Big boned: How fat storage and other adaptations influenced large theropod foraging ecology

Abstract

Dinosaur foraging ecology has been the subject of scientific interest for decades, yet much of what we understand about it remains hypothetical. We wrote an agent-based model (ABM) to simulate meat energy sources present in dinosaur environments, including carcasses of giant sauropods, along with living, huntable prey. Theropod dinosaurs modeled in this environment (specifically allosauroids, and more particularly, Allosaurus Marsh, 1877) were instantiated with heritable traits favorable to either hunting success or scavenging success. If hunter phenotypes were more reproductively successful, their traits were propagated into the population through their offspring, resulting in predator specialists. If selective pressure favored scavenger phenotypes, the population would evolve to acquire most of their calories from carrion. Data generated from this model strongly suggest that theropods in sauropod-dominated systems evolved to detect carcasses, consume and store large quantities of fat, and dominate carcass sites. Broadly speaking, selective forces did not favor predatory adaptations, because sauropod carrion resource pools, as we modeled them, were too profitable for prey-based resource pools to be significant. This is the first research to test selective pressure patterns in dinosaurs, and the first to estimate theropod mass based on metabolic constraints.

Fig 1. Phenotype heredity.

Allosaur Agent offspring phenotypes were determined based on a hereditary gradient of Advantage Factors, which were bound to an expression magnitude. In this example, the parent Agent will have offspring with an 80% chance of having “medium” tailfat, but 10% chance to be small, and 10% to be big.

Fig 2. Phenotype compositions.

In year 140, most surviving Allosaur Agents on the final step had improved hearing and bite_force attributes compared to the founding population, but detection_range and tailfat were negatively selected. Year 144 resulted in a population with improved detection_range, while hearing and bite_force in contrast were negatively selected.

Fig 3. Reproductive probability distribution of high tailfat individuals.

Kernel density estimation for the lifetime reproductive success distribution of Allosaur Agents. Allosaur Agents with tailfat above 35 kg per feeding event had more than 0.28 likelihood probability of producing 1 offspring during their lifetimes, whereas in all others, the probability was ca. 0.12, i.e., a factor of 2.33 times higher for tailfat > 35kg than for all other tailfat phenotypes. This fitness pattern agrees with other empirical and modeled population data, regardless of life-history scheme [66].

Fig 4. Reproductive probability distribution of high bite_force individuals.

Reproductive success distribution of high bite_force individuals. Agents with high bite_force were less likely to produce offspring than all others.

Fig 5. Reproductive success of Allosaur Agents vs. tailfat expression.

Reproductive success of Allosaur Agents vs tailfat expression.

Fig 6. Reproductive success of Allosaur Agents vs carcass detection_range.

Average Allosaur Agent carcass detection_range vs. lifetime reproductive success. Carcass Object detection_range was consistently higher in the most successful Allosaur Agents, demonstrating a clear selective advantage for theropods that could locate sauropod carrion better than their peers. These data suggest it therefore is likely that animals like Allosaurus used olfaction to detect food opportunities in a similar manner.

Fig 7. Reproductive success of Allosaur Agents vs. bite_force.

The bite_force attribute determined the likelihood an Allosaur Agent would successfully kill a Prey Agent upon encounter. Selective pressure did not favor this attribute in our model, and given that allosaurs and most large allosauroids were poorly adapted to overpower prey, we think it is likely that the mechanisms at work in this model were broadly similar to those present in Mesozoic ecosystems.