2009 NORTHEASTERN NATURALIST 16(1):141–147 Persistence of Dragonfly Exuviae on Vegetation and Rock Substrates Maria A. Aliberti Lubertazzi1,* and Howard S. Ginsberg1,2 Abstract - Surveys of dragonfly exuviae have been used to assess rare species’ habitats, lake water quality status, and wetland restoration programs. Knowledge of the persistence of exuviae on various substrates is necessary to accurately interpret exuvial surveys. In 2006, we recorded exuvial persistence at defined areas in a variety of small freshwater wetlands in Rhode Island. Exuviae were field-identified, labeled with small daubs of nail polish, and observed every three weeks from June through September. Overall, exuvial persistence displayed exponential decline, disappearing rapidly during the first few weeks, and more slowly thereafter. The initial rate of decline was similar for most species, but differed in some taxa. There was no significant difference in exuvial retention on emergent vegetation vs. rock substrate. Introduction Dragonflies (Odonata: Anisoptera) are hemimetabolous insects that spend the majority of their lives as aquatic nymphs. Upon transformation to adulthood the last-instar nymphs emerge from the water, and ecdysis occurs when a suitable substrate has been found. The shed nymphal exoskeleton—or exuviae (singular and plural; Needham et al. 2000)—is left behind. Dragonflies have traditionally been studied by conducting adult or nymphal (= “larval”; Needham et al. 2000) surveys. Exuvial surveys, however, hold the potential for substantial, direct analyses of the dragonfly communities because exuviae indicate that the individuals sampled developed in the wetland of interest. Furthermore, exuvial surveys have low impact on the local population because live individuals are not removed or disturbed, and volunteers can be easily trained to collect them. Pupal midge exuviae (Diptera: Chironomidae) have been used for stream water quality assessment (e.g., Ruse 1995), and dragonfly exuviae have been used to study seasonal emergence patterns of individual species (Corbet 1999). However, there have been relatively few surveys of dragonfly exuviae as potential faunistic monitoring tools at diverse emergence sites. Pollard and Berrill (1992) conducted intensive exuvial surveys to assess lake water quality status, Foster and Soluk (2004) used exuviae to monitor an endangered dragonfly species, and K.H. Gaines (unpubl. data) used exuviae to census the dragonfly populations of rare, fragile desert pothole ecosystems. Exuvial surveys of dragonflies have also been used to evaluate wetland restoration (D’Amico et al. 2004) and habitat quality of recently constructed 1Department of Plant Sciences/ Entomology, University of Rhode Island,Woodward Hall, Kingston, RI 02881. 2USGS Patuxent Wildlife Research Center, Coastal Field Station, University of Rhode Island, Woodward Hall – PLS, Kingston, RI 02881. *Corresponding author - mariaaa@mail.uri.edu. 142 Northeastern Naturalist Vol. 16, No. 1 wetlands (Chovanec and Raab 1997) in Europe. To accurately interpret these and similar studies, it is necessary to know how long the exuviae persist on rock and vegetation substrates. Knowledge of exuvial persistence will help to determine optimal sampling frequency, and can be used to calibrate population estimates based on exuvial surveys. In this paper, we assess exuvial persistence for several dragonfly taxa on rock and vegetation substrates. We then discuss the effects of exuvial persistence on the interpretation and limitations of exuviae-based studies of odonate biology. Methods Five Rhode Island wetlands were chosen in June 2006 for the exuviae retention study, based on the presence of discreet potential emergence sites. Three sites (CCRIWarwick, Phelps Pond, Slater-Gazebo Pond) contained anthropogenic rock-like structures (e.g., stone walls, concrete supports) that emerged directly from the water. Two sites (Strathmore, BristolSk) with abundant emergent vegetation were also selected (Table 1). Four of the five sites were initially visited between June 15–27, then July 5–25, July 28–August 15, and August 18–September 7; all but one site (CCRIWarwick) were visited a fifth time between 1–13 