Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

289

EVALUATION OF CASTORBEAN (RICINUS COMMUNIS L.) MUTANTS FOR GENETIC PARAMETERS AND

CLUSTER ANALYSIS

Ghulam Sarwar, Hafiz Munir Ahmed and Jafar Hussain*

ABSTRACT

Twenty seven desirable mutants of castorbean were evaluated in M5 generation at Nuclear Institute for Agriculture and Biology, Faisalabad, Pakistan during 2007-08. Estimation of different genetic parameters alongwith correlation coefficient, path and cluster analyses were done in castorbean mutants on quantitative traits. The results showed positive and significant genotypic correlation of main spike length, capsules of main spike, capsules of secondary spike and spikes per plant with seed yield. These traits alongwith branches per plant also showed positive but non-significant phenotypic correlation with seed yield. Days to mature, number of branches and capsules of main spike showed positive direct effects coupled with positive genotypic correlation with seed yield. Therefore, direct selection of these characters might be helpful for increasing seed yield. Plant height showed high heritability(bs) (54%) alongwith high genetic advance (19.87) seemed to be governed by additive genes while days to mature exhibited high heritability(bs) (45.10%) with low genetic advance (6.61) governed by non-additive type of genes (dominants, epistasis or their interaction). Complete linkage elucidea- distances classified the population into three main clusters. Cluster-II showed maximum average seed yield (187.50 g/plant) followed by cluster I (138.78 g/plant). The variation (standard deviation) was also higher in cluster-II (20.84). The high average seed yield (187.50 g/plant) in cluster-II was due to high average values of main spike length (50.29 cm), secondary spike length (33.81 cm), capsules of main spike (135.21), capsules of secondary spikes (74.13) and spikes per plant (7.94) as compared to other two clusters. Aforementioned characters also showed positive and significant correlation with seed yield. So mutant lines possessing more number of branches, more number of spikes and higher number of capsules in main as well as secondary spikes may be promoted further for increasing seed yield.

KEYWORDS: Ricinus communis; genotypes; mutants; agronomic characters;

performance ; Pakistan.

*Nuclear Institute for Agriculture and Biology, Faisalabad, Pakistan.

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

290

INTRODUCTION Castorbean (Ricinus communis L.) is an industrial oilseed crop. Its seeds contain about 45-58 percent oil. The castor oil is tremendously used in petrochemicals, pharmaceuticals, cosmetics, textile, soap, leather, paint, varnish, ink, nylon, and plastic industries. This oil is traditionally associated with medical and veterinary uses, for example in obstetrics and dermatology, and also as purgative and laxative. Its current use as biofuel production has magnified its importance. Moreover, castor oil does not freeze at high altitude. So due to this particular property, it can be used in aeroplans, helicopters and other vehicles operating at high altitudes or in low temperature zones. Further, it is the best lubricant for jet engines. The shell of castorbean is used in organic termite control in soil, and its seed cake can be used as manure (16, 19). In Pakistan, castorbean is grown on 3204 hectares with an annual production of 2089 tons. Its average seed yield is 652 kg per hectare, which is very low (3) as compared to main castor-growing countries of the world like India, (1266 kg), China (909 kg) and Brazil (850 kg) (2). Although castor is a low input-requiring crop and can be grown on marginal land, yet farmers are not inclined to grow it in Pakistan due to lack of suitable high yielding genotypes/varieties that may fit in country’s cropping system. Much of the harvested crop is derived from semi-wild, tall indeterminate plant growth habit with non-synchronous maturity period. Several undesirable traits of castor plants act as deterrents to species’ that need to be improved to encourage its commercial production as a high income cash crop. To develop high yielding genotypes that fit in the present cropping system it is imperative to create genetic variability for selecting desirable variants. Gerard et al. (12) reported relatively low genetic diversity compared to genetic diversity in other plant species. Induced mutation is an important supplementary approach for creating variability (34). Further, it has the potential to break undesirable linkages between traits. Baldanzi et al.( 5) reported the induction of permanent dormancy of axillary buds to stop iterative growth which is at the base of perennial form of castorbean through the use of induced mutation. Chemical mutagens were also successfully used for creating variation in M2 generation and the greatest increase in variation was observed in yield components (7). Knowledge of genetic variables is important for breeding ideotypes with emphasis on certain traits. Selection based on yield components is advantageous if different yield-related traits are well documented with respect to genetic behaviour (13, 21, 30). Seed yield is a complex and multifaceted

Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

291

trait, and represents the ultimate expression of different yield factors. Path coefficients can be used to assess the contribution of component traits towards seed yield, allowing the traits to be focused upon to be chosen. Cluster analysis was used to classify the similar/ dissimilar mutants into distinct classes using Euclidean distance computed from raw data, which allows the breeder to select genetically diversified material.

