modified interactions between salamander life ... - Semantic Scholar

ecol 79

Monday Jan 05 03:28 PM Allen Press • DTPro

REPORTS

740

203 Mp 740 File # 03sc

Ecology, Vol. 79, No. 2

Ecology, 79(2), 1998, pp. 740–745 q 1998 by the Ecological Society of America

MODIFIED INTERACTIONS BETWEEN SALAMANDER LIFE STAGES CAUSED BY WILDFIRE-INDUCED SEDIMENTATION J. LAWRENCE KERBY

AND

LEE B. KATS1

Natural Science Division, Pepperdine University, Malibu, California 90263 USA

Abstract. A 1993 wildfire and subsequent landslides modified many streams in the Santa Monica Mountains of southern California (USA). Prior to the fire at Cold Creek Canyon, adult California newts (Taricha torosa) frequently preyed on conspecific eggs and larvae. Post-fire landslides increased the number of stream pools containing terrestrial earthworms. Earthworms were more common in adult newt diets after the fire, and conspecifics were absent. More earthworms and fewer conspecifics were present in the stomachs of adult newts in streams at burned sites than at unburned sites. In laboratory experiments, newt larvae used refuges significantly less in the presence of combined chemical cues from both newt adults and earthworms as compared to adult-newt cues alone. These data suggest that cannibalism is reduced in the presence of increased alternative prey items and that larvae can detect this reduced predation risk. Key words: alternative prey; California newt; cannibalism, California newts; chaparral wildfire; diet preferences; earthworms; mudslides; newt cannibalism; salamander life stages, interactions; Santa Monica Mountains, California, USA; Taricha torosa.

INTRODUCTION Ecological communities are commonly affected by natural and periodic disturbance events. These disturbances may kill organisms, change competitive outcomes, alter habitats, or introduce new species to existing communities (for a review see Sousa [1984]). If a disturbance suddenly adds a new species to a community it may be possible to directly observe the effects of this new species on the interactions of existing species, particularly if the dynamics between existing species had been observed before the introduction (Adler and Morris 1994). The effects of wildfire as a disturbance on chaparral plant communities are well studied (Whelan 1995). Wildfires also impact animal communities but studies examining animal–fire interactions generally consider only short-term population changes in small mammals and birds (e.g., Lawrence 1966). Despite studies on the physical changes to stream habitats following fire (Booker et al. 1993), little is known about the effects of wildfire on aquatic organisms (for an exception see Young and Bozek [1993]). Wildfire may indirectly affect aquatic systems by altering the surrounding terrestrial habitat. Fires may cause loss of vegetation, producing soil instability and landslides that change stream geomorphology (Minshall et al. 1989). After Manuscript received 8 November 1996; revised and accepted 18 April 1997. 1 Address correspondence to this author.

rains, suspended sediments are often higher in streams flowing through burned areas (Minshall and Brock 1991). If organisms are introduced into the stream as a result of sedimentation, interactions among members of the stream community could be modified. We have taken advantage of a major chaparral wildfire to examine if fire indirectly impacts the feeding ecology of the California newt (Taricha torosa). Adults breed in perennial streams that flow through the chaparral communities of the Santa Monica Mountains of southern California. Adult T. torosa spend most of the year on land and return to the streams to breed each spring (Stebbins 1972). Females oviposit on sticks and the undersides of rocks; egg masses are estimated to hatch 30–45 d later (Elliott et al. 1993). While many adult T. torosa return to land after breeding, some remain in stream pools for 3–6 wk after oviposition (Elliott et al. 1993). During this time, adult T. torosa are known to cannibalize both larvae and egg masses (Kaplan and Sherman 1980, Marshall et al. 1990, Kats et al. 1992). Cannibalism is apparently a major selection pressure in T. torosa because newt larvae hide from chemical cues of conspecific adults (Elliott et al. 1993, Kats et al. 1994). Adult T. torosa are also opportunistic predators on a variety of aquatic prey and terrestrial organisms that fall into streams (Ritter 1897, Hanson et al. 1994). In many vertebrates, the presence of an alternative food source reduces the frequency of canni-

