Environmentally Cued Hatching in Reptiles

Integrative and Comparative Biology, volume 51, number 1, pp. 49 61 doi:10.1093/icb/icr043 SYMPOSIUM Environmentally Cued Hatching in Reptiles J. S. Doody 1 School of Biological Sciences, Monash University, Clayton, Victoria 3800, Australia From the symposium Environmentally Cued Hatching across Taxa: Embryos Choose a Birthday presented at the annual meeting of the Society for Integrative and Comparative Biology, January 3 7, 2011, at Salt Lake City, Utah. 1 E-mail: sean.doody@monash.edu Synopsis Evidence is accumulating for the widespread occurrence of environmentally cued hatching (ECH) in animals, but its diversity and distribution across taxa are unknown. Herein I review three types of ECH in reptiles: early hatching, delayed hatching, and synchronous hatching. ECH is currently known from 43 species, including turtles, crocodilians, lizards, snakes, tuatara, and possibly worm lizards. Early hatching caused by physical disturbance (e.g., vibrations) is the most commonly reported ECH across all groups; although it apparently serves an antipredator function in some species, its adaptive value is unknown in most. Delayed hatching, characterized by metabolic depression or embryonic aestivation, and sometimes followed by a hypoxic cue (flooding), occurs in some turtles and possibly in monitor lizards and crocodilians; in some of these species delayed hatching serves to defer hatching from the dry season until the more favorable conditions of the wet season. Synchronous hatching, whereby sibling eggs hatch synchronously despite vertical thermal gradients in the nest, occurs in some turtles and crocodilians. Although vibrations and vocalizations in hatching-competent embryos can stimulate synchronous hatching, cues promoting developmentally less advanced embryos to catch up with more advanced embryos have not been confirmed. Synchronous hatching may serve to dilute predation risk by promoting synchronous emergence or reduce the period in which smells associated with hatching can attract predators to unhatched eggs. Within species, advancing our understanding of ECH requires three types of studies: (1) experiments identifying hatching cues and the plastic hatching period, (2) experiments disentangling hypotheses about multiple hatching cues, and (3) investigations into the environmental context in which ECH might evolve in different species (major predators or abiotic influences on the egg, embryo, and hatchling). Among species and groups, surveys for ECH are required to understand its evolutionary history in reptiles. The probability of ECH occurring is likely influenced by a species s life history, ecology, behavior, and interrelationships with other species (e.g., sizes of predator and prey). More broadly, the discovery of embryo embryo communication as a mechanism for synchronous hatching in crocodilians and turtles indicates that the social behavior of (nonavian) reptiles has been underestimated. Introduction In the Origin of Species, Charles Darwin foreshadowed the importance of embryonic variation by recognizing that organisms are liable to variations at all ages (Darwin 1859). Indeed, he asserted that the instincts and structure of the young could be slowly modified as surely as those of the adult; and both cases must stand or fall together with the whole theory of natural selection. Thus, even as the theory of natural selection was being offered to explain the diversity of life, its importance in molding embryonic traits in different environments was being considered. Darwin s consideration of the evolution of embryos mainly concerned phenotypic change due to natural selection. We now know that different environments also more directly induce changes in an individual s behavior, morphology, and physiology (West-Eberhard 2003). Such changes, collectively termed phenotypic plasticity, can buffer against evolutionary change but can also interact with environmental conditions to produce genetic change Advanced Access publication June 9, 2011 ß The Author 2011. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved. For permissions please email: journals.permissions@oup.com.

50 J. S. Doody (Huey 2003; Price et al. 2003). But is this phenotypic plasticity general to embryos? A recent advance that widens our consideration of phenotypic plasticity to embryos is environmentally cued hatching (ECH; Warkentin and Caldwell 2009). The novel contribution here is that animals can increase their fitness by altering timing of hatching to environmental conditions using phenotypic plasticity in behavior, morphology, and physiology (Warkentin and Caldwell 2009). For example, animal embryos can hatch early in response to predatory attacks, they can delay hatching until environmental conditions are more suitable, or they can synchronize hatching with their clutchmates. We are only beginning to understand the diversity, costs benefits, and adaptive role of ECH in animals: reported incidences in animals are widespread but isolated, and details of mechanisms and costs benefits are not well known (Warkentin and Caldwell 2009; Warkentin 2011). A major hindrance to our discovery of how and why ECH has evolved is thus taxonomic. How do ECH, its cues and environmental and evolutionary contexts differ among taxonomic groups? Reptiles do not currently feature prominently in our understanding of ECH. However, the nests, eggs, and embryos of reptiles are poorly known relative to other vertebrates due to the secretive egg-laying habits of reptiles (Doody et al. 2009). The current evaluation of reptiles thus likely underestimates their contribution to ECH in animals. Nevertheless, reptiles make good subjects for studying ECH for two reasons. First, reptiles typically suffer high embryonic mortality due to predation: for example, predation upon nests of freshwater turtles typically spans 50 100% (e.g., Congdon et al. 1987; Schwanz et al. 