:' ms ': ' is I. * ‘-. an. .- . . . . ' o o ”I. ‘ nIT . «L ‘1 . g. ." g be ' . 1+ .0 .nI I I." n] ’5’": n I“: n . I ' '. , o I : ' I“; w I . .ll ‘IIn'HV' I I0 I. I. no ‘ ”I ' ’n¢ . n V. ‘I "."' I II I I ' 4' ' ° I ' I H I‘. ;: WII -II."I II': I“ If...“ n I.“.I -»I. I ‘5 ,. WI ”II, ‘: 'V' I 'II'II‘n .II I I . I ”.3 'I I) III.,I:"'I'I_ I n~ II ' II ' v' I\* " I: ' . ' I i" ‘34 ' . SIS“... 1".” I" ‘1.“ I I ‘ I "I I I '. ’lewnw , I“ ."‘ n I" I n n - . n 'I‘I'IEI‘W |' .' n '1 I"fl I '-"l I I‘ ”I .l .J I. II'I III.“ I ‘I I "I‘ I .l.‘ "I, I I' I' ‘ ”I I ‘ '4". \'I|’ ' LIN“ I, ,~ .n.. I I .1; T ",I' I'II. -I‘,'I . II n I I, I I I ' .I’II'I‘J 'wl'W-II I“ .‘z. .. AI" I."..‘|. . .I 'I I I .‘ " '- .-...‘7-'I.;,n'a w . ‘~*"-'II'I'I-':' I. , .‘ . I: I. . ‘ “I .. ‘I I ':"\‘\".\III n I‘III-I ‘ ‘ “I 'I 'I' I.‘ I"'.' u I" ”H ‘I“‘ ‘W . .4 r... . I -..~; 1......Ile H I- . I_'|"..P ‘01 ' $32.1“. |'rlz‘ . 'L “H'I‘I IIIII_‘.': II “III: I ,| h." I ' It. I. I I .‘I". ‘1‘ I T3 II :In: Iii": " I.‘ I” :II‘ I I. H. 3:,ll? :I' ';" ':;n 2., I'I‘I" ""w'W'V'lu 'I' ' ’V . " .l‘ » . ;.. I . I I .n .- I '1! i. [In In (,' I; ‘V‘. I6; .I M'III: I"I‘I‘| IIIII' "' iIIIII‘ I In I II;- I w». .w: ' has. . '- I): '- "'WIIIH I 'I III“ WWIII‘I‘l-HIHI ' I‘IEII'J IInII H. I I In I" ‘ III .III “diva .‘,- 'v 'I'I‘III'I l " I ,' | I'. . , . .‘1'. ‘:' , 2.".‘..-I"«'rl'I‘ 7 ' 1 .. My»... 't W; ‘I “ IC'W' "II . I; ' V I :.\'l"""":. _".",..'|I:, ' I 5".‘I'II1Iw i'.'.-I.'IH'.'. 1'; if . II: I ‘ ".02." -I '"H?II;.I|H“:S|I I'l .: II“ ”HI. I ‘nI‘ . ‘l' I.. "I (“inf-I! "5‘1|"-"I ' [I .I" ‘IIIIII' I r“ n, ‘ II. '.. I '1," -'.-rl I, I ”I H 'IIl' 'IIIII IUI, | II I I. I; 1». III'III " II I . I -. I . .. :Ii’I'II'IIIInIiIIJ. I’m-III !:III’5IM'I .l ' f 'n .‘u I - 'LI‘I‘” I“III.r'I'i'w.'i~‘-"~'s" my” a I! ’7 ’3' .‘I m- “III-IN |II!1_.f-. 1 '-I_.'.| . I" r 3 f ' ‘ ' ' “ ' 1"" I'M”; ~:7: .. VIP. .. In" "| I ."1 na'-r n‘ I..Ind“: I I'.I:‘ IIII'I‘ ,. “ - .' n ,4 ‘HIIg . ‘IC .I' ,. ‘ J ~ -. .1 I ’ll [fry] II '4' I: ‘I II, filling“ I ". W I I H ‘W I’ : 2' I tips “2.1: an . ,. ~r . iI."~...-,'.' ' II .,.;1' gr. 1‘ ' . 2 -"'.....v‘Ii.'»-**‘:.= ’. I ' 1" ‘,.: I | .. my. - ....' ... r ' “III; . I .- MI 'I‘m - w- I w .- . .. ~- . . ‘.’- n: «"3: ., m .:.r' .- :1, L 'I' ._.H W!» .‘ 0 . . . _- I ....';;;,';‘ 5}. 3;,- M“? ,:. '. ‘ . nI ..' I "It ”A ‘ ‘ 'n I- I .u,‘ .f'll'l ny'ka'bju '.nl I " INJH; I, - ”III“ "~' ...I.“. . .. .. 1 .‘I‘Hrfy" l-J' mama. Iv ' wrn ' ' ' I -. I " ’M’II‘M. w. . n‘..-.' IIxI k”III“ ‘an “'1: ”.3 ’ ‘n‘ I. “It‘ll" ' I ' " 'v I ‘I' I w, «I! o w: .'. .W-Ii I“ i IV “5%?fo "'I" . ~ «- 9., ., ,. , I I I. I , I». I. , I 0 0 or 'r, ”W” . .TI I"I " I ‘II .‘I; ‘. III IV I ‘I'. - - 4' ..u,. " . mum Iv! ,.! "I"- "' ‘ h n." 1*" ' I II'.~‘I;I:\NN 7 31. Pods black eye group blue eye group purple eye group speckled/mottled eye group brown/tan eye group red eye group green eye group per peduncle: average of three observations 32. Locules per pod: average of three observations 33. Seeds per pod: average of three observations 34. Pods per plant: total number of pods produced per singIe-row pIot divided by the number of plants 35.100 seed weight: weight in grams of 100 randomly selected seeds Shannon-Weaver Diversity Index Data was analyzed on a per-character basis by simple calculations of means, ranges, standard deviations and frequency distributions. This yielded information of genetic diversity for both quantitative and qualitative data on,a national (geographic) level. Afterwards, five 33 regions within the country were delineated as separate collection sites (see appendix A) based on geographical clustering and actual distances. Qualitative traits were class-coded for frequency analysis. Traits of a quantitative nature were divided into discrete categories and coded so that class frequencies could be obtained. Divisors differed depending on the range of the character under analysis, but were constant across all regions for each character. During the actual collection, differences were noted between sampling sites for such characteristics as seed color and size. Analysis of the diversity contained within each of these sites was accomplished by using the Shannon- Weaver diversity index (H'L. Individual indices were calculated for each site and for all descriptors. Only 219 of the 336 accessions were included in the latter analysis because of uncertainty regarding origin of some seed samples. Proper use of the ShannoneWeaver diversity index is based on random samples from a larger population. As collection sites were chosen at random and farmers within those sites were also chosen at random, it is felt that the data recorded meet that requirement. The equation for diversity measurement is given by the formula H' =-'Z:';pi log pi (1) where; s = the number of genotypic classes or species 34 pi . the proportion of the total number of individuals comprising the ith class or species and, Zpi . 1 Since the log of any value <1 is negative, the formula is preceded by a negative constant (usually 1) which results in a positive value of H'. The base of logarithms is open to choice but in most cases the base e should be used (Poole, 1974). H' is an upward biased estimate and is correctable providing the true value of s is known, since the magnitude of the bias depends on it (Pielou, 1966). However, this value is rarely known. When using natural logarithms, the expected value of H' may be found by the formula E(H')='-2.pi lnpil- s-1 (2) M 2N where; N = the size of the sample The variance of the estimate of H' is found by 8 3 var(H') :1; pi In2 pi - (3:; pi ln pi)-2 + s - 1 N 2 N2 (3) In large samples, the first term is usually sufficient. The approximate standard error of H' is given by 35 m. 8 s.e. = pi 1n2 pi - (g3 pi ln Pi)2 N (4) 1:1 V When H'rms been estimated for two or more collections, pair-wise comparisons can be made with a t-test where t = H1' - H2. [var(H1')‘; var(H2')_]‘y2 (5) and the null hypothisis is Ho: H4' x Hz'. The approximate degrees of freedom of the test is calculated by df = [var(H1') + var(Hg')]2 var(H1')2 /N1 + var(Hg')2 /N2 (6) the number of individuals in the first sample the number of individuals in the second sample where; N1 N2 Cluster Analysis Further analysis of the data was performed using the Clustan program for multivariate data analysis (Program Library Unit, Edinburgh University, 1978). Nineteen variables were used for the clustering procedure and 213 accessions were included in the analysis. Accessions having missing values for any of the variables were excluded. Ward's method, based on error sum of squares, was 36 chosen as the hierarchy Option. After analysis, the number of clusters selected for data interpretation was based on eigenvalues greater than two. RESULTS AND DISCUSSION Analysis pf Genetic Variability Both quantitative and qualitative traits showed a broad range of variabiliity on a nation-wide basis (Tables 1 and 2). Examination of the genetic variability present throughout Botswana reveals similarities, for most traits, to the extent ofcdiversity previously found in the world collection of Vigna unguiculata held at IITA in Nigeria (IITA, 1974). Such a wide range of diversity in a small geographic area the size of Botswana (582,139 sq. km) suggests a large amount of gene exchange and recombination. Many factors undoubtedly contribute to this diversity, including natural hybridization, natural and artificial selection and gene exchange between one geographic area and another. Analysis on a regional basis reveals the predominance of certain plant traits over others (Tables 3 and 4) which, quite possibly, results from farmer preferance for certain genotypes. Farmers do intentionally select certain seed and plant types over others for propagation for specific purposes. Throughout our collection trips, farmers were 37 38 Table 1. Number of accessions, means and ranges for quantitative characters. Character Accessions ‘i’ Range No. main branches 218 3.50 t 1.62 1 9 Nodes per main stem 219 4.73 t 1.32 2 9 Peduncle length (cm) 218 14.37 t 5.84 3.00 36.33 Days to 50% first flower 215 87.50 1 39.40 39 168 Days to 951 ripe pod 213 139.32 1 44.93 64 189 Pod forming period 213 53.27 3 35.95 12 147 Plant height (cm) 219 13.80 i 3.54 6.66 30.00 Plant width (cm) 219 20.48 t 6.10 4.66 42.00 Pod length (mm) 214 129.34 1 32.12 37.50 208.66 Pod width (mm) 214 7.79 1 1.66 4.00 11.67 Leaflet length (mm) 219 89.54 t 1.70 54 160 Leaflet width (mm) 218 53.79 t 1.13 28 100 10 seed weight (g) 219 1.55 t .50 5.00 3.50 Seed length (mm) 219 7.28 i 1.54 4 12 Seed width (mm) 219 6.19 t 1.25 4 9 Seed thickness (mm) 219 4.48 i 0.93 3 7 Pods per peduncle 214 1.82 t 0.74 1 4 Locules per pod 214 13.95 t 2.24 5 21 Seeds per pod 216 10.26 i 3.02 2 16 39 Table 2. Number of accessions and ranges for qualitative characters. Character Accessions Range Growth habit 219 1 7 Twining habit 219 1 3 Pod attachment 215 1 3 Raceme position 218 1 3 Determinacy 215 1 2 Flower pigmentation 211 1 6 Plant pigmentation 219 1 4 Pod pigmentation 215 1 6 Pod shape 217 1 3 Leaflet shape 219 1 4 Pod shattering 217 0 1 Seed crowding 219 1 4 Testa texture 219 1 5 Eye pattern 219 1 6 Eye color 219 1 7 0 u. =0. 8.0— an an... 8.2 2 3.. 3.0— a. gun 3.3 no. .7. n... 3 I no. a": .73 an... 3 .7: 3.: «a a... 3.3 .. .7. 8.3 no. .73 3.: 2 I no. .833 ..u .3 2 7. a... 2 a; 2.. 3 .-u no.” 2: 7. a; 3 :13... .1 38 ..a 3.. 3 To. 2.. .u To a... .. one 3.. no. a... on.» an a'. cools: 100. on. 8.. 3 ..n ~... .~ a... =.. .. .o. 3.. 2: a... 3.. 3 al. a...) .6. :0. a... a 3.. o... .u :0. no.” 0. 2:. 3.. no. 3:. 3.» an .8. 4...: s00. . 31¢ a... an «.70.. 3.. .n 0.70.. a... o. «.72. tn." no. ..n.¢.u a... 2 an. .332. so: 2 87:. 8.: 2 .7... 3.0. .n «on: 8.8 .0 «cc: .~.nn no. out: 3.: 2 .8. 3.... .033 9...». 3.8. an an...» 2.: .u 0...: 3.: 3 3.0: an... no. cannon a... 3 al. 4...: .0301. 3.2.8.. g... 3 8.2.8.. 1.. 3 8.2.3.. .1. .. 3.2.8.. 2.. 8. 3.3.8.. 3.. 2 a8. a...) ‘8 3.878.: 8.9: 3 3.878.: :3: an 3.873.: 0...: .. 3.98.8.8 8.": «3 3.873.: 2.63 3 .8. Joule. 8; 8.2.8.: 2.: a. 3.2.2.... 3... .u 8.3.5.... on... .. 3.2.0... on... 8. 8.3.2.: 8.2 a. .8. 3.... «'2 8.2.8.. 3... 3 2.3.8... I...— .n 3.97.... 3.3 .. 8.3.8.. 8... .3 8.30.... 2.3 2 .8. in... «no: .203 3... 3 37: a... an :73 an... a. :73 3... no. «3.... 3... 3 3...! .33.. so; 0:... 8.... 3 :7.- 283 um 03.... 2‘: a. 0.7.. :3: «3 :7: 8.... an '0. 8: «no 0. one. :70. 8.3. 2 .3... 0...: .n 0.7: 8... a. 3.7—. 3.: «3 ..T.» 3.2 3 you... :8. «3 3 {on 8.3.3.. 8.3 3 3.3.8.. 3.2 .~ 8.2.3.. 3.: a. 2.3.3... 1.: .8 247.... 8.2 3 16. no... 3!... 7n ..6 3 a... 3.. .u ..u n... .. .-a o... no. ..a a... 2 I... a... no. on! ..n a... an ..a 3.. .~ 7. a... .. a... 3.. 2: .-u a... 3 .333. 3.! .0- 0...- n. .883... 0.:- u. coo-.33. on... m 08.3.... 0.8- m. 8.....04 one-d . .u. 33.33. 3.3.16 J51. .|I.1. Jan... a E .. .33.: h. 3.3:... 09:33: so. no...» .8 one- .33....»- .o ~30... .n o...“ QQ‘O‘O FOO-GHQ QQQ!‘ HHHH \O¢V¢fl¢fl¢V\o~¢In¢V-c eI—I—IpquanIpI—Ioa v\«I«I«Ioa¢>«Iv\na«I «s—q—aoaw1r4—Ip4wao4 \Otfltfl¢fl¢V\D~Q\D¢W~3r4~3-¢\OI\ eraerIrI—IFIFIFIoIc>pIp¢—IFI o. o. co o. no o. no mm o. H. N. o. co we me o I H «OH 0 I H mOH c I H MOH . I H MOH H I o moH . I H moH N I H «OH 9 I H «OH n I H «OH o I H HOH N I H mOH n I H MOH m I H NOH n I H MOH h I H noH “MMMNO MMHQmw‘O llHr-Il NHHNHH HNOHHHH MH MH MH nH mH MH mH «H MH mH MH nH MH nH mH uOHou ohm cuuuuam ohm uuauxuu Mummy wsHvsouu comm mnHuouumsu vow 09mg» uunmoH oamnm com aOHuuuamBmHa com QOHuauaoaan uamHm aoHumucoaan uuaon humcHauouoa cowuHmoa vacuum unmanumuum won .22 .553 uHaan suaouo umawm uQOHuuuuu< owsmm unoammuuu< owamm maowmmuuo< owcwm macammoou< manna m:o.mmuuu< n conmm N couwum H :3me wouumumau .maOHmou an muouumumsu o>HuuuHHm=w we. mmwcmu can chHmmouum mo Honaaz .. UHAMH H2 asked why they grew a particular mixture and what special attributes were possessed by the components of the mixtures which made them desirable. Almost invariably, the farmers could identify particular seed types, often calling them by name, and cite the characteristics of these genotypes which contributed to its significance in the mixture (i.e. good drought tolerance, aphid resistance, resistance to pod bugs, eth. Correlation coefficients between metrical traits among cowpea landraces are shown in Table 5. Correlations were noted between days to 50% first flower, pod forming period and most other metrical traits. A range of 12 to 147 days for pod formation found among the accessioons reflects the variability present in the landraces for growth habit, since erect, determinate types often mature quickly while prostrate, indeterminate types require several weeks to complete their reproductive phase. The observed variability in reproductive period and consequent correlation with other plant traits is one genetic aspect of cowpea which may be related to the social and ecological structure of the country. Mixtures of fast maturing determinate varieties with late maturing, indeterminate types provides some insurance to farmers that some yields will be obtained, either in leaf or grain harvest, even if only scant seasonal rainfall is received. Indetermiate types supply a source of leaf vegetable over .u.o>.auosoou ....po. u..~..¢.¢ut «9.9 .0. «9.0 on. .- «new...nu..cc.o . 3 I no. .10.. cc.n~.o .— 105 .0. 00.300. .n~.o 0.0.: a. ennui... no. .10; no... n.°.oI can.~.on .. encode... .00. s.e.o «no.0 0.....ol ...~..c an a...) v... ..o... ..e... .....o. ......o ..o...o .. ..n... v... on... ..o.o ......o. ...¢..o ..n...o ..n...o .. .a...» ..c. o. ..e... ..o.o ....o. ......o ......o on... ......o .. a»... ....... ..o.o .....o ..oo..o. ......o ......o ......e ......o ......o .. a..¢.. ....... «<4 ....... .....o ......o. ......e ......o ......o ......o ...o..c ..n...o o. a.... so. nu. ..n.... ..n.... ..o... ...o..o ..o...o ......o ..a...o o...o ......o ..o...o . ..un.. so. ..o.o ....o .o...o. ......o ......o ..n...o ......o ......o ..o.... ......o ...o..o . 1.... ..... .oo... on... .o... ....o ..o.o 9.... ....o ......o ..oon.c ..o.o s.e.. ..n...o . .g.... ..... ooo.o ..o.o ..o... ea... ..c... s.e.o- on... ......o ....¢ ..o.o ..o.o ....o ..c... . ...... ...-... so. .o... ..... ...n..o. ...¢..o ......o .....o ......o ......o ......o ..oo... ..o...o ......o o...o. ......o . so. .... ... .. .... ....o ....o ..u...o. ....° ......o ..o...o ..o...o ....o ....n.o ..o...o ...o..o ..n...o ..o... ......o. ..o...o . ......o.. .o. a. ...a ..co..o ..o... ......o ....o- ......o- ....o. ..eo..au ....¢. ...o..o. ......o. ..o.o ..o...o- ....o ....o. ..o...o. ..ao..o. . ...... ..u..... ....o ....... .....o. ...... ......o ..o.