The public of the Murray valley in the summer of 1951 did not invent a connection between two diseases. It noticed one, and got the mechanism wrong. This piece is about the human disease, where it had come from, how the Walter and Eliza Hall Institute pinned it down, and why it turned out to be riding the same mosquito as the rabbits.

X disease

In the late summer of 1917, and again in 1918, an encephalitis of a kind nobody could name broke out across the inland of eastern Australia, in New South Wales, Victoria and Queensland. It began suddenly with fever, headache and drowsiness and went rapidly to stupor, paralysis and often death, and it took children more than adults. The doctors who saw it called it Australian X disease because they could call it nothing else. Gray Anderson, going through the records in 1954, knew of 110 cases in 1917 and seventy-five in 1918 and thought there must have been others never published; Burnet’s round figures were sixty and eighty. About half of those it touched died. In 1918 Cleland, Bradley and Campbell in Sydney took brain material from fatal cases and produced a fatal encephalitis in monkeys and sheep by injecting it into their brains, before the days of mouse inoculation or the chick embryo, and before anyone could keep a virus alive long enough to send it anywhere. Burnet’s judgement in 1952 was that this was the first recognition of what are now called the insect-borne encephalitides, and the first isolation of one of their viruses, and that the published descriptions and the surviving histological sections matched what his institute was seeing in every respect.

There was a smaller outbreak in 1925, about a dozen cases, ten of them at Broken Hill. Then nothing for twenty-six years.

1951

The first case of the new outbreak fell ill in December 1950 in south-western New South Wales. Thirteen followed in January 1951, twenty-seven in February, two in March and two in April: forty-five reported severe cases, thirty-four of them in northern Victoria, ten in western New South Wales and one in South Australia, with, Anderson thought, probably others in the same region that went unreported. The patients were between seven weeks and sixty-nine years old; twenty-eight of the forty-five, 62 per cent, were under fifteen, and thirty-one were male. The disease came on suddenly with malaise, headache, fever, lethargy, vomiting, giddiness and a stiff neck, and moved fast: by the time a patient reached hospital he was often semi-comatose and making involuntary purposeless movements, with patches of paralysis that made swallowing or breathing difficult. Death or recovery usually came within two weeks. Recovery was not always complete. Nineteen of the forty-five died, 42 per cent. In infants the course was quicker and convulsions were common; two older children broke out in a rash, which recalled three patients of 1917 and 1918. Many adults in the valley came to their doctors that summer with mild “neurological” symptoms that nobody could assign with confidence to the virus or to anything else.

The diagnosis of each case was clinical. Laboratory work was done on thirty-eight of the forty-five, and in seven of them the virus itself confirmed it. Macfarlane Burnet had put two of his staff on the outbreak as soon as it was reported, Anderson on the epidemiology and Eric French on the virus, and French grew it from the brains of fatal cases, using inoculation onto the chorioallantoic membrane of the chick embryo, which in at least two instances succeeded where mouse inoculation failed. It behaved in experimental animals as the agent of 1918 had behaved in monkeys and sheep. It fell clearly into the group of mosquito-borne encephalitis viruses that includes Japanese B, St Louis and West Nile, with a standing equal to each of them. They called it Murray Valley encephalitis. Its type strain was soon being studied in America as well as in Melbourne.

The valley and the flood

Burnet gave the American Public Health Association a map to think with: the Murray–Darling basin, which covers all of the south-eastern quarter of the continent but the coastal fringe, with the Darling’s headwaters in sub-tropical summer-rainfall country to the north and the Murray fed by winter rain to the south. Two towns mattered to his argument. Broken Hill, the mining city, had been the centre of the 1925 outbreak. Mildura, the irrigation town where the Darling meets the Murray, had a dozen clinical cases in 1951 in a population of about twenty thousand.

Serology turned the dozen into something much larger. Among adults bled in one section of the district round the confluence, thirty per cent had complement-fixing antibody to the virus, which, since that antibody fades, meant recent infection; eighty per cent of domestic fowls on farms within a mile of the river had neutralising antibody, and fifteen per cent of those further off; horses, dogs and wild water birds, cormorants, herons, ducks and swans, showed the same. Infection that could be found in the blood had been, Burnet reckoned, five hundred to a thousand times as common as infection that reached a hospital. “We can feel certain,” he said, “that an extremely high mosquito infection rate must have occurred at some period in the season.”

