Crop Science 41:15-19 (2001)
© 2001 Crop Science Society of America
CROP BREEDING, GENETICS & CYTOLOGY
Recurrent Selection for Grain Yield in Two Spanish Maize Synthetic Populations
M.I. Valesa,
R.A. Malvarb,
P. Revillab and
A. Ordásb
a Dep. of Crop and Soil Science, Oregon State Univ., Corvallis, OR 97331
b Misión Biológica de Galicia, CSIC, Apartado 28, 36080 Pontevedra, Spain
Corresponding author (aordas{at}cesga.es)
 |
ABSTRACT
|
|---|
The most common maize (Zea mays L.) heterotic pattern in Europe is European flint x U.S. dent. Northern Spain x southern Spain has been proposed as an alternative heterotic pattern. Three maize synthetic populations, namely EPS6 from northern Spain, EPS7 from southern Spain, and EPS10 formed by early American populations, were produced in Pontevedra, Spain. Because of their low yield, the two Spanish synthetic populations were subjected to three cycles of intrapopulation S1 recurrent selection for grain yield. Our objectives were to evaluate the effect of selection on grain yield and other agronomic traits, and to determine the changes in heterosis and general (GCA) and specific (SCA) combining ability brought about by selection. The three original synthetic populations, the three cycles of selection of the two Spanish synthetic populations, and the crosses among the original and the improved cycles were evaluated in two locations in northwest Spain in 1994 and 1995. Yield significantly increased with selection in both Spanish synthetic populations. The GCA improved with selection in both EPS6 and EPS7. Heterosis and SCA did not change significantly with selection. The third cycles of each Spanish synthetic would be the most appropriate maize populations to start a reciprocal recurrent selection to improve the heterotic pattern northern Spain x southern Spain.
Abbreviations: GCA, general combining ability • SCA, specific combining ability
|
INTRODUCTION
|
|---|
MOST MAIZE HYBRID SEED in the USA derives from the Corn Belt Dent race, and only a small portion of the variation available in this race is used (Goodman and Brown, 1988). As it is well known, the most productive single-cross hybrids generally include one parent that traces its ancestry back to a Reid type and the other to a Lancaster type (Darrah and Zuber, 1986; Smith, 1988).
In Europe, the situation is somewhat similar. Hybrids based on the heterotic pattern `Reid x Lancaster' are extensively grown. Also, hybrids obtained by crossing U.S. Corn Belt Dent with European flint inbred lines are used in some areas, and only a few flint lines are used extensively in this continent (Moreno-González et al., 1997; Smith et al., 1991; Messmer et al., 1992, 1993). Collection and evaluation of new European populations would be useful for a possible incorporation of this material to the genetic pool of current selection programs.
Ordás (1991) studied the heterotic relationship between Spanish and Corn Belt maize germplasm in a diallel cross among four Corn Belt populations and five Spanish races. Using heterosis for yield as the measure of dissimilarity, a phenogram showed the existence of three groups of populations: U.S. Corn Belt, northern Spain, and southern Spain. Based on the results of this work, three composites were formed: one with germplasm from northern Spain (EPS6), another with germplasm from southern Spain (EPS7), and finally, a third composite was constituted with early U.S. Corn Belt germplasm (EPS10).
Preliminary results from the crosses among the maize germplasm from northern Spain x southern Spain, northern Spain x U.S. dent, and southern Spain x U.S. dent were promising, and might lead to the development of superior hybrids (Ordás, 1991). The two Spanish populations, EPS6 and EPS7, could be included in a comprehensive breeding system (Eberhart et al., 1967). Because of their low yield, an intrapopulation selection program was started in both populations to improve grain yield independently before starting an interpopulation selection program. Three cycles of S1 recurrent selection for yield were performed in both Spanish synthetic populations. Heterosis, GCA, and SCA in the heterotic patterns northern Spain x southern Spain and Spanish populations x U.S. dent could change from S1 recurrent selection for yield in both Spanish synthetic populations. This must be determined before starting a program of reciprocal selection following the lines of the comprehensive breeding system (Eberhart et al., 1967).
