It was know before Phelps. Bill Ricker, of Ricker Curve fame, looked at the changes in size over time for the five species of Pacific salmon. Fishing definitely was correlated with changes in size. Most of the time, it made them smaller. Sometimes, larger. In that case, fishing for sockeye, with the smaller sockeye gear, selected out the small chum. So, the chum run actually got larger.

Back in the 80s, Skagit coho were routinely in the hole and required protection. State and tribes fished the Bay and river for chum with, I believe a 6 or 6.5 inch minimum mesh. Caught lots of chum and no coho. December 1, end of salmon management, start of steelhead they dropped to 5 inch. Bam, coho showed up in the catch. For a month they hadn't been there, because they were smaller than the chum.

There are lots of by catch issues out there....

It's genetics. The genes that aren't killed are what you use. Very intensive selection. I know lots of folks complain, or did, that hatcheries breed "non-biters". Not intentionally, but if the 10% of the return that you use to spawn was too small for the nets, too early for the intensive fishery, and didn't bite a hook just what do you think the kids will look like?