Ontogenetic Niches and the Development of Body Shape in Juvenile Sea Turtles
Abstract
Many marine organisms (invertebrates, fishes) produce large numbers of small offspring (larvae) that differ in appearance and behavior from adults. That outcome occurs because larvae as they grow occupy one or more ontogenetic niches that select for phenotypes that differ from those that promote adult survival. Our purpose in this study was to determine whether similar associations also occur in marine turtles. To find out, we examined relationships between juvenile appearance (body shape) and the ontogenetic niches occupied by 3 species of marine turtles: leatherbacks (Dermochelys coriacea) and 2 cheloniids (green turtles, Chelonia mydas; loggerheads, Caretta caretta). Our measurements indicate that juvenile body shape differs the least from adult body shape in leatherbacks and the most from adults in the 2 cheloniid species. We present evidence from other studies showing that juvenile leatherbacks occupy ontogenetic niches that resemble adult niches, whereas during their ontogeny, both juvenile green turtles and loggerheads make radical niche shifts as they transition between oceanic and neritic habitats. These results are thus consistent with the hypothesis that many morphological and behavioral characteristics expressed by juvenile marine turtles, like those of other larval forms, are best understood as evolved responses shaped by the characteristics of their ontogenetic niches.
Oceans, bays, and estuaries currently serve as habitats for 7 species of marine turtles divided into 6 species in the family Cheloniidae, or the “hard-shelled” turtles, and 1 species in the family Dermochelidae, the leatherback (Dermochelys coriacea; Pritchard 1997). All show unusually high fecundity for a reptile, necessitated by high nest and juvenile mortality (Hendrickson 1980; Bolten 2003; Heithaus 2013). Nests can fail because eggs are left unprotected on beaches where they are vulnerable to predators, inundation from the tides, and wash-outs by storms. Hatchlings emerging from surviving nests undergo a long and dangerous offshore migration and have few defenses from predators other than to avoid detection. These circumstances select for females that invest little in each hatchling but instead invest in large clutches, frequently produced during any one breeding season (van Buskirk and Crowder 1994). In that respect, sea turtle reproduction resembles the pattern shown by many r-selected marine organisms.
All population models for marine turtles predict that juvenile mortality rates decline with growth (Bolten 2003; Heppell et al. 2003). The reasons are difficult to determine because hatchlings and juveniles disperse widely either to the continental shelf (flatbacks, Natator depressus) or deeper offshore oceanic habitats (all other species; Bolten 2003) where their growth patterns, behavioral survival strategies, and interactions with predators are difficult and inconvenient to study. Advances in marine turtle husbandry (Higgins 2003; Jones et al. 2011), however, make it possible to rear marine turtles in captivity, measure how they grow, and observe how they respond behaviorally to prey and predators when transported to and released in natural (open ocean) habitats (Salmon et al. 2004; Smith and Salmon 2009).
Laboratory studies have shown that as they grow, the cheloniid species develop an armored morphology. Most species widen faster than they lengthen (Salmon and Scholl 2014) and some develop prominent spines on their vertebral, costal, and marginal carapace scutes (Salmon et al. 2015, 2016). Carapace widening may reduce the time required for small turtles to reach a refuge width from gape-limited predators (Salmon and Scholl 2014) while spination probably makes small turtles more difficult and dangerous to attack.
This study was inspired by the absence of armoring features in laboratory-reared leatherbacks. Here, we contrast differences in the development of juvenile morphology between leatherbacks and 2 cheloniid species (green turtles, Chelonia mydas; loggerheads, Caretta caretta), both of which were the subject of previous morphological studies (Salmon and Scholl 2014, Salmon et al. 2015). We hypothesize that the observed differences occur because the ontogenetic niche(s) experienced by each species select for different morphologies (Werner and Gilliam 1984; Snover 2008). We provide new data on leatherback growth to confirm that this species shows relatively minor changes in morphology (Jones et al. 2011), whereas shape changes significantly with armoring in the 2 cheloniids. Finally, we summarize evidence from other studies documenting that juvenile and adult leatherbacks occupy similar oceanic ontogenetic niches. In contrast, during development green turtles and loggerheads shift between oceanic and neritic niches (Bolten 2003; Snover 2008) where survival probabilities may be improved by changes in body shape.
