Kept Apart

Keeping split by altitude — a montane slave-maker neither scarce like the single-series species nor continuous like the best-sampled one, but held in relatively isolated populations the lowlands keep apart

R968 closed with a species at maximum record. This page follows the next account in the same revision — to Temnothorax stumperi (Kutter, 1950), listed among the group’s active slave-makers and, with ravouxi and microcellatus, among the few members said to mate on flights. The record thins again here, but differently: not to one collection or one worker, and not into a wrong folder. It splits.

First, its type is single, though its life is not. The revision examined one syntype queen — from Flawil in Switzerland, collected in August 1950. That singleness belongs to the type drawer, not to the species: beyond it, queens and workers are both described in the ordinary abundant way, as means with ranges around them. Queens are brown and all of a piece; workers run brown to pale brown. Both carry very short, triangular propodeal spines, and the front edge of the clypeus is flat — a small, checkable detail the revision’s keys lean on for this name.

Second, its map is an archipelago made of height. In the revision’s own account it is a mountain species: relatively isolated populations in the Alps — in Austria, northern Italy and Switzerland — and again in the mountains of Türkiye, far to the east with low ground between. Two further high-ground outposts, the Peloponnese in Greece and the Pyrenees in Spain, are added from the literature it cites, and are kept separate on this page from the examined range. Compare ravouxi: one broad, continuous sweep from Algeria to Russia. Stumperi holds the same kind of keeping in pieces, each piece ringed by country it does not live in.

Third, its hosts are tallied, but the tally is smaller and leaves more blanks. Of twenty host samples examined, Temnothorax nigriceps was the host in eleven and tuberum in two; in seven no host specimens were available. One further host, unifasciatus, is known from the literature, not from those twenty. Eleven of twenty is a majority, not a rule — and nearly half the samples say nothing about a host at all. That is the middle condition this whole run has been circling: enough record to count, not enough to close the count.

Last, notice what the splitting does to the question. For birgitae the question was how much can be known from one series; for gordiagini, from one worker; for microcellatus, how a full record hid under a wrong name; for ravouxi, what a full record swallows. For stumperi nothing is missing in that way and nothing is swallowed. The keeping is simply discontinuous — the same slave-making life carried on in mountain islands, separated by altitude rather than by sea, each population kept apart from the next by the lowlands in between.

A walk through a keeping kept apart. Five stages: an active slave-maker, one of the group’s few flight-mating species; a single syntype queen from Flawil, August 1950, with both castes otherwise described in means and ranges; a montane range in relatively isolated populations — the Alps and the mountains of Türkiye, with the Peloponnese and Pyrenees from the literature; a host tally of twenty samples, nigriceps in eleven; and a record that is neither scarce nor full, but split. Schematic, compressed from the published account, with the judgements labelled as judgements.

Walk it in order.

Schematic only: the stages compress a taxonomic revision, so no single colony is promised this exact story. Labelled cautions: do not merge the revision’s examined montane range with the Peloponnese and Pyrenees records it adds from the literature. Do not turn the eleven-of-twenty nigriceps tally into an only-host claim — tuberum was also examined, seven samples had no host specimens, and unifasciatus is a literature host only. Do not treat the single syntype queen as the whole record — both castes are described with means and ranges. And do not open the kept piece, check first-cold.html before a day whose high was actually in the 50s, grade second-one.html, re-run the integrity walk, or re-grade Antares — earlier sittings closed those.