 ; 17-94 1 11 ·55 10·05 9·33 I 25 ·)3 25,/0,5 1 2-36 54·22 8·76 55 166'2 5-68 4·73 2-44 31·63 50·94 10-26 6· 15 H8 , 4·9 1 I' 2·53 32·84 52·89 6·83 1 25.10.51 50 219-5 5·98 4·14 2·02 32-97 50·15 I 10·72 I 5 78 I H5 I 4-28 2-09 I 34·08 ; 51·83 7'72 TABLE 557 JEBEL MEOEIN[S ANALYS[S OF ARlSTlDA MUTADIUS HAY, 1950 CoNPQSmON Of DRY MA1TI.R 1 --I I I'D ICESTIBLE CoEFFICIENTS ~ 0 GESTl NUTRIENTS I I IM Ois- IO ;g;"';c ' crude ' Ether I crude ' N-Free Siii"';-'j Organ'ic cr~d~--I-E;her : Crude N-Fr.. Organic IC ~d.i:ther crude,-N---Fr-.-. S.E. T.D.N. lure Matter Protein Extract. Fibre Extract. Ash SiOI I Matter Protein Extract. ! Fibre Extract. Matter Protein IE xtract. Fibre Extract . ~__ t _~_I_~_:_~_W_U_i_~'4_- _~_~_-_r~~_~_:4~_T_;_;~i_[_~_~_-~· _3_·_~_-~r_~_·_~_5+i _16_~_~-_+r_~_3_~_; +' _n_~_4; '_~_;_~-_~1_~_~~__:_-~[~__'_H_'~' _4_~_~_~_-~-_~_:~_41_T_~_::_ ~ ! ~~;r;~' CUt after Howeringi 3·33 85·01 T[~2: 1·32 31'16 44·12 14'93 IH2 5[,04 Nil 39·10 63·10 ~'95 43-43 1 Ntl f52F 2405 r:r::-:53- 1 [ ,09 I I l TABLE 558 PERCENTAGE OF ARiST/DA MUTADILIS HAY CONSUMED BY SHEEP (CUt at Je~1 Megeinis, [950) ______ __M_ a_t_cn.~_ _____~ ----S-h)-t'P--~-.---S-~-p----I------~~_ __ Hay cut before flowering ... 92·41 [00·00 96·21 Hay cut after flowering ... 11-61 91'19 87·10 'J!''''' . • "T" TABLE 559 Jl!BEL MEGEINIS ANALYSIS OF HAY SAMPLES, 1951 (on dry matter basis) Ash Silica Sample SiO, I Arls/ida muloblfis .. . I 20-79 17-14 I 20-74 18-89 Sor,Jlllm purpuuD-serlceum 45-92 ( 1~;;--1--9-94 51-54 11-00 7-()6 HyparrMnlo pseudocymbaria 100 41-40 25 48 -47 t:_~_.-_:_:_:_ Cymbopogon ntrValu$ 25. 8.51 100 44·63 19-80 11 ·19 I 25.11.51 54·87 7-58 4-51 SABA-ASUDA EXPERIMENT The villages of Saba and Asuda are approximately 30 km. south-west of Kosti and about the same distance from the Nile and from the nearest large well-centre at Gedid. At Saba there is one well-field and at Asuda, about 3 km. away, there are two. The country around this centre is fairly typical qoz catena with sandy, almost treeless qoz, producing short annual grasses, alternating with flat plains covered partly by fairly thick Acacia seya/ and Acacia melli/era and partly by open grassland (see Vol. I, Chap . 2, pp. 148-50). Thousands of cattle, sheep, and goats were known to water at this centre in the dry season and it was sufficiently far from any large watering-centre to be considered an entity. It was thought that the stocking intensity of the area could be measured by counting the number of animals watering at this centre throughout the dry season and relating this number to the area of grassland grazed by anima ls from thi s centre. Greater accuracy could have been obtained by fencing a known area and grazing it witb a controlled number of livestock; but this method would have required staff, materia ls, livestock, and time, none of which were available in sufficient quantity. During the dry season before the 1950 rains a recorder was stationed at each of the three groups of wells with instructions to record daily the number of cattle, sheep, and goats watering at the centre. In add ition there was one ranger whose duty it was to join a herd as it left the centre after watering and accompany it unti l it returned to water the following day or two days later. At the grazing pl ace farthest from the well-centre the ranger blazed a tree to mark the spo!. This routine, though simple, did not work smoothly at first, but from February onwards it went well. In this way a series of points soon circled the watering-centre indicating the perimeter of the area grazed, and the perimeter was gradually pushed out until, in April, the area grazed was estimated to be 400 sq. km. Distances from the centre to the perimeter were measured on the mileometer of a motor vehicle and, owi ng to the impossibility of moving in a straight line, this method was liable to error. No distances were measured to the north-west because thick bush made it impossible