Feeding ecology.– Mammals are the predominant prey for R. scalaris in the south-eastern Iberian Peninsula (nearly 95% of the diet in mass), which set the species among the most stenophagous snakes in western Palaearctic, at least when prey is considered at the level of the traditional vertebrate classes (revision in Böhme, 1993,1999; Schleich et al., 1996). Birds are the secondary prey; all are captured during their reproductive period (spring time), a general pattern in Mediterranean Elaphe s.l. (Filippi et al., 2005). Our results on the diet of the study species coincide in general with previous data (Valverde, 1967; Vericad and Escarré, 1976; Cheylan and Guillaume, 1993).
We also observed a high percentage of stationary prey in our species’ diet, suggesting an active foraging strategy (Schoener, 1971), as usual in Elaphe s.l. species (Schulz, 1996; Mullin and Cooper, 2000). Evidently, the species frequently raids both mammal and bird nests, although the importance of bird eggs as food was probably underestimated because of difficulties in detecting egg remains in studies of snake diets. As deduced from the nesting ecology of its main prey, R. scalaris searches for prey especially in burrows (nests for most of the mammal prey; unpub. data on the authors on radio-tracked individuals).
Although comparative analyses are necessary to define a snake as a specialist predator (Rodríguez-Robles and Greene, 1999), the fact that 49.6% of the prey mass of the diet consists of prey taken in their nest (Table 1) implies that R. scalaris somewhat specialises on nest predation. To our knowledge, of all species of Elaphe s.l. in western Palaearctic, R. scalaris has the highest percentage of nestling prey in its diet (e.g. Luiselli and Rugiero, 1993; Rugiero et al., 1998; Filippi et al., 2005). Nest predation is an efficient foraging strategy for snakes. A single nestling, either bird or mammal, accounts for much less prey mass than an adult prey; however, nestlings are not normally alone and the individuals of a brood or a litter considered together would equal the biomass of an adult of the same species, or even more. Moreover, a nestling struggles less effectively (or not at all) than an adult and handling time may be less for nestling prey than for adult prey (Rodríguez-Robles, 2002). In rat snakes Elaphe helena it has been experimentally proven that time to subdue and ingest a prey increases with the size of the prey (Mehta, 2003). An additional advantage of this is that eggs or nestling are food resources seldom shared with other predators in the study area (Valverde, 1967). As deduced from some predictors of body condition in this species, such as the high level of fat bodies (the highest in western Mediterranean colubrids; Feriche, 1998; Fahd, 2001) and the high reproductive frequency (86% of adult females reproduce every year; Pleguezuelos and Feriche, 2006), R. scalaris is well fitted to this specialised diet.
Most snakes exhibit spectacular ontogenetic change in body dimensions that provides an excellent point of departure for testing the existence of ontogenetic shifts in diet. Rhinechis scalaris is not an exception, since it can increase its body length by up to sixfold and its body weight by up to 120-fold over its ontogenetical stages (unpub. data of the authors). The most interesting finding is that R. scalaris did not follow the general rule in the ontogenetic dietary shift of medium-sized snakes – that is, the shift from ectothermic prey (mainly lizards) to endotherms (mainly mammals; Saint Girons, 1980; Pleguezuelos and Moreno, 1990; Rodríguez-Robles and Greene, 1999), a shift that has particularly been observed within Elaphe s.l. (Fitch, 1999; Filippi et al., 2005). According to our findings, the feeding habits of R. scalaris parallel those of large-sized snakes belonging to other families, such as Corallus hortulanus (Henderson, 1993) and Python regius (Luiselli and Angelici, 1998). Thus, juveniles are capable of consuming endotherms as adults do, but as the predator’s gape size is constrained, their prey are rather small mammals such as Suncus etruscus (one of the smallest mammals in the world) or even smaller (e.g. nestling mammals). Since none of the newborn snakes found in the field had gut contents, we suggested that they enter the first hibernation period relying only on their large vitellogenic reserves, which fuel growth until the spring of the next calendar year (see also Pleguezuelos and Feriche, 2006).
The high percentage of nestlings in the diet is probably also responsible for the small proportion of variance explained by the correlation between R. scalaris body length and prey length. Large snakes, up to 900 mm SVL, continue to feed on very small prey, such as newborn mammals and bird eggs (Figure 1). Hence, the lower limit of prey length does not increase with snake length (Figure 1). Snakes larger than 900 mm SVL appear to ingest rather large prey, introducing considerable variation in the relationship between prey and snake length. Another aspect that arises from the prey size/predator size relationship is that R. scalaris takes relatively small prey. The predator size/prey size relationship in Mediterranean Elaphe rat snakes seems rather constant, both at intraspecific and interspecific levels, and set at about 10% of the snake biomass (Capula and Luiselli, 2002; Filippi et al., 2005). We failed to find major sexual differences in the prey type consumed, in accordance with similar SVL and body mass in both sexes (Pleguezuelos and Feriche, 2006).
Rhinechis scalaris showed a strong preference for head-first ingestion, as is usual in snakes (Rodríguez-Robles, 2002). In rat snakes, prey size has significantly affected prey-handling behaviour (Mehta, 2003). The differences in relative prey size between prey swallowed head-first and those swallowed tail-first suggest that this preference must be guided by reduction of resistance and hence, energy cost and handling time (Brown et al., 2003; Mehta, 2003).
Correlates between morphology, diet and foraging mode.– In snakes, studies on the natural diet also provide information on functional morphology (Schwenk, 2000). For instance, teeth characteristics are probably more related to diet than in some other groups of reptiles or vertebrates, as snakes do not use their trophic structures in conspecific rivalry (Shetty and Shine, 2002). Here, we attempt to relate tooth number with feeding ecology in R. scalaris. The number of maxillary teeth in R. scalaris was among the lowest published for 40 species of Elaphe s.l. (reviewed by Schulz, 1996). Reduction in tooth number is a common trend in fossorial and diet-specialised (i.e. oophagous) snakes (Gans, 1961; Broadley, 1979; Scanlon and Shine, 1988). However, R. scalaris is not fossorial and its diet does not permit it to be classified as oophagous. Simply, this species does not feed upon slippery prey (fish, some reptiles), which require many teeth, or soft contoured prey (adult birds), which implies the need for numerous long teeth (Edmund, 1969). Rather, this species specialises in the consumption of small mammals and probably has the minimum number of teeth needed to seize adult mammals, since highly developed dentition is not necessary for seizing motionless prey (eggs, nestling birds, mammal litter). The reduction in tooth number in R. scalaris agrees with their specialised diet on stationary prey, at least within Elaphe s.l. (Schulz, 1996). More data on the link between diet and tooth number are necessary in other taxonomic groups of snakes, an aspect rather well known in many lizard groups (Edmund, 1969).
Being an efficient active forager, as revealed by a high proportion of specimens with prey (Valverde, 1967) and by the highest level of fat bodies among West Mediterranean colubrids (Feriche, 1998; Fahd, 2001) implies some costs during the foraging activity. Studying the spatial ecology of the species by radio-tracking, Blázquez (1993) found that the mean home range of active individuals was 1.83 ha, larger than in other Iberian colubrids (Malpolon monspessulanus: 0.39 ha, Blázquez, 1993; Natrix maura: 0.18-1.77 ha, Santos and Llorente, 1997). When a snake moves widely, it increases the risk of predation or casualty (Huey and Pianka, 1981; Secor, 1995; Bonnet et al., 1999). Hence, in the study area, although R. scalaris represents only 20.6% of the snakes encountered in the field (n = 1697), this species accounts for 39.5% of the prey consumed by the short-toed eagle Circaetus gallicus, a raptor specialised in catching snakes (Gil and Pleguezuelos, 2001). This is most striking if we take into account that short-toed eagles are exclusively diurnal and R. scalaris somewhat nocturnal (Cheylan, 1986). Indirect evidence of heavy predation pressure is the proportion of individuals with damaged tails (Turner et al., 1982), even though it could indicate inefficient predation (Medel et al., 1988). In the study area, other snakes (Coronella girondica, Hemorrhois hippocrepis, Macroprotodon brevis, Malpolon monspessulanus, Natrix maura, Natrix natrix, Vipera latastei) exhibited tail breakage in proportions ranging from 1.1 to 16.4% (n ranged from 91 to 382, depending on the species; unpub. data of the authors). Thus, the percentage of R. scalaris individuals with tail breakage proved to be the highest of snakes in the study area, suggesting high predation pressure. The elaborate ontogenetic shift in dorsal pattern of this species (Pleguezuelos et al., 1990) may have also arisen for the concealment of a moving widely species. Thus, our data suggest the existence of a significant mortality cost of active foraging by R. scalaris.
We conclude from indirect evidence (high percentage of stationary prey in the diet, many individuals with damaged tails) and direct evidence elsewhere (Blázquez, 1993), that R. scalaris is an active searcher and that this foraging mode may be correlated with some natural-history traits of the species (for example high feeding rate, abdominal fat reserves, frequency of reproduction; this study; Pleguezuelos and Feriche, 2006). A pattern of wide movement in a snake may explain an elevated risk of predation (Gil and Pleguezuelos, 2001). However, unfortunately it may also be correlated with a high vulnerability to anthropogenic mortality (Bonnet et al., 1999), as presently occurs in R. scalaris (traffic casualties).
Received: 8 May 2006.
Accepted: 23 March 2007.