Cane toads, mosquito fish, mongoose — these are some of the images Australians may (or not) conjure at the mention of biocontrol.
To be fair, there have been some absolute cock-ups. And almost without exception, proper scientific procedure wasn't followed in their making.
Cane toads, for example, were released by an entomologist working for a sugar cane industry research body, and against the advice of other entomologists.
Next to no research was done on the cane toad's host specificity — whether the toad would stick to its intended prey, or go chasing greener pastures.
In fact, nobody actually tested whether the cane toad ate the cane beetle it was supposed to eradicate.
But what about when biocontrol is done right? We tend not to hear about the times it has gone well. And there are plenty of times where it's gone really well.
When it comes to weeds alone, Australia has completed 58 control programs, as of 2017.
Here are a few examples that demonstrate the power of biocontrols that are done right, and some of the challenges researchers face in finding suitable candidates.
Skeleton weed and the rust fungus
First recorded in Australia in 1917, skeleton weed was considered Australia's worst plant pest by the 1950s.
Although it looks fairly innocuous, it wrought economic damage on a huge scale, particularly in Australia's eastern wheatbelts.
"It's got these really wiry flower structures and ... the harvesting machine gets stuck on them," said Mariana Campos, CSIRO's team leader of ecosystem change and ecology.
"[The plant] contains latex as well, so it all gets gummy. Not only does it stop the harvester from actually working, it also contaminates the grain."
Wheat crop yields were cut by up to 80 per cent, farmland was abandoned, and estimated losses in the 1960s were around $30 million a year — more than $200 million in today's value.
In 1966, researchers from CSIRO began investigating possible pathogens in the skeleton weed's native range around the Mediterranean.
Scouring Mediterranean Europe east to Iran, they came up with an initial list of nine possibilities.
Then they spent several years testing for host specificity, which CSIRO's team leader in plant diagnostics Gavin Hunter said is the "core tenet" of biocontrol today.
"That is the bedrock upon which we operate as control researchers," Dr Hunter said.
"In terms of skeleton weed, I think they tested a total of [around] 57 non-target species against the candidate biocontrol agents to prove they were [species] specific."
With a laboratory eventually set up in Montpellier, France, they narrowed their list down to four possibilities. The first of those was released on the Australian east coast in 1971, followed by three more.
Three of the four became established and one in particular, a rust fungus called Puccinia chondrillina, decimated the weed. The two others, a mite and gall midge, worked to further weaken the plant.
In all, CSIRO estimates the program saved the Australian economy more than $1.4 billion, with a cost-to-benefit ratio of 112-to-one.
In Australia, there are three "biotypes" of skeleton weed — they're all the same species, but one has narrow leaves, another intermediate, and the third broad leaves.
The biocontrols were so specific that they only had a significant effect on the worst of the three, the narrow form.
All three types of skeleton weed still persist in Australia's east, albeit in much lower amounts.
But the west took a different approach, opting to suppress the weed through physical and chemical means rather than going all out on a biological control.
That program is ongoing in Western Australia, via an annual levy paid by grain producers. The WA Department of Primary Industries estimates that program costs around $3.4 million per year, at a cost-benefit ratio of one-to-three over 30 years.
Prickly pear and cactoblastis
Prickly pear was first introduced to Australia by Governor Phillip in 1788, for the purpose of setting up a cochineal dye industry.
At the time, red dye was valuable and was used, among other things, to create the distinctive red coats worn by the New South Wales Corps or Rum Corps.
The dye was extracted from a cochineal scale insect — Dactylopius coccus— which feeds on the spiky leaves of the cactus.
But the cochineal insect didn't take well to Australian conditions, and the dye industry floundered.
Not so the prickly pear, which was prized for its hardiness, bright flowers and edible fruit.
By 1840 it had found its way to Chinchilla in Queensland, and by 1920 it had taken over around 60 million acres of farmland, mostly in Queensland and New South Wales.
Arsenic was used to try to control the weed in such quantities that a mine was opened in Queensland to meet increased demand for the poison.
But the cactus continued advancing, swallowing more than 2 million acres a year at its peak.
Farms were abandoned, cash was offered for a solution (but never collected), and everything from flamethrowers to tanks and enlisting returned soldiers for a military assault was floated.
The eventual solution was nowhere near as dramatic but far more devastating — the combination of a moth and some scale insects.
Following the release of the cactus-eating Cactoblastis moth larvae in 1926, "within 10 years, the once dense-fields of common prickly pear lay rotting or had vanished completely", according to the CSIRO.
By 1932, 7 million hectares of land was reopened to farming.
Today, most of the credit goes to the Cactoblastis moth, but biocontrol researcher Michael Furlong of the University of Queensland says it was a team effort.
"The important thing is it wasn't just Cactoblastis ... there were cochineal insects — small-scale insects — introduced as well, and if it hadn't been for them, the Cactoblastis likely wouldn't have worked," Professor Furlong said.
"The cochineal insect was extremely important, and is still very prevalent in areas around western Queensland today."
Myxo, calici and rascally rabbits
Myxomatosis and calicivirus are probably the best known biocontrols to have been used in Australia.
But there was a long time before these viruses came onto the scene. Until then, building fences and "rabbit drives" were the only tools available.
Rabbit drives involved herding often thousands of rabbits at a time into enclosures, before clubbing them to death en masse.
Fence-building was a huge operation, and had some limited success, but weaknesses were invariably found, and rabbits made their way across.
