Wind by temperature estimates were positive in early UK moth samples but uncertain in later samples in 1993-2015

We tested whether wind changes the temperature slope of recorded moth catch. Four conditional Poisson models used 28,486 nights in England and Scotland in 1993-2003 and 2004-2015. They controlled trap weeks, curved main weather terms, displaced weather, rain, radiation and moon phase. Wind by temperature terms were +0.01272 and +0.01719 in the two early samples. Their four-test intervals excluded zero. Later terms were -0.00481 and +0.00688, with intervals that included zero. The early English result became uncertain without curved main terms or with positive catches only. Conservative source-quality exclusions kept the primary pattern. Thus a clear interaction did not repeat across both periods. These are catch associations, not population effects. Missing later zero catches, two Scottish sites, sensor error and few year clusters limit inference. Code, aggregate results, a figure, source hashes and external data restoration instructions are included. Pollution and climate-change attribution were not tested.