September. The first visit to the fifth site (Slater-Gazebo) occurred when sufficient water was present (August 21), and there was one follow-up visit on 13 September. Areas with emergent structures (e.g., cement planks, stands of cattails, etc.) were selected on the initial visit. These sample substrates were thoroughly examined for dragonfly exuviae, which were then visually identified to species- or genus-level and marked with daubs of brightcolored nail polish. We used photographs and detailed diagrams of the sample substrates to record location and species-group of each individual. Exuviae that were present at subsequent visits received additional daubs of nail polish, with each visit represented by a unique color. Table 1. Sample sites for 2006 dragonfly exuviae retention study. See text for explanation of taxon categories. Site Name (Town) Survey Substrate Dates Taxa Rock Substrates CCRIWarwick Stone wall 27 June, 25 July, EPI-LIBEL, (Warwick) 15 August, 7 Sept SYVISE Phelps Pond 3-sided cement 20 June, 10, and 28 July CEEL, EPI-LIBEL, (West Greenwich) structure 18 August, 1 Sept SYVISE Slater-Gazebo Pond Cement decorative 21 August, 13 Sept TRAMEA (Pawtucket) stone wall Vegetation Substrates Strathmore Typha, Sagittaria 15 June, 5 and 29 July, Anax junius only (Narragansett) 18 August, 12 Sept BristolSk Juncus, Phragmites, 23 June, 10 and 31 July, CEEL, EPI-LIBEL (Barrington) Typha 21 August, 13 Sept 2009 M.A. Aliberti and H.S. Ginsberg 143 All exuviae data were compiled after the last visit of the season. The number of color-coded individuals of each species-group was quantified by visit. Loss of exuviae from substrates was characterized by fitting curves to the proportion of exuviae remaining through time using EXCEL. We measured time in terms of the number of time periods since the exuviae were first marked. The initial visit was counted as number 1, with each time period (between visits) being about three weeks. This sampling period was utilized because this study was part of a larger project (M.A. Aliberti Lubertazzi, unpubl. data) in which dragonfly populations were sampled at numerous sites with roughly three weeks between visits to each site. Data were analyzed using BIOMstat, version 3.3 (Rohlf and Slice 1999). Differences in declines of exuviae of different species through time were analyzed using R x C tests (row by column G-tests) of independence (Sokal and Rohlf 1985), and differences in persistence on rock vs. vegetation substrates were tested by 3-way ANOVA using log-linear models (presence x substrate x time period). Results Species-groups consisted of the following: CEEL = primarily Celithemis elisa (Hagen) (Calico Pennant; Libellulidae), SYVISE = Sympetrum vicinum (Hagen) (Autumn Meadowhawk)/S. semicinctum (Say) (Band-winged Meadowhawk) (Libellulidae), TRAMEA = Tramea sp. (gliders; Libellulidae), and EPI-LIBEL = Epitheca-Libellula (Corduliidae: Libellulidae). Some exuviae of the genera Epitheca (baskettails; Corduliidae) and Libellula (skimmers; Libellulidae) are often of similar size, and are not easily separable in the field, especially when remaining attached to the substrate. The exuviae of Anax junius (Drury) (Common Green Darner; Aeshnidae), the only species that was marked at the Strathmore site, are not analyzed separately here; however, they were included in the presence vs. substrate vs. time analysis (Fig. 1). The interval between site visits was roughly 3 weeks (overall average = 21 ± 3.2 days SD). Exuviae were initially lost rapidly from the sample substrates, with declines leveling out after the first few weeks (Fig. 2). The declines for CEEL, EPI-LIBEL, and SYVISE exuviae gave close fits to an exponential decline model (Table 2). We had only two sample times for TRAMEA (initial sample Table 2. Statistical models for loss of exuviae from substrate. See text for explanation of taxon categories. Proportion remaining = (coefficient)e(exponent)(# periods) Taxon Coefficient Exponent R2 CEEL 1.3370 -0.5263 0.951 EPI-LIBEL 1.2265 -0.2642 0.885 SYVISE 1.2345 -0.4803 0.758 TRAMEA 4.5455 -1.5141 144 Northeastern Naturalist Vol. 16, No. 1 Figure 1. Overall proportion of exuviae remaining through time (number of 3-week periods