2010). There is thus no shortage of natural selection ready to seize at any novel variant that would increase reproductive success. Second, reptiles are generally socially simple. They rarely exhibit parental care after oviposition (Shine 1988; Soma 2003) and show little advanced social behavior relative to other vertebrates (J. S. Doody and D. Chapple, submitted for publication). Negligible sociality reduces the potential number of ad hoc hypotheses for egg-related phenomena while offering potential insights into the evolution of social behavior. The latter is relevant because embryo embryo communication a form of social behavior and a mechanism for synchronous or early hatching has recently been demonstrated in reptiles. Herein I review evidence for ECH in reptiles. The assessment is perhaps premature; for a few species we have detailed studies, but for most species we have only anecdotal evidence for ECH. Regardless, the evidence is persuasive for the widespread occurrence of ECH in reptiles and thus the potential for a key function of ECH in the fitness of reptilian populations and a role for reptiles in understanding the evolution of ECH in animals. I review three major types of ECH: early hatching, delayed hatching, and synchronous hatching. I discuss adaptive hypotheses when evidence permits. I also discuss ECH cues including embryo embryo communication, an emerging phenomenon in reptiles otherwise known only from birds. Finally, I outline directions for future research on ECH in reptiles; specifically, where we might look for different types of ECH in different groups of reptiles, and why. What constitutes ECH and this review? ECH is variation in the time, age, or developmental stage at hatching, facilitated by an extrinsic cue, which allow animals to improve survival by balancing the costs and benefits of hatching as risks vary (Gomez-Mestre et al. 2008; Warkentin and Caldwell 2009). Broadly speaking, there are a number of intrinsic phenomena that influence timing of hatching, including the timing of laying, acquisition of energy, embryonic diapause, developmental rate, and temperature (Deeming and Ferguson 1991). In the present review, I focus primarily on extrinsic cues that influence an embryo from the onset of hatching competence onwards. Some definitions are necessary for exploring the diversity of ECH. Spontaneous hatching refers to hatching that occurs without an external cue (Gomez-Mestre et al. 2008). Ewert (1991) terms this normal hatching time in a review of temporary suspension of development in reptiles and birds. Timing of spontaneous hatching can be normally distributed (Gomez-Mestre et al. 2008). In some species of animals embryos can hatch early, and the earliest possible time of hatching is referred to as the onset of hatching competence (Gomez- Mestre et al. 2008). In species with early hatching the plastic hatching period thus spans the onset of hatching competence and the period during which spontaneous hatching occurs (Gomez-Mestre et al. 2008). In contrast, hatching in some species does not occur at the onset of hatching competence or during a period of spontaneous hatching, but rather embryos delay hatching, awaiting some environmental cue to signal a more favorable hatching time (Warkentin and Caldwell 2009). Thus, in delayed hatching, the plastic period of hatching spans the onset of hatching competence and some point

Induced hatching in reptiles 51 days to months later. There are two caveats here: first, in these species an artificial cue is generally required to initiate hatching near the onset of hatching competence because such cues do not occur in nature at that time. Second, embryos delaying hatching for too long lose their ability to hatch due to exhausted energy reserves (Ewert 1991; Martin 1999). Early hatching generally involves an external cue (e.g. vibrations, temperature change) that will trigger hatching earlier than the spontaneous hatching period (Gomez-Mestre et al. 2008). For example, physical disturbance associated with handling eggs may simulate: (1) vibrations made by predators attacking eggs, (2) disturbance from sibling hatchlings, or (3) prehatching movements of sibling embryos (Shine 2002; Warkentin 1995; J. S. Doody et al., submitted for publication). Delayed hatching herein refers to embryos that defer spontaneous hatching, waiting to hatch later during a more favorable period (e.g., triggered by flooding). In a review on reptiles and birds, Ewert (1991) distinguished between delayed hatching and embryonic aestivation: delayed hatching tends to last only a very brief period, a few days in birds to three weeks in reptiles, while aestivation occurs when the embryo remains in the egg for very long periods. In both cases metabolism either plateaus or decreases, although in aestivation metabolism may be lower (Ewert 1991). Because delayed hatching and embryonic aestivation essentially differ only in the magnitude of the delay and the extent of metabolic depression, herein I consider embryonic aestivation to be a subset of delayed hatching, rather than a parallel category. Thompson (1989) and Ewert (1991) considered asynchronous hatching within a clutch as delayed hatching. In this review I consider only cases in which the entire clutch delays hatching, thus distinguishing it from synchronous hatching. Synchronous hatching must be more carefully defined. Although asynchrony of oviposition can produce developmental asynchrony in birds (Brua 2002), in reptiles the eggs are laid in one brief bout. Thus, all eggs in a clutch should hatch at the same time, assuming negligible individual variation in developmental rate. However, thermal gradients occur in shallow-nesting reptiles with large three-dimensional clutches (Whitehead 1987a; Thompson 1988), and developmental asynchrony occurs because embryonic development is temperature-dependent (Ewert 1979). Shallower eggs within these nests experience higher