I had tied a wide range of October caddis pupa for years, but one late-October day on the Rogue, I caught and harvested a hatchery steelhead and cut it open. Its stomach and gullet were packed with fresh October caddis pupa that were darn near still wiggling. It was a moment I will never forget because I truly saw what they looked like, and in the past, I had tied off of what I assumed they looked like. Two things that stood out to me were the defined banded segmentation, and the other was that they were 1/3 again larger than I expected. Within a week, I had fine-tuned a pattern that looked just like them, sunk fast, and outperformed all of my previous attempts to imitate a key insect in a phase that is hard to observe.