1‘ 0.1 V VI. - .. lilu1l I . ...!5 . 5‘. 0... an: 0.03.2.3. ... 0.... 0, . ., I . .4 . ... p. . . .. .... . ... . do...” .....ofko {06“. .n. g: “PH-pix VHRfiiPfig o» 23.0.. .....i.».k. :. 3...“...20. . . é . ....3..L.s‘.0 $3)... ....0..- 3... :7. ..0. . . o ”8 .v,..l..000 '0 0. .0... 0- .3-.0 .03.... .20.. '0...-. O... 0.0.0.000 0 0.4.0.10”- 0...) ... .1. 100...... 0.0. .... . . ...0 ..0...‘(..... . Z . I. I . . $1.... .00 0-. ...-3...}... . .0: cil... :- 0 0.. 0‘... .....‘0 1:0.- .00 3 .T: T... . ... 2201.. .... ... 2... ..w ... 0., . ¢ . n . \I 000 0 ...-0... 0 .!.S..r..:8‘0...0.3..0.|0. 1.0.20.0 . 13.00:} S... J ’- V0.0 0...)..000...0...0.. 20.- . . .v n 1.. . . 0. . . .5. 0.00 .0: X .01! 09.0.4005! _ . .. 34.00.. ...0 0 . ’03.: 00..., :‘.’.0. 0-,... 00.00.00000....0}. . ....0‘....0?" . ....I ...00. liar! 0000.,5 .v' I‘finag0a . 3.27... .030. 02:00. . ...: . ..0....r.....).!0.1..1210... 7 . ... . . . _ . . . ...... 1...)...4fuiso _ o 0, 9-0... 01.1.0 . i 0.530 ! g . ..V 03.0 0.0.0.020...“ v 300.....07. 04... _ 5.0.0.2303: 0.0.. ..- 0.. 10.0.... . ... 0 ... l . . . . . . . .... .. .9 N01 \L. .2 40, ....04900. .03.}. a. .90 0 . 00 o 0 30.2,...03 3.0.0 “A 0.! .0 00.“... -0-.0.10...A0 .- ......” . . . . . .. 0.9 0'. ..V. 0. .. a...‘ 0.. . .001. Va... ... J0... .0.V-. ...-0.0. “.0 vn I H1... . 01.3.0.0... 0.7.0.: 0.00M. 0.... 30.0... 01 a .... 0.0.9.0.... 0. 0.01.401... 0 ...- ..017 , f . . . . _ . . . .. . . . . . . ...u. 210......2. 3300099 . . 0 0.....vv 0 9.591%... ...z «.../0. )0”. .. ”0.0.: 0 . o 0 . . . 9.0-.r’1. 0...! 0-..0..00...’0..-..0. €320.00... ..1.!0...70...0. .... . C . ... .... 0 0.0.0. o. . .. . . . . .13.. .0. . ~ 30. .0. 0.. s r g . . . .. T 0 V .0 ...-f a... . ... - “Mn..." .20. . 100.000.! 1. ...-1010",.‘70. .1 .60..“ Q... 7.0-3.030'. {nigwouvuoz .0... . 0 02.0.0... .. 0.0 300. . .n 1 . . . . . . . .. .. 0 10090.30 ¢Y§000u0 0 “10.0“". 2.1. ...-00.. ”a. $0001.... . 3... A. ...-0. a 0 $1.... 00¢... .0 3‘... . . . . .. .. A. . . 0 t . . .0 {vii-N0.“ . .. .30....0 ...V..0..390 .03... . 5...! I. . .0 3}.» 50.0 'L I. 0.003. q , $ ...-610.0 .00 .‘.3- ‘0. 00) .00. ‘I Z . ‘ . 7030 .70 30.33.31: 0 0.0” .. ... .0 1.... p.180gfl,¥nl..34 7. .0 o. 00:.- 0...‘ . 7‘ . 0 . L . m» 0 ...... ”0.0.00 0.. 0 . . .700... «00.0.00‘.‘..'0n . ..0‘...‘ ..00’30000001. .1 00000.- i . . . 0. 0n . L’fl’k“ 0. 00.00... .n‘0.‘. 3000:2300: 01.30.9390 ...-00.00... 7: $00.30 ha ’7. o 0 . c '0... o. _ . . 0 0V. . 2 . .1. 0:00.00... ...: .0001. 0.0.$!3!.00.3 0300.....03007....01.0 0:... 0'0... 0. , . 30.. .l: .l'. . . '0 .. £03. 0. ~ 0.- . ...." L406 . 0.394. . 30". 000000 0. 20000.... 0.. ‘10330....C¢ ... .00. . 0 - . .4 2...... 0;. I. ....0 . ..0 7.40.: w 0.. If.“ 00.00.. ~10..- ..... 0. ......0'40. ‘00. .," 0.005.... . 9 .300 “0%....- r. 0.0 . .3. $0 0.0 0.0 ... ... :5 C... . v... ...: ....- ..3W . .. .. .2 . ... . . ... It .... 0. 31:70.0. . 2‘00...‘ . 0.23:0. .. . .. 3.0.9.0.. . 33900.. . .0 ...000._1 ’10... 0 00.00. 001.00. .... . .94... .4”. '30.‘ .01... .0 30.0....0 04.0... 0.0.9.}!!! 0 J0. 000.02,.‘. ..o..0o....0 ....0. .. v . : :1). .000... 0....3....13)2$..£.: . . m. .0... . : .0 . .... r. . I. .05.“?! 3000.0 3 ..0. «.0 .0...¢00...l.; VI. 0.3.1030". 00 3.0.. 0.2 D 00. f0... ...-0.0.I...‘ a”. ....”Iamfnvuuf u. 0 .9005” 0.. I. 0.1. 3’0 .0 ... (0 f 00-. 0.300000! 0' .0. \o’IéfiL.‘ . - m. . . . x. ...0‘ 0’”. 0000 0 I... .1 _ . ._!'0.\...,‘ '0, ...—$320.... 0.0.0.3... .2..0_..!. .(SIQ. 3.07.03 3.0 ... . .1 0. .01.... .... .0 v . .. v .04 .....1'. Q .0000.- .. .0000...- ..I’.0 a . .1000l’0.-.w .. ”n Vol-“..‘uobli‘ 0.0.0.00000—0..30.r0‘.. .. .. 0. .30. u ’7‘. ...‘nwfl 0.0 .01 ... 10.3.003'.’ ... 1.000.. 0.00 3.0.10 0.. 00.00.70.000 Qanrh.. . . “00.30....030: .0 ’ 010.91.... . 9.00 70.330.90.302! \00 ..v ..01010 .000 ..10’0‘40‘0...‘ .0 a“... . I. . . .. ..It... 04 0 .... 5.! 1.0003 .0 0 .9014 , 00 . 0.. 3.0.06. . . . .. .0... o... . . . .. , . . . . . . 0. 0"" 5.030% .v L00.w0l’000100.00h|.’.0 0. Q . . .00 . '0 my”. “M'LQ.AI” .00.l..0000o‘ 0.. u .0.00H..~.€0~00¢00.3uu2..070 .0...h?0w1‘.30 040270.“ “firvfi..000a30.'0.0.0 "0.0 ..I“!. h .0". 00.-. I . . . . . . . . . .I .0 O a I. 0 P . ‘0‘. .0 ‘0. . . 0i. .0 . . . 0. 0. ‘ ..90 Q. . 0 ..., .. . I. v . . . . . . . . . 00...": 00.0.3.» I c 0.0.... . .0“. 9300, 0d” "-0.000‘...u.30“.0..~..o i. . x . . 0 . 0 0v... ..9. ...Or...0.0.30u.. 0. 2. 32.0.0 1... 0 0 . .r.» . .. . . . . . . (0‘.v.rfl lu. 402...... 0‘...01..50..‘. 0 0.00.00.20.39... J _ . . . ..0. 1.02:. I. . 0... 0.30.2.0 3 . . . . . . . , . . . ‘Vtua “0‘ . , _ _ . , V ...0 I. .0 1......0‘0... .. . . . . . . . . . . . . . . 0‘0. Joiniufduf‘ .. .. . . . .. _ . . . ... Oufi ....fo 9‘0?! .......031(. .30 002.930.... 0. ....u 300.. ..0. . ..i,‘r’ho 300'. "00.0000... ._ _. 5a . ”a”... W'. . 30... YVJfifiaer; ...."3v. 0‘... 0. .903‘53 005 00. 0.00. o. v 0.2 . .0.”M00..p.l.p.!a._“.o:?!.a. 3.0.” 0.0“"037” 0. 3.. ’3‘ 0.0.0 . ..14.. 00 [0.301.\00..I. .‘(\”u." q. ..."! . .000; 0.0.9.... .... o .0000...- . I 031:0- '. 0 ‘0... u. ..r 0... 0; 0.0. . ......0. .. 7... 00'3qu LN“ 00. "0 ‘Ju .0. “w"uww- ‘I—l..0.0a ..n“”...¢000000or-07. 0.. h"0uu0'-.i30~0.90.0.0.7.: .a.t.n0a..\ : 00. 0.30.30.3‘: .0. 0 0 0. 000.00. a .0 .00. 0 I . 0 ..00 0‘ 0 00000.00 . . . -0 I 000. 0 0.00.0?! y'a-fifi .... u 0.00 0.”... .304!.U«w\0.0..’. ...ulh. Jfi.»l-&0"0.0?Imhrficu 000 300.30 0.- . tn" . ... 00.4. ..0 .0.‘\ 3.1.0.030 '0‘. . . . .0... 0,. 13:000.. «$5..‘I¢,...3¥...0 ....7......~.0...0.0..0.0?at .,.0">.0Q,.04l .... 0:1...) to... ..lo .03... ..0... :0 5.0.0 .0 .0... . . u...» 0... 0". :00... ...-0,0 0100.00.I0‘..0000. ..0 . 0| 000000.000...”0r..00‘0000 . Tot-tpoulo’ .0!.0.‘o0.'0...0-0000o 00: ’0‘.(3 . . .. . . .. , 0... .0. 0 ..w ..l. ’0“. . \I ...._;.Jt..o.l.t..... .3 . 0.060....9220 .00.}... ...0....:..ao£....- .. f... 32...-.. N: $330.99.... . . . 32...... .....t... 9 1.. . .. . . .. . . . . . . . ...! 9.0.. .hrjr0 . ... gt“. J 5.3.0.0. 000.) o . .— .000o.0 0000! «401.09 ...V 00-. (...-.... 0. 00610.0 ... 0.0.0.3003?! -..-1‘ . . 320.0 (0...... 0.0. l ... .... 111:0,930. . . . .. . . . _. , . I 0 .030. 0, I. 00...”. 37.05.33] ’3......1..0¢ & 10...}...20... .00.“..033 3...: 0.... .... .0...’0....... .... .33... J9 .....V. ..1. t... . -.. ......Qo). .. . .0... ... ...l. ... . .0 F0 00 .Lr l.‘ .0 . . uh ....«ffv .0. . ...... ... 300.0. $3.10.... .00. 3.03.. .... 3......0. o’. ... .... 1 3.0 . 1.00; ...-Ct $905.12!!! . .039. {18.700.00.210 0.0 ...-.70.?- . -..00....0 . 0'. q. 8. 1... . . 4 . . \. I . 0.0a... 0.0. 0!. 301.0.- .... t I a .... 0.. ........No r0» 0"?0 €..000000....t..0'.. 20"... .v . a;?110‘.!0.?000 ....V.000....O....0. . u '20 v.00. . .. ’00.) 00.01.. ..00 ‘0' v .'D .....T ... ...-0.0.3020. .2.. t)... .. z .... . .2 .I . . 501.4.C;0. 0.0000 0.7.0300 ...-.Y0100009300....9.. .0: 00. .00 b ...... 30.0.0.1! 5...... .00 . ”4.0.1.370... .o . . .5. I... .30 ...’..0W.0.0 .... 0“ 0.00.. ..3. ...-0:“: 0-0»... 03.0.1.6..le . .... 0.0 90.. :HVL.H.UJMM.Jfl.~u0.u. ”... Jon: ”.00 00.0.1.1. ..3'!-2..0..3.00.'0\oh.300 I ...... .0. ...) 0.0. , . . . 30.0%,). a.“ v! .O 0 vv .0 d v ' 0 0 a .. .I . 00 O. 0 . Q . l .. I0 . . .... . . .3’Cnfiuv 0.3....u. .9"! a. ......"4 .0 o 0.0 ...... .n..u"l0-. 9:31.. 0. .p .0. . 0.0.. ..?0\..000... 0".mu..i..w 00.1.... 0. i0... .. .0..0.I0..sx.4.(.... ......1000.. 13330.! ‘0 1.1/0.3... 2.3.3.00 o. : ... .... . 0‘00! . . .00. N. . . ... . 2...... . I. 0.....9I: . I . ......t. .Yi". $7.03.. .. ...- ..n1-..3u ..13’00....0..0-0....:3?~ ...... 0 3-00003000- u“ 1. n .. 0‘ , . .. hi «.... 0.: 000.17.010.14? 0.. !0 )0 .11.}. NH“..- . .....0. .0 ... 90 .N ...!..‘Jld 00.55.003.00. .r'.é.0.0 .fipifl.00:.' .0051... 0.30.0 It. .000.- 0 . 0"... 90 000000.-\00 7.030.. 0. 0' 000.. ‘ 00— I ...\.-.00 95. .0 0 ‘4‘- 0 .nd ...Jncwhuoo LOG.“ u, ! , u . . . 1x0 ....) 1\0.‘000... "n10... 0......300. ...“.3? 00“....3 0.0. '03.}; ....0: $00.20.... . 0.0.. > 0.. . a .l . .0! ’3'. v '0. (II 0. ’01.: . . ...-.9... u . 0.0, \.0\..0.0000 «Hal.» . ._ n .uvufiblb. -... 0.. 0 2).. 1.5)...5000fiu9’... ...... “on. u “00% 140.0 “0.” ..9 0.7. .0... 0 .. .0. ...... ..I... 0.0.0.0.. .......v.09\.‘7 .30.?! .3404100020" 6.0. 100.... ...0 V .... . .. 9'! .... . 0.. 0. 90.0. . 9‘0 @3. .0 00 .Qvuf... 10.... 0.0. .. . 'o.0.0|0.0.0..9..0w". 0’ ..0 0 . . .. . u . . . . 0,00 .01.} .Louo.'0o.o..0l10 00 0~ .-.—5.0 0“ 0-0.‘ 0.0 0.00: 0. .30.i3. 40 0 000.0430... . . ‘0. $0.030 .. . 0.10lsnn...0h0......r . x . ...!» 0 . 0.10 1...:1”.......vih 00... 33.0.3.3...0... .... 00 ..0 «0 L 0.‘~.00’ dl‘lll...0. . 0 2 .3. 0".” . . 0 . . . . . . ‘ . . 70...... .....00000; 9. .0. «20.3000... 00.00.03... .. ‘0..l.0.0.. 0.. '50.! 3.1.00.9...00. 32.2.33}...1...}...‘00, {10.10.0003- .0... ... .... 0.30. .0.. o 0. . . 9 ...-.00.. . 0... 700.0 .0- s . D .00.?0...!.:...0.‘.0.l0... ..9. 0.....1? . I. I , 90:10.00 0.. v. ......7..'...... .....O....o.3. . u. .10.}... P Dyan. ...0 003..., 04...: ...- 2?... 0.0-3.0. ... .- 20 90.0.3.0...0. 72.0 . .0000 0. .. . 00 .0! ...... ...l 0.... . 30.0.11 .00. J33303s00 "OL’Ib 0:000. ‘0033000uukY‘Jno.01h..‘ . It... .00..0.00 0.00. .. 00.010‘ . .03" 5‘ .00.... n... V 00. ... 00.00.10000: o. .10....n100004040 .0.J0 .0. 5.2,...1 ... 33.030. 00.. 0.0. u. u 3.0.1.4. ... .0. .02....0 o (0...... ......0......‘0. $000.43.. ... 20-031.! 900983.: . . . . . O n \ . 0 0.0.0: . .00.! 30.16.1703. 33.030.31.00 1 . 00 0.3- ...... 1. 3.0.00.0 . . . . k. “.10 )0 .3).. .00 140'. o ...‘00. o0 . ‘00... 2 337...... .. .....0. 0.0...90u ...0 35.0... .023— 9.0.. 3.0 . ...0 0.0.0.0.!Qi’0fif: 0.0.. 0.! .... ..0... 0. .. V. . I .... ..0. 0.0.0.0.... 0 14 0.‘1.....0._..l8..-0. .... 0......3. 5 .... .... . .‘f. . . 3.1.0.0033; 7... . ...1...0f?0.0 30. 0.0.0.0 . :- ... 0.. ,4 £220..-. 0.00.2527... .4. .. y....l..0..3 01.0.00]... 0 .- 0... ...-1C, ’70,... 00‘..'V..0....I0 0' 3'0. (“00.0 0.0.. ,d-020VVMI. 0 . VOL-v“: 3...... I“... 3... ..I...3.00. .. 3.5..‘3 . .0.) 2.. It... “so... .w0 I... 0.00)....9. 81.. . . . .. ... 0. 51 . .I.4 ....7’9. 53...; 10.0.. .200... .. {00.3.0030 .. .0100‘. 001.000... ....II..I¢...£.0.\‘.I00 'I.... ‘1000 0.0 '00....ooor: . 00 $00.0 Z ’0.- .. ‘ 0 0 . . 7. . A l?.....!!..0..3$. 0.00.... 2 .10 .... 50.00.1000 0. 0 01.01.... 3.. . . . J .0 C .08.... $000.11: \ 0'0 . I . r.- :. 13v 000... .. L30. .00..”4 0...}. £30.00 . ...-0.-...‘310 5.0. 0,003.00 ..., ll- . u. p I? 300.... 00...?! 3.07.. It... 1:1-.Is3503... 0.3.1900!!! $050.3...3. 0.. .00. ,. o . 0 .10... 1...!0u0..1.¥..e ...... 3.9%....” _unu§’fl.:! ....I.. 0.0... 0 . 0.0.304. . ..0 0.30.0.9... 0 30.0.0. 0.70.31.12.10: . 0... ......0 .0. 0. 0.0.0.3170 . ....0...(0 . 0.- 0‘..0 3290.90. .0... .004... ‘0. .0. . ...... .90.. .5. 9:39..- WD. . ...0 .3101...ch0. .— .....al. a... 0‘0-_0403000..I. . ‘30....03 .0.I00..l.0 . . 0..., '0, . ... 1.4.70. I00! .0... ...-8.19.01... .0. . . 9-320.... .40. u h. . ..‘0 ’0 0 0.! ..0 0:0: ...." 0.. {0.0. 030. 0.0.09.3... .10....0..0.0... 0.. ,0... 00.1.3. 0.. 010.....r... .90.:0000... . ”.0 V 0.3.3.0.}... 0.. ~90 3‘ ..Q. 0 .00 o 0040 .0. r.’ 390).. .01}...u0.00. . 4.. .330... .7 . ...... ...; 00.. . ... 0.. “20.: 0.0.3.. 00.. l. 0.. _ 53.00 .0. o I ”.0“; .00’.‘ ...-0’. 0 .0. u .. 00).. .0000. ”...... 0070.001-..) .u‘.’\0 .0.r. '0. .00..‘.‘.’0 . 9105.00 0-,.00301000'0- Hr..00 (0;. .000000\ .00.. 00030.5( 0. I... 20.! . . . . . ...0 00.0. 0.00 . . 0 1.0 0.0! 0000. 0u‘0 .00. 0.00... (l0.0. . , ”,0g‘zh: 0.0..- 10.0.... 0 .0.’0 -190}... , . 01.90.35! ....9'!.000.u...00.. 0. . . 0 l0. .0 :00 2,. . ’0 0.0 0. v 0.\~.~.t0. .10". 0.0. 0.: ...0!‘ 30.00.50 I. .. .3...IJ.JIHZ.2I.I.T ......vz. 3.21.3...1. ...-:0... . 0......018001‘1 ! . m» «x... i . .u . n P .olrnl. .052 ..v..lu.’.0d'...0«0.w““.0 I I O C 4 L D 0 , 0.000.301". 0 ‘0’ 3.00 0 ‘3 n 0 0043.‘. I, 0.00. .10. .00 . . .0. 0.“.08'. p0. 0.7.0.11. "at...u00u.0..k.00u~.luu . . li...0...\0...?_r0flu..0:l(..«310...$..0.000b. 2'0 10.... I... 0:... v. ...}.tocy . ‘00.. 00“. 10110.. ...-0.52.... .- . u 3 .0 .... . 1 $0 0...! .. 0.3010 ...‘9’030..0J.0..0..’ .C- . 390.0. . ...5'100 0‘00). 00.0.0.0 .’¢. .0100. :2 .r. 0. 0.0. I - .31... 0.... 40!... .030 0.2.400. .0..'l0.‘0 '0...0:.1 3.0.0.1932... . .. 0.10.0.3... . ...0 .02... c ‘300‘ 0' 0.03.0. . £2.05 . 0...... 1111... :. ”00000.1 .. 03..-...0'0! 16.. ...-.0. .00 . ...-.0. 7:30.390 . 0.0.00. .1 ,n ...... 00001? .J. I. ....0 ...}: .0-0...'-..r it... ...I. .000? )0 1.09. P..I0.... . 002.00 0:0... . 4:000? 80.:‘1!¥300....UY0.38 ....gr....€.....0o.. 30%;}. . 000.0. . :- . 30.. . a 0. ':.000.0.0z‘0|0050 0 A .0001 .0 0.00 .0000?! 0.000;... 0000......00000. 0.‘»J" 00.0000. 1.0.".0‘: I . .. ...-......000... P . 0.. O .0. ‘80... ......33 90.0.00?! 0.......0..0..0..I. ..VI000...’0..0‘ . 0..., 30. . ... 0 .1.._1..l§. 7.9.0.30....IL.....8..,0‘0:.0,....Q. 0.. 0'. ..-. T 23.0.0000 ... .094‘!..0.00..3.500 Ir . {3.01:}... 20.2.00...- v'l..3. ....‘,.. ”-200. ..--0.000'00. 0 00.0.0 .‘01‘3..l§.0.... 0, 001090....33’. 0.300 .030000102 ‘0-- .. ’0‘. ..000000-00‘.a...0.: .iriu. 00.0.00 30"? Q....100~l' 0 :3 ....L...L .. X . 0"}00-3 . . ...). .004030. ..-I 430.01‘.‘00..Jo‘! .030. $.00»... 00000.12: .0! ..000 . 0. 0.. 0.09.0.20 0 'r. .Y0 0 0 7.0.0.003! . 0. 0.0.0 0.0 300.0... .19. . I. '0 '30. .000: .90. .040. . D... ' 0’! 0. ...... 0.0.0 9.". . r 13.010.000 ..I..l.0.l¢ 5.0.. 03;: 0|... 0.. {0.0.3.801‘ 0.00.; v .. — 3......03J'... ...... hr’toll1c...’..0:04..n.§ I, 0‘. .- .0 I. ‘. I . I! ,‘v- ,0. I O. . . .00-.-.011). ...)... .‘I' .. 0. 20.0. . ..0. .0. . {.03. .v 0 '02.... . 309‘..‘0.00.000. :00... D 0.00,!'.....x.0'-n 000... .0000 ‘0 ..l V0a...” 0200-0. ‘NGIM.NMW N0. .0 0. .1 0.0. :00. 0.0 . 5.0.00 0‘“..‘... .0 .0 0.00.0. .. 0.40.9030 .30.}...110 1.... . ....0.. . 0:0 00" :00 0 .00! 40.. . . .0 050.00? 03;} . 0.....'....3:90.‘. v.0 .....T‘o-ngaolofa-,..0.0..003 .....0 .0V.,..3.‘ 0.€0.0 .~ 00:00. .0.,.00 0.0 0“ q .. .031. . 0.010 .s . . . 0a .‘0... ..000...70..0\0 . 0.1.10 .....90"! ..0.-HN.0I00.0V0‘. 0.0 .. .‘0.0.l 0‘ ...0....!.. 30.0 0...}... c.5903. ..0.\V’!0 . 0 .51L0’0 .'l.‘.0v. .d.‘ ...-.01.": .wl ‘0.....0..0.O. 10.9. :03 , 0‘ 0... .010)" pl...0.0‘ ...0 .030?! ...-000.10.: . . | . , . 0 . :00 . ...!310: . ...-a0 .0 .P.0... 50.200...- ..v ...QE..|L0Z~.. .0 .0!.0.... 0050.00,... . 0. 7’10. 0’ ’5 .. ... ,0. . . . l.......... 0., ...l... 0.0 . . I ..\ 10.0 ’3:f§...0‘. ...... ’L230“?! VJ“.‘ . If0.’0~.0.5§09.00.1....0 ’0: . -000. .0 «0.000.000. 3 .0 0 0. . 9Q .0v0. .r. . . .,... .I\ .3...9........ .... . 10...... £30 . ... .00". ..0‘00 . . . 9.0.10. ’0. .."~.0 i ...-0.0 :00... 7031.0... ....0010...:........ 315.9... .... 0.2.... c. . .w... 2.3.8.0. 00001.0..1. 000.. £00,. 0000.00.20‘. 3.0. 0.0... . . \.'.0. £0.01 . ...0. 00.3....000 .. 0.0.3. . 2.9.3.201... ...-...}... 1. 0V.‘ ...000...90 0 . . .‘a...00.0..700.0.0 0.... 0.901.113. 0.0.. 0.0.7... 0..., .00 0|..- .40 0.. v0.0.2.'l:.0 0.0... ...,.I av ....-.0 0 . V 0v '0 :0 . 00.0 .. .... ..0L00...0.0010.r.’0...‘ 00,0 :9. 'l . 3 0...... 0 ..Po..0...10.1 ! o) . 0“. LI0.0.#0N0.JI .rx.‘.0.0 ..-...i... . .- .. fl' "060.30, ‘5‘. 0 ’0 .00. 'I.V0l000.0‘. 0'. !\‘0 \. . I1. 04"21’00. 0 . . . 0‘. . #3.. 000 01000010.... t 0.. P. 3.0.. 0300.!r..g\: '-'. V~I’.0.... ..l......r 0 0000.01.13 ..I. 00”.- . .0 . .300" .‘0'. 00-0,. . V‘ .n I! .... ..s. ’. .. . ...... 0.0‘0...000-...00.V.. . 3‘.» . 1 .. .090. 0.0 ....v 0'. 0 ..Qio‘no‘siob.‘ I0 7...-0 ... . 7.10.. . ‘ 10.00 3|? 0.. .3 . . . . ....afj..¢d<.~.l.n.w3 . . .. .. . . . .. . . . . .. .....T ...0... ...lool)’: 0... . . . .. . . , . . . . n _ n . . ...- 0 . . . . . .. .. . 10..-: .0..0;.!I..30-;00.". ...}: ... 9- ) 110.020...- .90 00 ... .'V.1 9 0.4.... .. ... . ...-.... . . ......)0 .s9109X00p .0; .309... . 10.....0.‘ ... 80.007100. ,0. .....‘l. v00: . 00. 00. .C l. 0.. 0.'.§.’.. .. 0.0..) r. .0.€!)..|£Oli.0......0 .0 )0...).I.-.. ...P‘Os 30.0.. 9.0.0.)... 0.0. 0.0. 0,”) . . 0o. . I {.9131 '20.... .. .0!. t‘) I... y .. 0...! I.Wt..000‘ ... ..fi. 00.9. . .. . 1‘0... 0 0:09 .0’!‘IQ!¢\3.9' 1.000. : I m 00... 0.0. ‘i‘ 4-... . 'a‘... . LII 0.»... \‘0'0.0.00I.0J\$.0v . 0...”.0 “03 ..).00.0.0.0. ..0-‘....§n..0..3~00.\_.... .0. . . 0. . . lb... .. .*0..0.000 ’..00.l\'§u . ‘o..0<‘.000000..0. .9000 0.0....01 0.0.70.0... ”I . I. 0. 0o. .... ...0...l. ... I. )0 s...0.?. 0‘. -..-.... ! 0000!.0.I040\0.0.0" \ ... 0...00.00.I.’0.0.00.0l0.\‘ _ ax . . ..00 ... 19.1.0.9... 01.0\! ...... 00.... ..w . . . . ....0 «3.055.. s $9300... a 0 0.. 0.0.0.0 .0....§. . f! . . .0.. . . w 0. .0.\.. ..300“§ 0.0. .v . ’ P’... f. 0,1 200....00‘0 '0 0;\0 00!..0 I 0.00h0.‘ ,. . 019......- .I\.4.. ..olv.i...0., ...-0 0....'0.0.i 0.... 3.0.20.0].0- l’.....0...! 0 . .00.. 'a.....O.-.3..00 .0...0..0:0.0,V0)’0000l.0...lr. -.VQI.)00.\-:1 . ... 0. 00.10 ...... P .. . m; 2,3... 0.31 . ... \0...‘0.. v.3.lv“.0_0 v .....‘I... “Joli..a.l'o .0" . .00000 ..0... )0...0..1_.o.0. .0. \. .v .0 0'. .. 0'. I... ,I.l\i..!.‘a......10..0l0 . . .. . . . . . . . o .97.... ..0...70..0..\ :0. .. . . . . ,. . . . -00.... $.00 0.0.w . .04.‘.0 _ 0'00...¢.0.00 . ‘0 0.0.0 4 . . . .. . . 1 .0 . ...0. . . .....10.....0..0... 0:... . V . . 0 I.‘ .g.’.v!..\. V1000- . v . . . . . . . .. . .. . . . 0.00 0.0.. .0...0 '1. . . .. .. . ,‘0!!! .0‘00F’0. ..0 of. 22.3.... _.0...p0. .0 000... 0.0 0.... \.. 0.0.00. .0...r ... >0 .....- .00.- .fi... 0000.. 0......(nu......v.o.0 1010.... 0.0. 0 ,0 .0. . 711.0. .. ..031 0 . 0.... 00.0.‘30. . . . \. . . 0o. 0.... .90.?0 . .. o. 0-"... . 'g)..'fv ‘13:...‘07‘ . .0\P. 330.5. ’09. .3..!)\ 00.. I‘m-7:) .. 00).... . .. 0,0.10..0,..?;:0...'... 0.. . ‘ 0‘0...1’0..l04.!. .. . 9‘0 . ...z.’ ..03 330.0. .... \n . A‘.00 .0. 00 .... 00. a .00 ’..... 010.0. .113... 10. .00. .1..9o00...i...-01.000.003.0I...‘ 0.0V ..0 I. .. .- 0..0‘.0. .r: 1.. a 0‘ . .0. l0 . . .. 0 .. ... 0.. ..0 01.0.0.0... . 0 .7..v,..100....1. I.\0... O....'IID.$...¢. 2' ". .J'..a.1\ 03?.-03...0\. J. 0 .0. 0.30 . . .. .... 00.... v.0..." ...-r .‘Ia..‘00000..0. . 0 ‘ . 0... 0.01.3.9! . .01....) 40.....0; r100...!4\...) . .0_. 9. ... 0.. . 02004.50. ..0 0. .0 1N0 ’3’..\0.. \. ~ . . .. .. . ... . . . $0. 0 I \.l . v ..0. . 0-50. . .00 ...I\.0\0 II... .’0§00\0..... 0. I . y .0 . a. . ...0 .. ...-.... ...... 0.900. 3.31.... . ..i...:..v..!. J: . ‘. . 0 0.. 0 0.. ......0a. 0.. s .01. ! . 0‘0". 0.. “0.. 0.7. ...... 0.. .... 0 ..0, 00.0"“ 0 v. 0 . . .. .0.\ .0. s I. \t. .00.. l-v’lf”,\ . 0 0. 1.. 030‘ 6' . v . . . . 1’ . 0. ' 0.0.3.... ...-000.. ...-..0; ... 0.3!..0 . u ....-.0 ....c r.. .. ... .0 . $0I_0\0.l 00.....\..00..0.0.0: ‘0‘0‘001.ul.:0...00 000..-...0...0.00 00 .0000"... ... a...r .00.!- , ...00....~0 . _ ...I. . . 0.. 70.0.0 , .0..~0......l€.3.03.0.1..0003.0.1.o....‘1.V... ..300!‘«...0.0....i ... ! . . . ‘ICI. ..D - _ 00 0 0. . I , {If . o . .. I . ..-qu . ... ...0...4.:...IJ.d....1.0.l7|..h.~2!.0.....3.9 000...; .... ...h 0.0-... . I .u .10. 0.!!— ... 0t .00.... 0‘0. 0 ‘0 0|... 4 . 0 I . . . 3.. 1‘! H .10.\’v...‘1 ..0 .104... ...0‘... .... 0.. 0).!!! .. .000... 01.90., .. -6... o 0.. . .10 ’0‘00.0..... .Qfi. ..\0...|r.. 70.. .. --. ..!0.\..30 ‘\ I. V. .\I.00.|‘.0-...0. ..I.0 0’0I. I 0.0900.‘l. -00. I .... ..0 1.0. . .... OI. :0 0-’....0.. 0. 00.7! 00.. ..(..0)0..,05§.. ... )9. . 0. 0 .0 .¥0§\‘\.09- 0.0. .s I! .....‘0. ..-... .01.. . .. .... .000 0 .\0....00.0; 0.f..0 9v.0§'lu..20.tv ..0\\. .00. 0 .. 010.00.09.13! . o .0 0.0.. ....I00..-.9...0. ...0900... . . . 0--....) . . 30....409) ...L0.00.0.. . .. .\. ... - . .... 01.00.". .... 0.0 0 m... .r .. .....OAQ— .I . IV .0 . ... . . . ‘.'..0.‘ 0. .u I . . ......0. . . 0.0 I . ...-15...;0 $000.0.0‘. . 9. 0.. 0.. c. 0.0 0.. . (0.00. . -00.. .. . v. 0.... 0 1.... .. . 0.... . .v. . 00.. 0.3.00.3‘ . ... ....1. 3.0.2... . .. . I 0. ......fiu.....h “32.3....“ .. :0... 000.10. .0». ... ...—... . <0 0.. 0‘ ...... ...!) ...00. . . 00.0 300.0 .00.. 00.00 lfl0L0 . . L'i‘llv . V017. 0. ..0!0.0L.... ...l.... . 30:0 Kt..-0.0.. .0..00\.000... . 0 .l0 :1... 00.'..... .0.. 0 . .0wl!0.0 .- u. ..I .20 rJo.0:....:.0. .0. ..‘00.... . . 0.... 00.0.1). 010.000 1.0.... .. ,.0‘o.000..',-\ ..0... I... .35.... . .0 ....otv. "...... u! .-..C O. .V . . . This is to certify that the dissertation entitled FAMINE, MIGRATION, RESE'I'I'LEMENT, AND RECOVERY: CASE STUDIES FROM NORTHWESTERN NIGERIA presented by John Grolle has been accepted towards fulfillment of the requirements for PhD. Geography degree in EMA . ‘ MajoWl July 12, 1995 Date MS U L! an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Michigan State 1 University PLACE ll RETURN BOX to rename thle checkout from your record. TO AVOID FINES return on or betore date due. DATE DUE DATE DUE DATE DUE MAY 02 200' 118 51.6 II II MSU to An Affirmative Action/Emil Opportunity Institulon Walls-9.1 FAMINE, MIGRATION, RESETTLEMENT, AND RECOVERY: CASE STUDIES FROM NORTHWESTERN NIGERIA Volume I BY John Grolle A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geography 1995 influential dePriVation COUEinmm : product i 'v'e migr a“ " on lmpcvt . nu ABSTRACT FAMINE, MIGRATION, RESETTLEMENT, AND RECOVERY: CASE STUDIES FROM NORTHWESTERN NIGERIA BY John Grolle Mass migrations from agricultural villages in the West African Sahel have occurred during famines. One highly influential model of famine impacts holds that intense deprivation forces many farmers inexorably along a response continuum that culminates in the liquidation of all productive assets, including land, and permanent out— migration. This process purportedly results in the impoverishment and displacement of a large proportion of the Sahel’s farming population. The present study tests the validity of the continuum model by posing two basic questions: 1) After a famine ends, do farmers that migrated in distress return to their villages of origin? 2) Should they not return, then what becomes of them? The research that addresses these questions is centered upon three interrelated theoretical concerns: 1) the retreat and advance of human settlement in response to arid and wetter climatic phases; 2) the dynamics of famine, and the b v "’ ‘1! at..'—;>ged.‘y -. This 52:13:: a v w'n Ana Warhol-oi aAeuA .1 Sanei cf i‘x ; zone popula potential for farming households to avoid or recover from famine impacts by migrating; 3) the ratchet effect, allegedly the greatest obstacle to recovery from famine. This study also has major implications for famine early warning and mitigation programs, and for deveIOpment policy and praxis in West Africa. Fieldwork was undertaken in villages in the semiarid Sahel of Nigeria, and in settlements in the sub—humid Sudan zone populated by famine refugees. Interviews with groups and individuals were the primary sources of data. The continuum model is deterministic and seriously flawed. Migration during famine is not the culmination of an inevitable and irreversible slide into penury. Many former famine refugees have returned to their Sahelian villages of origin and reclaimed their farms. Former refugees who have settled permanently in the Sudan zone have achieved impressive levels of prosperity without the assistance of governments or international aid organizations. Most of these households have attained self— sufficiency in staple foods while maintaining access to‘ farmland in the Sahel. Major challenges for contemporary famine early warning systems include: 1) the “triggering” causes of famine are more complex than previously thought, and could easily escape detection by high-tech methods; 2) households often migrate early on during a crisis. Paw-71“ rvn‘ uvh v b C ~f‘!'\' U ‘92:! ran- lUU" 11H 9". w luh‘A V" I - - -54..- .4 Copyright by JOHN RAYMOND GROLLE 1995 C rerugee sup that It ‘ -L DiStrf J91 to< .t O liYa 4s ACKNOWLEDGEMENTS Several hundred people made this study possible. I am immeasurably indebted to Illela—area rural producers for their tolerance, cooperation, and extraordinary kindness. Although invaluable insights were imparted by scores of individuals, Mallam Hamidou Gari, Mallam Mamman Gateri, and Sarkin Noma A deserve special thanks. Alhaji Magaji Barau Dansadau’s thoroughly remarkable knowledge of Dansadau District greatly facilitated fieldwork. To the heads of refugee families in Dansadau District villages, I owe an exceedingly profound debt of gratitude. I appreciate the support of modern and traditional governmental authorities in both study areas. Mallam Bagudu Natilli Kalgo constantly proved himself to be indispensable as a field assistant. His assertion that refugees resettle and recover from famine was the crucial impetus for continuing fieldwork in Dansadau District. On four occasions when Mallam Bagudu was unable to join me in the field, Mallam Adamu Al-Hassan filled in most ably and with great alacrity. At Usman Danfodio University, Sokoto, Dr. Mohammed Iliya was a source of perpetual and invaluable advice, support, and encouragement. My association with Dr. Iliya a1 his 3: ‘ ‘ AH " “a“ . a"a, ' 5‘ Age ..vv we 9' h ' Ffi'yavs‘ ... UA¢-'I-‘ “- mark” 5 .'.. , "‘ ee‘S-aS‘ enthusiasr -‘ “e- «L. \ “A. ‘ A~§He.‘Le_\ ‘ “ V i' J as . a" eeqr‘ual and his colleagues in the geography department greatly enhanced my interpretive capacity and knowledge of the research environment. I also thank Dr. Peter Abdu for his counsel and interest in my work. Dr. Haruna Birniwa of the Department of Nigerian Languages offered important advice on the Sokoto dialect of the Hausa language. I am most grateful to Dr. Joseph Ijere and the University of Maiduguri for critical institutional support. Dr. David Campbell, my advisor at Michigan State University, guided this work to completion. His enthusiastic criticisms at key junctures resulted in major improvements. I thank Dr. John Hunter for his unparalleled enthusiasm for research in rural Africa, and for his instruction in medical geography. Dr. Asefa Mehretu's location theory courses, and more recently, his work on domestic water and fuel supplies, have significantly influenced my perceptions of rural poverty and famine. Dr. Jay Harman’s seminar on research design in geography was superb. I also have benefited immensely from his review of my research proposals and from his instruction in climatology. Title VI Foreign Language and Area Studies fellowships, Graduate Office fellowships, and research and teaching assistantships made my work at Michigan State possible. The fieldwork was funded by a Fulbright-Hays fellowship and by awards from Michigan State’s African Studies Center, vi (r (I) (I) (\ c‘hay-Pr; 4'79 vuv‘L a- -n“: .— ~1 - g :fl-tAy-q -| uvl-~v‘ -51; (I) ‘ ... ...V- ‘.‘ C ..a ‘nn. ’ " . i..a- .‘ -C‘ V“ 4 ‘Q. ‘t 4 . A‘.‘~ “1“ fig CYQASA ‘- h "‘ u .-..a .3: ‘ A data“; ~D‘ ha ‘ §.. r: ‘u n beaSeT Graduate School, and office of International Studies and Programs. I am very grateful to Mr. Keith Shaw and Dr. Barbara Shortridge for preparing the maps and other figures. For editorial support I thank Tina Blue and Dr. Curtis Sorenson. I have been enriched by my Master’s advisor, Dr. David Helgren, and his tutorials on Quaternary Studies and other aspects of physical geography. Dr. Byron Emery, my first mentor in geography at the University of Toledo, suggested that I join the Peace Corps. I will be forever grateful for his advice. The people who taught me Hausa deserve a very special acknowledgement: Mallam Solomon Kama, Malama Zoera Isacca, and Dr. Frank Wright at Michigan State University; and in Niamey, Mallam Abdou Loushe, Mallam Yazi Dogo, Dr. John Hutchison, Dr. Jennifer Yanco, Dr. Hassana Alidou, and Dr. Ousseina Alidou. Throughout this and other endeavors Ousseina has provided trenchant criticism tempered by almost ceaseless encouragement. She is a source of boundless inspiration and joy. Ousseina, na gode Allah sosai da sosai saboda shi ya sa muka hadu! Allah ya ba mu lafiya da alheri a cikin duka aiki da fata da muke. Amin. vii TABLE OF CONTENTS VOLUME I LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES LIST OF HAUSA LANGUAGE AND OTHER TERMS CHAPTER ONE: INTRODUCTION Climate Impacts and Return Migration Famine . . The Ratchet Effect . The Dissertation in Outline Summary CHAPTER TWO: FIELD RESEARCH: SITES AND METHODS Hausaland Sahel Study Area Natural Features Human Features . The Sahelian Study Villages Amarawa . . . Tudun Gudali Mai Ruwa . . Gidan Alkasim and Affiliated Hamlets Sudan Study Area Natural Features Human Features Research Methods Methodology Options Research in the Sahel Research in the Sudan Data Limitations Summary viii xii xv xix XX 11 13 16 18 21 24 31 33 33 37 41 43 46 49 51 53 57 62 63 65 CHAPTER THREE: PERSPECTIVES ON FAMINE AND MIGRATION . . . . . . . . . . . . . . . . . . . . . 67 Definitions . . . . . . . . . . . . . . . . . . . 69 Causes . . . . . . . . . . . . . . . . . . . . 70 Vulnerability . . . . . . . . 73 The Progression of Famine through Time . . . . . . 80 The Spread of Famine across Space . . . . . 90 The Spread of Disease and Death across Space . . . 96 Summary . . . . . . . . . . . . . . . . . . . . 103 CHAPTER FOUR: CHRONOLOGY OF FAMINES AND CLIMATIC AND ENVIRONMENTAL VARIATION . . . . . . . . . . . 106 Early Settlement . . . . . . . . . . . . . . . . 107 The Period 1899—1910 . . . . . . . . . . . . . . 112 The 1911-20 Period . . . . . . . . . . . . . . . 115 The 1920-29 Period . . . . . . . . . . . . . . . 119 The 1930-40 Period . . . . . . . . . . . . . . . 125 The 1941-51 Period . . . . . . . . . . . . . . . 134 The 1952—59 Period . . . . . . . . . . . . . . . 144 The 1960—67 Period . . . . . . . . . . . . . . . 153 The 1968-80 Period . . . . . . . . . . . . . . . 158 The 1980-85 Period . . . . . . . . . . . . . . . 163 Summary . . . . . . . . . . . . . . . . . . . . 164 Notes for Chapter Four . . . . . . . . . . . . . 169 CHAPTER FIVE: FOOD PRODUCTION SYSTEMS, ASSETS, FOOD SELF-SUFFICIENCY, AND OFF-FARM INCOMES IN THE SAHEL STUDY AREA . . . . . . . . . . . . . . . . . 171 Food Production Systems . . . . . . . . . . . . 172 Upland Farming . . . . . . . . . . . . . . . . . 173 Fadama Farming . . . . . . . . . . . 176 The Kalmalo Irrigation Scheme . . . . . . 179 Crop By- Products and Natural Vegetation . . . . 184 Livestock Rearing . . . . . . . . . . . . . . . 186 Labor . . . . . . . . . . . . . . . . . . . . . 190 Pests . . . . . . . . . . . . . . 193 Farmer- Herder Conflicts . . . . . . . . . . . . 195 Socioeconomic Class and Assets . . . . . . . . . 197 Climatic Variation, Yields, and Self- Sufficiency . . . . . . . . . . . . . . . . . . 209 Seasonality . . . . . . . . . . . . . . . . . . 220 Locally Generated Non-Farm Incomes . . . . . . . 222 ix fl K .— mu! “'7‘“ youfi: . -983. - l I K m e 3. e i . 1 mt V... m L 8 an .‘t' m V: 9&1 | V545}. “KHLDII Extra-Local Incomes: Long—Distance Dry Season Migration, Long—Distance Trade, and . “Commuting” . . . . . . . . . . . . . . . . . . 227 Summary . . . . . . . . . . . . . . . . . . . . 231 Notes for Chapter Five . . . . . . . . . . . . . 234 CHAPTER SIX: FAMINE-RELATED MIGRATION AND THE 1983-85 FAMINE IN THE SAHEL STUDY AREA . . . . . . 237 Makaji (1904) . . . . . . . . . . . . . . . . . 240 Saketariya (1913-14) . . . . . . . . . . . . . . 240 Shekara Kyamro (1931—32) . . . . . . . . . . . . 245 Takesharbo/Mazarkwela (1942-43) . . . . . . . . 247 Muda (1953—54) . . . . . . . . . . . . . . . . . 250 Mai Zobe . . . . . . . . . . . . . . 263 Muna Sane/Shekara Karmami .(1972-74) . . . . . . 270 The 1983-85 Famine Buhariyya . . . . . . . . . . 282 Migration During Buhariyya, 1983-85 . . . . . . 294 The Fulani of Tudun Gudali Mai Ruwa . . . . . . 305 Lakoda and Other Study Area Settlements . . . . 306 Summary and Discussion . . . . . . . . . . . . . 311 VOLUME II CHAPTER SEVEN: RECOVERY FROM FAMINE IN DANSADAU DISTRICT . . . . . . . . . . . . . . . . . . . . . 320 Farming Systems, Crop Surpluses, and Off-Farm Incomes . . . . . . . . . . . . . . . . . . 321 Farming Systems . . . . . . . . . . . . . . . . 322 Crop Surpluses . . . . . . . . . . . . . . . . . 325 Off-Farm Incomes . . . . . . . . . . . . . . . . 325 Settlement History, Famines, and Historical Links to the Sahel from Pre-Colonial Times to C. 1950 . . . . . . . . . . . . . . . . . . . . 326 Settlement History, Famines, and Influx from the North, 1953-90 . . . . . . . . . . . . . . . 330 Refugees in Seven Dansadau Communities . . . . . 335 Dan Gurgu . . . . . . . . . . . . . . . . . . . 336 Mai Goge . . . . . . . . . . . . . . . . . . . . 344 Babban Doka . . . . . . . . . . . . . . . . . . 346 ’Yan Sawaiyu . . . . . . . . . . . . . . . . 350 Mai Tukuniya (Gobirawa) . . . . . . . . 354 A Neighborhood in Dansadau Village . . . . . . . 361 Nasarawa (Rugan Fulani) . . . . . . . . . . . . 364 Maganawa: Refugees and Pioneers . Migrants’ Status in Villages of Origin Response to Drought and Famine Subsistence in Maganawa Before the First Harvest Recovery Summary and Discussion CHAPTER EIGHT: SUMMARY, CONCLUSION, AND IMPLICATIONS Implications for Famine Early Warning (FEWS) and Rural Development APPENDICES LIST OF REFERENCES xi 367 373 375 378 379 383 390 393 402 472 at“... of (AJ 5.1 5.2 5.3 5.4 f. ‘ .