ISSN: 0098-4590 «2 “Florida Scienti Summer, 2003 Volume 66 CONTENTS Effect of Season and Shoot Removal on Root Carbohydrate Storage in a Subtropical Invasive Shrub, Rhodomyrtus tomentosa ................ Jennifer E. Possley, Kaoru Kitajima, and Randall K. Stocker 157 A Survey of Epifauna Among Macrophytes in a Southwest Florida Estuary Paul J. Rudershausen, James V. Locascio, and Lourdes M. Rojas 168 Register of an Exceptionally Large Redbellied Pacu, Piaractus brachy- pomus (Teleostei, Characidae), in East-Central Florida, with Gonad a UE OMS Fen, SSNS OU cuidslld dws aupivs saacdaeeedeserbetonmmnndens Ramon Ruiz-Carus and S. Barry Davis 184 A Pedestrian Road Survey of the Southern Hognose Snake (Heterodon Mea rteriAnGO COUNtY, FIOTICa :......252...2.0cco.ecceeeecnnseecconecneneds Kevin M. Enge and Kristin N. Wood 189 The Distribution of Hemidactylus (Saura: Gekkonidae) in Northern RET FE 2 OE an ea ee Se ne eRe Josiah H. Townsend and Kenneth L. Krysko 204 Cues Used by the Golden Mouse, Ochrotomys nuttalli, to Assess the SME MEMINT OT NMOSEIMALIC POY 23.61... soe cs oh diene cecscecdseeeteces cen ones Fred Punzo 209 A Comparison of Diagnostic Methods Used to Detect Intestinal Parasites in Residents of the Province of Chalatenango, El Salvador ............ Michael A. Silverman, Raymond K. Zlamal, Carlos Cruz, Patricia Dipatrizio, Carol J. Palmer, Donald E. Burris, Scott Schatz, and Harold E. Laubach 217 The Madagascar Giant Day Gecko, Phelsuma madagascariensis grandis Gray 1870 (Sauria: Gekkonidae): A New Established Species in Florida Kenneth L Krysko, A. Nichole Hooper, and Coleman M. Sheehy III 222 Pathologic Findings in Stranded Atlantic Bottlenose Dolphins (Tursiops truncatus) from the Indian River Lagoon, Florida .......................04. Gregory D. Bossart, René Meisner, René Varela, Marilyn Mazzoil, Stephen D. McCulloch, David Kilpatrick, Robin Friday, Elizabeth Murdoch, Blair Mase, and R. H. Defran 226 The Suitability of Shrews (Blarina carolinensis, Sorex palustris) for Papetimcurs On Spatial Learning Tasks. .......,......00.0.c0.ciececeseenoeseees Renn FELICE OL SCICUCES WICKAIIS(S 2.2... 0.s0scs.cccvoeceacssdeaachssecessveoess Os RE NE EN, 200, 98s Jud Jy Fic ow Ju shicosded juss dads oiitlseaversccavoassees 252 FLORIDA SCIENTIST QUARTERLY JOURNAL OF THE FLORIDA ACADEMY OF SCIENCES Copyright © by the Florida Academy of Sciences, Inc. 2003 Editor: Dr. Dean F. Martin Co-Editor: Mrs. Barbara B. Martin Institute for Environmental Studies, Department of Chemistry, University of South Florida, 4202 East Fowler Avenue, Tampa, Florida 33620-5250 Phone: (813) 974-2374; e-mail: dmartin@chumal.cas.usf.edu Business Manager: Dr. Richard L. Turner Department of Biological Sciences, Florida Institute of Technology, 150 West University Boulevard, Melbourne, Florida 32901-6975 Phone: (321) 674-8196, e-mail: rturner@fit.edu http://www. floridaacademyofsciences.org The Florida Scientist is published quarterly by the Florida Academy of Sciences, Inc., a non-profit scientific and educational association. Membership is open to in- dividuals or institutions interested in supporting science in its broadest sense. Ap- plications may be obtained from the Executive Secretary. Direct subscription is avail-. able at $45.00 per calendar year. Original articles containing new knowledge, or new interpretations of knowl- edge, are welcomed in any field of science as represented by the sections of the Academy, viz., Biological Sciences, Conservation, Earth and Planetary Sciences, Medical Sciences, Physical Sciences, Science Teaching, and Social Sciences. Also, contributions will be considered which present new applications of scientific knowl- edge to practical problems within fields of interest to the Academy. Articles must not duplicate in any substantial way material that is published elsewhere. Contri- butions are accepted only from members of the Academy and so papers submitted by non-members will be accepted only after the authors join the Academy. Instruc- tions for preparations of manuscripts are inside the back cover. Officers for 2003—2004 FLORIDA ACADEMY OF SCIENCES Founded 1936 President: Dr. Cherie Geiger Treasurer: Mrs. Georgina Wharton Department of Chemistry 11709 North Dr. University of Central Florida Tampa, FL 33617 Orlando, FL 32816 President-Elect: Dr. John Trefry ° Department of Oceanography Florida Institute of Technology 150 W. University Boulevard Melbourne, FL 32901 Past-President: Barry Wharton Executive Director: Dr. Gay Biery-Hamilton Rollins College 1000 Holt Ave., 2761 Winter Park, FL 32789-4499 Rebecca Amonett, Secretary e-mail: GBiery-Hamilton@osc.org HDR Engineering, Inc. Program Chair: Dr. Jeremy Montague 2202 N. Westshore Boulevard Department of Natural and Health Sciences Suite 250 Barry University Tampa, FL 33607-5711 Miami Shores, FL 33161 Secretary: Dr. Elizabeth Hays Barry University Miami Shores, FL 33161-6695 Published by The Florida Academy of Sciences, Inc. Printing by Allen Press, Inc., Lawrence, Kansas Florida Scientist QUARTERLY JOURNAL OF THE FLORIDA ACADEMY OF SCIENCES DEAN F. Martin, Editor BARBARA B. MartTIN, Co-Editor Volume 66 Summer, 2003 Number 3 Biological Sciences EFFECT OF SEASON AND SHOOT REMOVAL ON ROOT CARBOHYDRATE STORAGE IN A SUBTROPICAL INVASIVE SHRUB, RHODOMYRTUS TOMENTOSA (2) (1) JENNIFER E. Posstey’*, KAoru KitasimMA™, AND RANDALL K. STOCKER “University of Florida, Department of Agronomy, Center for Aquatic and Invasive Plants, 7922 NW 71“ St., Gainesville, FL 32653 University of Florida, Department of Botany, PO Box 118526, Gainesville, FL 32611 ABSTRACT: Rhodomyrtus tomentosa (downy rose myrtle) is a shrub native to southeast Asia that invades pine flatwoods of central and southern Florida, where it can form a dense, nearly monodominant understory. We studied the phenology, shoot extension and carbohydrate allocation of this little-studied invader at the University of Florida and in Jonathan Dickinson State Park (JDSP) in 1998-2000. R. tomentosa flowered April through June and produced fruits August through September. Shoot extension was greatest in September and lowest January through May. Tissue concentration of carbohydrate reserves in roots was relatively high (up to 300 mg/g) and did not change substantially with season, except for a June peak in total nonstructural carbohydrate (TNC) concentration in lateral roots. Lateral roots had a significantly higher concentration of TNC than root crowns (lignotubers) throughout the year. Only TNC of lateral roots showed a positive correlation with plant size. The consistently high TNC concentration in roots of R. tomentosa indicates its high resprouting potential throughout the year. Based on these results, we recommend that shoot removal by cutting, burning or herbicide is optimally done in early August before fruits ripen and become dispersed. Shoot removal then would result in rapid shoot regrowth and depletion of root TNC, because August—October is the main season for shoot extension. Repeated shoot killing 6-12 months later will maximize the chance of killing plants completely. Key Words: Rhodomyrtus tomentosa, Jonathan Dickinson State Park, invasive non-native plant, phenology, shoot extension, root total nonstructural carbohydrate THE evergreen shrub Rhodomyrtus tomentosa (Aiton) Hassk. has naturalized in southern and central Florida, USA (Fig. 1). It grows 2-4 m tall, produces edible * Author to whom all correspondence should be addressed: Fairchild Tropical Garden, 11935 Old Cutler Rd., Coral Gables, FL 33156. 17 158 FLORIDA SCIENTIST [VOL. 66 Rhodlom yr tus domert QGus Fic. 1. Rhodomyrtus tomentosa. Line drawing by A. Murray, Center for Aquatic and Invasive Plants, University of Florida, Gainesville. berries, and does not exhibit clonal growth. R. tomentosa invades Florida’s sandy pine flatwoods, where populations often form a dense, nearly monodominant understory. Less commonly, it invades Taxodium (cypress) stands near Naples, FL (Alexander, 1981). It is listed by the U.S. Fish and Wildlife Service as a facultative wetland species (USFWS, 1996). No. 3 2003] POSSLEY ET AL.—SUBTROPICAL INVASIVE SHRUB 159 A member of the Myrtaceae family, Rhodomyrtus tomentosa is known in Florida by the common names downy rose myrtle or downy myrtle. Two varieties of R. tomentosa are recognized: one is from southern India and Sri Lanka (var. parviflora (Alston) A.J. Scott); the other, which is the subject of this study, is from southeast Asia (var. tomentosa) (Scott, 1978). Countries to which R. tomentosa vat. tomentosa (hereafter R. tomentosa) is native include China, Vietnam, Cambodia, Laos, Thailand, Burma, Malaysia, the Philippines, and several Indonesian islands (Scott, 1978), as well as Okinawa and Ishigaki of Japan’s Ryukyu Islands (Walker, 1976). It is often erroneously reported as being native to Australia (e.g., Long and Lakela, 1976; Morton, 1976; Nelson, 1996; Wunderlin, 1998; Austin, 1999), where it is cultivated (Craven, 1999). In addition to Florida, R. tomentosa is also found outside its native range in Hawaii (Hosaka and Thistle, 1954; Haselwood and Motter, 1966; Smith, 1985). In Florida, R. tomentosa is “extremely invasive” and “‘difficult to control” (Nelson, 1996). In central Florida, it can be more aggressive than Schinus terebin- thifolius (Brazilian pepper) (Austin, 1999). In fact, Alexander (1981) compared R. tomentosa to S. terebinthifolius in its pattern of escape, local population build-up, and rapid spread into the wild. Sizable populations of R. tomentosa can be found in several counties, most notably Martin, Palm Beach, Collier, Lee, and Charlotte. It has been repeatedly introduced to Florida for use as an ornamental, as hedgerows, or to harvest the fruit (Watkins and Wolfe, 1968; Alexander, 1981; Cox et al., 1997). The earliest University of Florida herbarium specimen of R. tomentosa (#17903) was collected in 1898, near Oneco in Manatee County. A second Manatee County specimen (UF #17904) from 1916 labels the plant as a “‘roadside escape.” R. tomentosa was reportedly introduced to Highlands and Lake Counties as early as 1905, where it had escaped cultivation by 1906, and subsequently spread to most counties of central and southern Florida by the late 1970s (Austin, 1999). Finally, one author (Alexander, 1981) noted that R. tomentosa grows as far north as Putnam County. These records indicate that range expansion throughout peninsular Florida may be possible. The counties of Broward, Lee, and Palm Beach, as well as the city of Jupiter have ordinances against R. tomentosa . It is listed as a Category-I exotic pest plant by the Florida Exotic Pest Plant Council, meaning it is “‘altering native plant com- munities by displacing native species, changing community structures or ecological functions, or hybridizing with natives’ (FLEPPC, 2001). The Florida Fish and Wildlife Conservation Commission recognizes R. tomentosa as a major problem on some of their management lands (Boyter, 1994). In addition, the Florida Depart- ment of Agriculture and Consumer Services, Division of Plant Industry places R. tomentosa on its list of prohibited noxious weeds, making it unlawful for anyone to introduce, possess, move, or release R. tomentosa without a permit (FDACS, 1999). Despite recognition as a noxious weed, little ecological information on R. tomentosa exists (but see Langeland and Burks, 1999; Stocker and Possley, 2001). Existing populations of R. tomentosa are dense, and its range is expanding (Thayer and Ferriter, 1994). However, because control is expensive, South Florida Water Management District has done little to control R. tomentosa on its lands due 160 FLORIDA SCIENTIST [VOL. 66 to funding limitations (Thayer and Ferriter, 1994). Scientific information may help design more cost effective measures for controlling R. tomentosa. Our first goal in this study was to investigate the reproductive phenology and timing of shoot extension of R. tomentosa. We recorded phenology and shoot extension of plants at JDSP from September 1999 through August 2000. Our second goal was to investigate the effect of season and shoot removal on total nonstructural carbohydrate (TNC) in roots of R. tomentosa. In woody perennial plants, carbo- hydrates usually accumulate in root tissue and act as energy reservoirs (Lambers et al., 1998). The study of TNC allocation to roots is especially important in invasive plants because TNC storage allows vegetative regrowth in response to shoot re- moval. Knowledge of seasonal patterns of TNC may aid in predicting the most efficient and effective times to administer control methods of weeds (Conway et al., 1999; Becker and Fawcett, 1998; Katovich et al., 1998). MeETHODs—Description of study site—Field studies were carried out in a pine flatwoods near the northern border of Jonathan Dickinson State Park (JDSP) in Martin County, Florida, USA (27°1'N, 80°11'W). During June to October of a normal (non-drought) year, it is common to find standing water throughout these flatwoods. The R. tomentosa population at JDOSP (which was removed upon completion of this study) was relatively free of human traffic, though there was a seldom-used access road through the middle of the stand. Phenology—Phenology of R. tomentosa was assessed in the field at JDSP every 4 (when fruiting or flowering) to 6 (when vegetative) weeks from September 1999 through August 2000. Presence/absence of new vegetative growth, buds, flowers, and fruit on individual plants were recorded. Fruit presence was further classified as either green or ripe. Dried fruits from previous years’ reproduction were not included in the study. On each visit, fifty different plants were randomly selected by walking through the population and choosing every fourth plant. Only mature plants (reproductive or over | m tall) were counted. Plants were selected from as wide and diverse an area as possible, however removal of R. tomentosa by JDSP in March 2000 restricted the population to an area including only a few hundred plants. For analysis, the percentage of plants in each stage was calculated. The population was considered to be in a certain phenological stage if >5% of plants exhibited it. Shoot extension—Every 4 to 6 weeks between 28 August 1999 and 20 August 2000, shoot growth was measured for five marked branches on each of seven mature, 2 m tall R. tomentosa at JDSP (35 total branches). New flushes of green, non-lignified shoots were marked and measured on 28 August 1999 (typically <10 cm). The rate of shoot extension (cm/day) was calculated for each month by dividing the change in branch length by days since last measurement to account for slightly different time intervals between data collection. ANOVA (analysis of variance) was used to test for the effects of month and individual plant on growth. Differences between monthly growth were examined with Duncan’s multiple range test. Root carbohydrate—Bimonthly, from September 1999 to August 2000, root samples were collected from six of seventy-two mature R. tomentosa plants that were randomly pre-assigned to different sampling dates. To facilitate root excavation, we selected the smallest reproductive plants at the site; generally 1—1.5 m tall. To determine initial carbohydrate content, 5-cm-long sections of lateral root (near the root crown) were collected from each of the six plants, whose above-ground shoots were cut immediately following the sampling. Total fresh mass of shoots was measured in the field upon shoot removal. Two months later, the shoot regrowth was collected, dried and weighed, and the root crown (lignotuber) was harvested. In June and August 2000, lateral root samples were collected along with root crowns to examine the effects of shoot removal on lateral root TNC. No. 3 2003] POSSLEY ET AL.—SUBTROPICAL INVASIVE SHRUB 161 To prepare root tissue for analysis, samples were placed on ice to minimize respiratory loss of carbohydrates. Upon returning to Gainesville (approximately 5 hours), bark was removed and samples were dried in a 65°C oven for approximately | week. Dried samples were ground to pass through 1 mm mesh. Total nonstructural carbohydrates were extracted and analyzed in a three-step process: (1) simple sugars were extracted with ethanol, (2) starch and complex sugars were enzymatically hydrolyzed to glucose with amyloglucosidase, and (3) the concentration of glucose from the ethanol extraction (for sugar) and enzymatic digest (for starch) were determined colorimetrically with a phenol-sulfuric acid reaction (modified from Dubois et al. 1956, Ashwell 1966, Marquis et al. 1997). Microsoft Excel (Microsoft Corp., 1996) was used in regression analyses and SAS (SAS Institute Inc., 1996, 2000) was used in analyses of variance (ANOVA). RESULTS—R. tomentosa formed reproductive buds February through May, flowers April through June, and ripe fruits in August and September (Fig. 2A). Plants exhibited shoot extension in all months, although from March to June, new leaves were mostly limited to those associated with flower buds. Shoot extension exhibited a clear seasonal pattern over the 12 months of this study (Fig. 2B). Mean growth rates were the highest (1.5 cm per month) in September 1999, after which they decreased through March and began to rise once again in June. ANOVAs showed that month significantly affected shoot extension (Figg69 = 11.83, P < .0001), but shoot extension rate did not differ significantly among individual plants (Table 1). Tissue TNC concentration of R. tomentosa roots changed with season (Fig. 2C). One-way ANOVA showed that the effect of month on TNC was significant for lateral roots (Fy. 25 = 4.53, P = 0.0069) but not for root crowns (F4, 25 = 1.81, P = 0.1583). For all sampling dates, lateral roots had significantly higher TNC concentration than root crowns. Lateral root TNC peaked in June, and then decreased again by August. Root crown TNC fluctuated less. TNC concentration in lateral roots was positively correlated with plant size (Fig. 3) (P = 0.023), but this was not the case for root crowns (P = 0.82). Shoot removal significantly reduced lateral root TNC concentration (P = 0.0003), with a greater response to shoot removal in June than in April (Table 2, Fig. 4A). The dry mass of resprouts was very small and was not significantly affected by season (Fig. 4B). On average, plants analyzed in August happened to have greater aboveground biomass than those in June (Fig. 4C). DiscussiIoN—Fleshy-fruited members of the Myrtaceae family (subfamily Myrtoideae) have a tendency to flower at the dry/rainy season transition (Lughadha and Proenca, 1996). R. tomentosa fit this pattern, as peak flowering occurred in May, when Florida’s rainy season begins. Fruits were ripe in August and September, several months prior to the peak of winter bird migration, a potential factor in long- distance dispersal. TNC concentrations in R. tomentosa roots were relatively constant over the course of a year. Lateral root TNC increased slightly while shoot extension was minimal and flowering and fruit initiation were active (February—June). Some of this stored carbohydrate was apparently mobilized to aboveground shoots to support fruit maturation in August and shoot extension and new leaf production in September. However, relatively small fluctuations in root TNC in this evergreen 162 FLORIDA SCIENTIST [VOL. 66 (A) Ss (4) So = &S s (B) 3 . ‘ae ae S O N/D- D/J F M A M 350 (C) 300 - 400 — —— Lateral roots —— Root crowns Avg. TNC conc. (mg/g) on (o) Sep Dec Feb Apr Jun Aug Fic. 2. (A) Reproductive phenology, (B) shoot extension (N = 35 branches), and (C) TNC con- centration (mg glucose/g) of R. tomentosa lateral roots and root crowns (N = 6) at JDSP from September 1999 to August 2000. Columns sharing the same letter(s) are not significantly different (Duncan’s multiple range test at P = 0.05). Mean + s.d. Root crown TNC was not determined for September. species probably indicate that current photosynthates, and possibly TNC stored in stems, support seasonal reproductive and vegetative activities, as has been found for other tropical evergreen shrubs (Marquis et al., 1997; Tissue and Wright, 1995). Root crowns do not seem to be the main carbohydrate storage organ; they had lower concentrations of TNC than lateral roots. Further, total mass of root crown No. 3 2003] POSSLEY ET AL.—SUBTROPICAL INVASIVE SHRUB 163 TaBLeE 1. ANOVA table showing the effects of month and individual plant on average monthly growth of R. tomentosa branches, September 1999 through August 2000. Growth measurements were divided by days since last measurement to standardize for variations in visit intervals. (N = 7 plants * 5 branches = 35 branches). Source of variation df Sum of squares F-value 22 Month 10 0.01260 11.83 <0.0001 Plant 6 0.00096 1.50 0.1929 Error 76 0.00639 appeared much smaller than the lateral roots, although we did not determine total mass of lateral roots due to the logistic difficulty of completely excavating the extensive lateral root systems. Root crowns do not act as the main storage organ in other species, e.g. Lythrum salicaria (purple loosestrife) (Katovich et al., 1998). Ardisia crenata, another invasive Florida shrub, likewise maintains higher concen- trations of TNC in lateral roots than root crowns throughout the year (Kitajima, unpublished data). This lower TNC concentration of root crowns is likely due to higher degree of lignification in root crown tissue than lateral roots, reflecting the older age of root crown tissue than lateral roots, on average. 300 - . Dd) O 200 - V7 rs Zz. ov eee r e (A) 6 | (e) = 100- © | c | © — 0+ r 7 or 0 500 1000 Shoot fresh mass (g) 300 - 200 - (B) °° or % Root crown TNC (mg/g) e o ¢ 0- 0 500 1000 Shoot fresh mass (g) Fic. 3. Correlation of root TNC concentration with fresh mass of shoots for (A) lateral roots ¢ = .213, P = .02 ) and (B) root crowns (r° = .002, P = .82) of R. tomentosa. 164 FLORIDA SCIENTIST [VOL. 66 TABLE 2. Two-way ANOVA table for the effects of shoot removal (before vs. after shoot removal) on R. tomentosa lateral root TNC concentration. Source of variation df Sum of squares F-value P Shoot removal 1 293 97S 18.74 .0003 Error 22 34,507.44 It is possible that root TNC concentration of R. tomentosa exhibited only a mild seasonal change because of the drought in February to June of the study year. While temperate plant growth usually responds to changes in temperature and photoperiod, tropical shrubs (such as R. tomentosa) often are more influenced by water levels (Tissue and Wright, 1995). The drought of 2000 may have caused R. tomentosa at JDSP to grow less than it would in years with normal precipitation; consequently, root TNC may not have been depleted. Alternatively, the drought may have compromised the photosynthetic production and accumulation of TNC during this period in the study year. The ability of a plant to resprout after shoot damage is usually proportional to root TNC concentrations (Miyanishi and Kellman, 1986). Shoot removal appeared to reduce lateral root TNC concentrations more dramatically in June than in April (Fig. 4A). Given that, it is surprising that resprout biomass was not significantly different between the April and June shoot removals (Fig. 4B). One possibility is that drought reduced the respiratory demands from the roots of plants whose shoots had been removed in April. Roots of June-cut plants, in contrast, may have used stored TNC for their metabolism. Pre-cut lateral root TNC concentration was positively correlated with total shoot mass (Fig. 4, A and C). This suggests that larger plants will have greater resprouting vigor following shoot loss due to fire, cutting or other control measures. Management implications—The information reported here may help design an effective R. tomentosa management strategy in Florida. The reproductive phenology study suggests that control by shoot removal is more effective before August in order to prevent dispersal of mature seeds and establishment of another cohort of seedlings. Any public information campaigns may be best launched in late April or May, when the showy flowers are abundant. While R. tomentosa populations are currently prevalent in pine flatwoods of southern Florida, a successful control program is possible. Large populations take decades to form, since R. tomentosa does not exhibit clonal growth and individual growth rates are not extremely rapid. For the most effective control of R. tomentosa in natural areas, we recommend control via a systemic herbicide (see Stocker and Possley, 2001) or by repeated top-kill measures. Based on these results, we recommen that shoot removal by cutting, burning or herbicide is optimally done in early August before fruits ripen and become dispersed. Shoot removal then would result in rapid shoot regrowth and depletion of root TNC, because August—October is the main season for shoot extension. Repeated shoot killing 6-12 months later will maximize the chance of killing plants completely. No. 3 2003] POSSLEY ET AL.—SUBTROPICAL INVASIVE SHRUB 165 350 ; 300 - ~-_~ NO oO (2) 4 200 - 150 - 100 - 50 - 0 = TNC concentration (mg/g =) April June June August (B) ) _ N (=) on = on NO on (o%) lt | | | | =) Dry mass of new shoots (g) April cut June cut June harvest August harvest © jo) oO (C) ~“N i=) oO —l = NN Oo O oO oO es il = (eo) fo) (eo) L oO jo) io) Paes. Fresh mass of shoots (g) BSS Oo (>) 0 + - = April cut June cut Fic. 4. The effects of shoot removal on TNC concentration for R. tomentosa in spring (April) and summer (June) 2000. (A) Mean lateral root TNC concentration before (black) and after (white) shoot removal (N = 6). Post-cut measurements of TNC were determined 2 months after the shoot cutting treatment (e.g., June for April cut and August for June cut, respectively). (B) Dry mass of resprouts from stumps, 2 months post-shoot removal. (C) Fresh mass of shoots when cut. Mean = s.d. ACKNOWLEDGMENTS—We thank Chad Reddick and Bernice Lo for laboratory assistance; J.B. Miller, Mark Nelson, and the staff of Jonathan Dickinson State Park for allowing us to conduct research and assisting with logistics; Galin Jones for assistance with data analysis; Dorothy Brazis, Dan Clark, Neil Hill, Jim Myers, Steve Smith, Jackie Smith, and Mike Ward for help with field work; and David Sutton and George Tanner for helpful comments on the manuscript. 166 FLORIDA SCIENTIST [VOL. 66 LITERATURE CITED ALEXANDER, T. 1981. An exotic plant pest. Palmetto 1(1): 2-3. Arron, W. 1789. Hortus Kewenis: A Catalogue of the Plants Cultivated in the Royal Botanic Garden at Kew, Volume 2. Longman, Hurst, Rees, Orme, and Brown, London. ASHWELL, G. 1966. New colorimetric methods of sugar analysis. VII. The phenol-sulfuric acid reaction for carbohydrates. Methods Enzymol. 8: 93-95. AusTIN, D. F. 1999. Displacement of native ecosystems by invasive alien plants— the Florida experience, or How to destroy an ecosystem. Pp. 3-17. Jn: Jones, D. T. AND B. W. GAMBLE (eds.), Florida’s Garden of Good and Evil: Proceedings of a joint conference of the Florida Exotic Pest Plant Council and the Florida Native Plant Society. June 4—7, 1998, Palm Beach Gardens, FL. BEcKER, R. L. AND R. S. FAwcetr. 1998. Seasonal carbohydrate fluctuations in hemp dogbane (Apocynum cannabinum) crown roots. Weed Sci. 46: 358-367. Boyter, H. H. 1994. Non-indigenous species on lands within Florida’s wildlife management area system. Pp. 206-208. In: Scumitz, D.C. AND T. C. BRowN (eds.), An Assessment of Invasive Non- indigenous Species in Florida’s Public Lands. Florida Dept. of Environmental Protection Technical Report No. TSS-94-100. Conway, W. C., L. M. Situ, R. E. SOSEBEE, AND J. F. BERGAN. 1999. Total nonstructural carbohydrate trends in Chinese tallow roots. J. Range Management 52(5): 539-542. Cox, J. A., T. G. QUINN, AND H. H. Boyter. 1997. “Management by Florida’s Game and Fresh Water Fish Commission.” Pp. 297-316. Jn: StmBeERLOoFF, D., D. C. SCHMITZ, AND T. C. BROWN (eds.), Strangers in Paradise: Impact and Management of Nonindigenous Species in Florida. Island Press, Washington, D.C. CRAVEN, L. 1999. Australian National Herbarium, Canberra, Pers. Comm. Dusois, M. M., K. A. GILLEs, J. K. HAMILTON, P. A. BEBERS, AND F. SmitH. 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28: 350-356. FLEPPC. 2001. Florida Exotic Pest Plant Council’s 2001 list of invasive species. Brochure available online at http://www.fleppc.org/O1list.htm. FDACS. 1999. Introduction or Release of Plant Pests, Noxious Weeds, Arthropods, and Biological Control Agents. Florida Department of Agriculture and Consumer Services—Division of Plant Industry. Tallahassee, FL. Chapter 5B-57. HASELWoop, E. L. AND G. G. MoTrer (eds.). 1966. Handbook of Hawaiian Weeds. Institute for Technical Interchange/East-West Center, Honolulu, HI. Hosaka, E. Y. AND A. THISTLE. 1954. Noxious Plants of the Hawaiian Ranges. Extension Bulletin 62. University of Hawaii, Honolulu, HI. Katovicu, E. J. S., R. L. BECKER, C. C. SHEAFFER, AND J. L. HALGERSON. 1998. Seasonal fluctuations of carbohydrate levels in roots and crowns of purple loosestrife (Lythrum salicaria). Weed Sci. 46: 540-544. Lampers, H., F. S. CHAPIN, AND T. L. Pons. 1998. Plant Physiological Ecology. Springer-Verlag, Inc., New York, NY. LANGELAND, K. A. AND K. C. Burks. 1999. Identification and Biology of Non-native Plants in Florida’s Natural Areas. University of Florida Press, Gainesville, FL. Lona, R. W. AND O. LAKELA. 1976. A Flora of Tropical Florida: A Manual of the Seed Plants and Ferns of Southern Peninsular Florida. Banyan Books, Miami, FL. LuGHADHA, E. N. AND C. PROENCA. 1996. A survey of the reproductive biology of the Myrtoideae (Myrtaceae). Ann. Missouri Bot. Gard. 83: 480-503. Margulis, R. J., E. A. NEWELL, AND A. C. VILLEGAS. 1997. Non-structural carbohydrate accumulation and use in an understorey rain-forest shrub and relevance for the impact of leaf herbivory. Funct. Ecol. 11: 636-643. MicrosorT CorPoRATION. 1996. Microsoft ® Excel 2000. MIYANISHI, K. AND M. KELLMAN. 1986. The role of nutrient reserves in regrowth of two savanna shrubs. Canadian J. Bot. 64: 1244-1248. Morton, J. F. 1976. Pestiferous spread of many ornamental and fruit species in south Florida. Proc. Florida State Horticulture Society 89: 348-353. No. 3 2003] POSSLEY ET AL.—SUBTROPICAL INVASIVE SHRUB 167 NELson, G. 1996. The Shrubs and Woody Vines of Florida: a Reference and Field Guide. Pineapple Press, Inc., Sarasota, FL. SAS INnstiTuTE INCORPORATED. 1996, 2000. SAS 8.1. SAS Institute, Inc., Cary, NC. Scott, A. J. 1978. A revision of Rhodomyrtus (Myrtaceae). Kew Bull. 33(2): 311-329. SmitH, C. W. 1985. Impact of alien plants on Hawaii’s native biota. Pp. 180-242. In: STONE, C. P. AND J. M. Scorr (eds.), Hawaii’s Terrestrial Ecosystems: Preservation and Management. Proceedings of a symposium held June 5—6, 1984 at Hawai’i Volcanoes National Park. Cooperative National Park Resources Studies Unit, University of Hawaii, Honolulu, HI. STOCKER, R. K. AND J. PossLey. 2001. Comparing application methods and herbicides for control of Rhodomyrtus tomentosa. Ecol. Restor. 19(1): 34-36. THAYER, D. AND A. FERRITER. 1994. Invasive non-indigenous vegetation management on South Florida Water Management District lands. Pp. 146—157. In: Scumitz, D. C. AND T. C. BRown (eds.), An Assessment of Invasive Non-indigenous Species in Florida’s Public Lands. Florida Department of Environmental Protection, Tallahassee, FL, Tech. Rept. No. TSS-94-100. Tissue, D. T. AND S. J. WriGuT. 1995. Effect of seasonal water availability on phenology and the annual shoot carbohydrate cycle of tropical forest shrubs. Funct. Ecol. 9: 518-527. USFWS (United States Fish and Wildlife Service). 1996. National list of vascular plant species that occur in wetlands: 1996 national summary (1996 National List). Available online at: http://www.nwi. fws.gov/bha/list96.html (Oct 2002). WaLker, E. H. 1976. Flora of Okinawa and the Southern Ryukyu Islands. Smithsonian Institution Press, Washington, D.C. WATKINS, J. V. AND H. S. Wo tFe. 1968. Your Florida Garden. University of Florida Press, Gainesville, FL. WUNDERLIN, R. P. 1998. Guide to the Vascular Plants of Florida. University Press of Florida, Gainesville, FL. Florida Scient. 66(3): 157-167. 2003 Accepted: October 25, 2002 Biological Sciences A SURVEY OF EPIFAUNA AMONG MACROPHYTES IN A SOUTHWEST FLORIDA ESTUARY (2) Pau J. RUDERSHAUSEN)*, James V. Locascio™, AND LourpEs M. Rosas” ‘Florida Center for Environmental Studies, Palm Beach Gardens, FL ™ Sanibel-Captiva Conservation Foundation, Sanibel, FL “Archbold Biological Station, Lake Placid, FL ABSTRACT: A study was conducted by dipnet to compare community structure and abundance of major taxa of epifauna among Halodule wrightii, Thalassia testudinum, Syringodium filiforme, drift algae, and unvegetated bottom in Tarpon Bay, Sanibel, FL. Sampling was conducted roughly once per week from January, 1999, through January, 2000. Samples from unvegetated bottom contained a significantly lower average number of total organisms than any of the four macrophytes. On average, gastropods were the most abundant taxa collected from each species of seagrass, caridean shrimps were the most abundant taxa collected from drift algae, and mysids were the most abundant taxa collected from unvegetated bottom. Among the five habitats, Halodule contained the highest average number of gastropods, mysids and pink shrimp, Farfantopenaeus duorarum. Thalassia contained the highest average number of pagurid crabs and syngnathid fish. Drift algae contained the highest average number of amphipods, caridean shrimps, and the isopod, Harrieta faxoni. For each macrophyte, regression did not reveal consistent relationships between abundances of major taxa, salinity, or percent macrophyte cover. Community composition differed significantly among all five habitats. Numbers of specimens from major taxa were generally found in greater abundance in summer than in winter. Key Words: Thalassia, Halodule, Syringodium, seagrasses, epifauna, macrophytes SEAGRASSES have long been recognized as important components of healthy estuaries. They are sources of refuge and food for a wide array of invertebrates and small fishes (Thayer et al., 1975; McRoy, 1977; Orth and van Montfrans, 1987) and frequently support faunal densities many times those of unvegetated substrates (Virnstein et al., 1983; Orth et al., 1984). Between 70 and 90% of harvested species in the Gulf of Mexico depend on seagrass meadows for at least part of their life cycles (Lindall and Saloman, 1977). Comparisons of macrofauna found in seagrass beds and adjacent bare substrates have shown that diversity and abundance is higher (Thayer et al., 1975; Orth, 1977; Virnstein et al., 1983; Bell et al., 1984; Orth et al., 1984; Blaber et al., 1992), and predation is lower in vegetated than unvegetated habitats (Virnstein et al., 1983; Summerson and Peterson, 1984). Seagrass species composition and biomass, habitat selectivity, and differential survivorship are factors that likely influence the abundance of epifauna in seagrass meadows (Virnstein and Howard, 1987a). Additionally, unattached drift algae found in some seagrass systems may provide food, habitat, and refuge for a variety of benthic organisms * Present Address: North Carolina State Univ., 303 College Circle, Morehead City, NC 28557. 