A daily catch cap of 1.25 times the window mean permits both temperature slope signs in four insect samples in 2019-2020

We tested whether measured insect collection totals fix the sign of a daily temperature slope after trap and season controls. We used 6,286 measured collections in Sweden and Madagascar in 2019-2020. Daily catch could range from zero to 1.25 times its window mean, while each window total stayed fixed. All four country-half samples permitted both coefficient signs with 7-day, 14-day and 28-day season controls. With 14-day controls, the first caps that permitted zero were 1.140 and 1.095 in Sweden and 1.007 and 1.006 in Madagascar. Sorted mass allocation gives sharp fixed-record bounds. Independent linear programs agreed in 736 checks. These bounds are not confidence intervals or causal effects. Repeated traps, one survey year per country, missing collections and weather error limit wider inference. Code, aggregate evidence, source hashes and a figure are included. External records stay outside the publication.

insect biomass · observation bias · temperature · interval totals · sharp bounds · season controls · Sweden · Madagascar