TIGER BEETLES OF CONNECTICUT PROJECT REPORT   note: this report to the CT DEP incorporates all the data from the 1996 report to The Nature Conservancy and includes new data. The appendices are absent from this version, however, material examined data can be found on each species' webpage with the exception of sensitive data dealing with the localities of protected species (that have been omitted from this on-line version). In other respects this report should be identical to the report submitted to the CT DEP. -Derek Sikes, 1999   Connecticut Tiger Beetle Status Survey (Carabidae: Cicindelinae) 1996-1998 Conservation status, Taxonomy and Ecology       Derek S. Sikes Department of Ecology and Evolutionary Biology University of Connecticut Storrs, CT 06269 phone: (860) 486-6365 FAX: (860) 486-6364 dss95002@uconn.uconnvm.edu         Submitted to: The Connecticut Department of Environmental Protection       22 October 1998 Table of Contents Abstract Introduction Methods Collections Databasing Field Surveys Results and Discussion Tabular Summaries Species Accounts (not below but on separate pages in this website) Conclusions and Suggestions for Future Work Acknowledgments Literature Cited     Abstract Of the 14 species of tiger beetles known from Connecticut, 6 (C. repanda, C. sexguttata, C. punctulata, C. scutellaris, C. duodecimguttata and C. rufiventris) appear to be stable enough within the state to currently be of no conservation concern. Of the remaining 8 (57%) species, 3 are extinct within the state (C. dorsalis, C. purpurea, one of which may be a non-native: C. limbalis), 2 are federally listed as Threatened (C. dorsalis, and C.puritana), and 1 (C. formosa generosa), known from 2 apparently small populations that were found during 1997 and 1998, is the second rarest tiger beetle in the state behind C. puritana. Populations of 2 (C. hirticollis and C. tranquebarica) of the remaining 3 species are present in far fewer numbers than where known from historical records. The single remaining species (C. marginata) oddly, has few to no historical records in Connecticut but remains of conservation concern due to its rarity. The top species of conservation concern, exclusive of extirpated species, are (in order from highest to lowest): C. puritana, C. formosa generosa, C. hirticollis, C. tranquebarica, and C. marginata. With the addition of data from the 1997-1998 field seasons this ranking has not changed from that of 1996 with the single exception that C. rufiventris, now known from almost as many post-1980 populations as C. duodecimguttata, should probably no longer be considered of conservation concern. The most significant additions since the 1996 report include 2 recently discovered, but apparently small, populations of C. formosa, a species feared to be in danger of extirpation from CT. New populations of C. hirticollis (1), C. marginata (1), C. rufiventris (4) and possibly C. tranquebarica (2) were also found. This report combines data from 1996-1998 and represents over 60 site visits, 80 hours in the field exclusive of travel time, over 2,000 miles driven, 60 populations of Cicindela species surveyed, over 130 tiger beetle voucher specimens collected, and the creation of a database of 1632 specimen records that include data from 16 collections. During 1997-1998, with a research focus restricted to just six species of conservation concern, I made 22 site visits, spent 25 hours in the field exclusive of travel time, drove 880 miles, documented 21 populations of Cicindela species, built an on-line key to larval tiger beetles of Connecticut, and checked six additional collections for Connecticut records. In addition to summary statistics presented on the fauna, I provide for each species discussion of current and past status, synonymies, conservation status ranking for federal, state and Heritage Program rankings, identification diagnoses for adults and larvae, ecological and biological information, distribution maps in Connecticut, graphs of collection events versus time (1850-1998), and specimens examined. This report in its entirety, with additional high-quality color images of the species, and a key to adults and larvae has been prepared as a World Wide Web site (http://viceroy.eeb.uconn.edu/ctb/home.html). Introduction Tiger beetles represent a disproportionate number of the invertebrates on Federal and New England state lists of rare and endangered taxa and prior to 1996 were the focus of disparate data-capture efforts in Connecticut. Among the endangered, threatened, and special concern biota of Connecticut are listed seven species of tiger beetle (Anon 1998), representing 50% of the Connecticut cicindeline fauna. Tiger beetles are one of the best collected groups of invertebrates, providing an excellent historical record and are thus a superb focal taxon for the conservation of their unique habitats. This report is a summary of Connecticut Cicindela species' conservation status that should allow more informed management of this taxon and its associated habitats. Prior to the 1996 survey, knowledge of tiger beetle species of conservation concern was limited to the three state listed species at that time (C. dorsalis, C. puritana, and C. rufiventris). Of Connecticut's 14 known tiger beetle species, after surveys were complete in 1996, five were estimated to be of no immediate conservation concern, the other nine include species that are extirpated, federally protected, rare to very rare and/or apparently declining. Among these nine, there are three species considered extirpated: Cicindela dorsalis dorsalis an inhabitant of broad, sandy, marine beaches (federally endangered), Cicindela limbalis (known from only 3 specimens in CT) and Cicindela purpurea (known from 17-31 populations in CT and 102 specimens, and last recorded in 1973). One of the remaining six species, Cicindela puritana, is a federally protected species still extant in Connecticut. The status of this species is being monitored by P. Nothnagle and