The objectives of this study were: (i) to evaluate the effect of selection on grain yield and other agronomic traits after three cycles of S1 recurrent selection for grain yield, (ii) to determine the changes in heterosis, GCA and SCA with the selection procedure, and (iii) to identify the best Spanish synthetic populations to start reciprocal recurrent selection.
|
MATERIALS AND METHODS
|
|---|
Three synthetics were used for this study: EPS6, EPS7, and EPS10. The synthetics EPS6 and EPS7 have already been described and represent germplasm from northern (humid) and southern (dry) Spain, respectively (Ordás, 1991). EPS10 was made from the early Corn Belt synthetic populations AS-A, AS-B, AS-3(HT)C3 (Peterson et al., 1976), and the well-known, open-pollinated variety Minnesota no. 13 (University Farm strain).
The synthetic populations EPS6 and EPS7 have been subjected to three cycles of intrapopulation S1 recurrent selection for yield. The original populations were selfed and 100 random S1 lines per se were tested. The top 10 yielding S1 lines were intermated in all possible combinations to produce the population of the first cycle of selection (Ordás, 1991). Three cycles of S1 recurrent selection for yield had been completed in both populations following the same scheme. EPS6(S)C0, EPS6(S)C1, EPS6(S)C2, and EPS6(S)C3 are the initial populations and the populations obtained after the first, second and third cycles of selection in EPS6, respectively. EPS7(S)C0, EPS7(S)C1, EPS7(S)C2, and EPS7(S)C3 are the corresponding populations for EPS7. From now on, EPS6(S)C and EPS7(S)C will represent the cycles of S1 recurrent selection of EPS6 and EPS7, respectively.
In 1993, EPS6(S)C, EPS7(S)C, and EPS10 were multiplied by sibmating. Each plant was used only as a male or as a female, but never as both. For each population, 10 rows with 15 plants per row were sown. At least 50 ears were obtained for each population.
The 24 interpopulation crosses represented as EPS6(S)C x EPS7(S)C, EPS6(S)C x EPS10, and EPS7(S)C x EPS10 were produced in 1993. The crosses were made in paired rows, with 15 plants per row. Five sets of paired rows were used for each of the 24 interpopulation crosses. Each plant was used as a male or as a female but never as both. The number of ears obtained in the crosses ranged from 14 in EPS6(S)C2 x EPS7(S)C2 to 52 in EPS7(S)C1 x EPS10. Pollinated ears from each of the five pairs of rows for each variety cross were harvested and shelled in bulk.
EPS6(S)C, EPS7(S)C, EPS10 and the 24 interpopulation hybrids were evaluated along with three hybrid checks, (A639 x A638) x W182B [Minhybrid 7301, with a Minnesota relative maturity (RM) rating of 90 d], A632 x W117 (RM 100 d), and A619 x A632 [Minhybrid 4201, RM 110 d] in a 6 by 6 triple lattice. Experiments were carried out in 1994 and 1995 at two locations in Pontevedra, in northwestern Spain (Salcedo, 42°25' N, 4°57' W, 20 m above sea level, and Barrantes 42°30' N, 8°46' W, 50 m above sea level). Each experimental plot was hand planted and consisted of two rows spaced 0.80 m apart, with 29 two-plant hills spaced 0.21 m apart. Plots were overplanted at both locations and thinned, obtaining a final density of 
60000 plants ha-1 at both locations. The cropping techniques carried out during these field trials were similar to those normally used for maize cultivation. Data taken for each plot were silking (d), plant and ear height (cm), stalk and root lodging (%), grain moisture at harvest (g H20 kg-1), grain yield (weight of grain expressed as Mg ha-1, adjusted at a kernel moisture of 140 g H2O kg-1), number of ears per plant, ear length (cm), and number of ear rows.
