Mark-recapture models have been widely used in ecology to estimate pop
ulation sizes of animals. In contrast, estimation of plant population
size has usually been assumed to be much easier. However, detection of
individuals is difficult for perennial plants, such as the rare prair
ie plant Mead's milkweed (Asclepias meadii), which does not produce ab
oveground parts every year and lives in dense vegetation where nonflow
ering stems are hard to observe. In these cases, a count of the number
of plants observed in a particular year may greatly underestimate the
true population size, just as a count of animals in traps does not ad
equately estimate the total number of animals in an area. Using a fami
ly of closed population models (CAPTURE), we applied mark-recapture me
thodology to estimate population size of A. meadii. Over a 4-yr period
, a total of 129 patches (aggregated collections of stems) was observe
d, with 124 flowering in at least one year. In any one year, however,
the number of flowering patches ranged from 15 to 105. Using model M-t
h of CAPTURE with these data, the estimated number of patches capable
of flowering was 219. Although the confidence interval is broad (95% c
onfidence interval of 175-302), these results emphasize that the obser
ved number of patches in any one year, or even over a 4-yr period, und
erestimates the actual population size.