How do we know...? (Part IV: Chip on your shoulder)

The limb gets a lot of attention because it's the part of the body in direct contact with the ground (or in the case of some animals, avoiding contact with the ground). But equally important is the apparatus that connects to the limbs - the girdles! What we call the shoulder girdle is scientifically termed the 'pectoral girdle,' while what we call the hip or the pelvis is termed the 'pelvic girdle.' This week's post will look at the pectoral girdle!

Temnospondyls have four to five different elements in their pectoral girdle: the interclavicle, the clavicle, the cleithrum, and the scapula + the coracoid / a composite scapulocoracoid. Stem tetrapods (like Acanthostega in the middle below) have an anocleithrum inherited from our fishy forerunners, but this is absent in all temnospondyls. Conversely, some extra bones are found in modern amphibians (like Salamandra on the right below). Several of these will sound unfamiliar to people who only know mammals because we (and other mammals) lack cleithra and interclavicles.

Comparison of the pectoral girdle and upper forelimb in (A) the living lungfish a representative extant sarcopterygian fish Neoceratodus; (B) the stem tetrapod Acanthostega; and the living salamander Salamandra (source: Molnar et al., 2017).

To help contextualize things, the diagram to the left shows the progressive evolution of the sternum in mammals. The clavicle of early tetrapods becomes thin and long and is now what we call the "collarbone." The interclavicle is found in monotremes like the platypus and the echidna but is lost in therians (marsupials + placentals like humans). The cleithrum has long been lost at a much earlier point in the evolution of amniotes. However, mammals acquire another set of sternal bones, sternebrae (there are several) and the manubrium, which are unpaired elements at the center of the chest that replace the interclavicle in position. These are the bones right at the center of your chest, but they are not found in early tetrapods.

The diagram below shows an expanded version from the base of Synapsida (the mammal lineage), where the cleithra were already lost, but the clavicles and the interclavicle were still around like in  other early tetrapods (source: Buchholtz et al., 2020)

---

Clavicle

The clavicle is one pectoral element still found in humans, but it looks nothing like what we see in temnospondyls. In humans, our collarbone is a long and slender bone with a slight curvature. The equivalent in temnospondyls serves the same function, connecting the shoulder (scapula) to the sternum/central part of the chest.

There is no easy way to describe the temnospondyl clavicle, but sort of like a cross between a soup ladle and a spatula isn't bad, by which I mean that there's a dorsal process sticking up that forms an obtuse angle with a small, flat semi-circular ventral portion that underlies the interclavicle. The figure to the right shows the clavicle of Lydekkerina, a Triassic temnospondyl. Part K shows the clavicle from below (note that it can be ornamented like the skull roof). Parts M and N show the clavicle from both side views (source: Pawley & Warren, 2005). Hopefully you can see what I mean by a soup ladle / spatula cross...

A comparison of clavicles in lateral view. (A) the colosteid (non-temnospondyl) Greerpeton; (B) the eryopid Eryops; (C) the dvinosaur Dvinosaurus; (D) the stereospondylomorph Sclerocephalus; (E) the trematosaur; Benthosuchus sushkini; (F) the capitosaur Paracyclotosaurus davidi; (G) the trematosaur Lyrocephaliscus euri; (H) the metoposaurid Anaschisma browni; (I) the chigutisaurid Siderops kehli. The arrow points to an anterior thickening on the dorsal process that occurs in most stereospondyls, which are thought to be aquatic (source: Warren & Marsicano, 2000).

The main differences between temnospondyls are in the "handle" and the "scoop" of this ladle metaphor. The ventral "scoop" can be massively expanded (both width and length), which always corresponds with an increased size of the interclavicle and increased ornamentation on both elements. Usually, these taxa often have a more robust dorsal "handle" as well, which can sometimes have a long blade-like structure connecting it to the base (like in F, G, and H, in the figure to the left). The blade-like development tends to correlate with size and is only found in Triassic temnospondyls. These larger and more robust clavicles tend to be associated with aquatically inclined temnospondyls. Terrestrial temnospondyls tend to have smaller clavicles with more slender parts that make it appear more like a melon ball scooper than a soup ladle + spatula cross. There are a few instances where aquatic temnospondyls have a lateral line groove cutting across the clavicle (like on the skull or lower jaw), but this is less common and absence of such a groove does not imply terrestriality.

Interclavicle (2.1) and clavicle (2.2–2.6) of the Permian dvinosaur Trimerorhachis insignis (source: Pawley, 2007).

Aquatically inclined

Clavicle of the Permian eryopid Eryops megacephalus (source: Pawley & Warren, 2006).

Terrestrially inclined

---

Interclavicle

The interclavicle (which I'm told sometimes sounds like 'inner clavicle' when I don't enunciate) is the only unpaired element of the temnospondyl pectoral girdle. This is typically one of the larger elements and sits right under the chest in the position that we would call the sternum in humans. The interclavicle isn't super variable in shape, usually being diamond to rhomboidal, but it can be a little variable in whether there is a long process sticking forward or backwards. It is essentially entirely flat and plate-like. Because the interclavicle directly articulates with the clavicles, as seen below, any diagnostic features that can inform on the lifestyle of these animals are also likely found in the clavicle. These include the relative size and degree of ornamentation. While we often associate robust bones with terrestriality, too much weight can be bad on land, which is why terrestrial taxa tend to have proportionately smaller and lighter clavicles and interclavicles (the weight gets loaded on the scapula and the limb). Conversely, the massive (and often quite thick) elements of aquatically inclined taxa may have served as ballast, similar to how their internal bone structure is more dense (see last week's post).

