Birds, along with humans, are the sole bipedal living animals. But unlike humans, birds’ forelimbs are highly modified and transformed into wings. As a result, all activities that birds perform must be done exclusively with their hind limbs and beaks.

Foot morphology of birds has evolved significantly over approximately 100 million years from the anisodactyl foot ancestral condition, i.e., three toes facing forward and one toe facing backward. Anisodactyl feet are still seen nowadays in birds like chickens or passerines. The evolution of birds’ feet led to various foot types adapted to fulfill several needs or locomotor skills. For example, some birds like ostriches have didactyl feet, i.e., they have lost two of the four digits and are specialized in running. Other birds like parrots have zygodactyl feet, i.e., they display a particular backward disposition of the fourth toe, allowing them to perch, climb, and manipulate objects more easily. There are also aquatic species like ducks with palmate feet, i.e., they have feet webbing between their toes, facilitating swimming and diving.

These adaptations allow birds to conquer a wide variety of habitats and adopt diverse lifestyles all over the world. Because of the importance of birds’ hind limbs, their study has been a topic of great interest among ornithologists, paleontologists, and morphologists historically. Scientists of the late 19th and early 20th centuries dedicated their efforts to describing the osteology and myology of birds’ hind limbs. To better understand hind limb anatomy and function, a plethora of studies employing advanced tools, such as geometric morphometric, finite element analysis, and muscular physiological cross-sectional area have been conducted over the past years.

Ten years ago, an innovative methodology known as Anatomical Network Analysis (AnNA) was developed by the researchers Diego Rasskin-Gutman and Borja Esteve-Altava. This approach in studies on morphology focuses on the connectivity patterns between anatomical parts. AnNA has its foundations in network science, an interdisciplinary field that uses the concepts of graph theory in the study of the behavior and dynamics of complex systems. Network science is widely applied in several areas such as telecommunications, epidemiology, economics, and social media. 

As a group of scientists investigating birds' morphology and evolution, we have adopted AnNA as one of the many tools in our research. Particularly in this work, we wanted to answer the following research question: what new insights into the evolution and function of the avian foot architecture can emerge from applying AnNA?

In doing so, we built networks of birds’ foot musculoskeletal systems considering bones and muscles as nodes, and articulations and muscle attachments as connections. We included 62 species from most major bird lineages. These species span a variety of primary lifestyles (terrestrial, arboreal, aquatic, and hyperaerial), foot types (anisodactyl, didactyl, tridactyl, heterodactyl, zygodactyl, ectropodactyl, semizygodactyl, pamprodactyl, syndactyl, lobate, semipalmate, palmate and totipalmate), nest attendance types (precocial and altricial), and diverse locomotor and manipulative skills (walking, running, hopping, wading, perching, climbing, swimming, diving, hanging upside down, grasping). 

To analyze these networks, we obtained different node and network parameters, such as connectivity degree, number of nodes, number of connections, density, among others, using the igraph package of the programming language R. With the obtained data, we estimated the complexity -or simplicity- of the foot. Also, we explored the foot distribution and grouping in a phylomorphospace, i.e., a projection of a birds’ tree into a multidimensional space where birds’ foot phenotypes are plotted. Finally, we mapped the parameters as characters onto a time-calibrated phylogenetic tree to estimate the ancestral parameters states in the hypothetical networks of birds’ ancestors.

Our analyses revealed some interesting and unexpected results. Here are the top 3 highlights:
(1) The tarsometatarsus and the digital flexor/extensor system have the highest connectivity degree, i.e., a node parameter showing how interdependent an anatomical part is in connection with others, and its limitations for evolutionary change. This may relate to the avian body plan, i.e., the layout characterizing the body structure of the clade Aves, as these anatomical parts already present in ancestral dinosaurs are largely conserved in modern birds (Neornithes).
(2) We found no link between more complex foot networks and the ability to perform specialized skills like climbing, powerful grasping, digital dexterity, and hanging upside down, nor to perform a higher number of different tasks. This means that the simplicity of the foot network does not limit its potential functions. Also, we found a trend toward the simplification of foot networks on a macroevolutionary scale: in general, basal species have complex networks, and derivate species have simple networks.
(3) Foot networks are phylogenetically constrained, i.e, limited for change, and conserved across all birds despite their diverse foot types. This may be the result of stabilizing selection during evolution, acting specifically in the maintenance of foot network connectivity rather than on foot type variation.

Our findings raise new research questions and invite us to explore birds’ feet from an Evo-Devo perspective. For this reason, our next efforts are dedicated to using AnNA in studies on the connectivity patterns during the development of the foot musculoskeletal system of birds.