— V \ U .. \ r- 4.3 u: x.) l [1 (I) LIST OF TABLES Characteristics of Strategies for Coping with Food Deficit Famine Chronology for the Sahel Study Area Colonial-Era Famine Chronologies for Northern Nigeria and Niger Population Change in Eight Districts of Northern Sokoto Province, 1952-1954 Farm Holdings by Socioeconomic Class, Amarawa Village, June 1988-March 1989 Farm Holdings of Ten Equivocally Stratified Household Heads, Amarawa Village, June 1988- March 1989 . . . . . . . . . Livestock Holdings of a Sub—Sample of 50 Household Heads by Socioeconomic Group, Amarawa Village, March-May, 1989 490 Upland Farms, Amarawa Village, Means of Acquisition by Socioeconomic Class Gidan Alkasim Hamlets, 55 Tudu, Debagi, and Rafi Farms: Means of Acquisition . . Tudu, Debagi, and Rafi Farms, Tudun Gudali Mai Ruwa, Means of Acquisition by Group, September 1989-March 1990 . . Farm Holdings by Group, Tudun Gudali Mai Ruwa, October 1989-May 1990 . . . . . . . . . Farm Holdings, Gidan Alkasim Hamlets, September 1989-March 1990 Reported Grain Yields, Amarawa Reported Grain Yields, Tudun Gudali Mai Ruwa xii 85 108 109 152 200 201 202 203 205 206 207 208 211 212 tea-db... 11 Repcr: " r1 1 514 tase : we Y L “‘V. ‘R j r- 1" 5.1: 5812': VA a 1787-: 514 a i: . " CEL‘-\ 91,, Kuwa, 515 Self-~ Hamle‘ 6-1 FaM‘v‘ in-.. «an I\) 31 l) f 6.3 64 Alta: Shake 6.5 Alta 6.6 6,7 x, 6.8 6.9 TABLE Reported Grain Yields, Gidan Alkasim Hamlets Case Studies of Self-Sufficiency in Grains, Niger and Northern Nigeria Self-Sufficiency in Grains, Amarawa Village, 1987-89 Self-Sufficiency in Grains, Tudun Gudali Mai Ruwa, 1987-90 Self-Sufficiency in Grains, Gidan Alkasim Hamlets, 1987-90 Famine Chronology for the Sahel Study Area Alternative Foods Reported Consumed During Makaji, c. 1904 Alternative Foods Reported Consumed During Saketariya, 1913-14 Alternative Foods Reported Consumed During Shekara Kyamro, 1931-32 Alternative Foods Reported Consumed During Takesharbo, 1942-43 Alternative Foods Reported Consumed During Muda, 1953-54 Alternative Foods Reported Consumed During Mai Zobe, 1965-67 Alternative Foods Reported Consumed During Buhariyya, 1983-85 Response to Food Shortage by Socioeconomic Class, Amarawa Village, 1984-85 Response to Food Shortage, Tudun Gudali Mai Ruwa, 1984-85 Response to Food Shortage, Gidan Alkasim Hamlets, 1984-85 Ratios, Reported Permanent Out-Migration: Out-Migration/Subsequent Return Reported Domestic Situation of Migrants During Famines xiii 212 215 217 217 218 238 241 242 246 251 252 265 287 289 290 291 312 313 Immigration to Dansadau District and Vicinity Summary of Available Data on Refugees Maganawa Informants’ Farmland Holdings in Villages of Origin Prior to 1983-85 Maganawa Informants‘ Income Sources c. 5 Years Before Departure Response to Food Shortage of Maganawa Informants Who Migrated from the North 1973-74 Response to 1983-85 Food Shortage, “Northern” Migrants to Maganawa Village . . . . Response to 1983-85 Hardships, “Southern" Migrants to Maganawa . . . . . . Sources of Income, Maganawa Village, 1989-90 Reported Agronomic Problems in Maganawa Ratios of Permanent Out-Migration: Reported Return Migration xiv 333 337 374 374 376 376 377 381 381 388 FIGURE LIST OF FIGURES Nigeria, Hausaland, and the Two Study Areas The Sahel Study Area Rainfall and Temperature in the Sahel Study Area . . . . . . . . . . . . Tudun Gudali Mai Ruwa Gidan Alkasim and Affiliated Hamlets Sudan Study Area Gusau Mean Monthly Rainfall, 1942-1984 Temporal Sequence and Organization of Responses to Food Shortage General Location Map of Sokoto Province/State Grain Prices in Sokoto City, 1950-54 Grain Prices in Gusau, 1950-54 1953 Daily Rainfall Series for Birnin Konni, Niger Birnin Konni Monthly Rainfall, 1965-67 and 1933-91 Monthly Means . . . . . . . . Birnin Konni Monthly Rainfall, 1971-73, and 1933-91 Monthly Means . . . . . . . . Birnin Konni Monthly Rainfall, 1983-85 and 1933-91 Monthly Means . . . . . . . . Birnin Konni Monthly Rainfall, 1987-89 and 1933-91 Monthly Means . . . . . . . . Famine-Related Immigration to Amarawa Village, 1913-1954 XV 19 22 28 38 42 44 48 88 110 145 146 150 156 157 162 210 243 Pu. . rhlv rhe FIGURE 6.2 Out-Migration from Amarawa Village and Subsequent Return, Shekara Kyamro (1931-32) . . 248 6.3 Reported Permanent Out—Migration during Famines from Gidan Alkasim Hamlets . . . . . . . 253 6.4 Reported Permanent Out-Migration from Amarawa during Muda (1953-54) . . . . . . . . . . . . . 258 6.5 Out-Migration from Amarawa Village and Subsequent Return, Muda (1953-54) . . . . . . . 260 6.6 Reported Permanent Out-Migration from Amarawa Village during Mai Zobe (1965-67) . . . . . . . 266 6.7 Out-Migration from Amarawa Village and Subsequent Return, Mai Zobe (1965-67) and Muna Sane (1972-74) . . . . . . . . . . . . . . . . . 268 6.8 Reported Permanent Out-Migration from Tudun Gudali Mai Ruwa (TGMR) during Famines . . . . . 271 6.9 Famine-Related Immigration to Amarawa, 1972-85 . 274 6.10 Reported Permanent Out-Migration from Amarawa Village during Muna Sane (1972-74) . . . . . . . 277 6.11 Out-Migration from Tudun Gudali Mai Ruwa (TGMR) during Famines and Subsequent Return . . 280 6.12 Self-Sufficiency in Grains, Amarawa, 1984-85 (Wealthy and Middle Class) . . . . . . . . . . . 283 6.13 Self-Sufficiency in Grains, Amarawa, 1984-85 (Peasants and Extremely Poor) . . . . . . . . . 284 6.14 Self-Sufficiency in Grains, Tudun Gudali Mai Ruwa, 1984-85 . . . . . . . . . . . . . . . . . 285 6.15 Reported Permanent Out-Migration from Amarawa Village during Buhariyya (1983-85) . . . . . . . 296 6.16 Out-Migration from Amarawa Village and Subsequent Return, Buhariyya (1983-85) . . . . . 299 6.17 Total Reported Family Out-Migration from Lakoda Village during Buhariyya (1983-85) . . . 308 6.18 Total Reported Family Return Migration to Lakoda Village after Buhariyya (1983-85) . . . . 309 xvi A3.1 A3.2 A3.3 A3.4 A3.5 A3.6 A3.7 A4.1 A4.2 A4.3 A4.4 A4.5 Migration to Dan Gurgu Village, 1953-88 Migration to Mai Goge Village, 1953-88 Migration to Babban Doka Village, 1973-89 Nigerien Refugees' Migration to ’Yan Sawaiyu Village, 1984-86 Migration of the Dakoro Refugees, 1973-74 Migration to Maganawa Village, 1984-89 Summary of Available Data on Refugees’ Time of Departure during Famines of 1953-54, 1972-74, 1983-85, and 1987 . . . . . . . . . . Summary of Available Data on Attainment of Food Self-Sufficiency by Refugees from Famines of 1953-54, 1972-74, 1983-85, and 1987 Birnin Konni Annual Rainfall, 1933-91 Birnin Konni May Rainfall, 1933-91 Birnin Konni June Rainfall, 1933-91 Birnin Konni July Rainfall, 1933-91 Birnin Konni August Rainfall, 1933-91 Birnin Konni September Rainfall, 1933-91 Birnin Konni October Rainfall, 1933-91 Levels of Self—Sufficiency in Grains, Amarawa (Wealthy) Levels of Self—Sufficiency in Grains, Amarawa (Middle Class) Levels of Self-Sufficiency in Grains, Amarawa (Peasants) Levels of Self-Sufficiency in Grains, Amarawa (Peasants Cont.) Levels of Self-Sufficiency in Grains, Amarawa (Extremely Poor) xvii 341 343 347 352 355 368 385 387 410 411 412 413 414 415 416 417 418 419 420 421 FIGURE A4.6 A4.7 A4.8 A4.9 A11.1 A11.2 A11.3 A11.4 A11.5 A11.6 A11.7 Levels of Self-Sufficiency in Grains, Tudun Gudali Mai Ruwa (Wealthier) Levels of Self-Sufficiency in Grains, Tudun Gudali Mai Ruwa (Other Hausa) Levels of Self-Sufficiency in Grains, Tudun Gudali Mai Ruwa (Buzu) Levels of Self-Sufficiency in Grains, Gidan Alkasim Hamlets Immigration to Amarawa Village during Non- Famine Periods, 19108-19408 Reported Permanent Out-Migration from Amarawa Village during Non-Famine Periods, c. 1940- 1988 Out-Migration from Amarawa Village during Non- Famine Periods and Subsequent Return, c. 19408-1988 Immigration to Amarawa Village during Non- Famine Periods, 19508-19808 Reported Permanent Out-Migration during Non- Famine Periods from Tudun Gudali Mai Ruwa, 19608-19808 Out-Migration from Tudun Gudali Mai Ruwa during Non-Famine Periods and Subsequent Return Reported Permanent Out-Migration from Gidan Alkasim Hamlets during Non—Famine Periods (late 19708) xviii 422 423 424 425 452 455 457 459 462 464 469 APPENDIX 10 11 LIST OF APPENDICES Vegetation in Sahel Study Area Hausa Language Terms for Atmospheric and Climatic Phenomena Birnin Konni Annual and Monthly Rainfall, 1933-91 Levels of Self-Sufficiency in Grains, Sahel Study Area Villages Grain Transactions by Socioeconomic Class, Amarawa Village, 1988-89 Data Pertaining to Self-Sufficiency in Grains, Sahel Study Area Villages Local Sources of Off-Farm Incomes, Sahel Study Area Villages Participation in Seasonal Migration, Sahel Study Area Villages Seasonal Migration Destinations, Sahel Study Area Villages Occupations at Seasonal Migration Destinations, Sahel Study Area Villages Long-Term Migration in the Sahel Study Area Not Related to Famine . . . . . . . xix 402 408 410 417 426 427 435 439 444 447 451 . 1. .t .. . I. w” 3.. “u «a 4‘ 5. :3 FL a. nu no a .d 2 .. a: Mu. r. Y. 2 vs .1... *4 Vi 2 z a v .~u Y. Wu u ”in a a a a Any .lu .Iu .1: u u ..H In ‘ . ‘d ..H. 3 33 Do a: no :8 a: DJ :8 5.5 LIST OF HAUSA LANGUAGE AND OTHER TERMS Acca: “hungry rice," a small cereal-like grass Adarawa: Hausa-speaking people from the Adar (Niger) region Arawa: an agricultural people who speak Hausa and Zarma- Songhai and inhabit the Dogon Dutci (Niger) region and extreme northwestern Nigeria Arziki: wealth, assets Au: buying grain, often by the measure AWPP: Accelerated Wheat Production Programme Bammi: palm wine Bargaji: a variety of fadama soil Barna: destruction (often of crops by livestock) Bazara: the hot, dry season (c. March-c. May) Boringo: hard, rocky soil Buka: a hut Burgu: giant bush rat Burtali: a livestock right-of-way Buzaye: plural of Buzu Buzu: formerly “vassals” of Tuareg pastoralists Cin Rani: dry season migration Daji: uncultivated land; any land outside of a settlement Daki: a room Damba: a relatively dark, clayey soil Damo: iguana lizard Para Fats rir l a 3 a 0. § r h. At .lq .1! a «Q Rh 6 F “(VB Pl Damuna: the rainy season (c. June-c. September) Dan Tauri: a member of a traditional martial fraternity Dari: the cold season (c. November—c. February) Debagi: hard rocky soil Dilancin Bisashe: local livestock trade Dogarai: the body guards of traditional rulers Doruwa: the locust bean tree, or its seeds/pods Fadama: low-lying, clayey, moisture-retentive soils often suitable for dry season farming Fako: hard, rocky soil Falle: the practice of borrowing one bundle of grain during the rainy season, to be repaid with two bundles at the harvest Fara: locust or grasshoppers Fatauci: long-distance trade FCFA: Franc Communaute Financiere Africaine; during the period of fieldwork one French franc was the equivalent of 50 FCFA (one U.S. dollar bought approximately 280 FCFA) Fili: an open space Fulani: a pastoral/agro-pastoral people inhabiting the western and central Sahara, Sahel, and Sudan ecological zones Fura: millet porridge Gandu: a family farming unit; a large farm cultivated by this unit Garin Rogo: cassava flour Gayya: communal work party Gero: millet Gida: compound Gidan Zamani: a compound constructed of cement or cement- clay mixture, often having metal sheeting as its roof xxi Gobirawa: Hausa-speaking people from the Gobir region Godiya: thanks Goro: cola nuts, tribute Gwaiba: guava Haraji: the head tax Harawa: cowpea hay Hatsi: millet Hijira: religiously sanctioned migration in search of refuge Hunturu: Harmattan wind (northeast trade wind) Hurmi: livestock grazing reserve IFPRI: International Food Policy Research Institute Intaya: “hungry rice,” an small cereal-like grass Jangali: the livestock tax Jan Gargari: a bright red soil Janjare: a variety of red sorghum Jan Talakawa: extremely poor people Jari: profit Jigawa: sandy upland soil Jimbiri: the pod containing unripe COWpeas Kadanya: shea butter tree or its fruit Kara: sorghum or millet stalks Karkara: a village’s farming area Kayan Miya: sauce ingredients Kobo: one hundred kobo equal one Naira (N) Konnawa: Hausa-speaking people from the Birnin Konni region Kuda Tsando: the tse-tse fly, the vector of trypanosomiasis (a disease that afflicts both humans and livestock) xxii ' {J‘L-EEL’CE I" 1 ee~ .Ru '7 aka 7.. Au 7.. .\d e' .1?“ Kudu: south Kuka: the baobab tree or its leaves Kulle: Islamic wife seclusion Kurmi: southwestern Nigeria, coastal West Africa Kwano: an enamel bowl; a c. 2.5 kilogram measure of grain; metal sheeting Kwarkwada: an alleyway Laka: clay or mud Lissafi: arithmetic, counting, reckoning Lisso: fine sandy upland soil Lungu: an alleyway, recess Magariba: sunset, sunset prayers Mai Gida: a family or compound head, a patron Mai-Mai: the second weeding Mai Shela: the “town crier" Maiwa: late millet Malamin Sarki: a traditional ruler’s scribe and administrative assistant Malka: period from middle to end of rainy season; torrential rainfall Masakaita: “middle class" people Masara: corn Mashekari: pastoralists’ base camp Masu Arziki: wealthy people Masu Dan Hali: wealthy people Murzuna: a kind of caterpillar or grub Naira/Ne the principle unit of Nigerian currency; one U.S. dollar bought N4 in early 1988 and $38 in late 1990 Purdah: Islamic wife seclusion xxiii Rafi: a small stream; an irrigated plot Rai-Rai: sand; sandy upland soil Rani: the dry season (c. November-c. May) Rufewa: a clay granary Ruga: pastoralists’ base camp Rumfa: a shed Sabra: fallow farms Samari: men in their teens or early twenties Sarauta: traditional authority Sarki: a traditional ruler Sarkin Aske: the head of the barbers’ guild Sarkin Fawa: the head of the butchers’ guild Sarkin Noma: the “king of farming," a traditional office Sassabe: preparing land for cultivation Shigifa: a small rectangular building Shigowa: entry, immigration Tabki: a lake Talakawa: peasants, commoners Tarmani: fire ants Taruruwa: driver ants Tsawo: a three-ply rope often used to tie bundles of grain Tuareg: a pastoral people inhabiting the central Sahara, Sahel, and Sudan ecological zones Tudu: hill; uplands; rainy season farmland Tumu: roasted millet from the earliest ripening heads Turbe: the first weeding Tuwo: a “paste” made of millet, sorghum, or corn, and served with a sauce xxiv 7...1 Haw .., £017: Udawa: Buzu agropastoralists/pastoralists Unguwa: a neighborhood, ward Wake: cowpeas Yunwa: hunger, famine ‘Yan Kasa: people indigenous to a particular area ‘Yan Tebur: petty traders ‘Yar Unguwa: a spinach-like snack food Zakkat: the Islamic tithe Zana: a large coarse mat Zarma/Zabermawa: an agricultural people who speak Zarma- Songhai and inhabit western Niger Zaure: the entrance room of a compound Zomo: rabbit XXV CHAPTER ONE: INTRODUCTION Mass migrations from agricultural regions of West Africa’s semiarid lands occurred during the drought-induced famines of the early 19708 and mid-19808. Famine-related migrations by farmers are considered to be of two general types: temporary labor migration, usually practiced by adult males, and distress migration, which often involves entire families. One highly prominent model holds that a relationship exists between the two types, whereby intense deprivation forces households inexorably along a response continuum that begins with temporary movements by selected individuals and culminates in the liquidation of all family assets and permanent out-migration (Watts, 1983). A8 hypothesized in this model, the prospects of households returning to their villages of origin and reestablishing pre-famine levels of productivity are poor. Drought and famine therefore result in the pauperization and permanent redistribution of large segments of the farming populations of West Africa's drylands. Although other models of famine impacts have been advanced, the continuum model has had by far the greatest influence on recent theoretical studies and on the conceptual and operational foundations of famine early warning and mitigation programs. The present study tests the validity of this model by posing two principal questions: AV? w.‘ YD .‘ DJ 2 1) After a famine ends, do farmers that migrated in distress return to their villages of origin? 2) Should they not return, then what becomes of them? The research that addresses these questions is centered upon three interrelated theoretical perspectives. 1) The impacts 9: climatic variations 9Q humankind. Archaeological evidence and written records indicate that over the past several millennia droughts have forced people to evacuate semiarid lands. Abandoned settlements were then reestablished during subsequent phases of climatic amelioration. Evidence that this process continues in contemporary semiarid West Africa would question the validity of the continuum model. 2) The denouement gf famine. The research seeks to document the existence and examine the nature of a famine recovery phase. The revelation that people recover from famine would effectively challenge the denouement posited in the Watts (1983) model — the impoverishment and permanent displacement of a large proportion of West Africa’s dryland farmers. 3) 1_§ rachet effect. Periodic famines purportedly cause the transfer of assets, particularly land, from poorer farmers to the relatively rich. This ratchet effect therefore is viewed as the main obstacle to return migration and recovery from famine. But the possibility exists that 3 farming households are able to dismantle the ratchet by migrating. This study also has major practical implications for famine early warning and mitigation programs. The continuum model, with its portrayal of famine-related distress migration as the culmination of a process of destitution, has had a significant influence on such programs. But perhaps distress migration is undertaken early on in a crisis period in a deliberate attempt to preserve productive assets. Perhaps also some migrating households have secured new and even better assets at their destinations. This finding would be significant for rural development policy and praxis because often the best development initiatives are those designed to facilitate indigenous coping strategies. (Climate Impacts and Return Migration The role of climate in human affairs is extremely cxomplex and often controversial. Bryson (1988) contended tliat many, if not most, of the major cultural developments <>f7 the past 10,000 years were related to climatic variations [arnd their impact on people’s economic base, and to migration. Many other historical studies of climate impacts h“3\re elucidated only very weak links between climate and SeVeral socioeconomic variables (Wigley e; _l., 1985). 1*t4:empts to analyze the present-day impacts of climate on IIEunankind emphasize the building of sophisticated models and manipulations of secondary numerical data. Yet one of the , - ea: 3-1 waf‘f‘ t|~bi4 -. ”V“ U... he; . uh“ .1 has. Bu (M- V‘\ '5? () 4 leading proponents of this approach has acknowledged that all such models have substantial conceptual and methodological inadequacies (Downing, 1988). Despite the recondite character of the links between climate and humankind, it should be possible to determine whether a prominent human response to past phases of climatic amelioration — return migration to the southern margin of the Sahara Desert — has continued in the twentieth century and particularly during the post—colonial period. Throughout the past two million years of earth history the Sahara, according to Huzayyin (1956, in Wood and Knight, 1975), has functioned much like a sponge, squeezing out jpopulations during arid climatic variations, and absorbing them during favorable periods. Mortimore (1988) echoed this <3bservation in remarking that since "earliest times," human <3ccupation of the southern Saharan borderlands has been cflaaracterized by two coexisting tendencies. One has been a pmarvasive southward movement into the sub-humid savannas, tile other a steadfastness in maintaining or reestablishing CCNnmunities farther north. Archaeological research supports truese propositions. The frequency distribution of radiocarbon dates on charcoal, a good proxy indicator of 11E“nan.occupance, reveals that over the past 12,000 years p0p'ulations have advanced into and retreated from the Sahara 111 :response to alternating centuries-long moist and arid Iphuases (Gehy and Jakel, 1974). Extensive and detailed Eixkzhaeological investigations at Dhar Tichitt, south—central 5 Mauritania, have discovered strong evidence for this process during the period c. 1,800 B.C. — c. 400 B.C. (Stemmler, 1980). In the past several centuries shorter cycles of advance and retreat can be inferred from written records and oral histories. From the 16th through the 18th centuries, Tubu pastoralists withdrew during dry intervals to the relatively moist plateaus and valleys of the Tibesti highlands, from which they dispersed once rainfall increased and pastures improved. In eastern Niger, similar migration cycles embracing the Air and Termit massifs, the Fachi oasis, and the desert of Tenere, also apparently took place (Maley, 1973). Lovejoy and Baier (1975) described how cycles of advance and retreat proceeded between the savannas and the Sahara's southern margins. Their major focus was the 'Tuareg, a pastoral people for whom mobility, economic ciiversification, and a rigid system of social stratification unere the keys to coping with climatic variation. During ctroughts, the Tuareg and their Buzu vassal farmers withdrew flnom semiarid lands, taking refuge in subservient aszricultural communities that had been established EPrweviously in the savanna zone. When climate improved, Tl4areg pastoralism, trade, and agriculture expanded again tlCmrard the Sahara, but the savanna communities were ‘irrvariably maintained in anticipation of future droughts. DLliring benign climatic phases, farmers outside the immediate cDr‘bit of Tuareg hegemony also migrated north to establish 6 agricultural settlements. According to Mortimore (1988), the advance and retreat of farmers in response to droughts and wetter phases has yet to be satisfactorily examined. Delehanty (1988), however, determined that northward agricultural colonization in south-central Niger during the mid-twentieth century was attributable in part to favorable climatic conditions. Does northward return migration occur today in semiarid 'West Africa as a consequence of favorable climatic 'variations? Or has there occurred a kind of historical rupture, in that farmers who migrate during periods of (drought and dearth rarely, if ever, return to their villages (of origin once agroclimatic conditions improve? Have the ratchet effect and other recent socioeconomic changes .intervened to alter or eliminate this ancient link between snocieties and climate? Return migration by refugees would hue among the most significant observable impacts of a Exositive climatic variation. Return migration would illdicate that the cyclical process of desert-edge regeneration has continued despite the far-reaching SC>Cioeconomic changes wrought by the colonial intervention. W This study investigates the possibility that migration thITing famine is not necessarily the culmination of a prcDcess of pauperization, but is instead a strategy for The preserving productive assets or for acquiring new ones. :Ef these themes. Famine may be brought about by a chance 9 negative concatenation of poor leadership, social pathology, external political pressure, and environmental perturbations. Positive concatenations may lead to demographic growth, economic development, and territorial expansion. Watts (1983) found the ideas of the Annales school and the conjuncture to be especially appropriate for his exegesis of drought and famine in Northern Nigeria. I contend, though, that the structure — the centuries-long process of capitalist penetration — receives in his analysis more emphasis than the conjuncture. The underscoring of structure accounts for the determinism that Richards (1990) and others perceive in Watts’s research. Hi8 model of the "temporal sequence and organization of responses to food shortage," which culminates in "permanent out-migration" (Watts, 1983:436), epitomizes this determinism. How famine progresses through time is exemplified by the sequences in which people resort to different coping strategies. Corbett (1988) reviewed several empirical studies of coping strategies, including Watts's (1983), and from them deduced a general three-stage model that begins with insurance mechanisms (changes in agronomic practices, sale of small livestock, reduction of food consumption, collection of wild foods, interhousehold transfers and loans, increased petty commodity production, labor tnigration, sale of non-productive assets), proceeds to Siisposal gfi productive assets (sale of livestock, sale of 10 agricultural tools, sale or mortgaging of land, credit from merchants or moneylenders, reduction of food consumption), and ends with destitution and distress migration. Distress migration has been shown to be an important means by which famine spreads across space (Cutler, 1984; Kumar, 1990), and for this and other reasons has previously been characterized in most analyses as an entirely negative consequence of famine. As indicated by Corbett’s model, destitution is for many the inevitable outcome of famine, and distress migration its signal symptom. In contrast, Wood and Knight (1975) hypothesized that distress migration might actually be a "strategic withdrawal," and Jodha (1975) suggested that a final stage in the temporal progression of famine might entail return migration, recovery, replanting, and reconstitution of reserves. Burton 8; al. (1978) considered migration to "new lands" to be an effective strategy for "modifying" famine conditions. This research demonstrates, at least in the context of case studies, that out—migration during famines is not inevitably a step toward pauperization. The data show that out-migration is not always permanent, that migrants maintain access to farmland, and that some migrants have achieved impressive levels of prosperity at their destinations. 11 The Ratchet Effect The ratchet effect is a concept that presupposes the differential impacts of famine on different socioeconomic groups and the irreversibility of Corbett’s stage two and stage three coping strategies. For those who accept the validity of the ratchet effect, the "sell cheap, buy back dear" character of famine/post famine transactions means that households that sell productive assets have little hope of regaining them. According to this concept, famine enhances extant patterns of inequality because it is the rich who buy and the poor who sell. This study contends that migration during famine is a means of escaping the purported ratchet effect. Watts (1983:464) stated that during the early 19708 famine "many thousands in Nigeria were quite literally forced to abandon their homes." But instead of "liquidating their Lilliputian assets," perhaps some of these farmers, through strategic withdrawal, were able to maintain them. Perhaps others, through permanent out-migration, secured new and even better assets in "new lands." Watts’s (1983) data on famine impacts in one Northern Nigerian village appear to demonstrate convincingly the ratchet effect. Of household heads who in 1973-74 resorted to labor migration, borrowed grain or money, sold domestic assets, pawned farmland, and sold farmland, a disproportionate number were from the lowest of three socioeconomic strata. The households that migrated were 12 also from this poorest group. On the other hand, household heads of the wealthiest stratum hired labor, sold grain, lent grain or money, and purchased assets, including land. In attempting to generalize these findings, Watts referenced Jodha’s (1975) research in semiarid Rajasthan, India. Jodha showed for a sample of 15 poor farming households how six drought—induced famines from 1939 to 1964 caused the progressive disposal of land and livestock and forced increased participation in wage laboring. He also showed in monetary terms the magnitude of the sell cheap — buy back dear dilemma; from a year of dearth to a subsequent normal year, prices of commonly sold assets increased 50 to 300 percent. As a consequence, only a small fraction of depleted assets were recovered during this normal year. The most convincing evidence for the ratchet effect are records and interview data from three villages that document drought-caused transfers of land from the poor to the relatively wealthy over a ten-year period. Despite the relevance of this aspect of the Rajasthan study to Watts’s model, other information in that study suggests that out- migration by families may be an important strategy for maintaining productive assets, especially livestock. Such migrations are practiced mostly by small landholders, are of several months duration, and cover distances as great as 500 kilometers. The efficacy of migration as a famine coping strategy depends on opportunities for employment or sharecropping and the condition of pastures in the 13 destination areas. During the post-drought year, Jodha did not collect data on families that had migrated, so the extent to which migration enabled them to maintain their productive assets and recover from famine is not demonstrated. Mortimore (1989:192) argued that in Northern Nigeria the ratchet effect has yet to be convincingly demonstrated "from the longitudinal micro-studies of interpersonal transactions that alone would settle the question whether mechanisms for economic recovery, which do exist, are ever effective enough to balance it in the medium term." These mechanisms for recovery are embedded in what Mortimore calls "resilience," the bases of which are access to farmland, spatial diversification of economic effort, and mobility. This mobility is seasonal migration, or circulation. But perhaps the form of migration depicted as distress migration in the models of the temporal progression of famine might actually be another facet of resilience. The Dissertation in Outline The next chapter begins with a general description of Hausaland, an ethno-linguistic region occupying much of the central Sahel and Sudan ecological zones of West Africa. 'Then the physical and human geography of the two Hausaland lxocalities where I conducted field research — the Sahel En:udy area, at c. 13°45’ N latitude, and the Sudan study area, at c. 11°15’ N latitude — are described in detail. In like remainder of chapter two I discuss data reliability Efi- 14 representativeness, the central problem of social science field research; the approaches other researchers have developed to contend with this problem; and the field methods that I developed and employed. Chapter three is a review of the literature on famine and famine-related migration. The issues considered are: 1) the differences between the Western and African definitions of famine; 2) the ultimate and catalytic causes of famine; 3) spatial variations in vulnerability to famine; 4) the progression of famine through time; 5) famine-related migration as a means by which famine zones expand; 6) famine-related migration as a means by which disease and death spread across space. The topics of return migration and recovery from famine have never been addressed. Consequently the time scale over which these processes might unfold is not known. I therefore had to develop in chapter four a twentieth century chronology of famine and environmental change for northwestern Nigeria and the Sahel study area. This chronology results from a synthesis of: 1) information provided by groups of elders in several villages and hamlets; 2) the available monthly and daily rainfall series; 3) information from archival documents. Chapter four also elucidates spatial variations in the incidence of and vulnerability to famine, and demonstrates that the catalytic causes of famine are more complex than previously thought. 15 Chapter five examines livelihood systems in the Sahel study area. Farming, integrated with livestock rearing, is the preponderant food production system. Most farmers are not able to provision their families from their own production for an entire year. The chronic lack of food self-sufficiency and, as is demonstrated for the years 1987- 89, the considerable interannual variation in levels of food self-sufficiency, require that families earn incomes from non-farm economic activities. These constant forms of coping are the bases for famine coping strategies, including migration. Chapter six answers the two principal research questions in the Sahel study area. The emphases are famine- related permanent out—migration and return migration by refugees. Data on immigration to study area villages during famines are also presented. These three types of migration events are analyzed for each famine in the twentieth century chronology. Migration data for the 1983-85 famine are augmented with data on coping strategies, asset liquidation, and other famine impacts. Supplementary data for famines that occurred prior to the 19808 include reports of recourse to wild or famine foods. Chapter seven answers the two principal research questions in the Sudan study area. After a brief discussion of livelihood systems, the area’s settlement history is examined with an emphasis on the longstanding links between the Sudan and Sahel. Research in communities populated 16 entirely or in part by refugees addressed the issue of recovery from famine. Data on refugees’ return migration from the Sudan study area are presented. The final chapter summarizes the results of field research and concludes that the continuum model is inaccurate and deterministic. I then discuss the significance of this study for famine early warning systems and for rural development policy in West Africa. Summary Fieldwork was focused by the two principal research questions regarding return migration and the status of refugees who remain permanently at their destinations. The discovery that return migration has occurred, and that other famine refugees have successfully rebuilt their lives in new lands, would challenge the validity of the ratchet effect and refute the prevalent View of distress migration as the end-point in a process of destitution. Return migration would also indicate that the cyclical process of desert-edge regeneration continues as a result of short-term climatic ameliorations. That refugees are capable of achieving prosperity without the assistance of governments and international aid organizations has major implications for agricultural development initiatives in West Africa. A better understanding of famine-related migration is also critical for effective famine early warning systems. CHAPTER TWO: FIELD RESEARCH: SITES AND METHODS This chapter discusses Hausaland and the two areas within this large territory of West Africa where I undertook field research. The physical and human geography of the semiarid Sahel and sub-humid Sudan study areas are described in detail. Because the Sahel study area has a long history of famine and population mobility, research there focused primarily on famine impacts and especially out-migration. In the Sudan study area, fieldwork sought answers to the question concerning the status of people who have migrated permanently from Sahelian villages during famines. Virtually no information about the Sudan study area was available from published sources. A discussion of data reliability vs. representativeness, the most difficult methodological problem in social science research, is followed by a review of selected researchers’ approaches for contending with it. The tentativeness of the interview and the resulting data is discussed, as well as the sensitiveness of famine impacts and especially famine-related migration. Frequent informal interviews, formal interviews with both groups and individuals, and interviews with local experts were the main research methods employed in this study. In both study areas research was conducted in several settlements. 