METHODS
Leatherback Collection and Husbandry.
Leatherback hatchlings were collected between June and August, 2011–2013, as they emerged from nests on the beach at Boca Raton, Deerfield Beach, and Juno Beach, Florida (lat 26°–27°N, long 80°W). Turtles were held in quarantine for 5–7 d until deemed healthy, and then transferred to shallow tanks (100 × 62 × 30 cm deep). The seawater supply for these tanks was first circulated through a ultraviolet sterilizer and protein skimmer, and then cooled using a chiller to 23°–25°C as we found that leatherbacks thrive best at slightly lower, open ocean summer temperatures.
Because leatherbacks do not recognize barriers such as tank walls (Jones et al. 2011), each turtle was confined by a short length of monofilament line (a “tether”) to the central area of the tank. Tethers allowed each turtle to swim and dive in any direction, but were short enough to prevent any contact with the side or bottom of the tank. Turtles were fed to satiation 3 times daily on a low-calorie diet of minced squid or fish embedded in gelatin, carbohydrate in the form of French bread, and reptile vitamins and minerals.
Cheloniid Husbandry.
We compared our leatherback data with the data gleaned from Florida East Coast green turtles and loggerheads, reared using similar procedures. Details are provided elsewhere (Salmon and Scholl 2014).
Turtle Measurements.
We measured morphology in 4 ways, quantifying straight carapace length (SCL) and width (SCW), body depth (BD), and body volume (BV).
Body volume was measured during the 2015 nesting season using a sample of 26 leatherbacks from 6 nests, 37 green turtles from 6 nests, and 46 loggerheads from 7 nests. Volume measurements were initiated with the hatchlings, and repeated at 2–3-wk intervals for up to 14 wks as the juveniles grew. To obtain the data, we filled a 1000-ml Pyrex beaker with filtered seawater until it overflowed from the spout. After the flow stopped, a turtle, held by the tips of its fore flippers while they were extended over its head, was briefly (2–3 sec) lowered vertically into the beaker until its head was completely submerged. Each turtle had to remain quiescent during its test; if not, it was tested again. Each turtle's volume (in cm3) was then estimated by the amount of seawater from a graduated cylinder required to again fill the beaker to overflow capacity.
Body shape data were collected from another sample of leatherbacks obtained during the nesting season between 2008 and 2010 and during 2013. Our sample consisted of 31 turtles from 11 nests. All turtles were reared for up to 14 wks. As hatchlings and at weekly intervals thereafter, each turtle's growth (SCL, SCW, and BD) was measured using vernier calipers (accurate to the nearest 0.1 mm). Body depth was measured at its maximum from a lateral view, which in leatherbacks is made across the body about one-third of the shell length behind the anterior margin of the carapace (Fig. 1).
Citation: Chelonian Conservation and Biology 16, 2; 10.2744/CCB-1274.1
Because juvenile leatherbacks are difficult to rear in captivity, only 8 of the 31 turtles were observed for the full 14 wks. Data from the 23 remaining turtles were used only while they were growing and feeding normally. Turtles that stopped eating were treated with antibiotics and then immediately taken offshore and released.
Comparable measurements of the cheloniid turtles were made from samples of 60 turtles of each species (5 hatchlings from each of 12 green turtle and 12 loggerhead nests) also reared in captivity for 14 wks. Details are provided elsewhere (Salmon and Scholl 2014).
Leatherback Growth Patterns.
To determine if leatherback hatchling proportions were maintained by juveniles as the turtles grew, we compared the shapes of juvenile turtles (based upon their SCW:SCL proportions) with those of hatchlings (Salmon and Scholl 2014). We calculated a baseline measurement, the hatchling ratio, by measuring each hatchling's SCW and SCL, then dividing the average SCW by the average SCL for the entire hatchling sample. Because hatchlings are longer than they are wide, that ratio was < 1.0. We then multiplied that ratio by the SCL of each growing juvenile turtle to determine an expected SCW for that SCL, based upon the assumption that growth was isometric (hatchling proportions were maintained). The expected SCW was then compared statistically with the observed SCW of the growing juvenile actually measured, using paired t-tests (Zar 1999). If the 2 measurements differed statistically, then we concluded that growth was allometric (Huxley 1932).