the CT branch of The Nature Conservancy, thus it is not included in this project. The remaining five species were the focus of the 1997-1998 research, although some effort to relocate Cicindela purpurea was made. The highest priority of the 1997-1998 project was to assess the status of C. formosa which was feared to have been recently extirpated or experiencing a decline towards local extinction in Connecticut. Among the remaining four species, the state listed C. rufiventris was found to be more common than expected as a result of the 1996 research and was determined to be the lowest priority among the nine species of immediate conservation concern. Cicindela hirticollis, Cicindela tranquebarica and Cicindela marginata, were the three remaining species of conservation concern. All three of these species were documented during 1996 but they were each known from fewer than six populations since 1980 and were thus thought to be either very rare or declining, or both. In Rhode Island, it had been suggested that the majority of that state's cicindeline fauna was in greater danger than previously believed, as described in the following excerpt from a letter written on 22 November 1995 by Chris Raithel of the RI Fish and Wildlife to Tom French of the Natural Heritage and Endangered Species Program of MA.   "-several of the tiger beetle species revealed themselves to be conservation priorities once we compared the fauna across the state. C. dorsalis and patruela are certainly gone. C. limbalis, tranquebarica, purpurea, rufiventris hentzii and nearly formosa exist as single populations on "islands". C. duodecimguttata and marginata are known from less than a handful of sites each, with extensive searching, and even hirticollis has unquestionably declined (or is at least localized enough to worry about). The question is not whether the Tiger Beetles belong on our list, but whether they deserve federal attention. Like the herps and some other groups, there is a lot of attention for some few species while the rest of the fauna is arguably not in much better shape. I suspect that a modern survey of southern New England would reveal a much broader pattern of decline than is presently suspected, as the existing information in most cases is quite old." -22 November 1995, Chris Raithel.   It may be that cicindelines in general are predisposed towards endangerment. This could be due to the tight associations of many tiger beetle species with ephemeral, patchy, disclimax communities dependent on periodic disturbances such as fire, erosion or flooding. These habitats are generally rare in Connecticut and are often overlooked as worthy targets for conservation action. Tiger beetles may prove valuable to the protection of such habitats. Many tiger beetle habitats are characterized by sparse vegetation cover, and although botanically often unremarkable, such habitat is frequently home to a rich assemblage of invertebrates. There are indications that many of these invertebrate taxa flourish in such habitats--therevid flies, asilid flies, sphecid wasps, mutilid wasps and anthicid beetles, in addition to cicindelines, co-occur in these sparse-cover habitats. Given that a large proportion of tiger beetles are presently perceived as imperiled, it is not improbable that these associated invertebrate faunas will contain species of high conservation status. Additionally, historic records of tiger beetle populations that are currently extinct may be used to identify suitable sites for future re-establishments, sites inhabited by other rare taxa, or simply to identify valuable sites to maintain for natural re-establishment. This report provides recommended state Heritage Program rankings for CT tiger beetles species. Because tiger beetles have been more thoroughly sampled over time and space by amateur and professional collectors than most invertebrate groups, abundance data obtained from museum specimens can be more reliably translated into S-ranks. Pre-1996 state rankings for Connecticut tiger beetles were incomplete, uninformative, or unreliable (J. O'Donnell, pers. comm.). Methods Databasing of Collections: I captured data from 16 collections, that, combined with data obtained by other means, totaled 1,632 Connecticut tiger beetle specimens databased. During 1997-1998 six collections not examined during the 1996 report were checked for records of rare species (URI, BU, JSC, CJRC, RNFC, DFEC- Table 1), which resulted in the addition of 70 specimen records to the database. The University of Rhode Island (URI), and Boston University collections (BU) lacked Connecticut records. The Jay Shetterly collection (JSC) lacked records of rare species. Table 1. 16 Collections whose specimens' label data have been captured with the number of Cicindela specimens databased. Collection name Coden* No. specimens Christopher J. Raithel Collection CJRC 1 SUNY -Syracuse DFEC 2 Raul N. Ferreira Collection** RNFC 2 Field Museum of Natural History FMNH 3 Michael A. Valenti Collection MAVC 22 University of Michigan Collection UMMZ 36 Thomas D. Schultz Collection TDSC 48 David L. Wagner Collection DLWC 50 American Museum of Natural History AMNH 51 Jay Shetterly Collection JSC 65 Museum of Comparative Zoology MCZC 68 Christopher T. Maier Collection CTMC 99 CT Agricultural Experiment Station CAES 106 Douglas J. Comboni Collection DJCC 244 Peabody Museum of Natural History PMNH 373 Univ. Connecticut Collection, Storrs UCMS 453 *Codens correspond to the first four characters of each specimens' unique specimen code in the database. Coden format follows Arnett et al. (1993). **Only specimens of C. formosa and C. rufiventris were databased from RNFC.   Field Surveys: Six rare species were targeted for field surveys (C. purpurea, C. formosa, C. hirticollis, C. marginata, C. tranquebarica, C. rufiventris). The highest priority target for field surveys was the species C. formosa, and the lowest was C. purpurea. Additional sites not visited during 1996 of the remaining four target species were also surveyed.   