Josh Tewksbury has just finished a series of very nice posts chock full of good advice for giving a job seminar. The one thing I’d add is that your job talk can’t be too superficial. Yes, you need to give the audience the big picture of your science rather than just narrowly focusing on one specific project. But you also need to “give the audience some meat” (as a colleague of mine once put it). You shouldn’t just skim over too many projects, that doesn’t allow your audience to appreciate and evaluate your science. Rough rule of thumb: go into at least some depth on 2-3 projects or lines of research.
A blast from the past: you know how everybody working on phylogenetic community ecology or “trait-based” community ecology these days interprets coexistence of closely-related or phenotypically-similar species as indicating “habitat filtering”? As I explain in this old post, they should really stop doing that. (UPDATE: this link works now)
Another blast from the past: The new cohort of graduate students who started this fall are starting to develop their research projects. So here’s an old post I did on weak reasons for choosing a research project. Like “no one’s ever studied X”, or “lots of people are currently interested in X”. Lots of people appeal to such reasons to motivate and justify their research. Don’t be one of them!