ecol 79


REPORTS

March 1998

balism (Elgar and Crespi 1992). Environmental factors (e.g., prey abundance) may affect cannibalism by adult T. torosa. In 1993 a chaparral wildfire burned one of our primary study sites in southern California. After the fire we made preliminary observations that unstable stream banks resulted in the input of earthworms into streams. Given that adult T. torosa are known to feed on earthworms (Stebbins 1972), the purpose of this study was to examine whether an influx of earthworms into a stream would modify interactions (Wooton 1994) between life stages of T. torosa. In addition, if the rate of cannibalism varies from year to year depending on habitat changes related to wildfire, T. torosa larvae would benefit by hiding less during periods when cannibalism is low or absent. We use diet analysis, field surveys, and a laboratory experiment to determine the effects of chaparral wildfire-induced sedimentation on cannibalism in T. torosa. METHODS

Field study We collected data in three perennial streams in the Santa Monica Mountains, Los Angeles County, California, USA. One of these streams, Cold Creek Canyon, burned in a wildfire on 2 November 1993. Prior to the 1993 fire, Cold Creek had not burned in over 20 yr. During the winter rains of 1993–1994 mudslides and sedimentation changed Cold Creek dramatically. Two other sites (Newton Creek Canyon, Trancas Creek Canyon) have not burned in the last 10 yr and served as controls. We used water lavage to collect stomach contents of adult Taricha torosa during the spring and summer (April–July) of 1992–1996 in Cold Creek. Stomach contents were also collected at Trancas Creek Canyon in the spring and summer of 1995. We recorded gender, snout–vent length (SVL), and the location of each adult T. torosa examined. Stomach contents were examined under a microscope and all items were counted, classified taxonomically to order, and preserved in 70% ethanol. Total number of stomach lavages with and without specific prey items (newt conspecifics, earthworms, coleopterans, ephemeropterans) were compared before and after wildfire. Stomach contents in pre-fire years (1992, 1993, 1994) were combined and compared to combined post-fire years (1995, 1996) using G tests. Stomach contents (1995) from newts in Trancas Creek were compared to Cold Creek using a Fisher’s exact test. We also surveyed streams in both burned and unburned sites for earthworms (Eisenia rosea; Gates 1942, Eddy and Hodson 1982). Six pools in each unburned site were monitored over the summers of 1995 and 1996 in four separate surveys each year. Twenty-

203 Mp 741 File # 03sc

741

one pools were similarly monitored in the burned site. We recorded the dimensions of each pool and carefully searched for earthworms with a dive mask and/or an underwater periscope. Only earthworms found on the surface of the stream bed were included in each count. All pools monitored contain adult T. torosa during the breeding season. The number of pools containing earthworms from the two streams that did not burn were combined and compared to Cold Creek using a G test. To be confident that any observed interaction modifications between T. torosa life stages were related to earthworm influx we monitored the availability of T. torosa larvae and egg masses before and after the Cold Creek fire. Total T. torosa egg-mass numbers decreased after the Cold Creek fire; however, because of changes in stream morphology and pool sizes after fire, densities of egg masses in stream pools increased slightly and egg masses remained conspicuously oviposited in stream pools (Gamradt and Kats 1997). The density of adults returning to the streams in years before or after fire also did not differ (Gamradt and Kats 1997). During the same time period that we were sampling T. torosa stomach contents we counted T. torosa larvae in six stream pools in Cold Creek before (1993) and in seven stream pools after (1994) wildfire.

