The BDI might be a little weird since you aren't dealing with a clinical population, which is really what its for. A common mistake many researchers make is to use it for diagnostic purposes (drives me crazy when I see that!). Its meant to measure depression severity among depressed individuals. I don't know the area so I'm not sure to what extent an androgen-induced depression might differ from typical MDD. As long as your interests are entirely in the presence or absence of symptoms and not in whether they meet some kind of clinical threshold, the BDI is probably a good choice.
Have relatively little experience with the Hamilton, but I've heard good things about that one as well.
Only other one I've used is the CES-D, which I hate with a fiery passion. Soooooooooo many false positives. We used it as a screening measure with one study (because the IRB insisted...don't ask). Its just very sensitive to response styles and what we got seemed to be much more reflective of how the individual tended to respond rather than the presence of any actual symptoms. I think eventually we ended up finding some obscure study that had changed the scale headings on it that we could use as an excuse to change ours so we could stop excluding perfectly healthy individuals who were flagged as depressed.