Whales, culture, and graphs

Whales, apparently, learn from one another. That’s the message of this paper in Science last month on the transmission of feeding tactics among humpback whales. There are lots of fascinating things in the article, some of which are news to me but have been known for a long time to people who pay attention to these things: whales sometimes feed by blowing bubbles around schools of fish, corralling them into easily swallowed bunches, and they sometimes, beforehand, whack the water’s surface with their tail fins, probably to further disorient the fish. It’s this last behavior that’s the subject of the paper: Observing the interactions of several hundred whales via more than 70,000 sightings and analyzing the occurrence of the “lobtail” tactic of fin-slapping shows its cultural “diffusion.”

The paper has been nicely commented on by others (e.g. link), to which I’m unlikely to add anything directly relevant that’s of interest. Two indirect comments come to mind, though:

1. The data for this paper come from the observations of about three decades of whale-watching excursions, collected by others and analyzed by the paper’s authors. It provides yet another example of mining insights from increasingly abundant “big data.” I briefly noticed this paper a few weeks ago, but then I actually thought about it for more than a few seconds when one of my students presented it at the “three minute journal club” section of our weekly group meetings. Coincidentally, the other paper he covered, though on the totally different topic of leaf morphology, also involved the new analysis of large, old datasets. Perhaps I’ll write about that one some other time.

2. Like many papers on network analysis, this one includes a truly awful graph:

This isn’t meant as a criticism of the content of the paper, but I really do wonder what the authors hope to convey with this figure beyond “We use networks! You can tell because there are nodes and edges.” I see this in lots of papers these days. Normally, one uses a graph (“graph” in the conventional sense) to convey some trend among data, or to encapsulate complex numerical information in an easily absorbed form. Neither of these aims are met, it seems to me, by many papers involving graphs (in the mathematical sense) — who can tell anything meaningful from the shrapnel of dots and lines? Their inclusion seems more to be a signal to the reader that the authors are aware that graph theory is neat, rather than something more substantive. Of course, I’d be happy if I’m wrong, and if there’s a message contained in the above figure, someone should write! And again: the paper is fascinating, and my poking at the figure is a minor quibble.

On a lighter note, S. and I painted whales yesterday, shown here:

Mine is rather clumsy, and has a horrible sense of scale. His shows a whale poking out of the water to lick a coconut, a feeding behavior not yet observed in nature.