Variable Rates of Parasitism on Ornate Box Turtles (Terrapene ornata) in Eastern Kansas by the Chelonian-Specific Ectoparasite Cistudinomyia cistudinis
Abstract
Understanding which factors make an individual susceptible to parasitism provides insight on how a parasite infects a host. Numerous factors such as habitat characteristics, body size, and sex have shown differential impacts on parasitism rates. Trends in parasitism rate often exist for similar taxonomic groups, so expanding research to new species helps test the robustness of such trends. A chelonian-specific ectoparasite, Cistudinomyia cistudinis (Diptera: Sarcophagidae), has been documented in case studies dating back a century. Almost all reports of infection by this flesh fly are of a single or few turtles, so there is a lack of information regarding the population dynamics of infection. This lack of information means little is known about which factors make a turtle more or less likely to be parasitized. In this work, we compared the occurrences of parasitism in ornate box turtle (Terrapene ornata) populations across multiple seasons from 2 different habitats in northeastern Kansas (North Topeka and Lawrence). At the population level, we found a significantly higher number of parasitized turtles in the Lawrence population. Within the Lawrence population, we found females were more likely to be parasitized. This study sets baseline infection rates for a long-lived vertebrate host that can help better understand the reasons for parasitism both among and within populations.
The susceptibility of a host to be infected by parasites can be influenced by several factors, including body size (Watkins and Blouin-Demers 2019), sex (Salvador et al. 1996; Reimchen and Nosil 2001), and habitat characteristics (Durden et al. 2004; Mellor and Rockwell 2006; Hussain et al. 2013; Álvarez-Ruiz et al. 2021; Morena-Rueda 2021). Larger individuals, which have a corresponding larger surface area to harbor parasites, are often found to have greater parasite numbers than smaller conspecifics (Watkins and Blouin-Demers 2019). Sex of the individual can influence parasitism rates in multiple ways, with sexually dimorphic rates of parasitism attributed to different sex-specific hormone levels (Salvador et al. 1996) or sex-specific niche differentiation (Reimchen and Nosil 2001). Beyond characteristics of the individual, habitat composition can also play a role in parasitism rates, as some habitats are more conducive to the presence of parasites than others (Durden et al. 2004; Mellor and Rockwell 2006; Hussain et al. 2013; Álvarez-Ruiz et al. 2021; Morena-Rueda 2021). The numerous and varied factors influencing parasitism likely mean a single trait or habitat characteristic is not solely responsible for increased parasitism. As such, there is a need for more studies, especially in free-ranging animal hosts, to better understand the underlying reasons for variation in parasitism within and across populations.
At the individual level, parasite load can have a direct impact on an individual’s fitness, e.g., via host death (Rainey 1953; King and Griffo 1958) or decreased reproductive output (Kleindorfer and Dudaniec 2016). At the population level, the death of some individuals through parasitism may be viewed as a population-limiting factor (Watson 2013). In a long-lived species, understanding rates of parasitism across years can help understand how host-parasite interactions are impacted in response to many factors, including genetic changes in both the host and parasite (Barribeau et al. 2014) and changes in immune response over time (Møller et al. 2003).
Outside of individual and population-specific factors, increased parasite load in 1 population allows for some inferences about the environment to which the host population is exposed. A higher rate of parasite infection will be reflective of climatic conditions that are better suited for the parasite (Herrmann and Gern 2010) or the presence of specific features, such as vegetation, needed to support parasites (Mize et al. 2011). The influence that the environment has on parasite abundance may be correlated with the amount of time spent on the host. Changes in environmental conditions can have differing levels of effect on parasites, as parasites that spend a smaller percentage of their lives on the host are more sensitive to changes in environments (Smith et al. 2023). Studying longer-lived hosts and their subsequent parasite infection rates can lead to increased knowledge on how environmental changes are influencing the host-parasite dynamic.
Specialist dipteran parasites can target groups of species belonging to the same taxonomic family (Moreno-Rueda 2021). A species of flesh fly, Cistudinomyia cistudinis (Townsend) (= Sarcophaga cistudinis), seems to be exclusively parasitic to chelonians (Knipling 1937). Adult C. cistudinis females larviposit onto chelonians, with larvae that are ectoparasites, feeding on the hosts until falling off to pupate (Knipling 1937). Documentation of C. cistudinis infestation in chelonians dates back over a century (Wheeler 1890). Numerous papers describe infestation of different chelonians by these flesh flies, including the gopher tortoise (Gopherus polyphemus; Knipling 1937), eastern box turtle (Terrapene carolina carolina; King and Griffo 1958), three-toed box turtle (T. c. triunguis; Jackson et al. 1969), and ornate box turtle (T. ornata; Rainey 1953). Even a chelonian species collected from the Galapagos Islands, which is outside the known range of C. cistudinis (Knipling 1937), was parasitized in a Texas zoo after relocation. Early work forcibly attempting to initiate an infestation on wounds of nonchelonian organisms, including goats, sheep, and an alligator, found no infestation on nonchelonian hosts (Knipling 1937). The ability to infest chelonian hosts and the inability for infestation on organisms outside the group demonstrates the high level of specialization for C. cistudinis on chelonian hosts.