168 No. 3 2003] RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES 169 (Heck, 1979), and may serve as an alternate habitat for some species. Several major taxa of epifauna have been found in greater abundance in drift algae than in seagrasses (Hooks et al., 1976; Virnstein and Howard, 1987b). The study site, Tarpon Bay, is located in southern Pine Island Sound and considered part of the Caloosahatchee River Estuary. Tarpon Bay, located just west of where the Caloosahatchee River empties into the Gulf of Mexico, is influenced by releases of fresh water through Franklin Lock and Dam (S-79) roughly 40 km upstream from the mouth of the river. S-79 introduces the potential for alteration of fundamental water quality parameters, macrophytes, and fauna in the estuary. Variations in salinity throughout the estuary are likely greater in magnitude and different in seasonality now than before completion of S-79 in the mid-1960’s. The impression of local residents is that seagrass coverage has declined in the lower estuary (Wilzbach et al., 1999). Although human population and development in the lower Caloosahatchee River watershed are increasing rapidly, little baseline biological data have been collected. We undertook this study to evaluate fauna associated with macrophytes in the lower portion of the estuary. Specific objectives were to compare abundance and composition of epifauna among five habitats: unvegetated sandy mud, drift algae, and three morphologically distinct species of seagrasses We also examined the relationship of the abundance of epifauna with salinity and macrophyte cover. MeETHODS—AIl sampling took place in Tarpon Bay, Sanibel, FL (26.4°N, 82.1°W). Tarpon Bay is a 13-km* embayment lying within J.N. “Ding” Darling National Wildlife Refuge. The bay is largely fringed by red mangrove, Rhizophora mangle. The three seagrass species commonly found in Tarpon Bay include narrow bladed Halodule wrightii, wide bladed Thalassia testudinum, and cylindrically bladed Syringodium filiforme, which has a lower surface area: biomass ratio relative to Halodule and Thalassia (Virnstein and Howard, 1987a). Although patchy in distribution, seagrass covers much of the shallow bottom fringing the bay. Seagrass species are consistently separated by depth, with Halodule shallowest (mean depth of occurrence at mean low water ~50 cm), Thalassia at intermediate depths (55 cm), and Syringodium in the deepest waters inhabited by seagrasses in Tarpon Bay (82 cm) (Wilzbach et al., 1999). Consistent with observations of subtropical seagrasses by Phillips (1960), in Tarpon Bay Halodule is found in sandy substrate while Syringodium is more common in silty, organic substrate. No studies have quantified the abundance and relative distribution of seagrasses in Tarpon Bay. Several species of drift algae are seasonally abundant in Tarpon Bay. These include Gracilaria tikvahiae, Gracilaria caudata, Acanthophora spicifera, Caulerpa fistigiata (drift form), and Solieria filiformis (C.J. Dawes, 1998). Drift algae represents a spatially variable, ephemeral habitat that often covers Thalassia and Syringodium during the winter months (personal observation). Additionally, we found a blue-green algae, Lyngbea majuscula, often attached to Halodule and Syringdium during the summer. Local abundance of drift algae in Tarpon Bay is frequently determined by the direction, intensity, and duration of wind and wave action (personal observation). The phenomenon of contiguous beds of Halodule, Thalassia, and Syringodium is common along western and southern fringes of Tarpon Bay. While we attempted to randomly select sites in these areas, we also took care to choose sites that provided sufficient sampling area, consistently abundant cover, and uniform depth. We also made an effort to sample monospecific beds of macrophytes, although Thalassia was occasionally covered by drift algae during winter months. Some sample locations were marked with PVC pipe 24 h prior to sampling where water depth or poor clarity prevented rapid visual location of suitable sites. Sampling was conducted with a dipnet having 800 X 900-11m nylon mesh and dimensions of 305 mm wide X 254 mm high at the mouth. While the dipnet may be ineffective at collecting large, highly motile 170 FLORIDA SCIENTIST [VOL. 66 epifauana such as pink shrimp, Farfantopenaeus duorarum, blue crabs, Callinectes sapidus, and some fishes, it does allow for comparison of epifaunal communities among macrophytes. Dipnetting adds a surprise element to capturing taxa such as xanthid and pagurid crabs whose burrowing habits prevent effective sampling with some other types of gear (Sogard and Able, 1991). Sampling occurred during daylight on a weekly basis, 53 times, from January 19, 1999 through January 29, 2000. Each week the dipnet was deployed by a researcher afoot pushing it along the sediment surface for 30 seconds in each of four bottom types: Halodule, Thalassia, Syringodium, and unvegetated sandy mud. We also sampled drift algae for 22 weeks: from the beginning of the experiment until its disappearance at the end of April and then from its reappearance at the beginning of December until the end of the experiment. All four (or five) samples were taken in a randomly selected order within 30 minutes. Salinity and water temperature were taken in conjunction with each set of samples. While dipnetting in water of different depths introduces the potential bias that more area is covered by the sampler in shallow water than in deep water, we took care to sample at uniform speeds among samples. We estimate that each dipnet transect averaged 10 m in length. After collection of each sample, macrophyte cover was visually estimated (using mask and snorkel when necessary) to the nearest 5%. We did not measure the relative coverage of species of drift algae or quantitatively characterize sampling locations by marcrophyte biomass or plant surface area. Depth, measured in conjunction with each sample, varied from 0.2 to 2.0 m. Upon collection, macroscopic fauna was enumerated. Organisms were identified to the lowest practical taxon and released. We preserved sub-samples of animals that could not be identified during the enumeration process. Sessile invertebrates attached to seagrass blades that entered the dipnet were not counted. For each habitat we computed the average number of organisms in several common taxa: amphipods, caridean shrimps, gastropods, Harrieta faxoni, mysids, pagurids, Farfantopenaeus duorarum, and syngnathids. Data were transformed (logo (x+1)) prior to using analysis of variance (ANOVA) to test for differences in mean abundances of major taxa among the four habitats sampled over all 53 weeks. We employed the Kruskal-Wallis test when heteroscedasticity was identified using the F,,,,-test (Hartley, 1950), and in all tests of mean faunal abundances that included drift algae (where n < 30). Multiple comparisons, when appropriate, were conducted with the Tukey test and its nonparametric analog. The Kolmogorov-Smirnov (K-S) two-sample test was used for comparisons of ratio data. The K-S test analyzes whether a pair of data are drawn from populations having the same distribution by using differences between cumulative percentages to determine a test statistic (Tate and Clelland, 1957). Using the K-S test, one can compare general taxonomic compositions between bottom types despite disparity among sample sizes of specimens (Brook, 1978). We used backward stepwise regression to evaluate the relationship between the abundance of each major taxa collected from each habitat (dependent variable) and salinity and macrophyte cover (independent variables). Tolerance for each regression was set at 0.01, with « to remove a variable from the model = 0.05. In the case of unvegetated bottom, we ran simple regressions that included only salinity as an independent variable. ResuLts—Temperature ranged from 16.6 to 32.3°C, and salinity ranged from 19.0 to 37.3 parts-per-thousand (ppt) during the study period (Fig. 1). Temperature and salinity were slightly, but not significantly, inversely correlated (r =—0.063, p < 0.5). A total of 33,397, 24,407, 12,258, and 1,767 organisms were collected from 53 samples of Halodule, Thalassia, Syringodium, and unvegetated habitat, respectively. A total of 12,362 organisms were collected from 22 samples of drift algae. We identified 7 species of fishes and 51 species of invertebrates from the collections (Appendix 1). The average number of organisms per sample (+ SE) captured from Halodule, Thalassia, Syringodium, unvegetated, and drift algae was 630.1 + 152.6, 460.5 + 55.4, 231.3 + 28.6, 33.3 + 4.6 and 561.9 + 86.5, respectively (Fig. 2). Unvegetated samples contained a significantly lower average number of total organisms than each seagrass species. For each major taxon except mysids, No. 3 2003] RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES 171 40 —@— Salinity --©-. Temperature 35 30 25 20 Salinity (ppt) and Temperature (C) Sampling week (numbers) and month (letters) Fic. 1. Temperature and salinity in Tarpon Bay from January 19, 1999 to January 29, 2000. X-axis is ordered by month beginning with January, 1999. unvegetated samples contained a significantly lower average number of organisms than at least one of the seagrasses (Fig. 2). Drift algae habitat was available from the start of sampling in January, 1999 through April, and again in December and January (n = 22). Using the Kruskal- Wallis test due to small sample sizes, we compared the average abundance of fauna collected from drift algae with concurrent collections from the other four habitats. Drift algae contained a significantly greater average number of total organisms, amphipods, caridean shrimps, and Harrieta faxoni than one or more of the seagrasses, and a significantly lesser average number of mysids than Halodule. Drift algae also contained a significantly greater average number of organisms than unvegetated for all major taxa except mysids (Fig. 3). Gatropods were the most numerically abundant taxa from all samples combined (45,304 total, 63.1% of all organisms collected) and from each seagrass. Three taxa, caridean shrimps, gastopods, and mysids, comprised 95.2%, 87.8%, 84.7% and 79.5% of all organisms collected from Halodule, Thalassia, Syringodium, and unvegetated samples, respectively. Caridean shrimps, amphipods, and gastropods comprised 91.9% of all organisms collected from drift algae. Abundance of major taxa generally increased during summer months (Figs. 4-8). To compare taxonomic composition among the five habitats using the K-S test, we subdivided collections into nine taxonomic groups (Table 1). In every case, the 2 FLORIDA SCIENTIST 1000 25 2A. All taxa 800 20 600 15 400 10 200 5 0 0 800 n 2D. © = Gastropods OS 600 = se) = <— 400 O i ® Q 200 E = Zz 2G. Pagurids 2B.Amphipods 2E. Harrieta faxoni F. duorarum Bottom type [VOL. 66 250 2C. Caridean 200 shrimps 150 100 50 0 2|.Syngnathids H T “She Fic. 2. Average abundance (+ S.E.) and comparison of the eight most abundant taxa among Halodule (H), Thalassia (T), Syringodium (S), and unvegetated (U) samples (n = 53). 173 RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES No. 3 2003] 140 800 shrimps Cc 0 ® AS) — © O O o 600 3B.Amphipods 120 G x & x < 5 100 600 oe 400 200 0 60 8 40 20 400 200 Mysids SF 140 120 100 8 Harrieta faxoni SE ” ne) Oo QO O = — ” 14) O QO 0 oO Oo i=) i=) oO =) oO co co wt N = SJENPIAIPUI JO JOqUINNY 3G. Pagurids F. duorarum 3H Alec DALY H T S DA U H T S DAU Bottom type Average abundance (+ S.E.) and comparison of the eight most abundant taxa among Fic. 3. Halodule (H), Thalassia (T), Syringodium (S), drift algae (DA), and unvegetated (U) samples (n = 22). 174 FLORIDA SCIENTIST [VOL. 66 Halodule Thalassia Syringodium Unvegetated Drift algae All taxa (log,,+1) Sampling week (numbers) and month (letters) Fic. 4. Temporal patterns in abundance of all taxa combined from each habitat. Y-axis is logarithmically scaled (log, 9 + 1). X-axis is ordered by month starting with January, 1999. maximum difference in taxonomic composition between two habitats was highly significant (p < 0.001). Percent macrophyte cover had a significant, positive relationship with: all taxa combined from Thalassia; gastropods from Halodule, and Thalassia; and caridean shrimps from drift algae. Percent macrophyte cover had a significant, negative relationship with mysids from Halodule. Salinity had a significant, negative relationship with: all taxa combined from all habitats except Halodule; caridean shrimps from all four macrophytes; gastropods from Syringodium and unvegetated samples; and mysids from Haludule. Independent variables used for stepwise regressions generally explained little variability in the abundance of major taxa collected from each habitat (7 values) (Table 2). Discussion—Our comparisons of vegetated and unvegetated samples agree with Virnstein and co-workers (1983), who found 13 times the abundance of epifauna in seagrass than over bare sand, McLaughlin and co-workers (1983), who found populations of F. duoranum and caridean shrimps to be an order of magnitude higher in Thalassia beds than in barren areas, and Lewis (1984), who found that Halodule and Thalassia held greater numbers of individuals and species than unvegetated areas. Fishes collected during this study were dominated by small, cryptic species such as syngnathids, which is consistent with ichthyofaunal compositions in other seagrass areas (Stoner, 1983; Pollard, 1984; Bell and Pollard, 1989; Jenkins et al., No. 3 2003] RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES 175 Halodule Thalassia Syringodium Unvegetated Drift algae Amphipods (log,,+1) Sampling week (numbers) and month (letters) Fic. 5. Temporal patterns in abundance of amphipods from each habitat. Y-axis is logarithmically scaled (logy) + 1). X-axis is ordered by month starting with January, 1999. 1997). Significant differences in epifaunal community structure, as was our finding, indicates that major taxa actively select advantageous habitats and/or have different rates of survival among macrophytes. Like other seagrass studies in Florida (e.g. Livingston, 1976; Heck, 1979), seasonal peaks of epifauna generally occurred in summer. Thickly vegetated stands of Halodule found in the summer in Tarpon Bay (personal observation) provide abundant habitat among densely packed blades, which became heavily inhabited by small gastropods (Fig. 7). This rise in gastropod numbers indicates that changes in physical complexity of a seagrass canopy affect associated fauna (Stoner, 1982; Orth et al., 1984). High weekly variability within habitats may be a consequence of the changing efficiency of capture using a dipnet to sample several macrophytes that changed in biomass and blade density. However, high variability in our capture of major taxa (standard errors, Figs. 2 and 3) is consistent with other shallow subtropical embayments where animal populations increase and decrease rapidly (Hoese and Jones, 1963). Variability in our catches may also be due to the fact that populations of nekton in shore zones of estuaries have high rates of migration (Leber, 1985) and turnover (Knudsen and Herke, 1978). Despite possible differences in sampling efficiency, we regard the large differences in abundance and species composition among habitats as real, and believe that because we maintained 176 FLORIDA SCIENTIST [VOL. 66 Halodule Thalassia Syringodium Unvegetated Drift algae Caridean shrimps (log,,.+1) 0 10 20 30 40 50 Sampling week (numbers) and month (letters) Fic. 6. Temporal patterns in abundance of caridean shrimps from each habitat. Y-axis is logarithmically scaled (logjg + 1). X-axis is ordered by month starting with January, 1999. a rigorous sampling schedule our results are highly representative of macrophyte use by epifauna in the study area. Differences in abundance and species composition among habitat types occurred despite factors tending to homogenize these communities. Monospecific seagrass beds in Tarpon Bay are often long and narrow (<4 m wide), with beds of different species often separated by only small areas of sand or muddy sand. Small distances between habitat types may facilitate migrations of motile epifauna between morphologically distinct seagrasses. The high mobility of epifauna may also be due to the fact that either drift algae or seagrasses can serve as habitat for many of these species (Virnstein and Howard, 1987b). Windrows of drift algae may serve as seasonal corridors through which epifauna travel between seagrass _ beds. Additionally, small seagrass patches, such as those found in Tarpon Bay, have a greater proportion of edge than large patches, with a resulting increase in the probability of collecting transients from surrounding habitats (Heck, 1979). Our findings suggest that drift algae is an important, ephemeral habitat in Tarpon Bay. Virnstein and Howard (1987b) found in the Indian River Lagoon, Florida that while gastropods were significantly more abundant in seagrass than in drift algae, crustaceans were more abundant in drift algae than in seagrass. Hooks and co-workers (1976) found greater abundances of caridean shrimps and xanthid crabs in drift algae than seagrasses in northwest Florida. While a dipnet samples drift No. 3 2003] RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES 177 4 Halodule Thalassia Syringodium Unvegetated Drift algae Gastropods (log,,+1) Sampling week (numbers) and month (letters) Fic. 7. Temporal patterns in abundance of gastropods from each habitat. Y-axis is logarithmically scaled (log;g + 1). X-axis is ordered by month starting with January, 1999. algae differently than seagrasses, our results parallel those of Hooks and co-workers (1976) in that drift algae samples contained, on average, more caridean shrimps and amphipods than concurrent samples from each seagrass, and fewer gastropods than each seagrass, except for Halodule. The refuge of drift algae may be effective for small nestling crustaceans that are capable of occupying spaces between branches of the algae (Virnstein and Howard, 1987b). Seagrass biomass/physical complexity is related to seasonal changes in water temperature (Nelson, 1980), and seagrass invertebrates are most abundant where plant biomass is highest (Hooks et al., 1976). Based on other studies (e.g., Hooks et al., 1976) that have found a direct relationship between plant biomass and species abundance, we suspect that the higher rates of capture of euryhaline organisms in the summer are more likely due to increases in seagrass biomass than due to decreasing salinity at this time of year. We believe that the summer increases in the abundance of fauna in Tarpon Bay are more likely the direct result of increases in seagrass biomass than increases in temperature per se. In addition to their ability to withstand wide changes in temperature and salinity, many taxa we collected are well adapted to survive in seagrasses. For instance, we collected taxa that typically cling to blades of seagrass, including the pagurid crabs (McLaughlin et al., 1983) and caridean shrimps such as Tozeuma carolinense (Rouse, 1970). Roberts (1968) commented that an adaptation that Pagurus pollicaris 178 FLORIDA SCIENTIST [VOL. 66 Halodule Thalassia Syringodium Unvegetated Drift algae Mysids (log,,.+1) Sampling week (numbers) and month (letters) Fic. 8. Trends in abundance of mysids from each habitat. Y-axis is logarithmically scaled (logig9 + 1). X-axis is ordered by alternate month starting with January, 1999. and Pagurus longicarpus have for vegetated habitats is that they feed on detrital material, and that other marine pagurids likely have similar modes of feeding. Cryptic coloration represents another adaptation to live among seagrasses. Although wide variability in body coloration is displayed by species such as Hippolyte zostericola and 7. carolinense (personal observation), individuals of these species that we collected were often cryptically camouflaged with seagrasses and drift algae. Future research should investigate diel differences in community structure in seagrass beds in this estuary. Do major taxa exhibit fidelity towards particular macrophytes, or do habitat preferences fluctuate with time of day? Some taxa may be numerically dominant at certain times of day but by reducing their height in the canopy and/or burrowing in the sediment they may contribute little to samples. Decapods, for example, have been found to be more active at night than during the day (Heck, 1979; Heck and Orth, 1980), and some decapod species are probably collected with greater efficiency at night (Heck and Thoman, 1984). Because many benthic seagrass invertebrates increase their activity and rise higher in the seagrass canopy at night, these taxa may be more available to capture at that time (Greening and Livingston, 1982). Studies in the dynamic Caloosahatchee River Estuary should focus on quantifying densities of fauna, as well as relationships between abundance of epifauna and: macrophyte biomass, plant structure, and shoot density. Seagrasses are 179 RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES No. 3 2003] LOLI 8°6 vLI £0 ¢ 80 vl OTC 9¢ Cab, 6CL © 0 9) 0'8C S6V COL O8I GL al % ‘ON poyejosoauy) Z9ETI Ce Lev €0 Iv v0 8y oT 007 Z0 NG 07 Lv ars 189 ae L8L8 ESI r68I % ‘ON ovale Yu 8SCCI 9°C 6l€ 9°0 89 vl ibje)| iE 6tV 6°C LSet c0 6S 6 81 966$ 6CE 8cOV JO S18 % ‘ON wniposull1Ks LOVVC Lal 8SC c0 Tel 60 LCC 6S cerl VC c6S 60 VCC cs SIScl [ve (Abs: 6°C CIL % iON DISSD]DY I LOECE L’‘0 OSC c0 18 Gil 86€ © 0 vil 0°6 SOOe © 0 COT LSh 86C9C CL LVS 61 ceo % ‘ON ajnpo]vH [e1OL BXP] “OSTA sepryyeusuds WNADAOND “+f ‘dds snunsvq vooepisAy] 1uoxv{ “HY epodonsey Bopue) evpodiydwy “MOTO PdIST] IV UOTJISOAUIOD STLUOUOXR} UT SOOUdIOFJIP OJ Sd} oPdures-OM}] AOUTTUUG-AOIOSOW]OY IY) UT posn eXR], “oRS[e IJLIP WoOIJ sUONdaT[OD ZZ puv WOYOG payeJosaAuN pure ‘WNIPOsULAG ‘DISsD]DY J, ‘a/NPO]VET WoOI, SUOTDET[OO JoudIp ¢¢ UT vUNL] Jo ATeuUTUING "| AsV], 180 FLORIDA SCIENTIST [VOL. 66 TABLE 2. Results of stepwise regressions of major taxa (dependent variable) on salinity and percent macrophyte coverage (independent variables). R-square and p-values are listed for each regression. The sign of relationships (positive or negative) is indicated next to respective p-values. Dependent variable were logarithmically transformed (log; 9 + 1) prior to each regression. Dependent Salinity Macrophyte cover variable ie p p Halodule All taxa 0.000 0.205 (—) 0.461 (+) Caridea 0.112 0.014 (—) 0.394 (+) Gastropoda 0.177 0.567 (—) 0.002 (+) Mysidacea 0.236 0.101 (+) 0.000 (—) Thalassia All taxa 0.401 0.011 (—) 0.000 (+) Caridea 0.559 0.000 (—) 0.249 (—) Gastropoda 0.288 0.162 (—) 0.000 (+) Syringodium All taxa 0.117 0.012 €—) 0.437 (—) Caridea 0.144 0.005 (—) 0.496 (—) Gastropoda 0.098 0.022 (—) 0.829 (+) Drift algae All taxa 0'559 0.000 (—) 0.319 (+) Amphipoda 0.000 0.982 (+) 0.552 (+) Caridea 0.783 0.000 (—) 0.042 (+) Unvegetated All taxa 0.120 0.011 €) = Gastropoda 0.098 0:022 () = Mysidacea 0.117 0.012 (—) = affected by increases in turbidity and nutrients and, as such, are threatened directly by development and indirectly by changing land use patterns (Shephard et al., 1989). We believe that artificial exacerbations of salinity levels in Tarpon Bay and neighboring estuarine waters will have primary effects on seagrass abundance and distribution, and secondary effects on the abundance of associated epifauna. While we found that the 17 ppt drop in salinity in June had little effect on numbers and compositions of fauna we collected, the short duration of this study limits our ability to detect long-term changes wrought by such stressors to either macrophyte habitat or animal populations. Relative roles of macrophyte biomass, salinity, and water quality in shaping animal abundance and community composition should be topics of long-term future research. Current theory suggests that subtropical estuaries are naturally stressed environments whose physico-chemical features fluctuate widely and whose biota is dominated by few species (Hooks et al., 1976). A major question remains as to how well this theory holds with respect to long-term artificial manipulations of salinity and other human-generated stressors in the Caloosahatchee River Estuary. No. 3 2003] RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES 181 ACKNOWLEDGMENTS—We thank South Florida Water Management District and the Florida Center for Environmental Studies for their financial support, and the Wallace Foundation for providing 2 months salary for the first author. We are grateful to Dr. Robert Virnstein of the St. Johns River Water Management District for his thorough reviews of earlier versions of the manuscript. We also thank Beth Cook, Mike Boerema, and Bill Simons for their voluntary assistance to the project. LITERATURE CITED BELL, S. S., K. WALTERS, AND J. C. KERN. 1984. Meiofauna from seagrass habitats: a review and prospectus for future research. Estuaries 7: 331-338. BELL, J. D. AnD D. A. POLLARD. 1989. Ecology of fish assemblages and fisheries associated with seagrasses. Pp. 565-609. In: Larkum, A. W. D., A. J. McComps, AND S. SHEPARD (eds.), Biology of Seagrasses: A Treatise on the Biology of Seagrasses with Special Reference to the Australian Region. Elsevier, Amsterdam. BiaBer, S. J. M., D. T. BREwer, J. P. SALINI, J. D. KERR, AND C. CONACHER. 1992. Species composition and biomasses of fishes in tropical seagrasses at Groote Eylandt, northern Australia. Estuarine Coastal Shelf Sci. 35: 605-620. Brook, I. M. 1978. Comparative macrofaunal abundance in turtle grass (Thalassia testudinum) communities in south Florida characterized by high blade density. Bull. Mar. Sci. 28: 212-217. Dawes, C. J. 1998. Univ. South Florida, Tampa, Pers. Comm. GREENING, H. S. AND R. J. LIvINGSTON. 1982. Diel variation in the structure of seagrass-associated epibenthic macroinvertebrate communities. Mar. Ecol. Prog. Ser. 7: 147-156. Hart ey, H. O. 1950. The maximum F-ratio as a short cut test for heterogeneity of variances. Biometrika 37: 308-312. Heck, K. L., JR. 1979. Some determinants of the composition and abundance of motile macroinvertebrate species in tropical and temperate turtlegrass (Thalassia testudinum) meadows. J. Biogeogr. 6: 183- 200. AND R. J. ORTH. 1980. Structural components of eelgrass (Zostera marina) meadows in the lower Chesapeake Bay—Decapod Crustacea. Estuaries 3: 289-295. AND T. A. THOMAN. 1984. The nursery role of seagrass meadows in the upper and lower reaches of Chesapeake Bay. Estuaries 7: 70-92. Hoese, H. D. AND R. S. Jones. 1963. Seasonality of larger animals in a Texas turtle grass community. Univ. Texas Publs. Inst. Mar. Sci. 9: 37-47. Hooks, T. A., K. L. HEck, AND R. J. Livincston. 1976. An inshore marine invertebrate community: structure and habitat associations in the northeastern Gulf of Mexico. Bull. Mar. Sci. 26: 99— 109. JENKINS, G. P., H. M. A. May, M. J. WHeaTLEy, AND M. G. Hottoway. 1997. Comparison of fish assemblages associated with seagrass and adjacent unvegetated habitats of Port Phillips Bay and Corner Inlet, Victoria, Australia, with emphasis on commercial species. Estuarine Coastal Shelf Sci. 44: 569-588. KNUDSEN, E. E. AND W. H. HEerkKE. 1978. Growth rate of marked juvenile Atlantic croakers, Micropogon undulatus, and length of stay in a coastal marsh nursery in southwest Louisiana. Trans. Am. Fish. Soc. 107: 12-20. Leprr, K. M. 1985. The influence of predatory decapods, refuge, and microhabitat selection on seagrass communities. Ecology 66: 1951-1964. Lewis, F. G., Ill. 1984. The distribution of macrobentic crustaceans associated with Thalassia, Halodule and bare sand substrata. Mar. Ecol. Prog. Ser. 19: 101-113. LINDALL, W. N., JR. AND C. H. SALOMAN. 1977. Alteration and destruction of estuaries affecting fishery resources of the Gulf of Mexico. Mar. Fish. Review 1262: 1-7. LivINGSTON, R. J. 1976. Diurnal and seasonal fluctuations of organisms in a north Florida estuary. Estuarine Coastal Mar. Sci. 4: 373-400. McLaucuiin, P. A., S-A. F. TREAT, AND A. THORHAUG. 1983. A restored seagrass (Thalassia) bed and its animal community. Env. Cons. 10(3): 247-254. 182 FLORIDA SCIENTIST [VOL. 66 McRoy, C. P. 1977. Seagrass ecosystems: research recommendations of the International Seagrass Workshop. Inter. Decade Ocean. Exploration, Univ. Alaska Institute of Marine Science, Fairbanks, Alaska, 62 pp. NELson, W. G. 1980. The biology of eelgrass (Zostera marina) amphipods. Crustaceana 39: 59-89. OrTH, R. J. 1977. The importance of sediment stability in seagrass communities. Pp. 281-300. In: CouLL, B. C. (ed.), Ecology of Marine Benthos, University of South Carolina Press, Columbia, SC. , K. L. HEck, JR., AND J. VAN MONTFRANS. 1984. Faunal communities in seagrass beds: a review of the influence of plant structure and prey characteristics on predator-prey relationships. Estuaries 7: 339-350. AND J. VAN MONTFRANS. 1987. Utilization of a seagrass meadow and tidal marsh creek by blue crabs Callinectes sapidus. I. Seasonal and annual variations in abundance with emphasis on post- settlement juveniles. Mar. Ecol. Prog. Ser. 41: 283-294. PuHiLuips, R. C. 1960. Observations on the ecology and distribution of Florida seagrasses. Professional Paper ser. No. 2. Florida State Board Cons., St. Petersburg, Florida, 72 pp. PoLLarD, D. A. 1984. A review of ecological studies on seagrass-fish communities, with particular reference to recent studies in Australia. Aquat. Bot. 18: 3-42. Roserts, M. H., Jr. 1968. Functional morphology of mouth parts of hermit crabs, Pagurus longicarpus and Pagurus pollicarpus. Chesapeake Sci. 9: 9-20. Rouse, W. L. 1970. Littoral crustacea from southwest Florida. Quart. J. Florida Acad. Sci. 2: 127-152. SHEPHARD, S. A., A. J. Mccoms, D. A. BULTHUIS, V. NEVERAUSKAS, D.A. STEFFENSOEN, AND R. WEST. 1989. Decline of seagrasses. Pp. 346-393. In: LaRKuM, A. W. D., A. J. McComp, AND S. SHEPARD, (eds.), Biology of Seagrasses: A Treatise on the Biology of Seagrasses with Special Reference to the Australian Region. Elsevier, Amsterdam. SOGARD, S. M. AND K. W. ABLE. 1991. A comparison of eelgrass, sea lettuce macroalgae, and marsh creeks as habitats for epibenthic fishes and decapods. Estuarine Coastal Shelf Sci. 33: 501-519. STONER, A. W. 1982. The influence of benthic macrophytes on the foraging behavior of pinfish, Lagodon rhomboides (Linnaeus). J. Exp. Mar. Biol. Ecol. 58: 271-284. . 1983. Distribution of fishes in seagrass meadows: role of macrophyte biomass and species composition. Fish. Bull. 81: 837-846. SUMMERSON, H. C. AND C. H. PETERSON. 1984. Role of predation in organizing benthic communities in a temperate-zone seagrass bed. Mar. Ecol. Prog. Ser. 15: 63-77. Tate, M. W. AND R. C. CLELLAND. 1957. Nonparametric and shortcut statistics. Interstate Printers and Publishers, Inc., Danville, IL, 171 pp. THAYER, G. W., S. M. ADAmMs, AND M. W. Lacrorx. 1975. Structural and functional aspects of a recently established Zostera marina community. Pp. 517-540. In: CRONIN, L. E. (ed.), Estuarine Research. Vol. 1. Academic Press, New York, NY. VIRNSTEIN, R. W., P. S. MIKKELSEN, K. D. CAIRNS, AND M. A. CAPONE. 1983. Seagrass beds versus sand bottoms: the trophic importance of their associated benthic invertebrates. Florida Scient. 46(3-4): 363-381. AND R. K. Howarp. 1987a. Motile epifauna of marine macrophytes in the Indian River lagoon, Florida. I. Comparisons among three species of seagrasses from adjacent beds. Bull. Mar. Sci. 4: 1-12. AND . 1987b. Motile epifauna of marine macrophytes in the Indian River lagoon, Florida. II. Comparisons between drift algae and three species of seagrasses. Bull. Mar. Sci. 4: 13-26. Wivzpacn, M. A., K. W. Cummins, L. M. Rosas, P. J. RUDERSHAUSEN, AND J. V. Locascio. 1999. Establishing baseline seagrass parameters in a small estuarine bay. Pp. 125-135. In: BorTONE, S. (ed.), Seagrasses: Monitoring, Ecology, Physiology and Management, CRC Press, Boca Raton, RES Florida Scient. 66(3): 168-183. 2003 Accepted: October 25, 2002 No. 3 2003] RUDERSHAUSEN ET AL.—EPIFAUNA IN MACROPHYTES APPENDIX 1. Species identified from dipnet collections. Amphipods Cymadusa compta Annelid Bispira melanostigma Bivalves Anomalocardia auberina, pointed venus Lucina nassula, woven lucine Pitar simpsoni, Simpson venus Tellina mera, mera tellin Decapods Alpheus normanni, green snapping shrimp Hippolyte curacaoensis Latreutes fucorum Pagurus longicarpus Palaemonetes vulgaris Pinnixa sayana Tozeuma carolinense, arrow shrimp Echinoderm Echinaster sentus, thorny starfish Fishes Anarchopterus criniger, fringed pipefish Hippocampus zosterae, dwarf seahorse Lucania parva, rainwater killifish Symphurus plagiusa, blackcheek tonguefish Gastropods Acteocina canaliculata, barrel bubble Bittiolum varium, variable bittium Costoanachis semiplicata, dove shell Dentimargo aureocincta, golden marginella Fasciolaria tulipa, true tulip Microeulima hemphillii, brown eulima Nassarius vibex, common eastern nassa Pilsbryspira leucocyma, white-knob drillia Stellatoma stellata, stellate mangelia Terebra protexta, fine-ribbed auger Turbonilla dalli Zebina browniana, smooth risso Isopods Erichsonella attenuata Mysids Anchialina typica Lembos websteri Laevicardium mortoni, Morton egg cockle Macoma constricta, constricted macoma Pseudomiltha floridana, Florida lucine Farfantepenaeus duorarum, pink shrimp Hippolyte zostericola Pagurus annulipes Palaemonetes intermedius Palaemonetes pugio Rhithropanopeus harrisii Chilomycterus schoepfi, striped burrfish Lagodon rhomboides, pinfish Microphis brachyurus, opossum pipefish Astyris lunata, lunar dove shell Cerithium muscarum, flyspeck cerith Crepidula fornicata, Atlantic slipper shell Dentimargo eburneola, tan marginella Haminoea succinea, amber glassy bubble Modulus modulus, Atlantic modulus Olivella pusilla, tiny dwarf olive Pyrgocythara plicosa, plicate mangelia Terebra dislocata, eastern auger Turbonilla arnoldoi Turbonilla incisa Harrieta faxoni Taphromysis bowmani 183 Biological Sciences REGISTER OF AN EXCEPTIONALLY LARGE REDBELLIED PACU, PIARACTUS BRACHYPOMUS (TELEOSTEI, CHARACIDAE), IN EAST-CENTRAL FLORIDA WITH GONAD AND DIET ANALYSES 9 ) ) RAMON Ruiz-Carus’” anp S. BARRY Davis? “Florida Fish and Wildlife Conservation Commission, Florida Marine Research Institute, 100 Eighth Avenue SE, St. Petersburg, FL 33701-5020 University of Florida Herbarium, Florida Museum of Natural History, P.O. Box 110575, Gainesville, FL 32611-0575 ABSTRACT: The capture of an exceptionally large redbellied pacu in Brevard County is documented. The redbellied pacu measured 841 mm total length and is the largest specimen on record in Florida. Gonad histology showed a mature female with oocytes in late maturation stage, or class 4. The condition of the oocytes also indicated that female red bellied pacu may possibly be synchronous spawners. Stomach contents consisted of cabbage palm seeds, sour orange seeds, and clupeid fish bones. The size and advanced stage of sexual maturation of the redbellied pacu described in this study suggests that the species is able to withstand Florida’s winter cold and to reach reproductive stage. Key Words: Introduced, non-indigenous, exotic, species A LARGE piranha-like fish was caught on November 16, 2000, at Turkey Creek (27.9856 N, 80.6601 W) in Palm Bay, Brevard County, Florida. Mr. David Delaware caught the fish using hook and line baited with bread. He kindly donated the specimen to the Florida Fish and Wildlife Conservation Commission’s Florida Marine Research Institute for further study. The fish was identified as Piaractus brachypomus (Cuvier, 1818) (Characidae, Serrasalminae), a South American species native to the Orinoco and Amazon rivers (Machado-Allison, 1982). This fish is commonly called the ‘“‘morocoto”’ in Venezuela (Roman, 1983) and “‘pirapitinga”’ in Brazil (Britski et al., 1999). In the aquarium trade, it is also known as the redbellied pacu, and its occurrence in Florida has been previously documented (FWC-Florida Marine Research Institute unpublished records; Loftus and Kushlan, 1987; Massette, 1993; Fuller et al., 1999). Ten voucher specimens of P. brachypomus collected from the wild exist in Florida museum collections: 4 deposited in the FWC-Florida Marine Research Institute and 6 in the Florida Museum of Natural History. In addition, several newspaper articles have described the capture of pacu from the wild in Florida. The fish on which most of these newspaper reports were based however, were not identified by an ichthyologist, or were based on unclear photographs. Con- sequently, the identity of the fish in these newspaper reports is questionable. Piaractus brachypomus reaches a maximum size of 850 mm total length (TL) and a maximum weight of 20 kg in Venezuela and Brazil (Saint-Paul, 1986). In 184 No. 3 2003] RUIZ-CARUS AND DAVIS—LARGE REDBELLIED PACU 185 Bolivian Amazonia, the redbellied pacu attains a shorter maximum length (710 mm TL) and reaches only 14 kg maximum weight (Loubens and Panfili, 2001). The largest fish previously recorded in Florida measured 563 mm TL (FWC-Florida Marine Research Institute, unpublished records). The present report documents the largest redbellied pacu collected and recorded in Florida. MeETHODs—Taxonomic identification was based on Cuvier’s (1818) original description, Cuvier and Valenciennes (1849), and the review of Machado-Allison (1982). Systematics adhere to Eschmeyer (1998), and counts and measurements made correspond to those used by Machado-Allison (1982). Mea- surements are expressed in millimeters (mm) or as a percentage of total length (TL). A portion of the gonad was fixed in 15% buffered formalin, embedded in plastic, and serially sectioned to 4 um. The sections were stained with hematoxylin and eosin. The stomach contents were examined with a dissecting microscope; identification of seeds was done by direct comparison with seeds from the University of Florida Herbarium. The maturation stage was determined according to criteria de- fined by Taylor and co-workers (1998). The specimen was cataloged in the FWC- Florida Marine Re- search Institute’s collection; the catalogue number is FSBC 19358. RESULTS—Piaractus brachypomus, FSBC 19358.— Diagnostic characters: opercle elongated dorsoventrally, longer than wide. Opercle height (h) 114, opercle width (w) 38, opercle ratio (h/w) 3; opercle shape not semilunar (Fig. 1A). Number of lateral line scales 102; scales not modified in their posterior margin and no supplementary scales present in membranes. Maxillary with teeth. Meristics: dorsal fin If+ 15; pectoral fin I+ 16; pelvic fin I+ 7; anal fin III + 24; gill rakers on first arch 36; ventral scutes 48. Morphometrics: TL 841; SL 739; body depth 368. Body depth : TL 43%; head length 214; head length : TL 25%; eye diameter 20; eye diameter : head length 9%; preorbital length 44; length of pectoral fin 102; length of pelvic fin base 120; length of anal fin base 172; caudal peduncle width 86. Other morphological characteristics: the teeth in the upper jaw were modified tricuspid and in two series; the internal series comprised 4 teeth, and the external series consisted of 10. The lower jaw teeth, 10, were tricuspid and in a single row, becoming progressively smaller posteriorly; the medial teeth were incisor-like, and the lateral ones were conical (Fig. 1B). The total weight was 20.6 kg (45.5 Ib). The fish showed no evidence of disease or parasites. Gonad histology: the analysis showed a mature female with oocytes in late maturation stage, or class 4 (Fig. 1C). The condition of the oocytes also indicated that female P. brachypomus may possibly be synchronous spawners. Stomach contents analysis: the stomach contents consisted of approximately 870 ml (dry volume) of fruit seeds and an undetermined number of complete and fragmented fish bones. Most seeds (95% of volume) were from cabbage palm, Sabal palmetto (Arecaceae), and these seeds were broken and had a brown-black seed coat and white endosperm. All remaining seeds (5% volume) were from sour orange, Citrus aurantium (Rutaceae), these were entire and had a white, wrinkled seed coat. The fish bones were those of an unidentified herring or shad (Clupeidae) ap- proximately 160 mm TL. 186 FLORIDA SCIENTIST [VOL. 66 at Fic. 1. Piaractus brachypomus, FSBC 19358, 841 mm TL and 739 mm SL. A) Lateral view: notice that opercle (O) is not semilunate. B) Frontal view: mouth showing the diagnostic redbellied pacu teeth (T). C) Gonad histology: oocytes in late maturation stage. Hematoxylin and eosin, 10X. D) Ventral view: gonad (G), stomach (S), and stomach contents (C) are visible. No. 3 2003] RUIZ-CARUS AND DAVIS—LARGE REDBELLIED PACU 187 DiscussioN—The number of lateral line scales of P. brachypomus FSBC 19358 is higher than the number reported by Cuvier (1818), Cuvier and Valenciennes (1849), and Machado-Allison (1982), but within the range reported by Roman (1983). It has been assumed that two tropical species of pacu, P. brachypomus and Colossoma macropomum, have not become established in Florida because of their inability to tolerate low temperatures. However, Milstein and co-workers (2001) demonstrated that P. mesopotamicus, a closely related species, tolerates the short winters of Israel. Furthermore, the International Game Fish Association records indicate that in Florida pacus showed a steady increase in maximum size during the period from 1993-2001 (Kelley, 2002). The size and advanced stage of sexual maturation of the redbellied pacu described in this study suggests that the species is able to withstand Florida’s winter cold and to reach reproductive stage. It is interesting to note that a nonnative fish was able to survive in Florida by including in its diet sour orange, another nonnative species. The sour orange tree is native to southeastern Asia and was introduced by the Spaniards into St. Augustine and the Indian River region in Florida. Sour orange trees can still be found in Everglades hammocks on the sites of former Indian dwellings (Morton, 1987). Recorded observations of redbellied pacu in Florida include specimens that were captured as well as sightings made in Dade, Hillsborough, Pinellas, Citrus, Holmes, Alachua, Marion, Martin, Escambia, Duval, and St. Lucie counties (Florida Marine Research Institute unpublished records; Loftus and Kushlan, 1987; Massette, 1993; Fuller et al., 1999; Robins, 2002). Many of the sightings however, are ques- tionable because of doubtful identification. To alleviate this recurrent problem, we would like to emphasize the morphological differences useful in distinguishing the species of pacu that have been reported in Florida. Piaractus brachypomus, or Cuvier’s Mylete of short opercle, can be separated from Colossoma macropomum, based on the proportion and shape of the opercle (Cuvier and Valenciennes, 1849). The opercle of P. brachypomus is short and not semilunate; the ratio of opercle height to opercle width (h/w) is =2.0 (Fig. 1A). In C. macropomum the opercle is semilunate and its h/w ratio is 1.3—1.8. Additionally, the distal portion of the adipose fin (Fig. 1A) is rayed in C. macropomum but is not rayed in P. brachypomus. The adipose fin rays are evident already in C. macropomum about 55 mm SL (Machado- Allison, 1982). There are also a few reports of Piaractus mesopotamicus taken in Florida, but most of these records should be considered unconfirmed (Fuller et al., 1999). Piaractus mesopotamicus can be distinguished from the other two pacu by the number of scales in the lateral line (Fig. 1A). Usually, P. mesopotamicus has >110, C. macropomum has 68-80, and P. brachypomus 95-110. The redbellied pacu’s numbers include those reported by Roman (1982). ACKNOWLEDGMENTS—We thank D. Delaware for his donation. We are indebted to C. Plybon, FWC- FMRI, who did the histology work. We thank K. Perkins, University of Florida Herbarium; P. L. Fuller, Florida Caribbean Science Center; G. Kelley, International Game Fish Association; R. H. Robins, Florida Museum of Natural History; and FWC-FMRI colleagues L. French, R. Cody, L. Barbieri, R. Reese, G. McRae, G. Henderson, J. Leiby, and J. Quinn. The suggestions of two anonymous reviewers improved the original manuscript. 188 FLORIDA SCIENTIST [VOL. 66 LITERATURE CITED Britskl, H. A., K. Z. S. DE SILIMON, AND B. S. Lopes. 1999. Peixes do Pantanal. Manual de identificagao. Embrapa. Servicgo de Produgao, SPI, Brasilia, Brazil 184 pp. Cuvier M. LE B. 1818. Sur les poissons du sous-genre Myletes. Mem Mus. Natl. Hist. nat. 4: 444-456, pls. 21-22. AND M. A. VALENCIENNES. 1849. Histoire Naturelle des Poissons, vol. XXII, pp. 199-201. Chez P. Bertrand, Paris. EscHMEYER, W. N. 1998. Catalog of Fishes. 2nd ed. California Academy of Science, San Francisco, CA. 2905 pp. FuLter, P. L., L .G. Nico, AND J .D. WILLIAMS. 1999. Nonindigenous fishes introduced into inland waters of the United States. Amer. Fish. Soc. Spec. Pub. No. 27. 613 pp. KELLEY, G. 2000. International Game Fish Association, Dania Beach, Florida, Pers. Commun. Lortus, W. F. AND J. A. KUSHLAN. 1987. Freshwater fishes of southern Florida. Bull. Fla. Stat. Mus. Biol. Sci. 31: 147-344. LouBENS, G. AND J. PANFILI. 2001. Biologie de Piaractus brachypomus (Teleostei: Serrasalmidae) dans le bassin du Mamoré (Amazonie bolivienne). Ichthyol. Explor. Freshwaters 12(1): 51-64. MACHADO-ALLISON, A. 1982. Estudio sobre la subfamilia Serrasalminae (Teleostei, Characidae). Parte 1. Estudio comparado de los juveniles de las ‘‘cachamas” de Venezuela (géneros Colossoma y Piaractus). Acta Biol. Venez. 11(3): 1-101. MasseTTE, B. 1993. Florida’s next game fish—the pacu? Fla. Game and Fish (Tallahasssee) February: 28— 31, 60-61. MiisTEIN, A., M. ZORAN, Y. PERETZ, AND D. JOSEPH. 2001. Low temperature tolerance of pacu, Piaractus mesopotamicus. Environ. Biol. Fishes 58: 455-460. Morton, J. F. 1987. Fruits of Warm Climates. Florida Flair Books, CD-ROM, November 1, 2000. 504 pp. Rosins, R. H. 2002. Florida State Museum of Natural History, Ichthyology Department, Gainesville, Pers. Commun. RomMAn, B. 1983. Las pirahfas y demas caracidos. Coleccién los peces de Los Llanos de Venezuela. Fundacion Cientifica Fluvial Los Llanos, Caracas, Venezuela. 207 pp. SAINT-PAUL, U. 1986. Potential for aquaculture of south-American freshwater fishes: a review. Aquaculture 54: 205-240. Tayor, R. G., H. J. Grier, AND J. A. WuitTTINGTON. 1998. Spawning rhythms of common snook in Florida. J. Fish Biol. 53: 502-520. Florida Scient. 66(3): 184-188. 2003 Accepted: October 25, 2002 Biological Sciences A PEDESTRIAN ROAD SURVEY OF THE SOUTHERN HOGNOSE SNAKE (HETERODON SIMUS) IN HERNANDO COUNTY, FLORIDA Kevin M. Ence? anp Kristin N. Woop” Florida Fish and Wildlife Conservation Commission, 5300 High Bridge Road, Quincy, FL 32351 Florida Fish and Wildlife Conservation Commission, Chinsegut Nature Center, 23212 Lake Lindsey Road, Brooksville, FL 34601 ABSTRACT: A pedestrian survey of snakes was conducted for 1022 days on 6 km of rural roads through xeric upland habitats in Hernando County, Florida. Two hundred twenty-eight snakes of 18 species were recorded, 93.4% of which were dead on road (DOR). The southern hognose snake (Heterodon simus) was the second most frequently observed species, and all 39 H. simus were DOR. The local abundance of H. simus was notable, because it has apparently experienced population declines throughout much of its range. Heterodon simus is seldom trapped during drift-fence surveys in Florida, but data on commercial collection for the pet trade indicate that it is still locally common in at least three areas of Florida. Mean annual mortality of H. simus was 2.3/km/yr and for all snakes was 12.8 per/km/yr, despite low traffic volume on the roads surveyed. Two-thirds of DOR H. simus remained on roads for <1 day, although one carcass lasted 12 days. Peak months for H. simus highway mortality were June and November, and hatchlings comprised 96% of roadkills October-December. Sections of roads bisecting ruderal habitats (lawns, old fields, and improved pastures) accounted for 48.7% of H. simus observations. Heterodon simus can apparently persist in areas of fragmented and altered upland habitats, although cumulative road mortality may be a significant factor, especially for hatchlings. Key Words: Southern hognose snake, Heterodon simus, road survey, xeric uplands, Hernando County, Florida THE value of the longleaf pine (Pinus palustris) community to many rare amphibian and reptile species in the southeastern Coastal Plain of the United States is now widely recognized (Campbell and Christman, 1982a; Means and Grow, 1985; Noss, 1988; Dodd, 1992; Guyer and Bailey, 1993; Dodd, 1995; Enge and Wood, 2001). The longleaf pine community was historically the predominant community in the southeastern Coastal Plain, but only 14% of it remains (Means and Grow, 1985) due to fire suppression and conversion to agricultural fields, timber plantations, and residential and urban areas (Ware et al., 1993). Longleaf pine forests on xeric, sandy sites are often referred to as sandhill or high pine habitat (Myers, 1990). In Florida, sandhill habitat has declined by 88% since European settlement and now covers only 2.4% of the state; one of the largest remaining patches of sandhill habitat in peninsular Florida is found on the Brooksville Ridge (Kautz et al., 1993), where this study was conducted. The southern portion of the Brooksville Ridge is a sandy, upland area ca. 97 km long and 16-24 km wide (White, 1970). Destruction, degradation, and fragmentation of sandhill habitat have caused population declines 189 190 FLORIDA SCIENTIST [VOL. 66 of several upland snake species, including the southern hognose snake (Heterodon simus) (Guyer and Bailey, 1993; Tuberville et al., 2000). Under natural conditions, the pyrogenic sandhill habitat probably burned every 2—5 years, but habitat frag- mentation and fire suppression by man have allowed many sandhill habitats to be invaded by hardwoods and to succeed to xeric or even mesic hammocks (Myers, 1990). The shady conditions and reduced grassy and herbaceous ground cover in hammocks are less favorable for snake species adapted to open longleaf pine forests (Guyer and Bailey, 1993). Heterodon simus once inhabited xeric, sandy habitats in six southeastern states, but it may no longer occur in Alabama and Mississippi (Tuberville et al., 2000). Heterodon simus has been recorded from 34 of Florida’s 67 counties, but records exist from only 11 counties within the past 15 yr (Tuberville et al., 2000). However, Florida has not been surveyed for this species, and relatively little information is available on its current population status (Tuberville et al., 2000). Because H. simus is small, cryptic, and fossorial, its presence and abundance are often difficult to determine, especially when populations are disjunct and widely distrib- uted (Tuberville et al., 2000). Vehicular surveys (1.e., road cruising) have been widely used to survey snake communities (Klauber, 1939; Fitch, 1949; Dodd et al., 1989; Mendelson and Jennings, 1992; Rosen and Lowe, 1994; Sullivan, 2000), and road cruising appears to be the best method to detect H. simus (Tuberville et al., 2000; Beane, 2002). Although a pedestrian survey is more time-consuming than road cruising the same route, it should be more successful at detecting small species like H. simus. We obtained information on habitat use, relative abundance, and seasonal activity of H. simus by conducting a pedestrian survey on five rural roads in Hernando County, Florida (Enge and Wood, 2002). Additional information on H. simus in Florida was obtained from drift fence surveys (Enge, 1997) and commercial collectors for the pet trade (Enge, in press). MeETHODS—We recorded alive on road (AOR) and dead on road (DOR) snakes while walking a 6.0-km route approximately 7 km east of Brooksville, Hernando County, Florida, that consisted of five rural roads: Camp Castle, Haddon, Preston, Rich Barn, and Weatherly (Fig. 1). This route was walked 1022 days (equivalent to 2.8 yr) between 3 June 1998 and 20 December 2001, which represented 79% of the 1297 available days. Over 95% of our surveys occurred prior to 0800 hr, but some surveys occurred after 1600 hr. We recorded the location and habitat type on both sides of the road for each snake. We measured the total length (TL) or snout-vent length (SVL) of each DOR snake found after the first two months of the study. Dead snakes were left on the road to determine the number of days before they disappeared or became unrecognizable to species. We obtained hourly traffic volumes Friday—Monday from automatic counters installed 29 June—2 July 2001 on Preston Road and 13-16 July 2001 on Camp Castle and Weatherly roads. We determined the extent of roadside habitats using a Rolatape® distance measuring wheel in October 2001, and the length of each road was determined from U.S. Geological Survey color digital orthoquad (DOQ) maps (Table 1). Four of the survey roads connected to form a circular route; Haddon Road and the northern portion of Preston Road were dead-end roads (Fig. 1) and were walked twice, although we only measured each road once. Haddon Road is a short, one-lane road that serves as a driveway for seven houses before ending at a private gate, and the northern portion of Preston Road ends at a gate leading to an abandoned Camp Quarry. All roads were paved except for Rich Barn, which was limerock. No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 19] Chinsegut Z Nature N ay Center SS “8 Ho g \A yee Ma Fic. 1. Locations of pedestrian survey route (the route is highlighted in the inset) and three drift fence survey areas in Hernando County, Florida. We determined the roadside habitat occurrence of each snake by assigning half a snake to the habitat type on each side of the road. Snakes found at the junction of two roads were assigned half of a snake to each road and a quarter of a snake to the habitat type on each side of the road. We grouped the seven habitats present into natural and ruderal communities. Natural communities included sandhill, xeric hammock, modified sandhill, and modified xeric hammock habitats. Sandhill vegetation typically consists of widely spaced longleaf pines with a sparse understory of turkey oak (Quercus laevis) or other deciduous oak species and a fairly dense ground cover of wiregrass (Aristida beyrichiana), other grasses, and forbs. Xeric hammocks typically have sparse shrub and ground cover, and a canopy predominated by several evergreen oak species (e.g., QO. virginiana, Q. geminata, Q. hemisphaerica) or pignut hickory (Carya glabra). Sandhill habitat along the route included overgrown sandhill vegetation undergoing succession to xeric hammock habitat due to fire suppression. Modified sandhill and xeric hammock habitats along the route had been altered substantially by man but retained some natural overstory and understory vegetation. Typically, much of the shrub layer and ground cover had been cleared or mowed in these modified habitats, which often consisted of wooded yards. Ruderal communities included old fields, lawns, and improved pastures. Some lawns and improved pastures retained scattered oaks or longleaf pines, but the grass was short and nonnative. Oldfield habitats lacked an overstory and included overgrown pastures and mowed areas of former sandhill or xeric hammock vegetation in vacant lots or along power lines. Spatial Analyst in ArcView 3.2 was used to calculate the acreage of land-cover types present in a 5-km/* area around the survey route and in all of Hernando County using data from the Florida Fish and Wildlife Conservation Commission’s Office of Environmental Services. Elevations in the study area ranged from 18 to 37 m above sea level. Most soils were Arredondo or Candler fine sands on 0-5% slopes. RESULTS—Heterodon simus was the second most common of 18 snake species observed along the route (Table 2). All 39 H. simus observed were DOR, and 93.4% of all 228 snakes observed were DOR (Table 2). In proportion to its length, Camp Castle Road had the most DOR snakes and almost three times more DOR H. simus 192 FLORIDA SCIENTIST [VOL. 66 TABLE 1. Dimensions, mean number of DOR snakes observed per km, mean density of houses, and mean traffic volume along five roads in Hernando County, Florida. Length Width No. No. No. No. Road (m) (m) snakes/km 4H. simus/km houses/km vehicles/day Camp Castle 670 5.6 asi) 18.7 17.9 589 Haddon 380 2.5 18.4 533 18.4 NA Preston 1895 9 44.8 4.2 15.3 790 Rich Barn! 1036 Sul 36.2 35 5.8 NA Weatherly LoW3 7 24.1 6.6 liye 1! 560 ' limerock instead of paved. than did any other road (Table 1); this species accounted for 35% of all snake observations on Camp Castle Road (Table 2). Mean annual mortality was 12.8 snakes/km and 2.3 H. simus/km on the five roads combined. These mortality estimates should be considered underestimates because DOR snakes undoubtedly disappeared before being detected, particularly since we typically surveyed in early morning before H. simus and other diurnal species were active. Most DOR diurnal snakes had part of a day and all night to be removed by scavengers, consumed by ants, or displaced by motor vehicles before the route was walked the next morning. Of 38 DOR H. simus left on the road, 66% were not found the following day, including one carcass that disappeared within 1 hr after it was found. Seven carcasses remained <2 days, two carcasses remained <3 days, and two carcasses remained <4 days. In July 1998, a 5-cm segment of H. simus lasted nine days, and another carcass lasted 12 days before disappearing. When first found, most DOR H. simus were already dry and flat, and some were twisted. Red imported fire ants (Solenopsis invicta) and native ants were often observed on carcasses, and some carcasses were rapidly skeletonized by fire ants. We might have accelerated the disappearance of some carcasses from roads by detaching them from the pavement in order to measure them. During the survey, H. simus was found on the route only June-August and October-December, but it was observed elsewhere in Hernando County in April and May (Fig. 2). The earliest record we have is for a DOR juvenile H. simus found on Haddon Road on 26 February 2002 after the study ended. Heterodon simus exhibited bimodal seasonal activity, with peaks in June and October-November (Fig. 2). Adjusting for differences in monthly sampling intensity, the mean number of H. simus observed annually was highest in June (N = 3.3) and November (N = 3.9). Young-of-the-year (110-159 mm SVL or 126-203 mm TL) comprised 62% of all H. simus observations along the route and 96% of all October-December observations (Fig. 3). Ninety-two percent of H. simus found on Camp Castle Road were neonates or juveniles (Fig. 4). Hatchlings were occasionally found in close proximity, including three found on 25 October within 20 m of each other and two found on 7 November within 10 m of each other. Traffic volume was highest on Preston Road, and Camp Castle and Weatherly roads had similar traffic volumes (Table 1). Traffic volume on these three roads averaged =20 vehicles/hr from 0600 to 2200 hr EDT, with a peak hourly flow of No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 193 TaBLE 2. Number of DOR (AOR) snakes observed while walking five roads in Hernando County, Florida, for 1022 days from 3 June 1998 through 20 December 2001. Half snakes are those that were found at the junction of two roads. Camp Rich Species Castle Haddon Preston Bam Weatherly Total Cemophora coccinea 0 0 l 0 0 1 Coluber constrictor 1 2 16 (2) 8.5 Sos) (CU) 8)3) (8) Crotalus adamanteus 0 0 1 0 0 1 Diadophis punctatus 3 1 7 1 0 (1) a) Elaphe guttata 3 0 2 1 4 20 E. obsoleta 0 0 eS) 0.5 3 P) Heterodon simus WS) 2 8 325 (3) 39 Lampropeltis triangulum D2 0 5 l 1 9 Masticophis flagellum 0 0 1 0 1 2 Micrurus fulvius 2) l 2 (3) 1 1 (1) 7 (4) Nerodia fasciata 1 0 0 0 0 1 N. floridana 0 0 l 0 0 1 Opheodrys aestivus 5 0S 11 17 (1) 6.5 40 (1) Rhadinaea flavilata 0 0 1 0 0 | Storeria dekayi DS) 0 7 (1) I@) tes) 12 (3) Tantilla relicta 2 0 8 3) S) (iL) 22 (3) Thamnophis sauritus 0 0 2 0 0 2 T. sirtalis 1 0.5 0.5 0 0 2 Unidentified 1 0 0 0 2 3 Total 36 7 85 (6) 37.5 (5) ATs) (G5) = PAS) ls) 240 vehicles from 1700 to 1900 hr. Traffic volume was not recorded on Rich Barn and Haddon roads, but Haddon Road probably had the least traffic because it was a short, dead-end road that serviced only seven houses. The three roads monitored for traffic volume had similar house densities, and the limerock Rich Barn Road had the lowest house density (Table 1). The mean segment length of roadside habitats, which indicates the degree of habitat fragmentation, did not differ significantly among the seven habitat types (Kruskal-Wallis ANOVA on ranks; H = 2.59, 6 df, P = 0.86). The mean length (+ SD) was 160 m = 132 for all 76 habitat segments; it was longest for sandhill (196 m + 194) and shortest for modified sandhill (122 m + 70). Modified sandhill and sandhill had the fewest roadside segments (three and six, respectively), and lawn and modified xeric hammock had the most segments (21 and 16, respectively). The mean length of the 10 road segments containing natural habitats on both sides was 129 m = 75, whereas the mean length of the five ruderal segments was 262 m + 335, but mean lengths of natural and ruderal segments did not differ significantly (Mann- Whitney rank sum test; T= 43.0, P = 0.76). Percent coverage by the six major terrestrial land-cover types (excluding hydric habitats) in a 5-km? area around the survey route differed significantly from that of the county (X* = 229.6, 5 df, P < 0.0001), with the survey area containing more sandhill habitat (34%) and agricultural land (25%) and less pinelands (1%) and upland hardwood forest (i.e., mesic hammock; 4%) than the county. Other common 194 FLORIDA SCIENTIST [VOL. 66 —— on survey route ---- Off survey route Number of hognose snakes Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov’ Dec Fic. 2. Number of H. simus observed monthly on and off the survey route in Hernando County, Florida, from 3 June 1998 through 20 December 2001. land-cover types in the survey area were urban/barren land (21%) and mixed hardwood/pine forest (1.e., xeric hammock; 13%). The survey area consisted of 48% xeric upland habitats and 46% ruderal habitats, whereas the county had 30% xeric upland habitats and 34% ruderal habitats. However, if hydric habitats are excluded, the county had 37% xeric upland habitats and 42% ruderal habitats. Of the four longest roads, the highest snake diversity (17 species) was recorded on Preston Road (Table 2), which also had the highest roadside coverage of natural habitats (59%) and 97% of the remaining sandhill habitat (Table 3). Two additional species, the dusky pigmy rattlesnake (Sistrurus miliarius barbouri) and short-tailed snake (Stilosoma extenuatum), were recorded on Preston Road in October 2002. The proportion of H. simus found adjacent to various habitats was not significantly different from roadside coverage of these habitats GE = 8.58; 6 dit P7020); although 63% of H. simus were found adjacent to the three ruderal habitats, which covered 50% of the road route (Fig. 5; Table 3). Other snake species were not found along roads in proportion to roadside habitat coverage (X* = 64.3, 6 df, P < 0.0001); more snakes than expected were found adjacent to xeric hammocks (22%) and old fields (20%), and fewer snakes than expected were found adjacent to lawns (16%) and improved pastures (4%) (Fig. 4). In contrast, H. simus was frequently found adjacent to lawns (32%) and improved pastures (14%) (Fig. 5; Table 3). All roads had at least 35% coverage by ruderal habitats, with Camp Castle and Weatherly, the two most productive roads for H. simus, having at least 60% coverage (Table 3). Natural habitats occurred on both sides of roads 21% of the time, and 21% of H. simus were found along natural road segments. Ruderal habitats also occurred 21% of the time on both sides of roads, but 46% of H. simus were found along ruderal road segments (Fig. 4). All other snake species with at least 10 observations occurred <20% of the time along ruderal road segments. During drift fence surveys in terrestrial habitats on the Brooksville Ridge in Hernando County at Chinsegut Nature Center, Chassahowitzka Wildlife No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 195 500 A TL onroute ASVL on route @ SVL off route 400 300 200 Length of hognose snake (mm) 100 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Fic. 3. Total length (TL) or snout-vent length (SVL) and observation dates of H. simus in Hernando County, Florida. Six adult snakes found in June and July were not measured. Management Area, and the Croom Tract of Withlacoochee State Forest, we trapped 576 snakes of 24 species (Enge and Wood, 1999-2000, 2001, unpubl. data). The only two H. simus trapped were in mesic flatwoods habitat in Chassahowitzka Wildlife Management Area and in sandhill habitat in the Croom Tract. DiscussiIoN—We found June, October, and November to be the most productive months for finding H. simus on roads in Hernando County. We did not observe an April—May peak, which is usually attributed to males searching for mates (Gibbons and Semlitsch, 1987). However, eight H. simus were observed in March—April and 6 in August-September on the northern portion of the Brooksville Ridge in southwestern Alachua County (Ashton, 2002). A radiotelemetry study in South Carolina found that H. simus spent 69% of their time underground but could be found on the surface year-round, at which time they were relatively sedentary, except during peak activity periods May—June and in October (Tuberville, 2002). A radiotelemetry study in North Carolina found that H. simus spent 66% (N = 343) of their time underground or completely concealed by surface cover during their period of surface activity, which was late March-early November (Beane, 2002). Most H. simus have been collected May—June and in October in South Carolina (Gibbons and Semlitsch, 1987) and May—June (35%) and September—October (47%) in North Carolina (Palmer and Braswell, 1995). A road survey in North Carolina found 77% 196 FLORIDA SCIENTIST [VOL. 66 Fic. 4. Locations of 39 H. simus (solid circles = adults, open circles = juveniles) observed on five rural roads in Hernando County, Florida. Stippling