How Comparable are 2D and 3D Methods for Measuring Bone Laminarity in Developing Pigeons

The cortical bone of birds contains microscopic tunnels called vascular canals, which can be classified based on orientation as: longitudinal, radial, oblique, or circumferential. Circumferential canals are hypothesized to reflect adaptation to flapping flight. This hypothesis is supported by histological surveys of birds that show the proportion of circumferential canals (laminarity) is elevated in bones thought to experience habitual torsional loads. However, the most detailed data on torsional loading during avian flight is from the pigeon, a species in which the adults show low wing bone laminarity. Nevertheless, there may still be support for biomechanical adaptation if the laminarity is greater in adults than in juveniles incapable of flight. Alternatively, if the pattern is reversed, laminar bone may simply be a juvenile trait. Laminarity has largely been analyzed using 2D histological sections, which may not accurately represent complex 3D structure. High‐resolution micro‐CT is an imaging technique that allows 3D reconstruction, providing greater accuracy and objectivity in the measurement of canal orientation. Here, we compare 2D and 3D methods to clarify how wing bone laminarity develops in pigeons. We sampled a growth series of four pigeons that were collected from a previous study. A cortical strip (1 × 3 mm) was cut from the right humerus, ulna, and radius of each bird. The strips were scanned with high resolution microCT. 3D analysis involved segmentation, autoskeletonization, and measurement of 3D canal orientation using Avizo. For 2D analysis, five evenly spaced orthoslices were digitally extracted from each strip. Canals were fitted with ellipses and trigonometric equations were used to approximate orientation. In both 2D and 3D methods, canals were categorized and laminarity was calculated. Principal components analysis was performed to convert highly correlated variables (i.e., mass, bone length, and section modulus) into principal components (PCs). The relationship between principal components and laminarity was assessed using beta regression modeling. 2D analysis overestimates laminarity among juvenile specimens. Therefore caution is warranted when applying 2D analysis to complex canal networks. Nevertheless, both methods show that the laminarity of each element significantly decreases with maturity (PC 1). Therefore, torsionally loaded bone is not necessarily more laminar. It seems laminarity is influenced by additional factors that warrant further investigation. Support or Funding Information College of Graduate Studies Intramural Funding (Erin L. R. Simons & Andrew H. Lee); Kenneth A. Suarez AZCOM Summer Research Fellowship (Alexandra Brown, Kittaporn Jonglertjanya, Rylee S. McGuire)