search
for
 About Bioline  All Journals  Testimonials  Membership  News


African Crop Science Journal
African Crop Science Society
ISSN: 1021-9730 EISSN: 2072-6589
Vol. 6, Num. 2, 1998, pp. 119-127
African Crop Science Journal, Vol

African Crop Science Journal, Vol. 6. No. 2, pp. 119-127, 1998

HERITABILITY AND GENOTYPE X ENVIRONMENT INTERACTION FOR YIELD AND COMPONENTS OF A YIELD MODEL IN SEGREGATING POPULATIONS OF GROUNDNUT UNDER SEMI-ARID CONDITIONS

B. R. NTARE and J. H. WILLIAMS1

ICRISAT -Kano, Sabo Bakin Zuwo Road, PMB 3491, Kano, Nigeria;
Email: ICRISAT-W-Nigeria@cgnet.com

1Peanut CRSP, Georgia Experiment Station, Griffin GA 30223, USA

(Received 24 February, 1998; accepted 1 April, 1998)

Code Number:CS98013
Sizes of Files:
      Text: 68K
      Graphics: No associated graphics files

ABSTRACT

Limited information is available on heritability and genotype x environment interaction of physiological traits linked to yield in groundnut (Arachis hypogaea L.). This study was conducted to estimate the heritability of components of a simple physiological yield model identified as crop growth rate (C), partitioning (p) and duration of reproduction growth (DR), determine relative contribution of genotype (G), environment (E) and G x E interaction to the variation in the yield model parameters and determine their potential as indirect selection criteria for yield. Forty populations were evaluated in replicated trials in F2, F3 and F4 at three contrasting locations between 1992 and 1994. Physiological traits were estimated from final biomass as well as data on flowering and maturity. Heritabilities were estimated from variance components. Differences were observed among populations for pod yield and yield model components. Relative to the genetic variance, G x E interaction and error variances were larger for yield, C, p and DR. Heritabilities were low, 0.11 for yield, 0.03 for C , 0.22 for p and 0.14 for DR. Indirect selection for yield via p was estimated to be 22 % more effective over direct selection.These results reveal that p had larger heritability than yield, while C seems to be the factor largely responsible for low heritability of yield. The results also indicate that selection for yield and physiological components in segregating populations may be difficult.

Key Words: Arachis hypogaea, breeding, heritability, peanut, selection

RÉSUMÉ

Il y a peu d'information sur l'héritabilité et le degré de l'interaction génotype x environnement parmi les composantes physiologiques intégrées du rendement de l'arachide (Arachis hypogaea L.). La présente étude a été menée dans le but d'évaluer l'héritabilité des composantes d'un modèle de rendement physiologique simple identifié comme taux de croissance de la culture (C), répartition (p) et durée de la reproduction (DR), de déterminer la contribution relative du génotype (G), de l'environnement (E) et de l'interaction G x E à la variation des paramètres du modèle de rendement et d'évaluer les potentialités des composantes d'un modèle de rendement comme critères de sélection indirecte pour le rendement. Quarante populations ont été évalués en F2, F3, F4 dans trois sites différents en 1992 et 1994. Les composantes physiologiques du rendement ont été évaluées à partir du rendement final et de la biomasse ainsi que des données de floraison et de maturité. Les héritabilités ont été estimées à partir des composantes de variance. Il y avait des différences entre les populations pour le rendement en gousses et les composantes du modèle. En fonction de la composante génétique de la variance, l'interaction G x E et les erreurs de variance étaient plus grandes pour le rendement, C, p et DR. Les héritabilités générales étaient faibles, 0,11% pour le rendement, 0,03 pour C, 0,22 pour p et 0,14 pour DR. Il a été estimé que la sélection indirecte pour le rendement à partir de p est de 22% plus efficace que la sélection directe. Ces résultats indiquent que p a une plus forte héritabilité que le rendement, alors que C semble étre le facteur largement responsable de la faible héritabilité du rendement. Les résultats montrent aussi que la sélection hëtive des générations pour le rendement par rapport aux composantes de modèle dans les populations en ségrégation pourrait s'avérer difficile.

