USGS
Pwrconly.gif (1828 bytes)


PWRC

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

Back to Showcases

Back to NAAMP Page
Effects of Artificial Cover Objects
on Salamander Counts

CHRISTOPHER A. BINCKLEY1, PAIGE EAGLE1, LAURA MONTI2, BODO PLESKY1, KATRIN WERNER1, DANNY BYSTRAK3, MIKE QUINLAN4, PATRICK CONGDON5, GENE SPEARS6, NATE CARLSON7, AND SAM DROEGE1

1 USGS Patuxent Wildlife Research Center (Patuxent), 12100 Beech Forest Dr., Laurel, MD 20708-4038
2 9 Knapp St., Apt. 1, Somerville, MA, 02143
3 Mallard Lane, Lothian, MD 20715
4 12508 Caswell Lane, Bowie, MD 20715
5 Star Route 2, Box 319, La Honda, CA 94020
6 Lees-McRae College, Natural Science and Mathematics, Banner Elk, NC 28604
7 W 301 N. Second St., Spring Valley, WI 54767

INTRODUCTION

Accurately measuring the abundance of organisms is essential for conducting basic ecological and conservation studies. This has become especially important for amphibians due to recent increases in reported population declines (Blaustein et al., 1994) and with long-term studies documenting large natural population fluctuations (Pechman et al., 1991; Pechmann and Wilbur, 1994). There are a diversity of monitoring techniques for amphibians, all of which have their own inherent biases in counts of individuals. One monitoring technique for terrestrial salamander populations is the use of artificial cover objects (ACOs). This technique has been applied in only a handful of studies and thus remains largely untested (Fellers and Drost, 1994). Which types and sizes of ACOs will optimize population counts is unknown for most, if not all, species of terrestrial salamanders. Wood ACOs have been the primary test objects used to date (Stewart and Bellis, 1970; Fellers and Drost, 1994; Davis, 1997; Bonin and Bachand, 1997), while two studies have used other ACO types (tin and wood, Grant et al., 1991; tar paper, stones, and wood, Taub, 1961).  In experiment 1, we tested the effects of 9 different ACO types on salamander counts. We chose primarily non-wood ACOs because they were untested as potential cover objects, would not decay, and could be easily obtained. Experiment 2 examined 3 types of wood ACOs, 2 of which allowed a greater influx of environmental moisture. Finally, a third experiment tested the effect of repeatedly checking wood ACOs on the number of salamanders captured.

MATERIALS AND METHODS

Experiment 1. Nine different types of ACOs were tested at 10 sites throughout the continental United States. ACOs used were 2 red bricks (RB2-.04 m2), 8 red bricks (RB8-.16 m2), 2 z bricks (TB2-.04 m2), 8 z bricks (TB8-.16 m2), cement blocks of 2 sizes (B1-.09 m2 and B2-.16 m2), a dinner plate (DP), a dark green plastic trash bag (PB-1.0 m2), and an oak wood board (Wood-.08 m2). Twelve arrays, with 1 array consisting of 1 of each of the 9 types of ACOs, were constructed at each site. ACO type, species of salamander, and its corresponding SVL (mm) were recorded when each site was checked. Number of checks and date of check varied between sites. Count data was analyzed by site by removing days when no salamanders were caught under any ACO, and with and without correcting for the size of the ACO.

Experiment 2. Twenty five of 3 types of oak wood ACOs (75 total) were randomly placed 6 m apart in a predominantly beech (Fagus grandifolia) forest at Patuxent. ACOs used were (A) a square wooden board (.09 m2); (B) 4 equal sized smaller square blocks placed together to form a larger square with the same area as an A board (.09 m2); and (C) 8 equal-sized smaller square blocks placed together to form a larger square with twice the area as an A or B board (.18 m2). From September 16, 1997 to May 14, 1998, all ACOs were checked weekly. We hypothesized that B and C boards would receive more moisture, thus increasing salamander counts. All salamanders (P. cinereus only) were measured for SVL (mm) and their color phase was recorded.

