The "peers" would have to be other researchers of eighteenth-century gunsmiths--those who, by some measure of consensus, were the most respected researchers in the field. As in any peer review system, the reviewer won't be able to assess everything in a particular article, but he/she could certainly weigh in on whether the writer has engaged with relevant material, overlooked major contrary claims, etc. The anonymity of the review system, while subject to abuse, for sure, does enable a reviewer to criticize a shoddy article in ways that a friendly editor of a journal (esp. in a small field) may be uncomfortable doing.
By the way, there is plenty of published research on barns, pottery, weaving, tools, painters, and dress that has undergone peer review. It's overstating things to say that one would be "hard pressed to find any" work in those fields that is "peer reviewed." But please don't take that to mean that all the excellent work in those fields has been peer reviewed: there is plenty of excellent work in those fields that hasn't been peer reviewed. I published a piece in a peer review journal on a colonial painter, a few years ago, and the journal sent it to five reviewers, both historians and art historians. Their suggestions required more work from me; and this certainly improved the article.
It's certainly true that peer reviewed work gets refuted or modified: it nearly always does, at some point, which is how fields progress or grow or whatever you want to call it (stay in business, perhaps). But this phenomenon is just the definition of a field of research. It isn't a knock on peer review--as if, without peer review, this phenomenon wouldn't occur.
Like you say, there is superb research in this field. And then there is poor research, also in print. The ideal situation would involve some sort of peer review that would try to produce as much of the former as possible, through suggestions or recommendations after a reading of the first draft that would be incorporated into later drafts, and to prevent as much of the latter as possible from seeing print in its initial draft-state.
But I am a newcomer to this field and am uncomfortable with the position I seem to have put myself in this thread. I only meant to suggest, after JWH1947 described (in his initial post) an article big on claims and little on substantiation, that peer review is a system that, for all its flaws, can be successful in dealing with this.