and more variable temperatures, both of which reduce developmental time relative to deeper eggs (Georges et al. 1994). Yet hatching and emergence can be more or less synchronous, indicating the evolution of some additional phenomenon that offsets developmental asynchrony (Thompson 1989; Spencer et al. 2001). For this review, synchronous hatching is thus restricted to reptilian species with known or suspected thermal gradients in their nests (see also Spencer and Janzen 2011, for details). A semantic issue affecting the defining of synchronous hatching concerns the degree of synchronicity and its adaptive function. If the embryos within a clutch hatch within 24 h of one another, is this synchronous hatching? Although the boundary between asynchronous and synchronous hatching is nebulous, we would assume an adaptive function for any mechanism that counteracts asynchrony due to thermal gradients. Nevertheless, care should be taken in classifying in the intervals between the extremes of synchronous and asynchronous. The evidence for ECH gathered for the present review was gleaned from scientific papers, popular articles, personal observations, and personal communications from reputable herpetologists or other scientists. In many cases the evidence is anecdotal and may prove to be biologically insignificant. For example, hatching of eggs induced by mechanical disturbance or handling may not reflect a response to increasing risk of predation in nature. However, my assessment provides a starting point (study species) for empirical studies and experiments by offering testable hypotheses regarding adaptive value, context (e.g., flooding), cues, and costs benefits of ECH in reptiles. It is possible in some species that cues normally associated with early hatching are also used to end delayed hatching or promote synchronous hatching, or vice versa. For example, vibration-induced hatching, normally associated with early hatching, could trigger hatching after a period of quiescence (delayed hatching) (J. S. Doody et al., submitted for publication). Thus in some cases the type of ECH assigned in the present review could change for a given species. Evidence for ECH in reptiles Environmentally cued hatching is reported for 41 species of reptiles, including 17 species of turtles, 13 of lizards, 5 of snakes, 4 of crocodilians, 1 tuatara, and possibly 1 worm lizard (Table 1). This number almost certainly grossly underestimates the prevalence of ECH in reptiles because ECH (1) is a recently discovered phenomenon in animals and (2) has rarely been explicitly investigated in reptiles. Indeed, most reports are published or unpublished anecdotes of early hatching in which authorities did

52 J. S. Doody Table 1 Evidence for ECH in reptile embryos Species ECH type Hatching cue Source Turtles Apalone spinifera EH PD, ant attack a S. Doody, personal observation A. mutica EH PD S. Doody, personal observation Aspideretes hurum DH Hypoxia Whitaker and Andrews 1997 A. gangeticus EH, DH PD, hypoxia Whitaker and Andrews 1997; Whitaker 2000 Carettochelys insculpta DH, EH, SH Hypoxia, PD, vibrations, sound Webb et al. 1986; Doody et al. 2001; J. S. Doody et al., submitted for publication Chelodina longicollis SH? R. Spencer, unpublished data Chelonia mydas EH PD, vibrations Bustard 1972 Chelydra serpentina EH PD J. Iverson, personal communication Chrysemys picta SH? Colbert et al. 2010 Emydura macquarii SH, DH a? Spencer et al. 2001 Kinosternon arizonense DH Hypoxia, temperature change Ewert 1991 K. leucostomum DH Hypoxia Horne 2007 K. scorpioides DH Hypoxia b, temperature change Ewert 1991; Iverson 2010 Lissemys punctata EH, DH Vibrations, sound Vijaya 1983a, 1983b; Ewert 1985 Melanochelys trijuga DH a? Ewert 1991 Pseudemys spp. EH PD J. Iverson, personal communication Sternotherus minor DH a? Ewert 1991 Lizards Anolis sagrei EH Hypoxia, PD Losos et al. 2003 Bassiana duperreyi SH Temperature Radder and Shine 2006 Brachylophus vitiensis DH? Morrison et al. 2009 Christinus marmoratus EH PD M. Thompson, personal communication Iberolacerta monticola EH? (fungal infection) Moreira and Barata 2005 Lampropholis delicata EH PD, vibrations S. Doody, personal observation Lampropholis guichenoti EH PD, vibrations S. Doody, personal observation Plestiodon fasciatus EH PD L. Vitt, personal communication Plestiodon inexpectatus EH PD L. Vitt, personal communication Plestiodon laticeps EH PD L. Vitt, personal communication Plica plica EH PD Vitt 1991 Varanus rosenbergi DH? D. Brown, personal communication Varanus varius EH, DH PD D. Brown, personal communication Snakes Coluber mormon EH PD L. Vitt, personal communication Elaphe obsoleta EH PD L. Vitt, personal communication Mastigodryas bifossatus EH PD L. Vitt, personal communication Micrurus fulvius EH PD L. Vitt, personal communication Pituophis catenifer EH PD L. Vitt, personal communication Crocodilians Alligator mississippiensis EH, SH/DH c PD, hypoxia, temperature change T. Rhen, personal communication; M. Thompson, personal communication; Booth and Thompson 1991; Ewert 1991 Crocodylus johnstoni EH, SH/DH c PD M. Thompson, personal communication; Whitehead and Seymour 1990 C. niloticus EH, SH PD, vocalizations Blake 1974; Vergne and Mathevon 2008; Myburgh and Warner 2010 C. porosus EH, SH/DH c PD, vocalizations Magnusson 1980; Whitehead and Seymour 1990 Tuatara Sphenodon punctatus EH? (fungal infection) M. Thompson, personal communication Worm lizards Rhineura floridana EH d PD Carr 1949 Notes: EH¼ early hatching, DH ¼ delayed hatching, SH ¼ synchronous hatching, PD ¼ physical disturbance,? ¼ hatching cue unknown. a Pipping embryos attacked by fire ants (Solenopsis invicta). 3 b Eggs placed in substrate with excess water or nearly submerged. c Inferred from extended declines in metabolism following a peak period (Whitehead and Seymour 1990; Ewert 1991). d Eggshells may have been broken by disturbance rather than vibrations inducing hatching.

