Titanosaur nesting strategies

The taphonomy of behavior emphasizes the need for defining trace fossils in terms of both classical ichnology and modern behavioral biology (Plotnick, 2012). Furthermore, the concept of behavioral fidelity, expressed as the “extent to which trace fossils preserve original behavioral signals” (Plotnick, 2012), allows behavioral interpretations contrastable to currently existing data. Considering solely the sedimentological evidence, none of the above-mentioned nesting sites display a diagnostic fossil nest structure. In the absence of direct evidence of preserved fossil nests, eggs and their spatial grouping, eggshell morphologies, coupled with observations of surrounding sediments, provide the only data with respect to nesting moisture content and heat, as those are the main extrinsic parameters that are paramount for hatching success. Pore canals allow the diffusion of gases and water vapor through the eggshell (Paganelli, 1980). Their size, geometry, and number reflect a specialization to the habitat where nesting occurs (Williams, Seymour & Kerourio, 1984; Cousin, 1997; Grellet-Tinner, Fiorelli & Salvador, 2012). Eggshell ornamentation (or lack thereof) is another morphological characteristic from which a nesting paleoenvironment can be hypothesized. Cousin (1997) and Cousin & Breton (2000) used the characteristics of the nodular appearance of eggshells from the Late Cretaceous of France as an indicator of the substrate, or nesting material, surrounding the eggs. This character (Grellet-Tinner, Chiappe & Coria, 2004; Grellet-Tinner & Zaher, 2007; Grellet-Tinner et al., 2011) was regarded as a specialization, increasing gas conductance through the pores that are located around and at the base of each node, by preventing nesting debris from plugging their apertures (Sabath, 1991). The densely packed nodular structures typically observed on the outer surfaces of pristine titanosaur eggshells provide a substantial increase of surface in contact with the surrounding environment, suggesting they could also have acted to buffer acidic erosion during long periods of incubation, an interpretation consistent with GGT and colleagues’ (G Grellet-Tinner, 2014, unpublished data) observations on modern megapode eggshells and associated nesting environments in Australia. Gas diffusion occurs through the pores and can be quantified by the eggshell’s G H 2 O (Seymour, 1979). It is commonly obtained for modern birds and reptiles by measuring water loss in a known vapor gradient across the shell (Ar et al., 1974). G H 2 O for fossil eggs can be estimated from eggshell thickness and pore-system geometry and so forms a valuable proxy for assessing the moisture content in dinosaur nests, their environments (Seymour, 1979; Seymour & Ackerman, 1980), and nesting strategies (Ar et al., 1974; Seymour, 1979; Birchard & Kilgore, 1980; Seymour et al., 1987; Grellet-Tinner, Chiappe & Coria, 2004; Deeming, 2006; Grellet-Tinner, Fiorelli & Salvador, 2012).

Titanosaurs could not have used the classic contact incubation strategy typical of most modern dinosaurs (Seymour, 1979; Werner & Griebeler, 2011; Ruxton, Birchard & Deeming, 2014), thus must have relied on external environmental heat for incubating their eggs. Interestingly, the extant Australasian megapodes are one of the most intriguing avian dinosaur families because they exhibit a practice unique among modern birds, regarded as a reversal character, of incubating their eggs by utilizing only environmental heat sources rather than body heat (Booth & Thompson, 1991; Jones & Birks, 1992; Del Hoyo, Elliott & Sargatal, 1994). Moreover, they display the most diversified incubation behaviors among ground nesting archosaurs (Jones & Birks, 1992; Del Hoyo, Elliott & Sargatal, 1994; Harris, Birks & Leaché, 2014), which are associated with species specific nesting strategies and nesting sites (Boles & Ivison, 1999; Harris, Birks & Leaché, 2014). These are: (1) mound-building; (2) burrow-nesting using geothermal heat; (3) burrow-nesting using solar-heated beaches; (4) burrow-nesting using decaying tree roots; (5) mound parasitism. As such, Megapodiidae potentially offers a valid proxy for titanosaur reproductive behaviors.

