Large and stout snakes commonly consume large prey and use rectilinear crawling; yet, whether body wall distention after feeding impairs rectilinear locomotion is poorly understood. After eating large prey (30–37% body mass), all Boa constrictor tested could perform rectilinear locomotion in the region with the food bolus despite a greatly increased distance between the ribs and the ventral skin that likely lengthens muscles relevant to propulsion. Unexpectedly, out of 11 kinematic variables, only two changed significantly (P<0.05) after feeding: cyclic changes in snake height increased by more than 1.5 times and the longitudinal movements of the ventral skin relative to the skeleton decreased by more than 25%. Additionally, cyclic changes in snake width suggest that the ribs are active and mobile during rectilinear locomotion, particularly in fed snakes, but also in unfed snakes. These kinematic changes suggest that rectilinear actuators reorient more vertically and undergo smaller longitudinal excursions following large prey ingestion, both of which likely act to reduce elongation of these muscles that may otherwise experience substantial strain.

Macrostomatan snakes, such as boa constrictors, can consume very large prey as a result of their highly mobile skull bones (Cundall and Greene, 2000; Cundall and Irish, 2022) and their distensible skin and soft tissues (Close et al., 2014; Jayne et al., 2022). Unlike limbed vertebrates, in which the propulsive structures are effectively isolated from the gut, the axial structures of snakes have the dual function of accommodating a meal and propelling the animal. Thus, consuming a large meal could be disadvantageous because the localized increase in mass and diameter might impede locomotion by slowing movement velocity, reducing endurance or making some locomotor modes that require axial bending impossible (Crotty and Jayne, 2015; Ford and Shuttlesworth, 1986; Garland and Arnold, 1983). Therefore, the possible detrimental mechanical consequences of a large meal may vary with the locomotor mode used by a snake (Jayne, 2020).

Rectilinear locomotion is the only type of snake locomotion that uses movement of the skin rather than axial bending to generate propulsive forces. This motion is driven by the costocutaneous muscles moving the skeleton relative to the skin, often with a straight-line motion of the body (Lissmann, 1950; Newman and Jayne, 2018). Rectilinear locomotion may be a favorable method of locomotion for snakes after eating if large prey impedes axial bending (Crotty and Jayne, 2015). During the propagated movements of rectilinear locomotion, some ventral scutes are lifted and pulled forward while other scutes have static contact with the substrate (Newman and Jayne, 2018). These localized lifting movements might avoid the high frictional forces associated with lateral undulatory locomotion during which the entire length and weight of the snake simultaneously generate sliding frictional resistance (Hu and Shelley, 2012). Furthermore, after a large meal the distance between the tips of the ribs and the belly scales increases (Cundall and Greene, 2000). In contrast to the fibers of the major epaxial muscles that are used in most locomotor modes (Jayne, 1988a,b), the costocutaneous muscles that connect the ribs to the skin used in rectilinear locomotion (Newman and Jayne, 2018) would have to lengthen. For these reasons, rectilinear locomotion may be crucial for snakes after ingesting large prey.

Despite previous kinematic and electromyographic studies of rectilinear locomotion in Boa constrictor (Lissmann, 1950; Newman and Jayne, 2018), nothing is known about how the kinematics of this mode may change in the region of the body that is distended by a large prey bolus, or whether rectilinear locomotion is possible at all. To test how ingesting large prey affects rectilinear locomotion, we analyzed the kinematics of boa constrictors both before and after ingesting large prey (30–40% of the snake mass). When greater forces are required to move, either because of increased weight or for moving uphill, limbed animals commonly take shorter strides with several attendant changes in kinematics (Birn-Jeffery and Higham, 2014; Watson et al., 2009). Hence, just the increased mass after feeding could be expected to alter the kinematics of snake locomotion. Additionally, a large prey bolus in snakes seems likely to alter the kinematics of locomotion as a result of deforming the body wall and the shape of the snakes, and changing the length and orientation of some muscles that are important for propulsion.

Animals

Our experiments used six captive-bred boa constrictors (Boa constrictor Linnaeus) with similar mass and snout–vent length (SVL) (mean±s.e.m. 1.48±0.07 kg, 1.35±0.02 m; Table 1). All snakes were housed in cages with incandescent light bulbs where they could thermoregulate their daytime body temperature between 25 and 33°C. All husbandry and experimental procedures were approved by Brown University's Institutional Animal Care and Use Committee.

Table 1.

Snake and prey dimensions

Snake and prey dimensions
Snake and prey dimensions

Procedures and apparatus

We fed the snakes prey ranging from 30% to 37% of their body mass (mean±s.e.m. 0.50±0.03 kg) and with a mean cross-sectional area of 20.71±0.70 cm2, which were determined using scaling equations in Jayne et al. (2022) (Table 1). To reduce the chance of regurgitation and to facilitate handling the snakes, the experiments were conducted between 24 and 48 h after feeding. To elicit rectilinear locomotion, we placed the snakes on a 366 cm long foam platform with a vertical wall. The middle floor of the platform had an acrylic plastic window that provided a ventral view of the snake (Fig. 1A) and we placed strips of packing tape every 10 cm on the acrylic window to reduce the amount of slipping. To encourage locomotion in the desired direction, we placed a cardboard box to act as a hide at one the end of the platform. Heat lamps beside the platform allowed the snakes' body temperatures to remain within the preferred range of active body temperatures (26–34°C) (Brattstrom, 1965), but temperature was not quantified.