Results and Discussion Table 2. 40 surveys for tiger beetles during 1996. Sorted by County. Fairfield Co. Long Hill Trumbull Pequonnock Valley 18 Jul 1996 Fairfield Co. Sherwood PointWestport Sherwood Island State Park 26 Jun 1996 Hartford Co. South Windsor junction of 5 & 291 15 Jun 1996 Litchfield Co. Barkhamsted Pleasant Valley, jctn 318 &181 Farmington River 21 Sep 1996 New Haven Co. Guilford Leetes Island Island Bay,Shell Beach 02 Jul 1996 New Haven Co. Branford Branford, Landfill S. of Tabor Cemetery 19 Aug 1996 New Haven Co. Guilford Guilford Guilford Point beach 02 Jul 1996 New Haven Co. Guilford Leetes Island salt marsh S of road 02 Jul 1996 New Haven Co. Madison Madison Hammonasset Beach St. Prk 02 Jul 1996 New Haven Co. New Haven West Haven Sandy Point (beach-penisula) 26 Jun 1996 New Haven Co. North Haven North Haven St. Casimirs Cemetery 14 Aug 1996 New Haven Co. North Haven North Haven Valley Service Rd. 14 Aug 1996 New Haven Co. North Haven North Haven W of Wharton Brook State Park 14 Aug 1996 New London Co Voluntown Voluntown North shore, west end Beach Pond 23 Jun 1996 New London Co. Voluntown Voluntown Pachaug State Forest 23 Jun 1996 Tolland Co. Mansfield 0.5kmN Eagleville sand & gravel pit, nr Rt 32 02 Jun 1996 Tolland Co. Mansfield S of Eagleville gravel pit W of Rt 32 02 Jun 1996 Tolland Co. Mansfield Chaffeeville @ bridge over Fenton Riv. 19 May 1996 Tolland Co. Mansfield Chaffeeville Gravelly-sandy hillside 22 Aug 1996 Tolland Co. Mansfield Chaffeeville just S of bridge over Fenton Riv. 19 May 1996 Tolland Co. Mansfield Chaffeeville nr. Fenton Riv., Chaffeeville rd. 19 May 1996 Tolland Co. Mansfield Chaffeeville nr. Fenton R. / Chaf Rd. 08 Jun 1996 Tolland Co. Mansfield Mansfield Depot e side Willimantic R. 02 Jun 1996 Tolland Co Mansfield Mansfield Hollow SP beach E of Basset bridge 19 May 1996 Tolland Co. Mansfield Mansfield Hollow SP boat launch @ Basset Bridge 06 Jul 1996 Tolland Co. Mansfield Mansfield Hollow SP N. of Chapin's Pond nr Chaf. rd 17 Jun 1996 Tolland Co. Mansfield off Rt 89 sand / gravel pit 02 Jun 1996 Tolland Co. Mansfield Rt 89, nr SE School open, sandy area near rd 02 Jun 1996 Tolland Co. Mansfield S. Mansfield Depot gravel pit S. of Willimantic River 04 Jul 1996 Tolland Co. Mansfield S. Mansfield Depot gravel pit S. of Willimantic River 06 Jul 1996 Tolland Co. Mansfield S. Mansfield Depot Gravel/sand pit S. of Willimantic R. 12 Jun 1996 Tolland Co Mansfield S. Mansfield Depot sand pit, just S of Willimantic Riv. 04 Jul 1996 Tolland Co. Tolland Tolland Gravel pit E. of Weigold road 15 Jun 1996 Tolland Co. Tolland Tolland Sand/gravel pit E of Weigold Rd 19 Aug 1996 Tolland Co Willington S. Willington gravel pit W of Rt 32 15 Jun 1996 Windham Co. 0.75 km E of Oneco Pine barren 02 Jun 1996 Windham Co. 0.75 km W of Oneco sand / gravel pit 02 Jun 1996 Windham Co. S. Windham sand pit, rt 32 15 Jun 1996 Windham Co. Pomfret N. of Abington Windham-Tolland 4-H Camp 07 May 1996 Windham Co. W. of Killingly town line off Rt 101, sand pit 02 Jun 1996   Table 3. 22 surveys for tiger beetles during 1997-1998. Sorted by County. Fairfield Co. Westport Sherwood Isl. St. Prk. 45.52.750N, 6.40.0E 26 April 1997 Fairfield Co. Westport, Burying Hill Beach 1 August 1998 Fairfield Co. Bethel, West Redding 45753mN, 6285mE gravel pit off Picket's Ridge Rd. 1 August 1998 Fairfield Co. Stamford, Cove Island Beach 1 August 1998 Middlesex Co. East Hampton Hurd State Park ca. 46.00.000mN-45.99.500mN 30 July 1998 Middlesex Co. Middletown,W side CT River 46.00.000mN, 7.03.000mE jctn Freeman Rd& Canal Access rd. old gravel pit 30 July 1998 Middlesex Co. Middletown W side CT River, oppisite Hurd State Park long, clean beach45.99.500mN, 7.03.750mE 30 July 1998 Middlesex Co. Haddam,Beaver Meadow rd gravel hillsides 0.3mi N of jctn w/ Woods Rdand Beaver Meadow, ca. 45.90.750mN, 7.04.500mE 1 August 1998 New Haven Co. North Brandford, 1.5mi N of Northford, red rock quarry on Rt 17 nr Coe Rd. 1 August 1998 New London Co. Groton, Bluff Pt. St. Prk. 45.78.900mN, 7.48.100mE small mud beach 2 August 1997 New London Co. Preston [jctn of rt 12 &2A just east of the Thames riv gravel/sand pit along river north of Happyland south of Norwich State Hospital,46.95.500mN, 7.44.500mE] 15 May 1998 New London Co. Preston [jctn of rt 12 &2A just east of the Thames riv. gravel/sand pit along river north of Happyland, south of Norwich State Hospital,46.95.500mN, 7.44.500mE] 28 May 1998 Tolland Co. Mansfield: Gurleyville, Gurleyville, gravel pit W of Fenton River near jtcn of riv and Gurl. Rd. 4633250mN, 730100mE 9 July 1997 Tolland Co. Mansfield: Gurleyville gravel pit W of Fenton River near jtcn of riv and Gurl. Rd. 4633250mN, 730100mE 17 July 1998 Tolland Co. 3 mi N Mansfield Center Rt 195 old farm road 30 July 1998 Tolland Co. Mansfield: Gurleyville, gravel pit W of Fenton River near jtcn of riv and Gurl. Rd 4633250mN, 730100mE 1 August 1997 Tolland Co. Willington S. Willington gravel pit W of Rt 32 2 August 1997 Tolland Co. Willington dump on rt 74 2 August 1997 Tolland Co. Mansfield: Gurleyville gravel pit W of Fenton River near jtcn of riv and Gurl. Rd. 4633250mN, 730100mE 17 July 1998 Tolland Co. Mansfield Center- Mansfield Hollow, sand pit nr Chapin's pond 46.28.300mN, 7.32.750mE 15 May 1998 Tolland Co. Mansfield, Storrs: Horsebarn hill dirt roads 8 May 1998 Windham Co. nr Rt 6. large gravel pits 46.23.900mN, 7.33.600mE 2 August 1997   Table 4. Sites/populations of Connecticut tiger beetles. N= number of vouchers taken from population, if N=0 then no population was found. A "?" in the Surveyed column indicates that the site description was too ambiguous to be relocated, a "na" indicates the population was unknown in 1996. C. scutellaris and C. duodecimguttata were not targeted for surveys in 1997-1998. C. dorsalis site/population last observation surveyed in 1996/97-98 N Fairfield Jenning's Beach 2 July 1950 no/no This table contains detailed locality data of protected species. For detailed information on these specimens contact the Connecticut Chapter of The Nature Conservancy, 55 High St., Middletown, CT 06457, or the Connecticut Department of Environmental Protection. C. rufiventris site/population last observation surveyed in 1996/97-98 N Long Hill Pequonnock Valley 18 July 1996 yes/no 4 West Redding Gravel Pits 45753N6285E 1 August 1998 no/yes 1 Haddam Beaver Meadow rd, gravel hillsides 29 July 1989 no/yes 0 West Rock State Park 18 July 1986 yes***/no 1 0.2 mi NE Northford S of Rt. 17. gravel hill 28 August 1988 no/yes 1 Branford landfill S. of Tabor Cemetary 19 August 1996 yes/no 1 Mansfield: 0.55 mi N of Fenton R.