INTRODUCTION
In many endothermic species (birds and mammals), offspring remain with their parents until they achieve body sizes relatively close to adult size. In such species, the ecology of young individuals after the termination of parental care may be broadly similar to the ecology of their parents. In most ectotherms, however, offspring begin their independent life at a size far below mean adult size: often, an order of magnitude smaller than their parents (Shine Reference SHINE2005). In such cases, we might expect the ecology of juvenile animals to differ substantially from the ecology of adult conspecifics. For example, neonates may be exposed to greater risk from predators, eat different types of prey, be less capable of buffering rates of thermal and hydric exchange with the environment, and be less able to tolerate long periods without feeding (Hjelm et al. Reference HJELM, PERSSON and CHRISTENSEN2000, Lankford & Targett Reference LANKFORD and TARGETT2001, Pizzatto & Shine Reference PIZZATTO and SHINE2008, Seebacher & Alford Reference SEEBACHER and ALFORD2002, Werner & Gilliam Reference WERNER and GILLIAM1984). Such size-related shifts may generate ontogenetic variation in behaviour, growth rates, diel cycles, habitat use, movement patterns and home ranges (Harestad & Bunnel Reference HARESTAD and BUNNEL1979, Hjelm et al. Reference HJELM, PERSSON and CHRISTENSEN2000, Keren-Rotem et al. Reference KEREN-ROTEM, BOUSKILA and GEFFEN2006, Pizzatto & Shine Reference PIZZATTO and SHINE2008, Schoener & Schoener Reference SCHOENER and SCHOENER1984, Werner & Gilliam Reference WERNER and GILLIAM1984, Wilson et al. Reference WILSON, HEINSOHN and LEGGE2006). Understanding how animals occupy and use space is important for conservation purposes (Groom et al. Reference GROOM, MEFFE and CAROLL2006). For an ectotherm, that understanding should include information on ontogenetic shifts in spatial ecology. For example, we can only assure conservation of a population by protecting the areas where individuals live at all stages of their lives.
Most snakes have very secretive habits, but the increasing availability of miniature radio-transmitters has yielded substantial progress in understanding patterns of movements and habitat use in many species (Brown et al. Reference BROWN, SHINE and MADSEN2005, Glaudas et al. Reference GLAUDAS, ANDREWS, WILLSON and GIBBONS2007, Macartney et al. Reference MACARTNEY, GREGORY and LARSEN1988, Shine Reference SHINE1987, Wilson et al. Reference WILSON, HEINSOHN and LEGGE2006). The extent of the home range varies among and within species depending on factors such as reproductive status, sex, body size, season, prey availability and geographic location (Brown et al. Reference BROWN, SHINE and MADSEN2005, King & Duvall Reference KING and DUVALL1990, Macartney et al. Reference MACARTNEY, GREGORY and LARSEN1988, Madsen & Shine Reference MADSEN and SHINE1996a, Shine Reference SHINE1987, Whitaker & Shine Reference WHITAKER and SHINE2003). However, small (young) snakes are rarely included in such studies (Macartney et al. Reference MACARTNEY, GREGORY and LARSEN1988). One obvious problem is that methods used for large snakes (such as surgical implantation of transmitters) often are inapplicable to small ones (Macartney et al. Reference MACARTNEY, GREGORY and LARSEN1988, Shine et al. Reference SHINE, SUN, KEARNEY and FITZGERALD2002). Another problem is that young snakes are difficult to find in the field: their small size facilitates hiding, their greater vulnerability to predators may favour secretive behaviour, and their higher rates of heating may limit time of exposure during basking (Asplund Reference ASPLUND1974, Sun et al. Reference SUN, SHINE, ZHAO and TANG2001). Such biases affect most mark-recapture studies of snakes (Pike et al. Reference PIKE, PIZZATTO, PIKE and SHINE2008). Thus, although newborn snakes are often seen immediately after birth/hatching when they leave the nest, they then tend to disperse and are often not encountered again until they reach larger size. Aiming to understand what young snakes do in their first months of life we studied the spatial ecology of hatchlings of the water python (Liasis fuscus Peters 1873: Pythonidae) in the wet-dry topics of northern Australia, comparing our results to those available for conspecific adults in a neighbourhood area. We hypothesize that (1) like the conspecific adults, hatchling water pythons use the floodplain as main habitat, take refuge in soil cracks and under the vegetation, (2) differently from the adults and due to their smaller size they do not undertake very long distance dispersion/migration and (3) as typical tropical ectothermics they are exposed to very high temperatures and their movements are limited/related to the weather conditions.