Fig. 1.

Experimental apparatus and kinematic methods. (A) A Boa constrictor (no. 6 in Table 1) before (below) and after (above) feeding, performing rectilinear locomotion on the experimental apparatus. (B,C) Lateral views of a snake (no. 4 in Table 1) before (B) and after (C) ingesting large prey show how maximum snake height was determined and the location of the marker used to calculate changes in cyclical height (H), and illustrate how relative longitudinal skin length (Xrel) was calculated. The estimated rib tip location can be visualized by the contrast in lateral skin color from light to dark (yellow dashed line) in both A and C. (D,E) Ventral views of the same snake before (D) and after (E) ingesting large prey show how maximum snake width was determined and the location of the marker used to calculate changes in snake width (W), and illustrate how the length between two marked scutes was calculated (L). The scale bar in E relates to B–E.

Fig. 1.

Experimental apparatus and kinematic methods. (A) A Boa constrictor (no. 6 in Table 1) before (below) and after (above) feeding, performing rectilinear locomotion on the experimental apparatus. (B,C) Lateral views of a snake (no. 4 in Table 1) before (B) and after (C) ingesting large prey show how maximum snake height was determined and the location of the marker used to calculate changes in cyclical height (H), and illustrate how relative longitudinal skin length (Xrel) was calculated. The estimated rib tip location can be visualized by the contrast in lateral skin color from light to dark (yellow dashed line) in both A and C. (D,E) Ventral views of the same snake before (D) and after (E) ingesting large prey show how maximum snake width was determined and the location of the marker used to calculate changes in snake width (W), and illustrate how the length between two marked scutes was calculated (L). The scale bar in E relates to B–E.

We recorded simultaneous lateral and ventral views of movement with two digital video cameras operating at 30 images s−1 with a resolution of 4096×2048 pixels (Fig. 1). To provide discrete landmarks on the skin for motion analysis, we attached small square pieces of gaffer tape with a white dot (Fig. 1). We placed the markers on the skin at a mid-dorsal location and five additional locations on the side of the snake in the region of the stomach (Fig. 1B,C). Additionally, two ventral scutes were marked, with a single unmarked scute between them (Fig. 1D,E), as in Newman and Jayne (2018). In fed snakes, we noticed a distinct line of contrast of light and dark skin color in the distended region resulting from different scale patterns and skin strain (Fig. 1A,C). Upon palpation, we determined that this was the approximate region of the tips of the ribs (Fig. 1C). Using stills from video, we estimated the approximate rib tip elevation from the substrate (Table 1).

Motion analysis

Video recordings of locomotion were conducted for each snake both before and after feeding. For each snake, we analyzed at least three trials (several successive cycles of movement) per feeding treatment, resulting in 6–24 cycles of movement analyzed per individual, for a total of 70 and 84 cycles of movement before and after feeding, respectively. Marker position was tracked using XMALab 1.5.3–1.5.5 (Knörlein et al., 2016) to produce two-dimensional coordinates that were subsequently analyzed in Igor Pro 7 software (WaveMetrics, Lake Oswego, OR, USA) and RStudio (RStudio Team, Boston, MA, USA). We calculated 11 kinematic variables based on the positions of the markers, most of which were previously described by Newman and Jayne (2018). Each of these kinematic variables is summarized in Table 2 and representative data are shown in Fig. 2.

Fig. 2.

Representative traces of kinematic data. Representative plots show one cycle of movement for an unfed (open circles; left axes) and fed (filled circles; right axes) snake. Shaded regions indicate static contact for both fed (dark gray) and unfed (dark and light gray) snakes. The unfed snake (no. 2 in Table 1) had a similar cycle duration of 5.6 s to the fed snake 5.7 s (no. 4 in Table 1), and data were down-sampled from 30 Hz to 10 Hz to enhance the visibility of individual symbols. (A,B) Snake height (A) and snake width (B) are shown following a correction based on the dimensions of each individual with cyclic changes in height (ΔH) and width (ΔW) indicated by dashed black lines; differences in magnitude of these traces are apparent between feeding conditions. (C) The distance between two marked scutes (L) is shown along with the method used to calculate changes in L per cycle (ΔL). (D) The relative longitudinal skin movements (Xrel) are shown along with the method for calculating change in Xrel during retraction during static contact (ΔXrel). (E) Forward x-coordinate progression of the snake (X) is shown along with the methods for calculating backwards slipping and forward distance per cycle (ΔX). The average forward velocities for the illustrated fed and unfed trials were 0.45 and 0.63% SVL s−1, respectively.