° ..c... ......o ..n...o ......o ..... ..o...o ......o ..o.o ..n.... ......o ..a... . a... u... .. ..go. .....o ..o...o ....o. ..o.° ......o ....o ....o ..s...o ......o ......a ......o ......o ......o ..o.o ...o..c ......o ..o... ......o . ..s..... ...: o. .. .. o. .. .. n. .. .. o. o . . . a c . . .......;u ..ou...ao. oogaeo .0 on...» o>.uou.uuoav ocean-A ..oo.u.-coo snug-non... .o u.u.-: .n o...p 1414 an extended period of time, as well as grain. In addition, they provide a wealth of forage material for cattle, which is the main source of revenue to Botswana farmers. Shannon-Weaver Diversity Measurements Phenotypic diversities, as measured by the Shannon- Weaver index (H0, for fifteen qualitative plant characters are presented in Table EL Differences among the five regions were obtained for all traits under investigation. Generally, index values were quite high, and for only a few plant characters (ie. plant and pod pigmentation in region 1 and pod shattering in regions 4 and 5) was monomorphism expressed. Averages ofl¥ values over all characters within areas showed region 2 and 3 to contain the highest amounts of phenotypic diversity (1.0196 and 0.9550, respectively) while region 1 had the lowest value UL7180). Diversity index values were used to calculate expected values of H' (E(H')) and variances (Var(H')) shown in Table 7. From this information, pairwise comparisons were calculated between H' values of different regions for all characters studied. The resultant t-tests (Table 8) indicate which regions differ significantly with respect to phenotypic diversityu.Although index values obtained for growth habit were high for all regions, significant H5 ~osm.o “mo~.o onmm.o eaHo.H oma~.¢ gm HoHo.H nxmm.H mogn.~ smam.a anem.H uoHoo «mm ~oao.o ~osw.o aaa~.H mnem.a HoHo.H sumuumm new o~mo.o Hoao.c sas~.o Nash.o cams.o ouasxou «some “Hmm.a mBsN.H amon.~ namH.H ooon.a meuusouu swam o o san~.o qun.o HHB~.o wcfiuouuwnm com oHHa.o mshm.o maam.o Nflsm.o NmNs.o «amen uufiuaos mooe.o nomo.o smem.c o-o.¢ Hofio.a mama» com smmo.H meaq.o cmm~.o mono.H o couuoucmewua com ammm.o omae.o mmam.° ~sm~.o o coaumuaoswua sagas cmma.o “HmN.H msflq.fl Hoqo.H HoHo.H acauaucoswaa umaoflm Nosm.o eaco.o shoo.o humo.o na~s.o suacusuuuun ammo.a moms.o Hmww.o masc.a Nome.o cofiuumoa assume HmNa.o «Nmo.o oesm.o ~mew.¢ «Nao.fl acusnuauua nos ~a~s.¢ s~a~.o econ.o aess.o ~H-.c sung; sausage s~se.a maso.fi Haas.” Nqnn.fl ~me~.a “as“; assets n s n ~ H souuuuano :owmmm .muououumno o>auuuw~u=v you noowvca hufiuuo>wv um>aoSn=occmsm .o manna 46 Table 7. Shannon-weaver diversity indices, expected values and variances for qualitative traits. Character N* S** H' E (H') Var (8') Region 1 Growth habit 13 7 1.2657 1.0349 0.0138 Twining habit 13 3 0.2712 0.1943 0.0337 Pod attachment 13 3 1.0123 0.9354 0.0109 Raceme position 13 3 0.6902 0.6133 0.0005 Determinacy 13 2 0.4293 0.3908 0.0291 Flower pigmentation 13 6 1.0101 0.8178 0.0137 Plant pigmentation 13 4 0 - - Pod pigmentation 13 6 0 - - Pod shape 13 3 1.0101 0.9332 0.0137 Leaflet shape 13 5 0.4292 0.2754 0.0291 Pod shattering 13 2 0.2711 0.2326 0.0337 Seed crowding 13 4 1.3060 1.1906 0.0116 Testa texture 13 5 0.6870 0.5332 0.0482 Eye pattern 13 6 1.0101 0.8178 0.0137 Eye color 13 7 1.3777 1.1469 0.0327 gegion 2 Growth habit 103 7 1.5542 1.5251 0.0047 Twining habit 103 3 0.6469 0.6372 0.0067 Pod attachment 102 3 0.8652 0.8554 0.0042 Raceme position 103 3 1.0695 1.0598 0.0005 Determinacy 103 2 0.6927 0.6878 0 Flower pigmentation 101 6 1.6401 1.6302 0.0028 Plant pigmentation 103 4 0.7542 0.7369 0.0043 Pod pigmentation 102 6 1.0363 1.0118 0.0111 Pod shape 102 3 0.6220 0.6122 0.0013 Leaflet shape 103 5 0.8412 0.8218 0.0076 Pod shattering 103 2 0.5542 0.5493 0.0023 Seed crowding 103 4 1.1373 1.1227 0.0042 Testa texture 103 6 0.7772 0.7529 0.0061 Eye pattern 103 6 1.5835 1.5592 0.0005 Eye color 103 7 1.5194 1.4903 0.0016 Table 7. (Continued). 147 Character N* S** H' E (H') Var (H') Region 3 Growth habit 66 7 1.5901 1.5446 0.0035 Twining habit 66 3 0.3064 0.2912 0.0099 Pod attachment 64 3 0.9960 0.9804 0.0031 Raceme position 66 3 0.8881 0.8729 0.0052 Determinacy 62 2 0.6674 0.6513 0.0008 Flower pigmentation 61 6 1.4143 1.3733 0.0113 Plant pigmentation 66 4 0.8955 0.8728 0.0044 Pod pigmentation 6S 6 0.7830 0.7445 0.0203 Pod shape 65 3 0.8484 0.8330 0.0007 Leaflet shape 66 5 0.8798 0.8495 0.0060 Pod shattering 65 2 0.1394 0.1317 0.0055 Seed crowding 66 4 1.3059 1.2822 0.0022 Testa texture 66 6 0.7694 0.7315 0.0060 Eye pattern 66 6 1.2999 1.2670 0.0069 Eye color 66 7 1.5408 1.4953 0.0045 Region 4 Growth habit 24 7 1.0775 0.9525 0.0017 Twining habit 24 3 0.7924 0.7509 0.0139 Pod attachment 23 3 0.6324 0.5889 0.0264 Raceme position 24 3 0.6365 0.5948 0.0044 Determinacy 24 2 0.6616 0.6408 0.0025 Flower pigmentation 24 6 1.2817 1.1775 0.0091 Plant pigmentation 24 4 0.6750 0.6125 0.0225 Pod pigmentation 23 6 0.4703 0.3616 0.0323 Pod shape 24 3 0.6365 0.5948 0.0044 Leaflet shape 24 5 0.3768 0.2935 0.0173 Pod shattering 23 2 0 - - Seed crowding 24 4 1.2678 1.2053 0.0076 Testa texture 24 5 0.6161 0.5328 0.0258 Eye pattern 24 6 0.8602 0.7560 0.0148 Eye color 24 7 1.5373 1.4123 0.0153 148 Table 7. (Continued). Character N* S** H' E (H') Var (H') Region 5 Growth habit 13 7 1.4126 1.1818 0.0257 Twining habit 13 3 0.4292 0.3523 0.0291 Pod attachment 13 3 0.9251 0.8482 0.0250 Raceme position 12 3 1.0397 0.9564 0.0100 Determinacy l3 2 0.5402 0.5017 0.0198 Flower pigmentation 12 6 0.9830 0.7747 0.0589 Plant pigmentation 13 4 0.8587 0.7433 0.0262 Pod pigmentation 12 6 1.0984 0.8901 0.0801 Pod shape 13 3 0.6663 0.5894 0.0040 Leaflet shape 13 5 0.9110 0.7572 0.0175 Pod shattering 13 2 O - - Seed crowding l3 4 1.3517 1.2363 0.0048 Testa texture 13 5 0.6870 0.5332 0.0482 Eye pattern 13 6 0.6902 0.4979 0.0005 Eye color 13 7 1.0101 0.7793 0.0137 *N 8 number of accessions evaluated; **S - number of phenotypic classes per character. "9 Table 8. Student's t values for pairwise comparisons by regions of 8' values for qualitative plant characters. Region 1 2 3 4 5 Growth Habit 1 2.1211* 2.4664* 1.5117 0.7391 2 0.3964 5.9588** 0.8121 3 7.1085** 1.0387 4 2.0244 5 Twining Habit 1 1.8692 0.1686 2.3889* 0.5856 2 2.6428** 1.0137 1.1506 3 3.1503** 0.6218 4 1.7515 5 Pod Attachment 1 1.1971 0.1378 1.9670 0.4602 2 1.5309 1.3308 0.3505 3 2.1170* 0.4230 4 1.2910 5 Raceme Position 1 11.9945** 2.6212* 0.7671 3.4108** 2 2.4027* 6.1857** 0.2908 3 2.5679* 1.2296 4 3.3600** 5 Determinacy 1 1.5438 1.3770 1.3068 0.5015 2 0.8895 0.6209 1.0835 3 0.1010 0.8862 4 0.8129 5 Flower Pigmentation 1 4.9045** 0.5623 1.7987 0.1006 2 1.9016 3.2854** 2.6454* 3 0.9284 1.6278 4 1.8257 5 50 Table 8. (Continued). Region 1 2 3 4 5 Plant Pigmentation 1 _ _ - - 2 1.5149 0.4838 0.5984 3 1.3441 0.2104 4 0.8324 5 Pod Pigmentation 1 - - - - 2 1.4295 0.27169** 0.2056 3 1.3634 0.9954 4 1.8735 5 Pod Shape 1 3.1688** 1.3475 2.7769* 2.5842* 2 5.0625** 0.1921 0.6085 3 2.9672** 2.6562* 4 o. 3251 5 Leaflet Shape 1 2.1506* 2.3848* 0.2433 2.2319* 2 0.3310 2.9430** 0.4406 3 3.2953** 0.2035 4 2.8636** 5 Pod Shattering 1 1.4921 0.6652 - — 2 4.6967** - - 3 - - 4 _ 5 Seed Crowding 1 1.3421 0.0009 0.2757 0.3569 2 2.1075* 1.2013 2.2600* 3 0.3849 0.5474 4 0.7534 5 Table 8. (Continued). Region 1 3 4 5 Testa Texture 1 0.3539 0.2606 - 2 0.0709 0.9020 0.3871 3 0.8597 0.3539 4 0.2602 5 Eye Pattern 1 2.0191* 0.8879 2.6845* 2 3.2968** 5.8475** 28.2486** 3 2.9849** 7.0876** 4 1.3744 5 Eye Color 1 0.8456 0.7308 1.7065 2 0.2740 0.1389 4.1174** 3 0.0251 3.9338** 4 3.1120** 5 *,**Significant at the .05 and .01 level respectively. 52 differences in H' were obtained between most regions except 5, whose value (HHVL4126) was intermediate between the two extremes, region 4.(1U=1.0775) and region 3 (10:1.5901). Such differences in diversity are noted by a lower degree of polymorphism for a given trait, as in region 4, or a high degree of phenotypic diversity with some alleles expressed more frequently than others, as in region 3. Figures 1 and 1a demonstrate the relative frequencies of the various classes of plant growth habit occurring on a country-wide and regional basis. The position of cowpea racemes, and consequently pod production, is an important plant breeding characteristic in developing varieties for commercial production. Botswana landraces of cowpea show considerable variability both within and between regions for this trait. Significant differences were obtained between nearly all regions (Table 8), with region 2 displaying the greatest variability (Figures 2 and 2a). Other plant characters contributing significantly to differences among regions were pod shape (Table 8 and Figures 5 and 3a) and leaflet shape (Table 8 and Figures 4 and 4a). No differences in H' values were found for determinacy, plant pigmentation or testa texture, indicating the similarities in frequency distributions for these characters across all regions (Figures 5 and 5a). Interesting regional differences were obtained among 53 50 - Total Alteration: = 219 .2 4° " 5 so .- £5 20 '- 10 - l 1 [ I I code: 1 2 3 4 5 6 1 Growth habit Figure 1. Dispersion of cowpea growth habit among all acessions. :1 50 '- v ‘o - 5‘ so - 0 : 2° - l H I 2 10 l h — I 1'11 11 L I 7 I ’III'T‘IIIIIITIITI [1 133113511337 13511357 Region I 2 3 4 5 Growthhabit Figure 1a. Cowpea growth habit distributions by regions. 54 5° " ' Total accessiont = 213 40!- l‘roqnency ( % i 8 I 10!- uflc l 2 3 Raceme Position Figure 2. Distribution of raceme position for all cowpea accessions. I'requencytx) 8883888 I I I I I I I n -— u -— ' | l 1 ud— ”N ”d o-s-I 2 3 a.” - “IN Region 1 Raceme Position Figure 2a. Distribution of raceme position by regions. 55 gor- ‘I'otal Accession = 217 —~ 50" g V g. 40" a 0 E n. 30" 20" 10'- l I «is 1 ‘ 2 3 Pod Shape Figure 3. Frequencies of the various classes of pod shapes observed among cowpea accessions. . 70" at so- E 40-- §. 304 0 a; 20+- 10— J - l I I r F l 1 1 2| 3 1 2. 3 1 2; 3 1 2. 3 1 2 3 Region 1 2 3 4 5 Pod Shape Figure 3a. Distribution of pod shape classes according to regions. 56 75" Total accession = 219 60" 8 45" 8‘ 8 3' 30'- h. 15" . l I I I «Me 1 it 3 1 Terminal Leaflet Shape Figure 4. Observed variability in terminal leaflet shape of Botswana cowpea landraces. .. .OF' 1 3' 10- ‘0'. a 4o- 1 0 :2: 30" 20' I l I I l . l I II I I I I I I I I I I I l I l 2 13 I l 2 13 4 1 2 13 t l 2 3 I 1 2 ,3 4 Mafia 1 2 3 4 5 Terminal Leaflet Shape Figure 4a. Distribution of leaflet shape according to regions. 57 15" Total Accessions = 219 ‘0 I- g 45 h- 8- 8 . he 30 .- 15- n l . I . I I an: 1 2 3 4 5 Testa Texture Figure 5. Distribution of the various classes of seed coat texture for all accessions of cowpea. Aso am ”so gso 34o sso In. so "II I l I 1 I t .. .. .. .: .1. 135 135 135 135 135 Regionl z 3 4 5 Testa Texture Figure 5a. Similarities in the distribution of seed coat texture throughout all regions. 58 11' values for seed eye pattern (Table 8 and Figures 6 and 6a). Almost all regions were significantly different from each other, with region 5 displaying the lowest (HH4L6902) and region 2 the highest (HHKL5835) diversity. Among all regions, eye patterns 1 and 6 appeared to be most prevelant (Figure 6a) corresponding to solidly pigmented and red hilumed seed, respectively. Because of the distinct pigmentation of the hilum, as opposed to the seed body, detection of "preferred seed types" in cowpea becomes complicated, especially among landraces where a multitude of combinations exist. This is further reflected in Table 8 and Figures 7 and 7a. All regions show high diversity and are not significantly different from one another, except for region 5 (HHKLO101) which differs from regions 2,3 and 4. Only 3 of the seven color classes known to exist in Botswana landraces are found in region 5 (Figure 7a). Further investigations would be needed to determine if the lower diversity of eye color found in this region is a resultzof direct selection by farmers for preferred seed types, since no such trends are evident in any other regions. For the nineteen qualitative cowpea characters under investigation, the highest average 11' value was obtained for region 3 (HHVL0893) while region 4 had the lowest value (H'=O.8673, Table 9). Calculations of E(H') and Var(Hfl) (Table 10) and corresponding t-tests (Table 11) 59 50- Total accessions = 213 a~ ‘40- i . g; 30" 8 E a: 20" I I I I «M: 1 2 3 4 5 6 Eye Pattern Figure 6. Distribution of seed eye pattern in cowpea landraces. .. 80 P 3: 10" “I 5‘ so- :‘i “" a: 30 P 20' l Illl 1 In J I I l I I ”I r I I l I l I l 2 4» 6 2‘4 6 2 q. 6 2 '4 6 2 4 6 Region 1 2 3 4 s EyePattern Figure 6a. Distribution of seed eye pattern by regions. 15.- I‘rfitufl! (1H 60 Total accessions = 219 ‘- Bye Color Figure 7. Seed eye color frequencies of Botswana COWpea landraces. $ 70 P -50- so.- E 340- am- ” l l H I I I I I l I I 1 II IIIIIIIIIIIIIIIIII I 3 5 1' l 3 5 1 I 31 5 1' l 3 5 1’ 1 3 55 1 Region 1 2 3 4 5 ByeColor Figure 7a. Distribution of eye color by regions. 