Published in Crop Sci. 43:2065-2070 (2003).
© 2003 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP BREEDING, GENETICS & CYTOLOGY
Best Types of Maize Hybrids for the Western High Plains of the USA
F. R. Guillen-Portala,
W. K. Russell*,b,
D. D. Baltenspergerd,
K. M. Eskridgec,
N. E. D'Croz-Masonb and
L. A. Nelsonb
a Northwestern Agricultural Research Center, 4575 MT HWY 35, Kalispell, MT 59901
b Dep. of Agronomy and Hortic., Univ. of Nebraska, Lincoln, NE 68583
c Dep. of Biometry, Univ. of Nebraska, Lincoln, NE 68583
d Panhandle Res. and Ext. Center, Univ. of Nebraska, Scottsbluff, NE 69361
* Corresponding author (krussell3{at}unl.edu).
 |
ABSTRACT
|
|---|
Maize (Zea mays L.) production in the western High Plains of the USA occurs under dryland and irrigated conditions. Thus, an ideal maize hybrid for this region would have stable and high average performance across environments that are highly variable in productivity. Stability of a hybrid in this research was defined as its across-environmental variance. Nine commercial single-cross hybrids and 36 related double crosses were evaluated for grain yield in four irrigated and eight dryland environments in Nebraska, Colorado, and Wyoming. Environmental means ranged from 2.10 to 12.01 Mg ha-1. The average superiority of the single crosses over the double crosses in mean grain yield was 11.5%. However, the double crosses were generally more stable than the single crosses. This was true regardless whether stability was measured across all 12 environments or only across the eight dryland environments. The greater stability of the double-cross hybrids in the dryland environments was the result of fewer extremely low or high yields compared with the single crosses. Hybrids were ranked considering both stability and mean performance with a safety-first index. Values of this index depended on the value chosen for acceptable minimal yield. Across all 12 environments, a double-cross hybrid ranked the best when the minimal acceptable yield was <2.68 Mg ha-1, and across the eight dryland environments this happened when the minimal acceptable yield was <1.88 Mg ha-1. Possible use of other types of heterogeneous maize hybrids in the Western High Plains are discussed.
|
INTRODUCTION
|
|---|
THE WESTERN HIGH PLAINS of the USA is a region that includes the Nebraska Panhandle, northeastern Colorado, and southeastern Wyoming. Total maize production in this region in 2000 was >700000 ha (Colorado Agricultural Statistics Service, 2002; Nebraska Agricultural Statistics Service, 2002; Wyoming Agricultural Statistics Service, 2002). The climate of the western High Plains is characterized by low moisture. For example, in the Nebraska Panhandle the average annual precipitation is <450 mm (Owenby and Ezell, 1992). Nonetheless, acreage of maize in the Nebraska Panhandle increased by 35% from 1995 to 1999, and in 1999, maize was the second most important crop in the Panhandle (Nebraska Agricultural Statistics Service, 2002). Similar increases occurred in other parts of the western High Plains. Although most production of maize in this region occurs under irrigation, most of the recent increase in acreage has been nonirrigated (dryland) production. The average grain yield under dryland production in the Nebraska Panhandle from 1995 to 1999 was 3.50 Mg ha-1, whereas the average yield under irrigation production during this same period was 8.50 Mg ha-1. For only dryland production, there was >100% difference between the lowest and highest average annual maize yields from 1995 to 1999. Therefore, the ideal maize cultivar adapted to the western High Plains would perform well under both high-yielding, irrigated environments and variable, lower-yielding, dryland environments.
Currently, nearly all maize cultivars grown in the USA are single-cross or modified single-cross hybrids. The primary reason for the transition from heterogeneous stands of open-pollinated varieties and to double-cross hybrids and then to homogeneous stands of single-cross hybrids during the mid-20th century was the higher mean yield of the latter. Eberhart and Russell (1969) compared the performance of 45 single-cross hybrids with a balanced set of 45 related double-crosses. The mean yield of the 10 best single crosses was 7.10 Mg ha-1, whereas the five best double crosses averaged 6.96 Mg ha-1. Superiority in mean yield of single crosses compared with either three-way or double crosses is expected on the basis of single-locus additive and dominance types of gene action (Hallauer and Miranda, 1988). Another cause of the observed superiority of single crosses may be more uniform plant-to-plant competition. Considerable evidence indicates that the effect of neighborhood competition is undercompensatory. That is, gains in productivity realized by plants that are competitively advantaged are less than losses in productivity from competitively disadvantaged plants (Hoekstra et al., 1985; Jedel et al., 1988).
