On a small lake near Kamloops, British Columbia, researchers tagged 157 rainbow trout, fished the water every single day for a month, released everything they caught, and watched the catch rate fall off a cliff. It dropped by roughly three-quarters over the month, with the population held constant and not a single fish removed. The bulk of the collapse happened in the first week to ten days. Nobody changed flies to trick the fish. The fish simply got harder to catch.

That study — Askey and colleagues, published in 2006 — is the reason to take “hopper fatigue” seriously, and explains more about your late-August hopper bank than any fly-shop bin can. The catch rate on Little Pantano started high and fell off through the month, most steeply in that first week to ten days. The researchers couldn’t explain the pattern without what they called learned hook avoidance, and their best-fitting model split the population into two classes: a smaller group of highly catchable fish that got caught first, and a tougher, less-catchable class that made up the rest. Each time a fish was caught, it grew markedly less likely to be caught again.