Clay and pastry caterpillars in science education

Pastry caterpillars: a tool to teach evolution by natural selection

Last week Jay Fitzsimmons over at the Canadian Field-Naturalist Blog sent me an article from the new issue of The American Biology Teacher that describes how teachers can use clay caterpillars to teach ecology and evolution to their students. I whole-heartedly embrace the dissemination of techniques that can help teachers educate students about ecology and evolution, so I wanted to highlight this and a few other logistical tips that would help teachers who might consider running a fake caterpillar lab exercise with their class.

Here is the citation & link to the paper:

Barber, N. (2012) Clay caterpillars: a tool for ecology & evolution laboratories. The American Biology Teacher 74(7): 513-517.

Jay sent this paper along because he knows about my work using pastry caterpillars to study the protective value of eyespots on caterpillar-like prey (Hossie and Sherratt 2012). The pastry caterpillar system has been used countless times to study a range of topics, most notably in the study of apostatic selection (Allen and Anderson 1984), polymorphisms (Wennersten and Forsman 2009), and countershading (Rowland et al. 2008). Clearly this experimental approach can be modified to examine a wide variety of questions relating to natural selection.

How it works:

  • Fake caterpillars are made using pastry dough (flour + lard) dyed with food colouring, or modelling clay. The colour, shape, and size of your caterpillars and the proportion of each type that you deploy depends on the question you want to examine.
  • Caterpillars should be roughly 5-30 mm long and 3-6 mm wide
  • Caterpillars are deployed in the field either on short-cut lawns, painted wooden boards, or pinned to tree branches where they will be susceptible to "predation" by wild birds (e.g., sparrows, chickadees, starlings etc.)
  • After a pre-defined amount of time (e.g., 24 h), return and record which prey were attacked or removed


  • Often it is good practice to train and/or attract birds ahead of time in the site where you will be conducting the experiment using birdseed or small pieces of plain pastry dough.
  • Birds are typically most active in the early morning. Consider deploying very early morning or the evening before.
  • Consider using areas where predation by animals other than birds might not be as big an issue.
  • Some "background" predation is likely to occur (i.e., from ants, squirrels, chipmunks, mice etc.), but this is usually not a big problem. You can choose to count them as killed, survived, or eliminate them from the proportion survived calculation (i.e., 50 deployed, 10 pecked, 2 eaten by ants -> 38/48 = 79% survived).
  • Presenting prey on elevated platforms or pinning them to tree branches can minimize the effects of ants and small mammals.
  • Small mammals will be a bigger issue if you leave prey deployed on the ground overnight.

Barber presents this experimental approach as a framework through which teachers could examine ecological and evolutionary questions with their students. Ecological questions could relate to the abundance and distribution of animals in various habitats, whereas evolutionary questions might examine how natural selection can shape appearances or behaviours in a population. As Barber sees it, the main advantage of this kind of lab is its flexibility in allowing students to develop and test hypotheses according to their own interests. Barber states:
The possible hypotheses for students to explore are almost limitless and, with no “wrong” answers, interpreting results can be a valuable critical-thinking exercise.
Actually this is not a new idea. In 1993 J. A. Allen and his colleagues presented a similar technique for using pastry caterpillars to teach similar concepts in the Journal of Biological Education:

Allen, J. A., Cooper, J. M., Hall, G. J. and McHenry C. (1993) 'Evolving pastry': a method for simulating microevolution. Journal of Biological Education 27(4): 274-282.

This paper provides project ideas specifically designed to teach evolution by natural selection. Real evolution takes a very long time to observe, but it can be simulated in the short-term (i.e., over a couple of days) through the successive deployment of various fake caterpillars in the field. The key is to use the proportion of caterpillars surviving after one day of "predation" to set up the proportion of caterpillars from each treatment in the following deployment, then repeat. For example, imagine a study where you were interested in the evolution of camouflage colour in prey. You might deploy 100 prey on tree branches, 50 brown ones and 50 green ones. After 24 h you check the attack rate, and you see 20 brown ones killed and but only 5 green ones killed. That means that 30 brown (30/50 = 60% of brown prey), and 45 green (45/50 = 90% of green prey) survived. Prey need to survive to reproduce so the proportion of prey in the next generation depends on how many of each treatment survived. Thus, on the next day you deploy 100 more prey, but this time the ratios are different: 60 green prey are deployed ( = [45 / (30+45)] = 0.6*100) and only 40 brown prey ( = [30 / (30+45)] = 0.4*100). Each successive deployment represents a generation. If this process is repeated for 5-10 generations you should be able to see a change in the proportion of each phenotype in your population over time. The more generations you run the better data you will get.

Of course, this design can also be modified to ask an almost limitless number of questions and gives students a first hand snapshot of how evolutionary process work. Allen et al. (1993) includes a general procedure and the basic methods for field experiments using fake caterpillars, similar to Barber (2012), but Allen et al. go an extra step and provide modified methods and example results for 6 different projects that teach various principals in evolution by natural selection, such as:
  1. Directional selection
  2. Spread of new mutations
  3. Crypsis in uniform backgrounds
  4. Crypsis in varied backgrounds
  5. Disruptive selection

Further thoughts about fake caterpillar experiments:

I have had a fair amount of experience making similar caterpillars. Between multiple experiments during my PhD I have now made over 1500 model caterpillars. Most of these models were used to study how well eyespots can protect caterpillars from wild bird predators. Either pastry or modelling clay can be used for these experiments with the main difference being that the edible pastry provides a food reward comparable to that of a similar sized caterpillar. Clay caterpillars are effectively inedible and once birds realize this it could reduce their incentive to search for them. Another advantage of pastry is that it is easier to obtain the desired colour for your fake caterpillars. Simply adding different amounts of food colouring to dye the pastry dough is easier than mixing different colours of clay together. That being said, when teaching a class of younger students it might be easier and less messy to just use modelling clay.