Selecting a cultivar for the diverse and variable western High Plains environments requires that farmers consider stability of performance in addition to mean or maximum performance. Numerous researchers have reported that the genotype x environmental variance for double crosses is less than for single crosses (Sprague and Federer, 1951; Eberhart and Russell, 1969; Weatherspoon, 1970), although this observation has not been documented for more recently developed elite hybrids. The cause of this greater stability of the double crosses probably was their greater genetic heterogeneity. This attribute of a genetically heterogeneous population has been referred to as genetic homeostasis (Lerner, 1954) or population buffering (Allard and Bradshaw, 1964).
When both mean yield and stability are selective factors, then a statistic or index that incorporates both variables is needed to obtain a single ranking of the cultivars. Numerous statistics have been proposed, and many of these have been reviewed by Kang and Pham (1991) and Hühn (1996). The limitation of these statistics is that they do not explicitly allow for weighting of mean performance and stability based on attitudes of farmers toward risk. These attitudes determine which cultivars farmers will prefer. Producers who are more risk averse will weight stability more heavily than those who are less risk averse. Eskridge et al. (1991) described a safety-first index for ranking cultivars that was based on a safety-first model proposed by Roy (1952). In this index, risk aversion is accounted for by specifying a minimal acceptable value of yield. Another advantage of this safety-first index is that any one of several measures of stability can be used.
Numerous measures of stability have been proposed. Lin et al. (1986) grouped these into three types. Type 1 stability parameters measure a genotype's variance in performance across environments. Type 2 stability parameters measure the responsiveness of a genotype to environments relative to other genotypes in the test. The coefficient obtained by regressing a genotype's performance at each environment on the productivity of that environment (Finlay and Wilkinson, 1963; Eberhart and Russell, 1966) is a Type 2 measure of stability. A Type 3 measure is the residual mean square from the regression model (Eberhart and Russell, 1966; Perkins and Jinks, 1968). Both Type 2 and Type 3 stability measures are model-based, and their relative importance depends on the other genotypes in the test. Thus, they are empirical, not biological measures of stability. In contrast, the Type 1 measure is independent of other genotypes in the test and is analogous to the biological concept of homeostasis (Lin et al., 1986).
The goal of this research was to determine for elite, modern maize germplasm whether heterogeneity conferred greater stability for grain yield across a representative set of environments in the western High Plains of the USA. If this was true, then a second question was whether this advantage in stability resulted in greater values of a stability-first index. Comparisons were made between elite, single-cross hybrids and a set of related double crosses. The implications of the safety-first rankings on the types of maize hybrids best suited for commercial production in this region are then discussed.
|
MATERIALS AND METHODS
|
|---|
The cultivars were nine commercial single-cross hybrids (Cargill 3677 and 4111; DeKalb Brand DK449 and DK442; Mycogen 2500, 2275, and 2395; Ottilie RO2431 and RO2439)1 and all possible 36 double-cross hybrids (reciprocals excluded) produced by crossing the single crosses. The commercial single crosses were chosen because they exhibited good adaptability to the Western High Plains in dryland trials conducted at the University of Nebraska Panhandle Research and Extension Center. They were denoted as H1 through H9, according to the order listed above, and the double crosses as H1 x H2, H1 x H3, etc.
All hybrids were evaluated at the same six locations in both 1998 and 1999 (Table 1). At Mitchell, NE, 76 mm of water was applied by gravity irrigation approximately every 10th day, beginning at the V4 stage of plant development (Ritchie et al., 1993). At Sidney, NE, two sublocations were established each year by different irrigation treatments. In the irrigated treatment, about 25 mm of water was applied by overhead sprinkler at the V4 and V6 stages and again 10 d before flowering. No irrigation occurred in the dryland treatment or at any other location.
Thus, there were four irrigated environments (1 through 4) and eight nonirrigated environments (5 through 12). In the statistical analysis, environments were treated as a random variable, the inference space being common types of maize production environments, including irrigated and dryland production, in the area encompassed by these six locations.
At each environment, the experimental design was a 7 x 7 triple lattice (four additional commercial hybrids were included in the test to increase the number of entries from 45 to 49). The experimental unit was a two-row plot measuring 7.6 m long with 0.76 m between rows at each environment except at Environment 7, at which the plots were only 6.0 m long. All trials were planted with a Wintersteiger Plotking precision space planter at 74000 kernels ha-1. With the exception of plots at Environments 1 and 2, plots were thinned about 30 d after emergence to a uniform stand of 37000 plants ha-1. Plots at Environments 1 and 2 were not thinned. Fertilizer was applied at customary local rates, and weeds and insects were controlled with chemical agents.