Induced hatching in reptiles 53 not formally place their findings into a functional, evolutionary, or adaptive context. It is therefore premature to consider the evolution of ECH across reptiles due to potential bias associated with small sample sizes. Early hatching was the most commonly reported type of ECH (27 spp.), followed by delayed hatching (17 spp.), and synchronous hatching (9 spp.). Six species show evidence of multiple ECH types. In some cases, we do not know if induced hatching reflects early hatching, delayed hatching, or synchronous hatching. Early hatching Our knowledge of early or accelerated hatching in reptiles and is mainly based on anecdotes from professional biologists, and mostly in response to mechanical stimulation arising from handling via humans, transport of eggs, or movements of sibling embryos or hatchlings (Table 1). Early hatching is reported in similar numbers across all groups of reptiles (Table 1). In nearly all cases, vibrations are thus assumed to be the mechanism, although chemical, visual, or auditory cues cannot be ruled out (Warkentin 2000). For example, observations of thunder triggering hatching could be mediated by vibrations of the substrate or sound (Vijaya 1983a; J. S. Doody, submitted for publication). Crocodiles can hatch in response to either physical disturbance or sound (Magnusson 1980; Vergne and Mathevon 2008), and in captivity premature hatching Nile Crocodiles can be induced by the vocalization of embryos from adjacent clutches (Blake 1974). Hatching was apparently induced by immersion in water in two eggs of the lizard Anolis sagrei (Losos et al. 2003), but mechanical stimulation may have confounded the experiment if eggs moved (or were moved) upon immersion. In a few species hatching was induced by a change in temperature after incubations in the laboratory (Table 1). Quantitative studies of three species have demonstrated early hatching in reptiles. Pig-nosed turtles (Carretochelys insculpta) can hatch in response to either hypoxia or vibrations in the laboratory, the former being more closely aligned with delayed hatching because embryos in nature await the onset of wet season flooding before hatching and emerging (Webb et al. 1986; Doody et al. 2001). However, vibrations can induce early hatching, and in laboratory experiments groups of (six) pig-nosed turtle eggs both hatched and emerged from experimental nests significantly faster than did solitary eggs when perfused with gaseous nitrogen or immersed in water (J. S. Doody et al., submitted for publication). Fig. 1 Hatching of a delicate skink, Lampropholis delicata, in a communal nest, after its egg was touched gently by the author. Vibrations can induce hatching up to 7 days early (25% early) in this species. Photograph by N. Pezaro. Therefore, although hypoxia is generally a cue to end delayed hatching, siblings vibrations apparently accelerate hatching in this species. In the lizard Iberolacerta monticola, embryos from eggs experimentally infected with fungus hatched 3 days earlier than did those in control eggs (Moreira and Barata 2005). In that study, the mechanism causing the early hatching is unknown but may be related to the fungus depressing oxygen levels or water potentials (Moreira and Barata 2005). Fungal infection also induced early hatching in a single tuatara egg (M. Thompson, personal communication). The fungal hyphae penetrated the yolk, and the embryo emerged at a smaller body size than did uninfected hatchlings and without residual yolk, which was left in the egg. The obvious adaptive function for this early hatching would be to escape fungal attack, which kills embryos in nature (Moreira and Barata 2005). Only one other study has determined the plasticity in the period of hatching: in the laboratory vibrations induced delicate skink (Lampropholis delicata; Fig. 1) embryos to hatch 2 4 days earlier than controls (J. S. Doody and P. Paull, submitted for publication). Not only does rolling the eggs of this species

54 J. S. Doody and those of Plica plica between the fingers cause early hatching, but hatchlings often launch themselves from the egg and immediately can sprint for up to a half meter (Vitt 1991; J. S. Doody, personal communication). Although costs of early hatching have been demonstrated for other animals (Warkentin 1995, 2000; Warkentin et al. 2001), few have been identified in reptiles. An exception is that premature hatchling Nile Crocodiles retain their yolk sacs and often die (Blake 1974). Early hatching Delicate Skinks are smaller than their spontaneously hatching clutchmates (J. S. Doody and P. Paull, submitted for publication). Hatching early is associated with smaller body size and earlier developmental stage in frogs and fish (Vonesh 2000; Kusch and Chivers 2004) and may mediate a survival cost (Warkentin 1995). Delayed hatching Ewert (1991) considered asynchronous hatching within a clutch as delayed hatching, although he acknowledged that thermal gradients in nests could be responsible in some cases. In this review, I consider only cases in which the entire clutch delays hatching. Delayed hatching is reported for a few species of turtles and crocodilians, but in the latter it is inferred from changes in metabolism (Table 1). The bestknown example is the tropical turtle Carettochelys insculpta, which has evolved a complex strategy for hatching. The eggs are laid in subterranean nests on riverbanks and sandbars during the early dry season, and embryos reach hatching competence in the late dry season about 60 90 days later (Webb et al. 1986; Doody et al. 2001). Embryos defer hatching, however, and instead aestivate for 10 25 days until the onset of the wet season (Doody et al. 2001). During aestivation metabolism is greatly reduced (Webb et al. 1986). Embryos hatch during the flooding of rivers, although heavy rainfall can induce hatching as well (Doody et al. 2001). In the laboratory, hypoxia induces eggs to hatch explosively within minutes (Webb et al. 1986; Doody et al. 2001); in most other turtle species the hatching process spans hours to days (Bustard 1972; Ewert 1985). The precise adaptive benefit to synchronizing hatching with the wet season is not known but is probably an increase in the availability of food and/or reduced risk of predation associated with flooded conditions (hatching at a time when available water is clear and shallow water during the dry season may be risky for small turtles) (Doody et al. 2001). Temporal buffering is probably the ultimate reason for embryonic aestivation in this species. Timing of nesting can vary by up to 5 weeks and is related to the magnitude of the previous wet season(s) (Doody et al. 2004). After small wet seasons late nesting causes considerable nest failure due to flooding associated with the onset of the next wet season, while after big wet seasons