I'd need to see some pretty strong evidence before believing that murder was main cause of death among males in "pre-civilised societies", and that evidence would be pretty hard to gather. That's quite a sweeping statement, especially given the wide variety of societies that would presumably be included in that category, and the range of ways in which it is possible to die without being murdered.

I haven't read the book in question, but I'd be very surprised if Diamond would claimed to quote "studies showing that in pre-civilized societies, murder is the single greatest cause of death in males". I'd have thought that at most he would have cited studies which suggested that, in some "pre-civilised societies", this may have been the case.