Preparation, articulation, and photogrammetry of a neonatal moose (Alces alces) skeleton

Abstract

Objective

Neonatal skeletal articulations for research and display purposes are uncommon due to issues surrounding incomplete bone maturation and reduced structural integrity that affect the bone preparation and articulation procedures. The present project was designed to add to the veterinary literature on neonatal moose (Alces alces) osteological specimens, document the procedures of preparing and articulating a delicate specimen, and construct a 3-dimensional (3D) scan of the articulated skeleton that could be used for scientific and veterinary research and study worldwide.

Animal

A neonatal moose that had succumbed to capture myopathy resulting from entanglement in a barbed wire fence was the sole subject of this project.

Procedure

The neonatal moose carcass was degloved, the bones processed and cleaned, the skeleton articulated, and the articulated skeleton rendered into a 3D model using photogrammetry.

Results

The result was an articulated neonatal moose skeleton. The articulation is on display at the University of Northern British Columbia (UNBC); the 3D model is available on UNBC’s website.

Conclusion and clinical relevance

This project adds to the literature and makes available tools to study neonatal ungulate (A. alces) allometry, morphometry, osteology, and skeletal articulation. The documented processes and 3D model can be used as references in veterinary and biological research, study, and instruction. In addition, the 3D model is available to download (open source) for future projects from UNBC (www.unbc.ca/roy-rea/moose-calf-articulation).

There is a lack of information on the preparation and analysis of neonatal mammalian skeletons. Such skeletons do not appear commonly accessible or curated, especially in intact or untarnished form, given the issues associated with fragility due to incomplete ossification (1–3). This circumstance presents challenges, limiting both knowledge about proper preparation and rearticulation of such skeletons and the resources that wildlife biologists and those in the field of wildlife medicine have for neonatal anatomical reference (4). The aim of this project was to help address this deficiency and provide a valuable resource to professionals in various fields.

A 1- to 2-week-old male moose (Alces alces) calf was discovered entangled in a barbed-wire fence near Prince George, British Columbia, by a group of hikers on June 6, 2017. At the time of discovery, the calf was alive but alone. The hikers freed the calf and took it to the Ospika Veterinary Clinic in nearby Prince George. The skin and connective tissues on the distal aspect of the calf’s right rear leg had been badly damaged. The veterinarians at the clinic treated the wound, bandaged it, and rehydrated the calf. When the calf was stabilized, co-author RR transported it to a rendezvous point where he met staff from the Northern Lights Wildlife Shelter in Smithers, British Columbia. The staff transported the calf to the shelter where it was given antibiotics and tended. The calf lived at the shelter for a few days, after which it succumbed to its injuries on June 9, 2017. The shelter staff believed the calf eventually died of capture myopathy related to its struggle while trapped in the fence. Following discussions with the shelter staff, the calf was frozen whole. Several days later, the calf was transported back to Prince George to the University of Northern British Columbia, where it was stored in a −20°C freezer until it was processed.

After the proper arrangements had been made with a local taxidermist, the calf was thawed and skinned. The cape was used to make a taxidermized mount of the calf (for display at the university) by Rod Gray (Bear Bone-z Taxidermy). The skinned carcass was subsequently disarticulated into individual components and frozen (Figure 1 A). Sections were removed from the freezer one at a time and thawed to prepare the articulation. The bulk of the flesh was removed using knives and scalpels, and the flesh discarded (5). Due to the delicate nature of immature unossified bones, various methods were used for cleaning (3). Larger leg bones were gently boiled in a large stock pot until protein was denatured and flesh could be readily detached from bone (5,6). Medium-sized leg bones were sorted into mesh bags and soaked in a dilute hydrogen peroxide solution (1 to 2%) for 3 h at a time and intermittently scraped to remove flesh. The more intricate spinal column and skull were placed into a dermestid (Dermestidae) beetle tank until the flesh was removed, and then placed into a fresh hydrogen peroxide solution. Once cleaned, all bones were placed into a 2% borax solution for 5 to 7 d for degreasing. Throughout this process, photographs were taken to document the procedures.

FIGURE 1.

Photographs representing different stages in our methodology: carcass processing, bone preparation and articulation, and 3D photogrammetry. A — Documentation of the knee during flesh removal. B — Drilling and aligning vertebrae. C — Orthographic and lateral depictions of the 592 photographs, taken at different heights and angles denoted by white boxes, used to digitally reconstruct the neonatal moose skeleton.

