Patuxent Wildlife Research Center

NAAMP III Archive - accepted papers
Home | Archive by Alphabetical Order | Archive by Category

A Test of the Suitability of Various Cover Objects for Monitoring Terrestrial Salamander Populations

Laura Monti

University of Maine
9 Knapp Street #1
Somerville MA 02143
email: lemmonti@aol.com

Introduction
Traditionally, methods used for censusing terrestrial salamanders have included pitfall traps, funnel traps, transect walks (low intensity visual encounter surveys) and quadrat searches. However, these techniques are not ideal for monitoring long-term population trends. The surface activity pattern of these animals is often dictated by immediate atmospheric conditions, creating temporal differences in pitfall and funnel trappings that may obscure long term trends. Transect counts can yield small numbers of salamanders and are weather-dependent Quadrat searches are destructive and can give results dependent upon microsite anomalies (Vogt and Hines 1982, Witham et al. 1993, Brooks 1996).

An alternative technique is currently being investigated by a number of researchers. It involves artificial cover objects (ACO's) placed in the forest and monitored regularly. Advantages to this method include: low time and labor costs, lack of habitat destruction, repeatability at the same site, potential for large sample sizes, minimization of between-observer bias, durability of the cover items, relative safety of the animals, and standardization of the number and area of cover items (Fellers and Drost 1994).

Before ACO's can be used in a large-scale monitoring effort, several issues must be addressed. Others are assessing such potential confounding factors as environmental variables and social interactions under the objects. Our concern was to test this method for a variety of species across North America and to determine the most effective type of cover item.

Study Sites:
After posting a call for volunteers in several herp-related newsgroups on the internet, a pool of seven groups of volunteers was gathered. Each group established sets of test arrays, described below. One set was established in a mixed evergreen forest in the Skyline Ridge Open Space Preserve in La Honda, California. Two were established in Maine; one in an oak-pine forest on the mid-coastal island of Arrowsic, the other in a beech forest bordered by a meadow near Bangor. In North Carolina, six arrays were set in an abandoned apple orchard overgrown with locust and maple, the other six in a beech-hemlock forest. A test site was established in the Ted Shanks Conservation Area in Missouri. In Spring Valley, Wisconsin, arrays were set out along a footpath bordered by willows. Finally, a site was installed in Woodbury, MN.

Materials and Methods:
Objects evaluated included: 4"X8" red smooth common bricks and face bricks (also called facade or z bricks), 12"X12"X2" and 8"X16"X2" cement construction blocks, dinner plates, plastic lawn and leaf bags, and wooden boards. An array was composed of a set of eight common bricks arranged in two rows of four bricks, a similar arrangement of face bricks, a set of two common bricks placed side-by-side, a similar pair of face bricks, a single 12"X12" cement block, a pair of 8"X16" blocks (placed to form a 16"X16" square), two slabs of wood at least 12" X 1.5", two plates placed 5cm below the surface of the forest floor, and a single 1m X 1m square dark plastic trash bag weighed down by rocks, sticks or jute staples. The eleven sets of cover items were arranged in two parallel rows; objects within the rows were three feet apart, the rows were also three feet apart (Fig. 1). Each site held 12 complete arrays, spaced at least ten feet apart. Because of some confusion over details and difficulty in obtaining some cover objects, not all arrays exactly fit this description.

Cover objects were checked approximately once a month throughout the summer and early fall. Volunteers attempted to perform counts 3-5 days after a significant rainfall (1/2" or more). Species, SVL, and the cover object it was under were recorded for each salamander encountered. Volunteers were asked to note signs of disturbance and the presence of ants and small mammal burrows under the objects.

Results and Discussion:
At several sites, very few salamanders were found under the ACO's. Three sites had no salamanders at all (the Bangor, Maine site, the Ted Shanks Conservation Area in Missouri, and the Minnesota site) These poor results underscore the necessity of verifying the presence of terrestrial salamanders in an area before beginning a monitoring program. An apparently suitable area may not harbor enough salamanders to warrant an array. Another study has found that cover objects require several months to "settle in" before redback salamanders (Plethodon cinereus) will use them to their full potential (see Fig. 2) (Monti unpublished data). In fact, early in this season, one volunteer found redbacks under natural logs, while nearby arrays were unoccupied. Later in the summer, however, the salamanders began using the ACO's. In light of these observations, I suggest any ACO-based monitoring program allow for an initial trial year to confirm the presence of salamanders and to allow objects to "settle in". Data from this year should not be used to estimate population density.

In terms of user-friendliness, the heavy cover objects had at least two detractors among the volunteers. I found the masonry too unwieldy to transport far. Another volunteer commented that the difficulty of carrying these heavy objects made it impossible to bring them to remote areas that might be of interest in a comprehensive monitoring program. The volunteer also found these items fragile in spite of their bulk. After an unfortunate incident in which I experimented with the airworthiness of a pickup truck on a bumpy road, I found several bricks and concrete blocks had shattered. A drawback to the wood, which I found to be more portable and sturdy, is its tendency to rot and need replacement. Some concern had been expressed that newer boards would not be utilized as extensively as weathered boards. In another study, however, I determined that old boards and new boards are equally attractive to salamanders; hence replacing boards should not significantly alter the array from a salamander's perspective (Monti, MS thesis-in-progress). Garbage bags tended to dry out underneath, according to the Wisconsin team and my own observations. However, bags worked well at the La Honda, CA site.

