Skelhorn and company tested this by taking a couple of different species of caterpillars that seem to mimic twigs (Opisthograptis luteolata is pictured). They exposed chicks (caterpillar predators) to hawthorn branches, which caterpillars live on and might be mimicking. Although they don’t say it, hawthorn is apparently very spiny, and I’m assuming that the chicks don’t like it very much. Some chicks got a “raw” hawthorn branch, some got a hawthorn branch wrapped in purple thread, and some got no branch.
As expected, when they gave the chicks caterpillars right out in the open, with nothing to conceal them, those that had previous experience with the unmodified hawthorn were slower to peck at the caterpillars and handled them for a longer duration.
Chicks with experience with the hawthorn wrapped in purple went for the caterpillars, presumably because the caterpillars didn’t look like the nasty branches. It all suggests that even if you’re readily visible, looking like something unpleasant might be able to give you a survival advantage.
A fun manipulation would have been to wrap the caterpillars in purple string, like the modified branches. You would predict that you should be able to get the effect to change so that the chicks that had seen the branches in purple string would avoid the coloured caterpillars, and go straight for the “raw” caterpillars.
There is one important qualifier buried in the supplementary material for this paper, though:
Caterpillars were refrigerated before use to reduce movement.
The authors are minimizing a very important cue that the chicks might use to distinguish between ouchy branch and tasty meal: movement. This makes it more difficult to extrapolate their results from the lab to what might be occurring in the field.
Even with that price they’ve paid, I’m sure they’ll come and join the masquerade...
Skelhorn, J., Rowland, H., Speed, M., & Ruxton, G. (2009). Masquerade: Camouflage Without Crypsis Science, 327 (5961), 51-51 DOI: 10.1126/science.1181931
Photo from here.