Preceding and following winter warmth gave no clear slope difference for four matched butterfly species in England, 1980-2023

We tested whether preceding winter warmth had a more negative link than following winter warmth to an adult-minus-larval butterfly index. We matched Peacock, Small Tortoiseshell, Meadow Brown and Ringlet at the same sites. Four models used 10,326 site-years at 869 sites in northern and southern England in 1980-2001 and 2002-2023. Models controlled site effects, time and seasonal weather. Preceding minus following winter slopes ranged from -0.337 to -0.063 log-index units per C. All four adjusted intervals included zero. Positive-index and long-record filters and site trends kept negative estimates, but all intervals still included zero. Removing one year could reverse later-period signs. This test gives no clear repeated winter timing contrast. It does not show equal responses or no winter effect. Selected complete records, few years, coarse weather, species differences and unmeasured survey bias limit inference. The package gives a matched panel, code, source hashes, a figure and numerical checks. We did not test air pollution or causal winter survival.

butterflies · winter temperature · observation bias · negative control · matched species · regional replication