Crop Science 40:1233-1237 (2000)
© 2000 Crop Science Society of America
CROP BREEDING, GENETICS & CYTOLOGY
Repeatability and Genotype x Environment Interaction of Coleoptile Length Measurements in Winter Wheat
Frederic Hakizimanaa,
Scott D. Haleyb and
E.Brent Turnipseeda
a Plant Science Dep., South Dakota State Univ., Brookings, SD 57007 USA
b Soil and Crop Sciences Dep., Colorado State Univ., Fort Collins, CO 80523 USA
shaley{at}lamar.colostate.edu
 |
ABSTRACT
|
|---|
Coleoptile length is often associated with fall stand establishment in winter wheat (Triticum aestivum L.). While coleoptile length evaluation and improvement is a common objective of wheat breeding programs, little information is presently available to guide effective evaluation efforts. The objectives of this study were to assess the association between kernel and test weight and coleoptile length, to determine the inherent precision of a coleoptile length screening procedure through repeatability analysis, and to assess the presence and nature of genotype x environment (G x E) interaction for coleoptile length. Seed samples from 45 winter wheat genotypes planted in a randomized complete block design at four locations in the 1994-1995 growing season were used to evaluate G x E interaction. A subset of samples of 15 lines from two contrasting environments (Selby, high kernel and test weight; Winner, low kernel and test weight) were used to estimate repeatability (R) from five repeated measurements of the same seed sample. Analysis of variance revealed highly significant (P < 0.01) G x E interactions for coleoptile length measurements from seed samples from the four locations. Spearman rank correlations between genotype means at the four environments were highly significant
, suggesting that G x E interactions were caused by differences in scale among environments and not by differences in genotype rankings among environments. Estimates of repeatability of coleoptile length measurements were high and not significantly different between samples from the two contrasting environments (Selby:
; Winner:
). Our results suggest that coleoptile length is a moderately heritable trait and that seed source for coleoptile length measurements has little or no effect on coleoptile length expression or the variation experienced with the evaluation procedures.
Abbreviations: AYT, advanced yield trial • G x E, genotype x environment interaction • R, repeatability
|
INTRODUCTION
|
|---|
THROUGHOUT THE NORTHERN GREAT PLAINS, recommended winter wheat seeding dates generally represent a compromise between optimum fall stand establishment and disease (e.g., wheat streak mosaic and barley yellow dwarf viruses) and insect avoidance (e.g., various cereal aphid species). In these areas, optimizing this compromise is especially critical because of frequent occurrence of winter-killing temperatures at northern latitudes and the synergistic effect of fall insect infestation and viral infection on winter survival (Endo and Brown, 1962; Kieckhefer and Kantack, 1988). Research has demonstrated the importance of the length of the coleoptile (protective sheath that covers the shoot during emergence) in achieving optimum fall stand establishment (Allan et al., 1959, 1962; Vogel et al., 1963), particularly when seed is planted deep to reach moisture in dry soils.
The association between coleoptile length and other plant characteristics has been investigated in several studies. Reports of reduced coleoptile length, and inferior fall stand establishment, were made simultaneously with the deployment of semidwarfing genes, such as Rht1 and Rht2 (Allan et al., 1959, 1961). The physiological basis of reduced plant height and coleoptile length in such germplasm has been demonstrated to be directly related to insensitivity to the hormone gibberellic acid (GA3; Allan et al., 1959).