The same procedures were used to determine how growth occurred in loggerheads and green turtles (Salmon and Scholl 2014).
Comparisons Among the Species.
To determine whether leatherbacks changed shape like loggerheads and green turtles, we compared the growth distributions generated by 1) BV (cm3), 2) SCW (mm), and 3) BD (mm) against SCL (mm) as the turtles increased in size.
To compare changes in BV with growth among the species, we normalized for the differences in size between the turtles. Leatherback hatchlings are larger than green turtle hatchlings, and green turtle hatchlings are larger than loggerhead hatchlings (Wyneken 2001). We standardized our measurements by determining the average BV (in cm3) at growth increments of 10 mm in SCL (starting at 10.0–19.9 mm, then 20.00–20.9 mm, etc.). Because the relationship between BV and SCL was exponential, we log transformed the BV data to yield an approximate straight-line relationship. These data met assumptions of normality (Shapiro-Wilk test, p > 0.05) and equality of variance (Levene's test, p > 0.05). We used a 1-way analysis of covariance (ANCOVA), with SCL as the covariate, to determine if the pattern of volume change with growth differed statistically among the species.
We compared the distributions generated by plots of the observed SCW and BD against SCL for the 3 species, using a single-factor analysis of variance (ANOVA; Zar 1999) to determine if they showed similar changes over 14 wks of growth. All data met the assumptions of normality (Shapiro-Wilk test, p > 0.05). When probabilities indicated that differences were significant, we used either a Tukey test (for samples with equal variance) or (when samples variances were unequal) a Tamhane's T2 post hoc comparison (Tamhane 1979) to identify the species responsible for the differences.
In loggerheads, the analysis of BD:SCL ratios was complicated by the positive allometric growth of spines during early ontogeny on the carapace and plastron (Salmon et al. 2015). Spination was especially evident on the vertebral scutes and the plastron ridges (Fig. 1). Because the spines do not contribute substantially to internal body volume, we elected to exclude them from our measurements of BD:SCL relationships. To do so we measured carapace and plastron spine lengths in 122 loggerhead turtles within an SCL range between 40.8 and 130.5 mm. A linear regression of SCL against spine length showed that the 2 variables were strongly correlated (R2 = 0.75). We used the resulting linear regression equation (y = 0.0277x − 0.0174) to estimate spine length for a turtle of a given SCL, where y was spine length and x was SCL. The appropriate spine length for a turtle of a given SCL was then subtracted from previous measures of BD that included the spines. The result was a “spineless” loggerhead whose BD:SCL ratio could be more appropriately compared with the ratios shown by green turtles or leatherbacks lacking spines naturally.
RESULTS
Body Shape in Leatherbacks.
The observed SCW measurements were significantly greater than the expected SCW measurements (Fig. 2; t30 = 5.077, p < 0.001) indicating that with growth, carapace width increased allometrically.
Citation: Chelonian Conservation and Biology 16, 2; 10.2744/CCB-1274.1
Comparisons Among Species.
Hatchling ratios were on average smaller for leatherbacks (n = 31; 0.71 ± 0.05 SD) than for loggerheads (0.78 ± 0.03 SD) or green turtles (0.77 ± 0.03 SD; Salmon and Scholl 2014), indicating that leatherback hatchlings had proportionally narrower bodies than either of the cheloniid species. Those differences were statistically significant (ANOVA, F2,148 = 47.60, p < 0.001). The hatchling ratios for loggerheads and green turtles did not differ statistically from one another but both differed significantly from leatherbacks (Tukey test, p < 0.001).
The 3 species showed no significant differences in BV with increasing SCL (Fig. 3; ANCOVA, F2,21 = 2.42, p = 0.113).
Citation: Chelonian Conservation and Biology 16, 2; 10.2744/CCB-1274.1
With growth, the 3 species differed in their SCW:SCL relationships. At the same SCL, growing leatherbacks were significantly narrower in SCW than the 2 cheloniid species (Fig. 4). These differences among the species were statistically significant (ANOVA, F2,147 = 23.06, p < 0.001). There were no statistical differences between the 2 cheloniid species but both differed significantly from the leatherback (Tahmane's T2 test, p < 0.001).
