Squirrel Middens Influence Marten Use of Subnivean Access Points

Martens (Martes americana) use the subnivean zone during winter to obtain prey. In Yellowstone National Park, we noted that martens used access holes differentially, and that a number of subnivean access holes were associated with red squirrel (Tamiasciurus hudsonicus) middens. In this paper we addr...

Full description

Bibliographic Details
Main Author: Bissonette, John A.
Format: Text
Language:unknown
Published: Hosted by Utah State University Libraries 1993
Subjects:
Online Access:https://digitalcommons.usu.edu/wild_facpub/1304
Description
Summary:Martens (Martes americana) use the subnivean zone during winter to obtain prey. In Yellowstone National Park, we noted that martens used access holes differentially, and that a number of subnivean access holes were associated with red squirrel (Tamiasciurus hudsonicus) middens. In this paper we address the question: Does the presence of red squirrel middens influence subnivean access point use by marten? Marten use of access points was influenced significantly by presence of middens. Thirty-three percent of used access points were associated with middens compared to 16% for unused access points (Kendall's tau t = -0.2293, P = 0.015). Martens appeared to key in on subnivean prey and used access holes with higher small mammal biomass (not including squirrels) more frequently (t-test t = -3.0697, df = 88, P = 0.002). When squirrels were included in the analysis, the relationship was strengthened (t-test t = -3.9723, df = 88, P = 0.0001). Scat analyses (n = 69) identified red-backed voles (Clethrionomys gapperi, 35.5%), red squirrel (21.1%), unknown sciurids (9.0%; total sciurids 30.1%) and lagomorphs (16.6%) as the major prey species. Coarse woody debris (CWD) levels around access holes did not differ regardless of whether squirrel middens were present or absent (t-test t = 0.1363, df = 88, P = 0.89). Likewise small mammal biomass levels did not differ whether middens were present or absent (t-test t = 0.9771, df = 88, P = 0.33). Access use was not related to percentage CWD when tested using a separate (unequal) variance estimate t-test (t-test t = -1.5959, df = 88, P = 0.11).