&Chaf. Rd. 22 August 1996 yes/yes 2 Mansfield Gurlyville W of Fenton River nr jtcn of riv &Gurl.Rd.4633250N730100E 9 July 1997 na/yes 1 Ledyard-Gayles Ferry [rt12nr Thames Riv.] 25 August 1997 na/yes 1 Meriden& Berlin, Cathole Mountain 41°33'55"N 72°48'01"W 1-15 August 1998 na/yes 2 North Branford, 1.5mi N of Northford red rock quarry 1 August 1998 na/yes 1 C. duodecimguttata site/population last observation surveyed in 1996 N Long Hill Pequonnock Valley 1 July 1989 yes 0 Long Hill: Trumbull, Old Mine Pk 30 May 1992 no West Rock State Park 10 May 1987 yes 0 Hamden near Lake Wintergreen 15 September 1987 no** Branford Landfill and Pond 7 May 1988 yes 0 3 mi NE Northford Rt. 17, Mill Pd Tavern 8 May 1988 no Wallingford Rt. 5 opp Wharton Brook SP 24 May 1988 yes 0 West Haven West R. Prk, shoreline of pool 10 August 1995 no Milford Silver Sands SP 14 August 1995 no Voluntown Pachaug St. Forest 17 April 1985 yes 2 Mansfield Depot Williamant. R., edge of R 3 July 1995 yes 1 C. scutellaris site/population last observation surveyed in 1996 N Newton-Botsford 4 September 1989 no Newtown: Sandy Hook 13 June 1993 no Shaker Pond area-Enfield ca. 4 km NE Thompsonville 3 June 1983 no ca. 5 km NNW of Farmington 12 May 1987 no Farmington, along RR tracks to east of Brickyard Road 27 April 1989 no South Windsor: Jctn of 5 & 291 27 April 1990 yes* 0 Woodbury 11 May 1985 no Barkhamsted, Farmington R., nr Rt 181 1 August 1991 yes 6 Cockaponsett St. For., Haddam, nr Beaver Meadow Rd.and Haddam exit off Rt 9 3 September 1989 no [Haddam], Geo. Dudl. Seym. SP 26 September 1990 no North Haven, Silman Road ca. 1 km N. of Conn. Hwy. 22 7 May 1987 no Wallingford. Rt. 5 opp Wharton Brook State Park 24 May 1988 yes 0 North Haven Valley Service Rd. 22 April 1989 yes 1 Mansfield Twp. Chapin's Pd. 6 July 1974 yes 3 Coventry Twp. nr. Eagleville Dam 25 April 1972 no Mansfield Twp Gurleyville 16 September 1972 yes? 2 Mansfield Storrs 24 April 1976 yes? 0 North Windham 15 April 1990 no [Sterling], Oneco 2mi W,sand pit 29 April 1990 yes 1 Eastford Phoenixville, sand pit 9 May 1993 no Oneco railroad bed 10 May 1993 no * the tiger beetle habitat at the junction of Rt 291 and 5 was destroyed by highway development. ** the population of C. duodecimguttata near Lake Wintergreen was reported destroyed due to development of the area. *** This note contains detailed locality data of protected species. For detailed information on these specimens contact the Connecticut Chapter of The Nature Conservancy, 55 High St., Middletown, CT 06457, or the Connecticut Department of Environmental Protection.       Table 5. 15 new specimen records representing 10 newly documented populations of 6 rare species, undocumented prior to the 1997-1998 survey. This table contains detailed locality data of protected species. For detailed information on these specimens contact the Connecticut Chapter of The Nature Conservancy, 55 High St., Middletown, CT 06457, or the Connecticut Department of Environmental Protection.     Populations, Specimens, and Collection Events   Populations: The ideal data required to best manage species for conservation purposes would be constantly updated full census data for each population within a region of interest. Although some conservation projects can obtain such data on population dynamics, few species' distributions are well enough known to begin such detailed investigations. Before single populations can be surveyed and censused there must be a fairly complete data set describing how many populations exist and where they are found. One wouldn't want to fund costly census work on a single population if a dozen other populations existed but remained undocumented. The first step is to map populations in time and space. This macro-view allows conservationists to determine if the total number of populations is declining over time and specifies which habitats are critical to the survival of the species in question. Tiger beetles, which require sandy, open habitats, at least in Connecticut are often restricted to small patches or islands of suitable habitat. Thus it is more tenable to count single populations of tiger beetles than it would be to count populations of a less restricted organism, such as the gypsy moth. The historical data that is available in the form of specimen labels in museum collections are rarely precise enough to allow unambiguous interpretations of population units. For example, 5 specimens labeled "New Haven" all collected on different dates by different people may represent 5 different populations, or they may represent only one population. Even if each specimen was collected at a different site, all 5 sites may fall within the boundaries of a single population. An additional factor that reduces the value of population unit data is that the greater the number of collection events the more variability there is in the interpretation. For example, data consisting of 4 collection events are reliably translated into population units but data consisting of over 100 collection events might be translated into population units differently by different people. It is very difficult to assess population units for this reason. Therefore for the purposes of this report I present population units as ranges from minimum to maximum. A population is a time-free unit, that is, unless the population blinks in and out of existence, it