Appendix A. Detailed methods.
Detailed field methods
Methods are the same as listed in LoGiudice et al. (2003). Host community parameters measured during the peak in larval abundance in late summer were incorporated into the model to predict NIP, which was compared to the observed NIP measured during the nymphal peak the following spring. Small mammals were trapped at approximately 3 week intervals from May through October on three unmanipulated 2.25 ha trapping grids with 2 traps at each station and 15m spacing. Deer and bird densities were determined by spotlight counts and point counts, respectively. All other densities were taken from the literature and considered typical moderate densities for our habitat type (see LoGiudice et al. 2003, Table 1). NIP was measured by collecting ticks from each grid by transect drag sampling (Falco and Fish 1992) and testing via direct immunofluorescence assay (Ostfeld et al. 2001). This method has been corroborated by highly similar NIP values estimated by PCR on the same populations of ticks (Brisson and Dykhuizen 2004).
Data for this study were collected as part of a larger project to assess the relationships between landscape variables, host community, vegetative community, and human LD risk. Forty forest fragments ranging in size from approximately 0.3 ha to 19.0 ha and separated by at least 1 km were chosen in three states with high numbers of LD cases, Connecticut (n=12 fragments), New Jersey (n=11), and New York (n=17).; Fragments were located in the counties in each state with the highest numbers of reported LD cases in 2001, dominated by non-oak deciduous trees (<40% oak canopy), and isolated by at least 80m of non-forest matrix. An excess of candidate fragments were identified and those used were chosen at random. Fragment sizes were determined using orthophotos and subsequently confirmed by GPS and fragments used were separated by at least 1 km.
Host Community Sampling
Sampling of host communities was limited to measuring species richness/identity and densities. Logistical considerations forced us to hold the other parameters in the model constant at the levels observed during the IES study in 2000/1 (LoGiudice et al. 2003). Our goal was to determine a relative activity level (which we will call "activity density") of each species on a given fragment to provide an estimate of the availability of each species to provide tick blood meals. Estimates of NIP in the spring of 2004 were generated from nymphs that fed on this host community as larvae in the late summer of 2003 but host community data were collected in 2003 and 2004. We use combined 2003/4 frequency data for larger mammals (mesopredators and deer) to improve our detection probabilities for these species (see below), which we consider unlikely to fluctuate wildly interannually. We used 2003 data only for small mammals, which can undergo large interannual changes in abundance. The small mammal community was sampled via trapping during peak larval abundance in August and early September (see details below) and density was estimated directly for the most trappable species (white-footed mice and eastern chipmunks). Larger mammals (mesopredators and white-tailed deer) were sampled via camera trapping (see details below).
For the less readily trapped small mammals (shrew and squirrel species) and the larger mammals, we created quasi-quantitative estimates by assigning observed densities to categories and using published values for similar habitats to create an activity density estimate (individuals/ha) for each category (Appendix B Table B1). We have assumed that failure to detect a species indicates that the species occurs at such low activity density that it makes a negligible contribution as a host for ticks. Although we detected 54 species in all (Appendix A), we report results for only the 14 species for which the model was parameterized unless otherwise noted. We also calculated the Shannon Diversity Index (H’; Magurran 1988) for each fragment, using only the species included in the host diversity model.
White-footed mouse, eastern chipmunk, short-tailed shrew (Blarina brevicauda), Sorex shrew (Sorex spp) and grey (Sciurus carolinensis) and red squirrel (Tamiasciurus hudsonicus) densities were assessed via trapping during the peak larval period in late-August, early-September 2003. When possible, a trapping grid of 1.1 ha was established in an 8 × 8 array with 15 m spacing on each fragment. In fragments too small or irregularly shaped for a full grid, traps were placed at 15 m spacing to cover the entire fragment up to 1.1 ha. A single Sherman live trap (9 × 8 × 23 cm) was placed at each grid intersection and a Tomahawk trap (15 × 15 × 48 cm) at every other intersection (30 m spacing). All traps were covered with plywood cover-boards for protection, locked open and pre-baited for 72 hours. Traps were set between 16:00 and 18:00 for three consecutive nights and checked the following morning between 7:00 and 12:00.
All captured animals were weighed, sexed, and assessed for reproductive condition, marked with Monel ear tags and released at point of capture. Shrews were marked by fur clipping. For mice and chipmunks we report densities based on MNA, using the size of the grid plus a 7.5 m buffer zone as the effective trapping area when trap stations were more than 7.5 m from the fragment edge (i.e., fragments larger than approximately 1.1 ha). This general approach has been criticized as overestimating densities in very small fragments because effective trapping area is limited by the size of the fragment when animals in these fragments may be using the surrounding matrix as habitat (Wolf and Batzli 2002). To avoid this bias, in September 2005 we conducted trapping in 5 non-forest matrix types that surround our fragments to determine the probability that a species would be captured in each matrix type. When the probability of capturing an individual of the target species in a given matrix was greater than 10% of the capture probability within the fragment, we considered that matrix type to be suitable habitat for that species and included the buffer zone in our calculation of effective trapping area. If the probability of capturing the species in the matrix was less than 10%, we did not include the portion of the buffer zone that extended into the matrix in our calculation of estimated trapping area. This conservative approach should eliminate any bias associated with very small fragments. As many fragments were bordered by multiple matrix types and the probability of capture in a given matrix varied with species, effective trapping areas were not constant across species for a given fragment.