Then the institute planned for a second season like the first, and it did not come. In the summer of 1951 to 1952 the work was extended: a three-month visit from William Reeves of the University of California, a very large collection of mosquitoes tested for virus, cormorant rookeries watched through the spring for infection in nestlings, sentinel fowls posted at strategic points in the Mildura district, and large numbers of volunteers bled before and after the summer. Nothing. “No evidence whatever had been obtained of the presence of virus in the Mildura area,” Burnet reported, and on the scale of the work he was prepared to go further and say there had been none. The virus could not overwinter in the valley. It was not endemic there. It came in from somewhere else when the conditions were right and went away again.

The conditions were rain. Anderson and Charles Miles, independently, went through the rainfall records and reached the same conclusion: in the year before each epidemic, 1917, 1918, 1925 and 1951, there had been abnormally heavy winter rain in western New South Wales and a heavy excess of early-summer rain, in November, over the headwaters of the Darling, with the November the more important. Since 1913 only five Novembers had been wet enough to flood, 1916, 1917, 1924, 1933 and 1950, and all but 1933 were followed by encephalitis in the valley. The second half of 1950 had brought almost unprecedented rain over Queensland and New South Wales; 1951 had been a normal year. Burnet’s speculation was that the virus lived permanently somewhere in the tropics, perhaps around the Gulf of Carpentaria, in birds and the mosquitoes that bite them, and came south down the rivers in a wet year as infected birds moved and the mosquitoes of each new stretch of water picked it up, a hundred miles at a time.

Strictly, he admitted, nobody had yet proved that the virus was carried by mosquitoes at all. But the circumstantial evidence was such that everyone on the investigation would have named one species as almost certainly the principal vector: Culex annulirostris, “a river bottom mosquito which feeds freely on water birds and rabbits.” It was the mosquito that Ken Myers had found dominating the Murray flats in January and February 1951, 88 per cent of the larvae he dipped from the lagoons, and the mosquito that Roman Mykytowycz had ground up to give a laboratory rabbit myxomatosis. The two viruses were unrelated. The flood that bred their carrier was one flood.

Parliament

The question stayed in the House of Representatives after the scare had passed. On 11 October 1951, the day Casey named the three men who had injected themselves, the member for Darling, Joe Clark, interjected that “the mosquitoes are spreading myxomatosis and encephalitis,” and Casey, who said the opinion was “not in accordance with the views of our most prominent authorities,” took the question on notice. The considered answer came from the Prime Minister on 22 November: CSIRO was “not itself taking any steps to increase mosquitoes in Australia with a view to promoting the spread of myxomatosis”; a campaign to eradicate mosquitoes in the Murray valley was a matter for the States; and “there is no direct proof that the mosquito is responsible for the spread of encephalitis although the available evidence supports this theory.” Clark had the mechanism half right, which was more than the rumour had managed, and the authorities had moved, in eight months, from “no connexion whatsoever” to “the available evidence supports this theory.” Both statements were true. They were about different things.

Sources

• S. G. Anderson, “Murray Valley encephalitis and Australian X disease,” Journal of Hygiene 52 (4), 447–468, December 1954. The outbreaks of 1917, 1918 and 1925, Cleland’s transmissions, and the 1951 epidemic in detail: the forty-five cases by month, State, age and sex, the nineteen deaths, the clinical course, and the laboratory confirmation. doi:10.1017/S0022172400036925
• Sir Frank Macfarlane Burnet, “Murray Valley Encephalitis,” an address to the American Public Health Association, Cleveland, 22 October 1952, American Journal of Public Health 42 (12), 1519–1521. The round figures, the map, Mildura’s dozen cases, the serology in people, fowls and birds, the 1951 to 1952 programme and its null result, the rainfall analysis, the five Novembers, the Gulf of Carpentaria, and Culex annulirostris. archive.org
• K. Myers, “Studies in the epidemiology of infectious myxomatosis of rabbits. II,” Journal of Hygiene 52 (1), 47–59, 1954, for the mosquito fauna of the Murray flats in 1951. doi:10.1017/S0022172400027248
• Commonwealth of Australia, Parliamentary Debates, House of Representatives, 11 October 1951 (Clark and Casey) and 22 November 1951 (Menzies’s answer). 11 October 1951, 22 November 1951