Citation: Chelonian Conservation and Biology 16, 2; 10.2744/CCB-1274.1
Leatherback body depth exceeded both green turtle and loggerhead body depth over a range of comparable sizes (SCL in mm; Fig. 5). Those differences were statistically significant (ANOVA, F3,207 = 69.44, p < 0.001). The loggerhead and green turtle distributions did not differ from one another but both differed significantly from the leatherback distribution (Tukey tests, p < 0.001).
Citation: Chelonian Conservation and Biology 16, 2; 10.2744/CCB-1274.1
DISCUSSION
Our data indicate that during the first 14 wks of development leatherbacks, loggerheads, and green turtles show statistically similar increases in volume as they increase in length (Fig. 3). Other measurements, such as those relating SCL to mass (Jones et al. 2011), also show similarities among the species. Those results have been interpreted as illustrating constraints imposed by similar body plans. However, we suggest such similarities mask the unique changes in body shape that occur among the species as a consequence of allometric growth. These patterns of change are complex, consisting of some features that are shared but differ in degree, and others that are unique to each species.
Features shared with growth by all 3 species include a disproportionate increase in width relative to length, which occurs in leatherbacks (Fig. 2) as well as in the cheloniids (Salmon and Scholl 2014; Salmon et al. 2015). However, bodies are proportionally wider in green turtles and loggerheads than in the leatherback (Fig. 4). Unique features among the 3 species include the development of spines on the carapace and plastron of loggerheads (Fig. 1; Salmon et al. 2015), the persistence of a smooth surface on the carapace of green turtles and leatherbacks (Fig. 1; Salmon et al. 2016), and an increase in leatherback body depth that exceeds those proportions in both green turtles and loggerheads (absent their spines; Fig. 5). The result in leatherbacks is a morphology in which volume is increased disproportionally to surface area as the animal becomes more “tubular” in shape, whereas in the cheloniids, surface area is increased disproportionally to volume as the animals become both wider and flatter.
With few exceptions (see below), these contrasts in growth and morphology remain as interesting features whose significance has not been studied. That situation arises in part because of the inaccessibility of juvenile turtles in the open ocean (some exceptions are reviewed below) and, as a consequence, a lack of knowledge about their biology. As a result we know almost nothing about how morphology is coupled with behavior in ways useful in such an ecological setting. In what follows, we present hypotheses that might explain how those relationships evolved, why they differ among species, and how they might they promote survival during life-history periods when young turtles are particularly vulnerable to predators.
Habitat Shifts and the Evolution of Complex Morphologies.
Sea turtles, like many other marine organisms that produce large numbers of small offspring, increase in size by several orders of magnitude as they grow to maturity. In many fishes and invertebrates, those conditions usually result in the evolution of complex life cycles, defined as those with larval stages uniquely adapted in morphology, physiology, and behavior to ecological niches that differ from those occupied by the adults (Ebenman 1992; Moran 1992, 1994). Typically, the larval stages are planktonic and disperse but eventually settle in benthic habitats where they complete growth to maturity. It is also common for larval stages that occupy similar planktonic niches to resemble one another but for adults, which generally occupy distinctly different niches, to be easily distinguished. These relationships provide evidence that among species with complex life cycles, selection favors divergence in body design (often achieved through metamorphosis) that results in life-history stages uniquely adapted for their functions to the habitats where they are found (Ebenman 1992; Moran 1992, 1994; Marshall and Morgan 2011). Conversely, when juvenile and adult niches are similar, juvenile body design more closely resembles adult body design (Ebenman 1992).
A second important principle centers on how selection governs choice of optimal habitat and the behavior expressed in that niche by the larval stages. As a general principle, the limited mobility and small size of larval stages result in high rates of mortality, most often as a consequence of interactions with predators, although starvation and exposure to unfavorable physical change can also take a toll. To counter predation pressures and increase their chances of survival, larvae must find the right balance between 2 competing factors: rapid growth and predator avoidance. Both have associated costs that can reduce survival and compromise fitness, i.e., rapid growth requires a search for food and feeding activity, but foraging activity can increase vulnerability to predators (Werner and Gilliam 1984; Snover 2008; Heithaus 2013). In addition, growth rates are maximized by selecting habitats where food is abundant, but those sites are also likely to be frequented by predators. Larvae must therefore optimize the trade-off between rapid growth and predator avoidance. When youngest and smallest, that balance may favor safety over growth but after an increase in size (accompanied by a decrease in vulnerability to predators) that balance can shift to foraging more aggressively, thereby expediting growth (Werner and Gilliam 1984; Snover 2008).