remains a single unit regardless of how many different collections are made from it. And, although many tiger beetle populations may blink in and out of existence, without evidence that they do we should default to the assumption that they do not. For example, a population sampled once a year for 5 years that generates 30 specimens is still a population count of 1, not 5. Specimens: The least ambiguity is obtained if the unit of interest is the specimen. These are simply totaled and no interpretation or external criteria are necessary. In general, one may roughly translate the number of specimens in collections to the commonness or abundance of a given species in the wild. It usually safe to assume that a species represented by large numbers of specimens in collections is easier to encounter in the wild. However, specimen data are plagued by a very high variability. Different collectors use different stop-rules. For example, one collector may collect only a single specimen of a common species, abundant in the area, and yet focus on collecting less common or rare species, thereby creating a collection in which the number of specimens per species is virtually inverse to the abundance in the wild. On the other hand, other collectors might take as many specimens, regardless of species, as they can within a certain time limit--thereby producing a collection in which the number of specimens per species is proportional to the species' abundances in the wild. For this reason specimen data cannot be interpreted as proportional to natural abundances. Collection Events: Because interpreting population units requires criteria that result in data not free from ambiguity, and because specimen units are subject to great variability that prevent their valid use as proxies for population units, a third unit that is intermediate should also be considered--that of the Collection Event. Collection events are defined as a single sample taken at a certain time. For example, 3 specimens taken from site A in 1993 and 5 taken from site A in 1995 would total 8 specimens, 1 population and 2 collection events. By incorporating the element of time, the ambiguity of whether it is one population or two is avoided. Unfortunately, collection events are not always independent events, so these data cannot be interpreted as random samples. For example, a collector might return to the same site repeatedly over a short time interval, or a collector might tell associates about a particular site. In either case there would be numerous collection events but only one independent population sample. Therefore collection events, although more independent than specimens are also not valid proxies for population units. Because each of these units has its advantages and drawbacks all three should be considered, understood, and interpreted accordingly. I have presented data broken out as pre-1980 and post-1980. The year 1980 roughly equates to a point at which tiger beetle populations in the state became of conservation concern. Therefore since 1980 there has been greater effort to locate rare species' populations than prior to 1980. As a result the data pre- 1980 and post 1980 should not be directly compared without consideration of the increase in search effort. This disparity in effort between pre-1980 and post-1980 is easily seen in Figs 1-14, which show graphically the number of collection events over time (from 1850-1996)     Table 6. Population data. Species sorted by rarity based on post-1980 population units. Last 5 species not surveyed in 1997-1998. species <1980 >1980 during 1996-98 total proposed S rank 1. C. limbalis 1-2 0 0 (accidental, non-native) 1-2 SX / SA? 2. C. dorsalis 4 0 0 (extirpated) 4 SX 3. C. purpurea 17-31 0 0 (extirpated) 17-31 SX 4. C. puritana 4-6 2-3 0 (not surveyed) 6-9 S1 5. C. formosa 14-22 2-5 2 16-27 S1 6. C. hirticollis 14 6 2 20 S1 7. C. tranquebarica 23-34 2-7 4 25-41 S1 8. C. marginata 2-3 6 5 8-9 S1 9. C. rufiventris 0 11 9 11 S3 10.C. duodecimguttata 10-13 12-15 3(not surveyed in 97-98) 22-28 S3 11.C. scutellaris 8-11 22-27 6(not surveyed in 97-98) 30-38 S4 12.C. punctulata 43-62 34-36 4(not surveyed in 97-98) 77-98 S5 13.C. sexguttata 48-93 48-58 6(not surveyed in 97-98) 96-151 S5 14.C. repanda 41-50 43-63 13(not surveyed in 97-98) 84-113 S5 Discussion of populations: Extirpated species: There are no post-1980 data for C. dorsalis, C. purpurea, and C.limbalis. Cicindela dorsalis and C. purpurea are assumed extirpated from Connecticut, the last observed records being: 2 July 1950 and 15 May 1973, respectively. Cicindela limbalis is known from 3 specimens in Connecticut: two labeled "New Haven 1866-1867" and one labeled "New Haven 18 Sep 1933". Cicindela limbalis requires clay soils and it is assumed that Connecticut lacks sufficient habitat of this type. The data for this species indicate that although it may have survived in CT for over 60 years, it might be an accidental, non-native species (i.e. a species incapable of sustaining a continual presence within the state under natural circumstances). Cicindela limbalis is known from numerous populations in New Hampshire, Maine and Vermont and at least one from Rhode Island (Block Island). Cicindela formosa. This species, known from five populations since 1975, is the second rarest tiger beetle in CT. Of these five populations one is known to be destroyed (South Windsor: jctn of 5 & 291) and the remaining four have not yielded vouchers since their initial discoveries despite numerous resurvey attempts. Two of these populations were discovered in 1997-1998, confirming, at least that this species remains extant in CT. However, at least two and perhaps three of these four sites were determined to be no longer appropriate for this species due to habitat destruction or alteration.   