17 18 Hausaland The general locus of field research is Hausaland, an ethno-linguistic region encompassing south-central Niger and most of northwestern and north-central Nigeria (Figure 2.1). The home of perhaps 40 million Hausa-speaking people, the landscape is nearly everywhere a gently undulating plain broken only occasionally by inselbergs. According to Hill (1972) rural Hausaland is unique in West Africa, for nowhere else on the sub-continent do such vast tracts of permanently cultivated, manured farmland exist. While its population possesses a remarkable degree of linguistic, religious, and cultural homogeneity, Hausaland is a region of considerable climatic and environmental diversity. Northernmost Hausaland extends beyond 15°N latitude into the southern Sahara; the southern boundary is somewhat arbitrarily put at about 10%°N in a well vegetated sub-humid ecological zone. Excepting areas under orographic influence, the rainfall gradient is among the steepest in the world. Along a 500 kilometer south to north transect, average annual rainfall deteriorates rapidly from approximately 1200mm to less than 250mm, the theoretical lower limit for rainfed agriculture. The inter-annual variability of rainfall along this transect is also remarkable, from 20 percent in the far south to over 50 percent in extreme northern Hausaland. Several systems for delimiting the ecological zones of West Africa have been offered. To simplify, one might state l9 03/999922, ““359 " llela-Umam/fl ‘ / / /////// 49:... Niamey Birnin Konni M SAHELSEWX * \ H ¢6/{///M ///a/ / // / ‘ .ilorin 'arN NIGERIA . . I? .Ibadan .Makurdl /' CAMEROON .1 {1' I‘/ I Study Areas —--— lntemational Boundaries 0 50 1(‘111502m250Kiornetere Figure 2.1: Nigeria, Hausaland, and the Two Study Areas 20 that the southern half of Hausaland is situated in the sub- humid Sudan zone, the northern half in the semiarid Sahel, and the northern and southern margins penetrate, respectively, the Sahelo-Saharan and Soudano-Guinean ecological zones. This ecological diversity, and the substantial climatic variation characterizing the agriculturally marginal north, suggest that Hausaland would be an appropriate region for investigating this study’s two principal research questions: 1) After a famine ends, do farmers that migrated in distress return to their villages of origin? 2) Should they not return, then what becomes of them? Within Hausaland should be found both the origins and the destinations of famine migrants. If return migration does indeed take place, it should be detected in the Sahel; farming families that do not return would most likely be found in the Sudan. My decision to work in Hausaland was also a pragmatic one. Before embarking on field research I had studied the Hausa language formally for three years. I was also fortunate to have visited Hausaland previously on three brief occasions, so I had more background to build upon there than in most other regions of Africa. Practical considerations limited the research to rural areas in Nigerian Hausaland, yet many of the resulting data pertain to Niger. 21 Sahel Study Area According to many maps published during the 19708, the Sahel zone existed in Nigeria only in far northeastern Borno State. Farther west, in Hausaland, the Sahel was depicted as lying entirely within Niger. Most recent assessments indicate that Sahelian conditions extend much farther south into the northern tier of Nigerian states, yiz., Sokoto, Katsina, Kano, and Borno (personal communications, faculty of the Arid Zone Study Center, University of Maiduguri, 1988; see e.g. Olaniran and Sumner, 1989). The Sahel study area comprises several villages in Illela Local Government Area (LGA), Sokoto State (Figure 2.2). As one of the northernmost localities in Nigeria, at c. 13° 45’N latitude it is situated farther north than Niamey and Maradi, and farther north than those regions that collectively contain a majority of Niger’s population. All of the study villages are located within 15 kilometers of the international border. Before 1989 Illela LGA was a part of Gwadabawa LGA; Gwadabawa town, the headquarters of the District Head (Uban Kasa), Sarkin Gobir of Gwadabawa, remains the seat of traditional authority. A tarred road connects Birnin Konni, Illela, Gwadabawa, and Sokoto. The Sahel study area might be considered to be on the periphery of the Sokoto Close Settled Zone (CSZ) or, according to Abdu (1987), within a northern extension of this zone. Population densities are significantly higher than in other regions of northern Sokoto State, but do not 22 I tadama lands * ' ° , I . m... water deficit , ’ o settlement / — tarred road ., 0 l 2 3 4 5 1 1 1 i l g / / l <2? i 0 o 13'50'N $ / I | . g \ Hurmn Lakoda I ~ ‘0 l \ 9 l /€' l n-Ehihsw . e ‘ s // 'Runoamwe.. —“~w’ o ‘ . . Out: 0 w———‘°d' . ‘ / ' ' . O / Atebe 3 W . ’ O . . Denbdu O Tudun Gudali Mei am on; Mm. .. ' in. mm' Gidan Nah-you. .1039 M“. g ........ Ezo'w 5°25'w Figure 2.2: The Sahel Study Area 23 attain the levels observed in the core of the CSZ surrounding Sokoto city. Land shortage has been reported as a major problem since the 19508 (Prothero, 1957), and Udo (1971) included parts of Gwadabawa and Illela LGAs in one of Nigeria’s nine areas of chronic food deficit. The colonial tax burden, and the inability of a large proportion of farmers to produce enough food to feed their families, have contributed to widespread participation in dry season migration (Swindell, 1984). The available publications indicate that at least four famines have occurred in the Illela area since 1950. A mid- 19508 famine was reportedly followed by a localized crisis in 1959. The famine of 1966 was also supposedly local in scope (Abdu, 1975; Watts, 1983). According to Van Apeldoorn (1978), harvests in both 1973 and 1974 were near-total failures. In addition to the region’s Sahelian setting, its rather long history of population mobility, and its relatively recent experiences of famine, there is another reason for selecting for study villages in Illela LGA. In 1975, Peter Abdu conducted research for his undergraduate thesis on "drought-caused migration" in villages near Illela town. He reported that 12 of 13 villages had experienced mass emigration during the early 19708 crisis, and in three villages established that out-migration by entire families had occurred. It therefore seemed that Illela LGA would be an excellent locale for attempting to determine whether 24 refugees from drought and famine return to their villages of origin during subsequent phases of climatic amelioration. Natural Features The relief is typical of the gently rolling plains of Hausaland. The average altitude is about 230m above sea level (ASL). The highest inselbergs are near Lafani (380m ASL), Dango (370m ASL), and Tozai (350m ASL). Numerous smaller inselbergs dot the landscape. Soils. In the uplands, or tudu, soils have formed atop fixed Pleistocene dunes. When traversing these tudu areas on foot, one usually experiences a sensation similar to walking on a sandy beach. Farmers distinguish at least three upland soil types. Sandy rai-rai soils, known elsewhere in Hausaland as jigawa, predominate. Of secondary importance are the sandy lisso soils, whose particles are finer than rai-rai. According to informants, lisso "does not like" abundant rainfall, and under some circumstances yields may actually be higher in droughty years. Debagi and fako are terms generally applied to hard, rocky soils. Although fako has been translated as barren land or wasteland, good yields are frequently realized with favorable rainfall. For some farmers, fako and debagi are synonymous. Others consider debagi to be less rocky, and therefore more productive, than fako. Boringo is another local term for rocky soils. Fako and debagi are far more common in the eastern, higher parts of the study area. In the lowlands, fadama are darker, clayey, moisture retentive 25 soils often suitable for both dry and wet season cultivation. Fadama are generally restricted to lake perimeters and riverine areas, but may also occur in depressions, and in and near ephemeral or dry stream and lake beds. Copious rainfall causes crop failure most readily on bargaji, a lighter-colored variety of fadama soil. Hydrology. Two lakes, Kalmalo and Baile, are the study area’s most important hydrological features. Both lakes are highly significant to the basic maintenance of households and in local economies. Fadama lands, extending from Lake Kalmalo’s eastern and southern shores, cover an area of approximately 15 square kilometers. A species of small, black catfish and one talapia species constitute an impressive fishery in this lake, the depth of which varies seasonally from two to three meters. Although the fadama lands adjacent to the very shallow (< 1 meter) Lake Baile are relatively small, it is a crucial source of domestic water for Dango, one of the largest villages in the study area, and, at least during the dry season, for about a dozen other settlements. Other study area settlements bereft of a perennial water source rely on wells in neighboring communities (Figure 2.2). The water course depicted in the south-central sector of Figure 2.2 was a rafi, or small stream. Water has not flowed in this valley, or occupied the tiny lake bed at its western terminus, since 1983. A swath of fadama lands centered on the streambed is on 26 average about 60 meters wide. The hydro-pedological situation is reportedly similar along the water courses portrayed in the northeastern portion of Figure 2.2. Vegetation. Probably more than 90 percent of the Sahel study area is farmland. The approximate extent of a hurmi, or livestock grazing reserve, is depicted in the north— central sector of Figure 2.2. Cultivation on this parcel, known locally as Hurmin Lakoda, has been prohibited by local authorities since before the colonial intervention. The other major uncultivated parcel, known as Gandun Gwamnanti or Dalia Gwamnanti (government reserve), is situated to the southwest of Lake Kalmalo and is not included in Figure 2.2. Burtali, or livestock right-of-ways, lead from most study area settlements to Hurmin Lakoda and Gandun Gwamnanti. In the uplands, small arable uncultivated zones are sometimes found along the boundaries between the karkara (farming areas) of neighboring villages. Natural vegetation in the study area is typical of the southern Sahel. Copses of trees are confined to Gandun Gwamnanti and stream valleys, where the most important fruit tree is the gwaiba (guava). Elsewhere trees usually stand alone. Shrubs and grasses occur in the margins of karkara, along farm boundaries, in and near irrigation ditches on the Lake Kalmalo fadama, and in the few other uncultivated zones, including those on the outskirts of villages. (Appendix 1 presents species of vegetation in the study 27 area, and their reported and potential uses to the local population.) Fauna. The number of fauna species and their populations are rather low in the study area. Antelope and giraffe have not been sighted since the 19608. Perhaps only the avifauna are sufficiently diverse to hold the interest of wildlife enthusiasts. Buggy (giant bush rat), zgmg (rabbit), and damp (iguana lizard) make only occasional contributions to diets. In the western portion of the study area, apparently only one man seriously pursues hunting and trapping as an income-generating activity. Climate. Figure 2.3 summarizes the study area’s basic climatic parameters. Rainfall data for the years 1933-91 were recorded at the Birnin Konni meteorological station, just across the international border in Niger. The temperature data are from Niamey, which is at nearly the same altitude (200 meters) and latitude (13°29’N) as Illela. (Appendix 2 presents a range of terms for Sahelian atmospheric phenomena, and imparts an appreciation for the ways in which Hausa-speaking people discuss weather and climate.) On average, all but 50mm of the mean annual 530mm of rainfall occurs during gamuga (June-September), the critical season for food production. Several characteristics of Sahelian rainfall would surprise an observer accustomed only to mid-latitude climates. More than 90 percent of rain events are nocturnal. Emanating almost always from the 28 50°Co (122F) 40°03 (104F) 30° - (3691:) I,"'""‘—-e~\ o e’ ‘e—-—-..-..-e-_-.\ 2 _ ’I O ‘\ (goal-1’” ° ° . . ' . '~~-. o . ° 10;: ‘ . ° (50F) , o — mean high temperatures (32°F) --- mean low temperatures x record high temperatures record low temperatures _ ------ ------ - - 150mm Mean Monthly Rainfall, . 1933-1991 T 1 I . . . J F M A M J J A S O N D ' missing observation: May 1933 Figure 2.3: Rainfall and Temperature in the Sahel Study Area Sources: Hulme (1992); Sivakumar (1986) 29 east, a large majority of rain events are preceded by incandescent displays of lightning, thunder, and very strong, even violent winds. The ensuing rainfall is often intense and generally lasts for less than two or three hours. The gentle, widespread, all-day or all-night rainfall so common in temperate regions is rare in the Sahel. Extreme spatial variation in the incidence and amount of rainfall is another distinctive characteristic. Often, one locality receives torrential rainfall while other settlements only a few kilometers away receive little or no precipitation. By the malka phase (August-September) of an adequate rainy season, the Saharan borderlands have been transformed into an astonishingly verdant landscape. The panorama from atop an inselberg inspires awe at the productivity of the erstwhile ergs. Yet with its cloying humidity and heat, damuna is the most uncomfortable season. One sweats almost constantly; even a brief errand on foot in the evening generates rivulets of perspiration; sweating during sleep is normal. Flies are profuse and annoying during the daylight hours (although most people accept this good-naturedly as a tentative sign of an above-average harvest). At night, frogs croak in large puddles, mosquitoes pullulate among the dense crop growth, and fire ants (tarmani) sometimes invade people’s sleeping quarters. October is a month of transition. Rain is not uncommon during the first week or ten days of the month. During some 30 intervals the southwest monsoon i8 perceptible, while at other times a weak hunturu, the northeast wind from the Sahara desert, hints of the upcoming cold and desiccation. Perhaps the concomitant diurnal variations in atmospheric pressure, coupled with high temperatures (October is the second annual period of temperature maxima), contribute to the headaches, earaches, and general ill-humor that often affect many individuals. Raga, the dry season (c. November — c. May), is divided into two nearly equal periods. For most people, dari, the cold season (c. November — c. February), is the most comfortable and enjoyable. Low relative humidities are certainly partly responsible for a surge of energy and a sense of well-being. Haaa, a mist or drizzle, occurs infrequently and is generally the only form of precipitation during the cold season. Insect populations markedly diminish. Blankets are imperative during the chilly nights. The earth progressively dries, natural vegetation withers, and crop stubble peppers the upland farms. The hunturu is sometimes strong enough to produce audible effects at night. As the cold season progresses, this wind becomes increasingly laden with fine Saharan dust, which produces a thin veneer on most surfaces. Dust coats faces and invests the hair, so that younger men, at first sight, may be mistaken for hoary elders. Spreading clothes to dry out-of- doors is consummately counter-productive. By mid-December it is easy to understand how European colonial officers 31 without knowledge or experience of rainy season conditions could have mistaken the Sahel for an ever-encroaching Sahara desert. Bazara (c. March — c. May) is the annual period of temperature maxima. Low relative humidities in March and April, and the near-total absence of insects, make these months bearable. Sleeping under the stars is frequently very enjoyable. But the parched landscape appears most forbidding. Winds in March and in early April are from the east; afterwards, the southwestern monsoonal flow gradually becomes established. Although three rain events are typical in May, bazara frequently continues into June before another rainy season truly begins. flamaa Features Illela is the largest, and in terms of its population and economy, the most diverse settlement in the Sahel study area. A very vibrant Sunday market is the town’s most prominent attribute. The marketshed includes Niamey, Tahoua, Sokoto, and areas to the east in both Niger and Northern Nigeria. Like most major markets in the Sahel, Illela’s offers an extensive range of goods and services. The Illela market has an air about it similar to what one might imagine of a Medieval European carnival, or perhaps expect of a contemporary American county fair. Illela probably encompasses more than one thousand habitations. These range in quality from the large two- story cement block house constructed by one of the 32 wealthiest men in northern Sokoto State, to the sorghum- stalk huts, located just across the road, that serve as shelter for several families of fair-skinned Tuaregs. A large majority of habitations are typical Sahelian compounds made of clay bricks. Most of their male inhabitants would cite farming as their primary occupation. Illela’s Ibo community was constructing a large church during the years 1988-90. Mosques are numerous and vary widely in their size and material composition. A health clinic exists, but patients are usually required to procure medicine from one of the town’s pharmacies. Illela also has a courtroom, a police station, primary schools, a government girls’ secondary school, and the offices of the customs and immigration services. The Sokoto State electric power grid includes Illela, but service is limited to a few hours one or two days per week. The motor park on the south side of town is the hub of the study area’s transportation system. From here one can catch a taxi (usually a Peugeot station wagon, a micro-bus, or a one-quarter ton pick-up truck) bound for Sokoto, Gada (the first large town to the east of the study area), or Birnin Konni, Niger. Motorcycle taxis convey passengers to some of the study area villages not situated on paved roads. Warau and Kadadi are, respectively, the study area’s second and third largest settlements. Both have government schools, a few shops, and small markets. Warau has a health clinic. But in comparison to Illela, these settlements are, 33 in the opinion of most informants, no more than expansive villages. Compact villages of approximately 30 to 300 compounds are the most pervasive form of human settlement in the Sahel study area and throughout northern Hausaland. Several of the larger villages in the area have small ancillary settlements of sedentary or semi-sedentary Fulani or Buzu people. Hamlets comprising fewer than 30 compounds are not quite as numerous as the villages. One or two compounds, standing in relative isolation, are occasionally encountered as well as the base camps (mashekari or gaga) of other Fulani agro-pastoralists. An analysis of the Sahel study area’s settlement history must take into account the violence of French colonial conquest, the differential tax burdens in colonial Niger and Nigeria, flight from French forced labor, the movements of pastoral peoples, and twentieth century famines. Taa Sahelian Sgaay Villages AEEEEEQ A preponderance of Amarawa’s 220 habitations are traditional clay brick compounds. Inside the compounds’ walls are shigifa, typically small rectangular buildings with walls and roofs of clay. Kyaag, or metal sheeting, serves as roofing material for perhaps 10 percent of these structures. Only about one-fourth of the compounds have a aaaga, an entrance room in which the household’s men receive visitors. Five compounds in Amarawa are considered gidan 34 zamani ("modern" compounds) because their walls and rooms are made of cement or a cement-clay mixture. The walls of approximately 20 compounds are made mostly, or entirely, of kaga (sorghum stalks) or aaaa (a large coarse mat woven from the grass gamba). Some shigifa without enclosing walls are inhabited by widowed or divorced women. Several teenage boys who prefer to live some distance from their families have constructed shigifa known as dakin samari. Clay granaries, or rufewa, are the most numerous structures in the village. More than one half of the rufewa are situated outside the walls of compounds. Amarawa is well endowed in domestic water sources. One hundred meters is the greatest distance anyone must negotiate to retrieve water, and a few compounds even have wells inside their walls. Ramfa, which are sheds made of guinea corn stalks and/or grass mats, have several functions. Two of Amarawa’s tailors set up their sewing machines under these structures, and a well built shed shields the one functioning grain— grinding machine from the elements. A large, centrally located rumfa serves as a gathering place for younger men. The village’s rather informal motor park has a rumfa for storing a barrel of gasoline. Here during the early morning and late afternoon hours four or five young men gather to work on their motorcycles. For the sum of one Naira they will usually provide taxi service to Illela. 35 Four motorable roads of compacted soil run through Amarawa. The main road, trending north—south, nearly bisects the village. The three other roads, two of which form the boundaries of Amarawa’s ungawa (wards), have an east-west orientation. Narrow alleyways (kwarkwada and laaga) wend their way among the compounds of the three unguwa, lending a labyrinthine quality to some sections of the village. For purposes of tax collection, censuses, and elections, unguwa are the smallest demographic units in Northern Nigeria. Amarawa’s smallest ward, situated at the northern end of the village, is known as ungawar dogarai (the ward of the village head’s "bodyguards"). At the opposite end is ungawar LEQE ("the neighborhood of the hill"), which is recognized as the village’s newest section. Unguwar tudu and unguwar dogarai each has a ward head, an older man who occasionally is called upon to settle disputes but whose primary responsibility is to collect annual taxes. (Haraji, the head tax, was ostensibly N20 for every male 18 years of age or older; jangali, the livestock tax, was an per head of cattle and 50 kobo for each sheep or goat). Tax collection in the largest, central unguwa was handled by the village head. Situated near the northern and southern margins of Amarawa, £111, or open spaces, are centers of social interaction and village commerce. The most prominent commercial activity is the prepared foods trade. Children 36 normally conduct this trade on behalf of their mothers, who usually remain inside the compounds. Rice and cowpeas, boiled chick-peas, kaaai (fried cakes made of cowpea flour), and sweet sesame seed balls are usually available in the morning. The most common evening fare is gaya (paste) made of red sorghum, served with a baobab leaf- or okra-based sauce. Small fried fish are almost always available, as is freshly grilled goat meat and kilishi, a kind of beef jerkey. Perhaps the most important staple food, faga, a millet and milk porridge, is rarely sold in Amarawa. The availability of most other prepared foods in these micro- markets is strongly seasonal. Carrots are sold in March; corn is roasted to order in July; boiled cassava, taro, and sweet potatoes are available during most of the rainy season; roasted locusts may be purchased after the harvest. Guavas, oranges, mangoes, and sugar cane can also be had in season. Snacks such as roasted peanuts and Lya; ungawa (clumps of boiled, seasoned tafasa) are often sold by young girls who circulate in and near the fili. A few older men sell cola nuts and small measures of grain and cowpeas. Each fili has two or three petty traders (Lyaa gagag) who display miscellaneous items on small tables. Their stock usually includes cigarettes and matches, hard candies, sugar cubes, assorted cookies and crackers, salt, Maggi cubes, bar soap, small quantities of powdered laundry detergent, batteries, and mosquito coils. One young man prepares and sells mugs of coffee or tea from a table. 37 From about seven until eight A.M., and again for about an hour before magariba (sunset) prayers, as many as one hundred men and children gather at these fili. Children play, sing, and sell the prepared food. The men joke, discuss local affairs, their farmwork, or other economic activities. Many of these men, and a few of the children, buy their meals and eat, seated in groups, on the margins of the open spaces. Village commerce is not limited to the two fili. Fruit, fish, cola nuts, and sometimes firewood are sold in front of a few compounds, usually by older men and children. Items not normally offered for sale out-of-doors, such as kerosine, onions, and other condiments, can be obtained by sending a child to purchase them from within one of the compounds. Commerce conducted from within or in front of compounds or at the micro-markets represents only a small proportion of the available income-generating activities. Other sources of income will be discussed in chapter five. mama-snags The settlements known collectively as Tudun Gudali Mai Ruwa are portrayed in Figure 2.4. The name denotes attributes of the settlements’ site. The nuclear village and the tiny hamlet to its northwest have been built atop a hillock (tadu). Eleven or twelve years ago water (gawa) was flowing (gadali) in the small stream (rafi). The streambed and some of the adjacent fadama lands afford opportunities 38 85¢ as. =85 88.: 3.8 28¢ :Nsm Ema“. Sam: I 52:00 mEmnmn. cam omnEmmzm m xomc. 639082 I ......................... .............. ............... .............. Eoop 39 for dry season farming. Water for domestic uses is perennially available. The habitations in the nuclear village are approximately 40 in number. Their material composition is considerably less varied than Amarawa’s habitations. Neither gidan zamani (cement compounds) nor buka (stalk huts) exist. Two widow’s habitations are clay brick shigifa (rooms) without enclosures. All of the other habitations are compounds with shigifa and surrounding walls. Three or four compounds have metal-roofed shigifa; a similar number have rather large sections of their walls patched with mats (zana) or stalks (kara). The numerous granaries are situated inside the walls of compounds. A large open space (£111) is located near the center of this settlement. Two sorghum stalk sheds (rumfa) on the periphery of the fili provide shelter from the hunturu wind or the midday sun during bazara. Under these structures men often gather to converse while manufacturing gaaya, a three- ply rope. An excellent well with a one meter high cement collar is located at the center of the fili. Nearby the only immediately observable commerce in the village is practiced by one or two meat sellers and one or two petty traders with small tables. Other forms of commerce, including the cooked foods trade, are conducted from within the compounds. Meals usually are not eaten in the Open. The main fili is a place for meeting, sitting, and socializing. From here two short alleyways (kwarkwada) lead 40 among the compounds on the north side of the village. At the end of one of these is a very small fili where a few men gather, making use of unhewn logs as benches. A marginally motorable road connects Tudun Gudali Mai Ruwa’s nuclear settlement to Makwala, the closest village to the west, where on Sundays people often can board a small pick-up truck bound for the Illela market. An unusually large high-walled compound is situated to the south of this road. The other Tudun Gudali Mai Ruwa settlements shown in Figure 2.4 are reached most easily by footpaths. The two Hausa hamlets to the northwest and southeast of the main village both comprise three compounds with solid clay brick walls and shigifa. The three Buzu (Bugaje) compounds to the west, and the one compound situated to the southeast of the nuclear settlement, have clay brick shigifa but are enclosed only partially by sorghum stalks (kara). The Fulani settlements are referred to locally as mashekari or gaga. Although both of these terms connote impermanence, the Tudun Gudali Mai Ruwa Fulani all have well constructed clay brick shigifa. The Fulani shigifa in the southeast settlement are surrounded by clay brick walls, as are some in the southwest quadrant of Figure 2.4. The remainder either have walls of sorghum stalks or are without walls. The single Hausa habitation to the south is a clay brick shigifa partially enclosed with stalks. 41 QAQEQ Alkasim aaa Affiliated Hamlets The hamlets depicted in Figure 2.5 may be considered a "village area." Gidan Alkasim’s headman is, at least nominally, the traditional authority for all four hamlets, and has the responsibility of collecting taxes. The six compounds that constitute Gidan Alkasim have clay brick walls enclosing clay brick shigifa and granaries. Three capacious granaries are situated on the outskirts of the hamlet. Most people in the Sahel study area refer to Gidan Alkasim as Zangon Buzaye because its inhabitants are ethnically Buzu (Bugaje). The people who live in the northwestern hamlet, Zangon Tsakuwa ("the encampment on gravelly ground"), also identify themselves as Buzu. This settlement comprises about ten compounds, arranged in two parallel rows, with walls, shigifa, and granaries constructed entirely of clay brick. The 15 compounds of the southwestern hamlet, known simply as Zangon Yamma ("the western encampment"), are arranged similarly and have the same material composition. A majority of Zangon Yamma’s people are ethnically Hausa; the others may consider themselves Buzu. The eastern hamlet is called Katanga ("compound wall"), or, less frequently, Debagi ("rocky soil"). Its walled compounds, approximately seven in number, are solidly built of clay brick. Three of these are quite large and have entrance rooms (aaaga). The people are all Hausa. Sarkin Noma A, Katanga’s elderly yet vigorous patriarch, spends most of his days either working 42 8286: 628:5 2m 888.: Sec 8.0. 8:9“. $221 35 52:00 mEmomm 95 25.5826 .... .... 18381 9.: «95th Eoop O @886 :09.va .19 .../.... Amy 5.0.9.2 520 t . an.“ I y I . sun.” EV magma» coocmN .. . . 43 on his farms or camped out, with a calabash of fiaga, in a sorghum stalk enclosure some 300 meters to the west of the hamlet. All of the hamlets have small, centrally located fili where men rest and socialize. No commercial activity is readily observable, apart from the occasional meat seller. People usually do not eat in the open. Each hamlet has a small clay brick mosque and good wells. Dry season farming is practiced on parcels in and near the stream bed. A dirt track, used by vehicles to convey people to the Illela market on Sundays, passes near Gidan Nashiyau. mmafl The Sudan study area consists of villages in Dansadau District of Akna LGA, Sokoto State (Figure 2.6). Situated at about 11°15’N latitude, the district is among the more sparsely settled areas of Hausaland. In the southwest Dansadau District abuts Niger State, and in the south and southeast, Kaduna State. Something of an inter-ethnic boundary also exists not far to the south of the study area, where the homelands of the Gwari and Kamuka peoples begin. Based on several criteria Dansadau District is located in the northern "Middle Belt" (Gleave and White, 1969). In the context of Northern Nigeria, Dansadau is a remote district. Before completion in December 1990 of the all-weather road linking Dansadau to Gusau, access during the rainy season was usually difficult, occasionally 44 1".I :— . s.. C. . teas .34“ . \ .