Fig. 2.

Representative traces of kinematic data. Representative plots show one cycle of movement for an unfed (open circles; left axes) and fed (filled circles; right axes) snake. Shaded regions indicate static contact for both fed (dark gray) and unfed (dark and light gray) snakes. The unfed snake (no. 2 in Table 1) had a similar cycle duration of 5.6 s to the fed snake 5.7 s (no. 4 in Table 1), and data were down-sampled from 30 Hz to 10 Hz to enhance the visibility of individual symbols. (A,B) Snake height (A) and snake width (B) are shown following a correction based on the dimensions of each individual with cyclic changes in height (ΔH) and width (ΔW) indicated by dashed black lines; differences in magnitude of these traces are apparent between feeding conditions. (C) The distance between two marked scutes (L) is shown along with the method used to calculate changes in L per cycle (ΔL). (D) The relative longitudinal skin movements (Xrel) are shown along with the method for calculating change in Xrel during retraction during static contact (ΔXrel). (E) Forward x-coordinate progression of the snake (X) is shown along with the methods for calculating backwards slipping and forward distance per cycle (ΔX). The average forward velocities for the illustrated fed and unfed trials were 0.45 and 0.63% SVL s−1, respectively.

Table 2.

Definitions of kinematic variables

Definitions of kinematic variables
Definitions of kinematic variables

Statistical analysis

For descriptive purposes, we show graphs of data for all 154 observations over a wide range of velocities (Fig. 3). For all these data, several variables had significant correlations with velocity within each feeding treatment (Fig. 3). Hence, to test for a significant effect of feeding and avoid the confounding effects of velocity while accounting for the replication within each individual, we selected a subset of 70 cycles (dark circles in Fig. 3) without a significant difference in velocity between the two feeding treatments (Table 3). This subset had six observations per individual per feeding treatment, except for one individual with just four observations prior to feeding. To test for the main effect of feeding, we used a two-way mixed-model analysis of variance (ANOVA) with type III sums of squares in which feeding (2 levels) was a fixed and crossed factor and individual (6 levels) was a random and crossed factor (hence Ffeed=MSfeed/MSfeed×individual) (Table 3). We used P<0.05 as the criterion for statistical significance for all tests. The relationship between the dependent variables and snake velocity for all trials was determined using linear regressions in Prism 10 software (GraphPad Software, Boston, MA, USA) (Table 4).

Fig. 3.

Effects of velocity on kinematic variables before and after feeding. Each plot shows the entire dataset (n=154). Dark (n=70) and light (n=84) circles indicate the data were included or excluded in the ANOVA testing for an effect of feeding, respectively (Table 3). For the entire dataset within each feeding treatment, regression lines are only shown for significant effects (supporting statistics are in Table 4). (A) Duration of static contact. (B) Forward distance per cycle (ΔX). (C) Movement frequency. (D) Duty factor. (E) Cycle duration. (F) Change in snake width (ΔW). (G) Backwards slipping. (H) Change in snake height (ΔH). (I) Change in ventral skin length (ΔL). (J) Change in the relative longitudinal position of the most ventral marker relative to the mid-dorsal marker during the retraction in static contact (ΔXrel).

Fig. 3.

Effects of velocity on kinematic variables before and after feeding. Each plot shows the entire dataset (n=154). Dark (n=70) and light (n=84) circles indicate the data were included or excluded in the ANOVA testing for an effect of feeding, respectively (Table 3). For the entire dataset within each feeding treatment, regression lines are only shown for significant effects (supporting statistics are in Table 4). (A) Duration of static contact. (B) Forward distance per cycle (ΔX). (C) Movement frequency. (D) Duty factor. (E) Cycle duration. (F) Change in snake width (ΔW). (G) Backwards slipping. (H) Change in snake height (ΔH). (I) Change in ventral skin length (ΔL). (J) Change in the relative longitudinal position of the most ventral marker relative to the mid-dorsal marker during the retraction in static contact (ΔXrel).

Table 3.

Mean values and ANOVA results for the subset of data lacking a difference in velocity between fed and unfed treatments

Mean values and ANOVA results for the subset of data lacking a difference in velocity between fed and unfed treatments
Mean values and ANOVA results for the subset of data lacking a difference in velocity between fed and unfed treatments

Table 4. Regression statistics for snake velocity and kinematic variables for all trials

Table 4. Regression statistics for snake velocity and kinematic variables for all trials
Table 4. Regression statistics for snake velocity and kinematic variables for all trials

All snakes tested were able to perform rectilinear locomotion after ingesting large prey. Measurements of scute movements confirmed that the motions driving rectilinear locomotion included the distended region of the prey bolus (Movie 1). For the subset of data with a similar range of velocities used to assess the effects between feeding, the main effect of feeding condition was significant (P<0.05) for only two of 11 variables: ΔXrel and ΔH (Table 3). Values of ΔXrel (Fig. 3J) were significantly less after feeding (mean±s.e.m. 15.70±1.32 mm) than before feeding (21.48±1.98 mm) (Table 3). In contrast, values of ΔH (Fig. 3H) were significantly higher after feeding (4.39±0.61 mm) than before feeding (2.63±0.43 mm) (Table 3).