Evidence for other trait associations with coleoptile length, either within Rht1/Rht2-based semidwarf germplasm or not, have not been so conclusive. In an early study, Allan et al. (1965) reported a lack of association between kernel weight and coleoptile length among GA3-insensitive semidwarf lines. Parodi et al. (1970), using a six-parent diallel cross analysis of coleoptile elongation in wheat, indicated that larger seeds produced longer coleoptiles than small seeds. In a greenhouse study on early plant development in barley (Hordeum vulgare L.), Kaufmann and Guitard (1967) concluded that seedling growth was most rapid from large seeds and slowest from small seeds, yet no direct influence of coleoptile length was noted. Kaufmann (1968), using low- and high-kernel weight samples obtained by sieving, also reported that seed size was not associated with coleoptile length in oat (Avena sativa L.) or wheat yet a positive association was observed with barley. Clarification of the existence of associations between coleoptile length and seed quality (as determined by kernel weight and test weight) would provide useful information with regard to choice of seed sources for use in coleoptile length improvement programs.
Genotype x environment interactions may be defined as the failure of genotypes to have the same relative performance from one environment to another (Baker, 1988; Yang and Baker, 1991). Such interactions are a significant challenge to plant breeders because they complicate breeding procedures and limit the usefulness of selection in any one environment (Easton and Clement, 1973). Knowledge of the presence and type of G x E interaction can help breeders make informed decisions to optimize breeding methods, selection intensity, and testing procedures (Baker, 1969). Studies dealing with G x E interaction have suggested that they are usually due to inconsistent genotypic responses to temperature, soil moisture, soil type, or fertility level from location to location and year to year (Liang et al., 1966). Variation in these environmental factors can therefore cause yield and its components (e.g., kernel number and kernel weight) to vary from one environment to another.
Knowledge of the influence of seed quality on the conduct of coleoptile length screening procedures would potentially help guide effective selection efforts. Therefore, the objectives of this study were (i) to assess the association between seed quality (as determined by kernel weight and test weight) and coleoptile length, (ii) to determine the inherent precision of a coleoptile length screening procedure through repeatability analysis, and (iii) to assess the presence and nature of G x E interaction for coleoptile length measurements.
|
Materials and methods
|
|---|
Source Material
Seed of 45 winter wheat genotypes (experimental lines and released cultivars) grown in the South Dakota Advanced Yield Trial (AYT) and harvested in 1995 was evaluated for coleoptile length. This set of genotypes included those carrying height-reducing genes (semidwarf) and those lacking height-reducing genes (tall). The trial was grown as a randomized complete block design, with three replications at each of nine locations in South Dakota. Seed samples from four of these locations (dryland environments at Selby, Winner, and Wall and an irrigated environment near Pierre) were selected for coleoptile length evaluation on the basis of variable grain filling patterns among the locations and yield and test weight data obtained after harvest. These four locations are widely separated geographically and growing conditions during the study were considered representative of the typical range of variation in growing conditions and grain filling patterns for winter wheat in South Dakota. A subset of 15 genotypes from the AYT at Selby and Winner were used to estimate repeatability from repeated measures of the same samples. These genotypes were chosen based on diversity for agronomic characteristics and prior coleoptile length information. The experimental unit for coleoptile length evaluation consisted of each of the three replications for the G x E analysis and a composite of the three replications from Winner and Selby for the repeatability analysis.
Coleoptile Length and Other Measurements
Coleoptile length was measured using a blotter-paper germination protocol (Dilday et al., 1990; J. Quick, Colorado St. Univ., 1993, personal communication). Forty seeds of each genotype per experimental unit were placed with the germ end down 
0.5 cm apart and 5 cm from the bottom on a wet germination towel (no. 76 germination paper; Anchor Paper Co., St. Paul, MN). A second wet germination towel was placed on the top of the seeds and the samples were rolled loosely and secured with a no. 14 rubber band. The samples were placed vertically in plastic trays and placed in a dark incubator at 4°C for 4 d to reduce dormancy. After 4 d, samples were placed in a second dark incubator at 15°C for 16 d. Coleoptile length of each sample was measured by placing a premarked plastic sheet over the sample and recording directly on a computer. Germination percentage varied among samples (data not shown); the average coleoptile length of the germinated seedlings (ranging from 30–40 per sample) was the replicate value used in the statistical analyses.
