Caña de Azúcar, Vol. 20(1):3-17. 2002
THE EFFECT OF INTRAROW PLANT SPACING ON THE EFFECTIVENESS OF FAMILY SELECTION IN SUGARCANE. REPEATABILITY ANALYSIS
1INIA-Yaracuy, Apartado 110, San Felipe,
Estado Yaracuy 3201 VENEZUELA.
2S. B. Milligan, United States Sugar Corp. P. O. Box 1207,
Clewiston, Fl 33440, USA.
Cross appraisal is performed in the Louisiana Sugarcane Variety Development Program (LSVDP) since 1992. It comprises the evaluation of progeny from the hybridization of two parents to measure its potential to produce elite individuals. Studies on the cross appraisal methodology have showed that intrarow plant spacing influences the variance and reliability of a trait. Thus, affecting the efficacy of progeny testing and family selection. The efficiency of selection is enhanced if a trait is repeatable across test environments. In order to evaluate the effectiveness of family selection at an early stage of selection under Louisiana environmental conditions, family repeatability of plant weight and the major components of plant weight were examined. This study shows the effect of two intrarow plant spacings in the estimation of family repeatability for a given trait between all possible pairs of three environments. Repeatability values at both intrarow plant spacings were reasonably high compared to other studies. Nevertheless, repeatability values were higher for wide intrarow plant spacing than for narrow spaced plants. Number of stalks per plant was the least repeatable trait. Repeatability values between all possible environment combination were high showing a low effect of the environment on the traits studied. Genetic associations calculated with data taken from wide spaced plants were generally higher than those associations determined with data taken from narrow spaced plants. This difference clearly shows, as it did at the phenotypic level, the effect of the spacing factor on the genetics of a trait.
Key words: progeny test, LSVDP, hybridization, environments
EL EFECTO DE LA DISTANCIA
ENTRE PLÁNTULAS DENTRO DE LA HILERA EN LA EFICIENCIA DE LA SELECCIÓN
DE CAÑA DE AZÚCAR.
A partir de 1992, el Programa de Desarrollo de Variedades de Caña de Azúcar de la Universidad del Estado de Louisiana (LSVDP, por sus siglas en inglés) implementó la prueba de progenie como su sistema de evaluación de familias de caña de azúcar. Esta metodología implica la evaluación de la progenie producto de un cruce biparental con el objetivo de medir el potencial de esa combinación para producir individuos elite. Estudios realizados para mejorar la eficiencia de esta metodología demuestran que la distancia entre plantas de caña de azúcar dentro de una misma hilera influencia la varianza y confiabilidad de un carácter, por consiguiente afectando la eficacia de la prueba de progenie y la selección de familias de caña de azúcar. Otros estudios indican que la eficiencia de la selección aumenta si el carácter bajo selección es repetible a través de diferentes ambientes. Con el objetivo de evaluar la efectividad de la selección de familias en etapas tempranas de selección, bajo las condiciones ambientales del área cañera del Estado de Louisiana, U.S.A., se examinó la repetibilidad familiar del peso de cepa de caña de azúcar y sus principales componentes. Este trabajo muestra el efecto de dos distancias de siembra entre plantas de caña de azúcar en la estimación de la repetibilidad familiar de un mismo carácter entre todas las combinaciones posibles de tres ambientes. Los resultados indican que los valores de repetibilidad en ambas distancias entre plantas fueron razonablemente altos en comparación con otros estudios. Aún cuando los valores de repetibilidad fueron altos, los valores obtenidos utilizando una separación de 82 cm fueron más altos que los obtenidos utilizando 41 cm de separación. La variable número de tallos por cepa de caña de azúcar fue el carácter menos repetible. Los valores de repetibilidad obtenidos entre todas las posibles combinaciones de tres ambientes fueron lo suficientemente altos para demostrar el bajo efecto del ambiente en los caracteres estudiados. Las asociaciones genéticas calculadas a partir de los datos obtenidos de plantas separadas por 82 cm fueron, generalmente, más altas que las asociaciones determinadas a partir de plantas separadas por 41 cm. Esta diferencia demuestra, tal como ocurrió a nivel fenotípico, el efecto de la distancia entre plantas de caña de azúcar dentro de una misma hilera sobre la genética de un carácter.
Palabras clave: prueba de progenie, LSVDP, hibridación, ambientes
Parent evaluation and family selection comprise the initial selection steps of the Louisiana Sugarcane Variety Development Program (LSVDP). Since 1992, LSVDP has used a replicated family appraisal method that is intended for both family selection and progeny testing. Studies on the cross appraisal methodology have showed that intrarow plant spacing influences the variance and reliability of a trait. Thus, affecting the efficacy of progeny testing and family selection (De Sousa-Vieira and Milligan, 1999). Those studies indicated that selection using spaced plants (82 cm) would be more accurate than selection using narrowly plants (41 cm).
