Study Location and Animals.—

We conducted this study from August 2017 to January 2018 at the St. Augustine Alligator Farm Zoological Park (SAAFZP), Florida, USA. The SAAFZP displayed over 600 individual crocodilians representing 24 extant species (Table 3). This study only included captive crocodilians, which is advantageous as captive specimens display a higher tolerance toward conspecifics and are therefore less likely to express agonistic behavior (Lang, 1987). The monitored crocodilians were typically isolated from conspecifics or paired with a respective mate or several potential mates, with the exception of A. mississippiensis housed communally and separated by size (juveniles, subadults, and adults) and Crocodylus siamensis housed communally (juveniles and adults). All 24 extant species were not equally represented in number of individuals, age, or sex. This study excluded hatchling and neonate crocodilians (<1 yr) to avoid discrepancy in determining provoked vs. unprovoked mouth gaping observations, as newborn crocodilians are hard wired from birth to display greater agonistic behavior than adults (Brien et al., 2013). We recorded data for adult, subadult, and juvenile crocodilians. The five species with the highest frequency of mouth gaping observations varied in total number of individuals available for the study, being A. mississippiensis, Cr. siamensis, Crocodylus intermedius, Crocodylus rhombifer, and Crocodylus halli (= Crocodylus novaeguineae; Murray et al., 2019) (Table 3). All monitored crocodilians were adults except Cr. siamensis (6 adults, 30 juveniles), Cr. halli (1 adult, 2 subadults), Mecistops cataphractus (6 adults, 4 subadults, 4 juveniles), Crocodylus palustris (1 adult, 2 juveniles), and an indeterminate number of A. mississippiensis (∼150 adults, ∼50 subadults, >150 unknowns).

Table 3.  List of all crocodilians studied with number of individuals (n), highest ambient temperature (T_a), and lowest T_a for each species observed mouth gaping, current geographic range, and International Union for Conservation of Nature (IUCN, 2020) conservation status of least concern (LC), vulnerable (VU), critically endangered (CR), conservation dependent (CD), and not evaluated (NE).

View Fullscreen

Habitats.—

Climate conditions of St. Augustine, Florida were humid subtropical, with average T_a between 20–32°C. All crocodilians were housed outdoors with groundwater wells supplying continuous running water pools maintaining temperature between 20–22°C.

Observations.—

We conducted the study ad libitum (i.e., not set to strict time restraints or specific enclosures) observing all 24 species at least twice a day, 5 d a week. Similar to previous event-based observational studies, in which duration of behavior was not recorded (Augustine et al., 2017), this examination took a systematic observational approach (Bakeman and Gottman, 1997). Metrics recorded were as follows: species, individual, date, time of day, enclosure number, sex, size (adult, subadult, juvenile), T_a, T_sb, T_sh, eyes open or closed, weather conditions (sunny, overcast, raining), and insolation position (body completely/partially/not in sun, head in sun/shade). We opportunistically documented any potential external stimuli during mouth gaping events, including interactions with other organisms and simultaneous gaping with environmental changes. Such opportunistic mouth gaping events can be useful for further testing.

Several factors constrained this study to omit observations made in the spring and early summer months of the year (February–July). First, increased bellowing is more frequent during the courting season (February–May at the SAAFZP) and is typically understood to be a sexual attractant (Carr, 1967; Garrick and Lang, 1977) or agonistic behavior to conspecifics (Kellogg, 1929; Oliver, 1955; Vliet, 1989). Mistaking an agonistic mouth gaping event for one unprovoked would have been likely. Second, T_b in female Caiman latirostris, and presumably all crocodilians, can vary in relation to reproductive condition, including seasonality (Bassetti et al., 2014). Vitellogenesis (i.e., yolk production in the oocyte) and oogenesis occur in the spring and extend into midsummer (May–July at the SAAFZP) as T_a increases (Joanen and McNease, 1980, 1989; Lance, 1989; Guillette et al., 1997), except for SAAFZP Cr. johnstoni and Gavialis gangeticus, which typically deposited eggs in March, SAAFZP Cr. intermedius, which sometimes exhibited courting behavior in October and November, and American Crocodiles (Crocodylus acutus), which have bred in January and February (Kushlan and Mazzotti, 1989). And third, aggressive interactions during the courting/nesting season can affect T_b in displaced subordinate crocodilians from normal basking areas (Seebacher and Grigg, 1997; Seebacher et al., 1999).

