Estimates of genetic correlations and correlated responses to selection in cashew (Anacardium occidentale L.)

Estimates of genetic correlations and correlated responses to selection in cashew (Anacardium occidentale L.)^*

Diógenes Manoel Pedroza de Azevedo¹, João Ribeiro Crisóstomo², Francisco Célio Guedes Almeida³ and Adroaldo Guimarães Rossetti²

¹ Embrapa Rondônia, BR 364 km 5,5, Caixa Postal 406, 78900-970 Porto Velho, RO, Brasil. Send correspondence to D.M.P.A.

² Embrapa Agroindústria Tropical, Rua D. Sara Mesquita, 2270, Planalto Pici, 60511-110 Fortaleza, CE, Brasil

³ Departamento de Fitotecnia, Universidade Federal do Ceará, UFC, 60451-970 Fortaleza, CE, Brasil.

ABSTRACT

The present study estimates variances and genetic and phenotypic correlations for five traits in 27 progenies of cashew trees (Anacardium occidentale L.). Data were obtained from a trial conducted in 1992 at Pacajus, Ceará, experimental station of Embrapa Agroindústria Tropical. The characters studied were plant height (PH), North-South and East-West canopy spreads (NSS, EWS), and primary and secondary branch numbers (PBN, SBN). All genetic and phenotypic correlations presented positive and significant values. Selection to increase or decrease the average of any one of the five characteristics of cashew plants in the progenies studied affected the average of the others. The 16-month-old canopy spread can be predicted from NSS or EWS since correlations between them were high. Correlations between PH and SBN were low, indicating that there is a good possibility of obtaining smaller plants without causing drastic reductions in SBN. PH and SBN showed, respectively, the lowest and highest genetic variance estimates relative to the corresponding population means.

INTRODUCTION

The cashew (Anacardium occidentale L.) is an important Brazilian crop. There are constantly increasing demands, for both domestic and export markets, for two of its principal products: the cashew nut and cashew nut shell liquid (CNSL), which is a high-grade industrial oil. Cashew farming is an activity of considerable socio-economic importance for Brazil, because it provides income for over 300,000 people and generates more than $100 million yearly in revenue for the country (EMBRAPA, 1990).

The common cashew trees currently cultivated in Brazil produced in 1990 about 190 kg of nuts per hectare. This yield is very low, probably because seedlings are obtained from seed lots with a poor genetic basis.

The cashew has received little research effort in the past. Galang and Lazo (1936) studied hundreds of productive and non-productive terminal and lateral sprouts, relating vegetative growth to fruit production. They found that productivity of the cashew plant is associated with leaf surface area and certain growth habits. Nambiar (1977) presented one of the principal reviews for the cashew crop. He found that cashew plant productivity is negatively correlated with leaf surface area and the size of the internodes. More recently, Parameswaran et al. (1983), cited by Barros (1988), observed that the number of fruiting branches per unit of canopy area is strongly correlated to cashew production. According to Ohler (1979), the relationship between plant area and productivity is important, because plants that develop vigorously require more space. Also, Ohler indicated that production per unit area is more important than the production per plant.

Rao (1974) found a positive correlation between cashew production and the percentage of hermaphroditic flowers. Similar results were obtained by Murthy et al. (1984) who found a significant correlation between the number of panicles and hermaphroditic flowers. Northwood (1966) observed that the number and weight of nuts per plant were associated with productivity, but the number of nuts per inflorescence was not. He concluded that the number of inflorescences produced per plant was the principal factor determining productivity. Auckland (1961) observed that more dense nuts had a greater rate of germination. In seedlings, positive correlations were found between the specific weight of nuts and the following variables: number of leaves, diameter of the stem, and length and width of leaves. Rao and Hassan (1956) found a positive correlation between the weight of nuts and kernels. A similar result was obtained by Faluyi (1987), indicating that selection for one of these characters affects the other. The objective of the present study was to estimate variances, genetic and phenotypic correlations along with correlated responses involving five traits in 27 progenies of cashew trees.

MATERIAL AND METHODS

The trial was set up in 1991 at Pacajus Experimental Station, which belongs to Embrapa Agroindústria Tropical, located in Ceará State. The experimental design was randomized blocks with 27 treatments replicated twice. The 27 treatments were: four progenies, descendents of early dwarf mother trees (CP 06, CP 76, CP 09 and CP 1001), four descendents of common mother trees (CP 07, CP 77, CP 96 and CP 12) and 19 progenies resulting from crossing of these mother trees, and the progeny of the common cashew, named 'B tonelada'. Each plot had five plants, spaced 7 m x 7 m. The data were obtained from plants 16 months old. The following parameters were noted and analyzed: plant height (PH), North-South and East-West canopy spreads (NSS, EWS) and number of primary and secondary branches (PBN, SBN).