61 Table 9. Shannon-Weaver diversity indices for quantitative characters. Region Character 1 2 3 4 5 No. main branches 0.8981 1.0331 1.1590 1.0378 0.6902 Nodes on main stem 1.0579 0.9115 1.0836 0.8152 1.0928 Peduncle length 0.6870 1.1871 0.9337 0.6365 0.8587 Days to 502 first flower 0.8587 0.6673 1.1546 0.9582 1.0579 Days to 95% ripe pod 0.9370 1.1810 1.3335 0.8961 0.2711 Pod forming period 1.2202 1.0484 0.8874 0.9297 1.1569 Plant height 0.6663 1.1204 0.9949 0.8294 1.0123 Plant width 0.9110 0.9376 1.0236 0.6792 0.7903 Pod length 0.9251 0.9913 1.2867 1.0671 0.8981 Pod width 0.9839 1.2417 1.1927 0.9990 1.2047 Leaf length 0.8587 1.1727 1.2791 1.0506 0.8981 Leaf width 0.9184 0.8280 0.8621 0.5117 0.8587 10 seed weight 0.9110 0.8047 1.1153 0.9181 1.0101 Seed length 0.6870 0.9082 0.8451 0.8366 0.5402 Seed width 0.6902 0.8814 1.0147 0.7781 0.9839 Seed thickness 1.2658 1.1584 1.3202 1.1524 1.2711 Pods per peduncle 0.6663 1.1235 1.0005 0.6765 0.8981 Locules per pod 0.6663 0.8043 0.9813 0.6365 0.4292 Seeds per pod 1.0579 1.1669 1.2279 1.0702 0.6197 H' 0.8877 1.0088 1.0893 0.8673 0.8706 62 Table 10. Shannon-Weaver diversity indices, expected values and variances for quantitative traits. Character N* S** H' E (H') Var (H') Region 1 No. main branches 13 5 0.8981 0.7443 0.0198 Nodes per main stem 13 4 1.0579 0.9425 0.0804 Peduncle length 13 5 0.6870 0.5332 0.0482 Days 502 first flower 13 4 0.8587 0.7433 0.0262 Days 951 ripe pod 13 4 0.9370 0.8216 0.0590 Pod forming period 13 4 1.2202 1.1048 0.0199 Plant height 13 5 0.6663 0.5125 0.0040 Plant width 13 4 0.9110 0.7956 0.0175 Pod length 13 5 0.9251 0.7713 0.0250 Pod width 13 4 0.9839 0.8685 0.0157 Leaflet length 13 6 0.8587 0.6664 0.0262 Leaflet width 13 4 0.9184 0.8030 0.0192 10 seed weight 13 4 0.9110 0.7956 0.0175 Seed length 13 3 0.6870 0.6101 0.0482 Seed width 13 3 0.6902 0.6133 0.0005 Seed thickness 13 5 1.2658 1.1120 0.0137 Pods per peduncle 13 4 0.6663 0.5509 0.0040 Locules per pod 13 5 0.6663 0.5125 0.0040 Seeds per pod 13 4 1.0579 0.9425 0.0062 Region 2 No. main branches 102 5 1.0331 1.0135 0.0029 Nodes per main stem 103 4 0.9115 0.8969 0.0054 Peduncle length 103 5 1.1871 1.1677 0.0039 Days 50% first flower 102 4 0.6673 0.6526 0.0065 Days 95% ripe pod 102 4 1.1810 1.1663 0.0032 Pod forming period 102 4 1.0484 1.0337 0.0063 Plant height 103 5 1.1204 1.1010 0.0011 Plant width 103 4 0.9376 0.9230 0.0050 Pod length 103 5 0.9913 0.9719 0.0039 Pod width 103 4 1.2417 1.2271 0.0024 Leaflet length 103 6 1.1727 1.1484 0.0032 Leaflet width 103 4 0.8280 0.8134 0.0023 10 seed weight 103 4 0.8047 0.7901 0.0056 Seed length 103 3 0.9082 0.8985 0.0023 Seed width 103 3 0.8814 0.8717 0.0024 Seed thickness 103 5 1.1584 1.1390 0.0041 Pods per peduncle 102 4 1.1235 1.1088 0.0032 Locules per pod 103 5 0.8043 0.7849 0.0046 Seeds per pod 103 4 1.1669 1.1523 0.0033 lab Che lJHHHrrtmh-imfiri’l! (A In H m—n In In 63 Table 10. (Continued). Character N* S** H' E (H') Var (H') Region 3 No. main branches 66 5 1.1590 1.1287 0.0022 Nodes per main stem 66 4 1.0836 1.0609 0.0046 Peduncle length 65 5 0.9337 0.8952 0.0034 Days 502 first flower 63 4 1.1546 1.1308 0.0025 Days 952 ripe pod 63 4 1.3335 1.3097 0.0017 Pod forming period 63 4 0.8874 0.8636 0.0098 Plant height 66 5 0.9949 0.9646 0.0030 Plant width 66 4 1.0236 1.0009 0.0053 Pod length 64 5 1.2867 1.2555 0.0057 Pod width 64 4 1.1929 1.1695 0.0029 Leaflet length 66 6 1.2791 1.2412 0.0050 Leaflet width 66 4 0.8621 0.8394 0.0042 10 seed weight 66 4 1.1153 1.0926 0.0035 Seed length 66 3 0.8451 0.8299 0.0065 Seed width 66 3 1.0147 0.9995 0.0025 Seed thickness 66 5 1.3202 1.2899 0.0074 Pods per peduncle 64 4 1.0005 0.9771 0.0024 Locules per pod 64 5 0.9813 0.9501 0.0127 Seeds per pod 65 4 1.2279 1.2048 0.0047 Region 4 No. main branches 24 5 1.0378 0.9545 0.0193 Nodes per main stem 24 4 0.8152 0.7527 0.0364 Peduncle length 24 5 0.6365 0.5532 0.0044 Days 502 first flower 24 4 0.9582 0.8957 0.0095 Days 952 ripe pod 22 4 0.8961. 0.8279 0.0167 Pod forming period 22 4 0.9297 0.8615 0.0328 Plant height 24 5 0.8294 0.7461 0.0107 Plant width 24 4 0.6792 0.6167 0.0011 Pod length 20 5 1.0671 0.9671 0.0032 Pod width 21 4 0.9990 0.9276 0.0076 Leaflet length 24 6 1.0506 0.9464 0.0038 Leaflet width 24 4 0.5117 0.4492 0.0122 10 seed weight 24 4 0.9181 0.8556 0.0242 Seed length 24 3 0.8366 0.7949 0.0100 Seed width 24 3 0.7781 0.7364 0.0211 Seed thickness 24 5 1.1524 1.0691 0.0136 Pods per peduncle 22 4 0.6765 0.6083 0.0015 Locules per pod 21 5 0.6365 0.5413 0.0051 Seeds per pod 22 4 1.0702 1.0020 0.0024 64 Table 10. (Continued). Character N* S** H' E (H') Var (H') Region 5 No. main branches 13 5 0.6902 0.5364 0.0005 Nodes per main stem 13 4 1.0928 0.9774 0.0009 Peduncle length 13 5 0.8587 0.6664 0.0262 Days 502 first flower l3 4 1.0579 0.8656 0.0062 Days 952 ripe pod l3 4 0.2711 0.1557 0.0337 Pod forming period 13 4 1.1569 1.0415 0.0315 Plant height 13 5 1.0123 0.8585 0.0109 Plant width 13 4 0.7903 0.6749 0.0362 Pod length 13 5 0.8981 0.7443 0.0198 Pod width 13 4 1.2047 1.0893 0.0229 Leaflet length 13 6 0.8981 0.7058 0.0198 Leaflet width 13 4 0.8587 0.7433 0.0262 10 seed weight 13 4 1.0101 0.8947 0.0137 Seed length 13 3 0.5402 0.4633 0.0198 Seed width 13 3 0.9839 0.9070 0.0157 Seed thickness 13 5 1.2711 1.1173 0.0180 Pods per peduncle 13 4 0.8981 0.7827 0.0198 Locules per pod l3 5 0.4292 0.2754 0.0291 Seeds per pod l3 4 0.6172 0.5018 0.0108 *N - number of accessions evaluated; **8 - number of phenotypic classes designated per character. 65 Table 11. Student's t values for pairwise comparisons by regions of H' values for quantitative plant characters. Region 1 2 3 4 5 Number of Main Branches 1 0.6111 1.759 0.7065 1.4592 2 1.7630 0.0315 5.8807** 3 0.8266 9.0221** 4 2.4703* 5 Nodes Per Main Stem 1 0.4896 0.0882 0.7101 0.1224 2 1.7210 0.4710 2.2842* 3 1.3255 0.1241 4 1.4374 5 Peduncle Length 1 2.1910* 1.0860 0.2202 0.6295 2 2.9658** 6.0436** 1.8929 3 3.365** 0.4359 4 1.2702 5 Days to 502 First Flower 1 1.0584 1.7466 0.5266 1.1067 2 5.1366** 2.2998* 3.4660** 3 1.7929 1.0367 -4 0.7957 5 Days to 952 Ripe Pod 1 0.9784 1.6093 0.1596 2.1871* 2 2.1786* 2.0196 4.7368** 3 3.2246** 5.6466** 4 2.7840* 5 Pod Forming Period 1 1.0614 1.9311 1.2654 .2792 2 1.2689 0.5995 0.5581 3 0.2046 1.3261 4 0.8960 5 66 Table 11. (Continued). Region 1 2 3 4 5 Plant Height 1 6.3587** 3.9275** 1.3452 2.8345** 2 1.9600 2.6789* 0.9858 3 1.4140 0.1476 4 1.2445 5 Plant Width 1 0.1773 0.7457 1.6996 0.5209 2 0.8474 3.3085** 0.7257 3 4.3050** 1.1452 4 0.5753 5 Pod Length 1 0.3894 2.0638 0.8456 0.1276 2 3.0149** 0.8996 0.6054 3 2.3278* 2.4335* 4 1.1144 5 Pod Width 1 1.9432 1.5325 0.0989 1.1238 2 0.6703 2.4270* 0.2471 3 1.8923 0.0735 4 1.1778 ‘5 Terminal Leaflet Length 1 1.8313 2.3800* 1.1079 0.1837 2 1.1750 1.4594 1.8107 3 2.4358* 2.4194* 4 0.9927 5 44g Terminal Leaflet Width 1 0.6165 0.3680 2.295* 0.2802 2 0.4230 2.6267* 0.1819 3 2.7362** 0.0195 4 1.7708 5 67 Table 11. (Continued). Region 1 2 3 4 5 10 Seed Weight 1 0.6994 1.4098 0.0384 0.5610 2 3.2560** 0.6569 1.4785 3 1.1849 0.8021 4 0.4726 5 Seed Length 1 0.9843 0.6760 0.6201 0.5630 2 0.6726 0.6456 2.4754* 3 0.0662 1.8801 4 1.7170 5 Seed Width 1 3.5505** 5.9245** 0.5981 2.3075* 2 1.9043 0.6739 0.7619 3 1.5401 0.2290 4 1.0728 5 Seed Thickness 1 0.8050 0.3745 0.6863 0.0298 2 1.5088 0.0451 0.7581 3 1.1579 0.3081 4 0.6677 -5 Pods Per Peduncle 1 1.0946 4.1775** 0.1375 1.5025 2 1.6437 6.5202** 1.4862 3 5.1882** 0.5770 4 1.5184 5 Locules Per Pod 1 1.4881 2.4375* 0.3124 1.3032 2 1.3457 1.7038 2.0433* 3 2.5844* 2.7004* 4 1.1210 5 68 Table 11. (Continued). Region 1 2 3 4 5 Seeds Per Pod 1 1.1183 1.6283 0.1326 3.3800** 2 0.6820 1.2808 4.6293** 3 1.8716 4.9053** 4 3.9429** 5 ot‘n unt flc cha mat pl: Che 9? la tn fr in a: Si 69 indicate significant differences between region 2 and most other regions in the number of days required from planting until 50% of the plants within a plot produced their first flowers. Region 2 had the lowest diversity for this character, reflected in a: higher frequency of early- maturing genotypes (Figure 8a). The number of days required from flowering until plants reach physiological maturity (95% ripe pod) is related to the determinate status of cowpea (Figures 9, 9a, 10, 10a). Extended pod-production period is a useful plant characteristic under Botswana's semi-arid conditions with erratic and scant rainfall. On a nationwide scale, most landraces (59.6%) require more than 128 days to complete their reproductive period (Figure 9). Genotypes derived from region 5 show a particularly strong tendency towards indeterminate types with extended pod production periods and have a low diversity index of HH4L2711, which is significantly different from all other regions. Other quantitative characters found to have less variability in specific regions were leaflet width in region 4 (Figures 11 and 11a), and seeds per pod in region 5 (Figures 12 and 12a) which differed from all other regions (Table 11). 70 §==3130 Total accessions = 215 401-- 30- 10— W \‘ i t! s s 13 O s N 3 Days to 50% First Flower Figure 8. Distribution of the days to 502 first flower among cowpea landraces. V§§§k F ‘\\\\\\\I V§§§ Frequency 1%) assessss‘ l I I I I I I I I «12020040120200 2 I I I T' '1 120 200 40 120 too 40 120 200 3 4 3 g Days to 50% First Flower Figure 8a. Regional distribution of the days to 507. first flower. 71 3 = 139.32 50 F Total accessions = 213 a. 40'- 3 8‘ 30-- 3 g- 0 E: 20 - \‘ .LW .10— N 7 code: 60 90 120 150 180 210 Days to 95% Ripe Pod Figure 9. Frequency distribution of pod production period for all accessions. Frequency '( 'I. 1 888388388 l I I' I I I I"I I somozzoeo Ioozzoso 140 22060 14022060 uozzo Resin- 1 z 3 4 5 Days to 95% Ripe Pod Figure 9a. Distribution of accessions based on days required to complete reproduction (by regions). Figure . R . Susanne: Rey Figure Figure 10. Frequency 1 so I 3 I 72 - Total Accessions = 215 so - 3? 4o *' E I: 20F- 10" 1 win 1 2 Determinacy Determinate growth patterns observed among cowpea landraces. Figure 10a. Distribution of determinate classes by regions. Frequency ( so I 30 10 73 i = 53.29 Total accessions = 218 I I I I 25 45 55 85 105 Terminal Leaflet Width (mm) Figure 11. Range of variability in leaf width measured among Botswana A .0 ' r. g g g I: 7! Z7 g g g 25 55 I05 25 55 I05 25 55 I05 25 55 .105 25 55 I05 Region 1 2 3 4 5 Terminal Leaflet mm. (mm) Figure 11a. Distribution of leaflet width among cowpea landraces. 74 % ... z. / 7 Seeds Perz Pod Figure 12. Variability in the number of seeds per pod formed in cowpea landraces. 23 " 60 3‘50 3 40 €- 2 h ¢ 10 / 77% 4 ,, / z I I» I I I If I II I I If I If I I 431215204 31215204 51215204 81216204 5121620 Ihghn 1 2 3 4 5 SecdsPerPod Figure 12a. Distribution of seeds per pod by regions. 75 Cluster Analysis A cluster analysis was used to determine the pattern of genetic similarities between individuals within the germplasm collection of cowpea landraces. Three main clusters were identified (correlation of similarity = 21.732, Figure 13). Geographically, these clusters overlapped in their distribution (Figure 14), and between cluster dissimilarities were not correlated with latitude, rainfall or soil type. Table 12 lists the most common attributes of each cluster. Variables were ranked for each cluster according to the size of their F-ratios. Small F- ratios are good diagnostics for indicating similarity between individuals within a specific cluster for a given variable. Thus, pod width and 10 seed weight distinguished clusters 2 and 3 and the number of days to 95% Pipe pod was found to be an important variable in distinguishing between cluster 1 and 2. However, it is the set of variables and their ordering of importance which gives each cluster its unique properties. Cluster groupings were also not found to be correlated with a particular seed color, although variables related to seed size and number did appear to be important in cluster identification. The wide diversity of landraces grown throughout Botswana reflects local farmer preferences and natural 76 .moanmauw> o>HumuHuomso ma do momma moomuwama mma3oo mamzmuom mo Emuwouvcon .mH ouswwm mooanmooo< 809300 mN.NH NN.NH om.¢H -.na mm.HN Karietrmrs go nuaror;;aoo oa.- Fig“ 77 BOTSWANA L AL I u Lso ‘ 1 x \ I" \ \\‘<( “ as 1 " ‘ 1 2 ' / I 24 3 I u‘ \ .se L _1100mi L____Jroo km a: so i. :1 so Figure 14. Geographic clustering of COWpea landraces. 78 Table 12. Cluster diagnosis of means, standard deviations and F-ratios for the 4 most similar variables per cluster. Cluster Standard Number Variable Mean Deviation F-Ratio 1 Days to 952 ripe pod 170.09 24.19 0.2903 Peduncle length (cm) 13.14 4.14 0.5030 Plant width (cm) 25.13 4.83 0.6442 Pods per peduncle 1.47 0.62 0.6861 2 Days to 952 ripe pod 165.12 25.67 0.3270 Plant height (cm) 11.45 2.04 0.3402 10 seed weight (g) 1.88 0.34 0.3728 Pod width (mm) 9.02 1.05 0.3896 3 Days to 502 first flower 58.71 15.37 0.3270 10 seed weight (g) 1.20 0.35 0.3402 Pod width (mm) 6.67 1.08 0.3728 Seed thickness (mm) 4.10 0.67 0.3896 79 selection for genotypes most adapted to the many variable, and sometimes unfavorable, environmental conditions. The reasons that farmers continue to grow landraces under such conditions become evident. The greater adaptability of landraces and the protection afforded them by the mixture of components is essential to farmers faced with nutrient- poor soils, low rainfall and many disease and insect problems. The special gene combinations that have evolved under such systems of cultivation in Botswana may confer, to the population as a whole, greater stability. This is in contrast to elite cultivars bred for commercial production where stability is based on a single genotype. It is no wonder, then , that such variation still exists in Botswana, despite the cultivation of commercial cowpea purelines in neighboring Zimbabwe and South Africa. Farming remains a low-technology practice in terms of the application of fertilizer, irrigaiton or pesticides. There is, perhaps, no better protection against the vagaries of nature then to plant a range of genotypes throughout the environment, each possessing various levels of resistance to the array of possible difficulties, such as insects, drought and disease, quite capable of eliminating populations of pure line varieties. The question of why individual farmers use mixtures of differing levels ofeiiversity remains unanswered. If extremely diverse and unpredictable environments are only 80 manageable by planting highly heterogeneous mixtures, then one would expect a correlation between environmental instability and landrace mixture diversity; One would also expect such mixtures to evolve at a greater rate, since the presence of many more genes increases the probability of mutation, drift or natural selection which change population allelic frequencies and presents opportunities for genetic recombination. Such questions need far more study than they have heretofore received in order to provide for a more complete understanding of the dynamics of landrace populations. The degree of polymorphism exhibited by this collection for all characters studied is quite large. Only upon examination on a regional basis were differences, presumably preferences, for particular plant traits evident. Although artifical selection is practiced for certain desirable seed and plant traits, the extent of this selection is not known. It is also not known if farmers deliberatly select against certain seed types appearing in the mixtures, such as off-types resulting from natural hybridizatione Such genotypes were found in most samples collected and are presumed to originate as a result of outcrossing which, under experimental conditions in Botswana, range from 0.5 to 2.0%. If these rarer genotypes are not selected against by farmers, then their incorporation into a landrace mixture depends on the many 81 environmental adversities with which they must contend. In terms of plant traits which displayed the largest variability, growth habit ranked high with an H' index value of 1.380. Through their history of cultivation, cowpea landrace mixtures have evolved a gamut of plant types, each with some special property enabling it to exploit the micro environment in which it is found. This was no where as well pronounced as in Pelotshetlha and Kanye in southeastern Botswana where some samples collected from farmers' fields contained seed all of a beige color, yet plant types were quite diverse. However, selection for a particular seed color was not observed. Shrfact, seed color had the highest 11' value of all qualitative traits ”$97). The information concerning genetic properties of these cowpea landraces may serve as ea useful basis in establishing a national plant breeding program in Botswana. Foremost, the amount of diversity for each of the many quantitative and qualitative traits measured provides knowledge to the breeder as to the extent of variation in the available gene pool. In this case, a satisfactory basis has been defined for a breeding program based on plant architecture or yield components. Further evaluations would be needed in order to identify the extent Iif resistance genes harboured within the collection to the many'insects and diseases which threaten cowpea production. 