represents predominantly forested areas. of 328 simus in October and 13% in September, but much of the sampling effort was concentrated during these months (Beane, 2002). Our peak month was November, probably because warmer weather permitted a longer activity season than in more northerly states. Four years of data on 104 H. simus collected for the pet trade in Florida showed sales every month of the year, but most sales occurred May—July (52%) and in October (16%) (Enge, unpubl. data). We found only one adult but 24 hatchling H. simus October-December, often in close proximity. Jensen (1996) found five DOR hatchlings on a 0.6-km stretch of road during a 3-day period in October in Okaloosa County, Florida. Our smallest hatchlings were shorter than those found by Jensen (1996), but the twisted condition of some of them precluded accurate measurements. Young-of-the-year comprised 62% of all H. simus observations along the route, which suggests that road mortality No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 197 TABLE 3. Percent coverage of seven habitat types along five roads in Hernando County, Florida. Percent of DOR H. simus found adjacent to various roadside habitats appears in parentheses. Habitat Camp Castle Haddon Preston Rich Barn Weatherly Total Sandhill 0 (0) lis -Bil(64) (GQ) 0 (0) ila) (©) Xeric hammock 0 (0) 125) Sel) 19 (0) 8 (8) 12 (4) Modified sandhill 12 (4) 15 (0) 0 (0) 0 (0) 2 (4) 316) Modified hammock 27 (24) 28 (0) 17 (25) 25 (43) 18 (12) 20 (20) Old field 6 (4) 0 (0) 10 (16) 33 (50) 10 (23) 12 (17) Residential lawn 52 (60) 85) @0) Zady) 14 (0) PI PZ) D5. (62) Improved pasture 2 (8) 0 (0) 6 (6) 8 (0) 31 (31) 13 (14) may be a significant factor for dispersing hatchlings, as has been noted for other snake species (Bonnet et al., 1999). All H. simus found on our route were DOR. Eighty- seven percent of 328 H. simus found on roads in North Carolina were DOR, despite much of the survey being conducted during hours of peak activity (Beane, 2002). Almost two-thirds of DOR H. simus were observed only once before disap- pearing, presumably due to removal by scavengers, consumption by ants, or dis- placement by motor vehicles. The rapid disappearance of DOR H. simus suggests that the actual highway mortality was higher than 2.3 H. simus/km/yr, especially considering that we surveyed in early morning prior to peak activity, which apparently occurs from 1000—1600 hr EDT in North Carolina (84% of 57 sightings) (Beane, 2002). The most intensive sampling of the snake community probably occurs during periods of peak traffic flow when higher mortality rates presumably occur (Seigel, 1986; Mendelson and Jennings, 1992). During the peak activity period for H. simus, an average of 30.5 vehicles/hr traveled the three roads that were monitored for traffic volume. Preston Road was the most heavily traveled road and had the most DOR snakes, but Camp Castle Road had the most DOR snakes in proportion to its length. In rural northern Alabama, road mortality of snakes was not correlated with traffic volume, possibly because snake populations had already declined along heavily traveled roads or carcasses disappeared more quickly from these roads (Dodd et al., 1989). The rural roads we surveyed had relatively low traffic volume (<1000 vehicles/day), whereas a study that estimated the ecological effects of roads assumed that primary roads in rural areas were used by 10,000 vehicles/day (Forman, 2000). Road surveys from moving vehicles tend to miss seeing short, slender snakes (Fitch, 1949; Pough, 1966; Dodd et al., 1989), but our pedestrian survey detected four such species (Tantilla relicta, Diadophis punctatus, Storeria dekayi, Rhadinaea flavilata) and hatchlings of larger species. Heterodon simus is a short, stout species that can be seen while driving, although small or flattened DOR individuals are often difficult to discern. Some motorists intentionally run over or avoid running over snakes, but we suspect that most motorists do not see the H. simus they run over, particularly hatchlings. Heterodon simus is seldom trapped during drift fence surveys (Enge, 1997), possibly because its fossorial habits and sedentary nature make it less susceptible to trapping than more surface-active species. However, this behavior does not account 198 FLORIDA SCIENTIST [VOL. 66 Mm Habitat coverage [1 Southern hognose snake [=] Other snake species Percent Sandhill Xeric Modified Modified Oldfield Residential Improved hammock — sandhill hammock lawn pasture Fic. 5. Percent roadside coverage of various habitats and percent occurrence of H. simus and all other snake species adjacent to these habitats along five rural roads in Hernando County, Florida. for all the observed differences in the relative abundance of H. simus obtained from road surveys versus drift fence surveys. Some preferred prey species, particularly toads (Goin, 1947; Conant and Collins, 1991; Palmer and Braswell, 1995; Beane et al., 1998), might be more abundant or susceptible to capture in the fragmented, degraded natural habitats or ruderal habitats along the road route than in the natural, intact habitats on nearby public lands. The southern toad (Bufo terrestris) and eastern spadefoot (Scaphiopus holbrookii), which were abundant along the road route despite the scarcity of wetlands, are common residents of lawns (Neill, 1950), although S. holbrookii appears especially susceptible to habitat degradation as- sociated with residential development (Delis et al., 1996). However, xeric upland habitats on public lands contained more diverse lizard and anuran prey for H. simus, and drift fences were typically situated near ephemeral wetlands that supported large anuran populations (Enge and Wood, 1999-2000, 2001). Almost half of our H. simus were found along completely ruderal road seg- ments, which comprised only 21% of the route. Tuberville (2002) and Beane (2002) also found H. simus in ruderal habitats, and Tuberville and co-workers (2000) con- cluded that it was apparently less habitat specific than many animals endemic to longleaf pine forests. A well-developed herbaceous layer, which may be important to H. simus and its prey, was present in modified sandhill habitat, old fields, and many lawns along the route. Lawns occurred along >50% of Camp Castle Road, where almost three times more H. simus/km, primarily juveniles, were recorded than on any other road. Open habitats might be favored for nesting, which would explain the high percentage of juvenile H. simus recorded on roads adjacent to lawns and pastures. More H. simus might be found on roads adjacent to lawns and pastures because snakes encountering these habitats exhibit increased activity because the short grass and reduced surface debris provide little cover. No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 199 The eastern hognose snake (Heterodon platirhinos) was not found during road or drift fence surveys in Hernando County, despite this species being more abundant and ubiquitous than H. simus in Florida. Conversations with knowledgeable residents of Hernando County yielded numerous sightings of H. simus, but no sightings of H. platirhinos. During drift fence surveys in Florida, 37 H. platirhinos were trapped in 10 habitats, whereas 17 H. simus were trapped in three habitats (Enge, 1997, unpubl. data). According to Conant and Collins (1991), H. simus is found in sandy woods, fields and groves, dry river floodplains, and hardwood hammocks. The latter two habitats might be considered atypical, although 11 H. simus were trapped in loblolly pine (Pinus taeda) cabbage palm (Sabal palmetto) hammocks at St. Marks National Wildlife Refuge, Wakulla County, Florida (U.S. Fish and Wildlife Service, 1980). Heterodon simus has also been observed using mesic hammock habitat in Alachua County, Florida (Goin and Goin, 1953). The remaining drift fence captures in Florida were two snakes in mesic flatwoods and four snakes in sandhill habitat (Enge, 1997, unpubl. data). Most other observations of H. simus in Florida have been in sandhill habitat (Stevenson and Crowe, 1992; Jensen, 1996; Florida Natural Areas Inventory, 2001; Enge, pers. obs.). In south- western Alachua County, Florida, 10 H. simus were observed on sand roads through sandhill habitat, and four.H. simus were observed on a limerock road adjacent to an improved pasture within 50 m of a xeric hammock and sandhill pond (Ashton, 2002). None of the five surveyed roads appears on the 1954 topographic quadrangle map, although an unimproved dirt road ran to Camp Quarry, and a railroad line was present where Rich Barn Road is now. Residential development is rapidly occurring along our surveyed roads, and Hernando County has been one of the three fastest growing counties in Florida each of the past three decades (Bureau of Economic and Business Research, 2000). Over 45% of our survey area consisted of ruderal habitats, which was >10% higher than the overall coverage of ruderal habitats in the county. We suspect that snake populations, especially slow-moving species like H. simus or active species with large home ranges, will experience future declines in the area due to cumulative road mortality and increased traffic, as has been observed elsewhere (Rosen and Lowe, 1994; however, see Sullivan, 2000). Although H. simus remains common in the survey area, habitat fragmentation and road mortality have apparently already reduced populations of large snake species typical of xeric uplands: eastern coachwhip (Masticophis f. flagellum), Florida pine snake (Pituophis melanoleucus mugitus), eastern indigo snake (Drymarchon corais couperi), and eastern diamondback rattlesnake (Crotalus adamanteus). These four species comprised only 1.3% of the 228 snakes observed on roads but 4.7% of the 576 snakes trapped by drift fences in terrestrial habitats on nearby public lands, despite the bias of drift fences against catching large snakes (e.g., Campbell and Christman, 1982b; Greenberg et al., 1994; Enge, 2001). In contrast, these four large species comprised 46.7% of snake sightings (N = 246; 21 species) over a 5-yr period on a more lightly traveled, 8-km limerock road through more intact sand- hill habitat in Alachua County, Florida, where only 25.6% of snakes were DOR (Ashton, 2002). 200 FLORIDA SCIENTIST [VOL. 66 Coverage of xeric upland habitats in our survey area was 47.8% compared to only 30.5% in Hernando County, but upland habitats in our survey area were fragmented, whereas large tracts of contiguous upland habitat were present on public lands elsewhere in the county. Although 46% of habitats in the survey area had been converted to agricultural or residential use, sandhill remained the predominant habitat type with 34% coverage. Lack of fire in the remnant patches of natural sandhill habitat, which were most prevalent along Preston Road, resulted in a dense shrub layer and reduced ground cover. Prescribed burning of sandhill fragments in the area is no longer feasible due to the proximity of houses, but clearing of understory vegetation and mowing in modified upland habitats, old fields, and lawns may partially mimic fire and create the open conditions favored by foraging or nesting H. simus. Besides highway mortality, a factor that might contribute to the decline of H. simus 1s predation on eggs or young by red imported fire ants (Mount, 1981), which have invaded much of the species’ range (Tuberville et al., 2000). In undisturbed sandhill habitat, red imported fire ants are almost completely restricted to pond margins and roadsides, but fire ants are numerous in ruderal habitats and heavily disturbed forests (Tschinkel, 1988). Ruderal habitats in our study area were infested with fire ants, which frequently skeletonized DOR snakes. Human persecution and exploitation pose additional threats to H. simus. Heterodon simus is intentionally killed by ignorant persons due to its “bluff” display (Myers and Arata, 1961) and superficial color and pattern similarities to the dusky pigmy rattlesnake, which may be mimetic (Neill, 1963). Residents in this part of Florida often call H. simus “‘ground rattlers’’ (Griswold, 2000). The demand for this species in the pet trade has increased recently, and hatchlings have sold for $150 each (MacInnes, 2001). In 1990-94, at least 135 wild-caught H. simus from Florida were reported sold in the pet trade (Enge, in press). Based upon commercial collecting data, H. simus is apparently still locally common in the Panhandle in Calhoun and Jackson counties, along the entire Brooksville Ridge from Gilchrist County to Pasco County, and along the Atlantic Coastal Ridge in Brevard, Indian River, and St. Lucie counties (Enge, unpubl. data). Interestingly, Tuberville and co- workers (2000) documented no observations of this species from any of these Panhandle or Atlantic Coast counties in over three decades, and their last record for Hernando County was in 1963. However, the least comprehensive data were obtained from Florida, and many of the museum records did not have associated dates (Tuberville, 2002). The expanding network of roads threatens wildlife populations due to increased highway mortality, greater accessibility of remote areas to collectors, and fragmen- tation of individual home ranges and regional populations (Lodé, 2000; Trombulak and Frissell, 2000). The tolerance of H. simus for fragmented ruderal habitats might allow populations to persist in one of Florida’s fastest growing counties. How- ever, snake populations will be detrimentally impacted in the long term by the con- tinued destruction and degradation of natural habitats, along with the proliferation of roads and residential communities. The road with the highest diversity of snake species had the greatest roadside coverage of natural habitats, particularly sandhill. No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 201 In addition to highway mortality and direct human persecution or exploitation, H. simus is potentially threatened by declining populations of amphibians, their primary prey, and depredation of eggs and hatchlings by imported fire ants. ACKNOWLEDGMENTS—We are grateful to Jeff Beane (North Carolina State Museum of Natural Sciences), Ray Ashton (Ashton Biodiversity Research & Preservation Institute), and Tracey Tuberville (Savannah River Ecology Laboratory) for sharing their unpublished data. We thank Darrell Davis, Engineering Division, Hernando County Public Works Department for installing traffic counters on roads. Beth Stys graciously conducted land-cover analysis. Paul Moler, Jeff Gore, Jeff Beane, Tracey Tuberville, and Kenney Krysko reviewed this manuscript. LITERATURE CITED ASHTON, R. E., JR. 2002. Ashton Biodivers. & Res. Preserv. Inst., Newberry, FL. Pers. Comm. BEANE, J. C. 2002. North Carolina State Mus. Nat. Sci., Raleigh, NC. Pers. Comm. . T. J. THorP, AND D. A. JACKAN. 1998. Heterodon simus (southern hognose snake) diet. Herpetol. Rev. 29: 44-45. BONNET, X., G. NAULLEAU, AND R. SHINE. 1999. The dangers of leaving home: dispersal and mortality in snakes. Biol. Conserv. 89: 39-50. BUREAU OF ECONOMIC AND BUSINESS RESEARCH. 2000. Florida Statistical Abstract 2000. Univ. Florida, Gainesville, FL. CAMPBELL, H. W. AND S. P. CHRISTMAN. 1982a. Field techniques for herpetofaunal community analysis. Pp. 193-200. In: Scott, N. J., JR. (ed.), Herpetological Communities. Wildl. Res. Rept. 13, U.S. Fish Wildl. Serv., Washington, DC. AND S. P. CHRISTMAN. 1982b. The herpetofaunal components of Florida sandhill and sand pine scrub associations. Pp. 163-171. In: Scort, N. J., JR. (ed.), Herpetological Communities. Wildl. Res. Rept. 13, U.S. Fish Wildl. Serv., Washington, DC. CONANT, R. AND J. T. Couns. 1991. A Field Guide to Reptiles and Amphibians of Eastern and Central North America (3rd Edition). Houghton Mifflin Co., Boston, MA. 450 pp. Deis, P. R., H. R. MuUSHINSKy, AND E. D. Mccoy. 1996. Decline of some west-central Florida anuran populations in response to habitat degradation. Biodivers. Conserv. 5: 1579-1595. Dopp, C. K., Jr. 1992. Biological diversity of a temporary pond herpetofauna in north Florida sandhills. Biodivers. Conserv. 1: 125-142. . 1995. Reptiles and amphibians in the endangered longleaf pine ecosystem. Pp. 129-131. Jn: LARog, E. T., G. S. Farris, C. E. PucKETT, P. E. DoRAN, AND M. J. Mac (eds.), Our Living Resources: A Report to the Nation on the Distribution, Abundance, and Health of U.S. Plants, Animals, and Ecosystems. U.S. Dept. Interior, Natl. Biol. Serv., Washington, DC. . K. M. Ence, ann J. N. Stuart. 1989. Reptiles on highways in north-central Alabama. J. Herpetol. 23: 197-200. EncE, K. M. 1997. A standardized protocol for drift-fence surveys. Florida Game and Fresh Water Fish Comm. Tech. Rept. No. 14, Tallahassee, FL. . 2001. The pitfalls of pitfall traps. J. Herpetol. 35: 467-478. . In Press. Commercial harvest of amphibians and reptiles in Florida for the pet trade. In: MeEsuHaKA, W. E., JR. AND K. J. Bassitr (eds.), Status and Conservation of Florida’s Amphibians and Reptiles. Krieger, Malabar, FL. AND K. N. Woop. 1999-2000. A herpetofaunal survey of Chassahowitzka Wildlife Management Area, Hernando County, Florida. Herpetol. Nat. Hist. 7: 117-144. AND . 2001. Herpetofauna of Chinsegut Nature Center, Hernando County, Florida. Florida Scient. 64: 283-305. AND . 2002. A pedestrian road survey of an upland snake community in Florida. Southeast. Nat. 1: 365-380. Fitcu, H. S. 1949. Road counts of snakes in western Louisiana. Herpetologica 5: 87-90. FLORIDA NATURAL AREAS INVENTORY. 2001. Element occurrence database, Tallahassee, FL. 202 FLORIDA SCIENTIST [VOL. 66 ForMAN, R. T. T. 2000. Estimate of the area affected ecologically by the road system in the United States. Conserv. Biol. 14: 31-35. GrBBons, J. W. AND R. D. SeEMuITScH. 1987. Activity patterns. Pp. 478-513. In: SEIGEL, R. A., J. T. COLLINS, AND S. S. Novak (eds.), Snakes: Ecology and Evolutionary Biology. Macmillan, New York, NY. Gow, C. J. 1947. A note on the food of Heterodon simus. Copeia 1947: 275. AND O. B. Gorn. 1953. Temporal variation in a small community of amphibians and reptiles. Ecology 34: 406-408. GREENBERG, C. H., D. G. NEARY, AND L. D. Harris. 1994. A comparison of herpetofaunal sampling effectiveness of pitfall, single-ended, and double-ended funnel traps used with drift fences. J. Herpetol. 28: 319-324. GrIswoLbD, W. G., IV. 2000. The hognose snake (Heterodon simus): All you ever wanted to know. Reptiles Mag. 8(6): 32-45. Guyer, C. AND M. A. BAILEy. 1993. Amphibians and reptiles of longleaf pine communities. Pp. 139-158. In: HERMANN, S. M. (ed.), The Longleaf Pine Ecosystem: Ecology, Restoration and Management. Proc. Tall Timbers Fire Ecology Conference No. 18, Tall Timbers Res. Stn., Tallahassee, FL. JENSEN, J. B. 1996. Heterodon simus (southern hognose snake). Hatchling size. Herpetol. Rev. 27: 25. Kautz, R. S., D. T. GILBERT, AND G. M. MAuLDIN. 1993. Vegetative cover in Florida based on 1985-1989 Landsat Thematic Mapper Imagery. Florida Scient. 56: 135-154. KLAuBER, L. M. 1939. Studies of the reptile life in the arid Southwest. I. Night Moree on the deserts with ecological statistics. Bull. Zool. Soc. San Diego 14: 1-79. Lope, T. 2000. Effect of a motorway on mortality and isolation of wildlife populations. Ambio 29: 163-166. MacInnes, R. 2001. Glades Herp Inc., Ft. Myers, FL. Pers. Comm. Means, D. B. AND G. Grow. 1985. The endangered longleaf pine community. ENFO (Florida Conserv. Found.) Rept. 85: 1-12. MENDELSON, J. R., I] AND W. B. JENNINGS. 1992. Shifts in the relative abundance of snakes in a desert grassland. J. Herpetol. 26: 38-45. Mount, R. H. 1981. The red imported fire ant, Solenopsis invicta (Hymenoptera: Formicidae), as a possible serious predator on some native southeastern vertebrates: Direct observations and subjective impressions. J. Alabama Acad. Sci. 52: 71-78. Myers, C. W. AND A. A. Arata. 1961. Remarks on “defensive” behavior in the hognose snake Heterodon simus (Linnaeus). Quart. J. Florida Acad. Sci. 24: 108-110. Myers, R. L. 1990. Scrub and high pine. Pp. 150—193. In: Myers, R. L. AND J. J. Ewe (eds.), Ecosystems of Florida. Univ. Central Florida Press, Orlando, FL. NeEwL, W. T. 1950. Reptiles and amphibians in urban areas of Georgia. Herpetologica 6: 113-116. . 1963. Polychromatism in snakes. Quart. J. Florida Acad. Sci. 26: 194-216. Noss, R. F. 1988. The longleaf pine landscape of the Southeast: almost gone and almost forgotten. Univ. Michigan School Nat. Resour., Endangered Species Update 5(5): 1-8. PALMER, W. M. AND A. L. BRASWELL. 1995. Reptiles of North Carolina. Univ. North Carolina Press, Chapel Hill, NC. 412 pp. Poucu, F. H. 1966. Ecological relationships of rattlesnakes in southeastern Arizona with notes on other species. Copeia 1966: 676-683. Rosen, P. C. AND C. H. Lowe. 1994. Highway mortality of snakes in the Sonoran Desert of southern Arizona. Biol. Conserv. 68: 143-148. SEIGEL, R. A. 1986. Ecology and conservation of an endangered rattlesnake, Sistrurus catenatus, in Missouri, USA. Biol. Conserv. 35: 333-346. STEVENSON, D. AND D. Crowe. 1992. Geographic distribution: Heterodon simus (southern hognose snake). Herpetol. Rev. 23: 91. SULLIVAN, B. K. 2000. Long-term shifts in snake populations: a California site revisited. Biol. Conserv. 94: 321-325. TROMBULAK, S. C. AND C. A. FRISSELL. 2000. A review of ecological effects of roads on terrestrial and aquatic communities. Conserv. Biol. 14: 18-30. No. 3 2003] ENGE AND WOOD—HOGNOSE SNAKE ROADKILL 203 TSCHINKEL, W. R. 1988. Distribution of the fire ants Solenopsis invicta and S. geminata (Hymenoptera: Formicidae) in northern Florida in relation to habitat and disturbance. Annals Entomol. Soc. Amer. 81: 76-81. TUBERVILLE, T. D. 2002. Savannah River Ecol. Lab., Aiken, SC. Pers. Comm. . J. R. Boprr, J. B. JENSEN, L. LACLAIRE, AND J. W. GiBBoNS. 2000. Apparent decline of the southern hog-nosed snake, Heterodon simus. J. Elisha Mitchell Scient. Soc. 116: 19-40. U.S. FisH AND WILDLIFE SERVICE. 1980. St. Marks National Wildlife Refuge: forestry management and non-game wildlife. Final report. Natl. Fish and Wildl. Lab., Gainesville, FL. Wark, S., C. Frost, AND P. D. Doerr. 1993. Southern mixed hardwood forest: the former longleaf pine forest. Pp. 447-493. In: Martin, W. H., S. G. Boyce, AND A. C. ECHTERNACHT (eds.), Biodiversity of the Southeastern United States. Lowland Terrestrial Communities. John Wiley and Sons, New York, NY. Wuite, W. A. 1970. Geomorphology of the Florida Peninsula. Florida Bur. Geol. Bull. 51: 1-164. Florida Scient. 66(3): 189-203. 2003 Accepted: November 13, 2002 Biological Sciences THE DISTRIBUTION OF HEMIDACTYLUS (SAURIA: GEKKONIDAE) IN NORTHERN PENINSULAR FLORIDA JOSIAH H. TOWNSEND AND KENNETH L. KRYSKO Florida Museum of Natural History, Division of Herpetology, P.O. Box 117800, University of Florida, Gainesville, FL 32611, USA ABSTRACT: Four species of Hemidactylus geckos have been introduced into Florida. Presently, only the Indo-Pacific (H. garnotii) and Mediterranean (H. turcicus) geckos occur in northern peninsular Florida; however, their distributions and ecological status are poorly known. We combined records from the literature, systematic collections, and field surveys to summarize the distributions of H. garnotii and H. turcicus in northern peninsular Florida. Herein, we document 16 previously unreported county records and one significant distributional record for these two species. Key Words: Gekkonidae, Hemidactylus garnotii, Hemidactylus turcicus, gecko, introduced species, Florida FLORIDA has a well-documented exotic herpetofauna (Duellman and Schwartz, 1958; King and Krakauer, 1966; Wilson and Porras, 1983; Dalrymple, 1994; Butterfield et al., 1997; Townsend et al., 2002). Prominent members of Florida’s introduced herpetofauna are geckos, primarily the genus Hemidactylus of Old World origin. Four species of Hemidactylus have been introduced in Florida: the com- mon house gecko (H. frenatus Duméril and Bibron, 1836), Indo-Pacific gecko (H. garnotii Duméril and Bibron, 1836), tropical house gecko (H. mabouia [Moreau de Jonneés, 1818]), and Mediterranean gecko (H. turcicus [Linnaeus, 1758]). Hemi- dactylus frenatus is native to Africa, Asia, Australia, and Polynesia (Welch, 1994), and has been introduced in North America and throughout Central and South America. In Florida, H. frenatus is currently limited to Key West and Stock Island, Monroe County (Meshaka et al., 1994b), and Ft. Myers, Lee County. Hemidactylus garnotii is native to northeastern India, southern China, the Malay Peninsula, Indonesian archipelago, Philippines, and Oceania (Welch, 1994). King and Krakauer (1966) first reported H. garnotii as established in Miami-Dade County, Florida, prior to 1964, and they stated that its introduction likely resulted from being transported by researchers upon their return from the 1960-1963 International Indian Ocean Expedition. Hemidactylus garnotii expanded its range rapidly and has subsequently been recorded in the Florida Keys and southern peninsula (Wilson and Porras 1983), and northern peninsula (Stevenson and Crowe, 1992; Reppas, 1999; Lindsay and Townsend, 2001). Hemidactylus mabouia is found in central and southern Africa, the east coast of South America from Uruguay north to Suriname, the Amazon Basin west to its headwaters in Bolivia, Peru, and Ecuador, and the Caribbean in Trinidad 204 No. 3 2003] TOWNSEND AND KRYSKO—GECKOS IN NORTH FLORIDA 205 and Tobago and throughout the Lesser Antilles (Powell et al., 1998). Presently, H. mabouia is found throughout the Florida Keys (Lawson et al., 1991) and the southern peninsula (Meshaka et al., 1994a), with records as far north as Orange County (Butterfield et al., 2000). Hemidactylus turcicus is native to coastal regions of the Mediterranean in Europe and Africa, the Red Sea in Egypt, Somalia, Arabian Peninsula, and Persian Gulf, to western India along the Indian Ocean (McCoy, 1970). The introduction of this species has been reported from Arizona, California, Louisiana, Texas, and along the Gulf of Mexico south to the northwestern Yucatan Peninsula (Lee, 2000). Fowler (1915) first reported the introduction of H. turcicus (as H. mabouia, ANSP 18035) in 1910 from Key West, Monroe County, Florida. Hemidactylus turcicus has subsequently been recorded throughout the Florida Keys (Duellman and Schwartz, 1958), southern peninsula (King and Krakauer, 1966), northern peninsula (King, 1958; Meylan, 1977; Wise, 1993; Townsend and Reppas, 2001; Townsend et al., 2002), and panhandle (Nelson and Carey, 1993). Presently, Hemidactylus garnotii and H. turcicus are the only two members of that genus known to occur in northern peninsular Florida, but their geographic distributions are poorly known. While conducting recent surveys, it became appar- ent that H. garnotii and H. turcicus were even more widespread in the northern peninsula than had previously been reported in the literature. Herein, we document the geographic distributions of these two species in northern peninsular Florida. MeETHOps—Surveys were conducted from 14 June to 15 November 2001 throughout northern Florida. Additionally, we obtained records from the literature and systematic collections throughout the United States. These geckos are nocturnal and easily observed around lights at night in urban settings (King, 1958; Punzo, 2001); thus, we surveyed the outside walls of buildings after dark between 2000 and 0200 hrs. Voucher specimens were deposited in the Florida Museum of Natural History (FLMNH), University of Florida (UF collection). The region referred to herein as the northern peninsula follows Enge (1997). Source acronyms follow Leviton and co-workers (1985). RESULTS—Five new distributional records were collected during our surveys: one county and one mainland record for Hemidactylus garnotii, and three county records for H. turcicus. The county record for H. garnotii was from Bradford County, Hampton, on the wall of a store at the junction of SR 301 and CR 18, on 27 September 2001 (UF 128022). The mainland record for H. garnotii was from Levy County, Bronson, on the wall of a store at the junction SR 500 and Gilbert Street, on 15 October 2001 (UF 128024). The Levy County record for H. garnotii 1s significant because this species had previously only been recorded in Levy County from Cedar Key (Table 1). New county records for H. turcicus were from Bradford County, Starke, on the wall of a shopping center north of the junction of SR 301 and CR 100- A, on 27 September 2001 (UF 128023); Clay County, Keystone Heights, City Municipal Building, on 27 September 2001 (UF 128020-—21); and Marion County, Ocala, on the wall of a shopping center at the junction of US 441 and SW 7th Street, on 16 Oct 2001 (UF 128025-—28). A search of systematic collections produced twelve previously unreported county records, including Hemidactylus garnotii from Alachua (UF 87825-—27), Baker (UF 94949), Hernando (UF 99771-72), Levy (UF 87721), Orange (UF 206 FLORIDA SCIENTIST [VOL. 66 TABLE 1. Distributions of the Indo-Pacific (Hemidactylus garnotii) and Mediterranean (H. turcicus) geckos in northern peninsular Florida. Specimens with no citation under source have not been previously reported in the literature. Species County Locality Date Source H. garnotii Alachua Gainesville 4 Sep 1993 UF 87825-27 Baker Macdenny 6 Nov 1994 UF 94949 Bradford Hampton 27 Sep 2001 UF 128022 Citrus Camp Cove 9 May 1991 Stevenson and Crowe, Campground 1992 (UF 80802) Flagler The Whitney 13 Jun 2001 Lindsay and Townsend, Laboratory 2001 (UF 124688) Hernando Hernando Beach 15 Jul 1995 UF 99771-72 Levy Cedar Key 22 Aug 1993 UPIS7721 Levy Bronson 15 Oct 2001 UF 128024 Orange Orlando 27 Mar 1983 UF 53909 Putnam Palatka 10 Jan 1991 UF 79999 Seminole WDW Environmental 20 Mar 1980 UCF 1314 Protection Lab St Johns Anastasia Island 23 Jun 1988 UF 69310 Volusia South Daytona 22 Nov 1998 Reppas, 1999 (UF 116050) H. turcicus Alachua Gainesville 25 Oct 1956 King, 1958 (UF 8917) Bradford Starke 27 Sep 2001 UF 128023 Citrus Inverness 2,O0ct 11999 Townsend ef ai., 2001 (CAS 210987) Clay Keystone Heights 27 Sep 2001 UF 128020-21 Columbia Ellisville 25 Jun 2001 Townsend and Reppas, 2001 (UF 124750) Duval Jacksonville Apr-May 1970 Meylan, 1977 (UF 37275) Hernando Weeki Wachee 15 Jul 1995 UF 99770 Gardens Levy Cedar Key 22 Aug 1993 UF 87725-—26 Marion Ocala 16 Oct 2001 UF 128025-28 Orange Orlando 20 Oct 1983 MCZ 166912 Putnam Melrose 30 Aug 1997 UF 123265 Seminole Sanford 29 Oct 1981 UCF 1321-23 St Johns Madeira Heights 24 Jan 1993 Wise, 1993 (UF 86816) 53909), Putnam (UF 79999), Seminole (UCF 1314), and St. Johns (UF 69310) counties, and H. turcicus from Hernando (UF 99770), Orange (MCZ 166912), Putnam (UF 123265), and Seminole (UCF 1321-23) counties (Table 1). Hemi- dactylus turcicus had been recorded in 1997 from Cedar Key, Levy County (Means, 1999), however museum records indicate the presence of H. turcicus on Cedar Key as early as 1993 (UF 87725-26). DiscussioN—In Florida, Hemidactylus exhibits a stratified diffusion invasion No. 3 2003] TOWNSEND AND KRYSKO—GECKOS IN NORTH FLORIDA 207 pattern (Shigesada and Kawasaki, 1997), likely a result of man-assisted trans- portation along trucking routes (Davis, 1974; Godley et al., 1981; Meshaka, 1995). This pattern is similar to the one exhibited by the introduced brown anole (Anolis sagrei) in Florida (Campbell, 1996). Moreover, H. garnotii has apparently replaced the longer-established H. turcicus in much of southern Florida (Meshaka, 1995; Butterfield et al., 1997), and it appears that H. mabouia may now be displacing H. garnotii in those same areas (Krysko, pers. obs.). With the colonizing success that Hemidactylus has had in Florida until now, these geckos will likely continue to disperse throughout the state in future years. ACKNOWLEDGMENTS—We would like to thank Deborah T. Vergara for invaluable field assistance. Boyd Blihovde, Billy Griswold, Steve "Picklebarrel" Johnson, Anthony T. Reppas, and Louis A. Somma collected some of the specimens reported in this paper. Stacy Kubis (UCF), Jose P. Rosado (MCZ), Robin Lawson (CAS), and Ned Gilmore (ANSP) provided collections records and Todd S. Campbell provided helpful comments on a draft of this manuscript. LITERATURE CITED BUTTERFIELD, B. P., W. E. MESHAKA, JR., AND C. Guyer. 1997. Nonindigenous Amphibians and Reptiles. Pp. 123-138. In D. SimBertorr, D. C. SCHMITZ, AND T. C. BROWN (eds.), Strangers in Paradise. Impact and Management of Nonindigenous Species in Florida. Island Press, Washington D.C. 467 pp. , 1. Fox, J. GARNER, K. CARTER, AND J. B. HAuGgE. 2000. Geographic distribution: Hemidactylus mabouia. Herpetol. Rev. 31: 53. CAMPBELL, T. S. 1996. Northern range expansion of the brown anole (Anolis sagrei) in Florida and Georgia. Herpetol. Rev. 27: 155-157. DALRYMPLE, G. H. 1994. Non-indigenous amphibians and reptiles in Florida. Pp. 67—78. In D. C. SCHMIDT AND T. C. Brown (eds.), An Assessment of Invasive Non-indigenous Species in Florida’s Public Lands. Florida Department of Environmental Protection Tech. Rep. TSS-94—100, Tallahassee. 303 pp. Davis, W. K. 1974. The Mediterranean gecko, Hemidactylus turcicus, in Texas. J. Herpetol. 8: 77-80. DUELLMAN, W. E. AND A. SCHWARTZ. 1958. Amphibians and reptiles of southern Florida. Bull. Florida State Mus. Biol. Sci. 3: 181-324. EncE, K. M. 1997. Habitat occurrence of Florida’s native amphibians and reptiles. Florida Game and Freshwater Fish Comm. Tech. Rep. No. 16, Tallahassee. 44 pp. Fow.er, H. W. 1915. Cold-blooded vertebrates from Florida, the West Indies, Costa Rica, and eastern Brazil. Proc. Acad. Natur. Sci. Philadelphia 67: 244-269. Gop Ley, J. S., F. E. Lourer, J. N. LAYNE, AND J. Rossi. 1981. Distributional status of an introduced lizard in Florida: Anolis sagrei. Herpetol. Rev. 12: 84-86. Kine, W. 1958. Observations on the ecology of a new population of the Mediterranean gecko, Hemi- dactylus turcicus, in Florida. Quart. J. Florida Acad. Sci. 2: 317-318. AND T. KRAKAUER. 