MATERIALS AND METHODS
Study species
The water python Liasis fuscus is a large nocturnal snake (up to 3 m in snout-vent length (SVL)) widely distributed in tropical Australia (Cogger Reference COGGER1992). Many aspects of the ecology of adult and sub-adult pythons have been studied in great detail in the Adelaide River floodplain in the Northern Territory, Australia. The dusky rat (Rattus colletti) is the major prey of these snakes (Shine Reference SHINE and Uricheck1993) and during the wet season the pythons follow the rats in a migration of up to 12 km (Madsen & Shine Reference MADSEN and SHINE1996a). Python breeding is highly seasonal, with reproductive frequency and output largely dependent on prey availability (and thus, maternal body condition), but also affected by availability of specific thermal conditions within nesting burrows (Madsen & Shine Reference MADSEN and SHINE1996b, Reference MADSEN and SHINE1999, Reference MADSEN and SHINE2000; Shine & Madsen Reference SHINE and MADSEN1997). Most female pythons show fidelity to their breeding sites (Madsen & Shine Reference MADSEN and SHINE1998). The only study to date on hatchling water pythons was conducted in the laboratory, and focused on the influence of incubation temperature on the snakes' phenotypes at hatching (Shine et al. Reference SHINE, MADSEN, ELPHICK and HARLOW1997). However, the large size of newborn water pythons (mass 28–33 g and 410–430 mm in SVL, Shine et al. Reference SHINE, MADSEN, ELPHICK and HARLOW1997) makes them a good candidate for studies on spatial ecology by radio-tracking in the field.
Study area
Our study site (Beatrice Hill Farm) comprises 2600 ha of predominately alluvial clay soil (blacksoil) in the floodplain of the Adelaide River, approximately 60 km south-east of Darwin in Australia's Northern Territory. Most areas of the floodplain can be accessed by all-terrain vehicles, even during the height of the wet season. The climate is generally hot year-round but with pronounced seasonal variation in rainfall: more than 78% of the 1402 mm annual precipitation falls during monsoonal downpours from December to March. Vegetation in the area is composed mainly of native and introduced plants: Imperata cylindrica (L.) Beauv., Dichanthium eriantha (R.Br.) A.Camus, Digitaria eriantha Steud., Urochloa mutica (Forsk.) T. Q. Nguyen, Urochloa humidicola (Rendle) Morrone & Zuloaga, Setaria apiculata K. Schum., Hymenachne amplexicaulis (Rudge) Nees (occurring in thick mats floating on waterways), and the sedges Eleocharis spiralis (Rottb.) Roem. & Schult. and E. dulcis (Burm.f.) Trin. ex Hensch.
Radio-tracking
Water python eggs from five different clutches, found under hay bales stored in a building on the study site, were incubated in moist vermiculite at approximately 30 °C. A total of 16 hatchlings were retained and kept at room temperature (about 30 °C) in individual plastic boxes (32 × 20 × 10 cm) lined with paper towel and with a dish containing water ad libitum. After the first slough each young snake was offered a mouse every week. The first ten snakes to attain sufficient size (50 g body mass) were chosen for radio implantation. These 10 individuals were all females and comprised two offspring from each of the five litters. Each temperature-sensitive radio-transmitter (model LTM, Titley Electronics, Australia) weighed 2.5 g and was calibrated in a water bath ranging from 10 °C to 40 °C. We used the program Curve Expert 1.3 to find the best fitted curve describing the relationship between temperature and pulse rate of the transmitter. The transmitters were surgically implanted in the peritoneal cavity of the animals under halothane anaesthesia and using sterile procedures. After a recovery period of 7 d we offered each animal a mouse. After they fed successfully, the snakes were released on the tenth day post-surgery. All snakes were released during the night at the hay shed where they had been collected as eggs. The first snake was released on 16 February and the tenth snake released on 7 June, in 2007. This delay was due to the differences in hatching times (from 20 December 2006 to 14 January 2007), body masses (mass at birth varied from 16 g to 34.5 g) and growth rates of the snakes (0.13–0.38 g d−1, mean number of days to reach 50 g was 104.2 ± 28.1 SD, range = 58–145 d).