early nesting causes embryos to reach the onset of hatching competence during the dry season (Doody et al. 2001, 2004). Aestivation thus decouples the highly variable seasonal timing of nesting from the timing of hatching, thereby allowing turtles to both nest early and hatch later under more favorable conditions. Synchronizing hatching with the onset of the wet season is difficult due to factors creating asynchrony in the timing of embryos reaching hatching competence, including (1) annual variation in the timing of nesting (e.g., Carettochelys), (2) annual variation in the timing of the first wet season rains and thus an unpredictable length of the dry season, and (3) within-year variation in timing of nesting due to multiple clutching. Delayed hatching offsets these factors. There are a number of reptilian species that appear to conform to the pattern of timing their hatching with the onset of the wet season, but for which the strength of the evidence for delayed hatching varies. For example, in the iguana Brachylophus vitiensis two clutches with dates of ovipositions 63 days apart hatched within 2 days of one another (Morrison et al. 2009). Although an increase in developmental rate associated with a seasonal increase in air temperatures can reduce the difference in duration of incubation between early and late clutches (Doody 1995), this thermally based catch-up is not sufficient to offset a 63-day difference in incubation. Rather, embryos in the early Brachylophus clutch almost certainly reached hatching competence weeks prior to embryos in the later clutch but delayed hatching until the environmental trigger of heavy rainfall occurred (Morrison et al. 2009). There are a number of species that appear to synchronize hatching with the onset of the wet season and that possess long and variable incubation times, suggesting delayed hatching. These include the tortoises Geochelone sulcata, G. pardalis, and Astrochelys radiata (Cloudsley-Thompson 1970; Wilson 1968) and the monitor lizard Varanus olivaceus (Auffenberg 1988). The pattern of nesting during the dry season and hatching with the onset of wet season flooding (e.g., in Carettochelys) occurs in the turtle Lissemys punctata, another riverine species, and in all six species of hole-nesting crocodilians (Vijaya 1983b; Thorbjarnarson and Hernandez 1993a). Hatching in response to flooding or hypoxia may occur but has

Induced hatching in reptiles 55 not been confirmed in these species (in this scenario crocodilians would possess both the ability to hatch in response to flooding and maternal assistance in hatching). Hatching in response to flooding/hypoxia does occur in the turtles Aspideretes gangeticus and A. hurum, however. Although these two riverine species do not nest exclusively during the dry season, hatching is associated with rainfall and flooding, and eggs immersed in water in the laboratory hatched almost immediately (Whitaker and Andrews 1993). Reptilian embryos showing a peak of oxygen consumption at about 90% of development followed by a modest decline in oxygen consumption could either possess delayed hatching, or be preadapted to evolve it (Thompson 1989; Ewert 1991). Researchers have considered this pattern to underpin delayed hatching within clutches, thereby facilitating synchronous hatching (Thompson 1989; Whitehead and Seymour 1990). The turtles Melanochelys trijuga and Sternotherus minor and the crocodilians Crocodylus johnstoni, C. porosus, and Alligator mississippiensis may delay hatching based on this pattern of embryonic metabolism (Whitehead and Seymour 1990; Booth and Thompson 1991). Synchronous hatching Two experiments on reptilian species with thermal gradients in their nests indicated that less-advanced sibling embryos in the presence of more developed siblings pipped or hatched earlier than expected, suggesting synchronous hatching in nature (Spencer et al. 2001; Colbert et al. 2010). In the turtle Emydura macquarii embryos placed with more advanced siblings hatched about 5 days earlier than control embryos but still lagged behind their more advanced siblings by about 5 6 days (Spencer et al. 2001). Similarly, embryos of the turtle Chrysemys picta placed with more advanced siblings pipped about 2 days earlier than did control eggs, although pipping in these eggs still lagged behind their more advanced siblings by about 4 days (Colbert et al. 2010). It thus appears that a threshold of developmental asynchrony exists beyond which embryos cannot fully catch up to their more developmentally advanced clutchmates (Colbert et al. 2010). Asynchrony in the above experiments was 7 days for E. macquarii and 11 days for C. picta (Spencer et al. 2001; Colbert et al. 2010). The missing piece of the synchronous-hatching puzzle is the mechanism or cue involved. How do less developmentally advanced embryos recognize their more advanced clutchmates? Vocalizations by embryos can facilitate synchronous hatching in birds and crocodilians (Woolf et al. 1976; Vergne and Mathevon 2008), while vibrations by embryos may perform the role in turtles (J. S. Doody et al., submitted for publication). In both experiments on synchronous hatching, incubating turtle eggs were not in direct contact with one another (Colbert et al. 2010; R. Spencer, personal communication). Perhaps the vibrations of embryos were transmitted through the incubation medium (vermiculite). Suggested cues are auditory sounds associated with pipping, vibrations, or changes in heart rate, O 2, or CO 2 consumption (Colbert et al. 2010). Nitrogenous wastes within late-stage eggs may serve as olfactory cues for embryos (F. Janzen, personal communication). Simple experiments manipulating different possible sensory pathways among eggs should reveal the role of possible cues. Playback experiments using vocalizations or vibrations would confirm the ability of embryos to receive and respond to cues transmitted by sibling embryos and their ability to distinguish informative cues from irrelevant (background) information (Caldwell et al. 2009). The classic adaptive hypothesis for synchronous hatching in species with thermal gradients in their nests is predator dilution (Spencer et al. 2001; Colbert et al. 2010). However, other than in sea turtles there is currently little evidence for synchronous emergence in reptile nests. In several species, emergence occurs one at a time and/or spans more than 1 day (Congdon et al. 1983; Butler and Graham 1995; Doody et al. 2001; but see Tucker 1997). Moreover, field experiments revealed no effect of dilution on survival of Trachemys scripta hatchlings (Tucker et al. 2008). An alternate hypothesis is that asynchronous hatching releases smells that would attract predators to the nest who would then consume unhatched eggs (Lack 1968; Congdon et al. 1983; Vitt 1991). In support of this view, most predation in turtle nests occurs during two distinct periods: within 24 h of laying and during hatching (Ernst and Lovich 2009). During these two periods, eggs smell like egg-laying fluids and hatching fluids, respectively. Finally, in some deeper-nesting species such as sea turtles the ability to escape the nest may be improved by the joint efforts of many hatchlings over the separate efforts of individuals (Carr and Hirth 1961). Costs of synchronous hatching are virtually unknown. However, Colbert et al. (2010) found that sibling turtles that hatched earlier exhibited longer righting times compared to later-hatching sibs, and concluded that this difference in neuromuscular development (motility) reflected a cost to early