The remaining steps of skeletal articulation were derived from Post (7). The bones were further processed in a solution of 4 L of 5% ammonia and 237 mL of dish detergent mixed with 16 L of water. During this processing, any soft cartilaginous parts were removed from bones to prevent decay in the articulation. Once any remaining grease and soft tissue were removed by the solution (~2 wk), the bones were dried in an oven heated at 35 to 65°C and removed individually as they dried. The bones were then separated by segment (into each of the forelimbs, hind limbs, head, and spine). A 2-dimensional layout of the bones was arranged and photographed to ensure that all bones were in their proper places, with an accounting of each. Once accounted for, the brittle young bones proved difficult to drill for the insertion of standard-use 12-gauge articulation wires. To overcome this issue, the bones were submerged in a “bone-hardening solution” consisting of 2 parts white Elmer’s School Glue to 1 part water for 15 s and then hung to dry; at times, more than 1 coat was needed for more brittle bones, and/or additional glue was painted over the bones to help keep them intact. Once dry, bones were predrilled for wire insertions and hot glue used to articulate the skeleton. Those bones that were still brittle, despite the hardening treatment, were repaired with Aves FIXIT Sculpt (Aves Studio, Hudson, Wisconsin, USA).

The spine was first articulated and mounted onto 2 metal posts attached to a wooden base with a stainless-steel threaded rod that was 1/3 the diameter of the smallest vertebra (Figure 1 B). The rod was suspended between and through each vertebra. The order in which bones of the skeleton were mounted, after articulation of the spine, was as follows: the ribs, then the skull (drilled through the base and filled with hot glue before mounting onto 1 of the 2 primary 2.5-centimeter hollow pipe posts), followed by the legs. After the main skeleton was articulated, transparent caulking was used to mimic the cartilaginous areas that were removed; blowing-bubble liquid was used to smooth the caulk. Last, acrylic paint was used to colour-match the skeleton to the patched sections composed of Aves FIXIT Sculpt, using 3 to 4 coats of paint.

After the skeleton articulation was assembled, photogrammetry was completed. Photogrammetry is the process of using photographs to create a 3-dimensional (3D) model using trigonometry on pixels. We used RealityCapture version 1.2.1.116300 RC software (Epic Games, Cary, North Carolina, USA) to process the photographs into a 3D model. This process estimates the position (and orientation) from which each photograph was taken and records the position of every feature identified in the photographs. Offsetting the photographs was required to calculate depth, similar to how humans use stereoscopic vision to perceive depth and 3 dimensions.

We initially scanned several bones individually but then uncovered an issue with this approach. First, the individual 3D models of each bone would not necessarily be at the same scale since photograph positions are estimations; this is important for comparing the sizes of individual bones. In addition, to continue with this initial approach, we required several more photographs and much more processing time than needed to obtain a single scan after articulation. Scanning after articulation solved both the efficiency and scale issues. Since all the bones were being scanned at the same time, in the same photographs, fewer individual 3D models needed to be rendered and each bone would be the correct size in relation to the others.

Excellent photogrammetry requires attention to detail in photography. Large, sharp photographs are required, as is consistent lighting, a large depth of field, and low sensor noise (static) (8). Photographs with higher resolution (megapixels) provide more details that can be extracted to build the model (8). Sensor noise distorts the image, reducing the accuracy of any features one wishes to capture (8). Modern cellphone cameras have a large depth of field and many megapixels, but smaller sensors introduce more noise into the photographs. To optimize the photographs, we used a Nikon Z5 mirrorless camera with a 24 to 70 mm lens set at 24 mm F13 1/3 s ISO 200 and mounted on a tripod.

For the final 3D scan, 592 photographs of the moose calf were captured, resulting in a point cloud of 6.0 million 3D points and a 3D model of 15.1 million triangles (Figures 1 C ,2) (9). We took photographs in a series of orbits around the skeleton at different heights and distances. To ensure we did not miss any areas, we rebuilt the 3D model after every orbit. This process ensured we covered any blind spots; for example, on top of or underneath the calf (8).

FIGURE 2.

A gradation from individual 3D points (the point cloud) to the complete 3D model of a neonatal moose calf skeleton, containing 6.0 million 3D points and a 3D model of 15.1 million triangles.

This project presented a unique set of challenges due to the lack of previous references on cleaning and rearticulating a neonatal skeleton. A multidisciplinary approach was used to produce a complete neonatal moose calf skeleton display accompanied by a 3D photogrammetric model that is now available online (open source) (10). The process was, to some extent, trial-and-error-based and can serve as a basis for similar future endeavors. A complete skeletal model, especially one that is easily accessible online, can serve information exchange across a variety of fields such as wildlife biology, comparative vertebrate anatomy, wildlife rescue medicine, and veterinary medicine/surgery.