Scientists have been arguing for more than twenty-five years about whether it is a good idea to collect voucher specimens from particularly vulnerable biological populations. Some think that, obviously, scientists should not be harvesting (read: killing) organisms from, for instance, critically endangered species. Others think that, obviously, it is the special job of scientists to collect precisely such information before any chance of retrieving it is forever lost. The character, extent, longevity, and span of the ongoing disagreement indicates that this is likely to be a hard problem to solve. Nonetheless, the aim of this paper is to help field biologists figure out what do to when collecting a voucher specimen risks extinction. Here I present and assess varying practices of specimen collection for both extant (i.e., neontological) and extinct (i.e., paleontological) species in order to compare and contrast cases where extinction risk both is and is not a problem. When it comes to taking vouchers from extant species at some risk of extinction, I determine that those advocating for conservative approaches to collection as well as those advocating for liberal information-gathering practices have good reasons to assess things in the way they each do. This means that there is unlikely to be a decisive, one-size-fits-all response to this problem. Still, progress can be made. We can acknowledge the risks of proceeding in either manner, as well as the uncertainty about how best to proceed (which will be deep in some cases). We can proceed as thoughtfully as possible, and be ready to articulate a rationale for whichever procedure is used in any particular case.