Mots Clés: Arachis hypogaea, heritabilité, arachide, sélection

INTRODUCTION

Genotype x environment (G x E) interaction has been widely studied (Crossa, 1990). Knauft and Wynne (1995) reviewed the effects of G x E in groundnut. Most of these studies used data sets based mostly on observations on yield. The use of physiological models offers a means of identifying traits linked with yield and may contribute to improvements in breeding efficiency. This approach has been considered to be more useful, in that the complex holistic trait 'yield' is partitioned according to a biological model into major functional components (Shorter et al., 1991).

Wallace et al. (1993) suggested that indirect selection for yield will be most effective when applied to traits that already integrate most of the genetic and environmental effects that lead to yield. Bandyopadhyay et al. (1985) evaluated the genetic potential of F2 progeny from single and three-way groundnut populations using physiological traits such as leaf area, specific leaf weight and leaf dry weight and components of yield. They concluded that the use of a selection index based on both physiological and yield components in groundnut can be made as early as the F2 generations. They also found that a selection index based on physiological and yield components was more efficient than an index based on yield components alone. Prabhu et al. (1990) also advocated for the use of physiological traits such as leaf area in selecting for relative yield performance in groundnut. Subbarao et al. (1995) argued that integrated traits are more useful in crop improvement programmes than single traits.

A simple yield model, such as that proposed by Duncan et al. (1978) provides a framework for understanding yield variation among different genotypes in variable environments. This model considers that yield (Y) can be defined as the product of crop growth rate (C, in g dry matter m-2 day-1), the length of the reproductive growth (DR , days) and partitioning (p, coefficient) of new material to pods.

Thus, Y = p C DR                                     (1)

These model components integrate many physiological processes. The model has been exploited in analysing a number of agronomic and variety trials (Greenberg et al., 1992; Ntare et al., 1993; Ndunguru et al., 1995). Its application in early generation of breeding is as yet unproven.

The growth analysis measurements summarised in equation 1 must be economically feasible for the large number of lines, progenies and environments required for effective selection. There are difficulties of measuring physiological traits on individual plants without either destroying the plant or incurring great cost. Fortunately, methods that allow largely nondestructive growth analysis with reasonable accuracy have been developed (Williams and Saxena, 1991; Williams et al., 1996). For a relatively minor investment in extra data collection, an improved understanding of genotypic variation in physiological components of yield can be gained.

Examples of heritability and G x E interaction studies on integrated physiological traits in groundnut are few. The first report on the effects of G x E interactions in early generation testing in groundnut is that of Coffelt et al. (1993). Hubick et al. (1988) estimated the magnitude of genetic and genotype x environment interaction sources of variation in transpiration efficiency (W) and carbon isotope discrimination (CID) in 16 groundnut genotypes. They reported broad sense heritability of 81 % for CID and no G x E interaction. In a review of heritability of physiological components of yield, many estimates for broad sense heritability were close to 60 %, a few were < 20 % or >70 % (Wallace et al., 1972). This demonstrates genetic control over physiological components of yield and an opportunity to select desired levels of expression.

Information on heritability and genotype x enviroment of integrated physiological traits linked with yield is essential to develop effective selection strategies to improve yield of groundnut in variable environments. Our objectives were to estimate the heritability of components of a simple physiological yield model using final harvest data, determine the relative contribution of genotype, environment and G x E interaction and determine the potential of the yield model components as indirect selection criteria for yield.

MATERIALS AND METHODS

Site characteristics. Field experiments were conducted at three locations in Niger from 1992 to 1994. The first site was located at the ICRISAT Sahelian Centre, Sadore (lat 13° 15' N, long 2° 17' E, alt 240 m) near Niamey, with average annual rainfall of 560 mm from June to September. The soils at Sadore are sandy loams to loam sand texture and are classified as sandy, silicious, Isohypothermic Psammentic Paleustalf. The second location was Gaya (lat 11° 59' N, long 3° 30' E, alt 160 m) with an average rainfall of 850 mm. The soil is an alfisol (clayey-skeletal, mixed Isohypothemric family of Udic Rhodastalf). The third location was Tara (lat 11° 59' N, long 3° 30' E, alt 200 m and average annual rainfall of 700 mm). The soil is classified as Haplic Acrisol with 90% sand in the top soil. Gaya and Tara are 30 km apart. Monthly rainfall, dates of sowing, flowering and harvest are presented in Table 1.

