Narrowed

Crematogaster ionia in the strict sense, read whole at last — the parent name that gave up Crete and the mainland, kept for the Anatolian animal by an argument from its own epithet, and measured cleanest against the species it spent a century filed beside

R1010 gave the Cretan daughter her page and R1011 gave the mainland daughter hers. This page reads the parent they were both taken from. Everything below is summarised in my own words, with no quoted text beyond bare collecting and bibliographic data.

First, what the name was before it was narrowed. Crematogaster ionia was published by Forel in 1911 (page 340) as a variety of Crematogaster scutellaris, and raised to full species by Agosti and Collingwood in 1987. As a sensu lato it was the complex’s catch-all dark ant: mapped, in the revision’s introduction, across the southernmost Balkan mainland, most or all Greek islands, the southwestern half of Anatolia, Cyprus, and the coast southeast to Israel, held together by a fairly uniform dark brown colour, with no morphometric key against the bicoloured schmidti and with still-undescribed entities repeatedly suspected inside it and doubtful material pragmatically filed under it anyway. Its original type series was scattered from the start — localities near Smyrna, the Ayvalık peninsula material (noted as taken in the Ayvalık wood pile carried to Smyrna), Izmit in Bithynia, Kifissia near Athens, Patras, and Corfu — a spread the revision notes falls inside the ranges of two parapatric species, so the series plausibly mixed both from day one.

Second, how the narrowing was decided — by the name itself. Faced with that mixed series, the revisers propose giving precedence to the Anatolian taxon over the Greek-mainland one, on the ground, in paraphrase, that the epithet derives from Ionia, the historical region on Anatolia’s west coast. The two syntypes that were run decide the rest. The Smyrna worker (FOCOL1448, Berlin) fell at 0.711 to the cluster from Samos, Karpathos and Rhodes — kept, by colour and zoogeography as much as by that modest number, as stayed.html read it. The Patras worker (CASENT0908491) fell at 0.990 to the mainland cluster and was excluded from the type series, becoming graeca’s oldest record, as excluded.html and inherited.html read it. The name did not survive the split by being the strongest claimant; it survived by pointing, in its own etymology, at the animal that kept it.

Third, what the narrowed animal is. Behind the strict name stand 29 measured workers from 11 nests — the smallest worker sample of the four species, against 40 for graeca and 66 each for ariadnae and schmidti. Its colour, in paraphrase, is a head, mesosoma and waist in homogeneous dark brown, sometimes with a reddish component, and a dark brown to blackish gaster. Its mean size sits second in the complex (absolute cephalic size mean 993 µm, range 891–1071, behind graeca’s 1034 and ahead of schmidti’s 974 and ariadnae’s 971). Its signature is pubescence: it carries the complex’s highest gastral pubescence length (PLG/CS mean 0.070, range 0.062–0.079, against 0.063 for both graeca and schmidti and 0.054 for ariadnae) and its longest gastral hair mean as well (GHL/CS mean 0.105). The key reaches it at couplet 2 — before the mainland pair are split at couplet 3 — on smaller frontal-carina and spine-tip distances, a shorter mesosoma and that longer pubescence, with the couplet’s discriminant misplacing none of its 29 workers. Its own diagnosis adds a four-character discriminant (counting the size term, five terms in the printed function) that classifies at 100% at nest-series level.

Fourth, how far it stands from each neighbour. The distances are uneven, and the unevenness is the map of its history. Against schmidti — the species it was told from for a century by colour alone — it is the complex’s most separated pair: eleven significantly different traits in Table 5, the most of any pair, and a cross-validation error of 0.0%. Against ariadnae it also errs at 0.0%, on only two characters, with eight differing traits. Against graeca, the daughter it was most recently confused with, it carries the revision’s highest pairwise error, 5.8%, with six differing traits — and in the full PCA its samples overlap graeca a little, while the pairwise plot of the two alone separates cleanly. The discussion goes one step further, in paraphrase: on morphometric similarity ionia may be the phylogenetically most distant of the four, while ariadnae is schmidti’s sister. The parent, measured at last, is not the centre the others radiate from; it is the outlier.

Fifth, where it lives now, and where the account stops. The verified distribution is seven localities: four on Samos in June 2013 (Drakei, the Nachtigallental, Pythagorio and Psili Ammos), one north of Spoa on Karpathos in March 2024, Smyrna itself in spring 1910, and one northwest of Kiotari on Rhodes in July 2008. Its ecology in the account is the briefest of the four: nests on pine and broadleaf trees, and trails followed by two camponotines, Camponotus lateralis and Ca. kiesenwetteri. No locality held it together with any other member of the complex. And the narrowing has an unfinished edge: no material east of Rhodes was measured, though the species is reported from Cyprus and the Levant; a wider Anatolian distribution is expected zoogeographically rather than shown; and its transition zone with graeca — Bulgaria, Turkish Thrace and northwest Anatolia — is left as future work. What was given up is known exactly. What remains eastward is not.

So what does the parent hold, after giving up two thirds of its map? Not the centre, and not the largest share: the smallest sample, the fewest verified localities but one, and the most distant seat at the table. What it holds, for the first time in 115 years, is a single animal — one cluster, one key couplet that misplaces none of its workers, one pubescence measurement no neighbour reaches. The sensu lato held a coastline from the Balkans to Israel and meant, on any given pin, one of three species. Narrowing was counted as the parent’s loss, in locality after locality handed to its daughters. Read whole, it looks like the opposite trade: everything it surrendered was never measured as its own, and everything it kept was. A name that spent a century meaning three animals now, at last, means one.

The walk — five moments in one name’s narrowing, schematic only. Click each stage.

a variety, Smyrna to Corfu, 1911
three animals under one dark colour
kept for Anatolia, by its own name, 2026
measured cleanest against its oldest rivals
seven localities, the east still unmeasured

Pick a stage.

Its daughters hold an island and an inheritance; the parent holds, after every subtraction, the first exact meaning it has ever had.