Sunday, March 16, 2014

Nope, couldn't find an answer.

To the post below that is. I was thinking that skin colour change in trout, is mostly to do with overall hue, colour intensity or something like that, without affecting the spot distribution and pattern itself. Actually I don’t know if spots change with changes in overall colour as I couldn’t find anything about it and I sort of lost the will a little - it is not the kind of stuff, at this reduced level, I am interested in. There was a study1 that looked at identifying young salmon and trout from the melanophores, the little spots under the eye that parr exhibit. Apparently it works, you can tell the difference between individuals based on the number and distribution of spots, but since the study was published in 1994 and has only been cited three times subsequently it seems doubtful that this has been picked up and used with any regularity. And the spots disappear as the fish get older anyway so - fat lot of good for adults.

Maybe I should go back a bit.

The reason I wondered whether trout could be told one from the other on the basis of their spot patterns is due to one of those reading rambles I go on, a reading ramble that started with wasps. See there was this study2 a few years ago, quite a few years ago now, that I thought very neat. I put it aside to use in a post and, as you can see promptly forgot about it until now. The study highlights the fact that even ‘lower’ animals can learn and remember really quite subtle differences in appearance - an interesting point if you wonder whether trout can tell the difference between a real and an artificial fly.

Image courtesy reference cited in footnote 2.

These are a species of paper wasps, Polistes fuscatus and, as I am sure you have noticed, each of the individual wasps faces in the picture are slightly different. These are real animals (Abigail, Belinda, Cathy, Doris ….) not computer constructs for the purposes of the study. It turns out that this particular species of paper wasp is very social and to avoid unnecessary conflict or confusion about roles and who is who and who is dominant to who and who is feeding the bloody cat tonight, they learn to recognise and distinguish individual members of their colony by the different patterns on their faces. Closely related but solitary nesting paper wasps, like Polistes metricus, don’t have the variable face patterns you can see above and do not recognize other individuals one from the other.

And so I thought, cool, imagine that. This kind of individual recognition is a complex and specialised form of cognition requiring flexible learning and memory. If a little wasp can learn, remember and apply these memories to accurately discriminate patterns, imagine how that would apply to a trout discriminating between real and artificial flies. Naturally it is a bit more complicated than directly assuming that because some wasps can perform this kind of task trout may have and be able to apply this ability in a different context. But leaving that aside the study made me think about individual recognition a bit more. The interesting point is showing that not only can wasps discriminate between individuals but that they do so by using ‘facial pattern recognition’. Where it lead to was a review article3 and this image.

Image courtesy reference cited in footnote 3

At the top we have the wasps again showing their different faces but in addition you can see variation in fiddler crabs, ruffs, queleas, swallows, murre eggs, hunting dogs and finally humans. All of this information is in the caption of course and you may note the mistake in attribution of the human faces. The point here, despite the human error, is that some animal groups, particularly very social animals or those that live very close together, contain individuals with markings that are unique to that individual and that these markings can be learnt and remembered by others.

It is well known that fish, trout included, can learn and remember the identity of other conspecifics. They are not truly social in the way the paper wasps are but trout, as with many other fish, often live in close proximity to each other and compete for places to live, compete for food and compete for mates. This competition is energetically costly and dangerous if it distracts competing individuals from feeding and other predators or sources of danger. So trout in stream pools organise themselves into a strict linear hierarchy with the most dominant individuals in the best lies/feeding lanes and the most subordinate fish in much more reduced circumstances. Much of the ranking is based on size - an easy way to assess the strength of a neighbour. But within a pool there are usually a number of trout of the same or a very similar size. If they couldn’t learn and remember who they had won a fight against and who they had lost a fight against they would be at it continually and squandering all that energy and all those feeding opportunities mentioned before.

They don’t do that of course, the fitness consequences are too dire. What they do is simply remember who is better, who is worse, from the conflicts they have and they makes sure they don’t fight again with higher ranked individuals. Which is all rather clever. But there is more. A recent study4 has shown that trout can use ‘transitive inference’ to minimise conflicts even further. Transitive inference is a type of social learning that uses known outcomes to deduce unknown ones. In a simple situation a trout can watch conflicts and infer the dominance order even if it hasn’t seen all combinations of the conflict. For example, working out a dominance hierarchy of three trout where A is higher than B which in turn is higher than C requires that the watching trout only see a conflict between say A and B and B and C to establish the A > B > C order. It doesn’t need to see the A versus C fight. If the observing trout has already had a punch-up with either A or B or C it can insert itself into the heirarchy without having to challenge the other two. That is very neat and has now been shown to hold for at least five order hierarchies. Transitive inference was once thought to be restricted to humans and not thought to feature in complex social systems like that of trout. Not so now.