TABLE 1. Monthly rainfall (mm), dates of sowing, flowering and maturity at three locations in 1992 - 1994

Variable

Sadore

Gaya

Tara

1992

1993

1994

1992

1993

1994

1992

1993

1994

Rainfall mm)

June

85

86

145

0

81

138

102

69

165

July

164

197

153

189

148

232

162

206

94

August

227

229

306

265

241

319

228

186

319

September

53

21

126

93

121

0

66

133

18

Total

529

533

730

547

591

689

558

594

759

Planting

Sowing date

2 June

17 June

13 June

6 July

7 June

6 June

-

2 July

8 June

Flowering

30 June

15 July

15 July

2 Aug

5 July

2 July

-

30 July

10 July

Last harvest

30 Sept

10 Oct

10 Oct

15 Oct

5 Oct

30 Sept

-

20 Oct

30 Sept

Parental lines and crosses. During the dry season of 1991, single crosses were made between Spanish and Virginia parental lines in the glasshouse at the ICRISAT Sahelian Centre, Sadore, for an on-going breeding programme for yield and adaptation in West Africa. The parental lines which were predominantly Spanish and cross combinations are presented in Table 2. The only Virginia lines were ICGV 87121, ICG MS 42 and M13 (introduced from India) which mature in 120 days. Lines 55-437, 796 and TS 32-1 are of short-duration (< 100 days sowing to maturity) which are widely grown in the Sahel and have high partitioning (Greenberg et al., 1992). Chico is an extra- early maturing (< 90 days) germplasm line and J11 and JL 24 are early-maturing lines popular in India. ICGV 86007, ICGV86015, ICGV SM 83005 and ICGV 87123 are advanced medium duration (110 days) lines developed by ICRISAT. Lines with the prefix 'ICGV' and 'ICGVSM' are elite lines developed by ICRISAT in India and Malawi. These genetic materials are representative of those that would be used by groundnut breeders in semi-arid environments in Africa. The F1 was grown in the field at Sadore to produce F2 seed.

Field experiments. In the 1992 crop season, forty F2 populations (Table 2) were chosen based on availability of sufficient F2 seed needed for a replicated trial ( a minimum of 300 seeds). The F2 populations and nine parental lines (excluding Chico, ICGV 87003 and ICGV SM 83005) were grown at Sadore and Gaya. A basal dose of 100 kg ha-1 of single superphosphate was incorporated into the soil by broadcasting during land preparation. Individual plots consisted of three rows, 3 m long and 0.5 m apart. Within-row spacing was approximately 10 cm. Plots were kept weed-free by regular manual weeding. The experimental design was a 7 x 7 triple lattice. Plots were regularly observed to record the date at which 50% of the plants had started flowering. The beginning of the pod development was taken as 15 days after the date of 50% flowering. At maturity, all plants in a plot were hand-harvested. Three two-seeded mature pods were harvested from each plant without selection. These were later bulked over locations to form seed for the subsequent generations. The remaining pods (including immatures) were separated from the haulms and bulked together with pods recovered from the soil. Pods and haulms (including recoverable fallen leaves) were sun-dried separately and weighed. Maturity was indicated by the blackening of internal shell wall (Williams and Drexler, 1981).

TABLE 2. Populations used in the study

Males

Females

796

55-437

TS32-1

Chico

J11

JL24

LCGV
87003

ICGV
86015

ICGV87121

x

x

x

x

x

x

x

x

ICGV87123

-

x

x

x

-

x

-

-

ICGV SM83005

x

x

-

-

x

x

-

x

ICGMS42

x

-

x

x

x

x

-

-

M 13

x

x

x

-

-

x

x

-

ICGV86015

x

x

x

x

x

x

-

-

J 11

x

x

x

-

-

-

-

-

JL24

x

-

-

-

-

-

-

-

55-437

x

-

-

x

x

-

-

-

Total

8

6

6

5

5

6

2

2

The trials were repeated in the F3 (1993) and F4 (1994) generations at Sadore, Gaya and Tara. In these trials, each plot consisted of four rows, 4m long and 0.50 m apart. Land preparation, fertilizer application, seeding rates, flowering date determination and harvest procedures were the same as for the F2 populations.

