- Email: [email protected]
A comparative study on the paths of five anura species
Behavioural Processes 41 (1997) 193 – 199
A comparative study on the paths of five anura species Delfi Sanuy a,*, Pierre Bovet b a
Departament de Produccio´ animal, Uni6ersitat de Lleida, A6. Ro6ira 177, 25198 Lleida, Catalonia, Spain b FAPSE, Uni6ersite´ de Gene`6e, 9 route de Drize, 1227 Carouge, Switzerland Received 17 June 1996; received in revised form 6 June 1997; accepted 8 June 1997
Abstract The movements of five species of European toads were recorded in the daytime and at night in a 3 × 3 m area. The paths obtained were analyzed according to the first order correlated random walk model developed by Bovet and Benhamou with which it is possible to characterize each path with two independent indices: its sinuosity and its speed. The analysis showed that the day/night variable affected the sinuosity of the paths but not their mean speed. Significant differences between species were found to exist, however, in the case of both the sinuosity and the speed. It is worth noting that these interspecific behavioural differences did not match the classical phylogenetic classification, as shown by multiple comparison tests. © 1997 Elsevier Science B.V. Keywords: Anura; Animal movements; Chronobiology; Path analysis; Species differences; Toad
1. Introduction Studies of animals’ paths can be of great interest from the ethological point of view, provided that the behaviour in question can be described and quantified. This requires accurate methods of measurement which fit the theoretical ecological models (Huntingford, 1984). Amphibians were the first vertebrate species to become able to travel on the ground (the anura in particular studied here are the most terrestrial of all the Southern European vertebrate species). * Corresponding author.
Their highly adaptive anatomical and physiological characteristics (skeleto-muscular structure, epithelial hydrostatic balance, ecothermia) and behavioural patterns depend a great deal on their surroundings (Brattstrom, 1968, 1970; Fitzgerald and Bider, 1974; Putmann and Bennet, 1981; Wisniewski et al., 1981; Gittins, 1983; Smits and Crawford, 1984; Semlitsch, 1985; Duellman and Trueb, 1986; Miller and Zoghby, 1986; Sinsch, 1988, 1990). The locomotor behaviour of these species is one of the many characteristics which can be said to adapt to changes in the environment, and can therefore be taken to constitute an index to these animals’ adaptability.
0376-6357/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 3 7 6 - 6 3 5 7 ( 9 7 ) 0 0 0 4 5 - 4
194
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
Table 1 Characteristics of species, origin and their ecological communities Specie
Specie size (mm) From
Ecological site
Alytes obstetricans Pelobates cultripes Bufo bufo Bufo calamita Bufo 6iridis sp. balearica
40–50 70–100 70–210 50–90 70–90
Caricion nigrae and Adenostylo – Valerianetum – pyrenaicae Agropyro – lygeion Quercetum rotundifoliae Buxo Queretum Pubescens Apietum nodiflori
Pyrenees Monegros (Z) Pyrenees Pre-Pyrenees Majorca
A toad released in an unknown place allowing no protection will tend to move look around for shelter. The animals’ movement from the release point can be measured in terms of one space-time variable—speed—and another space variable — sinuosity—(Bovet and Benhamou, 1988; Sanuy, 1990; Benhamou and Bovet, 1991). Although numerous studies have been carried out on animals’ locomotor speed, the sinuosity of their paths has been investigated in only a few studies, such as those by Bovet et al. (1989) on ants and by Martin et al. (1989) on newts. Against this background, the aim of the present study was two-fold. On the one hand, it was proposed to determine what differences might exist between several anura species’ locomotor behavior to a new environment and to examine whether these differences (if any) fitted the established phylogenetic scheme. For this purpose, experiments were carried out with five species of terrestrial anura: Pelobates cultripes, Alytes obstetrican, Bufo bufo, Bufo calamita and Bufo 6iridis. Individuals of these species become active at nightfall or during the night, and usually remain underground or hide under stones or in other suitable hollows providing, not only protection from predators, but also optimum conditions for regulating their hydrostatic balance. The phylogenetic scheme of the five species investigated here is taken from Duellman and Trueb (1986) and modified as of Llorente et al. (1996). Since toads carry out their activities at dusk and at night—thereby reducing their energy expenditure and the risk of being caught by a predator (Putmann and Bennet, 1981; Taylor et al., 1982; Paladino, 1985; Duellman and Trueb, 1986)—it was proposed in addition, to examine
whether any differences existed between the trajectories recorded at night and during the day.
2. Materials and methods
2.1. Animals The animals characteristics, recollected site and ecological characteristics site are show in Table 1. The animals were housed individually in terraria and fed with cricket and Tenebrio larvae and beetles. These experiments began as soon as possible after the animals were captured (1–2 days at most). Each animal underwent a single test every day.