after marking exuviae) on rock vs. vegetation substrates. Figure 2. Proportion of exuviae remaining at each follow-up site visit; A) CEEL; B) EPI-LIBEL; C) SYVISE; D) TRAMEA. Time period = 3 weeks. and a second sample three weeks later), but we fit the data to an exponential decline model for consistency with the other taxa (Fig. 2). The initial rate of decline (proportional decline after one period) differed among species groups (R x C test: G = 28.015, df = 3, P = 0.0000036), with 2009 M.A. Aliberti and H.S. Ginsberg 145 CEEL, SYVISE, and TRAMEA (G = 4.710) and SYVISE and EPI-LIBEL (G = 6.003) forming non-significant subsets. Thus, short-term retention of EPI-LIBEL exuviae differed from that of CEEL and TRAMEA. There was no significant 3-way interaction between exuvial presence x substrate x time period (3-way ANOVA using log-linear models: G = 2.194, df = 3, P = 0.533), and in each time period, exuvial presence was independent of substrate type (G = 2.263, df = 4, P = 0.6875). Therefore, persistence of exuviae did not differ on rock vs. vegetation substrates (Fig. 1). Discussion Loss of exuviae was rapid over the first three weeks for all species, but differed among species groups, with least decline in EPI-LIBEL. EPI-LIBEL species tend to be larger in size than CEEL and SYVISE, but smaller than TRAMEA. Therefore, we detected no consistent relationship between body size and persistence. Persistence of exuviae did not differ significantly on vegetation vs. rock emergence substrates. Our results suggest that there are no consistent effects of dragonfly body size or substrate type on exuvial persistence, but more comprehensive sampling with larger sample sizes and additional taxa might reveal subtle differences that we did not detect. We did not specifically study position of the substrate, but that aspect might be important because wind action can be stronger on more exposed compared to sheltered areas. For example, at two of the study sites with rock substrates, we noted that exuviae tended to persist longer in areas protected from the wind. Benke and Benke (1975) performed daily collections of exuviae to provide a close measurement of the total number and diversity of successfully emerging dragonflies along a stretch of shoreline. Wissinger (1988) also utilized daily collections of exuviae in his survey of the dragonfly fauna in an Indiana farm pond over several years. In addition to virtually year-round surveys of nymphs, he collected exuviae daily for one field season, and every 3 days the next year. An attrition experiment indicated a 15% discrepancy in emergence quantification between 3-day and 1-day intervals when sampling exuviae. Our results are consistent with Wissinger’s because they also indicate rapid declines through time. Interestingly, Wissinger’s species emergence phenologies from one wetland are very similar to those compiled from 3-week exuvial surveys at multiple wetlands in Rhode Island (M.A. Aliberti Lubertazzi, unpubl. data). The relatively rapid loss of exuviae in this study suggests that non-daily exuvial surveys typically record only a partial sample of the individuals of a species-group emerging from a given wetland. Furthermore, exuvial samples might be biased toward certain species groups, because certain taxa differed from others in the rapidity of loss from the substrate. For most taxa in our study, more than half of the exuviae were lost in three weeks. Therefore, species with brief and synchronous seasonal emergences could be underrepresented (if emergence occurred soon after a sample) or overrepresented (if emergence occurred just before a sample) in samples taken three weeks apart. D’Amico et al. (2004) sampled exuviae and adults at 10 sites every 146 Northeastern Naturalist Vol. 16, No. 1 two weeks, and found similarities, but also some differences, between the exuvial and adult surveys. Collection of both types of data allowed a more comprehensive interpretation of odonate population status at treated and reference study ponds, even with samples taken only every two weeks. Nevertheless, our results suggest that samples should be taken as frequently as possible to reduce unknown biases in detection of individual