In all the generations, the dried pods and vegetative weights were used in final calculations. Pod dry matter was multiplied by a correction factor of 1.67 (Duncan et al., 1978) to adjust for the differences in energy requirement for producing vegetative vs. pod dry matter. Crop growth rate (C), pod growth rate (R) and partitioning (p) were estimated from each plot as follows:

C = W / tm                                      (2)

R = WR / (tm- tV + 15)                    (3)

p = R / C                                        (4)

where tm is the time in days from sowing to maturity, tV is the time in days from sowing to 50% flowering and 15 represents the days between flowering and start of podding. W is the adjusted biomass [haulm yield + (pod weight x 1.67)] and WR is the adjusted pod mass. Reproductive duration was the difference between maturity and flowering dates.

Data analysis. All statistical analyses were performed using GENSTAT (Genstat 5.3 Committee, 1993) procedures. The data were analysed in two ways. First, lattice designs were analysed separately for each location and then a combined analysis over environments (locations-year combination) was made using adjusted treatment means. Assuming a fully random model, the data for populations alone were analysed by the GENSTAT method of residual maximum likelihood (REML) to estimate variance components and their approximate standard errors. The components of variance were used to estimate heritability on a mean basis (Singh et al., 1993). Heritability (h2 ) in a single environment was estimated as follows:

h2 = s2g (s2g + s2 / r ) -1                                     (5)

where s2g = genetic variance, s2 = error variance, and r = replications. Across locations, heritability was claculated as:

h2 = s2g (s2g + s2 g x e /n + s2 / nr)                           (6)

where, g x e = G x E interaction variance, and n = number of locations. Variance components were calculated by equating mean squares to their expectations. We calculated the predicted ratio of indirect to direct response to selection as Q= rg (hx/hy), where rg is the genetic correlation between yield and physiological components, h x is the square root of the heritability of trait x and hy is the square root of heritability for yield (Falconer, 1989)

RESULTS

Environments (location and year combinations) mean squares were highly significant for pod yield, C, p and DR (Table 3), indicating significant differences among year-location means as indicated by the range and magnitude of mean values in Table 4. Parents and populations mean squares were both significant, indicating that parents and populations differed for the traits measured. The parents x environment interaction was not significant for pod yield and p but was significant for C and DR . On the other hand, populations x environment interaction was significant for yield and model components. The lowest pod yields were recorded at Sadore in all generations (Table 4). Mean pod yields at Gaya were nearly four times those at Sadore. Pod yields at Tara were double those at Sadore, but lower than at Gaya. C followed the same trend as for pod yield.

TABLE 3. Analysis of variance for parental lines and populations over environments ( years-location combination)

Source of variation

D.F.

Pod yield

C

p

DR

- - - - - - - - - - - - - - - - - - - - - - MS - - - - - - - - - - - - - - - - - - - - - -

Environments (E)

7

84.40**

79.25**

3.73**

14253.19**

Reps/Environment

16

0.32

0.32

0.05

29.48

Parents

8

0.18**

0.14**

0.09**

172.51**

Parents vs Populations

1

3.43**

0.57**

0.309**

124.36**

Parents vs Populations x E

7

0.28**

0.09

0.023

88.35**

Populations

39

0.38**

0.20**

0.083**

105.27**

Parents x E

56

0.10

0.12*

0.013

25.23**

Populations x E

273

0.13**

0.15**

0.014**

31.64**

Residual

768

0.07

0.08

0.008

11.05

*, ** Significant at 0.05 and 0.01 level of probability, respectively

TABLE 4. Mean ranges for pod yield (t ha-1), crop growth rate (C, kg ha-lday-1), partitioning (p, coefficient), and reproductive duration (DR, days) of populations and parents in F2, F3, and F4 at three locations