2.2. Experimental set-up The experiments were carried out in December and January in a room measuring 9×6.5 m. A constant temperature of around 20°C was maintained day and night. A square 3 ×3 m in size was drawn on the cement floor and divided into a grid consisting of 25× 25 cm squares. The experimenter remained hidden from sight, about 4 m from the grid. When the experiments were carried out at night, a faint red light (that gave 0.03 lux on the ground) was switched on so that the toad’s movements could be detected. After releasing a toad in the centre of the grid, its movement was recorded along with any stops it made during the trial (see Fig. 1) and the time taken to reach the boundary of the grid. 302 paths were recorded. Since 25 trials were discarded because the animals remained motionless for more than 30 min before beginning to
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
move, only 277 of the paths were analysed. 65% of these paths were recorded during the daytime (i.e. between 7 a.m. and 6 p.m.) and 35% at night (between 6 p.m. and 7 a.m.). Table 2 gives the exact number of trials run with each member of each species by day and by night. All the statistical analyses were done on an individual basis (see below). Note that the animals marked (*) were eliminated in the analysis of sinuosity because of the shortness of their paths at night, and that the animals marked (**) were eliminated in both the analysis of speed and sinuosity because of the lack of data at night. The number of animals used for each analysis is given in Table 4 for the five species.
2.3. Path analysis The paths were quantified using a 50 cm Summagraphics digitizing table (measurement accuracy to within 0.1 mm). The numerized paths were discretized, i.e. transformed into a series of points (xi, yi ) where i =1...n, forming a broken line. The data was afterwards processed by computer using a geometrical procedure called
195
Table 2 Number of trials undergone by each member (a, b,..., u) of each of the species studied Species
Animals
D
Pelobates cultripes
a b (*) c (*) d e f g h i j k l (**) m n o p q r s t u
9 6 5 7 8 12 15 14 13 9 9 7 6 13 5 5 12 6 6 6 6
Alytes obstetricians
Bufo Bufo
Bufo calamita
Bufo 6iridis
N
(1)
(1) (1) (1) (2) (3)
1 1 1 4 1 5 4 5 5 4 3 0 3 9 10 8 10 6 6 6 6
(1) (2) (2) (2)
(1) (1) (5) (2)
D: daytime trails, N: night trails. Figures in brackets refer to trials in which the animals remained motionless for more than 30 min at the beginning of the trial. The animals marked (*) and (**) were eliminated in statistical analysis (see text).
rediscretization. This procedure was based on the correlated random walk model developed by one of the present authors (Bovet and Benhamou, 1988). By introducing a pre-determined step length P, the original path was transformed into a path consisting of a set of points (Xj, Yj ) where j= 1...N, and where each segment has a constant length equal to P. Hence for j=1... N−1, one always obtains P= (Xj + 1 − Xj )2 + (Yj + 1 − Yj )2
Fig. 1. Path taken by one individual belonging to the species Bufo bufo (individual b in Table 2). Stopping-points along the animal’s path (); the discretized (numerized) path (- - -); the rediscretized path (——), with step length P= 25 cm.
The value of P is required to be sufficiently small to yield a large enough set of angles for statistical analyses to be valid, and sufficiently large to prevent the alignment of path segments from giving rise artefactually to a large number of null angles. The set of N points obtained after rediscretization was used to calculate a set of N−2 rotation
196
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
angles between successive path steps. The data analysis dealt with the statistical distribution of the rotation angles. These data were tested using Kuiper’s test (Batschelet, 1981) in order to check the normality of the angular distributions thus obtained. The null hypothesis, involving the compatibility of the angular distribution with the normal circular distribution, is a prerequisite for Bovet and Benhamou’s correlated random walk model (1988) to be applicable. In this model, it is the successive step directions which are correlated and not the angles between successive steps, which on the contrary are uncorrelated. It is therefore necessary to check that the correlation between successive angles does not differ significantly from zero. The Jupp –Mardia test (Batschelet, 1981) was used to check whether the successive angles can be said to be independent random variables. If either of the above two criteria is not satisfied, the correlated random walk model cannot be applied. In this case, it was necessary to carry out further path rediscretizations, raising the steplengths until both criteria were satisfied, or until the number of angles became too small for the statistical tests to be applicable. In the latter case, the first order correlated random walk model must be rejected. If the data fit the model, the path can be characterized in terms of the index of sinuosity S developed by Bovet and Benhamou using the formula S = 1.18s/ P where s is the standard deviation of the angular distribution and P the step length adopted. From the length of the discretized path it is then possible to calculate the index V, corresponding to the animal’s speed, V = L/T where L is the path length and T the actual travelling time, i.e. not including the time spent motionless at the starting-point. It is worth noting that these toads’ patterns of movement were very sporadic, since they frequently stopped. These frequent stops affected considerably the rate of travel, since they
amounted on average to almost 80% of the total travelling time.
