The trials of 1950 produced one result that mattered, and it was a negative one: on every site the disease went in, spread a little, and died. A negative result is only worth having if the counting behind it is good, and the counting is the part of the work that Ken Myers was proudest of when he wrote it up in 1954. This piece is about the method.

The sites

Following the earlier work at Gunbower, on the Murray in Victoria, which had begun in January 1950, and alongside a large trial that Bernard Fennessy ran at Wymah in the foothills east of Albury, Myers opened four sites in the Corowa district in August 1950. He chose them to stand for three of the four kinds of country the rabbit lived in there. Two were stands of native pine, Callitris glauca, on light sandy soil at Balldale: site A, seventeen acres of warren with 150 warrens in it, and site B, a smaller stand of pine and box on sand, with twenty-four warrens in a little over an acre and a half. One was a rocky granite hill at Coreen, with twenty-six warrens in seventeen acres under sparse eucalypts. One was the old goldfield at Rutherglen, where the diggers had left mullock heaps in open cleared pasture, and where fourteen or fifteen warrens on thirteen acres held seven hundred rabbits in the first trial and nine hundred in the second. The fourth kind of country, the flats and frontage of the Murray itself with its billabongs and swamps, was not used, because the annual flood was rising when the trials began. It was the country where the disease would break out.

On every site the rabbits lived in warrens. There were none of the surface-living rabbits that harbour in hollow logs elsewhere in Australia, which would have made the counting harder.

Counting rabbits

Myers had noticed in earlier work that rabbits in an undisturbed population, when they first come up from underground in the early evening, spend twenty or thirty minutes on the surface or within a few yards of the warren. That was the window. Counts were made of the colonies on the surface of their warrens, or, where timber made that impossible, along a standard transect laid out on a large-scale map of the warrens to give the best coverage from the available cover, walked at a slow even pace. Every count on a site was made in precisely the same way, always in the early evening of a warm sunny day, and the observer watched for a few minutes first to be sure the animals had not been disturbed and were behaving normally.

Nobody knew the absolute number of rabbits on any site, so there was no direct check on the counts. What Myers could show was that they were steady. On successive favourable evenings at Balldale A in mid-August the counts were 289, 334 and 314; at Coreen in early September, 372, 412, 401 and 420; at Rutherglen in late August and early September, twelve counts between 652 and 747 with a mean of 702 and a coefficient of variation of 1.1 per cent. He set these beside the counts that Frank Tidswell had made in 1907, on Broughton Island off New South Wales, during the last large Australian experiment with a rabbit disease, when the New South Wales government had tried a pasteurella in a fenced population; Tidswell’s counts, made at longer intervals and in whatever weather, varied by 5.6 to 6.9 per cent. Myers took the lower variability of his own as the reward for the precautions.

He also explained why it mattered so much. “In our myxomatosis work,” he wrote, “it is necessary to distinguish clearly between a 95 % and 99.5 % mortality in a rabbit population, while the distinction between a 90 and 95 % kill is rarely of great importance.” Overstate a ninety per cent kill by five points and the count of survivors is wrong by a hundred per cent; confuse a 98 per cent kill with 99.5 and the error in the survivors is three hundred. Where the sites were small and the viewing exceptionally good, as at Rutherglen, he believed the counts came close to the absolute number of rabbits.

Inoculating

Rabbits for inoculation were taken by ferreting at Coreen and Balldale, and by digging out and live-trapping on the feeding grounds at Rutherglen. Each was injected under the skin with one millilitre of a glycerinated extract of the liver and lungs of a rabbit that had died of the disease, supplied by the New South Wales veterinary research station at Glenfield and descended from the strain B that Sir Charles Martin had used at Cambridge in the 1930s. The stock was checked at Balldale: three caged wild rabbits given a ten-thousandth dilution showed typical lesions on the seventh day and died on the ninth. Care was taken that the inoculated animals were scattered through the whole population, and the bare patch of fur at the injection site let the observers confirm afterwards that they had been. At Balldale and Coreen each inoculated rabbit was earmarked with a combination of punches that recorded its warren. At Rutherglen, where the counts were daily, individual rabbits could be recognised by size, colour, the burrow they used and the stage of their disease, and the daily sightings of sick animals were tabulated by letter, capitals for adults and small letters for kittens, so that the reader can follow rabbit A through six days of sickness in warren complex 1 until it was seen no more.

