Locust and tumour invasion
Interestingly it seems that the reason that locusts combine into swarms is out of fear of being eaten by other locusts. Although locusts are normally herbivores, when food becomes scarce some of them might resort to cannibalism.
And that is interesting in the sense that these highly aggressive (from the way they travel and eat) swarms are made of the less aggressive individuals that band together and escape from the cannibals, with devastating consequences. This is rather similar to what some researchers are finding in tumours, both experimentally and theoretically. As was discussed in previous posts (here or here for instance) reverting to a glycolytic metabolism allows tumour cells to survive in environments poor in oxygen while harming their neighbours. It has been hypothesised that motile tumour cells are more likely to appear if there is an increase in the acidity of the environment, in many cases as a consequence of the existence of these glycolytic cells (I should give a disclaimer and say that one of my papers is devoted to study this using game theory). Could it be a similar phenomena? that is, that less aggressive cell types escape from the more aggressive ones and that the consequence is also devastating (since these motile cells are responsible for the invasion and metastasis that characterise the last stages in cancer). Probably not very useful as an analogy: I am not sure how these researchers measured the fear of locust to their cannibal colleagues but it is safe to assume that tumour cells do not have much of that. Still fear is an instinct for self preservation that has a parallel in the microcosms of a tumour.