We used a Raytek MT6 Infrared Thermometer “Thermo gun” (Raytek, Santa Cruz, California, USA) equipped with a 1 : 10 heat-source diameter-to-distance ratio to record T_sb, centered midway down the osteoderm columns at a 90° angle, and T_sh, centered on the cranial table at a 90° angle. We did not record T_sh and T_sb on individuals at a distance >1.8 m because of potential inaccuracy, especially regarding posterior width of the cranial table that can be <18 cm in 2.7 m snout–vent length (SVL) crocodilians (Hall and Portier, 1994). We did not obtain T_sh on crocodilians <2.7 m SVL unless we could safely procure temperature readings <1.8 m from the cranial table. We obtained T_a at the time of the mouth gaping observation from a nearby weather station (KFLSTAUG19 at 29.894490°N, −81.296396°W, datum WGS 84), controlled by a Davis Vantage Vue monitoring station (Davis Instruments, Hayward, California, USA) utilizing MeteoBridge software (version 1.x, Lentföhrden, Germany).

Survey.—

We distributed an electronic mail survey to 390 zookeepers and crocodilian biologists asking two questions based on their expertise: 1) What factors contribute to unprovoked mouth gaping in crocodilians?; and 2) What are potential advantages of unprovoked mouth gaping in crocodilians? This survey was sent out on 23 April 2018 and left open for a collection period of 30 d.

Statistical Analysis.—

This study employed statistical analyses using the open-source statistics software “R” (R Core Team, 2014). In addition to summary statistics, nonparametric tests were applied as data failed to meet assumptions of normality and homoscedacity, regardless of transformations. A series of Wilcoxon Signed Rank Tests (nonparametric paired t-test) were performed to detect differences between T_a and T_sb, T_a and T_sh, as well as T_sb and T_sh (α = 0.05). False Discovery Rate (FDR) adjustments (Benjamini and Hochberg, 1995) were applied to t-test values following Dalmasso et al. (2005). Analyses included pooled data comprising all species in addition to comparisons of species-specific differences for the top five by number of observations (n): Cr. siamensis (n = 109), Cr. rhombifer (n = 85), Cr. intermedius (n = 71), A. mississippiensis (n = 57), and Cr. halli (n = 48); while observations for each individual were paired for T_a, T_sb, and T_sh.

Linear mixed effects models were constructed using the R package lme4 (Bates et al., 2015) to evaluate the effects of T_a and T_sb on T_sh using a repeated measures design. The models included T_a and T_sb as fixed effects, with species and individual as random effects: T_sh ∼ T_a + T_sb + (1|species) + (1|individual). Models were also constructed to evaluate the relationships between the magnitude of temperature differences (T_d) between T_sh and T_a considering eye position (open or closed), head position (sun or shade), body position (sun or shade), sex, weather (sunny, overcast, rain) and T_a's. Fixed effects included eyes (E), head position (H), body position (B), sex (S), and weather (W) with species and individual random effects: T_d ∼ E + H + B + S + W + (1|species) + (1|individual). Models were selected based on QQ plots of residuals and Bayesian Information Criterion (BIC) values and P-values following Maximum Likelihood Ratio (LRT) testing.

Temperature Results.—

Across all crocodilians collectively, paired Wilcoxon Signed Rank Tests (WSRT) indicated median T_sh were not significantly higher than T_sb while mouth gaping (Z = 0.299, P = 0.1815; Fig. 2). Nonparametric t-tests on species-specific patterns revealed Cr. halli, Cr. intermedius, and Cr. rhombifer demonstrated statistically higher T_sh, while no significant differences in T_sh and T_sb were detected for A. mississippiensis and Cr. siamensis (Table 4).

View Figure

• Download as Powerpoint
• Download Figure

Table 4.  Temperature range, mean, standard deviation (SD), and Wilcoxon Signed Rank Test statistical analysis for surface body (T_sb), surface head (T_sh), and ambient (T_a) temperature in five species with the highest frequency of mouth gaping observations. Sex ratio of males (M) to females (F) and eye frequency data reported for each species.

