Conceived and designed the experiments: AL GK. Performed the experiments: AL. Analyzed the data: AL. Contributed reagents/materials/analysis tools: AL HKK. Wrote the paper: AL GK. Collection and maintenance of the research animals: AL HKK.
The authors have declared that no competing interests exist.
The common chameleon,
Changes in body orientation in response to external stimuli are fundamental to animal motion and locomotion and require the perception of one’s location in relation to the relevant stimuli. Frequent examples are provided by visually guided responses, including cases in which animals perform highly accurate spatio-temporal corrections of their body or organ position relative to a moving stimulus. Such position corrections, often referred to as “station keeping”
Chameleons (Chamaeleonidae, Reptilia) are slow-moving, predominantly arboreal lizards that capture insect prey with a long tongue. Chameleons rely on cryptic coloration and slow motion to approach prey and to reduce visual detection by potential predators
When a threat appears on the side of a branch opposite to that on which a chameleon is perched, the chameleon will often remain motionless
In this avoidance response, as with other motor responses of animals with bilateral morphological symmetry, the question of laterality arises. Is there an effect of direction of the stimulus’ approach on the motor patterns displayed by the chameleon? While the spatio-temporal patterns of eye use have been analyzed
Bilateral symmetry in vertebrates is widely expressed morphologically and anatomically
Lateralization of motor functions has been described in all poikilotherm groups–fish, amphibians and reptiles. For example, in the mosquitofish (
(A) Oblique view. (B) Schematic overhead view. The experimenter, positioned behind the camera (a) acts as the threatening stimulus. Chameleon (x), vertical pole (b), incandescent bulbs (c), pole rotation cords (d), visual barrier (e), screen (f).
An overhead view of the sagittal plane of the head of a chameleon (C) when perched vertically on a pole (P), in relation to the threat (T); α – the angle in relation to the threat, β – the angle in relation to both threat and pole.
In amphibians, lateralized motor patterns are found in limb use in toads (e.g.,
Among reptiles, lateralized visuo-motor behavior has been found in the aggressive behavior of
Several theories have addressed the possible functions of lateralization. One line of reasoning is that in animals with laterally placed eyes, a feature common to most vertebrates, lateralization reduces inter-hemispheric conflicts. Such conflicts may arise when two stimuli are perceived simultaneously, one by each eye
Lateralization may occur at the individual level, population level, or both. An equal distribution of lateralization in the population (termed “anti-symmetrical”) implies that approximately one half of the individuals are biased toward one side while the other half is biased toward the opposite side. An asymmetrical distribution occurs when a significant proportion of the population is biased toward a given side
In this study we analyze the motor responses of chameleons faced with an approaching threat in an attempt to provide insight into the dynamics of the response and to assess whether it is lateralized at the individual level, population level or both. In their natural habitats, chameleons mostly move in relatively thick, homogeneous vegetation. Threats, such as predators, may appear from any distance or direction with equal probability. Having an avoidance response that is lateralized, i.e., biased toward a given side, may be detrimental to survival. We therefore hypothesize that the avoidance response of the chameleon will not be side-dependent.
A single frame from a sampled video sequence is depicted. (A) Unmodified image showing the ventral view of the chameleon holding onto a narrow pole, with its eyes protruding from both sides of the pole. (B) Body surface of the chameleon with the areas exposed on each side of the pole (hatched) used for the determination of respective surfaces. The caudal border of the area analyzed (broken horizontal line) is determined on the basis of 3×maximal head width, from the rostral end.
A chameleon perched on a vertical pole (P) and the threat, as viewed from above. (A) The chameleon is positioned opposite (ca.180°) the threat, in an initial state. (B) The position of the chameleon during, or immediately following, pole rotation. A given side of a chameleon is termed the “leading side” if the threat approaches from that side (i.e., the left side of the chameleon during left-approaching threat, as shown here, or the right side of the chameleon during right-approaching threat). The side opposite the leading side in each test is termed the “following side.”
The research was conducted at the Dept. of Biology, University of Haifa, Oranim Campus in Tivon, Israel, between November 2006 and November 2009. Collection, maintenance, and experimentation with the chameleons were performed under permits from the Israeli Nature and Parks Authority (permit 2011/11411) and the University of Haifa ethics committee. Methods are provided here in brief; further details can be found elsewhere
Each tested chameleon was exposed to a threat that approached it from its left or right side in the following manner: the chameleon was placed on a vertical wooden pole that was between 3 mm and 14 mm in diameter. The pole could be rotated on its long axis either clockwise or counter-clockwise. Once the chameleon had settled, the pole was rotated in a 30° step (at ∼15°/s) in a given direction (Phase 1) and was then left stationary, allowing the chameleon to respond (Phase 2). The two phases were termed a “run” and each test comprised three consecutive runs. The experimenter acted as the “threat”, standing stationary ca. 120 cm from the pole so that the pole’s rotation resulted in relative movement of the chameleon toward the threat. Clockwise rotations resulted in a “left-approaching” threat toward the chameleon, while counter-clockwise rotations resulted in a “right-approaching” threat. The poles were either wide or narrow relative to the ventral width of the head of the tested chameleon. The wider pole allowed the chameleon to view the threat only monocularly at any given moment, whereas the narrower pole allowed the chameleon to view the threat both monocularly and binocularly. Each chameleon was tested once with a left-approaching threat and once with a right-approaching threat. Each test comprised three consecutive runs in the given direction. The tests were video-recorded with the camera positioned in front of the experimenter, 120 cm from the tested chameleon and at its level. From this position, the camera’s view was of the chameleon’s ventral side (
To determine whether the chameleon’s correction of position is a vestibular-driven compensatory response, we performed two control experiments: (a) the pole was rotated without a visual threat and (b) the threat was rotated, while the pole was kept stationary. In control experiment (a), the vertical pole was placed inside an opaque-white plastic sphere, 35 cm in diameter. The chameleons (n = 4), when perched on the pole, could view only the pole and its base but no obvious threats. In each test, the pole was rotated in succession 10 times clockwise and 10 times counter-clockwise at an angular velocity of ∼15°/s [see
Provided are the head angles relative to a moving threat under angular velocities of 15°/s, 35°/s and 70°/s. Each data point (mean ± SE) is from six readings (three per chameleon).
The degree of body exposure is depicted during three consecutive runs (respectively, triangles, circles, and squares) along with their mean (continuous line). The images are of the chameleon as viewed by the observer (the “threat”) at the respective time points.
Video sequences were edited using Adobe Elements™ software. A specially written program (SIPL Lab, Technion, Israel) sampled the sequences at intervals of four frames (i.e., 160 ms) and provided the size of the surface of the chameleon’s body (in number of pixels) that was exposed on each side of the pole (
The tested population comprised chameleons of different sizes, which would have resulted in unequal effects on the statistical tests of the pooled data, since smaller chameleons would have lower exposures by definition. To normalize the data, the tested population was divided into four groups according to head width. Each chameleon in each group had its exposure measurements multiplied by a computed factor which took into account the head width relative to the pole width used in each test. Consequently, data were normalized to the size of the largest chameleons.
The temporal aspect of the response, “latency to final exposure,” was calculated by counting the number of frames from the moment of termination of the pole rotation to the moment (frame) when the chameleon had reached its final exposure and remained still. The data extracted for each sampled frame in each run represented the exposed surface (in pixels) for each chameleon and for each side of the pole, within the above-defined area. In each test, only the side that approached the threat during a given pole rotation, termed the “leading side,” was used for analysis (
In each exposure of each individual in a given run, three measures were considered: 1) the exposure at the onset of the pole rotation, “Initial exposure”; 2) the exposure at the very end of the rotation, “End of rotation exposure” and 3) the final exposure at the very end of the run, “Final exposure”. Each of the three values was averaged over the three consecutive runs of any given test.
The data were analyzed using repeated measures MANOVA with pole width and direction of threat approach as the main effects. As individuals could be classified as side-biased on a given pole width (see Results), two further analyses were required for each of the biased groups separately.
The ventral surface exposed to the threat, on a narrow or wide pole, at the onset of pole rotation (Initial), end of pole rotation (Rotation end), and end of test (Final).
The ventral surface exposed to a right- or a left-approaching threat at the onset of pole rotation (Initial), end of pole rotation (Rotation end), and end of test (Final).
The latency to final exposure on a narrow or wide pole, under a right- or left-approaching threat (N = 17).
(a) Rotation of the chameleon on the pole within the opaque-white sphere elicited no apparent change of position: the chameleon maintained its position on the pole and rotated with it, clockwise or anti-clockwise [see
Distinct spatio-temporal motor patterns were observed in the exposure of the chameleon’s body. In Phase 1, an initial increase in ventral body exposure was observed and in Phase 2, there was a decrease in exposure (
Ventral surface exposure (mean ± SE) on narrow poles in response to right- or left-approaching threats in chameleons of the right-biased group (10.1, N = 14) and of the left-biased group (10.2, N = 10). Exposure readings are at 200-ms intervals, (A) at the onset of pole rotation, (B) at the end of pole rotation, and (C) at the end of the test.
Exposure (mean ± SE) during right- or left-approaching threats, for the right-biased (N = 14) and left-biased (N = 10) groups, in tests on narrow poles at the onset of pole rotation (Initial), end of pole rotation (Rotation end), and end of test (Final).
Latencies (mean ± SE) to final exposure of chameleons of the right-biased and left-biased groups, under right- or left-approaching threats, on narrow or wide poles (respective number of chameleons tested: 14, 10, 10, 7 for groups from left to right).
A repeated measures MANOVA, with pole width and direction of threat approach as the main effects (
For each chameleon and for a given pole width, a comparison was performed of the mean values of each of the three parameters (i.e., Initial exposure, End of rotation exposure, and Final exposure) between tests of right-approaching threat and left-approaching threat. The means were calculated from the values of a given parameter over the three consecutive runs comprising each test. If two or three of the parameters provided a lower mean value than the parameters for the comparable test on the opposite side, that individual chameleon was considered “side biased.” Of the individuals tested on narrow poles, a proportion of 0.75 were either all higher or all lower in all three spatial parameters than the comparable values in the opposite threat-approach direction. For the wide pole tests, the proportion of individuals was 0.76.
On narrow poles, the proportion of chameleons displaying a bias to right-approaching threats was 0.583 (14/24), whereas the proportion displaying a bias to left-approaching threats was 0.416 (10/24). On wide poles, the proportions were 0.588 (10/17) for the right-side bias and 0.411 (7/17) for the left-side bias. Right-side-biased or left-side-biased individuals were found throughout the tested population in tests on both narrow and wide poles. When examining the entire population, no side bias was observed due to the existence of two sub-groups, each biased toward a given threat-approach direction. Consequently, a further repeated measures MANOVA was performed for each pole width, with the direction of threat approach as a main effect and bias group as a covariate factor. The results showed that, for all three spatial parameters, there was a significant effect of the direction of threat approach (Initial exposure: F(1,23) = 26.273, p<0.001; End of rotation exposure: F(1,23) = 30.437, p<0.001; Final exposure: F(1,23) = 16.486, p<0.001). Moreover, the interaction between the direction of threat approach and the bias group was significant for all three spatial parameters (Initial exposure: F(1,23) = 26.426, p<0.001; End of rotation exposure: F(1,23) = 30.858, p<0.001; Final exposure: F(1,23) = 15.549, p = 0.001). No differences were found between right- and left-approaching threats in the latency to final exposure, with bias group as a covariate (F(1,23) = 0.58, p = 0.454). Similarly, the interaction between direction of threat approach and bias group was not significant (F(1,23) = 0.503, p = 0.485).
Because the interaction between threat-approach direction and bias group was significant with respect to the three spatial parameters, a separate repeated measures MANOVA was conducted for each of the bias groups (i.e., left and right) and for each pole width (i.e., narrow and wide).
Ventral surface exposure (mean ± SE) on wide poles in response to right- or left-approaching threats in the right-biased (13.1, N = 10) and in left-biased (13.2, N = 7) groups. Exposure readings are at 200-ms intervals, (A) at the onset of pole rotation, (B) at the end of pole rotation, and (C) at the end of the test.
Exposure (mean ± SE) under right- or left-approaching threats, for chameleons of the right-biased (N = 10) and left-biased (N = 7) groups, in tests on a wide pole at the onset of pole rotation (Initial), end of pole rotation (Rotation end), and end of test (Final).
In the tests on narrow poles, for the right-biased group, all three spatial parameters were significantly lower for right-approaching threats than for left-approaching threats (Initial exposure: F(1,13) = 16.721, p = 0.001; End of rotation exposure: F(1,13) = 18.049, p = 0.001; Final exposure: F(1,13) = 10.853, p = 0.006) (
In tests on wide poles, for the right-biased group, all three spatial parameters were significantly lower in tests on right-approaching threats than on left-approaching threats (Initial exposure: F(1,9) = 14.412, p = 0.004; End of rotation exposure: F(1,9) = 12.269, p = 0.007; Final exposure: F(1,9) = 17.058, p = 0.003) (
When a threat stimulus [
At the population level, lateralization was observed in the Final exposure, with better concealment (i.e., lower exposure) when threats approached from the right, under both monocular and binocular viewing. No such lateralization was observed in the Initial exposure or in the End of rotation exposure. In comparison, eye use under these conditions
That the chameleons were either “right biased” or “left biased”, as judged by their individual performance, points to the existence of two similar-sized sub-populations. Comparisons between responses to right- and left-approaching threats within each sub-population revealed significant differences in the three examined spatial parameters, but not in the temporal parameter (i.e., latency to final exposure). Each sub-population was lateralized with respect to a given threat-approach direction: individuals of the “right-biased” sub-population were better concealed from threats approaching from the right, while individuals of the “left-biased” sub-population were better concealed from threats approaching from the left. This bimodality of response is expressed in the exposure values (mean and SE) of any given sub-population to threats approaching from the right that do not overlap with the exposure values to threats approaching from the left. The divergence in response is depicted in
Here, the fact that the tested population is weakly lateralized (and only in the parameter of Final exposure), yet comprises sub-populations that are strongly lateralized (in all three parameters), makes sense in light of their relatively homogeneous natural habitats. In these arboreal habitats, the chances of confronting a threat from a given three-dimensional position are equal, and the chances of a threat confronting a left- or right-biased chameleon are also equal. This may be regarded as an evolutionary solution to preventing predators from using lateralization to their advantage.
A noticeable difference between the right- and the left-biased groups was observed with respect to final body exposure. The Final exposure values of the right-biased group were lower (i.e., better concealment) for right-approaching vs. left-approaching threats, for all pole widths. The Final exposure of the left-biased group on a wide pole was kept relatively low for both threat-approach directions. This, together with the similarity of exposure levels to right-approaching and left-approaching threats on a narrow pole, underlies the overall better concealment of the entire population when responding to right-approaching threats. Our results show that under monocular viewing (a wide pole), the left-biased group exerted a relatively similar and efficient avoidance response to threats from both sides, a feat not accomplished by the right-biased group. In other words, the performance of the left-biased group did not mirror that of the right-biased group. Under binocular viewing (on a narrow pole), the responses of each of the side-biased groups to threats from the right or left differed significantly, under all conditions. Comparably, in humans, left-handed individuals using their left hand perform better in given tasks than do right-handed individuals using their right hand
An individual belonging to a given biased group in narrow pole tests could respond as belonging to the opposite bias group during wide pole tests. The population thus comprises individuals with a stable bias toward a given side and individuals with a transient bias, depending on whether the visual input is monocular (wide pole) or binocular (narrow pole).
The motor responses observed here were accomplished by the chameleon’s using the pole as the axis of rotation. Although there are numerous examples of lateralized limb use, we did not consider this as a lateralization factor. The chameleons’ responses were also analyzed in terms of eye use
Obvious disadvantages of lateralization will occur when, for example, the probability of encountering prey or predator is similar for both sides of the body. In that case, having one side of a sensory system less efficient in identifying or responding to stimuli will be deleterious to the organism’s survival
Many species, such as ground-dwelling birds and amphibians, show lateralization at the behavioral level and live in a world dominated by two distinct visual domains. One is the nearby surfaces, such as ground or water, where food is found and social interactions occur. The other is the above-head space, where avian predators are likely to appear. Thus, for the fiddler crab (e.g.
Lateralized motor patterns include pawdness in bufonids during body righting or removal of disturbances
(WMV)
(WMV)
(WMV)
We are deeply indebted to Yossi Baydatch for initiating the chameleon research. We thank Ido Izhaki and Keren Or-Chen for statistical advice. Nimrod Peleg, Yaara David, Oded Yeruhami and Yuval Bahat were most helpful in providing the computer analysis software. Tova Katzir was very helpful in the language editing of the manuscript and Camille Vainstein in the proof reading.