Aim Plant macrofossils in fossil woodrat (Neotoma) middens are the primary source of information on late Quaternary biogeographical history of plants in arid and semi-arid regions of North America. Macrofossil records from middens are playing particularly important roles in documenting spatial and temporal patterns of plant migrations and invasions since the last glacial maximum 21,000 years ago. However, relatively few actualistic studies comparing contents of modern middens with surrounding flora and vegetation have been carried out. The primary aim of this study is to assess the reliability of midden assemblages in detecting the presence of tree, shrub and several other plant species growing on the surrounding landscape. The secondary aims are to determine whether probability of occurrence of species in middens is related to abundance in vegetation, and whether representation of individual species in middens is contingent on presence/absence or abundance of other species.
Location Our five study sites were bedrock escarpments or canyons in the central Rocky Mountains (north-eastern Utah, central Wyoming and south-central Montana). All sites were in conifer woodland or forest/woodland variously dominated by Juniperus osteosperma, J. scopulorum, Pinus edulis, P. flexilis and Pseudotsuga menziesii. Neotoma cinerea is the only woodrat species in the region.
Methods Macrofossil assemblages from 59 modern middens (all showing clear signs of recent or ongoing woodrat activity) were compared with floristic composition of vegetation within 50 m of the middens, and with percent cover of vegetation within 30 m of the middens.
Results Coniferous trees and shrubs were well-represented in middens, occurring consistently even when abundance in the local vegetation was very low. Juniperus osteosperma and J. scopulorum were particularly well-represented, occurring in middens regardless of local abundance. Other conifers (P. edulis, P. flexilis, P. ponderosa, Pseudotsuga menziesii) were occasionally absent from middens when abundance in vegetation was low (< 20% canopy cover). Occurrence of dicot shrubs and graminoids was less consistent.
Main conclusions Middens constructed by N. cinerea are highly reliable sensors of presence/absence of J. osteosperma and J. scopulorum, and hence can be used to infer invasions and past biogeographical distributions of these species. The middens are also reliable in registering populations of other conifers, although presence in middens may be contingent on local abundance. Additional comparative studies are needed to develop a sound empirical basis for using middens of N. cinerea and other species to infer past presence/absence of plant species on the landscape, and to explore the vegetation-sensing properties of midden assemblages.