How do we know...? (Part III: Inner Workings)

Last week, I covered the external anatomy of limbs, which is often one key datapoint used to infer the lifestyle of temnospondyls - after all, limbs change quite a lot depending on whether you need to walk around on land or merely paddle in the water. But external anatomy isn't the only useful thing about a fossilized limb. Thin-sectioning, or the process of slicing a fossil or bone and making a thin section out of it for examination under a microscope allows us to examine microscopic structures that offer insights into how that bone grew. This in turn can be extrapolated to how the entire animal grew and is a powerful source of data for testing hypotheses about lifestyle and ecology more broadly. In today's post, I'll be covering histology and microanatomy, the two main attributes of the internal bone that we can derive from fossils.

Histological section of the femur of the Triassic metoposaurid Panthasaurus (source: Teschner et al., 2020).

Histological sections of the humerus of the Triassic capitosaur Cherninia (source: Mukherjee et al., 2020).

Microanatomy

Broadly speaking, microanatomy refers to the structure and compactness of the internal bone, visible only in thin section. It differs from histology in not focusing on which tissues are doing what. You might alternatively think about it as "what does the cross section look like?" Is there a large opening in the center? Are the outer layers completely compact or are they perforated by numerous spaces? One of the advantages of bone microanatomy is that studying it doesn't require a thin section, only an exposed cross-section; this is one reason why the study of bone microanatomy in extinct animals goes back to the 19th century, well before modern thin-sectioning methods had been invented.

A few key terms

• Medullary cavity: this is a large hollow opening in bones where the bone marrow is contained. In limbs, it runs more or less the entire length of the shaft (between the ends). The medullary cavity can be "free" (entirely open, without bone) or it can be partially to completely trabecular. Trabecular refers to...
• Trabeculae: in generic term, a trabecula is just a partition between spaces. In the sense of bone, trabeculae are relatively thin tissue structures that often form a network (the trabecular network) to create a large series of pores within the medullary cavity; the bone marrow is then housed within these pores. This often creates a "spongy" texture that can be relatively dense (lots of trabeculae) to disperse (few trabeculae).
• Cortex: the cortex is usually defined relative to the medullary cavity, although the distinction between them can be blurred with remodeling and more developed trabecular networks. Essentially, this is the more compact region of bone (if a trabecular network is present) or the only region of bone (if none is present).
  

Scientists quickly realized that different types of microanatomy seemed to correlate with lifestyle in modern animals. Aquatic animals often had much a thicker and more compact cortex (like the platypus on top in the figure to the left), while terrestrial animals had more delicate cortices with large open medullary cavities (like the echidna on the bottom).

There are, of course, nuances to this, as well as intermediary or highly unusual conditions that look nothing like these two end-members. The general trend of increasing compactness in aquatic animals is often tied to buoyancy control; in order to avoid constantly bobbing up to the surface, increasing the amount of bone as ballast is one way to regulate this. In contrast, highly active swimmers like those found in the open ocean (rather than near a beach) or those that dive may actually exhibit the opposite trend - lightened bones that help them to move quickly and to be maneuverable; weigh too much and you'll never come back up... However, these lightened bones will look very different from a terrestrial animal, which maintains a relatively light bone mass by keeping the medullary cavity open (parts G-I in the figure on the right). Deep divers instead develop a think cortex but a fairly extensive trabecular network (parts A-C in the figure on the right).

Comparison of different limb bone microanatomies. A-C = southern elephant seal, a deep diver; D-F = the coypu (a large semi-aquatic rodent); G-I = the wolf (terrestrial). Source: Canoville et al. (2021).

---

So what do we know about the microanatomy of temnospondyl limbs? A fair bit! A large number of predominantly Permian and Triassic taxa have been sampled, which appear to encompass a broad range of species and ecologies. The taxa below are considered good examples of taxa with bone microanatomy that reflects a terrestrial lifestyle: an open medullary cavity and a thin to moderately thick cortex.

