Lines of Arrested Growth

The recognition that vertebrate bones preserve cyclical growth marks has a long history in comparative anatomy. Frank E. Peabody's 1961 paper "Annual growth zones in living and fossil vertebrates" in the Journal of Morphology provided the first systematic survey of annual growth zones across a wide range of living and fossil vertebrate taxa. This work established the foundational observation that bone growth in many vertebrates is not continuous but punctuated by periodic interruptions.

The specific term "lines of arrested growth" (LAGs) was introduced by the French bone histologist Jacques Castanet in 1974, distinguishing these sharp cessation lines from the broader, more diffuse annuli that represent periods of merely reduced (rather than halted) growth. In 1990, Francillon-Vieillot, de Buffrénil, Castanet, and colleagues published a comprehensive classification of vertebrate skeletal tissue microstructure in the volume Skeletal Biomineralization, formally standardizing the terminology for LAGs, annuli, zones, and other histological features. This standardization was further consolidated by Castanet, Francillon-Vieillot, Meunier, and de Ricqlès in 1993 in their chapter "Bone and individual aging" within the Bone series, which remains a key reference for skeletochronological methodology.

Cortical bone grows appositionally outward from the periosteal surface. During periods of active growth, rapidly deposited fibrolamellar bone with abundant vascular canals forms wide zones. When environmental or physiological conditions become unfavorable, bone deposition slows dramatically or ceases entirely, producing a LAG at the periosteal surface. Once growth resumes and new bone is deposited over the LAG, it becomes embedded within the cortex as a permanent record of that growth hiatus.

Under plane polarized light (PPL), LAGs appear as distinct thin dark lines within the cortex. Annuli, by contrast, are diffuse bands of slower-growing tissue—typically parallel-fibered bone within a woven matrix—characterized by reduced vascular canal density and flattened osteocyte lacunae. Annuli are often more readily identified in cross-polarized light (XPL) or circularly polarized light (CPL) due to the higher birefringence of their organized collagen fibers. Importantly, LAGs and annuli are not mutually exclusive: a single growth mark can transition from a LAG to an annulus at different positions around the bone cortex.

Recent work has identified additional complexity in growth mark morphology. Some cortical growth marks (CGMs) occur as "multiplets"—two or more closely spaced parallel lines bounding thin zones with little or no vascularization between them. The biological significance of multiplets remains debated: they may represent multiple brief growth cessations within a single year, or they may reflect distinct annual events.

The systematic application of LAG-based skeletochronology to dinosaurs was pioneered in the early 2000s. Erickson, Curry-Rogers, and Yerby published a landmark 2001 paper in Nature that used LAG counts from long bone thin sections of multiple dinosaur taxa to construct the first mass–age growth curves for dinosaurs. Their results demonstrated that dinosaur growth rates far exceeded those of living reptiles and, in some lineages, approached or matched those of modern birds and mammals.

In 2004, Erickson and colleagues extended this approach to tyrannosaurid dinosaurs, again in Nature. Analysis of seven Tyrannosaurus rex specimens revealed ages ranging from approximately 15 to 25 years, with the largest and oldest specimen—FMNH PR2081, popularly known as "Sue"—estimated to have died at a minimum age of approximately 28 years. The growth curve showed that T. rex experienced a dramatic growth spurt between roughly 14 and 18 years of age, gaining up to approximately 2.1 kg per day during peak growth, before leveling off as the animal approached adult size.

Erickson's 2005 review in Trends in Ecology and Evolution synthesized these advances and coined the phrase "a microscopic revolution" to describe the transformation that osteohistology had brought to dinosaur paleobiology. Using LAG data, researchers could now reconstruct not only individual ages but also growth strategies, the timing of sexual maturity, and population-level demographic parameters.

Dinosaur growth curves are typically modeled using sigmoidal functions (logistic, Gompertz, von Bertalanffy, or Richards curves). The process involves several steps. First, long bones—preferably femora or tibiae, as their circumferences can be converted to body mass estimates—are sectioned transversely at the diaphyseal midshaft, where the cortex is thickest and the medullary cavity is smallest, preserving the most complete growth record. The circumference traced by each LAG provides a proxy for body size at that age. By counting inward from the periosteal surface, each successive LAG corresponds to a progressively younger year of life.

Because medullary cavity expansion destroys the innermost (earliest) LAGs, the absolute age at the first preserved LAG must be estimated. This is commonly done through superimposition of growth series from multiple individuals of different sizes, or by retrocalculation methods that extrapolate inner LAG spacing. Once absolute ages are assigned, circumference-at-age data points are fitted to a sigmoidal curve using nonlinear regression.

The spacing between LAGs is informative beyond simple age counting. Wide inter-LAG spacing indicates rapid growth during that year, while narrow spacing indicates slow growth. In many later-branching dinosaurs, inter-LAG spacing progressively decreases toward the periosteal surface, reflecting declining growth rates as the animal approaches adult size—consistent with asymptotic growth models. The outermost cortex may exhibit an External Fundamental System (EFS), a band of closely spaced LAGs within slow-growing lamellar bone, signaling that the individual had reached skeletal maturity.

A 2026 study by Woodward and colleagues presented the most comprehensive histological analysis of Tyrannosaurus rex species complex ontogeny to date, based on femoral and tibial transverse sections from 17 individuals. This study introduced several methodological innovations: it included annuli visible only under cross-polarized light and counted individual lines within multiplets as separate annual events. When all visible growth marks were incorporated, the best-fit sigmoidal growth curve yielded an asymptotic age of approximately 35–40 years—roughly 15 years later than the earlier estimates of Erickson et al. (2004). This suggests that T. rex grew more gradually over a longer lifespan, with a protracted period of subadult development. The authors also found no strong link from extant vertebrate data supporting the idea that the growth curve inflection point corresponds to sexual maturity, challenging a widely cited interpretation.

Despite its widespread adoption, LAG-based skeletochronology faces significant methodological challenges.

Loss of early LAGs through medullary remodeling: As bone matures, the medullary cavity expands and secondary Haversian remodeling can obliterate inner cortical LAGs. This means the number of observed LAGs is a minimum estimate of age, and corrective methods are required.

Non-annual LAGs: A single year may produce zero, one, or multiple growth marks depending on the organism's environment and physiology. Hibernation and estivation in the same year, for instance, could produce two LAGs, while an exceptionally favorable year might produce none. This introduces systematic uncertainty into age estimates.

Methodological discrepancies: Schucht, Klein, and Lambertz (2021) conducted the first systematic comparison of petrographic ground sections (used primarily in paleontology) and stained microtomized sections (used in neontology) applied to the same bones. They found that the two techniques yielded divergent growth mark counts in the majority of specimens, and even when counts agreed, the spatial distribution of marks often did not correspond. The authors concluded that "much more research on the fundamental methodological side of skeletochronology—especially regarding the general nature and microscopic recognition of GM—is required."

Intraskeletal histovariability: Different skeletal elements from the same individual can yield different LAG counts and spacing patterns. Woodward, Horner, and Farlow (2014) documented this extensively in Alligator mississippiensis, and it is now standard practice to recommend sampling the same element (preferably the femur or tibia) across specimens for meaningful population-level comparisons.

Studies of early branching sauropodomorphs have revealed that LAGs do not always conform to the channelized, asymptotic growth pattern seen in many later-branching dinosaurs. Sander and Klein (2005) demonstrated that Plateosaurus trossingensis exhibited extreme growth plasticity: individuals of the same body size could differ substantially in LAG count and spacing, indicating that growth rate was highly responsive to environmental conditions. Chapelle et al. (2021) found a similar pattern in Massospondylus carinatus from the Lower Jurassic of South Africa, with high intraspecific variability in LAG number, spacing, and the occurrence of multiplet LAGs. These findings suggest that growth plasticity was a plesiomorphic condition in dinosaurs—potentially a preadaptation that allowed early dinosaurs to thrive in unstable post-extinction environments—while more deterministic growth trajectories evolved later in lineages such as ornithischians and coelurosaurs.

LAGs are not restricted to ectotherms or extinct animals. They are well documented in living amphibians, where phalangeal skeletochronology is a standard method for age determination in population ecology studies. Sea turtle skeletochronology, based on LAGs in humeral cross-sections, is employed by NOAA and other conservation agencies for demographic monitoring of endangered species. Crucially, Köhler et al. (2012) demonstrated in Nature that LAGs form annually in the long bones of extant ruminant mammals—endotherms living in seasonal environments. This finding overturned the longstanding assumption that LAGs are characteristic of ectothermic physiology, showing instead that they are a widespread feature of seasonal bone growth across Tetrapoda regardless of metabolic strategy. Consequently, the presence of LAGs in dinosaur bone cannot be used as evidence for ectothermy.

Because traditional thin-sectioning is destructive, researchers have explored non-destructive imaging alternatives. Propagation phase-contrast synchrotron radiation micro-computed tomography (PPC-SRμCT) has been used to visualize growth marks in three dimensions without physically sectioning the specimen. While this technique has shown promise in certain contexts—including embryonic dinosaur bone—it does not yet match the resolution and contrast of conventional thin sections for reliably identifying all types of growth marks, particularly subtle annuli and LAGs in heavily remodeled cortex. Standard medical CT and even high-resolution industrial micro-CT generally lack the contrast necessary to distinguish growth marks, though they can be useful for identifying the optimal sampling location (growth centre) prior to sectioning.

LAGs represent one of the few means by which direct temporal information can be extracted from the fossil record. Through LAG-based analyses, paleobiologists have established that non-avian dinosaurs grew at rates far exceeding those of any living non-avian reptile, providing strong evidence for elevated metabolic rates. LAG data have been used to construct the first life tables and survivorship curves for dinosaur species (Erickson et al., 2006, Science), estimate incubation periods by counting growth lines in embryonic teeth (Erickson et al., 2017, PNAS), and determine that basal birds such as Archaeopteryx grew at rates comparable to non-avian dinosaurs rather than modern birds (Erickson et al., 2009). Despite ongoing debates about the reliability of absolute age estimates, the integration of LAG analysis with other osteohistological observations—tissue type, vascularity, remodeling intensity, and presence or absence of an EFS—continues to provide an unparalleled window into the life histories of animals that have been extinct for tens or hundreds of millions of years.