species. One of the advantages of surveys conducted with exuviae is that species can be detected whose other life stages are difficult to collect in the field. For example, Benke and Benke (1975) found that one of the most abundant dragonfly species emerging from their study pond (Libellulidae: Perithemis tenera (Say) [Eastern Amberwing]) was not common in extensive nymphal surveys of the pond. Thus, some common species might not be detected in wetlands if their aquatic life stage inhabits hard-to-sample areas, such as profundal zones. The adults of some river species are rarely seen near emergence areas, and collection of their exuviae has provided useful documentation of their presence, abundance, and habitat use (Orr 2006). Ruse (1995) reported a similar phenomenon in comparative samples of chironomid larvae and pupal exuviae in chalk-gravel streams, where the exuvial surveys documented species whose larvae inhabit macrophyte stands that are minimally included in larval surveys. Benke and Benke (1975) found a >90% pre-emergence mortality rate for most species in intensive surveys of a pond’s nymphal dragonfly population from hatching through late-instar stages. Hence, exuvial surveys can be the best measure of a pond’s adult productivity. We should point out, however, that exuviae of some species can be located in cryptic sites, or in sites distant from the water, and can thus be difficult to detect. By collecting exuviae, surveys of rare taxa have documented successful reproduction of an individual species at a site. For example, Somatochlora hineana Williamson (Hine’s Emerald Dragonfly; Corduliidae), whose nymphal life stage lasts several years before adult emergence, can be sampled with low-impact population surveys by collecting exuviae (Foster and Soluk 2004). Thus, surveys targeting either habitat (i.e., wetlands) or species status (e.g., establishment, conservation, restoration, etc.) can potentially benefit from exuviae-based sampling. Acknowledgments Special thanks to M. Freeman and S. Droege for constructive comments on early drafts of the manuscript. K.H. Gaines provided excellent advice and inspiration with regard to the merits of focusing on exuviae. Literature Cited Benke, A.C., and S.S. Benke. 1975. Comparative dynamics and life histories of coexisting dragonfly populations. Ecology 56:302–317. Chovanec, A., and R. Raab. 1997. Dragonflies (Insecta, Odonata) and the ecological status of newly created wetlands: Examples for long-term bioindication programmes. Limnologica 27:381–392. 2009 M.A. Aliberti and H.S. Ginsberg 147 Corbet, P.S. 1999. Dragonflies: Behavior and Ecology of Odonata. Cornell University Press, Ithaca, NY. D’Amico, F., S. Darblade, S. Avignon, S. Blanc-Manel, and S.J. Ormerod. 2004. Odonates as indicators of shallow lake restoration by liming: comparing adult and larval responses. Restoration Ecology 12:439–446. Foster, S.E., and D.A. Soluk. 2004. Evaluating exuvia [sic] collection as a management tool for the federally endangered Hine’s Emerald Dragonfly, Somatochlora hineana Williamson (Odonata: Corduliidae). Biological Conservation 118:15–20. Needham, J.G., J.M.J. Westfall, and M.L. May. 2000. Dragonflies of North America, Revised Edition. Scientific Publishers, Gainesville, fl. Orr, R. 2006. Volunteers collect dragonfly cast skins for survey along Potomac River. P. 109, In Natural Resource Year In Review: 2005. A Portrait of the Year in Natural Resource Stewardship and Science in the National Park System. National Park Service, US Department of the Interior, Denver, CO. Pollard, J.B., and M. Berrill. 1992. The distribution of dragonfly nymphs across a pH gradient in south-central Ontario lakes. Canadian Journal of Zoology 70:878–885. Rohlf, F.J., and D.E. Slice. 1999. BIOMstat for Windows. Statistical Sofware for Biologists. Version 3.3. Exeter Software, Setauket, NY. Ruse, L.P. 1995. Chironomid community structure deduced from larvae and pupal exuviae of a chalk stream. Hydrobiologia 315:135–142. Sokal, R.R., and F.J. Rohlf. 1985. Biometry. W.H. Freeman and Co., New York, NY. Wissinger, S.A. 1988. Life history and size structure of larval dragonfly populations. Journal of the North American Benthological Society 7:13–28