View Fullscreen

Median T_sb and median T_sh were both significantly higher than T_a across all crocodilians in all weather conditions (paired one-sided WSRT: T_sh, Z = 3.527, P < 0.0001; T_sb, Z = 3.507 P < 0.001) and held true for each of the five selected crocodilian species (Table 4). Median T_sh and median T_sb were significantly different from T_a across 39 mouth gaping observations in overcast/raining conditions or at night (paired two-sided WSRT: T_sh, Z = 1.052, P < 0.0001; T_sb, Z = 1.014, P < 0.0001). The mean T_a for all mouth gaping occurrences was 23.8°C (n = 370) while minimum and maximum observations for T_a were 3.9°C and 32.2°C (Table 3).

Across all species (values represent means ± SE), T_sh of 29.2°C (±0.01) was consistently higher than T_sb and T_a of 28.9°C (±0.01) and 23.8°C (±0.01). Variation in T_sh is best explained by T_sb, as T_sh increases by 0.8°C relative to T_sb, but only by 0.1°C for T_a (LMM fit with LRT: χ² = 48.4, P < 0.001 and χ² = 14.9, P < 0.001 for T_sb and T_a, respectively).

Insolation, Weather Condition, Eye, and Size Results.—

Crocodilian insolation posturing on sunny days varied greatly between the five selected species, with fully exposed (both head and body) as low as 14.7% in Cr. intermedius and as high as 52.9% in Cr. siamensis. The five selected species exhibited varied body position results during mouth gaping activity in different weather conditions (Fig. 3). The closed-eye occurrence frequency while mouth gaping was 13% across all crocodilians (n = 491). Size classes were not equally represented because of the opportunistic study design; however, adult Cr. siamensis in this study mouth gaped more often (87.16%) than did juvenile Cr. siamensis (12.84%).

View Figure

• Download as Powerpoint
• Download Figure

Across species, T_d between T_sh and T_a averaged 5.4°C (± 0.01 SE). Eye position (open or closed) was a significant predictor for T_d, as closed eyes reflect a temperature increase of 1.9°C relative to eyes open (LMM fit with LRT, χ² = 28.7, P < 0.001). Head and body position were also significant predictors, with head in the sun yielding an increase of 4.5°C compared to the shade and body in the sun yielding an increase of 1.6 °C compared to shade (LMM fit with LRT, χ² = 28.8, P < 0.001 and χ² = 13.1, P < 0.001 for head and body position, respectively). Interestingly, sex was not a significant factor and was removed from the model (LMM fit with LRT: χ² = 3.2, P = 0.2). Conversely, weather was a significant variable, most notably between sun and rain with a decrease of 6.2°C followed by sun to overcast decreasing by 4.6°C (LMM fit with LRT: χ² = 38.8, P < 0.001).

Opportunistic Behavioral Results.—

A single Cr. palustris (#91284) accounted for a disproportional amount of observations, with 4.28% of the total (n = 561), whereas two Cr. acutus and two C. yacare accounted for 0.00% of the total. On multiple occasions, crocodilians showed no behavioral response to flies landing on their tongues (Fig. 4), and one individual adult Cr. porosus (#A03162) frequently gaped underwater in the absence of fish (Fig. 5). An anecdotal example of weather/excitatory events coinciding with a behavioral response occurred after Hurricane Irma hit the SAAFZP. On 12 September 2017, after the storm passed and weather conditions were sunny with T_a = 28.9°C, staff filled the half-empty pools back to their original levels. Concurrent with filling the main Alligator Lagoon pool (which housed 36 A. mississippiensis), nearly every individual began bellowing, which is seldom observed outside of spring. We observed many A. mississippiensis exiting the pool once filled at ∼1320 h, and >18 specimens maintained long-term mouth gaping over the course of a few hours. We additionally observed mouth gaping on a fair-weather night after 2000 h with T_a = 26.1°C on 18 September 2017. Observations included an adult female Crocodylus mindorensis, adult male Cr. halli (T_sb: 28.6°C; T_sh: 28.9°C), subadult male Cr. halli (T_sb: 25.8°C; T_sh: 25.8°C), and an adult male Cr. siamensis (T_sb: 30.3°C; T_sh: 29.7°C).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Survey Results.—

We received 18 responses for a response rate of 5% from the electronic questionnaire. A range of answers that have not been published in previous literature included: facilitated breathing (during high relative humidity or with respiratory illness), improved low-frequency auditory ability, dental health, relaxation (especially regarding muscle-relaxing drugs), and inattentive behavior (Table 5). Responses that coincided with published literature were not included in this study.

Table 5.  List of unpublished potential advantages of unprovoked mouth gaping behavior from an electronic mail questionnaire.

View Fullscreen

Temperature and Mouth Gaping.—

Two crocodilian species did not exhibit significantly higher T_sh than T_sb (Table 4). Johnson (1974) also observed no differences between T_b and T_h in artificially heated and thermocoupled Cr. novaeguineae and Cr. porosus. The magnitude of difference (T_d) between T_sh and T_a was mainly a function of head position (4.5°C difference between sun and shade) rather than body position (1.6°C between sun and shade).

Interestingly, gaping A. mississippiensis T_sb, T_sh, and T_a (Table 4) were below the preferred optimum T_b (33.4–35°C) (Colbert et al., 1946; Johnson, 1974; Lang, 1979). This is somewhat surprising, as thermoregulatory studies indicate skin temperature is typically higher than core T_b in reptiles (Andrews, 2008; Carretero, 2012; Halliday and Blouin-Demers, 2017). Conversely, T_sb's and T_sh's of the four crocodile species (Table 4) were within the preferred optimal T_b (28–32°C) for both captive and wild hatchling Cr. acutus (Lang, 1975, 1979) and the maximum mean T_b (26.9–29.2°C) for Cr. niloticus (Downs et al., 2008).

Mouth gaping can be effective for cooling areas of the body (Cloudsley-Thompson, 1969; Spotila et al., 1977; Terpin et al., 1979; Lang, 1987). However, nonuniform insolation positions (Fig. 3) indicated crocodilians do not mouth gape exclusively to cool off. Huchzermeyer (2003) suggested sunlight radiation on the interior buccal surface area (see Fig. 1B) facilitated potential T_h and T_b elevation. Yet, shadow cast from the maxilla often interfered with this potential (e.g., Fig. 1C,D), and the small surface area of slender-snouted Crocodylia (e.g., Fig. 1A) also limited this effect. Solar radiation is unnecessary to increase T_b via mouth gaping, provided T_a > T_b and high relative humidity (Diefenbach, 1975). Further research is needed in a controlled setting to decouple these relationships. However, the diversity of insolation observations in the current study (Fig. 3) is consistent with wild Cr. johnstoni, which employ an array of behavioral postures to maintain operative temperatures (Seebacher, 1999).

We were unable to adequately test for size as a factor because of the limited number of size classes with few observations for juveniles. A more pronounced surface area-to-bulk ratio of larger animals provides a greater capacity for heat storage (Cott, 1961; Grigg and Gans, 1992), which may explain why, if larger crocodilians require a longer time to internally reach optimal T_b, adult Cr. siamensis mouth gaped more often than their juvenile counterparts.

We report surface temperatures and not core T_b, lending caution when interpreting results. For example, skin temperature is typically higher than cloaca temperature in warming Common Gartersnakes (Thamnophis sirtalis) (Halliday and Blouin-Demers, 2017) and lizards (Andrews, 2008; Carretero, 2012). However, T_b and T_h studies are invasive and can alter behaviors, requiring some experiments to stimulate gaping by propping the mouth open (Spotila et al., 1977). Core T_b requires researchers to surgically implant dataloggers/thermocouples (Johnson, 1974; Glanville and Seebacher, 2006; Downs et al., 2008; Bassetti et al., 2014), administer them orally as pseudogastroliths (Loveridge, 1984; Grigg et al., 1998), or perform cloaca probing (Cott, 1961; Brisbin et al., 1982; Loveridge, 1984; Hutton, 1987). Conversely, we utilized a Thermo gun so natural behaviors would not be compromised from handling stress following cloaca capsule insertion or datalogger attachment. Additionally, SAAFZP crocodilians were accustomed to daily human presence and were typically <1.8 m from the Thermo gun, allowing for accuracy beyond that of a wild scenario. We also acknowledge that while T_a and direct solar radiation influence crocodilian thermoregulation, heat from the ground may also radiate via conduction, which would alter individual behavior (Bassetti et al., 2014). Additionally, effects of wind and solar conductance can affect the true thermal condition (i.e., T_a) for each mouth gaping event (Chappell, 1981).