>-\\ ¢~~.<~ \- ‘W JFK“ \ _ has}; , - x \.\ “‘3.“ ‘ has»: on}; .2 ; .zv \x a). Y . ‘\\g§:}>$ .. . “3.. V?§\V§I\\ .‘it Unguwar Gwandu (abandoned) 17. Mai Tukuniya Dan Goge 25' Kuian Dan 33' M“ 1. Maganawa 10. Karauci 19. Gobirawa 2. Mai Tukuniya (Gobirawa) 11. Kwiambana (abandoned) 20. Baban Doka 3. Han Han 12. Dan Gurgu 21. 'Yan Sawaiyu 4. Unguwar Mai Jan Kai 13. Naduka 22. Nasarawa(ll) 5. Mai 6099 14. Ruwan Toia 6. Nasarawa ('yan kasa) 15. Tunga 23' Randa 7. Nasarawa(Fiugan Fulani) 16. Mai Tukuniya Baban Gona 24' Madada 8 9 Unguwar Barau 18. Dandala 26. Marabou Figure 2.6: Sudan Study Area 45 somewhat hazardous, and at times impossible. Taxis and other vehicles operated in relays between swollen watercourses, where vigorous young men garnered impressive incomes by conveying passengers and goods across on their heads and shoulders. They are doubtless the only people not entirely pleased with the new road and bridges. The choice of Dansadau District as the Sudan study area was influenced to a great extent by my research assistant. Mallam Bagudu has an astonishing, almost visceral knowledge of rural Sokoto State. His intuition suggested the strong possibility that answers to my second research question could be found in Dansadau’s villages. Information from people in the Sahel study area also contributed to my decision. Apart from intuition and reports from a few people in Illela LGA, the selection of Dansadau District was otherwise uninformed. This was especially true of the area’s recent settlement history. Neither Mallam Bagudu nor I had ever visited this part of Sokoto State, and our colleagues at Usman Danfodio University were not aware of any research that had been conducted there. References in the literature to Dansadau District are few. Usman (1978) and Adamu (1978) briefly discuss the precolonial history of Kwiambana, once a large and important city founded before 1300 A.D. but now almost completely abandoned. The Zamfara valley, to the north of Dansadau District, has been mentioned as a 46 destination for many seasonal migrants from northern Sokoto State (Prothero, 1957). Natural Features The average attitude of this gently undulating land is approximately 510 meters above sea level. Inselbergs having 30 to 60 meters of relative relief are the most distinctive physiographic phenomena. figila. Here too a general dichotomy exists between lowland and upland soils. Lowland soils that would be suitable for dry season agriculture are referred to as laka, the Hausa word for clay or mud. A few bogs occur in depressions and near streams. The upland soils are comparatively hard. Jigawa has the highest sand content and the lightest color. Crops do relatively well on this soil even in years with below average rainfall. Damba, a darker, clayey soil, is the most productive soil for sorghum when rainfall is abundant. On the bright red soil iaa gargari, high yields of sorghum and millet can also be produced during a good rainy season. Hard rocky soils, such as fakg and gabagi, are rare. Hydrglogy. Numerous streams and four small rivers flow through the study area. Log bridges afford the only means of crossing many of these watercourses during the rainy season. Small ponds are the only other hydrological features. Only a few settlements do not have a perennial source of domestic water. 47 Vegetation. The Kwiambana Forest Reserve covers about half of the Sudan study area. Outside the reserve, an estimated 25 percent of the land is under cultivation. Large tracts are relatively densely wooded. Acacia, neem, and baobab are the most conspicuous tree species. Guava are the most numerous fruit trees. Other trees of economic and nutritional importance are the shea butter tree (kadanya; Balyrosperum Parkii) and the wine-yielding palm (itacan gamma; Raphia vinifera). Meadowlands offer grazing for livestock. Eaaa_. Wild animal populations, including lions and hippopotami, are reportedly high enough to prevent people from farming in some areas along the margins of the forest reserve. In 1989 a rampaging hippopotamus caused major crop damage, and monkeys are persistent pests. For some men hunting and trapping are important economic activities. Climate. The people of the Sudan study area also distinguish three seasons: damuna (the rainy season), dari (the cold season, with hanguru, the northeast Harmattan wind), and bazara (the hot dry season preceding the rains). Figure 2.7 presents mean monthly rainfall in Gusau for the years 1942-84. Located approximately 110 km north of the study area, Gusau is the closest station for which a consistent multi-year rainfall record was obtainable. The mean annual total of 962 mm is more than 400 mm greater than in the Sahel study area. The rainy season begins sooner, often at the end of April, and 25 mm is the average for 48 bc-occcoccn-uunc---------------------------------- l.------------------------------------------------- p-----------n.--------sumo------------------ ”IVA" 4’ I 81¢, 32?? .55' ’. . . - *C----------------------------------------- p----------------------------------- b........-.-..--..-—-.---.—- 'An’l:' 2:. ----------- ' ' M ' ' ' f” {Rf-we}; w' I. 'Q ‘35 ; j '\ 9. <4?" 36'»; x - 55%;? V — 300mm 1- 250mm 1— 200mm ----------------------150mm --~ 100mm 3v... I." I . . I-i-r'xvri-Z'Z-Zfrf " 23%?“ «mp-a. .- A. {0:53 51::- s r 50mm - I. 'fd'e' 'I-‘u'o' «- N mean annual total = 962 mm Figure 2.7: Gusau Mean Monthly Rainfall, 1942-1984 Source: Ahmadu Bello University Department of Geography (ND) 49 October. Both dari and bazara are consequently shorter. Although no other meteorological data were obtainable, bazara and damuna are decidedly cooler and more humid, and the Harmattan season warmer and its winds weaker, than in the Sahel. Human Features Figure 2.6 is based on two topographic maps published in 1970 by the Federal Survey of Nigeria. During the period 1970-90, considerable changes in the area’s settlement geography have taken place. More new settlements have been established, and probably more have been abandoned, than are depicted. Most of the new settlements are populated by migrants from areas farther north; most, but not all of the abandoned settlements were once inhabited by Lyaa gaaa, the people indigenous to the area. Dansadau village, the largest settlement in the Sudan study area, comprises approximately 400 compounds, most of which have clay brick walls surrounding rooms made of clay bricks. The palace of Sarkin Kudu, the district’s traditional ruler, is the largest structure in Dansadau. There also exist a courtroom, a police station, a small hospital, a large primary school and a secondary school, a compound for lodging guests, and an office staffed by Malamin Sarki, the traditional ruler’s scribe and administrative assistant. Several motorable roads of compressed soil traverse the village. Along the main thoroughfare, two women cook and 50 sell food, butchers grill and sell meat, and petty traders display their wares on tables. Various snack foods are available, including peanuts, fruits, and cooked cassava and taro. Fish are conspicuously absent. People do not normally eat meals in the open. Dansadau’s market is held on Fridays. Although it is only a small fraction of the size of Illela’s Sunday market, the range of goods available is nearly the same. The livestock section is especially small. A motorpark pa; aa does not exist. In the mornings taxis bound for Gusau and Birnin Gwari (a large town located 40 kilometers to the southeast) wait near the market. Except on Fridays, it may take an hour or more for one to fill and be on its way. The other settlements in Dansadau District are small compared to the nuclear villages of the Sahel study area. During the rainy season settlements consisting of three or four relatively dispersed compounds are nearly invisible amid the dense growth of crops and natural vegetation. A few other settlements are more compact, and may contain as many as 100 compounds. A minority of habitations are constructed entirely of clay brick; most have clay brick rooms enclosed by sorghum stalks. Semi-permanent Fulani hamlets, or ruga, now are numerous. 51 Research Methods The central methodological problem of social science field research can be summarized as data reliability ya. the rapresentativeness of case study findings. The researcher must perforce focus his/her efforts on specific communities, but generalizations from these data about processes affecting wider regions and populations will always be called into question. On the other hand, if the spatial scope of fieldwork embraces a large region and numerous villages, the researcher may well end up with a copious amount of largely unreliable survey data. This is because many, if not most topics of interest to social scientists are very sensitive. Household and community demographics, assets and cash incomes, farm productivity and some types of migration are all very often intensely private matters. Questions regarding these topics may engender exaggerated, obfuscatory, or equivocal responses. Furthermore, rural folk are often suspicious and sometimes even fearful of an outsider, especially one whose presence in their community depends on the permission of modern and traditional governmental authorities. Be they linguistic or otherwise, the researcher’s own shortcomings, his/her status as an outsider, and the sensitivity of many areas of inquiry are major reasons for the tentativeness and unreliability of many field data. For a variety of reasons migration in Hausa society, the focus of this study, is one of these very sensitive 52 subjects. Migration is so complex, and informants sometimes so reticent or embarrassed, that accurate data may be extraordinarily difficult to obtain. Owing partly to these methodological obstacles Hill (1972), rather than presenting any major conclusions, chose to impart her insights on migration by offering a set of eight hypotheses. Watts (1983) identified credit, grain sales, and migration as very sensitive research topics, and in Katsina town encountered famine refugees’ extreme reluctance to discuss their experiences. According to him, "The sensitiveness of matters pertaining to ‘origins’ in Hausa society is such that it is often impossible to collect data on refugee families" (Watts, 1979 373). An additional and more recent obstacle is people’s fear of deportation. In 1982 and 1983, and again in 1985, the Nigerian government expelled tens of thousands of foreigners, many of whom had been attracted by the expanded economic opportunities of the oil boom years. The Western press, centered in Lagos, focused on the plight of Ghanaian deportees, but many other West Africans were also affected, including those from Chad, Mali, and Niger. The difficulty of gathering data on migration and famine is confirmed by my own experience. In Amarawa, both ward heads insisted that famine-related migration had never occurred. Several other prominent elders, and approximately half of the informants interviewed individually, also held this view. Group interview participants in Tudun Gudali Mai Ruwa acknowledged only that their settlements had been 53 destinations for famine refugees, and elders in the Gidan Alkasim hamlets cited just two incidents of household distress migration during famines. There is no panacea for the problem of data reliability, nor is there a comprehensive solution to the issue of representativeness. The best approach seems to be one that relies on patience, a substantial commitment of time, strong language skills, and the services of an outstanding field assistant. Mallam Bagudu and I confronted the data reliability/representativeness contradiction by proceeding slowly and circumspectly while broadening our horizons beyond "the village study." Mathodology 0 ion Much has been said about the tentativeness of the interview and the resulting data. Nevertheless the interview remains, and must remain, the principle source of data for social scientists. Compounds, granaries, health clinics, and schools can be counted; who lives in the compounds, who owns the granaries, who attends the clinics and schools must be learned by asking people. Farmland can be measured and yields estimated; only interviews might ascertain who receives the produce and the tenurial arrangement under which a farmer cultivates a particular parcel. Anthropometric measures can be taken, but the resulting data say little about the underlying causes of undernutrition. 54 Because of the central significance of the interview in most social science research, Briggs (1986) has called for a critical and contextual evaluation of the interview as a communication event. He asserts that a great deal of misapprehension stems from researchers’ inability to understand that the interview is an unusual form of communication, at odds with the ways in which people normally exchange information. In both formal and informal interviews researchers impose alien communicative norms on their informants. To attempt to deal with this problem researchers must strive to identify people’s resources for conveying information and the extent to which interviewing strategies might conform or be adapted to them. This is a formidable task. Briggs offers no single solution, but the incongruousness of the interview in relation to people’s normal modes of interacting should always be kept in mind. Recognition of monumental methodological difficulties — a sort of synergistic interplay of sensitiveness, suspicion, cultural and linguistic perplexities, and the tentativeness of the interview — has been important in influencing many researchers to undertake single village case studies. Watts (1983) prefaces the presentation of his field data with a catalogue of caveats, or "Hausa hazards." His decision to work in one village was informed in part by these methodological concerns for data reliability. Land tenure and land transactions are so sensitive and complex that Ross (1987) and his wife needed 18 months to investigate 55 thoroughly these issues in a hamlet of approximately 500 people. Demographic data from one village in Nigerien Hausaland were gathered only slowly and with the promise of anonymity by the Faulkinghams, a husband and wife team (published in Faulkingham and Thorbahn, 1975). Hill’s (1972) seminal Hausa ethnography is also a village study. Many additional examples of fieldwork undertaken in a single settlement could be cited. Almost without exception these researchers have expressed concern about the appropriateness of a village as a unit of study and the general issue of representativeness. Several researchers have struck what to me seems a tenable compromise for contending with the data reliability ya. representativeness problem. The essence of their approach is a series of formal interviews, including group interviews, and frequent informal discussions with a limited number of people in several villages. Extensive surveys consisting of very brief interviews in these and other villages may enable researchers to make inferences regarding the representativeness of their principal informants. For example, Painter’s (1986) work on labor migration from western Niger focused primarily on a total of seven "principal contact households" in two villages. These data were supported by extensive (N=424) but very brief interviews in six neighboring villages. Sowers’s (1986) approach to his research on migrant Fulani agro-pastoralists in Burkina Faso was similar. The survey encompassed more 56 than 40 Fulani settlements; 20 households from one settlement were selected for repeated interviewing. In a study concerning the influence of extra-village incomes on farming systems in the Gusau region of Sokoto State, Iliya (1988) supported his series of in-depth interviews (10 household heads from each of four villages) with rapid rural assessment surveys. Sutter (1982) also adopted the frequent interview approach, supervising a total of more than 80 series of interviews in two Nigerien villages in different ecological zones. (However, no field data were collected outside these villages.) The frequent interview approach offers several important advantages. Through repeated contacts a researcher has the opportunity to build a rapport with informants, to refine his/her interviewing strategies, and to probe more deeply some of the more sensitive issues. Eventually he or she will also be in a stronger position for judging the reliability of informants’ responses. Direct observation, and informal interaction with individuals not included in the sample, may provide opportunities for cross- checking information. Data from interview series in several villages, together with those from an extensive survey, should yield significant insight into the representativeness or otherwise of specific individuals and communities. Unfortunately this approach is not entirely suitable for addressing this study’s principle research questions on famine and migration. Frequent interviewing necessitates 57 small samples; return migrants in the Sahel, or refugees settled in the Sudan, could easily be excluded from such a sample. Even if a few were found in the small group selected for repeated interviewing, very little could be inferred from them about the pervasiveness and implications of this form of migration. Convincing some migrants to participate in a series of interviews might also prove difficult. Rapid rural assessment surveys would result in poor quality data. The sensitiveness of migration amidst drought and famine and especially the fear of deportation would justify reticence or prevarication on the part of many informants. I therefore had to develop a different approach for contending with the data reliability ya. representativeness contradiction. This incorporated: 1) the frequent informal interactions of participant-observation research; 2) series of interviews with expert informants; 3) series of group interviews; 4) series of in-depth interviews with individual household heads; 5) single in-depth interviews with individual household heads. I was thus able to corroborate and overlay data from several different field sources. Some information gained during relatively brief forays into the National Archives, Kaduna, was also important. Eessarshinthegallel In the Sahel study area my initial and principal research site was Amarawa. Amarawa is one of the three villages where Abdu (1976) determined that entire families 58 had emigrated during the early 19708 drought and famine. This was an important consideration because in some villages emigration during drought may be extremely rare (Faulkingham and Thorbahn, 1975; Watts, 1983; Mortimore, 1989). By June 1988 I had successfully negotiated the protocol for gaining access to the field. The representative of the Sarkin Gobir of Gwadabawa introduced me to the village head, who arranged for my accommodation in Amarawa and presented me to several groups of elders. I undertook no formal interviewing during the following six months. Instead I concentrated on refining my Hausa language skills while meeting as many people as possible, either in the village or on their farms, to explain the purpose of my presence and to ask for their future cooperation. I participated frequently in discussions with groups and individuals. I learned a great deal during this period, but I also misunderstood or misinterpreted many things. Nevertheless, some of my notes later proved valuable for cross-checking data and as a source of more general information. In the absence of an aspatial sampling frame, such as a tax list or census records, I decided to produce a map of Amarawa. This was accomplished with the assistance of Mallam Sadiq Yelwa, the chief of the cartographic unit at Usman Danfodio University, and the two neighborhood (ward) heads. Amarawa comprises 220 habitations. However, the household heads of nine of these habitations reportedly spend only a very small amount of time in Amarawa each year, 59 and were never available for interviewing. Their wives and children are the only inhabitants. An additional three compounds are living quarters for the wives and children of three wealthy men. Three large compounds were under construction and uninhabited during the period of fieldwork. Six shigifa are the habitations of widowed or divorced women, and young unmarried men live in two others. Because of the anxiety caused by censuses and their putative unreliability, and in light of the Faulkingham’s research on demographics and drought, I abandoned the idea of making my own headcount. Over the next 13 months, with the assistance of Mallam Bagudu, I attempted in-depth interviews with household heads from as many of the habitations as possible. My sample eventually included 162 Amarawa household heads. The rate of non-response was disappointingly high. I later learned, however, that Amarawa comprises about 450 taxpayers. Officially a taxpayer is any male over 18, although younger men often pay taxes. (Many in fact want to do so in order to obtain a tax receipt, a document frequently inspected by authorities at seasonal migration destinations.) A man does not usually attain the status of mai gida (family or compound head) until he reaches his mid or late twenties. It seems reasonable, therefore, that the sample included at least 50 percent of all household heads who are at least quasi-permanent residents of Amarawa. I 60 succeeded in securing at least one follow-up interview with 110 informants from this sample. In Amarawa during this time I had additional frequent formal and informal interviews with expert informants including the two ward heads, the Sarkin Noma (the "king of farming," a traditional office), the heads of the barber’s and butcher’s guilds, the village head’s scribe (Malamin Sarki), the town crier (Mai Shela), a primary school teacher who is an Amarawa indigene, the resident meteorological technician, and several other knowledgeable and very helpful individuals. I also conducted group interviews that sometimes included some of these people. Upon completion of the research, the sample of 162 household heads was stratified according to four socio- economic categories identified by the informants: the wealthy (maaa arziki or maaa gag gala), the "middle class" (masakaita), commoners or “peasants” (Lalakawa), and the extremely poor (iaa talakawa). The method used was pioneered by Hill (1972) and adOpted in research by Watts (1983) and Lennihan (1987). Through the deliberations of indigenous informants each household head is assigned to one of the above groups. I organized two such "stratification sessions," and the assessments of socio-economic status differed for ten household heads. The groups were unable to make a determination for fourteen household heads. Midway through the research in Amarawa I began interviewing in the two other settlements. Neither Tudun 61 Gudali Mai Ruwa nor Gidan Alkasim was included in Abdu’s 1975 study. Interviews in Amarawa had provided some general information about these settlements, but no specifics concerning drought/famine impacts and migration. It seemed appropriate to include in my Sahel study settlements about which I had no prior knowledge. Gaining access was relatively easy; I was introduced to the traditional authorities and many of the people by the Mai Shela of Amarawa, and initially the people seemed willing to participate. In Tudun Gudali Mai Ruwa and Gidan Alkasim I began with a series of group interviews, and then proceeded to conduct single in-depth interviews with individual household heads. The Tudun Gudali Mai Ruwa sample included 54 household heads out of about 150 taxpayers. I interviewed in all Tudun Gudali Mai Ruwa settlements. Gidan Alkasim has a total of 100 taxpayers. Owing to suspicion and fear only three individual interviews were possible in Zangon Yamma, the largest hamlet comprising 40 taxpayers. In the other three hamlets I succeeded in interviewing 18 household heads. The research in Tudun Gudali Mai Ruwa and Gidan Alkasim has an element of a "top-down" orientation, because in most cases it was necessary to be introduced to individual informants by the village head. The interviewing also .1acked the depth possible with repeated contacts, and formal :3tratification interviews were not practicable. 62 In seven surrounding villages I organized series of formal group interviews. Through informal interviews in these villages and elsewhere I learned that much of the resulting data concerning famine and migration were unreliable. In the village of Lakoda, however, informants were very forthcoming about migration during the 1983-85 period. 3w arc intheSuclan. In Dansadau District an intensive village study similar to my work in Amarawa was not possible. Limitations on time precluded follow-up interviews with most individual household heads. I relied on series of interviews with expert informants, series of group interviews, single interviews with groups, and single in-depth interviews with individual household heads. I pursued these four strategies concurrently except during the initial phase of fieldwork in April 1990. During this time expert informants and groups provided detailed accounts of the settlement history of Dansadau District. Based on this information I selected three villages where I hoped to find answers to the second principal research question. In Maganawa, Unguwar Mai Jan Kai, and Mai Tukuniya, I began with series of group interviews. I then proceeded with individual interviews: 21 household heads/61 taxpayers in Maganawa, 9 household heads/20 taxpayers in Unguwar Mai Jan Kai, and 6 household heads/61 taxpayers in Mai Tukuniya. 63 In a settlement of Fulani farmers I was able to interview two of seven household heads. Group interviews in eleven other communities provided important additional information on migration and drought, corroborated previous accounts of the district’s settlement history, and helped me in addressing the issue of representativeness. Data Limitations The most prominent bias in the fieldwork results from the exclusion of women. Although Islamic wife seclusion (galla or gargah) is not widely practiced in the study areas, attempting to interview married women would have been a serious affront to community mores. My research assistant and I were also uncomfortable with the idea of interviewing widows and divorcees. We decided that it would be best to err on the side of caution, and not take any risks that might compromise research on the principal foci of the study. Informal, usually brief discussions with post- menopausal women were sometimes valuable in cross-checking information. The individuals who declined to be interviewed represent an additional possible source of bias. Judging from the material composition of their compounds, however, it seemed that this group included both wealthier and poorer household heads. The size of samples and sub-samples are not always consistent among the tables that summarize field data 64 because: 1) some informants were forthcoming on some topics, yet reticent regarding others; and 2) follow-up interviews revealed inconsistencies that required the deletion of some of the information provided by some informants. I am confident that the data on migration from the Sahel study area (both famine-related and that occurring during non-famine years) are the best obtainable considering the sensitivity of the subject and the imprecision of memory. Individual informants who reported out-migration specified, among other particulars, the names of the household heads involved. Cross-checking was therefore possible. In Amarawa, these accounts were cross-checked further during group interviews. However, data on migrations that occurred prior to c. 1960 are probably less comprehensive and accurate than the migration data from the past thirty years. In the Sudan study area there was virtually no doubt concerning informants’ status as refugees. The only misgiving I have is that some informants identified as their origins localities, and not specific villages. Village names might have been valuable in attempting to determine the characteristics of settlements from which peOple migrate during famines. 65 Summagy The two principal research questions were addressed in two localities. The Sahel study area, located in northern Illela LGA, receives on average 530mm of rainfall per annum. Farms on both the sandy uplands and clayey lowlands are cultivated every year. Population densities are relatively high, uncultivated arable land is generally not available, and fallowing is extremely rare. Published sources indicated that the Illela area has a long history of population mobility and famine. Research in Illela-area villages focused primarily on return migration by famine refugees. Dansadau District, constituting the Sudan study area, probably has an average annual rainfall total of at least 1000mm. Most of the land is not under cultivation, and population densities are relatively low. Apparently no previous field research had been undertaken there. Fieldwork focused on the second principal research question concerning migrants who have not returned to their villages of origin. The only possible way of finding answers to the research questions was to ask people. But questions concerning famine-related migration are extremely sensitive. In both study areas, interviews with groups, individuals, and key informants were accomplished in several settlements. This approach enabled me to address the gaga reliability yaa raprasentagivaness contradiction while seeking answers to 66 this study’s principal research questions on the excruciating experience of famine and its aftermath. The next chapter examines key theoretical and pragmatic issues pertaining to famine, especially famine-induced migration. CHAPTER THREE: PERSPECTIVES ON FAMINE AND MIGRATION The present chapter is a focused review of the literature on famine. Discussions of definitions and causes, vulnerability to famine, and the progression of famine through time and across space will: 1) situate this study’s two principal research questions upon a broader theoretical canvas; 2) elucidate the practical importance of these questions; and 3) provide contexts for the empirical findings of field research. The greatest emphasis is on distress migration which, according to the continuum model, is a last-ditch famine-coping strategy undertaken only after household assets have been liquidated in an attempt to procure food. The African definition of famine suggests that this is by no means always true. Moreover, an alternative conceptualization of the temporal progression of famine cites “migration to new lands" as a strategy for “modifying” conditions of intense hardship, and reports from Ethiopia indicated that two groups of refugees recovered from famine. Yet published accounts also have implicated distress migration as a cardinal agent in the expansion of famine zones. The focus on migration, as well as the issues of definitions, causes, and vulnerability, are of critical significance to attempts to develop effective programs for early detection and mitigation of famine in Sub-Saharan Africa. 67 68 Sub-Saharan Africa is, to borrow Dando’s (1980) term, the quintessential "famine region" of the late twentieth century. Well publicized famines assumed sub-continental proportions in the 19708. According to a United Nations list, people in twenty African countries were suffering from famine again in 1985 (Timberlake, 1985). Western news media focused attention on the agony of Ethiopia, yet famine also plagued millions in West, East, and Southern Africa. In 1993, news media were reporting daily on the strife that plunged Somalia into a protracted, devastating famine. The 1993 famine in the Republic of Niger has, in contrast, received virtually no coverage in the major North American news media. The 1993 Nigerien famine demonstrates how poorly understood, how pathetic, and how intractable the bane of famine in Africa is. Excruciating famines wracked Niger in the early 19708 and from 1983 to 1985; in 1987, famine decimated livestock holdings and uprooted families in some parts of the country. Despite these recent famine episodes, and despite a relatively dense deployment of expatriate "famine experts" and an eight-year-old famine early warning system (FEWS), an estimated 200,000 Nigeriens were experiencing famine in 1993. Reports referred to distress migrations at a "sustained rate," and to deaths "due to insufficient food and inadequate medical care" (Voix du Sahel radio network, in Foreign Broadcast Information Service-AFR-93-123, June 29, 1993, and Foreign Broadcast 69 Information Service-AFR—93-137, July 20, 1993). This crisis arose not from the unpredictable impacts of armed conflict, but from events whose famine-precipitating potentials should have been recognized and monitored: a poor rainy season, locust attacks, and rodent infestations. This most recent failure to arrest famine attests to the continued existence of vast lacunae in Western knowledge of Sahelian human ecology and the dynamics of famine. Definitions The concern with defining famine has raised two questions of considerable theoretical and practical significance: 1) How should famine be differentiated from the chronic poverty and malnutrition affecting a large proportion of Africa’s people? 