I assessed the influence of microhabitat variables on the abundance and spatial structure
of small mammals across three types of forest, deciduous (Colonial Point), hardwood-pine (the
Burn Plots) and cedar swamp (Reese’s Swamp), at the University of Michigan Biological
Station (UMBS), in Cheboygan County, northern lower Michigan. Traps have been placed in
these forests twice a year for 18 years since 1989 by Professor Philip Myers, and I based my
study on his trapping records. I created a four by four meter square plot centered on each trap
station and measured habitat variables on each plot in late September and early October 2006.
Over 2,000 captures of sixteen small mammal species were recorded in 6,480 trapnights
over eighteen years. Four species were commonly captured: eastern chipmunk (Tamias striatus),
white footed mouse (Peromyscus leucopus); short-tailed shrew (Blarina brevicauda); and
southern flying squirrel (Glaucomys volans). I eliminated Tamias because its trapping records
were clearly strongly influenced by daily weather, and I focused on the remaining three. For P.
leucopus and B. brevicauda, I included trapping records for only the most recent four years, as
those species are short-lived, and ecological variables such as woody debris and ground cover
change from year after year. For G. volans, which is longer-lived, I used all 18 years of trapping
records. Also, G. volans is much less common than P. leucopus or B. brevicauda, and sample
sizes were too small to analyze unless all years were included.
The preferences of each species across three transects located in the different forests were
evaluated by statistical methods including Poisson regression models, principal component
analysis, analysis of variance and analysis of covariance. The spatial structures of the
populations of each species were examined by spatial autocorrelation analysis using Moran’s I.
In deciduous and pine-hardwood forests, white footed mice were found to be habitat
generalists. In the cedar swamp, on the other hand, they were likely to be habitat specialists
limited by food availability and predation risks. A relatively high spatial autocorrelation of the
number of captures/station was found only for P. leucopus in the hardwood forest in the fall.
Spatial distribution of food in patches over the series of adjacent trap stations may explain for
this pattern, but none of my measurements test this hypothesis. Short-tailed shrews preferred
deciduous forest that is characterized by many trees, much leaf litter and large variety of ground
cover species. In the cedar swamp, they preferred sites with large trees, moss/lichen cover on the
ground, much woody debris and many snags. Limiting factors for them were likely to be soil
moisture and food availability. Most southern flying squirrels were found in the deciduous forest,
where they were not habitat selective. In the pine-hardwood forest, G. volans preferred sites
with greater diversity of trees and larger trees. No preference for snags was found. In the cedar
swamp, no G. volans was found over the eighteen year study.||en_US