Kernel weight was determined on samples of combine-harvested seed. One-hundred seeds from each experimental unit were counted, placed in an oven at 60°C overnight, and weighed to the nearest milligram. Test weight (grain volume weight) was estimated from a small sample (0.47 L) from the combine-harvested seed.
Statistical Analysis
For the G x E analysis, data were analyzed as a randomized complete block design for each location. Following confirmation of error variance homogeneity (according to Gomez and Gomez, 1984), a combined analysis across environments was conducted with both genotypes and environments considered as fixed factors. Fischer's protected least significant differences (LSD) were used for comparison of means within each environment.
For the repeatability analysis, data from each location were analyzed as a randomized complete block design with repeated measurements considered as blocks. The repeatability (R), and its standard error (SE), of coleoptile length measurements were estimated using a one-way analysis of variance procedure (Becker, 1992)
 | (1) |
where 
2B is the between-genotype variance and 2E is the within-genotype variance.
 | (2) |
where k is the number of measurements per genotype and n is the number of genotypes.
The correlation among genotype rankings for coleoptile length measurements at different locations was estimated using Spearman's rank correlation analysis (Steel and Torrie, 1980). A linear Pearson correlation matrix was used to estimate the association between coleoptile length and kernel weight and test weight.
|
Results and discussion
|
|---|
Genotype x Environment Interaction
The analysis of variance revealed highly significant (P < 0.01) genotypic differences in coleoptile length, kernel weight, and test weight from seed samples harvested at each location. Marked differences in grain quality (as determined by kernel weight and test weight) were observed among the four environments (Table 1)
. The environment with the poorest grain quality traits was Winner, where grain filling was adversely affected by leaf disease pressure from leaf rust [Puccinia recondita Roberge ex Desmaz. f. sp. tritici (Eriks. & E. Henn.) D.M. Henderson] and tan spot [Pyrenophora triticirepentis (Died.) Drechs.] combined with drought and heat stress late in the grain filling period. Although differences in kernel weight and test weight among the four environments were apparent, the expression of coleoptile length (mean and range) from samples harvested at these locations was not drastically affected by these differences (Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1 Mean, minimum, and maximum for coleoptile length, kernel weight, test weight, and plant height of 45 winter wheat genotypes grown in four South Dakota environments
|
|
The combined analysis of variance revealed highly significant environment, genotype, and G x E interaction effects for coleoptile length, kernel weight, and test weight. Yang and Baker (1991) recognized two main categories of G x E interaction, crossover and noncrossover interaction, with changes in genotype rank order or differences in response scale between environments determining the nature of the interaction. To assess the nature of G x E interaction for coleoptile length in our study, Spearman's rank correlation coefficients were estimated using genotype means for coleoptile length from the four environments. Each of the six correlation coefficients was highly significant
, suggesting that the G x E interaction for coleoptile length was largely due to differences in scale among the environments.
To determine if the nature of G x E interactions was consistent when considering tall or semidwarf genotypes separately, combined analyses of variance were done within the tall
and semidwarf
height classes. In both instances, highly significant G x E interaction effects were observed for coleoptile length. Spearman's rank correlations within tall and semidwarf height classes, using genotype means for coleoptile length from the four environments, were nearly identical (rs 
0.90) to those observed when considering tall and semidwarf entries as one set of genotypes. This suggests that the lack of crossover G x E interaction for coleoptile length is consistent regardless of the presence of major semidwarfing genes.
Simple Pearson correlation analysis was used to assess the association between coleoptile length and grain quality traits (as determined by kernel and test weight). Associations between coleoptile length and kernel weight were not significant in any environment, whereas associations between coleoptile length and test weight were significant (P < 0.05 or < 0.01) in each environment (Table 2)
. Thus, kernel soundness (as estimated through test weight) appeared to have a greater influence on coleoptile length than kernel weight. This finding is in agreement with previous reports (Allan et al., 1965; Kaufmann, 1968), indicating that kernel weight is not associated with coleoptile length. A high degree of similarity between these analyses and analyses done within height classes (data not shown) suggested that the presence or absence of major semidwarfing genes has no effect on the association between coleoptile length and either test weight or kernel weight.