Many of the objects were utilized by ants, small mammals, spiders, scorpions and other creatures that might deter salamanders. Some species constructed burrow entrances; others built nests or webs. It was feared that aggressive ants, predatory shrews and the like might prevent salamanders from occupying the objects. A graph of the number of objects noted to harbor ants suggests that the masonry was more likely to attract these potential salamander foes (see Fig. 3). However, ants, as well as spiders, worms, beetles, centipedes and snails, were found at least once with a salamander under the same object. One researcher has determined from food studies that ants are the preferred prey item for several species of plethodontids. At her study site in the Allegheny National Forest she has frequently encountered ants and salamanders under the same board (Dr. L. Ordiway, email comm.). It appears, then, that most of the other organisms found under the objects are not completely exclusive of salamanders.

Unfortunately sparse data made statistical analysis difficult. The following represents a basic attempt at quantitative examination of the data available. I thought about other ways of addressing this data (count data techniques like loglinear modelling), but did not investigate them closely. Helpful comments regarding more appropriate tests would be most appreciated. It should be noted that the objects were treated as equals in this analysis; no adjustments were made to compensate for size differences or for differences in the number of parts an object had (i.e two bricks versus eight bricks).

To get an idea of the overall preference regardless of geographic location, all sites where P. cinereus were encountered (to eliminate between-species differences) were considered. These sites were: the two Maryland sites; Spring Valley, WI; Arrowsic, ME; and Banner Elk, NC. Using the largest count for each site, data for all twelve arrays were lumped and a single-factor ANOVA was run with SYSTAT v.5. A significant difference between objects was detected (p=0.002) and a Lillefore's test on the residuals indicated that they were normally distribute. Based on the raw data (Fig 4), it appears that these salamanders favored the wood boards.

A second series of analyses were run to determine the preferences of different species within sites. For this analysis, ANOVA's using a repeated measures approach were employed. For Desmognathus ochrophaeus at the Banner Elk site, there was a significant effect of object on number encountered (p=0.06). Neither date, nor interaction of date and object were significant (p= 0.194, 0.312 respectively). Preferred objects were the eight thin and the eight regular bricks (Fig. 5). Unfortunately, sample sizes were too small to allow similar analysis of the remaining data.

Qualitative observations based on visual representations of the counts suggest the following: Batrachoseps attenuatus appeared to show some preference for wood boards in La Honda, CA. However, the bags worked the best for ensatinas (Fig. 6). For Jug Bay, P. cinereus show a strong preference for the wood boards (Fig 7a). The other Maryland site also reflects this trend (Fig. 7b). The Arrowsic data suggests some preference of P. cinereus for the wood boards and both types of bricks in the two-brick formation (Fig. 8). The Wisconsin redbacks were found most frequently under regular bricks. (Fig. 9). In summary, it seems that wood boards are preferred by redbacks, with bricks a close second. Garbage bags do not work well in the east, but were very successful in California. The heavy cement blocks were not worth the extra effort required to lug them around.

In conclusion, I will use that classic phrase "further study is needed". I recommend continuing this project next summer. If the current volunteers are willing to continue this generous donation of time and effort, we stand to have a much longer field season next summer. Also, the initial "breaking in" period will hopefully have ended and the objects will be utilized to a greater extent. One of the volunteers has developed an incredible data management program that will facilitate record-keeping and analysis. I suggest posting this program (provided the volunteer in question is willing to share his work) along with another request for volunteers at the AMP website. Although any additional sites will probably not yield much data initially, it would be helpful to get some information on more species in other parts of the country.

Acknowledgements
I wish to extend my deep gratitude to all the volunteers who maintained arrays. These folks were completely unselfish of their time and energy. They participated in this program solely to benefit the salamanders. I also want to thank Sam Droege for his unending energy, patience and helpful advice.

Literature Cited
Brooks, R.T. 1996. Terrestrial salamander monitoring. North American Amphibian Monitoring Program online document. //amphib/naampsala.html

Fellers, G.M. and C.A. Drost. 1994. Sampling with artificial cover. pp.147-150 In Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. W.R.Heyer, M.A. Donnelly, R.W. McDiarmid, L.C. Hayek and M.S. Foster, Eds. Smithsonian Institution Press: Washington, D.C

Vogt, R. and R.L. Hine. 1982. Evaluation of techniques for assessment of amphibian and reptile populations in Wisconsin. pp. 201-217 In N.J. Scott (Ed). Herpetological Communities.U.S. Fish and Wildlife Service Wildlife Research Report. Washington, D.C.

Witham, J.W., E.H. Moore, M.L. Hunter, Jr., A.J. Kimball and A.S. White. 1993. A Long-Term Study of an Oak-Pine Forest Ecosystem: Techniques Manual for the Holt Research Forest. Maine Agricultural Experiment Station, University of Maine, Technical Bulletin 153.

U.S. Department of the Interior
U.S. Geological Survey
Patuxent Wildlife Research Center
Laurel, MD, USA 20708-4038
http://www.pwrc.usgs.gov/naamp3/naamp3.html
Contact: Sam Droege, email: Sam_Droege@usgs.gov
Last Modified: June 2002