It may be unrealistic to think of marine turtle hatchlings and juveniles as larvae but nevertheless we suggest that the principles described above are usefully applied to better understand why and whether juvenile growth patterns, morphology, and behavior diverge among marine turtle species (Snover 2008). A conspicuous example arises when one compares juvenile with adult morphology between leatherbacks and the 2 cheloniid species. While allometric growth results in juvenile green turtles and loggerheads that differ in appearance from conspecific hatchlings and adults, juvenile leatherbacks remain conservative in color and body shape and show a strong physical resemblance to both hatchlings and adults (SCL:SCW ratio; Stewart et al. 2007; Jones et al. 2011). Why should that be?
Why Do Juvenile Leatherbacks Resemble Adult Leatherbacks?
We suggest that the resemblance between juvenile and adult leatherbacks is a consequence of the similarities between their niches because even though these life-history stages differ by several orders of magnitude in size (adult mass can be 6000 times more than hatchling mass: Zug and Parham 1996), leatherbacks at all stages of growth specialize on low-calorie gelatinous prey. In addition, the behavioral interactions between predators and prey are much the same regardless of their size relationships. We also suggest that the distribution of gelatinous prey in the open ocean, as well as its ephemeral nature in space and over time, requires the turtles to continuously search for regions where this food supply is available. The energetic demands of that search require leatherbacks to be highly streamlined and efficient swimmers that exclusively use lift-based mechanisms for locomotion at any size (Davenport 1987; Wyneken 1997). They possess those physical characteristics as hatchlings and retain them as larger juveniles and adults.
Long-distance open ocean migrations and searches for food are characteristic of leatherbacks. The locations where gelatinous prey are abundant are not well predicted by the usual measures of surface chlorophyll, nor are they associated with weed lines in current convergence zones frequented by small cheloniids (Witherington 2002; Witherington et al. 2012) or associated with gyres (Saba 2013). Oceanographic data suggest that the most productive locations for gelatinous prey are coastal and associated either with upwelling that occurs adjacent to the continental shelf, or with seasonal stratification changes (Saba 2013). In temperate regions in the western Atlantic, the largest concentration of gelatinous zooplankton occurs during the summer months after phytoplankton blooms associated with turnovers. Peak leatherback nesting in the Caribbean and western Atlantic mostly occurs between April and June (Eckert et al. 2012), enabling most hatchlings, carried initially by oceanic currents from nesting beaches to the south, to reach northern coastal locations when jellyfish are abundant and surface water temperatures are most favorable (≥ 26°C) for small turtles (≤ 100 mm SCL). Those are the locations where juvenile leatherbacks are most often seen (Eckert 2002).
For juveniles, growth rates are rapid even though food is of low nutritional value. Jones et al. (2011) suggest that leatherback assimilation efficiencies are high, supporting those growth rates. Estimates are that young turtles must consume, on average, more than 100% of their body weight per day to sustain such high growth rates (Lutcavage and Lutz 1986). The rapid increase in BD (Fig. 5) we document here for juveniles is likely the first manifestation of the remarkable ability of leatherbacks to change body shape as they store fat (Davenport et al. 2011). Such a capacity has obvious benefits when foraging activity must be nearly constant to locate rich but patchy concentrations of low-energy prey.
These demands suggest that juvenile leatherbacks (in contrast to green turtles and loggerheads of the same age) should select oceanic habitats that favor growth over those favoring concealment, that is, locations where their foraging efforts are most likely to be successful. Risk to survival may be minimized by other life-history features, such the production of larger hatchlings, fast rates of growth as juveniles (Jones et al. 2011), and the exceptionally high reproductive effort required of leatherback females to produce large numbers of offspring during a given breeding season (van Buskirk and Crowder 1994). It is also possible that predators are less frequently encountered in the open ocean where young leatherbacks search for prey, though there are no data available to support this hypothesis.
