Metabolic rate in common shrews is unaffected by temperature, leading to lower energetic costs through seasonal size reduction

Abstract

Small endothermic mammals have high metabolisms, particularly at cold temperatures. In the light of this, some species have evolved a seemingly illogical strategy: they reduce the size of the brain and several organs to become even smaller in winter. To test how this morphological strategy affects energy consumption across seasonally shifting ambient temperatures, we measured oxygen consumption and behaviour in the three seasonal phenotypes of the common shrew (Sorex araneus), which differ in size by about 20%. Body mass was the main driver of oxygen consumption, not the reduction of metabolically expensive brain mass. Against our expectations, we found no change in relative oxygen consumption with low ambient temperature. Thus, smaller body size in winter resulted in significant absolute energy savings. This finding could only partly be explained by an increase of lower cost behaviours in the activity budgets. Our findings highlight that these shrews manage to avoid one of the most fundamental and intuitive rules of ecology, allowing them to subsist with lower resource availability and successfully survive the harsh conditions of winter.

Keywords: Dehnel's Phenomenon; metabolism; oxygen consumption; winter adaptation.

Figure 1.

Shrew mass (g) and standardized braincase heights measured from X-rays. All capture data from 2014 to 2018 are shown in small circles, and the individuals presented in this study are shown as large circles. Both shrew body mass and braincase height differ by season for the animals in this study.

Figure 2.

Activity budgets as per cent of total observations that shrews spent in four activities. Shrews in winter spend more time eating and looking for food.

Figure 3.

Shrew absolute metabolic rates (a–c) and relative metabolic rates (d–f) predicted by body mass, relative braincase height and ambient temperature. Margin plots show the density of each variable in the row or column. Mass, season and temperature had significant effects on absolute metabolic rates on their own and the lines in (a) and (c) show the predicted relationship from the models that account for random slope of season, random intercept of individual and a continuous time autocorrelation structure within the data.