Limbs, Shoulders, Necks, and Trunks: An Investigation of the Neck-Trunk Boundary in Squamates

Introduction

At least 30 squamate lineages have convergently evolved the elongate, limbless body plan most well known in snakes1, This makes squamates a useful group in which to study the factors that drive and constrain evolution. Though many aspects of the evolution of limblessness have been well studied, and compared between groups, body elongation, a defining part of this transition, is poorly understood.

The presacral region becomes elongate in all limbless and limb-reduced squamates, but it is unclear whether the neck region or the trunk region (torso) or both contribute to elongation. In other words, do limbless squamates have long necks or long trunks, or are both long? Furthermore, is elongation a similar process in all limbless groups?

Problem

One obstacle to studying body region evolution in limbless squamates is the difficulty of defining body regions.

In squamates, the boundary between the neck and the trunk region is usually defined as the last vertebra with a rib connecting to the sternum. In limbless squamates however, there is usually no sternum left!

A CT scan of a lizard torso in ventral view. The kite-shaped sternum is connected to three pairs of ribs. The last rib connects to vertebra 11, which designates the beginning of the trunk region. The neck is ten vertebrae long. — Here, the sternum (orange) is connected to a several ribs, but the last (indigo) connects to the eleventh vertebra (V11), making the neck (yellow) ten vertebrae long.

Proposed Solutions

How can one define the neck-trunk boundary in groups that lack a rib connecting to the sternum? Here are three proposed morphological markers.

1. The presence of hypapophyses. These projections (marked in yellow) on the ventral sides of anterior vertebrae are found on the neck vertebrae of pentadactyl lizards and could distinguish the neck.

The head and vertebrae of a CT scanned lizard skeleton in a ventrolateral view. The first nine vertebrae have flattened projections extending ventrally. They are laterally flattened and located in the middle r towards the posterior end of the vertebrae. They taper in size with the largest being about as long as the neural arches. Some are roughly shaped like the end of a hockey stick, but other species have other shapes like rectangles. — Hypapophyses are the ventral projections colored yellow.

2. The position of whats left of the pectoral girdle (shoulder blades, collar bones, sternum). This would indicate where the forelimbs would be if they were to develop and where the trunk region could begin. Here, it is marked in orange at V6.

At the level of the 6th vertebra, two pairs of pectoral girdle bones are present. They are almost mirror images of each other. the scapulocoracoids, or shoulder blades are sticks that slant posteriorly and medially and curve ventrally around the body. the clavicals are located medial to the scapulocoracoids and are small spheroids. — The pectoral girdle, here in orange, can hardly be recognized as a scapulocoracoids and clavicles, or shouderblades and collar bones

3. A shift in overall vertebral morphology. Neck vertebrae are shaped differently than trunk vertebrae. The difference in vertebral shapes could represent the boundary between the neck and trunk.

Method

CT data of lizards were downloaded from Morphosource.org and segmented in VGStudio or 3D Slicer. One lizard was used per species.

The hypapophyses were counted and the position of pectoral girdle elements (if they existed) were noted.

Shape Analysis

To assess overall vertebral morphology I used a geometric morphometric analysis (GMM). This is a statistical method used to calculate the similarity between complex shapes. Homologous points shared by all a specimen’s vertebrae were marked by landmark points. The positions of these points together represent the overall shape of the vertebra.

The shapes were then compared using a geometric

morphometric analysis. All the landmark data from a given specimen were compared and a Principle Component Analysis (PCA) was conducted.

A composite figure of three elements. the first is a CT scanned vertebra with orange dots at different points such as the posterior tip of the neural arch, a bump where the rib attaches, and the anterior extent of a prezygapophysis. the second component is a CT scanned skeleton of a lizard with a reduced pectoral girdle. each vertebra is a different color and each color matches a dot in the third component. the third component is a plot with axes labeled PC1 and PC2. the dots are labaled V3-V12 and are aranged in a upside down V shape in an order that matches the order of the vertebra of the skeleton. The apex of the V is V7 and is labeled "start of trunk." V — Each dot in the PCA plot represents the shape of a vertebra. The axes aren't important, all you need to know is that the closer dots are to each other, the more similar their shapes are. The shift in morphologyalong the axial column is described by the shift in the line of the dots. The vertebra at the apex is the first trunk vertebra.

Results

Here’s the long description. Buckle in its going to be long. This tree of squamates depicts the seven major groups to contain limb-reduced and limbless members. I will say which species are pentadactyl, typically shaped lizards. All others are limbless The earliest diverging group is Dibamidae with two species. Anelytropsis papillosus which has a geometric nec- trunk boundary at V7, pec girdle at V7 and hypapophyses that extend to V8. Dibamus novaeguinea has a GMM boundary at V8, pec girdle at V7, and hypapophyses to V8. The next earliest diverging group is Gekkota. It has two major branches, one with two pentadactyl species and one with two limbless species. The pendadactyl species are Gekko gecko with a GMM boundary and pec girdle at V9, and hypapophyses that extend to V7. Nephrurus asper has a GMM boundary at V7 and ho hypapophyses. Now for the limbless gekkotans. Aprasia clairae which has the GMM boundary and pec girdle at V7 and hypapophyses to V8. Lialis burtonis has all three markers at V6. The next diverging group is a clade containing cordylids and scincidae. I donly had one cordylid in my analysis: chamaesaura anguina. It has a GMM boundary at V6, pec girdle at V7 and hypapophyses to V6. Within scincidae, there is a divergence between a group comprised of Lerista species, and one of other skinks. Within Lerista, L. bougainvilii, a pentadactyl species is the earliest diverging and it has a GMM boundary at V8, pec girdle at V9, and hypapophyses to V6. Lerista apoda has a GMM boundary at V11, pec girdle at V9, and hypapophyses to at least V11, but I the scan I had was only 11 vertebrae long and its likely the hypapophyses went to V13. Lerista stylis has a GMM at V6, pec girdle at V8, and hypapophyses to V8. The other branch of Scincidae has a split with two clades. The first I will describe has Typhlosaurus caecus as the earliest diverging member. It has a GMM boundary at V9, hyapophyses to V12, and no pectoral girdle. The others in this subclade are two acontias species. A. gracilicauda has a GMM boundary at V7, and a pec girdle at V6, and hypapophyses to V10. A. garipensis has a GMM boundary at V9, hypapophyses to V10 and no pectoral girdle. The other subclade is comprised of the following. Melanoseps loveridgei which has a GMM boundary at V8, pec girdle at V7, and hypapophyses to V8. Ophiomorus raithmai has a GMM boundary at V8, pec girdle at V9, and hypapophyses to V8. The remaining squamates are comprised of four clades. There is a split in the tree with one clade containing Amphisbaenians and Gymnophthalmidae and the other containing Anguidae and Serpentes. Here are the Amphisbaenians. Rhineura floridana is the earliest diverging. it has a GMM boundary at V9, no pec girdle and hypapophyses to V9. the next earliest diverging is bipes canaliculatus. It has a GMM boundary to V7, a pec girdle at V5 and hyappophyses to V5. the next earliest diverging is Blanus cinereus. It has a GMM boundary at V6, a pec girdle at V4, and hypapophyses to V6. Bonus cinereus split from a clade comprised of Amphisbaena fuliginosa and agamodon anguliceps. Amphisbaena fuliginosa has a GMM boundary at V5, a pec girdle at V3 and hypapophyses to V7. Agamodon anguliceps has a GMM boundary at V6, pec girdle at V3 and hypapophyses to V14. The cade that split form amphisbaenians is Gymnophthalmidae. The earliest diverging in this group is the pentadactyl Cercosaura ocellata which has a GMM boundary at V8, pec girdle at V9, and hypapophyses to V6. the two limb reduced gymnophthalmids, Bachia dorbignyi and Calyptommatus sinebrachiatus both have a pec girdle at V8 and hypapophyses that extend to V8. their geometric morphometric boundary is less of a boundary and more of a transition with vertebrae from V6 to V9 all showing both cervical and trunk characteristics. The clade that splits from the amphisbaeniand and gymnophthalmids contains Anguidae and Serpentes. The earliest diverging anguid is the pentadactyl Elgaria coerulea, which has a GMM boundary at V9, pec girdle at V9, and hypapophyses to V5. the two limb reduced species are as follows. Pseudopus apodus has a GMM boundary at V6, a pec girdle at V5, and hypapophyses to V6. Ophiodes striatus has a GMM boundary at V9, a pec girdle at V6, and hypapophyses to V5. I used two snake species, a blind snake and a colubrid. The former, Typhlops arenarius has a GMM boundary at V8, no pec girdle, and hypapophyses to V5. the latter, pantherophis guttatus has a GMM boundary at V7, no pec girdle, and hypapophyses extending to V54. — A phylogenetic tree of the specimens in the sample with each of the three proposed neck-trunk boundary markers illustrated. Anterior is to the left.

Discussion

Neck lengths vary between species and markers. All three markers are generally loosely associated with each other but seldom co-occur. However, in some species, the hypapophyses extend well past the other markers.

Clades show similar patterns

Related species show similar patterns suggesting phylogenetic history influences elongation. For example, the PCA plots for Gymnophthalmids show a gradual shift from neck to trunk morphology (top) rather than the typical sharp angle (bottom).

Necks tend to get shorter as bodies become longer

No specimen had a pectoral girdle past V9, and only a few limbless skinks had a GMM boundary posterior to the plesiomorphic position. This suggests a trend towards a decrease in neck length and that presacral elongation occurs only in the trunk for those metrics.

Variation leads to questions about body patterning

The neck-trunk boundary is thought to be patterned in development by the anterior extent of expression of the HoxC6 gene. However, this explanation is insufficient to account for species like Agamodon anguliceps, with possible boundaries at V3, V6, and V14. Which of these markers, if any does HoxC6 pattern? In Pantherophis guttatus, HoxC6’s anterior extent is V114, but none of the markers in this study occur there.

Given the variation in these data, there not be a true neck-trunk boundary. Though this makes body region determination difficult, it may be the key to revealing nuances in vertebrate body patterning.

Citations

1. Camaiti M., et al. (2022) A database of the morphology, ecology and literature of the world’s limb-reduced skinks.

2. Head, J., Polly, P. (2015) Evolution of the snake body form reveals homoplasy in amniote Hox gene function.

3. Zheng, Yuchi, and John J. Wiens. (2016) Combining phylogenomic and supermatrix approaches, and a time-calibrated phylogeny for squamate reptiles (lizards and snakes) based on 52 genes and 4162 species.

4. Woltering, Joost M., et al. (2009) Axial patterning in snakes and caecilians: evidence for an alternative interpretation of the Hox code.

Acknowledgements

Thank you to my husband Patrick, my family, and everyone who put their data on Morphosource.

Krista Koeller