Kang et al., (1984) indicated that the effectiveness of selection of sugarcane clones would be enhanced if the character under selection is repeatable across environments. Falconer (1966) regarded two plants of the same clone as two individuals or as one individual replicated twice. He suggested the terms clonal repeatability and individual repeatability, the former when it refers to different plants of the same clone, the latter when it refers to different parts of the same individual. It was also Falconer (1952, 1989), who suggested the use of genotypic correlations for a given trait between two different environments as a measure of genotype by environment interaction but, as Kang et al. (1984) pointed out, genetic correlations in this fashion have not been commonly used in sugarcane breeding programs.
Hanshe and Brooks (1965) working on sweet cherry used the terms temporal and spatial repeatabilities for measurements taken on the same tree during different years and measurements taken in the same year among trees of the same genotype. Mariotti (1974) introduced this approach as a measure of space repeatability in sugarcane.
Most of the work in repeatability has been done to investigate trait repeatability from stage to stage in a sugarcane breeding program (Ladd et al., 1974; Miller and James, 1975). Since sugarcane is vegetatively propagated, a high degree of repeatability should be expected from stage to stage if environment has no effect, however that is the exception rather than the rule. Generally, high repeatabilities have been reported for stalk diameter (Brown et al., 1968; James and Miller, 1971; Mariotti, 1974; Ladd et al., 1974), whereas Smith and James (1969) and Kang et al. (1984) reported stalk number to have high clonal repeatabilities.
While phenotypic associations give a measure of clonal repeatability, genetic associations can give evidence of the kind of mechanisms involved in the expression of a trait. When the same trait is measured in two different environments, it can be regarded not as one trait, but as two different traits (Falconer, 1989). If genetic correlation for a given trait at two different environment is low, Falconer (1989) explained, then the physiological mechanisms acting upon the trait on those environments would be to some extent different, and consequently the genes required for high performance of that particular trait are to some extent also different.
Kang et al. (1984) reported intermediate to high (0.62 to 0.92) genetic correlations for stalk number indicating nearly similar genetic expression of stalk number at different environments. Although Mariotti (1974) reported almost the same genetic correlations values as Kang et al. (1984) did (0.62 to 0.91), he concluded that a different gene action seemed to be working in different environments. Genetic mechanisms in determining stalk weight at most locations were similar in both Kang et al. (1984) and Mariotti (1974) studies. The latter also indicated high genetic associations for stalk diameter. Individual repeatabily of a trait between the same clone in different crops has also being studied. High individual repeatabilities have been reported for stalk number (Tai et al., 1980; Kang et al. 1984), and for stalk weight (Tai et al., 1980; Milligan, 1988).
High phenotypic associations between the same trait at two different environments would indicate good repeatability, which may evidence that improvement made in one environment will be translated in improvement in the second environment. High genetic associations between the same trait at two different environments would indicate that the genetic mechanism(s) conditioning that trait in one environment would be the same acting at the second environment (Falconer, 1989).
The purpose of this research was to estimate family repeatability by using the phenotypic correlation coefficients for a given trait between all possible pairs of three environments and two intrarow plant spacings, and to conduct a study to analyze the mechanisms conditioning in the expression of a trait by using genetic correlations.
MATERIALS AND METHODS
Field experiments and data collection
Twenty-five randomly selected biparental families from the Louisiana Agricultural Experiment Station (LAES) 1993 crossing series were used in this study (Table 1). Crosses were made by research personnel at St. Gabriel Research Station in St. Gabriel, Louisiana.
Seeds were germinated in flats under greenhouse conditions in January 1994. Approximately three weeks after germination, seedlings were transplanted to SpeedlingTM trays with 3.8 cm2 cells and cultured in the greenhouse. The progeny were then transplanted to the field in April 1994 at the St. Gabriel Research Station, and to the USDA Ardoyne Farm near Chacahoula, Louisiana. Seed and progeny for the same crosses were again planted and transplanted in January and April at the same farms in 1995.
Individual plants from each cross were planted in a randomized complete block design using two blocks with a split plot treatment arrangement where the main plots were intrarow plant spacings of 41 cm (standard at LSVDP) and 82 cm on rows 1.8 m apart. Subplots were families. Each subplot consisted of two rows with up to 16 randomly selected seedlings in each row.