3. Results A set of discretized paths was obtained as described above, based on the animals’ actual paths. As an illustration, Fig. 2 gives all the paths taken by one individual belonging to the species Bufo bufo (the paths recorded at night are marked N), and Fig. 1, one of these paths, which has been digitized using the graphic table and rediscretized with a step length P= 25 cm.The step length P= 25 cm chosen was the smallest step for all the species considered which was compatible with the correlated random walk model. With this step length and the given number of paths (see Table 3), it was not possible however to calculate the sinuosity of each path because for many of them there were less than five turning angles (this being the minimum number required to be able to carry out the calculations). In order to ensure the validity of our statistical analyses, it was therefore decided to combine all the turning angles produced by a single animal during its various movements by day and by night. Even with this method two animals (marked * in Table 2) had to be eliminated in the analysis of the sinuosity because their only path was too short. Table 4 gives the value of the sinuosity (in rad/ m) recorded by day and by night in each species, along with each animal’s mean speed (in cm/s) by day and by night. Table 5 gives the results of the analysis of variance carried out on these two sets of data. A further a posteriori analysis was carried out on the differences in the two variables between species, using the Student–Newman–Keuls (SNK) method (Sokal and Rohlf, 1969) (Table 6). The results of the analysis of variance showed that the daytime data differed significantly from the night-time data in the case of the sinuosity but not in that of the speed: the animals’ paths were more rectilinear at night. On the other hand, the interactions between the day/night and species factors were not significant, whereas the species to which the animals belonged significantly affected
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
197
Fig. 2. Complete set of paths taken by one individual toad (individual j in Table 2) belonging to the species Bufo bufo. Paths taken at night are marked N. Path three is very short and does not appear in the drawing.
both the sinuosity and the speed. The results of a priori comparisons based on Duellman and Trueb’s (1986) phylogenetic classification were not in good agreement with this classification, however: it was not possible to differentiate clearly in this way between either the three families of toads or between the three Bufo species. Nor was it possible on the basis of a posteriori analysis of the
Table 3 Number of trials excluded because involving less than five turning angles on the basis of a 25 cm step length P Species
Total number of trails
Number of excluded trails
Pelobates cultripes Alytes obstetricans Bufo bufo Bufo calamita Bufo 6iridis
43 73 41 72 48
6 35 13 1 0
animals’ movements to differentiate between the species: in the case of the speed, only B. calamita and B. 6iridis were clearly distinguishable from the other species, and in the case of the sinuosity, only P. cultripes stood apart from the others.
4. Discussion After analysing the data on each animal and under each experimental condition (day or night), the sinuosity of the toads’ paths was investigated using the correlated random walk model developed by Bovet and Benhamou (1988). Taking a rediscretization step length of 25 cm, the model was found to be applicable for measuring the trajectories of the five anuran species studied here. It would be interesting, however, to investigate longer paths (both temporally and spatially speaking), with which it should be possible to analyse each trial separately.
198
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
Table 4 Mean speed (in cm/s) and sinuosity (in rad/ m) of each species by day and by night (for each species the number of analysed animals appears in brackets) Speed
Day
Night
Mean
Sinuosity
Day
Night
Mean
Pelobates cultripes (5) Alytes obstetricans (4) Bufo bufo (3) Bufo calamita (4) Bufo 6iridis (4)
0.950 0.578 0.403 3.925 3.653
0.536 1.305 0.337 2.280 4.830
0.743 0.941 0.370 3.103 4.241
Pelobates cultripes (3) Alytes obstetricans (4) Bufo bufo (3) Bufo calamita (4) Bufo 6iridis (4)
0.324 0.270 0.238 0.169 0.223
0.157 0.220 0.248 0.171 0.108
0.241 0.245 0.243 0.170 0.166
Mean
1.929
1.868
Mean
0.241
0.178
The correlated random walk model, which is a probabilistic model, shows up the intrinsically random nature of the toad movements recorded here, but also made it possible to characterize these animals’ paths in terms of a very simple index, their sinuosity. By introducing another simple numerical index, the speed, it was possible to describe the toads’ movements concomitantly (but separately) in terms of both their spatial and temporal features.
The results of the present study indicate that the day versus night factor did not significantly affect the animals’ speed, whereas this factor significantly affected the sinuosity of their paths. This was actually the case in only three out of the five species studied: the Pelobates cultripes, Alytes obstetricans and Bufo 6iridis all produced more sinuous paths during the day than at night. The differences observed here between the species do not fit the recognized phylogenetic scheme:
Table 5 Results of the analysis of variance carried out on the data in Table 4 Sum of squares
df
Speed Between subjects Species (A) Pc/Ao/Bb, Bc, B6 Bb/Bc/B6 Subjects within species Within subjects Day/Night (B) A×B B×Subjects within species (residual) Total
165.3980 90.2050 29.9050 52.7610 75.1930 25.7558 0.0378 9.6409 16.0770 191.1538
19 4 2 2 15 20 1 4 15 39
Sinuosity Between subjects Species (A) Pc/Ao/Bb, Bc, B6 Bb/Bc/B6 Subjects within species Within subjects Day/Night (B) A×B B×Subjects within species (residual) Total
0.0959 0.0507 0.0215 0.0249 0.0453 0.1387 0.0350 0.0384 0.0652 0.2346
17 4 2 2 13 18 1 4 13 35
Mean square
F
P
22.5513 14.9525 26.3805 5.0129
4.50 2.98 5.26
B0.05 B0.10 NS
0.0378 2.4102 1.0718
0.04 2.25
NS NS
0.0l38 0.0107 0.0125 0.0035
3.64 3.09 3.58
B0.05 B0.10 B0.10
0.0351 0.0096 0.0050 0.0067
6.98 1.91
B0.025 NS
The effects on the speed and the sinuosity of the following factors are given: day versus night, the species, interactions between day versus night and species, along with the corresponding totals and residual factors.