2) Is Africans’ concept of famine different from Westerners’? Mesfin’s (1984:9) definition calls to mind the soul- searing images transmitted to Western homes at the dinner hour. "Famine is the most negative state of food consumption under which people, unable to replace even the energy they lose in basal metabolism, consume whatever is stored in their bodies.... Famine is general and widespread, prolonged and persistent, extraordinary and insufferable hunger lasting for several months and affecting the majority of the rural population over an extensive area resulting in total social and economic disorganization and mass death by starvation." According to Watts (1983:13), famine is "a societal crisis induced by the 70 dissolution of the accustomed availability of, and access to, staple foods on a scale sufficient to cause starvation among a significant number of individuals." Starvation is the key word in these decidedly Western definitions of famine. For Africans, however, the essence of famine is a threat to their way of life, not necessarily a threat to their lives. The liquidation of productive assets and destitution are thus the key concerns (Walker, 1989; de Waal, 1989). In light of this basic difference between Western and African definitions of famine, it seems reasonable that Sahelian farmers would migrate early on during a crisis period in a deliberate attempt to preserve productive assets. That mass migrations occur before asset liquidation becomes widespread has major implications for famine early warning systems that monitor what are supposedly indicators of destitution, such as grain and livestock prices and the quantity of household assets for sale at regional markets. Causes Since the early 19708 scores of studies have sought to explain the scourge of recurrent famine in Africa. Despite the disparate theoretical, ideological, and disciplinary perspectives represented in this large corpus of work, nearly all analysts agree that famine is both pre-disposed and catalyzed. Warfare and various environmental perturbations are identified without much controversy as catalytic causes. The ultimate, or pre-disposing cause of 71 famine is probably its most polemical aspect. That the poorest peOple apparently suffer disproportionately from famine inspires many of these often acrimonious debates. Adherents to dependency theory charge that capitalism, unevenly introduced in the wake of colonialist penetration, is ultimately responsible for twentieth century famines. According to Watts (1983), taxation, monetization, and compulsory cash crop production undermined the "moral economy," a set of technical, social, and political arrangements designed to guarantee subsistence to everyone. In pre-colonial times, the ultimate cause was equivalent to the catalytic cause; famines occurred only as a result of a particularly virulent environmental perturbation or military aggression, and all but the extremely wealthy faced death. This view of pre-colonial famine causality is concordant with Fur rural producers’ perception of the etiology of contemporary famines. The Fur affirmed to de Waal (1989) that the local rural economy is sound and the national economy insignificant; that severe drought, exacerbated by desertification, is liable to bring about famine under any circumstances; and that the rich and poor both suffer. The Fur view contradicts the "external view" that outside political economic forces have eroded traditional capacities for coping with drought, thereby instituting increased marginality and insecurity for the poor. De Waal suggested that the most complete and accurate explanation of famine’s ultimate cause would incorporate 72 elements of both views. His most significant finding in support of the Fur view is that deaths during the 1984-1986 famine occurred in nearly equal proportions among relatively wealthy and poorer households. This unexpected revelation challenges a very basic tenet, that famine is and always has been essentially a class phenomenon. Another tenet, that the ultimate cause of famine is the same as the ultimate cause of poverty, has also been challenged. Iliffe (1987) suggested that famine is the major cause of poverty in Africa’s savanna environments, and that famine should be attributed to climatic and political insecurity. Mellor and Gavian (1987) also state that the underlying cause of famine is crop failure brought about by warfare or climatic perturbations. Delehanty (1988), who conducted fieldwork in Niger during the 1984-1985 famine, was in sympathy with "external" dependency explanations of the ultimate cause of famine. Yet in his view these explanations were largely irrelevant in the case of the 1984 drought, which was so overwhelming that even the wealthy were forced to evacuate farmsteads. Dreze and Sen (1989) also acknowledged that the "erosion of the moral economy" explanation is rendered meaningless during a severe community-wide environmental onslaught. Anyone entering the debate on the relative importance of ultimate and catalytic causes immediately confronts an old dilemma. By emphasizing catalytic causes, one courts accusations of environmental determinism; a near-total focus 73 on social structure abnegates humankind’s capacity to adapt and innovate, and is therefore also deterministic. I believe that the importance of catalytic causes has been consistently understated by dependency theorists and others for whom social structure is the dominating concern. But I also believe that one ultimate cause is paramount: the failure of policymakers to comprehend that freedom from the threat of famine is the most basic pre-condition for development and the alleviation of poverty. This assertion accords with Iliffe’s view, that famine causes poverty. Vulnerability Vulnerability is currently the key concept for theoreticians and for practitioners involved in famine early warning and mitigation programs. In addressing the question "Who is vulnerable to famine?" analysts focus on the risks and consequences of famine, rather than on its causes. Downing’s (1991) analysis of vulnerability is among the most prominent. Reflecting the strong influence of Sen’s (1981) "entitlement theory," the principal concern in his analysis is with vulnerable socioeconomic groups. Specification of household incomes, and incorporation of these data into numerical models that would signal "entitlement failure" for poor people, represent the main methodological orientation. Downing also designates "regional food shortage" and "individual food deprivation" as additional "domains" of famine vulnerability. He acknowledges that vulnerable groups may be concentrated in 74 specific regions, but does not emphasize the potential importance of identifying spatial patterns of vulnerability. Data limitations severely restrict the practical significance of the individual deprivation domain. Vaughan’s (1987) research on the 1949 famine in Malawi is unsurpassed in demonstrating the critical influence age and gender had on vulnerability. Socioeconomic status was important, the most vulnerable group comprising households normally dependent on wage labor to meet a substantial portion of their staple food requirements. Oral testimony consistently revealed that young children and the elderly suffered more than other age groups, and that women abandoned by their husbands were also highly vulnerable to the impacts of famine. Of these, the inability to nurse infants and prostitution for food were the most shameful and degrading. Adult men, however, reportedly lost the most body weight and died most frequently. Vaughan argued that intra-family relationships can be as important as class divisions, and that any analysis of famine vulnerability must take both factors into account. In their model for situating famine vulnerability in COgnitive space, Watts and Bohle (1993) address age, gender, and virtually every other conceivable concern. The model is based partly on entitlement, but seeks to redress the major Shortcomings of Sen’s (1981) theory: its neglect of historical process, manifested by the narrow focus on immediate determinants of vulnerability (food price 75 increases, hoarding, the collapse of labor markets); and its failure to embrace gender, age, and ethnicity, owing to the pre-eminence afforded occupational and socioeconomic status. The space of vulnerability is represented by a ternary diagram, with "entitlement," "empowerment," and "political economy" designating its sides. This diagram is capable of portraying any or all vulnerable entities, including regions, groups, and individuals. Another attribute is the ability to represent changes in vulnerability over time. Conceived on an eminently theoretical plane, the model is intellectually very appealing. Whether any practical application is possible remains to be seen. I contend that identifying spatial patterns of famine vulnerability would be of greater practical significance than approaches based on statistical scenarios and data from household income surveys. Reliable data on household economics are often exceedingly difficult to obtain, requiring a considerable commitment of time for series of interviews with a relatively small sample of informants. On a theoretical level, a focus on household economics, or entitlements, is of dubious relevance in rural Africa. As de Waal (1989) demonstrated, people who face famine frequently elect not to use what entitlements they have to purchase food. Instead they endure hunger, eat wild foods, and migrate in order to preserve productive assets. Persistent spatial patterns of drought, a prominent catalytic cause, have been detected. Nicholson (1985, 1986) 76 and Farmer and Wigley (1985) have identified at least four sub-continental scale patterns of drought and abundant rainfall in Africa. Vermeer (1981) discerned patterns of climatic stability and mutability in Mauritania during the 19608 and 19708. The persistence of these patterns and others elsewhere in the Sahel has been confirmed by Tucker _; al.’8 (1991) research on remotely sensed vegetation indices for the period 1980-1990. Palutikof (1986) has proposed that drought-prone areas and "spatial compensation zones" (characteristically drought-free areas) could be identified through analyses of decades-long rainfall series and soil moisture data. On several different spatial scales, enduring patterns of the incidence of famine can also be identified. In India, recurrent famine has been most typical in the sorghum- and millet-growing regions of the Deccan plateau in Gujurat and Rajasthan (Currey, 1979); other regions having an especially pronounced liability to recurrent famine include Maharashtra (Subramanian, 1975; McAlpin, 1983), Bihar (Singh, 1975; Dreze, 1990a), and Bengal (Currey, 1979). In contrast, some regions of southern India have apparently been free of famine for periods spanning many generations (Murton, 1984; McAlpin, 1983). Nicholson's (1980, 1981) Chronologies for western and northern Africa also indicate that sub-continental scale patterns of famine incidence have persisted over several centuries. 77 Numerous persistent national scale patters of famine incidence have been detected. Through an exhaustive analysis of archival materials, Mesfin (1984) determined that during the years 1955-1977 Ethiopian famines recurred in five regions (Tigray, Eritrea, the middle Awash valley, and the southeastern lowlands of Hererghe and Bale). A narrow famine-free corridor persisted between Gonder and Kefa. Iliffe’s (1990) historical study of famines in Zimbabwe identified the climatically and hydrologically capricious Zambezi valley and semiarid southern Mashonaland as the regions most persistently prone to famine. On the other hand, hardship in the highveld region was only rarely worse than an intense seasonal food shortage. In Niger, the Zinder region and the Zarmaganda have been consistently more famine-prone than most of the rest of the country. Yet within these regions, very significant local scale and inter-village differences in famine impacts have been recorded (Fuglestad, 1983). Local scale differences in famine impacts in northern Kano State, Nigeria, have been discussed by Mortimore (1989). The severity of famine has consistently increased from west to east across the state. Within the western Indian state of Maharashtra, the "great famine tracts" that Mann described in 1955 also experienced the worst hardship during subsequent drought-induced famines (Dreze, 1990a). The severity of famine has also varied considerably from village to village. In a study of the early 19708 78 drought and famine in the Department of Maradi, Niger, Campbell (1977) elucidated extreme inter-village differences in the prevalence of distress migration to food distribution centers. The Faulkingham’s detailed fieldwork in villages near Madoua, Niger, further illustrates the significance of these differences. Crude death rates in their principal study village actually declined from 1969 to 1973, and virtually no distress migration of families occurred during the drought. Temporary labor migration to Nigeria was the most important strategy for coping with extreme dearth. In contrast, between 55 and 80 percent of the populations of four neighboring villages reportedly emigrated. Village headmen regarded these migrations as permanent (Faulkingham and Thorbahn, 1975). Intervillage differences in famine impacts have been identified in localities in northern Cameroon (Campbell, unpublished data), in Zimbabwe (Dreze, 1990b; Campbell a; al., 1991), and in the Senou region of Mali’s interior delta (personal communication, M. Bangaly, Save the Children Fund-U.K director, Bamako, 1992). De Waal’s (1989) research during the 1984-86 famine in western Sudan demonstrated a close relationship between location and mortality. Death rates in three of ten case study settlements were significantly different statistically than the calculated region-wide rate. In two settlements, mortality was higher (p < 0.01, p < 0.05) and in one settlement lower (p < 0.05). De Waal suggested that the influence of location may mask the effect of intra-village 79 differences in socioeconomic status. Access to safe drinking water and the most basic medical care were two highly significant spatial variables. A strong association between location and famine vulnerability is suggested by recent nutritional studies. In Kenya, important differences in the prevalence of childhood malnutrition have been linked to ecological parameters and agricultural production systems. Malnutrition rates were highest among children in coffee- growing households in Eastern province, lowest in children whose parents cultivate drought-resistant crops in semiarid zones, and moderate in the children of monocrop corn farmers. A statistically significant difference in nutritional status between children from relatively wealthy and poor households was demonstrated only in Kenya’s Nyaza province. Elsewhere, differences in nutritional status between children from different socioeconomic strata were 1., 1986). Seyoum a; a1. (1986) not discernible (Haaga a; also determined that household income had no significant influence on nutritional status in two Ethiopian communities. In Kenya and Malawi, Kennedy and Peters (1992) found that incomes of a; facto female-headed households were the lowest of all socioeconomic groups considered in their field research. Nevertheless, the nutritional status of children in these households was significantly better than children in higher income male-headed and fig jaga female- headed households. Better child feeding practices and other 80 nurturing behaviors can improve nutritional status despite household poverty. Anthropometric measures of randomly selected children from villages situated within a c. 150 square kilometer area of Kwara State, Nigeria, demonstrated striking inter-village differences in nutritional status. The villages’ locations and ecological characteristics are responsible for the disparities (Ebomoyi, 1987; Ebomoyi a; al., 1991). A detailed dietetic investigation in neighboring communities in Mali permitted a similar inference (Martin, 1985). The persistence of inter-village patterns of malnutrition and famine vulnerability has not been demonstrated by longitudinal studies. I argue that specification of these patterns would provide the strongest possible foundation for an effective famine early warning system. This endeavor would also define development initiatives that would substantially reduce the threat of famine and alleviate poverty. The knowledge and participation of people who have experienced famine in the past, and who remain vulnerable to famine, are essential to achieve these goals. Tag Progression 9f Famine through Time Definitions, causes, and vulnerability are issues often subsumed in studies that focus on the temporal dimension of famine. These studies have provided inspiration and support to proponents of widely differing theoretical perspectives. When a famine begins, and what happens to whom as a crisis 81 deepens, should also be of great interest to experts on famine early warning and mitigation. People’s behaviors, including their strategies for coping with famine, best illustrate the progression of famine through time. Different frameworks for conceptualizing and analyzing famine-related behaviors have been advanced. Dirks (1980) has elaborated a very broad framework consisting of three phases: alarm, resistance, and exhaustion. Intensified sharing, or an expansion of the "moral economy," typifies the advent of the alarm phase. As conditions become worse, people often adhere more strictly to the rules of their religion, and may engage in fasting or other rituals. Increasing restlessness and irritability may soon lead to political instability. Sales of productive assets and mass emigration are behaviors signalling the end of the alarm phase. The resistance phase includes a range of what might be considered anti-social behaviors. Hiding and hoarding food, and cessation of household members’ interactions with the community (involution of the household), mark the demise of the moral economy. Raiding may occur, but the probability of organized rebellion diminishes. The exhaustion phase is characterized by the total individuation of society. Elders may be sent out to die, children sold or abandoned. According to Dirks’ sources, famines that occurred in Ireland, Eastern Europe, and South Asia between 1845 and 1945 progressed through all three phases. Only two of the more recent famines reached the exhaustion phase: 82 the 1973-75 crisis in the Horn of Africa, and the 1964-65 famine among the Ik people of Uganda. Exhaustion phase behaviors, however, can be discerned in Vaughan’s (1987) analysis of the 1949 famine in Malawi, and in Clarke’s (1978) account of the early 19708 Sahelian famine. Perhaps during all famines the exhaustion phase has been reached in some communities or among certain groups. Several frameworks commence with categories that encompass what are clearly long-term adaptive behaviors, but stop short of including categories equivalent to Dirks’ resistance and exhaustion phases. Thomas a; al. (1987) attempted to organize and describe famine-related behaviors as response modes, response levels, and response sequences. These categories overlap in some rather perplexing ways. Response modes include various forms of agronomic adaptation, the initiation and perpetuation of social support networks, distress migration, and passive acceptance of the consequences of food deprivation. Response levels are cooperative group behaviors, which also may include appeals to patrons and distress migration. The response sequence category is epitomized by a continuum of strategies that people pursue as a famine progresses. A key concern is whether strategies such as pawning, asset liquidation, and distress migration are reversible after a famine ends. Rather than elucidating the temporal progression of famine, this diffuse framework imparts an appreciation for its complexity. 83 Jodha and Mascarenhas (1985) have focused their analysis on one aspect of this complexity, the apparent dichotomy between adaptive behaviors and famine-coping strategies. Mixed cropping, crop/livestock mixed farming, toposequencing, asset accumulation, and various forms of inter-household cooperation, are adaptive behaviors that are deeply imbricated in rural life. Two forms of short—term adjustments, risk/loss minimization and risk-loss management, are employed sequentially as a crisis progresses. The first set of strategies are principally agronomic, and include midseason adjustments in planting or weeding operations, reducing hired labor, and attempts to salvage crops or crop residues. Postponing social commitments, reducing food consumption, wage laboring or labor migration, and asset liquidation are the most significant examples of risk/loss management strategies. Table 3.1 is perhaps the most lucid and comprehensive framework for conceptualizing the temporal progression of famine. It includes adaptive behaviors and coping strategies, as do the frameworks of Thomas a; a1. and Jodha and Mascarenhas, but extends to incorporate elements of Dirks’s exhaustion phase. This framework also situates adaptation, coping strategies, and serious famine impacts within four general spheres of interaction. Coping strategies should be a central focus of research aimed at developing a famine early warning system. To identify the sequences in which rural people resort to various coping 84 strategies, analysts must shift back and forth among the columns of Table 3.1. An analysis of more than thirty community-level studies among African farmers, agro-pastoralists, and pastoralists has identified broad similarities in coping strategy sequences (Campbell, 1986, 1990b). From a review of several of these studies, Corbett (1988) deduced a three-stage model of the temporal sequencing of famine-coping strategies: 1) Insurance Mechanisms (changes in agronomic practices, sale of small livestock, reduction of food consumption, collection of wild foods, inter—household transfers and loans, increased petty commodity production, labor migration, sale of non-productive assets); 2) Disposal pf Productive Assets (sale of livestock and agricultural tools, sale or mortgaging of land, credit from merchants or moneylenders, reduction of food consumption); 3) Destitution (distress migration). An effective famine early warning system would call for outside assistance before rural producers resort to stage two coping strategies. Timely delivery of aid would insure the basis of people’s future livelihoods. Corbett’s model is generic. Recent research has demonstrated that the repertory, efficacy, and sequencing of famine-coping strategies vary from community to community (Campbell ap _l., 1991). This variation should be a critical focus of assessments of famine vulnerability. Site-specific ecological conditions, population density, 85 3an ..fi Hm corsm 5am “83695 SH. d .882. .....mmnsEU Hm0m30m .886 8:8 8:85:38 53 macro 66 ......sz Emma Sui mmocmaaomcoo comm 82.855 doze. 2 $82. memo. BEG $226.9 2 on coco-Eu ”memo. $50 mmocoaaomcoo 92m Seam moan. Soc 9 9992 8:998 523 £0282. .mummmm coo. .mqoto sow comesE >3 SEQ scams .6 888:8 .882 839m 952 3 :26me 808 28, 59.9 606208 65:9: ”30:8“. 933 a: 2:5 56.96 DEERE ”398%. 8.6305 ”92m .20 25.0 ”mm: mmm=_> S=EE umUchm seems *0 00:0: 3000 :50)th _8_=_oa .mEoEcESCw o_Eocoow Edam guano noon. 5.? acaoo .8 $5325 .0 83338220 #6 033. 86 settlement history, and several socioeconomic and cultural characteristics may all have a profound influence on the availability of coping strategies. Another key focus is the dynamic nature of coping strategy parameters. Deteriorating agroclimatic resources and the cumulative effects of several poor harvest may have reduced the number and effectiveness of coping strategies in many communities. For example, in some areas of Ethiopia people had coped with up to six consecutive years of harvest failure before suffering severe famine impacts during 1973-74. In contrast, the famine of 1984-86 was triggered by only one or two deficient harvests (Walker, 1989). In some communities, on the other hand, new coping strategies may be emerging as a result of long-term agronomic and socioeconomic adaptations (Richards, 1985, 1986; Mortimore, 1989). Most of the existing works on coping strategy sequences do not take into account this possibility. Another shortcoming of many of these works is that they do not directly address the denouement of famine, or acknowledge the possibility of a famine recovery phase. This criticism also applies to the broader frameworks describing the progression of famine through time. Jodha and Mascarenhas’s framework ends with coping strategies; l.'8 discussion of response sequences concludes Thomas a; by raising the issue of reversibility. The total individuation of society is the terminal phase in Dirks’s framework. Although Table 3.1 suggests that migration may 87 be an early and effective famine coping strategy, the other outcomes specified are coups d’etat, desertification, crop and livestock losses, and death. For tens of thousands of Africans, death has certainly been the denouement of recent famines. But for those who survived, was destitution enduring and inevitable, and was massive out-migration the signal symptom of this outcome? Watts’s (1983; Figure 3.1) representation of coping strategy sequences strongly implies that stage three of Corbett’s model is indeed irreversible. Figure 3.1 is based on field data that purportedly demonstrate the "ratchet effect," a concept that concerns the differential impacts of famine on different socioeconomic groups and the irreversibility of stage two and stage three coping strategies. Because of the "sell cheap, buy back dear" character of famine/post-famine transactions, households that sell productive assets have little hope of regaining them. Apart from death, the denouement of famine is therefore, according to Watts, the pauperization and redistribution of a large proportion of Africa’s people. A few empirical case studies, however, have indicated that this conclusion is seriously flawed, and that it epitomizes a rather crude form of radical determinism. In 88 09.11.82 ates ”momaow 3001809 3218.8 Soolammc 20029.. 302 comes: “mmzmxocn $029.. coon 10.: 0E:. >>O.. 3 o m 0.002 058.01 w. m Ex 80: 59m Botom w a m M. 303995 .038 Loam. ._0m 1. ... w W 30299.5 @599 00300 >5 w W. S .A 0.00502. __mEm to 0.0m m. . 8 90092 $00895ch8 50: m >009: .0 A952 0229: 99m Bottom m 3 a 9030 0:09:00 to 0.0m 8 0098.9 to 9.6005 009E193 gum 002996-50 EmcmEcmd >>O.. 10.: 39.25 noon. on 39333. to cos-H.590 use cocoa—00w .92“th "tn 2:9... 89 Ethiopia, distress migration by one group of people during the early 19708 famine was actually the first step in a successful indigenous resettlement program. The refugees’ success was attributed to their avoidance of relief camps, to various forms of inter-family assistance, and to diplomatic accommodations with neighboring people at their destination (Turton and Turton, 1984). One group of migrants who arrived in the Northern Nigerian village of Dagaceri during the 1984-85 famine had maintained possession of their donkey, and by early 1986 had built a house and secured access to farmland (Mortimore, 1989). In Darfur, Sudan, de Waal (1989) determined that temporary household migration during times of famine was indeed an effective strategy for maintaining assets and recovering from famine. However, migration resulting in agricultural settlement in new lands is typically undertaken only during non-crisis periods by relatively well-off farmers. These case studies confirm Jodha’s (1975) speculation regarding the existence of a famine recovery phase. The Ethiopia study goes a step further, indicating that far from leading inevitably to pauperization, the denouement of famine can even entail positive developments. Although this finding contradicts mainstream views on the probable outcome of famine, it is consonant with several scholars’ ideas on the subject of climate and the human career. Butzer (1982) suggested that climatic variations in 90 pre-history provided the impetus for novel adaptive behaviors and a boon to cultural evolution. Childe (1928) contended that desiccation during the late Pleistocene forced the concentration of humans, flora, and fauna around a decreasing number of permanent water sources. The viability of hunting and gathering would have diminished as a result of increased competition for circumscribed resources, while the compulsory intimacy between humans, wildlife, and plants would have afforded the opportunity for domestication and the development of pastoral and farming systems. That refugees from a desiccating Sahara brought agriculture to the Nile Valley during the fifth millennium B.C. is a distinct possibility (Hassan, 1988; Butzer, 1976). Tag Spread _£ Famine Across Space Pioneering research on natural hazards in industrialized countries described how losses from disasters cascade through society. People directly affected by disasters usually do not bear the consequences alone; some effects are transferred from the disaster area to larger regions and populations through social and economic linkages (White and Haas, 1975). The category "share the consequences" in Table 3.1, and some of the coping strategy sequences discussed previously, indicate that this also occurs in Africa. When food first becomes scarce, inter-household transfers and other moral economy practices usually take 91 place within communities. In some Sahelian villages this may happen every year in response to seasonal food shortages. Periodic episodes of intensified sharing reportedly contribute to a "levelling effect" among a community’s socioeconomic classes (Hill, 1972; USAID, 1992). Levelling is accentuated as a food shortage intensifies, and communities progressively lose their ability to support needy households and individuals. For some people remittances from relatives on labor migration may offset this deficiency throughout the course of a famine (Faulkingham and Thorbahn, 1975; Swinton, 1988; Campbell a; al., 1991). Other migrants may face stiff competition in labor markets. Their wages may decline; they may go for weeks without work; they may cease sending money to their families, or perhaps worse, they may return home empty-handed. An impaired moral economy, attenuated labor markets, and the loss of other coping strategies often force people to seek sustenance from beyond their home villages. Although distress migration might be an effective strategy for preserving assets or securing new ones, it is also a major means by which famine spreads across space. Simply put, the demand for food (or cash) attempts to move to the supply (Walker, 1989). The inability of governments and aid organizations to identify food deficit communities explains in large measure why the reverse rarely obtains in Sub-Saharan Africa. Famine, whether viewed as asset 92 liquidation or elevated rates of morbidity and mortality, diffuses on different spatial scales. Mobility is partly responsible for the spread of famine at the local level. One form of mobility, associated with small scale trade in staple foods, contributes to a levelling effect within localities. Some individuals travel to nearby communities or markets to procure small quantities of food to sell in their home villages. These people, as Dreze and Sen (1989) pointed out, bear little resemblance to the urban-based, often portly, and much maligned grain merchants. They are usually poorer women who are willing to undertake a sometimes rigorous journey in an attempt to feed dependents with their small profits. Webs of inter-village affiliations, of which kinship ties are the most significant filaments, represent culturally and religiously sanctioned conduits for mutual aid. As a famine progresses, people from the hardest hit villages begin requesting subsistence support from relatives and patrons in nearby settlements where, at least momentarily, food is available. Occasional opportunities for labor, remunerated either in kind or in cash, may also attract migrants, as might a relative abundance of wild foods, raw materials for crafts production, firewood and fodder for trade, and water for drinking and washing. Migrant families often take up residence in these neighboring villages. Their presence hastens the depletion 93 of water and vegetative resources, as well as their hosts’ granaries. The spread of famine from village to village has been documented. Chastanet (1983) described the diffusion of famines in the cercle of Bakel, northeastern Senegal, during the period 1914-1927. The destinations of migrants from famine-stricken villages in Niger were reportedly relatives’ villages not far across the border in Nigeria (Faulkingham and Thorbahn, 1975). In northern Kano State, Nigeria, Mortimore (1989) determined that inter-village migration was pervasive during the early 19708 famine. The moral economy and the web of inter-village affiliations has contributed to the spread of famine on local scales in the Seno Delta region of Mali (personal communication, M. Bangaly, Save the Children Fund-U.K. director, Bamako, 1992). In some localities, host villages may be capable of sustaining refugees from neighboring communities over the entire course of a famine. In other localities the limits of inter-village migration and the moral economy are realized as hosts’ granaries and wild food reservoirs approach depletion. The erstwhile donors may then be forced to seek similar assistance from father afield. At this juncture, a locality has become a famine epicenter. If numerous local epicenters develop, famine will expand to become a regional, or perhaps even a national phenomenon. 