View this table:
[in this window]
[in a new window]
|
Table 2 Pearson correlation coefficients for coleoptile length vs. kernel weight and test weight of 45 winter wheat genotypes grown in four South Dakota environments
|
|
Repeatability
The main objective of this portion of our study was to evaluate the inherent precision of the coleoptile length screening procedure using samples from environments with vastly different seed quality attributes. Seed samples of 15 genotypes from the two environments showing the most extreme differences in kernel and test weight (Selby and Winner) were used to estimate repeatability of the coleoptile length measurements. The genotypes (mostly released cultivars) selected for the repeatability analysis were genetically diverse and represented the full range in variability observed for coleoptile length, kernel weight, and test weight at both environments (Table 3)
.
View this table:
[in this window]
[in a new window]
|
Table 3 Height class and coleoptile length, kernel weight, and test weight of 15 winter wheat genotypes from seed samples harvested at Selby and Winner, SD
|
|
The one-way analysis of variance (Table 4)
revealed highly significant between-genotype variance at both locations. Repeatability estimates, derived from partitioning of variance components from expectations of mean squares, were relatively high and not significantly different between the two locations
. These results suggest that the precision of the screening procedure used in our study was largely unaffected by differences in seed quality used for coleoptile length screening.
View this table:
[in this window]
[in a new window]
|
Table 4 One-way analysis of variance of coleoptile length of 15 winter wheat genotypes from seed harvested at Selby and Winner, SD**
|
|
Knowledge of repeatability can also provide information on the relative level of heritability of a character. Repeatability establishes the upper limits of heritability (Falconer, 1981) as:
 | (3) |
where VG is the genotypic variance, VP phenotypic variance, VA additive genetic variance, VG/VP is the broad sense heritability, and VA/VP is the narrow sense heritability.
As repeatability estimates were relatively high at both locations in our study, actual heritability estimates for coleoptile length would also be expected to be relatively high. These results are similar to those of Chowdhry and Allan (1963), who indicated that due to the high heritability for coleoptile length, selection for coleoptile length would be effective and could be practiced in early segregating generations.
|
Conclusions
|
|---|
The presence and nature of G x E interaction and repeatability of coleoptile length measurements were estimated to potentially provide information that would be useful for applied screening efforts to improve coleoptile length in wheat. Observed G x E effects were determined to be due to differences in scale (range of coleoptile length expression) among environments and not due to changes in genotype rankings for coleoptile length. Genotypic variation in kernel weight was not associated with coleoptile length expression, whereas kernel soundness (as determined by test weight) was determined to be positively associated with coleoptile length.
Estimates of repeatability (the theoretical upper limit of heritability) for coleoptile length were relatively high, suggesting that response to selection for coleoptile length would be realized in improvement programs. Interestingly, relatively large differences in seed quality between locations had no effect on repeatability estimates derived from measurements from the contrasting seed samples. The seed source used for coleoptile length measurements appears to have little or no effect on coleoptile length expression or the variation encountered with the evaluation procedures.Liang Heyne Walter 1966
|
NOTES
|
|---|
Joint contribution Dep. of Plant Science, South Dakota State Univ. and Soil and Crop Sciences Dep., Colorado State Univ. Research partially supported through funding from the South Dakota Wheat Commission. South Dakota Agric. Exp. Stn. Journal Series no. 3142. Manuscript submitted by the senior author in partial fulfillment of the requirements for the M.S. degree in Agronomy at South Dakota State Univ.