Experiment 3. Fifty pairs of oak wood ACOs (.09 m2) were randomly placed in a beech (F. grandifolia) dominated forest at Patuxent for a total of 100 boards. The 2 boards constituting a pair were only several cm apart, while the pairs themselves were 6 m apart. On day 1, one board of each pair was randomly checked by flipping a coin. On day 2, both boards from pairs 1-25 were checked while the same board from day 1 was checked for pairs 26-50. The same board chosen on day 1 was checked again on day 3 for pairs 26-50, while on day 4, all boards were checked for pairs 26-50. All P. cinereus were measured for SVL (mm) and color phase was recorded. Sites were visited every 2 weeks.

RESULTS

Experiment 1. Of the 10 sites, four sites (Laurel, MD (Patuxent); Jug Bay, MD; Banner Elk, NC;  La Honda, CA) exhibited high enough captures of salamanders in order to conduct statistical analyses. The number of salamanders captured under each type of ACO at each site, with and without correcting for ACO size, is shown in Fig. 1. At Patuxent, arrays were checked 37 times and 148 P. cinereus were captured. Jug Bay arrays were checked 33 times and 317 P. cinereus were captured. At Banner Elk, NC, 10 checks caught 109 salamanders (Desmognathus ochrophaeus-74.3%, P. cinereus-19.3%, P. jordani-5.5%, and P. glutinosus-0.9%), while La Honda, CA, arrays were checked 8 times for a total of 104 salamanders (Batrachoseps attenuatus-85.6%, Ensatina eschscholtzi-6.7%, and Taricha torosa-7.7%).

At Patuxent, ACO type, size-corrected ACO type and array #, all had significant effects on salamander counts (Fig. 1a). Similar results were found at Jug Bay, MD (Fig. 1b). Using D. ochrophaues data from Banner Elk, NC, ACO type had a significant effect on salamander counts, but this effect was removed when ACO size was corrected for (Fig. 1c). At Banner Elk, array # had no effect on the number of salamanders captured. At La Honda, CA, ACO type, size-corrected ACO type, and array #, all had significant effects on B. attenuatus counts (Fig. 1d).

Exp1as.jpg (14939 bytes) Exp1bs.jpg (16447 bytes)

Figure 1 a,b,c,d - Total # of salamanders captured with and without correcting for the size of the ACOs. The sites include (a) USGS Patuxent Wildlife Research Center (Patuxent) - Laurel, MD, (b) Jug Bay, MD, (c) Banner Elk, NC, and (d) La Honda, CA. ACOs used were 2 red bricks (RB2-.04 m2), 8 red bricks (RB8-.16 m2), 2 z bricks (TB2-.04 m2), 8 z bricks (TB8-.16 m2), cement blocks of 2 sizes (B1-.09 m2 and B2-.16 m2), a dinner plate (DP), a dark green plastic trash bag (PB-1.0 m2), and an oak wood board (Wood-.08 m2). In general, larger ACOs caught more salamanders, except in MD where wood ACOs captured the greatest numbers of P. cinereus. When size of ACO was corrected for, the "effectiveness" of the larger ACOs decreased dramatically.

Exp1cs.jpg (16417 bytes) Exp1ds.jpg (16249 bytes)

Experiment 2. 402 P. cinereus (294 red and 108 lead) were captured under 3 types of wood ACOs (Figs. 2, and 3). A 3-way ANOVA comparing season (fall and spring; F = 23.4, df = 1,66, p < .001), ACO type (A,B,C; F = 19.4, df = 2,66, p < .001), and color morph (F = 56.6, df = 1,66, p < .001) revealed that all of these factors had significant effects on salamander counts when not correcting for size of the ACO. Both A (p=.02) and C (p<.001) boards captured significantly more salamanders than B boards but did not differ from each other (p=.58, Tukey-Kramer). The season * ACO type interaction (p=.04) and ACO * morph interaction (p=.006) were both statistically significant (Fig 3).