94 For many people in famine epicenters, few coping strategies remain. Day labor is completely unavailable, all livestock have been sold or slaughtered, wild foods are impossible to find, and household possessions may have been liquidated in a buyer’s market. Water for domestic purposes may become very scarce or unobtainable. In the absence of outside assistance, households may have only three choices: bear the consequences (Table 3.1); sell or mortgage farmland, and attempt to procure food with the often meager proceeds; or migrate to areas less severely affected by famine, to seek employment or gratuitous relief. Household distress migration is a fundamental means by which famine spreads on regional or national scales. A8 famine victims migrate, they carry the famine with them (York, 1985). Displaced households create increasing and eventually intolerable burdens in areas less severely affected by famine (Glantz, 1987). The spread of famine on sub-national scales can be inferred from Brooke’s (1967) exhaustive analysis of nineteenth and twentieth century famines in Tanzania, and from Becker’s (1986) research on famines in Senegambia during the 17th and 18th centuries. Historical data from central and southern Angola also indicate that famines commenced in some areas and not in others, and subsequently spread across space (Dias, 1981; Miller, 1982). According to Chastanet (1983), the greater the duration of famines in 95 the Senegalese Sahel, the more the "area of survival" expanded, forcing desperate households to migrate farther from their home villages in search of food. The role of migration in enlarging famine zones can also be detected in Colson’s (1979) account of her research among the Tonga people of southern Zambia. In northern Ethiopia, Cutler (1984) demonstrated that the 1972-74 famine diffused with a "ripple effect" entrained by mass migrations from localized epicenters. Food prices at markets on epicenter peripheries rose in response to the increased demand and concurrent reduction in supply brought about by people fleeing famine. Famine subsequently expanded southward from these areas into Ogaden between 1973 and 1975 (Goyder and Goyder, 1988). The ripple effect produced by migration from famine epicenters was also evident during the early 19808 crisis in Ethiopia. Pankhurst (1992) identified five areas in the central part of the country through which famine sequentially spread. By mid-1983, even Gonder, which Mesfin (1984) had included in Ethiopia’s famine-free corridor, had been overwhelmed by a massive influx of famine refugees from western Wollo (Kumar, 1990). De Waal (1989) also observed family migration from Darfur famine zones to areas where grain was available. But refugees purchased grain only infrequently, having relied instead on wild foods and on self-discipline to reduce food consumption. Depletion of 96 wild food reserves in the receiving areas and increased competition in low-status, low-wage occupations (collecting firewood and fodder, crafts, porterage, begging) may have propagated an expansion of famine beyond de Waal’s research sites. Mesfin (1984) detected national—scale patterns in the spatial progression of famine. In a section of his book compellingly entitled "The March of Famine," Mesfin explained how Ethiopian famines spread in an irregularly proceeding but clearly identifiable clockwise in-spiral from epicenters situated in the country’s peripheries. The analysis spans two decades (1958-77), and does not deal specifically with the role of distress migration in perpetuating the spread of famine. On this time scale, famine diffused because it was able to follow the path of least resistance, one corresponding to areas where subsistence production predominated, and commercial agriculture remained undeveloped. In many areas, earlier famines had increased vulnerability to famine in the 19708. The Spread pf Disease and Death Across Space Distress migration is a key element of the health crisis model, the most convincing explanation of famine mortality. The already precarious health and nutritional status of people afflicted by famine deteriorates further during the ordeal of migration. As refugees take up residence in overcrowded squatter settlements and relief 97 camps, they are exposed to a host of life-threatening infectious diseases. Refugees themselves may introduce diseases to other groups of migrants and to the local indigenous population. Migration, rather than food price rises, is the key to understanding the different patterns of increased mortality that resulted from eighteenth century famines in Western Europe. Post (1990:261) asserted that "almost without exception those states that passed through the most severe crisis mortality also witnessed a sharp rise in the incidence of unemployment, labor migrations, itinerant vagrancy, and mendicancy." During the early 17408 famine, Prussia and Austria experienced increases in mortality that were an order of magnitude lower than those recorded for eleven other West European countries. In Prussia the persistence of a feudal social structure, which prohibited subjects from migrating but which also obligated landlords to provide food aid, was largely responsible for the low famine mortality. England, in comparison to the other countries struck by famine in the early 17708, experienced an almost negligible increase in mortality rates. Post (1990) attributed this to the development of public and voluntary relief organizations and the influence their activities had in minimizing distress migration. Due perhaps in part to the European experience, the relationship between migration and famine mortality was 98 recognized during the British Raj in South Asia. A major goal of famine relief works was to check migration and the spread of disease (Srivastava, 1968). Since independence the Government of India’s anti-famine apparatus has achieved remarkable proficiency in arresting the temporal and spatial progression of famine. Although its efficiency in providing fodder and drinking water to drought-affected regions has recently been assailed (Dhanagare, 1992), and although its success in halting asset liquidation has been at best equivocal (Dreze, 19908), the system of gratuitous relief, fair price shops, institutional credit, and relief works has, according to Torry (1986), virtually eliminated distress migration and famine mortality. India’s most impressive effort at mitigating famine was undertaken in Maharashtra State from 1970 to 1973. Maharashtra suffered three consecutive years of drought and harvest failure that were at least equal in severity to the contemporaneous disaster in the West African Sahel. A dense network of relief works, attended by as many as five million people, minimized overcrowding, the distances traveled, and the incidence of household distress migration. Most analysts agree that famine-related deaths were averted (Subramanian, 1975; Dreze, 1990a). The Kenyan government’s timely actions to mitigate famine in 1984 and 1985 have been repeatedly praised. Appropriate data were gathered and analyzed, and the need to 99 import corn was identified. This and the mobilization of domestic transport were achieved relatively quickly. Food at subsidized prices was widely available through well established commercial distribution systems (Cohen and Lewis, 1987; Dreze, 1990b). These efforts were apparently very successful at preventing distress migration py entire families and "starvation deaths." Hunger, increased rates of malnutrition, and asset liquidation were nonetheless in evidence, especially in more remote areas (Dreze, 1990b; personal communication, D. Campbell, 1994). In the West African Sahel the pathetic record of deaths during famines frequently includes accounts of distress migration. Mass migrations, epidemics, and mortality were widespread during the severe droughts and famines of the 17408 and 17508 (Lovejoy and Baier, 1975). In Senegambia during the pre-colonial period, many famine refugees experienced a fate arguably worse than death as victims of the Atlantic slave trade (Curtin, 1983). The most ghastly famine of the colonial era occurred from 1913 to 1914. Distress migrations and death were pervasive in Hausaland and throughout the Sahel. Baier (1974) deduced that in the gapgla of Zinder (Niger), 85,000 of a total population of 350,000 perished. A French colonial administrator estimated that an additional 32,000 people fled to Nigeria (Fuglestad, 1983). Of the 4,000 people who reportedly died in a small area just to the north 100 of Katsina (Nigeria), a majority were famine refugees from Niger (Van Apeldoorn, 1981). The drought and famine of 1926-27 caused major distress migrations from many parts of Nigerian Hausaland. Watts (1983) documented extensive abandonment of farmsteads in several districts of northern Katsina Emirate during 1927. Infant mortality rates throughout the emirate rose from 278/1000 in 1926 to 438/1000 in 1927. (De Waal, 1989, has argued that these data are seriously flawed.) The synergism among drought, wartime food expropriations, and more general economic adversity led to famine in the Sahel during 1942 and 1943. Massive population displacements and death resulted. A flight of correspondence in the Nigerian National Archives, Kaduna, affords a glimpse of the suffering of a relatively small group of famine refugees (NAK SOKPROF 324. 793/Vol II). A telegram dated July 26, 1943 reported "large numbers of men, women, and children from [six northern Katsina districts] wandering Gusau, Chafe districts without food, money, or strength to work. Any famine relief measures taken here will only attract others. Suggest relief measures at home towns only solution. Grateful early action as Gusau, Chafe cannot support them." On August 9 some "rough figures" concerning the situation in the small town of Chafe were transmitted: of a total of 780 famine refugees, 506 were reported "too weak to move," but 274 were "walking." An 101 additional 252 refugees were reportedly encamped in Gusau, a town approximately 50 kilometers northwest of Chafe. By August 29 the refugee community in Chafe had increased to about 1000. A month later "scouts" from this community returned to report the near total unavailability of food in the home villages. The file on the famine refugees ends with the assumption that "there is little doubt that all the immigrants will settle here . . . crops are good, so there is no longer any question of the people starving, though it is perhaps rather hard on the peasantry who are their hosts." Deaths were not specifically reported, but the observation that people were "too weak to move" obviously indicates serious illness, possibly resulting from the stress of long-distance migration on foot or unsanitary conditions in the encampments. The writer’s assertion that relief measures in the encampments "would only attract others" seems cynical, and his call for relief in the "home towns" may reflect his anxiety about administering a major food distribution operation. But if relief had been implemented in the affected localities several months earlier, the incidence of distress migration, morbidity, and death would have been much lower. No one knows how many people died during the early 19708 famine in the West African Sahel. Watts (1983) quoted the most frequently cited statistic, that 100,000 people died in 1973 alone, and suggested that as many as 250,000 102 perished during the entire famine episode. Citing the reports of survey teams and Caldwell’s (1977) assessment of their data, de Waal (1989) argued emphatically that 250,000 famine-related deaths is a gross overestimate, and that famine mortality, to the extent that it occurred, was likely attributable to infectious diseases that killed people (especially children) toward the end of the crisis. Estimates of famine mortality in the 19808 are equally conjectural and controversial. Delehanty’s (1988:xiii) contention that "tens of thousands died" is based on first- hand experience in rural Niger from 1984 to 1985, and on conversations with others who were working elsewhere in the Sahel of West Africa. That mass migrations occurred during both the 19708 and 19808 famines is incontrovertible. Perhaps more than one million people were forced to flee famine-stricken settlements in the early 19708. In Niger during the 1984-85 famine, an estimated 400,000 people migrated in search of food and water (Timberlake, 1985). Medical researchers have recently focused on the connections between distress migration, the squalid conditions in squatter settlements and relief camps, and famine deaths. Their studies lend support to de Waal’s (1989) conclusion that mass migrations and conditions at the destinations to a great extent account for deaths during famines. Shears and Lusty’s (1987) study of famine migration and communicable disease epidemiology in northeast 103 Africa concluded that the major causes of morbidity and death were: 1) lack of basic health services for migrating people; 2) exposure of refugees to new disease risks present in different ecological zones; 3) increased disease transmission rates due to crowding in relief camps; 4) elevated susceptibility to diseases as a consequence of higher malnutrition rates. Owing to these causes, mortality rates in Ethiopia and southern Sudan during the acute phases of distress migration reached levels as high as 60 times the "expected" rates (Toole and Waldman, 1990). According to Toole a; a1. (1988), early detection of large scale population movements and provision of adequate rations to refugees would greatly reduce death rates by reducing malnutrition and the related heightened susceptibility to disease. Summary The African concept of famine is distinct from Western definitions in that the principal emphasis is the danger of destitution, not starvation unto death. Bearing in mind this basic distinction, it should seem reasonable that African farmers will migrate in a deliberate attempt to preserve productive assets and avoid the ratchet effect. If this strategy is successful, return migration would likely occur after agro-climatic conditions improve. The evidence strongly suggests that, in fact, the pre-colonial process of 104 advance and retreat from the southern Saharan margins continues even today. This revelation, and the discovery that the denouement of famine also can entail improvements in the material well- being of refugees, would substantiate a famine recovery phase. Moreover, documenting escape from the ratchet effect and a still-functioning moral economy would in large measure refute two mainstays of the radical interpretation of famine causality. Indeed, the moral economy is partly responsible for the spread of famine across space. This points up a fundamental contradiction with which famine early warning and mitigation programs must contend. Migration is an effective strategy for preserving productive assets and for securing new ones. But refugees also place demands for food and water on people along their route and at their destinations, and may thus propagate additional waves of distress migration. Furthermore, the ordeal of migration may result in severe illness and death for some refugees and their hosts. This contradiction, and potential strategies for contending with it, will be addressed in the final chapter. The next chapter presents a twentieth century <:hronology of environmental variations and famines in northwestern Nigeria and especially in the Sahel study area. Evidence of spatial variations in the incidence of famine is examined, in addition to reports that appear to demonstrate 105 the role of population movements in enlarging famine zones. The catalytic causes of famines are scrutinized. CHAPTER FOUR: CHRONOLOGY OF FAMINES AND CLIMATIC AND ENVIRONMENTAL VARIATION Historical context is essential to any assessment of famine impacts and especially this study’s principal research questions on famine-related migration. The length of the famine recovery phase, which in the case of refugees might entail return migration, has never been thoroughly investigated. It is therefore necessary to consider the entire record of twentieth century famines and their role in inducing population movements and resettlement. The primary goal of the present chapter is to develop a chronology of climatic and environmental variation and famine for the Sahel study area in particular, and more generally, for Sokoto State and adjacent regions. The chronology provides the indispensable context for examining data on distress migration and other famine impacts (chapters six and seven). Additional goals of this chapter are to examine what historical records reveal about the complexity of the catalytic causes of famine and to bring to light evidence of spatial variations in vulnerability to famine and of the spread of famine across space. The historical evidence, compiled for this chapter through a synthesis of archival documents, rainfall records, and series of interviews with village elders, shows that governments frequently have failed to recognize the early signs of famine. The inability to detect incipient famine, which persists to this 106 107 day, stems in part from ignorance about the catalytic causes of famine and famine-related migration. The reliability of famine early warning systems could be greatly enhanced by judicious applications of the lessons of history. The data presented in this chapter also suggest that areas of particularly high vulnerability to famine lack perennial sources of domestic water and dry season farmland. Such crucial baseline information would help to identify potential famine epicenters. At least ten famines have afflicted the people of the Illela area since 1900 (Table 4.1). These famines, in addition to the violence and oppression of colonial conquest and rule, have played major roles in this area’s settlement history. Eaply Settlement In pre-colonial times, the Illela area was part of the Emirate of Birnin Konni. A majority of the study area’s residents identify themselves as Konnawa, and many elders and people in their middle years bear the facial scars of this sub-division of Hausa ethnicity. The second most numerous group are the Adarawa, whose historical home region, Adar, is centered on the Nigerien towns of Tahoua and Illela-Umani. Gobir, a region to the east and north of the study area, is the original homeland of the Gobirawa people, the smallest group of Hausa-speaking people in the Illela locality. Older Gobirawa also have characteristic facial scars. Tuareg-controlled farming villages, populated 108 Table 4.1: Famine Chronology for the Sahel Study Area Year Name 1904 Makaji 1913-14 Saketariya 1920-22 Kyalle Zumunka, ’Yar Kirit, Buda Bai, 'Yar Kuzut 1926-27* Not Recalled 1931-32 Shekara Kyamro 1942-43 Takesharbo, Mazarkwaila 1953-54 Muda 1965-67 Mai Zobe 1973-74 Muna Sane, Shekara Karmami, 'Yar Taralle 1983-85 Buhariyya * Evidence from archival sources only 109 8me 000256.28 802 .mt<>a 88.528 08:5 802 mtg: 802 .9285 802 .0C<>e 00:005c9w 803 to coca—2 0030555 00000585 00000535, _80._ co. 52810.89 0280.80. .83 to. .83 to. .83 to. 602 .5529 2.2.3.2 .0802 4402 4002 ~40. $2 :2 82 82 82-52 82 .82 9.2 52-22 .82 82 52 R2 R282 2.2 82 32.82 2.2 22 .82 22 R2 22 2.2.2.2 R2 32 82 22-22 82 R282 82 82 :2 R2 22-22 22 22 R2 22 82 R2 22-22 22 22-22 22 82 S2 32 8252.82 Ozm. 00 000.2 . 025. 8.09.2 <53 99.00 000.2 25200 200.1562 026.0555. 22.9.5. zmwemmz, .092 can 90092 505.52 to. 0030.25.50 05:5“. 9m-_a_:o_oo "ué 030k 110 + Argungu Shae-n . Tm 0 Tuna. .Bamm Kebbn +Gwandu O 3.“? .Tdun '5 Zamfara River ‘ Naomi. l‘," 0“ Mom MM . .Mm , Wat-nu Bonn Karma I o 0 Dansadau Km . 50 100 150 Kiomoton J Figure 4.1: General Location Map of Sokoto Province/State 111 at least in part by Buzu "vassals," had been established in the general vicinity. A cluster of five of these villages was situated 35 kilometers north of Birnin Konni, and the large villages of Galmi, Gunfaro, and Kaura were located approximately 50 kilometers to the east and north (NAK SOKPROF. 6. 86/1904). The area around Goronyo, about 30 kilometers southeast of Illela, was described by British colonial officers as a southern outpost of the Nigerien Buzu (NAK SNP. 16. 670). Fulani mashekari (base camps) also existed in and around the study area before the colonial intervention. A ruin in the Gidan Alkasim hinterlands is probably the oldest observable evidence of human occupancy in northern Illela LGA. It consists of stone and clay remnants of walls that once enclosed an area of perhaps 3,500 square meters. The identity of the structure’s builders is a mystery to everyone. One group of elders speculated that it was a quasi-permanent military outpost of the Hausa state of Kebbi, whose zenith of power antedated the 1804 Jihad. Kalmalo, situated across the lake from Amarawa, is most likely the oldest extant settlement. P. G. Harris, author of the 1939 Sokoto Provincial Gazetteer, described it as "large and flourishing" at the time of the Jihad. Dango, situated near the top of an inselberg, and Lakoda, whose elders claim that their ancestors' occupation is of such antiquity that they have forgotten their founder's name, are examples of large and old villages. Alhaji I, the head of 112 Illela, maintains that his town was founded in the eighteenth century. Illela nevertheless remained relatively small and insignificant until recent decades. Amarawa also existed before the advent of European colonialism. Recitations by Malamin Sarki D.M. (the "chief’s scribe") of the succession of village heads and the number of years and months each held power indicated that Amarawa was founded in the 1860s. Rungumawa, another large village, also existed in the pre-colonial period (NAK SOKPROF. 635. 3336 and 110. C. 36). The Period 1899-1910 The devastation wrought by the Voulet—Chanoine expedition of 1899, during which Birnin Konni was sacked and burned, caused the first recollected influx of immigrants into the study area. A subsequent major influx, often referred to as "shigowa Adarawa" (the immigration of people from Adar), occurred as a result of Makaji, the first famine in living memory. By counting and marking in the sand amid deliberations (lissafi), three groups of elders independently determined that Makaji took place in 1904. Drought was probably a proximate cause. According to Fuglestad’s (1983) sources, the period 1899-1904 was marked by climatic instability and insufficient rainfall. Amarawa elders believe that the hamlet of Gidan Alkasim was founded as a result of Makaji. The years of tenure ascribed to Gidan Alkasim’s three headmen (including the present one, Mallam G) support a near turn—of-the-century 113 founding. The elders of Gidan Alkasim, however, said that they did not know whether drought and famine prompted their parents’ migration. They recounted that Mallam G's father, Alkasim, and the other founders were agropastoralists in Taboye, a village in Niger located approximately 75 kilometers northeast of Illela. During one of their annual southward pastoral excursions, they decided to establish a settlement in the void that existed between Dango and Amarawa. Permission was readily granted by Marafa, the Sarkin Gobir of Gwadabawa and a grandson of the Shehu, Usman Dan Fodio. A power vacuum stemming from the struggle between the French and the Tuareg may have made this migration possible. The founding of Gidan Alkasim may have resulted from a conjuncture: the "push" of drought-induced hardship, the "pull" of a well watered uninhabited area, and the opportunity to escape Tuareg suzerainty. The establishment of Zangon Tsakuwa, and the Buzu hamlets near Rungumawa, Lakoda, and Tozai, might be attributable to this or other similar conjunctures. In the years immediately following Makaji, large differences in tax rates motivated people to migrate from Niger into Northern Nigeria and the study area. A Mr. Brooks, on tour to the northwest corner of Gwadabawa district in December, 1910, determined that the tax rate across the frontier in "French territory" amounted to 4,000 cowrie shells per adult male, while a levy of 1,800 cowries 114 was being collected in Nigeria (NAK SOKPROF. 2121. S. 2347B). Brooks’s report offers a glimpse at agro-ecological conditions in an area from five to 100 kilometers west of Illela and Amarawa. He and his entourage arrived in Kalmalo from Gwadabawa via what was then the only "road" in the general vicinity of the study area. Brooks remarked on the 1 "fine tabki" (lake), and judged its dimensions as three 1.- miles long and one and one-half miles wide. From Kalmalo, .Jn. Brooks’s group proceeded by "bush paths" west-southwest along the Niger border for about 100 kilometers. In addition to gero (millet), maiwa (late millet), and wake (cowpeas), farmers were cultivating agga, a small cereal— like grass also known in Hausa as intaya, and in English as "hungry rice" (Dalziel, 1916). Cash crop production was limited to a "few" peanuts and "one or two small patches" of cotton. The chief difficulty in this area was a dearth of water for domestic purposes. The residents of Kurdula village were carrying water from Nakigaya, a "considerable" distance. The people of Kiso had been returning annually to Boto owing to a lack of water during the dry season. Salawa had been abandoned, its people having moved to Cilas, another village in the locality where water was available. The father of Sarkin Noma A and his family emigrated from Kutufari, a village in the vicinity of the aforementioned settlements, either a few years before or a year or two after Brooks’s tour, and subsequently established Katanga. /_'T:~1 "I k 115 Whether a turn-of-the-century dry phase and Makaji were in any way responsible for desiccation and the abandonment of villages is a matter for speculation. The 1911-20 Period The ghastly Sahel-wide famine of 1913-14 is known throughout the study area as Saketariya. Elders’ lissafi in 1989 consistently determined that it had occurred 75 years ealier. Saketariya has two meanings: that men were forced repeatedly (sake) to leave their homes to seek and gather (tariya) food , or that men divorced (sake) their wives and dependents and left to gather food. Elders recall, or recall being told, that Saketariya was associated with a terrible drought. The data available for the central Sahel suggest that rainfall had been poor in the years 1910-1912, and especially in 1913. In his 1913 annual report for Sokoto Province, E.J. Arnett, the British Resident, stated that there had been a shortage of rainfall and that the harvest was generally bad (NAK SNP. 10. 104P). Although the first five paragraphs of this document have disintegrated, it appears that the British administration had no appreciation of the peril faced by their subjects. Arnett’s 1914 report gives a surprisingly detailed and insightful, if understated, account of the famine that followed the 1913 harvest failure. According to Arnett, 500,000 people in the northern districts were affected by "partial" failure of village crops, while harvests in the 116 east and south of the province were above average. The Resident contended that Sokoto Province could have dealt adequately with the food shortage had it not been for the vast influx of refugees from Niger. Elders in the study area remembered that Sakatariya caused a second episode of "shigowa Adarawa" and "shigowa Buzaye." An additional concern of the British administration was the unprecedented number of Buzu camel, donkey, and ox caravans moving about the province in early 1914. Grain purchases by these traders reportedly caused prices to go up by "leaps and bounds," and the people of northern Sokoto suffered as a consequence. Arnett listed three groups who were "very bad off:" 1) the poorer, mainly agricultural classes; 2) Fulani who either did not farm, or who had insufficient croplands; 3) "chiefs" and their dependents who were engaged in administrative work, and who therefore had to purchase food. The problem of indebtedness was also acknowledged. Many farmers would have to sell a large proportion of the "generally excellent" 1914 crop in order to repay creditors. Illela-area informants recalled that fall; — borrowing one bundle of grain during the "hungry season," to be repaid with two bundles at harvest time — was common during Saketariya. In the aftermath of the 1913—14 famine, refugees from Niger were regarded as a problem. Their presence in Nigeria was "not desirable," perhaps because they were thought to 117 have fled French conscription for World War I. Accordingly, 225 refugees were "rounded up...and deported" in 1915, as were "a number" the following year (NAK SOKPROF. 635. 3336 and NAK SNP. 10. 148p). In 1918, a delegation of Tuareg "chiefs" from the Kel Gress confederation visited Sokoto to request help in recovering their escaped Buzu "slaves." Apparently by this time the British had softened their stance on refugees, refusing to accommodate the Kel Gress except in the case of Buzu who had settled in Sokoto Province during the previous two years. One hundred fifty men and their families were deported. Arnett reported that the Buzu's flight had been precipitated by the harsh demands of their Kel Gress overlords, from whom the French had been requisitioning supplies to reprovision their beleaguered Saharan outpost at Agadez. The Resident opined that the Buzu preferred the "easy life" in Nigeria but also attributed their immigration to the "desiccation" affecting "French territory" (NAK SNP. 10. 100B). In the study area at least two settlements were established at about the time of Saketariya. The parents and grandparents of the Fulani based in the western Tudun Gudali Mai Ruwa meshekari moved here during the famine from their encampment on the fadamalands near Gaidau (Figure 2.2). Tozai was founded by Zanke, the village's first head, and a group of families that had been ordered to move from Tarike (a village 15 kilometers to the southwest) by Sarkin Gobir Gwadabawa Marafa following a serious clash with Fulani 118 herders. Malamin Sarki read from a document written in glam; (Hausa in Arabic script) which recorded the succession and years of tenure of Tozai’s headmen. The document indicates 1915 as the year the village was founded. These are the earliest recollected examples of what Sidikou (1974) documented in western Niger: that the stresses and hardships of famine force migration and the establishment of new settlements within localities. Harvests in 1915 and 1916 were reported as "excellent" and "exceptionally good" (NAK SOKPROF. 635. 3336 and SNP. 10. 148p). Arnett's Sokoto Province report for 1917 stated that rice and sorghum crops had failed in many districts, although millet appeared to be good everywhere. Rainfall was below average in Sokoto (477mm) and in Niamey (375mm). Grain prices remained low until Christmas, and then rapidly doubled and stayed high until July 1918. Food shortages were reported in the northern and central districts in the first months of 1918. In and near the study area, farmers caught practicing trans-border cultivation (noma-jide) had their crops confiscated and some were imprisoned for brief periods in Birnin Konni (NAK SNP. 33. 555 and SNP. 10. 100P). Informants in the study area did not recall a famine occurring three or four years after Saketariya. Perhaps harvests had been adequate. Considerable quantities of wheat had been produced recently in Kaffe, a large village approximately 35 kilometers east of Illela where, owing to a shortage of imported wheat in 1918, the Native Authority 119 purchased five tons. Crops were generally good in Sokoto Province in 1918 (NAK SNP. 10. 100P). The 1919 annual report mentioned partial harvest failures in some districts as a result of poorly distributed rainfall (NAK SNP. 10. 10P). The 1220-29 Period The next famine that afflicted the people of the study area spanned the period 1920/21-1923/24. Group interview participants in Gaeti and Lakoda determined that droughts, meager harvests, and famines occurred in three consecutive years seven to nine years after Saketariya. The names applied to these years of adversity include Eyele Zumunka (ignore your age mates), ’Xe; Kirie, Beee Bei, and ’Xer KBQEL- This famine period, however, was not recollected by any other group of elders in study area settlements. The 1923 Sokoto Province Report related that people in Kwarre (a large village located midway between Sokoto and Gwadabawa) and its vicinity, as well as people in districts farther west, had suffered their third consecutive year of total harvest failure (NAK SOKPROF. S2. 