for a vehicle to get through . From Table 561 below it will be seen that the average number of Animal Units watering daily at this well-centre was 4,146, indicating a stocking rate of 10 A.U. per sq. km., or 1 A.U. per 24 feddans . It must be remembered that groups of cattle sometimes came into this area both from the ri ver and from the well-centre at Gedid, so that the stocking intensity may have been slightly higher th an indicated by our records. Another point of note is that the stocking rate of I A.U . to 24 feddans applies to 24 feddans of land, which includes bush, cultivations, villages, and small encampments; but as a working figure it is sufficiently accurate for our, purpose. TABLE 560 SA BA-AS UDA WELL-CENTRE NUMBER OF ANIMALS RECORDED WATERING, DRY SEASON 1950 SABA ASVOA I AsUD A (small) (large) Weekly Periods Callie I[ S~~ and Callie 1"- ~h~P- and ~-. ::-- 1 -;b~ :~- ____ . ___. _ __ L Goats Goats Goats Feb')!~h':14th .. . i. 16,28l '-:'848 --,:--::'] - -:: I ~~ 1 9 I 1,4l6 -22nd-28th . .. I3 ,Ol4 I 11,008 1I,.I4l I 1,48l 1,493 1,629 T::-'---':~-1~9~--- ;o~T-;;;~-r-~1--2,912 ! 3~ March 8th-14th 18,352 16,478 i 4,826 l,446 1,689 1,094 22nd-28th 16,880 12,740 l l,I61 II 6,091 1 4,842 3,672 TOtal~_.~_~ 39,2 18 9,987 I 11,537 6,l31 4,766 I I 1,946 --:1- -I,-32- l-+-- 8-,4-8-8-+--:-- 8-,81--7 -April8 th-14th 13,456 1,4l6 22nd-28th ::: 13,780 1,3l6 1,349 1,831 7,924 9,035 Total 27,236 2,8 12 3,29l I 3,156 16,412 17,852 1052 TABLE 56 1 SADA-ASUDA WELL-CENTRE AVERAGE NUMBER OF AN IMA LS WAT ER ING DAILY Monlh Cattle Sheep and Goats Animal Units February 4,392 2,293 4,679 March 3,696 3,966 4,192 April 3,35) 1,70 1 3,566 ____. ________. ..1..-- ___ _ Average through the dry season . 3,8 13 2.653 4, 146 4. IRRIGATED PASTURES The following experiments were designed primarily to give some indication of the amount of water required to keep a pasture green and productive throughout the dry season and to provide information on the management of irrigated pastures (stock-carrying capacity, hay- making, etc.). Other experimental work in connection with hay-making has been described in Chapter 4, Vol. I. Finally some results are given of the preuminary work carried out in con- nection with introduced and indigenous grasses and legumes wllich were studied to determ ine their suitability for use in irrigated pastures. All the experiments and investigations were conducted at tile Malakal Experimental Station (East Bank), and the reader is therefore referred to pp. 990-1 of this volume for an acco unt of the soil and climatic conditions under which _t he experiments were carried out. WATER-DUTY: RATE AND FREQUENCY Two closely related experiments were set out 10 determine the optimum rate and frequency of applying water to pasture, in the dry season , in order to maintain grass growth. EXPERIMENT A This experiment was replicated four times, and conducted over a term of two yea rs (1950- 1951 and 1951-1952) on a natural SeTaria incrassala type pasture. This is the dominant type of pasture around Malakal and has the following grass composition: Setaria incrassola Dominant grass. Ischaemun1 brachyatherun1 . Hyparrhenia spp. (H. mla and H. elisso/llfo) I Panirwn porrphyrhizos Pennisetum ramosum A ll occurring in varying proport ions. Cymhopogoll sp. Andropogon gayanus Cyperus sp. Small plots -fli-feddan in area (33 m. X 12·75 m.) were marked out and enclosed within small banks so that water could be impounded. Each plot was separated from its neighbour by an open space 1·5 m. wide along its longitudinal axis, and by the minor irrigation channels along its shorter axis. Each plot was given one of the follo wing irrigation treatm ents during the dry season: 200 rns. water per feddan applied every 7 days. 200 m'. 14 " 200 m3. 2 1 " 200m', 28 400 rn3• 7 " 400m3 • 14 " 400m3. 2 1 400 rn3. 28 " 600 013• 14 600m3, 21 " . 600 m'. 28 Control (no water applied). The above treatments were replicated four times, so that in all there were 48 plots. 