Monarch butterflies are one of the best studied non-pest lepidopterans, serving as a model for migration, chemical ecology, and insect conservation. Despite the intensity with which the larvae and adults have been studied, the cryptic pupal stage is often difficult to study in the wild. It is perhaps due to this difficulty that researchers have largely overlooked monarchs’ interactions with a pupal parasitoid, Pteromalus cassotis. Using field experiments in the northern U.S. and observational data from wild collected pupae in the southern U.S., we report occurrences of this host-parasitoid interaction at sites Minnesota, Georgia, Oklahoma, Texas, and Wisconsin. At sites in Minnesota, rates of parasitism of experimentally placed monarch were highly variable, ranging from 60% in 2010 to 0% in 2013 and 2014. Observations of wild-collected pupae suggest that rates of parasitism may near 100% at some sites in the southern U.S. The number of wasps emerging from a single host ranged from 1–425 (mean  =  71). Later dissections of hosts revealed that, in some cases, dead parasitoids remained inside the host as larvae, pupae, and/or adults. Within a host, wasp sex ratios were typically female biased (median  =  91% female), as is common in gregarious parasitic hymenopterans. Infected monarch pupae at a site in Oklahoma produced more wasps per host, more male-biased sex ratios, and had higher survival than hosts from other sites. We discuss the possibility that P. cassotis is a specialist on monarchs and perhaps closely related species, based on monarchs’ sequestered cardenolides, published host records, and evidence for correlated population dynamics of this host and parasitoid.