This guest post is from Leonardo Maiorino, a vertebrate paleontologist with a particular interest in understanding the evolution of the skull in horned dinosaurs. Leo was at the helm of a recent paper in PLOS ONE (I was a co-author), so I invited him to write up a post for the blog. — A. Farke
Anatomical and behavioral differences distinguish males and females in many extant and extinct animals. For instance, male peacocks have a large and flashy tail, whereas females are smaller and less brightly colored. Male lions have a mane and are larger than females. Red deer male sport antlers, lacking in females. This phenomenon is called sexual dimorphism and represents a product of sexual selection. It represents a key factor in the success of breeding within many species, as originally stated by Darwin, and mate choice.
Despite obvious sexual dimorphism in some modern animals, sexual dimorphism is hard to identify in extinct organisms such as dinosaurs, often due to a very small sample size. Having a few fossil skulls with different morphologies (crest or horns) and relating them to sexual differences thus appears quite hasty. For instance, hypothesized males and females of hadrosaurs (Parasaurolophus or Lambeosaurus) were later shown to be species separated in time and in space. Variation in the number of chevron bones (wishbone-shaped bones under the tail vertebrae) in Tyrannosaurus, previously attributed to sex differences between males and females, are instead the product of non-sexual variation.
Ceratopsians, popularly known as “horned dinosaurs”, represent another group for which paleontologists have inferred sexual dimorphism. Protoceratops andrewsi is a well know ceratopsian from the Late Cretaceous of the Gobi Desert, Mongolia, discovered almost a century ago during the American Museum Expeditions of the 1920s. Protoceratops is a small (<2 m total body length) quadrupedal animal, characterized by a skull with a thin, bony frill projecting over the neck and lacking the prominent horns that characterized ceratopsids such as Triceratops. More than one hundred well-preserved skulls and skeletons (many from various life stages) have been unearthed, providing a lot of information on its paleoecology and paleobiology. This large sample size allows a statistical test on whether shape differences of skulls could be truly related to sex.
Back in 1975, Peter Dodson used skull measurements from 24 specimens to investigate growth changes in Protoceratops as well as possible sexual dimorphism. He assigned the sex to each specimen on the basis of its position on scatter plots (above or below a trend line). Based on this analysis, Dodson hypothesized that male Protoceratops had a broader and taller frill, broader skull above the eyes, and taller nasal bump than seen in females. This analysis was quite advanced for its time, but given advances in statistical methods, morphometrics, and computing power, as well as the discovery of many new fossils, we decided to look at the problem again.
Geometric morphometrics is a suitable method to analyze shape and size differences among groups of animals. Where previous work used basic measurements, we digitized the shape of 29 skulls in top and side views, based on photographs. We assigned skulls to male or female using the previously outlined criteria. Geometric morphometrics allows the study of shape variation within sampled specimens and capturing the overall geometry (instead linear measurements) of biological structures by means of simple analyses. How can we detect shape information? It is easy: just digitizing spots on images (called landmark) on homologous anatomical features for the all skulls (in our case) in top and side view. By so doing, we have now a large sample constituted by geometries that represent the shape of the skull.
We asked ourselves, if males and females have different skull shape, are we able to distinguish the two groups using geometric morphometrics and statistical tests?
We plotted all specimens in a X-Y graph. One graph for the Protoceratops individuals in side view, and one other for the individuals in top view. We refer to those X-Y graphs as “morphospace” because they represent the shape variation of the specimens in the space. Those two graphs highlighted that males and females overlap each other; the two groups are not well separated in the plots. This suggests that males and females possess similar skull shape. Moreover, we used statistical tests to verify if what we can see in the graphs is statistically reliable and thus if the shape difference between males and females is well-founded.
No matter how we plotted or analyzed the data, we weren’t able to find a clear, statistically significant difference between males and females. Therefore, we think that the previously hypothesized criteria are not suitable to distinguish the two sexes in Protoceratops andrewsi easily. Frill height and width as well as the width of the skull over the eyes do not represent sex-discriminant features in our sample. The occurrence of the horn above the nose would seem partially related to sexual dimorphism, even if this evidence is not well-rendered.
We suggest that cranial shape variation previously interpreted as sexual dimorphism is related instead to shape changes occurring during the animal growth and perhaps also related to the shape variation occurring among the adults of this ceratopsian dinosaur. Maybe, new fossils will give the chance to confirm our results or reject them.
Maiorino, L., A. A. Farke, T. Kotsakis, and P. Piras. 2015. Males resemble females: re-evaluating sexual dimorphism in Protoceratops andrewsi (Neoceratopsia, Protoceratopsidae). PLoS ONE 10:e0126464.