For the entire dataset, six kinematic variables had significant relationships with velocity (Fig. 3A–F, Tables 3 and 4). Forward distance per cycle (ΔX) and movement frequency, for both fed and unfed snakes, increased with increasing movement velocity (Fig. 3B,C, Table 4). In contrast, the duration of static contact, duty factor and cycle duration decreased with increasing movement velocity for both fed and unfed snakes (Fig. 3A,D,E, Table 4). Cyclic changes in snake width (ΔW) also decreased with increasing movement velocity, but this relationship was only significant for unfed snakes (Fig. 3F).

The evidence regarding whether a large prey bolus led to slower rectilinear locomotion was rather ambiguous. For example, 24 of 29 of all the cycles with velocity <0.5% SVL s−1 were from fed snakes, and 14 of 16 cycles with velocity >0.9% SVL s−1 were from unfed snakes (Fig. 3). Although the mean velocity of the fastest cycle for each individual snake was higher prior to feeding (mean±s.e.m. 1.03±0.07% SVL s−1) than after feeding (0.83±0.06% SVL s−1), this difference was not statistically significant (two-tailed P=0.122, paired t=1.85, d.f.=5) (Fig. 3). For the entire dataset, the average velocity for each individual after feeding (0.61±0.04% SVL s−1) was less than that prior to feeding (0.74±0.05% SVL s−1), but this difference was not statistically significant (two-tailed P=0.0523, paired t=2.53, d.f.=5).

The distension of the body after feeding was substantial (Fig. 1A), and maximum height (Fig. 1B,C) of the snakes was significantly lower before feeding (mean±s.e.m. 51.9±0.9 mm) than after feeding (83.7±1.7 mm; two-tailed P<0.0001, paired t=25.96, d.f.=5). Similarly, the average maximum body width was significantly lower before feeding (43.5±0.6 mm) than after feeding (71.2±1.1 mm; P<0.0001, paired t=17.63, d.f.=5). The approximate locations of the ribs were visible in the distended area (Fig. 1A,C), and among the six individuals at the greatest diameter, we estimated that the tips of the ribs were elevated from the ventral surface by 30–44 mm as a result of the prey bolus (mean 36 mm). Despite these changes in height and width dimensions of fed snakes, the width of a measured ventral scute (24.47±0.43 mm) in the area containing a prey bolus was not significantly different from that prior to feeding (24.28±0.28 mm) (P=0.462, paired t=0.797, d.f.=5).

Although large prey ingestion resulted in a considerable ∼33% increase in mass and ∼60% increase in both local maximum height and width, rectilinear locomotion was still possible in the region containing a prey bolus and, unexpectedly, only two of 11 kinematic variables were significantly different when compared with snakes prior to feeding (Fig. 3, Table 3). These results are surprising given the expectation that the short-term increase in mass and body wall distention from ingesting large prey might limit mobility and movement velocity, and prevent certain types of locomotion from being performed, as suggested in other studies (Crotty and Jayne, 2015; Ford and Shuttlesworth, 1986; Seigel et al., 1987). Lower movement velocity in fed snakes may also be influenced by other factors, such as a potential decline in the shortening velocity of the muscles actuating locomotion, if higher forces are required to move a more massive snake (Hill, 1938). Many of our highest observed velocities were from unfed snakes (Fig. 3), but the difference between maximal velocities of unfed and fed snakes was not quite statistically significant (P=0.0523). Although our data only indicate preferred movement velocities, this result is interesting when considering the implications for predation and survival; snakes with a large prey bolus present a larger target for predators than normal sized snakes, so moving to safety quickly is probably important. The minimal effects of feeding on large prey (30–37% body mass) observed for the velocity of rectilinear locomotion suggest that this mode may mitigate some of the costs imposed by a large prey bolus, such as decreased movement velocity. However, it is important to note rectilinear locomotion is one of the slowest forms of snake locomotion (Capano, 2020; Newman and Jayne, 2018). Furthermore, we note that our experiments were not specifically designed to test maximal performance.

During rectilinear locomotion in fed snakes, noticeable cyclic changes in height (ΔH) of the region containing a prey bolus were observed that matched the frequency of ventrolateral skin movements and have not been documented in previous studies (Lissmann, 1950; Newman and Jayne, 2018) (Fig. 3). Indeed, ΔH was one of the two variables that had a significant main effect of feeding condition (Table 3). We can think of two possible explanations for these changes in height. First, the timing of the decrease in height coincides with the activity of the costocutaneous inferior (CCI) muscles (Newman and Jayne, 2018), which extend from the ribs to the scutes and thus produce a component of force in the dorso-ventral axis (Fig. 4). It is possible that these muscles may compress the body of the snake when they are active. Indeed, greater oscillations in height were observed in the region containing a prey bolus when compared with non-distended regions in unfed snakes. However, it is in this distended region containing a prey bolus where one would expect dorsoventral compression might be more difficult. Another possibility is that the snakes used localized lifting of the vertebrae. Even without a gap between the belly and the ground, lifting forces could still decrease the load on the belly and the attendant frictional force that would oppose dragging the skin forward in preparation for the next propulsive movement. Snakes are capable of lifting localized regions of the body as is well documented for sidewinding when the most dorsal epaxial muscles lift the body up off the ground (Jayne, 1988a). Occasionally, subtle localized lifting also can occur during lateral undulation and cause a more favorable distribution and orientation of friction forces (Hu and Shelley, 2012). Unlike our study, some previous studies (e.g. Marvi et al., 2013) have observed a small gap between the belly of the snake and the ground during rectilinear locomotion, but it was not clear whether this was from arching the vertebrae or locally deforming the shape of the body. Future experiments with force plates could resolve whether snakes that maintain belly contact with the ground during rectilinear locomotion also locally reduce the vertical force between the belly and the ground.

Fig. 4.

Right lateral views of the hypothesized effect of the food bolus on the length and orientation of the propulsive muscle for rectilinear locomotion. The costocutaneous inferior (CCI) muscle in the stomach region is shown schematically in red extending for the snake before (A) and after (B) ingesting large prey. The skeleton was illustrated from X-ray images of unfed snakes during movement. In fed snakes, the average estimated elevation of the rib tips (36 mm) was similar to the resting muscle length (Lm) of similarly sized unfed Boa constrictor (Newman and Jayne, 2018). This led us to estimate a 45 deg angle of the CCI muscle relative to the substrate and a Lm of 51.7 mm after feeding. A purple dashed line indicates the relative longitudinal skin length range (Xrel) for each condition. The value of ΔXrel post-feeding is 73% of the value pre-feeding, closely matching the 71% prediction resulting from the muscle being oriented at 45 deg (cos 45=0.71). Force component diagrams show the resulting decrease in the horizontal component (Fh, black) and increase in the vertical component (Fv, blue) due to the change in angle of the muscle component (Fm, red).

Fig. 4.

Right lateral views of the hypothesized effect of the food bolus on the length and orientation of the propulsive muscle for rectilinear locomotion. The costocutaneous inferior (CCI) muscle in the stomach region is shown schematically in red extending for the snake before (A) and after (B) ingesting large prey. The skeleton was illustrated from X-ray images of unfed snakes during movement. In fed snakes, the average estimated elevation of the rib tips (36 mm) was similar to the resting muscle length (Lm) of similarly sized unfed Boa constrictor (Newman and Jayne, 2018). This led us to estimate a 45 deg angle of the CCI muscle relative to the substrate and a Lm of 51.7 mm after feeding. A purple dashed line indicates the relative longitudinal skin length range (Xrel) for each condition. The value of ΔXrel post-feeding is 73% of the value pre-feeding, closely matching the 71% prediction resulting from the muscle being oriented at 45 deg (cos 45=0.71). Force component diagrams show the resulting decrease in the horizontal component (Fh, black) and increase in the vertical component (Fv, blue) due to the change in angle of the muscle component (Fm, red).

Although the main effect of feeding condition was not significant for cyclic changes in snake width (ΔW), these changes occurred with the same frequency, but opposing sign, as cyclic changes in height (ΔH), especially for fed snakes (Fig. 2A,B). This suggests that in fed snakes, the ribs may have cyclical lateral movements during rectilinear locomotion in the region of the prey bolus (caliper rotation as in Capano, 2020). While the same pattern was observed in unfed snakes, the magnitude of cyclic changes in both height and width was lower (Fig. 2A,B). Yet, these cyclic changes in width suggest that the ribs also move during normal rectilinear locomotion in unfed snakes. This idea supports the notion from recent studies that ribs may play an active role during movement (Capano, 2020; Capano et al., 2022) and is in contrast to previous hypotheses that the ribs are stationary during rectilinear progression (Capano, 2020; Lissmann, 1950; Newman and Jayne, 2018). A limitation to our measurement of cyclic width changes is that it required snakes to maintain constant contact between the lateral portion of their body and the vertical wall (Fig. 1D,E). For example, if a snake flared the marked segment of its body outward from the wall, the resulting data would indicate an increase in snake width. Lapses of contact with the vertical wall occurred in some trials, but when large outward movements were visible, those trials were not analyzed. Still, we acknowledge that small movements of the snake away from the wall could have been incorporated into the final analysis. Finally, while the data seem to suggest that the ribs are at least partially mobile during rectilinear progression, the benefit of lateral caliper rotation of the ribs to the kinematics is not obvious; in fact, this motion would act to lengthen the CCI muscles (Capano, 2020). More likely, it is a result, particularly in the fed snakes, of the volume inside the snake shifting as the CCI muscle acts to compress the snake dorso-ventrally.

While changes in cyclic height and width were accompanied by significant differences in maximum height and width, we did not observe any change in width of ventral scutes for unfed and fed snakes. This result is consistent with our expectation that there would be very low compliance, even for the scutes within the distended region containing a large prey bolus, because the ventral scales of boa constrictors are quite thick and stiff (Fig. 1D,E). Indeed, it is largely the skin in the hinge regions between the stiff scales that is capable of unfolding to accommodate distention from stresses imposed by large prey boli. The stiffness and relatively large width of the ventral scutes likely benefits rectilinear locomotion by providing a reasonably rigid structure for the propulsor muscles (CCI) to act upon.

The only other variable with a significant main effect of feeding condition was the change in relative longitudinal skin length (ΔXrel), which was lower after feeding (mean±s.e.m. 15.70±1.32 mm) than before feeding (21.48±1.98 mm) (Table 3). Interestingly, we believe this kinematic pattern and rib elevation measurements provide some insight into the primary actuators of rectilinear locomotion (CCI muscles). As the costocutaneous muscles are attached to the skin, stretching the skin will affect muscle fiber orientation which likely influences muscle fiber length as well (Fig. 4). After the snakes in our experiment were fed, we estimate that the CCI fibers are oriented ∼45 deg relative to horizontal (Fig. 4B), whereas in unfed boa constrictors, the fibers of the CCI are effectively oriented longitudinally (Fig. 4A). Our estimate of 45 deg comes from the assumption that the CCI muscle represents the hypotenuse of a right triangle where the average estimated elevation of the rib tips (36 mm; Table 1) is about the same as the resting muscle length (Lm) of similarly sized Boa constrictor (Newman and Jayne, 2018) (Fig. 4B). This more vertical CCI reorientation following feeding would change the orientation of muscle force (Fig. 4B) and supports the idea that CCI muscles in the region of the prey bolus could compress the snake following feeding (Fig. 4). Notably, this more vertical orientation could also reduce the strain placed on the CCI muscle. Using trigonometry and assuming that the longitudinal span (bottom of the right triangle) of the CCI muscle is the same in unfed and fed snakes, we estimated that the 36 mm vertical elevation of the CCI after feeding would cause the resting length of the CCI muscle to be 1.41 times longer than in the unfed snake (51.7 mm; Fig. 4B). Additionally, for every 1 unit of change in the CCI length after feeding, the change in horizontal location of the muscle insertion would be only 0.71 (cos 45) (Fig. 4B). Interestingly, this predicted 29% reduction in muscle extension corresponds very closely with the empirical 27% reduction in ΔXrel following large prey ingestion (Table 3). Although the observed decrease in ΔXrel after feeding could help reduce lengthening of the CCI (Fig. 4B), this kinematic change would not be sufficient to allow the CCI to function over the same range of strain as before feeding if the position of the ribs remains the same. In contrast, a substantially more vertical orientation of the ribs could accomplish this, but an increasingly vertical rib orientation would decrease the useful (horizontal) force (Fh in Fig. 4). According to the well-established length–tension relationship of vertebrate skeletal muscle, muscle force declines to zero when stretched to long lengths (Gordon et al., 1966; Ramsey and Street, 1940). Given the apparent preservation of function after feeding, further work investigating CCI muscle shortening patterns and contractile physiology could reveal unusual muscle properties that may compensate for a likely large range in muscle strain.

Similar to Newman and Jayne (2018), we found that as velocity increased, both frequency and the forward distance traveled by a ventral scute per cycle (ΔX) increased significantly (Fig. 3B,C), whereas both the duration of static contact and duty factor decreased significantly (Fig. 3A,D). Thus, fed and unfed snakes used a similar combination of modulating both the frequency and amplitude of movements to modulate velocity. Increasing the velocity of rectilinear locomotion by increasing movement frequency and distance traveled by the ventral scutes (ΔX) is analogous to increasing the frequency of limb oscillations and stride length in limbed vertebrates. Similar to previous findings for both rectilinear locomotion of boa constrictors (Newman and Jayne, 2018) and a wide variety of limbed locomotion (Birn-Jeffery and Higham, 2014; Watson et al., 2009), the proportion of the cycle with static contact with the ground decreased with increased velocity. At least on artificial surfaces, backwards slipping of the ventral scutes during the propulsive cycle often occurs during rectilinear locomotion (Lissmann, 1950; Newman and Jayne, 2018). Ingesting large prey increases the weight of a snake and, therefore, frictional forces between it and the substrate. It might be expected that slipping would decrease significantly in fed snakes as a result of increased friction. Yet, the smaller amount of backward slipping per cycle after feeding (2.2 versus 3.5 mm) was not significantly different between feeding conditions. While the lack of significant differences of kinematic variables between feeding conditions was surprising, it is important to note that rectilinear locomotion is made up of muscle activity and movements that are propagated down the length of the snake (Newman and Jayne, 2018), unlike ventilatory rib movements, which are not propagated (Capano et al., 2022). Thus, an interesting but unresolved issue for rectilinear locomotion is whether snakes could propagate muscle activity posteriorly up to the distended region in a scenario where body segments might lack muscle activity, followed by the resumption of motor pattern propagation posterior to the food bolus. This could not be the case for either lateral undulation or sidewinding locomotion because the very definitions of these modes include the propagation of lateral bending along the entire length of the snake (Jayne, 2020). Similar localized axial movements have been observed during ventilation when snakes only move a subset of the ribs surrounding the lung (Capano et al., 2022). While the snakes in our study performed rectilinear locomotion after feeding using all body segments as skin movements propagated seamlessly through and beyond the distended region, a scenario may exist, perhaps with very large prey (>37% body mass), where extreme distention could prevent active CCI muscle activity. Here, the series elastic component of the CCI muscle could function purely as a passive structure at such long lengths, similar to the highly extensible intermandibularis muscle of northern watersnakes (Nerodia sipedon) (Close et al., 2014). Thus, whether or not undistended regions of a snake are capable of compensating for a distended region to produce rectilinear locomotion is another interesting question for future study.

