Skip to main content

The functional diversity of marsupial limbs is influenced by both ecology and developmental constraint

Cite this dataset

Pevsner, Spencer; Grossnickle, David; Luo, Zhe-Xi (2022). The functional diversity of marsupial limbs is influenced by both ecology and developmental constraint [Dataset]. Dryad.


Extant marsupials are less ecologically diverse than placentals, and this is reflected by placentals exhibiting a greater diversity of locomotor modes, including powered flight and fully aquatic swimming. One proposed explanation for this discrepancy is that the development of more disparate marsupial forelimbs is prevented by the neonate’s crawl to the pouch, which requires precocious forelimb development for climbing adaptations. To test predictions of this Developmental Constraint Hypothesis, we pursue a comparative morphometric study on osteological traits of mammalian limbs, with an emphasis on functional differentiation of marsupial limbs among locomotor modes. We apply multivariate analyses to a large dataset of limb metrics and a diverse sample of mammals, with the placental sample limited to taxa whose locomotor modes are exhibited in marsupials. Overall, we do not find consistent evidence in support of the Developmental Constraint Hypothesis. Diprotodontia serves as an exception, with comparisons of their forelimbs to hind limbs supporting the Developmental Constraint Hypothesis. Our results suggest that developmental constraints on marsupial forelimbs may have limited marsupial diversity to some degree. Despite this, the marsupial locomotor groups show unexpectedly high levels of morphological differentiation relative to placentals of the same locomotor modes, indicating that ecological functions may overcome developmental constraints on a macroevolutionary scale.


This dataset was collected by taking photographs of the skulls, girdles, and limb elements of mammals. The limb and girdle skeletal elements examined included the scapula, humerus, radius, ulna, metacarpals, manual phalanges, pelvis, femur, tibia, fibula, calcaneum, metatarsals, and pedal phalanges. Linear measurements were collected from photographs with scale bars using ImageJ. The data included here are these raw measurements, and not the log-shape ratios used in downstream analyses. The method by which the log-shape ratios were calculated is detailed in the Materials and Methods section of the main text of this study.

Usage notes

The units for all measurements included in this dataset are in millimeters.