Osteohistology and growth dynamics of the Brazilian noasaurid Vespersaurus paranaensis Langer et al., 2019 (Theropoda: Abelisauroidea)

Femoral histological class I (CP.V 4136)

The complete femur CP.V 4136 is the smallest specimen of the sample. It is 182 mm in length, 16 mm in antero-posterior diameter and 45.86 mm in circumference (Table 1). We sectioned the diaphysis of this specimen at two points of the midshaft (Fig. 2D). In these sampled locations, the microstructure and the growth record did not vary along the diaphysis. It is the only specimen to have two annuli rather than any LAG (Fig. 3C; Table 2), suggesting growth with a periodic reduction in the rate of bone deposition, but without interruption. The deposition of annuli rather than well marked LAGs can occur in very young individuals (Woodward, Padian & Lee, 2013). Three cortical zones are delimited between the two annuli. The innermost zone is the thickest on average (1.37 mm), showing an initial pulse of proportionally faster growth, followed by two zones (0.72 and 0.45 mm) that decrease slightly in width outward (Table 2). The medullary cavity occupies 44.29% of the cross-sectional area, and a very tenuous endosteal lamella is visible (0.04 mm in thickness). Resorption cavities were not observed in CP.V 4136. Secondary osteons were scarce in CP.V 4136 and limited to a small area in the medial-posterior deep cortex. The cortex exhibits a high birefringence under polarized light (Fig. 3D). Osteocyte lacunae are flattened, preferentially-oriented, and arranged parallel to one another, forming a parallel-fibered matrix. Within the annuli, the osteocyte lacunae are even more flattened and preferentially-oriented. The cortex presents reticular and longitudinal channels, the later often consisting of primary osteons. Primary osteons are highly abundant in the deep cortex. The degree of vascularization (size and density of the channels) is homogeneous throughout the entire cross-section. In the subperiosteal region of this specimen, some channels are not encircled by bone matrix, such that they are open to the outside (Fig. 3B). An age retro-calculation plus the growth marks number allows us to estimate the individual should be around its fifth year of life at the time of death (Table 3). The minimal remodeling of the deep cortex, a tenuous endosteal lamella, plus the annuli number and age estimation, support a juvenile status for CP.V 4136.

Femoral histological class II (CP.V 4142 and CP.V 4206)

The femora CP.V 4142 and CP.V 4206 are nearly equal in size, varying between 16 and 18 mm in antero-posterior diameter and 47.90–50.11 mm in diaphyseal circumference (Table 1), respectively. CP.V 4206 is severely taphonomically damaged and partially embedded in the matrix, with most of the shaft preserved as impression. The specimen conserved the proximal portion of the diaphysis (the region between the insertion of the M. iliofemoralis externus and M. caudofemoralis longus fossa) and the distal end (most of the fibular condyle). CP.V 4142 represents a proximal half of the femur shaft, with a small degree of compression. The femora CP.V 4142 and CP.V 4206 exhibit two marked LAGs, typifying FHC II (Figs. 4D and 4E). The cortical zones of the CP.V 4206 and CP.V 4142 are nearly equal in thickness, and the zone thickness decreases toward the subperiosteal cortex in both specimens (Table 2). Although CP.V 4136 (FHC I) described above also exhibits two cyclical growth marks (annuli), we prefer to discriminate CP.V 4206 and CP.V 4142 in a different class for the following reasons. In CP.V 4142 and CP.V 4206, the medullary cavities have 17.94 and 24.94 mm of circumference, which corresponds to 46.64% and 50.36% of the cross-sectional area, respectively (Table 1), a little larger than in FHC I. The medullary cavities in CP.V 4142 and CP.V 4206 are covered by thicker endosteal lamellae (0.07 mm) than seen in FHC I (0.04 mm). CP.V 4206 exhibits almost all cortical area obliterated by secondary osteons, especially in the medial-posterior zone of the cross-section, differing from CP.V 4136 (FHC I), which exhibits little secondary remodeling, at least in the sampled section. Curiously, CP.V 4142 also lacks secondary osteons, which suggests variation of secondary remodeling even in femora displaying similar size. Longitudinal and primary osteons are rare, and all vascular channels are confined within the cortex in FHC II, differing from FHC I, which exhibits a high amount of primary osteons and open vascular channels in the subperiosteal cortex. As in FHC I, the osteocyte lacunae are flattened, preferentially-oriented, and parallel one to another, forming a parallel-fibered matrix. FHC II specimens have plexiform channels in the innermost cortex, graduating to a reticular arrangement toward the outside (Fig. 4D). Age retro-calculation of CP.V 4142 allow us to estimate that the members of FHC II would be in their fifth year of life at the time of death (Table 3). Despite the same age inferred for FHC I, the set of histological features of CP.V 4142 and CP.V 4206 suggests a more mature ontogenetic status for FHC II.