In the second Rutherglen trial, from the seventh day after the inoculations, when the inoculated rabbits showed the disease, the whole population was driven underground throughout the daylight hours, in the hope of increasing contact between the sick and the healthy.

Predators and insects

Nobody tried to measure the rate of predation directly. The sites were searched instead, daily at Rutherglen, for the signs. A fox left the carcass decapitated; the little eagle left plucks of fur; the feral cat left the pelt cleaned and turned inside out, and at Rutherglen and Coreen the cats carried their catches to lairs, where the pelts could be counted and removed on each visit. Barn owls were caught in the act by their regurgitated pellets, which John Calaby identified. Fresh scats and the smell of a fox were noted wherever they were met. Predation was actually seen only twice in the whole programme, both times at Rutherglen, both times by the little eagle. The reckoning was that at Rutherglen a cat and a pair of little eagles took two or three small kittens a day between them, mostly healthy, that a fox visited three times and killed one healthy adult, and that about ten kittens a week died above ground of coccidiosis. Over the fourteen weeks of the trial known predation and coccidiosis accounted for about four hundred kittens. Myxomatosis, caught naturally, killed about seventy rabbits.

Biting insects were counted by the number coming to bite the observer’s bare arm in a minute, and rabbits in the field were watched for insects on them. Later, mosquito larvae were sampled: fifteen minutes in any one breeding place, five on surface-resting larvae, five deep-sweeping with a net-covered scoop, and five collecting from the surface after the water had been muddied, which often drove up larvae the first two methods had missed. The only ectoparasite found on the rabbits was the louse, in small numbers on every animal examined.

What the counts said

The results on all the sites were, in Myers’s words, “essentially similar, the disease failing to gain momentum, and either dying out or approaching this end-point during the period of observation.” At Rutherglen the second trial was followed day by day. Sixty-six rabbits were inoculated on 27 October 1950. Four natural generations of the disease followed, each smaller than the last, with maximum numbers of infected rabbits of about thirty-five, then sixteen, twelve and eight, and the peaks eight or nine days apart, which fits what was known about when a sick rabbit becomes infective. The ratio of infection from generation to generation was 0.53, 0.46, 0.75 and 0.66. When intensive observation ceased on 15 December the population had grown from seven hundred to a thousand. Three coloured rabbits, one yellow and two piebald, were seen moving in and out under the fence by the burrows beneath it, so the enclosure was not quite closed.

The insects were counted too. Through August, September and October the only winged blood-feeder taken at Rutherglen was the day-biting mosquito Aedes theobaldi, in very low numbers; towards the end of November the first few Anopheles annulipes came biting in the evening. At Balldale the day-biting Aedes were dense, twenty to the arm in ten minutes, and did nothing. Myers’s conclusion, written with the knowledge of what came next, was that the Aedes were inefficient vectors, which was later confirmed, and that the Anopheles at Rutherglen had been below the density at which transmission takes off. The final trial had every condition that would favour spread by contact, by mouth or by breath: a very high density, no selective predation of the sick, and most of the sick staying in their warrens until they died. The ratio of infection still came out at about six-tenths. “As subsequent developments show, the one factor lacking was transmission by flying insects.”

Sources

• K. Myers, “Studies in the epidemiology of infectious myxomatosis of rabbits. II. Field experiments, August–November 1950, and the first epizootic of myxomatosis in the Riverine Plain of south-eastern Australia,” Journal of Hygiene 52 (1), 47–59, March 1954. Everything in this piece: the sites and their dimensions, the counting method and its variability against Tidswell’s, the remark on 95 and 99.5 per cent, the Glenfield extract and the caged rabbits, the earmarks and the lettered sightings, the driving underground, the signs of predators, the insect counts and the larval sampling, the coloured rabbits, and the four generations. doi:10.1017/S0022172400027248
• F. Tidswell, Report on the Broughton Island Experiments, New South Wales Legislative Assembly, 1907, and Virus for the Destruction of Rabbits, Department of Public Health, Sydney, 1908, as cited and tabulated by Myers; not read directly.