The results of the present study suggest that rectilinear locomotion varies minimally following the ingestion of large prey, likely due to the CCI muscle maintaining function despite being strained along the body wall it lines. While studies characterizing the properties of snake muscles are scarce, many unique features of snake cranial morphology associated with large gape (macrostomy) and eating large prey are well known (Close et al., 2014; Cundall and Greene, 2000; Jayne et al., 2022). Presumably, the ability of a snake to move its distended body after feeding is also important, but compared with cranial structure and function, the effects of feeding on post-cranial locomotor function have received scant attention (Cundall, 2019). One unique feature of rectilinear locomotion is that skin movements propel the animal, but the skin and associated muscles also must stretch considerably to accommodate a large meal. This direct effect of eating a meal on the locomotor apparatus of snakes differs radically from the situation in limbed animals whose propulsive structures are not altered by feeding. Rectilinear locomotion also differs from other modes of snake locomotion because it involves neither axial bending nor overcoming static friction forces across the entire body of a snake to start movement (Capano, 2020; Hu and Shelley, 2012). These attributes of rectilinear locomotion coupled with the results of this study, which suggest that rectilinear locomotion kinematics are resilient to the acute changes caused by body distention from a prey bolus, highlight its potential importance when snakes may have a reduced ability to bend following feeding. It is noteworthy, however, that all snakes, unlike limbless lizards, have costocutaneous muscles and, therefore, a potential for performing rectilinear locomotion, including basal lineages of non-macrostomate snakes that eat small prey. Altogether, rectilinear locomotion may be a locomotor mode that uniquely mitigates some challenges posed by a large prey bolus; hence, it may have facilitated the evolution of large prey ingestion in macrostomatan snakes.

We thank two anonymous reviewers whose suggestions greatly improved the manuscript, Richard Marsh and Hannah Weller for helpful discussion, Erika Tavares for proofreading, and Lucy Campbell and Yoshihiro Yajima for assistance with data collection.

Author contributions

Conceptualization: J.C.P., J.G.C.; Methodology: J.C.P.; Formal analysis: J.C.P., B.C.J., A.D.W.; Investigation: J.C.P., A.D.W.; Resources: B.C.J., T.J.R.; Data curation: J.C.P.; Writing - original draft: J.C.P.; Writing - review & editing: J.C.P., B.C.J., A.D.W., T.J.R.; Visualization: J.C.P.; Supervision: B.C.J., T.J.R.; Funding acquisition: T.J.R.

Funding

This project was supported by the National Science Foundation grant 1832795 to T.J.R.

Data availability

The data used in this study are available from the Dryad digital repository (Petersen et al., 2024): https://doi.org/10.5061/dryad.hmgqnk9q5