56 J. S. Doody hatching. Some precocial birds similarly shorten incubation times and subsequently exhibit reduced motor skills (Vince and Chin, 1981; Cannon et al. 1986). In an example that does not fit the above synchronous hatching strategy, late-term embryos of the skink Bassiana duperreyi enter temporary torpor during cold periods (e.g., at night in a cool climate). This torpor apparently allows embryos in shallow nests to synchronize hatching with aboveground conditions that facilitate successful emergence from the nest (Radder and Shine 2006). The cue for this diel-sensitive hatching is temperature, which varies considerably across a 24-h period in the nests of most reptiles (e.g., Whitehead 1987; Thompson 1988). It is worth noting that synchronizing hatching with environmental conditions is different from, but not mutually exclusive of, synchronizing hatching with sibling embryos. Future directions for revealing ECH in reptiles The present review indicates that reptiles have evolved a diversity of responses to challenges facing the embryo. This diversity may form patterns across reptilian groups; different species or groups may face different egg predators or possess life histories or occupy habitats that expose them to greater, or different, kinds of risks than experienced by others. It is difficult to predict which reptilian groups might have evolved particular types of ECH with information from so few species. Nevertheless, we can make some testable predictions as to which groups might possess certain types of ECH based on attributes such as seasonal climate, nest structure, nesting habitat, egg size relative to the predator s body size, and propensity to nest communally. We can also predict the cues to which embryos may have evolved the ability to recognize and develop appropriate responses. In back-filled subterranean nests of turtles, crocodilians, tuatara, and some lizards, embryos would presumably be exposed to stimuli (e.g., vibrations, sounds, smells) emitted from siblings, predators, and possibly weather (rainfall, thunder). In the more exposed nests of snakes and many lizards (e.g., in leaf litter, logs, mammal burrows, rock crevices) embryos would experience stimuli from other animals besides predators and siblings, stronger weather signals, and probably different consequences of weather (e.g., movement of eggs due to temporary flooding). In these species we might expect embryos to have evolved the ability to distinguish between dangerous and benign cues (Caldwell et al. 2009). Finally, although postovipositional parental care occurs in 55% of reptiles (Shine 1988; Somma 2003), in those species embryos could potentially recognize and respond to cues emitted from their attending mother. Where should we look for predator-induced ECH in reptiles? We might predict that the timing of hatching could advance or retreat based on pre- and posthatching risk (Sih and Moore 1993; Warkentin and Caldwell 2009). Animals with complex life cycles such as amphibians and some insects often face completely different predators, and thus predation pressures, in the egg versus the juvenile stage. For example, in the frog Agalychnis callidryas, eggs attached to a leaf are exposed to risk of predation by snakes and wasps and risk of infection from fungi, while newly hatched tadpoles face risk of predation via insect larvae in the pond below (Warkentin 1995, 2000; Warkentin et al. 2001). Reptiles possess a simpler life cycle in which the same predators of the eggs might also consume hatchlings. Nevertheless, hatchingcompetent reptilian embryos under direct attack from predators or pathogens should hatch early, barring constraints, regardless of their probability of predation as juveniles. Thus, when considering predator-induced hatching, we should probably look for early hatching in response to predators, pathogens, or cues that mimic those dangers (e.g., vibrations). Avoidance of predation by hatching early could evolve if (1) embryos have reached the onset of hatching competence, (2) predation risk is typically high in the period of hatching competence, (3) the embryos have a chance of escaping predation by an egg predator, and/or (4) if the clutch could satiate the predator. Most of what we know about egg predation in reptiles is from turtles, because their conspicuous nesting tracks in the substrate facilitate discovery by humans. Indeed the literature abounds with quantitative studies of predation on turtles eggs. Predation is typically very high, often approaching 100% (Congdon et al. 1987). Although the majority of predation occurs within 24 h of laying, presumably due to olfactory cues left behind by mothers and freshly laid eggs, a second peak of predation may occur when fully developed embryos begin to hatch (Ernst and Lovich 2009). Embryos of amniotic eggs are protected by a parchment or calcareous layer through which oxygen, carbon dioxide, and water can diffuse, but the release of olfactory cues from

Induced hatching in reptiles 57 the embryo is apparently minimal. Once the embryos pip the eggshell, however, chemical odors that can attract predators are evident even to humans with relatively poor olfactory sensitivity (Lack 1968; Vitt 1991). At this potentially vulnerable stage, early hatching (in response to vibrations from predators, for example) could facilitate escape. However, the hatching embryo must also be capable of thwarting capture by the predator. Sluggish, slow-hatching embryos that are small relative to an agile predator would likely not escape. Many reptiles are slow to hatch; turtles and lizards can take several hours to 2 days to completely emerge from the egg after pipping (Fitch 1956; Ewert 1985). In contrast, the rapid escape of some lizard embryos in response to handling indicates the potential for a successful antipredator function of early hatching (Vitt 1991). A final consideration is satiation of the predator, which can be influenced by the size of eggs relative to the size of the predator, by clutch size, and by communal nesting (either within a nest or among nests). Thus, a larger egg predator might be