Desperate times led to the attempted introduction (not by CSIRO) of mongoose as a biocontrol during the 1880s.
Around 1,000 mongoose were released in South Australia, Victoria and New South Wales, but thankfully all failed to take hold.
Mongoose today are considered among the world's most destructive invasive pests.
By 1887, New South Wales was offering a 25,000 pound reward for anyone who could come up with a novel idea to exterminate the pests.
In the 1890s, a virus that was killing rabbits in a lab in Uruguay was discovered.
But there wasn't much appetite to take a risk on releasing a virus here, and the idea was shelved.
It wasn't until 1950 that public sentiment, combined with research and political support, saw the introduction of myxomatosis.
Myxomatosis is a disease caused by the myxoma poxvirus, and spread by mosquitoes.
CSIRO vertebrate management systems group lead Tanja Strive said wildlife biologists don't really like to put a number on how many rabbits there were, because it's always fluctuated, but that best estimates suggest about half a billion in Australia at their peak.
"They talk about the 'grey blanket' before myxomatosis," Dr Strive said.
"[Myxomatosis] was slow to get going initially, but then there was good rainfall and suddenly there was an enormous number of mosquitoes.
"Then it really took off and had a massive knockdown of the rabbit population — over 90 per cent in some areas."
But the honeymoon was short-lived.
The longer rabbits lived, the longer they could pass on the virus, so evolutionary pressure was favouring more resistant rabbits and less virulent myxomatosis.
"Quickly, in two or three years, the effectiveness was starting to wane.
"By the mid-1980s, the numbers were not as high as pre-myxo, but pretty high again."
In the meantime, CSIRO had been searching for alternatives, and an emerging virus had appeared in rabbits in China.
Calicivirus, also known as rabbit haemorrhagic disease virus, attacks the internal organs of the animal and is usually fatal within 30 hours.
Following four years of laboratory testing, in the mid-1990s, CSIRO set up a quarantine research station on Wardang Island in South Australia, when the virus they were testing — calicivirus — escaped on feeding flies.
"Once that got off the island and into the almost completely naive Australian [rabbit] population, again it basically wiped out over 90 per cent of rabbits in some cases," Dr Strive said.
"So it was a raging success again, but over time, much slower than myxo, we've seen some resistance developing."
Which is inevitable in disease biocontrol.
Another new strain of calicivirus made its own way here around 2014, and another was deliberately released in 2017, both of which knocked down the population.
As well as saving the Australian economy billions, the rabbit biocontrol program has also had a knock-on effect on cat numbers, which has enabled some small native mammal species to recover.
CSIRO is currently researching new viruses in rabbits both here and overseas, and looking at ways to improve the impact of calicivirus and myxomatosis.
"We're pursuing a number of avenues," Dr Strive said.
"Ideally we'll have a new tool, or virus, or strategy ready every 10 to 15 years because what we don't want is for numbers to get away again."
'Palm killers' and Oryctes rhinoceros nudivirus
The coconut rhinoceros beetle or "palm killer" is an example of how biocontrol can be a moveable feast.
The coconut rhinoceros beetle was first detected in Samoa just over 100 years ago, far outside its native south-east Asian range, according to Professor Furlong.
"It's a large beetle that attacks coconut palms.
"It first got into the Pacific in 1909, devastating coconuts there, particularly in Samoa."
In the early 1960s, a virus was found in rhinoceros beetles in Malaysia and was introduced to Samoa with very promising early results.
"It's not immediately lethal [but] ... it reduces their overall fitness — how much females can move and fly, and it reduces the number of eggs they lay, and that leads to overall population suppression."
The Oryctes virus is also transmitted easily between beetles and spreads rapidly.
For years the virus played an integral role in suppressing beetle numbers across the Pacific.
But in the mid-2000s, the virus's efficacy seemed to be waning.
Attempts to infect beetles on Guam failed, leading to a hypothesis that the Guam beetles were immune, and the coconut rhinoceros beetle (Guam) type was named.
Similar beetles designated as the Guam type showed up in Papua New Guinea, Hawaii, and Palau, the Solomon Islands in 2015, and Vanuatu in 2019.
In response, countries like the Solomon Islands have declared local states of emergency and launched offensives to try to stymie the beetle's spread.
A worst-case scenario of 50 per cent palm losses, would devastate the local oil palm, copra and coconut oil industries, which provide 10 to 15 per cent or more GDP to some Pacific countries.
The beetle also attacks bananas, pineapple and other crops, leading to warnings for Australia to do more to help control the beetle for its own, and its neighbours' benefit.
But Kayvan Etebari, a molecular biologist at the University of Queensland, doesn't believe the Guam type is immune, or that the Guam type has spread across the Pacific.
"There is no scientific evidence to prove that particular [Guam] population is resistant," Dr Etebari said.
He is working to understand why the virus is no longer suppressing beetle numbers, and said he thinks the virus-insect relationship has changed.
"When we screen the beetles, most of them have the virus infection. So if this is the case the first question is, 'Why is the virus not killing the insect?'
"That's why I believe the virus and insect has changed to a more 'friendly' relationship.
"[But] that can change again. It's not going to stay the same for another 10 years."
In the meantime, he's looking for another virus: "But finding a novel virus is not an easy job."
He's also working with the Oryctes virus to see if it might be able to do the job the cane toad failed to do.
"We want to use the same virus to see if it can control the cane grub, which doesn't have any biological control agent at the moment."