1966. The exotic herpetofauna of southeast Florida. Quart. J. Florida Acad. Sci. 29: 144-154. Lawson, R., P. G. FRANK, AND D. L. Martin. 1991. A gecko new to the United States herpetofauna, with notes on geckos of the Florida Keys. Herpetol. Rev. 22: 11-12. Lee, J. C. 2000. A Field Guide to the Amphibians and Reptiles of the Maya World: The Lowlands of Mexico, Northern Guatemala, and Belize. Cornell Univ. Press, Ithaca, New York. 402 pp. Leviton, A. E., R. H. Gress, Jr., E. HEAL, AND C. E. DAwson. 1985. Standards in herpetology and ichthyology: Part I. Standard symbolic codes for institutional resource collections in herpetology and ichthyology. Copeia 1985: 802-832. Linpsay, C. R. AND J. H. TowNsenpb. 2001. Geographic distribution: Hemidactylus garnotii. Herpetol. Rev. 32: 193. MEANS, R. C. 1999. Geographic distribution: Hemidactylus turcicus. Herpetol. Rev. 30: 52. 208 FLORIDA SCIENTIST [VOL. 66 Mccoy, C. J. 1970. Hemidactylus turcicus. Cat. American Amphib. Rept. 87: 1-2. MesHAKA, W. E., JR. 1995. Reproductive cycle and colonization ability of the Mediterranean gecko (Hemidactylus turcicus) in south-central Florida. Florida Scient. 58: 10-15. , B. P. BUTTERFIELD, AND J. B. HauGE. 1994a. Hemidactylus mabouia as an established member of the Florida herpetofauna. Herpetol. Rev. 25: 80-81. : , AND . 1994b. Hemidactylus frenatus established in southern Florida. Herpetol. Rev. 25: 127-128. MEYLAN, P. 1977. Geographic distribution: Hemidactylus turcicus. Herpetol. Rev. 8: 39. NELson, D. H. and S. D. Carey. 1993. Range extension of the Mediterranean gecko (Hemidactylus turcicus) along the northeastern Gulf Coast of the United States. Northeast Gulf Sci. 13: 53-58. PowELL, R., R. I. CROMBIE, AND H. E. A. Boos. 1998. Hemidactylus mabouia. Cat. American Amphib. Rept. 674: 1-11. Punzo, F. 2001. The Mediterranean gecko, Hemidactylus turcicus: Life in an urban landscape. Florida Scient. 64: 56-66. Reppas, A. T. 1999. Geographic distribution: Hemidactylus garnotii. Herpetol. Rev. 30: 110. SHIGESADA, N. AND K. KAWASAKI. 1997. Biological Invasions: Theory and Practice. Univ. Press. Oxford, New York, New York. 205 pp. STEVENSON, D. AND D. Crowe. 1992. Geographic distribution: Hemidactylus garnotii. Herpetol. Rev. 23: 90. TOWNSEND, J. H., K. L. Krysko, A. T. REpPAS, AND C. M. SHEEHY. 2002. Noteworthy records of introduced reptiles and amphibians from Florida, USA. Herpetol. Rev. 33: 75. AND A. T. Reppas. 2001. Geographic distribution. Hemidactylus turcicus. Herpetol. Rev. 32: 193. WELCH, K. R. G. 1994. Lizards of the World: A Checklist. 1. Geckos. KCM Books, Bristol, England. 165 pp. Wiison, L. D. AND L. Porras. 1983. The ecological impact of man on the south Florida herpetofauna. Univ. Kansas Mus. Nat. Hist. Spec. Publ. 9: 1-89. Wise, M. A. 1993. Geographic distribution. Hemidactylus turcicus. Herpetol. Rev. 24: 109. Florida Scient. 66(3): 204-208. 2003 Accepted: November 26, 2002 Biological Sciences CUES USED BY THE GOLDEN MOUSE, OCHROTOMYS NUTTALLI, TO ASSESS THE PALATABILITY OF APOSEMATIC PREY FRED PUNZO Department of Biology, Box 5F, University of Tampa, Tampa, FL 33606 Asstract: Studies were conducted to investigate the palatability of 7-spot ladybird beetles (Coccinella septapunctata) to adult males of the golden mouse, Ochrotomys nuttali, as well as to determine what cues were used by these mice to avoid contact with these beetles. Ladybird beetles were presented to captive mice either intact (unbled) or after being bled to remove the reflex hemolymph discharged after being bitten (reflex bleeding), and in various treatment groups in which the beetle’s color pattern, taste, or odor were added to a palatable control bark beetle, either alone or in various combinations. Unbled 7-spot ladybird beetles were unpalatable to O. nuttalli, and were rejected by a majority of these mice even after reflex-bleeding. Twenty-eight and 40.7% of the unbled and reflex-bled beetles were rejected based on visual cues (color pattern of the elytra) alone. The overall survivorship of 7-spot beetles after an encounter with a mouse was high for both bled (82.7%) and unbled (79.3%) insects. Only 15.7% of the unbled beetles were eaten following an attack as compared to 5.6% of the reflex-bled beetles. Seventy-five and 84.6% of the bled and unbled beetles, respectively, survived an attack (chewing) indicating that these insects can survive even after some degree of mandibulation by these mice. Although golden mice responded to visual, olfactory, and gustatory cues, either singly or in combination with one another, odor alone resulted in a lower level of rejection. When presented with the color pattern of the beetle alone, there was an initial high rate of rejection by these mice which then decreased significantly after three trials. Key Words: Aposematic prey, avoidance, cues, Ochrotomys nuttalli THE survival capacity of any animal is dependent upon a number of factors including the possession of adaptive mechanisms to avoid toxic substances. This is especially true for those species that feed on a wide variety of foods (broad trophic niche) because many animals and plants have evolved chemical defenses to reduce predation or grazing (Hughes, 1990; Cowan et al., 2000). The presence of toxic compounds in plant and animal tissues places any animal that utilizes indiscriminate selection of food at considerable risk (Rozin and Kalat 1971; Lasker and Metzker 1998). The ability to recognize permanent or seasonal changes in food availability, composition, and potential toxicity based on previous experience (learning) can con- tribute to overall fitness (Punzo, 2003). Thus, it is not surprising that the ability to learn aversions to foods containing unpalatable or toxic substances has been demon- strated in a wide variety of taxa, both in the field and in the laboratory (see reviews by Garcia et al., 1985; Batsell et al., 2001). Some animals depend primarily on olfactory cues for food aversion learning (Roper and Marples, 1997), while others utilize tactile 209 210 FLORIDA SCIENTIST [VOL. 66 (Hughes, 1990), visual (Regan et al., 1996), or gustatory (Galef et al., 1994) cues, or various combinations of these sense modalities (Sloltnick et al., 1995). With respect to mammals, the house mouse (Mus musculus) is known to utilize visual cues associated with a variety of aposematic prey including adult and larval butterflies (Danaus plexippus) and moths (Arctia caja) (Marsh and Rothschild, 1974). A more recent study has shown that deermice (Peromyscus maniculatus) from Arizona learned to avoid the aposematic rainbow grasshopper (Dactylotum variegatum) after several encounters (Neal et al., 1994). The golden mouse (Ochrotomys nuttalli) is an omnivore that is found throughout the southeastern region of the United States (Whitaker, 1996). It occurs in a variety of habitats including greenbriar thickets, boulder-strewn hemlock slopes, hedgerows, and swamps. It is unusual in that it is semi-arboreal and uses its prehensile tail and the tubercles on its feet to climb into bushes and trees where it moves about easily in search of fruits, acorns and other types of seeds, and insects (Gingerich, 1999). Because it is semi-arboreal, it presumably encounters the 7-spot ladybird beetle, Coccinella septapunctata (Coleoptera: Coccinellidae), an aposematic species that is common in the habitats preferred by this rodent. This beetle was introduced into the United States from Europe and now firmly established throughout the eastern re- gion of North America (Milne and Milne, 1980). This brightly colored insect con- tains toxins including the alkaloids coccinelline and precoccinelline (Pasteels et al., 1973) and pyrazines which give the beetle a characteristic taste and odor when its body wall is punctured by a potential predator (Moore et al., 1990). Although these toxins are distributed throughout its body, they are highly concentrated in the reflex hemolymph that is released from the tibia/femur joint of the legs when the insect is attacked (Hollaway et al., 1991). Thus, this aposematic beetle presents a potential predator with a combination of olfactory, gustatory, and visual cues. Pilot studies in the laboratory have indicated that golden mice learn to avoid C. septapunctata after 1-3 encounters (Punzo, unpubl. data). Previous experiments have shown that C. septapunctata is toxic to several species of birds and a variety of avian predators (Marples et al., 1989) as well as some insects (Punzo, 2000) can learn to avoid C. septapunctata after initial encounters. This study was conducted to determine what cues this mouse uses as discriminitive stimuli for food aversion learning. MATERIALS AND MeEtTHODs—Subjects and housing conditions—The 7-spot ladybird beetles (C. septapunctata) used in these experiments were collected from 2 sites in Leon County, Florida, and maintained in an environmental chamber (Precision Model 85A, Boone, Iowa) at 20°C, 75-80% RH, and a 14L:10D photoperiod regime. They were fed on a diet of pea and rose aphids (Macrosiphum spp.) and provided with water ad libitum. I collected reflex hemolymph from 10 beetles (from each collection site) using the procedure described by Punzo (1990) and tested it for uniformity of alkaloid content using gas chromatography as described by Hollaway and co-workers (1991). No significant differences were found in the alkaloid content between the 2 collection sites (Mann-Whitney U-test, W = 97.2, n= 10, 10, P=0.511). During the course of behavioral testing, each 7-spot ladybird beetle was presented to a mouse only once, although some of the beetles were bled 2 or 3 times. In such instances, there was always 8—10 days between bleedings so that the insects could have ample time to replenish their hemolymph and toxins. No. 3 2003] PUNZO—GOLDEN MOUSE CUES 211 TaBLeE 1. Treatment groups showing the defense cues of 7-spot ladybird beetles (Coccinella septapunctata) presented to golden mice (Ochrotomys nuttalli) in the 2 sessions of the experiment. Treatment group Session | Session 2 1 color, taste, smell, toxic color, taste, smell, less toxic 2 taste, smell, non-toxic taste, smell, color, non-toxic 3 smell, non-toxic smell, color, non-toxic 4 color, non-toxic color, smell, non-toxic The golden mice used in these experiments were adult males (total length: 164-170 mm; weight: 79-81 g). All animals were third or fourth-generation offspring from adults originally collected in Leon County, Florida, during 1999-2000. They were maintained individually in plastic rodent cages (35 X 20 X 20 cm) with a substrate of woodchips and sand, and provided with a tube-lick water bottle. The mice were fed on a diet of rolled oats, commercial rat chow (Ralston Purina, St. Louis, Missouri), and palatable adult bark (control) beetles (Scolytus quadrispinosus) comparable in size to a 7-spot ladybird beetle, and common in areas where O. nuttalli is found. The palatable beetles ensured that all mice had prior experience with capturing and eating insects. The room was well ventilated and was maintained at 22— 23°C, 68-75% RH, with a photoperiod regime of 14L:10D. The mice were offered a 7-spot beetle 14 days before the start of testing to standardize the time period between the start of an experiment and the last encounter with a ladybird beetle. Previous pilot studies had indicated that this time period was sufficient for the aversion of these mice toward these beetles to wane (Punzo, unpubl. data) and that they would resample the beetles when experiments began. Because mice in one of the subsequent treatment groups would receive bread dough (see experimental procedures), one day before the start of testing each mouse received a small piece of bread dough (Pillsbury Dough, Canton, Ohio) to familiarize them with the taste, smell, and texture of this food. All mice readily ate this food item. Experimental procedure—The procedure used was modified from that described by Marples and co- workers (1989) in a study investigating differences in toxicity between two species of coccinellid beetles. Mice were divided into 4 treatment groups (n = 15/group). Each group was exposed to treatments in 2 training sessions of 10 days each. The purpose of the first session was to assess the effects of each part of the 7-spot beetle’s defense system. The second session was designed to determine what change (if any) in response occurred when a new defense cue is added. Mice were tested individually in a plexiglass arena (35 X 25 X 20 cm) and were presented with food items placed in a glass petri dish (6.0 cm in diameter) located in the center of the arena. The floor of the arena contained sand as a substrate. Each ladybird beetle was presented to the mice only once. The treatment presented to each group of mice is shown in Table 1. In session 1, group | mice were presented with an entire untreated 7-spot ladybird beetle and were thus exposed to a prey item that had the appearance, taste, and smell of this toxic species. Group 2 mice received a control beetle (S. quadraspinosus) that had been treated with the reflex hemolymph from C. septapunctata. This was ac- complished by diluting the reflex hemolymph of C. septapunctata in twice its volume of water and dipping the control beetles in the mixture. Thus, the non-toxic control beetle had the odor and taste of a ladybird beetle, but not its appearance. Because the ‘“‘taste’’ of a 7-spot beetle is present in its hemolymph (which also carries its odor), it was not possible to present a mouse with the taste of a 7-spot beetle alone. Group 3 mice received an untreated (non-toxic) control beetle that possessed only the odor of a 7-spot beetle. This was achieved by putting a beetle in a glass petri dish (6.0 cm in diameter) with small holes in its bottom and placing a crushed ladybird beetle beneath it (Marsh et al., 1994). The bottom of the petri dish was rendered opaque with white paint so that the mouse could not see the beetle. All treatments were presented in this type of petri dish. Mice in group 4 were exposed to a palatable control beetle that resembled a 7-spot ladybird in color pattern, size, and shape. The 2 elytra of a 7-spot ladybird beetle were attached to the control beetle using bread dough which was shown to be readily eaten by these mice and yet could also be shaped to fill the space between the ladybird elytra and those of the control beetle. To remove any traces of odor or 212 FLORIDA SCIENTIST [VOL. 66 TABLE 2. The numbers of beetles rejected by Ochrotomys nuttalli for 10 trials for 2 training sessions. N = 15 individuals tested per trial. Al (unbled 7-spot ladybird beetles, Coccinella septapunctata, offered for 10 trials; then A2, 7-spot beetles that had been reflex-bled for the remaining trials); B1 (edible control beetles carrying the taste of 7-spot beetles in the form of reflex hemolymph offered for 10 trials; then B2, carrying both the taste and color pattern for the remaining trials); C1 (edible control beetles in the presence of the odor of 7-spot ladybirds offered for 10 trials; then C2, both the odor and color pattern for the remaining trials; D1 (edible control beetles with the color pattern of 7-spot ladybirds offered for 10 trials; then D2, carrying both the color pattern and the odor for the remaining trials. Each beetle was presented to the mice only once. See text for details. Number of beetles rejected treatment/sessions 1 and 2 Trials Al A2 Bl B2 Cl CZ D1 D2 1 11 13 2 12 3 8 11 12 2) i 14 5) Tell 0 5 12 7 3 14 ih yd y) 1 6 10 9 4 IS 14 Je 7. 0 8 4 il 5) 72 12 8 8 2 9 6 4 6 14 13 0 6 0 6 B) 8 a 13 2 0 8 1 5 3) i 8 13 11 0 8 0 6 8) 5 9 15 13 l 7 il 7 4 8 10 14 8. 1 6 0 9 2 6 Total: 134 126 16 82 y) 69 60 V3 Percent: 89.3 84.0 10.7 54.6 6.0 46.0 40.0 48.6 taste, the ladybird elytra were dipped in chloroform, and thoroughly dried before they were attached to the control specimen. Pilot studies had shown that neither the dough or chloroform treatment adversely affected the feeding behavior of the mice. A one-week interval was employed between the end of session 1 trials and the start of session 2. During session 2 trials, group 1 mice (which had received an entire 7-spot ladybird in session 1), now received 7-spot beetles that had been bled to reduce the alkaloid content of their hemolymph. Hemolymph was removed through the use of microcapillary tubes as described by Punzo (1989). In addition, bleeding was done immediately before the start of a trial and was continued until the ladybird beetle could no longer exude reflex hemolymph. As a result, these beetles were less well defended than the unbled ladybirds. Group 2 mice had color added to the taste treatment, and mice in groups 3 and 4 were now exposed to a combination of odor and color (Table 1). During each trial, a bark beetle was presented with each 7-spot ladybird beetle to ensure that a mouse would approach the petri dish at the center of the arena and see the ladybird. A trial was terminated when the mouse walked away from the petri dish for the second time. This decreased the probability that a mouse would handle a ladybird as the result of some displacement activity. We recorded whenever a mouse ate a bark beetle or a ladybird, or if it touched a beetle with its snout (contact) but did not bite it (no damage), or chewed it for the first time (thereby piercing the body wall), and whether or not it rubbed its nose in the sand. Beetles that were uneaten but had been contacted or chewed were kept in a petri dish for 3 days to determine mortality resulting from encounters with mice. All statistical analyses (Mann-Whitney U-test; Chi Square, X*; Spearman rank correlations) fol- lowed procedures described by Sokal and Rohlf (1995). ResuLts—Unbled and reflex-bled ladybird beetles—Golden mice avoided 7- spot ladybird beetles in 82.6% of trials with unbled beetles and in 77.3% of trials with reflex-bled individuals, and continued to do so throughout both training sessions (Table 2, Al and A2). The 7-spot beetles continued to exhibit a chemical No. 3 2003] PUNZO—GOLDEN MOUSE CUES Le, TABLE 3. Survival data for unbled and reflex-bled 7-spot ladybird beetles (Coccinnella septapunctata) when presented to male golden mice, Ochrotomy nuttalli. Numbers in parentheses represent percentages. See text for details. Mouse behavior Unbled beetles (n = 150) Reflex-bled beetles (n = 150) No contact with beetle 42 (28.0) 61 (40.7) Attacked beetle 108 (72.0) 89 (59.3) Number rejected after attack 91 (84.3) 84 (94.4) Number of beetles surviving 77 (84.6) @3)@5,0) Number of beetles dying 14 (15.4) Din 5.0) Number eaten after attack WE (Sz) 5 (5.6) defense (rejected by mice) even after reflex-bleeding. In fact, survival data (Table 3) indicated that 25% of bled beetles attacked by mice were killed as compared to only 15.4% for unbled beetles OS = 6.7, df=1, P < 0.05). In addition, the mice avoided contact altogether with 28 and 40.7% of the unbled and bled beetles, respectively. Overall survivorship after an encounter with a mouse was high for both the bled (82.7%, n = 150) and unbled (79.3%, n = 150) ladybird beetles. Of 108 unbled beetles that were attacked, only 15.7% were eaten; and only 5.6% of the reflex-bled beetles were eaten (X* = 6.2, P < 0.05) (Table 3). Seventy- five and 84.6% of bled and unbled beetles, respectively, survived an attack involv- ing chewing, indicating that these insects can survive some degree of damage by mice. Odor, taste, and color treatments—Odor on its own resulted in a low level of rejection by mice (Table 2, C1); however, odor and color pattern combined (Table 2, C2) resulted in a significant increase in rejection rate (Mann-Whitney U-test, W = 149.6, n= 10,10, P = 0.01). There was no significant difference in rejection rate when odor and taste treatments were tested separately (Table 2B and C, Mann- Whitney U-test, W = 162.1, n= 10,10, P =0.20). When the color pattern of the 7- spot beetle was added to the taste (Table 2, B2) and odor (Table 2C) treatments, the level of rejection was significantly increased (X* = 81.2, df= 1, P < 0.01). The mice initially avoided taste and color cues at the same level that they avoided intact 7-spot beetles (Fishers exact test with Yate’s correction, X* = 0, df = 1, P= 1.00). In addition, the mice rubbed their noses in the sand (84.6%) to a greater extent in the presence of taste cues than in their absence (62.3%) (X? = 75.6, P < OOL): When mice were presented only with the color pattern of 7-spot beetles (Table 2, D1), there was an initial high rate of rejection that decreased significantly after the third trial (Spearman rank correlation: r, = —0.797, P < 0.01). Discussion—These experiments clearly demonstrate the unpalatability of the 7-spot ladybird beetle to golden mice. To our knowledge, this is the first analysis of cues used to assess aposematic prey in this rodent. Males of O. nuttalli respond to a combination of stimuli associated with C. septapunctata including gustatory, olfactory, and visual cues. Because the rejection rate was low when mice were presented with the taste of these beetles alone, gustatory cues may be less important than visual and odor cues for food aversion learning in this rodent. 214 FLORIDA SCIENTIST [VOL. 66 These results are in general agreement with several studies on food aversion learning in birds. When unpalatable mealworms (T. molitor), that had been painted red, were offered to birds, Miihlmann (1934) found that although gustatory and visual cues alone would result in some rejection, avoidance rates were significantly higher when both cues were presented together, and higher still when intact 7-spot ladybird beetles were offered. Almost 41% of these mice avoided contact with reflex-bled beetles (Table 3), indicating that even after bleeding these beetles still retain some noxious properties. Previous studies have shown that 7-spot beetles need at least one week to replensh their supply of reflex hemolymph after bleeding (Hollaway et al., 1991). This sug- gests that sufficient toxin is retained in the insect’s body to maintain some degree of protection even when the animal cannot release reflex hemolymph when attacked. It has been reported that the odor associated with 7-spot beetles is to some degree associated with the presence of pyrazine compounds (Moore et al., 1990). Guilford and co-workers (1987) have suggested that pyrazines function as ‘alerting odors’ that facilitate association/avoidance learning in animals. Although the odor of beetles alone did not result in a strong avoidance response by these mice, its combination with color and taste when these beetles are attacked under natural conditions may very well strengthen a mouse’s response to color pattern and taste. In this case, odor may represent another category of conditioned stimulus (or un- conditioned stimulus, if the capacity is innate) for the type of learning known as classical conditioning (Davey, 1989). It has also been shown that some animals pos- sess an innate capacity to respond to specific chemosensory cues associated with either prey or predator species (Kats and Dill, 1998; Punzo, 2002a,b, 2003). In conclusion, the 7-spot ladybird beetle has a multifaceted defense system which allows O. nuttalli and other potential predators to avoid the insect at various stages of the hunting sequence. Some responses of the predators to chemosensory or visual cues possessed by this beetle may be innate while others are the result of classical conditioning and/or association (avoidance) learning. The survivorship of a predator may be higher if it can learn to avoid aposematic prey using visual cues alone because it will not have to sample (i.e.-chew) the prey and thereby expose itself to even small levels of toxic substances. Ochrotomys nuttalli clearly avoids C. septapenctata based solely on visual cues, although the rejection rate is higher when a suite of cues are present. ACKNOWLEDGMENTS—I would like to thank O. Bettis, J. Layne, and D. F. Martin for commenting on an earlier draft of the manuscript, L. Hane for procuring some of the reference literature, and B. Garman for consultation on statistical analyses. A Faculty Development Grant from the University of Tampa made much of this work possible. LITERATURE CITED BATSELL, W. R., J. D. BATESON, G. Y. PASCHALL, AND D. L. GLEASON. 2001. Taste preconditioning augments odor-aversion learning. J. Exper. Psychol. (Anim. Behav. Proc.) 27: 30-47. Cowan, D. P., J. C. REYNOLDS, AND E. L. HILL. 2000. Reducing predation through conditioned taste aversion. Pp. 281-299. In: GosLinc, L. M. and W. J. SUTHERALAND (eds.), Behavior and Conservation. Cambridge Univ. Press, New York, NY. No. 3 2003] PUNZO—GOLDEN MOUSE CUES PA'S Davey, K. L. 1989. Ecological Learning Theory. Metheun, London. Ga.er, B. G., M. MAINARID, AND P. VALsEccul. 1994. Behavioral Aspects of Feeding: Basic and Applied Research in Mammals. Harwood Acad. Publ., Singapore. GarciA, J., P. S. LAsiTeR, F. BERMUDEZ-RATTONI, AND D. A. DEEMS. 1985. A general theory of aversion learning. Pp. 8-21. In: BRAVEMAN, N. S. AND P. BRONSTEIN (eds.), Experimental Assessments and Clinical Applications of Conditioned Food Aversions. Annals of the New York Academy of Sciences, Vol. 443, New York Acad. Sci., New York, NY. GINGERICH, J. L. 1999. Florida’s Fabulous Mammals. World Publs., Tampa, FL. GuILForD, T., C. Nicot, M. ROTHSCHILD, AND B. P. Moore. 1987. The biological roles of pyrazines: evidence for a warning odour function. Biol. J. Linn. Soc. 31: 113-128. Ho.iaway, G. J., P. W. DE JoNG, P. M. BRAKEFIELD, AND H. DE Vos. 1991. Chemical defence in ladybird beetles (Coccinellidae). I. Distribution of coccinelline and individual variation in defence in 7-spot ladybirds (Coccinella septapunctata). Chemoecol. 2: 7—14. HuaGues, R. N. 1990. Behavioral Mechanisms of Food Selection. Springer, Berlin, Germany. Kats, L. B. AND L. M. DILL. 1998. The scent of death: chemosensory assessment of predation risk by prey animals. Ecoscience 5: 361-394. LasKER, M. AND K. METZKER. 1998. Food avoidance learning in squirrel monkeys and common marmosets. Learn. Memory 5: 193-203. Marpies, N. M., P. M. BRAKEFIELD, AND R. J. Cowie. 1989. Differences between the 7-spot and 2-spot labdybird beetles (Coccinellidae) in their toxic effect on a bird predator. Ecol. Entomol. 14: 79-84. Marsu, N. AND M. ROTHSCHILD. 1974. Aposematic and cryptic Lepidoptera tested on the mouse. J. Zool. 174: 89-122. , W. VAN VEELAN, AND P. M. BRAKEFIELD. 1994. The relative importance of colour, taste, and smell in the protection of an aposematic insect Coccinella septapunctata. Anim. Behav. 48: 967-974. Miine, L. AND M. MILne. 1980. Field Guide to North American Insects and Spiders. Alfred A. Knopf, New York, NY. Moore, B. P., W. V. BROWN, AND M. ROTHSCHILD. 1990. Methylalkylpyrazines in aposematic insects, their host plants, and mimics. Chemoecol. 1: 43-51. MUHLMANN, H. 1934. Im modellversuch kiinstlich erzeugte mimikry und ihre bedeutung fiir den ‘nachahmer’. Z. Morph. Okol. 28: 259-296. NEAL, P. R., M. STROMBERG, AND K. A. JEPSON-INNES. 1994. Aposematic coloration in Dactylotum variegatum (Orthoptera: Acrididae): support from vertebrate feeding trials. Southwest. Nat. 39: 21-25. PASTEELS, J. M., C. DERoE, B. Turscu, J. C. BREKMAN, D. DALOZE, AND C. HooreLe. 1973. Distribution et activités des alcaloides defensifs des Coccinellidae. J. Insect Physiol. 19: 1771-1784. Punzo, F. 1989. Composition of the hemolymph of mygalomorph spiders. Comp. Biochem. Physiol. 93A: 757-760. . 1990. The hemolymph composition and neurochemistry of the spider wasp, Pepsis formosa (Say) (Hymenoptera: Pompilidae). Comp. Biochem. Physiol. 96A: 341-345. . 2000. Desert Arthropods: Life History Variations. Springer, Heidelberg, Germany. . 2002a. The effects of early experience on subsequent feeding responses in the tegu, Tupinambis teguixin (Squamata: Teiidae). J. Environ. Biol. (in press). . 2002b. Evidence for the use of chemosensory cues by the alligator snapping turtle, Macroclemy temminckii, to detect the presence of musk and mud turtles. Florida Scient. 65: 134-138. . 2003. Chemosensory responses to odors in the masked shrew (Sorex cinereus). Northwest. Nat. (in press). REGAN, B. C., F. Vienot, P. C. CHARLES-DOMINIQUE, S. PFEFFERKORN, B. SIMMEN, C. JULIOT, AND J. D. Mo toon. 1996. The color signals that fruits present to primates. Invest. Ophthal. Vis. Sci. 37: 2997-3001. Roper, T. J. AND N. M. Marptes. 1997. Odor and color as cues for taste-aversion learning in domestic chicks. Anim. Behav. 53: 1241-1250. 216 FLORIDA SCIENTIST [VOL. 66 Rozin, P. AND J. W. KALAT. 1971. Specific hungers and poison avoidance as adaptive specializations of learning. Psychol. Rev. 78: 459-486. SLOTNICK, B. M., R. F. WESTBROOK, AND F. M. DarRLinc. 1995. Odor and taste aversion learning. Adv. Biosci. 93: 465-472. SOKAL, R. R. AND F. J. ROHLF. 1995. Biometry. 3rd ed. W. H. Freeman, San Francisco. WHITAKER, J. O., JR. 1996. A Field Guide to the Mammals of North America. Alfred A. Knopf, New York, NY. Florida Scient. 66(3): 209-216. 2003 Accepted: December 27, 2002 Medical Sciences A COMPARISON OF DIAGNOSTIC METHODS USED TO DETECT INTESTINAL PARASITES IN RESIDENTS OF THE PROVINCE OF CHALATENANGO, EL SALVADOR Micuact A. SILVERMAN”), RAYMOND K. ZLAMAL™,, CaRLos Cruz”, Patricia Dipatrizio™, CAROL J. PALMER“? (4) , DONALD E. Burris*”’, Scotr ScHatz™, AND HAROLD E. Lausacu?* ‘Martinsburg Veterans Affairs Medical Center, Martinsburg, WV 25401, ‘324 Combat Support Hospital, Perrine, FL 33101 and Nova Southeastern University, Colleges of © Allied Health, Medical Sciences, Optometry and Oceanography, Ft. Lauderdale, FL 33328 ABSTRACT: Fecal specimens from 50 patients in the Province of Chalatenango, El Salvador were examined for parasites using zinc sulfate flotation of fresh feces and by direct smears of feces collected in Parasafe, Proto-Fix or Unifix fixatives following trichrome or acid-fast staining. The overall prevalence of parasitic infections in the patients was 64% (32/50). Two parasites, Entamoeba histolytica/dispar and Ascaris lumbricoides were found in large numbers, 20% (10/50) and 36% (18/50), respectively, and were used for the comparative studies. Trichrome staining of direct smears was most effective in identifying E. histolytica/dispar in specimens of all three fixatives and acid-fast staining did not reveal any amoeba. Ascaris lumbricoides eggs were consistently found using all three fixatives and with both staining methods. Zinc sulfate flotation of fresh feces was most effective for identification of A. lumbricoides eggs and trichrome staining of feces collected in Parasafe, Proto-fix or Unifix fixatives for the detection of E. histolytica/dispar. Key Words: Parasitic infections, parasite identification, Ascaris, Entamoeba, El Salvador THE prevalence of intestinal parasitic helminths and protozoa in the province of Chalatenango, El Salvador has never been reported on, mainly because of the lack of availability of laboratory facilities in this rural area. Previous studies of parasitic infections in El Salvador, however, have documented infections with Entamoeba histolytica/dispar (Silverman et al., 1998; Spencer et al., 1981; Reinthaler et al., 1988a), Ascaris lumbricoides (Spencer et al., 1981; Reinthaler et al., 1988a), Trichuris trichiura (Spencer et al., 1981; Reinthaler et al., 1988a), Giardia duodenalis (Reinthaler et al., 1988a), Cryptosporidium parvum (Reinthaler et al., 1988a; Reinthaler et al., 1988b), Hymenolepis nana (Spencer et al., 1981; Reinthaler et al., 1988a) and hookworms (Spencer et al., 1981; Bloch and Rivera, 1977). These studies used fecal concentration methods along with direct smears to sample for parasites. * Corresponding author: harold@nova.edu WN} 218 FLORIDA SCIENTIST [VOL. 66 The apparent difference between non-concentration and concentration methods of parasite identification in fecal samples was the basis for the present study. The objective was to assess the prevalence of parasitic infections in this human population of the Province of Chalatenango in El Salvador, using different laboratory recovery methods for parasites in stool samples. It was postulated that the use of newer identification techniques would increase the accuracy of the detection of parasitic infections. Data demonstrate that the use of different methods of preservation of fecal material enhances the finding of parasites following staining and examination with light microscopy. MATERIALS AND METHODS—During the month of March 2000, the 324" Combat Support Hospital of Miami, Florida performed a 10-day Medical Readiness Training Exercise in the Chalatenango Province of El Salvador, Central America. The research was performed in the three villages of San Rafael, Dulce Nombre de Maria and El Paraiso. These villages were selected because the residents had poor living conditions and lacked available health care. The people were presumably exposed to intestinal parasites due to their close contact with domestic animals, consumption of contaminated water and poor sanitary practices due to the lack of hand washing facilities. The patients in this study were relatively healthy when they were given physical examinations but had some obvious problems of malnutrition, weight loss, arthritis and hypertension. Fecal samples were collected from 50 patients randomly selected from several hundred residents with varying symptoms such as abdominal problems, skin and joint disorders, respiratory diseases and eye infections, and stored in 50-ml screw-cap plastic tubes. The age, gender and socioeconomic status of the patients were not recorded. Fresh feces were examined using zinc sulfate flotation (Ash and Orihel, 1991). Briefly, approximately 0.5 ml of fresh feces was mixed with 20 ml of saturated zinc sulfate solution, specific gravity of 1.20, and poured through gauze in order to fill a 10-ml vial. A cover slip was placed on top of each vial and the mixture was allowed to set for 15 minutes. The cover slips, containing adhered fecal debris, were removed, placed on glass slides and observed for parasites. Five-gram aliquots of feces from each of the samples were preserved in three different fixatives, Parasafe (Scientific Device Laboratory, Inc., Des Plaines, IL), Proto-Fix (Alexon-Trend, Ramsey, MN), and Unifix (Medical Chemical Corporation, Santa Monica, CA) and transported to the United States where they were examined. Following fixation, three thin fecal smears were made from each of the three different fixed specimens. Smears were made on glass slides and allowed to dry for 24 hours at 37 degrees C. One of each of the three fixed fecal smears was examined, unstained, for parasites, a second following trichrome staining, and a third following acid-fast staining (Clark, 1981). All preparations were examined for protozoan and helminth parasites using 100 and 400x magnification with a bright field microscope. An ocular micrometer was used for measurements of protozoan cysts, trophozoites and helminth eggs. The presence of parasitic forms in each preparation was recorded and differences between methods (positive or negative for a parasite) were analyzed using McNemar’s test. Values of P < 0.05 were considered to be significant. RESULTS—Fecal samples examined for helminth eggs and protozoans were found to contain eggs of A. lumbricoides and Trichuris trichiura and trophozoites and cysts of E. histolytica/dispar and Giardia lamblia (Table 1). Because the fecal specimens contained large numbers of A. Jumbricoides and E. histolytica/dispar with low numbers of G. lamblia and T. trichiura, the use of different fixatives and staining methods were compared using the specimens containing the helminth, A. lumbricoides and the protozoan, E. histolytica/dispar. Of the 50 stool samples examined, 20% (10/50) were positive for E. histolytica/ dispar using smears of the three transport media: Unifix, Parasafe, Protofix and 2.0% No. 3 2003] SILVERMAN ET AL.