Almost all reports of C. cistudinis infestation on chelonians consist of observations on only a single (McMullen 1939; Rainey 1953) or small group of turtles (Knipling 1937). An exception is the unpublished note of D.G. Hall (mentioned in Kipling 1937 and Peters 1948), which reported a 25% parasitism rate in 136 turtles over a 3-year period in Georgia. Unfortunately, no details were provided for which species of chelonian was being parasitized and in what type of habitat or area. Therefore, there is a need to expand studies on flesh fly parasitism of chelonians in populations across multiple years to better understand the factors that may lead to varying levels of parasitism.
In this work, we conducted flesh fly parasite surveys on ornate box turtles from 2 populations in northeast Kansas. These 2 populations are relatively close distance-wise (≈60 km) but are in distinctly different habitat types. We aimed to address questions regarding the role of size, sex, and habitat differences in parasitism rates. Specifically, we determined whether larger individuals harbored more parasites, whether 1 sex was more commonly parasitized than the other, and whether we observed a difference in parasitism rates based on habitat differences/population locality. As ornate box turtles can live to be over 30 years old (Metcalf and Metcalf 1985), this work helps set a baseline for which future studies can build on to better assess how this host-parasite dynamic is shaped across multiple years.
METHODS
Populations Sampled. —
Ornate box turtle populations were examined in 2 different northeastern Kansas sites, henceforth called Lawrence and North Topeka (Table 1). We systematically examined all turtles located with radio transmitters (transmitters: Wildlife Materials 2380 or Holohil 2B-RB; radio receiver: Wildlife Materials TRX 1000S with a 3-element Yagi antenna) for flesh flies between 2020 and 2022 for our Lawrence site and between 2021 and 2022 for our North Topeka site. We also opportunistically examined all turtles without transmitters (including juveniles) for flesh flies as we found them while conducting our normal telemetry and survey work (Table 2). Until we ran out of radio transmitters to use, turtles with a mass exceeding 280 g were given a radio transmitter so that we could monitor potential flesh fly infection throughout the summer. Turtles were visually checked, independently, by a minimum of 2 trained observers (trained by B.R.) for evidence of flesh flies, including myiasis, scarring, and live larvae. If flesh fly larvae were detected, the larvae were extracted when possible from the turtle using a pair of small forceps and stored in a freezer (either dry or in 100% ethanol). All turtles were weighed in grams (Pesola M1000 Digital Pocket Scale) and measured (curved carapace length [CCL]) using a soft tape measure at time of handling regardless of whether they had flesh flies or not. Turtles were typically examined once per year when originally found unless they had a radio transmitter (see Table 2 for details), in which case they were handled and checked for flesh flies an average of 3 times per active season: beginning (April/May), middle (June/July), and end (September/October).
Parasite Identification Based on Larval Genotype. —
Molecular methods were used for species identification. A total of 60 larval specimens collected from 15 unique turtles (ranging from a minimum of 1 larva extracted from a turtle to a maximum of 24 larvae from 1 turtle) were analyzed for molecular species identification. Tissue was removed from either the midportion or anterior half of each larval specimen. DNA was extracted using 2 different methods. The first method (n = 10 specimens) included using 200–220 μl of 10% Chelex (BioRad, Hercules, CA) solution and incubating overnight at 56°C. After overnight incubation, samples were incubated at 99°C for 10 minutes. The second method (n = 50 specimens) incorporated robotic DNA extraction on the Automate Express. Using the BTA Lysis Prepfiler Kit (ThermoFisher, Waltham, MA), samples were incubated in 220 μl of BTA Lysis buffer and 10 μl of Proteinase K overnight at 56°C. Lysates were run on the AutoMate Express using the BTA protocol and eluted in a final volume of 40 μl. All DNA extracts were amplified with primers C1-J-1751a and C1-N-2191 following the thermal cycling protocol in Wells et al. (2001). PCR products were visualized using the methods described in Smith and Cook (2020), and samples were cleaned up and sequenced as described in Mercader et al. (2022). Edited sequences were searched in BLAST.