Site/population

Pod yield

C

p

DR

Sadore

F2 Population

0.17 - 1.09

7 - 39

0.24 - 0.56

84 - 99

Parents

0.28 - 0.64

17 - 26

0.22 - 0.45

82 -100

F3 population

0. 10 - 1.42

4 - 26

0.23 - 0.76

69 -75

Parents

0. 17 - 0.53

7 - 22

0.22 - 0.70

80 - 98

F4 population

0.35 - 2.12

10 - 52

0.51 - 0.75

73 - 80

Parents

0.60 - 1.39

17 - 34

0.61 - 0.77

75 - 95

Gaya

F2 population

1.75 - 3.41

44 - 90

0.58 - 0.96

70 - 97

Parents

2.44 - 3.04

55 - 88

0.70 - 0.99

80 - 93

F3 population

2.09 - 3.29

52 - 86

0.47 - 0.65

75 - 87

Parent

2.7 7- 3.24

71- 82

0.60 - 0.66

80 - 93

F4 population

1.69 - 3.16

48 - 82

0.48 - 0.84

70 - 84

Parent

2.15 - 2.69

28 - 70

0.64 - 0.82

79 -88

Tara

F3 population

1.40 - 2.65

40 - 52

0.50 - 0.93

73 - 88

Parents

1.73 - 2.51

44 - 54

0.50 - 0.93

74 - 95

F4 population

1.53 - 2.61

44 - 63

0.51 - 0.75

66 - 83

Parents

2.23 - 2.86

45 - 59

0.71-0.94

70 - 90

The importance of the source of variation is indicated by the relative magnitude of the variance components (Table 5). Environments had notably larger influence on pod yield and C compared with p and DR in F 2 and F3 generations. In the F4, environment had greater effects on C than on the other traits. The large influence of environment relative to genotypes was in part due to variable conditions at Sadore. For example, the coefficients of variation values (%) for pod yield at Sadore were 43 in F2, 48 in F3 and 45 in F4 compared to 13, 17, and 10 at Gaya. At Tara, corresponding values were 13 in F3 and 12 in F4.

Heritability for yield and model components were generally low (Table 5). Overall, heritability for p was larger than for yield. Heritability for C was nearly zero, while heritability for DR was similar to that of yield.

If any trait is to be used as an indirect selection criterion for yield improvement, heritability of such a trait should be greater than the heritability of yield. In this study, this requirement was met by partitioning (Table 5). The predicted ratio of indirect to direct response to selection (Q) was 1.22, indicating that indirect selection for yield via partitioning would result in a 22 % increase over direct selection for yield. The prediction ratios for C and DR were less than one.

TABLE 5. Variance components, heritability estimates and their standard errors (39 D.F.) for yield, C, p, and DR in F2, F3 and F4 generations

F2 generation 1992

F3 generation 1993

F4 generation 1994

Overall

Parameter+

Yield

C

P

DR

Yield

C

p

DR

Yield

C

p

DR

Yield

C

p

DR

s2G

0.008

32.0

0.002

11.65

0.003

1.42

0.002

0.35

0.024

4.82

0.003

3.40

0.010

2.96

0.003

3.06

SE

0.0062

15.61

0.0010

62.44

0.0044

2.238

0.0012

0.197

0.0078

3.781

0,003

1.251

0.0034

1.938

0.0007

0,998

s2GE

0.014

35.51

0,001

14.58

0.018

9.23

0.004

0.007

0.035

3.24

0.005

4.08

0.011

18.37

0.002

6.57

SE

0.0071

14.00

0.0011

5.930

0.0071

4.684

0.0013

0.255

0.0058

5.568

0.003

0.949

0.0045

3.934

0.0004

0.929

s2E

0.943

79.6

0.009

34.42

0.066

48.27

0.012

4.068

0.089

72.61

0.003

4.96

0.068

65.01

0.008

11.88

SE

0.0054

9.951

0.0011

4.220

0.0066

4,904

0.0012

0,404

0,0092

7,437

0.003

0,509

0.0043

4.082

0.0005

0.733

h2

0.12

0.22

0.15

0.19

0.03

0.00

0.12

0,08

0.22

0.06

0.46

0.27

0.11

0.03

0.22

0.14

SE

0.091

0.096

0.081

0.094

0.051

0.039

0.062

0.043

0.059

0.045

0.071

0,081

0.038

0.022

0,048

0,041

+ Variance component (s2) are for genotypes (G), genotype x environment (GE) and error (E). Heritabilities are on a mean basis for 2 locations in F2 and 3 locations in F3 and F4 The overall is based on 8 environments (location - year combinations)