The aim of this study was to select better ideotypes with high yield potentials and grouping of castor mutants into different classes/clusters to be utilized properly and efficiently according to the significance/importance of their different traits.

MATERIALS AND METHODS Twenty seven mutants of castorbean were derived after irradiation of two castor genotypes (DS30 and C176) with gamma rays ranging from 100-800GY. These mutants were evaluated and planted in M5 generation at Nuclear Institute for Agriculture and Biology, Faisalabad, Pakistan during 2007-2008 replicated thrice. Four rows (10 m long) of each entry were planted keeping inter and intra row spacing of 75 and 90 cm, respectively. Four plants from each entry were randomly selected and data on single plant basis were recorded at maturity for different morphological traits viz. days to mature, plant height, number of branches per plant, main spike length, secondary spike length, capsules of main spike, capsules of secondary spike, spikes per plant, 100 - seed weight and seed yield per plant. The data were subjected to analysis of variance (32) and genetic parameters, correlation coefficient and path analysis were worked out following Dewy and Lu (9) and Sing and Chaudhry (29) keeping seed yield as dependable variable. Cluster analysis was performed following Johnson and Wichern (14) and Afifi et al (1).

RESULTS AND DISCUSSION

The data on days to mature showed positive and highly significant genotypic correlation (Table 1) with main spike length (1.0680) and secondary spike length (0.6827). Positive and highly significant association of plant height was exhibited with secondary spike length (0.6677) and main spike length (0.4085). Branches per plant also showed highly significant and positive correlation (0.9342) with spikes per plant. Main spike length showed positive and highly significant relationship with secondary spike length (1.4887), capsules of main spike (0.6015) and capsules of secondary spike (0.9179). Secondary spike length only showed positive and highly significant association with capsules of main spike (0.8449). Capsules of main spike

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

292

Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

293

exhibited positive and highly significant correlation with 100 seed weight (0.6937) and seed yield per plant (0.5875). Number of spikes per plant only showed significant and positive correlation with seed yield per plant. Branches per plant showed positive and highly significant phenotypic correlation (Table 1) with spikes per plant. Main spike length exhibited positive and significant correlation with secondary spike length (0.4773) and capsules of main spike (0.485). Secondary spike length only showed highly significant and positive correlation with capsules of secondary spike (0.5309). Overall main spike length, capsules of main spike, capsules of secondary spike and spikes per plant showed positive and significant genotypic correlation with seed yield (Table 1). In case of phenotypic correlation none of the characters under study showed significant association with seed yield. However, phenotypic correlation of secondary spike length, capsules of secondary spike, spikes per plant, branches per plant, and days to mature with seed yield had somewhat higher and positive values. This indicated that traits like branches per plant, main spike length, spikes per plant and capsules of main and secondary spikes are more important and can play a significant role for improving seed yield if selection is based on these traits. Moshkin (17) found a close relationship between seed and number of spikes, number of capsules, and size of seed in castor. There was a positive relationship with height of plant and green mass. Number of seeds from one plant had the greatest direct effect on seed yield. The direct effect of central spikes and lateral spikes is much lower. Similarly plant height and green mass had no direct effect. Hundred seeds weight had slight but positive effect. Phenotypic correlation coefficient values are always higher than genotypic which may be due to pleotropic action of genes on different characteristics (25). Association study of different plant traits (35) revealed the positive association of seed yield with height and main shoot length and negatively associated with capsules number per spike and 100-seed weight, whereas spikes number per plant and capsules number per spike were negatively associated. Lima et al. (19) found that castor bean seed yield was positively correlated with plant height and number of capsules per plant. Reddy et al. (23) studying 56 castor germplasm lines observed correlation of seed yield with earliness, capsule weight and number of capsules per plant under rainfed and irrigated conditions. Aswani et al. (4) reported that capsules in primary spike, number of spikes per plant, number of days to 50 percent flowering and maturity, length of primary spike and 100-seed weight were the major yield contributing characters in castor.