Statistical Analysis. —
We used a Fisher’s exact test to compare the number of turtles infected and not infected with flesh flies between the North Topeka and Lawrence populations in 2021 and 2022 when our sampling efforts were equivalent. To determine the severity of observed infections, we calculated the average number of flesh flies with 1 standard deviation per turtle. As a result of the low observed infection rate in North Topeka, we focused on within-population comparisons using our observations collected over 3 field seasons at the Lawrence site. Within the Lawrence population, we used a logistic regression with a Firth correction using the Brglm package in R (R Core Team 2023) to determine whether infection probability (infected or uninfected) of adults (> 280 g) was related to the size (mass or curved carapace length [CCL]), sex (M/F) and/or year (2020, 2021, or 2022) within this population. Our model was parasite infection (yes/no) ∼ Size + Sex + Year.
RESULTS
Demographics Based on Field Site. —
In 2021 and 2022, when we surveyed both populations, a total of 13 of the 130 (10%) unique turtles we encountered and examined exhibited at least 1 parasitic flesh fly. The infection rate in Lawrence (11/42 = 26%) was much higher (11.5 times) than the infection rate in North Topeka (2/88 = 2%) during this time frame, and this difference was statistically significant (p < 0.05). When we include our 2020 data from Lawrence, the infection rate for all adults examined at this site was 18/49 (37%). The average number of larvae for all infected turtles (n = 20) across all years was 16.2 ± 15.72 (SD), with an infection intensity ranging between 1 and 56 flesh flies (Fig. 1). We had 2 additional turtles (M031 and F069) with flesh flies, but we were unable to safely extract any of these flesh flies to get a count estimate. In turtles from which we could remove flesh flies, we found some individual gas tracks (infection burrows through the tissue) contained more than 10 larvae in the same space.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 23, 1; 10.2744/CCB-1586
For the Lawrence population, the average size in terms of mass (g) was 422.33 (± 69.97) and 380.68 (± 51.22) for females and males, respectively. Assessing body size using CCL (mm), the average size was 141.15 (± 7.12) and 135.8 (± 6.38) for males and females, respectively. Using only the Lawrence population for all years (2020–2022), our logistic regression using mass (g) as our metric for body size showed that sex (p = 0.03) was a significant predictor of parasite infection, while mass (p = 0.41) and year (p = 0.10) were not. In these models, males had an infection rate (0.24; 6/25) that was half that of females (0.50; 12/24). When we replaced mass with CCL as our metric for body size in our logistic regression, we found sex again was a significant (p = 0.035) predictor of parasite infection probability, and while year trended toward significance (p = 0.08), CCL did not (p = 0.49). The infection rate in 2022 (0.60) was double the rate in 2020 (0.30) and 2021 (0.33).
Across the turtles’ active season from March until September, we found flesh flies in every month (Fig. 2). August, with 7 flesh fly infections, was the month with the most flesh flies detected. Conversely, March, June, and July all had the lowest number of flesh flies, with only 1 found during each of these months.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 23, 1; 10.2744/CCB-1586
Molecular Identification of Parasitic Larvae. —
Of the 60 samples sequenced, all matched C. cistudinis (GenBank accession number MG967779.1). All but 7 of the specimens had 100% sequence identity to MG967779.1. The 7 other specimens had over 99.7% sequence identity to MG967779.1, with the only difference being a single base (G instead of A). The resulting substitution is synonymous.
Repeated Chelonian Parasitism. —
A total of 2 turtles, both from the Lawrence population, had parasites during separate field seasons over the course of the study period.
DISCUSSION
Here we present baseline C. cistudinis infection rates across 2 ornate box turtle populations in different years. Previous work describing flesh fly infestation of the ornate box turtle (McMullen 1939; King and Griffo 1958), including from northeastern Kansas (Rainey 1953), focused on individual turtles. As a long-lived species where many members of each population will be examined in subsequent field seasons, this information will be a useful reference point for future comparative studies that probe specific reasons for why parasitism rates are higher in some individuals than others.
Our 2 study populations are geographically close but occupy distinctly different habitats in terms of vegetation type, cattle grazing, and human disturbance. We detected clear differences in infection rate, which we believe is likely attributable to these habitat differences. One factor that may contribute to the different parasitism rate is the increased human disturbance at the Lawrence site in comparison to the North Topeka site. The Lawrence site includes cropland that is regularly tilled, fields that are hayed, mowed lawns, several roadways, numerous houses, and prairie-burning practices throughout the year, including when the turtles are aboveground. Likely due to these disturbances, large parts of the prairie are overrun with wild raspberry (Rubus occidentalis) and poison ivy (Toxicodendron radicans), clear signs of disturbance. Previously, it has been shown that parasitism rates of ectoparasites can increase in response to increased human habitat disturbance in other ectotherms (Payne et al. 2020). Others have found lower parasitism rates in disturbed areas (Carbayo et al. 2019); however, this lower rate means the level of human disturbance is not solely responsible for the increased parasitism, which may also be true across our sites.