82 In addition, regional analysis of the collection proved useful in detecting any preferences for gene combinations which might have resulted through generations of cultivation of cowpea by agriculturalists. Diversity indices, such as the one used in this study, provided information as to the amount (number of phenotypic classes) and extent (frequency of each class) of diversity on a per character basis which was used to compare regions. One of the most important factors in a plant breeding program is the selection of parents. Diverse parents, when crossed, are expected to produce a greater array of, though not necessarily more desirable, segregants than parents which are closely related. Through the use of cluster anlaysis, which classified the available germplasm into relatively homogeneous groups, accessions of divergent clusters might now be crossed and the progeny screened for useful or superior recombinants. CONCLUSIONS An evaluation of cowpea (Vigna unguiculata ) germplasm col lected throughout the arable portion of Botswana revealed extensive genetic diversity among landraces grown by local farmers. Of the nineteen quantitative and fifteen qualitative plant characters evaluated, all were polymorphic in nature. The country was divided into five regions corresponding to geographically isolated collection 31 tes. Analysis of genetic diversity found within and be tWeen regions indicates that some selection for various p lant characters is taking place between the different regions. Calculation of diversity index values (11') for a l 1 traits averaged across all regions indicates that plant 8P0Wth habit and seed color had the highest genetic valPSI-ability of all qualitative characters. A cluster analysis of all accessions on a nation-wide basis indicates that three main groups of cowpea genotypes e)(‘ist‘n Clustering was not correlated with north—south c ‘1 1r1al patterns associated with latitude, soil type or rainfall. Given the variable environmental and climatic conditions in Botswana it may be that micro-environmental 2 S v o 1- ution was more important in forming the genotypes which 83 814 clustered together within these three groups. Selection of diverse parents from the three clusters may be useful in a breeding program to generate F2 populationsfrom which new , useful recombinants might result. LITERATURE CITED LITERATURE CITED Adams, MAI. 1977. An estimation of homogenity in crop plants, with special reference to genetic vulnerability in the dry bean, Phaseolus vulgaris L. Euphytica 26:665-79 Adams, MJI. and J.V. Wiersma. 1978. An adaptation of principal components analysis to an assesment of genetic distance. Research Report 347. Michigan State University Agric. Exp. Sta. East Lansing. Allard, a.w., S.K.Jain and P.L. Workman. 1968. The genetics of inbreeding populations. Advances in Genetics 14:55—131. Allard, R.W. 1970. Population structure and sampling methods. In: fignetic Resources ig Plggpg :Thgig exploration and Conservation Ieds. O.H. Frankel and E. Bennett), Blackwell Scientific Publ., Oxford and Edinburgh. PP 97-107. Bennett, E. 1970. Tactics of Plant Exploration. In: Genetic Resources E Plants : Their exploration and Conservation (eds. O.H. Frankel and E. Bennett), Blackwell Scientific Publ., Oxford and Edinburgh. pp 157-179. Blackhurst, H.T. and J.C. Miller. 1980. Cowpea. In: Hybridization pf Crop Plants. (eds. W.F. Fehr and H.H. Hadley}. American Society of Agronomy-Crop Science Society of America, Madison WI. pp 327-337. Brittingham, W.H. 1950. The inheritance of date of pod maturity, pod length, seed shape and seed size in the Southern Pea, Vigna sinensis. Proceedings of the American Soc. for Hort. Sci. 56:381-388. Broich, S.L. and R.G. Palmer. 1980. A cluster analysis of wild and domesticated soybean phenotypes. Euphytica 29:23- 32. Burkill, 1.11. 1953. Habits of man and the origin of the cultivated plants of the Old World. Proc. Linnean Soc. London 164:12-42. 85 86 Chandra, S. Comparison of Mahalanobis's method and metroglyph technique in the study of genetic divergence in Linum usitatissimum Lu germplasm collecitons. Euphytica 26:141-148. Cooley, wxw. and P.R. Lohnes. 1971. Multivariate Data Analysis. John Wiley and Sons, Inc. N.Y. 364 pp. Corley, WJ;.1969. Some preliminary evaluations of Vigna plant introducitons. Georgia Agr. Exp. Sta. Bull. 165. pp 5-23. Doku, E;V. 1970. Variability in local and exotic varieties of cowpea (Vigna unguicualta (Ln) Walp) in Ghana. Ghana Journal Agri. Sci. 3:139-143. Edye, INA”, W.T. Williams and AJJ. Pritchard. 1970. A numerical analysis of variation patterns in Australian introductions of Glycine wightii (G. igvanica). Aust. J. agric. Res. 21:57-69. FAO Plant Production and Protection Paper No.9. 1977. Food legume crops: improvement and production. FAO, Rome Faris, DJL.1964~ The chromosome number of Vigna sinensis (L.) Savi. Can. J. Genet. Cytol. 6:255-258. Faris, D.G. 1965. The origin and evolution of the cultivated forms of Vigna sinensis. Can. J. Genet. Cytol. 7:433-452. Grivetti, luE. 1978. Nutritional success in a semi-arid land: examination of Tswana agro-pastoralists of the eastern Kalahari, Botswana. The AmJJ. of Clin. Nutr. 31:1204-1220. Gunn, (LR. 1973. Recent nomenclatural changes in Phaseolus L. and Vigna Savi. Crop Science 13:496. Hawkes, J.G. 1981. Germplasm collection, preservation, and use. In: Plant Breeding 11, Iowa State University Press, Imam, AAG. and RJW. Allard. 1965. Population studies in predominantly self-pollinating species. VI. Genetic variability between and within natural populations of wild gats from differing habitats in California. Genetics 51:49- 2. International Institute of Tropical Agriculture. 1974 Cowpea Germplasm Catalog. IITA, Ibadan, Nigeria. 87 Jain, S.K. 1969. Comparative ecogenetics of two Avena species occurring in Central California. Evolutionary Biology 3:73-118. Jain, S.K., C.O. Qualset, G.M. Bhatt and K.K. Wu. 1975. Geographical patterns of phenotypic diversity in a world collection of Durum wheats. Crop Science 15:700-704. Knowles, P.F. 1969. Centers of plant diversity and conservation of crop germplasm: Safflower. Economic Botany 23(3)=524-329- Lee, J. and P.J. Kaltsikes. 1973. The application of Mahalanobis's generalized distance to measure genetic divergence in Durum wheat. Euphytica 22:124-131. Lewontin, R.G. 1972. The apportionment of human diversity. Evolutionary Biology. 6:381-398. Lloyd, M. and R.J. Ghelardi. 1964. A table for calculating the 'equitability' component of species diversity. J. Animal Ecology 33:217-225. Lush, W.M. 1979. Floral morphology of wild and cultivated cowpea. Economic Botany 33(4):442-447. Lush, W.M. and H.C. Wein. 1980. The importance of seed size in early growth of wild and domesticated cowpea. J. Lush, W.M., L.T.Evans and H.C. Wein. 1980. Environmental adaptation of wild and domesticated cowpeas (ligna unguiculata (L.) Walp.). Field Crops Research. 3:173-187. Lush, W'.M. and L.T. Evans. 1981. The domestication and improvement of cowpeas (Vigna unguiculata (L.) Walp.). Euphytica 30:579-587. Mackie, W.W. 1946. Blackeye beans in California. University of Calif. Agri.Exp.Sta. Bulletin No. 696. Berkley, Californaia. Maronna, R. and P.M. Jacovkis. 1974. Multivariate clustering procedures with variable metrics. Biometrics Mehra, K.L., C.B. Singh and K.S. Kohli. 1970. Phenotypic diversity and breeding of forage cowpea. XI International Grassland Congress. (ed) M.J.T. Norman. Surfers Paradise, Queensland, Australia. pp 293-299. 88 Mital, S.P., B.S. Dabas and T.A. Thomas. 1980. Evaluation of the germplasm of vegetable cowpea for selecting desirable stocks. Indian J. Agric. Sci. 50(4):323-326. Moore, C.S. 1975. Relative importance of rootstock and scion in determining growth and fruiting in young apple trees. Ann. Bot. 39:113-123. Murdock, G.P. 1959. Africa: its people and their culture. McGraw-Hill, N.Y. 456 pp. Narayan, R.K.J. and A,J, Macefield. 1976. Adaptive responses and genetic divergence in a world germplasm collection of chick pea (Cicer arietinum L.). Theor. and Appl. Genetics 47:179-187. Ojehomon, 0.0. 1968. The development of the inflorescense and extra-floral nectaries of ViLna unguicualta (L.) Walp. J. of W. Afr. Sci. Asso. 13(2):93-109. Ojomo, 0.A. 1971. Inheritance of flowering date in cowpeas (Vi na unguiculata (L.) Walp.). Trop. Agric.(Trinidad) Pielou, E.C. 1966. Species diversity and pattern diversity in the study of ecological succession. J. Theoret. Biol. Pielou, E.C. 1966. The measurement of diversity in different types of biological collections. J. Theoret. Biol. 13:131-144. Piper, C.V. 1913. The wild prototype of the cowpea. USDA and Bureau of Plant Industry Circular 124. pp 29-32. Poiakova, H. and A. Blum. 1983. Land-races of wheat from the northern Negev in Israel. Euphytica 32:257-271. Poole, R.W. 1974. Ag Introduction 39 Quantitative Ecology. McGraw-Hill Book Co., N.Y. pp 101-125. Rachie, K.O., K. Rawal, J.D. FranCKowiak and M.A. Akinpelu. 1975. Two outcrossing mechanisms in cowpea, liggg unguicualta (L.) Walp. Euyphytica 24:159-163. Reyment, R.A., R.E.Blackith and N.A. Campbell. 1984. Multivariate Morphometrics. Academic Press, London. 253 PP. Rawal, K.M. 1975. Natural hybridization among wild, weedy and cultivated Vigna unguicualta (L.) Walp. Euphytica 24:699-707. 89 Rawal, K.M., K.O. Rachie and J.D. Franckowiak. 1976. Reduction in seed size in crosses between wild and cultivated cowpea. J. of Hered. 67:253-254. Riggins, R,, R.A. Pimentel and D.R. Walters. 1977. Morphometrics of Lupinus nanus (Leguminosae).I. Variation in natural populations. Systematic Botany 2:317-325. Roy, R.S. and R.H. Richharia. 1948. Breeding and inheritance studies on cowpea, Vigna sinensis. J. Am. Soc. Agron. 40(6):479-489. Sauer, C.0. 1952. Agricultural Origins and Dispersals. American Geographical Soc., N.Y. 110 pp. Scott, A.J. and M. Knott. 1974. A cluster analysis method for grouping means in the analysis of variance. Biometrics 30:507-512. Sellschop, J.P.F. 1962. Cowpeas, Vi na unguiculata (L.) Walp. Field Crops Abstracts 15(4):259-2 6. Sen, N.K. and J.G. Bhowal. 1960. Cytotaxonomic studies on Vigna. Cytology 25:195-207. Sen, N.K. and J.G. Bhowal. 1962. A male sterile mutant cowpea. J. of Hered. 13(1):44-46. Shannon, C.E. and W. Weaver. 1964. The Mathmetical Theory pf Communication. U. of Ill. Press, Urbana. pp 8-17. Singh, C.B., K.L. Mehra and K.S. Kohli. 1971. Evaluation of a world col leciton of cowpea, Vigna unguicualta (L.) Savi., for pod characters. SABRAO Newsletter 3(1fi11-16. Smartt, J. 1976. Tropical Pulses. Longmans, London. pp 23- 25. Stebbins, G.L. 1957. Self fertilization and population variability in the higher plants. Am. Naturalist 91(861):357-354. Summerfield, R.J., P.A. Huxley and W. Steele. 1974. Cowpea (Vigna unguiculata (L.) Walp.). Field Crop Abst. 27(7):301- 312. Vavilov, N.L. 1955. Selected writings on the origin, variation, immunity and breeding of cultivated plants. tight, introd Industz 90 Wight, W.F. 1907. The history of the cowpea and its introduction into America. USDA and Bureau of Plant Industry Bull. 102. pp 43-59. INTRODUCTION Genetic Variability for Dinitrogen Fixation ‘Among African Cowpea Biological N2 fixation is largely responsible for the supply of N needed for food crop production in the tropics (Kang et a1., 1975). Among the food legumes, peanut (Arachis hypogaea), cowpea (Vigna unguiculata) and dry bean (Phaseolus vulgaris) are the most important protein sources in Africa (Sinha, 1977). Cowpea in particular probably has received the least amountLOf attention from researchers. Little is known of the variation in cowpea yields from farmers in Africa (Summerfield et a1., 1978). Average yields are estimated to be between 100 and 300 kg/ha and total crop failures are common (Rachie and Roberts, 1974). Estimations of the amount of atmospheric nitrogen fixed by cowpea range from 73 to 240 kg/ha"1 a"1 (Sinha, 1977). The evaluation of germplasm collections of Vigpg unguiculata has demonstrated the wealth of genetic variability which exists in this crop (IITA, 1974). The possibility of selecting genotypes for increased dinitrogen fixation has recently gained attention (Zary et al., 1978a; Graham, 1983). Enhancing the efficiency of symbiotic N2 91 92 fixation through selection and breeding may provide a means for improving yields for cowpea producers in Africa. Factors Affecting Variability ;Q.Ng Fixation ig_Cowpea Defining the extent of genetic variability for dinitrogen fixation in a crop species is often complicated. This is because heterogeneity may be eXpressed in both the host and the symbiont. Complex interactions between host, Rhizobium and environment are known to occur (Graham, 1982). Under non-symbiotic conditions, phenotypic variation in a crop species results from a combination of genetic differences, environmental effects on plant growth and reproduction and genotype x environment (GE) interactions. When considering a host—Rhizobium relationship, additional factors must be examined such as Rhizobium x host and Rhizobium x environment interactions or even second order interactions (Summerfield et a1., 1978). Environmental Factors Limiting Symbiosis Environmental factors often play a key role in limiting the expression of genetic variablity for N2 fixation in cowpea. Extreme environmental conditions may 93 have varying effects on the symbiotic system, possibly affecting the host plant.and Rhizobium in different ways (Minchin et a1., 1981). Thus, the range of environments for the symbiosis may be narrower than that of the nitrogen-fed plant (Lie, 1981). Environmental factors are thought to operate mainly through