The name-bearing type of Crematogaster schmidti itself, in Vienna — too damaged to run in a revision built on numbers, matched to its own species by colour and place alone, while a stand-in body scored 0.945 in its stead
R1004 followed the type that scored lowest and stayed anyway, at 0.620, and ended on the one that has no score at all. This page follows that one alone. Everything below is summarised in my own words, with no quoted text beyond bare collecting and bibliographic data.
First, what the specimen is. It is the type worker of schmidti, the worker the name itself rests on. Its home locality is given as the Vipava Valley, in what was then Krain, in modern Slovenia, and the revision’s distribution list credits its collecting to F. Schmidt. It is housed in Vienna, in the NHMW, and its AntWeb pictures carry the identifier CASENT0919655. Its labels, in paraphrase, tie it to Mayr’s collection and to Mayr’s own determination of it under the name’s first genus, Acrocoelia: Mayr published schmidti in 1853, at page 149, in that genus, and moved it into Crematogaster two years later.
Second, why there is no number. The revision is a machine for producing numbers: 201 workers from 68 nests, fifteen traits apiece, four clusters, and six type workers run as wild cards whose winning scores span 0.990 down to 0.620. This worker is not one of the six. In the schmidti account the revisers state, in paraphrase, that it was not available for morphometric investigation because of damage. Pictures of it exist — the account points to them — but pictures were not enough here, and that distinction matters, because two of its neighbours in the same account were measured from their AntWeb pictures alone. Damage, not absence, is what kept this body out of the table: the specimen is present, photographed, labelled, housed — and unmeasurable.
Third, what matched it instead. Two witnesses, the same two that kept doing the work all through this sequence, but here they work entirely alone. Colour: the revision’s redescription gives schmidti a typically reddish head, mesosoma and waist against a dark brown to blackish gaster, and its diagnosis sets that bicoloured pattern against the other three species’ usually uniform dark colouring. Place: the Vipava Valley is not merely inside schmidti’s range, it is the northern, Balkan heart of it — the species the revision maps farthest north of the four, from Slovenia southward. Set a damaged, bicoloured worker down in its own type locality and there is, in practice, nowhere else for it to go. On those two grounds, in paraphrase, the revisers found that it matches schmidti strongly. No probability is attached. The sentence that carries the central name’s own anchor is the one sentence in the type accounts with nothing after the decimal point.
Fourth, the stand-in. The number the type could not supply was supplied, one step to the side, by another body. A further worker from Dalmatia — Ragusa, modern Dubrovnik, collected by Reitter — is housed in Geneva as CASENT0908487, labelled as a cotype, and its label records, in paraphrase, that Mayr gave it and held it to be absolutely identical to the type. Measured digitally from its AntWeb images, it ran as a wild card at 0.945 for schmidti. So the cluster does touch the name’s neighbourhood with a real score — but the score belongs to a worker whose claim to stand for the type is a sentence on a label, written by the name’s own author. The anchor’s anchor is an assertion of identity, and the revision lets it stand, printed plainly beside the missing measurement it replaces.
Fifth, what this type actually holds. Start with what it does not hold up. The four-species structure does not lean on it: the clusters were formed from the 201 measured workers, and they would stand unchanged if this drawer in Vienna were empty. What the type holds is the attachment between that structure and a name from 1853 — and that name has needed holding. In Mayr’s own 1855 work schmidti was sunk under scutellaris; Emery raised it in 1891; Forel demoted it to a subspecies in 1904; Collingwood raised it again in 1961, and that same year Fromantin and Soulié sank it again; it has stood at species rank since Arnol’di and Dlussky in 1978. Six changes of rank in a century and a quarter, every one of them a judgement about a name whose bearer, all along, was a worker too damaged, in the end, to be measured by the first revision that ever tried. The name survived its demotions the way it survives this revision: somebody looked at the animal, or at where it came from, and said that it matched.
So what does a type hold when it cannot be run? It holds continuity, and nothing more — but continuity is the whole job. A wild-card score can show that a body belongs to a cluster; it cannot show that the cluster is entitled to an old name. Entitlement runs through the type, and the type runs, here, on being looked at. At 0.990 a type was excluded; at 0.711 and 0.620 types were kept by arguments built around their numbers. At no number at all, the keeping is the argument, undiluted: a colour pattern, a valley, and a stand-in at 0.945 whose label says its author found it identical. That is not a weaker version of the same holding. It is the holding with the measurement stripped away, showing what was underneath it in every other case — that a name is kept, finally, by people deciding that a place and a pattern are enough, and writing the decision down where the next revision can find it.
The walk — five moments in one worker’s working life, schematic only. Click each stage.
Pick a stage.
The excluded worker had a number and lost its name; the lowest worker had the worst number and kept its synonymy. This one has no number, and it is the name itself that it keeps — proof, if the sequence needed it, that the numbers were never what a type is for. A type is for being returned to. Even damaged, even only looked at, it is the one body the name can always be brought back to, and brought back it has been, for a hundred and seventy-three years.