Developmental Plasticity in the Life History of a Prosauropod Dinosaur

See allHide authors and affiliations

Science  16 Dec 2005:
Vol. 310, Issue 5755, pp. 1800-1802
DOI: 10.1126/science.1120125


Long-bone histology indicates that the most common early dinosaur, the prosauropod Plateosaurus engelhardti from the Upper Triassic of Central Europe, had variable life histories. Although Plateosaurus grew at the fast rates typical for dinosaurs, as indicated by fibrolamellar bone, qualitative (growth stop) and quantitative (growth-mark counts) features of its histology are poorly correlated with body size. Individual life histories of P. engelhardti were influenced by environmental factors, as in modern ectothermic reptiles, but not in mammals, birds, or other dinosaurs.

Virtually all dinosaurs studied to date show a primary bone type known as fibrolamellar complex in their long bone wall (13). This bone type indicates fast growth that must have been sustained by a metabolic rate well above that of modern reptiles, if not as high as that of mammals (14). Dinosaurs for which such data are available grow along a species-specific growth trajectory with little individual variation in rate of growth and final size (3, 58), as in mammals (9) and birds (10). Here, we show that the most common early dinosaur had a life history in which its growth was affected by environmental factors such as climate and food availability [developmental plasticity (11)].

P. engelhardti is found in several mass accumulations of medium to large individuals in the Norian of central Europe, such as Trossingen (southern Germany) and Frick (northern Switzerland) (1215). Plateosaurus belongs to a group known as prosauropods, which flourished from the Upper Triassic to the Lower Jurassic, representing the dominant herbivores in faunas of this age worldwide (15). Prosauropod interrelationships are controversial (1518), but prosauropods and sauropods together form a monophyletic group, Sauropodomorpha. At a maximum length of 10 m and a corresponding mass of nearly 4 tons, Plateosaurus was one of the larger bodied prosauropods. Together with some other prosauropods, this dinosaur was the first to reach the large body size generally attributed to dinosaurs, and the first high browser to evolve.

We sampled the histology of long and girdle bones of P. engelhardti from Trossingen and Frick (19) (table S1). Plateosaurus long bones are characterized by a large medullary cavity (49% to 58% of shaft diameter) and relatively thin bone walls (Fig. 1). The cortex is sharply set off from the medullary cavity with little or no secondary cancellous bone and rare resorption spaces in the compact bone. The primary bone of the cortex is dominated by growth cycles of fibrolamellar bone, ending in a line of arrested growth (LAG) (Fig. 2). Vascular canals are primarily circumferential, and vascularity decreases toward the LAG (Figs. 1 and 2). Growth-cycle width decreases substantially toward the outer bone surface (Fig. 1A). In one group of specimens, fibrolamellar bone is the last tissue type to have been formed (Fig. 2). We assigned the ontogenetic stage of “fast growth” to these specimens, because fibrolamellar bone deposition indicates a high growth rate. A strong decrease in growth rate is documented in the last bone tissue deposited in many other specimens, in which growth cycles in fibrolamellar bone in the outer cortex become narrow and less vascularized (Fig. 2). We assigned the “slow growth” stage to this second group. A third group of specimens survived to an even later ontogenetic stage, as evidenced by lamellar-zonal bone with closely spaced LAGs and poor to absent vascularization in the outermost cortex (Fig. 2). This tissue type is also known as an external fundamental system and documents a growth plateau, i.e., that final body size had been reached. Individuals in this group were thus scored as “fully grown.”

Fig. 1.

Histology of P. engelhardti long bones, polished sections, normal light. (A) Cross section of a humerus shaft (right humerus NAA F 88/B640, 44.5 cm long, Frick). Arrows mark LAGs. Note the decrease in LAG spacing toward the periphery of the bone. Scale bar, 1 cm. (B) Core section (left femur IFG compactus, 74 cm long, Trossingen) with growth-mark count. Numbers mark LAGs. IFG, Institut für Geowissenschaften, Universität Tübingen, Germany; NAA, Naturama, Aarau, Switzerland. Scale bar, 3 mm.

Fig. 2.

Ontogenetic stages in the histology of P. engelhardti long bones. (A) Polished section (left femur IFG compactus, 74 cm long, Trossingen) showing laminar fibrolamellar complex with LAGs. This tissue type indicates fast growth of the cortex. (B) Thin section from the same specimen showing the same bone tissue. (C) Polished section (right humerus NAA F 88/B640, 44.5 cm long, Frick) showing fibrolamellar bone with LAGs followed by lamellar-zonal bone in the outer cortex. This tissue type indicates slow growth. (D) Thin section (left tibia MSF 1, 51 cm long, Frick) with closely spaced LAGs in outermost cortex. This tissue type indicates that the animal was fully grown. Bone surface is beyond [(A) to (C)] or at top (D) of image. IFG, Institut für Geowissenschaften, Universität Tübingen, Germany; MSF, Sauriermuseum, Frick, Switzerland; NAA, Naturama, Aarau, Switzerland. Scale bars, 1 mm.

Surprisingly, we found such fully grown individuals virtually across the whole size range sampled (19). Some individuals had reached final size at 4.8 m body length (BL), whereas others attained 10 m BL (Fig. 3). Similarly, the “fast growth” and the “slow growth” stages were also found at widely differing body sizes (Fig. 3). Size at the slow growth stage is close to final body size because not much bone tissue was added to the circumference of the bone during this stage.

Fig. 3.

Relation between body size, age, and ontogenetic stage in specimens of P. engelhardti, based on bone histology. Proxy for body size is femur length. Data for plots are in table S1. (A) Relation between age and body size in the Trossingen specimens. For each specimen, observed LAG count (black symbols) and estimated total LAG count (blue symbols) are plotted. Estimated total LAG count is equivalent to individual age in years. Note poor correlation between age and size. (B and C) Relation between growth stage and body size in the samples from Trossingen (B) and Frick (C). Note poor correlation between growth stage and size in both samples. The large variation in final body size is best appreciated if the slow growth individuals, which would not have grown much larger, are considered together with the fully grown individuals.

Life history was quantified by applying skeletochronology to long and girdle bones (19). We estimate that the youngest specimen in the sample was 9 years old, whereas the oldest had reached 26 to 27 years (Fig. 3). This specimen had attained nearly final size but was still growing slowly. The minimum age for a fully grown specimen was 12 years. However, in agreement with our observations about the qualitative growth record, we found a poor correlation between body size and age (Fig. 3). This is most obvious from the “slow growth stage” sample from Trossingen (n = 19, r = 0.383, no significant correlation at P = 0.001) and the “fully grown” sample from Frick (n = 5, r = 0.114, no significant correlation at P = 0.001).

Two lines of evidence, growth stage assignment and skeletochronology, thus indicate that growth rate and final size varied strikingly in individuals of the species P. engelhardti. Hypotheses to explain this extreme variability include: (i) that more than one biological species is represented by the material identified as P. engelhardti, (ii) that the material represents a single species with a strong sexual size dimorphism, (iii) that our methods for detecting growth stage and age are inadequate, and (iv) that P. engelhardti had strong developmental plasticity in life-history parameters. We view the last interpretation as the most credible. We reject hypothesis (iii), methodological problems, because termination of growth (Fig. 2) can be detected histologically with confidence (3, 7), although our skeletochronological age estimates may be inaccurate because of the lack of juvenile Plateosaurus (Fig. 3). We also reject hypothesis (ii), sexual dimorphism, because terminal body size shows a unimodal distribution, albeit with a high standard deviation (Fig. 3), and not the bimodal distribution expected for sexual morphs. Sexual dimorphism in aspects of morphology and in metric characters has been postulated for Plateosaurus but cannot be proven (20, 21). Finally, the morphology and systematics of Plateosaurus have been intensively studied in recent years, and all authors agree that there is only evidence for one species among the Plateosaurus fossils from the rich bone beds of central Europe (1315). This leads us to reject hypothesis (i).

Additional support for hypothesis (iv), strong developmental plasticity, comes from the observation that the Plateosaurus individuals from Frick show the same great variability in final size and growth rate as the individuals from Trossingen (Fig. 3). The individuals from Frick are smaller on average and show a smaller final size. The Frick bone bed may represent a population of generally smaller stature, possibly due to a less favorable habitat or because the Trossingen and Frick rocks have different ages.

Outgroup comparison based on long-bone histology indicates that the strong developmental plasticity is plesiomorphic for archosaurs and was retained in the crocodile lineage (1, 2, 4). Virtually all basal archosaurs and pseudosuchians have lamellar-zonal bone with numerous and distinctive LAGs (4). Ornithodirans (pterosaurs and dinosaurs), on the other hand, had lost developmental plasticity, as indicated by the predominance of the fibrolamellar complex (1, 2, 4, 22, 23). The strong developmental plasticity of Plateosaurus is a reversal to an ancestral condition and contrasts with all of the more derived dinosaurs and two other basal dinosaurs, the prosauropod Massospondylus (7, 24) and the basal saurischian Herrerasaurus [although only two specimens of Herrerasaurus were sampled (2, 22)]. Recent phylogenetic analyses (1518) agree that Massospondylus and Plateosaurus are closely related, suggesting that different species of prosauropod dinosaurs varied in their degree of developmental plasticity. It may be no coincidence that a similarly unexpected reversal to an ancestral condition, i.e., quadrupedality, was recently discovered in embryos of Massospondylus (25).

In extant amniotes, strong developmental plasticity is correlated with a low metabolic rate and behavioral thermoregulation [the ectotherm strategy (26)], resulting in widely differing growth rates and final sizes in individuals of the same species of, e.g., turtles (27, 28), lizards (29, 30), and crocodiles (31, 32). The observed strong developmental plasticity in Plateosaurus would suggest that it also was an ectotherm. This disagrees with the dominance of the fibrolamellar complex in the long-bone cortex of this dinosaur [which is not known from modern ectotherms in natural habitats and indicates high growth rates (1, 33)] and its advanced dinosaurian locomotor apparatus. The early dinosaur P. engelhardti possibly represents the initial stage in the evolution of metabolic thermoregulation (endothermy) in dinosaurs, in which endothermy was decoupled from developmental plasticity.

Supporting Online Material

Materials and Methods

Table S1


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article