219). The 1922 report included an account of famine-impelled distress migration from western Sokoto districts to areas in the east of the province, including Zamfara and Kaura Namoda. Some refugees reportedly pressed on as far as Zaria, Kano, and Bauci (NAK SOKPROF. 19. 265). Shenton and Watts (1979) contended that hunger throughout Northern Nigeria in 1920—21 was exacerbated by economic recession, depressed export crop 120 prices and the difficulty households faced in paying taxes, the influenza pandemic, and the concomitant heightened susceptibility to cerebro-spinal meningitis. Tens of thousands of people were estimated to have perished in Kano Province during this period of conjunctural hardship. The death toll in Sokoto apparently has never been estimated. The 1924 harvest in Sokoto Province was reported as "excellent," although abruptly occurring freshets caused localized failures of rice crops. The 1925 province report indicated that despite poor weather early in the farming season, late rains resulted in "bumper crops" everywhere but in northwestern Sokoto (NAK SOKPROF. 57. 234 and SOKPROF. 60. 474). Much like Arnett in 1913, Mr. G.A. Woodhouse, the new Resident at Sokoto, appears to have underestimated the gravity of harvest failures in 1926. Sokoto rainfall totalled 531mm, the lowest since 1917. The 60mm recorded in July is the worst in the entire 1916-1989 monthly rainfall series. Woodhouse remarked on the general inadequacy of this rainy season, adding that deficiencies were most acute in the west and northwest of the province. Nevertheless millet was reported "up to average." Only those farmers who failed to plant after the copious rainfall of April 30 had "practically no gero." But sorghum was scarce around Sokoto. Although grain reserves were supposedly being augmented to guard against potential shortages, Woodhouse concluded that, with the exception of a few far below iii- 121 average areas, grain reserves from the "last three fat years" existed "everywhere in the productive area" (NAK SOKPROF. 64. 229). In his 1927 Sokoto Province report Resident G.W. Webster went into considerable detail about that year’s famine, which affected most of Northern Nigeria (NAK SNP. 17. 6863). Webster had realized that as a result of "very patchy" 1926 harvests, food shortage was "almost universal" in the province. He maintained that the farmers’ own grain stores together with the Native Authority’s increased local reserves would have been sufficient to prevent real hardship but for large scale hoarding and profiteering. Efforts to commandeer hoarded grain, and to prohibit falle loans, were reportedly successful. Prices fell immediately, and it was said that everyone was able to purchase food. But a decrease of 57,000 cattle, attributed largely to distress slaughtering, was recorded in Sokoto division. In the western and southern divisions of Argungu and Gwandu, slight increases in "normal" Native Authority grain reserves and restrictions on the export of foodstuffs limited famine to a "few small areas." A rinderpest epidemic and the "unsatisfactory conditions" of food shortage presumably diverted herders from these divisions. As a result, 16,911 fewer cattle were enumerated in 1927 than in the previous year, and jangali revenues were reduced commensurately. On the other hand, the tax assessment for the Zuru area increased from £4,643 in 1926 to £5,245 in 1927. The 122 increase was attributed to immigration from regions to the north, including Gwandu, and to additions to village livestock holdings. Despite "bumper crops of most kinds in all areas" in 1927, food prices did not fall after the harvest. Webster ascribed this to improved transportation infrastructures, including the rail line. The efforts of the colonial government to mitigate the 1927 famine in the northern provinces have been assessed as woefully inadequate. Adequate provisioning of the millions affected by famine would have entailed transportation costs well beyond the fiscal means of the British administration (Shenton and Watts, 1979). Yet, considering the proportionately very small number of people who benefited from the famine relief scheme, expenditures were, according to Watts (1983), absurdly high. The Report 9e Famine Relief ie the Northern Provinces i 1927 (NAK SOKPROF. 74. 416) indicated that the sum allocated to Sokoto Province, £10,029, was second only to Kano. But aid was lacking in areas that, even from the colonialists' perspective of reduced tax revenues, were worst hit. In Gwandu and Argungu divisions, from which households reportedly emigrated during the famine and in which wholesale decreases in cattle populations were recorded, expenditures were "nil" (NAK SNP. 17. 6863). The report on famine relief is interesting in several respects. Apparently for the first time, colonial administrators appreciated the disastrous potential of two 123 consecutive harvest failures. The report’s authors also realized that some imported foods, such as rice, may not be appropriate for famine relief. Deaths from consuming unripe millet were acknowledged as having occurred in Sokoto Province in 1927 as well as during previous famines. The report offers a hint of the possible consequences of famine in Northern Nigeria for the people of Niger. A famine relief work in Gumel Emirate (northern Kano Province) was not well attended because many people crossed the border into Niger to live temporarily with relatives. Lower grain prices were reportedly the main inducement for their migration. Even though 1926-27 is included in Painter’s (1986) and Delehanty's (1988) Chronologies of famine, Watts (1983) assessed harvests in Niger as having been fair, and distress migration to Nigeria as minimal. Perhaps harvests were also adequate in the study area. No informants specifically recalled this famine, and apparently no name had been attached to it. One elder in Makwala stated that "Tashi ta Tikittibale," a distress migration ee meeee from villages in the northeastern quadrant of the study area to a locality near Shinkafe (13°05’N, 6°31’E), occurred two or three years before the Sultan of Sokoto Tambari’s flight to Niger in 1931. One Amarawa elder recalled that this exodus took place during Tudu Malami’s tenure as village head (c. 1927- c. 1934). Most elders, however, associated the migration to Tikittibale with several subsequent colonial-era famines. 124 If Tashi ta Tikittibale occurred in 1927, perhaps it was the first in a chain of distress migrations. The Tudun Gudali Mai Ruwa nuclear settlement was founded during Muhammadu Tambari’s reign as Sultan of Sokoto (1924-31). The exact date could not be determined because of disagreements concerning the tenure of the first two village heads. The first settlers emigrated from unspecified Konnawa villages in Niger to escape French forced labor. Their sons could not recall whether their migration was also compelled by food shortages. Shortly after Tudun Gudali Mai Ruwa was established, some of the local Fulani began their northward wet season transhumance. According to one Fulani elder, this was necessary and possible because the expansion of land under cultivation reduced pasture resources, and the Tuareg ("Asbenawa") could no longer prevent excursions into the Sahara. In 1928 the road from Gwadabawa to Kalmalo was improved so as to be motorable, and people in the study area saw motor vehicles for the first time. An Amarawa informant born at about the time of this event has as his lakabi ("nickname") "me; meee," implying that his birth was associated with the appearance of an automobile. With few exceptions 1928 harvests were reported as "excellent." Abnormally heavy August rainfall caused localized failures of millet crops in southern Sokoto Province (NAK SNP. 17. 9270). The following year crops of all kinds also were judged to have been "excellent" 125 throughout the province, and grain prices decreased markedly relative to 1928 post-harvest levels. The 1930—40 Period The famine of 1931-32 is recalled throughout the study area as Shekara Kyamro. This name connotes that harvest failures on upland farms were so complete that only the woody grass kyamro survived. Informants consistently recounted that this year (Shekara) was a "yunwa fara" (a "hunger" caused by locust), that swarms emanated from the west, and that it occurred during Hassan dan Ma’azu's reign as Sultan of Sokoto (1931-38). A minority stated confidently that Shekara Kyamro was either the year of or one year after his turbaning (coronation). The published accounts of this famine focus on western Niger (Salifou, 1975; Fuglestad, 1974, 1983). Archival records for Sokoto concentrate on areas in the south of the province. That locust attacks resulted in major harvest failures and severe hardship in northern Sokoto apparently had never been established. The 1930 Sokoto Province report by Mr. H. F. Blackwell recounted that the first flights of locust had been sighted in September 1929 near Kwarre (c. 15 kilometers north of Sokoto) travelling west. In November another flight was observed near Argungu. Between April 14 and May 4, 1930, swarms appeared in Jega, Birnin Kebbi, Gusau, and Sokoto. Blackwell concluded that harvests were good or very good overall despite continued locust infestations during the 126 rainy season, which in Sokoto yielded 744 well distributed millimeters. But he also cited what amounts to a rather long list of exceptions. Crops were damaged in Anka, Bukwium, Gummi, Kebbi, Kwiambana, and Zurmi districts. Famine conditions prevailed in parts of the Ka River valley, as well as in the Zuru area, where the 1929 harvest had been poor and "much" of the 1930 sorghum crop and 70 percent of the eeee crop had been demolished. Northern Argungu was seriously affected, but "abundant" crops in the rest of the division would more than compensate for this shortfall. Wet season crops at the Kwarre experimental farm had not been "very successful" because of locust infestations and flooding (NAK SNP. 14. 818). The province report for 1931 by Mr. C. A. Woodhouse is not nearly as contradictory (NAK SNP. 16. 670). The monsoon arrived tardily to northern Sokoto, but otherwise rainfall was well timed. All parts of Sokoto Province reported abundant harvests, and Woodhouse saw no potential for food shortage during 1931—32. A vigorous anti—locust campaign had been carried out during the early rainy season. Locust were attracted and killed by, inter alia, millet bran and corn cobs laced with sodium arsenite. (Ferric hydroxide, an antidote, was kept on hand to treat accidental poisoning.) Because too many livestock died from eating the toxic "bait," this method was abandoned in favor of trenching, Africans’ traditional strategy for combating locust attacks 127 (NAK SOKPROF. 283. 376B). Locust swarms did "remarkably little" damage to crops (NAK SNP. 16. 670). Pre-harvest hunger had necessitated the opening of famine relief works in the Zuru area. Despite this assistance, farmers were too weak to cultivate during the 1931 rains and famine dragged on. High food prices obtained in parts of the province, while grain had to be distributed in Bukwium, Anka, northern Argungu, and Tangaza during the 1931 "hungry season." In Woodhouse’s view the 1931 harvest ended the few localized food shortages. The Zuru area famine was the only possible exception. But in the 1932 Sokoto Province report prepared by Acting Resident Mr. L. S. Ward, one finds strong evidence that the British administration seriously misjudged the magnitude of 1931 harvest failures and the extent of famine (NAK SNP. 18. 921). In 1932 farmers were supposedly so worried by the prospect of a third consecutive year of locust attacks and harvest failures that they greatly expanded the amount of land devoted to peanuts. More than 14,000 tons were exported from the rail heads at Gusau and Kaura Namoda, an increase of 6,600 tons over 1931. Cattle and sheep were "absurdly cheap" before the 1932 harvest. Tax remissions, ranging from -0.9 percent in Gwadabawa to -28.5 percent in Kwiambana, were necessary in 32 of Sokoto Division’s 48 districts. Taxes had to be reduced by nearly 10 percent in Argungu division despite a recorded increase in population of 1,400. The assessment for the Yelwa area 128 was reduced by 13 percent, but in the area around Zuru, where farmers had been too weak to cultivate, the remission amounted to only two percent (NAK SOKPROF. 2127. S. 2403). Distress migrations were reported, but only from Gwandu division, the Zuru area, and the area south of Yelwa. Immigration from "French country" was noted consistently during the 1930—32 famine episode. Argungu, for which was recorded a decrease in population in 1931 and then an increase in 1932, was reportedly the principal destination for "starving Zabermawa and Arawa" (NAK SNP. 18. 921). But by this time the British had decided that refugees should be welcomed, and to the extent possible, farmland allocated to them. Ward related that what was first thought to be an influx of Nigerien refugees into Gwadabawa district was in fact a temporary flight from French military conscription. This information was not totally correct. Although apparently no new settlements were established in the study area by refugees from the 1930-32 famine, Tozai elders recollected that about ten Zarma families joined their village at this time. Moreover, a large village south of the study area was reportedly founded in the early 19305 by refugees from Niger. Its people speak Zarma to this day. The famine apparently did end after the 1932 harvest, which, owing to good rainfall (766mm in Sokoto), the nearly complete absence of locust and freedom from epidemics, was judged to have been "excellent," "abundant," and "very 129 satisfactory." In Argungu, where food shortages and asset liquidation were acknowledged to have occurred during the 1932 hungry season, food prices fell to levels 25-30 percent below those of the preceding year (NAK SNP. 18. 921). The 1932 harvest marked the beginning of what may generally be considered a famine-free decade. The Sokoto Province reports for these years indicate fluctuations in harvest quality, but no food shortages comparable to those of the 19105, 1920s, and the 1930-32 interval. With some exceptions, the province administration recorded gains in human and livestock populations. In 1933, John H. Carrow, "Commander," reported that locust again appeared throughout Sokoto Province, but damage was slight and harvests "well up to average." Extremely low cereal prices prevailed in many parts of the province that year (NAK SOKPROF. 598. 3133). Annual rainfall in Sokoto was 805mm. The 414mm recorded in Sokoto in 1934 was rather well distributed, and although Carrow reported the fifth annual locust visitation, damage was minimal and a "fairly good" harvest was had. Cereal prices, however, "rallied." The low rainfall total was responsible for the desiccation of wells and the poor condition of trees (NAK SOKPROF. 705. 3554). The 767mm that fell in 1935 in Sokoto redressed this problem, and although yields were somewhat reduced by saturation, a "good average" harvest was realized. The effects of famine had "disappeared" in the Zuru area, and 130 the tax assessment was increased accordingly (NAK SOKPROF. 722 . S757) . Rainfall in 1936 was exceptionally heavy (1025mm in Sokoto). The 476mm recorded in August probably informed Carrow’s judgment that rainfall had been poorly distributed "from the farmer’s point of view." Rice crops failed north of Sokoto and in the Birnin Kebbi area, but on the whole the harvest was deemed better than in either 1934 or 1935 (NAK SOKPROF. 883. 4464). Despite a deficient June (36mm), 1937 rainfall in Sokoto totaled 738mm, and good food crops were produced. Harvests in the northernmost districts were surprisingly good (NAK SNP. 17. 29664). In his 1938 province report, Carrow did not comment directly on the quality of food crop harvests. He instead lamented the low prices of cotton and peanuts, which had reached their nadir in October 1938. A September 1938 report from another archival source indicated that the millet harvest had been good, and the prospects for sorghum "excellent"1 (NAK SOKPROF. 456. 1964). But during the bazara season of 1938, a devastating epidemic of cerebro- spinal meningitis killed an estimated 37,000 people in Sokoto division and 8,000 in Argungu (NAK SOKPROF. 1023. 5104). Certainly a very large proportion of the victims were children. Mr. R. D. Ross, the new Resident, reported 1939 harvests as "good" and "excellent," while the rise in the prices of export crops was "encouraging." The far northern 131 districts of Gwadabawa, Gada, and Tangaza had enjoyed their best millet crop since 1932. No epidemics were reported. A population decrease of nearly 13,000 in the province was attributed to a better census and the emigration of "semi- nomadic settlers near the French border" (NAK SOKPROF. 1142. 5452). Ross considered 1940 rainfall, which totaled 651mm, to be below average. Excepting failures of rice crops, food crop production was "fair" (NAK SNP. 17. 33155). Colonial officers’ reports during the years 1933-40 convey a sense of urgency about environmental degradation in the northern tier of Sokoto Province districts. Stebbing’s (1935) alarmist view of a relentlessly advancing Sahara desert probably inspired these concerns. In 1936 Carrow described Gwadabawa as a once-fertile and productive region now "embarrassed" by soil erosion and deforestation. Only 100 bundles of grain could be produced on farms that formerly yielded 400. Men had to embark on long—distance dry season migrations in order to feed their families and pay taxes, and a "certain degree" of permanent emigration had been noted (NAK SOKPROF. 883. 4464). The report of the Anglo-French forestry commission, submitted in 1937, to some extent allayed fears of naturally occurring desert encroachment. Its authors instead fixed blame for land degradation on "harmful and wasteful" farming practices (NAK SNP. 17. 29664), a view colonial administrators would also modify in later years. The 1938 province report related that efforts to institute "mixed 132 farming" (animal traction, chiefly with oxen) in the northern districts had been abandoned. Village firewood plantations were prescribed for Gwadabawa district. Dung and crop residues could then be left on farms to enhance soil fertility instead of being used for cooking fires (NAK SOKPROF. 1023. 5104). Additional administration prescriptions for improving the allegedly prodigal agronomic practices included delaying the first two farming operations (sassabe and Lerbe) and preserving undergrowth and termite mounds. The abundant 1939 harvests in the northern districts were attributed in part to these and other measures (NAK SOKPROF. 1142. 5452). After the very mediocre 1940 rainy season, however, Resident Ross seems to have come around to a more moderate and, in light of Tucker ee e1.’s (1991) recent work, more accurate view of environmental degradation and dryland ecology. "Natural regeneration on light sandy soils" he wrote, "can only occur in years when rainfall is unusually heavy and well distributed" (NAK SNP. 17. 33155) . A major and related concern of the colonial administration was immigration into frontier districts from Niger. In 1933, for example, a "considerable number" of recently settled Nigeriens were reported in Gwadabawa district. The administration suggested rather benignly that their settlements be amalgamated in a new "village area" (NAK SOKPROF. 598. 3133). (This was not accomplished until two decades later, and then only partially.) By 1936 133 official opinion had turned against the immigrants, "whose wasteful and ignorant methods of farming lead to excessive deforestation and degradation of the soil (NAK SOKPROF. 883. 4464)." What attracted Nigerien farming families to Gwadabawa and the study area was, according to the administration, the large number of deep wells sunk by the geological department. These humanitarian efforts had therefore inadvertently exacerbated deforestation and soil erosion by attracting too many people. Study area elders offered strikingly different accounts of immigration and the agro-ecological situation during the reign of Hassan dan Ma’azu as Sultan of Sokoto (1931-38). Nigeriens migrated to the study area not on account of the wells, but to escape French forced labor. They established hamlets in the hinterlands of existing villages, including Dan Tudu near Amarawa. Abrugel, Zango, and Lisawol were other hamlets founded by these refugees, as well as the northern Hausa satellite settlement of Tudun Gudali Mai Ruwa. Probably the Hausa hamlet to the south of Tudun Gudali Mai Ruwa’s nuclear settlement, and Gidan Alkasim's Zangon Yamma were also established by refugees at this time. Examples of the immigrants’ villages of origin are Magozo, Shiyya, Malawa (near Giden Ider), and Takar. Residents of the host villages by and large accepted the migrants without consternation. This was true even in the case of Dan Tudu, whose founders constructed compounds on an established farm. The extent of uncultivated bushlands and fallowed land was 134 greatly reduced by the immigration and resettlement, but wells avowedly never dried up. That the 1932-40 period was free from famine is consonant with study area informants’ perceptions. Several bounteous harvests in the years following Shekara Kyamro were achieved without alterations of agronomic practices. The Birnin Konni rainfall series, which begins in 1933, indicates that rainfall was adequate and in some years abundant (Appendix 3). Despite a relatively low annual total, 1934 rainfall was fairly well distributed. Perhaps the monthly totals for July 1935 and August 1938 were somewhat deficient. The rainfall record, rather than British tutelage, probably explains the great harvests of 1937 and 1939. The rains of both years began early, ended late, and were well distributed. The 1941-51 Period The famine of 1942-43 was the third widespread excruciating colonial-era famine. It will be recalled that 1940 harvests were judged to have been only "fair." According to Resident Carrow, harvests in 1941 were below average. Although annual rainfall was adequate (580mm in Sokoto), its distribution reportedly was not favorable to agriculture. The below average total of 433mm for Birnin Konni seems to have been fairly well distributed (Appendix 3). Food prices had risen in Sokoto Province, but the large sums paid to farmers "on military measures" created "relative prosperity" (NAK SNP. 33. 155). 135 Rainfall in 1942 was reportedly deficient for the third consecutive year. In Sokoto 499mm were recorded. Monthly rainfall totals at Birnin Konni were slightly lower than the previous year’s. Elders recall that a very late replanting was necessary. As a result millet either failed or was "dwarfed," but the sorghum harvest was not too bad. Carrow reported that millet crops had been "light" in the north and northeast of Sokoto division. On the whole, the province produced a "good average crop" of sorghum, and cowpeas "yielded well everywhere." In 1942 the government’s requirements for grain and peanuts could no longer be satisfied through purchases on the open market. Requisitioning therefore became necessary. By the end of the year, a total of 427 tons of peanuts, 3,550 tons of corn, and 3,400 tons of millet had been amassed in Gusau. Although food prices "soared," Carrow acknowledged "only a slight shortage in some of the commoner foodstuffs." A sharp increase in the incidence of smallpox was attributed to the succession of droughty years2 (NAK SOKPROF. 1262. 6014). In his 1943 Sokoto Province report, Acting Resident Captain H. B. Leonard made no mention of famine despite the correspondence on the plight of refugees cited in chapter three. The only hint of hardship was that the bumper crop of 1943 enabled farmers to replenish granaries depleted by military requisitions and three previous poor harvests. Money was "pouring into peasants’ pockets," and taxes had 136 been collected "in record time." The deleterious effects of guinea worm had been noted in the northern districts. In Kaura Namoda, thirteen women had set up an industry for manufacturing cassava flour, which was in good demand among the local people3 (NAK SOKPROF. 1284. 6110). The 1942-43 famine is known in the study area as Takesharbo or Mazarkwaila. This latter name means "brown sugar;" its significance could not be ascertained. The former term suggests that the famine caused unripened millet to be cut down (sharba) and brought home. Takesharbo is most frequently associated with Fithitla ee Jamus, a Hausa pun used in referring to World War II.4 Some elders attributed this famine less to climatic misfortune than to government actions. The most pernicious and bitterly remembered of these was "aikin dole ee gudjiya," or the forced labor of peanuts. Because this crop was not cultivated locally, men were forced either to trek or to travel by donkey to peanut-producing areas. Talata Mafara, approximately 150 kilometers southeast of the study area, was the most frequently cited destination. Here men had to seek wage labor in order to be able to purchase their quota, which one elder recalled was two 20-liter tins of shelled peanuts. Informants suggested that cash compensation was paid, but that traditional authorities "intercepted" these sums. The conjuncture that led to the famine Takesharbo therefore involved forced migrations of 300 kilometers, the 137 diversion of men’s labor from their own farms, and a failed 1942 millet crop. The next famine to affect the study area has been named ghege after a local boxer who "conquered all." The precise year of Shago could not be determined. Informants’ recollections encompassed the period from three years to nine years after Takesharbo (i.e. 1946-52). Some elders stated that if Shago occurred, then they had forgotten it. This famine is most often described as a period of intense seasonal hunger that lasted from the second weeding (mei; mei) until the harvest. Archival sources recount wide variations in the quality of harvests during the years 1944- 51. These reports and the Birnin Konni rainfall series suggest that Shago occurred during either the 1946-47 or the 1950-52 period. The 1944 Sokoto Province report stated that despite localized harvest failures, millet and sorghum crops had been good and that only rice was in short supply (NAK SOKPROF. 1316. 6212). In Sokoto, the distribution of the 655mm of rain was less than ideal, as only 51mm were recorded in June. Although July was droughty in Birnin Konni, the 466mm recorded there was otherwise well distributed. The 1945 harvests were variable: millet was "fair," rice "poor," sorghum "good." "Abnormally heavy" rains (782mm in Sokoto) caused damage to buildings and bridges. The total for Birnin Konni was above average, but the 40mm 138 recorded in June suggests that the farming season began late. Acting Resident McCabe related that the peasants were happy because the end of the war marked the end of military conscription and government requisitioning (NAK SOKPROF. 1364 . 6319). The 1946-47 interval was the first documented example from Sokoto Province of a famine triggered by poorly timed and excessive rainfall. It was not to be the last. The 1946 total for Sokoto was 969mm, with 115mm falling in June, 196mm in July, 355mm in August, and 234mm in September. In Birnin Konni, the highest June total on record, 159mm, was followed by near—average totals in the succeeding three months. The number of rain days, however, was unusually high. July was punctuated by three four-day rainless intervals, but the longest break in August was between the 27th and 3lst (ORSTOM, 1976). Labor demands for weeding operations must have been extreme. Resident McCabe’s 1946 province report signalled major harvest failures, especially in low-lying areas. Food was reported to be in good supply, even though prices failed to fall after the 1946 harvest. McCabe attributed this to a lack of textiles and other consumer goods on which people could spend rising incomes. Migration to southeastern Sokoto Province from northern districts was described as a trend (NAK SNP. 17. 41874). The next year Senior Resident Mr. B. E. Sharwood-Smith reported that the 1946 harvest had been a failure (NAK SNP. 17. 43499). In the first months of 1947, grain prices 139 reached record high levels, and hardship was almost universal. Reports reached the administration that "whole communities were reduced to supplementing their meager stocks of farm produce by recourse to herbs, grasses and lily bulbs." Sharwood—Smith did not specify the areas in which this occurred, but food prices were highest in the western part of Sokoto Province. The young men trekking to seasonal migration destinations were reportedly "as numerous as ever." Population pressure was causing permanent migration to areas farther south. A second population "drift" was discernible from northern districts to the Zamfara Valley. The 1946 harvest failure and the famine that followed motivated farmers to plant more food crops and fewer cash crops. This was welcome in the short term as inflationary pressures were expected to be reduced as a result. Although the 598mm of rainfall recorded in Sokoto in 1947 were not well distributed (85mm in June, 86mm in July, 277mm in August, and 96mm in September), harvests were apparently adequate. Grain was "everywhere abundant" and prices fell accordingly (NAK SNP. 17. 43499). The 515mm in Birnin Konni seem to have been adequately distributed, with no major breaks apart from a nine-day rainless period from June 15th to the 24th (ORSTOM, 1976). At the end of 1948 Sharwood-Smith reported that food had been relatively plentiful "by Sokoto standards, which are not high." He had been told that a sharp price increase 140 at Gusau markets during the summer months could be attributed to a spate of heaving buying motivated by the sighting of a comet, "reputedly the inevitable harbinger of a bad harvest." Nevertheless, harvests overall apparently turned out to be good. The trade in livestock skins declined markedly from July to December. Healthy rises in cash crOp prices and good grain harvests obviated the need for farmers to raise money by selling their animals (NAK SNP. 17. 43499). The 690mm of rainfall in Sokoto in 1948 apparently did not cause serious crop damage. The monthly distribution was not very different from 1946 except in the case of September rainfall which, at 155mm, was 80mm less than the corresponding month of the famine-catalyzing year. Argungu, where 387mm of rainfall in August "perhaps" caused serious damage to millet crops, may have been an exception. An unspecified number of Zarma people who had settled in this division returned to western Niger. Their return migration was supposedly prompted by news of good harvests in their homelands (NAK SOKPROF. 1643. 7225/S.1). Only 481mm were measured in Birnin Konni during the 1948 rainy season. The monthly totals suggest that it was reasonably well distributed. Rainless periods, however, occurred during June 24-July 1, July 17-24, and August 16—25 (ORSTOM, 1976). A 1948 report from the "Experimental Unit" farm at Tarike (c. 15 kilometers south of Illela) indicated that no practical and economical means existed for improving soil fertility in Gwadabawa district. Fallowing, either 141 with or without manure, was effective, but high population densities (115,744 people on 1,018 square miles, i.e. 114/square mile) severely limited the potential for employing this strategy. The area was devoid of forest cover (NAK SOKPROF. 1177. 5590). Sharwood—Smith’s saturnine assessment was that Sokoto Province’s soils were steadily deteriorating as its population continually increased. In 1947 the enumerated population of the province exceeded two million for the first time (NAK SNP. 17. 43499). Birnin Konni rainfall in 1949 amounted to only 440mm. If farmers planted after the 39mm rain event on May 19 (ORSTOM, 1976), they probably had to plant again in mid or late June, or even during the first week of July. Annual rainfall in Sokoto was 493mm (monthly totals are not available). According to Sharwood-Smith, nature had been "wayward," and famine at one point seemed imminent. The sorghum crop was poor in Sokoto Province, but millet crops "survived" and in the western districts were reported as "well above average." In the Senior Resident’s view the threat of famine had therefore receded at least until May or June of 1950. Censuses found that the province’s population had grown by only 3,000 instead of the "usual" annual increase of 20,000. A disastrous cerebro-spinal meningitis epidemic had claimed an estimated 15,000 lives. Only five medical officers were stationed in the province, yet sulfa drugs reportedly saved two of every three infected individuals. The epidemic seems to have evoked empathy in 142 Sharwood-Smith, who argued that on a proportional basis it had had the same impact on Sokoto Province as World War II had had on Great Britain. Another insightful observation concerning the Sokoto Close Settled Zone presaged Esther Boserup’s frequently cited theory: ". . . the standard of farming practices improves in direct relation to the pressure on the land . . . the peasants will improve their methods only when forced to do so." Despite these forced improvements, migration had been detected from densely settled zones to "more empty lands" (NAK SNP. 17. 