The height of the plant was taken from the drill level until the highest point of the highest leaf of the principal branch. The primary branch was defined as that originating from the principal stem, and the secondary branch as that originating from the primary branch.

The parameters were analyzed by using variance and covariance components according to Vencovsky (1987) and Kempthorne (1973), in order to obtain the estimates of genetic and phenotypic correlations. Tables I and ^II show the models of variance and covariance analysis, and the respective expected mean squares of the products. The integer results were transformed to (Steel and Torrie, 1960). The estimates for variance and the genetic coefficient of variation were obtained from Vencovsky (1987):

Genetic variance

Environmental variance

Phenotypic variance

Genetic coefficient of variation

Genetic (r_g) and phenotypic (r_F) correlations between two x and y traits were estimated according to the procedure described by Kempthorne (1973):

The evaluation of the indirect effect of selection was obtained through the estimates of correlated responses (Falconer, 1981):

where: i = the dependent constant of the selection intensity applied among the treatments. A 20% selection intensity was considered (i = 1.3998).

RESULTS AND DISCUSSION

These results were obtained from an experiment conducted in one environment. Because of this, they only classify the effects of the year and location. According to Allard (1971), this procedure does not permit to avoid the genotypic interaction with the environment.

The observed mean squares of the variance analyses, the averages, the environmental variation coefficients, and the minimum significant differences for the characters are in Table III. Highly significant differences (P < 0.01) among the progenies for NSS, EWS, and PBN, and (P < 0.05) for SBN were found.

The coefficient of variation for PBN was low (Table III), which indicated good experimental precision (Gomes, 1987). The high coefficient of variation for SBN may be due to the young age (16 months old) at evaluation time and to the genetic variation among and within progenies.

Estimates for the genetic variances among progenies , environmental variances among progenies , phenotypic variances , and the genetic coefficient of variation (CVg) of the characters evaluated are presented in Table IV. The largest estimates of genotypic variance in relation to the average of the character were found for SBN, while PH gave the lowest estimate, indicating that this character had the lowest genetic variability.

The genetic and phenotypic correlations among the five characters studied were positive and significant (Table V). The high values of the estimates make it evident that selection to increase the mean of any one of the characters evaluated would have repercussions on the remaining characters, causing indirect increases in their means.

The correlations between vigor and fruit production of the cashew plant are important. Dasarathi (1958), cited by Nambiar (1977), brought attention to the fact that cashew trees with excessive vegetative growth and long internodes are less productive than those within the average or with low values. This information could be useful for premature selection, when using progenies in the juvenile phase. Results obtained from other perennial species have shown the existence of correlations between fruit production and diameter of the canopy (Glendinning, 1966; Garcia and Nicolella, 1985; Sing Dhaliwal, 1968).

The values of genetic gains expected in various characters, when selection is applied to other characters (correlated response; Table VI), are important, because at times the selection of a complex character with low heritability may be required. In this case, selection can start from simple characters, but with greater heritability since they are related to the principal character (Falconer, 1981; Mariotti, 1985). The characters PBN and PH (Table VI) had the highest and the lowest expected genetic gains for the remaining characters. In a reciprocal manner, PH received the lowest influence from direct selection on the remaining characters. These results are a product of the low genetic variability exhibited by plant height (Table IV). High genetic covariance estimates benefit indirect selection (Falconer, 1981; Mariotti, 1985). According to these authors, indirect selection is also benefited by the following factors: higher genetic correlations and differences in heritability of the characters involved, principally when , y being the character of direct selection.

When selection was applied to the NSS and EWS characters, the genetic gains expected in the remaining characters were similar (Table VI). These results were the product of the greater genetic covariance presented by canopy spread in the progenies studied and genetic correlations between canopy spread and each of the remaining characters (Table V).

ACKNOWLEDGMENTS

The authors are grateful to Embrapa Agroindústria Tropical for the research grants received during the period in which the work was carried out and to Arnaldo Bianchetti for the English review.

RESUMO

Neste trabalho são estimadas variâncias, correlações genéticas e fenotípicas e respostas correlacionadas, envolvendo cinco caracteres em 27 progênies de cajueiro (Anacardium occidentale L.). Os dados foram obtidos em Pacajus-CE, num ensaio conduzido no Campo Experimental da Embrapa Agroindústria Tropical, em l992. Os caracteres estudados foram altura de planta (PH), envergaduras norte-sul (NSS) e leste-oeste (EWS) e número de ramos primários (PBN) e secundários (SBN). Todas as correlacões genéticas e fenotípicas obtidas foram positivas e significativas. A seleção para aumentar ou reduzir a média de qualquer um dos cinco caracteres estudados nas progênies de cajueiro afetou indiretamente a média dos outros quatro caracteres. A envergadura da copa aos 16 meses pode ser representada por NSS ou EWS, tendo em vista que a correlação entre elas foi elevada. As correlações envolvendo PH e SBN foram baixas, indicando ser possível obter plantas de porte reduzido sem causar drástica redução no SBN. PH e SBN apresentaram, respectivamente, a menor e a maior estimativa da variância genética em relação à média do caráter.