—DIAGNOSIS OF INTESTINAL PARASITES 219 TABLE |. Parasites found in 50 stool specimens of residents of the Province of Chalatenango, El Salvador. Number of positive fecal samples/ Parasites % found using all methods* Ascaris lumbricoides 18 (36)** Entamoeba histolytica/dispar 10 (20)** Giardia lamblia 2 (4) Trichuris trichiura 2 (4) * Samples positive for parasites were positive on any of the tests performed and the possibility of false positive findings was considered to be zero. ** A_ lumbricoides and E. histolytica/dispar were found in significantly greater numbers than G. lamblia or T. trichiura. (1/50) using zinc sulfate flotation (Table 2). Cyst stages were mainly found using the transport media. Apparent differences were not noted between direct smears of Parasafe and Protofix fixatives but Unifix slides had significantly higher numbers of positive specimens. The positive samples found using Parasafe and Protofix were not the same samples but were a part of the six found using Unifix. Trichrome staining of fixed specimens demonstrated more amoebae using Unifix and Parasafe fixatives than with the use of Protofix. Acid-fast staining of all three fixative preparations did not reveal any amoeba but zinc sulfate flotation of fresh feces had one specimen containing cysts of E. histolytica/dispar. The greatest number of specimens containing A. lumbricoides, 30% (15/50) was found when zinc sulfate flotations of fresh fecal specimens were examined (Table 2). None of the fixed specimens individually had as many positive findings as the zinc sulfate flotation method. A comparison of direct smears to trichrome stained specimens using Unifix was significantly higher when compared to the other two fixatives. Trichrome staining revealed only a few specimens positive for A. lumbricoides demonstrating a non-significant difference between the different fixatives. Acid fast staining produced similar results to those of direct smears. Discussion—In this survey of intestinal parasites, stool specimens obtained from people living in this rural area of El Salvador confirmed a high prevalence of E. histolytica/dispar in the region. The overall prevalence rate of 20% was greater than the previous finding of 10% using stool samples that were fixed with polyvinyl alcohol or by formalin-preservation (Silverman et al., 1998). The higher prevalence of parasitic infections detected in the present study could have been due to the use of multiple fixatives and stains that resulted in the preparation of several samples and slides of each specimen. Infection with A. lumbricoides was also found to not only be endemic but was present in a large percentage of the population of this region of El Salvador. Previous findings using formalin as a fixative showed 10% (Silverman et al., 1998) and 18% (Reinthaler et al., 1988a) prevalence rates for A. lumbricoides that were significantly less than the finding of 36% in the present study. It is suggested that the different results could be due to differences in the laboratory methods used in each of the two studies. 220 | FLORIDA SCIENTIST [VOL. 66 TABLE 2. Identification of Entamoeba histolytica/dispar and Ascaris lumbricoides using different fixatives. Specimen Number of positive fecal samples/% found using different fixatives preparation Unifix Parasafe Proto-fix E. histolytica/dispar Direct smear Gra) = r@) 1 (2) Trichrome stain 8 (16)* 4 (8) i@) Acid-fast stain 0 (0) 0 (0) 0 (0) A. lumbricoides Direct smear 7 (14)** 3 (6) 5 (10) Trichrome stain 2 (4) 1L@) 1 (2) Acid-fast stain 5 (10) 3 (6) 3 (6) * A significantly higher percentage of parasites was found in Unifix specimens than in Parasafe or Proto-fix specimens. ** A significantly higher percentage of parasites was found in Unifix specimens than in Parasafe specimens. There was a considerable difference in the numbers of positive specimens for both A. lumbricoides and E. histolytica/dispar using direct smears of the three different fixatives and zinc sulfate flotation of fresh feces. The 83% (15/18) positive A. lumbricoides specimens found with zinc sulfate flotation (a concentration method) was considerably higher than comparative results using Unifix, Parasafe or Protofix fixatives (non-concentration methods). Both direct preparations and trichrome staining of Unifix preserved feces resulted in an increase in the identification of E. histolytica/dispar infections. Results suggest that performing trichrome staining on Unifix fixed smears was the most effective method for identification of amoebae in fecal samples. From these data, the authors recommend that Unifix along with other methods should be tested before deciding on one product to use. Possibly, using a combination of several methods is best. The existence of parasitic infections in people living in rural El Salvador and in other parts of the world will be better understood when studies are performed comparing different survey methods. Data presented here suggest that the num- bers of reported cases of parasitic infections are under-estimated and further epidemiological studies are needed to determine the true prevalence of parasites in this region of El Salvador. Moreover, studies comparing infections in urban vs. rural inhabitants would also yield valuable information that could assist the country’s ministry of health in effectively targeting appropriate public health measures aimed at reducing intestinal parasite burdens in the population. ACKNOWLEDGMENTS—Support for this research was provided by a grant from the Nova Southeastern University, Health Professions Division. We thank the healthcare professionals in the El Salvadoran military for their valuable assistance with this study. LITERATURE CITED Asi, L .R. AND T. C. ORIHEL. 1991. Parasites: a Guide to Laboratory Procedures and Identification, ASCP Press, Chicago, IL. No. 3 2003] SILVERMAN ET AL.—DIAGNOSIS OF INTESTINAL PARASITES 221 CiarK, G. 1981. Staining procedures, Williams and Wilkins, Baltimore, MD. Biocnu, M. AND G. H. Rivera. 1977. Hookworm disease. Epidemiology. Diagnostic test. Rev. Int. Invest. Med. 6: 171. REINTHALER, F. F., G. LINCK, G. KLEM, F. MASCHER, AND W. SIXL. 1988a. Intestinal parasites in children with diarrhea in El Salvador. Geo. Med. 18: 175-180. : 3 : , AND . 1988b. Cryptosporidiosis in children with diarrhoea from slum areas in San Salvador. Ann. Trop. Med. Parasitol. 82: 209-210. SILVERMAN, M. A., D. BARNES, R. ZLAMAL, S. SOUTHWORTH, R. MCKINNEY, H. LAuBAcH, D. DILEO, AND M. Bercer. 1998. Medical readiness training exercise in El Salvador, Central America, 1996. Mil. Med. 163: 519-523. SPENCER, H. C, J. J. SULLIVAN, H. M. MATHEWS, M. SAUERBREY, M. BLOcH, W. CHIN, AND G. R. HEALY. 1981. Serologic and parasitologic studies of Entamoeba histolytica in El Salvador, 1974-1978. Am. J. Trop. Med. Hyg. 30: 63-68. Florida Scient. 66(3): 217—221. 2003 Accepted: December 31, 2002 Biological Sciences THE MADAGASCAR GIANT DAY GECKO, PHELSUMA MADAGASCARIENSIS GRANDIS GRAY 1870 (SAURIA: GEKKONIDAE): A NEW ESTABLISHED SPECIES IN FLORIDA KENNETH L. Krysko‘’, A. NICHOLE Hooper™, AND COLEMAN M. Sueeny III? ‘Florida Museum of Natural History, Division of Herpetology, P.O. Box 117800, University of Florida, Gainesville, FL 32611 College of Veterinary Medicine, Campus Box 100125, University of Florida, Gainesville, FL 32610 ABSTRACT: During recent surveys between March and August 2002, we found established pop- ulations of the Madagascar giant day gecko (Phelsuma madagascariensis grandis) in the Florida Keys, Monroe County. We recorded 29 individuals on Little Torch Key and three individuals on Grassy Key. Additional records were obtained from Grassy Key, Big Pine Key, and Plantation Key. Both genders and all size classes were recorded on each island illustrating that this species is presently established as an element of Florida’s introduced herpetofauna. Population monitoring, documentation of ecological impacts on Florida’s native flora and fauna, and/or eradication efforts should be conducted. Key Words: Phelsuma madagascariensis grandis, Madagascar, Gecko, Lizard, Introduced, Reptile, Florida Keys FLorIpDA has the largest number of non-native amphibian and reptile species in the United States (Butterfield et al., 1997). Its diverse habitats and suitable climates from the subtropical southern peninsula and Florida Keys north to the subtemperate panhandle have facilitated exotics in becoming established and expanding their ranges. While conducting surveys in the southern peninsula and Florida Keys over the last decade, we have uncovered numerous geographic distributional records, misidentified species, and new exotic species with established populations (Krysko and Decker, 1996; Reppas et al., 1999; Krysko et al., 2000; Krysko and King, 2002; Townsend et al., 2002; Krysko et al., 2003). Herein, we report four established populations of the Madagascar giant day gecko, Phelsuma madagascariensis grandis Gray 1870, in the Florida Keys, Monroe County. MetTHops—Records of Phelsuma madagascariensis grandis are based on captures and observations during six survey days in the Florida Keys in 2002: 3 March; 3, 5, 7-8 May; and 8 August. Additional records were acquired based on observations by colleagues. Because this gecko species is diurnal (Henkel and Schmidt, 2000), searches were conducted during the daytime. Dorsal patterns are unique for each individual, and an attempt was made to photograph, estimate total length (TL), and note location of each individual for identification purposes. Only individuals that could be distinguished from others were counted in our overall total. We divided size classes into six categories based on estimated TL, including <8 cm (hatchling), 8-11 cm, 12-16 cm, 17-20 cm, and >20 cm (adult). Gender was distinguished in 222 No. 3 2003] KRYSKO ET AL.—_NEW GECKO IN FLORIDA jaja TABLE 1. Estimated size classes and UF voucher specimens and photographs of Madagascar giant day geckos (Phelsuma madagascariensis grandis) we recorded in 2002 on Little Torch Key and Grassy Key, Monroe County, Florida. Note that adults are >20 cm TL, and individuals with no UF # were neither photographed nor collected. Date N <8cm 8-ll cm 12-16cm 17-20 cm Adult 9 Adult 3 Little Torch Key 3 Mar 9 130736-37 No# 131554, IB O735;MBia5S, No # 131555, 133942 5 May 8 132483, No# 132486, 132484, 132487 132485 13248889 7 May 1 No # 8 May 6 133938 132492 132493, 132490 132491 No # 8 Aug Dees 583233939 133940 133943 133941 Grassy Key 3 May 3 132482, 132481 No # adults by the noticeable presence or absence of endolymphatic chalk sacs, which are used for calcium metabolism and egg shell formation in females (Tytle, 1992). Captures were made by hand, and voucher specimens and photographs were deposited in the Florida Museum of Natural History (FLMNH), University of Florida (UF collection). RESULTS—We recorded 32 Phelsuma madagascariensis grandis on Little Torch Key and Grassy Key. Numerous individuals were also recorded on Grassy Key (McCleary, 2002), Big Pine Key (Decker, 2002; pers. obs.) and Plantation Key (Kavney, 2002). Between 3 March—8 August 2002, we recorded 29 P. m. grandis consisting of both genders and all age classes on Little Torch Key (24°39.944'N, 81°23.411'W) (Table 1). Individuals were frequently seen on white mangrove (Laguncularia racemosa) and buttonwood (Conocarpus erectus) trees, buildings, and bird cages bordering mangrove-lined estuaries. On 3 May 2002, we recorded two adult females and one adult male on the confluence of a gumbo limbo tree (Bursera simaruba) and telephone post in the parking lot of the Dolphin Marine Research Center on Grassy Key (24°46.260'N, 80°56.433’W) (Table 1). Additionally, on 14 June 2002 four juveniles (6 cm, 7.5 cm, two 9 cm TL) were observed in the shade of bushes and trees on these premises (McCleary, 2002). These individuals were not collected. In August 2001, three individuals were observed on buildings and wooden fences near Cunningham Lane just N of U.S. 1 on Big Pine Key (24°40.250’N, 81°21.336’W) (Decker, 2002). We verified one of these individuals as an adult female. The other two individuals were not collected. Another adult was observed basking on a utility pole at the junction of U.S. 1 and Wilder Road (Kavney, 2002). We have also observed numerous individuals being sold in a local pet store that were reportedly collected locally. 224 FLORIDA SCIENTIST [VOL. 66 In September 2002, eight newborns to adults of both genders were collected and numerous others were observed near the Indian Mound on Plantation Key (24°59.270'N, 80°33.021’W) (Kavney, 2002). DiscussionN—Phelsuma madagascariensis grandis was first reported from Hollywood, Broward County (Bartlett and Bartlett, 1999). These geckos were released or had escaped from a nearby reptile importer in the early 1990s, but never represented an established population. No voucher specimens were ever taken and no individuals were seen after numerous surveys of the area (Decker, 2002). We know of other intentional releases of this species in Miami-Dade County, but these individuals are not known to be reproducing. The closest population we identified (Grassy Key) exists ca. 240 km SW of the Broward County report, and our data provide evidence of the first verified established populations of P. m. grandis in the United States. The population of Phelsuma madagascariensis grandis on Little Torch Key appears to have originated from a single introduction. This species was likely introduced onto this island by an exotic animal hobbyist as geckos were frequently seen on buildings and exotic bird cages. Phelsuma m. grandis on Big Pine Key appears to have originated from a single source. A resident of the island released this species in certain areas for subsequent harvesting (Decker, 2002) as their offspring are collected and sold for resale at a local pet store. Phelsuma m. grandis on Grassy Key and Plantation Key were introduced onto both islands independently by different local residents. Phelsuma madagascariensis grandis feeds primarily on nectar and arthropods (Demeter, 1976; Tytle, 1992), but congeners have also been documented feeding on Hemidactylus geckos (Garcia and Vences, 2002). The tropical house gecko (Hemidactylus mabouia) was first reported introduced on Crawl Key (Lawson et al., 1991), and presently this species is probably the most abundant terrestrial vertebrate in the Florida Keys and may prove to be a food source for P. m. grandis. On 7 May 2002, an adult P. m. grandis was observed feeding on insects up until ca. 30 min after dark, upon which the gecko retreated into a crack on a wooden building on Little Torch Key. Some individuals on Little Torch Key were seen in the same vicinity on every survey day, suggesting that this species might be territorial like other Phelsuma species (McKeown, 1993). Phelsuma m. grandis have been known to live for >20 years in captivity (McKeown, 1993). Tytle (1992) reported a captive female ovipositing 27 eggs in one year, and during a nine-year span one captive female produced 68 clutches consisting of 120 eggs (Krysko, pers. obs.). This species is a non-gluer (1.e., oviposited eggs are not affixed to a substrate) (Osadnik, 1984), and its eggs are usually oviposited in pairs (Demeter, 1976; Osadnik, 1984; Tytle, 1992) within crevices, between strong leaves, or in the ground (Osadnik, 1984). Longevity, high fecundity, and abundance of prey are factors likely to facilitate population increases and range expansion of P. m. grandis in the Florida Keys. Population monitoring, documentation of ecological impacts on Florida’s native flora and fauna, and/or eradication efforts should be conducted. No. 3 2003] KRYSKO ET AL.—NEW GECKO IN FLORIDA 225 ACKNOWLEDGMENTS—We would like to thank Sean W. Morey (FLMNH) for field assistance; Frank and Susan Morey for accommodations in the lower Florida Keys; John N. Decker, Jim Kavney, Sean O’Donnell, Ryan J. R. McCleary, and Bob Ehrig for Phelsuma information in the Florida Keys. LITERATURE CITED BarTLeTT, R. D. AND P. P. Bartietr. 1999. A Field Guide to Florida Reptiles and Amphibians. Gulf Publishing Co., Houston, TX. 280 pp. BUTTERFIELD, B. P., W. E. MESHAKA, JR., AND C. Guyer. 1997. Nonindigenous amphibians and reptiles. Pp. 123-138. Jn: StmBERLoFF, D., D. C. ScHmiTz, AND T. C. BRown (eds.), Strangers in Paradise. Impact and Management of Nonindigenous Species in Florida. Island Press, Covelo, CA. Decker, J. N. 2002. 2503 NW 23 St #156, Boynton Beach, FL. Pers. Comm. DeEMETER, B. J. 1976. Observations on the care, breeding and behaviour of the giant day gecko Phelsuma madagascariensis at the National Zoological Park, Washington. International Zoo Yearbook 16: 130-133. Garcia, G. AND M. VENcEsS. 2002. Phelsuma madagascariensis kochi (Madagascar day gecko). Diet. Herpetol. Rev. 33: 53-54. HENKEL, F. W. AND W. Scumipt. 2000. Amphibians and Reptiles of Madagascar and the Mascarine, Seychelles, and Comoro Islands. Krieger Publishing Co., Malabar, FL. 316 pp. Kavwney, J. 2002. P.O. Box 1597, Islamorada, FL. Pers. Comm. Krysko, K. L. AND J. N. Decker. 1996. Tantilla oolitica (Rim Rock Crowned Snake). Geographic distribution. Herpetol. Rev. 27: 215. : , AND A. T. Reppas. 2000. Geographic distribution of Ramphotyphlops braminus (Brahminy blind snake). Herpetol. Rev. 31: 256. — AND F. W. Kine. 2002. The ocellated gecko (Sphaerodactylus argus argus) in the Florida Keys: An apparent case of an extirpated non-native species. Caribbean Journal of Science 38(1—2): 139- 140. . , K. M. ENGE, AND A. T. Reppas. 2003. Distribution of the Introduced Black Spiny-tailed Iguana (Ctenosaura similis) on the Southwestern Coast of Florida. Florida Scient. 66(3): 74-79. Lawson, R., P. G. FRANK, AND D. L. Martin. 1991. A gecko new to the United States herpetofauna, with notes on geckos of the Florida Keys. Herpetol. Rev. 22: 11-12. Mccieary, R. J. R. 2002. Department of Zoology, University of Florida, Gainesville, FL. Pers. Comm. Mcxeown, S. 1993. The General Care and Maintenance of Day Geckos. Advanced Vivarium Systems, Lakeside, CA. 143 pp. OsaDNIK, G. 1984. An investigation of egg laying in Phelsuma (Reptilia: Sauria: Gekkonidae). Amphibia- Reptilia 5(1984): 125-134. Reppas, A. T., K. L. Krysko, C. L. SONBERG, AND R. H. Rosins. 1999. Geographic distribution of E. Anolis distichus (Bark Anole). Herpetol. Rev. 30: 51. TOWNSEND, J. H., K. L. Krysko, A. T. Reppas, AND C. M. SHEEHY III. 2002. Noteworthy records of introduced reptiles and amphibians from Florida, USA. Herpetol. Rev. 33: 75. TytLe, T. 1992. Day geckos: Phelsuma The captive maintenance and propagation of day geckos. Vivarium 2: 15-19, 29. Florida Scient. 66(3): 222-225. 2003 Accepted: December 31, 2002 Biological Sciences PATHOLOGIC FINDINGS IN STRANDED ATLANTIC BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) FROM THE INDIAN RIVER LAGOON, FLORIDA (RENE MEISNER“), RENE VARELA‘?, MARILYN Mazzon™, oh ROBIN Fripay’”), (2) GREGORY D. BOSSART STEPHEN D. McCutLocu"'”, Davip KILPATRICK ELIZABETH Murpocu™, BiaiR Mase? ) AND R. H. DEFRAN “Division of Marine Mammal Research and Conservation, Harbor Branch Oceanographic Institution, 5600 U.S. 1 North, Ft. Pierce, FL 34946 Division of Dolphin Research, Harbor Branch Oceanographic Institution, 5600 U.S. 1 North, Ft. Pierce, FL 34946 Southeast Florida Science Center, National Marine Fisheries Service, 75 Virginia Beach Drive, Miami, FL 33149 ABSTRACT: This report describes for the first time the pathologic findings associated with mortality in I7 Atlantic bottlenose dolphins (Tursiops truncatus) which stranded in the Indian River Lagoon (IRL), Florida, between February 2001—July 2002. The cause of death could be determined in most necropsy cases, which demonstrates the importance of performing gross and microscopic necropsy examinations on freshly stranded marine mammals. Causes of death in decreasing order of frequency were infectious/ inflammatory disease, cachexia, asphyxiation, degenerative central nervous system disease and traumatic injury. In cases of inflammatory/infectious disease and cachexia, multisystemic disease was common, which reflects the complex and dynamic processes associated with death. Dermatologic disease was present in nine cases. In many cases, the combined histologic pattern of skin, lymphoid and other lesions suggested a state of altered immunologic homeostasis and subsequent immunologic dysfunction. Under- standing the pathologic features associated with mortality of stranded IRL dolphins is important for the future management of this species and provides an insight into the health of the IRL ecosystem as a whole. Key Words: Pathology, stranding, Indian River Lagoon, Tursiops truncatus THE Indian River Lagoon (IRL) is a unique shallow water coastal ecosystem located on the east coast of Florida, which extends 250 km from Ponce de Leon Inlet south of Daytona Beach to Jupiter Inlet north of West Palm Beach. The IRL is an aggregate of three estuarine water bodies, the Indian River, Banana River and Mosquito Lagoon. This geographic setting contributes to the exceptional biological diversity of the system. In 1990, the IRL was identified as an Estuary of National Significance (IRLNEP, 1996). As part of the five year development of the Indian River Lagoon National Estuary Program Comprehensive Conservation and Man- agement Plan IRLCCMP), key living resources were identified for implementation of management and protection actions. In part and as a high priority, IRLCCMP was * Current address: Department of Agriculture, 700 Camino de Salud NE, Albuquerque, NM 87106. 226 No. 3 2003] BOSSART ET AL__PATHOLOGY OF STRANDED DOLFINS Mjei| to undertake studies of wildlife diseases occurring in the IRL region, which were caused by human activities (IRLNEP, 1996). Mammals probably represent one of the least studied wildlife groups occurring within and along the IRL (Woodward-Clyde, 1994). While Atlantic bottlenose dolphins (Tursiops truncatus) represent one of the top-level predators within the IRL, this population has been relatively understudied and important structural details, such as its abundance and home range, remain indeterminate. One of the earliest studies of IRL dolphins was carried out from 1979-1981 and involved marking and subsequent monitoring of marked animals, mostly within the northern section of the lagoon (Odell and Asper, 1990). Repeated sightings of some marked dolphins over several years suggested that at least some members of this population could be considered IRL residents (1.e., living all or most of their life within the IRL system). A more recent study carried out from 1999— 2001 in the Indian River, between Sebastian Inlet and Ft. Pierce Inlet, found strong photographic evidence for residency among a subset of naturally marked dolphins, and even stronger evidence of long term residency within the IRL from sightings of 11 dolphins originally marked by Odell and Asper (1990) (Mazzoil et al., 2002a). For the years 1993-2000, dolphin strandings in the IRL represented ap- proximately 40% of reported bottlenose dolphin strandings along the east coast of Florida (Stolen, 2002). Volunteer and non-volunteer members of the Southeastern United States (SEUS) Marine Mammal Stranding Network have collected data from live-stranded and dead IRL dolphins over the last 25 years. These data are archived by the National Marine Fisheries Service and the SEUS Stranding Coordinator. Data on foraging ecology, age structure, stock identity and geographic patterns of mortality have been gathered through examination of stranded animals (Hersh et al., 1990; Barros, 1993; Stolen, 1998). Seasonal stranding patterns have revealed high numbers of stranded dolphins in the spring and summer in the IRL system. In addition, with the exception of 1999, strandings in the IRL increased and have remained high since 1996 (Stolen, 2002). Results from these studies have been critical to the understanding of dolphin habitat use and life history, and continue to be valuable sources of information for managers and interested scientists. However, there have been no comprehensive studies to determine the health status of IRL dolphins, making all the more alarming the recent occurrence of an unusual mortality event of dolphins in the northern portion of the IRL during the summer of 2001, in which at least 30 dolphins died over a two month period (Whaley, 2002). To date, the explanation for these mortality patterns is unknown largely because the pathologic aspects of disease associated with strandings and mortality of IRL dolphins have not been characterized. In this report, we document for the first time the pathologic findings associated with mortality in 17 IRL dolphins that stranded and died between February 2001—July 2002. The cause of death could be determined in most cases, which demonstrates the importance of performing gross and microscopic necropsy examinations on stranded marine mammals. Understanding the pathologic features associated with the mortality of IRL dolphins is important for the future 228 FLORIDA SCIENTIST [VOL. 66 management of this population and provides an insight into the health of the IRL ecosystem as a whole. MetHops—Thirteen Atlantic bottlenose dolphins (four males [M]/nine females [F]) with minimum apparent postmortem decomposition received complete gross and microscopic necropsies (Table 1). Age categories were estimated from total body length and a necropsy was performed on each case following recommended protocols (Geraci and Lounsbury, 1993). Tissue sections from the lung, heart, liver, spleen, multiple lymph nodes, thymus (if present), gastrointestinal tract, pancreas, kidney, adrenal gland, skeletal muscle and skin were collected for histologic examination from Cases 1-8; 10-13. Additionally, central nervous system tissue (brain and/or spinal cord) was collected from Cases 1-6, 8, 10 and 11; and reproductive tract (ovary or testicle, uterus) from Cases 1-5 and 8. A complete gross and partial microscopic necropsy was performed on another adult male dolphin (Case 9). Additionally, distinctive firm white raised cauliflower-like and sometimes ulcerated cutaneous lesions were the only tissues sampled from four adult dolphins (2M/2F; Cases 14-17) due to advanced autolysis of other tissues. Tissues were placed in 10% neutral buffered formalin, routinely processed, embedded in paraffin, sectioned at 5 um and stained with hematoxylin and eosin. Gomori’s methenamine silver (GMS) was a special stain used on all skin lesions. ResuLTts—The necropsy study group consisted of three neonates (1M/2F) and 10 adults (3M/7F). Causes of death and pathologic findings for this study group are given in Table 1. The four inflammatory/infectious disease cases involved lesions of lungs only or lesions of multiple organ systems. One neonate (Case 1) had a moderate to severe multifocal suppurative bronchoalveolar pneumonia associated with aspirated foreign material suggestive of milk or amniotic fluid. Infectious agents, meconium or squamous epithelial cells were not present. No other significant lesions were observed. Alternatively, severe multisystemic infectious disease was present in the three remaining cases in this category. The remaining neonate that died from inflammatory/ infectious disease (Case 2) had severe multifocal pyogranulomatous bronchointer- stitial pneumonia and chronic-active ulcerative tracheitis with intralesional pulmonary and tracheal coccobacillary bacteria and nematodes. Additionally, a severe multifocal ulcerative chronic-active dermatitis (with intralesional coccobacillary bacteria and invasive ciliated protozoa), moderate diffuse necrotizing enterocolitis and mild mul- tifocal nonsuppurative lymphocytic meningitis were present. This neonate also had severe hepatic lipidosis, serous fat atrophy at multiple sites and splenic and widespread lymph node lymphoid depletion. One adult (Case 3) had a moderate to severe multifocal nonsuppurative lym- phocytic meningitis, moderate chronic fibrosing interstitial nephritis and moderate to severe eosinophilic lymphadentis with severe lymphoid depletion. Moderate diffuse cardiomyocyte degeneration also was present. Infectious agents were not observed. The remaining adult (Case 4) in this category had a severe multifocal chronic cholangiohepatitis and a moderate multifocal chronic-active suppurative bronchoal- velor pneumonia (with intralesional nematodes). Additionally, multifocal to co- alescing and often deep cutaneous ulcers were present along the peduncle, lateral thorax and rostrum (Fig. 1). Microscopically, the cutaneous lesions represented a se- vere multifocal chronic-active necrotizing hyperplastic dermatitis (with intralesional coccobacillary bacteria and invasive ciliated protozoa) (Fig. 2). Mild diffuse chronic enterocolitis and multifocal chronic-active gastritis were also present. No. 3 2003] BOSSART ET AL.—PATHOLOGY OF STRANDED DOLFINS 229 TABLE |. Pathologic findings in 17 stranded IRL dolphins. Case No. Age Sex Cause of death Primary lesions 1 Neonate IB Inflam./Infec. Disease* Pneumonia 2 Neonate F Inflam./Infec. Disease Pneumonia, Tracheitis, Enteritis, Dermatitis (bacterial/protozoal) 3 Adult F Inflam./Infec. Disease Meningitis, Nephritis, Lymphadentis, Pneumonitis 4 Adult F Inflam./Infec. Disease Cholangiohepatitis, Pneumonia, Colitis, Dermatitis (bacterial/protozoal) =) Neonate M Cachexia Hepatic Lipidosis, Lymphoid Depletion, Thymic Atrophy, Colitis 6 Adult le Cachexia Hepatic and Pancreatic Atrophy, Lymphoid Depletion, Dolphin Pox is Adult ig Cachexia Hepatic Lipidosis, Pulmonary Nematodiasis 8 Adult FE Asphyxiation Pulmonary Edema, Hepatitis, Meningitis, Dermatitis (bacterial) 9 Adult M Asphyxiation Pulmonary Edema, Endocardiosis 10 Adult M Degenerative Disease Spinal Cord Axonal Degeneration, Lobomycosis 11 Adult M Trauma Acute Mandibular Trauma, Cerebral and Cerebellar Edema 12 Adult E Unknown Hepatic Atrophy and Lipidosis (mild to moderate) 13 Adult FE Unknown Abdominal Fat Atrophy (moderate) 14 Adult M Unknown Lobomycosis 15 Adult I Unknown Lobomycosis 16 Adult F Unknown Epidermal necrosis 17 Adult M Unknown Epidermal Hyperplasia and Vacuolar Degeneration * Inflammatory/Infections Disease. Cases 5—7 died of cachexia. These dolphins had poor overall body condition characterized by pronounced peduncular vertebrae, ribs and scapulae with moderate to severe depletion of postnuchal and abdominal fat stores. The gastric chambers were empty and no other gross lesions were noted. Microscopically, these dolphins had moderate to severe hepatic lipidosis and atrophy. Additionally, less consistent, con- current or individual lesions included moderate to severe lymphoid depletion of multiple lymph nodes (Figs. 3 and 4) and the spleen (Cases 5 and 6); moderate to severe acinar atrophy of the pancreas and moderate pyloric trematodiasis (Case 6); mild pulmonary nematodiasis (Cases 5 and 7); and severe thymic atrophy and mild lympho- plasmacytic colitis (Case 5). Additionally, Case 6 had intracytoplasmic eosinophilic inclusion bodies within epidermal keratinocytes consistent with dolphin pox infection. None of the latter lesions were considered severe enough either alone or in combination to have caused death. Two adult dolphins (Cases 8 and 9) died of apparent asphyxiation. In both cases, the larynx was ventrolaterally dislodged into the esophageal lumen by a fixed [VOL. 66 FLORIDA SCIENTIST 230 Multifocal to coalescing deep cutaneous ulcers on the peduncle of a stranded IRL dolphin Fic. 1. (Case 4). Fic. 2. Photomicrograph of Fig. 1. The section of skin has three intradermal protozoa (arrows) with vacuolated cytoplasm surrounded by numerous neutrophils and occasional macrophages. H&E stain. Bar = 190 pm. Ma) BOSSART ET AL.—PATHOLOGY OF STRANDED DOLFINS No. 3 2003] : '* ne %. Fic. 3. Photomicrograph of a normal lymph node from a dolphin. The cortex contains highly cellular lymphoid follicles (arrows) surrounded by paracortical regions both composed of lymphoid cells. H&E stain. Bar 240 um. mS See aoe Py any nA Cre that died of cachexia (Case haracterized by the loss of normal histologic architecture with marked Fic. 4 Severe lymphoid depletion Photomicrograph of a lymph node from a stranded IRL dolphin 1S C 6). = 240 um. Bar H&E stain d hypocellularity. i lympho DBD FLORIDA SCIENTIST [VOL. 66 esophageal obstruction. In both cases, the obstructions consisted of swallowed whole fish, which were firmly lodged into the adjacent esophageal mucosa by dorsal and/or pectoral fin spines. Both dolphins had severe pulmonary edema and congestion. Case 8 had concurrent moderate to severe multisytemic inflammatory/infectious disease consisting of multifocal chronic-active necrotizing dermatitis (with intralesional coccobacilli), diffuse chronic-active necrotizing cholangiohepatitis and diffuse chronic nonsuppurative meningitis. Case 9 did not have gross evidence of multi- systemic disease. However, a complete set of tissues was not examined microscop- ically thus multisystemic disease could not be ruled out. Only focal mitral valve endocardiosis was observed microscopically, which was considered an incidental lesion. Case 10 died with degenerative CNS lesions consisting of locally extensive cervical spinal cord axonal degeneration with spheroid formation and myelin loss. The lesion was considered severe enough to have interfered with normal CNS function although the etiology could not be definitively determined. This dolphin also had multiple firm white raised cutaneous nodules and plaques distributed along the distal mandible (Fig. 5). Microscopically, the latter lesions represented a moderate multifocal granulomatous hyperplastic dermatitis (Fig. 6) with intralesional agyrophilic yeast consistent with Lacazia loboi (formerly Loboa loboi) (Fig. 7) and moderate multifocal cardiomyocyte degeneration and myocardial and endocardial fibrosis. Case 11 was in good overall physical condition and had moderate cerebellar and cerebral edema with locally extensive hemorrhage and edema of the subcutaneous tissue of the mandible. Evidence of significant inflammatory or infectious disease was not present. These lesions were attributed to a mandibular traumatic insult of unknown etiology. The cause of death could not be determined in necropsy Cases 12 and 13. Case 12 had moderate diffuse hepatocellar atrophy and mild diffuse hepatocellar lipidosis. Case 13 had moderate serous fat atrophy of abdominal fat stores. Other changes pres- ent in both dolphins were mild or nonspecific. No evidence of significant inflam- matory or infectious disease was present. Cutaneous lesions sampled from Cases 14-17, that only had skin tissue collected, were diagnosed as severe multifocal granulomatous hyperplasic dermatitis with intralesional agyrophilic yeast consistent with Lacazia loboi (Cases 14-15), mild focal necrosis of the epidermis of unknown etiology (Case 16) and moderate focal vacuolar degeneration and hyperplasia of the epidermis also of unknown etiology (Case 17). DiscussioN—This is the first report of pathologic findings in stranded Atlantic bottlenose dolphins from the IRL; and demonstrates the diverse pathobiologic data, which can be produced from gross and microscopic necropsy examinations on recently stranded cetaceans. In only two cases (15%) from the necropsy study group could a cause of death not be determined. Additionally, the data provide the first glimpse of the causes of mortality of IRL dolphins and may provide important initial clues as to the health status of the IRL dolphin population and ecosystem as a whole. Based on data from past case reports of stranded and captive cetaceans (Bossart No. 3 2003] BOSSART ET AL—PATHOLOGY OF STRANDED DOLFINS BBB ae ee sh pees a 5 E aia Fic. 5. Multifocal white slightly raised cutaneous nodules and plaques distributed along the distal mandible of a stranded IRL dolphin (Case 10). és ws Fa < vt ee 2 hd é ly en, 2 ag ail é 4 op 2 — © % * 5 BE q BB ich Tf BG Fic. 6. Photomicrograph of Fig 5. The dermis is infiltrated by numerous macrophages and multinucleate cells (arrow) occasionally containing pale tan round organisms with refractive walls (arrowheads). H&E stain. Bar = 100 pm. DB emirns 234 FLORIDA SCIENTIST [VOL. 66 jie 2 f my es a * AN inf Cee 3l. ea * ny HE 7 . ike Pr ge Pee OD fy ae BONS aue Fic. 7. Photomicrograph of a dermal granuloma (Case 10). Notice numerous, round to lemon- shaped, thick walled yeast cells (arrows) (Lacazia loboi). GMS stain. Bar = 50 um. et al., 2001; 1991; Bossart, 1995; Miller et al., 2001), some typical, as well as unique, pathologic features were found in this study. Four dolphins (31%; Cases 1-4) died of inflammatory/infectious disease. In some dolphins, inflammatory lesions contained intralesional organisms indicating an infectious etiology. Multisystemic inflammatory disease was the most common pathologic finding with only one case of single organ inflammatory disease (Case 1), the latter being a result of probable aspiration pneumonia in a neonate. Multisystemic inflammatory lesions included pneumonia, enterocolitis and dermatitis (Cases 2 and 4); nonsuppurative meningitis, interstitial nephritis and lymphadentis (Case 3); cholangiohepatitis (Case 4); and tracheitis (Case 2). Based on the presence of intralesional organisms or the histologic pattern, all of the pneumonia cases were likely bacterial and/or parasitic in origin, although dolphin morbillivirus could not be excluded in Case 2. However, inclusion bodies typical of dolphin morbillivirus infection were not observed. Additionally, the histologic pattern of the meningitis in Cases 3 and 8 was suggestive of a viral etiology, although viral inclusion bodies were not seen. Terminal bronchopneumonia or bronchointerstitial pneumonia is a common finding in captive and free-ranging single stranded dolphins reported in other studies (McBain, 2001). In these dolphins, the cause of death was likely respiratory compromise and terminal septicemia. Two cases of multisystemic inflammatory/infectious disease (Cases 2 and 3) — No. 3 2003] BOSSART ET AL—PATHOLOGY OF STRANDED DOLFINS 23D also had concurrent degenerative lesions including hepatic lipidosis, serous fat atro- phy and cardiomyocyte degeneration, which are consistent with the protein-calorie deficiency of starvation or the inanition associated with chronic disease (Barker, 1993; Jubb, 1993). Additionally, lymphoid depletion, which in dolphins can lead to a predisposition for developing opportunistic disease (see below), was present in Case 2. The degenerative lesions were likely secondary and reflected the generalized debilitated state of these dolphins, which further demonstrates the complex and dynamic processes associated with death. The dermatologic lesions present in dolphins that died of multisystemic inflammatory disease were associated with intralesional bacteria as well as invasive ciliated protozoa. Such findings are unique in our experience. The cutaneous lesions may have acted as an entry portal for other infectious agents and are discussed in- depth below along with other unusual dermatologic findings from this study. The dolphins (Cases 5—7) which died of cachexia had a poor overall body condition and histologic lesions consistent with protein-calorie deficiency associated with starvation including hepatic lipidosis, hepatic atrophy and pancreatic acinar atrophy (Barker, 1993; Jubb, 1993). Often, severe lymphoid depletion and thymic atrophy were present. These lymphoid lesions can be associated with the pathologic stress of starvation producing altered immunologic homeostasis and subsequent immunologic dysfunction (Bossart, 1984; 1995; Valli, 1993). The low-grade inflam- matory and parasitic lesions of the lungs and gastrointestinal tract also support a state of altered immunologic function in these dolphins. Additionally, Case 6 had kera- tinocyte intracytoplasmic inclusions of the skin consistent with dolphin pox in- fection, which is a non-fatal recrudescent viral disease activated by prolonged physiologic or pathologic stressors (Kennedy-Stoskopf, 2001). In this case, star- vation may have caused viral recrudescence. The conditions which led to cachexia could not be determined from the gross and microscopic necropsy data. However, further investigations of the conditions which lead to cachexia are warranted, as it suggests a potential problem with the food supply quantity or quality for dolphins living in the IRL. Two dolphins (Cases 8 and 9) died of apparent asphyxiation due to an esophageal obstruction. While uncommon, asphyxiation due to airway blockage from a dislodged larynx caused by food prey fish has been reported in stranded dolphins (Frasier, 2002). Necropsy examinations produced two important findings in these cases. First, the fish lodged in Case 9 was identified as a blackchin tilapia (Sarotherodon melanotheron). This is a non-native established species in the IRL (Jennings and Williams, 1992). Asphyxiation by this method represents a novel mortality factor associated with an exotic prey species and may be accidental or reflect an aberrant shift in prey species of IRL dolphins. Second, Case 8 had significant underlying multisystemic infectious disease, which may have predisposed it to selecting an inappropriate fish size. The fish species causing the obstruction was identified as a striped mojarra (Eugerres plumieri), a perciform IRL resident that may be a prey species for IRL dolphins and is incriminated as causing similar obstructions in predator fish species (Gilmore, 2002). One dolphin (Case 10) died with degenerative cervical spinal cord disease. This dolphin was judged to be of advanced age based on significant tooth wear and 236 FLORIDA SCIENTIST [VOL. 66 cutaneous scars and many of the observed lesions were thought to be age-related. Degeneration in the cranial cervical spinal cord can interfere with innervation to the diaphragm and intercostal muscles resulting in respiratory failure and death (Bullock and Henze, 2000). The cause of this lesion could not be determined but it may have occurred as a result of compressive injury secondary to age-related degenerative vertebral joint disease. An in-depth gross or histologic examination of the vertebral column to confirm this suspicion was not carried out. This dolphin also had cutaneous lesions similar to other dolphins in this study, which are described below. Another adult male dolphin (Case 11) had evidence of acute mandibular trauma and associated acute cerebral and cerebellar edema, which likely caused CNS com- promise and death. No evidence of an underlying disease process was present and death was speculated to be the result of an accidental blunt trauma of unknown etiology. The high number of cases with dermatologic lesions is worrisome and supported by a preliminary photo-identification study of IRL dolphins, which indicates an approximate 50% incidence of skin lesions (Mazzoil et al., 2002b). Causes of the skin lesions in our study varied but were generally found to be of an infectious etiology. The cutaneous lesions reported could have compromised epidermal integrity or produced pathologic stress, potentially acting as entry portals for pathogenic organisms and making animals more vulnerable to natural infections or anthropogenic factors. Alternatively, the lesions may have reflected the pathologic outcome of a generalized unhealthy state and/or yet undefined common environmental factor(s). A previous study (Wilson et al., 1999) indicated significant linear relationships between dolphin epidermal disease determined by photographs and environmental variables. Specifically, dolphin populations from this study that lived in areas of low water temperature and low salinity exhibited higher skin lesion prevalence and severity. We have recently designed a dolphin health assessment study in order to further investigate the nature of skin lesions, as well as general changes, in IRL dolphin mortality patterns. We anticipate that this study, which is part of a capture/ release program scheduled for the spring of 2003, will allow us to characterize environmental variables and the overall health of free-ranging IRL dolphins, as well as shedding light on the health of the IRL ecosystem as a whole. Dermatologic lesions observed in this study also have pathologic significance. Six cases had an infectious etiology that were bacterial, combined bacterial and protozoal, or fungal in nature. Three cases of granulomatous hyperplastic dermatitis associated with intralesional yeast consistent with Lacazia loboi were diagnosed, which is the classic histologic presentation of lobomycosis, a fungal disease of dolphins and humans (Bossart, 1984). The pathogenesis of lobomycosis is unknown, primarily because the fungus cannot be cultured in the laboratory. However, in stranded dolphins, lobomycosis has been associated with immunologic compromise and mutisystemic disease (Bossart, 1984). An immunologic suppressive cofactor in the pathogenesis of lobomycosis needs further investigation as it could have far ranging implications for the IRL dolphin population as a whole. Two cases of necrotizing hyperplastic dermatitis with invasive protozoa were observed. These lesions had a similar gross appearance as the white firm raised No. 3 2003] BOSSART ET AL—PATHOLOGY OF STRANDED DOLFINS 25] cauliflower-like lesions of lobomycosis, which were initially grossly diagnosed as lobomycosis. The latter data may be important information for those attempting to make an etiologic diagnosis from gross observations only. Dolphins with these lesions had severe multisystemic infectious disease, which could have produced a state altered generalized or localized cutaneous immunologic surveillance resulting in secondary opportunistic dermatologic disease. This speculation is supported by data from another published study of similar cutaneous lesions in stranded dolphins (Schulman and Lipscomb, 1999), which determined that the incidence of dermatitis with invasive ciliated protozoa was greater in dolphins that died of dolphin morbillivirus infection. Dolphin morbillivirus infects lymphoid tissues causing immunosuppression and increased susceptibility to secondary infections (Kennedy-Stoskopf, 2001). The currently reported data provides additional evidence for the possibility of an im- munologic disturbance in stranded IRL dolphins and requires further investigation. ACKNOWLEDGMENTS—We thank the volunteer members of the Southeastern Marine Mammal Stranding Network for their help with dolphin necropsies. We also wish to acknowledge Dr. Grant Gilmore for his expertise on fish identification, Megan Stolen and Hubbs-Sea World Research Institute for necropsy data submission, Indian River County Law Enforcement and Fire Rescue units, and Harbor Branch Marine Mammal Division Volunteers: Lisa Denham, Paul Denham, Bill Stewart, Jeanette Hezlitt, Marc Oberneir, Maggie Reinhalter, and Jennifer Bossart for their tireless efforts in advancing the science of marine mammal medicine and pathology. LITERATURE CITED Barker, I. 1993. The peritoneum and retroperitoneum. Pp. 425-426. Jn: JuBB, K., P. KENNEDY, AND N. PALMER (eds.), Pathology of Domestic Animals, Vol 2, Academic Press, New York, NY. Barros, N. B. 1993. Feeding ecology and foraging strategies of bottlenose dolphins on the central east coast of Florida. Ph.D. dissert. University of Miami, Miami, FL. Bossart, G. D., D. K. ODELL, M. T. WALSH, J. D. LyNcu, D. O. BEUSsE, R. FRIDAY, AND G. YOUNG. 1991. The histopathologic findings in a mass stranding of pilot whales (Globicephala macrorhynchus). In: Marine Mammal Strandings in the United States. NOAA/NMEFS, 98: 85-90. . 1984. Suspected acquired immunodeficiency in an Atlantic bottlenose dolphin with chronic- active hepatitis and lobomycosis. J. Am. Vet. Med. Assoc. 185: 1413-1415. . 1995. Immunocytes of the Atlantic bottlenose dolphin (Tursiops truncatus) and West Indian manatee (Trichechus manatus latirostris): A morphological characterization with correlations between healthy and disease states under free-ranging and captive conditions. Ph.D. dissert. Florida International University, Miami, FL. , T. H. Remarson, L. A. DIERAUF, AND D. A. DUFFIELD. 2001. Clinical pathology. Pp. 383-430. In: DieRAuUF, L. A. AND F. M. D. GULLAND (eds.), Marine Mammal Medicine. CRC Press, Boca Raton, BL. BuLLock, B. AND R. HENzeE. 2000. Focus on Pathophysiology, Lippincott, Williams and Wilkins, Philadelphia, PA. Frasier, K. 2002. Animal Diagnostic Laboratory, University of Georgia, Tifton, GA., Pers. Commun. Geraci, J. R. AND V. J. LouNspury. 1993. Marine Mammals Ashore: A Field Guide for Strandings, Texas A&M Sea Grant, Galveston, TX. GILLMoRE, R. G. 2002. Dynamac Corporation, Kennedy Space Center, FL., Pers. Commun. Hersu, S. L., D. K. ODELL, AND E. D. Asper. 1990. Bottlenose dolphin mortality patterns in the Indian/ Banana River system of Florida. Pp. 155-164. In: LEATHERWOOD, S. AND R. REEVES (eds.), The Bottlenose Dolphin. Academic Press, Inc., San Diego, CA. INDIAN River LAGOON NATIONAL ESTUARY PROGRAM (IRLNEP). 1996. The Indian River Lagoon Compre- hensive Conservation and Management Plan. Final Draft. Melbourne, FL. 238 FLORIDA SCIENTIST [VOL. 66 JENNINGS D. P. AND J. D. Witiiams. 1992. Factors influencing the distribution of blackchin tilapia Sarotherodon melanotheron in the Indian River system, Florida. Northeast Gulf Sci. 12: 111-117. Juss, K. 1993. The pancreas. Pp. 410-411. Jn: Juss, K., P. KENNEDY, AND N. PALMER, (eds.), Pathology of Domestic Animals. Vol 2, Academic Press, New York, NY. KENNEDY-STOSKOPF, S. 2001. Viral diseases. Pp. 285-307. In: DrerAur, L. A. AND F. M. D. GULLAND (eds.), Marine Mammal Medicine. CRC Press, Boca Raton, FL. Mazzort, M., S. MccuLLocH, AND R. H. DEFRAN. 2002a. Occurrence and distribution of Atlantic bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon (1996-2001). Southeast and Mid-Atlantic Marine Mammal Symposium. April 12—14. Coastal Carolina Univ., Conway, SC. , AND . 2002b. Bottlenose dolphin epidermal disease and Xenobalanus occurrence. Technical Report, Harbor Branch Oceanographic Institution, Ft. Pierce, FL. 5 pp. McBatn, J. F. 2001. Cetacean medicine. Pp. 895-906. In: DigRAuF, L. A. AND F. M. D. GULLAND (eds.), Marine Mammal Medicine. CRC Press, Boca Raton, FL. Miter, D. L., R. Ewinc, AND G. D. Bossart. 2001. Emerging and resurging diseases. Pp. 15—25. In: DIERAUF, L. A. AND F. M. D. GULLAND (eds.), Marine Mammal Medicine. CRC Press, Boca Raton, FL. ODELL, D. K. AND E. D. Asper. 1990. Distribution and movements of freeze-branded bottlenose dolphins in the Indian River Lagoon, Florida. Pp. 515-540. In: LEATHERWOOD S. AND R. REEVES (eds.), The Bottlenose Dolphin, Academic Press, San Diego, CA. SCHULMAN, F. Y. AND T. P. Lipscoms. 1999. Dermatitis with invasive ciliated protozoa in dolphins that died during the 1987-1988 Atlantic bottlenose dolphin morbilliviral epizootic. Vet. Pathol. 36: 171-174. STOLEN, M. K. 2002. Hubbs-Seaworld Research Institute, Orlando, FL, Pers. Commun. . 1998. Age, growth, and mortality of bottlenose dolphins from the east coast of Florida. Masters thesis. Univ. of Central Florida, Orlando, FL. VALLI, V. 1993. The hematopoietic system. Pp. 101—265. Jn: JuBB, K., P. KENNEDY, AND N. PALMER (eds.), Pathology of Domestic Animals. Vol 3, Academic Press, New York, NY. WHALEY, J. 2002. Office of Protected Resources, National Marine Fisheries Service, Silver Springs, MD, Pers. Commun. WIson, B., H. ARNOLD, G. BEARZI, C. M. Fortuna, R. Gaspar, S. INGRAM, C. LiRET, S. PRIBANIC, A. J. READ, V. Ripoux, K. SCHNEIDER, K. W. URIAN, R. S. WELLS, C. Woop, P. M. THOMPSON, AND P. S. HaMMonb. 1999. Epidermal diseases in bottlenose dolphins: Impacts of natural and anthropo- genic factors. Proc. R. Soc. Lond. 266: 1077-1083. WOODWARD-CLYDE CONSULTANTS. 1994. Biological Resources of Indian River Lagoon-Indian River Lagoon National Estuary Program. Final Technical Report. Project No. 92F274C. Melbourne, FL. Florida Scient. 66(3): 226-238. 2003 Accepted: January 21, 2003 Biological Sciences THE SUITABILITY OF SHREWS (BLARINA CAROLINENSIS, SOREX PALUSTRIS) FOR EXPERIMENTS ON SPATIAL LEARNING TASKS FRED PUNZO Department of Biology, University of Tampa, 401 W. Kennedy Blvd., Tampa, FL 33606 ABSTRACT: Studies were conducted to assess the spatial learning ability of the shrews, Sorex palustris and Blarina carolinensis. Spatial navigation was assessed in S. palustris using the Morris apparatus whereby subjects were required to locate a visible or invisible escape platform in a tank containing water made opaque by the addition of a non-toxic white dye. Adult males learned to locate a platform that they could not see, hear, or smell provided it remained in a fixed spatial location relative to distal room cues. Learning resulted in a significant decrease in the amount of time required to locate the platform (escape latency), as well as a reduction in the distance taken to locate it (path length), over a 20-trial test period lasting 5 days. Learning occurred rapidly except for a test group in which the platform was below the water surface and whose position was randomly changed from trial to trial. Transfer tests revealed that a spatial location search strategy was utilized by those animals for whom the platform was invisible but in a fixed location. Spatial learning ability was evaluated in B. carolinensis using a complex maze. The number of blind alley errors decreased significantly over a 10-day training period, from a mean of 72.4 + 13.8 SD on day I, to 2.5 + 0.2 errors on day 10. Similarly, the time required to traverse the maze decreased significantly from 483.2 + 47.9 sec on day I to 109.4 + 11.9 sec on day 10. This represents the first demonstration of spatial learning ability in these soricids. The data are discussed in relation to spatial learning studies conducted on the Norway rat, and the adaptive significance of spatial learning to foraging animals is addressed. Key Words: Blarina carolinensis, distal and proximal cues, Sorex palustris, spatial learning LEARNING is generally defined as a relatively permanent change in behavior that occurs as the result of previous experience (Kimble, 1971). It is known that the degree to which an animal can modify appetitive behaviors in response to changes in physiological state (motivation) and environmental conditions may have a profound effect on overall survivorship (Davey, 1989; Punzo, 1996). Spatial learning tasks are known to represent ecologically relevant learning paradigms for invertebrates and vertebrates (Johnston, 1982; Punzo, 2000; Punzo and Madragon, 2002). The ability to associate specific locations in three dimensional space with the availability of some required resource has been shown to increase fitness in animals from a number of taxa (Gallistel, 1990; Gormezano and Wasser- man, 1992). For example, spatial learning can significantly reduce the amount of time spent in random searching patterns thereby maximizing foraging activities (Stephens and Krebs, 1986; Punzo and Torres, 2003). In spatial learning studies with mammals, a disproportionate amount of research has been conducted on inbred laboratory strains of house mice and Norway rats, Wpyg) 240 FLORIDA SCIENTIST [VOL. 66 with wild animals receiving much less attention (Gormezano and Wasserman, 1992; Collett and Zeil, 1996). With respect to shrews, there is little information regarding their capacity for learning in general (Punzo and Chavez, 2003) despite the fact that they are amongst the most ancient of mammals, first appearing in the late Eocene epoch some 38 million years ago (Innes, 1994). Shrews (Insectivora) exhibit some of the earliest mammalian features including a generalized body plan, plantigrade mode of locomotion, small and narrow skulls, a mandibular surface with a double articulating surface, and relatively small brains (Churchfield, 1990). In addition, their brain weight/body weight ratios place them at an intermediate position between marsupials and murid rodents (Jerison, 1973). For these reasons, one would think that shrews would have received sufficient attention in behavioral studies. Yet, just the reverse is true. The general reluctance of many investigators to utilize shrews as experimental subjects is most likely attributed to their high energy demands and daily feeding requirements, hyperexcitability and low breeding success under cap- tive conditions, and high levels of aggression. Over the past several years, I have had some success in maintaining and breeding several species of shrews in captivity including the least shrew (Cryptotis parva), southern short-tailed shrew (Blarina carolinensis), desert shrew (Notiosorex crawfordi) and northern water shrew (Sorex palustris). This has provided me with the opportunity to conduct several behavioral studies including an analysis of aggression between B. carolinensis and several species of sympatric rodents (Punzo, 2003a), use of landmark cues in N. crawfordi (Punzo, 2003b), and sprint per- formance and spatial learning in C. parva (Punzo and Chavez, 2003). The northern water shrew (S. palustris) is one of the larger species of shrews in North America, with adults typically weighing between 11—15 g (Innes, 1994). It ranges from southeastern Alaska throughout Alberta and British Columbia, and south into Montana, northeastern California, Utah, the north central states, and east to North Carolina (Whitaker, 1996). It frequents rockslide areas and boulders along fast-moving streams, as well as marshy areas and sphagnum moss around lakes and ponds (Foresman, 2001). It is semi-aquatic and feeds primarily on a variety of aquatic insects and terrestrial invertebrates, and to a lesser extent on small fish, larval salamanders, and mice (Beneski and Stinson, 1987). The southern short-tailed shrew (Blarina carolinensis) ranges from southern Nebraska and Illinois, south to the Gulf of Mexico (Whitaker, 1996). It prefers mesic habitats in bayheads and flatwoods, and typically inhabits small burrows under decaying logs and plant debris (Gingerich, 1999). It forages for food over the ground surface and feeds on a variety of terrestrial invertebrates including earthworms, snails, and arthropods (French, 1984). Because the learning abilities of shrews has not been investigated, these experiments were conducted to assess spatial navigation and learning ability in these 2 species of soricids. Owing to the semi-aquatic habits of S. palustris and its strong capacity for swimming (Beneski and Stinson, 1987), spatial learning was tested using the Morris apparatus. Morris (1981) introduced a methodology that has proven successful with small mammals that are capable swimmers such as Norway rats, and allows the investigator to analyze the use of certain cues to locate an escape No. 3 2003] PUNZO—SHREWS AND LEARNING TASKS 24] platform. The task requires the subject to locate a platform hidden just beneath the surface of a pool of water. The platform can be made invisible to the subject by having its surface painted white and the water made opaque by the addition of a non- toxic white dye. In other cases, the platform can be clearly visible to the subject by painting its surface black and keeping the water clear. Spatial learning in B. carolinensis was tested using a complex maze with a floor plan similar to that used extensively in studies on rodents and insects. Mazes are known to provide ecologically relevant tasks for animals that forage over the surface of the ground in search of food, shelter sites, and mates (Benhamou and Poucet, 1996; Punzo, 2002). MATERIALS AND METHODS—Animals and rearing methods—The adult males of S. palustris used in these experiments were the third-generation offspring of captive-bred adults that were originally collected from several localities in Marshall Co., ND during May and June, 1999. The distance between collection sites ranged from 6—18 km. Animals were collected using pitfall traps consisting of 4-L tin cans that were placed along the margins of streams and ponds as described previously by Brown (1967). In order to ensure that test subjects were not closely related, animals chosen for captive breeding were from sites at least 8 km apart. The methods used to maintain and breed S. palustris in captivity have been described elsewhere (Punzo and Torres, 2003). Briefly, each pair of males and females were housed in a 50-L aquarium provided with a tubelick water bottle. The aquaria were kept in a room maintained at 20° + 0.5°C with a 10L:14D photoperiod regime. The floor of each aquarium was covered with gravel to a depth of 5 cm. One half of the floor was provided with flat rocks stacked on top of one another to a height of 16— 18 cm. The other 1/2 of the tank was filled with filtered pond water to a depth of 12 cm, giving the animals an opportunity to swim as well as the ability to climb out of the water. A plastic nest box (13 x 13 X 10 cm) was placed on top of the rocks against one end of the aquarium, with a single entrance hole facing toward the water. The floor of this box contained dampened sphagnum moss and cotton wadding as bedding material. A small glass food dish was placed against one side of the tank at a distance of 8 cm from the nesting box. Animals were removed every 3 days, and the aquaria and gravel thoroughly cleaned, and the pond water and bedding material replaced. Water shrews were fed ad libitum on a diet consisting of an equal mixture of commercial cat chow (Ralston Purina, St. Louis, MO), rolled oats, and minced cricket (Acheta domesticus). The adult males used in this study weighed 16-17 g. Males of both shrew species were used because preliminary observations indicated that they adjusted to the mazes more readily than females in that they exhibited less hyperexcitability and distress during testing trials. Individuals of B. carolinensis used in this study were second-generation adult males whose parents were collected at severasl sites in Jackson Co., FL during April and May, 2001. Shrews were collected using the same pitfall traps as described above. Animals used in these experiments were from sites at least 9 km apart. The methods used to maintain B. carolinensis in captivity have been described previously (Punzo, 2003a). To summarize, they were housed individually in 38 X 21 X 18 cm plastic animal cages (Model RG-67-4337, Carolina Biological Supply, Burlington, NC) and provided ad libitum with food and water. The cages were kept under the same temperature and photoperiod conditions described above. They were fed a diet consisting of cat chow, mealworms (Tenebrio molitor), crickets (A. domesticus), and earthworms (Lumbricus terrestris). Adult males weighed between 12-13 g. Morris apparatus and complex maze—TYhe Morris apparatus used to test spatial learning in S. palustris was similar to that described previously by Morris (1981). Briefly, a circular, collapsible plexiglass tank (1.3 m in diameter, and 0.6 m in height) was filled with either clear distilled water, or water made opaque by the addition of non-toxic Crayola white dye. This dye has been used successfully in spatial learning experiments with mice and rats (Castro and Rudy, 1987). Water level was maintained at 0.40 m, and water temperatures in the tank ranged from 20—21°C. The tank was filled and emptied via a drain system. The tank was located in the middle of a room (7 X 7 m) that included a tall bookshelf on the north-facing wall, cabinets on the east-facing wall, a barren white-painted west-facing wall, and 242 FLORIDA SCIENTIST [VOL. 66 40 cm Fic. |. Diagrammatic representation of the Turner complex maze used to assess spatial learning ability in adult males of Blarina carolinensis. The maze consisted of 5 blind alleys, a start box (S), and a goal box (G). See text for details. a south-facing wall containing a window. All of these served as distinct extramaze (distal) cues. Two lines of black string were placed at right angles across the top of the tank in such a way as to divide the tank into 4 equally-sized quadrants, with their points of attachment to the circumference of the tank facing N, S, E and W. All observations were conducted through a one-way mirror to minimize disturbance to the test subjects. Two 9-cm diameter circular perspex columns were used as escape platforms. One was covered with a white vinyl material, and when placed in the tank it became completely submerged to a depth of 1.5 cm below the surface. Its white color made it invisible when the water in the tank was treated with white dye. The other platform was covered with a black vinyl material and protruded 1.5 cm above the water surface, making it clearly visible against the backdrop of milky water. The shrews could escape from the water by using these platforms. A diagrammatic representation of the complex maze used to test spatial learning in B. carolinensis is shown (Fig. 1). The maze had a floor plan similar to that originally designed by Turner (1913) for large arthropods and rodents, and contained a start box (Fig. 1, S), goal box (G), and 5 blind alleys. It consisted of a white viny] sheet floor attached to a plywood base, and was covered with a transparent plastic lid. The walls of the maze were constructed of aluminum, painted white, and were 10 cm in height. A cool fluorescent light was positioned directly over the center of the maze and all observations were conducted through a one-way mirror to minimize disturbance to test subjects. A small glass dish containing a meal- worm (positive reinforcement) was placed at the center of the goal box. The start box was provided with a movable restraining panel so that the subject was prevented from entering the main body of the maze until the panel was removed. The orientation of the maze was rotated 180° every