Tyrannosaurus, Osborn, 1905

Erickson, Gregory M., 2005, Assessing dinosaur growth patterns: a microscopic revolution, Trends in Ecology and Evolution 20 (12), pp. 677-684 : 680-682

One of the oldest questions in paleontology is how dinosaurs attained giant size. Increased phylogenetic resolution has enabled Carrano to conclude that dinosaurian lineages attained enormous proportions (3+ tons) on at least seven or eight occasions [ 48]. How did these events occur? Evolutionary theory offers three possible solutions: (i) acceleration, whereby growth rates increased from those present in their ancestors; (ii) delay in the onset of maturity; or (iii) through a combination of both processes [ 49]. A growth-curve study published by my research group focused on the evolution of the enormous Tyrannosaurus within Tyrannosauria [ 12] and showed that acceleration in growth rates of fourfold or more was the key to the great stature of this taxon ( Box 3 ). A subsequent study by Sander and colleagues looked at the same phenomenon within the Sauropodomorpha and interestingly revealed the same pattern [ 13].

These findings stand in contrast to three earlier osteohistological studies on gigantism in the dinosaurian outgroups Lepidosauria [ 30], Crocodyliformes [ 50] and Crocodylia [ 51], which all revealed retention of ancestral growth rates and delays in the onset of maturity. It will be interesting to see in the future if all cases of dinosaurian gigantism involved acceleration.

Of course, not all dinosaurs were large. By which heterochronic mechanism(s) did dinosaurs become smaller? There have been several explorations of this phenomenon using osteohistology. One of the most interesting relates to the idea that relatively small dinosaurs found in Eastern Europe (e.g. hadrosaurs and sauropods) are island dwarfs [ 52]. Sander and colleagues tested this hypothesis using diminutive sauropod specimens from Germany [15,16]. They believe that they have found evidence that individuals just 8–9 years of age show histological attributes (tightly packed growth lines called an external fundamental system or EFS that suggest growth was plateauing) indicating full adult size. The same EFS structures occurred at ages of 15 years or later in their giant relatives [ 15]. As in cases of dwarfism in proboscidians (elephants) on islands [ 53], it is believed that selection for smaller sizes enabled these animals to maintain viable population sizes with limited resources.

Along the same lines, a heated debate surrounds whether Nanotyrannus , a purported dwarf species of tyrannosaur, is in fact just a juvenile of Tyrannosaurus [54,55]. Anatomical studies have revealed juvenile attributes in support of the latter hypothesis [55,56]. However, if the dwarfing event simply involved early sexual maturation, immature features might still be expected. I recently aged one of these specimens ( Burpee Museum of Natural History, Rockford, BMR P2002.4.1 ) to see if it has adult histological features or falls outside the confidence interval for

Tyrannosaurus development [ 12].

The 11-year-old specimen plots on the growth curve. This lends support to it being a juvenile of the larger taxon (unless of course maturity occurred somewhat later in development) and suggests there was only one large carnivorous taxon in the Latest Maastrichtian of North America.

The rediscovery that birds are theropod dinosaurs [57,58] has led to interest in how early birds, such as Archaeopteryx , attained small size [10,11,43]. A survey of dinosaurian and avian osteohistological types and formative rates using Amprino’s rule led to the conclusion that the diminutive size of the first birds was brought about by a decrease in the length development compared with that of their dinosaurian ancestors [ 11]. It was posited that selection favored reduced body size because it enabled decreases in wing loading and improved powerto-weight ratios. However, the latest discoveries of Archaeopteryx -sized theropods, such as Microraptor and Sinosauropteryx have led by Xu, Hwang and colleagues to conclude that this miniaturization event actually occurred before the cladogenesis of birds and was not driven by selection related to flight [ 59 – 61]. How then did these dinosaurs become small? The aforementioned dinosaur growth-rate regression ( Box 4 ) and associated theropod growth curves suggest that absolute decreases in growth rates and truncated development facilitated dwarfing [ 10].

Late developmental patterns

In addition to revealing aspects of evolutionary changes in growth rates, osteohistology has utility for analyzing late developmental patterns in dinosaurs. Some of the first dinosaur growth studies reported the absence of EFS structuring that would indicate the specimens were full adult size [21,26,62]. It was theorized that these dinosaurs had an indeterminate growth strategy [19,21,26,62] (i.e. the capacity to grow appreciably throughout life; this is not to be confused with the more common use of this term in ecology, where it refers to sexual maturation before the attainment of maximal body size [63,64]). Subsequent osteohistological analyses have since revealed that EFS structuring is in fact commonplace and it appears that all dinosaurs had determinant growth strategies [10,12,19,21,22,24,31,65]. Growth curves graphically reveal these size plateaus and show that some species of sauropods and tyrannosaurs spent as much as 30% of their total lifespan as full-grown adults [10,12,23]. These growth curve studies also point to an interesting taphonomic conclusion: most specimens in museums are not full-sized adults. Perhaps this is to be expected given that for every specimen that reached late adulthood, many younger individuals perished and are more likely to be represented in the fossil record.

When did the dinosaurs reach somatic maturity (i.e. adult body size)? Growth curves derived from various laboratories reveal that this occurred at an age of 2.5–3.0 years in tiny theropods such as Shuvuuia [ 10], at ~4–12 years in small- to moderate-sized dinosaurs such as Syntarsus and Massospondylus [ 21], at ~16.0–18.5 years in large dinosaurs such as Albertosaurus and Tyrannosaurus [12,24], and at ~20–26 years in giant sauropods, such as Lapparentosaurus [ 62] and Janeschia [ 23].

It has been speculated that precipitous rate changes in growth curves [ 23] ( Box 3 ) or EFS bone structuring [19,26,63] ( Box 1 ) could reflect the onset of sexual maturity, whereby energy allocation shifted from growth to reproduction. However, pending evidence for definitive correlations with sexual reproduction, this deduction is tenuous. Sexual maturity in most animals, including living reptiles, occurs well before full adult size is reached [ 66]. In birds (which are avian dinosaurs), however, it occurs once growth has come nearly to a standstill [ 67]. Furthermore, in most vertebrates, there are multiple pulses in growth rates [ 18]. Discerning which pulse, if any, reflects the onset of sexual maturity remains unclear.

Figures

Box 1. Assessing dinosaur longevity from growth-line counts Making growth curves for dinosaurs requires estimations of longevity for various individuals throughout development. To do this, osteo- histologists sample bones from specimens spanning juvenile through adult developmental stages. The bones that are utilized show minimal remodeling (i.e. replacement or loss of the original bone during life because of metabolic, reproductiveι biomechanical, or skeletal repair considerations [9]) and therefore provide a nearly complete record of development.Traditionally, dinosaur researchers have used the femur, but other bones have been shown to be equally or more efficacious in some taxa. For example, multi-element sampling in tyrannosaurs has shown (Figure Ia) that the bones shown in blue [pubis, fibula, ribs, gastralia (i.e. belly ribs) and some post-orbital skull bones] work better than the femur in these animals [12,24]. The bones are sectioned transversely at mid-shaft using a slow-speed saw fitted with a diamond-tipped blade (Figure Ib). The sections (Figure Ic) are then affixed to glass petrographic slides and sanded/polished for viewing with polarized and/or reflected microscopy. On very large specimens for which making entire cross- sectional slidesisdifficult,researchers polish the cut facesof the bones to reveal ‘polished lines’ that reflect hardness differences between indivi- dual growth lines [20,23,31]. Alternatively, Sanders has developed a method whereby a diamond-tipped drill coring-bit is used to extract a cylinder of bone from which polished-line preparations or petrographic slides can be made [23]. Aging is conducted by making total growth line counts within elements [28]. Here (Figure Ic), a thin-sectioned, dorsal rib from TYrannosaurus rex (Field Museum of Natural History, Chicago, FMNH PR 2081) shows growth lines (arrows) that were counted to age this specimen. Each line represents a period of slowed and/or potential stoppage in tissue deposition [9]. The highly vascularized regions between the rings are known as zones and represent periods of active growth [8,9]. The inset box denotes a region late in development composed of tightly stacked growth rings known as an external fundamental system that indicates when growth slowed precipitously 22,27]. At this point in development, somatic maturity and full-adult size was reached. (Figure I redrawn and reproduced with permission from [12]. Scale bar=10 mm.)

Box 3. Making dinosaur growth curves Longevity estimates are coupled with size data (from direct measures of length or mass estimates from bone circumferences) for dinosaurs to make age-versus-size growth curves.In Figure Iaι femur length was used for the sauropod Janeschia to produce a simple growth curve. This plot was used for determining the age of sexual maturity (hypothesized as occurring when the growth rates initially slowed) and somatic maturity (full adult size as indicated by the asymptote [23]). The curve can also be used to assess linear growth rates at various points in development. (Redrawn and reproduced with permission from [23].) Age-versus-mass growth curvesι such as the one shown in Figure Ib for North American tyrannosaurs [12], are generally sigmoidal in shapeι except in cases where older adult animals are not represented and the asymptote is absent. Timing to somatic maturityι mass standardized maximal growth rates and other life-history parameters can be assessed and used in both interspecific comparative and evolutionary contexts from this type of curve. (Redrawn and reproduced with permission from [12].)

Box 4. Comparison of maximal growth rates in dinosaurs to those in extant vertebrates In Figure I, Maximal growth rates for dinosaurs were deduced from age-mass growth curves for six dinosaurs represented by lettered boxes (from smallest to largest: Shuvuuia (Sh), Psittacosaurus (P), SYntarsus (Sy), MassospondYlus (Ms), Maiasaura (Ma) and Apato- saurus (A); [10]). These data were then plotted with similar data for major living vertebrate clades [5]. A regression line was fitted to the dinosaur data spanning the bounds of known dinosaur size. The results show that whole-organism growth rates for dinosaurs were faster than those of living reptiles of equivalent size. This finding supports qualitative conclusions to the same effect based on tissue- level signal and the reasoning underlying Amprino’s rule [11,26]. However the data do not conform to theories that dinosaurs grew in the same manner as living birds, mammals [6], or at rates between reptiles and birds and/or mammals [8]. Rather, dinosaur growth rates show a unique scaling trajectory. The regression [Maximal growth rate 0.002215 (M)0.925; R2 0.961] also reveals that small = adult = size in dinosaurs involved decreases in absolute growth rates [10]. One caveat of this being the first plot of its kind is that only a small sampling of species was available at the time. The addition of further data, particularly for large sauropods, where Sander [23] has found what appear to be considerably lower growth rates than those reported by Curry [25], might force reanalysis of the aforementioned trends. (Redrawn and reproduced with permission from [10].)