I agree with the conclusion that Amazoninae and Proboscigeridae may be seen as in prevailing use relative to, respectively, Androglossini and Microglossidae. Note that Joseph et al 2012 (who revived Androglossini and retained Microglossidae) did not, actually, perform any global literature search to assess name usage at all. Rather, they deliberately restricted their assessment of usage to a smallish set of “mainstream” (= arbitrarily selected) references (40 in total -- not all published in the sense of the Code), all of them strictly ornithological and more or less oriented towards taxonomy. In the case of Amazoninae at least, most of the authors who adopted the name in the 1975-2012 period were using Wolters’ system, and did not concern themselves directly with bird systematics : they were aviculturists, veterinary scientists, students of parrot ectoparasites, etc., who used the name at least in part as an anchor to make the data they published discoverable. These authors's works were of course absent from Joseph et al's "current usage references". (There were in fact still more uses of Amazoninae, by such authors, than listed by Gregory & Sangster. Enough, actually, to make that, in 2012, the name should have been made a
nomen protectum relative to Androglossinae under Art. 23.9.)
I would like to question the notion that "replaced", as used in Art. 40.2, is to be understood as requiring that a name was
proposed to take the place of the older family-group name.
(Such names would arguably be new replacement names (
nomina nova), except that
nomina nova are generally problematic in the family group name because (1)
nomina nova,
by definition, inherit the (nominal) type of the name they replace, and (2) family-group names
must be formed from the stem of the name of their type. If you replace a family-group name with a new one, formed from the stem of another nominal type genus, the name so formed will unavoidably have a different nominal type, and thus
cannot, actually, be interpreted as a
nomen novum for the original name.)
Note that Art. 40.2 further states that it is "the substitute name" that is to be maintained. A "substitute name" is "Any available name, whether new or not, used to replace an older available name." (Glossary). This definition implies that a name can become a substitute name at any point in its history -- it's not something that is directly linked to how the name was introduced.
FWIW, I regard Amazoninae as having been established by
Des Murs 1886, not by
Mathews & Iredale 1920: at this date, there is of course no way that the name was proposed "under Article 5 of the Règles (CIPNZ, 1905)". Anyway, the idea that Mathews & Iredale 1920 could have proposed Amazonidae with the replacement of Androglossinae in mind -- while Sundevall's name had not been used for almost half a century -- seems far-fetched to me at best. This is particularly true in the light of the fact that Mathews & Iredale 1920 indicated clearly that another name had been used in the recent past for several of their (e.g., "Palaeornithidae olim" for their (newly proposed !) Psittaculidae, on the same page as Amazonidae), and they did not do this in the case of Amazonidae.
One significant problem with Amazonini may be that Sittacinae Sundevall 1844 (although usually used for macaws and relatives) arguably applies to the same group, as a result of a 1855 type designation by Gray, which makes
Sittace Wagler 1832 a synonym of
Brotogeris Vigors 1825. (At least this is the earliest valid designation that I can find. Designations that have been claimed in Gray 1840 or 1841 do not exist -- in these works, Gray cited
Sittace in the synonymy of three different genus-group names for which he designated three different types; these citations must be read as indicating that the three types fell within the very broad original original circumscription of
Sittace, none of them can be interpreted as a type designation for
Sittace. Bonaparte restricted
Sittace to macaws before 1855 but without any type designation. The family-group name has been used a couple of times in the early 20th C -- not under the assumption that
Sittace is a synonym of
Brotogeris, of course, but this remains the same name and would prevent making it a
nomen oblitum.)
Lucky you.
(The citation of what -- correct me if I'm wrong -- seems to be an unpublished (in the sense of the Code) version of the work, in an analysis of usage for the purposes of nomenclature, is a bit surprising, however. Not that it matters really, of course -- the printed version is presumably validly published and uses the name as well.)
Interestingly, many of the attributions of pre-1861 names in this paper (Table I) are novel, and are evidently taken straight from a file I once shared with Steven. Yet, I don't even get a vague acknowledgement for this.