Abstract
Museum records can document long-term changes in phenology, species interactions, and trait evolution. However, these data have spatial and temporal biases in sampling which may limit their use for tracking abundance. Often museum records are the only historical data available, and Boyle and colleagues make long-term abundance estimates for the Eastern North American Monarch butterfly (Danaus plexippus) and its milkweed hostplant (Asclepias spp.) using 1,191 and 31,510 records from 1900-2016, respectively. They conclude that Monarch and milkweed abundance started to decline in the mid-20th century, before the adoption of herbicide-resistant crops that are often blamed for losses of Monarch hostplants. Using the same data, I argue that the Monarch trend changes with the choice of taxa used to standardize Monarch records. The abundance trend after dividing Monarch records by butterfly (Rhopalocera) or Nymphalidae records, instead of by Lepidoptera as in Boyle et al. (2019), shows no mid-century peak corresponding to the milkweed trends. One reason the Monarch trend reported by Boyle and colleagues changes when standardized by other taxa is the declining proportion of butterflies within Lepidoptera records from a peak of 40% in the mid-20th century to less than 10%. This reanalysis shows that changes over time within the taxa used to standardize records matter, in addition to potential sampling biases in the species of interest.