third day of trials as described by Gormezano and Wasserman (1992). Extramaze cues were identical to those described above for the Morris apparatus experiments. Spatial learning procedures in the Morris apparatus—Adult males (4.5—5 months of age) of S. palustris obtained from 8 different litters were randomly assigned to one of 4 groups, with each group containing 8 subjects. All procedures followed those described by Morris (1981) for rats. To summarize, there was a pretraining period during which all shrews were placed into the tank over a 2-day period, for 4 min on each day. The tank contained water made opaque by the white dye, and no escape platform was present. This allowed the shrews to become accustomed to swimming in the apparatus. No. 3 2003] PUNZO—SHREWS AND LEARNING TASKS 243 This was followed by escape acquisition trials which began on day 3. Animals in group | (cue + place) were allowed to escape the water by climbing onto a platform whose black surface protruded 1.5 cm above the water surface. The platform was placed in a fixed location throughout the training session for a given subject. These locations were designated as N, S, E, and W (compass positions), and counterbalanced across groups. At the start of each trial, a shrew was lowered into the water in such a way that it was facing N, S, E or W. An electronic timer was started and the amount of time required by the animal to locate and climb onto the platform was recorded (escape latency). The subject was allowed to remain on the platform for 60 sec, and then removed from the apparatus and placed back into the water at a different location for the start of trial 2. A sequence of starting locations was chosen whereby each shrew started from 2 locations over the course of the 8 training trials/day, with the sequence varying randomly from day to day. Subjects in group 2 (place) were treated in a similar fashion except that a platform with a white surface was used, and its surface was hidden 1.5 cm beneath water level. The opaque water made the surface of the platform invisible to the human eye at water level. Animals in group 3 (cue only) were trained using the black surface, above-water platform, but the position of the platform inside the tank was moved unpredicatbly (N, S, E or W) from trial to trial so that it occupied 2 locations across the 8 training trials. Group 4 animals (place-random) were allowed to escape onto the white submerged platform, whose location was placed in positions as previously described for group 3 subjects. This group allowed the animals to learn the location of the platform while also providing a behavioral check that the submerged platform provided no local cues. An additional 8 trials were conducted on day 4, with the sequence of start locations (for all groups) and end locations (for groups 3 and 4) changed. Day 4 began with a series of 4 trials and the recording of escape latencies as before. In addition, these trials were recorded on videotape from above the tank, using a Panasonic MVC recorder so that the movement paths of the subjects could be analyzed (path analysis). The path lengths were determined using a Umax Astro scanner and Alta 2400VS software (Alta Vista, CA). Prior to path length analyses, the fourth training trial of day 5 was immdediately followed by one of 2 test procedures (test A and B). The 4 test groups were subdivided, with 4 shrews from each group participating in each test. Animals were assigned in such a way that for the cue + place (group 1) and place (group 2) groups, there was now | subject trained to each compass position in each test condition. Test A was conducted to determine any spatial bias that might be exhibited by a trained subject. Whichever platform had been used during the preceeding 4 trials of day 4 was removed from the tank and the animal once again placed into the apparatus for a single 60-sec period. The movements of the animal were recorded on videotape and the amount of time spent in each quadrant of the tank was determined. Test B was also an assessment of spatial bias, but in this case the platform remained in the tank so that the subjects could learn a new spatial position. Following the 4 training trials of day 5, the platform was moved to a position diagonally opposite to its training position for groups | and 2, or placed in a fixed location for the first time for a sequence of 4 trials (groups 3 and 4). The escape latencies and path lengths were recorded as described previously. Thus, these testing protocols enabled the shrews to escape from water under conditions where an escape platform was either visible or invisible, and occupied either a fixed or semirandom position within a familiar space. Procedure in complex maze—All maze procedures followed those previously described by Punzo and Madragon (2002). To summarize, at the beginning of each trial a 4-5 month-old male shrew (B. carolinensis ), food-deprived for 24 hr, was placed in the start box and allowed to remain there for 3 min prior to testing. A trial began when the panel door was removed from the start box allowing the animal to enter the maze, and terminated when it reached the goal box (or after 30 min, if it failed to do so). Each animal fed on a small (<7 mm) mealworm (reward) upon entering the goal box. The walls and floor of the maze were washed with 70% ethanol after each trial. At the end of a trial, the subject was returned to the start box and the next trial initiated after a 5-min intertrial interval. The parameters used to assess performance in the maze were time/trial and the number of blind alley errors. Maze learning was assessed in 15 males, and each subject received 7 trials/day over a 10-day training period. Statistical analyses—A\\ statistical procedures followed those described by Sokal and Rohlf (1995), using Statistica software for Windows (Statsoft, Inc., Tulsa, OK). Data from animals tested in the Morris 244 FLORIDA SCIENTIST [VOL. 66 apparatus were analyzed using analysis of variance (ANOVA) and the Kruskal-Wallis test. In maze experiments, the Spearman rank correlation (r,) procedure was used to assess the relationship between: (1) days of trials and number of errors, and (2) days of trials and time to traverse the maze. ResuLts—Spatial learning/navigation in S. palustris—All shrews readily adapted to test conditions and were capable of swimming in the tank without diffi- culty. During pretraining, 25 of the 32 shrews (78%) swam around the perimeter of the tank, often making contact with the side walls. This was followed by swimming out into open water and crossing the tank several times during the initial trials on days 1 and 2. With respect to escape acquisition, the shrews readily learned to climb onto the platform on days 3, 4, and 5. All of the animals exhibited shaking movements of their bodies, typically followed by grooming. Animals did not appear to react any differently to the platform whether it was above or below the water surface. Acquisi- tion of escape behavior occurred rapidly for 3 of the 4 groups, with group 4 (place- random) exhibiting poorer performance than groups | (cue + place), 2 (place), and 3 (cue only) (Table 1). A Kruskal-Wallis test (a nonparametric analogue of single-classification ANOVA) (Sokal and Rohlf, 1995) by ranks of first trial escape latencies of all 4 groups was significant (H[3] = 16.13, P < 0.02), which confirmed the longer escape latencies of groups 2 and 4 (tested with the platform submerged) as compared to groups | and 3 (with the platform surface above water). This suggests that prior to any learning, the submerged platform was more difficult for these shrews to locate. An analysis of the paths taken by these shrews on the final training trial (trial 20) indicated that group 4 subjects were no more likely to head toward the escape platform upon their introduction into the tank than in any other direction. The paths were transcribed and an estimate of the direction of movement was calculated at the arbitrary point of one radius of the pool (0.65 m) into their track as described by Morris and co-workers (1982). The median deviation from a heading directly toward the platform was 94.6° for group 4, and 0° for groups |, 2, and 3. Random performance would yield a value of 90°. A Kruskal-Wallis test showed these differences to be significant (H[3] = 14.02, P < 0.05). Although these results do not prove that the platform was entirely invisible, they do suggest that whatever proximal cues may have been present were minimal and could not account in any significant way for the directional learning exhibited by these shrews. It should also be mentioned that the subjects in group 4 did learn that escape was possible, but without any definitive directionality in their behavior. The better performance of group 2 as compared to group 4 should be interpreted with this in mind. An ANOVA on the escape latencies of trials 1-8 for groups 1-3 showed a significant effect of groups (F2; = 10.11, P < 0.05) and trials (F7 147 = 14.02, P < 0.02), and no significant interactions (F,4147 = 1.42, P > 0.60) (Table 1). Orthog- onal comparisons between groups indicated no significant difference between groups | and 3 (F;.2; = 1.12, P > 0.50); however, both of these groups learned to escape somewhat faster than group 2 (F; 2; = 8.02, P < 0.05). No. 3 2003] PUNZO—SHREWS AND LEARNING TASKS 245 TaBLE 1. Escape latencies (in sec) over trials 1-20 for males of Sorex palustris in the Morris apparatus. Values expressed as means + SE. See text for details. Escape latency (sec) Place—Random Place Cue only Cue + Place Trial (Group 4) (Group 2) (Group 3) (Group 1) 1 HOMGp== 21EI Vs 2 WAS 34.3 + 9.4 Sul e267), 2 OSes 1186 44.3 +73 B74 se 7) 26.4 + 3.3 3 3 Bon aaed Zit 60.4 + 8.5 30,5) 28 Soi 21.4 + 6.8 4 GE] 95 35:4 = 5:8 19:8 = 3:8 14.7 + 4.5 5 ANTE = NO2 KOS 22 SL) PIB yes) SS) GE Shae: 6 66:47 1377 40.4 + 7.8 DirAwa= 29 23.5) = 164 7 By to) as Fs) Sy) LS yn ss wa) 24.4 + 4.5 20.6 + 3.8 8 OIoy ae ay YA) 22 ABD Dea see eo, 14.1 + 4.2 2 DON 2EaS:5 29.4 + 4.3 NOES se 403 ISS) 205) 10 DOL == LA. S028 Bh 7) AS, Se D5) 16.4 + 2.7 11 47.6 + 8.8 26.4 = 2.7 20.4 = 3.7 14.5 + 42 D2 oes 10/2 34.6 + 3.1 17.8 + 2.4 [S55 e==9326 13 36:8°22 6.2 3015722 522 Pe oo) EY Bi) Wiss) 229) 14 46.8 + 8.7 25.6 + 2.4 3 Spee] 8.6 + 0.4 15 3015, = .6:4 9) Ose II WQS 2S) Ouse 42 16 Gitte == Al 37, less 2s Ox 5) 23S) 8.6 + 3.4 17 38.8 + 9.2 DIA 23 MOR O5 9.4 + 2.5 18 50:7 + 9.4 NO. 22 ley Ih Pilista 2D Sije==eal 19 6O3u 357-9 10.4 + 2.5 8.8 + 0.4 11.4 + 2.4 20 p45 == 6.6 9.6 = 1.4 Diletta) OS) 2209) For trials 17—20 for these same 3 groups, escape latency did not vary significantly across trials (F> 5; = 0.74), while the main effect of groups was significant (F2 2; = 7.17, P < 0.02) (Table 1). Similar to results observed in initial acquisition, group 2 exhibited longer escape latencies than groups | and 3 (F,.2; = 10.89, P < 0.02). There were no significant differences between groups | and 3 (F; 2; = 0.79, P= 0.50). Analysis of path length for trials 17-20 showed that group 4 (place-random) had significantly longer path lengths than the other 3 groups (Table 2). Similar to the latency data, there was no overlap in the scores of group 4 and the mean path lengths of any subject in groups 1—3. Thus, analyses were restricted to groups 1, 2, and 3, and they showed a significant effect of groups (F221; = 6.12, P < 0.05) as well as the stability of path length over trials (P > 0.60). Orthogonal tests showed that groups 1 and 3 (cue groups) required shorter path lengths to locate the platform (F;.2; = 10.06, P < 0.02) than the other 2 groups, and that their path lengths did not differ sig- nificantly (F; 2; = 0.92). The performance of the 4 subjects in each subgroup on test A (without platform present) is shown in Table 3. Group 2 (place) shrews exhibited a strong directional bias toward their respective training quadrant (TR), whereas a weaker bias was shown by group 1 animals (cue + place). In addition, there was no indication of directional bias in groups 3 and 4. An ANOVA conducted on the time spent in each of the 4 quadrants (SW, NW, NE and SE) by each of the 4 groups showed 246 TABLE 2. Path lengths (in m) for Sorex palustris in the Morris apparatus over trials 17—20. Values FLORIDA SCIENTIST represent means + SE. See text for details. Cue + Place Place Cue only Place—Random Trial (Group 1) (Group 2) (Group 3) (Group 4) 17 1.23 + 0.4 2.66 + 0.3 V-42 22°02 4 21 ES 18 0.94 + 0.2 2.98 2102 1.29 + 0.4 5.25 + 0.9 19 O78. 02 1525102 12 S02 3.98 + 0.4 20 O72 =103 1.28 + 0.1 0.81 + 0.3 5:05 22 hal Path length (m) [VOL. 66 a significant groups X quadrants interaction (F¢,36 = 5.02, P < 0.02). The time spent in the TR by group 2 was greater than the time spent in the same quadrant by group 1 (F, 36 = 6.11, P < 0.05). However, the time spent in the TR by group 1 did not depart significantly from the null hypothesis value of 15 sec (P > 0.50). Path lengths were not measured in test A. The performance of the 4 animals in each subgroup on test B (with platform) is shown in Table 4. Group 2 animals required more time to locate the platform as compared to other groups, and group 3 animals required the least amount of time. Similar to test A results, group 2 subjects swam to and about the place where the platform had been located previously until they eventually located the platform at its new location. Learning the new location of the platform then proceeded quickly. The performance of groups 3 and 4 is of interest because, in this test, it was their first experience with the platform placed in a fixed position for a series of 4 trials. Group 4 exhibited a decrease in escape latency across the 4 trials. An ANOVA on escape latency showed a significant effect of groups (F312 = 13.03, P < 0.01) and trials (F3 36 =4.18, P < 0.05), with no significant interaction. Subsequent orthogonal com- parisons showed that group 2 animals escaped to the new platform location significantly more slowly than group 4 (F; ;2 = 16.93, P < 0.01), confirming the place-bound swimming tendencies of group 2 subjects, especially on the first trial. The escape latency of group 1 was not slower than group 3 (P > 0.50). Path length analyses showed a performance pattern similar to that found for escape latencies (Table 4). Group 2 subjects exhibited the longest path lengths over all trials, and group 3 the shortest. Complex maze learning in B. carolinensis—The results showed that B. carolinensis is capable of learning a complex maze. The number of errors and time to traverse the maze decreased significantly over the 10-day training period (Table 5). There was a significant correlation between days of trials and number of errors (tr, =—0.781, P < 0.02), and days of trials and running time (r, =—0.757, P < 0.02). These shrews exhibited a reliable and sustained level of locomotor activity within the maze and showed no obvious signs of hyperexcitability. At the beginning of each trial, they began to move about and entered the main body of the maze almost immediately after the panel on the start box was removed. No. 3 2003] PUNZO—SHREWS AND LEARNING TASKS 2AT TABLE 3. The duration of time (in sec) spent by Sorex palustris in each quadrant of the Morris apparatus in test A. Duration of each trial was 60 sec. TR is the training quadrant; A/L and A/R are the adjacent quadrants to the left and right, respectively; OP is the opposite quadrant for groups 1 (cue + place) and 2 (place). The data for other test groups 3 (cue only) and 4 (place—random) are organized with respect to compass positions. Only group 2 exhibited a strong spatial bias. Values represent means + SE. See text for details. Quadrant Group 2 Group 1 Group 3 Group 4 A/L See) Be Si) 12.2 + 4.5 T/R Sie = 98 Ns) 2= DL0) A/R 1-8 39 BO 22-8 OP 6:72 2/9 38) 28 Zl) SW ZO Wess Onl SIO ate AD NW 14.5 + 4.8 GOP =se325 NE 14.8 + 4.3 14.8 + 3.3 SE NOOR 226 125) eae et Discussion—These results show that S. palustris and B. carolinensis are both capable of learning spatial tasks. To my knowledge, this represents the first demon- stration of spatial learning in these soricids. These results also indicate that, despite the general reluctance of many investigators to conduct behavioral experiments with shrews (owing to their hyperexcitability, high levels of aggression, and high nutri- tional demands), S. palustris and B. carolinensis can be used as animal models for behavioral experiments. With respect to S. palustris, it is an adept swimmer and readily adjusted to conditions within the Morris apparatus. This shrew was able to locate an object that it could not see, hear, or presumably smell, by locating its position in a familiar 3- dimensional space. When they were allowed to approach the platform from any direction, their path lengths indicated a strong directional tendency. Although their performance was better when they could see the platform, localization with respect to a hidden platform was learned relatively rapidly. The fact that animals exhibited shorter path lengths and escaped more quickly when searching for a platform visible above the water when compared to group 2 animals (place) suggests that although distal cues may be sufficient to bring the animals to the general location of the platform, the presence of proximal cues does improve their performance. This type of spatial navigation ability could have important ramifications for the survival of this shrew in the wild. For example, S. palustris could use distal cues, such as the height of certain trees or the contour of certain rocks or boulders along the margins of rivers or streams, while foraging, and then use proximal cues such as the shapes of rock crevices or substrate color/texture patterns for the subsequent location of shelter sites or food. The performance of S. palustris in the Morris apparatus was comparable to that demonstrated by Norway rats (Morris et al., 1982; Castro and Rudy, 1987). Thus, the performance of soricids in spatial learning tasks is in line with that exhibited by other mammals. Previous research with murid rodents, canids, and primates has led investigators to argue that animals form a type of spatial (cognitive) map of the environment that represents the relative locations of distal and proximal cues in 248 FLORIDA SCIENTIST [VOL. 66 TABLE 4. The escape latency (sec) and path length (m) over trials 14 in test B for Sorex palustris in the Morris apparatus. Values represent means + SE. Group 2 subjects (place) searched where the escape platform used to be located in the first test trial and then rapidly learned to approach the new platform location using a slower route. Group 1 =cue + place; Group 3 = cue only; Group 4 = place—random. See text for details. Group | Group 2 Group 3 Group 4 Escape latency (sec) Trial 1 i354 47.3 + 8.5 SV PES ii) 14.3 + 5.5 Trial 2 42 = 37, 245 + 4.4 Peismeme Mss) 12:9 229326 Trial 3 15225276 211 "6:6 MOB 255) 13.7 26 Trial 4 Oe Al hy Eee 8.8" Dey O38) a Path length (m) Trial 1 3240" 15 7i35| Bak 1.18 + 0.4 1.55° 07 Trial 2 Pigs eae 4.51 + 1.6 0:37 7-103 Py ses, (Os) Trial 3 eT 12054 3.44 + 1.2 OFS 102 1.46 + 0.6 Trial 4 OLS Sess 0:2 PRIA el), 0.64 + 0.2 0.84 + 0.2 relation to the location of specific objects (O’Keefe and Nadel, 1978; Benhamou and Poucet, 1996). In the Morris apparatus animals may first establish their own location with reference to this cognitive map, giving them a search image of where the platform is located. It has been suggested that many animals are capable of using such search images, formed with the help of previous experience, to locate required resources (Stephens and Krebs, 1986; Collett and Zeil, 1996). The performance of B. carolinensis in the Turner complex maze was also comparable to that reported for murid rodents (Kimble, 1971; Davey, 1989). For many animals, including shrews, the availability of resources such as food may vary both spatially and temporally over the course of a day or several days. If, however, this spatial and/or temporal pattern remains consistent on a daily basis, an efficient forager should learn to visit those sites where food abundance is highest thereby reducing the amount of time spent in random search (Johnston, 1982; Punzo, 2000) TABLE 5. Number of blind alley errors and time (sec) required to traverse a complex maze for Blarina carolinensis. Values represent means + SD (N = 15) for 7 trials/day, over a 10-day training period. See text for details. Day Number of errors Time to traverse maze (sec) l IPA ZENBRS 483.2 + 47.9 2 5638s 1e4 3438 = 26 3 61.6 + 14.1 304.8 + 33.1 4 Soe = 8.8. 224.9 == ies 5) 19.6 + 6.9 233.5 eal 6 | es penton) 198.3 + 16.8 7 Siar a= 2a 184.4 + 20.1 8 (ag rue 165) 131.3 + 14:5 9 DA 128.9 + 17.4 10 Died pote) 109.4 + 11.9 No. 3 2003] PUNZO—SHREWS AND LEARNING TASKS 249 and optimizing foraging efficiency (Stephens and Krebs, 1986; Able 1991). The same would apply to learning the most efficient (1.e., shortest path length) escape routes and shelter sites. In conclusion, the results of this study show that S. palustris and B. carolinensis are capable of spatial learning, an ability that would allow them to remember locations that may be characterized by higher prey densities, or are suitable as shelter sites or escape routes thereby increasing overall fitness. In addition, S. palustris was able to locate an escape platform when it was visible (proximal cues) or invisible (distal cues). Thus, they were able to learn to approach a spatially fixed object that they could not see, hear, or smell. The ability to form spatial maps (spatial memory) gives the animal information about its environment and makes it possible to distinguish various locations from one another. The extent to which an animal can utilize spatial maps depends upon the complexity of the neural substrates involved in a given species and the degree to which various objects in its environment have their own distinctive features. ACKNOWLEDGMENTS—I would like to thank L. Harvey, D. F. Martin, B. Martin, and anonymous reviewers for commenting on an earlier version of the manuscript, C. Farmer and E. Pedrosa for assisting in the maintenace of the animals and running some of the experiments, and L. Hane who helped in procuring some of the research literature. A Faculty Development Grant from the University of Tampa made much of this work possible. LITERATURE CITED ABLE, K. P. 1991. Common themes and variations in animal orientation systems. Amer. Zool. 31: 157— CYR BeENngESKI, J. T. AND D. W. STINSON. 1987. Sorex palustris. Mammal. Species 296: 1-6. BENHAMOU, S. AND B. PouceT. 1996. A comparative analysis of spatial memory processes. Behav. Proc. 35: 113-126. Brown, L. N. 1967. Ecological distribution of six species of shrews and comparison of sampling methods in the central rocky mountains. J. Mammal. 48: 617-624. Castro, C. A. AND J. W. Ruby. 1987. Early-life malnutrition selectively retards the development of distal- but not proximal-cue navigation. Develop. Psychobiol. 20: 521-537. CHURCHFIELD, S. 1990. The Natural History of Shrews. Cornell Univ. Press, Ithaca, New York, i178 pp. Co.tetTT, T. S. AND J. Ze. 1996. Flights of learning. Curr. Dir. Psychol. Sci. 5: 149-155. Davey, G. 1989. Ecological Learning Theory. Routledge, New York, NY. 392 pp. FORESMAN, K. R. 2001. The Wild Mammals of Montana. American Soc. Mammalogists, Special Publ. No. 12, Allen Press, Lawrence, KS, 278 pp. FreNcH, T. W. 1984. Notes on the distribution and taxonomy of short-tailed shrews (genus Blarina) in the southeast. Brimleyana 24: 121-128. GALLISTEL, C. R. 1990. The Organization of Learning. MIT Press, Cambridge, MA, 624 pp. GINGERICH, J. L. 1999. Florida’s Fabulous Mammals. World Publs., Tampa, FL, 128 pp. GORMEZANO, I. AND E. A. WASSERMAN. 1992. Learning and Memory. L. Erlbaum, Hillsdale, NJ, 539 pp. INNES, D. G. L. 1994. Life histories of the Soricidae. Pp. 111-136. Jn: MerRriTT, J. F., G. L. KIRKLAND, AND R. K. Rose (eds.), Advances in the Biology of Shrews. Carnegie Mus. Nat. Hist., Special Publ. No. 18, Pittsburgh, PA. JERISON, H. J. 1973. Evolution of the Brain and Intelligence. Academic Press, New York, NY. 482 pp. JOHNSTON, T. D. 1982. Selective costs and benefits in the evolution of learning. Adv. Study Behav. 12: 65— 106. 250 FLORIDA SCIENTIST [VOL. 66 KimsLe, G. A. 1971. Hilgarde and Marquis Conditioning and Learning. 2nd ed. Appleton, Century and Crofts, New York, NY. 578 pp. Morris, R. G. M. 1981. Spatial localization does not depend on the presence of local cues. Learn. Motiv. 12: 239-260. , P. GarRRuD, J. RAWLINS, AND J. O'KEEFE. 1982. Place navigation impaired in rats with hippocampal lesions. Nature 297: 681-683. O’KEEFE, J. AND L. NADEL. 1978. The Hippocampus as a Cognitive Map. Oxford Univ. Press, Oxford, 432 Pp. Punzo, F. 1996. Localization of brain function and neurochemical events associated with learning in insects. Rec. Trends Comp. Biochem. Physiol. 2: 9-16. . 2000. Desert Arthropods: Life History Variations. Springer, Heidelberg, 321 pp. . 2002. Spatial associative learning in the crevice spiny lizard, Sceloporus poinsetti. Texas J. Sci. 54: 45-50. . 2003a. Behavioral interactions between the shrew, Blarina carolinensis and several species of Florida mice. Florida Scient. (in press). . 2003b. Natural history and ecology of the desert shrew, Notiosorex crawfordi from the Chihuahuan Desert, with notes on captive breeding. Mammalia (in press). AND S. MADRAGON. 2002. Spatial learning in Australian skinks of the genus Ctenotus. Amphibia- Reptilia 23: 233-238. AND S. CHAVEZ. 2003. Effect of aging on a spatial learning task and running speed in the least shrew, Cryptotis parva. J. Mammal. (in press). AND B. Torres. 2003. Undernutrition and spatial navigation in the northern water shrew, Sorex palustris. Behav. Proc. (in press). SOKAL, R. R. AND F. J. RoHLF. 1995. Biometry. 5th ed. W. H. Freeman, San Francisco, CA. 818 pp. STEPHENS, D. W. AND J. R. Kress. 1986. Foraging Theory. Princeton Univ. Press, Princeton, NJ, 712 pp. TURNER, C. H. 1913. Behavior of the roach (Periplaneta americana) in an open maze. Biol. Bull. 25: 348— 365. WHITAKER, E. O., JR. 1996. Field Guide to the Mammals of North America. Alfred A. Knopf, New York, INY= 937 pp: Florida Scient. 66(3): 239-250. 2003 Accepted: January 28, 2003 FLORIDA ACADEMY OF SCIENCES MEDALISTS 1963 Dr. Archie Carr 1964 Dr. Werner A. Baum 1965 Dr. Alex G. Smith 1966 Dr. Karl Dittmer 1967 Dr. Alfred H. Lawton 1968 Dr. Sidney Fox 1969 Dr. F. G. Walton Smith 1970 Dr. Pierce Brodkorb 1971 Dr. Maurice A. Barton 1972 Dr. Lloyd M. Biedler 1973 Dr. Ruth S. Breen 1972 Dr E. T. York, Jr. 1975 Dr. Alex E. S. Green 1976 Dr. Robert N. Ginsburg 1977 Dr. Michael Kasha 1978 Dr. John Edward Davies 1979 Dr. Stanley S. Ballard 1980 Dr. Thomas D. Carr 1981 Dr. Harold J. Humm 1982 Dr. George B. Butler 1983 Dr. Karen Steidinger 1984 Dr. Yngve Ohrne 1985 Dr. William Sears 1986 Dr. E. Dwight Adams 1987 Dr. Larry Hench 1988 Dr. Gene C. Ness 1989 Dr. Frank Bradshaw Wood 1990 Dr. Martin Uman 1991 Dr. Frank Millero 1992 Dr. Gregory Choppin 1993 Dr. D. O. Shah 1994 Mrs. Barbara B. Martin Dr. Dean F. Martin 1995 Dr. James N. Layne 1996 Dr. Leslie Sue Lieberman 1997 Dr. George R. Newkome 1998 Dr. Joseph J. Krzanowski 1999 Dr. Abraham Kandel 2000 Richard F. Lockey, M.D. 2001 Dr. Clinton J. Dawes 2002 Dr. John H. Trefry 2003 Dr. Christian A. Clausen University of Florida University of Miami University of Florida Florida State University University of South Florida University of Miami University of Miami University of Florida Mound Park Foundation Florida State University Florida State University University of Florida University of Florida University of Miami Florida State University University of Miami University of Florida University of Florida University of South Florida University of Florida Florida Dept. Natural Resources University of Florida Florida Atlantic University University of Florida University of Florida University of South Florida University of Florida University of Florida University of Miami Florida State University University of Florida University of South Florida University of South Florida Archbold Biological Station University of Florida University of South Florida University of South Florida University of South Florida University of South Florida University of South Florida Florida Institute of Technology University of Central Florida 251 Biology Meteorology Astronomy Chemistry Medicine Biochemistry Marine Science Zoology Medicine-Teaching Physiology-Biophysics Botany Agriculture Physics Geology Molecular Biophysics Public Health Optics Astronomy Marine Biology Chemistry Biology Chemistry Anthropology Physics Engineering Biochemistry Astronomy Electrical Eng’n Oceanography Chemistry Chemical Eng’n Chemistry Chemistry Zoology Anthropology Chemistry Pharmacology Computer Science Medicine Biology Marine Science Chemistry REVIEW Faraday, Michael. 2002. The Chemical History of a Candle (reprint edition). Dover Publications, Inc., New York, NY. 223 pp. $9.95+SH. (ISBN 0-486- 42542-8) MICHAEL Faraday (1791-1867) was a largely self-educated Sandemanian, who developed into one of the greatest experimental scientists of his time. He gained employment at the Royal Institution on March 1, 1813 on the basis of a copy of notes he took during public lectures there given by Sir Humphrey Davy. He transformed himself from a laboratory assistant to Director (in 1825). In addi- tion to his personal experiments in electricity and magnetism, Faraday became a popular lecturer. He was especially gifted in giving lectures to children. This book describes six lectures before juvenile audiences in the Christmas holidays of 1860-61. He examined in interesting detail the chemistry of a burning candle. This interesting book provides the transcripts of those six lectures with many il- lustrations of the demonstrations that accompanied his lectures. Some are demon- strations that I have done for various general chemistry classes over the years. Faraday managed to get an impressive amount of information out of a burning candle, and he used the burning candle as a springboard to consider a type of en- ergy (candlelight), components of the atmosphere, capillary action, and so on. It also provides an interesting set of demonstrations. It provides a good insight into actions of an outstanding experimentalist for those interested in the history of sci- ence.—Dean F. Martin, University of South Florida, Tampa, FL. Zao INSTRUCTION TO AUTHORS This information is available at two web sites: (1) IES site: http:www.chumalcas.usf.edu (click on “Centers and Institutes”, then select “Institute for En- vironmental Studies”, then select “Florida Scientist”). (2) FAS site: http://www.floridaacademyofsciences.org (select “Florida Scientist’). It is possible to search for papers and abstracts of papers for the past six years: Go the FAS site (above), select “Allen Press”. When the link is complete, select “Advanced Search” and follow the directions. The Florida Academy of Sciences Web page, indicated above, allows interested persons to join the Florida Academy of Sciences and pay dues. FLORIDA ACADEMY OF SCIENCES CORPORATE MEMBER Tierra Verde Consulting, Inc. INSTITUTIONAL MEMBERS Archbold Biological Station FFWCC (Gainesville) Army Corps of Engineers Technical Florida Marine Research Institute Library, Jacksonville Hillsborough Community College Disney’s Animal Kingdom NIOZ-Netherlands Institute for Sea Research Duke University Science Library, University of Chicago Florida Community College at Jacksonville South Florida Water Management District Florida Fish and Wildlife Conservation University of North Florida Commission (Panama City) University of Washington FFWCC (Ocala) US EPA Library FFWCC (Olustee) Virginia Institute of Marine Science FFWCC (Lakeland) Virginia Polytechnic University FFWCC (West Palm Beach) WDI Florida FFWCC (Tallahassee) Membership applications, subscriptions, renewals, and changes of address should be addressed to the Executive Secretary, Florida Academy of Sciences, Orlando Science Center, 777 East Princeton St., Orlando, FL 32803. Phone: (407) 514-2079 Send purchase orders, payments for reprints and publication charges, orders for back issues and other journal business matters to the Business Manager, Dr. Richard L. Turner, Department of Biological Sciences, FIT, 150 W. University Blvd., Melbourne, FL 32901-6975 [(321) 674-8196; e-mail rturner@fit.edu]. SONIAN INSTITUTION LIBRARIES WEI 42162 will PUBLICATIONS FOR aie by the Florida Academy of Sciences Complete sets. Broken sets. Individual numbers. Immediate delivery. A few numbers reprinted by photo-offset. All prices strictly net. Prices quoted in- clude domestic postage. Some issues may not be available. All are $20 per volume or $5 per issue, except for symposium issues. PROCEEDINGS OF THE FLORIDA ACADEMY’ OF SCIENCES (1936-1944) Volumes 1—7 QUARTERLY JOURNAL OF THE FLORIDA ACADEMY OF SCIENCES (1945-1972) Volumes 8-35 FLORIDA SCIENTIST (1973—) Florida ’s Estuaries—Management or Mismanagenien? "01am Symposium FLORIDA SCIENTIST 37(4)—$5.00 Land Spreading of Secondary Effluent—Academy Symposium FLORIDA SCIENTIST 38(4)—$5.00 Solar Energy—Academy Symposium FLORIDA SCIENTIST 39(3)—$5.00 (includes do- inven instructions) Anthropology—Academy Symposium FLORIDA SCIENTIST 43(3)—$7.50 Shark Biology—Academy Symposium FLORIDA SCIENTIST 45(1)—$8.00 Future of the Indian River System—Academy Symposium FLORIDA SCIENTIST 46(3/4)—$15.00 Second Indian River Research Symposium—Academy Symposium FLORIDA SCIENTIST 53(3)—$15.00 Human Impacts on the Environment of Tampa Bay—Academy Symposium FLORIDA SCIENTIST 58(2)—$15.00 Please send payment with order. If required, an invoice will be sent on purchases over $20 from a recognized institution. Please include Florida sales tax if shipped to a Florida address, unless a copy of a valid Florida Consumer’s Sales Tax Certificate is enclosed.