1. The generating and using of increasingly complex guesses as to how organisms are related to each other.
2. Something you have to do these days to study evolution.
That time has come. For five years in a heavily phylogenocentric lab in a museum mostly focussed on phylogeny in a department deeply into phylogenetics at a time when phylogenetics is nearing (I hope) the zenith of its trendiness, I have avoided really learning how to do phylogenetics. I can talk at length and in detail about the philosophical underpinnings of phylogenetics, I have read books and papers and taken classes on the subject, but I have never actually sat down and applied that knowledge. This is partly because of my inherent and unreasonable dislike for everything trendy and partly because I find that the most boring research talks in the universe are the straight phylogeny talks ("And then we sequenced 4327 base pairs of CR32.5 and SLD19423 from these twelve hundred taxa. Notice that on this taxa here there is a deletion, and I'll spend ten minutes talking about how we dealt with that. Now I'll spend half an hour talking about how we generated the priors for our Baysian analysis. And look, this taxon you have never heard of is closely related to this other taxon you have never heard of. Who would have thunk it? Someone wake that guy with the funny hair").
But my phylogenetic inexperience is based on more than simple obstinacy. I don't think that way. My predilection is to think of evolution in terms of selection, mutation, drift and so on. Phylogenetics at its core doesn't care WHY there are differences between organisms, phylogenetics is focussed on the methods for gathering and analyzing data on HOW these taxa are different from each other, and on drawing trees of relationships. In many papers, the tree itself is the goal, and maybe they do an analysis or two showing how useful their tree is.
I, knowing I needed to learn some phylogenetic software packages eventually, but deeply not wanting to, have backed myself into it. So I have taken a taxon for which the tree already exists (primates) and gathered from the literature (or had my students gather) a bunch of variables for as many species on that tree as possible. We have data on sex biased dispersal, social system, who provides care to the young and so on for about 90 species, and data on sex-biased longevity for 119. A huge amount of work over some years has gone into this, and there is no way I can weasel out of writing papers based on it. But there is also no way I can publish this in a decent journal without controlling for the effect of phylogeny. What "controlling for the effect of phylogeny" means takes a little bit of explaining. There is a tendency for related species to have similar traits, whether or not there is any adaptive mechanism driving that similarity. The common ancestor had that trait and both the descended populations inherited that trait from that ancestor. Humans and chimps have similar genetic sequences, and our common ancestor was surely very similar to both of us. This is termed 'phylogenetic inertia.'
Anytime one does a comparative analysis these days, one has to explain how we know that the observed pattern isn't just an example of phylogenetic inertia. Imagine one thought there was a causal relationship between being large and having hooves. One could find ten big species, notice they all have hooves, and ten species, notice none of them have hooves. But if those ten hooved species were all in the cow family, and the ten small species were all in the vole family, one would not have proved anything about hooves and largeness except that Bovidae have both and Cricetidae neither. So one has to make sure one is not being fooled by similarities due to evolutionary relatedness, or in the parlance, 'control for phylogeny.'
I need to control for phylogeny, and therefore will learn a few phylogenetics programs. But I don't have to like it, and I am going to make my students learn it too.