the host component of the symbiotic relationship since the host is exposed to both above and below ground environmental stresses. Adaptation 'mo environmental changes may occur through phenotypic plasticity of the nitrogen-fixing legume, or through genetic variability present in a plant population. Some of the more important factors affecting N2 fixation under field conditions are soil moisture, light, temperature, pH and available soil nitrogen. Soil Moisture Of all. legumes grown in the tropics, cowpeas are highly desirable because of their ability to survive and reproduce under limited rainfall. Currently, there is much interest in the development of superior cultivar-Rhizobium combinations which are able to maximize the N2-f1xation potential of cowpea in semi-arid regions. The effect of moisture stress on the nodulation capacity of Vigna unguiculata depends on the growth stage 94 at which the stress is imposed. Moisture stress imposed during the seedling stage adversely affects the production of root hairs (Lie, 1981). Since these are the normal sites of infection of the Rhizobium bacteria, symbiotic association may be decreased (Sharifi, 1984). Repeated exposure of cowpea to water stress prior to flowering resulted in decreased nitrogenase activity and nodule weight (Summerfield et ale, 19760. Ultimately, total seed weight, seed number and fruit weight per plant were reduced. Imposed water stress after flowering did not reduce yield and nitrogenase activity was less affected. Sprent (1972) has demonstrated a high degree of correlation between soil-water content and nitrogen-fixing activity in Vicia faba. Slow drying of soil under natural field conditions resulted in a reduction of nodule activity. Root nodules need a constant supply of water in order to export the products of fixation. Therefore, the effect of soil moisture stress on nitrogen fixation is believed to be a direct one. However, reduced supplies of photosynthates from wilted leaves may also affect the process since nodules are dependent on carbohydrates produced in the leaves (Hardy et a1., 1975; Lawn et a1., 1974). Soil moisture stress may also affect the differential survival rate of rhizobia present in the soil profile. Boonkerd and Weaver (1982) demonstrated that strains of cow; 501 com: rhi symt amor from 1967 suck COWp with Zabl effec geno 3380' cOns; CCodi t—o H- n; 3“ (‘1- 95 cowpea rhizobia have different capacities for survival in soil exposed to moisture and temperature conditions commonly found in tropical regions. However, the surviving rhizobia may not necessarily be the most effective symbiotic strains. It is thought that genetic differences among rhizobia to survive in adverse soil conditions result from the geographical origin of the species (Wilkins, 1967), although Osa-Afiana and Alexander (1982) found no such evidence in their studies of cowpea rhizobia. When cowpea cultivar Calilfornia no. 5 blackeye was inoculated with different strains of rhizobia and exposed to drought, Zablotowicz et a1., (1981) found some strains to be more effective in recovery of fixation ability than others. It is evident that both the host and rhizobia genotypes are important in forming a sucessful symbiotic association under moisture stress conditions and must be considered together when breeding cowpea for semi-arid conditions. The main effect of light on the symbiosis is through photosynthesis. Carbohydrates produced in the leaves are partly used for the development and functioning of the nodules. Photosynthates are presumed to be the key 96 limiting factor in nitrogen fixation under field conditions (Hardy and Havelka, 1976).lfl;has been demonstrated that improvements in photosynthesis such as increasing light intensity and/or 002 concentrations also improve nodulation and nitrogen fixation (Bethlenvalvay et al., 1978a). Pate (1966) has shown that the upper leaves in field pea (Pisum arvense LJ supply assimilates to the shoot apex while most of the assimilates produced in the lower leaves move downward to the root and nodules. Under field conditions, self-shading and mutual-shading reduce the light available to lower leaves. When a closed canopy develops the effect is even more pronounced (Sprent, 1976; Hardy and Havelka, 1976). lntercropping cowpea with maize (Zea mays) and sorghum (Sorghum bicolor) is a common practice among subsistence farmers in Africa. Such cultural practices may contribute to decreases in nitrogen fixation in cowpea by shading. Reduction in light intensity of cowpea grown under experimental conditions resulted in a decrease in root dry weight (Dart and Mercer, 1965). More primary root nodules were formed under two- thirds light conditions while full light conditions induced more secondary root nodules. When a 50% light interception treatment was applied to cowpea cv. Prima, seed yields were reduced by 25% due to the production of less pods. An additional effect of light on nodule formation has been demonstrated with phytochrome. Treatments of 97 alternating red and far-red light have shown that far-red light inhibits nodulation substantially while irradiation with red light decreases the inhibitory effect (Lie 1969). Under natural conopy conditions in Medicago sativa, an excess of far-red light was due to the preferential filtering of red light by chlorophyll (Robertson, 1966). This means that under field conditions biological nitrogen fixation may be reduced through shading which affects photosynthesis as well as root-nodule formation. Temperature High temperatures are known to affect plant metabolic processes such as respiration, photosynthesis and transpiration. In a symbiotic associaion, dinitrogen fixation may be indirectly affected via these plant processes. Studies comparing nodulated and N-fertilized cowpea cv. K2809 exposed to warm days (33°C) and cool nights (19°C) resulted in decreased dry weights and seed yields of the nodulated plants (Minchin et a1., 1980). Soil temperatures of 21°C have been shown to reduce root growth and possibly root hairs, resulting in decreased infection by rhizobia (Dart et a1., 1965). The most direct effect of temperature extremes appears to be on the growth and survival of rhizobia. High soil in In dii tem 815 was of r101" __.n U) (I) 98 temperatures inhibit the growth of rhizobia in the field. Temperatures of 35°C proved to be detrimental to the survival of two cowpea rhizobial strains TAL309 and 3281 (Boonkerd and Weaver, 1982L.But:differences among cowpea rhizobia in tolerance to high temperatures and dessication in soils.was reported by Osa-Afiana and Alexander (1982). In their studies, temperatures between 29-35°C resulted in differential surviving ability. However, when temperatures were increased to 40°C, no bacteria survived. At lower temperatures (15-20°C), growth of rhizobia is also inhibited (Dart et a1., 1965). Pisum sativum cv. Iran was found to be "resistant" to nodulation by a large number of rhizobial strains when grown at 20°C but nodulated normally at 26°C (Lie, 1971). The effect of low temperatures appears to be due to a delay in infection, since nitrogen fixation itself is not cold-sensitive (Lie, 1981). Recent studies indicate that elevated temperatures may also affect the nodulation process via the bacteria. Zurkowski and Lorkiewicz (1979) exposed several strains of .5; trifolii to high temperatures which resulted in the [reversion of some strains to non-nodulating mutants (Nod’). TPwo strains of R; trifolii, 24 and T12, were examined in detail. The inability to nodulate in these mutant strains wens due to the absense of plasmid pWZZ, indicating the Pldasmid-mediated control of nodulation in R; trifolii. 99 Studies on nitrogen fixation in legumes have shown that the activity of the nodule bacteria is directly affected by soil pH (Mengel et a1., 1978; Keyser et a1., 1979). Generally, rhizobia prefer a soil pH between 6.0 and 8.0 and may fail to develop in strongly acid soils (Allen and Allen, 1950). Experiments conducted with cowpea cv. California Blackeye no. 5 produced maximum nodule and pod number per plant within the pH range of 646 to 7.6 (Hwan-E Joe and Allen, 1980). Nodule size decreased at a pH of 7.5 and above , and roots became more fibrous. At acid pH levels 442 to 5a8, inoculated plants had 83% fewer nodules than those grown in the optimum range. Similar results of non-viability of rhizobia at acid pH levels were obtained with Medicago sativa (Munns, 1968). Acidity reduced nodule numbers at pH 5J5and nearly prevented nodulation at pH 4..5 However, root growth and production of root hairs were not.affected within this pH range. Also, a pH of 4.4 inhibited root hair infection by rhizobia and subsequent root curling was prevented, while raising the pH to 5m4 allowed these processes to continue. After root curling occurred, the pH could be lowered again to 4.4 without hindering the normal completion of the nodulation process. 100 Soil Nitrogen In the tropics cowpeas are normally grown on soils deficient in nitrogen. Yet some soils may contain sufficiently high amounts of available N to affect the nodulation process. Studies on the relationship between nitrogen and nodulation in cowpea are largely based on applied fertilizer N. Generally it is the timing and amount of applied N which affects legume symbiosis (Eaglesham et al:, 1983). Cowpea usually form nodules about 9-11 days after germination. Investigations on the nitrogen dependence of growing cowpea have shown that seedlings are dependent on reserves within the cotyledons (Ndunguru and Summerfield, 1975) which are normally shed within an few days after emergence. Since dinitrogen fixation is not detectable for about 15 days after planting (Summerfield et a1., 1976), this means that young plants may undergo a 'nitrogen hunger' period (Minchin et a1., 1980; Summerfield et a1., 1976; Pate and Dart, 1961). Applications of small starter doses (20-30 mg/plant) of fertilizer N has been suggested as a means to boost N2 fixation and yields in cowpea (Minchin et a1., 1980; Eaglesham et a1., 1977; Dart and Mercer, 1965; Pate and Dart, 1959). The benefits seem partly to arise from increased root production which results in an increase of infecti and Da fertil Iiidon H: vary ( have b plants (Somme J of P. nodul mutan pFESEI in a 1 101 infection sites for soil Rhizobia (Kang et a1., 1975; Pate and Dart, 1961). Applications of large amounts of fertilizer N depress nodulation (Tewari, 1965; Dart and Wildon, 1970; Pate and Dart, 1961). However, genotypic responses to applied N fertilizer vary (Eaglesham et a1., 1983) and some nodulating genotypes have been found to be vegetatively equal to non-nodulated plants, even producing significantly greater seed yields (Summerfield et a1., 1977; Dart et a1., 1977). Jacobsen and Feenstra (1984) used mutagenic treatment of P; ggplypp var. Rondo to induce variability for nodulation in this species. A monogenic and recessive mutant was discovered which nodulates efficiently in the presence of 15 mM KNO3- This mutant nodulated abundantly in a nitrogen-free medium as well. In Africa, where cowpea production is mostly on a subsistence level, it is not possible for the farmer to alter the many environmental factors which affect nodulation and yield. Irrigation systems are often too expensive or water sources may be limiting. Day time air temperatures may reach highs of 35°C while soil temperatures may rise above 42°C. Such conditions decrease nitrogen fixation (Dart and Mercer, 1965; Minchin.et:al., 1980). Soil amendments to alter pH or nitrogen status normally are not practiced among subsistence farmers. The practical solution to improving dinitrogen fixation and 102 yields in cowpea consists in breeding and selecting within the range of genetic variability present in landraces and cultivars. Host Factors Affecting Dinitrogen Fixation Although differences for N2 fixation exist among legume host-strain symbiosis (Neves et a1., 1981; Bethlenfalvay et.al., 1978b; Pate and Dart, 1961), it is the host which appears to play a more important role in controlling the symbiosis (Graham, 1982). The variable nodulation responses observed within legume hosts and cross-inoculation groups have provided material for genetic studies and screening legume host genotypes for increased N2 fixation has been accomplished using indigenous soil Rhizobium (Westerman and Kolar, 1978; Lie, 1981; Nangju, 1980). Cowpeas are thought to have originated in Africa and cultivated for several generations. Therefore, indigenous rhizobia-host associations are readily formed and in most cases inoculation is not needed (Rachie and Roberts, 1974). Vorhees, in 1915, first noticed genetic differences among cultivars of soybean (Glycine max) for nodulation ability. Plots containing strains of R; japonicum were planted with six cultivars of soybeans. All cultivars except one, 'Haberlandt', nodulated normallyu Thus, 103 selectivity for nodulation of host genotypes was demonstrated. Further studies revealed involvment of host genetic factors at various stages of the symbiotic process. For convienence, these processes are divided into initiaiton of nodules, nodule development and effectiveness of nodulation. Initiation pf the Symbiosis Strain specificity between rhizobia and soybean genotypes was reported by Caldwell and Vest (1968). Plants grown for sucessive years in different locations were predominately nodulated by Rhizobium japonicum strains of specific serogroups. Nangju (1980) studied soybean nodulation response to indigenous rhizobia in Airica. Exotic cultivars from the U.S.A. failed to produce nodules unless inoculated, while those from Indonesia nodulated quite well with indigenous strains of Rhizobium. Specificity in nodulation has also been reported in Lupinis sp. (Lange 1961). A collection of indigenous Australian rhizobia revealed that L; digitatus, L; albus and L: pilosus grouped together on susceptibility to nodulation while L; luteus and L: angustifolius were rarely nodulated by the different sample strains. Genetic control of nodule initiation has been 104 identified in various legume crops. Simple recessive genes for failure to nodulate with compatible bacteria have been reported in soybean (Weber, 1966), peas (Lie, 1971), and red clover (Nutman, 1954b). Nodulating and non-nodulating soybean isolines have been used to study the plants' role in nodulation (Weber, 1966). Recent work by Dazzo (1980) has shown that plant lectins appear to be involved in bacterial associations. The lectin-recognition hypothesis states that infection site recognition