Big, neon-red sockeye salmon with green heads and hooked jaws stream north along the Canadian bank, massed in slack water, shooting into the current upstream, swimming low among the boulders, then rising and resting in the shallows before they tackle the white water just ahead. Already, their carcasses, faded to gray, litter the rocks and gravel of an island toward the other side. The fish form a river within the river, a submerged current of red. They just keep coming. I recall the third stanza of Yeats' "Lapis Lazuli":

On their own feet they came, or on shipboard,

Camel-back, horse-back, ass-back, mule-back,

Old civilizations put to the sword. . . 

Millions of fish are streaming up the Adams River, east of Kamloops in interior British Columbia. The Adams isn't a big river. The water is very clear. There may be 9 million fish. You see them.

The return of the salmon to rivers large and small is, of course, a Northwestern commonplace. But can this seasonal appearance of big fish in small streams, of bright red males competing to fertilize the eggs pinkly visible in the gravel, their bodies half out of the shallow water, chasing each other across the rocks, ever be really commonplace? Certainly, there's nothing commonplace about seeing this many fish. They make you think this is the way it used to be. And this is the way it should be.

The Adams flows south through Shuswap Lake to the south fork of the Thompson River, then to the Fraser and ultimately to salt water, 250 miles away. In a good year, the Fraser River sockeye run dwarfs any other salmon run south of Alaska, and the Fraser's largest spawning population swarms into the Adams River. Every fourth year, when the sockeye run peaks, locals hold an October “Salute to the Sockeye” at Roderick Haig-Brown Provincial Park. (Haig-Brown was a writer, conservationist and fly fisherman who died in 1976.) Each morning, yellow school buses bounce into the grass parking area. Young kids ride their parents' shoulders to the river bank. People push wheelchairs along the dirt trails. Indians sell fry bread from trailers. It's an event.

Look at the fishie, says an excited grade-schooler on a field trip, spotting a red torpedo shape in the water. A teenager on a field trip of his own peers into the shallows and exclaims, they're feakin' big! They are indeed, although even the largest is a lot smaller than the dark bulk of the occasional chinook that lurks in the clear water beside them. (Now that Puget Sound and Columbia River chinook populations have been pushed to the list of threatened and endangered species, Fraser River chinook have become the Southern Resident Killer Whales' — i.e., Puget Sound orcas' — main nosh.)

These fish have followed clockwise gyres through the Pacific, swimming south past Southeast Alaska, then either east or west of Vancouver Island, passing through the Strait of Juan de Fuca or the Johnstone Strait on their way to the Fraser.

An estimated 34.5 million sockeye returned to the big river this year. An estimated 24 million of them were destined for the Shuswap area, primarily to the Adams. The government won't yet take a stab at the number that made it all the way back there, but even if you take out a percentage for Indian and other fishing and for natural deaths along the way, you still have an awful lot of fish.

This is the largest Fraser River run since 1913, the year before a rockslide caused by blasting for railroad construction in Hell's Gate, near Hope, B.C., largely blocked the Fraser to upstream migration. (More sockeye return regularly to Alaska's Bristol Bay, but they're heading for nine different river systems.) Viewed another way, no one younger than about 100 can remember seeing so many sockeye in the Fraser or its tributaries.

No one knew this huge run was coming. And no one can explain it. “Once those fish leave fresh water, they're just in a black box,” says University of Washington Professor of Aquatic and Fishery Sciences Ray Hilborn. “There's no research on the ocean.”

Back in the 1970s, Hilborn explains, when Japanese ships were drift netting American salmon on the high seas, “the U.S. And Canada put a lot of energy into proving those were American and Canadian fish.” At the time, scientists used scale patterns to prove fish came from North America or Asia.

Now, researchers could use sophisticated DNA analysis to find out which river, or even which part of a river system, fish came from. But no one is doing so. Without the political conflict, Hilborn says, “there's just no money for it.”

Without the research, it's hard to figure out what's going on. Are there any good hypotheses for the size of this year's run? “I haven't seen them,” Hilborn says.

This year's sockeye returns have varied widely, depending on the river system. Very few fish came back through Lake Washington. But on the Columbia River, more sockeye — nearly 400,000 — have showed up at Bonneville Dam than at any time since the dam was completed in 1938. Snake River sockeye are doing relatively well, too — at least compared to recent decades, and at least for 2010. Those fish, which spawn in Idaho's Sawtooth Mountains, were the first Columbia River system salmon listed under the Endangered Species Act. The population was basically extinct when it was listed in 1991; that year, only one fish made the 900-mile journey up to Redfish Lake. Four years ago, only 52 reached Lower Granite Dam, and only three made it all the way.. Now, the count at Lower Granite has topped 2,000. Wild salmon advocates attribute that not only to favorable ocean conditions but also to the water that a U.S. district court has ordered the feds to spill over the dams every spring for the past five years.

Last year's Fraser River run was a disaster. It capped 20 years of declining runs, and it was so much worse than anyone had expected that the Canadian government formed the Cohen Commission of Inquiry into the Decline of Sockeye Salmon in the Fraser River. There was a lot to inquire about: In August, around the time the Cohen Commission announced its research program, commercial fishermen started hauling in more Fraser River sockeye than any of them had ever seen, taking 10,000 or even 15,000 fish in one set of a seine.

They were doing it on both sides of the border. Fraser River sockeye have always made up most of the “Puget Sound” catch. At the turn of the century, fishermen rowed flat-bottomed skiffs up to the San Juans from the lower Sound to net them, camping on the beaches of Lopez and San Juan, and rowing out each morning to fish. The first salmon cannery on Puget Sound was built in 1877 at Mukilteo, for easy access to the Fraser runs. The big canning companies built fish traps on the points and islands where the bulk of the fish swam through.

They pushed out the natives who had fished there for millennia. In 1899, the assistant secretary of the Smithsonian Institution, Richard Rathbun, wrote that fish traps had displaced the native reef net fishery at Point Roberts. “After the completion in 1894 of the continuous line of traps commanding the approaches to the big reef,” he wrote, “its value for (native) reef-net fishing seems to have been in great part destoyed and the Indian catches declined so much in consequence as to render the old-time occupation practically unprofitable.”