Received for publication October 8, 1999.
|
REFERENCES
|
|---|
- Allan R.E., Vogel O.A., Craddock J.C. Comparative response to GA3 of dwarf, semidwarf, and tall winter wheat varieties. Agron. J. 1959;51:737-740.[Abstract/Free Full Text]
- Allan R.E., Vogel O.A., Craddock J.C. Effect of gibberellic acid upon seedling emergence of slow and fast emerging wheat varieties. Agron. J. 1961;53:30-32.[Abstract/Free Full Text]
- Allan R.E., Vogel O.A., Peterson C.J. Seedling emergence rate of fall-sown wheat and its association with plant height and coleoptile length. Agron. J. 1962;54:347-350.[Abstract/Free Full Text]
- Allan R.E., Vogel O.A., Russel T.S., Peterson C.J. Relation of seed and seedling characteristics to stand establishment of semidwarf wheat selections. Crop Sci. 1965;5:5-8.
- Baker R.J. Genotype x environment interactions in yield of wheat. Can. J. Plant Sci. 1969;49:743-751.
- Baker R.J. Tests for crossover genotype x environmental interactions. Can. J. Plant Sci. 1988;68:405-410.
- Becker W.A. Manual of quantitative genetics, 5th ed Pullman, WA: Academic Enterprises, 1992.
- Chowdhry A.R., Allan R.E. Inheritance of coleoptile length and seedling height and their relation to plant height of four winter wheat crosses. Crop Sci. 1963;3:53-58.
- Dilday R.H., Mgonja M.A., Amonsilpa S.A., Collins F.C., Wells B.R. Plant height vs. mesocotyl and coleoptile elongation in rice: Linkage or pleiotropism?. Crop Sci. 1990;30:815-818.[Abstract/Free Full Text]
- Easton H.S., Clement R.J. The interaction of wheat genotypes with specific factor of the environment. J. Agric. Sci. (Cambridge) 1973;80:43-52.
- Endo R.M., Brown C.M. Survival and yield of winter cereals affected by yellow dwarf. Phytopathology 1962;52:614-627.
- Falconer D.S. Introduction to quantitative genetics, 2nd ed New York: Longman, 1981.
- Gomez K.A., Gomez A.A. Statistical procedures for agricultural research, 2nd ed New York: John Wiley & Sons, 1984.
- Kaufmann M.L. Coleoptile length and emergence in varieties of barley, oats, and wheat. Can. J. Plant Sci. 1968;48:357-361.
- Kaufmann M.L., Guitard A.A. The effect of seed size on early plant development in barley. Can. J. Plant Sci. 1967;47:73-78.
- Kieckhefer R.W., Kantack B.H. Yield losses in winter grains caused by cereal aphids in South Dakota. J. Econ. Entomol. 1988;81:317-321.[ISI]
- Liang G.H., Heyne E.G., Walter T.L. Estimates of variety x environmental interactions in yield tests of three small grains and their significance on breeding programs. Crop Sci. 1966;6:135-139.
- Parodi P.C., Paterson F.L., Nyiquist W.E. A six-parent diallel cross analysis of coleoptile elongation in wheat, Triticum aestivum L. Crop Sci. 1970;10:587-590.[Abstract/Free Full Text]
- Steel R.G.D., Torrie J.H. Principles and procedures of statistics, a biometrical approach, 2nd ed New York: McGraw-Hill, 1980.
- Vogel O.A., Allan R.E., Peterson C.J. Plant and performance characteristics of semidwarf wheats producing most efficiently in eastern Washington. Agron. J. 1963;55:397-398.[Abstract/Free Full Text]
- Yang R.C., Baker R.J. Genotype x environment interactions in two wheat crosses. Crop Sci. 1991;31:83-87.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
G. Bai, M. K. Das, B. F. Carver, X. Xu, and E. G. Krenzer
Covariation for Microsatellite Marker Alleles Associated with Rht8 and Coleoptile Length in Winter Wheat
Crop Sci.,
July 1, 2004;
44(4):
1187 - 1194.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
