Robert, the paper by Amadon (1949) does not help much here. The procedure described by Amadon is connected with the phenetic conception of the subspecies. At that time the subspecies was treated as a geographic aggregate of the species without any relevance for the evolution (origin) of new species (Mayr 1982). Nowadays, after the rise of the 'population thinking' geographic forms (subspecies) may (are) treated as evolving (sub-)units. At the first glance it sounds trivial but actually its quite different because these two approaches are based on distinct scientific concepts (class-concept versus entity-concept; see Reydon 2005). It is (from an epistemological point of view) not possible to mix them up. Besides this philosophical reason the Amadon procedure is also not connected with the 'problem' discussed here, because it treats with geographic variation only (subspecies in the 'old fashion'). But A. citrinellus and A. labiatus occurs in the same lake (syntopic). The application given by Amadon are much outdated and formulas are not in use anymore. We have now more sophisticated statistical procedures which are applicable to molecular and morphological data.
As you cite Mayr, the term 'race' is nowadays restricted (or synonymized) to 'geographic variation'. Hence, it is not applicable to A. citrinellus and A. labiatus as both are occuring in the very same lake. I guess you are talking about what modern biologists (ecologists, taxonomists) are calling a 'ecomorphological morph'. So, I think it is much better to use the current terminology instead of using a burden word like race.