Snakes were located daily at different times of the day and night. At each location we recorded the time of the day, the pulse rate of the transmitter and the coordinates using a hand-held GPS receiver. If the snake was visible, we also recorded its activity and behaviour. We scored the vegetation type, height of the vegetation and depth of the leaf litter. If the snake was in a soil crack, we recorded the minimum width of the opening and minimum depth. To compare these microhabitat attributes to those available in the general vicinity, we measured these same variables in a location 3 m from the snake, in a random direction (determined by spinning a pen on a clip board).
We calculated the bearing of each movement for each snake using Hawths' Tools 3.26 for ArcView (ArcMap 9.1) and used Rayleigh's test to assess the directionality of this parameter in Oriana 2.02. Principal components analysis (PCA) was carried out on microhabitat measures. We compared the first two PCs between the used and non-used locations with a Mann–Whitney test (Zar Reference ZAR1999) using the software Statistica 6.0. We compared the proportion of location-points of snakes that were in the floodplain with the proportion of random points in this habitat using a Z ratio test. Using the Hawths' Tools, we obtained the random points by selecting 1000 points inside a circular area of radius equal the maximum distance that the snakes dispersed from the point of release.
We compared movements of hatchlings to those of the six non-pregnant adult females and four adult males studied by Madsen & Shine (Reference MADSEN and SHINE1996b, and unpublished data) from 1991 to 1993. As the adults were studied at a much larger spatial scale (tracking was done from a helicopter much of the time, Madsen & Shine Reference MADSEN and SHINE1996b) it was not possible to precisely measure movements shorter than 10 m. Thus we considered any relocation smaller than 10 m as ‘no movement’ for both adults and hatchlings. Considering only the data from consecutive days, we compared movement parameters between adults and hatchlings using climatic characteristics as covariates. The climatic variable was represented by the first axis (explaining 52.5% of variation) of a PCA analysis that included mean, maximum and minimum values for air temperature, air pressure, relative air humidity and rainfall. Weather data for this PCA analysis were obtained from the Bureau of Meteorology from daily measurements taken every 3 hours (00h00–21h00) in Darwin Airport for the studied years (1991, 1992, 1993 and 2007), from which we calculated weekly means. The movement parameters used as dependent variables were: (1) Mean weekly distances moved including no movements; (2) Mean weekly distances moved excluding no movements; (3) Mean of the total weekly distance moved (include no movement); and (4) Proportion of days moved per week. Because distances moved may depend on the body size of the snake, we also calculated size-corrected movement parameters. Because there was no overlap in body sizes between hatchling and adult snakes, it was not appropriate to include body size as a covariate in the analyses. Instead we divided each movement by the ln-transformed body length of the snake that made it. This calculation transformed movements from distances measured in m to distances measured in body lengths. Snake ID was used as a random factor nested within age-class (adult vs. hatchling) in all analyses to limit pseudo-replication.
Body temperatures
We recorded body temperature (obtained by the transmitter pulse) for the tracked hatchling snakes mostly in the early morning (07h00–10h00) but some snakes had temperature recorded during other times of the day, mainly in the late dry season when air temperature was more variable. We related mean monthly values of body temperature and concurrent air temperature (obtained from the Bureau of Meteorology, for Middle Point Village, in the vicinity of the Beatrice Hill farm) using linear regressions. For the comparisons we divided the data into wet (November–March) and dry (April–October) seasons. Published temperature data for adults (Shine & Madsen Reference SHINE and MADSEN1996) were used for comparative purposes in the discussion. As the frequency distribution of body temperature with respect to time of day was very similar in both groups, the mean overall temperatures of hatchlings and adults were compared by t-test, based on the mean values of each individual snake.
