Journal list menu

Volume 15, Issue 3 e4716
Open Access

Evidence of handedness in turtles

Caroline Honan

Corresponding Author

Caroline Honan

Department of Biological Sciences, Louisiana State University, Baton Rouge, Louisiana, USA


Caroline Honan

Email: [email protected]

Search for more papers by this author
John B. Hopkins III

John B. Hopkins III

Center for Wildlife Studies, Camden, Maine, USA

Search for more papers by this author
Matthew W. H. Chatfield

Matthew W. H. Chatfield

School of Biology and Ecology, University of Maine, Orono, Maine, USA

Search for more papers by this author
First published: 14 March 2024

Handling Editor: Debra P. C. Peters


Although research supporting cerebral lateralization, or handedness, in non-human vertebrates is expanding, reptiles represent one of the least studied. While assessing wood turtle shells, we noticed more dents, scrapes, chips, and other forms of damage on the right side of turtles than on the left. Asymmetrical injury has been attributed to cerebral lateralization in other taxa, and current research supports handedness in testudines. After creating a “scute damage index” to catalog damage location, we confirmed turtles had more damage on the right side of their shells than on the left. We offer four hypotheses that explain this pattern.

Wood turtles (Glyptemys insculpta) range widely in upland habitats during the summer months, thus exposing them to a wide variety of threats, including predators, vehicles, farm equipment, and other turtles (Harding & Bloomer, 1979; Saumure et al., 2007). In conjunction with a study that investigated the ecology and health of wood turtles in Maine, we assessed damage to the carapace (top) and plastron (bottom) of their shells (see Figure 1).

Details are in the caption following the image
Diagrams showing the categorization of scutes for analysis. (A) Scutes are labeled by the type of scute. Scutes labeled with “M” are marginal scutes, those with “P” are pleural scutes, and those labeled with a “V” are vertebral scutes. The plastron is not shown, as all plastron scutes are gular. (B) Scute locations include left (L), right (R), and center (C).

While analyzing our collection of photographs of 159 wood turtles, we observed an interesting pattern of shell damage and wear: turtles seemed to have more dents, scrapes, chips, and other forms of damage on their right side. This was especially pronounced on the anterior right margin of their shells. We hypothesized that cerebral lateralization, or handedness, is the likely cause for this pattern of damage (Babcock, 1993; Clapham et al., 1995; Reist et al., 2011).

The brain is split into two hemispheres, each of which regulates different functions. To prevent conflicting outputs, one hemisphere is dominant and can inhibit outputs from the other, ensuring the animal responds to stimuli appropriately (Sainburg & Eckhardt, 2005). Hemisphere dominance is thought to manifest physically as handedness (Bisazza et al., 1998), which is evident in motor skill biases, such as showing preference for either the left or right side when feeding, mating, or escaping predators. If an organism has a favored limb or an innate tendency to use limbs on one side of its body over the other, its turning direction may be influenced when escaping a predator, resulting in where bodily injuries occur, or, in our case, where shell damage will concentrate over time.

Once thought to be a uniquely human trait, evidence of handedness has been recorded in several non-human vertebrates in the past few decades, including toads (Bufo bufo; Bisazza et al., 1998), anoles (Anolis carolinensis; Deckel, 1995), tree lizards (Urosaurus ornatus; Hews & Worthington, 2001), cottonmouths (Agkistrodon piscivorus; Roth, 2003), humpback whales (Megaptera novaeangliae; Clapham et al., 1995), parrots (Psittaciformes; Harris, 1989), and many species of primates including chimpanzees (Pan troglodytes) (Sainburg & Eckhardt, 2005) and common marmosets (Callithrix jacchus) (Forrester et al., 2013).