expected to consume all 12 eggs in a clutch of typical freshwater turtle nest but not all 80 eggs in a typical sea turtle nest. The scenarios can be more complex, however: if a predator has just consumed 12 eggs in each of two nearby nests, then partial predation of the clutch in a third nest is certainly possible. Moreover, raccoons and monitor lizards may return to crocodilian nests repeatedly until all the eggs are eaten (Joanen 1969; Magnusson 1982). Careful consideration of such scenarios is required to evaluate the influence of the life histories both of predator and prey on predator-induced hatching. As such, determining the chief egg predators is a necessary first step in interpreting the evolution of predator-induced ECH in a given species. Specialist predators of reptile eggs such as snakes (e.g., Simoselaps, Cemophora, Phyllorhynchus) provide good starting points in the search for predatorinduced ECH. Searching for ECH in response to fungal attack may be easier. Embryonic mortality due to fungal infection occurs in lizards and turtles and probably in other reptiles (Rand 1980; Phillot and Parmenter 2001). High incidences of mortality in sea turtle eggs (Sarmiento-Rameriz et al. 2010) suggest the potential for early hatching in response to fungal infection in those species. Where should we look for abiotic-driven ECH in reptiles? Abiotic-driven ECH includes contexts such as flooding, desiccation, the diel cycle, or changes in temperature or photoperiod (Warkentin and Caldwell 2009). Although predicting which species might respond in these ways is challenging, the occurrence of one type ECH in response to hypoxia via flooding of eggs is somewhat more predictable. Eggs of terrestrial reptiles such as desert species that are not exposed to flooding conditions may not have evolved flooding-induced ECH, whereas riparian-nesting species such as all freshwater turtles and crocodilians and some species of lizards and snakes could experience mortality due to nest flooding and might evolve ECH. Flooding is the leading cause of mortality in many freshwater turtle species and in some crocodilians (Magnusson 1982; Moll and Moll 2004). If the timing of flooding is somewhat predictable, some of these species may delay hatching during a dry period and hatch during the nest flooding that is often associated with the onset of the wet season. Thus, we should search for flooding-induced ECH in species in which hatching coincides with the onset of the wet season, and in which nests are vulnerable to flooding. It is also possible that species that do not synchronize hatching with the onset of the wet season but nevertheless experience a reasonable probability of nest flooding, could evolve flooding-induced ECH. Experiments comparing embryos responses to hypoxia between species hatching during the wet versus the dry season or between those typically exposed to flooding versus those not typically exposed would be a starting point. Aside from flooding it is clear that many reptiles (e.g., monitor lizards, crocodiles, and some turtles) living in regions with wet and dry seasons synchronize their hatching with the onset of the wet season and that hatching-competent embryos of at least some of these species probably delay hatching during the dry season. Thus, future investigations should include species with long and variable incubation periods and those in which the appearance of hatchlings seems to be synchronized with the onset of the wet season (e.g., Auffenberg 1988). Where should we look for synchronous hatching in reptiles? Reptiles with eggs or nests large enough to possess thermal gradients could benefit from evolving a mechanism to counteract those gradients. Nests of turtles, crocodilians, and larger oviparous snakes and lizards would be expected to exhibit thermal gradients. Because of the abovementioned potential adaptive benefits of synchronous emergence, synchronous hatching may be ubiquitous.

58 J. S. Doody Alternatively, there could be costs to synchronous hatching that outweigh its benefits, or situations in which synchronous hatching is of little benefit. For example, in many communally nesting lizards many clutches of eggs are laid in contact with one other but are deposited serially over a period of months; under such a scenario synchronous hatching of the entire communal clutch is not possible. Finally, there could be constraints to synchronous hatching evolving. Conclusions In summary, ECH may be widespread in reptiles but is virtually unstudied. Moreover, the diversity of types of ECH may be comparable to that found in other vertebrates, although it is premature to compare diversity among taxonomic groups. The adaptive value of ECH in reptiles is at times reasonably clear (pig-nosed turtle, Iberian Rock lizard) and other times not (Macquarie River turtle, brown anole). Early hatching may be common in turtles, lizards, snakes, and crocodilians and thus possibly ubiquitous. Interpreting the evolution of predatorinduced early hatching, in particular, is hampered by our limited knowledge of egg predators in reptiles due their secretive nesting habits. Delayed hatching is apparently less common and is reported for only turtles and crocodilians but may occur in some lizards and crocodilians that time their hatching with the onset of the wet season. Synchronous hatching in species with appreciable thermal gradients in their nests is known from only a few turtles and perhaps a few crocodilians. Three types of studies are needed to advance our knowledge of ECH in reptiles: (1) experiments identifying cues for hatching and the plastic hatching period, (2) experiments disentangling hypotheses about multiple hatching cues, and (3) investigations of the environmental context in which ECH might evolve for different species (major predators, abiotic influences on the egg and embryo). The temporal window in which hatching occurs defines ECH in a species, and comparative studies would be valuable for determining how this window evolves (Gomez-Mestre et al. 2008). Once the cues for hatching have been identified, determining the length and timing of the window can be straightforward. However, comprehensive experiments are needed to investigate multiple cues, and in some cases experiments are needed to disentangle hypotheses. Relatedly, in some cases we know more about the cues than we do about their role (e.g., vocalizations in crocodile embryos). Across-species surveys for ECH are required to understand its evolutionary history in reptiles, and the probability of discovering ECH is likely influenced by life history, ecology, behavior, and the environmental niche occupied by each species. Having insightfully predicted that the embryos were subject to natural selection, Darwin would have been excited to learn that embryos of reptiles and other animals could improve