RESULTS
In 2007, the tracking period of newborn snakes varied from 24 to 110 d (mean = 53.9 ± 23.4 SD, N = 10 snakes) between 16 February (late wet season) and 24 September (dry season). One snake was lost during the wet season, 24 d after release, when it entered an inaccessible area of the floodplain. Three snakes died during the tracking: one was predated by a water rat, Hydromys chrysogaster E. Geoffroy, 1804 (all the soft organs were eaten and there were many marks of rodent incisors along the body) after 51 d; another was consumed after 62 d by an adult water python (found with the transmitter in its stomach); and the other died from bacterial infection. The other snakes were tracked until their transmitters failed. One snake was observed to contain recently ingested prey (inferred from bodily distension) on two occasions.
Habitat and microhabitat of newborn snakes
All hatchling pythons moved from the relatively high ground and pasture area of the release point eastwards onto the floodplain. None moved westwards into the savanna habitat and main pasture (Figure 1). Snakes used the floodplain more than expected by chance (Z = −6.61, P < 0.0002, N = 427 locations) and despite the apparently directional movement (Figures 1 and 2), Rayleigh's tests indicated that only one snake moved in a non-random direction (snake 11, Z = 3.14, P = 0.043, N = 89 locations; for all others Z < 2.0, P > 0.1; sample sizes varying from 21 to 72 locations). Some snakes were observed swimming or partially immersed in the water. The snakes were found moving mainly at night, and less often in the early morning. During the day snakes were rarely visible, typically hidden amongst vegetation, down soil cracks or inside farm machinery. Occasionally (during mornings in the mid-dry season), snakes were seen partially exposed, apparently basking. The first two axes in the PCA analysis explained 70.2% of the observed variation for microhabitat use. Crack width and depth were the main components in the first axis, and water and vegetation depth were the main components in the second. Snakes chose larger cracks (wider: U = 2108, P < 0.0001; and deeper: U = 1601, P < 0.0001) and sites where vegetation was higher (U = 14588, P = 0.009) than those available in non-used sites (Table 1). Mean water depth did not differ significantly between used and non-used sites (U = 612, P = 0.689, Table 1).
Movement patterns of radio-tracked hatchling water pythons Liasis fuscus in the Adelaide River floodplain. Each symbol represents one individual and each point is a day-location. Dotted area = floodplain; white area = savanna; hatched area = lagoon; cross-hatched area = pasture; ellipse = release point.
Direction of movement in four selected hatchling water pythons Liasis fuscus in the Adelaide River floodplain. Each line represents one individual and the arrow-heads indicate the direction of the movement. Dotted area = floodplain; white area = savanna; hatched area = lagoon; cross-hatched area = pasture.
Characteristics of shelter-sites used by hatchling water pythons, compared with randomly chosen nearby sites that were not used by the snakes. Sample sizes (N) refer to the number of measurements for each microhabitat trait.
Movements and area of use
The PCA on climatic conditions produced a primary axis which explained 52.5% of the variation. This axis was comprised (eigenvectors > 0.2) mainly by mean and minimum air temperature, mean and minimum relative humidity, mean, maximum and minimum air pressure and mean maximum and minimum rainfall. With the exception of eigenvectors for air pressure, all the others had positive values.
The ANCOVA analysis showed that mean weekly distances moved differed between adults and hatchlings (Table 2): adults travelled longer absolute distances (252 ± 50.3 m wk−1) than hatchlings (66.3 ±41.2 m wk−1; Figure 3a) and these dependent variables were also influenced by climatic conditions (Table 2). Although larger (adult) snakes moved longer distances than smaller (hatchling) snakes, the smaller animals actually moved further in terms of body lengths per unit time. When movement distances were converted to body-length units, neonates moved 1.34 times further (84.8 ± 1.12 body lengths wk−1) than did adults (66.0 ± 1.15 body lengths wk−1, Table 2; Figure 3b). These body-size-corrected movements were also influenced by climatic conditions (Table 2). Inclusion of zero-distance movements in the analyses revealed no significant difference between hatchlings and adults, in terms of absolute means (Table 2; hatchlings: 33.3 ± 24.8 m wk−1, adults: 89.6 ± 25.1 m wk−1) or total absolute distances travelled (i.e. sum of all distances: Table 2; hatchlings: 204 ± 46.6 m wk−1, adults: 195.1 ± 46.1 m wk−1; Figure 3a). In both cases, weather conditions influenced the distances travelled (Table 2) and the number of days the snakes were tracked also influenced the total distances travelled (Table 2). In terms of distance relative to body size, the total distance travelled was marginally significantly higher in hatchlings (Table 2), with most of the difference in distances travelled (hatchlings: 329 ± 32.4 body lengths wk−1, adults: 89.8 ± 32.1 body lengths wk−1; Figure 3b) due to the number of days the snakes were tracked (Table 2) and weather conditions (Table 2).