Evidence of handedness in turtles is also growing. For example, Pacific leatherback turtles (Dermochelys coriacea) demonstrate population-level “right flipperedness” in which the majority of the population uses their right flipper when digging nests (Sieg et al., 2010). Righting behavior is also often lateralized, with trends appearing either within an individual's behavior, as seen in painted turtles (Chrysemys picta) (Smith et al., 2017), or within entire populations as with green sea turtles (Chelonia mydas) and olive ridley turtles (Lepidochelys olivacea; Malashichev, 2016). Tortoises (Testudo hermanni) also tend to have lateralized righting responses, and when retested 10 months later, they display the same directional preference (Stancher et al., 2006). Other evidence of cerebral lateralization in turtles includes motor lateralization in eye and limb use (Sovrano et al., 2018) and lateralized sleeping positions where the head is tilted consistently to one side (Spiezio et al., 2022).

Turtles are an excellent group for studying handedness in wild animals because damage to their shells is direct evidence of predation events, long-term wear, injury from collisions with agricultural machinery, or other threatening circumstances. Turtles are also ideal candidates for studying injury and shell damage because they are long-lived and have high capture rates.

To assess carapace and plastron damage, we developed the scute damage index (SDI) based on a modification to the carapace mutilation index in Saumure et al. (2007). Instead of assessing damage to each of the four quadrants (sensu Saumure et al., 2007), we assessed the damage on all 50 scutes individually using the following scale:
  • 0 = No obvious damage aside from minor wear and scrapes. The entirety of the scute is still intact.
  • 1 = Some damage, such as minor chipping or denting. Small pieces of the scute may be missing but the scute is still largely intact.
  • 2 = Severe damage, such as fractures, often with sections missing.
SDI scores were calculated for each individual by adding all of an individual's scute scores together. The range of possible SDI scores an individual can receive ranges from 0 (if every scute received a score of 0) to 100 (if every scute received a score of 2). Most shell injuries appeared as chips or cracks in the shell, possibly from being chewed on or dropped by a predator.

We used the measuring tool in ImageJ (Schneider et al., 2012) to produce polygons for each scute. An example wood turtle shell used for data collection was taken from and used to provide relative sizes for each scute (Conservation Plan for the Wood Turtle in the Northeastern United States, 2021). We then divided these relative areas by the sum of all polygons to find the proportion of area coverage for the carapace and plastron separately. Each scute was identified by its placement and assigned a position of either right, left, or center (Figure 1).To account for the number of χ2 tests (38 for the carapace and 12 for the plastron scutes), we used a Bonferroni-corrected χ2 test to examine the relationships between damage frequency on individual scutes and the scute area coverage for males, females, juveniles, and all turtles together. Sex was determined using a combination of straight-line carapace length (adults >180 mm), plastron concavity, and presence of eggs.

This modified χ2 test resulted in three possible outcomes for each scute: greater than expected, less than expected, or average damage.

We recorded a total of 234 damaged scutes (i.e., those receiving an SDI score of 1 or 2). A total of 60 turtles (38%) had at least one damaged scute. We found that carapace scutes (n = 171) were more damaged than plastron scutes (n = 63). We further learned that females' shells were damaged significantly more often than males' shells (χ2 = 6.7, df = 2, p = 0.036) or juveniles' shells (χ2 = 12.8, df = 1, p < 0.001): 40 of 84 females (48%), 17 of 50 males (34%), and 3 of 25 juveniles (12%) had damaged sells. In terms of average SDI values and SE, males had significantly lower SDIs (50.2 ± 0.2) than females (51.4 ± 0.3) or juveniles (51.2 ± 0.3).

We also discovered that scutes on the right side of turtle shells were significantly more damaged than scutes on the middle or left side (χ2 = 7.6, df = 2, p < 0.001). We found that 12 of the 171 damaged scutes (7%) on turtle carapaces were cervical scutes and 10 of the 171 (6%) were first right marginal scutes, whereas the right gular scutes of the plastron made up 5 of 63 (8%) of damaged scutes on turtle plastrons; all three of these scutes had more damage than expected (p < 0.001). Five scutes also had less damage than was expected by area coverage (p < 0.001). Of the 171 damaged carapace scutes, the second and third pleural scutes accounted for 2 and 3 (2% each), 7 of them were the fourth vertebral scute (4%), and 4 of them were the fifth vertebral scute (25%). Of the 63 damaged plastron scutes, 5 of them were the left fifth scute (8%). Forty-one scutes were neither more nor less damaged than expected (carapace: χ2 = 284.9, df = 37, p < 0.001; plastron: χ2 = 43.0, df = 11, p < 0.001) (Figure 2).