That’s a bit of a digression but a good one to show again how easy it is to underestimate the cognitive processes that even ordinary animals like trout routinely perform day to day. Back to the original thought. If trout learn to recognise individuals in their social set what is it they are actually remembering. Paper wasps, at least the social P. fuscatus above, use the difference in face pattern to remember who is who in their colony. But just because trout vary doesn’t necessarily mean that their colour variation and spot distribution is so that Maude can tell Joe from Gerald.

It’s clear that one of the most important reasons that trout vary is so that they can match the background they are swimming over to help avoid predators. And they can rapidly change this colour intensity if they move from a dark background to a light background (or vice versa). Even then it might be that some colour or pattern signaling would still convey useful signals to neighbours while also allowing effective camouflage. But still there are signals and there are signals. Take a look at this.

Image courtesy reference cited in footnote 3

The queleas variation in plumage colour is thought to be a signal of individual identity in a bird that is highly social. Hence, there are considerable differences between individuals (each point in the graph is one bird) in the range of colours they exhibit, their intensity and the placement of those colours on the bird (see middle image). Male cardinals don’t go in for any of that scruffy tat. They don’t care if no one knows their name, they are after plumage that is intense, bright, stunning and they are after it in the main to impress the opposite sex. Imagine if human females only really liked human males sporting large biceps and six-pack abdomens. Those human males who wanted to attract females would be training their bodies to converge on the best form for arm and stomach muscles. So cardinal plumage is designed to signal the quality of the individual not its specific identity and their plumage converges on an ‘ideal’ (the points are clumped to the top left of the graph) rather than the variation shown by queleas.

The variation we see in trout therefore, particularly at certain times of the year and particularly in males whose change of colour (along with developing other signals such as an enlarged adipose fin and a kype) might simply be a signal of quality rather than identity. We know this is true - one only has to look at the lengths the outrageous charr goes to see how much investment there is in skin colour signals to advertise fitness. But as with variation in colour for background matching it isn’t clear to me that basic pattern, spot pattern changes too. For example although these two brook trout are a good example of background matching - one to a very dark forested, shady and seemingly bottomless stream and one to an open, gravel-bedded river... may be that those pink, mauve-haloed spots do not vary, that they are honest identification signals.

Or it may be not that at all. It may be smell, perhaps they sniff out the difference between themselves. Or sound. Like the bleating of sheep they fart air from their swim bladders in individually tuned syncopation. Or body shape.

And so I come full circle with no answer at all. Are spot patterns used by trout to tell one individual from another? Dunno. But it was a good ramble through a fascinating corner of the literature. And it does once again illustrate how complex other animals are and how dangerous it is to simply dismiss them as “bags of chemicals” just because they are not us - as well as illustrating how other animals core behaviours are ours too. Transitive inference indeed. Blimey. Leaniz, C. G. et al. (1994) Individual recognition of juvenile salmonids using melanophore patterns. Journal of Fish Biology, 45, 417-422.
2.Sheehan, M. J. & Tibbets, E. A. (2011) Specialized face learning is associated with individual recognition in paper wasps. Sciecne, 334, 1272-1275.
3. Tibbets, E. A. & Dale, J. (2007) Individual recognition: it is good to be different. Trends in Ecology and Evolution, 22, 529-537.
4. White, S. L. & Gowan, C. (2013) Brook trout use individual recognition and transitive inference to detemine social rank. Behavioral ecology, doi:10.1093/beheco/ars136.


T.J. Brayshaw said...

"Imagine if human males thought all human females liked large biceps and six-pack abdomens."

If? Wait. What?

T.J. Brayshaw said...

With respect to how this transfers (get it?) over to flies, fly design and real food, I think the thing to remember is that selection may choose for recognition in social situations but not in the context of foraging. In fact, recall that in the wasp study, even the non-social wasp was as good as the other in learning patterns and caterpillars (prey). It just lacked "face specialization" cognitive abilities, not other visual or cognitive abilities.

Eccles said...


"If, Wait. What?"

Oh, well did I ... yes .... oh dear, hmmm.

Eccles said...

... and to your second point, yes of course I mentioned the ability might not transfer but perhaps I should delve a bit more into their ability to distinguish between the types of their caterpillar prey.