Results of a statistical comparison (ANCOVA) of movement traits between free-ranging adult and hatchling water pythons (Liasis fuscus). The ANCOVA included weather conditions and number of days tracking as covariates, and snake behaviours (movement parameters) as dependent variables. NA = non-applicable. Sample sizes: Nadults = 93, Nhatchlings = 85 records, when zero-movements are excluded Nadults = 43, Nhatchlings = 75 records. PC1 is the first principal component of the principal component analysis (PCA) for the weather variables.
Characteristics of weekly movements of adults and hatchlings water pythons in the Adelaide River floodplain. Absolute distances travelled (a). Distances travelled in body lengths (i.e. corrected for body size) (b). Error bars represent SE. Data for hatchlings from the present study, and for adults from the previous radio-tracking study by Madsen & Shine (Reference MADSEN and SHINE1996a).
Adult snakes were more sedentary, staying for longer in the same place, than were the neonates (proportion of days moved = 0.48 ± 0.06 days wk−1 for hatchlings, and 0.31 ± 0.06 days wk−1 for adults: Table 2) and the proportion of days on which snakes moved were related to weather conditions (Table 2).
Thermal data
Air temperature varied more over the course of the day during the dry season than during the wet season (Figure 4a, b). Most of the newborn snakes were tracked during the dry season, when they often experienced a considerable (but short-term) decrease in environmental temperature, especially in the early mornings (Figure 4d).
Hourly variation in temperature. Mean air temperature in the wet (November to March) (a) and dry (April to October) seasons (b) in Middle Point Village, on the Adelaide River floodplain, in 2007. Mean body temperatures of hatchling water pythons in the wet season (mean values for all individuals pooled together) (c), in the Adelaide River floodplain. Due to the small sample size data on body temperature during the wet season are not illustrated.
Mean body temperatures were positively correlated with mean air temperatures (Figure 5a, r2 = 0.85, P = 0.003, N = 7, mean values per month) per month, and minimum body temperatures with minimum air temperatures per month (Figure 5b, r2 = 0.73, P = 0.01, N = 7, mean values per month). However, the relationship between maximum body and maximum air temperatures was not statistically significant (Figure 5b, r2 = 0.45, P = 0.097, N = 7, mean values per month). In general, snake body temperatures were slightly higher than the recorded air temperatures, with most records between 27.5 °C and 29 °C for snake body temperatures and between 23 °C and 24.5 °C for air temperature (Figure 6). The overall mean temperature of hatchlings was 29.3 °C ± 0.65 °C (Ntotal = 545), not significantly higher than 27.8 °C ± 2.72 °C (Ntotal = 773), the overall mean temperature of the adults (t = 1.70, df = 18, P = 0. 106).
Relationships between monthly snake body temperatures and air temperatures at the time the snakes were tracked in the Adelaide River floodplain. Mean temperatures (a). Maximum and minimum temperatures (b). Data points for body temperatures represent means of means, maximum and minimum body temperature of each individual in each month. Error bars represent SD. Air temperatures were recorded by the Bureau of Meteorology in Middle Point Village at the time of snake location.
Frequency of observed body temperatures of hatchling water pythons (a) and air temperatures during the radio-tracking period (b).
DISCUSSION
Habitat- and microhabitat-use by newborn snakes
The hatchling water pythons that we studied moved to and inhabited floodplain areas, which are typically used by adults as well (Madsen & Shine Reference MADSEN and SHINE1996a). The young snakes also temporarily used areas highly modified by human activities. Especially during the wet season, the young pythons frequently inhabited aquatic areas for several days at a time. Differences in microhabitat availability according to the seasons may change the habitat use by the snakes. Our sample sizes, especially during the wet season, are too small to properly address this question, but in general snakes sheltered under the vegetation and spent time in water bodies (either permanent or temporary) in both seasons. However, we can expect that snakes will spend more time in the water during the wet season (as most of the habitat is flooded) and will take refuge inside the soil cracks only in the dry.