Individual analyses of variance were computed for each lattice design. If the relative effectiveness of the lattice design was smaller than 105% the traits were analyzed as randomized complete blocks (Cochran and Cox, 1957). Orthogonal partitions of the treatment mean squares were made for the populations [EPS6(S)C, EPS7(S)C, EPS10], interpopulation crosses, and hybrid checks. The populations were divided in EPS6(S)C and EPS7(S)C. Each partition of the mean square for entries was tested with the corresponding partition of the mean square of the interaction. The means were compared by the LSD method (Steel and Torrie, 1980). All the analyses of data were carried out with the SAS package (SAS Institute, 1989). The general and SCA, as well as the midparent heterosis (Falconer, 1981) were computed in the interpopulation crosses.
|
RESULTS AND DISCUSSION
|
|---|
The average grain yield was 6.9 and 7.3 Mg ha-1 in Salcedo in 1994 and 1995, and 5.7 and 4.2 Mg ha-1 in Barrantes in 1994 and 1995, respectively. The causes of the low yield obtained in Barrantes, especially in the last year, can be attributed to strong corn borer infestation (caused primarily by Sesamia nonagrioides Lef., and secondarily by Ostrinia nubilalis Hbn.). In Barrantes in 1995, average stalk damage was 74%, ear damage was 85%, and stalk lodging was 37%. Cordero et al. (1998), monitoring the corn borer pest also observed a high plant damage in Barrantes in 1995.
The combined analysis of variance over years and locations did not show significant differences among genotypes probably because of a significant genotype x location interaction (data not shown). The unusual results of Barrantes in 1995 due to the high corn borer attack were probably the cause of the genotype x location interaction. In conditions of high corn borer attack the effects of selections were masked by that interaction. It is important to consider that stress conditions may affect gains realized by breeding programs.
After eliminating the environment of Barrantes in 1995, Bartlett's test showed homogeneity of the experimental error variances for grain yield. The combined analysis of variance across the other three environments (Table 1) showed significant differences among genotypes for grain yield and a significant genotype x environment interaction. The variation due to genotypes was divided into sources of variation due to populations, to interpopulation crosses, and to hybrid checks. Significant differences were found within the populations while the environment x populations interaction was not significant. In the populations, significant differences were found for grain yield within EPS6(S)C and EPS7(S)C. In EPS6(S)C there was a highly significant linear regression of grain yield on cycles of selection, while in EPS7(S) there were significant linear and highly significant quadratic regressions of grain yield on cycles of selection. No significant differences were found within the interpopulation crosses or hybrid checks because of the significant interaction with environment. The combined analysis of variance across three environments showed that the S1 recurrent selection program has been effective for improving grain yield in both Spanish synthetic populations. The two Spanish synthetic populations obtained after the first, second, and third cycles of selection had significantly better grain yield than the original populations. In EPS6, grain yield increased linearly during the S1 recurrent selection program, improving 0.5 Mg ha-1 per cycle (13.0% cycle-1). In EPS7 the gain in grain yield per cycle was lower (9.8%) and followed a quadratic trend, decreasing a little bit, but not significantly, in the third cycle of selection (Table 2, Figure 1)
. Grain yield gains per cycle were very similar to those obtained by Ordás (1991). The average rate of response per cycle in S1 recurrent selection for grain yield was 3.5% (Hallauer and Miranda, 1988; Moreno-González and Cubero, 1993; Weyhrich et al., 1998a). The response to selection for grain yield was better in EPS6 than EPS7. This could be explained by the fact that the former population has been selected in the same environment as their constituent varieties.