These two paragraphs contain detailed locality data of protected species. For detailed information on these specimens contact the Connecticut Chapter of The Nature Conservancy, 55 High St., Middletown, CT 06457, or the Connecticut Department of Environmental Protection. Cicindela hirticollis. Although there are six populations of this species recorded since 1980, all of these were (re)surveyed in 1996-1998 and only two of these sites yielded specimens of this species (Table 4). Of the other four sites, Hurd State Park appeared to lack suitable habitat, however, the remaining three sites appeared to contain suitable habitat. Cicindela tranquebarica This species is known from two to seven populations since 1980. The great range in estimated population numbers compared to the other species is due to the geographic pattern of these populations.This paragraph contains detailed locality data of protected species. For detailed information on these specimens contact the Connecticut Chapter of The Nature Conservancy, 55 High St., Middletown, CT 06457, or the Connecticut Department of Environmental Protection. Cicindela marginata. Although this species is known from six populations since 1980 only four and perhaps five of these six appear stable. This paragraph contain detailed locality data of protected species. For detailed information on these specimens contact the Connecticut Chapter of The Nature Conservancy, 55 High St., Middletown, CT 06457, or the Connecticut Department of Environmental Protection. Cicindela rufiventris This species has no pre-1986 records from CT which is remarkable for CT Cicindela, most of which have many records during the 1800s (with the exception of C. marginata). Since the discovery of this species in CT in 1986 new populations have been found at an average rate of almost one per year, and now this species, once considered among the rarest tiger beetles of the state, is among the more common species (albeit the most rare of the common species). This species is not difficult to relocate and populations checked over 10 years apart appear to have not changed. Due to the preference of this species for rocky outcrops and soils it is likely that many tiger beetle collectors overlooked this species and that as more of these habitats are surveyed more populations will be found.     Table 7. Collection Event data. Species sorted by rarity based on post-1980 population units (see Table 6). species <1980 >1980 during 1996-1998 total 1. Cicindela limbalis 2 0 0 (extirpated, non-native?) 2 2. Cicindela dorsalis 4 0 0 (extirpated) 4 3. Cicindela purpurea 60 0 0 (extirpated?) 60 4. Cicindela puritana 14 5 0 (not surveyed) 19 5. Cicindela formosa 21 7 2 28 6. Cicindela hirticollis 17 11 2 28 7. Cicindela tranquebarica 44 18 6 62 8. Cicindela marginata 3 16 5 19 9. Cicindela rufiventris 0 27 11 27 10.C. duodecimguttata 14 26 3(not surveyed in 97-98) 40 11.Cicindela scutellaris 15 37 6(not surveyed in 97-98) 52 12.Cicindela punctulata 71 52 6(not surveyed in 97-98) 123 13.Cicindela sexguttata 142 74 10(not surveyed in 97-98) 216 14.Cicindela repanda 52 96 16(not surveyed in 97-98) 148   Table 8. Specimen data. Species sorted by rarity based on post-1980 population units (see Table 6). species <1980 >1980 during 1996 total 1. Cicindela limbalis 3 0 0 (accidental, non-native) 3 2. Cicindela dorsalis 7 0 0 (extirpated) 7 3. Cicindela purpurea 102 0 0 (extirpated?) 102 4. Cicindela puritana 40 7 0 (not surveyed) 47 5. Cicindela formosa 40 10 2 50 6. Cicindela hirticollis 37 31 4 68 7. Cicindela tranquebarica 79 35 12 114 8. Cicindela marginata 5 56 11 61 9. Cicindela rufiventris 0 64 16 64 10.C. duodecimguttata 23 63 4(not surveyed in 97-98) 86 11.Cicindela scutellaris 30 118 20(not surveyed in 97-98) 148 12.Cicindela punctulata 92 79 13(not surveyed in 97-98) 171 13.Cicindela sexguttata 208 91 13(not surveyed in 97-98) 299 14.Cicindela repanda 104 301 63(not surveyed in 97-98) 405   Table 9. . Phenology, based entirely on data from Connecticut specimens. Species sorted by rarity based on post-1980 population units (see Table 6). Number of collection events per month, March-November. Numbers in bold indicate month with most collection events. M A M J J A S O N C. limbalis 1 C. dorsalis 1 2 1 C. purpurea 3 15 5 2 7 7 2 C. puritana 1 3 8 1 C. formosa 2 4 10 2 5 5 C. hirticollis 1 1 5 3 5 7 C. tranquebarica 11 15 5 1 3 8 1 C. marginata 1 2 11 2 3 C rufiventris 3 12 12 C. duodecimguttata 1 6 8 3 5 8 6 1 C.scutellaris 14 12 7 1 4 11 C.punctulata 5 66 20 16 C sexguttata 27 95 37 22 14 1 C. repanda 16 32 26 16 26 26 4 These phenology data correspond to previous knowledge regarding these species' life histories. There are essentially three patterns: (1) species whose adults are active from spring through summer and fall, (2) spring and fall only with little to no activity during the summer, and (3) summer active extending into the fall. Cicindela punctulata, C. rufiventris, C. marginata, and C. puritana are all type 3 -- active during the summer and early fall but not the spring. Cicindela duodecimguttata, C. repanda, C. sexguttata, and, apparently, C. hirticollis and C. formosa are type 1 -- active from spring through the fall with no obvious drop in numbers during the summer. The true spring-fall active species (type 2) show an obvious drop in numbers during the summer: C. purpurea, C. tranquebarica, and C. scutellaris. The remaining species (C. limbalis, and C. dorsalis) are too poorly known in the state to establish phenologies.   