Readers’ wildlife photos (with more science)

We have another great installment in Professor Bruce Lyon‘s series of researches on the biology and behavioral evolution of American coots (Fulica americana). He added this in his email: “This one is a bit different since it focuses on one very specific study and quite a bit on a former student of mine but I think your readers might enjoy a behind-the-scenes view of how science is often done. Lay people often have a Hollywood view of how science is done (antisocial nerds in lab coats, everything turning out just as the complex math on the blackboard” predicts).

Bruce’s notes are indented.

Another installment about our work with American coots (previous posts here, here, here, here). Today I focus on an amazing discovery made by my former PhD student Dai Shizuka (now a professor in Nebraska). In my PhD thesis work, I concluded that American coots do not recognize brood parasitic chicks. With a combination of key natural history insights and elegant experiments, Dai proved me wrong and I was delighted about the outcome. This research was the single most exciting project I have ever been involved with.

The story has a few messages that may interest readers—the discoveries themselves, but also the important realization that science is often a messy meandering path to enlightenment. Experiments often fail, but sometimes in ways that yield important new insights. In addition, incidental natural history observations that initially seem weird and puzzling can turn out to be vitally important for figuring out what is going on. And a bonus coincidence: Dai Shizuku was a postdoc with Jerry’s colleague Trevor Price at Chicago and was based in the lab right next door to Jerry for a year.

Below: Dai Shizuka at our BC site.

Below: My initial conclusion that coots don’t recognize parasitic chicks was not surprising given what we know about the more famous interspecific brood parasites like cuckoos and cowbirds. Virtually all of their hosts fail to recognize parasitic chicks, even when the difference between host and parasite is ridiculous, as the photo below of a common cuckoo (Cuculus canoris) and Eurasian reed warbler (Acrocephalus scirpaceusshows. The cuckoo chick looks nothing like the host and is several times larger than its foster parents (photo from the web). This lack of chick recognition has long been an evolutionary puzzle, particularly because the hosts are often extremely good at recognizing cuckoo eggs that differ only subtly from their own eggs.

Below: Back to within species parasitism in coots: can you spot the parasites? I bet you can’t because I would not be able to do so without notes on which chicks came from which eggs. This is an experimental brood that contains some chicks that are not the host’s own chicks (I know this because we added them to the brood).

Dai did not set out to study chick recognition; this would have been a crazy idea for a PhD project given the lack of recognition in systems like cuckoo hosts. Instead, he was interested in the function of the extreme hatching asynchrony coots show and he set up some exploratory experiments in which he created synchronously hatching broods by swapping chicks between nests.

Below: We hatch all of the coot chicks in incubators at the field cabin, which makes it easy to swap chicks among broods for experiments. We collect eggs from their nest as soon as they show signs of hatching (slight cracks on the shell) and return the hatched chicks to a nest (home or foster nest depending on the experiment) on the day they hatch. Due to the hatching asynchrony, we have to visit the same nest several days in a row to get all of the hatching eggs and releasing chicks back to the nest. The photo below shows Dai processing chicks in our three incubators.

Swapping offspring among families (called cross fostering) has a long history in evolutionary biology; it is used in heritability studies to control for environmental affects on traits. A critically important assumption is that foster parents treat the cross-fostered offspring as they would their own, but of course coots have a mind of their own. Dai came to me one day with both bad and good news. The bad news: the cross fostering experiments failed as a method to investigate hatch patterns because the fostered chicks had significantly worse survival than the host’s own chicks. The good news: this survival pattern suggested that coots might be able to recognize some parasitic chicks. Cross fostering is like adding parasitic chicks to a brood; the birds are given chicks of another pair, and when these chicks do worse it would most likely be due to discrimination by the foster parents. These preliminary results called for a full-scale experimental study to investigate chick recognition!

The following year we repeated the same experiment but with a much larger sample size. After a hell of a lot of work, we found no evidence for recognition. The year after that, more experiments and more toiling, but again not a shred of evidence for recognition. We even did a second experiment that involved cross fostering chicks while maintaining the normal pattern of hatching asynchrony, but this too failed to show evidence for recognition.

Now, a sane person would have given up at this point and concluded that the birds really don’t recognize parasitic chicks. However, Mother Nature kept teasing us with tantalizing signs that these birds really can distinguish their own chicks from brood parasites. While conducting brood observations from a floating blind, both parents of a brood I was observing suddenly turned on one of their chicks and began to peck at it really aggressively in a way that differed from normal spankings. There were several bouts of these attacks and it seemed they occurred whenever the chick called out. I eventually lost track of the chick, but Dai found it dead in the marsh a few hours later. Its tag revealed something really interesting—the chick was from a brood parasitic egg! This was the first time we had witnessed direct infanticide and it involved a brood parasite. The video below shows this attack:

Second, adoption is very rare in our population, so it was bizarre when a coot pair adopted three chicks from their neighbor. Checking the chick ID’s in our records revealed that all three adoptees were from parasitic eggs the adopting female had laid in the neighbor’s nest. The parasitic chicks had come back home to their biological parents! If this story involved humans, Hollywood would be asking for a movie script about now. This observation suggested that the parents recognized the chicks as their biological kids even though they had hatched in the neighbor’s nest.