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

294

Path analysis Highest positive direct effects (Table 2) were observed in case of days to mature (1.4453) followed by capsule of main spike (0.8886) and number of branches (0.4469). The highest positive and significant genotypic correlation with seed yield was shown by main spike length (0.7727) followed by capsules of main spike (0.5875), capsules of secondary spike (0.5226) and spikes per plant (0.4241). Days to mature and secondary spike length also showed positive but non-significant genotypic correlation with seed yield (Table 2). Higher positive indirect effects of days to mature were noted via main spike length (1.5436), followed by secondary spike length (0.9867) and capsules of secondary spikes. Plant height showed positive indirect effect via spike per plant. Main spikes length exhibited maximum indirect effect via number of branches (0.7774) followed by spikes per plant (0.5030). Capsules of main spike showed higher indirect effect via secondary spike length (0.7508) and main spike length and capsules of secondary spike. Maximum indirect effect of spikes per plant was estimated via secondary spike length (0.6327) followed by main spike length (0.5346), capsules of main spike (0.369) and capsules of secondary spike (0.3134). Days to mature, number of branches and capsules of main spike showed positive direct effect alongwith high positive genotypic correlation. These characters may be selected directly for seed yield improvement. Other characters showed either negative or very negligible direct effects. Genotypic correlation indicated that main spike length, capsules of main spike, capsules of secondary spike and spikes per plant showed positive and significant genotypic correlation with seed yield. Further, these traits alongwith branches per plant also showed high and positive values of phenotypic correlation with seed yield. Hence, more importance should be given to these traits at the time of selection to improve seed yield. Days to mature, number of branches and capsules of main spike showed positive direct effects coupled with positive genotypic correlation with seed yield. It indicated that selection based directly on these characters may be helpful for increasing seed yield. For improvement of other traits restricted and simultaneous selection may be performed. In some situations, where correlation coefficient is positive but direct effect is negative or negligible, the indirect effects seem to be cause of correlation. In such situation, the direct casual effects are to be considered simultaneously. In some cases, the direct effect is positive and high alongwith negative correlation, under these circumstances, a restricted simultaneous selection model is to be followed i.e. restrictions are to be imposed to nullify undesirable indirect effects to make use of direct effect (28).

Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

295

It is concluded that at the time of selection, main emphasis may be given to number of branches, spike number, capsules of main and secondary spikes for seed yield improvement rather than to length of main and secondary spikes. Genetic parameters Genotypic coefficient of variation (Table 3) were higher in case of spikes per plant (19.35%) followed by branches per plant (15.67%) and plant height (15.37%). The lowest genotypic coefficient of variation values were observed in 100-seed weight (3.00%), secondary spike length (4.33%), days to mature (5.61%) and main spike length (5.85%). Phenotypic coefficient of variation values were higher but with some different sequence than genotypic coefficient of variation, with highest value found in spikes per plant (37.24%) followed by seed yield (30.75%), capsules of secondary spike (29.99%), branches per plant (29.83 %) and capsules of main spike (26.63%). The highest heritability (bs) (Table 3) was estimated in plant height (54.0%) followed by days to mature (45.1%). Branches per plant and spikes per plant exhibited almost similar values of heritability (27.6 and 27%). The genetic advance as percent of mean was higher in case of plant height (19.87) followed by spikes per plant (17.69) and branches per plant (14.50). Days to mature (45.1% heritability) had very small amount of genetic advance (6.61%). Spikes per plant indicated 27.0 percent heritability alongwith 17.69 percent genetic advance. Characters like plant height which showed high heritability alongwith high genetic advance seemed to be governed by additive type of genes. Contrary to this, days to mature which had high heritability but low genetic advance termed to be governed by non-additive type of genes (dominant, epistasis or their interaction) (13, 21, 26, 27). Sachli (24) reported range of variation for various parameters like oil content (1.6 - 2.2 %), green mass (42.2-45.6%), 1000-seed weight and plant height (10-20 %), number of internodes, height of stem and seed yield (20-40 %) in castorbean. The coefficient of variation was also affected with respect to plant spacing. The range of coefficient of variation increases with narrow sowing for all characters except height of the stem. In seed yield and oil content, the increase in negative variability was statistically significant. Moshkin (18) and Sachli (24) estimated high heritability for two varieties of castor in case of oil content, seed hull and 1000-seed weight and low for plant height, number of internodes and green mass. Bhatt and Reddy (6) observed range of heritability (0.152 - 0.089) for seed yield per plant and number of nodes on primary raceme in 30 varieties of castorbean. Expected genetic advance was high for height and number of days to flowering. High

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

296

Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

297

positive phenotypic correlation with seed yield per plant was shown with all eight characters studied except number of days to flowering and number of nodes on primary raceme. Number of capsules for primary raceme and number of secondary branches have large positive direct effect on seed yield per plant. It indicated that semi dwarf lines with a great number of capsules per primary raceme and a moderate number of primary and secondary branches should be used as parents in a putative breeding programme. Dhapke et al. (10) studying castor genotypes, reported high heritability coupled with high genetic advance and genotypic variability, for branches per plant, capsules per main spike, length of pistillate region, capsules per plant and seeds per plant. This indicated the influence of additive gene action providing scope for further selection. Solanki and Joshi (31) found additive gene effects for inheritance of length of primary spike, capsules per primary spike, 100-seed weight and capsules per spike while spikes per plant and seed yield per plant were mainly governed by non-additive gene effects. Characters like number of capsules, main spike length and number of spikes per plant showed high heritability combined with high genetic advance. It showed that these characters were governed by additive type of genes and selection may be based directly on these attributes (13, 26, 27). It is also recommended to select for improved yield components under environmental conditions for which the genotype is best adapted (23). Cluster analysis Cluster diagram/tree diagram for 27 mutants of castorbean based on complete linkage elucidea- distances classified the population into three main clusters (Fig.). Cluster-I comprised nine mutants derived from two different parents i.e. DS30 and C176. Cluster-II consisted of 12 mutants obtained from same two parents. However, six mutants of cluster-III were derived from one parent only i.e. DS30. The cluster-I and cluster-II results indicated that mutants of both parents have some basic similarity, due to which these have been grouped together. Six mutants of cluster-III were dissimilar to some extent due to their significant variation in their characters although these are the derivatives of DS30 only. Tabrizi et al. (33) studying 100 safflower genotypes using multivariate statistical analysis found that material comprised 6 and 13 clusters based on principal and factor analysis, respectively. They also estimated high values of genotypic and phenotypic coefficient of variation for most of traits. Fortes et al. (11) evaluated the genetic divergence between castorbean (Ricinus communis L.) cultivars, from adaptative traits, yield components and cultivars productivity. Based on multivariate analysis, the genetic divergence was determined using cluster

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

298

analysis and principal components analysis. Three groups were formed, and one more divergent was not recommended for hybridization due to its low performance. Costa et al. (8) evaluated genetic divergence among castorbean (cultivars, to enable the choice of parents which make the formation of segregating populations possible. Accesses BRA 4871, BRA 2968, BRA 5550 and BRA 7722 Papo-de-gia, and cultivars BRS 188 Paraguaçu, BRS 149 Nordestina, IAC-80, Mirante-10 and Pernambucana Melhorada were evaluated. Characteristics analyzed were: days to flowering, number of racemes per plant, length of pistillate region of main raceme, plant height, potential yield, and seed oil content. The genetic divergence among accesses and cultivars was studied by multivariate analysis techniques, with canonical variables and cluster analysis, making use of mean euclidean distance. Two groups were formed: group I, formed by eight genotypes; and group II formed by one genotype, cultivar Mirante-10. In spite of being the more divergent, cultivar Mirante-10 should not be recommended for hybridization due to its low medium performance. The other cultivars also presented restrictions, as these were quite similar. The variables that more contributed to the genetic divergence were: days to flowering, length of pistillate region of main raceme, plant height, and seed oil content.

Cluster-II (Table 4) showed maximum average seed yield (187.50 g/plant) followed by cluster-I (138.78 g/plant). The variation (standard deviation) was Table 4. Mean and standard deviation of three clusters for ten characters.

Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

299

Cluster-I Cluster-II Cluster-III Variables

Mean SD Mean SD Mean SD Days to mature 106.11 7.39 113.79 6.70 111.63 3.71 Plant height (cm) 157.33 13.06 182.21 18.26 226.58 23.75 Branches/plant 6.25 0.89 6.23 1.33 5.21 1.22 Main spike length 44.58 3.45 50.29 5.74 50.17 6.71 Secondary spike length 29.56 2.75 33.81 2.36 33.08 3.80 Capsules of main spike 117.53 21.36 135.21 21.36 121.75 13.63 Capsules of secondary spike 64.42 14.33 74.13 11.79 59.75 8.91 Spikes/plant 7.92 1.43 7.94 2.27 6.33 1.33 100-seed weight (g) 22.92 1.58 23.27 1.29 22.88 1.39 Seed yield (g) 138.78 10.86 187.50 20.84 136.67 9.38

also higher in cluster-II (20.84). The high average seed yield in cluster-II was due to high average values of main spike length (50.29 cm), secondary spike length (33.81cm), capsules of main spike (135.21), capsules of secondary spike (74.13) and spikes per plant (7.94) as compared to other two clusters. As already mentioned that these characters also showed positive and significant correlation with seed yield, so mutant lines possessing more number of branches, more number of spikes and higher number of capsules in main as well as secondary spikes may be promoted further for increasing seed yield rather than the mutants having only more spike length. These may be utilized directly after thorough selection and testing or may be utilized in crossing programme for transfer of better characters in adapted varieties otherwise suitable for the region.

REFERENCES

1. Afifi, A., V.A. Clark and S. May. 2004. Computer-aided Multivariate

Analysis. 4th Ed. Chapman and Hall/CRC. New York. 2. Anon 2000. FAO Production Year Book. 54:123. 3. Anon 2006. Agri. Statistics of Pakistan. Ministry of Food, Agriculture and

Livestock, Islamabad, Pakistan. 4. Aswani, K., R. S. Sangwan and D.S. Jatasra. 2003. Correlation and

path coefficient analysis in castor (Ricinus communis L.) under dry land conditions. Indian J. Dryland Agric. Res. and Develop. 18:89-91.

5. Baldanzi, M., M. Fambrini and C. Pugliesi. 2003. Redesign of the castorbean plant body plan for optimal combine harvesting. Ann. Appl. Biol. 142: 299–306.

6. Bhatt, D. and T. P. Reddy. 1981. Correlation and path analysis in castor (Ricinus communis L.). Can. J. Gen. Cyt. 23: 525-531.

7. Bokhan, I. K. 1989. Variation in the quantitative characters of castor oil plant in M2 after treatment of seeds with chemical mutagens. Nauchno-

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

300

Tekhnicheskii Byulleten’ Vsesoyuznogo Ordena Lenina I Ordena Druzhby Narodov Nauchno-Issledovatel’ Skogo Instituta Rastenievodstva Imeni N. I. Vavilova. 188: 46-48.

8. Costa, M.N., W.E. Pereira, R.L.A. Bruno, E.C. Freire, M.B.M. Nobrega, M. Milani and A.P. Oliveira. 2006. Genetic divergence on castorbean accesses and cultivars through multivariate analysis. Pesq. Agropec. Bras. 41(11):1617-1622.

9. Dewy,D.R. and K.H. Lu. 1959. A correlation and path coefficient analysis of component of crested wheat grass seed production. Agron. J. 51:515-518.

10. Dhapke, S. K., P. W. Khorgade and M. N. Narkhede.1992. Estimation of genetic variability in castor (Ricinus communis L.). Agric. Sci. Digest (Karnal) 12:141-143.

11. Fortes, B. H., S.S. Alves and F. L.Gonzaga.2008. Genetic divergence in five castor bean cultivars. Pesq. Agropec. Bras. 43 :357-362.

12. Gerard, A., A. Williams, P. D. Rabinowicz, A. P. Chan, J. Ravel and P. Keim. 2007. Worldwide genotyping of castorbean germplasm (Ricinus communis L.) using AFLPs and SSRs. Gen. Resour. and Crop Evol. (Springer link, July 04, 2007.).

13. Johnson , H .W., H.E. Robinson and R.E. Comstock. 1955. Estimates of genetic and environmental variability in soybean. Agron. J. 47: 314-318.

14. Johnson, R.A. and D.W.Wichern. 1996. Applied Multivariate Statistical Analysis. Prentice Hall of India Private Limited, New Dehli-110001.

15. Lima, E. F., J. Santos and W. Dos. 1998. Genotypic, phenotypic and environmental correlations among agronomic characteristics of castorbean (Ricinus communis L.). Revista de Oleaginosas e Fibrosas. 2:147-150.

16. Maiti, S., M.R. Hegde and S.B. Chattopadhyay. 1988. Hand Book of Oilseed Crops. Oxford & IBH Publishing Co. (Pvt). Ltd. New Delhi.