1053 ~. The irrigation water was drawn from the Nile by means of a 6-inch pump and conducted to the experimental area by an unlined main channel. A v-notch and a gauge were erected in the channel , enabling the rate of water flow to be measured. The water was conducted to the plots by a series of minor channels from which it was allowed to flow on to the plots by openings made in the surrounding banks. The amount of water passing on to a plot was controlled, and the time required for the requisite amount of water to pass on to the plot was calculated from the rate of flow in the main channel. Differences as a result of varying treatments were estimated on the basis of the amount of herbage produced by the plots throughout the irrigation period. All cutting of the grass was done by hand with sickles, and fresh and air-dried weights were recorded. Analysis of the data was, however, limited to air-dried weights. As the result of experience in the first year, changes in experimental technique were found desirable in the second season; for this reason the results must be considered separately. FIRST SEASON (1950-51) The results showed a significant difference between groups of treatments, the general trend being an increase in yield of grass with higher applications of water. TABLE 562 EFFECT OF DIFFERENT WATERING TREATMENTS ON GRASS YIELDS, 1950-51 Yield of Number Total Amount Grass (Air· Treatment or of Dried Weights (Ac.tual) Applications Water Applied from Plot ml. p.r. 33 m. X 12·75 m.) in kg. 438 ml,p.r. every 14 days 12 5,250 145·1 598 14 12 7, 180 143·3 452 21 9 4,070 141'1 225 7 24 5,400 140-1 631 21 ... • 9 5,680 133·0 418 7 ... i 22 9,200 125-1 641 28 7 4,490 115·6 255 14 12 3,060 103·8 395 28 6 2,370 99·1 195 28 6 1, 170 91-4 274.. 21 9 , 2,470 80·5 Control ... 1 Nil 50·0 i Si.niro::'n, ~i~erence(p-o,o~)l----- r----- T--;~--- SECOND SEASON (1951-52) In the first season yields were based on the amount of grass cut at intervals from within the total area of the plot (33 m. x 12·75 m.). It was noticed, however, that there was a strong border effect, the grass around the fringes of the plots being taller and more vigorous in growth than that towards the centre. This was attributed to lack of competition around the outer edges and to the influence of a greater water supply caused by moisture seeping from the irrigation channels and adjoining plots. To overcome the effect of seepage the actual area from wbich measured cuts were taken in the 1951-52 season was reduced to 30 m. x 10 m., thus ignoring a border of approximately 1·5 m. in width around the plots. In the 1950-51 season eacb plot was cut wben the grass reached the early flowering stage, and between January and June each plot was cut approximately 7 times . Very often, bowever, cuts were taken either too early (well before flowering) or too late (well after flowering) and on occasions some plots were overlooked; this was due to the lack of experience of the operators. In the second season each plot was cut to a definite schedule. It was hoped that from the analysis of the herbage cut some information would be gained on the compositiol\. of the berbage at different growth stages and also some indication of berbage production trends. Both tbese are important from the point of view of the management of pasture. 1054 EaGh plot (30 m. X 10 m.) was therefore , ub-divid ed by marker peg' into 10 sub-pl ots, 6 m. X 5 m. in area, each of which was cut as fo llows: Sub-plol (a) cut every week. (b) .. 2 weeks. (e) " 3 (d) .. 