A humeral thin section of the Triassic lydekkerinid Lydekkerina (source: McHugh, 2014). At the diaphysis (mid-length of the shaft; A-D), the medullary cavity (MC) is open and large; the cortical bone (CB) is relatively thin. Conversely, a section taken at the metaphysis (one end of the humerus; E) has more trabeculae in the medullary cavity.

A humeral thin section of the Triassic amphibamiform Micropholis (source: McHugh, 2014). This specimen also has an open medullary cavity (MC) but with two large trabeculae cutting through the middle. The cortex (cortical bone = CB) is relatively thick and compact.

A femoral thin section of the Permian amphibamid Doleserpeton (source: B. Gee). This specimen has an open medullary cavity (no trabeculae) with a compact and relatively thick cortex.

An intimate moment shared between two individuals of Micropholis (credit: Gabriel Ugueto)

Terrestrially inclined

Conversely, below are examples of temnospondyls with microanatomy that indicates a more aquatic lifestyle: medullary cavities with a dense trabecular network throughout and a relatively thick cortext.

Femoral thin section of the Triassic metoposaurid Dutuitosaurus (source: Steyer et al., 2004).

Femoral thin section of the Permian dvinosaur Dvinosaurus (source: Uliakhin et al., 2020). mc = medullary cavity; cx = cortext.

Aquatically inclined

It's important to note that the distinctions are not binary, and there are some taxa that are often disputed, both on account of a mix of differing signals from the microanatomy and from other anatomical features on the skeleton that conflict with a histological interpretation. For example, the two figures below (from Canoville & Chinsamy, 2015) are from Lydekkerina. I showed some sections of this taxon from another study above, in the section displaying some terrestrially inclined temnospondyls. Well, these sections do show some features associated with terrestriality like a medullary cavity free of trabeculae, but the cortex of these bones is very thick, leading to a very small medullary cavity. This cortext is what we expect for an aquatic animal. Canoville & Chinsamy favor a terrestrial lifestyle based on both external features (e.g., a prominent adductor crest, something I talked about last week) and histology (absence of calcified cartilage, something I'll talk about more below), but suggested that it preferred to live in shallow ponds and had to retain the ability to move about well on land in order to find a new pond when the last one dried up. Aggregation in ponds might explain why Lydekkerina occurs so frequently in dense, multi-individual deposits that could represent the last of the ponds.

Femoral (A-G) and rib (H) thin sections of Lydekkerina.

Radial (A-D), radial/ulnar (E), and rib (F-H) thin sections of Lydekkerina.

A number of different studies have looked at whether it's possible to quantify the different microanatomical parameters of a thin section to more precisely estimate an extinct animal's lifestyle. This eliminates some of the subjectivity associated with determining the relative size, thickness, or compactness of a bone. For example, Quemeneur et al. (2013) compiled microanatomical data from the femur 7 species of mammals, 15 species of turtles, 56 species of lepidosaurs, and 27 species of birds, all of which can have their microanatomy correlated with their actual lifestyle because they are modern species. They measured compactness by dividing a thin section into many (many) polygons, as seen in the figure to the right, and then used these to assess compactness. This was in turn used to estimate the lifestyle of extinct temnospondyls and models the typical approach of using extant analogues or homologues to create a comparative reference foundation for inference in extinct taxa. This working group has actually produced an extensive number of papers on this topic of microanatomy and its relationship to lifestyle on different bones, such as the humerus (e.g., Canoville & Laurin, 2009, 2010), the tibia (e.g., Kriloff et al., 2008), the radius (e.g., Germain & Laurin, 2005), and ribs (e.g., Canoville et al., 2016).

A schematic thin section showing how the software BONE PROFILER divides the section into numerous polygons for assessment of compactness at small and large scales (source: Quemeneur et al., 2013).

Another method, by Sanchez et al. (2010), measured several parameters of a thin section, such as the circularity of the entire thin section and of the vascular canals exposed in section, and then used a discriminant analysis to assess the relative degree of terrestriality. Taxa in the top right are inferred to be more terrestrial, and taxa in the bottom left are inferred to be more aquatic.