View this table:
[in this window]
[in a new window]
|
Table 1. Analyses of variance of grain yield of the original and improved maize synthetic populations, the interpopulation crosses, and hybrid checks grown in three environments in Spain
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Means of yield and other agronomic traits evaluated in the original synthetic populations [EPS6(S)C0 and EPS7(S)C0], the populations after each of three cycles (C1, C2, and C3) of S1 recurrent selection for grain yield, the American synthetic population EPS10, and three hybrid checks grown in three environments in Spain
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 1. Grain yield of the original (C0) and the populations obtained after three cycles (C1, C2, and C3) of S1 recurrent selection for yield in two Spanish synthetic populations. The prediction equations are Y = 5.0 + 0.5C and Y = 5.1 + C - 0.3C2 for EPS6(S)C and EPS7(S)C, respectively. Y = yield of grain; C = cycle of selection
|
|
In a program of selection it is of paramount importance that there are no undesirable responses on correlated agronomic characters. Undesirable effects could appear as a consequence of inbreeding because the number of individuals used to advance each cycle of selection was not very high or because some traits change in an undesirable direction. No inbreeding effects were detected in our selection program. We can see, for instance, that plant height did not decrease along the S1 recurrent selection program (Table 1). Isozyme data (Revilla et al., 1997) also confirmed the lack of inbreeding depression in EPS6(S)C and EPS7(S)C. We used 10 S1 progenies to form the selected populations in each cycle of selection. Weyhrich, et al. (1998b) indicated that drift becomes a stronger force in altering allele frequencies than selection when fewer than 10 lines are recombined, and there does not seem to be an advantage, at least in the short term, to recombining more than 10 lines per cycle of selection.
Changes in other agronomic traits, when selection was done to improve grain yield following an S1 recurrent selection procedure, have been variable in similar breeding programs. The additive correlation among yield and other traits is low (Hallauer and Miranda, 1988), so no correlated responses among yield and other traits should be expected, at least on a short-term basis. Most of the changes in our study were in the desired direction (Table 2). Only a couple of undesired effects occurred. Silking was significantly delayed by 2 d in the third cycle of selection compared to the original population in both EPS6 and EPS7. Also, grain moisture increased significantly in EPS7 when comparing the original population with the population resulting after the third cycle of selection. Then, it seems that increases in grain yield were made at the expense of maturity. Although the maturity of both synthetic populations still remains in a range acceptable for the conditions of the northwest of Spain, it does not seem advisable to carry out more cycles of selection for yield without paying attention to maturity.
Grain yield in the interpopulation crosses increased linearly (Figure 2)
with the selection procedure. EPS6(S)C3 x EPS7(S)C and EPS7(S)C3 x EPS6(S)C showed significantly better grain yield than EPS6(S)C0 x EPS7(S)C and EPS7(S)C0 x EPS6(S)C, respectively (Table 3). A significant difference was also observed in GCA (Table 3) from the original cycle to the third one in both Spanish synthetic populations. SCA did not change significantly from C0 to C3 (data not shown). Improvements in grain yield were realized at the expense of the additive variance, which is in agreement with the loss in additive variance observed (data not shown).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 2. Grain yield of the crosses of the original (C0) and the populations obtained after three cycles (C1, C2, and C3) of S1 recurrent selection for yield in two Spanish synthetic populations. Each point in the graph is the mean of each synthetic population crossed to the corresponding four populations (C0, C1, C2, and C3) of the other synthetic
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Means for grain yield and general combined ability (GCA) in the interpopulation crosses between the two Spanish synthetic populations EPS6 and EPS7 over three environments in Spain
|
|
Recurrent selection methods can be used to develop heterotic groups or to improve existing ones. The mid-parent heterosis of grain yield (Table 4) did not change significantly from the cross of original populations EPS6(S)C0 x EPS7(S)C0 to the cross of the populations of the third cycle of selection for grain yield EPS6(S)C3 x EPS7(S)C3, and did not differ significantly from the heterosis observed among EPS6(S)C x EPS10 and EPS7(S)C x EPS10. This means that the heterotic patterns northern Spain x southern Spain and Spanish populations x U.S. dent suggested by Ordás (1991) continue to be appropriate after improving grain yield in the two synthetic populations EPS6 and EPS7. In a similar selection procedure, Garay et al. (1996) found a loss of heterosis for yield after selection. Álvarez et al. (1993) and Sinobas and Monteagudo (1996) proposed a heterotic pattern similar to Spanish populations x U.S. dent. The heterotic pattern northern Spain x southern Spain was first proposed by Ordás (1991) with the populations involved in this study, and S1 selection for grain yield has reinforced the suitability of this heterotic pattern. The two Spanish populations under selection belong to different parts of Spain. In a recent study, Revilla et al. (1998), using isoenzymes to study the genetic diversity of Spanish landraces of maize, suggested that landraces from southern Spain came mainly from Central America, and most landraces from northern Spain derived from germplasm introductions arriving from North America. The genetic distance due to the origin of maize (northern and southern Spain) (Revilla et al., 1998), and between the Spanish populations and U.S. dent germplasm (Ordás et al., 1994) could explain their heterotic behavior.
