Once a Manduca sexta decides that it has grown enough, it starts the process of pupation, wherein it metamorphoses into a moth. The first thing to go is the desire to eat: the caterpillar stops feeding, and eventually evacuates its gut (which creates a grand mess when rearing them on artificial diet in small cups).
After the caterpillar stops feeding it begins to look for a pupation site. Manduca sexta pupates in the soil, not on its host plant, so the caterpillar descends from the plant and walks around until it finds a suitable area of soil to dig in. During this stage it is called a wanderer, and it's easily identifiable because its cuticle becomes shinier/slicker, as well as somewhat transparent, exposing some of its internal organs. Manduca typically wanders at night, most likely to avoid predation.
On Saturday morning I opened up my cages and saw two wanderers, and by midnight another seven had started wandering. Since I had nothing better to do at 2am on a Sunday morning, I decided to take one of the wanderers outside and drop it in my garden patch to see what it would do. I'd never observed a wanderer wandering in the "wild" before, so I was unsure how long the process would take or what I would see. After promising that I would not let the caterpillar escape, I headed outside with it, as well as my camera and a flashlight.
Here's the wanderer just after I put it down on the soil:
M. sexta fifth instar wanderer at night.
The black shapes you can see on the top of the caterpillar (in its middle segments) are its hearts. Insects have an open circulatory system, and caterpillars have a number of hearts arranged in series traveling much of the length of their body. In fact, the easiest method of determining if a caterpillar is a wanderer is whether you can see its hearts: if you can, then it almost certainly is one. You can see the hearts beating when you hold a wanderer in your hand; as you look at the pictures below note the changes in the hearts' shape.
I plopped the wanderer down around 2:14 am, and after walking about a meter, it found the place it would eventually dig in (at 2:19 am). Here's the spot it chose:
Wandering fifth instar M. sexta.
Just after the picture above was taken, the caterpillar started digging. This was garden soil that hadn't been turned in a while, so the soil was compacted. The caterpillar dug with its head and front legs (which made it impossible for me to see how it was actually digging), though it also seemed to be expanding and contracting its body every now and then to help move the soil.
I won't post all the pictures I took of the process in this post, but here are a few to illustrate what happened:
In about 45 minutes the caterpillar had found a spot to pupate and pretty much finished digging itself in. At this point I decided that I should stop the cute little guy, mainly because I felt bad forcing the caterpillar to waste all its stored-up energy digging a hole I was just going to yank it out of. I carefully uncovered the caterpillar to see how it was arranged under the soil, and was surprised to find that it had twisted more than 270 degrees on its way underground.
An excavated M. sexta after it had dug most of its body into soil (uncovered immediately after the 3:01 am picture).
I put the caterpillar into a pupation cage so that it could go bury itself under some cellulose bedding and pupate in a controlled environment (where I can be sure that it won't escape as a moth). I expect all of the caterpillars will start wandering within the week, meaning that the end of the great tomato leaf harvest of 2004 is finally in sight.