Femoral histological class III (CP.V 4264)

The femur CP.V 4264 preserves about 80% of the shaft length (Fig. 2C). The head was not preserved, but the distal epiphysis is nearly complete. The diaphysis was severely damaged, and the anterior side of the bone was partially eroded. CP.V 4264 has an antero-posterior diameter of 14 mm and a circumference of about 58.37 mm (Table 1). Because CP.V 4264 preserved most of its diaphysis, it was sectioned in two points, at the midshaft and the distal half of the shaft (Fig. 2C). The femur CP.V 4264 is microstructurally similar to that of the FHC II, but exhibits three LAGs (Fig. 4F) instead of two.

The LAG number remains unaltered along the two sampled bone locations, and the spacing between the zones does not vary in thickness along the shaft (Table 2). The outermost LAG (3rd) is close to the outer cortex limit—0.07 mm (distal half) and 0.14 mm (mid shaft)—and seems to have been deposited near the time of the individual’s death. The medullary cavity corresponds to 55.63% of the cross-sectional area in FHC III and presents a surrounding endosteal lamella similar in thickness to that of FHC II (about 0.04 mm). Several resorption cavities are present in the perimedullary cortex of CP.V 4264, and they are significantly more abundant in the distal half of the bone shaft. The cortex consists of a parallel-fibered matrix with flattened and preferentially-oriented osteocyte lacunae. Most vascular channels are reticular, but the vascularization decreases toward the outer cortex, and simple longitudinal vascular channels become abundant at the subperiosteal zone. Primary osteons are common and more frequent in the two innermost cortical zones than the outermost one. The greater number of growth marks (three) indicates CP.V 4264 belongs to a more advanced ontogenetic stage than FHC I and II. The lesser vascularization plus the zones narrowing after the second preserved LAG suggests that CP.V 4264 was experiencing a decrease in growth typical of sexual maturity around the seventh year of life (age estimated with LAG number plus retro-calculation models) (Table 3).

Femoral histological class IV (CP.V 2277)

The femur CP.V 2277 is incomplete, and the section was taken from one quarter of the proximo-distal length of the diaphysis (Fig. 2B). The specimen exhibits 63.27 mm of circumference (Table 1) and 18 mm of antero-posterior diameter. Although distant from the midshaft, the CP.V 2277 thin-section does not exhibit the cancellous bone and discontinuous LAGs typically found in the metaphyseal zone, suggesting that the section was taken in the proximal diaphysis. Despite its incompleteness, we included the femur CP.V 2277 in the sample since it was the largest femur of Vespersaurus found so far (Fig. 2B). In this femur, the zones of parallel-fibered bone are separated by single, double, or triple LAGs, totaling eight zones and seven packs of LAGs (assumed as bi-and tri-annual growth marks) (Fig. 5D). The innermost zone ends with a pack of double LAGs. The second zone ends with three closely spaced LAGs, forming triple LAGs. The third and fourth zones end with double LAGs each. Three single LAGs separate the last outermost zones. The zones between LAGs decrease gradually in thickness centrifugally, especially after the third zone (third pack of LAGs; Table 2). The medullary cavity has 25.9 mm of circumference and occupies 41.50% of the cross-sectional area. Curiously, endosteal lamella were not observed in CP.V 2277, which could indicate that the medullary cavity was under endosteal resorption at the time of death (Chinsamy-Turan, 2005). The secondary osteons are restricted to the deep cortex, mainly in the anteromedial, anterior and anterolateral edges of the cross-section, just near to the ridge of M. caudofemoralis longus fossa. The outer cortex is rarely occupied by primary osteons in CP.V 2277. The vascular channels are entirely reticular-type but differ both in size and number between the zones (Figs. 5B–5D). The three innermost inter-LAG zones of CP.V 2277 are nearly equal in vascularization. The fourth zone presents reticular channels like the innermost ones, but in a smaller amount. Curiously, the fifth zone has a higher concentration of channels, but they consist of smaller vascular channels. The subsequent outermost zones comprise thinner and smaller reticular channels, showing a decrease in vascularization. The subperiosteal cortex is surrounded by an avascular tissue of the lamellar bone matrix, forming a lamellar-zonal bone (Fig. 5D). However, it is important to emphasize that CP.V 2277 (FHC IV) was sectioned near the ridge where the M. caudofemoralis longus fossa is located, broadening the cortical area. This femur might be the ontogenetically oldest specimen of the sample based on the highest LAGs number, extensive secondary remodeling, and size. Nevertheless, the absence of an EFS evidences that the bone had not reached full size. Taking into account that reduced growth rate is linked to the achievement of the phase of sexual maturity in dinosaurs (Sander, 2000; Lee & Werning, 2008), and that in CP.V 2277 this is more evidenced after the third growth pulse, we can infer that, based on femora, Vespersaurus could reach sexual maturity between the seventh (based on femur CP.V 4264) and tenth (based on CP.V 2277) years of life.