View this table:
[in this window]
[in a new window]
|
Table 4. Midparent heterosis of grain yield of the interpopulation crosses evaluated in three environments in Spain
|
|
S1 recurrent selection to improve grain yield could be continued in EPS6 because of the linear response, although it would not be advisable to do it, as stated before, without paying attention to maturity. In EPS7, it is not clear whether selection should be continued or not because of the quadratic effect observed. There were no significant differences among the different cycles of selection; nevertheless, to start an interpopulation recurrent selection following the scheme of the comprehensive breeding system (Eberhart et al., 1967), the populations resulting after the third cycle should be used because they are expected to contain fewer deleterious genes. They also have significantly higher yield than the original populations, greater GCA, and have maintained heterosis. An interpopulation selection method will continue improving both synthetic populations separately and at the same time finding the best combinations in the crosses among them to consolidate the heterotic pattern northern Spain x southern Spain, providing an alternative heterotic pattern to the commonly used in Europe, European flint x U.S. dent. These improved populations could also be useful to enhance the genetic basis of breeding programs conducted in other temperate areas beyond western Europe.
|
ACKNOWLEDGMENTS
|
|---|
M. Isabel Vales acknowledges a fellowship from the Diputación Provincial de Pontevedra.
|
NOTES
|
|---|
Part of a dissertation submitted by the senior author in partial fulfillment of the requirements for the Ph.D. degree at the Univ. of Vigo, Spain. Research supported by the Committee for Sci. and Tech. of Spain (project AGF95-0891) and the local government (Diputación Provincial) of the province of Pontevedra, Spain.
Received for publication December 10, 1999.
|
REFERENCES
|
|---|
- Álvarez, A., G. Garay, J. Jiménez, and J.I. Ruiz de Galarreta. 1993. Heterosis entre dos sintéticos de maíz expresada sobre caracteres morfológicos y reproductivos. Invest. Agrar. Prod. Prot. Veg. 8:333–340.
- Cochran, W.G., and G.M. Cox. 1957. Experimental designs. 2nd ed. John Wiley & Sons, New York.
- Cordero, A., R.A. Malvar, A. Butrón, P. Revilla, P. Velasco, and A. Ordás. 1998. Population dynamics and life-cycle of corn borers in south Atlantic European coast. Maydica 43:5–12.[ISI]
- Darrah, L.L., and M.S. Zuber. 1986. 1985 United States farm maize germplasm base and commercial breeding strategies. Crop Sci. 26:1109–1113.[Abstract/Free Full Text]
- Eberhart, S.A., M.N. Harrison, and F. Ogada. 1967. A comprehensive breeding system. Der Züchter 37:169–174.
- Falconer, D.S. 1981. Introduction to quantitative genetics. 2nd ed. Longman Inc., New York.
- Garay, G., E. Igartua, and A. Álvarez. 1996. Responses to S1 selection in flint and dent synthetic maize populations. Crop Sci. 36: 1129–1134.[Abstract/Free Full Text]
- Goodman, M.M., and W.L. Brown. 1988. Races of corn. p. 33–79. In G.F. Sprague and J.W. Dudley (ed.) Corn and corn improvement. 3rd ed. Agron. Monogr. 18. ASA, CSSA, and SSSA, Madison, Wisconsin.