In general, these quantitative methods align with previous concepts of temnospondyl lifestyle based on external anatomical features. Where they tend to struggle is with taxa like Lydekkerina that might have been more amphibious or that possibly underwent some kind of habitat shift during development (not necessarily metamorphosis). This was one idea put forward for the iconic temnospondyl Eryops from the Permian of Texas, which has long been debated with respect to its lifestyle. While microanatomy indicates a more aquatic animal, evidenced by extensive trabeculae in the medullary cavity, the external anatomy (e.g., well-ossified ends, prominent adductor crest and trochanter) are hallmarks of terrestriality.

Femoral thin section of Eryops, a Permian eryopid (source: Konietzko-Meier et al., 2016).

Reconstruction of the femur of Eryops (rotated 90 degrees clockwise; source: Pawley & Warren, 2006).

A somewhat amphibious Eryops enjoying a soak in the local pond (credit: Vladislav Egorov).

An Eryops retreating to the water after being attacked by the synapsid Dimetrodon (credit: Alain Beneteau).

A more aquatically inclined Eryops hunting a shark (credit: Doug Henderson).

Sizing down. Both small temnospondyls and small modern amphibians (most of them are small) are well-known for having unusually compact cortices. There are different names for various porosities, with compact (volume of bone > 50% of pore space) and cancellous (< 50% of pore space) being the most widely applied dichotomy. Cancellous is often divided into several subcategories, ranging from fine cancellous (least porous) to trabecular (most porous). Pores are formed by a number of different internal features, most of which relate to blood vessel distribution, secondary remodeling, or bone marrow distribution in the case of bone distributed within the medullary cavity.

Various thin sections of the temnospondyl Apateon showing cortices lacking primary or secondary osteons (source: Sanchez et al. 2010).

Various thin sections of the dinosaur Hypacrosaurus showing different degrees of remodeling. A, C, and F show densely packed secondary osteons, a feature of remodeling, in the cortex (source: Padian et al., 2016).

The unusually compact cortices of small-bodied amphibians is typically associated with their miniaturization, or prominent decrease in size from a larger ancestor (e.g., Skutchas & Stein, 2015). They thus have relatively few primary osteons (blood vasculature) and often lack secondary osteons altogether (remodeling), like in Apateon on the upper left, whereas these are common, pervasive features in larger and more metabolically active animals like the dinosaur thin sections to the upper right. The same pattern generally holds true for mammals; very small mammals tend to lack secondary osteons, and the density of these osteons increases with increased body size (e.g., Felder et al., 2017).

Histology

Histology can encompass microanatomy, and a lot of people use it in this way. Alternatively, it can be restricted to the specific tissues that make up different parts of the bone (or whether there are different tissues at all), which often requires more high-quality thin sections than are needed for assessing strictly microanatomy in which case a black-and-white silhouette or even just a broken surface of a bone can be sufficient. This is also where growth marks used to determine the relative age of an animal can come into play. But what can histology tell us about an animal's lifestyle?

Femoral thin section of the Permian stereospondylomorph Platyoposaurus (source: Uliakhin et al., 2021).

As it turns out, not that necessarily as much as we might expect. Histology can tell us a lot about a temnospondyl's general ecology, such as whether it grew fast or slow and for how long. It can tell us whether the animal experienced pronounced stress, and if so, how often and how it responded (e.g., did it entire cease growth or merely slow down, and did it do this once or twice per year). It can tell us whether the animal was mature or not. But most of this information is restricted to attributes other than whether an animal was aquatic or terrestrial. How old an animal could be doesn't necessarily tell us whether it was aquatic or not. The main study to use histological parameters to explicitly test lifestyle is Sanchez et al. (2010), which I discussed above; some of their parameters are more histological than microanatomical, such as the orientation of collagen fibers and the relative abundance of remodeling.