involves binding of specific legume lectins to unique carbohydrates found only on the surface of the appropriate rhizobial symbiont (Dazzo & Hubbell, 1982). This hypothesis is being tested using the R; trifolii-L repens (Clover) association. Recent studies have shown that mutation of certain plasmid genes of R; trifolii encoding essential nodulation functions result in alteration of the bacterial polysaccharides leading to significant loss of clover lectin-binding activity (Dazzo et al., in press). Nodule De ve lopment Host controlled factors affecting nodule development involve type, number, size, time to first nodule appearance, and nodulation patterns (Graham, 1982). Dart 105 (1975) classified nodule types into three distinct morphological forms which are determined by the host genotype. Elongated, cylindrical types normally occur on clove, peas and alfalfa and spherical types are found on soybean, cowpea and beans. Collar nodules are found in lupin and are unique in that the nodules grow around the root. Although various environmental factors may influence the number and size of nodules, some studies have clearly demonstrated genetic control of these characteristics by the host. Nutman's (1967) investigation of varietal differences in the nodulation of subterranean clover showed high nodule number per plant to be dominant over Sparse nodulation but of probable complex inheritance. Experiments using ngggplpg yplgggig demonstrated differences in number and weight of nodules produced for various Rhizobium strains tested (Graham, 1973). White clover (Trifolium repens I") inoculated with an effective rhizobial strain produced a wide range of variation in total numbers, size and weight of nodules when grown in a nitrogen-free medium (Jones, 1962). An inverse relationship exists between numbers of nodules and size of nodules (Nutman, 1967; 1981). It has been suggested that this is the means by which the host ensures a proper amount of nodule mass to support plant growth (Nutman, 1981). How a host regulates nodule mass is 106 still under investigation, but it is most likely influenced by plant hormones and environmental factors (Dart, 1975; Dart and Pate, 1959). The time from sowing to first nodule development has been studied with the objective of selection for early nodulation genotypes. Host-controlled differences in time to first nodule appearance have been reported in T; subterraneum and Stylosanthes spp. (Graham, 1982). Crosses involving early and late-nodulating parents in subterraneum clover displayed polygenic inheritance (Nutman, 1967). Jones (1962) obtained a range of variation for nodule appearance from 4 to 16 days in Trifolium repens cv s100 Nomark when inoculated with an effective rhizobial strain. Earliness of nodule formation was significantly correlated with harvest dry weight. Nodulation patterns on legumes have been attributed to host-regulated genes. In studies involving Lupinis, Lange and Parker (1960) distinguished three species (L; lueteus, L: mutabilis and L: digitatus) which exhibited preferential siting of nodules when inoculated with a Rhizobium strain, although a fourth species (p; angustifolius) did not. Distribution of nodules occurred either on the crown, tap root or lateral roots. Genetic regulation of nodulation patterns in species of Phaseolus and soybean was reported by Bhandari and Sen 1966). Three distribution patterns were identified: 11 FE 107 localized, diffused and mixed. A single gene was responsible for control of nodule siting in Phaseolus species. Crosses between L radiata var. N.P. 28 having a localized nodulation pattern and V; twilobata with a diffuse pattern displayed dominance for the diffuse pattern in the progeny. The ability to breed genotypes with preferential siting of nodules may be very important under semi-arid conditions where surface soil moisture is limited and surface temperatures are high. In an interesting investigation reported by Doku (1970), the effect of selfing and hybridization on the nodulation of cowpea was examined. Cowpeas are normally self-pollinated with occasional outcrossing. Upon examination of local cowpea varieties from Ghana, one variety was discovered in which a considerable degree of outcrossing occurred. Self pollination significantly increased the mean number and weight of effective nodules per plant as well as the mean weight of each effective nodule. Although the magnitude of these increases began to decrease by the S4 generation, a large increase was again observed when two selfing lines of the S4 generation were crossed. 108 Effectiveness pf the Rhizobium-Host Symbiosis The presence of root nodules does not predispose the legume host to nitrogen fixation. Nodules may be present but ineffective or inefficient. Ineffective nodulation means that infection on the root hair by rhizobium bacteria occurs and nodules are induced, however, nitrogen fixation does not occur (Caldwell and Vest, 1977). Investigations with red clover indicate that a single recessive gene (11, i1) is responsible for an ineffective nodulation response when inoculated with Rhizobium trifolii strain A (Nutman, 1954b). Homozygous i1 11 plants gave an effective response when inoculated with other strains of R; trifolii. Additionally, a recessive suppressor, m, was found which restores i1 11 plants to complete effectiveness with Rhizobium strain A. Caldwell (1966) reported three host genes controlling ineffective nodulation in soybeans. Inoculation of the cultivar 'Hardee' with Rhizobium japonicum strains of serogroups c1 and 122 yielded stunted plants with ineffective nodules. Inheritance studies of F1, F2 and F3 generations of 'Hardee' x normal nodulating genotypes indicated that ineffective nodulation is conditioned by a single dominant gene R32, This gene is distinguishable from R31 which conditions nodulating or non-nodulating response 109 of soybean to most all strains of & jgponicum (Caldwell and Vest, 1977). A third gene, R33, controls ineffective nodulation in 'Hardee' when inoculated with B; ,japonicum strain 33. Inefficient nodulation is described as normal nodule formation on the legume root with subsequent lack of nitrogen fixation. One such example involves the soybean cultivar 'Peking' which, when inoculated with _R_. japonicum strains of serogroup 123, ferms large normal-looking nodules yet nodule interiors remain white or light pink, and small, often insufficient, amounts of nitrogen are fixed. Plants of 'Peking' remain chlorotic and small, displaying typical symptoms of nitrogen deficiency (Caldwell and Vest, 1977) Genetic Variability for N2 Fixation Intraspecific variability for nitrogen fixation among legume host genotypes has been reported in several seed legumes including Vigna unguiculata (L.) Walp. (Zary et a1., 1978; Minchin et a1., 1978; Graham and Scott, 1982), Phaseolus vulgaris (Westerman and Kolar, 1978; Rennie and Kemp, 1981; Felix et a1., 1981; McFerson and Bliss, 1981), Cicer arietinum (Rupela and Dart, 1982), Pisum sativum (L.) (Bethlenfalvay and Phillips, 1979), Vicia faba (El- 110 sherbeeny et a1., 1977). Glycine wightii (Nicholas and Haydock, 1971), and Glycine max (Wacek and Brill, 1976). Screening of 100 cowpea genotypes, inoculated with mixed strains of Rhizobium, for variation in N2 fixation revealed significant differences among host plant genotypes (Zary et a1., 1978). Differences were obtained regardless of whether the criterion used for measurement was nodule mass, nodule number or nitrogen fixing activity measured by acetylene reduction. Consistent differences among performances of individual genotypes in replicated experiments demonstrated evidence of genetic control of the trait and the possibility of breeding for enhanced N2 'fixation in cowpea. Graham and Scott (1982) compared 12 cowpea varieties in.a time phase study for nodulation ability, dry matter production and N-accumulation under field conditions. Results indicated that when plants were completely dependent upon symbiosis for their nitrogen requirements, high and low N-fixers could be readily identified. A strong correlation was found between total plant N and nodule weight at 42 days after planting. This sampling date also represented the time of maximum dry matter accumulation. Studies involving Vigna mungo report the existence of genetically controlled intra-cultivar variability for dinitrogen fixation (Fernandez and Miller, 1983ab). Random plant samples were selected from two cultivars and two 111 hybrid populations were obtained by cmossing. Nodule number, weight and plant specific activity showed significant differences between F1's for both populations indicating that the parental cultivars were not genetically pure for these variables. In Phaseolus vulgaris at least three factors are thought to be contributing to the variability in N2 fixation: supply of carbohydrates to the nodule, relative rates of N uptake from soil and time to flowering (Graham, 1981). Climbing cultivars appear to transport a large portion of their carbohydrates to nodules compared with plants of other growth habits. Bush bean cultiyars were found to absorb soil nitrogen more rapidly than climbing types (Graham and Rosa, 1977) which might cause a decrease in carbohydrate supply to nodules, leading to lowered fixation. Maturity rates influence nitrogen fixation through competition of developing pods for photosynthates needed for nodule development. Thus, delaying flowering may be one way of increasing seasonal fixation (Hardy et a1., 1973). Felix et al. (1981) compared cultivars of common been from different geographic locations for nitrogenase and nitrate reductase activities under field conditions. Observations indicated that, on the average, tropical cultivars have a higher level of acetylene reduction and a lower nitrate reductase activity than temperate cultivars. 112 The increased nitrogenase activity of the tropical cultivars was due to a higher amount of nodules. These results support the idea of breeding cultivars for improved N2 fixation in tropical countries where the use of N- fertilizer may be limited. Screening for enhanced levels of nitrogen fixation and seed yield in common bean has already been accomplished (McFerson and Bliss, 1981; McFerson et a1., 1982). Populations were developed by the backcross-inbred method. A number of homozygous lines were developed similar in most characters to the recurrent parents. Evaluation of these lines under field conditions for nitrogen (CZHZ) fixation and yield resulted in considerable variation for both traits, and transgressive segregation was observed. Examination of eight varieties of Vicia faba in association with a standard strain of rhizobia and also with the application of N-fertilizer gave large differences in varieties (El-Sherbeeny et a1., 1976). Dry matter production, %N and total N uptake differences were apparent between varieties, Rhizobium and mineral N treatments. Similar results were reported by Nicholas and Haydock (1971) in their studies of Glycine wightii. Variations within lines of Q; wightii were measured by dry weight per nodule and time to nodule appearance. Fer’ most legume crops, exploration into the variability of N2 fixation and the possibility of breeding 113 for enhanced fixation is just beginning. Most of the studies conducted to date are based on improved cultivars. The genetic potential for dinitrogen fixation harboured within land races and wild relatives of common legume crops has yet to be investigated. Breeding Objectives Breeding programs involved.:h1 enhancement of dinitrogen fixation ultimately aim at improving yield. For grain legumes the desired goal is increased seed production while forage legume breeding is aimed at increased vegetative production. Since the improvement of dinitrogen fixation would involve various host characteristics, including morphological, physiological and agronomic traits, the task becomes somewhat complicated, although not impossible. Furthermore, one may proceed from either side of the symbiotic association - by improving the host or improving the bacteria. Provided that a good agronomic host is already available, searching for a strain of Rhizobium that will fix the greatest amount of N2 With a particular genotype is one possibility for improving fixation (Caldwell and Vest, 1977; Nutman, 1981; Graham, 1982). However, this procedure would not be feasible for cowpea since the host is susceptible to infection by a 114 large range of Rhizobium sp. within the cowpea cross- inoculation group. Furthermore, control over indigenous soil Rhizobium would not be possible. Genetic resistance to all but a select group of highly efficient Rhizobium strains has been suggested as another means of improving dinitrogen fixation. Cultivars developed by this method would have very specific adaptatibility and require the compatible strain to be present or applied as inoculum (Caldwell and Vest, 1977). Such work is now being investigated for soybean (Devine, 1977). In Africa, breeding for favorable gene combinations that are least likely to be affected by bacterial strains seems a plausible approach. Dart (1975) has suggested several methods for improving host characteristics which affect nodulation. Endogenous plant hormones and Rhizobium-produced hormones both contribute in regulating the quantity of nodule mass produced in a symbiotic association. One approach to improving fixation is to alter this balance so more nodules are produced. Placement of nodules, which is also a host regulated trait, may be altered so that the bacteroids are not exposed to adverse soil conditions such as high surface temperatures. While plants compensate for nodule number by increasing nodule size, plants with most nodules are thought to fare best. Breeding for an expanded root system would create more 115 potential infection sites by soil rhizobia. Equally important in selection for increased N2 fixation is plant architecture. Graham (1981) described differences in growth habits of Phaseolus vulgaris which affect several aspects of the dinitrogen fixation process. For example, climbing cultivars were found to transfer a greater amount of non-structural carbohydrates to nodules than plants of other growth habits. No such investigations have yet been conducted in cowpea, but one would need to consider such differences (should they exist) in light of farmer's preference for plant types. Improvements in plant architecture which would decrease shading to lower leaves and increase photosynthate supply to nodules is another alternative. Because