Experimental and Observational Studies of Patch Choice at Different Scales by the Crab Spider Misumena Vatia

We studied patch choice of adult female crab spiders (Misumena vatia) on common milkweed (Asclepias syriaca) at three different scales: among umbels (inflorescences), among stems, and among clones. Spiders occupied the umbel on a stem with the most insect prey (and most nectar—producing flowers) (=high quality) more often than predicted by chance and captured most of their prey there. However, even if sample patch choices (umbels) of greatly varying quality were provided, 30% of the spiders did not select high—quality umbels within a day, although choices required movements of only several centimetres. The distribution of spiders on these umbels most closely resembled that predicted if spiders responded to the number of times insect prey visited an umbel. Many spiders move to higher quality stems, upon which most prey were caught, as the last flowers on their own stems senesced. Significantly more spiders remained on high—quality stems than did those on stems with only senesced umbels; however, 25% of the spiders placed on high—quality stems nevertheless moved. Many spiders that moved did not immediately occupy stems of higher quality than the ones they had just left, but by continuing to move, their probability of eventually occupying a high—quality stem markedly increased. Thirty—five percent of the spiders on stems with only senesced umbels did not move within a day. Clones of milkweed were several metres apart, and there was no evidence that spiders moved between them during their last instar. We argue that stochastic factors (unpredictable visits to flowers by prey) may account for the nonuniform response in the simple umbel choice experiments. They may also account for part, but not all, of the variance in the stem choice experiments.