All other mammalian species were assigned to broad activity-density categories based on capture success and, in some cases, camera trapping. The actual densities assigned to the categories were taken from the literature (Appendix B Table B1). The number of categories for each species ranged from 3 to 6 depending on the variation in capture success and the published density range for each species.
During September 2003 and 2004 infrared triggered cameras were operated in each fragment for 14 consecutive days. Fragments greater than 2 ha were equipped with 2 cameras on opposite sides of the trapping grid and smaller fragments with 1 camera. In week 1, cameras were baited with a carnivore scent lure (Canine Call). In week 2, cob corn and raw chicken parts were added in a non-reward fashion to draw in as many additional species as possible (Foresman and Pearson 1998, Carbone et al. 2001). Camera data were standardized by the number of functional camera-nights to correct for malfunctions, and each fragment was assigned an activity index for each species based on the number of nights that the species was photographed and the maximum number of identifiable individuals detected in a given night.
White-tailed deer (Odocoileus virginianus) abundance was estimated, in part, via standardized pellet group counts. In late winter 2004, we conducted deer pellet counts on all fragments using a technique modified from Campbell et al. (2004). Final abundance estimates for white-tailed deer were determined through an index based on (in order of importance) 1) pellet group counts; 2) camera index (Jacobson et al. 1997, Koerth et al. 2000); 3) standardized observations by the field crews (frequency of encountering deer and group size); and 4) density estimates provided by the state wildlife departments.
Densities of forest bird species (including the 4 ground dwelling birds that were used in the model, American robin, ovenbird, veery and wood thrush) were assessed during the 2003 and 2004 breeding seasons via point counts and transect counts, respectively. We believe that the 2004 data are more accurate since point counts can overestimate occurrence on small fragments. Therefore we used 2004 data in the model with the assumption that densities of the 4 model species do not vary greatly interannually. Counts were conducted in the mornings between 5:00 and 10:00. Each fragment was surveyed 3 times in June with a minimum of 7 days between samples. Parallel transects were established running approximately 50m interior to the edge of the forest fragment (where fragment size allowed) and spaced 150m apart such that song sampling would cover the entire area of all fragments less than 10 ha. Transects in fragments greater than 10 ha were laid out to sample half of the fragment area. In each survey, transects were walked at a predetermined speed and the locations of all ground-foraging ground-nesting birds were mapped based on visual or vocal identification. These maps (n=3) were combined for estimating territory locations. Presence/abundance was based on the number of territorial males * 2 (i.e., + female). We used transects instead of point counts to avoid a potential bias created by the inclusion of edge species in small fragments and exclusion of those species in larger fragments. Densities were corrected for season as in LoGiudice et al. (2003).
Brisson, D., and D. E. Dykhuizen. 2004. ospC Diversity in Borrelia burgdorferi: Different Hosts Are Different Niches. Genetics 168:713–722. (doi: 10.1534/genetics.104.028738)
Campbell, D., G. M. Swanson, and J. Sales. 2004. Comparing the precision and cost-effectiveness of faecal pellet group count methods. Journal of Applied Ecology 41:1185–1196.
Carbone C., S. Christie, K. Conforti, T. Coulson, N. Franklin, J. R. Ginsberg, M. Griffiths, J. Holden, K. Kawanishi, M. Kinnaird, R. Laidlaw, A. Lynam, D. W. Macdonald, D. Martyr, C. McDougal, L. Nath, T. O'Brien, J. Seidensticker, D. J. L. Smith, M. Sunquist, R. Tilson, and W. N. W. Shahruddin. 2001. The use of photographic rates to estimate densities of tigers and other cryptic mammals. Animal Conservation 4:75–79.
Falco, R. C., and D. Fish. 1992. A comparison of methods for sampling the deer tick, Ixodes dammini, in a Lyme disease endemic area. Experimental and Applied Acarology 14:165–177.
Foresman, K., and D. E. Pearson. 1998. Comparison of proposed survey procedures for detection of forest carnivores. Journal of Wildlife Management 62:1217–1226.
Jacobson, H. A., J. C.Kroll, R. W.Browning, B. H. Koerth, and M. H Conway. 1997. Infrared-triggered cameras for censusing white-tailed deer. Wildlife Society Bulletin 25:547–556.
Koerth, B. H., and J. C. Kroll. 2000. Bait type and timing for deer counts using cameras triggered by infrared monitors. Wildlife Society Bulletin 28:630–635.
LoGiudice, K., R. S. Ostfeld, K. A. Schmidt, and F. Keesing. 2003. The ecology of infectious disease: Effects of host diversity and community composition on Lyme disease risk. Proceedings of the National Academy of Sciences 100:567–571.
Magurran, A. E. 1988. Ecological Diversity and Its Measurement. Princeton University Press. Princeton, New Jersey, USA.
Ostfeld, R. S., E. M. Schauber, C. D. Canham, F. Keesing, C. J. Jones, and J. O. Wolff. 2001. Effects of acorn production and mouse abundance on abundance and Borrelia burgdorferi infection prevalence of nymphal Ixodes scapularis ticks. Vector Borne and Zoonotic Diseases 1:55–63.
Wolf, M., and G. O. B. Batzli. 2002. Effects of forest edge on populations of white-footed mice Peromyscus leucopus. Ecography 25:193–199.