Birn-Jeffery
,
A. V.
and
Higham
,
T. E.
(
2014
).
The scaling of uphill and downhill locomotion in legged animals
.
Integr. Comp. Biol.
54
,
1159
-
1172
.
Brattstrom
,
B. H.
(
1965
).
Body temperatures of reptiles
.
Am. Midl. Nat.
73
,
376
-
422
.
Capano
,
J. G.
(
2020
).
Reaction forces and rib function during locomotion in snakes
.
Integr. Comp. Biol.
60
,
215
-
231
.
Capano
,
J. G.
,
Boback
,
S. M.
,
Weller
,
H. I.
,
Cieri
,
R. L.
,
Zwemer
,
C. F.
and
Brainerd
,
E. L.
(
2022
).
Modular lung ventilation in Boa constrictor
.
J. Exp. Biol.
225
,
jeb243119
.
Close
,
M.
,
Perni
,
S.
,
Franzini-Armstrong
,
C.
and
Cundall
,
D.
(
2014
).
Highly extensible skeletal muscle in snakes
.
J. Exp. Biol.
217
,
2445
-
2448
.
Crotty
,
T. L.
and
Jayne
,
B. C.
(
2015
).
Trade-offs between eating and moving: What happens to the locomotion of slender arboreal snakes when they eat big prey?
Biol. J. Linn. Soc.
114
,
446
-
458
.
Cundall
,
D.
(
2019
).
A few puzzles in the evolution of feeding mechanisms in snakes
.
Herpetologica
75
,
99
-
107
.
Cundall
,
D.
and
Greene
,
H. W.
(
2000
).
Feeding in snakes
. In
Feeding: Form, Function and Evolution in Tetrapod Vertebrates
(ed.
K.
Schwenk
), pp.
293
-
333
.
San Diego, CA
:
Elsevier Science
.
Cundall
,
D.
and
Irish
,
F.
(
2022
).
Macrostomy, macrophagy, and snake phylogeny
. In
The Origin and Early Evolutionary History of Snakes
(ed.>
D.
Gower
and
H.
Zaher
), pp.
437
-
454
.
Cambridge University Press
.
Ford
,
N. B.
and
Shuttlesworth
,
G. A.
(
1986
).
Effects of variation in food intake on locomotory performance of juvenile garter snakes
.
Copeia
1986
,
999
-
1001
.
Garland
,
T.
and
Arnold
,
S. J.
(
1983
).
Effects of a full stomach on locomotory performance of juvenile garter snakes (Thamnophis elegans)
.
Copeia
1983
,
1092
-
1096
.
Gordon
,
A. M.
,
Huxley
,
A. F.
and
Julian
,
F. J.
(
1966
).
The variation in isometric tension with sarcomere length in vertebrate muscle fibres
.
J. Physiol.
184
,
170
-
192
.
Hill
,
A. V.
(
1938
).
The heat of shortening and the dynamic constants of muscle
.
Proc. R. Soc. London. Ser. B - Biol. Sci.
126
,
136
-
195
.
Hu
,
D. L.
and
Shelley
,
M.
(
2012
).
Slithering locomotion
. In
Natural Locomotion in Fluids and on Surfaces
(ed.
S.
Childress
,
A.
Hosoi
,
W.
Schultz
and
J.
Wang
), pp.
117
-
135
.
New York, NY
:
Springer New York
.
Jayne
,
B. C.
(
1988a
).
Muscular mechanisms of snake locomotion: an electromyographic study of the sidewinding and concertina modes of Crotalus cerastes, Nerodia fasciata and Elaphe obsoleta
.
J. Exp. Biol.
140
,
1
-
33
.
Jayne
,
B. C.
(
1988b
).
Muscular mechanisms of snake locomotion: An electromyographic study of lateral undulation of the Florida banded water snake (Nerodia fasciata) and the yellow rat snake (Elaphe obsoleta)
.
J. Morphol.
197
,
159
-
181
.
Jayne
,
B. C.
(
2020
).
What defines different modes of snake locomotion?
Integr. Comp. Biol.
60
,
156
-
170
.
Jayne
,
B. C.
,
Bamberger
,
A. L.
,
Mader
,
D. R.
and
Bartoszek
,
I. A.
(
2022
).
Scaling relationships of maximal gape in two species of large invasive snakes, brown treesnakes and Burmese pythons, and implications for maximal prey size
.
Integr. Org. Biol.
4
,
obac033
.
Knörlein
,
B. J.
,
Baier
,
D. B.
,
Gatesy
,
S. M.
,
Laurence-Chasen
,
J. D.
and
Brainerd
,
E. L.
(
2016
).
Validation of XMALab software for marker-based XROMM
.
J. Exp. Biol.
219
,
3701
-
3711
.
Lissmann
,
H. W.
(
1950
).
Rectilinear locomotion in a snake (Boa occidentalis)
.
J. Exp. Biol.
26
,
368
-
379
.
Marvi
,
H.
,
Bridges
,
J.
and
Hu
,
D. L.
(
2013
).
Snakes mimic earthworms: Propulsion using rectilinear travelling waves
.
J. R. Soc. Interface
10
,
20130188
.
Newman
,
S. J.
and
Jayne
,
B. C.
(
2018
).
Crawling without wiggling: Muscular mechanisms and kinematics of rectilinear locomotion in boa constrictors
.
J. Exp. Biol.
221
,
jeb166199
.
Petersen
,
J.
et al. 
. (
2024
).
Effects of ingesting large prey on the kinematics of rectilinear locomotion in Boa constrictor [Dataset]
.
Dryad
.
Ramsey
,
R.
and
Street
,
S.
(
1940
).
The isometric length-tension diagram of isolated skeletal muscle fibers of the frog
.
J. Cell. Comp. Physiol.
15
,
11
-
34
.
Seigel
,
R. A.
,
Huggins
,
M. M.
and
Ford
,
N. B.
(
1987
).
Reduction in locomotor ability as a cost of reproduction in gravid snakes
.
Oecologia
73
,
481
-
485
.
Watson
,
J.
,
Payne
,
R.
,
Chamberlain
,
A.
,
Jones
,
R.
and
Sellers
,
W. I.
(
2009
).
The kinematics of load carrying in humans and great apes: Implications for the evolution of human bipedalism
.
Folia Primatol.
80
,
309
-
328
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information