None of these considerations are meant to minimize the differences in foraging capabilities between juvenile and much larger leatherbacks. Those differences include an increasing ability to generate and retain metabolic heat (Bostrom et al. 2010), coupled with a behavioral and physiological capacity to make longer and deeper dives (Salmon et al. 2004). The eventual result is niche expansion, characterized by the ability of larger turtles to seek assemblages of prey at greater depths and eventually, to forage at higher latitudes in cooler, more productive waters (Eckert 2002; Eckert et al. 2012). But while these enhanced capabilities affect where searches take place and may improve foraging efficiencies, they do not fundamentally change the nature of the interactions between these predators and their prey once food is spatially located (Salmon et al. 2004; Heaslip et al. 2012; Wallace et al. 2015). In addition, hungry posthatchling leatherbacks will without hesitation attack objects differing in size and shape, suggesting that they have no preferences for particular configurations of gelatinous prey that under natural conditions, are likely to vary widely in appearance. Indeed, recognition of gelatinous prey is apparently innate (Constantino and Salmon 2003). Our field study (Salmon et al. 2004) confirmed that feeding occurs in open water and is devoid of any attempt to minimize risk by fleeing when exposed to strange, large objects (like a photographer, Fig. 6).
Citation: Chelonian Conservation and Biology 16, 2; 10.2744/CCB-1274.1
Why Do Juvenile Green Turtles and Loggerheads Differ in Appearance from Adults?
In contrast to leatherbacks, juvenile green turtles and loggerheads differ strikingly in shape and/or spination from adults. We attribute those changes to the development of armoring features through allometric growth (Salmon and Scholl 2014). Carapace widening occurs in both species but is especially extreme in green turtles that as recruits to shallow foraging areas are often characterized as being shaped like “dinner plates.” As adults, both green turtles and loggerheads narrow in their width dimension, though, to the best of our knowledge, evidence in support of those observations has only been published for loggerheads (Kamezaki and Matsui, 1997). Adults of both species also have smooth shells (Wyneken 2001). So do juvenile green turtles. Adult loggerheads show no evidence that earlier in their ontogeny, their carapace was armored with prominent spines (Salmon et al. 2015).
In the western Atlantic, hatchling loggerheads and green turtles are feeding generalists (Bjorndal 1997; Jones and Seminoff 2013) that reside in weed lines, usually composed of Sargassum (Carr 1986; Witherington et al. 2012). Weed lines provide some measure of cover from aquatic as well as aerial predators, as well as opportunities for growth because suitable prey (in the form of hydroids, copepods, winged insects, polychaetes, gastropods, and bryozoans; Witherington 2002; Witherington et al. 2012) accumulate in these communities. But weed lines also attract known turtle predators (Luckhurst 2015) and so danger is likely to persist. That may be why these species during their early ontogeny show rapid changes in body shape and the development of structures that may have an antipredator function.
Survival for the hard-shelled turtles in weed lines requires a different balance between strategies that promote foraging and predator avoidance than in leatherbacks. The balance is most likely dominated in the smallest juveniles by inactivity and hiding, but later, with growth, shifts toward more risky foraging activity. Ontogenetic shifts of this nature are well documented in other aquatic organisms (Werner and Gilliam 1984; Snover 2008) but, unfortunately, have not been systematically or quantitatively studied in the cheloniid turtles. Field observations by Witherington and colleagues (Witherington 2002; Witherington et al. 2012) provide qualitative support for these ideas as they report that small turtles in weed lines minimize activity. Loggerheads resemble Sargassum in color and often float in a “tuck” position above or close to Sargassum mats (Smith and Salmon 2009). Small green turtles, however, possess a carapace color that differs from Sargassum and instead burrow into the mat before becoming inactive. Only their head is exposed (anachoresis; Smith and Salmon 2009). The 2 species also differ in how they respond to approaching threats. Posthatchling loggerheads remain inactive, whereas green turtles often flee by diving under the mat (M.S., pers. obs.; Carr 1967).