These intriguing natural history observations suggested that we were missing something with our experiments. The critical missing piece came from an observation of what appeared to be a completely deranged pair of coots. Once again I was in a floating blind observing a pair of coots with their very recently hatched chicks. Everything seemed to be normal but then a switch seemed to go off—one of the parents started to kill its own chicks, one by one, until they had all been killed off. I was stunned because we had never seen anything like this before and we were also certain that these were the biological chicks of these parents.  Here is a video showing the chick killing behavior (a different brood from experiments described below but it shows exactly the behavior I saw with the killer parents):

Several days later, over dinner, Dai made sense of why these birds went crazy, and he had an epiphany that solved everything. During dinner discussions, one of the field assistants recalled seeing the crazy killer parent coots feeding two chicks for a short period of time before their own eggs hatched. These birds had temporarily adopted their neighbors’ chicks before their own eggs had hatched, but the chicks returned home before the pair’s own eggs hatched. Dai reasoned that coots learn to recognize their chicks and that the short exposure to the wrong kids had caused these parents to learn from the wrong chicks and then when their own chicks hatched, to incorrectly see their own chicks as brood parasites.

This line of thinking led to the hypothesis that coots learn to recognize their chicks by imprinting on the chicks that hatch on the first day—they use these first hatched chicks as reference chicks to learn some recognition feature that they could then apply to chicks that hatch on the following days. The simple rule of using chicks that hatch on the first hatching day would be a pretty reliable rule since at most nests, only host chicks hatch on the first hatching day. The chicks’ features that the parents learn could involve vocalizations, smell or visual appearance—or a combination of these. To be useful, the features would have to vary among pairs but not within a pair’s own brood—just like the way they recognize eggs. This hypothesis could also explain our consistent failures up to that point. If parents learn from the chicks that hatch on the first day, then parents at nests where both hosts and parasites hatch on the first day would learn that both chick types are their own and would not be able to recognize parasite chicks. This was the design of all of previous experiments. Doh!

These insights came near the end of what was supposed to be the final field season of the project, but we decided that the idea was so interesting that it needed to be tested. So we went back one final year—it was do or die for showing that coots recognize parasitic chicks.

We conducted two main experiments to test Dai’s hypothesis. Since we hatch all chicks in captivity, we got to decide which chicks parents are exposed to, and in what order. In the ‘host first’ experiment, we gave birds their own chicks on the first hatch day, and then gave an equal mix of host and foreign chick on each day after that. Foreign chicks were chicks from a second brood and are the equivalent of brood parasites. The graphic below shows the experimental design. If parents use the day 1 chicks to learn which chicks are their own, then we can predict that the foreign (‘parasite’) chicks will survive less well than the host chicks (assessed only for chicks from days when both types hatch). For testing mechanisms like learning, real proof comes from tricking the birds to do the wrong thing. We therefore did a second experiment where birds were given only foreign chicks on day 1 and then an equal mix of host and foreign on later days (‘foreign first’ experiment). If the birds really do learn from the first hatched chicks then they should imprint on wrong chicks and discriminate against their own chicks.

After setting up the first few experiments, we were on pins and needles. Then, after surveying only three broods I knew that we had finally broken the curse. The results were so clear and strong that we now had convincing experimental evidence that coots can recognize parasitic chicks. Birds given their own chicks on the first day were able to recognize many of the parasitic chicks that hatched from day 2 onward, while birds given somebody else’s chicks on the first day treated their own chicks as parasitic. Evidence for recognition came from big survival differences between chick types that matched predictions. And these survival differences happened so quickly that we were barely able to see what exactly was causing them (the differences had already happened before our first survey on most broods). However, we were able to watch a few broods right at hatching to see what was happening. Just like the deranged pair of coots that killed all of their own kids, these birds very quickly dispatched many of the chicks that were the ‘wrong’ chicks. In three broods we only saw the female culling the chicks but in another brood both sexes were involved. Our experiment not only showed chick recognition, but how it occurs: parents learn something about the chicks that hatch on the first day. We do not yet know what information the parents learn but I strongly suspect that the information is in the chicks’ vocalizations.

We are aware that experiments like ours do raise some ethical considerations. However, keep in mind that we did not create unnatural behaviors that do not normally exist. Moreover, at our site each year thousands of coot chicks die naturally from starvation; it is not clear to me that this is a kinder fate than what happens to the brood parasite chicks.