It is clear our 2 study sites vary dramatically in infection rate, and we hypothesize this may be due to the Lawrence site’s having features such as more urbanization, more invasive species, more standing water, and more habitat disturbance via burning, which may make the area more conducive for the presence of C. cistudinis. Many members of the family Sarcophagidae, to which C. cistudinis belongs, are considered synanthropic (Buenaventura et al. 2018; Jafari et al. 2019). Unfortunately, little is known about the range, seasonality, and preferred habitat for C. cistudinis outside of known state records (Pape 1996). Cistudinomyia cistudinis is found along the East Coast and in the Great Plains (Pape 1996), indicating an ability to live in diverse climates. In terms of seasonality, we provide evidence of the presence of flesh fly infection throughout the active seasons of the ornate box turtle populations studied for this work (April through October). It has been hypothesized that larval infection coincides with openings in the skin caused by ticks (Knipling 1937). While no official tick surveys have been conducted at either field site, anecdotally there have not been any ticks present on turtles examined over this study period. A better understanding of how C. cistudinis infects turtles, as well as more information regarding range and habitat type most associated with this flesh fly, will lead to a better understanding of why certain habitat types are more prone to parasitism.
The data from Lawrence agree with other data showing aggregated distribution (Shaw et al. 1998), with few turtles carrying many parasites (n = 3 turtles with more than 30 parasites), while the majority of the turtles sampled did not have any parasites. North Topeka, on the other hand, had only a few individuals with any parasites, and the intensity of infection for these individuals was very low. The difference in maximal infection intensity across populations was not expected, and it will be interesting to keep monitoring this difference to see how it may change over time.
Within-population comparisons among the Lawrence population revealed that sex, but not body size, was a factor in different rates of parasitism. Female turtles were more than twice as likely to be parasitized as males. In numerous species, sex is a factor known to cause differences in parasitism rates, with examples of higher prevalence rates in males (Comas 2020) or in females (Durden et al. 2004). Differences in sex-biased infection are likely to be taxon- or even species-specific, as previous work in freshwater turtles shows males with higher ectoparasite infection rates for 1 species and females with higher ectoparasite infection rates for another species from the same genus (Santana et al. 2019). As we observed higher rates in females, it would be interesting to determine whether egg-bearing females have a reduced immunity response to flesh flies, as their energy would be likely be more directed toward follicular development than immunity. Anecdotally, at least 2 of our infected females have been observed nesting in the same year, and although this is a small sample size, it may be worth examining more in the future as a potential cost of reproductive investment in the ornate box turtle. It is possible other females also nested in the year they were infected; increased survey efforts could help corroborate this hypothesis.
Regarding the lack of differences in parasitism rates because of size, we used curved carapace length as 1 of our size measurements, a practice used in ornate box turtles (Reed et al. 2023) and other chelonian species (Codron et al. 2022). This measure accounts for the variation in shell shape, specifically with respect to how domed the shells can be. While not a direct measure of the amount of soft tissue that could harbor parasites, we assume turtles with larger shells would have increased soft tissue.
The work here provides the demographic data that could be further explored in subsequent research. We have set a baseline level of C. cistudinis infection rates for 2 ornate box turtle populations in northeastern Kansas. We found a clear difference in infection rate and habitat characteristics between the 2 populations, and further work will be aimed at attempting to clarify which features of the habitat of the Lawrence population make it more susceptible to infection. For example, 2 of the turtles from the Lawrence population have been parasitized in more than 1 field season. Continued tracking of instances of repeat parasitism and subsequent comparisons to turtles who are never infested will provide a dataset from which it would be possible to discern what makes certain turtles more or less susceptible to parasitic events. There is limited information beyond case notes involving few turtles for chelonian parasitism by C. cistudinis, so this work expands the current knowledge base by focusing on population-wide host-parasite demographics both within and among populations.
Histogram showing the infection intensity observed in the study as indicated by the number of larvae extracted per adult across both populations and all years.
Number of flesh fly detections per month for the Lawrence (n = 17) and North Topeka (n = 2; 1 each in May and September) populations. One adult parasitized is not included in this figure as the exact month of parasitism was not included.
Contributor Notes
Handling Editor: Peter V. Lindeman
