were treatment, period, hour, and hour × treatment,
and steer was included as a random term. The repeated
term was hour with steer × period serving as the subject,
and compound symmetry was used for the covariance
structure. For comparisons of methods related to
estimation of bacterial N flow to the duodenum, the
model included period, treatment, method, and treatment
× method and steer and steer × period × treatment
were included as random terms. The LSMEANS
option was used to calculate treatment means. Significance
among treatments was declared at P ≤ 0.05 and
tendencies were declared at 0.05 < P ≤ 0.15. Means
were separated using pair-wise t-tests with P = 0.05
when F-tests were significant, and with pair-wise t-tests
with P = 0.15 when F-tests tended to be significantStatistical Analysis