4 (el " 5 (f) 6 (g) " 7 (II) 8 (I) 9 U) " to Samples cut at the same intervals of time were bu lked together after fresh and air-dried weigh ts had been measured, and at the end of the season a sa mple of each (i.e. grass one week old, two weeks old, etc.) was submitted to the School of Agricu lture, Sham bat, for chemical analysis. Furthermore greater care was taken to ensu re that the watering treatment conformed as closely as possible to that set out in the design of the experiment. As in the previous season, the results showed a significant difte rence between certain groups of treatments, the same general trend observed in the ftrst season again being ev ident. TABLE 563 EFFECT OF DIFFERENT WATERING TREATMENTS ON GRASS YIELDS, 195 1-52 ' ; Yield of Grass N~rber Total ~;nOUnl ! (AJ:;i~~~ Treatment Applic.ltions \ wat~3 ~~f.lied i3 0 r~~ ~~o~.) 400 m'.p.f. every 7 days T~ ~- U&l -II i~k; 600 14 .. 200 7 24 4,800 34-8 200 14 12 lAOO I 30·5 600 21 •• 8 4,800 30·5 400 .. 14 11 4,800 27· 3 600 28 6 3.600 27 '1 400 21 8 ! ) ,200 1 27 ' 1 400 28 6 2,400 25-2 200 28 6 i 1.200 21·8 200 II 21 8 I 1.600 I 20-5 Control I Ni l 8-7 -, ---1----· Significant Difference (p- O'OSlj 7·2 The analysis of the grass cut from the plots is given in Table 564 below: TABLE 564 S ETARIA INCRASSATA ANALYSIS OF GRASS SAMPLES, 1951-52 (Composition of Dry Matter) Sample _1_C rude . Ether i Crude I- N-Free I I Silica ~::e~ _ i E~iract. tt Fibre Extract. ~~_ __ ~i~.2_._ -G-ra-scsu-,-a'- I-wee-kI-y-,-n--Ie-rv-als 1 8·2] -1-" )'59 ~1'32 -" ~48--iI- 1 8'39 14· 18 2 I 8-63 i 2·83 30-77 : 35·23 22-53 17-8 1 3 8-42 1 4 7-21 I 1·91 i 31 ·21 I 38-55 19-9 1 15-36 1-89 : 30- 17 38-00 22·72 17-8 1 5 7-3 1 1'77 ' 32-41 38·66 , 19·86 14-99 6 1 6-70 2-20 28·92 39-01 23 ·17 17-63 1 .1 7-09 1-94 28-90 I 38-03 24-03 17-96 8 6·82 1·58 32-61 37·39 · 21-60 16-90 .... 9 i 6·26 I-54 28-39 40·31 23-51 11-49 .. .. 10 5·78 1·53 )J-IS 38-62 20·92 15-83 1055 EXPERTMENT B This experiment, which was not replicated. was carried over a term of two years (1950-51 and 1951-52). A pasture composed only of Echillochloa pyramidalis was established and plots 33 m. x 12·75 m. were marked out as in Experiment A. The Echinocl,loa pyramMalis was collected from the riverain floOd-plain at Malakal. It was thought that this grass, which is a swamp grass, would have a higher water requirement than the Selaria illcrassala type grassland and the experiment was really put down to confirm this supposition. Different plots of Echilloch/oa p)'ramidalis were su bjected to one of the following treatments during the dry season: 300 m3 , water per fedda n applied every 7 days. t4 21 28 7 14 21 28 In all other respects the management was identical with that given to Experiment A. The results are again considered separately. FIRST SEASON (1950-51) In general terms. higher water treatments gave higher yields, although there was no definite steady trend in this respect. TABLE 565 EFFECT OF DIFFERENT WATERING TREATMENTS ON GRASS Y IELDS, 1950-5 1 ECHINOCHLO A PYRAMIDALIS Yield or Treatment Number Total Amount Grass (Air. (ba sed on average or or Dried Weights throughout dry season) Applications W:ltcr Applied from plols m~.p.f. 33 m. x 12·75 m.l in kg. 307 m'.p.r. every 21 days 8 2.460 227·0 605 7 ::: i 23 13,920 141·5 614 21 10 " .. · 1 6. 140 135·0 612 14 ... I 12 7,340 127·5 )10 14 " ... 1 12 3,720 123·0 288 7 " 3 12 28 " " .. ... 24 6,920 104·0 i 6 1,870 100·5 600 28 ,. , 6 3,600 72-8 SECOND SEASON (1951-52) All changes in management adopted in the seco nd season in Experiment A were also applied to Experiment B. Although again there was no definite trend, the higher applications of water generally produced the higher yields, thus following to a certain degree the results of the first season. TABLE 566 EFFECT OF DIFFERENT WATERING TREATMENTS ON GRASS YIELDS, 1951-52 ECHINOCHLOA PYRAMIDALIS Yield of Number Total Amount Grass (Ail- or or Dried Weights Treatment Applications Wlic ; a ll were imm'llure. Severa l o f the yo unger a nimals were entire on pu rchase a nd were ca~tra l cd hcforc Ihe trials commenced. Detai ls o r the va rious fodde r crops a re given in Chaplcr 5 o r Ihi s vo lu me (Pl'. 1003. IOUX). METII O!) Severa l P CIIS, cuc h large enough 10 hol~ Iwo bull ocks, were co nstructed at Malaka! of Im.:a l limher,