My recipe for pastry caterpillars: 

  • 1/3 cup of lard (I used Tenderflake)
  • 1 cup of flour (You will get better colour from food colouring if you use White flour)
  • 15 ml water
  • Food colouring

  • Cut the lard into the flour. 
  • Add desired amount of food colouring and the water
  • Hand mix and kneed the dough until the food colouring dough is a uniform colour and texture
  • The most common mistake is adding too much water. The dough will get very sticky instead of being easy to mould and work with. To remedy this you can work the dough and let it sit, but it make take a while for it to lose moisture. Adding more flour & lard mix might be quicker option.
  • Don't let the dough sit out too long or it will dry out. Freeze unused dough and prepared caterpillars in a ziplock bag until you are ready to deploy

In my paper on caterpillar eyespots I used a Play-Doh Mini Fun Factory (Hasbro Canada Corp., Longueuil, QC, Canada) to press out long strips of dough that I used to standardize the caterpillar size. A similar technique might be worth considering for teachers planning to make numerous fake caterpillars.

Deploying caterpillars:

A pastry caterpillar pinned to a tree branch. This one was attacked at both ends by a bird.

Older students could use dress making pins to
attach pastry caterpillars to  tree branches

My pastry caterpillars were pinned to tree branches using dress making pins, and this technique worked really well for securing the caterpillars to tree branches. Once I found a suitable branch I pressed the pin partway through it (make sure to buy a thimble!), then stuck the pastry caterpillar onto the pin. For better stability on the branch it helps if you mould the caterpillar to the branch a little bit. If pins seem a little too dangerous for your students Barber suggests using silicone which can be purchased at a hardware store. This would allow you to attach small caterpillars to tree leaves as well as branches. I do wonder about the possible toxicity of silicone to birds that eat or peck the baits though. With pins you don't have to worry about the possible toxic effects of adhesives, and they invariably remain in the tree when a bird removes or feeds on the fake caterpillar. You will need to remove pins after you are done, I suggest using needle nose pliers.

Aside: This brings up an interesting point. All research conducted by members of a University of College which involves any vertebrate animal (plus Cephalopoda) requires approval from an Animal Care Committee. Even my research on caterpillar eyespots where I was basically feeding pastry to wild birds needed such approval. I'm not sure whether elementary or secondary schools have some similar committees, but if they are involving vertebrate animals they probably should. Does anyone know what rules or restrictions exist that might affect elementary of secondary institutions looking to do this kind of laboratory exercise?

If none of these methods of attaching caterpillars to trees sound like a good option you can just place the pastry caterpillars directly on a short-cut lawn, pieces of astroturf, or painted wooden platforms:

Northern Cardinal feeing on plain pastry and bird seed on a raised platform during the training phase of an experiment 


The final step would be to do some kind of analysis. For elementary or high school it is probably sufficient to just have them build graphs. In the simplest version of this lab students could just compare how many prey from each treatment were killed after a single deployment (i.e., a single "generation"). This could involve a simple bar graph comparing the survival of each treatment.

If you are planning to teach evolution though it would be better to look at the change in proportion of each treatment across successive generations. Get the students to plot the proportion deployed at the start of each treatment, this will show them graphically how the frequency of each phenotype changed in their population over time. This is directly analogous to the change in frequency of the genes that carry those phenotypes in population (i.e., evolution!). You could also get them to plot the proportion of prey from each treatment that were attacked.

Here are two examples of plots from Allen et al 1993:

Additional thoughts for advanced students:

  • What are the limitations of using survival as a measure of fitness
  • What is apostatic selection?
  • What is a "search image"? How might the formation of a "search image" affect rates of predation?
  • If one treatment survives better on average, why might multiple phenotypes persist in a population?


Allen, J. A. and Anderson, K. P. (1984) Selection by passerine birds is anti-apostatic at high prey density. Biological Journal of the Linnean Society 23: 237–246.

Allen, J. A., Cooper, J. M., Hall, G. J. and McHenry C. (1993) 'Evolving pastry': a method for simulating microevolution. Journal of Biological Education 27(4): 274-282.

Barber, N. (2012) Clay caterpillars: a tool for ecology & evolution laboratories. The American Biology Teacher 74(7): 513-517.

Hossie, T. J., Sherratt, T. N., (2012) Eyespots interact with body colour to protect caterpillar-like prey from avian predators. Animal Behaviour 84(1): 167-173, doi:10.1016/j.anbehav.2012.04.027

Rowland, H. M., Cuthill, I. C., Harvey, I. F., Speed, M. P., & Ruxton, G. D. (2008). Can't tell the caterpillars from the trees: countershading enhances survival in a woodland. Proceedings of the Royal Society of London - B, 275(1651), 2539–2545

Wennersten, L. and Forsman, A. (2009) Does colour polymorphism enhance survival of prey populations? Proceedings of the Royal Society of London - B 276: 2187-2194


Post a Comment