(Received January 4, 1994)

• Allard, R.W. (1971). Princípios do Melhoramento Genético das Plantas Edgard Blucher, Săo Paulo, SP, pp. 381.
• Auckland, A.K. (l961). The influence of seed quality on the early growth of cashew. Trop. Agric. 38: 57-67.
• Barros, L.M. (1988). Biologia floral, colheita e rendimento. In: Cultura do Cajueiro no Nordeste do Brasil (Lima, O.M.S., ed.). Banco do Nordeste do Brasil. Escritório Técnico de Estudos Econômicos do Nordeste, Fortaleza, CE, pp. 486. (Estudos Econômicos e Sociais, 35).
• EMBRAPA (1990). Centro Nacional de Pesquisa do Caju (Fortaleza, CE). Relatório técnico anual 1987-88. Fortaleza.
• Falconer, D.S. (1981). Introduçăo ŕ Genética Quantitativa (Traduçăo de Silva, M.A. e Silva, J.C.). Universidade Federal de Viçosa, Viçosa, MG, pp. 279.
• Faluyi, M.A. (1987). Genetic variability among nut yield traits and selection in cashew (Anacardium occidentale L.) Plant Breeding (Zeitschrift für Pflanzenzüchtung) 98: 185-272.
• Galang, F.G. and Lazo, F.D. (1936). Fruiting as related to vegetative growth in cashew (Anacardium occidentale L.) Plilipp. J. Agric. 7: 21-33.
• Garcia, J.R. and Nicolella, G. (1985). Correlaçăo entre algumas medidas dendrométricas, origens genéticas e produçăo de frutos em cacaueiros. Rev. Theobroma, 15: 113-124.
• Glendinning, D.R. (1966). Further observations on the relationship between growth and yield in cocoa varieties. Plant Breeding 36: 591 (Abstract).
• Gomes, F.P. (1987). Curso de Estatística Experimental. 12th edn. Nobel, Săo Paulo, SP, Brasil, pp. 467.
• Kempthorne, O. (1973). An Introduction to Genetic Statistics. Iowa State University Press, Ames, pp. 545.
• Mariotti, J.A. (1985). Fundamentos de Genética Biométrica - Aplicaciones al Mejoramento Genetico Vegetal. Organizacion de los Estados Americanos, Washington, D.C., pp. 151.
• Murthy, K.N., Vijaya Kumar, K., Pillai, R.S.N. and Kumaram, P.M. (1984). Flowering behavior and correlation studies in cashew. In: Abstracts of the Proceedings of the International Cashew Symposium, Cochin, Kerala, India, 1984.
• Nambiar, M.C. (l977). Cashew. In: Ecophysiology of Tropical Crops (Alvins, P. de T. and Kozlovsky, T.T., eds.). Academic Press, New York, pp. 461-477.
• Northwood, P.J. (1966). Some observations on flowering and fruit setting in the cashew (Anacardium occidentale L.). Trop. Agric. 43: 35-42.
• Ohler, J.G. (1979). Cashew Koninklijk Instituut voor de Tropen, Amsterdam (Communication, 71), pp. 260.
• Rao, V.N.M. (1974). Crop in Improvement of Cashew. Report of the all India Summer Institute on Improvement and Management of Plantation Crops. Mimeographed, pp. 128-134.
• Rao, V.N.M. and Hassan, M.V. (1956). Variations in the seed characters of cashew (Anacardium occidentale L.). India J. Agric. Sci 26: 211-216.
• Singh Dhaliwal, T. (1968). Correlations between yield and morphological characters in Puerto Rican and Columnaris varieties of Coffea arabica L. J. Agric. Univ 52: 29-37.
• Steel, R.G.D. and Torrie, J.H. (1960). Principles and Procedures of Statistics, with Special Reference to the Biological Sciences McGraw Hill, New York.
• Vencovsky, R. (1987). Herança quantitativa. In: Melhoramento e Produçăo do Milho 5th edn. Fundaçăo Cargill, Campinas, SP, Chapter 5, pp. 137-209.

*

Part of a thesis presented by D.M.P.A. to the Universidade Federal do Ceará (UFC) in partial fulfillment of the requirements for the Master's degree.

Publication Dates

• Publication in this collection
  23 Feb 1999
• Date of issue
  Sept 1998

History

• Received
  04 Jan 1994