their survival by matching their timing of hatching to dynamic environmental risks. The present review should facilitate our understanding of how and under what circumstances ECH evolves in reptiles by stimulating researchers to examine ECH and its consequences more closely. In particular, researchers already working with reptile eggs but naïve to ECH will now be able to study it. More broadly, the discovery of embryo embryo communication in reptiles can teach us how social behavior has evolved in a group with relatively simple social behavior. Acknowledgments The author is particularly grateful to K. Martin, K. Warkentin, and R. Strauthman for the invitation to the symposium, which made this review possible. This review was stimulated by several valuable discussions with K. Warkentin. I thank D. Brown, J. Iverson, R. Shine, R. Spencer, M. Thompson, D. Trembath, L. Vitt, and D. Vogt for providing personal observations, published or unpublished data, or discussions. Funding This symposium was sponsored by the Society for Integrative and Comparative Biology and its Division of Animal Behavior. Travel to the symposium was supported by the National Science Foundation (IOS-136833 and HRD-0820175) and by travel grants from the Monash University Faculty of Science and School of Biological Sciences. References Auffenberg W. 1988. Gray s monitor lizard. Gainesville: University of Florida Press. Blake DK. 1974. The rearing of crocodiles for commercial and conservation purposes in Rhodesia. Rhodesia Sci News 8:311 4. Booth DT, Thompson MB. 1991. A comparison of reptilian eggs with those of megapode birds. In: Deeming DC, Ferguson MWJ, editors. Egg incubation: its effects on

Induced hatching in reptiles 59 embryonic development in birds and reptiles. Cambridge: Cambridge University Press. p. 325 44. Brua RB. 2002. Parent embryo interactions. In: Deeming DC, editor. Avian incubation: behaviour, environment, and evolution. Oxford: Oxford University Press. p. 88 99. Bustard R. 1972. Australian sea turtles. Their natural history and conservation. London: William Collins Sons and Co Ltd. Butler BO, Graham TE. 1995. Early post-emergent behaviour and habitat selection in hatchling Blanding s turtles, Emydoidea blandingii, in Massachusetts. Chelonian Conserv Biol 1:187 96. Caldwell MS, McDaniel JG, Warkentin KM. 2009. Frequency information in the vibration-cued escape hatching of red-eyed treefrogs. J Exp Biol 212:566 75. Cannon ME, Carpenter RE, Ackerman RA. 1986. Synchronous hatching of Darwin s Rhea eggs (Pterocnemia pennata). Physiol Zool 59:95 108. Carr A. 1949. Notes on the eggs and young of the lizard Rhineura floridana. Copeia 1949:77. Carr A, Hirth H. 1961. Social facilitation in green turtle siblings. Anim Behav 9:68 70. Cloudsley-Thompson JL. 1970. On the biology of the desert tortoise Testudo sulcata in Sudan. J Zool 160:17 33. Colbert PL, Spencer R-J, Janzen FJ. 2010. Mechanism and cost of synchronous hatching. Funct Ecol 24:112 21. Congdon JD, Tinkle DW, Breitenbach GL, van Loben Sels RC. 1983. Nesting ecology and hatching success in the turtle Emydoidea blandingi. Herpetologica 39:417 29. Congdon JD, Breitenbach GL, van Loben Sels RC, Tinkle DW. 1987. Reproduction and nesting ecology of snapping turtles (Chelydra serpentina) in southeastern Michigan. Herpetologica 43:39 54. Darwin C. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London: John Murray. Deeming DC. 2004. Post-hatching phenotypic effects of incubation in reptiles. In: Deeming DC, editor. Reptile incubation: environment, evolution and behaviour. Nottingham: Nottingham University Press. p. 229 52. Deeming DC, Ferguson MWJ. 1991. Egg incubation: its effect on embryonic development in birds and reptiles. Cambridge: Cambridge University Press. Doody JS. 1995. A comparative nesting study of two syntopic species of softshell turtles (Apalone mutica and Apalone spinifera) in southcentral Louisiana. Masters Thesis. Louisiana State University, Hammand. Doody JS, Freedberg S, Keogh JS. 2009. Communal nesting in reptiles and amphibians: evolutionary patterns and hypotheses. Quart Rev Biol 84:229 52. Doody JS, Georges A, Young JE. 2004. Determinants of reproductive success and offspring sex in a turtle with environmental sex determination. Biol J Linn Soc 80:1 16. Doody JS, Georges A, Young JE, Welsh MA, Pauza MD, Pepper AL, Alderman RL. 2001. Embryonic aestivation and emergence behavior in the Pig-nosed Turtle, Carettochelys insculpta. Can J Zool 79:1062 72. Ernst CH, Lovich JE. 2009. Turtles of the United States and Canada. 2nd ed. Baltimore, Maryland: Johns Hopkins University Press. Ewert MA. 1979. The embryo and its egg: development and natural history. In: Harless M, Morlock H, editors. Turtles: perspectives and research. p. 333 413. Ewert MA. 1985. Embryology of turtles. In: Gans C, Billet F, Maderson PFA, editors. Biology of the reptilia. Vol. 14 C, New York: John Wiley and Sons. p. 75 267. Ewert MA. 1991. Cold torpor, diapause, delayed hatching, and aestivation in reptiles and birds. In: Deeming DC, Ferguson MWJ, editors. Cambridge: Cambridge University Press. p. 173 92. Georges A, Limpus C, Stoutjesdijk R. 1994. Hatchling sex in the marine turtle Caretta caretta is determined by proportion of development at a temperature, not daily duration of exposure. J Exp Zool 270:432 44. Gomez-Mestre I, Wiens JJ, Warkentin KM. 2008. Evolution of adaptive plasticity: risk-sensitive hatching in neotropical leaf-breeding treefrogs (Agalychnis: Hylidae). Ecol Monogr 78:205 224. Horne B. 2007. The ecology of developmental timing in a neotropical turtle, Kinosternon leucostomum [dissertation]. [Oxford (OH)]: University of Miami, Ohio. Huey RB, Hertz PE, Sinervo B. 2003. Behavioural drive vs. behavioral inertia in evolution: a null model approach. Amer Natur 161:357 66. Iverson JB. 2010. Reproduction in the red-cheeked mud turtle (Kinosternon scorpoides cruentatum) in southeastern Mexico and Belize, with comparisons across the species range. Chelonian Conserv Biol 9:250261. Joanen T. 1969. Nesting ecology of alligators in Louisiana. Proceedings of the annual conference of the southeastern association of game and fish commissioners 23:141 151. Lack D. 1968. Ecological adaptations for breeding in birds. London: Methuen. Losos JB, Shoener TW, Spiller DA. 2003. Effect of immersion in seawater on egg survival in Anolis sagrei eggs. Oecologia 137:360 62. Martin KLM. 1999. Ready and waiting: Delayed hatching and extended incubation of anamniotic vertebrate terrestrial eggs. Am Zool 39:279 88. Magnusson WE. 1980. Hatching and creche formation by Crocodylus porosus. Copeia 1980:359 62. Magnusson WE. 1982. Mortality of eggs of the crocodile Crocodylus porosus in northern Australia. J Herpetol 16:121 30. Moll D, Moll E. 2004. The ecology, exploitation, and conservation of River Turtles. Oxford: Oxford University Press.