HOODED MERGANSERS - AS TEACHERS
By Paul Dye

When you've been around awhile, and have established yourself as an aviculturist, interesting things begin to happen. News of your talents and interests somehow makes it into unexpected places. You begin to get requests to advise and/or contribute to aviculturally related projects by people at Universities, Zoological facilities and wildlife agencies. All that study and problem solving that you had to apply to your own operation starts to pay off big time, and if there is anything even more rewarding than succeeding at a particularly difficult breeding project, it's got to be helping others to do the same.

Four such opportunities have come my way in the past couple of years, from assisting in study projects with waterfowl in my state and in the Brazilian wilds, to participating in recovery efforts with a species in peril. One I thought might be particularly interesting to other aviculturists is the use of Hooded Mergansers (Lophodytes cucullatus) to teach young fish to avoid avian predators prior to their release. It should have practical applications far beyond the National Marine Fisheries Service research facilities at Manchester, Washington, where the study was done using some mergansers from my farm, and may ultimately provide a ready market for surplus hooded mergansers and other merganser species at salmon and trout rearing facilities world-wide.

It doesn't take a rocket scientist to understand that salmon and trout fry scatter widely when released into streams and rivers. For this reason their first encounters with predatory birds often come at a time when very few others witness and learn from the loss of one of their own to these predators. The resulting losses can be prohibitively expensive, requiring the rearing and release of a great surplus of fish. It also may not be completely beneficial to the predators, creating a temporary glut of prey in one area, and luring them away from their normal behavior and habitats.

Salmonid (salmon and trout) fry are usually raised in tanks and raceways where they have very little exposure to the natural predators they will experience after release. Traditionally these tanks have been covered by netting to protect the fry from local predators. Simply uncovering the tanks to allow predators access doesn't provide the uniform exposure needed to assure that thousands of fry have been uniformly trained, therefore the percentage loss after release cannot be accurately predicted and an expensive surplus still must be raised.

The primary predator training sessions were conducted at the Manchester Marine Experimental Station by placing the Hooded Mergansers in cages placed in three of the six raceways. The fish in these three raceways recievd experimental predator avoidance training, while fish in the other three raceways served as untrained controls for the comparison study. The control raceways were always covered with bird-tight netting to ensure that unintended predator exposure did not affect the results. Except for predator avoidance training, the fish in both the trained and untrained groups received identical husbandry and were reared following standard salmon culture techniques.

The Hooded Mergansers were placed in cages constructed of a 1.6m long by 1.lm wide by 1.1m tall polyvinyl chloride (PVC) pipe frame that was completely covered with a 3.8 by 3.8cm mesh net. This size mesh allowed the chinook salmon fry to freely swim in and out of the cage, while confining the Hooded mergansers (and sometimes predatory fish) within the cage. When installed in the raceways, the top half of the cages were suspended above the water so that the mergansers would not drown. Cages containing no predators were often installed in the raceways so that the fish would learn to associate the predation with the Hooded Mergansers, rather than the presence of the cage.

The Mergansers were placed into the cages for seven 50 minute long training periods. In nearly all of the training sessions, the mergansers were removed from the raceway before they ceased fishing. This ensured 'that the salmon fry experienced nearly continuous negative reinforcement (training) from these predators.

The effect of predator avoidance training on post-release survival was evaluated with releases of study fish into two tributaries of the Curley Creek watershed in Washington. These releases consisted of representative samples of fish from each of the six raceways. The releases began three days after the last training session. The release site on each tributary was l..3 km upstream of the smolt collection weirs. The possibility of contagious behavior confounding the results was minimized by releasing fish from only one rearing group in each tributary on a given release day. Contagious behavior is a form of social learning where naive animals mimic the behavior that more experienced members of their group display to predators, food sources, and other new stimuli. The possibility of fish from different rearing groups meeting each other at the release sites was further reduced by allowing at least 48 hours to pass between releases. Tributary effects were controlled by alternating the tributary the fish were released into from one release to the next. A total of 511 control and 510 predator trained fish were released into the watershed during the 10 releases. The difference in recovery between the two treatments was compared.

In this study, the salmon rapidly learned to avoid mergansers. Prior to the introduction of mergansers to a training cage placed in a raceway, fry readily swam into and out of a cage. However, after three training sessions with mergansers, few fry continued to enter the cage. By the fifth session, almost no salmon entered the cage and nearly all the fish remained at least 15cm from the cage. Initially, the mergansers ate an average of more than nine fish per training session. However, this average rapidly declined to less than six fish per training session as fry became conditioned to avoid the birds.

CONCLUSION: The study confirmed that predator avoidance training with live predators can increase the post-release survival of hatchery-reared salmonids. The benefits of this training can be considerable with the post-release survival of predator trained fish being 26% higher than untrained fish. The predator avoidance training protocol used in this study required only a slight increase in operational costs. Dur4Lng predator avoidance training, less than two hours of personnel time are expended per day in handling mergansers. The birds rapidly learned to enter the training cage, the cage was easily transported to and from the raceways, and once the cage was placed in the raceway it did not appear to interfere with routine fish culture operations.

Predator avoidance training has a very favorable cost-benefit ratio. This ratio is based on the number of fish sacrificed in training (0.6%) that would have successfully migrated downstream compared to the increase in number of successful downstream migrants (26%) due to training.

While predator avoidance training is a useful tool for increasing post-release survival, it probably only needs to be implemented at those facilities that produce predator naive fish. Hatcheries with predators reliably entering their ponds due to a lack of bird netting and electric fences are probably already providing uncontrolled predator avoidance training.

Reference:
Maynard, D.J., A- LaRae, G.C. McDowell, G.A. Snell, T.A. Flagg and C.V.W. Mahnken 1997. Predator Avoidance Training Can Increase Post-rL-IL-ase Survival of Chinook S-alman. 49th Annual Pacific Northwest Fish Culture Conference Proceedings

Paul & Lynn Dye
Northwest Wildfowl Farm
10114 54th Place N.E., Everett, WA 98205 USA

Phone:(425)334-8223 Fax:(425)397-8136
E-mail: dye@greatnorthern.net