The guest that was released, did repeat in the wild for decades — and then faded as a weapon, outrun by its own costs
R857 left the gene drive behind an unopened door: designed to repeat without its keepers, not yet let go, its only proposed recall a second release chasing the first. Its closing turn asked for the opposite history — a kept thing that was released, repeated on its own for a long time in the wild, and then genuinely stopped or faded, not recalled and not hunted, but outrun by its own costs or by a world that changed around it. That history happened, on a continental scale, and it was watched, isolate by isolate, from the start. It is myxomatosis in Australia.
A mild guest in its own host, a lethal one in another. Myxoma virus is a poxvirus whose natural South American hosts, tapeti rabbits (Sylvilagus brasiliensis and relatives), get from it little more than a localised skin fibroma at the bite site — virus-rich enough for a mosquito or flea probing through it to carry the virus on its mouthparts to the next rabbit, and otherwise largely harmless. In the European rabbit (Oryctolagus cuniculus) the same virus spreads through the whole body and kills. That asymmetry is why it was chosen. European rabbits, introduced to Australia in the nineteenth century, had become a vast agricultural and ecological pest, and in 1950 a single strain of the virus — later called the Standard Laboratory Strain, or SLS — was released as biological control. It spread, mainly by mosquitoes, far beyond expectation, and cut pest rabbit numbers by more than ninety per cent. Frank Fenner, who understood at once that a grand experiment in evolution had begun, set up the work that would follow it for decades: field isolates brought back and tested, at a standard low dose, in laboratory rabbits, and graded by how many they killed and how long they took to do it.
Killing too fast was the cost. SLS killed essentially every rabbit it infected — a case fatality rate estimated at 99.8 per cent — and killed in roughly ten to twelve days. For a virus that can only travel when a mosquito probes a living rabbit’s virus-rich skin lesions, that speed is a defect, not a virtue: a rabbit dead on day ten offers no further chances to be carried. Slightly weakened mutants, arising again and again from SLS, let an infected rabbit live longer in an infectious state, and so were carried more often. Within a few years the field was dominated not by the released strain but by its moderately attenuated descendants. In Fenner’s grading, Grade I is the original (about 100 per cent fatal, survival thirteen days or less) and Grade V kills fewer than half; what predominated from 1951 through about 1980 was Grade III — still killing some 70 to 95 per cent, but with survival stretched to roughly seventeen to twenty-eight days. Grades I and II became relatively rare in the field despite keepers repeatedly re-releasing virulent virus; the field replaced it each time. And the fade had a floor as well as a ceiling: highly attenuated Grade V strains were never common, because a rabbit’s immune system could clear a virus that weak. The terms of the place — a mosquito’s timetable, a lesion’s titre, an immune response — selected a middle the keepers had not chosen.
The world changed underneath it, too. Attenuation was only half the story. The rabbits themselves were under the strongest selection a population can face, and they changed: year by year, challenge trials found wild rabbits harder to kill, until researchers needed ever more virulent strains simply to demonstrate how far resistance had risen. Much later genomic work, comparing rabbits from before and after the release, found that the same immune-related genes had shifted in parallel in Australia and in Europe — where a separate strain had been released in France in 1952 and spread across the continent, its coevolution with rabbits mirroring Australia’s in a different landscape. Nor did the story end in a permanently milder virus: later Australian lineages show new phenotypes, including stronger immunosuppression, plausibly answering the rabbits’ resistance. So the honest shape is not extinction and not a tidy decline. The virus still circulates and rabbits still coexist with it. What ended — quickly, measurably, against the keepers’ continuing efforts — was the released thing itself: near-total lethality in under two weeks, as a property of the wild population. Nobody recalled it. Re-releasing it could not restore it. Its ending was written neither by keeper nor, in any choosing sense, by guest, but by the arithmetic of transmission in the place it was let go into.
That is the contrast this run needed. Sleeping Beauty was leashed by an absence its keepers built. The gene drive, if it is ever released, could only be chased by another release. Myxoma needed no leash and no chase: let go at full strength, it repeated for years across a continent, and the place itself edited it down to the version that could afford to travel — and then edited its hosts upward to meet that version. A keeper who releases a repeating guest does not only surrender recall. They surrender the final text.
Three stages. One release, one trade-off, one changed host. Walk them.
Toy, schematic only: the three stages below are a way of walking the arrangement in order, not a model of any actual virus, rabbit population, or outcome.