Little is known of dispersal patterns in neonatal reptiles. All the hatchling water pythons in our study moved onto the floodplain soon after release. This was clearly non-random, although there was no statistical evidence of non-random directionality in movement for most individuals. The lack of significant directionality using the Rayleigh's test is due to the pattern of movements of these snakes: they tended to move around a small area for about 20 d and then travelled for about 4 d to another area and the Rayleigh's test analyses trends in the angles the animals travel. Previous studies have reported that hatchling lizards tend to move in a directional fashion (Imansyah et al. Reference IMANSYAH, JESSOP, CIOFI and AKBAR2008, Pérez-Buitrago & Sabat Reference PÉREZ-BUITRAGO and SABAT2007) and in the field, some hatchlings of massassauga snakes (Sistrurus catenatus: Viperidae) seem to follow the mother's trail (Jellen & Kowalski Reference JELLEN and KOWALSKI2007). Laboratory experiments show conspecific scent-trailing and discrimination in newborn pine snakes, Pituophis melanoleucus (Burger Reference BURGER1989), plains garter snakes Thamnophis radix (Graves & Halpern Reference GRAVES and HALPERN1988) and timber rattlesnakes Crotalus horridus (Brown & MacLean Reference BROWN and MACLEAN1983), but the nature of the cues that hatchling water pythons use to find the floodplain remains unknown.
The main activities (e.g. dispersion and foraging) of water pythons occur at night, both for adults (Madsen & Shine Reference MADSEN and SHINE1996a) and hatchlings (current study); the animals are hidden during the day. In the dry season, soil cracks are the main refuge not only for these young snakes but also for the adults (Madsen & Shine Reference MADSEN and SHINE1996a), as well as for other species of snake and for other animals (such as anurans and rats) (Brown et al. Reference BROWN, SHINE and MADSEN2005, pers. obs.). The snakes apparently select the cracks by size, preferring those wider and deeper than average, probably because it is easier to move and hide inside. As well as providing physical protection, the cracks also provide a thermal refuge and food, as rodents live and reproduce inside such cracks (Madsen & Shine Reference MADSEN and SHINE1996a). When not in a crack, the neonatal snakes were hidden in the base of vegetation, or under the soil surface where the roots or dry cover are tangled, making a protected, soft and camouflaged spot. Thus, the animals tend to select sites where vegetation cover is higher than average. Ontogenetic changes in microhabitat use are common in many animals, including reptiles (Asplund Reference ASPLUND1974, Heatwole Reference HEATWOLE, Gans and Tinkle1977, Imansyah et al. Reference IMANSYAH, JESSOP, CIOFI and AKBAR2008, Irschick et al. Reference IRSCHICK, MACRINI, KORUBA and FORMAN2000, Pérez-Buitrago & Sabat Reference PÉREZ-BUITRAGO and SABAT2007). Although there have been no studies on microhabitat selection by adult water pythons, our observations suggest that adults and hatchlings use the same kind of environment.
