Short-Circuiting Evolution

How Hatcheries Undercut Genetic Adaptation


Grand Coulee Dam (credit: Bonneville Power Administration)

Fresh out of high school, for a few summer weeks, I played God.

A returning female spawner having her eggs removed

The qualifications weren’t high. Nor did I feel particularly godlike. Dressed in a fisherman’s yellow bib and waders, I spent Monday mornings swirling the eggs and sperm of hundreds of returning salmon, in the shadow of Trinity River’s Lewiston dam.
A flick of my wrist brought into existence the next generation of Trinity river Chinook--60% of Klamath Chinook and 20% of California’s commercial salmon landings. This simple movement supplanted the complex stages of the salmon spawning that has evolved over the course of two million years. The swirl of sperm in buckets and laying the eggs in aluminum trays washed with iodised water substituted the three month spawning. Gone was nature’s selection of the most fit female, to build a durable redd, a perfect nested eddy which she has built by piling rocks with her tail--out of her nearly thousand stock sisters. Gone was the elaborate variation in male fish colors and forms designed to compete for mating rights; Gone was the need for aggression among adult male salmon to fight off jacks too young to spawn.

Eighty percent of all salmon in the Pacific Northwest are hatchery bred, meaning that interns and aluminum racks, troughs and feeding tubes make up the environment of the first three months of a salmon’s life rather than stream flow, availability of salmon carcasses or the life cycles of insect larvae. Till they hatch, eggs are held in aluminum racks and flushed with water that is mechanically-monitored for temperature, sterility and flow pressure. At two months, the fry are then shifted to incubation tanks in a warehouse, where they are lit by solar lamps and fed every twenty-four hours with fishmeal pellets shipped in from Vancouver, Canada. Then they are suddenly washed down river with a large dam release and need to fend for themselves for the next six years.

Hatchery worker laying fertilized eggs in trays to be hatched.

“Natural selection is gone,” says Paul Merz, a salmon fisherman from Charleston, Oregon, “the fish are now stupid like plastic wind up dolls.” Merz has fished for salmon commercially for thirty years. Many hours on the water guessing the depth salmon are schooling at, and adjusting troll lines accordingly, have made Merz a salmon psychologist by profession. Merz reasons that the devolution of salmon to sheer stupidity is the restructuring of their life cycle by hatchery programs. There are 48 hatcheries along the Pacific Coast of the United State,s and Merz hypothesizes that the heavy mechanization of hatchery salmons’ early life history and the obliteration of natural selection is the cause. The natural selection that Merz is referring to is the form of female nest site choice and tail strength, male aggression against other suitors and the egg’s ability to survive scouring and the alevin stage. But is physical localism really the key to genetic memory retention?

Not at the individual organism level, according to Laurie Weitkamp, an EPA salmon geneticist based at Hatfield Marine Center in Newport. “Reproduction is conserved in the evolutionary sense so salmon are probably recently evolved from freshwater fish.” In other words, localism is hard-wired, irrespective of life history.

If reproduction is conserved, habitat plays a big role in the evolutionary divergence of salmon species. The ability to return to their natal streamlet would then be a good test of the genetic integrity of a stock. Using this hypothesis, Weitkamp studied the distribution of 1.77 million coded wire tagged Coho from 90 hatcheries and 36 wild populations along the West coast of North America. She gathered her data from fishermen along the coast. Each fish caught with a nose tag would have its tag, a flesh sample and GPS position sent to Wietkamp. She identified 12 distinct ocean distribution patterns despite wide coastal river dispersal.

Both hatchery and wild Chinook and Coho along the east Pacific coast were returning, more or less stably, to their historical marine distributions meaning that their individual genetic memories are still intact. Regardless of the river that smolt are released on, they have a high likelihood of spending their years at sea in roughly the same areas that their genetic stock derives from. This means that Rogue River Coho will feed where Rogue river Coho fed, even if they are released on the Columbia River. This explains why salmon blocked off from historical spawning grounds will return to them if given the chance, even when their immediate ancestors have never seen those grounds.

For example, when Mount St. Helens erupted in 1980, volcanic sediment blocked the Toutle River system. Half the Toutle’s Chinook population spawned in the neighbouring Columbia tributaries.

At a species or population level, hatcheries still remain a threat to salmon though. Whatever genetic memory is retained at an individual level, salmon genetics at larger levels—ESUs, stocks, rivers, regions—are less diverse because of the coupled effect of dams and hatcheries.

Individual hard-wiring simply conserves existing genetic code. While localism is hard-wired at the individual level, it does little for the prospects of the ESU as a whole.

Hatcheries make no provision for appropriate ongoing adaptation because eliminating salmon as active members of the spawning process diminishes sexual selection.

Moreover natural selection selects for adaptation to unnatural hatchery conditions instead of selecting for those eggs that can survive gravel scouring or those fry that forage best.

Simultaneously, hatchery releases shock wild salmon populations as hatchery salmon are released all at once causing a sudden increase in competition for food in wild populations. Hatchery releases also mask overall population decline levels, which does little good for the ESU as a whole.