Grain yield was determined by mechanically harvesting each plot. Grain weight was adjusted to 155 g kg-1 moisture. An analysis of variance was done within each environment, and then lattice-adjusted means from these analyses were used to compute an across-environments analysis of variance with SAS Proc Mixed (Littell et al., 1996).
Hybrids were ranked by the safety-first index,
(Eskridge et al., 1991), where ISF(i) is the value of the safety-first index of the ith cultivar, 
i is the mean yield of the ith cultivar, d is the minimum acceptable yield of any cultivar, and Vi is a variance measure of yield stability for the ith cultivar.
Hybrids with large values of ISF, which result from high mean yields, low values of minimum acceptable yield, and/or high stabilities (low values of Vi), are desirable. The value a producer assigns to d will depend on several factors, including input costs, income risk attitude, and the anticipated market price of the grain that is produced. Two hybrids grown in the same environment could have different values of d if the cost of the seed differed or if the grain produced from each hybrid was used in a different way. Stability of the ith hybrid was defined as its across-environmental sample variance (Type 1 stability measure); namely,
where Yij = yield of the ith hybrid in the jth environment, and q = the number of environments.
This was the value used for Vi in the safety-first index. In the remainder of this paper, Vi will designate the across-environmental sample variance of the ith hybrid. When the Yij are normally distributed, the variance of Vi is given by
The relationship between i. and Vi was determined with Spearman rank correlation. Values of ISF(i) and Vi were determined for both the set of all 12 environments and for the set of eight dryland environments. The justification for considering the dryland locations as a separate set of environments is that a farmer typically will know before purchasing seed whether the environment will be irrigated or dryland.
|
RESULTS AND DISCUSSION
|
|---|
Environmental means ranged from 2.10 Mg ha-1 at Environment 11 to 12.01 at Environment 1 (Table 2). Even among only the dryland environments, the highest environmental mean yield was more than three times greater than the lowest environmental mean yield. Compared with the 1995 to 1999 average dryland maize yield in the Nebraska Panhandle of 3.50 Mg ha-1, three of the dryland environments had lower yields and five had higher yields. Total water (precipitation + irrigation) ranged from 197 to 720 mm (Table 2). One dryland environment, Environment 12, had more total water than two of the irrigated environments. While this is not a common phenomenon, the sporadic occurrence of thunderstorms producing heavy rainfall in this region occasionally results in dryland environments with abnormally high levels of precipitation. The two environments with the highest average yields, Environments 1 and 2, had the highest levels of total water, and the environment with the lowest yield, Environment 11, had the lowest level of total water. The correlation between total water and environmental mean yield across all environments was 0.73.
The total variation among all hybrids was highly significant at nine of the 12 environments and nonsignificant at Environments 3, 9, and 12 (results from within-environmental analyses are not shown). The single-cross mean yield was greater than the double-cross mean yield at every environment (Table 2), but the difference was not statistically significant at Environments 3, 9, 10, or 12. Across all environments, the average superiority of the single-cross over the double-cross hybrids in mean yield was 11.5%. This overall difference was highly significant (Table 3) and also was much greater than the difference of 2.9% reported by Eberhart and Russell (1969). However, the superiority of the best single cross over the best double cross was about 8% in both studies. One possible cause of the greater difference in average yield observed between single and double crosses in our study was close genetic similarity among some of the single cross hybrids. The parentage of the commercial single crosses is proprietary information, but genetic marker data of commercial hybrids (Smith and Smith, 1991) and a survey of germplasm sources used by U.S. commercial maize breeders (Goodman, 1998) suggest it is possible that the same inbred or closely related inbreds could have been parents of multiple single crosses. In contrast, in the Eberhart-Russell study the parents of the single crosses were 10 unique inbreds, and the double crosses were a balanced set of all possible double crosses that could have been developed from these inbreds.