- Hallauer, A.R., and J.B. Miranda Fo. 1988. Quantitative genetics in maize breeding. 2nd ed. Iowa State Univ. Press, Ames, Iowa.
- Messmer, M.M., A.E. Melchinger, J. Boppenmaier, R.G. Herrmann, and E. Brunklaus-Jung. 1992. RFLP analyses of early-maturing European maize germplasm. I. Genetic diversity among flint and dent inbreds. Theor. Appl. Genet. 83:1003–1012.
- Messmer, M.M., A.E. Melchinger, R.G. Herrmann, and J. Boppenmaier. 1993. Relationships among early European maize hybrids: II. Comparison of pedigree and RFLP data. Crop Sci. 33:944–950.[Abstract/Free Full Text]
- Moreno-González, J., and J.I. Cubero. 1993. Selection strategies and choice of breeding methods. p. 281–313. In M.D. Hayward, N.O. Bosemark and I. Romagosa (ed.) Plant breeding. Principles and prospects. Chapman & Hall, London, UK.
- Moreno-González, J., F. Ramos-Gourcy, and E. Losada. 1997. Breeding potential of European flint and earliness-selected U.S. Corn Belt dent maize populations. Crop Sci. 37:1475–1481.[Abstract/Free Full Text]
- Ordás, A. 1991. Heterosis in crosses between American and Spanish populations of maize. Crop Sci. 31:931–935.[Abstract/Free Full Text]
- Ordás, A., R.A. Malvar, and A.M. De Ron. 1994. Relationship among American and Spanish populations of maize. Euphytica 79:149–161.
- Peterson, R.H., J.L. Geadelmann, E.H. Rinke, and J.C. Sentz. 1976. Registration of six maize germplasm populations. Crop Sci. 16:605–606.[Free Full Text]
- Revilla, P., P. Soengas, R.A. Malvar, M.E. Cartea, and A. Ordás. 1998. Isozyme variation and historical relationships among the maize races of Spain. Maydica 43:175–182.
- Revilla, P., M.I. Vales, R.A. Malvar, and A. Ordás. 1997. Allozyme frequencies, heterozygosity and genetic distances following S1 recurrent selection in two synthetic maize populations. Theor. Appl. Genet. 95:1057–1061.
- SAS Institute, 1989. SAS/STAT user's guide. Version 6, 4th ed. SAS Institute Inc., Cary, NC.
- Sinobas, J., and I. Monteagudo. 1996. Heterotic patterns among U.S. Corn Belt and Spanish maize populations. Maydica 41:143–148.
- Smith, J.S.C. 1988. Diversity of United States hybrid maize germplasm: Isozymic and chromatographic evidence. Crop Sci. 28:63–69.[Abstract/Free Full Text]
- Smith, J.S.C., O.S. Smith, and S.J. Wall. 1991. Associations among widely used French and U.S. maize hybrids as revealed by restriction fragment length polymorphism. Euphytica 54:263–273.
- Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics. 2nd ed. McGraw-Hill, New York.
- Weyhrich, R.A., K.R. Lamkey, and A.R. Hallauer. 1998a. Responses to seven methods of recurrent selection in the BS11 maize population. Crop Sci. 38:308–321.[Abstract/Free Full Text]
- Weyhrich, R.A., K.R. Lamkey, and A.R. Hallauer. 1998b. Effective population size and response to S1–progeny selection in the BS11 maize population. Crop Sci. 38:1149–1158.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
G. Sandoya, A. Butron, A. Alvarez, A. Ordas, and R. A. Malvar
Direct Response of a Maize Synthetic to Recurrent Selection for Resistance to Stem Borers
Crop Sci.,
January 16, 2008;
48(1):
113 - 118.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