of the integrative relationship between host- Rhizobium-environment, breeding for improved fixation requires special considerations and may be best solved by cooperative efforts between breeders, microbiologists and agronomists. In 116 Selection Criteria and Methods Time pf Sampling Unlike breeding for disease resistance or improved grain quality, the symbiotic process is progressive throughout the life of the plant and intimately connected to its many growth stages and physiological processes. Rates Of N2 fixed are dependent on the stage of plant growth (Caldwell and Vest, 1977). An understanding of nodulation ontogeny for a particular legume crop is essential in order to determine periods of maximum fixation or fixation 'stress', thereby identifying optimum selection periods or stages where breeding efforts may be diverted. Previous studies in cowpea have identified such stages. Time phase studies have shown that fixation rates increase exponentially from the third to the sixth week at an average daily rate of 20% (Zary et al., 1978b). Maximum N2 fixation occurs about 42 days after planting and this period correlates well with maximum plant dry matter accumulation, total N and nodule weight (Graham, 1982). After 6 weeks, activity declines rapidly through pod fill to senescence (Zary and Miller, 1978b). The best time for screening cowpea, then, would be at full flower (Zary et a1., 1978ab;1980). In addition to seasonal patterns, 117 diurnal patterns of N2(C2H2) fixation have been observed in cowpea (Zary and Miller, 1980).. Maximum diurnal activity was found to peak at 1200 hours when measured at both 34 and 53 days after planting. These results indicate that selection of high N-fixing genotypes by the acetylene reduction method needs to be coordinated with both seasonal and diurnal patterns of fixation. Methods pf Measurement Selection of superior nodulating genotypes depends not only on the time of sampling but on the traits sampled and the method of measurement. Variation in N2 fixation among host legumes has been measured by acetylene reduction (Pacovsky et a1., 1984; Westerman and Kolar, 1978; Zary et al., 1978b; Lawn and Bushby, 1982), H2 evolution (Layzell et al.,1979; Pacovsky et al,,1984), total N (Rennie and Kemp, 1981; Graham, 1982), and quantitatively viz. plant dry matter, nodule weight and nodule number (Graham and Scott, 1982; Nicholas and Haydock, 1970; Rennie and Kemp, 1981 and others). No general agreement has been reached among legume breeders as to which method is most reliable, and decisions are no doubt influenced by time and resources. 118 Selection 2; Traits Various investigators have utilized different host characteristics in selecting genotypes with superior nodulation ability. Graham and Scott (1982) found a strong correlation between total N and shoot weight in cowpea. Presumably, varieties of a higher N-fixing ability can also accumulate'more dry matter so size and vigor of plant tops may be used as a measure of greater N-fixation. Total N was also highly correlated with nodule weight. The absence of correlation between total.hland nodule number suggests that itis nodule mass which is the more important criterion in assessing nodulation (Graham, 1981). Zary et al. (1978) used four indexing criteria in defining the extent of genetic variability among 100 cowpea genotypes. Specific activity was found to be the most consistant indicator of N2 fixation followed by nodule mass and plant.top dry weight. Nodule number was found to be the least consistent. Nodule weight has also been suggested as a suitable characteristic for selection in Glycine wightii (Nicholas and Haydock, 1970). Westerman and Kolar (1978) studied variations in dinitrogen fixation in Phaseolus vulgaris. Their results showed N2 (CZHZ) fixation to be related to average seasonal nodule weight and plant dry weight near physiological no 119 maturity. Moreover, total N uptake was signigicantly related to seed yields of the cultivars and cultivars with high seed yields also had high N2 (CZHZ) fixations. This suggests the possibility of selection based on yields only. Evaluation of Phaseolus vulgaris under two low temperature regimes showed the amount of dinitrogen fixed to be correlated with leaf area and leaf and shoot weight (Rennie and Kemp, 1981). Similar relsults were reported in studies involving four Asiatic liggg species (Lawn and Bushby, 1982). These studies indicate the need for measuring a number of host variables in initial screenig experiments and careful selection of the most reliable traits for further breeding purposes. Genetic Vulnerabi l ity Mostly all studies addressing the problem of dinitrogen fixation are based on data obtained from advanced or cultivated legumes. The number of plant introductions is usually small and those that adapt successfully to new environments are fewer yet. It is these successful introductions which are further used for breeding (Lie, 1981). The specificity of host genotypes for particular Rhizobium strains has been reported (Nangju, 1980). Narrowing of the genetic base of host species may 120 tend to favor certain strains of soil Rhizobium. Although the bacterium may live as a soil saprophyte, Nutman (1967) has reported that,its distribution.and multiplication is closely related to the presence of a compatible host. Over time, the disappearance of indigenous leguminous hosts may alter the population of soil Rhizobium. Moreover, the remaining symbiotic associations may not be those of highest dinitrogen fixing ability. Rationale for the Present Work Studies ofjgenetic diversity in landraces of cowpea are lacking. Given the amount of variability found in cultivated varieties for N2 fixation, one would expect even greater diversity in land races. This study attempts to define the range of variability for indigenous and exotic lines of cowpea currently grown in Botswana and to identify high and low nitrogen-fixers for use in future breeding programs. MATERIALS AND METHODS Origin gprlant Material One hundred lines of cowpea were grown in Botswana during 1982-83 to evaluate the nodulation capacity of each (Appendix F). Forty nine percent of the accessions were local landraces collected from Botswana farmers while fifty lines were derived from breeding material and world collections held at the International Institiue of Tropical Agriculture in Nigeria and SAFGRAD in Borkina Faso. The check, Blackeye, was originally'introduced into Botswana from California some time ago but has since undergone considerable adaptive changes. It has been one of the few commercial introductions of cowpea into Botswana and, until recently, the only seed source multiplied for planting. Field Plot Design A split plot design was used for the experiment with nitrogen treatments as the main plot and 100 varieties as sub-plots. Two replications were used and the entire 121 122 experiment was planted at two different locations in Botswana (Sebele and Mahalapye). Rows were spaced at 1.M with 20 cm within-row plant spacings. Plots consisted of a single 6 M row. Varieties were randomized within treatments and treatments were randomized within replications. A total of 2650 M2 area was required per experiment. Phosphate was applied pre-plant as superphosphate (P205 = 10.5%) at a rate of 250 Kg/ha. Planting at Sebele began on November 4 and terminated on November 5, 1982. At Mahalapye, planting occurred on December 1, 1982. Plots were sprayed twice to control pod-sucking bugs using NOGAS. (1 Ill/l H20 ULV). No Rhizobium inoculum was applied since one objective of the experiment was to determine genetic differences in nodulation between cowpea varieties under conditions of natural Rhizobia, particularly since inoculation practices are unknown among Botswana farmers. Nitrogen.wasrapplied.by sideldressing in the form of lime ammonium nitrate (LAN = 28%) at a rate of 100 kg N/Ha, seven days after planting. Description pf the Experimental Sites The two sites used for the experiment are located approximately 200 Km apart. Mahalapye (site 1) is located at longitude 26°48' and latitude 23°04'. Total rainfall 123 received during the 1982-83 growing season was 280 mm. Sebele (site 2) is located about 10 Km north of Gaborone, the capital city, at longitude 25°57'1and latitude 24°34H Rainfall received at Sebele during the course of this experiment exceeded 450 mm. Rainfall distributions for both sites may be found in appendices C and G. Sampling Procedures All varieties were sampled for nodule number, nodule fresh weight and dry weight, root weight, shoot fresh weight and dry weight at 2, 4 and 6 weeks after planting. For site 1, four random plants per variety were dug and evaluated at each of the three sampling times for both replications and for each treatment. Total values were recorded and later averages were calculated and used for analysis. At site 2, similar procedures were followed except that hardened soil conditions by the 4th week after planting prohibited the removal of plant roots without damage. Therefore, only shoot fresh weights and dry weights were recorded. By the 6th week after planting, the soils became even drier due to the lack of rainfall. Each plant which was sampled had first to be moistened around the roots by applying buckets of water. Since this entailed a great deal of extra time, only two plants per variety were 124 sampled. As before, total valuesifor all six variables were recorded and later averages were calculated for use in analysis. Since some local varieties are indeterminate in nature, harvesting continued over a period of several weeks. Five random plants per plot were harvested and the following data recorded: total number of pods per 5 plants, pod dry weight, pods per plant, total seed weight, seeds per pod, total seed number, yield per plant and 100 seed weight. Some insect damage occurred due to pod-sucking bugs despite insecticide treatments. In order to account for this damage, three measures of 100 seed weights were recorded: a random sample of damaged and non-damaged seed, a selected sample of non-damaged seed and a selected sample of damaged seed. Harvesting at site 2 was conducted in the same manner except that 10 random plants per plot were sampled for all varieties. RESULTS AND DISCUSSION The means and ranges for the six variables examined during the statistical analysis of the data at site 1 are summarized in Table 13. Only 60 of the varieties were included in the evaluation. Applications of nitrogen fertilizer at a rate of 100 Kg N/Ha produced no statistically significant differences from treatments with 0 Kg N/Ha. This effect carried through for all sampling periods at 2, 4 and 6 weeks after planting. The area occupied by the experiment had been sown with sorghum the previous season and soil analysis of mineral and mineralizable N were low (Appendix H). However, supression of nodule growth was visually'evident, particularly by the fourth and sixth week after planting. Genetic diversity for nodulation characteristics were expressed for all. variables measured and through all sampling periods. Thus, even at two weeks after planting, it was possible to categorize genotypes with respect to their capacity for nodulation (Table 14). Significant interactions were observed between varieties and nitrogen treatments for some of the variables at different times during the course of the experiment. Although these 125 693an see: N an 33 new 03 ha moi—33...:- 126 m: as a: a: m: a: - Ano.vama on..on.coo.n Aflo.nu oo.oo~-na.~n an..ana sua.n~.oa~.~ nn¢.u “an.oao cn~.o coo.n.° nnh.o can.o..ona.o o-.o~ z- on~.«ono~.~ Nuc.n~ ma.onnsoa.~c nao.~an oma.nuuoa~.~ ~ac.o na«.ono asc.o neo.oao no~.o ooc.~e.o -n.na z axes: o a: a: an m: an a: . Ano.eamn soc.oao~o.o no~.o can.nuooo.c -o.~ ono.coooa.o an..o oo~..-~oo.o ~oa.¢ coo.n~ucc~.~ as.“ cn~.osrocn.~ “ca.n~ z- on..ouo~o.o oo~.o ooo.ouon..o cao.~ ono.~.on~.o o...o oo5.n-~oo.o s.o.° coo.nn.o°n.a as.“ on~.~«u¢no.o oom.~u 3 ans»: N nun-d M ecu-u .NI eased m. ended H can: m. one: u uses-nun. “a. season was Leona flue nausea aaann noosm any sauna: coo: aAav season .m. season uasasz .Hauoz an: agave: amuse ounce: Agnew-415 nausea.— ueuuo one!- o use a an unannoun- 00533 no nous-eel noun-«woe no noun-u one one... .3 02:. 127’ .mxmma N uw muwv you OH NO Houumu a an vwfiaafiuaazm e .mam>fiuommmwu .me>uH NH vam m any um undefimwcwwm««.* HHH.mo oao.o~mH ~m~.m moo.o aoH.o HHm.om HHH a uouum Haw.mm thmu.oomm nom.n moo.o HmH.o eH~.HH hm > x z Hoo.~m ««mHm.oHom «Hooo.~H «Hoo.o «HHHN.o *«mmo.~nH hm H>V moHumHum> a~m.aHH HN~.- NHH.~ «mo.o mqn.~ mnw.w~m H a uouum mom.~mm emm.HHmm ~o~.o o¢o.o Hem.o~ omo.mmnH H sz :owouqu on.H¢ oe¢.coq e-.m mmo.o omo.N ¢-.OHH H mumUHHaom mxwm3 o noo.o me.o aoc.o o-.o oao.m mmm.«m mHH p uouum Noo.o HHN.o Hoo.o «m~m.o ««Hoq.¢ «Ham.mm Hm > x z «eoao.o ««ch.fi *«mmo.o ««-~.o ««eo~.m~ ««qu.oaa mm A>V mmwumfium> mmm.o QHH.O¢ mNH.o o~m.H m¢¢.mHH mmm.o~m H a uouum men.o ~wa.me «mm.o Bho.o eco.H~ moc.- H sz cowouqu “sm.c moo.mq HmH.o hum.n Hom.om m~o.mH H oumoHHamm mxmmz N Hwy Hwy Hwy Hwy Hmv nonasz Hg mouaom uamHms ustma ustas auanmz mucwHo: oHsuoz hue uoosm Ammum uoonm uoom hum «Hawoz ammum masvoz mmumamw cum: .Aomamawnmz ”H 3me wcwunwan you: 333 0 van N um muwumfiuouomuwso coaumaavo: pow 02622, no 39323 .3 «Home 128 interactions were not consistently significant, there is an indication that nitrogen treatments were having varying effects depending on the variety. Similar results were obtained at site 2 (Table 15 and 16) where sampling was carried out at two and six weeks after planting. Again, no significant differences were obtained between nitrogen treatments although, environmentally, this site was quite different from site 1 in terms of rainfall and soils (Appendices C & G). A combined analysis of variance over the two locations indicates that the plants performed differently between the two sites for some of the measured characteristics (Table 1?). Initially, this difference was mainly expressed in nodule number, fresh weight and dry weight. However, by the sixth week, differences in genotypic responses between environments began to show in plant biomass. Throughout the experiment, ,variety x location interactions were significant for most characters measured. In terms of the objectives of this experiment, the differences in environments allowed for the identification of superior- nodulating genotypes which displayed a stable performance, despite soil and climatic site differences. Simple correlations were calculated between all nodulation variables for both two and six week sampling data and for both locations. As expected, nodule number was highly correlated with nodule fresh weight throughout the 129 .02 no wouu-u a an .333»?! o...