17. Moshkin, V. A.1967. V kn.: Rukovodstvo po sektsii i semenovodstvu maslichnykh kultur (Manual of Breeding and Seed Production of Oil Crops) Moscow. In: Castor: Moshkin, V.A. (ed.). Amerind Publishing Co. (Pvt). Ltd. New Delhi 1986. p:1-315.

18. Moshkin, V.A. 1972. V sb.: Selektsiya i Semenovodstvo Maslichnykh ul, tur (VNIIMK). In: Breeding and Seed Production of Oil Crops. In: Castor. Moshkin, V.A. (ed.). Amerind Publishing Co. (Pvt). Ltd. New Delhi 1986. Pp:1-315.

19. Moshkin, V. A.1986. Castor. Amerind Publishing Co. (Pvt). Ltd. New Dehli . p: 1-315.

20. Narain, A. 1961. (In ‘‘Castor ‘’ editor Moshkin V.A. Amerind publishing Co. Pvt. Ltd. New Dehli 1986. pp: 1- 315). Ind. J. Genetics and Plant Breeding. Vol. 20, No. 2.

Evaluation of castorbean mutants for genetic parameters

J. Agric. Res., 2010, 48(3)

301

21. Panse , V.G. 1957. Genetics of quantitative character in selection plant breeding. Indian J. Genet. 17: 318-328.

22. Reddy, P.R., G.R.M. Sankar, M. Vanaj, C.H. Rao, S. Venkateswarlu and J. D. Eastin. 1999. Heritability and genetic advance in castor (Ricinus communis L.) over varied soils and dates of sowing. Indian J. Agric. Sci. 69: 371-375.

23. Reddy, P. R., M. Vanaj, G. R. M. Sankar, C.H. Rao, S. Venkateswarlu, and J. D. Eastin. 1999. Yield components in castor germplasm under irrigated and rainfed conditions. J. Maharashtra Agric. Univ. 24:36-41.

24. Sachli, I.K. 1972. V. sb.: Selektsiya i Semonovodstvo Maslichnykh Kul, tur. In: Breeding and Seed Production of Oil crops. VNIIMK. In: Castor. Moshkin, V.A. (ed.). Amerind Publishing Co. (Pvt.). Ltd. New Dehli 1986. Pp: 315.

25. Salikh, C.H. and M.O. Khider. 1975. Exper. Agri. Vol. 11, No.2. In: Castor. Moshkin, V.A. (ed.). Amerind Publishing Co. Pvt. Ltd. New Delhi 1986.Pp: 315.

26. Sarwar, G., and M.A. Haq. 2005. Radiation-induced variability for the improvement of yield and yield components in sesame (Sesamum indicum). Sesame and Safflower Newsletter. 20:42-47.

27. Sarwar, G. and M.A.Haq.2006. Evaluation of sesame germplasm for genetic parameters and disease resistance. J.Agric.Res. 44(2):89-96.

28. Sing, R. K. and S.N. Kakar .1977. Control on individual trait means during index selection. Proc. Third Congr. SABRO (Canberra). p: 22-25.

29. Sing, R. K. and B.D. Chaudhry. 1999. Biometrical method in Quantitative Genetic Analysis. Kalyani Publishers New Delhi. p: 1-303.

30. Singh, I. S., N. T. N. Hue and A.K. Gupta. 1995. Association and cause and effect analysis in some F2 populations of green gram. Leg.Res. 18: 137-142.

31. Solanki, S. S. and P. Joshi. 2000. Combining ability analysis over environments of diverse pistillate and male parents for seed yield and other traits in castor (Ricinus communis L.). Indian J. Genet. Pl. Br., 60:201-212.

32. Steel, R.G.D. and J.H. Torrie.1980. Principles and Procedures of Statistics. McGraw Hill Book Company Inc, New York.

33. Tabrizi, A.H.O., M. R. Ghannadha, M.R. Ahmadi, and S.A. Payghambari. 1999. Evaluation of some important agronomic traits of safflower using multi variate statistical methods. Indian J. Agri. Sci. 30:817-827.

34. Tepora, N. M. 1994. Mutation Breeding of Oilseed Crops. IAEA-TECDOC- 781. p: 149-157.

35. Yadava, T. P. and H. Singh.1973. Correlation studies in castor (Ricinus communis L). Oilseed J. 3:13-16.

G. Sarwar et al.

J. Agric. Res., 2010, 48(3)

302