We gathered over 200 skulls from different museum collections across Europe, including both European and American minks plus some European polecats and mink-polecat hybrids for comparison. The sample includes males and females of all species as well as adults and juveniles. Additionally, to further investigate intraspecific variation in each mink species, we collected specimens from the three current European mink populations (western, north-eastern, south-eastern), and both wild and feralized American mink from North America and Europe respectively.
Each specimen was then imaged using micro-computed tomography (microCT) at different facilities, and the resulting scans were segmented using a combination of commercial software to produce the 3D models that were used in the shape analyses (see below).
Geometric morphometrics (GM) is a methodological approach to study the form of a structure based on the coordinates of a series of homologous points, or landmarks, defining that structure. The sets of landmarks for all individual specimens are first scaled, rotated and translated in a process called Generalized Procrustes Analysis (GPA), which standardizes the effects of size, orientation and position in the sample, allowing the remaining shape data to be further analyzed (see below).
Thus, in order to quantify the morphology of the skull in European and American minks, the 3D coordinates of a set of landmarks (red points in the figure) were digitized on the left side of each cranium and mandible model rendered from the microCT scans. Additionally, semilandmarks were sampled across curves (blue points) and surfaces (green points) to provide further detail. Following GPA, the shape variation across our sample of mink skulls was studied.
Principal component analysis is used to explore the overall shape variation in a sample, as it generates a morphospace (or shape space) in which the axes represent the main shape changes across all individuals. For instance, the morphospace in the figure represents the shape variation in our mink crania, with each dot representing one European mink (blue), American mink (pink), European polecat (green) or mink-polecat hybrid (orange).
Other multivariate methods can be used with shape data to, for instance, study the relationship between shape and size (allometric regression) or other factors (partial correlation), quantify differences between groups (Procrustes ANOVA), or classify unknown specimens (discriminant analysis).
The results obtained so far can be found here, but keep an eye on the news section and our twitter account for updates on our findings!
Some of our partner organizations donated deceased mink specimens for us to study their muscle anatomy. European minks came from breeding programs and died from natural causes, while the American minks were acquired from sanctioned, humane cullings carried out to mitigate the threat they pose to the conservation of the native species.
To visualize both muscle tissues and bone in a non-destructive manner, the specimens were imaged using diffusible iodine-based contrast-enhanced computed tomography (diceCT). This enhancement to regular microCT uses a solution of potassium iodide (I2KI), in which specimens must incubate for some time, to increase the differential attenuation of X-rays among soft tissues. This has been shown to reveal patterns of muscle fibres and fascicles against the connective tissues. DiceCT images were used to reconstruct the morphology of all the jaw-closing muscles in 3D. This data, together with measurements of muscle fibre length, will allow us to accurately estimate muscle forces for each jaw muscle.
Finite element analysis (FEA) is a computer modelling technique developed to solve complex problems in engineering, but it has been adopted to study the mechanical properties of biological structures (such as bones). In FEA, a high-resolution 3D model of a geometrically complex object is virtually loaded to predict the patterns and magnitudes of stress, strain and deformation across that object.
Sensitivity analyses of material properties, muscle force magnitudes and orientations, and constraint locations, can be undertaken to provide a more accurate and realistic depiction of the loadings experienced by an object, in our case, mink skulls during feeding.
MINKS will create 3D FE models from the specimens used to reconstruct muscle architecture. This will allow us to compare the stress and strain patterns across the skull at a range of gapes and bite points, simulating how performance varies with prey of varying sizes and toughness along the dental row.