Biologists have long sought general principles to understand the ecology and evolution of organisms. The global distribution of animals is something that scientists have tried to explain for centuries and has important implications for crafting conservation strategies. One such explanation for the modern distribution of animals is Bergmann’s rule. The rule predicts that homeotherms (animals with constant body temperatures) living in cooler climates (typically higher latitudes) tend to be larger than those in warmer climates (lower latitudes). Commonly cited examples of this phenomenon include Arctic polar bears being larger than bears in more southern environments, like Southeast Asian sun bears and black bears, and Antarctic emperor penguins being larger than Galápagos penguins. However, it is uncertain whether the rule applies across long stretches of geologic time or has been a major driver of body size evolution. Our study tackles this question from a palaeontological and evolutionary perspective by applying phylogenetic methods and climate modelling to a large dataset of fossil occurrences and living species.

Dinosaurs are one of the most successful vertebrate groups. They evolved a remarkable range of body sizes, including the largest-known land animals, and lived in virtually every habitat on the planet. Today, they remain successful, as birds are the most diverse vertebrate group, outside of fish, with over 10,000 species. This success makes dinosaurs an ideal group to test Bergmann’s rule. We collected body size data for 339 Mesozoic dinosaur species, including some Mesozoic birds, and 62 Mesozoic mammals, as well as data on where they lived. Our dataset also includes dinosaurs from the Prince Creek Formation of Alaska, the only dinosaurs known to have experienced cold-weather winter conditions (Figure 1). However, there are three key things to consider when analysing fossil data.

First, we need to account for plate tectonics, as the current locations of fossil sites are not representative of where they were buried millions of years ago. For example, over 230 million years ago, during the evolution of the first dinosaurs, all the continents were merged into one supercontinent, Pangea. The continents we recognise today fragmented and drifted over the last 200 million years—the entire duration of dinosaur evolution. We, therefore, used a plate tectonic model that corrected for the past locations of continents and applied our phylogenetic and climate models to these corrected palaeo-coordinates of fossil occurrences (Figure 2).

Second, we also must factor in past climates, as the Mesozoic was much more temperate than the present. To do this, we collaborated with Dr. Alex Farnsworth (University of Bristol) and utilised the HadCM3BL model. This allowed us to estimate the local environmental temperatures at the time dinosaurs were living, which is essential for understanding the effect temperature may have had on dinosaur body size evolution.

Third, utilising fossil data calls for a sampling bias metric. This measures for fossil record biases, including if larger species are overrepresented in the fossil record or if biases exist against the preservation of smaller fossils. The geological processes that buried and preserved bones millions of years ago did not treat all animals equally. Smaller bones may be more prone to destruction, and when we palaeontologists find fossils today, we are more likely to overlook small bones in the field. Our sampling bias metric found that the observed variation in Mesozoic dinosaur and mammal body size was not biased by these preservational effects.

We tested for a relationship between size and temperature, but in an evolutionary context to account for the shared ancestry of species. This is an important step and criterion of Bergmann’s rule, as an animal’s body size is strongly influenced by that of their ancestors. For instance, you’d expect a polar bear to be more similar in body size to a black bear due to more recent common ancestry than to something much more distantly related, such as a crocodile. Our phylogenetic approach tests if the amount of change in body size along a branch or lineage correlates with the amount of change in local temperatures estimated from the HadCM3BL model. We would expect an inverse relationship between the two if decreasing temperatures were selecting for increases in body size. We found no signal for Bergmann’s rule in Mesozoic dinosaurs or mammals (Figure 3).

We have several hypotheses as to why this is the case. It may just be that there was not an extreme enough temperature gradient in the Mesozoic for Bergmann’s rule to function. However, based on our palaeotemperature estimates, this does not seem to be the case. It could also be that differences in body size were not selected for across species, but there may have been within-species variation explained by latitude or temperature (this is an alternative definition of Bergmann’s rule which has been proposed). Or perhaps Mesozoic dinosaurs and mammals encompass a greater degree of variation in thermophysiologies than what is seen in modern birds and mammals (i.e., more ectothermic and poikilothermic-leaning species), which has also been proposed for various ornithischian and sauropod dinosaurs.

After these initial analyses, we figured it would be worth adding a comparative analysis of living species. We collected body size, locality, and local environmental temperature ranges for 2,305 modern mammals and ran the same tests. To our surprise, we still didn’t find any signal for Bergmann’s rule (Figure 4). This came as a shock, as Bergmann’s rule is frequently cited as an explanation for why mammals such as moose, bears, and porcupines are bigger in Alaska than they are in more southern places like Montana or Colorado. Bergmann’s rule has long been considered a general ecological rule that applies to most mammal species.

We decided to keep investigating, so we added another dataset of 5,496 modern birds, the other major group of modern homeotherms, and the descendants of Mesozoic dinosaurs. The results were curious. We found no signal for Bergmann’s rule when we compared bird body mass to latitude, but we did find a small relationship between body mass and temperature. Our analyses demonstrated that about 13% of increases in bird body mass can be explained by decreasing temperature (Figure 4). 

We are still unclear as to the exact cause of this signal. We have a test that assesses an evolutionary signature of coevolutionary change that should apply reasonably to Bergmann’s rule. We made computer simulations with simulated data to show that it works, and it produces a signal in bears, but it doesn’t apply more broadly across homeotherms. So, what is going on? If Bergmann’s rule was as strong of an ecological principle as it has been historically treated, why isn’t there any signal when we compare body mass to latitude? If it’s something to do with bird physiology, why is there no signal for Mesozoic birds? Why didn’t we find any signal in modern mammals?

Preliminary investigations into our data suggest that Cenozoic climate change may have something to do with it. We found that the highest rates of avian body mass evolution occur in birds that rapidly speciated and dispersed in the last 23 million years. This is also consistent with decreasing avian body size associated with modern-day global warming.

Overall, our study shows the power of the fossil record in addressing longstanding ecological principles. We find that, although Bergmann’s rule has historically been considered a broadly applicable macroecological rule, it really doesn’t explain most extant homeotherm body size diversity.

The study was published in the journal Nature Communications on April 5^th.

Lauren Wilson recently graduated with her Master’s degree through the College of Natural Science and Mathematics at the University of Alaska Fairbanks, where she studied palaeontology. Lauren will start her Ph.D. in the fall under Princeton University’s Geosciences Program.

Jacob Gardner received his Ph.D. in Ecology and Evolutionary Biology at the University of Reading, UK, where he’s now a postdoctoral research associate researching human evolution.

This research was supervised by Montana State University professor Chris Organ.

Contact: lnkeller@alaska.edu, jacob.gardner@reading.ac.uk, organ@montana.edu