| Lefkovich matrices are a common tool for modeling
demography of species with complex life-cycles. However, matrices
with few cells greatly impair the ability to detect within-stage
variation. Large matrices reduce this problem, but require many
parameters that may be difficult to estimate with a finite data set.
Integral projection models provide a method to estimate transition
probabilities of large matrices using far fewer parameters. This
results in an extremely flexible matrix model that can be
parameterized accurately, even with a relatively small amount of
data. The biological implications of this method will be discussed
using models of two species of rare plants. |