Laboratory study The behavioral responses of larval T. torosa to chemical cues were examined using the gravitational flowthrough system described by Petranka et al. (1987) and modified by Elliott et al. (1993). This system consisted of three plastic tubs (45 3 32 cm) arranged in sequence at differing heights, with a 20-L bucket collecting effluent at the lowest position. Dechlorinated tap water from the uppermost reservoir tub flowed into lower tubs via aquarium tubing (inside diameter: 0.4 cm) at a rate of 0.5–0.6 L/min. The tub directly beneath the reservoir contained the differing treatments. The treatments consisted of a control and three chemical-cue tubs containing either one adult T. torosa (n 5 4 individuals; SVL 5 7.07 6 0.22 cm [mean 6 1 SE]), 16 earthworms (n 5 32 individuals; mean length 5 5.53 6 0.69 cm [mean 6 1 SE]), or a combination of the two. These densities of adult newts and earthworms were within those that we observed in field pools. Each treatment was replicated 10 times. Earthworms were placed in plankton mesh netting (1 mm) to prevent consumption by adult T. torosa. Previous studies had found that newt larvae do not respond to nylon netting alone (L. B. Kats, unpublished data); consequently, that treatment was not included in this experiment. In corresponding treatments, the T. torosa adult and earthworms were added to condition the water 30 min prior to experimentation. Below each treatment tub was located an experimental tub containing a smoked plexiglas refuge

ecol 79


REPORTS

742

203 Mp 742 File # 03sc


plate (20 3 30 cm) raised 1.0 cm from the bottom by screws fastened at the corners. Five naive, laboratoryraised T. torosa larvae collected from Cold Creek (n 5 200 individuals; mean length 5 15.38 6 0.07 mm [mean 6 1 SE], Harrison stages 41–43) were placed at the edge of each refuge plate in the tubs and given 10 min to acclimate. Larvae were used only once. Water was then allowed to flow for 2 min before data were collected. Percentages of the number of larvae in the open were recorded every 30 sec for 20 min. To minimize observer disturbance, larvae were watched via a video camera above the experimental tubs. The order of treatment replicates was randomly selected, with two sets of tubs running at the same time. The percentages of time that larvae spent in the open were arcsinetransformed for statistical analysis. We analyzed the data using a two-way ANOVA. RESULTS

Field study In Cold Creek Canyon adult Taricha torosa consumed significantly different numbers of earthworms (G 5 50.82, df 5 1, P , 0.01) and conspecific larvae (G 5 39.63, df 5 1, P , 0.01) before and after wildfire (Fig. 1A). Cannibalism that existed before the fire (1992, 1993) was absent in the two years immediately afterwards and then reappeared in 1996 (Fig. 1A). A corresponding increase in the frequency of earthworms in adult T. torosa stomachs occurred after the fire. We counted fewer newt larvae in stream pools in 1994 (post-fire) than we did before the fire (1993), but the difference was not significant (1993: density [mean 6 1 SE] 5 3.0 6 1.1 larvae/m2; 1994: density 5 1.44 6 0.32 larvae/m2; t 5 1.47, df 5 11, P 5 0.17). The presence of other major diet items in adult stomachs did not change before and after the fire (Coleoptera: G 5 1.07, df 5 1, P 5 0.30; Ephemeroptera: G 5 2.00, df 5 1, P 5 0.16; Fig. 1B). In 1995, the presence of earthworms and T. torosa larvae in stomach contents of adult T. torosa was significantly dependent on whether newts were in burned or unburned sites (earthworms: Fisher’s exact test P , 0.01; T. torosa larvae: Fisher’s exact test P , 0.01). Stomachs of adult T. torosa in the burned site contained more earthworms than stomachs of adults found at an unburned site (Fig. 2). Conspecifics were present in nearly 25% of adult T. torosa stomachs from an unburned stream site, while none were present in adults from a burned site. Diet analyses indicated that both male and female adults cannibalize T. torosa larvae or egg masses. In Cold Creek (1992, 1993), males represented 54.5% of the larval and 33.3% of egg mass cannibalism. In Trancas Creek (1995), males represented 66.6% of the larval and 66.6% of egg-mass cannibalism.

FIG. 1. Stomach contents of: Taricha torosa in Cold Creek, Los Angeles County, California, USA; n 5 no. of adult newt stomachs flushed. (A) Percentage of T. torosa stomachs containing earthworm (Eisenia rosea) and conspecific prey between 1992 and 1996. Fire occurred after the breeding season in late 1993. (B) Percentage of T. torosa stomachs containing coleopterans and ephemeropterans between 1993 and 1995.