Among the titanosaur nesting sites reviewed in this investigation, only two have been positively related with geothermalism: the well-documented Sanagasta and the South Korean eggs (Table 1). The compact arrangement of the Sanagasta eggs, in clutches of one or two layers, led Grellet-Tinner & Fiorelli (2010) and Fiorelli et al. (2012) to suggest incubation in excavated nests, a strategy already inferred by Paik, Huh & Kim (2004) and Paik, Kim & Huh (2012) for the eggs from Seonso Formation (Gyeongsang Basin, South Korea) (Table 1). The eggs from these sites have a similar morphology (but a thinner eggshell thickness in Gyeongsang) and they also share sedimentary and geological features of geothermal activities concomitant with the oviposition. This supports the hypothesis that geothermal heat sources could have been used by certain species of Cretaceous titanosaurs. Several megapode species—Macrocephalon maleo Muller, 1846, Eulipoa wallacei Gray, 1861 and many Megapodius spp.—use burrow nesting with vegetal decomposition, solar radiation and/or geothermalism (inclusive here of volcanism) as incubating strategies (Frith, 1956; Dekker & Brom, 1960; Jones & Birks, 1992; Del Hoyo, Elliott & Sargatal, 1994; Göth & Vogel, 1997; Dekker, 2007; Bowen, 2010; Harris, Birks & Leaché, 2014). The remarkable instance of opportunistic nesting in geothermal settings represents an adaptive case in which species avoid thermally heterogeneous nesting environments by exploiting geothermal conditions that maintain higher and more constant temperatures and moisture levels in egg clutches (Werner, 1983; Göth & Vogel, 1997; Chen, Kam & Lin, 2001; Wu & Kam, 2005; Guo et al., 2008; Huang et al., 2009; Sas, Antal & Covaciu-Marcov, 2010; Grellet-Tinner & Fiorelli, 2010; Grellet-Tinner, Fiorelli & Salvador, 2012). For example, the Malau megapode (Megapodius pritchardii Gray, 1864) digs pits more than 2 m deep to utilize underground geothermal heat (Frith, 1956; Del Hoyo, Elliott & Sargatal, 1994; Göth & Vogel, 1997). In addition, because burrow nests in geothermal ecosystems are less susceptible than mounds to predation, burrow-nesting megapodes can abandon their nests after burying their clutch (Dekker, 1989; Del Hoyo, Elliott & Sargatal, 1994), a strategy consistent with the titanosaur behavior inferred from the fossil record (Sander et al., 2008; Sander et al., 2011; Werner & Griebeler, 2011; Ruxton, Birchard & Deeming, 2014), where nesting sites were often located and synchronous with geothermal activities. Paradoxically, megapode nest-burrows are dug preferably in soft soils, in volcanic sands, environments that inhibit preservation of such structures in the fossil record because they easily collapse (Frith, 1956; Dekker & Brom, 1960; Roper, 1983; Bowen, 2010).

Although, the first reports of AM nesting site implied a monotaxic titanosaur assemblage, further refined to nemegtosaurid titanosaur (García et al., 2010), one of the co-authors (Eagle et al., 2015) has determined that the egg-laying titanosaurs in AM L#4 may represent a different nemegtosaurid species, certainly closely related to those nesting in Auca Mahuevo layers 1–3 (AM L#1–3) but displaying sufficient autapomorphies to justify a species variation (Table 1). This species variation is also supported by an environmental change. Celestite geodes and barite are ubiquitous in AM L#1–3 (Garrido, 2010a; Garrido, 2010b). These two minerals are readily produced in geothermal and evaporitic settings, both equally possible at AM due to its particular geology (Jackson, Schmitt & Oser, 2013). Moreover geochemical analyses reveal a higher concentration of magnesium and lithium in AM L#1–3 than AM L#4 (Eagle et al., 2015). These two minerals are prevalent in continental brines, which could also be formed under high evaporitic conditions or geothermalism. However, the evidence available suggests AM have been selected first by a certain nemegtosaurid species (AM L#1–3) for its presence of limited rivers in a semiarid environment and then, after a climatic change toward wetter conditions, replaced by another closely related species with more conspicuous nodular eggshell ornamentation, adapted to a more humid nesting environment (Table 1). Climatic and environmental changes are also documented in the clay fabric in the 4 AM layers (Jackson, Schmitt & Oser, 2013). Interestingly, among the modern Megapodiidae a similar niche partition is observed in Australasia. The mound-builder Alectura lathami Gray, 1831 (brush-turkey) nests in wetter environments in coastal Australian regions with respect to its congener Leipoa ocellata Gould, 1840 (malleefowl) that nests in semi-arid settings. Although both species are contemporaneous their geographical niche partitioning is related to vegetation and climatic differences. Such species specific nesting partitioning in modern Australia may explain the nemegtosaurid successive species replacement in AM from a dryer environment nesting adaptation, such as occurs in AM L#1–3, to a wetter setting, recorded upwards by the transition to the Allen Formation’s estuarine-coastal sediments (Garrido, 2010a; Garrido, 2010b). Combined with the lack of convincing geothermal evidence, although not entirely discounted due to the episodic explosive volcanism (Jackson, Schmitt & Oser, 2013), the horizons with egg clutches and the eggshell structures would suggest mound-building nesting strategies with a dryer climate for AM L#1–3 which display shallower nodular eggshell ornamentation and wetter for AM L#4 with pronounced nodular ornamentation (Table 1). The oological material from AM L#4 and Haţeg are very similar, sharing several synapomorphies including egg size, shape and eggshell microstructure (Grellet-Tinner, Chiappe & Coria, 2004; Grellet-Tinner et al., 2012). The complexity of their pore systems is consistent with a morphological adaptation to high moisture nesting environments, typical of burial conditions. Furthermore, in AM a semi-arid dry/wet climate coupled with episodic volcanism contributed to vertisol horizons development at the floodplain areas (Garrido, 2010a; Jackson, Schmitt & Oser, 2013). Rhizoliths and root traces as well as small fossil logs found there (Garrido, 2010a; Garrido, 2010b; Jackson, Schmitt & Oser, 2013) suggest a floristic abundance that would promote ideal conditions for mound-nesting behaviors.