Tibial histological class I (CP.V 4203, CP.V 4312a)

CP.V 4312a comprises a fragment of the proximal end with a circumference of 42 mm (Table 1), whereas CP.V 4203 is a highly fragmented and broken proximal diaphysis that prevents circumferential measurement. CP.V 4312a and CP.V 4203 exhibit the most immature growth record among the sampled tibiae, with three LAGs each. However, caution is necessary in this matter because the growth record could be biased by the proximity of the metaphysis in CP.V 4312a (Fig. 2H). In both tibiae, the innermost zones slightly decrease in thickness toward the subperiosteal cortex, but the outermost zones are exceptionally thicker than the penultimate ones (Table 2). The bones contain relatively thin compacta surrounding the medullary cavities in THC I. Medullary cavity exhibits 21 mm of circumference, occupying about 42% of the cross-sectional area in CP.V 4312a. Due to the fragmentary state of CP.V 4306, its medullary cavity circumference was not measured. The medullary cavities are partially covered by what we interpret as possible endosteal lamellae in CP.V 4306 and CP.V 4312a (about 0.10 mm in thickness), which consists of slightly thicker layer compared to the femora of FHC I and II (0.07 mm). The deep cortex is filled extensively by secondary osteons in CP.V 4312a, whereas CP.V 4306 exhibits fewer secondary osteons. Isolated secondary osteons within the primary bone in the outer cortex are common. Both specimens exhibit high birefringence under polarized light (Fig. 5D), and preferentially oriented and flattened osteocyte lacunae characterizing a parallel-fibered bone tissue. Primary osteons are abundant in both specimens of THC I with the deep cortices showing primary and secondary osteons side by side. Plexiform vascular channels occur in the deep cortex of both THC I specimen, but in CP.V 4306 they are replaced by a less dense and reticular vascular pattern toward the subperiosteal zone suggesting that this tibia reduced vascularity after the third LAG deposition. CP.V 4312a exhibits plexiform type occupying the outermost zone with any perceptible transition due to the high amount of secondary osteons in the cortical bone. Age estimation was not applied for the specimens belonging to THC I because the zones width required for retro-calculation are obscured by the remodeling, which underestimates the results (CP.V 4312a), or because the transversal section is highly incomplete for retro-calculation (CP.V4203). Although obtaining retro-calculated age is precluded, it is possible to observe that the zone widths and degree of vascularization slightly decrease outward in the cortex after the first zone in CP.V 4203 (Fig. 6C) (not easily recognized in CP.V 4312a), which suggests that the rate of bone deposition was slowing at the time of death. However, the supposed status of sexually mature individuals for this class remains in doubt.

Tibial histological class II (CP.V 4000, CP.V 4140)