Millable stalk number per plant, stalk length, and mid-stalk diameter were recorded in August 1995 from the progeny planted in 1994, and in August 1996 from the progeny planted in 1995. Data were collected in first ratoon cane, 16 months following each planting. Stalk length was measured from the stalk base to the first visible dewlap (leaf collar) of two random stalks in each plant. The same two stalks were measured for mid-stalk internode diameter using a caliper. Stalk weight was estimated as the volume of the stalk assuming a perfect cylinder with specific gravity of one (Miller and James, 1975; Gravois et al., 1991; Chang and Milligan, 1992):
where the density d = 1.0 gm cm-3, r = stalk radius (cm), and L = stalk length (cm). Plant weight was estimated as stalk weight times stalk number per plant.
Family mean data were computed on total number of stalks per plant, stalk length, stalk diameter, stalk diameter, and plant weight. Data were determined for the first ratoon crop of seedlings. Phenotypic and genetic correlations for a given trait between different environments were determined among the three possible pairs of environments and for each intrarow plant spacing.
The phenotypic and genetic correlations (r) have been estimated as follows:
where s i x y is the estimated covariance between the environment x and the environment y of the trait i, si x and si y are the standard deviations of trait i. Approximate standard errors of the genetic correlations were estimated as (Falconer, 1989):
where rg i x y is the genetic correlation between the trait i at environments x and y, respectively; S.E. hi x and S.E. hi y are the standard errors of the heritability of the trait i at environments x and y, respectively; and hi x2 and hi y2 are the heritabilities of trait i at environments x and y, respectively.
Approximate standard errors of the phenotypic correlations were estimate as (Efron and Tibshirani, 1993):
where rp i x y is the phenotypic correlation between the trait i at environments x and y, and N is the number of observations for trait i in each environment (N=50).
The terms spatial and temporal repeatabilities apply to the present study and are used as follows: phenotypic correlations for a given trait between Ardoyne farm 1995 and Ardoyne farm 1996 are a measure of temporal repeatability. Phenotypic correlations for a given trait between Ardoyne farm 1995 and St. Gabriel Res. Stn. 1995 are a measure of spatial repeatability, and phenotypic correlations for a given trait between Ardoyne farm 1996 and St. Gabriel Res. Stn. 1995 are a measure of spatial-temporal repeatability.
RESULTS AND DISCUSSION
The 1995/1996 winter freezes severely damaged the plant population at St. Gabriel Research Station. Although data were collected at this location, it was decided not to use them in the combined analysis since competition from plant spacing was a prime consideration in the experiment.
Repeatability values (Table 2), at both intrarow plant spacings, were reasonably high compared to other studies (Smith and James, 1969; Mariotti, 1974; Miller and James, 1975; Kang et al., 1984). Phenotypic repeatability is commonly used to determine gain in accuracy to be expected from multiple measurements of a trait (Falconer, 1989). When the repeatability value is high, multiple measurements are not necessary since they will give little gain in accuracy.
Except for stalk diameter, phenotypic repeatabilities were higher for correlations determined in wide intrarow plant spacing than for those computed in narrow intrarow plant spacing. This indicated that a higher level of accuracy was achieved when trait measurements were taken on plants spaced 82 cm apart compare to measurements taken on plants spaced 41 cm apart and that more measurements would be needed for plants spaced 41 cm apart to obtain the same level of precision as with wider spaced plants.
Among variables in narrow spaced plants, diameter, stalk weight, and stalk length were the most repeatable variables, plant weight was intermediate while stalk per plant was the least repeatable variable. In wide spaced plants, stalk length was the most repeatable variable, stalk diameter, stalk weight, and plant weight were intermediate and stalk per plant was the least repeatable variable. The high repeatability of stalk length is probably due that in wide spaced plants it no longer has to compete for light as it did in narrow spaced plants.
Kang et al. (1984) indicated that important information on genetic mechanisms conditioning a trait under different environments is missing when only phenotypic correlations coefficients are used. Genetic correlation coefficients for a given trait in two different environments ranged from intermediate to high (Table 3). Since genetic correlations measured in this way can give evidence on the kind of genetics mechanisms involved in the expression of traits (Falconer, 1989), it is safe to assume that the genetic mechanisms implicated in determining all five traits in all environmental combinations were similar as indicated by the relatively high genetic associations. Genetic mechanisms acting in the expression of stalk number per plant at Ardoyne farm varied depending on the year effect. Apparently, the harsh conditions of the 1995-1996 winter affected stalk number more severely than any other trait.