Although crocodilian mound-nests also average 1 m high and 3m in diameter (Joanen, 1969; Webb, Messel & Magnusson, 1977; Seymour & Ackerman, 1980; Waitkuwait, 1989), it is important to distinguish this type of vegetal mounds from those of the Australian brush-turkey megapodes, built from humus, soil and smaller amounts of true vegetal matter. Modern megapodes construct surprisingly large nest-mounds, relative to their egg and clutch sizes. Although a few mounds could reach 4 m in height, 18 m in length and 5 m in width, a typical brush-turkey mound commonly measures 1 m by 4 m (height and diameter respectively) and requires up to 5 tons of soil mixed with vegetal matter (Seymour & Ackerman, 1980; Jones & Birks, 1992; Del Hoyo, Elliott & Sargatal, 1994; Harris, Birks & Leaché, 2014). Conversely, nesting structures smaller than 0.75 m high and 2 m long are not functional (Jones & Birks, 1992; Del Hoyo, Elliott & Sargatal, 1994; Jones & Göth, 2008). In contrast, malleefowl mounds consist mostly of sandy material with terrigenous clasts. Malleefowl build their mound, and then dig out the center into which semi-arid to arid vegetation, like spinifex, is introduced before the rainy season (Frith, 1959; Jones & Birks, 1992; Jones & Göth, 2008; D. Booth, pers. comm., 2013). Eggs are laid around this center core that provides sufficient heat from vegetal decomposition to support embryonic development. This results in all eggs being placed relatively centrally within a mound (D. Booth, pers. comm.). Temperatures in the mound range from 27 to 38 °C, although eggs mostly lie in the range of 32 to 36 °C (Booth, 1987). In contrast, brush-turkey females burrow into the mound obliquely. The litter material does not collapse around the tunnel as sand would in a malleefowl mound (D. Booth, pers. comm., 2013). Hence, eggs can be dispersed throughout the place in large mounds and not just in the central core as in the malleefowl ones. One strategy would result in eggs grouped together in a compact clutch, while the other results in eggs isolated or lined up in small groups when the mound is eroded away. In comparison, the latter matches the pattern already observed in the titanosaur nesting sites of Southern France, where Cousin et al. (1989) and Cousin & Breton (2000) interpreted that small groups of eggs (2–4) are not randomly distributed but belong to circular “supergroups” of up to 15 eggs and 3.5 m in diameter (see Fig. 10 in Cousin & Breton, 2000), substantially separated from others (Cousin et al., 1989; Cousin & Breton, 2000). Additionally, the few-egg clutches found in close association in Coll de Nargó that were interpreted by Vila, Jackson & Galobart (2010) as partially preserved ∼ 25-egg clutches, show similar patterns. As such, the egg spatial distribution, in small clusters linearly to compactly grouped, but contained in round shaped areas of up to 2.3 m (see Table 1; Vila et al., 2010b), would either support burrow- or mound-nesting (Cousin & Breton, 2000).

The Upper Cretaceous eggs and eggshells found at the reviewed nesting sites of Southern France and Northern Spain share many characteristics. Regarding their similarities (Table 1), and considering the high conductance values of the Pinyes’ eggs, a highly humid incubation environment can be inferred for both locations. Although the evidence at hand suggests burial incubation, no record of organic matter or hydrothermal relics has been identified in their fine surrounding sediments. Regarding the morphological aspects, the minimal differences in size between eggshell pores and the fine sediments reported in both cases obscure the inference of vegetation as nesting material.

Although there are no published analyses on gas conductance for the Dholi Dungri eggs, the high values obtained from other Indian localities, ranging between 2,650 mgH 2 O/(day∗torr) and 3,490 mgH 2 O/(day∗torr) (Sahni et al., 1994), suggest a high moisture nesting microenvironment. Considering the nodular appearance of the eggshells as a good indicator of the surrounding nesting material (Cousin & Breton, 2000) the provision of substantial amounts of plant debris as constructing material cannot be discarded. Although the sediment itself is coarse enough to prevent pore obstruction, the well-developed ornamentation of the eggshells could be a mechanism to prevent the external eggshell erosion, by the acid formed by decomposing microbes and fungi during an extended incubation period, like in modern megapodes. The morphological and sedimentological evidences at hand, coupled with the interpretation of a semi-arid, tropical dry-wet climate for the Lameta Formation during Maastrichtian (Tandon et al., 1995; Tandon & Andrews, 2001; Wilson et al., 2010; Prasad & Sahni, 2014), allows us to infer an A. lathami-like nesting strategy, similar to that suggested for AM L#4 (Table 1). However, regarding the magnitude of the volcanism responsible of the Deccan Traps deposition, during Late Cretaceous, the geothermalism could also be regarded as an equally plausible alternative heat source for egg incubation. Interestingly, although Tandon et al. (1995) suggest the nesting activity in the Lameta Formation was contemporaneous with the first Deccan lavas, questions regarding its possible relationship still require further research.

Many species of modern megapodes are known to use mound-nesting strategies to incubate their eggs (Harris, Birks & Leaché, 2014). However, they are also known to revert to less conspicuous nesting strategies such as burrow-nesting (Dekker, 1989) because of their behavioral lability (Harris, Birks & Leaché, 2014). As such, although dusky megapode (Megapodius freycinet Gaimard, 1823) populations are categorized as mound-builders, they are also burrow-nesters and exploit geothermal resources for heat and moisture in New Britain and the Solomon Islands (Roper, 1983). This species demonstrates an interesting behavioral plasticity (Frith, 1956), yet still relying on environmental heat only. Additionally, the Moluccan megapodes (Eulipoa wallacei) lay their eggs in black volcanic sands and silt utilizing solar heat (Del Hoyo, Elliott & Sargatal, 1994; Dekker, 2007) but are also known to switch to geothermal and microbial decomposition. Because of the scarcity of suitable nesting areas, burrow-nester megapodes (e.g., M. freycinet, M. maleo, and M. pritchardii) commonly adopt colonial nesting behaviors and site fidelity (Del Hoyo, Elliott & Sargatal, 1994). Similarly, strong nesting environment selectivity and colonial nesting were related to the massive occurrence of eggs and egg clutches in similar lithofacies, in some of the best known titanosaur nesting sites (Sahni et al., 1994; Mohabey, 2001). Although some titanosaur species may have adopted nesting strategies relying on various styles of mound-nests, it is entirely plausible that a few of them might have reverted to such opportunistic behaviors when available, particularly in settings like Haţeg (Grellet-Tinner et al., 2012), India (Tandon et al., 1995) and AM (Jackson, Schmitt & Oser, 2013) where volcanism and related processes were coeval with nesting titanosaurs (Table 1).

Seymour & Ackerman (1980) emphasized that vegetal decomposition through microorganisms could be critical in terms of gas tension for mound-builder archosaurs. When clutches are large, as occurs with the ∼20 egg clutches of the Australian brush-turkey, the gas tensions as well as the temperature in the mound vary with the reworking of the mound by parents (Seymour & Ackerman, 1980; Del Hoyo, Elliott & Sargatal, 1994; Priddel & Wheeler, 2003). Assuming similar incubation strategies, it seems that the sizes of the sauropods could have prohibited active reworking of the mound to compensate for gas and temperature variations. Yet, such post-oviposition activities might have not been required, e.g., in modern alligators, that exhibit similar “en masse” laying behaviors.