Devil bait dispensers move Australia's bird flu bottleneck from the field to the vaccine bench

The reason a bait cannot simply carry the approved product is mechanical. Weakened or modified live viruses get into host cells readily enough to provoke an immune response, and a killed-virus formulation designed for a syringe does not, which the team gives as the reason it performs poorly in baits ^[10]. The nearest candidate on the right platform is American: a bird flu target inserted into a weakened human virus, which protected mice from bird flu and raised a similar immune response in cattle ^[1]. The authors note that design resembles both their devil tumour construct and the US rabies bait vaccine ^[2]. What that pair of results does not establish is protection in cattle. A comparable immune response is a measure of immunogenicity, not of survival, and neither a mouse nor a cow is a marsupial.

Scale is the other thing the rabies comparison brings with it. A billion baits since 1978 across 30 nations ^[3] averages roughly 21 million baits a year ^[18], near 700,000 per participating country per year ^[19], and that is the volume behind the rabies-free declarations ^[3]. The Tasmanian dispenser is a precision instrument by comparison, reading a microchip and a body temperature for each animal it serves and streaming the record home ^[6]. That is the right tool for a trial, and for settling a question the team is candid about not yet being able to answer, whether influenza would give a devil a fever ^[7]. Bulk coverage is a separate engineering problem, and the account does not report what fraction of a wild population the baits reach, only that several bait types were tested to raise uptake and that possums, quolls, pademelons and rats eat them as well ^[5].

The authors' case that a vaccine could arrive in useful time leans on COVID, where human testing began about two months after the novel coronavirus was sequenced ^[17]. Testing is not approval, and their own verdict on getting a wildlife bird flu vaccine approved within this outbreak is that it would be very challenging, though not impossible ^[21]. Meanwhile the delivery side is close to ready: with a suitable vaccine, they expect devils and other scavenging mammals could be covered quickly, and possibly land-scavenging bird colonies such as brown skuas, the first species found with bird flu in Australia ^[20].

Heard Island argues against assuming that vulnerability follows exposure. Six other bird and mammal species there were found infected, yet the mass die-off was confined to elephant seals, and one Australian fur seal has since died after infection ^[15]. That is the pattern behind the authors' admission that they cannot rank Australian species by risk in advance ^[12], and it favours spending on platforms that can be aimed at whichever species turns out to need them.