Extinction Probability Statistics Burgman et al. (1995) compare quantitative, repeatable methods to identify declines in populations from specimens stored in scientific collections based on occurrence data. These methods produce a "probability that a species is extinct". They stress that these quantitative estimates of extinction probabilities should be used in conjunction with other information and point out that these estimates are most likely to be useful as screening tools for large numbers of taxa (Burgman et al. 1995). Given the urgency of most conservation efforts and the patchiness of our knowledge, I see these estimates as one additional tool in the arsenal of conservation biologists. If these formulae are applied to large numbers of taxa they may help prioritize conservation efforts by targeting populations in need of surveying. Burgman et al. (1995) state that the use of this approach would improve the rigor and accountability in determining degree of threat and "provide an early warning system for a variety of extinction processes and observational conditions". These methods are particularly well-suited to be employed in conjunction with a corresponding field-survey program that targets populations of concern. For each species, I calculated both the probability of extinction based on data known as of 27 September 1995 (just prior to the 1996 survey) and the current probabilities (Table 10). Table 10. Solow's probability of extinction. Species sorted by rarity based on post-1980 population units (see Table 5). Solow's probabilities based on data known on 27 Sep 1995 (P<1996) are also presented. Probabilities greater than 25% are marked with asterixes. species first occurrence last occurrence (n-1) P<1996 1996 P current P 1. C. limbalis 1866 21 Sep 1933 1 0.4781* 0.4821* 0.4892* 2. C. dorsalis 15 July 1930 02 July 1950 3 0.9713* 0.9726* 0.9747* 3. C. purpurea 12 May 1869 15 May 1973 59 1.0000* 1.0000* 1.0000* 4. C. puritana 13 Aug 1850 17 Aug 1990 18 0.4756* 0.5367* ** 5. C. formosa 28 Aug 1897 18 Jun 1998 25 0.6736* 0.7469* 0.3244* 6. C. hirticollis 15 Jun 1915 26 Jun 1996 25 0.7997* 0.0754 0.0444 7. C. tranquebarica 03 May 1850 14 Aug 1996 57 0.7389* 0.0458 0.1002 8. C. marginata 01 May 1915 27 July 1996 16 0.5808* 0.0328 0.2045 9. C. rufiventris 20 Jun 1986 22 Aug 1996 19 0.9981* 0.1674 0.0058 10.C. duodecimguttata 03 May 1914 19 Aug 1996 39 0.0505 0.0493 ** 11.C. scutellaris 19 Sep 1936 21 Sep 1996 47 0.8024* 0.0128 ** 12.C. punctulata 28 Aug 1897 21 Sep 1996 122 0.1855 0.0200 ** 13.C. sexguttata 06 May 1850 29 Aug 1996 212 0.1399 0.0000 ** 14.C. repanda 30 Aug 1904 27 Sep 1996 144 0.0299 0.0000 **   **current data not available for calculation- expected P close to 1996 value.   Solow's Probability of Extinction (Burgman et al. 1995): P=(Ce/Ct)N P= probability of extinction. Ce= the number of time intervals between the start of observations and the last observation. Ct=the number of time intervals between the start of observations and when attempts to observe cease. N= (total number of observations for a species) -1. This formula assumes that observations are approximately random and are made independently of one another. For this reason I used collection events as observations rather than specimens (because specimens from single collection event are not independent of each other). As the observation period, Ct, increases, and in the absence of new records, the likelihood that the species is extinct increases. The results are conditioned by the number of times the species was previously observed. For example, a common species observed once a month for 100 years but not seen in the last year will have a greater probability of extinction than a rare species observed only once a year for 100 years and not seen in the last year. It should be understood that these statistics do not indicate rarity. The formula provides a probability for a binary option--either the species is extinct within the region of interest or it is not. Thus a very rare species that might be listed as endangered, if extant populations were recently documented, will not yield a high probability of extinction. For this reason the probability of extinction values presented in Table 10 do not entirely correspond to my rankings of rarity based on the number of populations observed since 1980. The most significant change from the values of 1996 is that for C. formosa. The probability of this species being extinct within the state dropped from near 75% to 32%.   The top five species based on extinction probabilities:   Cicindela limbalis, known from 3 specimens and two collection events, is given a probability of extinction near 50%--not 100% as one might expect. This is due to the very low rate of observation, i.e. 1 observation per fifty years. With this rate of observation we would have to wait until past the year 3000 without observations for the P value to approach 95%. Cicindela dorsalis turns up a probability of extinction within the state equal to 97%. This is very close to expected. Based on these data there is a higher chance of finding C. dorsalis in Connecticut than C. purpurea. However, the habitat requirements of C. dorsalis are far better known than those of C. purpurea and all suitable marine beaches in Connecticut have been surveyed, without success, for C. dorsalis. Cicindela purpurea was a commonly found species in the state prior to 1970. Due to the very high rate of observation, (ca. 1 observation for every two years for over 100 years), and the long run of no observations between 1973 and the present, this species' probability of extinction within the state is 100%. The habitat requirements of this species are poorly known, although there is evidence that this species thrives in areas developed for agricultural purposes. Because the amount of land devoted to agriculture in Connecticut has declined steadily since the mid 1800's it may be that this species, though present, has become rare enough to drop below detectable levels (David Wagner, pers. comm.). Certainly modern insect collectors in Connecticut rarely survey agricultural lands. We should not assume that this species is extinct within the state simply because it hasn't been seen for over 23 years -- recall that C. puritana avoided detection for 50 years (1939-1989). A careful survey based on a detailed understanding of this species' habitat requirements should be completed before we can be certain of its status. Cicindela puritana which is known to be not extinct within the state based on data from annual Nature Conservancy funded censuses of the known populations has a 54% probability of being extinct (based on specimen data which ceased being recorded in 1990). If the live-capture/sight-survey data, now held in the Natural Diversity Database, taken between 1990 and 1996 were included, this species would have a very low extinction probability. Cicindela formosa, a species that was known from 19 to 28 populations in Connecticut (Table 6.) but very few since 1963 has a 32% probability of being extinct. Although two populations were documented in 1997-1998, return visits to these sites failed to relocate this species and this P-value accurately reflects the apparently tenuous status of this species.   