We compared the frequency of earthworms in stream pools between two unburned sites and one burned site (Table 1). We found a significant difference in the numbers of pools with and without earthworms (G 5 11.77, df 5 1, P , 0.01). We also found a significant difference in the number of pools with and without earthworms in the burned site (Cold Creek) between 1995 and 1996 (G 5 20.45, df 5 1, P , 0.01; Table 1).

Laboratory study The significant T. torosa effect indicates that T. torosa larvae hid the most when they were exposed to effluent from conspecific adults (F1,36 5 5.14, P 5 0.03;

ecol 79


REPORTS

March 1998

203 Mp 743 File # 03sc

743

TABLE 1. umbers of pools with and without earthworms (Eisenia rosea) at three streams in the Santa Monica Mountains, Los Angeles, California, USA. Four separate surveys were conducted: Trancas, Newton, and Cold Creek (two years postfire) surveys occurred in 1995; Cold Creek (three years postfire) survey occurred in 1996. Earthworms Stream

Present

Absent

Trancas Creek (unburned) Newton Creek (unburned) Cold Creek (two years postfire) Cold Creek (three years postfire)

0 1 18 1

24 23 66 83

DISCUSSION

alternative prey item. While overall the number of T. torosa egg masses has decreased since the fire, eggmass densities did not decline (Gamradt and Kats 1997) and egg masses and larvae were readily available as prey items for adult T. torosa. On several occasions after the fire we observed adult T. torosa encountering relatively conspicuous egg masses and ignoring them. In addition, data from the burned site in 1996 indicates that cannibalism has reappeared although T. torosa egg-mass numbers are comparable to 1994 and 1995 (Gamradt and Kats 1997). Reappearance of cannibalism in 1996 is consistent with the reduced number of earthworms found in the stream as well as the low frequency of earthworms in stomachs of adult T. torosa. As vegetation on burned hillsides continues to recover we predict that conspecifics will again become a common prey of adult T. torosa. Comparisons between burned and unburned sites further support the idea that adult T. torosa quit consuming conspecifics

Wildfires are major disturbances to plant and animal communities (Lawrence 1966, Whelan 1995). In addition to the direct effect of fire on communities, changes in habitat associated with fires can affect communities for years after the actual burn. On 2 November 1993 a wildfire burned the chaparral vegetation of Cold Creek Canyon in the Santa Monica Mountains. The slopes were devegetated by fire, and the unstable hillsides eroded during winter rains. The rock, soil, and earthworms associated with the soil that were once supported by vegetation, roots, and leaf litter fell into Cold Creek under the force of gravity after fire (Spitler 1995, Gamradt and Kats 1997). Prior to the fire, conspecifics were a major dietary item of adult Taricha torosa in Cold Creek. Beetles (Coleoptera) and mayflies (Ephemeroptera) were also diet components, but their frequencies were similar before and after the fire. In contrast, conspecifics were completely eliminated from the diets of adult T. torosa in the two years after the fire. We also found an increase in the percentage of earthworms in the diet of adult T. torosa. Conspecifics were apparently replaced by an

FIG. 3. Response of larval Taricha torosa to chemical cues in laboratory flow-through experiments. Treatments are all possible combinations of the presence and absence of adult newts and earthworms. Data plotted are means.

FIG. 2. Percentage of Taricha torosa stomachs containing earthworm (Eisenia rosea) and conspecific prey in 1995; n 5 no. of adult T. torosa from unburned (Trancas Creek) and burned (Cold Creek) stream sites in Los Angeles County, California, USA.

Fig. 3). There was not a significant effect of earthworms (F1,36 5 2.63, P 5 0.11). However there was a significant interaction effect, indicating that larval hiding was dependent on whether or not earthworm cues were present (F1,36 5 6.42, P 5 0.016). Newt larvae hid more when conspecific adult cues were presented alone than when larvae were exposed to both conspecific cues and earthworm cues.

