Premium
Cultivar‐Environment Interaction Study of Alfalfa Yields in Oklahoma 1
Author(s) -
Taliaferro C. M.,
Denman C. E.,
Morrison R. D.,
Holbert D.
Publication year - 1973
Publication title -
crop science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.76
H-Index - 147
eISSN - 1435-0653
pISSN - 0011-183X
DOI - 10.2135/cropsci1973.0011183x001300060010x
Subject(s) - cultivar , yield (engineering) , forage , biology , regression analysis , regression , interaction , perennial plant , main effect , explained sum of squares , gene–environment interaction , variance (accounting) , agronomy , statistics , linear regression , mathematics , biochemistry , materials science , business , accounting , gene , genotype , metallurgy
Forage yield data from 14 alfalfa ( Medicago sativa L.) cultivars harvested for 3 years (1966 through 1968) at each of 5 Oklahoma locations were used to study CE (cultivar × environment) interactions. Mean squares, estimated variance components, and regression analyses statistics were used to draw conclusions. Statistically nonsignificant mean squares and small variance components for the first order CL (cultivar × location) and CY (cultivar × year) interaction components suggested minimal effects of locations and years per se on relative yields of alfalfa cultivars. The second order CLY (cultivar × location × year) interaction effect was of large magnitude and highly significant, indicating that CE interactions in alfalfa yields were attributable to factors other than years and locations. These results emphasized the necessity for multiple environment testing but gave little information which would elucidate how these tests should be distributed over years and locations. Cultivar regression coefficients, though in most cases not statistically different from unity, suggested some possible differential cultivar reaction to varying environments. Cultivars with regression coefficients above one had either a better than average response to high yield environments or a smaller than average response to low yield environments. Mean square deviations from regression (s 2 d ) revealed specific instances of yield instability which presumably resulted from CE interaction. Due to the perennial nature of alfalfa and the need to evaluate stand persistence in new lines, it would seem most feasible to extend tests over years using a minimum of test locations. Locations, however, should probably be divided to include both high‐ and low‐yielding sites. In Oklahoma, alfalfa strains and cultivars can very likely be adequately evaluated for forage yield ability by testing for 3 years (4 years including the year of seeding) at three to four selected locations.