Large scale geographic differences in magnitude of seasonal changes of braincase height and body mass in S. araneus
Our results from the literature review confirmed large variation in the intensity of Dehnel’s Phenomenon between populations. The data on decrease and regrowth of braincase height, brain mass and body mass inS. araneus are compiled in Table S1 (Supporting Information), and the list of corresponding references in the Supporting Information file. Mean ± SD decrease in braincase height from first summer peak to winter minimum was 13.4 ± 2.4% and regrowth from winter subadults to overwintered adults was 10.3 ± 2.8%. Proportion of braincase height decrease was positively correlated with latitude, longitude and their interaction (Fig. 1, Table 1) but not with altitude (Table 1). The large amount of variation in the intensity of Dehnel’s Phenomenon between natural populations is consistent with the great flexibility of skull changes observed in the captive experiments when ambient temperature was manipulated (Lázaro et al. , 2019). Specifically, when analyzing braincase height variation across populations at each age stage, we found a negative correlation of braincase height with longitude and with the interaction of longitude and latitude at both the small subadult and regrown adult stages (P (sub.-long.) < 0.05;P (sub.-long.:lat.) < 0.05; P (ad.-long.) < 0.05; P (ad.-long.:lat.) < 0.05), but no trends along with other variables. This means that braincase height of subadults and adults, but not of juveniles, decreased towards the Northeast. We found a positive correlation in the magnitude of the decrease of braincase height with temperature seasonality, annual temperature range, and precipitation seasonality. We also found a negative correlation with isothermality and mean temperature of the driest quarter (Table 2). This suggests a link between braincase height decline and the intensity of seasonality, with a positive gradient towards more continental climate. This is also supported by the correlations we found between the intensity of Dehnel’s Phenomenon and climate variables associated with seasonality. As it has been postulated that the size decrease stage of Dehnel’s Phenomenon occurs in anticipation of changes in climate and resource variability (which is influenced by climate) this supports the hypothesis that shrews shrink in preparation of winter to lower their energetic needs and thus resource demands.
Even more interesting is then that we did not find a correlation between braincase height regrowth and any geographical variable (Table1). Braincase height regrowth was only positively correlated with precipitation during the warmest quarter (Table 2). It is striking that adult body mass as well as the size of several crucial organs greatly exceed subadult and even juvenile mass, while the size of the brain and skull are largest in juveniles and only partially regrow after the winter decrease (Pucek, 1965). Similarly, energy expenditure is by far the largest in these regrown adults (Schaeffer et al. , 2020) even they are exposed to much higher ambient temperatures than the small winter subadults. The most important investment shrews face during the adult period is reproduction, associated with territory expansion, territorial fights, and massive enlargement of the testes in the males and the production of several large litters in the females (Vlasák, 1996, 1998). The correlation of brain case height with precipitation might then be linked to the availability of high quality food during regrowth and reproduction since the abundance of the main prey of common shrews, the common earthworm (Churchfield, Rychlik, & Taylor, 2012), is highly dependent on soil humidity. For both sexes reproduction appears to be a more or less terminal investment as most individuals die shortly after. Thus, regrowth with such a disproportional investment into mass instead of the brain might be driven by the demands of reproduction. This would mean that decrease and regrowth phases of Dehnel’s phenomenon have evolved under different evolutionary pressures, which are caused and modulated by independent factors. Decrease intensity would then mainly be determined by the physiological limits of shrews, and the regrowth by reproduction. Across all reviewed populations, mean body mass decreased by 21.2 ± 6.2% and regrew by 81.9 ± 18.2%. Similar to braincase height, the extent of body mass changes varied between populations in the analysis of literature data. We found a positive correlation between body mass decrease with longitude, and with the interaction of latitude X longitude, but not with altitude or in this case latitude alone (Fig. 1, Table 1). This matches results from a previous study where no correlation between winter body mass decline and latitude was found in S. araneus either (Ochocińska & Taylor, 2003). Again, matching results of skull measures, body mass regrowth were not correlated with any geographical variable.
When comparing the three size extremes, there was no geographical pattern in juvenile or adult body mass. However, there was a significant negative correlation of body mass of winter subadults with latitude, longitude and their interaction, i.e. winter subadults had lower body mass towards Northeastern populations (P (long.) < 0.05;P (lat.) < 0.001; P (long.:lat.) < 0.01). We found significant correlations between body mass decrease and most climate variables. In contrast, regrowth intensity was only negatively correlated with mean temperature during the driest quarter (the year quarter when precipitation is lowest) and, similar as in braincase height, positively correlated with precipitation seasonality (expressed as coefficient of variation, the more variation the more concentrated the precipitation on a period of the year; Table 2). Again these patterns support the hypothesis that different evolutionary drivers are responsible for the decrease - shrinking as an adaptation to save energy during cold periods with low resource availability, and the increase - growing a large body size well adapted for territory defense and to maximize reproductive output especially in females.
While the changes in body mass we describe are dramatic, seasonal fluctuations in body mass are common in mammals. For example, North American beavers (Castor canadensis ) lose 9-12% of their body mass during autumn and winter, mainly because of the metabolic use of their fat stores (Smith & Jenkins, 1997). During hibernation, marmots can lose 32% of their body mass (Lenihan & Vuren, 1996), and hedgehogs 15-28% (Haigh, O’Riordan, & Butler, 2012). These fluctuations are mainly caused by changes in fat tissues. However, none of these species exhibit the changes in skull and brain size, which are characteristic of Dehnel’s Phenomenon, as well as the changes in spine length (Saure & Hyvärinen, 1965; Hyvarinen, 1969). Thus, changes in body mass are not exclusive of Dehnel’s Phenomenon and should only be used to describe Dehnel’s Phenomenon in combination with other variables, as the combination all these morphological changes – body mass, brain mass, skull size and spine length – are inherent parts of the unique Dehnel’s phenomenon (see also general remarks below).