View this table:
[in this window]
[in a new window]
|
Table 3. Results from analyses of variance of grain yield of nine single-cross (SC) and 36 double-cross (DC) maize hybrids across all 12 environments and across eight dryland environments in 1998 to 1999.
|
|
In both the analyses across all 12 environments and across the eight dryland environments, the hybrid x environment interaction was highly significant (Table 3). In both sets of environments, the single-cross vs. double-cross comparison was a highly significant component of this interaction. Across all environments, the mean value of Vi was 10.41 for the single-cross hybrids and 7.82 for the double crosses (Table 4). Only three of the 36 double crosses had values of Vi that were greater than the mean Vi value of the single crosses. A similar situation occurred among the eight dryland environments. The mean values of Vi were 3.68 and 2.89 for the single and double crosses, respectively (Table 4), and only four of the double crosses had values of Vi that were greater than the mean single-cross value.
View this table:
[in this window]
[in a new window]
|
Table 4. Mean grain yields and across-environmental variances (Vi) of 45 maize hybrids across all 12 environments and across eight dryland environments, 1998 to 1999.
|
|
In comparing the stabilities of the single and double crosses, two issues must be addressed. One question is whether the greater stability of the double crosses occurred because the double crosses had fewer lower yields than the single crosses, fewer higher yields, or both. This question was investigated in two ways with only the data from the eight dryland environments. First, the frequency of extreme yields was determined for the nine single crosses and also for the nine double crosses with the highest mean yields. "Extreme" was defined as being in either the lower 10% (<2.13 Mg ha-1) or the upper 10% (>6.37 Mg ha-1) of the combined distribution of all hybrids. For the single crosses, six low and 15 high yields occurred, whereas for the nine selected double crosses only two low and 10 high yields occurred. Fewer low extreme values were observed for the selected group of double crosses even though the mean of the double crosses (4.67 Mg ha-1) was less than the mean of the single crosses (4.79 Mg ha-1). A fewer number of both low and high extreme yields by double crosses compared with single crosses also was reported by Jones (1958).
Second, the collective distribution of the single crosses and also the distribution of all the double crosses were checked for kurtosis and skewness with the g1 and g2 test criteria (Snedecor and Cochran, 1989). Both distributions were significantly leptokurtotic, but the double-cross distribution was more peaked. This was consistent with the observation of fewer extreme yields for the nine selected double crosses. The single-cross distribution was not skewed, but the distribution of the double crosses exhibited significant negative skewness. This means that the lower tail of the distribution was extended compared with the upper tail and that yields above the mean were more common than yields below the mean. On the basis of a safety-first criterion, this type of skewness seems desirable.
A second key issue is the relationship between mean yield and Vi. Lin et al. (1986) noted that a positive correlation between mean performance and Type 1 stability parameters frequently has been observed and may be a primary reason that breeders rarely have used this type of stability measurement. Across all environments, the highest-yielding hybrid (H2) had the highest value of Vi, and the lowest-yielding hybrid (H5 x H7) had the lowest value of Vi (Table 4). However, the relationship between yield and Vi in these data was not as strong as this observation indicated. The third highest-yielding hybrid (H3) had only the 10th highest value of Vi, and the highest-yielding double cross (eighth highest for yield overall) ranked only 15th for Vi. The Spearman rank correlation between mean yield and Vi was 0.52. Across the eight dryland environments, the value of this correlation was only 0.36. If mean yield and Vi are not strongly correlated, then selection for high-yielding double crosses with high stabilities should be possible.
On the basis of grain yield alone and across all environments, eight of the nine best hybrids (nine is 20% of the total number of hybrids) were single crosses (Table 5a). The value of d that most heavily weights the impact of Vi on the value of the safety-first index is 0. When d = 0, only one single cross was among the nine hybrids with the highest scores of the safety-first index. However, it seems unlikely that many producers would choose a minimum acceptable yield of 0. The mean yield across all environments and hybrids was 5.65 Mg ha-1. When d was set equal to this value, the number of double crosses ranked in the top 20% was two compared with only one when ranking was on the basis of yield alone. However, these two double crosses only ranked as the eighth and ninth best hybrids. The smallest value of d at which a double cross was ranked as the best overall hybrid was 2.68 Mg ha-1. This double cross was H6 x H7, which among all hybrids had the second lowest value of Vi, but a mean yield that was slightly below average.
Across the eight dryland environments, the best six hybrids based on yield alone were single crosses (Table 5). The next three best hybrids were double crosses. This situation did not change when the criterion was the safety-first index with d = 4.46, the mean yield across the eight environments and all hybrids. The smallest value of d at which a double cross was ranked as the best overall hybrid was 1.88 Mg ha-1. This double cross was H2 x H7. As was true for H6 x H7 across all environments, this double cross had below-average yield but had the second-lowest value of Vi. At d = 1.88, seven of the nine best hybrids were double crosses. Among the counties in the Panhandle of Nebraska, northeastern Colorado, and southeastern Wyoming that comprise the western High Plains region, county yield averages <1.88 Mg ha-1 occurred at a frequency of 25% from 1980 through 2002 (National Agricultural Statistics Service, 2002). Therefore, for risk-averse producers it is not unreasonable that a value of d as low as 1.88 may be appropriate.
