Keeping for one generation only — a parasite made only of queens, which kills the host queen it replaces and is reared by the workers she leaves behind
R959 closed with a resident that kept everything by keeping its host queen alive: Temnothorax minutissimus, workerless, accepted without violence, living its whole life inside a Temnothorax curvispinosus colony because that queen’s workers were the workforce it would need for as long as it lived. This page takes the turn it pointed to, to the relative that can afford the opposite move. Temnothorax kinomurai, a very rare inquiline from Japan, is also workerless — but in 2026 it was shown to have gone one step further. It has no males either. Every individual of the species is a queen.
First, the combination itself is the news. A typical ant colony divides life among reproductive queens, non-reproductive female workers, and males that die soon after mating. Entomologists already knew the pieces separately: some socially parasitic inquilines have lost the worker caste, and a few parthenogenetic ants have lost males. Hamaguchi and colleagues, reporting in Current Biology in February 2026, present kinomurai as the first ant known to have lost both at once — a species consisting exclusively of queens. It is known from only nine locations in Japan.
Second, the suspicion is old and the proof is new. Thirty-one mixed colonies of kinomurai and its host, Temnothorax makora, gathered between 1978 and 2020 in Honshu and Shikoku, each held a few dozen host workers alongside one to three parasite queens — some in the initially winged, gynomorphic form, others in a wingless, intermorphic form, a polymorphism of queens. None of those colonies contained a worker or a male that could be assigned to kinomurai, and forty years of informal captive rearing pointed the same way. To test it properly the team gathered six more colonies in 2022 to 2024, from Gifu and Kyoto, and reared their brood in laboratory nest boxes. The brood became 43 offspring, 24 gynomorphic and 19 intermorphic — every one a queen, no worker and no male among them. Examination of the genitals of 42 pupae also ruled out males disguised in queen-like form.
Third, the takeover deletes the engine minutissimus preserved. A young kinomurai queen usurps a small makora nest and, in the 2026 paper’s account, eliminates the host queen and some of the workers by stinging them. The survivors are then the whole workforce: they rear her offspring in a nest whose own queen will lay no more eggs. That is why the killing is affordable here and unthinkable for R959’s resident. A parasite that clones its daughters needs host workers for a single bout of rearing, not a living host queen renewing them for a lifetime. The workforce it inherits is therefore finite and diminishing from the day it wins. Popular reporting adds a separate, earlier field observation in which host workers themselves killed their own queen after an invasion; that detail belongs to those earlier observations, not to the mechanism the 2026 paper itself states, and the two should not be merged into one method.
Fourth, no mating is involved at all. The 43 laboratory queens were known to be unmated, and they were offered host colonies or groups of host workers and brood under controlled conditions. Only seven of them — three gynomorphic and four intermorphic — survived to produce offspring at all, a failure rate the authors note is typical for parasitic founding; the rest were killed by workers or died outside a host nest. The seven that succeeded laid unfertilised eggs that developed into 57 more queens, 41 gynomorphic and 16 intermorphic, and again there was no worker and no male. Dissection fits the story without overstating it: the sperm stores of kinomurai queens are much smaller than those of host queens and look rudimentary, and no sperm was ever found in them, while their ovaries carry fewer ovarioles than a host queen’s and more than a host worker’s. In the authors’ summary, in my own words: workerless parasitism combined with thelytokous parthenogenesis, females produced from eggs no male has fertilised, each queen able to produce both queen forms.
Last, it probably did not come from its host. It would be tidy to imagine so extreme a parasite budding directly off makora. The team’s preliminary family tree does not support that. Instead it places kinomurai closer to an even rarer and less studied slave-making species, Temnothorax bikara, which parasitises Temnothorax spinosior, a sibling species of the host. If a fuller analysis confirms that placement, the suggestion is that kinomurai arrived at workerlessness from slave-making ancestors that lost their workers, a route already demonstrated in two other lineages of the genus. Why losing workers is favoured at all is, the authors are plain, not yet fully understood; they link it in general to shifts away from outbreeding, strongly female-biased production, and the saving that comes from not investing in males when a queen is intolerant of the host queen and must rear her brood on a small, shrinking pool of borrowed workers. The arithmetic cuts both ways: every daughter is a potential founder, a hundred chances where a sexual queen’s brood would be divided, and yet in the laboratory most of those chances still ended at the door of the nest they tried to take.
A walk through a keeping that spends its engine. Five stages: a small host nest; a young queen usurps it; the host queen is gone and the survivors stay; cloned daughters are reared by a workforce that cannot be renewed; and the next usurpation, which most daughters will fail. Schematic, compressed from the published accounts.
Walk it in order.
Schematic only: the stages compress a field record spanning four decades, a laboratory rearing and a laboratory founding test, so no single nest is promised this exact sequence. Labelled cautions: do not give kinomurai a worker or a male, even as a rarity — the point of the 2026 work is that none appeared in 31 field colonies, in 43 laboratory offspring, or in 57 offspring of unmated queens. Do not merge the paper’s stinging queen with the workers-killing-their-own-queen observation reported from earlier field work; they are different accounts. Do not present the species as evolved directly from T. makora — the preliminary phylogeny points away from the host, towards T. bikara, and is labelled preliminary. Do not give it a second host past T. makora, or a range past the nine Japanese locations the paper reports, with field mixed colonies from Honshu and Shikoku. Do not read the seven-from-43 founding result as a field success rate — it was a controlled laboratory opportunity. And do not use the word clone to claim more than the work shows: unfertilised eggs developing into daughters, empty and reduced sperm stores, and no males produced, which is what thelytoky means here.