The appearance of the newly updated eBird taxonomy reminds me today that the venerable song sparrow name samuelis, dating to 1858, has been appearing recently in an emended form, samuelsis. I believe that the first such “correction” was made by Dickinson and Christidis in their fourth edition of the Howard and Moore Checklist, where they write that their change is “based on ‘Mr. Samuels’ mentioned in the original” description, by Spencer F. Baird. Denis Lepage’s “Avibase” followed suit a year later, and the emendation is adopted, too, in the HBW/Birdlife Checklist and in the most recent IOC list, which cites Howard and Moore explicitly in the matter of “internal information” in Baird’s account.
Dickinson and Christidis are, as usual, absolutely right on the facts. The co-types, both taken on May 9, 1856, in Petaluma, were shot by the visiting Massachusetts collector Emanuel Samuels, and it was his name that Baird assigned to them.
To my imperfect knowledge, Herbert G. Deignan was the first to take umbrage at Baird’s barbarous samuelis, which Deignan snorts was “probably used by Baird as euphonically preferable to the more proper samuelsi; unfortunately, as it stands it is not obviously dedicated to Emanuel Samuels at all, but apparently to some imaginary Samuel.” Deignan’s objection is grammatical: Samuelis is the genitive of “Samuel,” while “Samuels” would yield samuelsi—if it happened to be a Latin noun.
In a sense, Deignan was right, though it is easy to argue that the English family name “Samuels” has its historical origins in an English genitive meaning “of Samuel,” which would translate into Latin as precisely the Bairdian samuelis. And it is equally easy to argue that Baird considered “Samuels” a noun of the third declension, which would drop that final -s, like lux/luc-, rex/reg-, and so on, and form its genitive in -is.
If we accept either of those as possibilities, it is by no means certain that Baird’s samuelis was an error. Instead, the name can be squeezed into the provisions of ICZN 31.1.1 as a name formed from a modern personal name that has been latinized. Poorly latinized, to be sure, but ICZN 32.5.1 lets Baird off the hook even for that: “Incorrect transliteration or latinization, or use of an inappropriate connecting vowel, are not to be considered inadvertent errors” that must be corrected.
Unfortunately for Baird and for the original spelling (whatever its origin, in error or innovation), we cannot simply dismiss samuelsis even if we class it (as I think we ought) as an incorrect subsequent spelling of samuelis. For the ICZN magically makes “an incorrect subsequent spelling … in prevailing usage and … attributed to the publication of the original spelling … a correct original spelling.” With prevalence defined by the Code as “that usage of the name which is adopted by at least a substantial majority of the most recent authors concerned with the relevant taxon,” it’s clear that we’re stuck with samuelsis, adopted as it now is by all the major world checklists.
Why, though, did our pii correctores contrive samuelsis rather than simply adopt Deignan’s suggestion of samuelsi? I cannot see how the neologism satisfies the requirement of the ICZN that a species name formed from a personal name must be formed either in accordance with the rules of Latin grammar—yielding samuelsii, if we fussily latinize the name to “Samuelsius”—or by the simple addition of a terminal -i—yielding Deignan’s preferred samuelsi. The currently prevailing form samuelsis seems to be pulled out of thin air; it is at least utterly unpredictable on the rules of zoological nomenclature.
Stay tuned. I’ll find out what happened.