Just had occasion to look something up in a recent work of popular, yet scholarly, history based on a great deal of archival research, since I remembered that when I read it the author mentioned consulting a certain archive at a particular repository.
However, in the 'List of Archives' at the end, all we get are the names of the various repositories, and absolutely no indication of what actual collections in any given repository had been consulted. This is something that strikes me as extremely bad practice. Most of the repositories listed (except where the onsite records of a particular organisation had been consulted) hold very large numbers of archival collections, and in many instances several which the author of this work might have consulted.
I did, in the end, manage to find the information I was seeking through trawling through the footnotes of the chapter in which the archive was most likely to have been referenced. But this should not be necessary.
At least, however, the information had been footnoted in reasonable detail. It is extremely annoying to archivists when scholars cite some item - sometimes not even the actual file, but a single letter within a file or volume - with no reference, except perhaps to the very large collection within which it might be found, or with no indication at all as to the specific collection, only to the repository as a whole. Furthermore, if scholars have been given access to uncatalogued or partially-catalogued material, it behooves them to be particularly meticulous in contextualising any material they cite so that in due course it can be located by other researchers.