ecol 79


744

REPORTS

with the appearance of earthworms. Based on our studies on adult T. torosa both within a population (in Cold Creek before and after the fire) and our comparison of T. torosa in two separate populations (Cold Creek and Trancas Creek), interactions between adult T. torosa and larvae are modified with the input of earthworms. Ideally, adult feeding preferences could be assessed in controlled laboratory experiments; however, adult T. torosa feed readily in the laboratory only after extended captivity. Wildfire produced indirect ecological changes in a stream inhabited by T. torosa. Our data indicate that an influx of earthworms produced changes in feeding behavior of T. torosa and modified interactions between T. torosa life stages. Previous data from our study site in Cold Creek Canyon indicated that adult T. torosa cannibalize both larvae and egg masses (Kats et al. 1992). In this study, we sampled two separate T. torosa populations and found that both male and female adult T. torosa cannibalize larvae and egg masses. This is inconsistent with previous literature (Marshall et al. 1990). The fact that both male and female adult T. torosa prey on larvae further illustrates that adults are a predation pressure on T. torosa larvae. If predation pressure on larvae is reduced after a fire because alternative prey become available, then selection should favor T. torosa larvae that detect the alternative prey, recognize that overall predation risk is low, and reduce potential costly hiding behavior. Increased time spent hiding from adults might result in lost foraging time and delayed metamorphosis for the larvae. T. torosa larvae presumably must assess overall risk of predation based on the presence of adult conspecifics in pools. For many amphibians, waterborne chemical cues serve as the primary mechanism for predator detection (Petranka et al. 1987, Kats et al. 1988). Our data indicated that when larvae are exposed to cues from T. torosa and Eisenia rosea simultaneously, the larvae do not use refuge significantly more than in control treatments. This indicates that T. torosa larvae are not only able to detect the presence of adult conspecifics, but perhaps detect the reduction in predation risk from adults when earthworms are present in pools. Regardless of the mechanism mediating this change in hiding behavior (i.e., cues from earthworms may mask adult T. torosa cues), the implications are the same. Larval T. torosa are better off not hiding when earthworms are present, given that adult T. torosa feed primarily on earthworms. Historically, fires probably occurred far less frequently than today (Radtke et al. 1982). Increasing fire frequency may have increased selection pressure on T. torosa larvae to distinguish levels of conspecific predation risk from summer to summer. We realize that the strength of before-and-after studies has been debated because of the lack of replication

203 Mp 744 File # 03sc


and often the lack of good controls (Stewart-Oaten et al. 1992). In this study, we are confident that the patterns we observed are common after fire. Sedimentation is a common occurrence after wildfire (Minshall et al. 1989) and it follows that the input of earthworms would accompany mudslides and sedimentation in other burned streams. In addition, our observations on unburned areas confirm that earthworm introductions are related to fire. Future studies that include observations before and after wildfire could lead to a better understanding of how major natural disturbances have both direct and indirect effects on animal interactions. ACKNOWLEDGMENTS We are especially grateful to S. Gamradt, J. Goodsell, R. Rodriguez, S. Majoy, S. Elliott, K. Hanson, R. L. Buckskin, and M. Kerby for their help in collecting field data. D. Chivers, J. Kiesecker, A. Blaustein, A. Sih, and T. Vandergon provided helpful comments on the study. D. Green, R. Syrdahl, M. Thatcher, B. Hutchinson, and C. Vos Strache provided valuable logistic support. We thank J. Kitz and the Mountains Restoration Trust for supporting our research at Cold Creek. We also thank R. Sauvajot and the National Park Service for their support of the study. This study was funded by grants from the Pepperdine University Research Council and the National Science Foundation (BIR-9225034, SGER941029). LITERATURE CITED Adler, F. R., and W. F. Morris. 1994. A general test for interaction modification. Ecology 75:1552–1559. Booker, F. A., W. E. Dietrich, and L. M. Collins. 1993. Runoff and erosion after the Oakland firestorm; expectations and observations. California Geology 46:159–173. Eddy, S. A., and A. C. Hodson. 1982. Taxonomic keys to the common animals of the north central states. Burgess, Minneapolis, Minnesota, USA. Elgar, M. A., and B. J. Crespi. 1992. Ecology and evolution of cannibalism. Pages 1–12 in M. A. Elgar and B. J. Crespi, editors. Cannibalism: ecology and evolution among diverse taxa. Cambridge University Press, New York, New York, USA. Elliott, S. A, L. B. Kats, and J. A. Breeding. 1993. The use of conspecific cues for cannibal avoidance in California newts (Taricha torosa). Ethology 95:186–192. Gamradt, S. C., and L. B. Kats. 1997. Impact of chaparral wildfire-induced sedimentation on oviposition of streambreeding California newts (Taricha torosa). Oecologia: 110:546–549. Gates, G. E. 1942. Check list and bibliography of North American earthworms. American Midland Naturalist 27: 86–118. Hanson, K., J. Snyder, and L. B. Kats. 1994. Natural history note: Taricha torosa diet. Herpetological Review 25:62. Kaplan, R. H., and P. W. Sherman. 1980. Intraspecific oophagy in California newts. Journal of Herpetology 14:183– 185. Kats, L. B., J. A. Breeding, K. M. Hanson, and P. Smith. 1994. Ontogenetic changes in California newts (Taricha torosa) in response to chemical cues from conspecific predators. Journal of the North American Benthological Society 13:321–325. Kats, L. B., S. A. Elliott, and J. Currens. 1992. Intraspecific

