In PNAS, Saunders et al. (1) employ a hierarchical set of statistical models and variable selection to construct a predictive set of population input and environmental variables for winter colony sizes in Mexico. Due to limitations in availability of individual-colony census data, the authors focus their analysis on the years 2004 to 2015, which has two important implications (Fig. 1). First, this temporal window is after the large-scale adoption of genetically modified herbicide-tolerant crops (and corresponding herbicide use) topped out; thus, herbicide usage was no longer increasing. Second, this is the same window when the monarch population (as censused in Mexico) experienced the steepest and most persistent decline in recorded history, with the all-time low recorded in 2013 (Fig. 1). Saunders et al. report that winter abundances can be predicted by variation in the input of summer breeding monarchs, in addition to variation in an index of flower nectar availability (during the southern migration) and forest patch size (at the overwintering sites).

Monarchs use a wide variety of flowers for nectar, and milkweed is typically no longer flowering during the migration; therefore, the findings of Saunders et al. (1) support a long-standing but previously untested hypothesis of floral nectar limitation. The impact of overwintering forest cover at colony sites has also been suspected to be critically limiting to monarchs, but only recently has a quantitative link been made showing the limiting effects of degraded forests (12). From these factors, it can be concluded that during the decade of the steepest declines in monarchs, climate and its impacts on availability of floral nectar and the continued degradation of Mexican forests were critical factors. Nonetheless, why the summer breeding population of monarchs plummeted during the years 2004 to 2015 is still being debated. Was it a delayed response to the biofuel boom and enhanced usage of herbicides? Was it the 100-y drought experienced in Texas, which is a critical spring and fall bottleneck? Or was it some combination of stresses?

In the second study in PNAS, Boyle et al. (2) take the temporal long view, examining monarch and milkweed dynamics over the past 116 y. The extraordinary approach was to employ tens of thousands of museum records of both monarchs and milkweeds as indices of their population sizes. Although this approach will surely be criticized for all sorts of limitations, it importantly brings an entirely new set of data, and one of tremendous temporal depth, to the table. Boyle et al.’s key conclusion is that the monarch decline began over 60 y ago and was concordant with the abandonment of small-scale farms (which served as a reservoir for milkweeds). This result is concordant with the long-term decline of monarchs in California supported by observational data remarkably collected by a single researcher over a 45-y period (13). Such multidecadal declines have important implications, the first of which is that no single recent event, such as the advent of genetically modified crops, can be implicated in monarch declines. Indeed, a blue-ribbon panel of the National Academy of Sciences in 2016 came to this same conclusion about genetically modified crops and monarch butterflies (14).

Perhaps more importantly, the whole of land-use change, including agricultural practices, chemical inputs, habitat fragmentation, pollution, development, disturbance, and so forth, is what has been incrementally creeping up over the last century. Which of the many aspects of this long-term environmental crumbling is responsible for monarch declines is unclear. To me, one of the most disturbing findings of Boyle et al.’s study (2) is the strong and persistent declines of eight milkweed species over several decades. The common weedy milkweeds (Asclepias syriaca and A. speciosa) have enjoyed the benefits of agriculturalization through soil disturbance, fertilization, and creation of ditches resulting in larger populations. However, the nonweedy Asclepias, numbering well over 100 species in North America, appear to be in trouble.