Two other factors may strengthen the case for using double-cross maize hybrids for dryland production in the western High Plains. First, the highest-yielding double crosses that could be produced from the parental inbreds of the single crosses used in this study likely were not tested. Method B of Jenkins (1934) for predicting double-cross yields from single-cross yields suggests that the best double cross would result by crossing the two single crosses that are produced by crossing the pairs of inbreds that are most closely related. By crossing only commercial single crosses, the opposite was done in the production of the double crosses used in this research. Second, the cost of producing seed of double crosses may be less than the cost of producing single crosses because the single-cross female parents of double crosses would be expected to have a greater seed yield than the inbred female parents of single crosses. If this cost-savings is reflected in the price of the seed, then the value of d would be lower for double crosses than for single crosses.
Even when these factors are considered, however, we believe it is unlikely that commercial seed companies will undertake development, testing, and production of double-cross hybrids for this single geographic region. The size of the maize seed market in the western High Plains is probably not sufficiently large to allow the recovery of the additional costs associated with marketing a second type of hybrid. Farmers could produce their own double-cross hybrids with public inbreds. However, in addition to the cost of testing to identify the best double-cross hybrid, seed stocks of two single crosses and four inbreds would have to be maintained. This could be cost prohibitive. Other types of heterogeneous cultivars that may give a similar advantage as double crosses because of better stability include populations and population crosses. Although we know of no data comparing the performance of double crosses to populations, because of heterosis we believe the mean yields of the best populations would not be competitive with those of the best double crosses. Population crosses may overcome this problem, but mass production of population crosses would be difficult because the female parent is nonuniform.
We suggest a viable option may be a hybrid produced by crossing a single-cross hybrid (female) to an elite population (male). Such a hybrid would have a uniform female that would allow for easy detasseling and would have high seed yield. These characteristics would favor a low seed cost, which in turn would lower the value of d of this type of hybrid compared with single crosses. The parents could be selected not only for high per se performance but also for high heterosis. Public maize breeders in the USA have developed many improved populations, and periodically better versions of these or other improved populations are released. If the female single cross was produced from two publicly developed inbreds, then seed of this type of hybrid could be produced by local farmers. Finally, a single cross x population hybrid would be heterogeneous. The data from this research indicated that this property may give increased stability and higher safety-first index values, particularly in stressful, dryland environments such as those that occur in the U.S. western High Plains.
|
NOTES
|
|---|
Research partially funded by Pioneer Hi-Bred International. Nebraska Agricultural Research Division, Journal Series No. 13885.
1 The use of commercial products in this publication does not imply endorsement of the products named nor criticism of similar ones not mentioned. 
Received for publication October 31, 2002.
|
REFERENCES
|
|---|
- Allard, R.W., and A.D. Bradshaw. 1964. Implications of genotype-environmental interactions in applied plant breeding. Crop Sci. 4:503–508.[Free Full Text]
- Colorado Agricultural Statistics Service. 2002. Colorado agricultural statistics 2002. [Online]. [157 p.] Available at http://www.nass.usda.gov/co/pub/buletn02/2002-AllPages2.pdf (cited 25 Oct. 2002; verified 12 May 2003). CASS, Lakewood, CO.
- Eberhart, S.A., and W.A. Russell. 1966. Stability parameters for comparing varieties. Crop Sci. 6:36–40.[Abstract/Free Full Text]
- Eberhart, S.A., and W.A. Russell. 1969. Yield and stability for a 10-line diallel of single-cross and double-cross maize hybrids. Crop Sci. 9:357–361.[Abstract/Free Full Text]
- Eskridge, K.M., P.F. Byrne, and J. Crossa. 1991. Selection of stable varieties by minimizing the probability of disaster. Field Crops Res. 27:169–181.
- Finlay, K.W., and G.W. Wilkinson. 1963. The analysis of adaptation in a plant breeding programme. Aust. J. Agric. Res. 14:742–754.[ISI]
- Goodman, M. 1998. Research policies thwart potential payoff of exotic germplasm. Diversity 14:30–35.
- Hallauer, A.R., and J.B. Miranda, Fo. 1988. Quantitative genetics in maize breeding. 2nd ed. Iowa State Univ. Press, Ames.