45603). In his 1950 Sokoto Province report, Acting Resident Mr. A. T. Weatherhead wrote that the year had commenced with pestilence, and appeared for a time that it would end in famine (NAK SOKPROF. 1699- 7644). The pestilence was yet another cerebro—spinal meningitis epidemic, but as the province’s population reportedly increased by 21,000, its effects were judged to have been "far short" of the previous year’s outbreak. The foreboding time must have been the month of June, during which only 10mm of rainfall were recorded in Sokoto. But the rains rallied in July (149mm), became exceptionally heavy in August (443mm), and were abundant for September (193mm). The Sokoto annual total attained 853mm. Flooding in the Rima River valley caused a total failure of the rice crop, but other food crops apparently did well. Millet prices plummeted in October (Figures 4.2 and 4.3). 143 Birnin Konni received a well above average 742mm in 1950. The daily record again indicates that the definitive planting may have been delayed until the first week of July. Of the 44mm recorded in June, 32mm fell on the 5th of that month (ORSTOM, 1976). August rainfall was perhaps too abundant for some crops. A report telegraphed from the Gusau agriculture office in January 1951 indicated that food shortages were expected in the Isa and Sabon Birni areasS (c. 120 kilometers west-southwest of the study area) (NAK SOKPROF. 457. 1964). Apart from a low June total (62mm), the 590mm that fell in Sokoto in 1951 seem to have been adequately distributed. The Birnin Konni station recorded an annual total of 673mm. Again, grain crops may not have germinated successfully in June. Five rain events yielded 55mm in that month, but 31mm fell on the sixth and 13mm on the let. A dry spell then spanned the June 26—July 7 period (ORSTOM, 1976). Millet failed almost completely at the experimental farm in Tarike- Rainfall had been poor in July, at which time grain was reportedly unavailable in the countryside (NAK SOKPROF. 1733. 7778/S.1). (This farm, however, had been established on land that "native" farmers had abandoned because of its infertility; NAK SOKPROF. 1177. 5590.) A telegraphed report dated October 2, 1951, advised of partial local crop failures in northern areas (NAK SOKPROF. 457. 1964). In Argungu, an average millet crop was reaped amid local failures. Early planted varieties of sorghum did well, but 144 heavy late season rains reduced harvests of late planted varieties to average quantities. In some areas of Gwandu, crops were deemed to have been the best in many years, and were reported as "excellent" everywhere else in the division. Grain prices in Gusau and Sokoto fell during July—October 1951 (Figures 4.2 and 4.3). Their rather early rebound was thought to reflect the patchiness of the province harvest, although at least in Gusau, the disruption of rail transport and high export crop prices also played roles (NAK SOKPROF. 457. 1964). Over most of the first half of 1952, prices were higher than during the corresponding period of the previous year. The 1952—59 Period That the 19508 was a decade of abundant rainfall, and in retrospect, a climatic optimum, is manifest in numerous annual rainfall series. Nicholson’s (1989) and P. J. Lamb eh _T.’s (1986) rainfall departure indices for the entire Sahel and Sudan of northern Africa show a succession of well above average years. From 1952 through 1958, every annual total in Sokoto surpassed 700mm. In 1959, only 561mm were recorded, but 900mm fell in 1960. The pluviosity of the 19503 is also evident in the record of annual totals for Birnin Konni (Appendix 3). Many elders’ recollections of this period perhaps have been distorted by nostalgia. Jubilant shouts of "tuwon dawa" (sorghum paste) rang out across the upland farms, and yields were so copious that farmers could not harvest and store all of their crops. 145 SOKOTO -'- Sorghum --~ Millet 1950 1 950-54 Source: NAK SOKPROF. 457. 1964 Figure 4.2: Grain Prices in Sokoto City, 146 ~60 GUSAU -— Sorghum ~-»AMHet 1954 I 1951 l 1952 l 1953 1950 60— Figure 4.3: Grain Prices in Gusau, 1950-54 Source: NAK SOKPROF.457.1964 147 Yet in one year the capriciousness of the West African monsoon resulted in grievous and pervasive suffering. According to most elders, M222, the famine of 1953-54, was the most severe ever to strike the study area. Its name suggests a man warning his wife that he will divorce her if she persists in pressuring ($222) him to procure food. The 1953-54 famine was also widespread. It is known throughout Sokoto State as Muda. Although 1954 is included in Watts’s famine chronology for northern Katsina (Table 4.2), he described it elsewhere as a "food scare" and a "poor harvest" (Watts, 1983:346,568). Hill (1972), however, determined that harvest failure in 1953 led to the 1954 famine Uwar Sani in Batagarawa (only 25 kilometers from Watts’s study village). Mortimore’s (1989) chronology for northern Kano indicates that a famine occurred during the interval 1954-57. Oguntoyinbo (1981) ascribed famine in Kano and Borno to heavy rainfall in the years 1951-54, and Fuglestad (1983) stated that famine in Niger in 1953 or 1954 hey have been caused by too much rain.6 In Sokoto the 1952 rains began late (30mm in June) but were abundant during the three following months. The annual total was 873mm. The 990mm received in Birnin Konni is the highest in that station’s 59-year record. The monsoon was also tardy there, but totals for August and September are records (Appendix 3). This superabundance of rainfall did not trigger famine in 1952. The available evidence indicates that harvests were very good. Grain prices in 148 Sokoto fell to their lowest levels in the 1950-54 period (Figure 4.2). Higher post—harvest prices in Gusau were attributed to heavy rains restricting transport (Figure 4.3; NAK SOKPROF. 457. 1964) . The reason the 1953 rainy season was the catalyst of famine is difficult to determine from the monthly rainfall series. Sokoto received an annual total of 732mm, 140mm less than in 1952. The May total of 154mm was well above average. The 110mm for June, the 170mm for July, and the September total of 139mm could be considered good, but August’s 139mm was a bit droughty. Birnin Konni’s annual rainfall, 931mm, was 60mm less than the previous year’s total. May, June, and August totals were above average, but certainly not anomalous. The 299mm in July, however, may well have been excessive (Appendix 3). Study area farmers who departed from religious or teleological explanations of the cause of Muda stated that "ruwa sun hanna noma" ("rainfall prevented farming"). The daily rainfall series for 1953 suggests how this might have happened (Figure 4.4). Heavy rains in the first half of the farming season may have destroyed newly planted seeds and sprouting grain crops. According to Oguntoyimbo (1981)- heavy rainfall on the upland soils of Northern Nigeria results in the sandy stratum covering the seeds. Some farmers’ seed reserves were reportedly exhausted by numerous replantings- Heavy cloud cover and profuse weed growth 149 might also have contributed to harvest failures by retarding crops during critical developmental stages. Additional perspectives on the disastrous 1953 farming season are found in the available district and divisional reports (NAK SOKPROF. 1790. 8142/S.1). In Gusau, heavy early season rainfall caused flooding on fadama farms and extensive damage to dry season crops just before they were due to be harvested. Dampness during the period of early growth caused a "set back" to millet from which it never fully recovered. Sorghum, however, reportedly did recover and yields were above average in districts to the south of Gusau. In Gwandu division, very heavy early rains provoked fear of a "near famine." Millet failed nearly everywhere. Hardship was expected in areas near Birnin Kebbi. Resident Mr. K. P. Maddocks began his 1954 Sokoto Province report by relating that anxiety and hunger had caused a blight over the entire province during the first nine months of that year (NAK SOKPROF. 1843. 8318). The 1953 millet crop had been a failure. Food shortages had been acute and fear of famine nearly universal. Farmers had been too weak to cultivate, and there was much poverty and indebtedness. Three serious fights between Hausa farmers and Fulani herders had been made "more bitter by hunger." Censuses enumerated 64,460 fewer cattle than in 1953. Maddocks attributed this decrease to the Fulani having migrated to other provinces in search of less expensive grain. Although distress slaughtering was not mentioned, 150 80 70‘ 60j 50- 404 30j 1953 Total - 931 .3mm Figure 4.4: 1953 Daily Rainfall Series for Birnin Konni, Niger Source: ORSTOM (1976) 151 some Fulani groups had been forced "by economic necessity" to take up farming. Jangali revenues diminished, and assessments for other taxes either remained constant or were reduced. Five hundred tons of grain were distributed by the Native Authority. Grain prices skyrocketed after November 1953 (Figures 4.2 and 4.3). Distress migrations had taken place before this exceptional post-harvest price increase. Maddocks reported that migration to the south and east of the province had been extremely heavy. The magnitude of famine-related migration from northern and central districts might be inferred from Table 4.3.7 Although morbidity and hospital check—ins had increased, no famine-related deaths were reported. A major change in the settlement pattern of the study area took place in 1954. Most of the former refugee inhabitants of the hamlets established in the 19308 were forced by the Sarkin Gobir of Gwadabawa to found the nuclear village of Makwalla. Four families from Dan Tudu were relocated there, while the other four built compounds on the southeastern periphery of Amarawa. A few elders attested informally that the stresses and strife of Muda were responsible for this resettlement. The traditional authorities reportedly permitted hamlets in the orbits of Tudun Gudali Mai Ruwa and Gidan Alkasim to remain at their present sites. The only information regarding the study area in Maddocks’s 1954 report was that the road from Gwadabawa to Kalmalo had been confirmed as "motorable." Table 4.3: Population Change in Eight Districts of Northern Sokoto Province, 1952-1954 M 1952* 1953+ 1954+ Kebbi 27,454 25,090 23,983 Gummi 69,120 51,037 47,274 Bukwium 42,918 32,586 33,192 Anka 23,303 17,772 17,767 Tangaza 57,145 49,961 47,297 Gwadabawa 158,545 ND 140,293 Gada 73,024 62,359 65,647 Goronyo 68,415 58,981 57,701 ' Census results + Tax Returns ND = No Data SOURCE: NAK Sokprof. 1778.8059 (1952-55) 153 According to informants, the road from Illela and Amarawa to Gwadabawa remained a dirt track. Although the 1954 monsoon arrived late in both Sokoto and Birnin Konni, heavy rains from July to September resulted in good yields of food crops (but poor export crop production). Sorghum was "excellent." Grain prices fell after July, but remained relatively high during the immediate post-harvest period (Figures 4.2 and 4.3). According to Maddocks, they were likely to remain at these levels because farmers were building up their grain reserves (NAK SOKPROF. 1843. 8318) . Study area elders’ testimony indicated that another famine-free decade commenced at the 1954 harvest. A 1959 famine in the "Sokoto-Illela area" during which "no one died" was not recalled (Watts, 1983:375,571). Elsewhere in the central Sahel, famines reportedly occurred in the latter half of the 1950s (Table 4.2). The majority of elders recounted good or excellent harvest over this period. A minority view is that major gaps in food self-sufficiency began in the aftermath of Muda. The 1960-67 Period Several events of ecological and economic significance occurred in the study area during the early 19605. Near the village of Dan Boka in the 1960-61 dry season, ground water spontaneously issued forth in abundance from wells and depressions. This water was reputed to have had miraculous healing powers, and people traveled great distances to 154 obtain a quantity. One or two years after this event the spring (marmaro) near the village of Lafani dried up. Also in the early 19605, the low-lying lands near Lakoda and Tozai became incapable of supporting dry season agriculture. The people of Tozai were forced intermittently to retrieve water from settlements 2—3.5 kilometers away. Skekara Yar Bari, "the year of leaving wilted young crops," was probably the 1963 census year. An early season storm resulted in very heavy precipitation. (There is no indication of such a rain event in the Birnin Konni daily rainfall series.) Most farmers planted, and grain crops subsequently began to develop. But the incipient crops withered in the absence of additional early rains. Traditional authorities and religious leaders urged farmers not to "kill" these withered crops when tilling and replanting. Informants averred that those who heeded this advice were rewarded with bounteous harvests; those who ignored it got average or poor yields. Reportedly a rough gravel road linking Gwadabawa and Illela was completed in 1964. Watts (1983bz37) traced the origin of the early 19705 famine to a "largely unperceived" series of drought years in the 19605. It may indeed be difficult to discern drought in the available rainfall series. But the 1965-66 famine Me; gehe is indelible in the memories of the people of the Illela area. The timing of this famine is unmistakable: hardship was already intense when the January 1966 coup and 155 the assassinations of Ahmadu Bello (the Sardauna of Sokoto) and Tafawa Balewa (the prime minister of Nigeria) were reported. The catalytic cause of famine is once again to be found in the temporal distribution of rainfall, not in the annual total. Birnin Konni received an above average 572mm in 1965 (Figure 4.5). But the June total was a paltry 26mm, all of which fell between the seventh and 17th of the month. Only 68mm were received in July; dry spells occurred from the second to the ninth, and again from the 11th through the 19th. August rainfall, 191mm, was about average and fairly well distributed. The 221mm September total was not unprecedented. It would seem that an above average September might have compensated for deficiencies earlier in the season. Instead, according to study area farmers, it caused a near—total failure of the sorghum crop. All ten September rain events occurred from the first to the 2lst, and four of these were exceptionally heavy downpours; 41mm on the fourth, 29mm on the ninth, 33mm on the 12th, and 50mm on the 16th (ORSTOM, 1976). Millet crops were average or below average as a consequence of the very droughty June and July. For some farmers poorly distributed rainfall in 1966 resulted in the second consecutive year of lean harvests and the continuation of hardship into 1967. Mai Zobe was certainly much more than a localized food shortage. Its name, "characterized by a ring," implies that hardship encompassed an area of such extent that no one 156 38 5 05.3... ”8.30m 88.2 5.58.2 5-82 .25 $-82 ...esmm 258.2 .28. 555 8.... 58.”. 5:58 u .58 55 .8 u .52 882 52-82 R2 0 m. .... ...... .\ (I)... .n . .. s. . .. .‘Jf I I‘- .45.- . .m. I ,‘1 (3' 4}:- .\ .1»: +: 5 -\'II-' I. \‘IA' 02W% 2;. I. - v-}n_ . 5‘ we? ... {6" .6 v2 . . .}%?f {‘55 l.- I I "9.. :: 4;?! 4;: 3:. 1;»:4 age ' 4 a? [1% .4 O was 2.. .2 < ‘1. . $53. é '- ' «rm.- A ,5:- 4? 'u'.‘ I {u (3;. IMIJ'.‘ AS'I 'c'b'c' 240" '4. -? 'I I 3" , ;.-r‘ ' I :/ (25- 349' a u \‘I V. .: :~ :- .a: {at Mme-w 5‘? .3. ' I.‘-' '-'-VII.‘.‘.‘ vt' EE Km .1. .90... momw .l -----51--.- 5:33 rEEom. 3': - ..E-fi'l --:-:-:- I? 157 EEomm u .90... 0:092 ..mm Tamar Ow< 0 560mm u .90... mum F mum. w 4’ - -:-:r- I]. <3. . :- .l .‘ .=: ”’24: 4:99:44 I I. I . - 0 .‘II ”x I" #4! $9 ': . ’ I '.'fl.'h'f ._ /. V’ 'C‘ I... xv .4 ... . . . - . :45: > .1’ - vm 5‘22}! ' 792': 4;} cf; :pn :I ' '2.- (I. ' 1'39? '1 . 3' VI nr. 193- :i: 1”. u . .... EEmNm u .90... .82. 65.2. 828m 200.2 25:05. 5-89 0:0 ..K-Km. .359. 25:05. .58. 555 no... 050.”. 5?} .44“ 3:?» ' ‘IA-‘u’ a?" ..,v ’- u . . .45., i- E aims-<- 0' -. I"? . . 249' < . ‘34-. -.;.:.r.'.'.- A1? I?) 6- EEmmm u .90... Km? EEom .. H E602 -nocuiu---nL.I EEOmP -EEoom cou. stuu 60 ass maj 196 for was Bat acc thi tha Cas 158 could escape. Buzu agro-pastoralists interviewed in the study area attested that Mai Zobe was severe in villages 50- 60 kilometers north of Illela. USAID’s first disaster assistance to Niger was undertaken in 1966 in response to major harvest failures (USAID, 1992). The years 1964—65 and 1967-68 are included in Mortimore’s (1989) famine chronology for northern Kano State. Hill (1972:7) related that grain was "very scarce" during September and October 1967 in Batagarawa. The 1968-80 Period For hundreds of Sahelian localities, a diachronic account of agro-ecological conditions and famine would at this point begin to address the deterioration in climate that culminated in the 1972-74 calamity. This is not the case in the study area for two principal reasons. First, monthly rainfall series for Birnin Konni strongly suggest that climate did not deteriorate until at least 1971. Although June in 1967, 1969, and 1970 could be considered deficient, all other rainy season months during the period 1967-70 approached or exceeded the averages. Second, study area informants are nearly equally divided as to whether famine occurred in the early 19705. This was one of the most surprising findings of field research. All of Nigeria north of the 12th parallel was declared a disaster area by the federal government in 1973. Reports summarized by Van Apeldoorn (1978, 1981) recounted near-total harvest failures and distress migration in northern Sokoto and the study are alt mi. N5 dr in a}: Se T1 t1 159 area. Grain prices in Sokoto were reportedly higher than almost anywhere else in Northern Nigeria. By April 1973, millet had risen to N200/ton from a pre-drought price of HBO/ton (Watts, 1983). Prothero (1974) identified Gwadabawa District as one of the areas most severely affected by drought. The series of annual totals for Birnin Konni do indicate that drought prevailed during the early 19705 (Appendix 3). The 1971 monsoon arrived rather tardily; only above—average rainfall in August and a not-too-droughty September may have salvaged the farming season (Figure 4.6). The 1972 total, 329 mm, is the second lowest on record, with the worst July, the fifth worst August, and the sixth worst September in the 1933-91 series. Although August, September, and especially July rainfall was greater in 1973 than in 1972, the 290 mm annual total is the lowest ever recorded. Despite these exceptional rainfall statistics, and despite the published accounts of harvest failure and famine in and near Illela LGA, about half of the study area informants attested that there was no early 19705 famine. The timing is again unmistakable. The Kalmalo irrigation scheme was opened in 1972. The road from Gwadabawa to Amarawa and Illela was being tarred in 1972 and 1973. In January 1973 a currency change operation was initiated, and in the following months people began seeing Naira (the new unit of Nigerian currency) for the first time. Later that 160 year, a national census was conducted. Traditional authorities, including the Sarkin Gobir of Gwadabawa, were deposed for allegedly mishandling food relief. Many Amarawa informants, however, attributed these government interventions to traditional authorities’ failure to support them following their bloody conflict with the "Dossawa" Fulani. The president of Nigeria, Yakuba Gowon, visited the study area in early 1974, when famine supposedly was nearing its peak of intensity. The only coup in the history of Niger took place shortly thereafter. Group interview participants in Araba related that "akwai alama yunwa, amma da dama" ("there were signs of famine, but moderate"). No name had been given to this episode because there was no real hardship. Elders in Rungumawa stated that food was expensive, but neither drought nor hunger was bad. No famine occurred in Gaeti or Lafani, according to these villages’ elders. In Lakoda there was "wahalla" (hardship), but people had been able to stock grains and otherwise prepare for it through their activities at "wurin bida" (seasonal migration destinations). Groups of elders in the Tudun Gudali Mai Ruwa nuclear settlement and in Amarawa also asserted that since there was no hardship, there was no name for the 1972- 74 interval. A frequent general assessment was "yunwa Franci ta" (a famine of Niger). In the study area, "an yi hatsi" (millet was produced). Government grain distributions were seen as "help," not famine relief. They 1:] Cl C( re ti 01 W} ir CC me 161 could have gotten along without them. A 1974 UNICEF report (in Campbell, 1977:199) indicating that harvests in the Birnin Konni area "remained tolerable" through 1973 accords with these elders’ recollections. Other groups of informants, and approximately half of the individual informants in Amarawa and Tudun Gudali Mai Ruwa, refer to 1972-74 as Mhhe gehe ("we know"). There are two interpretations of this term: "we know hunger is coming," and "we know hunger is coming and we had best prepare for it." But in Tozai, 1972-74 is remembered unequivocally as famine. Here its inexplicable name is Tie; Taralle. Shekara Karmami, "the year of stunted crops," is a term applied either to 1972 or 1973 by some informants in the Gidan Alkasim hamlets and in the southern Hausa hamlet of Tudun Gudali Mai Ruwa. People in these settlements also know the name Muna Sane. That some study area households were not capable of preparing for hunger despite "knowing" of its coming will be demonstrated in chapter six. In the study area and throughout the Sahel, the relatively good rains of 1974 resulted in good harvests and the end of famine. Informants did not substantiate a report of major harvest failures that year in the study area (Van Apeldoorn, 1978, 1981). The 1975—80 period seems on the whole to have been favorable for agriculture. Annual totals in Birnin Konni were all higher than 400mm. In three consecutive years (1976-78) rainfall was above the 1933-91 mean (Appendix 3). June rainfall was abundant in several 162 .009. 06.5... ”0058 0:002 25:05. 5-009 0:0 60-009 ...0.:.0m 25:22 .::0v. :.:..m K... 050.”. EEomm u .90... EEmmm u .90... EEmmm u .90... EEhom u .90... 0:00.). Pam—-009. mmmw vmmw mmmp Om<......2 Ow< ...2< m<0 .2< Om<... :- I . . -.~:- 4*. ‘1' .. I 2'29} -' "XII". {'59 .4 3": I. 4 '9 '.‘ :2 ,ss\ o I. 'u'v‘. my: : - “$9 .$ . xix-xv- M <0 I. I, '3” -: I‘I: ' v.»- 5:.- 2.“: I I i i E E o In .43.. .Eifi iii-$51: 11'? . ...... w. ”M ... .\ ... .I 9.? {A ‘t" n. I 45:45::- 2° a" 0.. e- .3832: fl.‘ 4 . -------.-..EE:2 an ex Cc CL Wj 163 years, and only once (1975) was the total below 40mm. July and August, the two main rainy season months, had with one exception (July 1975) totals exceeding 100mm. The 1980-85 Period A run of five below average years commenced in 1981.8 That year was the first since 1973 to have an August total below 100mm. July 1982 was below 100mm, and September rainfall was the lowest since 1933. The first signs of famine were apparent to study area informants when Shehu Shagari began his second term as President of Nigeria in August, 1983. Rainfall in June (22mm) and August (72mm) of that year had been extremely poor, and partial harvest failures occurred (Figure 4.7). The 1984 rainy season apparently began well, but rainfall in July and August amounted to less than half of the mean values. Harvests were either partial or total failures. The interval 1983—85 is known throughout Sokoto State as Buhariyya because it corresponded to Muhammadu Buhari’s tenure as Nigeria’s military head of state. Buhariyya is described almost without exception as XEQEQ (famine). Shekara dan Buhari (the year of hardship and hunger associated with Buhari) is used to refer to 0. August 1984 - 0. August 1985, the period of greatest suffering. The hardship stemming from an acute shortage of grain was compounded by two adverse developments. First, the rapid currency change operation in April 1984 left many people with worthless money. Second, the Nigerian government cl tc Ge d. ml 0) 0) 164 closed the country’s land borders, and commerce at border markets such as Illela’s was curtailed as a result. Lake Kalmalo dried up during Buhariyya, reportedly because a small dam upstream from it in Niger had been closed. Water ceased to occupy the tiny lake and stream bed to the north and east of Amarawa. Some of the wells in Gaeti, and nearly all of Lakoda’s wells, became desiccated in 1983 or 1984. The year 1985 was very good for study area farmers despite an annual total less than either 1983 or 1984. The monthly records show well below average rainfall in June and July, but near average August and September totals. A daily rainfall series would probably demonstrate that rain events resulting in modest precipitation occurred at very favorable intervals. In any case harvests were good or excellent according to informants. There is nearly universal agreement that famine had ended by the time Ibrahim Babangida became the President of Nigeria on August 28, 1985. Summary Significant parallels and differences exist between the chronology of twentieth century famines in the Illela area and those for other regions of the central Sahel (Table 4.2). A near turn-of-the-century famine in western Niger and Katsina/Daura also occurred in the Illela area, as did the Sahel-wide famines of 1913—14 and 1942-43. In contrast, famines during the early 19205 were recalled by only two 91" de' th Th we 11 P1 he Ci 165 groups of elders, and in 1926—27, famine may not have developed anywhere in this part of northern Sokoto, despite the severity of its impacts elsewhere in Northern Nigerian. The locust-induced famine of 1931-32, included in the western Niger chronology, also caused severe hardship in Illela-area villages and in many other parts of Sokoto Province. A famine that apparently resulted in relatively minor hardship may have occurred during 1946—47, or, as was the case in regions to the east, during 1950—51. The 1953-54 famine was not only extremely severe but also regional in scope. So too was the 1965-67 famine. Evidence that the impacts of the mid-19805 famine were severe and widespread is also incontrovertible. That the early 19705 famine reportedly had uneven impacts is important to assessments of vulnerability on both regional and local scales. The Illela-Birnin Konni area apparently escaped severe hardship, while people in regions not far to the east and north experienced often intense deprivation. In some Illela-area settlements, however, this famine is remembered clearly, while groups and individuals elsewhere attested that famine did not occur. Elders in Tozai, a water deficit village, specified without hesitation or equivocation the famine ’Yar Taralle as having occurred in the early 19705. The possibility that lack of a perennial source of domestic water and dry season farmland 166 contributes to heightened vulnerability to famine will be examined further in the following chapters. Although informants’ memories sometimes may be imprecise and archival sources inaccurate, the available evidence suggests that spatial variations in vulnerability may have been considerable during colonial times. The 1926- 27 famine could have been similar to the famine of the early 19705. It was not recollected in the Illela locality, but appears in all of the Chronologies in Table 4.2. According to archival documents, early 19205 famines in Sokoto Province were most severe in the vicinity of Kwarre and in areas farther west. Within the Illela area, their impacts may also have been uneven. For the most part, the Illela area also may have escaped famine in 1950-51, even though reports from the experimental farm located just 15 kilometers to the south indicated a near-total unavailability of grain in the surrounding countryside. Famine induced by excessive rainfall in 1946 was severe in some unspecified parts of Sokoto Province, but apparently not so in the Illela area. This chapter has illustrated with concrete examples the validity of the axiom that the temporal distribution of rain events is more important than annual or even monthly totals. At least three famines were triggered by poorly timed and excessive rainfall. In 1965, low early season rainfall and torrential rainfall in September caused crop failures that led to famine in 1966. The l946-47'famine, and the famine 167 of 1953-54 — the most severe in the memory of most informants — were attributable to saturation. On the other hand, severe famine did not develop as a consequence of three ostensible drought years. Daily rainfall series, were they available, might partly explain why the Illela area escaped major harvest failures and intense famine in the early 19705, and why 1985 harvests were good despite an annual total less than either 1983 or 1984- Desiccation in the Illela area apparently did not commence until the early 19605, when two villages lost their dry season farmland and some wells first dried up. This problem became more acute in the early 19805, when a stream ceased flowing and wells in Lakoda and other Illela-area settlements no longer provided water. Other environmental changes were attributable in part to immigration. Nigerien refugees from French forced labor and famine increased population densities during the twentieth century. Fallowing became impossible, and Fulani herds could no longer pass the rainy season in the Illela area. Several archival documents lend support to the proposition that population movements result in the spread of famine across space. Arnett’s assertion in 1914 that refugees created major demands on Sokoto Province’s food supply, thereby perpetuating famine, may well be correct. The movement during 1927 of people from Northern Nigerian border communities to Niger in response to lower food prices 168 suggests that Cutler’s (1984) "ripple effect" occurred in West Africa in the colonial period. Inferences supporting the spread of famine are possible based on the 1943 correspondence on refugees and the 1952-54 population data. One report of return migration by refugees from the early 19305 famine was submitted in 1948. Empirical case-by-case documentation of return migration, permanent out-migration, the diffusion of famine, spatial variations in vulnerability to famine, and recovery from famine is presented in chapters six and seven. Chapter five develops additional and more contemporary contexts for the presentation of data on these phenomena. The chronic lack of self—sufficiency in staple foods and the considerable interannual variations in levels of food self- sufficiency require constant forms of coping. It is from these coping strategies that famine—coping strategies, including migration, arise. 169 NOTES FOR CHAPTER FOUR 1. The British had established a system for monitoring the progress of the agricultural season. At least from June 1938 to July 1939 (when the archival file ends), an African monitor was submitting monthly reports from northern Sokoto Province to the agricultural office in Sokoto. His reports were written in Hausa. Examples include: July, 1938, from Gwadabawa, "Daminarsu duka ta yi kyau" ("the rainy season overall is good"); and May 16, 1939, from Gwadabawa, Isa, and Sabon Birni, "Wasu sun yi shibka, wasu basu yi ba, gero shi ne alamar kyau" ("Some have planted, some have not, the millet is showing good signs"). (NAK SOKPROF. 456. 1964). Famine early warning systems would have benefited from such basic field reporting in June 1993, when the region north of Maradi was designated "slightly vulnerable" to famine when in fact thousands of people were suffering unmitigated hardship. A compelling debate could be provoked by the question "Were British colonial monitoring efforts more effective than those of the present?" 2. ’Yan rani, the Hausa word for smallpox, reflects Hausa speakers’ association of this disease with the dry season. 3. Hausa people consider cassava flour a low status and rather unappetizing food. It has been consumed widely during famines (see, for example, Laya, 1975). 4. The word fitina ("troublesomeness") is combined with Hitler. Germany is Jamus in Hausa. 5. The Sokoto Province Annual Reports for the years 1951-53 and 1955-60 were not available for examination. 6. The 1953—54 famine potentially has serious implications for Sahelian agriculture and food security in the next century. Climate monitors cannot overlook superabundant rainfall as a famine catalyst in semiarid lands. The three principal techniques used by atmospheric scientists to predict climatic change in a 215t century COf‘ warm world — computer—generated general atmospheric circulation models, palaeoclimatic reconstructions, and instrumental scenarios — all indicate strengthened monsoons and increased rainfall. 7. These data are taken from a proposal for a settlement scheme in southern Sokoto Province. That the 1953-54 famine and distress migration may have accelerated the north—south "drift" was not acknowledged by the proposal’s authors, although the accuracy of population figures was questioned. 170 8. I strongly suspect that the 366mm in the record for May 1982 is an error. HICHIGRN STRTE UNIV. LIBRQRIES Ill 1 6085 Ill ill 11 312 30 4 ll 1