- 003-? m2 m2 m2 m2 a: mz . “no.van onn.nn-onc.n oHH.~H coo..HH.oon.- aeo.a~ on..~H-oHH.~ Hna.o ann.HN.o unn.n can.nnuo neo.~ coo.Ho.o oqm.u z- con.chooH.c «no.a ooo.no~.oOn.oH «no.~o osH.nH.¢nw.c oo~.o ooo.on.o NoH.~ ooo.os.o «No.n ooc.~s.o 5H0.» z .xoca 0 m2 m: a: a: m2 u: u Ano.van con.oao¢o.o ohH.c coo.N-ooH.o ~o~.H coo.ouo-.o ooc.o nso.Hao oo~.o con.o-o coH.H con.oH.o Nu~.n z- can.a.o~c.o ouH.o can.~.on..o ao~.H oas.o-c-.o nae.o con.Huo o-.c coo.n-c Hna.o n~.~Huo huq.~ z .x..: N nun-u m. can: M 0.3 m. on...“ M Incas M :5 M use-use: an. usuHoa nun ”coca Hue usuH.= ;.uuu gonna Hue “Aswan “can name ugmHon .Huo uguHua tcaaaz «Have: nun oHauoz =..u~ .Havoz $30303 nanus-u.— 90qu .102.- 0 van a no uncuooouun cox—boo no vows-col noun-«nab no can...» can 330: .3 snack .mosam> mumswm cams pom «OH mo nouomw m hp wowanwuaazm .hHm>Huoonmou .mHo>oH NH van m on» an unmowmficwam««.« 130 omn.oH «Ho.mmo coo.e mom.oH oH¢.mn nhm.om omH a sauna o~e.~H HHH.on mmm.~ oom.HH 05H.me mom.ms mm > x z .«mmH.m~ «H-.¢HoH «.Nam.NH «ccm.mH .«mmH.mh «.om~.~w ma H>v moHuoHum> meo.mmm Hoo.omomm «mo.oe eH¢.HH ¢w~.o~ ooe.H~H H a sauna aom.wms muo.omHmH ~mm.me ocm.omH Ncm.mm~ OHo.oH H sz camouqu eaH.hmo Hmo.meHo cum.oHn HNm.~oH HmH.m-~ HHo.~Hom H muQUHHamm mv—mUB c Hoo.o “co.c NHo.o Nmo.o mmw.o «mo.s caH a sebum Hoo.o Huo.o oHo.o Noo.o mNo.H mmn.s mm > x z «neoo.o «aumq.o «eamo.o ««omo.o «as~m.~ «amam.n mm A>v moauofium> Hoo.ov mam.H os~.o o-.o smH.H amm.mmH H m uouum Hoo.o moo.o omo.o mmm.o Hmm.m Ho~.H~ H sz coonUHz ~oo.o Hm~.o o~o.o Ho~.w Has.oHH oom.cmm H «n.0HHamm memB N va Amy va Amv Amy nonasz mo mousom uanoz uszoa ucmez mocwHoz mustmz «Havoz %un uoonm Smuum uoosm uoom hum oasvoz :moum «Havoz nonmawm coo: .Aoaopomv N oufim you moaumwuouomumno :oHumHavo: mo oocmaum> mo manhuss< .cH ounce 131. .OOH %n voHHaHuHss mum moaHm>m oom.Hc “Hm.memH com.q moo.o mmo.o NHH.nn «mm a uouum nmm.~c ««~m~.-<~ mwm.m qoo.o Neo.o mo~.mm cm HZH moo.om ocm.o~nH nmq.m moo.o nmc.c amm.om on >2 mc¢.mm «emcm.~HmN «somc.n moo.o ««omo.o «awHH.mw on >H «enm.om «amqm.ommm «cHHm.aH «xooo.o «emHH.o «amco.emH on moHumHuw> ooo.m~n omo.mmmnH me.oH mmo.o NNH.H Ham.omu N a uouum sum.nmm wm~.cnm¢H HH~.mN h¢~.o nmq.m 59H.mmm H 2H moa.o «mm.mm «mm.o~ w¢m.o so~.0H om~.~oa H cowouqu smo.mq~ NHo.¢omsH ewm.QOH oeo.o HNo.H www.mceH N AHV mom «mom.weHeH «coc.nmmon¢ «omm.eoq mo<.o mmq.eH acmw.mom¢ H :oHuoooH oxoos o Noo.o HmH.o “Hc.o Noo.o mmo.o mqm.aH «mm A uouum Hoo.o mMH.o mHo.o ~oo.o ««ooo.o «ammm.qm mm >2H Noc.o HmH.o MHo.o ~oo.o «aco.o «Hom.w~ mm >z «moo.o «mqm.o mHo.o ««moo.o ««wco.o ««~mo.mq mm >H «amHo.o «amQH.H «wooH.o «emoo.o ««moH.o «aHHN.q~ mm mmHuoHum> HaH.o ooc.mc HaH.o oHo.o mmo.o «cm.Hmm N m uouum eoN.o «mm.qn MHH.o moo.o oa~.o mum.o H zH m-.o smH.Hm Hm~.o Hoo.o NmN.o eHm.cHH H ammouqu an.c mo~.- wac.o HHo.o sen.o Ho~.HHm N HHV gum mum.o o-.em omo.o «mmm.o «mmH.- «cow.o0HmH H :oHumooH axooz N usmHoz usmHoz uanos auanuz muanoz nonaaz we uuuaom hum zmoum uooz mun nooum 0H=voz uoonm uoosm oHsvoz oHsvoz moumsum com: .N woo H mouHm Scum mocuHum> mo mHmmHmcm vocHnaoo .NH oHnme 132 analysis (Tables 18 and 19). Nodule fresh weight and nodule dry weight were also highly correlated as were shoot fresh weight and shoot dry weight. Therefore, nodule number, nodule fresh weight and shoot fresh weight were chosen as useful variables for ranking genotypes. Root weight was not considered since plant extraction methods under field conditions, such as were used in this experiment, may be insufficient to remove the total root mass of deeply—rooted genotypes. Thus, fair comparisons between individuals could not be made. The sorting of genotypes acccording to mean values of the chosen variables allowed for identification of individuals which consistently ranked among the top or botton third of all accessions. Five genotypes were selected whose performance across sampling periods, treatments and environments were considered as consistently high. Blackeye, the check variety, was among the top performing genotypes. An introduction from IITA, IT81D-985, also ranked high. The three other accessions were local landraces (Tables 20 and 21). Two poorly nodulated varieties, Vita-5 and IT82E-8, were also identified. These lines are introductions originating from Nigeria. Two other lines, IT82E-17 and TVu 3236, were found to be low ranking in site 1 and 2, respectively. Comparison of mean values of nodule number, nodule fresh weight and shoot fresh weight for the high and low- 133 Table 18. Correlation coefficients between nodulation characteristicsaof cowpea grown with and without nitrogen (Site 1: Mahalapye) . Nodule Nodule Shoot Shoot Nodule Fresh Dry Root Fresh Dry Number Weight Weight Weight Weight Weight Variable (NN) (NFW) (NDW) (RW) (SFW) (SDW) 2 Weeks NN 0.408 b 0.705 0.485 0.433 0.442 (0.441) (0.526) (0.478) (0.720) (0.629) NFW 0.478 0.127 -0.162 -0.023 (0.520) (0.409) (0.348) (0.371) NDW 0.456 0.465 0.479 (0.361) (0.419) (0.452) RW 0.555 0.581 (0.511) (0.507) SFW 0.902 (0.895) 6 Weeks NN 0.797 0.721 0.341 0.285 0.192 (0.769) (0.640) (0.409) (0.402) (0.409) NFW 0.795 0.206 0.115 0.097 (0.835) (0.342) (0.334) (0.388) NDW 0.284 0.184 0.192 (0.298) (0.279) (0.328) RW 0.755 0.623 (0.726) (0.794) SFW 0.712 (0.887) aCorrelations exceeding .2732 and .3541 are significant at the 5 and 1 percent levels, respectively. b Values within parenthesis represent nitrogen-free treatments. 134 Table 19. Correlation coefficients between nodu1ation characteristics of cowpea grown with and without nitrogen (Site 2: Sebele). Nodule Nodule Shoot Shoot Fresh Dry Root Fresh Dry Nodule Weight Weight Weight Weight Weight Number (3) (g) (g) (g) (8) Variable (NN) (NFW) (NDW) (RW) (SFW) (SDW) 2 Weeksa NN 0.845 b 0.763 0.024 0.268 0.340 (0.902) (0.887) (0.273) (0.310) (0.210) NFW 0.873 0.030 0.192 0.271 (0.926) (0.239) (0.251) (0.193) NDW 0.032 0.125 0.223 (0.211) (0.237) (0.161) RW 0.372 0.304 (0.641) (0.576) SFW 0.922 (0.918) 6 Weeksa NN 0.732 0.568 0.195 0.027 -0.016 (0.793) (0.801) (0.269) (0.363) (0.314) NFW 0.619 0.197 0.071 0.030 (0.962) (0.199) (0.211) (0.189) (0.198) (0.215) (0.189) RW 0.662 0.475 (0.758) (0.672) SFW 0.694 aValues exceeding 0.2050 and 0.2673 are significant at the 5 and 12 levels, respectively. b Values within parenthesis represent nitrogen-free treatments. 135 .wstcou :H moHocoumHmcoocH oumochH mHmusucoumq :H chosenz n .oHumHuouomuono coHuostoc comm you consumma oconmmoom Hmuou 00 mo uso xcmu .mmmmuosow m mumoHvaH woumHH muonaszw e oH w m m 0 0H HH Homv Hacv m m NHlmNmHH w 3 m N e H 2 33 OH R «H 28 mumSeH mH ANNV mH mH e mH m N mH m N ¢ m ouH> 3oH co co Nm on em on on on oc on so mm mQHm on on em mm oo on me Nm no mm we Ammv oucoom an me me Nm mm mm me an on Ne Hm Nm NmHm NH 88 an 3. 2 8 3 an an as 3 8 «has»: as HONV AmHv an nAomv an we mm mm me as an mm¢nnHmHH szm z: z z: z z: z z: z z: z z: z conomoo< monaoo 2mm: 33»: $9.52 «H302 23m: 23»: “2:52 «Haeoz shown uoonm amoum oHavoz smoum uooam :moum oHsuoz exam: 0 exam: N .nAH ouHmv wcHucmHn nanus exam: 0 and N on monxuocow monaoo wouomHom mo wcchmm .oN mHnt 1136 .mcchmu cH moHoamumHmcoocH oumochH mHmmzucoumn :H muonaszo .mconmmooo Hmuou om mo uso xcwm n .mconmouum Hmuou ma «0 use Hammm mm Hm oN Anny mH Heme H oH m SH m on em~m=>a m Hesv HHoV mH HHHV am HNHV Hm Home n Home m manweH Aaev N mH H a H NH H NH m HHHV N m uuH> 3oH HHHV mm He ma as so am Hm as HNHV ma Ammo mHHm an as «m mm ma mm as mm am am on Ha ouooom so HNV no me as we Heme mm on mu we so HmHm Ammo HoHV mm am so as «a as am Ammo ma ms osmxoaHm mm mm HH Hm we Ha uHNov «a «H mm om mm mmauaHmeH cmH: 2. z z- z z- z z- z z: z z- z conmooo< monsou uanoz usmes umsaaz «Haeoz umemz unHms sensaz oH282 Smuuh Hoop—m Smmuh OHHH—uoz Smmhh Hoop—m fimwuh 0.2.602 nwv—mfl; o wag—063 N .aAN ouHmv wcHucmHa umuwm exam; 0 can N um mommuocmw moaaoo wouomHom mo wcchmm .HN mHnme 137 ranking genotypes demonstrates the differences in nodulation capacity between these two groups (Table 22, 23 and 24). Table 25 summarizes the differences for each location and treatment. Further investigations into the effect of the two nitrogen treatmentscnlthese individuals indicated that, for most genoytpes, nitrogen suppressed nodule formation and growth, as would be expected (Figures 15 and 16), although the effect of nitrogen on shoot fresh weight was not consistent. Since none of the landraces was identified as having a stable, poor performance, Tables 26 and 27 were constructed in order’to determine if, on thelaverage, the landrace population performed any differently than introduced varieties. In comparing treatment means between the two groups, no statistically significant differences were obtained for any of the variables examined. An examination of the differences in grain yield between the two sites was not meaningful since yields at site 1 were severly reduced due to an infestation of pod bugs and the drought occurring at site 2 caused many of the genotypes to shed their flowers and, hence, fail to produce ’pods. The inability to detect nitrogen treatment differences in this experiment may be attributable to different causes. Possibly the sensitivity of the F-test is such that the small degrees of freedom in the error term did notaallow 138 Table 22 . Mean values of nodules per plant at 2 and 6 weeks after ' planting for selected high and low nitrogen-fixing genotypes. 2 Weeks 6 Weeks Mahalapye Sebele Mahalapye Sebele Cowpea Accession N -N N -N N -N N -N High IT8lD-985 32.3 16.1 6.4 5.0 24.1 20.0 11.0 11.0 Blackeye 37.5 20.4 3.5 11.9 29.9 28.9 18.0 29.0 3152 17.0 21.6 5.3 4.1 22.4 30.8 16.0 11.8 3145 29.6 31.0 1.4 5.5 25.1 30.0 23.0 10.8 Low Vita 5 7.0 4.9 0 2.5 7.1 7.6 0 2.5 IT82E-8 9.4 9.4 0 5.8 2.3 8.5 0 9.5 TVu3236 1.6 1.0 6.5 3.3 IT82E-17 6.3 5.9 5.5 6.6 Lsn(.05)a 8.2 3.0 10.0 9.8 aLSD is for comparing means between any two values within the same column. 139 Table 23. Mean values of nodule fresh weight (g) at 2 and 6 weeks after planting for high and low nitrogen-fixing genotypes. 2 Weeks 6 weeks Mahalapye Sebele Mahalapye Sebele Cowpea Accession N -N N -N N —N N -N High IT81D-985 0.63 0.88 0.03 0.02 0.53 0.50 0.10 0.11 Blackeye 0.13 0.11 0.01 0.04 0.49 1.72 0.16 0.36 3152 0.06 0.06 0.01 0.02 0.28 1.10 0.07 0.09 3006—0 0.08 0.08 0.02 0.02 0.50 1.28 0.11 0.18 3145 0.09 0.14 <0.0l 0.02 0.30 1.76 0.18 0.10 Low Vita 5 0.03 0.04 - <0.01 0.08 0.50 - <0.01 IT823—8 0.03 0.03 - <0.01 0.04 0.30 - 0.07 TVu3236 <0.01 <0.01 0.03 0.02 IT823-17 0.07 0.09 0.04 0.36 Lsn(.05)a 0.01 0.462 aLSD is for comparing means between any two values within the same column. 1140 Table 24. Mean values of shoot fresh weight (g) at 2 and 6 weeks after planting for selected high and low nitrogen-fixing genotypes. 2 Weeks 6 Weeks Mahalapye Sebele Mahalapye Sebele Cowpea Accession N -N N -N N -N N -N High IT81D-985 3.8 2.0 1.9 1.4 142.4 196.4 118.0 91.5 Blackeye 4.0 2.5 2.0 1.9 160.8 205.8 47.8 73.3 3152 2.8 1.9 1.7 1.3 187.1 258.9 45.3 84.5 3145 4.7 3.0 1.8 2.0 303.6 260.8 84.5 82.5 Low Vita 5 1.9 1.0 0.8 0.9 163.6 100.0 31.5 86.3 IT823-8 2.6 1.3 1.1 1.2 133.3 118.1 25.3 32.0 TVu3236 0.9 0.8 58.3 68.3 1T823-17 2.1 1.3 144.8 115.6 Lsn(.05)a 0.8 0.4 59.2 36.6 aLSD is for comparing means between any two values within the same column. 14]. «HH.~H m~a.o oaH.HH Hm~.HH Nam.eH Nsm.om oam.mH osm.m~ uanma sue soonm oo~.Hc acm.mm ooa.em omw.mn mm~.HHH mm~.NHH oso.wmm ONH.NHH uzmHus £6600 soosm moa.m mHm.H wes.~ meo.o “NH.S Nos.“ ous.oH omo.a uamHaa boom aHo.o «No.0 Hoo.o meo.o meo.o «Ho.o mc~.o oHH.o uanoa s66 0H362 omo.o omo.o QSH.o HHH.o 5mm.o mmo.o ~N~.H o~s.o uanms amuse 6H.662 ooH.m NOH.~ om~.aH coo.NH New.“ 50¢.H osm.aN oOH.oN Hogans 6H262 @0303 o NHH.o osH.o sH~.o Hmm.o amH.o os~.o cm~.o «He.o sstma s06 soosm Nem.o mma.o oso.H osw.H oo~.H oo~.~ o6~.~ o-.m uanoa amuse uoonm moe.o Nem.o ohm.o mHo.o moe.o mms.o Nam.o omo.o ustus 366m Hoo.o Hoo.ov ooo.o Hoo.o ooo.o soo.o oHo.o mHo.o u£MHoa sue 6H262 oHc.ov oHo.ov «No.8 wHo.o mmo.¢ meo.o Hm~.o maH.o usmHoa gamut «Hacoz ooH.m mmm.o om~.e oo~.e mm“.o hem.“ omN.H~ omc.m~ banana 6H362 memB N z- z z- z z: z z- z oHanum> 36H an: anm oHonmm mammHmnmz .moHumHuouomumso :oHumHsvo: pom mommuocmw wcHXHu :mwouuH: 30H was :mH: mo mosHm> some anouo .nN mHnme 142 .mmmzuocmw wcHwa ammouuH: :st wouomHmm co uanmB shown mHsuoa so umuHHHuummnz mo uoowwm 05H .mH musmHm a In 0.30.. .2; $3.33 278:. 2: 8 2.. 2| 2 2.. 2 2: 2 «to n 3 V N ono N O O n «I «no 3 .1 U 3 3 co; u M 3 o .3. u L a on; ( a I... N‘L an. . .h..D OON 1143 .mmazuocmw wcfixfim :wwouufic 30H wouumamm do unwfima ammum cannon co umufiaauuownz mo uumwwm one oum~o». 2 P m..(fir. > .mma musmam gnuw Cm 1... rug. 10.. w... I... RIVVEIVV SWOOON (0) “19:31“ H5383 11H} .mwmzuocmw wafixwm :mwouufi: cw“: cmuomawm co unwams ammum uOOSm co umNHHHuumwlz mo uuwmmm any .QH muawfim n:- o-.oon :2. 2.3.23 279:. z- z a 2:3nuu - O-.OD H538: LOOHS N73“ 1H9I3M (5) 1&5 .mmmmuoamw waaxfim cowouufia 30H vmuumamm co unwfioz :mmum uOOSm do umuwaauummlz mo uummmm msa .moH musmfim (3)1H9l3M Hsaua oanoh. ml<._._> 2.. z 2.. z p - , a m \ w \ u \\ N \ 10a \ m. ‘\ o L too— L L L rom— loan a u—LOS . up.._HHHu lama 1M6 .ceaaou can. on» c«:»«: uoaun> can an. cooauop nanus ”sauna-co non nu and a .oo~ uo nouu-u - an voaunuuaaa on. use.) N a. gauge: snouu agave: no nuann> cues. m: m: m: mz m2 m2 . $3.33 www.cu nms.n~ can.ocu soo.-u nus.» qnn.o ena.o ecc.o nwo.c esa.o Hom.ma con.oH nounuvcua sne.au mnm.- cmo.onu own.auu who.” nan.o he~.o nno.o cam.c Hn~.o omo.oa mo~.~n ocoauuavouucn axes: o m2 m2 m2 2 m2 «z 3933 oa~.o nm~.o ~oa.~ o~w.~ oce.c om<.o Hoe. Hoo. con.c omq.o oo~.nn oo.ua concave-A oc~.o non.o coo.~ ~o~.~ e~¢.o cow.c doc. Moo. can.o onm.c osu.e~ s<.n~ acouuuavouucu axon: N z: z z: z z: z z- . z z: 2 2| 2 ”my nanus: va uzmuoz Amy unwwoa Inwv panama «Amv uswqua unnaaz ouavoz hug uoozm auuum uoozm uoom bun ouavoz nauum oaavoz .Aoamuansax "A ouamv ”caucnun nouns axon: o vac m u. noguouua> vuuavouucu van qua-hang“ nausea cooauop coa«uqnaou uOu acne. ucoauuuua .oN «finch lb?’ .casaoo mea may cacuws moaam> 03u mam cmmaumn mamma mafiHwQEOU pom ma 9mg A .mcmma mxmmz c was N uom oca mo nouowm m hp umaanwuana mum mm=Hm>m m2 m2 m2 m2 m2 m2 pamo.vnmg mmm.HH qu.OH «Hm.a~ qm¢.mo omH.h mqm.o m¢w.m Hmm.N mmw.m mmm.o «ca.m m-.m mmomuuamg oNH.~H ¢-.¢ Haq.m~ maa.oo mmh.e wm¢.m Hem.~ Noo.~ mHm.c cue.“ mHo.m omo.m msofiuosvouuaH mxmmz 0 m2 m2 m2 m2 m2 m2 Amo.vamg omH.o «NH.o mom.H ¢MN.H ~m¢.o ¢mq.o mHm.o NmN.o hqm.H mq¢.o mme.m mmq.~ mmumuccmq Nua.o th.o oo~.H oem.H «Ne.o -¢.o mm~.o «-.o m~¢.o ¢Ha.o