- Hoekstra, G.J., L.W. Kannenberg, and B.R. Christie. 1985. Grain yield comparison of pure stands and equal proportion mixtures for seven hybrids of maize. Can. J. Plant Sci. 65:471–479.
- Hühn, M. 1996. Nonparametric analysis of genotype x environments by ranks. p. 235– 271. In M.S. Kang and H.G. Gauch, Jr. (ed.) Genotype-by-environment interaction. CRC Press, Boca Raton, FL.
- Jedel, P.E., J.H. Helm, and P.A. Burnett. 1988. Yield, quality and stress tolerance of barley mixtures in central Alberta. Can. J. Plant Sci. 78:429–436.
- Jenkins, M.T. 1934. Methods of estimating the performance of double crosses in corn. J. Am. Soc. Agron. 26:199–204.
- Jones, D.F. 1958. Heterosis and homeostasis in evolution and in applied genetics. Am. Nat. 92:321–328.
- Kang, M.S., and H.N. Pham. 1991. Simultaneous selection for high yielding and stable crop genotypes. Agron. J. 83:161–165.[Abstract/Free Full Text]
- Lerner, I.M. 1954. Genetic homeostasis. Oliver and Boyd, London.
- Lin, C.S., M.R. Binns, and L.P. Lefkovitch. 1986. Stability analysis: Where do we stand? Crop Sci. 36:894–900.
- Littell, R.C., G.A. Milliken, W.W. Stroup, and R.D. Wolfinger. 1996. SAS system for mixed models. SAS Inst., Cary, NC.
- Lynch, M., and B. Walsh. 1998. Genetics and analysis of quantitative traits. Sinauer Assoc., Sunderland, MA.
- National Agricultural Statistics Service. 2002. USDA-NASS published estimates data base [Online]. [1 p.] Available at http://www.nass.usda.gov:81/ipedb/ (cited 8 Feb. 2003; verified 12 May 2003). USDA-NASS, Washington, DC.
- Nebraska Agricultural Statistics Service. 2002. Nebraska state statistical office publications [Online]. [1 p.] Available at http://www.nass.usda.gov/ne/nebrpubs.htm (cited 8 Feb. 2003; verified 12 May 2003). USDA-NASS, Lincoln, NE.
- Owenby, J.R., and D.S. Ezell. 1992. Monthly station normals of temperature, precipitation, and heating and cooling degree days, 1961–90—Nebraska. Rep. 81. U.S. Dep. of Commerce, Nat. Climatic Data Center, Asheville, NC.
- Perkins, J.M., and J.L. Jinks. 1968. Environmental and genotype-environmental components of variability. III. Multiple lines and crosses. Heredity 23:339–356.[ISI][Medline]
- Ritchie, S.W., J.J. Hanway, and G.O. Benson. 1993. How a corn plant develops. Spec. Rep. 48. Iowa State Univ. Coop. Ext. Serv., Ames.
- Roy, A.D. 1952. Safety-first and the holding of assets. Econometrica 20:431–449.
- Smith, J.S.C., and O.S. Smith. 1991. Restriction fragment length polymorphisms can differentiate among U.S. maize hybrids. Crop Sci. 31:893–899.[Abstract/Free Full Text]
- Snedecor, G.W., and W.G. Cochran. 1989. Statistical methods. 8th ed. Iowa State Univ. Press, Ames.
- Sprague, G.F., and W.T. Federer. 1951. A comparison of variance components in corn yield trials: II. Error, year x variety, location x variety, and variety components. Agron. J. 43:535–541.[Free Full Text]
- Weatherspoon, J.H. 1970. Comparative yields of single, three-way, and double crosses of maize. Crop Sci. 10:157–159.[Abstract/Free Full Text]
- Wyoming Agricultural Statistics Service. 2002. Wyoming Agricultural Statistics 2002. [Online]. [96 p.] Available at http://www.nass.usda.gov/wy/bulletin/bulletin2002.pdf (cited 25 Oct. 2002; verified 16 May 2003). WASS, Cheyenne, WY.
This article has been cited by other articles:
|
|
|
|
 
F. R. Guillen-Portal, W. K. Russell, K. M. Eskridge, D. D. Baltensperger, L. A. Nelson, N. E. D'Croz-Mason, and B. E. Johnson
Selection Environments for Maize in the U.S. Western High Plains
Crop Sci.,
September 1, 2004;
44(5):
1519 - 1526.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