ecol 79


March 1998

REPORTS

oophagy in stream-breeding California newts (Taricha torosa). Herpetological Review 23:7–8. Kats, L. B., J. W. Petranka, and A. Sih. 1988. Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology 69:1865–1870. Lawrence, G. E. 1966. Ecology of vertebrate animals in response to chaparral fire in the Sierra Nevada foothills. Ecology 47:278–291. Marshall, C. J., L. S. Doyle, and R. H. Kaplan. 1990. Intraspecific and sex-specific oophagy in a salamander and a frog: reproductive convergence of Taricha torosa and Bombina orientalis. Herpetologica 46:395–399. Minshall, G. W., and J. T. Brock. 1991. Observed and anticipated effects of forest fire on Yellowstone stream ecosystems. Pages 123–135 in R. B. Keiter and M. S. Boyce, editors. The greater Yellowstone ecosystem: redefining America’s wilderness heritage. Yale University Press, New Haven, Connecticut, USA. Minshall, G. W., J. T. Brock, and J. D. Varley. 1989. Wildfires and Yellowstone’s stream ecosystems. BioScience 39:707– 715. Petranka, J. W., L. B. Kats, and A. Sih. 1987. Predator–prey interactions among fish and larval amphibians: use of chemical cues to detect predatory fish. Animal Behaviour 35:420–425. Radtke, K. W.-H., A. M. Arndt, and R. H. Wakimoto. 1982. Fire history of the Santa Monica Mountains. Pages 86–90 in C. E. Conrad and W. C. Oechel, editors. Proceedings of the Symposium on the Dynamics and Management of Med-

203 Mp 745 File # 03sc

745

iterranean-type Ecosystems. USDA Pacific Southwest Forest and Range Experiment Station Technical Report PSW58. Ritter, W. E. 1897. The life-history and habits of the Pacific coast newt (Diemyctylus torosus Esch.). Proceedings of the California Academy of Science-Zoology 1:73–114. Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15: 353–391. Spittler, T. E. 1995. Fire and debris flow potential of winter storms. Pages 113–120 in J. E. Keeley and T. Scott, editors. Brushfires in California: ecology and resource management. International Association of Wildland Fire, Fairfield, Washington, USA. Stebbins, R. C. 1972. Amphibians and reptiles of California. University of California Press, Berkeley, California, USA. Stewart-Oaten, A., J. R. Bence, and C. W. Osenberg. 1992. Assessing effects of unreplicated perturbations: no simple solutions. Ecology 73:1396–1404. Whelan, R. J. 1995. The ecology of fire. Cambridge University Press, Cambridge, England. Wootton, J. A. 1994. Putting the pieces together: testing the independance of interactions among organisms. Ecology 75:1544–1551. Young, M. K., and M. A. Bozek. 1993. Postfire impacts on coarse woody debris and adult trout in northwestern Wyoming streams. Page 43 in D. Despain and P. Schullery, editors. Agenda and abstracts, conference on the ecological implications of fire in greater Yellowstone. National Park Service, Yellowstone National Park, Wyoming, USA.