Diagram showing the articulation between the clavicles and the interclavicles in metoposaurids (source: Sulej, 2002).

Different interclavicle morphologies plotted on a temnospondyl phylogeny (source: Witzmann & Voigt, 2015). The shorter and squatter interclavicles are more typical of terrestrially inclined taxa, while the longer, more diamond-like ones are more typical of aquatically inclined taxa.

I like to note just how big the interclavicles in particular can get. In large, aquatic taxa, they can be as long as the skull (exceeding 60 cm), fairly thick (>6 cm), and heavy (complete specimens can easily top 25 lbs). It's the large, dark grey element at the center of the chest in the reconstructions of Mastodonsaurus giganteus (Schoch, 1999) on the right.

Aquatically inclined

Terrestrially inclined

---

Cleithrum

The cleithrum is a weird element that often looks like it's a broken piece of something else (or that actually is broken). It can often present as a simple rod with slight curvature, like in Mastodonsaurus on the right (source: Schoch, 1999). Perhaps as a result of being hard to identify, even when complete, the cleithrum tends to be underrepresented in the fossil record of most temnospondyls. It has undergone substantial modification over the course of tetrapod evolution, from a bone that often co-ossified with the scapula in lobe-finned fish and the earliest tetrapods (A–G below) to a free element overlying the scapula in later tetrapods like temnospondyls (H–J below) to being entirely lost in most amniotes today (the cleithrum = 'cle' below; source: Molnar et al., 2017).

When you find a complete cleithrum, it is really easy to tell whether it belongs to a probable terrestrial temnospondyl or not. If it's a simple rod without substantial expansion, like Mastodonsaurus above, it likely belongs to an aquatic temnospondyl. Conversely, when the dorsal end is greatly expanded into a round plate-like process, like in Eryops below on the left  (source: Pawley & Warren, 2006), or greatly expanded to form an L-shaped element that overlaps the dorsal margin of the scapula, like in an indeterminate dissorophid below on the right (source: Gee & Reisz, 2018), you probably have a terrestrial taxon. Whereas larger tends to correlate with aquatic lifestyle in the clavicle and the interclavicle, the cleithrum serves as both a support strut for the scapula and for muscle attachment, so a larger cleithrum is more important for moving around on land under gravity.

---

Scapula + coracoid / composite scapulocoracoid

The scapula is what we term the "shoulder blade" in humans. Other tetrapods often have a corresponding element that sits below the scapula called the coracoid, but this is absent in placental mammals like humans. In temnospondyls, adults either have a compound element formed by the co-ossification of the scapula and the coracoid, aptly termed the 'scapulocoracoid,' or merely a scapula with an unfossilized, cartilaginous coracoid.

In the figure on the left, paedomorphic taxa like Trimerorhachis  (Part 1) and Dvinosaurus (Part 2) lack a coracoid, and the scapula appears as a short element, about as wide as it is tall (source: Pawley, 2006). This is interpreted as evidence of paedomorphosis where the development is truncated before the point at which the coracoid would ossify (this is a late stage ossification) and reflective of an aquatic lifestyle. Part of this is that the coracoid typically frames much of what we call the glenoid, which is the socket for the humerus. So if it fails to ossify, it part of that socket is cartilaginous. Conversely, when the coracoid ossifies and then co-ossifies with the scapula, it forms a tall, robust element with a fully ossified glenoid, like in Eryops (Part 4). With that said, the presence of a scapulocoracoid does not necessarily indicate that a taxon was fully terrestrial. It's probably better to frame these taxa as "terrestrially capable" and to look to other features for clarity.

For reference, the temnospondyl anatomy is typical of early tetrapods but not of later tetrapods, and it's quite different in modern amphibians. In frogs, the scapula and coracoid are both ossified, like you can see in the early frog Triadobatrachus and the modern frog Discoglossus below (source: Havelková & Roček, 2006). Modern salamanders typically have distinct but contacting scapulae and coracoids. Caecilians of course have no limbs, so their pectoral girdle is similarly undeveloped. These are hardly the only differences either; there are other marked changes to other parts of this girdle.

---

Summary

The pectoral girdle shows a lot of the same signals that we see in limb elements, which is what you'd expect for something that sometimes has to hold up half of the animal's weight while moving about on land. Proportions are the easiest way to infer lifestyle, but especially the cleithrum and the scapulocoracoid can exhibit major morphological differences as well (if you can them). One big caveat, as it often is with postcranial bones, is that immaturity can accidentally be confused for underdevelopment, and those immature forms are not necessarily practicing a different lifestyle than the adults. This remains one of the big challenges in paleontology because the fossil record will always be incomplete, and some taxa are probably represented only by juveniles. How we differentiate them from adults that just happen to be similarly underdeveloped is a major part of my research. Up next week: the other girdle - the hips!