The great tit (Parus major) needs to gain more than 10 % of its body weight in pure fat every evening, in order to survive a cold winter night (Image Credit: Frank Vassen, CC BY 2.0, Image Cropped)

Short-term insurance versus long-term bet-hedging strategies as adaptations to variable environments (2019). Haaland, T.R. et al., Evolution, 73, 145-157.

The Crux

Why do animals behave the way they do? Behavioral ecology is a field of research trying to explain the ecological rationale of animal decision making. But quite often, it turns out the animals are doing the ‘wrong’ thing. Why don’t all animals make the same choice, when there clearly is a best option? Why do animals consistently do too little or too much of something?

Sometimes it’s because we’ve got our explanation wrong, or we’re measuring the wrong thing. Other times, animals can have good reasons for behaving ‘suboptimally’. For example, if you’re faced with unpredictable conditions it’s often best to ‘play it safe’, if the penalty of doing too little of something is much worse than doing too much of it. Small birds in winter will build up fat stores that they burn every night in order to stay warm – preferably a bit more than what they actually think they’ll need! Being too fat makes it harder to fly away if a predator comes, but since you could die if you’re too cold that night, it’s better to be on the safe side. This strategy is known as ‘insurance’, for obvious reasons, and the more variable overnight temperatures are, the more extra insurance the birds will need.

This seems intuitive enough. But things change when we look at a longer time scale. We often think of evolution as a process that creates species that are perfectly suited for their environment. But if a bird continues to behave suboptimally (keeping themselves insured), could they be outcompeted by species who take a few more risks? That’s what we wanted to find out.

What We Did

We simulated the evolution of a hypothetical animal’s behaviour. It could be how much fat to store overnight, who to mate with, where to build your nest, when to migrate, etc. Basically anything that people working on animal behavior might study. The key thing is that we put in a skewed fitness function, i.e. doing slightly too much of something gives you much less fitness than doing slightly too little. And we let the optimal behavior (the peak of the fitness function) depend on the environmental conditions, which is variable both within and between generations. This allows us to find the optimal behaviors (how much insurance you need) over both the short term and long term, and identify when these might differ.

In addition to the mean behavior, we also let evolution adjust how variable this behavior should be. If the environment fluctuates a lot, the best strategy might be to actually create different offspring suited to different environments, so that at least some of them will do well no matter how the environment turns out. This ‘don’t put all your eggs in one basket’ makes sense from an evolutionary point of view, but it comes at a considerable fitness cost, since it means that every generation you’re producing a whole bunch of completely maladapted individuals.

Did You Know: Modelling Optimal Animal Behavior

Using mathematical models to find out the optimal strategy of an animal can be quite exciting. If you measure its behavior in nature, and it matches your model prediction, it’s a sign that you’ve really understood how the animal thinks about the problem. Or is it? Is the animal doing complex calculations about temperature predictability and predation risk while deciding how much to eat for dinner? Of course not. The key assumption is that we just say that evolution has had enough time to optimize this. We assume that there’s one gene simply determining the animal’s strategy, and that the optimal strategy has evolved. This shortcut is strictly speaking never true, but in many cases it works surprisingly well.

So our model combines these different short- and long-term adaptations to uncertainty. How do these strategies interact? Not much work has been done in this interface of behavioral ecology and evolution, so we didn’t really know what would happen.

What We Found

The best strategies in the long term (over evolutionary time) can differ a lot from the best strategies in the short term (over the course of a lifetime). For example, the “don’t put all your eggs in one basket” diversifying strategy works very badly in the short term, but quite well in the long term. However, within-lifetime insurance strategies work well also in the long term, and can actually reduce the need for diversifying offspring across generations.

Another interesting result consistently appearing is that evolutionary time scales require more insurance than what is optimal from the viewpoint of an individual animal. Thus, if an animal appears to be doing the wrong thing today, past environmental fluctuations could have shaped this behavior so it might actually be optimal in the long term.

When migrating, if arriving slightly too late means you do much worse than if you arrive slightly to early, what time should you decide too migrate? (Image Credit: Andreas Trepte, www.photo-natur.net, CC BY-SA 2.5)

Problems?

The strengths of this study are also its weaknesses. Most importantly, it’s a very general model, capable of explaining everything and nothing. Models will always vary from extremely specific (describing one thing perfectly, but of absolutely no relevance if you’re not interested in that one thing) to extremely general (describing just about anything, but of no relevance to any one specific thing). Our model is closer to the latter. For example, it leaves out a lot of complicating factors. What would happen if the animals could sense the environment in some way and adjust their strategy (which we of course know that they do)? This is a question for future studies.

So What?

Despite being very simplified, this paper ties importantly together different areas of theory that previously haven’t been well linked. There are both behavioral and evolutionary strategies for how to cope with varying environments, and we now know slightly more about how these are connected.

As we know, anthropogenic climate change is making environmental conditions more variable and unpredictable across the globe. Understanding how populations have adapted to environmental variability in the past might just help us predict how they will respond to future changes in these environmental patterns.