Kernel size, as it is related to kernel weight, is an important yield
component in oat (Avena sativa L.). The main objective of our research
was to evaluate the changes in kernel morphology following five cycle
s of recurrent selection for grain yield. Parents of Cycles 0, 1, 3, a
nd 5 and seven check cultivars were grown in row and hill plots at St.
Paul and Rosemount, MN, in 1989. Kernel area, perimeter, length, and
width were measured on 100-kernel samples from each plot using digital
image analysis (DIA). Grain yield, kernel weight, and kernel number w
ere also evaluated in these experiments. Grain yield was enhanced by 2
6.8 and 55.3% following five cycles of recurrent selection as measured
in row and hill plots, respectively. Kernel area increased by 9.8 and
10.1% in row and hill plot samples, respectively. Kernel perimeter, l
ength, and width increased by 4.8, 5.7, and 4.6%, respectively, in row
plot samples. In hill plot samples, kernel perimeter, length, and wid
th increased by 5.5, 6.8, and 3.7%, respectively. There were no signif
icant difference among cycles for two derived shape factors, FCIRCLE (
4 pi [kernel area]/[kernel perimeter]2) and kernel width/length ratio.
Values for kernel area, perimeter, length, and width from hill plot s
amples were significantly larger than those from row plots; however, s
election responses in the two plot types were quite similar. In this s
tudy, environmental influences on kernel morphology were minor. Kernel
weight was positively correlated with the two size measurements, kern
el area and width. Results from our study indirectly suggest that incr
eased kernel size, and not kernel density, largely accounts for the ob
served increases in kernel weight. These experiments also demonstrate
the utility of DIA in agronomic research.