Monarch butterfly trends reported in Boyle et al. (2019) are sensitive to unexamined changes in museum collections over time

Museum records provide an underutilized source of information for documenting long-term changes in phenology, species interactions, and trait evolution. However, non-systematic collection data must be treated carefully if they are to approximate abundance, as trends may be confounded with spatial or temporal changes in sampling effort. Boyle et al. (2019b) argue that the relative abundance of Eastern North American Monarch butterflies (Danaus plexippus) has been in a long-term decline since the mid-20th century, following a similar decline in milkweed (Asclepias spp.) herbarium records. I demonstrate that this reported abundance trend is biased by the choice to standardize Monarch records as a proportion of all Lepidoptera collected. The sampling of Lepidoptera has changed systematically over time to favor moths, causing the apparent trend in Monarch records. With the data standardized more appropriately, I show that the trend in Monarch records shows no mid-century decline and increases over recent decades. As the trend in Monarch museum specimens contradicts the recent trend in Monarch abundance documented from systematic population monitoring, I argue that these records are unreliable for abundance estimates. The conclusion in Boyle et al. (2019b) that Monarch declines started in the mid-20th century is unwarranted both because the trend is biased by sampling changes in museum records and because the trend in Monarch records, when corrected, does not correspond with real-world population abundance.