This species was once found mostly in northeastern Mexico and in the Rio Grande Valley of Texas but has expanded its range to include isolated populations in southern Louisiana and most of the eastern third of Texas to the Red River. It is often found in plants and other items transported from the Rio Grande Valley (AmphibiaWeb 2021). Our team did not detect E. campi on NPI and there are no other records, but there are three iNaturalist observations from Mustang Island and two from SPI. There are three museum records from SPI and 184 museum records from the mainland portion of the seven counties. It seems likely that E. campi will eventually show up on NPI or that it is already there and has not yet been reported.

Eleutherodactylus planirostris is native to Cuba, the Bahamas, the Cayman Islands, and the Turk and Caicos Islands (AmphibiaWeb 2021) but is now found in coastal areas around the Gulf of Mexico from Florida to the Yucatan Peninsula of Mexico. There are two museum records from SPI. The same records were posted on iNaturalist: https://www.inaturalist.org/observations/62578832 and https://www.inaturalist.org/observations/62578829). The species has not been recorded in the rest of the study area.

Osteopilus septentrionalis is native to the Bahamas, Cayman Islands and Cuba and has been introduced to other Caribbean Islands, most of Florida, and is spottily distributed in coastal areas from Georgia to Texas (AmphibiaWeb 2021). There is only one record for the species from the study area, an iNaturalist observation from Port Isabel in Cameron County. It was not found on the islands but is included here because of its potential to disrupt invaded ecosystems. It preys on a variety of animals including other amphibians and reptiles. In urban areas in Florida, O. septentrionalis, by predation and competition, has displaced some native frogs including (but not limited to) Hyla cinerea (green treefrogs) and H. squirella (squirrel treefrog). The University of Florida Department of Wildlife Ecology recommends capturing and humanely euthanizing any O. septentrionalis found (Johnson 2017).

During and after rains during warmer months in 2002 and 2003 our team recorded conspicuous choruses of H. cinerea at all calling-frog stations and collected one of the 14 museum specimens from NPI. The historical record indicates that the species is considerably less common in other years. Rabalais (1975) characterized the species as “uncommon.” I located 474 specimens from the mainland portion of the seven counties but only one specimen and nine iNaturalist observations from Mustang Island and no records from South Padre Island. Most observations during this inventory were in emergent wetlands during the breeding season.

The status of this species in southern portion of the study area is unclear. I heard choruses of H. cinerea at The Nature Conservancy’s Southmost Preserve, in Cameron County, in 2002. There are ten museum specimens from Cameron County but no iNaturalist records. There are 34 museum specimens and zero iNaturalist observations from Kenedy County. The newest specimen from Cameron County was collected in 1941, and the newest specimen from Kenedy County was collected in 1976. The only record of any kind from the mainland portion of the counties south of Nueces County is an iNaturalist audio recording made in 2015 near Port Mansfield in Willacy County: https://www.inaturalist.org/observations/1480951. The status of the species in southern Texas deserves further study.

There were no records for H. squirella from NPI until I photographed and audio recorded it during the 2002–2003 surveys. The calls were recorded in emergent wetlands during breeding season (Suppl. material 3; iNaturalist Observation https://www.inaturalist.org/observations/49370292), but I was unable to collect a specimen during those early surveys. There are two museum records from Mustang Island and 28 from South Padre islands but none from NPI. Sixty-six iNaturalist observations from NPI, 55 from Mustang Island, and ten from SPI based on photos and calls, were posted from 2015 to 2020. The species was not mentioned by Rabalais (1975) or by Baker and Rabalais (1978).

There is still only one mainland museum record collected from the study area south of Kleberg County of a specimen taken near Port Mansfield, in Willacy County in 2015. That specimen was also posted to iNaturalist where the collector commented that it was a single individual calling from a roadside ditch. The 28 museum records from Cameron County were all from a small area in the urbanized portion of southern SPI.

Our team did not observe or hear P. clarkii in 2002, but in 2003 I made audio recordings and collected a specimen (TCWC 93884) in temporarily flooded grasslands during drought-ending tropical rains. That specimen remains the only specimen of P. clarkii collected from the South Texas barrier islands, and among the eight research grade iNaturalist observations, the only observation based on a photo (https://www.inaturalist.org/observations/12007443). The species is not known from museum or iNaturalist records from SPI or Mustang islands. Only 18 of the 231 museum specimens from the study area have been collected since 1990. Rabalais (1975) categorized P. clarkii as “possible.” Baker and Rabalais (1978) did not mention it.

There are seven museum specimens of R. berlandieri from NPI and three from Mustang Island residing in the collections I surveyed. I found 584 specimens from the mainland portion of the seven counties. During the 2002–2003 surveys, I collected two specimens and audio-recorded the species at most ephemeral and man-made ponds in the 24 km south of Packery Channel.

I collected one specimen that was particularly noteworthy because of its locality: TCWC 93885 was collected ~ 80.5 km south of the southern end of Park Road 22. That locality is a narrow, arid, and sparsely vegetated part of the island between two wash-over channel depressions that sometimes hold freshwater but are periodically flooded with saline water from the Gulf of Mexico and hypersaline water from the Laguna Madre. Permanent or long-lasting freshwater, usually associated with R. berlandieri, is not present. When I last visited that site in June 2018, the ponds only supported halophytic vegetation around its edges, which were crusted with salt. The survival strategies and dynamics of that population segment deserve further study.

A 2018 iNaturalist observation, based on an audio file, is the first and only record of Rana catesbeiana on NPI (https://www.inaturalist.org/observations/8027494). Other experienced listeners have concurred on the identification of that call, but because of the poor quality of the recording, I cannot confirm with complete certainty that the call is that of R. catesbeiana. The NPI record is from the northern tip of the island, near Packery Channel. The species is relatively common in Aransas, Nueces, and San Patricio counties and much of the United States but uncommon in the counties adjacent to Padre Island. There were 77 museum specimens from the seven counties. The species requires relatively permanent freshwater for its long-lived larvae, therefore its habitat on NPI is probably limited to several small man-made ponds within the urbanized northernmost part of the island and three manmade ponds in the northern part of the Padre Island National Seashore. The lake that occurs near the middle of the island, from ~ 22.5 km south of Packery Channel to ~ 27 km south of Packery Channel is periodically dry but, in some years, might hold water long enough for R. catesbeiana reproduction. There is little to no habitat for the species in the southern 143 km of Padre Island. The species is a non-native and invasive in the western United States, where it competes with and preys on native fauna. Its recent arrival on North Padre and Mustang islands might be considered invasive.

There is one Rana sphenocephala specimen from Padre Island (TNHC 65562), which was also the first and only specimen from Kleberg County (Duran and Hall 2013). There are no other records for the South Texas barrier islands or from the inland portion of the counties adjacent to Padre Island south of Nueces County. There are 176 museum specimens from the mainland counties adjacent to Mustang Island. In 2013, I made an audio recording of R. sphenocephala (Suppl. material 4; iNaturalist observation: https://www.inaturalist.org/observations/314642), which was only the second verifiable record for the islands and the first within PINS. That recording may include R. berlandieri calling at the same time, but that is not clear. The species was not mentioned by either Rabalais (1975) or Baker and Rabalais (1978).

Dorsolateral ridges inset posteriorly at the groin in R. berlandieri usually distinguish it from R. sphenocephala, but dorsolateral ridges of some Rana specimens from NPI are not distinctly inset. We examined many specimens in the field and could never say that any of them were distinctly typical of R. sphenocephala. While morphological differences between the species were not consistently differentiating, their calls are quite different and perhaps better evidence of their occurrence on the island than specimens or photos. The audio file of R. sphenocephala calling may also include R. berlandieri calling at the same time. Hillis (1981) reported on sympatric populations of three ranid species and identified pre-mating isolating mechanisms, including staggered breeding times and habitat.

I did not find verifiable records of Scaphiopus couchii from the South Texas barrier islands. It is mentioned here because Baker and Rabalais (1978) reported that a specimen had been collected “from a freshwater pond area along Park Road 22 north of the entrance to the National Seashore,” but I did not locate that specimen. It may not exist or may have been re-identified. I found 792 S. couchii specimens from the mainland portion of the seven counties.

I located 30 museum specimens from NPI, nine of which were collected during the 2002–2003 surveys. I found five museum specimens from Mustang Island and 235 from the mainland portion of the seven counties. There are 61 iNaturalist observations from NPI, one from Mustang Island and none from South Padre Island. Scaphiopus hurterii was captured 34 times during the 2002–2003 surveys at four study sites, all in flooded grasslands within grassland/wetland matrices. It was audio-recorded at every stop on the calling-frog survey route. The species is conspicuous during and after rains but nearly undetectable by the casual observer during dryer times.

A photo taken by a PINS staff member in 2007 (TNHC 86867) is the only verifiable record of Chelydra serpentina from the South Texas barrier islands, though Baker and Rabalais (1978) reported that a live specimen had washed up on an NPI beach near the Malaquite Visitor Center. The species is not known from the counties adjacent to North and South Padre Island (Dixon 2013). I witnessed a PINS visitor attempting to release a red-eared slider (Trachemys scripta elegans) into the Gulf of Mexico surf in 2007 and suspect C. serpentina may have arrived on the island as a similar misguided rescue attempt. Baker and Rabalais (1978) speculated that freshwater Chelonians might be washed into the Gulf by flooding, then carried out of their native range by longshore currents. There is no evidence that C. serpentina occurs on the islands naturally.

There were no specimens or iNaturalist observations of M. terrapin from the South Texas barrier islands until one juvenile was collected, and another juvenile was photographed in an urbanized part of southernmost South Padre Island in February and March 2019 (Guadiana et al. 2020). The live specimen (photo voucher TNHC 114470; Fig. 5c), which was collected ~200 m west of the forebeach, is now housed in the Gladys Porter Zoo in Brownsville, Texas. The photo of the other specimen (TNHC 114630) appears to have been taken on the wet sand of the forebeach. Those localities are ~ 180 km south of the nearest localities in Corpus Christi Bay. Within the study area, there are 30 museum specimens from estuarine bays and marshes in Aransas, Nueces, and San Patricio counties. There are 80 museum specimens in all of Texas. Salt secreting glands allow M. terrapin to adjust to changing salinity, but there are no verifiable records from the hypersaline Laguna Madre, and it is rarely observed in 100% seawater; I found only one iNaturalist record (https://www.inaturalist.org/observations/21335833) and one museum specimen (TNHC 7421) from the seaward side of a Texas barrier island or peninsula: both were found on the Bolivar Peninsula, one near Rollover Pass and one near the mouth of Galveston Bay. Brackish marshes, usually associated with the species, are mostly absent along the shores of the Upper Laguna Madre and sparse along the shores of the Lower Laguna Madre. The species might enter the Upper Laguna Madre during times of lower salinity, then leave as salinity levels rise, a process characterized as “behavioral osmoregulation” by Dunson and Mazzotti (1989). The form that occurs in this part of the range is generally assigned to the subspecies M. t. littoralis (Texas diamondback terrapin), but the South Padre Island specimens could not be assigned to subspecies based on morphological characteristics described by Ernst and Lovich (2009) (Drew Davis, pers. comm.). Determining the subspecies by genetic or morphological means might provide a clue about how these animals arrived on SPI so far out of their native range.

There were three T. carolina museum records from the mainland portion of the study area and two iNaturalist observations from Mustang Island. The mainland occurrences are individual records separated by decades (Suppl. material 1), which suggests they are introductions, probably released pets. While the study area is on the edge of the range for the species, there is no evidence to suggest that it has established reproducing populations there.

There are no museum specimens of T. ornata from the South Texas barrier islands and no records of any kind from NPI or SPI. Three iNaturalist observations from Port Aransas on Mustang Island have been entered since 2013. Baker and Rabalais (1978) also reported having seen two Terrapene sp. road-killed on Mustang Island. I found 24 museum specimens from the adjacent counties. Box turtles are popular pets. In Texas they may be legally taken from the wild for non-commercial purposes. They often escape or are released back into the wild, far from where they were collected.

Trachemys scripta is easily observed on any visit to the three manmade ponds within the Padre Island National Seashore. Our team trapped 24 pond sliders in hoop traps in those ponds. No other turtle species were observed. Trachemys scripta was photographed but not collected. I located four museum specimens from NPI, two from Mustang Island, and 287 from the adjacent counties. Trachemys scripta from the study area is usually assigned to the subspecies T. s. elegans (red-eared slider), but there are seven iNaturalist observations (several of which appear to be the same individual) identified as T. s. scripta (yellow-bellied slider) from the South Padre Island Birding and Nature Center at the southern tip of SPI. The yellow-bellied slider is native to the eastern half of the US, but both subspecies have established introduced populations all over the US and Europe.

Baker and Rabalais (1978) reported that K. flavescens was common on Mustang Island and that it had been observed on the northern end of PINS, but the only verifiable records for K. flavescens from NPI I found were two iNaturalist observations of road-killed turtles from the northern end of the island entered on 22 and 24 September 2017. The nearest mainland record from those observation is 13.6 km across the Laguna Madre near Oso Bay, and the nearest mainland freshwater habitat for K. flavescens is 7.7 km across the Laguna Madre. Because Baker and Rabalais (1978) reported that the species was “common” on Mustang Island, and because there are seven museum records from Mustang Island since 1960 and 15 iNaturalist observations more recently, it is likely that K. flavescens is reproducing on Mustang Island. The species is not easily trapped in hoop traps, so our methodology might not have been adequate to determine if K. flavescens occurs as a reproducing population on NPI. There are 79 museum specimens from the seven counties and no records from SPI.

Several occurrences of Gopherus berlandieri on the South Texas barrier islands have been recorded, but there is little evidence that suggests the occurrences are natural. One specimen of a G. berlandieri (AMNH 9307) appears to have been collected on the southern end of Padre Island in 1917. Baker and Rabalais (1978) found a G. berlandieri with a painted carapace dead on the road on Mustang Island and a dead tortoise on the beach on NPI. Baker and Rabalais (1978) also reported that a live tortoise had been found just north of the Mansfield Channel, which divides North and South Padre islands; it is not clear whether they observed that animal themselves or if that was an anecdotal record. I did not find those specimens in natural history collections. Judd and Rose (2000) speculated that G. berlandieri populations occur on the barrier islands but offered no evidence of naturally occurring populations. In 2002, Frank Judd told me that he once found a live G. berlandieri on SPI but believed that was a human-aided occurrence and did not believe that tortoises occur on SPI naturally (Frank Judd, pers. comm.).

Little habitat for G. berlandieri is present on the South Texas barrier islands, and I could find only two reports of naturally occurring tortoises on deep sand: Neill (1958; citing pers. comm. with JR Dixon) reported a tortoise from “sand dunes” near Port Isabel on the mainland, and an iNaturalist observation posted in 2021 appears to show tortoise tracks on mainland sand dunes, also near Port Isabel (https://www.inaturalist.org/observations/74145383). While G. berlandieri often occurs on soils with a moderate sand content in the upper soil horizon, it is usually found on moderately clayey or loamy soils (Bury and Smith 1986; Kazmeier et al. 2001). Habitat of G. berlandieri at the Atascosa National Wildlife Refuge, adjacent to the Laguna Madre in Cameron County, was described by Bury and Smith (1986) as “lomas” (clay dunes). There is some evidence that the composition of vegetation on the islands also may not be suitable for tortoises, e.g., Scalise (2011) found cactus in 98% of scat of G. berlandieri that represented 29.8% of their diet. Cacti (mostly Opuntia sp.) occur on the barrier islands but not in the densities usually associated with G. berlandieri (Scalise 2011). Habitat for G. berlandieri appears to be limited on the South Texas barrier island and there is no evidence of reproducing populations, but the re-appearance of single individuals, probably via human-aided transport, is likely to continue to occur.

Ophisaurus attenuatus is one of the most frequently observed reptiles on NPI. I located 17 museum specimens and 1012 iNaturalist observations. During the 2002–2003 surveys, the species was trapped 14 times at all eight study sites between the north end of PINS and the 35-mile-marker. A phenomenon resembling a mass movement of glass lizards was reported on iNaturalist from 24 February to 20 June 2017, when Jon McIntire of Corpus Christi, Texas, reported 938 observations, mostly on Park Road 22 and Bird Island Basin Road on NPI. He reported that the surrounding grasslands had recently been control-burned. The species is less frequently seen on Mustang Island, with five iNaturalist observations and 16 museum specimens. There is only one iNaturalist record and no museum records for the species on SPI; potential observers on SPI rarely venture into the island far from the beach, so the species is probably more common on SPI than that single record might imply. There were 91 museum records from the mainland portion of the seven counties.

Hemidactylus turcicus is an introduced species native to the Mediterranean region. The earliest specimens I found from Texas were collected in Cameron County in 1953 (TNHC 23057–23060 It is now common on and around manmade structures on the South Texas barrier islands and across the southern United States. The only museum specimens of H. turcicus from NPI were collected in 1980 and 1982 (TCWC 93823 and 93845). We did not observe the species during the 2002–2003 surveys, but between 2017 and 2020, 38 iNaturalist observations from NPI were entered; all but two of those were observed in the northernmost residential part of the island. There are four museum specimens and seven iNaturalist observations from Mustang Island. I found 200 museum specimens from the mainland portion of the seven counties.

Holbrookia propinqua is probably the most abundant reptile on the South Texas barrier islands, certainly the most observable. I located 724 museum specimens from NPI and 1916 specimens from the seven counties. During the 2002–2003 surveys, it was trapped 93 times. Another 128 observations were recorded, and hundreds of casual observations by the survey team were not recorded. On NPI, it is most abundant in the back beach/foredunes ecological zone, but it is common on deep dry sand throughout the island. A primary component of H. propinqua habitat is deep sand, which is the primary soil component on the islands and of an ~ 800,000 ha area that extends westward from the Land Cut and includes parts of several counties in southern Texas. That area, commonly known as the Sand Sheet, has been altered by grazing, farming, and invasive, nonnative, grasses, mostly Kleberg bluestem (Dichanthium annulatum) and buffelgrass (Pennisetum ciliare), which has led to its decline on the mainland.

The catalogue of reptiles and amphibians for the collection kept at Texas A&M, Kingsville (TAMUK) contained an entry for H. (lacerata) subcaudalis (TAMUK 1879) collected in 1968 from the “Dunn Ranch” on NPI, but that specimen was missing when I examined the collection in 2002 and missing when the collection was transferred to AMNH in 2005. Prior to the creation of PINS, most of NPI was part of the Dunn Ranch. When PINS staff speak of the Dunn Ranch, they are generally referring to one of several sites where historical ruins remain. According to the collectors of the TAMUK specimen, the locality was probably the Green Hill site, ~ 1.3 km SW of the 25-mile marker (Thomas Shirley, pers. comm.) although it is possible that they were referring to the Black Hills site, ~ 1.5 km southwest of the 10-mile marker. Our team did not observe the lizard during a nonrandom visual encounter search at the southernmost site and did not trap it in a pitfall array ~ 2 km north of the Green Hill site. I conducted walking surveys at the site six times in subsequent years.

The questionable Dunn Ranch specimen was the only specimen of that species from the South Texas barrier islands. Axtell (1998) commented that the locality or identification of that specimen was probably erroneous. The species is known from just across the Laguna Madre in Kleberg County, along the clayey shores of Baffin Bay, but it avoids deep sand, so it is not likely to occur on NPI.

Both H. lacerata and H. subcaudalis have been the focus of a considerable amount of survey work and research following a 2011 ruling by the U. S. Fish and Wildlife Service that a listing of threatened or endangered, pursuant to the Endangered Species Act of 1973, may be warranted.

Rabalais (1975) listed P. cornutum as “possible” for PINS and Baker and Rabalais (1978) reported that it was “common” on Mustang and northern Padre Island in the 1950s and 1960s. I found seven museum specimens from Mustang Island; the most recent of those was 1967. I found one specimen collected on NPI in 1967 but no iNaturalist records. I located 126 museum specimens from the inland portion of the seven counties; of those, 108 had collection dates, and only 15 of the specimens with dates were collected after 1970, and only two were collected since 1987. There are five museum specimens and possibly an obscured iNaturalist record from a small natural island in Corpus Christi Bay. I received a photo of a horned lizard from that island taken in 2013 and have since received other anecdotal accounts of recent observations of P. cornutum on that island. Phrynosoma cornutum has apparently been extirpated from the South Texas barrier islands. It is listed as a threatened species in Texas.

Anolis carolinensis is an abundant fixture of urban backyards and woodlands throughout the southeastern United States. There are no museum records from NPI and only two from Mustang Island but there are iNaturalist records from all urbanized areas of the South Texas barrier islands. Across its range, A. carolinensis is rarely observed in undeveloped areas and is probably absent from undeveloped portions of the islands due to lack of perching structures. I found 84 museum records from the seven counties.

Anolis sagrei is another mostly urban species, native to Cuba and The Bahamas. It was first collected in the United States in Florida in 1935. The first records from Texas were specimens collected in Cameron County in 1986. I found nine museum specimens from the seven counties and 637 iNaturalist observations from the islands since 2009. One hundred nine of those came from NPI, all from the urbanized northern tip of the island.

Our team captured P. obsoletus ten times at five different localities from near the northern boundary of PINS to 56 km down the island, in wetlands, xeric grasslands, and dunes. We collected three specimens. There were seven specimens from NPI in museums and 47 specimens from the seven counties but none from South Padre or Mustang Islands. There is one iNaturalist record for Mustang Island based on a photo taken in 1985 (https://www.inaturalist.org/observations/2578244).

Figure 6. 

Eight reptiles that occur on North Padre Island a Plestiodon obsoletus (great plains skink b Scincella lateralis (little brown skink) c Aspidoscelis sexlineatus (six-lined racerunner) d Coluber constrictor (North American racer) e Heterodon platirhinos (eastern hognose snake) f Lampropeltis triangulum annulata (Tamaulipan milksnake) g Masticophis flagellum (coachwhip) h Nerodia rhombifer (diamondback watersnake).

I located four museum specimens and 31 iNaturalist observations of S. lateralis from NPI and two museum specimens and only one iNaturalist record from Mustang Island. There are no records from SPI. I found 82 museum specimens from the seven counties. Our team captured S. lateralis at seven localities from the northern boundary of PINS to 56 km down the island. All except one of the observations were in moist grasslands/wetlands. The exception was an individual that was trapped at study site 11, which was in sparsely vegetated foredunes.

In earlier versions of their online list, NPS had listed Plestiodon septentrionalis (prairie skink) as a species that occurs on PINS mostly based on a specimen in the NPS vertebrate collection (formerly PAIS 2025; now TCWC 93804) that had been identified as P. septentrionalis but has since been reidentified as S. lateralis. There is no evidence that the P. septentrionalis occurs on the barrier islands but no reason to think that it might not.

Aspidoscelis gularis was classified as “uncommon” by Rabalais (1975). It was not mentioned by Baker and Rabalais (1978). Our team did not observe A. gularis during the 2002–2003 surveys. I found two specimens (AMNH 8157, 8158), collected in 1916, and one iNaturalist record (observed in 2017) from South Padre Island. I found three specimens labelled A. gularis from NPI (UMMZ 54001 and 54004 and ASNHC 235), but upon examination of photos and consultation with the curators, I determined that all three were A. sexlineata. There is one specimen from Mustang Island (TNHC 50473). The Mustang Island specimen is too faded to identify to species, and there is insufficient evidence to dispute the locality, but in my experience, unique localities represented by single specimens are often misidentifications or mislocations. The species is common in the inland portion of the seven counties, where I located 274 museum specimens. In Duran (2004) I reported that I found a museum specimen (AMNH R-168649) 4 km south of the Port Mansfield on South Padre Island, but I later determined that specimen probably came from the mainland side of the Laguna Madre.

Aspidoscelis sexlineata was captured and photographed 46 times. Team members recorded another 16 observations during visual encounter surveys and made many more casual observations. I located 24 specimens in museum collections from NPI, eighteen from Mustang Island, five from South Padre Island, and 145 specimens from the mainland. The species was observed in the back-beach, foredunes, grasslands, emergent wetlands, and dunes.

Trauth (1992) described a subspecies of the six-lined racerunner, the yellow-headed racerunner (A. s. stephansae; originally A. s. stephansi, emended by Trauth 1995), from Kenedy, Willacy Brooks, and Jim Hogg counties. The localities for some specimens Trauth (1992) examined to describe A. s. stephansae were within a sandy ecological zone just across the Laguna Madre from NPI. Trauth (1992) described A. s. stephansae as, on average, smaller (< 70 mm) than A. s. sexlineata (eastern six-lined racerunner), with a distinctive yellow coloration on the face. Some of the specimens that we captured had bright yellow faces extending from the snout to the nape (Fig. 6c). The snout to vent length was < 70 mm for all specimens during the 2002–2003 surveys. Another characteristic by which Trauth and McAllister (1996) distinguished the subspecies was the relative position of horizontal stripes: In contrast to other A. sexlineata, the yellow-headed race purportedly has no vertebral stipe, paravertebral stripes converge just posterior to the rump and extend onto the anterior one fourth of the tail, dorsolateral stripes extend onto the tail, and lateral stripes blend into the bright ventrolateral surface of the tail. Based on limited analysis, I could not determine if the NPI specimens could be assigned to A. s. stephensae.

iNaturalist recognizes the taxon, but no observations of A. s. stephensae have been entered, while 12 specimens have been identified as A. sexlineata within the taxon’s purported range as defined by Trauth (1992). Because identification of the subspecies is based on external morphology and pattern, and apparently no one can identify it based on those characteristics, further analysis is needed to determine if this is a valid taxon.

I found eight museum specimens of Arizona elegans from NPI, six from Mustang Island, and one from SPI (as of 08 October 2020). Another iNaturalist observation from SPI was entered in April 2021 and was deposited into the TNHC collection. I found 36 museum specimens from the seven counties. Since the species is nocturnal and often fossorial, historical records are not good indicators of its relative abundance. During the 2002–2003 surveys, we observed one A. elegans. Between 2013 and 2020, seven iNaturalist observations were entered for Mustang Island and six were entered for NPI. Rabalais (1975) referred to A. elegans on NPI as “fairly common.” Baker and Rabalais (1978) referred to A. elegans as “common” on both NPI and Mustang Island. Glossy snakes in the study area are usually assigned to the subspecies A. e. arenicolus (Texas glossy snake).

I collected one road-killed C. constrictor (TCWC 93867) on Park Road 22, north of PINS where it passed through a flooded grassland/wetland. A team member observed one on the beach ~ 3 km south of the southern end of Park Road 22. Figure 6d is a PINS file photo of a C. constrictor on the beach, on wet sand near the surf. Coluber constrictor is a habitat generalist, but I did not find any mentions of C. constrictor on beaches in the literature. I found three iNaturalist records for C. constrictor on beaches in California, Florida, and Virginia. There are two museum specimens and eleven iNaturalist records from NPI but no records of any kind from Mustang or South Padre islands. There is an iNaturalist observation of a bird in flight carrying a C. constrictor; the photograph was taken on Mustang Island, but it is impossible to know where the bird caught the snake. There were 83 specimens from the mainland portion of the seven counties. Both Dixon (2013) and Werler and Dixon (2000) show the range of the subspecies, C. c. oaxaca (Mexican racer), extending from Mexico northeast along the coast to Aransas County, Texas. The racer I collected (TCWC 93867) was intermediate in key morphometrics between C. c. oaxaca and the C. c. flaviventris (yellow-bellied racer). Museum specimens of both subspecies are catalogued in each of the seven counties adjacent to the South Texas barrier islands. Werler and Dixon (2000) describe C. c. flaviventris as usually having seven supralabial scales and C. c. oaxaca as having eight, and both subspecies will usually have 17 or fewer dorsal scale rows at midbody. The specimen collected had a dark green dorsum and yellow ventrum, most like C. c. flaviventris, but it had six supralabials on one side and eight on the other and 15 dorsal scale rows, so it could not be assigned to subspecies based on those features. Burbrink et al. (2007) performed genetic analysis which appeared to indicate that C. constrictor may be composed of six independently evolving lineages not concordant with most recognized subspecies. No samples within the range of C. c. oaxaca were included in that analysis.

I found one museum record for the islands (TAMUK 5526), a road-killed snake collected on Park Road 22, just north of the PINS entrance station. That specimen is among numerous specimens from the TAMUK collection (later moved to AMNH) that have been lost. I spoke with the collector and former curators and confirmed that the record is legitimate (Donna Shaver, Allan Chaney, pers. comm.). There were 64 museum records from the seven counties: only three of those were from the counties adjacent to Mustang Island. The Nature Conservancy ecologist, Lee Elliott, captured a D. melanurus in the surf near Bob Hall pier on the northern end of NPI and released it in the dunes (Lee Elliott, pers. comm.). Specimens from that part of the range are mostly identified as the subspecies D. m. erebennus (Texas indigo snake), which is a large, highly mobile, and conspicuous species. The absence of anecdotal reports and road-kills may indicate that the snake is an occasional visitor to the island. It is a powerful swimmer that probably crosses the Laguna Madre occasionally and might populate the island if conditions there became preferable, but historical and recent evidence does not indicate that it is or has been a permanent resident.

Our team collected two road-killed specimens (TCWC 93872, TCWC 93873) and captured and examined three others during the 2002–2003 surveys. The road-killed specimens were surrounded by a grassland/wetland matrix. The trapped specimens were all from one location which was in a transition zone between sparsely vegetated foredunes and a more densely vegetated grassland on deep but stable sand. Photos of three of those specimens are the only iNaturalist records for the species from the South Texas barrier islands. There are eight museum specimens from NPI prior to the 2002–2003 surveys but no records of any kind from the South Texas barrier islands after that. Baker and Rabalais (1978) said that H. platirhinos on NPI was “one of the more common snakes on Mustang Island, particularly around the town of Port Aransas, which abounds with toads.” The historical record does not support the Baker and Rabalais (1978) assessment, as there are only three museum specimens and no other type of records from Mustang Island. The last of the Mustang Island specimens was collected in 1991. Whether the apparent decline of H. platirhinos on Mustang Island is related to the previously discussed extirpation of Bufo speciosus (a primary prey item), should be investigated further. I found 15 museum specimens from the mainland portion of the seven counties.

There are no records for Lampropeltis getula from NPI or SPI. One specimen from NPI (UMMZ 224256) had been labelled “L. getulus,” but in consultation with the curator of that collection, I determined that specimen was a mislabeled L. triangulum. Within the study area, two species or subspecies have been recognized: L. g. holbrooki (Stejneger 1903) and L. g. splendida (Baird & Girard 1853). Baker and Rabalais (1978) referred to L. g. splendida as “common” on Mustang Island and “possible” for NPI. Pyron and Burbrink (2009) proposed changes to the taxonomy and distribution boundaries of the Lampropeltis getula complex, which would elevate both subspecies to full species status. Widely followed taxonomic sources such as Crother et al. (2017), The Reptile Database (2021), and iNaturalist (2021) have adopted that taxonomy. Following that arrangement has created some confusion about the correct identification and taxonomy of kingsnakes on Mustang Island and the seven counties. Werler and Dixon (2000) noted that the seven counties adjacent to the South Texas barrier islands and much of central Texas lie within a broad area of intergradation between the subspecies, L. g. splendida and L. g. holbrooki, which led observers and collectors to identify specimens of both species or subspecies in the study area: Thirteen museum specimens were labelled L. getula ssp., 26 were labelled L. holbrooki, and twelve were labelled L. splendida. For this report I retained the subspecies arrangement of Blaney (1977) and used the verbatim museum labels for Table 2. There are 21 iNaturalist observations from Mustang Island identified as L. holbrooki and three identified as L. splendida. On Mustang Island, morphological characteristics of kingsnakes are sometimes more like L. (g.) holbrooki and sometimes more like L. (g.) splendida, but it seems unlikely that there are two independently evolving kingsnake species on the island. Variation in coloration and pattern displayed by kingsnakes on Mustang Island should probably be regarded as phenotypic variations among genetically similar individuals, i.e., morphotypes, not different species.

Our team observed four L. triangulum and took two as specimens (TCWC 93878, 93879) during the 2002–2003 surveys. Those observations were all within the grassland wetland matrix. Following the taxonomic arrangement prevalent at the time (Williams 1978), I labelled those specimens L. t. annulata. Ruane et al. (2014) proposed reducing the 14 previously recognized subspecies of L. triangulum to six species and redefined the distribution of those species. Widely followed taxonomic sources like Crother et al. (2017), iNaturalist (2021), and The Reptile Database (2021) adopted the Ruane et al. (2014) proposals. According to that arrangement, the species that occupied the seven counties would become L. annulata. Chambers and Hillis (2020) questioned the validity of the Ruane et al. (2014) analysis, stating that “over-reliance on the program Bayesian Phylogenetics and Phylogeography (BPP), without adequate consideration of its assumptions and of sampling limitations, resulted in over-splitting of species in this study.” Chambers and Hillis (2020) did not propose a new taxonomic arrangement but demonstrated that the BPP program can be used to support “virtually any geographic partition of samples in this potential continental cline as species.” Of the 14 specimens from NPI in museum databases, eight were originally labelled L. t. annulata, five were labelled L. triangulum ssp. and one was labelled L. t. gentilis. One museum specimen from Mustang Island was labelled L. t. annulata and three were labelled L. triangulum ssp. In Table 2, I retain the verbatim museum labels. On iNaturalist, which is following the Ruane et al. (2014) arrangement, two photos from Mustang Island have been identified as L. gentilis. The nearest mainland iNaturalist observation of L. gentilis is near Somerville, Texas, 323 km from the Mustang Island observations.

Masticophis flagellum was the most observed and most captured snake species during the 2002–2003 surveys. It was trapped 21 times at seven study sites, which spanned 94 km and four ecological zones: grasslands, emergent wetlands, sparsely vegetated foredunes, and dune/swell complexes. from Study Site 1 near Bird Island Basin to Study Site 17, near the Mansfield Channel. I took one specimen (TCWC 93880). There were 25 museum specimens from NPI, thirteen from Mustang Island, and 148 from the seven counties. There are 61 iNaturalist observations from NPI, 25 from Mustang Island, and five from SPI.

Nerodia rhombifer was observed four times and captured once near the man-made pond along the road to Bird Island Basin. There were two museum specimens from NPI and none from Mustang Island. There are four iNaturalist record for NPI, and none from Mustang Island. I found 163 specimens from the mainland portion of the seven counties. Mostly a fish eater and dependent on permanent fresh water, the species is probably only found on the northern end of NPI in and around the several manmade ponds.

Nerodia clarkii is regularly observed along the shorelines of Nueces, Corpus Christi, and Oso bays, and Mustang Island, but an iNaturalist observation of a road-killed specimen, near the northern end of NPI, just south of the Nueces County line, is the only verifiable occurrence of the species on NPI or in Kleberg County (https://www.inaturalist.org/observations/8105046). There are 33 iNaturalist observations and 18 museum records from Mustang Island. Forty-three more museum specimens have been taken around Corpus Christi Bay, in the counties of Nueces, San Patricio, and Aransas. The southern edge of the range of the species is an elastic boundary where the brackish water of Corpus Christi Bay meets the hypersaline water of the Laguna Madre. Nerodia clarkii could enter the upper Laguna Madre when higher precipitation lowers salinity, then retreat into Corpus Christi Bay when salinity rises, an example of “behavioral osmoregulation” (Dunson and Mazzotti, 1989).

There is only one record for Cemophora lineri from NPI or the South Texas barrier islands, an individual I captured in a grassland near the mid-point between the end of Park Road 22 and the Mansfield Channel in 2002 (photo voucher, TNHC 86866; Fig. 7a). Seemingly rare throughout the range, like other fossorial snakes, C. lineri may be more common than the sparse historical record implies; hundreds of square kilometers of suitable habitat within its range are privately-owned and largely inaccessible. It is listed as “threatened” by the state of Texas. Because of its rarity and the unique nature and atypical morphology of this lone barrier island specimen, a more detailed discussion follows.

Figure 7. 

Comparison of Cemophora lineri (Texas scarletsnake) specimen from North Padre Island with C. lineri specimen from San Patricio County a Cemophora lineri (Texas scarletsnake) from North Padre Island b Cemophora lineri from San Patricio County. Note that the colors of the NPI specimen are duller than those of the scarletsnake from the mainland.

Aufenburg (1948) collected the first specimens of what were probably C. lineri on U.S. Naval Air Station, Corpus Christi, on the mainland, near the northern end of NPI. At that time the nearest known occurrence of Cemophora sp. was > 1000 km from Auffenberg’s observations, so it is difficult to fault Aufenburg for stating: “There is no doubt that these snakes were accidentally brought in from Pensacola, Florida, from which station we received much air cargo.” It is not impossible that Auffenberg (1948) was right about the source of those snakes and the specimens were lost before they could be analyzed (Williams et al. 1966, citing personal communication with Auffenberg). Referring to that possibility, Brown (1950) called it “a doubtful species,” but most researchers later concluded that those specimens were probably from southern Texas (Williams et al. 1966; Williams and Wilson 1967).

Williams et al. (1966) found that scutellation and other morphological features of the only two Cemophora specimens from southern Texas (AMNH 75307 [holotype] and BCB 10993 [paratype]) were distinctly different from Cemophora specimens known from the southeastern United States and described a new subspecies, C. coccinea lineri (Texas scarletsnake). During a more extensive review of the genus, Williams and Wilson (1967) confirmed that morphometrics of C. (c.) lineri were different from the two previously recognized subspecies, C. c. coccinea (Florida scarletsnake) and C. c. copei (northern scarletsnake).

Williams et al. (1966) reported that one of the more distinct differences between C. (c.) lineri and the southeastern Cemophora was the significantly higher number of ventral scale rows (VSR) for C. (c.) lineri: They found that the VSR count of 59 specimens of C. c. coccinea was 158–185 (x̄ = 174.0) and VSR count of 180 C. c. copei specimens was 150–180, (x̄ = 165.3). For the two C. (c.) lineri specimens available for the original description, VSR were 188 and 195, (x̄ = 191.5). The comparisons led Williams and Wilson (1967) to hypothesize that C. (c.) lineri was more closely related to the more geographically distant C. c. coccinea than to the nearer C. c. copei, which they attributed to climactic conditions that led to a splitting of C. coccinea during the Pleistocene. Weinell and Austin (2017) proposed elevating the subspecies to C. lineri based on their genetic analysis; that analysis indicated that C. lineri diverged from the C. coccinea in the Pliocene or early Pleistocene and that C. lineri is monophyletic, while C. c. coccinea and C. c. copei are paraphyletic. Crother et al. (2017) and other taxonomic sources adopted that taxonomy.

In addition to their genetic work, Weinell and Austin (2017) performed a phenotypic analysis of five southern Texas specimens (including the NPI specimen: TNHC 86866) and the two specimens used by Williams et al. (1966) in the original description. In rough concurrence with Williams et al. (1966), they found that C. lineri differed most distinctly from C. coccinea in the number of VSR (178–195; x̄ = 186.1). The results of that analysis provided evidence that the forms are morphologically dissimilar as well as genetically distant from the southeastern Cemophora. In reporting the VSR count, Weinell and Austin (2017) included the VSR count for the NPI specimen (178). Excluding the NPI specimen from that analysis would leave its VSR count outside the range of variation of other known specimens of C. lineri but within the range of variation and near the mean for C. c. coccinea. There are no published data on C. lineri morphometrics which includes more specimens, but in unpublished notes, The Nature Conservancy zoologist, John Karges, analyzed eleven morphometric features of the ten C. lineri specimens residing in natural history collections in the late 1970s (John Karges, pers. comm.): the VSR count of the NPI specimen is still well outside of the range of variation in VSR in that larger dataset (183–197, x̄ = 188.5).

As it is with other reptiles and amphibians on NPI, the colors of the C. lineri I photographed on NPI are duller than those of other C. lineri I have observed in the study area on the mainland. Figure 7a is the snake I captured on NPI and Figure 7b is a C. lineri I photographed in San Patricio County.

During the 2002–2003 surveys, our team collected two road-killed P. emoryi of the four NPI specimens in museum databases (TNHC 85143, TCWC 93868). The road-kills were surrounded by inundated emergent wetlands within a grassland/wetland matrix. Our team captured and photographed three more specimens at two study sites; both sites were on the edge of an inundated emergent wetland surrounded by an extensive grassland/wetland matrix. Between 2015 and 2020, nineteen iNaturalist observations from NPI and 19 from Mustang Island were entered. Rabalais (1975) characterized the species as “possible.” It was not mentioned by Baker and Rabalais (1978). I found 147 museum specimens from mainland portion of the seven counties.

Figure 8. 

Eight reptiles that occur on North Padre Island a Pantherophis emoryi (Great Plains ratsnake) b Tantilla gracilis (flathead snake) c Thamnophis marcianus (checkered garter snake) d Thamnophis proximus (western ribbon snake) e Tropidoclonion lineatum (lined snake) from South Bird Island, just offshore of NPI f Rena dulcis (Texas threadsnake; PINS file photo) g Crotalus atrox (western diamond-backed rattlesnake) h Sistrurus tergeminus (western massasauga).

I found no museum specimens of P. catenifer from the barrier islands, but an iNaturalist record of a 2006 observation (https://www.inaturalist.org/observations/2580986) was entered in 2016. Subsequently, three iNaturalist observations from the northernmost, urbanized part of the island, were entered for snakes observed in 2015, 2019, and 2020. Rabalais (1975) referred to P. catenifer as “uncommon.” Given the sparse historical record, P. catenifer, a large-bodied snake that is probably capable of swimming across the Laguna Madre, may occur only as a vagrant on the islands, but further study might reveal that it occurs as a reproducing population. I located 74 museum specimens from the inland portion of the seven counties. In southern Texas, specimens are generally identified as the subspecies, P. c. sayi (bullsnake).

There is one 1982 museum specimen of S. dekayi from NPI, which is the only specimen from the South Texas barrier islands. Our team did not detect the species during the 2002–2003 surveys. There are now 13 iNaturalist records from NPI, one from Mustang Island, and one from SPI. I found 188 museum specimens from the inland portion of the seven counties.

Tantilla gracilis is largely fossorial, thus rarely observed, so they are probably more common on NPI than the three museum specimens and nine iNaturalist observations suggest. During the 2002–2003 surveys, our team captured three T. gracilis (TCWC 93900) at three study sites, all within the grassland/wetland matrix. There are no records for Mustang or South Padre islands. I found 52 T. gracilis museum specimens from the inland portion of the seven counties.

Thamnophis marcianus is common and conspicuous across most of its range. Our team observed this species three times. One road-killed specimen (TCWC 93901) was collected. I found eight museum specimens from NPI, fourteen from Mustang Island, one from Harbor Island (near Port Aransas), and 327 records from the inland portion of the seven counties. There are ten iNaturalist observations for Mustang Island and 33 from NPI. All observations are closely associated with wetlands and ponds within the grassland/wetland matrix. Its distribution is probably limited by the availability of freshwater to the northern 27 km of the island.

During the 2002–2003 survey our team captured Thamnophis proximus four times and collected six road-killed specimens (TCWC 93902–93908). All observations were closely associated with emergent wetlands within the grassland/wetland matrix. Baker and Rabalais (1978) reported that the species was more common on Mustang Island than on NPI, but museum and iNaturalist observations indicate that the species is and has been more common on NPI. I found 17 museum records and 413 iNaturalist observations from NPI. There was one museum specimen and 41 iNaturalist observations from Mustang Island. The distribution of this semi-aquatic fish-eating species on NPI is probably limited to the northern 27 km of the island by the availability of freshwater. Specimens that occur in the area are usually identified as the subspecies, T. p. orarius (Gulf Coast ribbonsnake).

A single museum specimen of Tropidoclonion lineatum (AMNH 171739) was collected in 1980 on South Bird Island, a 10.9 ha island in the Laguna Madre, ~ 75 m from NPI and ~ 2.0 km northwest of the western end of Bird Island Basin Road. There are no other records of T. lineatum from the South Texas barrier islands or from the seven adjacent counties (Werler and Dixon 2000; Dixon 2013). This locality is ~ 138 km from the nearest locality to the northeast in Calhoun County and ~ 156 km from the nearest locality to the northwest in Duval County. I could not locate the collectors (Richard R. Schmidt and C. Byrd) for comment. The species is spottily distributed from the north-central United States to south-central Texas.

Rena dulcis is known from the South Texas barrier islands from one photo taken by PINS staff in 2010. That snake was uncovered ~ 46 cm underground while digging a hole for a fence post. There is an iNaturalist record from the north end of Mustang Island, but the photo submitted with that record is not detailed enough to determine with certainty if the snake is R. dulcis or the non-native Indotyphlops braminus (Brahminy threadsnake; https://www.inaturalist.org/observations/1270983). There are 101 museum records from the seven counties. Rena dulcis is fossorial and not easily observed, so it is probably more common on NPI (and possibly Mustang Island) than the sparse historical record implies.

Baker and Rabalais (1978) reported that the western diamondback rattlesnake was “very common” on NPI and Mustang islands from the foredunes through the vegetated barrier flats. We received a few anecdotal reports and one photo during the 2002–2003 surveys, but our team did not observe C. atrox. I found three museum specimens and seven iNaturalist records from the northern end of NPI, seven museum specimens and 28 iNaturalist records from Mustang Island, one iNaturalist records from SPI, and 358 museum specimens from the seven counties. While it is clearly not uncommon, the historical record does not support the Baker and Rabalais (1978) contention that C. atrox is “very common.” I have avoided the use of subjective classifications of abundance, but historical records for C. atrox are sparse compared to species such as Masticophis flagellum or Holbrookia propinqua. Observing this large venomous reptile may be frightening and memorable, and observations may be reported and repeated out of proportion to its abundance.

During the 2002–2003 surveys, out team captured five S. tergeminus. I located 19 museum specimens and three iNaturalist records from NPI but no records from Mustang or South Padre islands. There are ten museum specimens and one iNaturalist observation from the mainland portion of the seven counties. Baker and Rabalais (1978) reported that S. catenatus (S. tergeminus) was common on NPI but unknown from Mustang Island. At the time of the 2002–2003 surveys, two subspecies, S. t. tergeminus and S. t. edwardsi were recognized, but the snakes observed were morphologically intermediate between those subspecies, per the diagnostic characteristics described by Werler and Dixon (2000) and could not be assigned to subspecies. Kubatko et al. (2011) and Ryberg et al. (2015) report that the genetic distance between S. t. tergeminus and S. t. edwardsi is not significant. Currently most authoritative sources do not follow the subspecies arrangement.