Shelter-building caterpillars

This post is a bit late in coming, but I wanted to share what I worked on for several summers with Doug Morse for my undergrad thesis. We did most of the work in 2008 and 2009, but finished up some during the summer of 2012 and it was published earlier this year.

Caterpillars, like many other animals, construct shelters for themselves, by some estimates over 60% of lepidoptera (butterfly and moth) species make some sort of shelter. Most often these are made by folding or rolling the leaves which they feed on into shelters. The shelters likely serve a variety of purposes: protection from dessication, protection from predators, reduced probability of dislodgement, etc. Interestingly, several studies had found that caterpillars in shelters suffered less predation but more parasitism and indeed, at least one author had speculated that since parasitism was often so high (90%+ sometimes!), that shelters must not be good protection.

an unidentified caterpillar folding up a Malvaceae leaf, Chiloe Island, Chile

That seemed odd to me; if parasitism was so prevalent, it must be a strong selective force and clearly some survive, so there must be some way to beat it. Working with a cool leaf-rolling moth, Herpetogramma thesausalis, on ferns in Maine, I ran experiments to test two hypotheses:

1) Shelters provide protection from parasitoids
2) Shelters are costly for the caterpillars to produce

an adult male H. thesausalis on sensitive fern, its primary host

To test the first hypothesis, I thinned shelters during the moth's pupal stage by carefully removing the outer layer of fern (the shelters usually involve 2-4 layers, so this did not expose the pupae inside). I used the pupal stage, as caterpillars would quickly rebuild the shelter when I thinned it - pupae are unable to. I found that parasitism almost doubled without that outer layer, though overall size of the shelter was only slightly smaller. Therefore, I found support for the first hypothesis - the first experimental evidence for this function of caterpillar shelters!

A cross-section of a shelter containing a parasitoid (Alabagrus texanus) cocoon.

For the second hypothesis, I divided caterpillars on ferns into three groups: the first I left alone (as controls), the second I dismantled their shelter every few days and the thirds I touched but did not destroy the shelters at the same interval. In essence, I was forcing one group to remake their shelters. Given my hypothesis, I expected that these caterpillars would either pupate smaller or later. Interestingly, the caterpillars delayed their pupation in order to rebuild shelters - an average of ~2 days later than the other groups - and they did not eclose smaller (or build smaller shelters!). Delaying their pupation puts them at increased risk for larval parasitism and delays their reproduction (the population is fairly synchronous, so this could be a big cost for some individuals). Hypothesis 2 supported!

A female ichneumonid (Itoplectis?) oviposits into a caterpillar shelter.

There is more to the story, it is a longer paper. But perhaps the most interesting thing is that there are two broad guilds of pupal parasitoids - those with long ovipositors (as seen in the above picture), which oviposit from outside the shelter, and those with short ovipositors which gnaw their way through the seams between leaves into the shelters. I suspect that larger shelters, with a longer distance between the outside and the pupa, are better protection against the former guild. However, in a larger shelter, there are more seams, which may make it easier for the latter guild to parasitize, an interesting trade-off, which I have only observational data to support.

Phaeogenes hebrus, the most common short-ovipositor species in this system.

The paper can be found here: http://link.springer.com/article/10.1007/s11829-013-9261-4#page-1 . Arthropod-plant interactions publishes lots of great - if limited interest - research, I read a very high number of the papers it publishes.

This would make a cool tattoo, eh?