Mariotti (1974) reported correlation values similar to the ones presented herein for stalk diameter, stalk length, number of millable stalks (comparable to number of stalks per plant), and cane yield (comparable to plant weight). Kang et al. (1984) also reported similar results for stalk weight and stalk number, but lower correlations for cane yield. Mariotti (1974) and Kang et al. (1984) values referred to replicated clonal plots while this study is reporting results obtained from first ratoon of seedlings and based on family means. Nevertheless, there is a good agreement among the three studies.
Genetic associations calculated with data taken from wide spaced plants were generally higher than those associations determined with data taken from narrow spaced plants. This difference clearly shows, as it did at the phenotypic level, the effect of the spacing factor on the genetics of a trait. Higher phenotypic and genetic correlation coefficients are obtained when a distance of 82 cm between plants within a row is used as opposed to a distance of 41 cm.
Repeatability values at both intrarow plant spacings were reasonably high compared to other studies. Nevertheless, repeatability values were higher in correlations determined on wide intrarow plant spaced plants than those computed on narrow spaced plants. Number of stalks per plant was the least repeatable trait. Repeatability values between all possible environment combination were high showing a low effect of the environment on those associations.
The authors wish to extend special thanks to the personnel at St. Gabriel Research Station and USDA Ardoyne Farm, Louisiana, their help made this research possible.
BROWN, A.D.H., J. DANIELS, AND B.D.H. LATTER. 1968. Quantitative genetics in sugarcane. I. Analysis of variation in a commercial hybrid sugarcane population. Theor. Appl. Genet. 38:361-369.
CHANG, Y.S., AND S.B. MILLIGAN. 1992. Estimating the potential of sugarcane families to produce elite genotypes using univariate cross prediction methods. Theor. Appl. Genet. 84:662-671.
DE SOUSA-VIEIRA, O., AND S. B. MILLIGAN. 1999. Intrarow plant spacing and family x environment interaction effects on sugarcane family evaluation. Crop Sci. 39:358-364.
EFRON, B., AND R.J. TIBSHIRANI. 1993. An introduction to the bootstrap. Chapman & Hall. New York, NY.
FALCONER, D.S. 1952. The problem of environment and selection. Amer. Nat. 86:293-298.
FALCONER, D.S. 1966. Comments on heritability and repeatability. Int. Soc. Sugar Cane Technol. Sugarcane Breeders Newsletter. 18:15-17.
FALCONER, D. S. 1989. Introduction to quantitative genetics. 3rd ed. Longman Scientific & Technical, England.
GRAVOIS, K.A., S.B. MILLIGAN, AND F.A. MARTIN. 1991. Indirect selection for increased sucrose yield in early sugarcane testing stages. Field Crops Res. 26:67-73.
HANSCHE, P.E., AND R.M. BROKS. 1965. Temporal and spatial repeatabilities of a series of quantitative characters in sweet cherry (Prunus avium L.). Proc. Amer. Soc. Hort. Sci. 86:120-128.
JAMES, N.I., AND J.D. MILLER. 1971. Selection in two seedling crops of four sugarcane progenies. Crop Sci. 11:245-248.
KANG, M.S., J.D. MILLER, AND P.Y.P. TAI. 1984. Clonal and individual repeatability of agronomic traits in sugarcane. J. Am. Soc. Sugar Cane Technol. 3:22-27.
LADD, S.L., D.J. HEINZ, H.K. MEYER, AND B.K. NISHIMOTO. 1974. Selection studies in sugarcane (Saccharum sp. Hybrids). I. Repeatability between selection stages. Proc. Int. Soc. Sugar Cane Technol. 15:102-105.
MARIOTTI, J.A. 1974. The effect of environments on the effectiveness of clonal selection in sugarcane. Proc. Int. Soc. Sugar Cane Technol. 15:89-95.
MILLIGAN, S.B. 1988. The genetic variance-covariance structure of the Louisiana sugarcane breeding population. Ph.D. diss. Louisiana State Univ., Baton Rouge (Diss. Abstr. 88-19963).
MILLER, J.D., AND N.I. JAMES. 1975. Selection in six crops of sugarcane. I. Repeatability of three characters. Crop Sci. 15:23-25.
SMITH, G.A., AND N.I. JAMES. 1969. Association of characters within and repeatability between years in progenies of four sugarcane crosses. Crop Sci. 9:819-821.
TAI, P.Y.P., J.D. MILLER, B.S. GILL, AND V. CHEW. 1980. Correlations among characters of sugarcane in two intermediate selection stages. Proc. Int. Soc. Sugar Cane Technol. 17:1119-1128.