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
199
Table 6 Diagram showing the results of multiple comparison tests, Student – Newman – Keuls Analysis (Sokal and Rohlf, 1969) on the data in Table 4 Speed
Sinuosity
All paths combined Daytime paths Night paths
Bb Pc Ao Bc B6 Bb Pc Ao Bc B6 Bb Pc Ao Bc B6
All paths combined Daytime paths Night paths
Ao Bb Pc Bc B6 Pc Ao Bb Bc B6 Bb Ao Bc Pc B6
The lines link up species in which the means did not differ significantly. Ao = Alytes obstetricans, Pc=Pelobates cultripes, Bb =Bufo bufo, Bc= Bufo calamita, B6=Bufo 6iridis.
no significant differences were found to exist for example between Pelobates cultripes and Alytes obstetricans. Interspecific differences were observed however among the members of the Bufo genus, since the common toad Bufo bufo produced more sinuous paths and travelled at a higher speed than the other members of the genus.
Acknowledgements The authors wish to thank Dr Jessica Blanc for translating the manuscript into English.
References Batschelet, E., 1981. Circular Statistics in Biology, Academic Press, London. Benhamou, S., Bovet, P., 1991. Biological functions of sinuosity regulation; a reply to BPR0819.W51Budenberg. Anim. Behav. 42, 159 – 160. Bovet, P., Benhamou, S., 1988. Spatial analysis of animals’ movements using a correlated random walk model. J. Theor. Biol. 131, 419–443. Bovet, P., Dejean, A., Granjon, M., 1989. Trajets d’approvisionnement a` partir d’un nid central chez la fourmi Serrastruma lujae (Formicidae: Myrmicinae). Insectes Sociaux 36 (1), 51 – 61. Brattstrom, B., 1968. Thermal acclimation in anuran amphibians as a function of latitude and altitude. Comp. Biochem. Physiol. 24, 93 – 111. Brattstrom, B., 1970. Thermal acclimation in Australian amphibians. Comp. Biochem. Physiol. 35, 69–103. Duellman, W. E., Trueb, L., 1986. Biology of Amphibians, McGraw-Hill Book Co., New York. Fitzgerald, G.J., Bider, J.R., 1974. Influence of moon phase
.
and weather factors on locomotory activity in Bufo americanus. Oikos 25, 338 – 340. Gittins, S.P., 1983. The breeding migration of the common toad (Bufo bufo) to a pond in Mid-Wales. J. Zool. 199, 555 – 562. Huntingford, F., 1984. The Study of Animal Behaviour, Chapman and Hall, London. Llorente, G., Montori, A., Angel, M., 1996. Els Amfibis i Re`ptils de Catalunya i Andorra, Kaetres ed., Barcelona. Martin, E., Joly, P., Bovet, P., 1989. Diel activity in the alpine newt Triturus alpestris. Biol. Behav. 14, 116 – 131. Miller, K., Zoghby, G.M., 1986. Thermal acclimation of locomotor performance in anuran amphibians. Can. J. Zool. 64, 1956 – 1960. Paladino, F.V., 1985. Temperature effects on locomotion and activity bioenergetics of amphibians. Reptiles and Birds. Amer. Zool. 25, 965 – 972. Putmann, R.W., Bennet, A.F., 1981. Thermal dependence of behavioural performance of anuran amphibians. Anim. Behav. 29, 502 – 509. Sanuy, D., 1990. Ana`lisi estadı´stica del desplac¸ament en cinc espe`cies d’anurs ibe`rics. Doctoral Thesis, Univ. Auto`noma, Barcelona. Semlitsch, R.D., 1985. Analysis of climatic factors influencing migrations of the salamander Ambystoma talpoideum. Copeia 1985, 477 – 489. Sinsch, U., 1988. Seasonal changes in the migratory behaviour of the toad Bufo bufo: direction and magnitude of movements. Oecologia 76, 390 – 398. Sinsch, U., 1990. Migration and orientation in anuran amphibians. Ethol. Ecol. Evol. 2, 65 – 79. Smits, A.W., Crawford, D.L., 1984. Activity patterns and thermal biology of the toad Bufo boreas halophilus. Copeia 1984, 689 – 696. Sokal, R.R., Rohlf, F.J., 1969. Biometrı´a, H. Blume, Madrid. Taylor, C.R., Heglund, N.C., Maloiy, G.M.O., 1982. Energetics and mechanics of terrestrial locomotion. J. Exp. Biol. 97, 1 – 21. Wisniewski, P.J., Paull, L.M., Slater, F.M., 1981. The effects of temperature on the breeding and spawning of the common toad (Bufo bufo). Brit. J. Herpet. 6, 119 – 121.
A comparative study on the paths of five anura species Delfi Sanuy a,*, Pierre Bovet b a
Departament de Produccio´ animal, Uni6ersitat de Lleida, A6. Ro6ira 177, 25198 Lleida, Catalonia, Spain b FAPSE, Uni6ersite´ de Gene`6e, 9 route de Drize, 1227 Carouge, Switzerland Received 17 June 1996; received in revised form 6 June 1997; accepted 8 June 1997
Abstract The movements of five species of European toads were recorded in the daytime and at night in a 3 × 3 m area. The paths obtained were analyzed according to the first order correlated random walk model developed by Bovet and Benhamou with which it is possible to characterize each path with two independent indices: its sinuosity and its speed. The analysis showed that the day/night variable affected the sinuosity of the paths but not their mean speed. Significant differences between species were found to exist, however, in the case of both the sinuosity and the speed. It is worth noting that these interspecific behavioural differences did not match the classical phylogenetic classification, as shown by multiple comparison tests. © 1997 Elsevier Science B.V. Keywords: Anura; Animal movements; Chronobiology; Path analysis; Species differences; Toad
1. Introduction Studies of animals’ paths can be of great interest from the ethological point of view, provided that the behaviour in question can be described and quantified. This requires accurate methods of measurement which fit the theoretical ecological models (Huntingford, 1984). Amphibians were the first vertebrate species to become able to travel on the ground (the anura in particular studied here are the most terrestrial of all the Southern European vertebrate species). * Corresponding author.
Their highly adaptive anatomical and physiological characteristics (skeleto-muscular structure, epithelial hydrostatic balance, ecothermia) and behavioural patterns depend a great deal on their surroundings (Brattstrom, 1968, 1970; Fitzgerald and Bider, 1974; Putmann and Bennet, 1981; Wisniewski et al., 1981; Gittins, 1983; Smits and Crawford, 1984; Semlitsch, 1985; Duellman and Trueb, 1986; Miller and Zoghby, 1986; Sinsch, 1988, 1990). The locomotor behaviour of these species is one of the many characteristics which can be said to adapt to changes in the environment, and can therefore be taken to constitute an index to these animals’ adaptability.
0376-6357/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 3 7 6 - 6 3 5 7 ( 9 7 ) 0 0 0 4 5 - 4
194
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
Table 1 Characteristics of species, origin and their ecological communities Specie
Specie size (mm) From
Ecological site
Alytes obstetricans Pelobates cultripes Bufo bufo Bufo calamita Bufo 6iridis sp. balearica
40–50 70–100 70–210 50–90 70–90
Caricion nigrae and Adenostylo – Valerianetum – pyrenaicae Agropyro – lygeion Quercetum rotundifoliae Buxo Queretum Pubescens Apietum nodiflori
Pyrenees Monegros (Z) Pyrenees Pre-Pyrenees Majorca
A toad released in an unknown place allowing no protection will tend to move look around for shelter. The animals’ movement from the release point can be measured in terms of one space-time variable—speed—and another space variable — sinuosity—(Bovet and Benhamou, 1988; Sanuy, 1990; Benhamou and Bovet, 1991). Although numerous studies have been carried out on animals’ locomotor speed, the sinuosity of their paths has been investigated in only a few studies, such as those by Bovet et al. (1989) on ants and by Martin et al. (1989) on newts. Against this background, the aim of the present study was two-fold. On the one hand, it was proposed to determine what differences might exist between several anura species’ locomotor behavior to a new environment and to examine whether these differences (if any) fitted the established phylogenetic scheme. For this purpose, experiments were carried out with five species of terrestrial anura: Pelobates cultripes, Alytes obstetrican, Bufo bufo, Bufo calamita and Bufo 6iridis. Individuals of these species become active at nightfall or during the night, and usually remain underground or hide under stones or in other suitable hollows providing, not only protection from predators, but also optimum conditions for regulating their hydrostatic balance. The phylogenetic scheme of the five species investigated here is taken from Duellman and Trueb (1986) and modified as of Llorente et al. (1996). Since toads carry out their activities at dusk and at night—thereby reducing their energy expenditure and the risk of being caught by a predator (Putmann and Bennet, 1981; Taylor et al., 1982; Paladino, 1985; Duellman and Trueb, 1986)—it was proposed in addition, to examine
whether any differences existed between the trajectories recorded at night and during the day.
2. Materials and methods
2.1. Animals The animals characteristics, recollected site and ecological characteristics site are show in Table 1. The animals were housed individually in terraria and fed with cricket and Tenebrio larvae and beetles. These experiments began as soon as possible after the animals were captured (1–2 days at most). Each animal underwent a single test every day.
2.2. Experimental set-up The experiments were carried out in December and January in a room measuring 9×6.5 m. A constant temperature of around 20°C was maintained day and night. A square 3 ×3 m in size was drawn on the cement floor and divided into a grid consisting of 25× 25 cm squares. The experimenter remained hidden from sight, about 4 m from the grid. When the experiments were carried out at night, a faint red light (that gave 0.03 lux on the ground) was switched on so that the toad’s movements could be detected. After releasing a toad in the centre of the grid, its movement was recorded along with any stops it made during the trial (see Fig. 1) and the time taken to reach the boundary of the grid. 302 paths were recorded. Since 25 trials were discarded because the animals remained motionless for more than 30 min before beginning to
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
move, only 277 of the paths were analysed. 65% of these paths were recorded during the daytime (i.e. between 7 a.m. and 6 p.m.) and 35% at night (between 6 p.m. and 7 a.m.). Table 2 gives the exact number of trials run with each member of each species by day and by night. All the statistical analyses were done on an individual basis (see below). Note that the animals marked (*) were eliminated in the analysis of sinuosity because of the shortness of their paths at night, and that the animals marked (**) were eliminated in both the analysis of speed and sinuosity because of the lack of data at night. The number of animals used for each analysis is given in Table 4 for the five species.
2.3. Path analysis The paths were quantified using a 50 cm Summagraphics digitizing table (measurement accuracy to within 0.1 mm). The numerized paths were discretized, i.e. transformed into a series of points (xi, yi ) where i =1...n, forming a broken line. The data was afterwards processed by computer using a geometrical procedure called
195
Table 2 Number of trials undergone by each member (a, b,..., u) of each of the species studied Species
Animals
D
Pelobates cultripes
a b (*) c (*) d e f g h i j k l (**) m n o p q r s t u
9 6 5 7 8 12 15 14 13 9 9 7 6 13 5 5 12 6 6 6 6
Alytes obstetricians
Bufo Bufo
Bufo calamita
Bufo 6iridis
N
(1)
(1) (1) (1) (2) (3)
1 1 1 4 1 5 4 5 5 4 3 0 3 9 10 8 10 6 6 6 6
(1) (2) (2) (2)
(1) (1) (5) (2)
D: daytime trails, N: night trails. Figures in brackets refer to trials in which the animals remained motionless for more than 30 min at the beginning of the trial. The animals marked (*) and (**) were eliminated in statistical analysis (see text).
rediscretization. This procedure was based on the correlated random walk model developed by one of the present authors (Bovet and Benhamou, 1988). By introducing a pre-determined step length P, the original path was transformed into a path consisting of a set of points (Xj, Yj ) where j= 1...N, and where each segment has a constant length equal to P. Hence for j=1... N−1, one always obtains P= (Xj + 1 − Xj )2 + (Yj + 1 − Yj )2
Fig. 1. Path taken by one individual belonging to the species Bufo bufo (individual b in Table 2). Stopping-points along the animal’s path (); the discretized (numerized) path (- - -); the rediscretized path (——), with step length P= 25 cm.
The value of P is required to be sufficiently small to yield a large enough set of angles for statistical analyses to be valid, and sufficiently large to prevent the alignment of path segments from giving rise artefactually to a large number of null angles. The set of N points obtained after rediscretization was used to calculate a set of N−2 rotation
196
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
angles between successive path steps. The data analysis dealt with the statistical distribution of the rotation angles. These data were tested using Kuiper’s test (Batschelet, 1981) in order to check the normality of the angular distributions thus obtained. The null hypothesis, involving the compatibility of the angular distribution with the normal circular distribution, is a prerequisite for Bovet and Benhamou’s correlated random walk model (1988) to be applicable. In this model, it is the successive step directions which are correlated and not the angles between successive steps, which on the contrary are uncorrelated. It is therefore necessary to check that the correlation between successive angles does not differ significantly from zero. The Jupp –Mardia test (Batschelet, 1981) was used to check whether the successive angles can be said to be independent random variables. If either of the above two criteria is not satisfied, the correlated random walk model cannot be applied. In this case, it was necessary to carry out further path rediscretizations, raising the steplengths until both criteria were satisfied, or until the number of angles became too small for the statistical tests to be applicable. In the latter case, the first order correlated random walk model must be rejected. If the data fit the model, the path can be characterized in terms of the index of sinuosity S developed by Bovet and Benhamou using the formula S = 1.18s/ P where s is the standard deviation of the angular distribution and P the step length adopted. From the length of the discretized path it is then possible to calculate the index V, corresponding to the animal’s speed, V = L/T where L is the path length and T the actual travelling time, i.e. not including the time spent motionless at the starting-point. It is worth noting that these toads’ patterns of movement were very sporadic, since they frequently stopped. These frequent stops affected considerably the rate of travel, since they
amounted on average to almost 80% of the total travelling time.
3. Results A set of discretized paths was obtained as described above, based on the animals’ actual paths. As an illustration, Fig. 2 gives all the paths taken by one individual belonging to the species Bufo bufo (the paths recorded at night are marked N), and Fig. 1, one of these paths, which has been digitized using the graphic table and rediscretized with a step length P= 25 cm.The step length P= 25 cm chosen was the smallest step for all the species considered which was compatible with the correlated random walk model. With this step length and the given number of paths (see Table 3), it was not possible however to calculate the sinuosity of each path because for many of them there were less than five turning angles (this being the minimum number required to be able to carry out the calculations). In order to ensure the validity of our statistical analyses, it was therefore decided to combine all the turning angles produced by a single animal during its various movements by day and by night. Even with this method two animals (marked * in Table 2) had to be eliminated in the analysis of the sinuosity because their only path was too short. Table 4 gives the value of the sinuosity (in rad/ m) recorded by day and by night in each species, along with each animal’s mean speed (in cm/s) by day and by night. Table 5 gives the results of the analysis of variance carried out on these two sets of data. A further a posteriori analysis was carried out on the differences in the two variables between species, using the Student–Newman–Keuls (SNK) method (Sokal and Rohlf, 1969) (Table 6). The results of the analysis of variance showed that the daytime data differed significantly from the night-time data in the case of the sinuosity but not in that of the speed: the animals’ paths were more rectilinear at night. On the other hand, the interactions between the day/night and species factors were not significant, whereas the species to which the animals belonged significantly affected
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
197
Fig. 2. Complete set of paths taken by one individual toad (individual j in Table 2) belonging to the species Bufo bufo. Paths taken at night are marked N. Path three is very short and does not appear in the drawing.
both the sinuosity and the speed. The results of a priori comparisons based on Duellman and Trueb’s (1986) phylogenetic classification were not in good agreement with this classification, however: it was not possible to differentiate clearly in this way between either the three families of toads or between the three Bufo species. Nor was it possible on the basis of a posteriori analysis of the
Table 3 Number of trials excluded because involving less than five turning angles on the basis of a 25 cm step length P Species
Total number of trails
Number of excluded trails
Pelobates cultripes Alytes obstetricans Bufo bufo Bufo calamita Bufo 6iridis
43 73 41 72 48
6 35 13 1 0
animals’ movements to differentiate between the species: in the case of the speed, only B. calamita and B. 6iridis were clearly distinguishable from the other species, and in the case of the sinuosity, only P. cultripes stood apart from the others.
4. Discussion After analysing the data on each animal and under each experimental condition (day or night), the sinuosity of the toads’ paths was investigated using the correlated random walk model developed by Bovet and Benhamou (1988). Taking a rediscretization step length of 25 cm, the model was found to be applicable for measuring the trajectories of the five anuran species studied here. It would be interesting, however, to investigate longer paths (both temporally and spatially speaking), with which it should be possible to analyse each trial separately.
198
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
Table 4 Mean speed (in cm/s) and sinuosity (in rad/ m) of each species by day and by night (for each species the number of analysed animals appears in brackets) Speed
Day
Night
Mean
Sinuosity
Day
Night
Mean
Pelobates cultripes (5) Alytes obstetricans (4) Bufo bufo (3) Bufo calamita (4) Bufo 6iridis (4)
0.950 0.578 0.403 3.925 3.653
0.536 1.305 0.337 2.280 4.830
0.743 0.941 0.370 3.103 4.241
Pelobates cultripes (3) Alytes obstetricans (4) Bufo bufo (3) Bufo calamita (4) Bufo 6iridis (4)
0.324 0.270 0.238 0.169 0.223
0.157 0.220 0.248 0.171 0.108
0.241 0.245 0.243 0.170 0.166
Mean
1.929
1.868
Mean
0.241
0.178
The correlated random walk model, which is a probabilistic model, shows up the intrinsically random nature of the toad movements recorded here, but also made it possible to characterize these animals’ paths in terms of a very simple index, their sinuosity. By introducing another simple numerical index, the speed, it was possible to describe the toads’ movements concomitantly (but separately) in terms of both their spatial and temporal features.
The results of the present study indicate that the day versus night factor did not significantly affect the animals’ speed, whereas this factor significantly affected the sinuosity of their paths. This was actually the case in only three out of the five species studied: the Pelobates cultripes, Alytes obstetricans and Bufo 6iridis all produced more sinuous paths during the day than at night. The differences observed here between the species do not fit the recognized phylogenetic scheme:
Table 5 Results of the analysis of variance carried out on the data in Table 4 Sum of squares
df
Speed Between subjects Species (A) Pc/Ao/Bb, Bc, B6 Bb/Bc/B6 Subjects within species Within subjects Day/Night (B) A×B B×Subjects within species (residual) Total
165.3980 90.2050 29.9050 52.7610 75.1930 25.7558 0.0378 9.6409 16.0770 191.1538
19 4 2 2 15 20 1 4 15 39
Sinuosity Between subjects Species (A) Pc/Ao/Bb, Bc, B6 Bb/Bc/B6 Subjects within species Within subjects Day/Night (B) A×B B×Subjects within species (residual) Total
0.0959 0.0507 0.0215 0.0249 0.0453 0.1387 0.0350 0.0384 0.0652 0.2346
17 4 2 2 13 18 1 4 13 35
Mean square
F
P
22.5513 14.9525 26.3805 5.0129
4.50 2.98 5.26
B0.05 B0.10 NS
0.0378 2.4102 1.0718
0.04 2.25
NS NS
0.0l38 0.0107 0.0125 0.0035
3.64 3.09 3.58
B0.05 B0.10 B0.10
0.0351 0.0096 0.0050 0.0067
6.98 1.91
B0.025 NS
The effects on the speed and the sinuosity of the following factors are given: day versus night, the species, interactions between day versus night and species, along with the corresponding totals and residual factors.