Juvenile loggerheads when larger develop spines on the vertebral and costal scutes that reach peak prominence when the turtles are ∼ 25 cm in SCL (Salmon et al. 2015). Thereafter, they gradually regress but spines on the posterior marginal scutes remain in turtles reared in the laboratory up to 40 cm in SCL (Salmon et al. 2015). With continued growth, the turtles abandon weed lines and are found swimming freely in the waters around the Azores and Madiera (Bolten 2003). They instead seek out shallow sites adjacent to seamounts and upwelling zones where productivity is high (McCarthy et al. 2010). Turtles feeding near the Azores had stomach contents indicating a dietary shift to larger prey, some that are and others that are not associated with weed lines but likely to be encountered in open, as well as deeper, water. These include squid, free-drifting cnidarians (Physalia, Velella), molluscs (Glaucus, Janthina), and crustaceans (Lepus, Halobates; Frick et al. 2009).
Thus, during their early development in the North Atlantic Ocean, loggerheads make at least 2 habitat shifts: from open water during hatchling offshore migration to weed lines, and from weed lines back to open water habitats as juveniles after they are transported eastward to seamount areas in the eastern Atlantic (Bolten 2003). Larger and older (≤ 2 yrs) green turtles in weed lines more actively forage in greater proportions both within (“manipulating” Sargassum with their front flippers) and beneath (in the water column; Witherington et al. 2012) the mat.
We interpret these observations as consistent with the hypothesis that with an increase in size both species devote a larger proportion of their time searching for food in the open, despite the risks of exposure to predators that accompany those efforts.
Conclusions
We suggest that these differences in shape and behavior between leatherbacks and cheloniid turtles occur because during growth, the hard-shelled turtles undergo significant shifts in their ontogenetic niches, and those habitats select for size-dependent changes in both morphology and behavior. In leatherbacks, however, the ontogenetic niche remains much the same with growth; it primarily expands rather than shifts. As we show here, posthatchling leatherbacks develop a more compact, cylindrical body that should also improve their physiological capacity to retain heat (Bostrom et al. 2010). Eventually, such a capacity enables them to feed in colder, more productive waters (Paladino et al. 1990; James et al. 2006; Wallace and Jones 2008, 2015) but during all stages of development, leatherbacks consume the same kinds of prey that are captured in much the same way in an open water habitat. Jones et al. (2011), in their study of leatherback growth, reached similar conclusions.
Thus, a comparison between marine turtles in how growth patterns differ during early ontogeny reveals significant contrasts, expressed by how species vary in morphology, physiology, and behavior. Those differences inform us about the ecological pressures shaping their adaptations during a period of development (the “oceanic” phase) when they are especially inaccessible, and hence difficult to study.
(Top) A juvenile leatherback at ∼ 14 wks. Note its deeper body compared with the more dorsoventrally flattened loggerhead (middle) and green turtle (bottom) of the same age. Loggerheads also develop carapace spines that are absent from the leatherback and the green turtle carapace. Photos: J. Wyneken and M. Salmon. (Color version is available online.)
Observed vs. expected measurements of straight carapace width (SCW) against straight carapace length (SCL) for growing leatherbacks (n = 31 turtles), showing that SCW increases allometrically.
Volume (log transformed) vs. straight carapace length (SCL), based upon 113 measurements for leatherbacks, 126 for green turtles, and 128 for loggerheads. The species show no significant differences in volume (by an analysis of covariance, with SCL as the covariate).
Observed straight carapace width (SCW) vs. straight carapace length (SCL) plots contrasting the 3 species. At a comparable SCL, leatherbacks are narrower than the 2 cheloniid species.
Body depth (mm) plotted against straight carapace length (SCL, mm) for leatherbacks, loggerheads (with spines excluded), and green turtles. At a comparable SCL, leatherback bodies are deeper than the 2 cheloniid species.
Response of a laboratory-reared juvenile leatherback during its first exposure to the open ocean and to natural prey (a jellyfish). Panels (left to right) show the turtle as it dove toward and then fed upon a moon jelly (Aurelia). It is fully exposed and seemingly oblivious to the presence of other large moving objects (such as the photographer). See Salmon et al. (2004) for details. (Color version is available online.)
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