Behavioral Factors Influencing Mouth Gaping.—

The literature contains many hypotheses for mouth gaping (Table 1) in addition to unpublished ones that include behavioral responses and physiological benefits (Table 5). Individual behavior/traits may factor into gaping. Although we did not document Cr. acutus or C. yacare mouth gape, Cr. acutus have mouth gaped in captivity and in the wild (Gamble, crocodilian trainer; Lloret, crocodilian biologist, pers. comm.), whereas C. yacare have mouth gaped in captivity (Gamble, pers. comm.). Loveridge (1984) noted that monitoring individual behavior, similar to the present study, could be more informative than following large groups of crocodilians, as individual personality differences may yield variability in gaping behavior.

Dominance behavior is plausible, although many crocodilians were isolated (e.g., all four Cr. rhombifer were isolated from conspecifics). Our models did not detect a relationship between sex and gaping behavior. Sleeping and inattentive behavior is also unlikely as results show a low ratio of closed-eye occurrences (13% [n = 491]; Table 4). Yet across species, modeling indicated gaping with closed eyes reflected elevated T_d by 2°C. Both head and body position were important factors explaining differences between T_sh and T_a while gaping; however, head position exerted a stronger influence. Insolation appears to play a role in the behavior, although limitations of this study did not determine exact mechanisms. Therefore, we were unable to detangle behavioral and thermoregulatory responses and warrants further study.

There is documentation of fish ‘cleaning' the teeth of submerged Cr. acutus (Guggisberg, 1972; Dinets, 2013a) and A. mississippiensis mouths (Darlington, reptile curator, pers. comm.). Our opportunistic observations of crocodilian interactions with other organisms suggest a potential relationship between crocodilians and fish/flies. If such an association exists with cleaner fish, it is interesting that the submerged Cr. porosus (#A03162) mouth gaped in the absence of fish.

Additional Factors Influencing Mouth Gaping.—

Weather may play a larger role than previous studies have indicated. We detected a significant relationship between T_d and weather, with substantial temperature differences between sun and rain. A sudden weather change can elicit mouth gaping, as indicated by behavioral responses in Cr. niloticus with respect to localized weather (Loveridge, 1984), seasonal weather, and behavior in A. mississippiensis (Brisbin et al., 1982) and posturing/movement in relation to weather by Cr. johnstoni (Seebacher and Grigg, 1997). Wild Cr. niloticus mouth gaped immediately after exiting water (Loveridge, 1984), similar to our documentation of >18 A. mississippiensis maintaining long-term gaping after exiting the pool following several days of reduced water. Curiously, we rarely recorded these 36 individuals gaping otherwise. This suggests excitatory events may produce changes in behavior such as gaping and vocalizations typically only heard during breeding season.

We report the second published nocturnal mouth gaping observation and the first for three species (i.e., Cr. mindorensis, Cr. halli, and Cr. siamensis). Loveridge (1984) observed nocturnal mouth gaping in a single Cr. niloticus. Dinets (2010) documented nocturnal stationary terrestrial behavior (lying on land) in five species of wild crocodilians, with no overlap of the aforementioned three species we observed. We did not note any probable stimuli to elicit these occurrences.

Recommendations for Future Research.—

Future research should consider factors of mouth gaping behavior in crocodilians beyond thermoregulation (Table 5). The range of environmental conditions and insolation likelihood rendered in this study, in addition to that of the literature, indicated gaping behavior is complex, making it difficult to draw clear cause and effect relationships. Future studies may benefit from a focus on natural history and behavior, as T_a and insolation likelihood fail to significantly factor into mouth gaping. This may reveal communication patterns associated with mouth gaping, as crocodilians are arguably the most behaviorally complex extant reptiles (Thorbjarnarson and Hernández, 1993). Long distance signaling is conserved across Crocodylia phylogenetic families (Dinets, 2013b), increasing the likelihood that gaping is a form of social behavior/communication.