The specimen CP.V 4000 comprises a crushed fragment of the tibial diaphysis (the circumference cannot be measured). CP.V 4140 consists of the distal half of a left tibia with a circumference of 41.29 mm (Table 1) and an antero-posterior diameter of 14 mm. Tibiae CP.V 4000 and CP.V4140 exhibit four LAGs each and they typifying the THC II. A tentative age of seven years old was inferred for CP. V 4140, but no estimation was possible for CP. V 4000 due to the incompleteness of the cross-section. A gradual decrease outward in the zones widths is observable in CP 4140 after the first LAG deposition, around the 4th year of life (Table 2), suggesting that the THC II individuals were gradually reducing their growth rate (Figs. 7C and 7G). In CP.V 4140, the medullary cavity has 12.42 mm of circumference, which corresponds to 27% of the whole cross-section (not calculated in CP.V 4000). The medullary cavity is covered by a thick layer of endosteal lamellae in these tibiae (about 0.30 mm in thickness) in both specimens, but in CP.V 4140, the endosteal lamella is expanded in the medial and lateral boundaries. In these regions (more evident on the lateral side), endosteal lamellae are filled by radial vascular channels. In the posteromedial perimedullary zone of CP.V4140, there are two layers of lamellar-zonal bone (lamellae) alternating with parallel-fibered bone (Fig. 7D). The endosteal lamellae are followed by a relatively thick cortex covering the whole medullary cavity in CP.V 4140 and CP.V 4000. Only in CP.V 4140, the anterior and posterior regions of the deep cortex exhibit trabecular bone with primary and secondary osteons (Fig. 7F). The deep cortices are extensively filled by secondary osteons, which occupy between 2/3 and 5/6 of the compacta in CP.V 4140, and between 1/2 and 2/3 in CP.V 4000, indicating intense remodeling (Fig. 7E). Osteocyte lacunae are flattened and parallel in the compacta of THC II specimens, suggesting some degree of collagen fiber orientation, which typifies the parallel-fibered matrix. For all specimens, the high birefringence under polarized light reinforces the predominance of parallel-fibered bone in the compacta. Primary osteons rarely occur in THC II, and reticular vascular channels are predominant. In CP.V 4000, there is a clear decrease in density and size of the channels toward the subperiosteal cortex, whereas in CP.V 4140 this is not evident due to the high cortical remodeling. This feature plus the narrowing of the zones toward the outermost cortex suggests the individuals of THC II were closer to attaining sexual maturity than those of THC I, but they had not reached full size at the time of death.

Tibial histological class III (CP.V 2385, CP.V 4130)

CP.V 4130 consists of atibia without the distal end, permitting sectioning at two points along the shaft (Fig. 2G). It has an antero-posterior diameter of 14 mm and a circumference of 41.18 mm (Table 1). CP.V 2385 represents a fragment of the distal end of a tibia with antero-posterior diameter of 15 mm and circumference of 44.88 mm (Table 1). This specimen does not present any sign of fusion between astragalus, calcaneus, or fibula (Fig. 2I). CP.V 2385 and CP.V 4130 both have five LAGs and are assigned to THC III. Only the members of this class have a subperiosteal cortex containing an EFS, which indicates that these individuals had completed their linear growth, at least in the tibiae (Figs. 8H and 8I). The EFS covers the entire cortex in transversal-section in both CP.V 2385 and CP.V 4130. There are four LAGs within the EFS in CP.V 2385 (totaling nine LAGs) and three in CP.V 4130 (totaling eight LAGs). Considering the retro-calculated age, we estimated 13 year of life for CP.V 2385 and 10 for CP.V 4130. Zones between LAGs decrease gradually in width toward the subperiosteal cortex in CP.V 4130 (Table 2). As described for the other cortical zones, THC III specimens exhibit a high-birefringence under polarized light (Fig. 8G). The osteocyte lacunae within the EFS are slightly more flattened than in the rest of the cortex. The EFS osteocyte lacunae are embedded within an avascular tissue of parallel-fibered matrix. The medullary cavities are surrounded by a thick compacta and are covered by thicker endosteal lamellae (about 0.40 mm in thickness) in THC III compared to the specimens of THC I and II. This could indicate the cessation of medullary cavity expansion.

The medullary cavity has 13.24 mm of circumference in CP.V 4130 (30.78% of the cross- sectional area) and 13.50 mm of circumference in CP.V 2385 (31.78% of the cross-sectional area). The deep cortex of CP.V 2385 exhibits a higher amount of secondary osteons than CP.V 4130. Primary bone consists of flattened and parallel tosteocyte lacunae. A small amount of a woven fibered matrix occurs in packs within the parallel-fibered matrix in the deep cortex, close to the secondary osteons concentration in CP.V 2385 and CP.V4130 (Fig. 8G). Along the lateral edge, the subperiosteal cortices are poorly vascularized and exhibit a layer of avascular lamellar-zonal bone. Primary osteons and simple longitudinal channels are common in the cortex of CP.V 2385, but rare in CP.V 4130. In the latter specimen, well-delimited zones of reticular and plexiform channels occur separately within the cortex. The density and size of vascular channels decrease toward the subperiosteal cortex in CP.V 2385 and CP.V 4130.

The decrease in vascularization and zone thickness outwards in the subperiosteal cortex after the first preserved LAG suggests that the specimens reached sexual maturity around three to five years of age, but they were able to grow for another four years until reaching somatic maturity, marked by the deposition of the EFS. Moreover, these specimens lived long enough to deposit three and four LAGs after entering the somatic maturity phase.