The changing pace of insular life: 5000 years of microevolution in the Orkney vole (Microtus arvalis orcadensis)

Abstract

Island evolution may be expected to involve fast initial morphological divergence followed by stasis. We tested this model using the dental phenotype of modern and ancient common voles (Microtus arvalis), introduced onto the Orkney archipelago (Scotland) from continental Europe some 5000 years ago. First, we investigated phenotypic divergence of Orkney and continental European populations and assessed climatic influences. Second, phenotypic differentiation among Orkney populations was tested against geography, time, and neutral genetic patterns. Finally, we examined evolutionary change along a time series for the Orkney Mainland. Molar gigantism and anterior-lobe hypertrophy evolved rapidly in Orkney voles following introduction, without any transitional forms detected. Founder events and adaptation appear to explain this initial rapid evolution. Idiosyncrasy in dental features among different island populations of Orkney voles is also likely the result of local founder events following Neolithic translocation around the archipelago. However, against our initial expectations, a second marked phenotypic shift occurred between the 4th and 12th centuries AD, associated with increased pastoral farming and introduction of competitors (mice and rats) and terrestrial predators (foxes and cats). These results indicate that human agency can generate a more complex pattern of morphological evolution than might be expected in island rodents.

Keywords: Dispersal; evolutionary rate; geometric morphometrics; island evolution; tooth shape; zooarchaeology.

Figure 1

Localization of the modern samples (abbreviated group names) and ancient samples (stars with site names) of Microtus arvalis in continental Europe and Orkney (insert) distribution area of the species (gray shading; Shenbrot and Krasnov 2005).

Figure 2

(A) Occlusal surface of Microtus arvalis right first lower molar (M₁): T, triangle; bra, buccal reentrant angle; lra, lingual reentrant angle. (B) Position of the 18 landmarks and 12 semilandmarks.

Figure 3

Variation in molar size (log‐transformed centroid size) in ancient (gray boxes) and modern (open boxes) Microtus arvalis samples (see Fig. 1 for localization and Tables 1 and 2 for sample details). Left, modern continental European samples include specimens grouped according to their genetic lineage (IT, Italian; EA, Eastern; WS, Western‐South; WN, Western‐North; CE, Central) and insular origin (Yeu, WSyeu; Noirmoutier, WSnoi; Guernsey, WNgue). Right, ancient Orkney specimens are grouped according to their geographic location in the archipelago.

Figure 4

Molar shape differentiation of modern and ancient Microtus arvalis samples from continental Europe (islands included) and Orkney. (A) Scatter plot of the two first principal components of the morphometric analysis. Each symbol with its group code (Tables 1 and 2) corresponds to the mean value of a modern or ancient group, bracketed by the 95% confidence interval. (B) Molar shape change associated with PC1 is depicted with a wireframe graph connecting the landmarks and semilandmarks. The gray wireframe represents the mean shape and the black wireframe represents the shape changes along PC1 in negative (−0.06) and positive (+0.06) directions.

Figure 5

Allometric trend in modern and ancient Microtus arvalis estimated from the regression of Procrustes coordinates on centroid size. Insert: the gray lines and open circles represent the average shape and the black lines and circles represent the predicted shape for a centroid size increase of 2 mm.

Figure 6

Molar shape variation in modern and ancient Microtus arvalis samples from Orkney. (A) Scatter plot of the two first principal components of the morphometric analysis. The modern samples are represented by empty symbols whereas the ancient samples are shown by gray‐filled circles. The dotted line separates northern from southern populations. (B) Molar shape change associated with PC1 and 2 is depicted by a wireframe graph connecting the landmarks and semilandmarks. The gray wireframe represents the mean shape and the black wireframes indicate the shape change of the PC score by 0.6 units in the positive and negative directions.

Figure 7

Morphological change over time in Mainland Orkney voles. (1) Molar size evolution depicted by mean size with confidence intervals. (2) Molar shape evolution depicted by mean scores along the second principal component axis detailed in Figure 6. (3) Allochronic shape changes expressed as Mahalanobis's distances (D) between each chronological step of the time series.