Movements and area of use
The area used by newborn water pythons was large and increased with the duration of tracking. This pattern suggests that the snakes were still dispersing over the course of our study; or alternatively, that they are nomadic at this stage of the life history. In the Adelaide River floodplain, adult Liasis fuscus engage in long-distance migration every year (Madsen & Shine Reference MADSEN and SHINE1996a). As observed in other reptiles (e.g. Komodo dragon Varanus komodoensis, Imansyah et al. Reference IMANSYAH, JESSOP, CIOFI and AKBAR2008; massassauga snake, Jellen & Kowalski Reference JELLEN and KOWALSKI2007), newborn water pythons travelled shorter distances than did older conspecifics. However, in contrast to hatchling Komodo dragons (which are about twice the body mass of our snakes, Imansyah et al. Reference IMANSYAH, JESSOP, CIOFI and AKBAR2008) the newborn pythons moved more often than did conspecific adults. In many species small snakes may have more difficulty finding prey of suitable size (Sazima & Martins Reference SAZIMA and MARTINS1990). This may be the case for small water pythons which depend almost exclusively on young dusky rats. Hence, newborn pythons may need to move around to look for food more often than do the adults or a more generalized predator such as a newborn Komodo dragon. Small snakes also may fit into a greater variety of refuges than would a larger snake. In the dry season, soil cracks are available all over the floodplain and a newborn python can easily hide in most of them. Thus, the cost of leaving a refuge may be lower for newborn snakes, and the risk of starving (especially in years of low rodent abundance, like 2007, T. Madsen unpubl. data, L.P. and G.B. pers. obs.) probably outweighs the benefits of staying for longer in the same spot. When travel distances are measured in terms of body lengths rather than absolute distances, newborn water pythons moved further than adults, similarly to hatchlings of Mona Island iguana, Cyclura cornuta stejnegeri (calculated from García et al. Reference GARCÍA, PÉREZ-BUITRAGO, ALVAREZ and TOLSON2007 and Pérez-Buitrago & Sabat Reference PÉREZ-BUITRAGO and SABAT2007). However, older Komodo dragons and massassauga snakes have been reported to travel relatively longer distances than do newborn conspecifics (calculated from Imansyah et al. Reference IMANSYAH, JESSOP, CIOFI and AKBAR2008 and Jellen & Kowalski Reference JELLEN and KOWALSKI2007).
Temperature
Water pythons typically display high body temperatures (Shine & Madsen Reference SHINE and MADSEN1996, and this study) and for most of the year, the main challenge for the animals may be to keep cool rather than to warm up (Shine & Madsen Reference SHINE and MADSEN1996). However, in the dry season, the snakes experienced low temperatures at night and in the early morning, inducing the hatchlings to thermoregulate by sun-basking during the day. Snake body temperatures were usually a few degrees higher than air temperature, probably due to thermal heterogeneity among microhabitats (Shine & Madsen Reference SHINE and MADSEN1996) as well as occasional basking. As observed for adult water pythons in the same area, minimum (but not maximum) body temperature was positively related to minimum air temperature (Shine & Madsen Reference SHINE and MADSEN1996).
Overall, keeping in mind that hatchlings and adult water pythons were tracked in different years, in adjacent areas of the floodplain and the transmitters were relatively heavier to the neonates (and these facts could affect the comparison between the two groups), the most striking result from our study was the broad similarity in spatial ecology and habitat use between hatchling water pythons and conspecific adults, despite the enormous disparity in body sizes between the two groups (mean body masses of 50 g vs 2 kg). That similarity in turn may be due to a broad similarity in the ecological requirements of hatchling and adult water pythons – for example, both depend on the same major prey species (dusky rats) and appear to have similar thermal preferences. The floodplain landscape provides relatively few types of refuge from the (otherwise lethally high) operative temperatures in this area, and thus both juvenile and adult pythons utilize soil cracks for shelter. The similarity in prey resources may also generate similarity in the spatial scale of movements, which presumably relate to the spatial dispersion of feeding opportunities. In summary, then, our study is encouraging: the age class most neglected in previous studies of snake ecology appears to be broadly similar in many respects to the age classes that have been more intensively studied, at least in a species that does not exhibit major ontogenetic changes in diet. By analogy, a recent analysis suggested that annual survival of juvenile snakes likely is more similar to those of adult conspecifics than has generally been assumed (Pike et al. Reference PIKE, PIZZATTO, PIKE and SHINE2008). Many more species need to be studied before we can frame generalities about such ontogenetic conservatism, but accumulating evidence suggests that at least in some species, the ecology of juvenile snakes may not be too dissimilar from that of their parents.
ACKNOWLEDGEMENTS
We are extremely thankful to Eric Cox from the Northern Territory Department of Primary Industry, Fisheries and Mines, for permission to track the snakes in the Beatrice Hill Farm area. E. Cox and Jason Stevens facilitated access to flooded areas during the wet season. Cathy Shilton identified the bacteria that killed one of the snakes, E. Cox and Taegan Calnan identified the grasses in the study area. Mattias Hagman helped in some of the night tracking and Matt Greenlees helped extract LP's bogged vehicle. This study was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES-Brazil, grant to LP) and the Australian Research Council (RS).
