Page 320 - Zoo Animal Learning and Training

292 12 Training Animals so They Can Return to the Wild

What can we do to prevent northern quoll
VetBooks.ir extinctions? One promising approach is to inducing chemical component). The quolls
that consumed the toads became mildly ill,
use conditioned taste aversion (CTA) to train
serious toad poisoning. Next, we offered
quolls not to eat cane toads. CTA is a power- and thankfully, none showed any signs of
ful form of learning that occurs when animals these trained quolls a small, live cane toad in
ingest a novel, toxic food, experience illness, an open topped jar, in front of a video cam-
and subsequently associate the smell or taste era, to see whether they would attack the
of that food with illness, and avoid ingesting toads. Remarkably, the trained quolls sniffed
the toxic food for long periods (Garcia et al. the toad, but refused to attack it, whereas our
1974). Unlike classical conditioning, whereby control toad‐naive quolls quickly consumed
animals require multiple trials to learn the the live toad. The aversion to live cane toads
appropriate response (Chapter 1), CTA usu- lasted a week in captivity, suggesting that the
ally involves one‐trial learning; that is, after a toad‐trained quolls might have the skills
single bout of illness, animals learn to avoid necessary to survive in toad‐infested areas.
the food that induced illness. After its initial Did toad‐aversion training confer survival
discovery, conservation biologists realised benefits to quolls in toad‐infested habitats?
that they could change the feeding behaviour To answer this question, 30 trained ‘toad‐
of predators by adding nausea inducing smart’ and 30 untrained ‘toad‐naive’ quolls
chemicals to meat baits. Initial pen trials on were fitted with radio collars and reintro-
coyotes (Canis latrans) were encouraging; duced to suitable rocky habitats in toad‐
coyotes that consumed sheep baits paired infested areas near Darwin, in northern
with lithium chloride became ill, and these Australia (Figure 12.2). A large team of vol-
coyotes subsequently refused to attack live unteers was enlisted to help follow each quoll
lambs (Gustavson et al. 1974). Likewise, during the first few hours after release, and
crows that ate green‐coloured chicken eggs locate them daily thereafter, so that we could
injected with a nausea‐inducing chemical determine their fate. The radio‐tracking
subsequently refused to eat untreated green work revealed that untrained quolls often
chicken eggs, but continued to eat white eggs encountered cane toads within hours of
(Nicolaus et al. 1983) Despite these encourag- release, and several of these toad‐naive quolls
ing results, the use of CTA to alter predator attacked large toads, and died from toad poi-
behaviour was embroiled in controversy soning. By contrast, the trained ‘toad‐smart’
(Gustavson and Nicolaus 1987). Most of this quolls encountered toads, sniffed them, and
controversy surrounded the use of CTA baits
to train wild coyotes to avoid attacking lambs;
field trials produced equivocal results
(Bourne and Dorrance 1982) and the use of
CTA was abandoned in favour of alternative
(and often, lethal) methods to control coyote
numbers (Gustavson and Nicolaus 1987;
Conover and Kessler 1994). Thus, until
recently CTA has largely been overlooked as a
tool for training animals.
To see if we could train quolls not to eat
cane toads, in 2009 my colleagues and I fed
toad‐naive, captive reared quolls a small,
non‐lethal sized (<2 g) dead toad infused Figure 12.2 Photograph of a trained northern quoll
with the odourless, tasteless, nausea‐induc- fitted with a radio‐collar. Both untrained and trained
northern quolls were released in the wet season at
ing chemical thiabendazole (a chemical used sites around Darwin, where toads were abundant.
to deworm livestock which has a nausea Source: Jonathan Webb.