DISCUSSION

Large location effects were apparent. The locations of our study are characterised by variation in amount and timing of rainfall relative to crop phenology (Table 1). Soil microvariability is known to be responsible for poor crop growth at Sadore (Brouwer et al., 1993). In addition, low rainfall and its poor distribution, high temperatures and hot winds during crop maturation compound the environnmental effects (Sivakumar, 1992). In all the years, rainfall stopped in September which created variable moisture during pod maturation. Under these conditions, environmental effects overrode genetic effects leading to low heritabilities. This highlights a need for a clear definition of target environments for selection between and within populations.

The low heritability estimates indicted the difficulty in controlling environmental influence and in reducing it so that genetic effects are effectively isolated. A number of factors could have affected the heritability estimates in this study. In theory, variation in C is dominated by environmental and management aspects (Williams and Boote, 1995). The evidence concerning partitioning indicates that genotypic differences are more important whereas environment is a less significant source of variation in partitioning (Greenberg et al., 1992; Ndunguru et al., 1995). Various environmental challenges have different impact on C, p, and DR. For instance, drought will influence C and p, calcium deficiency will influence p and foliar diseases will mainly influence C. The distribution of rainfall during crop growth was erratic and amount of raifall during pod addition and filling was also variable (Table 1). The estimation of C using final biomass gives an indication of the seasonal differences in crop resource use and resource-use efficiency. However, the method does not take into account differences in the distribution of that growth within the season. In this study, it would have been useful to have obtained estimates of C during flowering, although this would have been difficult and costly for such a large number of treatments. Variability due to differences in leaf loss at maturity and defoliation due to foliar diseases affect the measurements of C. Foliar diseases were more prevalent at Gaya than the other two locations (data not shown). Thus, the within-season variability may have interacted with the method of estimation of partitioning as well. This could occur because the growth of pods depends largely on photosynthesis.

Some caution should be exercised, however, regarding the generality of our results. This is because evaluations were made under management conditions close to those of small farmers who do not use purchased inputs in growing groundnut. Controlling foliar diseases using fungicides, application of gypsum to alleviate calcium deficiency and supplemental irrigation may give a different picture. In addition when generations are tested in successive years, the individual populations are subjected to different selection pressures which could lead to genetic shifts in the population that vary from generation to generation. Our results , however, point to a limitation which may occur when selecting for yield and physiological components in bulk progenies in early generations. In semi-arid environments, where year-to-year variation in rainfall is large, genotype x environment interaction becomes important. The high coefficients of variation and low heritability from a marginal location such as Sadore, suggests optimising agronomic practices and improved experimental design to decrease the error variance and consequently to increase the selection efficiencies.

In summary, genotype x environment interactions were evident and heritability was low for yield and physiological components. This suggests that selection for yield and integrated physiological components of yield in segregating populations is difficult. The results reveal that p had larger heritability than yield, while C seems to be the factor largely responsible for low heritability of yield.

REFERENCES

Bandyopadhyay, A., Arunachalan, V. and Venkajah, K. 1985. Efficient selection intensity in early generation index selection in groundnut (Arachis hypogaea L.). Theoretical and Applied Genetics 71:300-304.

Brouwer, J., Fussel, L. K. and Herrmann, L. 1993. Soil and crop growth variability in West African semi-arid tropics: a possible risk reducing factor for subsistance farmers. Agriculture, Ecosystems and Environment 45: 229-238.

Coffelt, T. A., Wynne, J. C. and Monteverde-Penso, E.J. 1993. Genotype x environment interaction effects on cultivar development. Oléagineux 48:133-137.