Season (F = 21.5, df = 1,66, p < .001), ACO type (F = 7.7, df = 2,66, p = .001), and color morph ( F = 59.6, df = 1,66, p < .001) all had a significant effect on salamander counts under size-corrected ACO type CA (p=.01) boards captured significantly more salamanders than B boards (Tukey-Kramer). The ACO * morph interaction (p=.002) was also statistically significant (Fig 3).

Exp22.jpg (71889 bytes) Figure 2 - Total # of P. cinereus captured under 3 different types of wood ACOs. ACO type
(A,B,C; F = 19.4,  df = 2,66, 
p < .001)  had a significant effect on salamander counts.  A (p=.02) and C (p<.001) captured significantly more salamanders than B boards when not correcting for size of the C board. Using size-corrected C ACOs in the analysis, A ACOs captured significantly more salamanders than B ACOs
(p=.01, Tukey-Kramer).

Ill1brd.jpg (47304 bytes)

Figure 3 - More "red-backed" P. cinereus were captured than "lead-backed" and the color morphs used the ACOs differently. The ACO type * morph interaction was significant using both the uncorrected size C data (p=.006) and the corrected size C data (p=.002). Red and lead morphs may exhibit different preferences for  thermal and hydric properties associated with different ACOs. ACO type usage changed with season (Season* ACO type interaction, p=.04) only when using the uncorrected size C data. P. cinereus were more often captured under C in the Fall and under A in the Spring. Exp23.jpg (51176 bytes)

Experiment 3. Preliminary results suggest that repeated checking of ACOs might reduce the counts of terrestrial salamanders after 3 days (Fig 4).

Exp34.jpg (50824 bytes)  

Figure 4 -Total # of P. cinereus captured after repeated checking of ACOs for 2 and 4 days compared to total # of P. cinereus captured after checking ACOs once. Only 3 tests were performed, making analysis impossible at this time. However, it appears that repeated checking of ACOs might decrease salamander counts after 3 daily checks.

Ill2brd.jpg (56936 bytes)

DISCUSSION

Experiment 1. More P. cinereus were found under wood ACOs than other ACO types at both the Patuxent and Jug Bay sites, even when area of ACOs was corrected for (Fig. 1). Wood ACOs attracted the most salamanders in CA, but several ACO types might be effective for D. ochrophaeus in NC. Stewart and Bellis (1970) used wood ACOs for sampling Desmognathus and Eurycea. ACOs that decay over time could cause abiotic and age variation in ACO experiments (Bonin and Bachand, 1997). Decaying ACOs has been a problem for other researchers (Grant et al., 1991; Monti, 1997). Our results argue against non-wood ACOs, but decay-induced bias may be inherent to this sampling technique. The question of interspecific variation in ACO preference still remains relatively unanswered and untested.

Experiment 2. Type A ACOs caught significantly more P. cinereus than B ACOs and more than C ACOs when the area of C was accounted for (Fig 2, 3). The slits in the B and C ACOs might have induced greater moisture and temperature variation. However, abiotic variables were not quantified under the 3 types of ACOs. The significant season * ACO interaction with the size-uncorrected data set was because more salamanders were captured under C in the Fall and under A in the Spring (Figs. 2, 3). Larger boards captured more salamanders during the period of greatest abundance (Fall) but not in the Spring. The significant ACO * color morph interaction was caused by more red P. cinereus using A ACOs and more lead P. cinereus using C ACOs (Fig. 3). This may be due to differences between the color morphs in physiology (Moreno 1989) and in preference for different microclimatic conditions under the different ACO types.

Experiment 3. Repeatedly checking ACOs might reduce the number of salamanders captured (Fig 4). These 3 experiments demonstrate that documenting the optimal ACO type, size, and check frequency for a particular species will reduce bias in population counts. This will increase the ability to detect population trends and allow predictions on how populations will respond to human-induced and natural changes in their environment.


USGS Patuxent Home Page