D. Sanuy, P. Bo6et / Beha6ioural Processes 41 (1997) 193–199
199
Table 6 Diagram showing the results of multiple comparison tests, Student – Newman – Keuls Analysis (Sokal and Rohlf, 1969) on the data in Table 4 Speed
Sinuosity
All paths combined Daytime paths Night paths
Bb Pc Ao Bc B6 Bb Pc Ao Bc B6 Bb Pc Ao Bc B6
All paths combined Daytime paths Night paths
Ao Bb Pc Bc B6 Pc Ao Bb Bc B6 Bb Ao Bc Pc B6
The lines link up species in which the means did not differ significantly. Ao = Alytes obstetricans, Pc=Pelobates cultripes, Bb =Bufo bufo, Bc= Bufo calamita, B6=Bufo 6iridis.
no significant differences were found to exist for example between Pelobates cultripes and Alytes obstetricans. Interspecific differences were observed however among the members of the Bufo genus, since the common toad Bufo bufo produced more sinuous paths and travelled at a higher speed than the other members of the genus.
Acknowledgements The authors wish to thank Dr Jessica Blanc for translating the manuscript into English.
References Batschelet, E., 1981. Circular Statistics in Biology, Academic Press, London. Benhamou, S., Bovet, P., 1991. Biological functions of sinuosity regulation; a reply to BPR0819.W51Budenberg. Anim. Behav. 42, 159 – 160. Bovet, P., Benhamou, S., 1988. Spatial analysis of animals’ movements using a correlated random walk model. J. Theor. Biol. 131, 419–443. Bovet, P., Dejean, A., Granjon, M., 1989. Trajets d’approvisionnement a` partir d’un nid central chez la fourmi Serrastruma lujae (Formicidae: Myrmicinae). Insectes Sociaux 36 (1), 51 – 61. Brattstrom, B., 1968. Thermal acclimation in anuran amphibians as a function of latitude and altitude. Comp. Biochem. Physiol. 24, 93 – 111. Brattstrom, B., 1970. Thermal acclimation in Australian amphibians. Comp. Biochem. Physiol. 35, 69–103. Duellman, W. E., Trueb, L., 1986. Biology of Amphibians, McGraw-Hill Book Co., New York. Fitzgerald, G.J., Bider, J.R., 1974. Influence of moon phase
.
and weather factors on locomotory activity in Bufo americanus. Oikos 25, 338 – 340. Gittins, S.P., 1983. The breeding migration of the common toad (Bufo bufo) to a pond in Mid-Wales. J. Zool. 199, 555 – 562. Huntingford, F., 1984. The Study of Animal Behaviour, Chapman and Hall, London. Llorente, G., Montori, A., Angel, M., 1996. Els Amfibis i Re`ptils de Catalunya i Andorra, Kaetres ed., Barcelona. Martin, E., Joly, P., Bovet, P., 1989. Diel activity in the alpine newt Triturus alpestris. Biol. Behav. 14, 116 – 131. Miller, K., Zoghby, G.M., 1986. Thermal acclimation of locomotor performance in anuran amphibians. Can. J. Zool. 64, 1956 – 1960. Paladino, F.V., 1985. Temperature effects on locomotion and activity bioenergetics of amphibians. Reptiles and Birds. Amer. Zool. 25, 965 – 972. Putmann, R.W., Bennet, A.F., 1981. Thermal dependence of behavioural performance of anuran amphibians. Anim. Behav. 29, 502 – 509. Sanuy, D., 1990. Ana`lisi estadı´stica del desplac¸ament en cinc espe`cies d’anurs ibe`rics. Doctoral Thesis, Univ. Auto`noma, Barcelona. Semlitsch, R.D., 1985. Analysis of climatic factors influencing migrations of the salamander Ambystoma talpoideum. Copeia 1985, 477 – 489. Sinsch, U., 1988. Seasonal changes in the migratory behaviour of the toad Bufo bufo: direction and magnitude of movements. Oecologia 76, 390 – 398. Sinsch, U., 1990. Migration and orientation in anuran amphibians. Ethol. Ecol. Evol. 2, 65 – 79. Smits, A.W., Crawford, D.L., 1984. Activity patterns and thermal biology of the toad Bufo boreas halophilus. Copeia 1984, 689 – 696. Sokal, R.R., Rohlf, F.J., 1969. Biometrı´a, H. Blume, Madrid. Taylor, C.R., Heglund, N.C., Maloiy, G.M.O., 1982. Energetics and mechanics of terrestrial locomotion. J. Exp. Biol. 97, 1 – 21. Wisniewski, P.J., Paull, L.M., Slater, F.M., 1981. The effects of temperature on the breeding and spawning of the common toad (Bufo bufo). Brit. J. Herpet. 6, 119 – 121.