Details are in the caption following the image
Damaged wood turtle shells and the resulting diagrams. (A) Carapace and (B) plastron diagrams depicting modified χ2 outcomes for the summed data set. Dark color indicated scutes that are more likely to receive damage, medium color is neither more nor less likely, and light color is less likely. In sum, the anterior right area is significantly more likely to accumulate injuries and wear. (C) Carapace of a female captured in June 2018 and (D) plastron of a female captured in September 2020. Both individuals were among the 159 turtles used in this study. Red arrows indicate examples of damage that were included in this dataset. Photo credit: Matthew Chatfield.
We formulated several hypotheses for the asymmetrical damage pattern that we observed.
  1. Turtles react to predators by turning to the left, exposing their right-hand side. Turning as a defensive posture has been documented in many turtle species (Dodd, 1978; Dodd & Brodie, 1975; Neill, 1971; Pope, 1935), although none have coupled this behavior with lateralization. Studies have been conducted, however, examining biases in other escape behaviors. For example, European pond turtles (Emys orbicularis) more often escaped to the left when faced with a simulated predator in three different scenarios (Pellitteri-Rosa & Gazzola, 2018).
  2. Turtles react faster to threats on the left, leaving them more susceptible to threats approaching on their right. In general, the right hemisphere is responsible for processing spatial awareness, escape responses, and intense emotions such as fear (Lippolis et al., 2002). Hemispheres and eyes are heavily contralaterally wired; thus, the left eye and the right hemisphere are strongly connected (Bisazza et al., 1998). Processing the predator through the right hemisphere may produce an escape response faster than if the same predator was processed through the left. Studies on escape behaviors of common toads and stripe-faced dunnarts (Sminthopsis macroura) show faster reaction times to predatory stimuli presented on their left side than their right (Lippolis et al., 2002, 2005).
  3. Turtles are more likely to make clockwise, right-hand turns as they navigate their environment, exposing their right side to greater wear and damage from gravel, rocks, logs, and other hard objects. Wood turtles can be severely injured during flooding incidents that wash individuals downstream, possibly ramming them into rocks or other hard objects (Jones & Sievert, 2009). Although it is unlikely that a turtle can direct itself in a large-scale flood, it may be possible in lighter rapids.
  4. Predators preferentially attack from their left, causing damage to the right side of turtle shells, therefore, the pattern we observed may be due to handedness in predators, not turtles. Just as studies have shown side preferences for escape behaviors, organisms often exhibit lateralized feeding behaviors. For example, humpback whales have a consistent side preference during rolling and lunging while feeding (Canning et al., 2011) and parrots often feed using their left foot (Harris, 1989). Interestingly, scarring on turtles from boat propellers—an injury source that is not lateralized—is equally distributed on a left-to-right axis, but not an anterior-to-posterior one (Hollender et al., 2018).

Our findings establish a pattern of damage that is consistent with handedness in wood turtles, but it is important to note that we remain unsure about the causes of injury or if they are all from the same source. Identifying the main cause that explains the pattern of asymmetrical damage we found, however, requires controlled experiments and field studies. For instance, simulating predation in a captive setting would allow us to test predictions deduced from hypotheses 1, 2, and 3, whereas a field study that tracks the daily movements of turtles could be helpful for investigating hypothesis 3.


We thank Derek Yorks and Phillip deMaynadier from the Maine Department of Inland Fisheries and Wildlife (MDIFW), Cheryl Frederick from the Center for Wildlife, and the many field technicians who have contributed to this project. We also give special thanks to Barry Woods (1952–2020) for his contributions to this analysis and his enthusiasm and dedication to teaching. This work was conducted under MDIFW Scientific Collection Permit numbers 2016-447, 2017-447, 2018-447, and 2019-447. Permission was granted by the Unity College Institutional Animal Care and Use Committee under approval numbers UC 051501 and UC 051701.


    The authors declare no conflicts of interest.


    Data (Honan, 2022) are available from Figshare: