Aim The status of geographical units within species and species complexes is debated for many taxa, with many molecular studies failing to detect phenotypically defined subspecies. The history and longevity of geographical patterns are also generally very poorly understood. We examine Holarctic red deer (Cervus elaphus and related forms), incorporating ancient DNA to ask whether the present phylogeography has persisted through climatic perturbations or is a relatively recent phenomenon. Location Holarctic (Europe, northern Asia and North America). Methods We obtained 21 modern and 30 Holocene and Late Pleistocene samples, which together with published data resulted in 180 individuals spanning 21 nominal extant and one extinct subspecies. Phylogenetic analyses were carried out on 748‐bp of mitochondrial DNA (cytochrome b and control region). Where possible, the morphology of DNA‐yielding ancient samples was examined to assess subspecies identity. Results Major clades within the red deer complex are upheld, but subspecies within them receive varying support. The ancient phylogeographical structure conforms in significant part to the modern situation, but some haplogroups no longer survive. Moreover, there have been substantial shifts in geographical ranges through time. Wapitoids spread as far west as Romania in the last glaciation, and elaphoids reached eastward to the Ural Mountains. A possible contact zone between the two lineages stretched from the Urals through the Crimea to Eastern Europe. Main conclusions Ancient DNA and morphology are strongly complementary in elucidating population history. Through the past 50 kyr, the major lineages of red deer, and some of the subspecies groups within them, have maintained their genetic and morphological integrity and their core geographical distributions, despite range expansions and contractions and likely contact between the haplogroups (with potential for hybridization).

One major challenge in the study of late-Quaternary extinctions (LQEs) is providing better estimates of past megafauna abundance. To show how megaherbivore population size varied before and after the last extinctions in interior Alaska, we use both a database of radiocarbon-dated bone remains (spanning 25–0 ka) and spores of the obligate dung fungus, Sporormiella, recovered from radiocarbon-dated lake-sediment cores (spanning 17–0 ka). Bone fossils show that the last stage of LQEs in the region occurred at about 13 ka ago, but the number of megaherbivore bones remains high into the Holocene. Sporormiella abundance also remains high into the Holocene and does not decrease with major vegetation changes recorded by arboreal pollen percentages. At two sites, the interpretation of Sporormiella was enhanced by additional dung fungal spore types (e.g., Sordaria). In contrast to many sites where the last stage of LQEs is marked by a sharp decline in Sporormiella abundance, in interior Alaska our results indicate the continuance of megaherbivore abundance, albeit with a major taxonomic turnover (including Mammuthus and Equus extinction) from predominantly grazing to browsing dietary guilds. This new and robust evidence implies that regional LQEs were not systematically associated with crashes of overall megaherbivore abundance.