Bone remodeling is a common phenomenon that occurs throughout an animal's life (including ours). Typically it leaves a distinct visual signature, reflecting the different nature of tissue formation, along the margin of the medullary cavity. In the image below on the left (of the Permian amphibamid Doleserpeton), the remodeling presents as what appears to be a different set of layers around the medullary cavity. The relative rate of secondarily deposited bone to that of primary bone resorption (basically the creation-destruction ratio) can lead to increased bone weight (when secondary deposition > resorption) or decreased weight (when deposition < resorption). High rates of secondary deposition are one means of adding ballast, an important physiological attribute for bottom-dwelling and shallow-water-residing animals that are trying to avoid bobbing up to the surface. The remodeling seen in Doleserpeton is typical of terrestrial taxa in which the medullary cavity remains open.

Remodeling can also present in the form of what we call 'erosion bays,' which are large openings formed by erosion of vascular canals that resorbs the divisions between them. These are often found at the margin of the medullary cavity and are identified by the presence of secondary bone deposition along the margin of the bay (part E in the figure above on the right from Sanchez & Schoch, 2013; figure is of Gerrothorax, a Triassic plagiosaurid thought to be very aquatic).

Some temnospondyls with extensive secondary remodeling and other weight-adding features, like Gerrothorax, are thought to have been bottom-dwelling ambush predators (credit: Alpha Nix).

A second feature that also relates to buoyancy control is called "calcified cartilage." This is apparently a pathological condition in humans, but in temnospondyls, it appears to have been an intermediary step between uncalcified cartilage and bone. Like secondarily deposited bone, calcified cartilage is denser than primary bone, so it would also act as ballast to help animals stay down underwater. Calcified cartilage is kind of hard to describe, but it has often appeared as a granular kind of texture and often occurs at the nodes between trabeculae. However, it can be quite variable throughout ontogeny and in different parts of the skeleton, and it seems to show up most often in the vertebrae, as in the figure to the left (source: Konietzko-Meier et al., 2014). The arrow in part E indicates some of this cartilage, and the tissue in part F is essentially all calcified cartilage. When temnospondyls have calcified cartilage, they are either really immature or almost certainly aquatic.

Can histology tell us about metamorphosis? This is a common question that I get asked. Significant life events often exhibit histological markers, such as hatching lines in reptiles (e.g., Hugi & Sánchez-Villagra, 2012), and birth and weaning lines in mammals (e.g., Castanet et al., 2004; Nacarino-Meneses & Köhler, 2018). Modern amphibian workers have sometimes identified the first growth mark as the line of metamorphosis (e.g., Hemelaar, 1985; Sinsch, 2015). However, unlike many  non-cyclical marks in other animals, the purported line of metamorphosis looks no different than any other subsequent growth mark. It is essentially identified only by virtue of being the earliest documented mark. We only know that this could represent metamorphosis because we have observational data showing that these species underwent metamorphosis, which is lacking for fossil taxa and thus hard to corroborate.

Examples of histology sections of the phalanx (finger/toe bones) in modern frogs. The line of metamorphosis (LM) is the growth mark closest to the medullary cavity (MC), with subsequent growth marks (LAG1, LAG2, etc.) further away. Endosteal bone (EB) represents secondary remodeling (source: Sinsch, 2015).

---

Summary

Thin sections of temnospondyl limb bones are informative for inferring their lifestyle. Part of this is related to the fact that limb bones are the most frequently sectioned elements for either extinct or extant vertebrates, and as a result, there are a lot of comparative data and a lot of data correlated directly with observations of an animal's lifestyle (like watching a diving seal). But really what most people are after for making the most robust and informed inferences is the microanatomy (the configuration of tissue), not the histology (the nature of the tissue). Aspects like whether the cortex is thin or thick and how compact it is, as well as whether the medullary cavity is infilled or not, have proven to not only be informative and readily assessed, but to also correlate fairly well with other datapoints like external anatomical features.

Up next week: the shoulder!