Effects of eyespots and body colour on caterpillar survival

What I did and how I did it: an experiment with pastry caterpillars

Me (Tom Hossie) examining a late 5th instar Papilio canadensis caterpillar that I reared last summer (2011)
Today the BBC Nature website posed a short piece on my research - see the write-up here. This write-up talks about the work I did in the summer of 2010 for the first chapter of my PhD thesis. This work was recently recently published in a journal called Animal Behaviour. You can see a copy of the article here via ScienceDirect. Note also that the photos from the BBC Nature piece are ones that I took during my research over the last two years.

A 5th instar Papilio canadensis caterpillar reared last summer (2011)
In this project I wanted to examine the relative protection aforded to caterpillars by eyespots, countershading, and any synergistic protective effects of these traits. This design was partly motivated by the question of whether eyespots could protect prey better than simply increasing their crypsis (e.g., via countershading), but also after seeing several examples of caterpillars with eyespots I had noticed that many of them were in fact countershaded as well. 

Aside: I should say that countershading is an animal colour pattern where the pigments on the dorsal surfaces (i.e., the animal's back) are darker than the ventral pigments. Most, but not all, of the hypothesized functions of countershading suggest a camouflage function. For example, the self-shadow concealment hypothesis proposes that having dark pigments where the body is naturally illuminated by the sun and lighter pigments in those areas that are naturally shaded may increase background matching, or weaken 3D cues used by predators to detect their prey.

My experiment had 4 treatments: solid-no eyespots, solid-eyespots, countershaded-no eyespots, and countershaded-eyespots. So, I made artificial caterpillars using flour and lard, dyed using with food-colouring (to add eyespots I used non-toxic paint). Once made, the "caterpillars" were pinned to tree branches (yes, caterpillars spend a non-trivial amount of time on tree branches) and monitored continuously over 90h. Specifically we were looking for peck marks left by birds or "caterpillars" that went missing altogether. Many other researchers have used a similar design to examine other questions, I just modified it to test the effect of eyespots on caterpillar-like prey.

The 4 prey types I used in this project
Field experiments are inherently ‘noisy’ meaning that because of a number of factors out of your control it can sometimes be difficult to detect significant patterns. Other researchers had shown previously that bigger eyespots resulted in greater protection of moth-like targets, and as this was the first time anyone had examined the protective effect of eyespots on caterpillars in this way we used relatively large eyespots to increase our chance of detecting an effect. Eyespot size and the specific eyespot pattern employed both could have influenced our results, but future work is needed to address these questions. Perhaps a more interesting point is the possible effect of morphological changes made by the caterpillar to change when it feels threatened. Many caterpillars with eyespots inflate their anterior body segments to change the shape of the “head” possibly to look more imposing or to increase their resemblance of a snake. We are investigating this question currently.

Update: We have now conducted this research examining the role of the "head" shape during the rest and defensive postures. We found that both inflated anterior body segments and eyespots can deter bird attacks. This work was published in the journal Animal Behaviour. See the full article here.

One of my field sites for this work. Kemptville, ON.
Sometimes my caterpillars were attacked by ants...
...and other times they were attacked by slugs (usually after a light rain)
A number of the pastry caterpillars were attacked by unwanted enemies of pastry. Ants were a big one, slugs were another. Both leave distinctive damage markings. Because we used a statistical technique called Cox-proportional hazards (a type of survival analysis) we were able to stop these unwated attacks from having an effect on our results by “censoring” them at the time when the unwanted attack occurred. Slugs were especially problematic after light rain. Along with the distinctive damage they leave on the pastry caterpillars you can tell the slugs were feeding on them because you can see the green food-coloured pastry in their stomachs! 

In this experiment, eyespots alone were not sufficient to protect caterpillars. In real caterpillars however other factors are also at play which likely increase the protective effect of eyespots. For example, in our pastry caterpillar experiment we weren't able to include the effect of behaviours expressed by caterpillars upon attack. We think that in caterpillars the protective effect of eyespots is enhanced by behaviours expressed upon attack that either startle off the attacker or resemble the movements of a threatening object like a snake. Also, eyespots may afford greater protection from insect-eating birds in the tropical ecosystems where a bird’s innate and/or learned wariness of eye-like features is heightened because of the greater diversity and abundance of dangerous things like snakes.

Interestingly, the countershaded-eyespots treatment - which were most like our local eyespot caterpillars (Papilio canadensis) - survived the best. This was an important finding because almost no empirical work had been done to examine the protective effect of caterpillar eyespots. Other work has shown that eyespots can increase survival of butterflies and moth-like targets, yet although eyespots are perhaps even more common in caterpillars only a handful of studies have ever attempted to examine the strength of their protective effect in this life stage. The most interesting result of this work is the synergistic effect of eyespots and body-colour. In isolation neither countershading nor eyespots significantly reduced predation on our pastry caterpillars, but together predation dropped significantly. To me, this suggests that seemingly independent defences are working together in ways that we don’t fully appreciate. Now that we have shown that eyespots can protect caterpillars we are looking to understand why some species evolve eyespots and not others.

Here is the full citation for this research:


  1. its seems to be a original one..Because both leaves and that insect are same in color ....Nice art done by you.keep to do these type of things..

  2. Hi Body paint,

    Thanks for the encouragement! Yes, the caterpillars are very near in colour to the leaves they feed on. When I was rearing live Papilio canadensis caterpillars they were often very difficult to find even though I knew where to look! I tried to match that colour in my experimental pastry caterpillars, the two-toned "countershaded" ones were probably most similar to the real caterpillars.

    You might also like the pictures of caterpillars that were painted for me by Mandy Stobo: