Understanding species' life history traits is important for developing and implementing sound management practices and assessing species conservation status (Purvis et al. 2000; Litzgus and Mousseau 2006; Sirsi et al. 2017). Reproductive biology is a vital aspect of a species' ecology, including key life history traits, such as clutch size, offspring size, and age and size at sexual maturity. These traits are related to body condition, available resources, and animal energy allocation and can be influenced by environmental factors and the quality of habitat (Lindeman 1996; Iverson et al. 1997; Litzgus and Mousseau 2006). Freshwater turtles are long-lived, latematuring vertebrates characterized by high adult survivorship and low egg/hatchling survivorship (Iverson 1991; Shine and Iverson 1995; Vitt and Caldwell 2014). Information on reproductive traits, such as sexual maturity and fecundity, has been essential for the development of population models for long-lived taxa, such as turtles (Crouse et al. 1987; Congdon et al. 1993; Heppell 1998). Furthermore, studying nesting seasonality can assist in designing protection programs during the months of the highest vulnerability for nesting females, eggs, and hatchling turtles (Bourjea et al. 2015; Sirsi et al. 2017).

Freshwater turtles of the genus Pseudemys are distributed mainly throughout the southeastern United States; however, they can be found in the east-central and southwestern United States as well as northern Mexico (Ernst and Lovich 2009). Of 8 extant species, reproductive biology has been extensively studied in 6 species, while the Alabama red-bellied cooter (P. alabamensis) and Rio Grande cooter (P. gorzugi) remain the least-studied species within the genus (Ernst and Lovich 2009; Lovich and Ennen 2013). The Rio Grande cooter represents the westernmost Pseudemys species and is found in the Rio Grande, Pecos, and Devils rivers of Texas (Bailey et al. 2014), the Pecos, Black, and Delaware rivers of New Mexico (Degenhardt et al. 1996), and the Río Bravo del Norte drainage, including the Salado, Sabrina, and San Juan rivers of Mexico (Legler and Vogt 2013; Pierce et al. 2016). Throughout their range, overall densities of P. gorzugi are low, although they can be locally abundant in some areas (Bailey et al. 2014; Mali et al. 2018). Historically, populations of the Rio Grande cooters have faced threats due to the commercial pet trade (Bailey et al. 2008; Dixon 2013; Mali et al. 2014b). However, recently identified threats include pollution, habitat degradation, and modification of rivers (reviewed in Pierce et al. 2016). In addition, shooting of turtles has been known to occur in New Mexico (reviewed in Pierce et al. 2016), and Waldon et al. (2017) found a female P. gorzugi with an ingested fishhook.

Because of the narrow range of occurrence and lack of data on the species' habitat requirements, P. goruzgi is listed as near threatened by the International Union for Conservation of Nature (2010). In New Mexico and Mexico, P. gorzugi is listed as threatened (New Mexico Department of Game and Fish 2018; Secretaría de Medio Ambiente y Recursos Naturales 2010), while in Texas, the species is of greatest conservation need (Texas Parks and Wildlife Department 2012). Currently, P. gorzugi is under review by the US Fish and Wildlife Service for potential listing as a federally protected species under the US Endangered Species Act (Adkins Giese et al. 2012; Pierce et al. 2016).

Although knowledge on P. gorzugi ecology has greatly increased in the past decade, information on some aspects of its natural history, such as reproductive ecology, are still lacking (Lovich et al. 2016; Pierce et al. 2016). Recent investigations have focused on the few existing areas where P. gorzugi still occurs in relatively high densities in New Mexico (Mali et al. 2018) and Texas (MacLaren et al. 2017). Nonetheless, aspects of P. gorzugi reproductive biology, such as courtship and nesting behavior, size at sexual maturity, and clutch size and frequency, are not known (Degenhardt et al. 1996; Ernst and Lovich 2009; Lovich et al. 2016). In New Mexico, only 3 gravid females have been reported to date (Degenhardt et al. 1996; Lovich et al. 2016; Letter et al. 2017). The nesting season in New Mexico is assumed to last from May to June and possibly extend into July, but no nesting female has been observed in the wild (Degenhardt et al. 1996; Lovich et al. 2016; Pierce et al. 2016).

In 2018, we conducted a capture–mark–recapture study of P. gorzugi on the Black River, New Mexico, as a part of the ongoing long-term monitoring program that focuses on survivorship, somatic growth rate, and general ecology of the species. In addition to standard capture data collection (i.e., morphometrics), we focused on studying P. gorzugi reproductive biology using X-ray and ultrasound technology. Specifically, we sought to address 4 aspects of the species' reproductive biology: 1) female size at maturity, 2) clutch size and mean egg size, 3) correlation between reproductive output and female size, and 4) reproductive seasonality.

METHODS

Study Site. — Our study site was located on the Black River, an ~87-km-long tributary of the Pecos River in Eddy County, New Mexico (Fig. 1). We focused on 2 accessible stretches of the river. The first stretch was ~1500 m long, located upstream near the Black River headwaters. This section is managed by the Bureau Land Management (BLM) for public recreation activities, such as fishing. The second stretch was ~3000 m long and was located about 30 km downstream from the first stretch. Most of the downstream portion of the river is located within natural gas and oil industry sites as well as private properties. Within these 2 stretches, the study areas were further divided into 7 sections. We divided the upstream stretch into equal segments of 750 m, while the downstream stretch was divided into 5 segments of 400–750 m, depending on river accessibility, physical features of the river, and landownership.

View Figure

• Download as Powerpoint
• Download Figure

Field Data Collection. — We captured turtles primarily using traditional hoop net traps. Each of the 7 segments was trapped at different weeks from mid-May to mid-August 2018. Given the elusive nature of these turtles, Mali et al. (2018) found that only high trap effort (i.e., 400 trap days per river kilometer) can yield sufficient capture–recapture data for P. gorzugi. Therefore, we set 20–50 traditional hoop net traps, depending on the length of the stretch, for 6 days, resulting in 120–300 trap days per river segment. Traps were 76 cm in diameter, fiberglass, single opening, single throated, and wide mouth with 2.5-cm mesh size and 4 hoops per net (Memphis Net and Twine Company, Memphis, TN). Two wooden poles were inserted between the first hoop and the fourth hoop to stretch the nets. We put a floating device in each trap to prevent captured turtles from drowning. We placed the traps in the river about 5 m apart and tied them to sturdy riparian vegetation. Traps were primarily baited with canned sardines, and the bait was replaced every other day (Mirabal et al. 2018). Bait was placed in plastic food containers with drilled holes, allowing turtles to detect the scent without being able to consume the bait. We additionally placed a piece of romaine lettuce leaf in traps to help lure larger individuals. In addition, we occasionally captured turtles throughout the duration of the study via snorkeling in a 1530-m2 pond located on private property near the Black River headwaters and Cottonwood Day Use Area, a stretch of the Black River managed by the BLM.

On capture, we recorded standard measurements: straight midline plastron length (PL) and body depth using dial or tree calipers and body mass. After taking measurements, we marked the turtles using 1 of 3 different marking techniques, depending on the size of turtles: marginal scute notching (Cagle 1939), inserting passive integrated transponder tags (turtles < 100 mm PL excluding hatchlings; Buhlmann and Tuberville 1998), or toe clipping (newly hatched turtles only), as suggested by Cagle (1939). Sex of turtles was determined using secondary sex characteristics, by which males possess longer front claws and the cloaca extends past the carapace (Ernst 1990). The majority of captured turtles lacking the secondary sex characteristics had a PL < 100 mm; therefore, for the purpose of this study, we released all juvenile turtles at their capture locations immediately after we marked them and took measurements. Adult turtles were retained for X-radiography and also released at point of capture (see below).

X-Radiograph and Ultrasound. — We transported female turtles with PL _ 100 mm to the Desert Willow Veterinary Services and Wildlife Rehabilitation Center, Carlsbad, New Mexico, where they were X-radiographed. The distance from our trapping sites to the rehabilitation center varied from 25 to 55 km. We used a portable MinXray 308 (MinXray Inc, Northbrook, IL) at 70-kV peak, 20 mA, and 0.06 sec for individuals with a body depth of less than 65 mm; individuals with a body depth of more than 65 mm were X-rayed at 70-kV peak, 20 mA, and 0.16 sec. If shelled eggs were present on an X-ray, we counted the number of eggs directly from the digital radiograph images. We measured the egg width to the nearest 0.01 mm using VetView software (University of Georgia, Athens). If shelled eggs were absent from female X-ray images, we then examined the turtles using Mindray Digi Prince DP-6600 ultrasound (Mindray Medical International Ltd, Shenzhen, China) to assess the development of oviductal follicles. To ultrasound a turtle, we pulled the turtle's hind leg and placed the ultrasound probe in the inguinal region (Kuchling 1989; Shelby et al. 2000). We examined both left and right sides of the turtle's inguinal region. Ultrasound transmission gel was used to amplify the ultrasonic signal. Recaptured individuals caught within a 1-wk period after the first X-ray (2-wk period in the case of the pond turtles) were not X-rayed or scanned with an ultrasound again for the safety of the turtles. After X-ray and ultrasound examinations, which took approximately 3 hrs, all turtles were returned to their respective capture locations.

Statistical Analyses. — The relationships among clutch size, egg size, and female size were evaluated using simple linear regression and reported with 95% confidence intervals for slopes, in which all of the parameters were log transformed to improve data interpretation and comparisons with other studies (King 2000; Ryan and Lindeman 2007). All parameters were normally distributed as indicated by Shapiro-Wilk tests. We used PL as the indicator of turtle size for all analyses (Congdon and Gibbons 1985; Iverson 2001; Mali et al. 2014a; Kern et al. 2015). Because of the orientation of eggs on the Xradiograph, it was difficult to obtain the length of eggs; therefore, we used egg width to represent egg size. In addition, egg width appears to be a more reliable indicator to predict the size of hatchlings in comparison to other parameters, such as egg mass and egg length (Wilkinson and Gibbons 2005). To demonstrate the seasonality of egg development, we used a bar chart to show the number of females with shelled eggs, egg follicles, or no egg development features per trapping period. For this evaluation, we used a conservative subset of females by including only females that were equal to or larger than the smallest female with follicles to ensure proportions included only those at assumed sexual maturity. We performed statistical analyses using R (version 3.4.2, R Core Team 2017).

RESULTS

We captured 165 female P. gorzugi, in addition to 132 males and 149 juveniles. We X-rayed 152 unique females. Each individual was X-rayed once, except for 3 females at the pond site that we recaptured approximately 1–2 mo after their first capture. The number of female P. gorzugi captured per day ranged from 1 to 14 individuals, with total captures of 2–36 P. gorzugi per day.

Because of the small size of some females, we could not ultrasound every individual. The smallest female that we were able to ultrasound had a PL of 173 mm, although 3 individuals with PLs of 175–177 mm and 1 individual with a PL of 228 mm appeared to have small gaps between the carapace and the plastron in the upper thigh area, which made it impossible to insert the ultrasound probe. Overall, we were not able to assess the reproductive status of 69 females (118–228 mm PL) using the ultrasound due to size constraints. Of all females examined using both radiograph and ultrasound (n = 86; Fig. 2), 16 contained shelled eggs (196–251 mm PL), 28 contained oviductal follicles (185–244 mm PL), and 42 did not contain any egg or follicle development structures (173–248 mm PL; Table 1). The smallest turtle containing egg follicles had a PL of 185 mm, and the PL of the smallest turtle containing shelled eggs was 196 mm.

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Reproductive outputs and body size of X-rayed female Rio Grande cooters (Pseudemys gorzugi) captured on the Black River, New Mexico, from May to August 2018. The individuals that produced a second clutch were considered as new individuals (n = 3).

View Fullscreen

Clutch size ranged from 5 to 14 eggs, with a mean of 9.3 eggs (n = 15). We could not obtain a complete count of clutch size from 1 individual with shelled eggs due to a thick layer of tissues, which was impenetrable by the Xray. This turtle was excluded from statistical analyses. Egg width ranged from 24.8 to 34.3 mm, with a mean of 30.1 mm (n = 144). There were significant positive linear relationships of both clutch size and mean egg width with female PL ( p = 0.036, r2 = 0.278, and p = 0.004, r2 = 0.457, respectively; Fig. 3a–b). However, the linear relationship between clutch size and mean egg width was not significant (p = 0.445, r2 = 0.042; Fig. 3c).

View Figure

• Download as Powerpoint
• Download Figure

Of all females with follicles or shelled eggs caught during this study (n = 44), 86% were caught at the end of May and mid-June (Fig. 4). We encountered the first female with shelled eggs on 15 May, the first day of our survey, while the last female with shelled eggs was encountered on 21 June. Females with oviductal follicles were found throughout the length of the study, but the number noticeably declined in July and August. The size of follicles ranged from 7.6 to 25.4 mm in diameter; according to Iverson (2001), follicles in P. concinna with diameters of > 10 mm and > 15 mm were considered mature and ovulatory sized, respectively. In June and July, we recaptured 3 individuals with PLs of 230, 231, and 237 mm containing follicles at the pond site. These individuals were first captured in May and contained shelled eggs. Given that we did not assess reproductive status prior to mid-May in 2018, we went back to one of our study sites from 10 to 15 March to 2019 to assess reproductive status. The attempt was largely unsuccessful due to the unusually cold weather conditions. We captured only 1 female (PL = 278 mm), but the individual contained oviductal follicles.

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Our study focused on one of the very few areas within the Pecos River system where P. gorzugi is known to occur at relatively high densities (Mali et al. 2018). Records of P. gorzugi reproduction in New Mexico were restricted to reports of 3 gravid females (Degenhardt et al. 1996; Lovich et al. 2016; Letter et al. 2017), and our findings on seasonality supported these previous observations. In 3 previous studies, gravid females containing shelled eggs were found from the end of May through June (Degenhardt et al. 1996; Lovich et al. 2016; Letter et al. 2017). Lovich et al. (2016) speculated that the nesting season of the species could potentially extend into July. In the present study, we found the first gravid female with shelled eggs on 15 May. Therefore, the ovulation cycle likely started between late April and the beginning of May or as early as mid-March. The last female with shelled eggs was found on 21 June, while 6 females carrying oviductal follicles were found in July and August, further suggesting that nesting can extend into August. Given the lack of information on the species' reproductive biology in New Mexico and the rest of its range, here we also use the research on other well-studied Pseudemys turtles to make comparisons of reproductive biology.

Although it is well known that P. concinna, P. floridana, and P. texana lay multiple clutches within a nesting season (Gibbons et al. 1982; Iverson 2001; Mali et al. 2014a), this reproductive characteristic has not been confirmed in P. gorzugi (Pierce et al. 2016). For our study, all sites except the pond and 1 segment of the downstream stretch were visited only once. However, we observed 3 individuals developing a second clutch in June and July after the detection of their first clutch (i.e., shelled eggs) in May; of the 3 females, 1 individual was identified with nearly shelled eggs with a mean diameter of 25.43 mm on 2 July (Fig. 2b). Therefore, we could safely assume that P. gorzugi also produce multiple clutches per reproductive season. Moreover, it is possible that other P. gorzugi with oviductal follicles in the end of June and July might be developing their second clutches.

The largest female P. gorzugi caught in our study had a PL of 251 mm (278 mm straight-line, notch-to-notch carapace length), while all captured females containing follicles or shelled eggs had PLs from 185 to 251 mm (207–278-mm straight-line, notch-to-notch carapace length). In the previous studies, the sizes of gravid females were 237 mm PL (Letter et al. 2017), 242-mm straight-line carapace length (Lovich et al. 2016), and 240-mm carapace length (Degenhardt et al. 1996). In comparison, the smallest nesting female P. floridana had a PL of 215 mm in Florida (Aresco 2004). In a population of P. concinna in Arkansas, the smallest mature female had a PL of 216 mm (Iverson 2001). Mali et al. (2014a) found the smallest nesting female of P. texana to have a PL of approximately 172 mm.

Of all female P. gorzugi captured (n = 165), only 49% were considered sexually mature by our conservative approximation of mature females having PL > 185 mm. We were able to scan with the ultrasound 3 females smaller than the smallest gravid female (173–181 mm PL) between 15 and 21 June, but we did not observe the presence of oviductal follicles. An absence of follicles in any scanned female could be due to females not being sexually mature (i.e., follicles being too small or < 4–5 mm in diameter for detection) or that a female has just laid eggs prior to the scanning (Kuchling 1989). Therefore, our estimates of size at maturity are conservative given that we were not able to confirm whether the individuals with the PL less than 185 mm were nongravid or had not reached sexual maturity.

Within the genus Pseudemys, clutch size can range from 6 to 29 eggs (Ernst and Lovich 2009). Our study reports the smallest and the largest clutch sizes for P. gorzugi to date: 5 and 14 eggs, respectively. We found a positive relationship between reproductive output (clutch size and mean egg size) and maternal body size. In comparison to other Pseudemys, Iverson (2001) found a positive relationship between clutch size and female size in a population of P. concinna in Arkansas, but there was no correlation between egg mass and female size. In both P. texana and P. floridana, larger females produced larger clutches; furthermore, larger P. texana produced wider eggs (Aresco 2004; Mali et al. 2014a).

In large-bodied turtles, the reproductive strategy is expected to follow the optimal egg size theory, meaning that the size of an egg should be at the optimum in each reproductive event regardless of female size, while the offspring number increases with the maternal body size (Congdon and Gibbons 1987; Wilkinson and Gibbons 2005). However, the relationship between egg size and the size of females does not always follow optimal egg size theory due to anatomical constraints (i.e., the size of pelvic aperture and caudal gap), especially in small turtles (Congdon and Gibbons 1987; Rollinson and Brooks 2008; Kern et al. 2015), although this might not be the case in some species (Lovich et al. 2012). Based on our findings, the population of P. gorzugi on the Black River appeared to experience morphological constraints that resulted in an increase in both clutch size and egg size as female size increases. Nonetheless, reproductive strategies either for increasing egg size or for increasing clutch size could be affected by other constraints, such as environmental condition or age (Clark et al. 2001; Wilkinson and Gibbons 2005).

Information on reproductive biology is an important aspect of species conservation (Horne et al. 2003; Sirsi et al. 2017; Lin et al. 2018). Knowledge of seasonality of egg development and deposition can aid management planners in determining appropriate times to implement protection for nesting females, hatchlings, and nest sites. With the current knowledge of P. gorzugi egg development, the protection of nests along the riparian area on the Black River should be emphasized during May and June. However, because we did not find any nest sites or nesting activities of P. gorzugi throughout our study, future efforts could focus on locating nests and monitoring nesting success of the species. Given that reproductive output may vary depending on the populations and geographic regions (Aresco 2004), comparing the reproductive ecology of different populations of P. gorzugi across its range will also contribute to a deeper understanding of the species' natural history, which can be used for planning appropriate conservation practices specific to each individual population.