Crossa, J. 1990. Statistical analyses of multilocation trials. Advances in Agronomy 44:55-85.

Duncan, W.G., Mcgraw, R. L. and Boote, K.J. 1978. Physiological aspects of peanut yield improvement. Crop Science 18:1015-1020.

Falconer, D.S. 1989. Introduction to Quantitative Genetics. 3rd Edn. New York: Longman Inc. 340pp.

Genstat 5 Committee, 1993. Genstat 5 release 3 Reference Manual. Oxford: Science Publications.

Greenberg, D.C., Williams, J.H. and Ndunguru, B.J. 1992. Differences in yield determining processes of groundnut (Arachis hypogaea L.) genotypes in varied drought environments. Annals of Applied Biology 120:557-566.

Hubick, K.T., Shorter, R. and Faraquahar, G. D. 1988. Heritability and genotype x environment interactions of carbon isotope discrimination and transpiration efficiency in peanut (Arachis hypogaea L.). Australian Journal of Plant Physiology 15:799-813.

Knauft, D. A. and Wynne, J. C. 1995. Peanut breeding and genetics. Advances in Agronomy 55:393-445.

Ndunguru, B.J., Ntare, B.R., Williams, J.H. and Greenberg, D.C. 1995. Assessment of groundnut cultivars for end-of-season drought tolerance in a sahelian environment. Journal of Agricultural Science Cambridge 125:79-85.

Ntare, B. R., Williams, J. H. and Bationo, A. 1993. Physiological determinants of cowpea seed yield as affected by phosphorus fertilizer and sowing dates in intercropping with pearl millet. Field Crops Research 35:151-158.

Prabhu, K. V., Arunachalan, V. and Bandyopadhyay, A. 1990. Nonparametric approach to multitrait selection for yield in groundnut (Arachis hypogaea L.). Theoretical and Applied Genetics 80:223-227.

Shorter, R., Lawn, R.J. and Hammer, G. L. 1991. Improving genotypic adaptation in crops-A role for breeders, physiologists and modelers. Experimental Agriculture 27:155-173.

Singh, M., Ceccarelli, S. and Hamblin, J. 1993. Estimation of heritability from variety trials data. Theoretical and Applied Genetics 86: 437-441.

Sivakumar, M.V.K. 1992. Emperical analysis of dry spells for agricultural application in west Africa. Journal of Climate 5:532-539.

Subbarao, G.V., Johansen, C., Slinkard, A.E., Nageswara Rao, R. C., Saxena, N. P. and Chauhan, Y. S. 1995. Strategies for improving drought resistance in grain legumes. Critical Reviews in Plant Science 14:469-523.

Wallace, D. H., Ozbun, J.L. and Munger, P. N. 1972. Physiological genetics of crop yield. Advances in Agronomy 24:97-146.

Wallace, D.H., Baudoin, J.P., Beaver, J., Coyne, D.P., Halseth, D.E., Masaya, P.N., Munger, H.M., Myers, J.R., Silberrnagel, M., Yourstone, K. S. and Zobel, R. W. 1993. Improving efficiency of breeding for higher crop yield. Theoretical and Applied Genetics 86:27-40.

Williams, E. and Drexler, J.S. 1981. A nondestructive method for determining peanut pod maturity. Peanut Science 8:134-141.

Williams, J.H. and Boote, K.J. 1995. Physiology and Modelling- Predicting the unpredicatable legume. In: Advances in Peanut Science, pp. 302-353.

Williams, J. H. and and Saxena, N.P. 1991. The use of non-destructive measurement and physiological models of yield determination to investigate factors determining differences in seed yield between genotypes of 'desi' chickpeas (Cicer arietum). Annals of Applied Biology 119:105-112.

Williams, J. H., Nageswara Rao, R.C., Dougbedji, F. and Talwar, H. S. 1996. Radiation interception and modelling as an alternative to destructive samples in crop growth measurements. Annals of Applied Biology 129:152-160.

Copyright 1998, African Crop Science Society

Home Faq Resources Email Bioline
© Bioline International, 1989 - 2024, Site last up-dated on 01-Sep-2022.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Google Cloud Platform, GCP, Brazil