The remaining species: All the remaining species were documented during the 1996-1998 surveys and thus are known to be not extinct within the state. These species all have a probability of being extinct under 25%. Although several of these species are rare and perhaps declining in number, this formula is not sensitive to rarity or decreases in population number--it only responds to (1) the rate of observations in relation to (2) the length of time since the species was last observed. For example, Cicindela hirticollis has a probability of being extinct of 4.4%; however, if this species had not been observed during 1998 it would have yielded a probability closer to 50%. The fact that one observation of this species was made on 11 June 1998 made the probability of it becoming extinct by 1 August 1998 (2 months later) very small. As a means of comparison, I calculated probabilities of being extinct based on data known on the date 27 Sep 1995 (Table 10). These values represent what would have been the best of our knowledge prior to the 1996 survey. By calculating the probability of extinction based on our knowledge as of 27 September 1995 it becomes clear that most species ranked as high priority survey targets (Table 10). Note that because populations were found in 1996-1998, the following species' probabilities of extinction all changed from over 58% to under 33%: C. hirticollis, C. tranquebarica, C. marginata, C. rufiventris, C. formosa, and C. scutellaris. Because no populations were found (or surveyed, in the case of C. puritana) for the following species, their probabilities of being extinct stayed the same or increased: C. limbalis, C. dorsalis, C. purpurea, C. puritana. Because C. dorsalis, and C. limbalis and probably C. purpurea, can be safely assumed to be truly extinct within the state and C. puritana is known to not be extinct, the only species of top survey priority is C. formosa generosa.   CONCLUSIONS and SUGGESTIONS FOR FUTURE WORK   The tiger beetle fauna of Connecticut has experienced a drop in species richness, having lost three species. If these data reflect actual trends within the state, it is likely that the few most imperiled species will disappear within perhaps the next twenty years. The habitat requirements of these species place them in particular risk because these habitats are both relatively rare within the state, small and thus easily purturbed by natural causes, and favored sites for use and development by society. A cautious interpretation of the data indicate that Cicindela formosa should be considered on the verge of local extinction. Cicindela puritana, C. hirticollis, C. tranquebarica and C. marginata have declined in numbers and appear to be very much imperiled. Cicindela rufiventris, although uncommon, is obviously more common than first thought upon its discovery in the state in 1986, and may even be increasing in numbers. It is impossible at this point to definitively address the cause(s) of decline of these species, although there are at least two major agents of change that either alone or in combination may exert a strong impact. Human alteration of habitat through uses such as the driving of off-road vehicles, or quarrying, and development such as road construction obviously can have an enormous impact. It should also be considered that within Connecticut there has been a steady increase in tree-cover since the early 1800s due to the decline in the amount of land devoted to agriculture. Because virtually all of the state's cicindeline species require open, unshaded habitats, this virtual doubling in the amount of land covered by forest since the early 1800s may have contributed to the decline of some or many of the species. Unfortunately, there are no historic data to tell us what the tiger beetle fauna of Connecticut was like prior to the deforestation that started in the 18th century.   This project leaves some questions unanswered. The most important question is undoubtedly : "Is Cicindela formosa a stable member of the Connecticut fauna?" It is likely that populations exist that were undocumented during the 1996-98 surveys of rare species not yet extinct within the state. Hopefully, this report provides an improved knowledge base with which to manage these species, however, given the high risk status of over half the fauna, future, perhaps periodic, state-wide status surveys are warranted. Given the appeal of tiger beetles to general collectors, the enlistment of amateurs (akin to the CT Butterfly Atlas Project), is likely to turn up many new populations. It should be emphasized that many valuable data were obtained by including tiger beetle larval stages in the surveys. In fact, three newly discovered populations (one of C. marginata, one of C. tranquebarica, and one of C. duodecimguttata) were sampled as larvae only--no adults were found. Future tiger beetle surveyors who fail to collect and identify larvae will be missing a valuable set of data. As part of this project I have completed an on-line identification key to the larvae of CT Cicindela species which should allow all future workers to use larval stages as vouchers for conservation research.     Home | Access the database | Key to adults | Key to larvae | About tiger beetles | Species Checklist | Note re: Identifications | Project Report | References | Acknowledgments Derek S. Sikes Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT 06269 dss95002@uconnvm.uconn.edu