In Denmark as well as in most other European countries, the amount of land covered by semi-natural grassland has decreased dramatically during the 20^th century concurrent with the general intensification of the agricultural production. To reverse this trend, actions are being taken in many places to either maintain or re-establish this biotope, and in particular, the meadow community. Today's nature conservation is a return to the old extensive agricultural methods, and includes grazing and hay cutting, as well as the abandonment of draining. Semi-natural grassland and in particular meadows constitute important habitat types for a large number of animal species in today's fragmented and intensively cultivated landscape in Europe.

Hay cutting and livestock grazing is known to affect a number of organisms, but the response to grazing may vary across classes of organisms and with the intensity of grazing [[1], and references therein]. The effect of haying and grazing on plant diversity and composition is well-documented [e.g. [2–4]]. Also, many avian species may respond to grazing, and certain grazing intensities may favour some species over others [e.g. [5]]. In contrast to this, only limited data on the response of the mammalian vertebrates to the application of these traditional farming methods is available [but see [6]]. However, meadow management in general may reduce small mammal species richness [7, 8]. Apart from being simple disturbances induced onto the flora and fauna, grazing and haying may change the physical environment, the plant composition and height. This may in turn influence the spatio-temporal distribution of the small mammals, and small mammal biomass has been found to decreases with grazing intensity [8, 9].

The Common shrew (Sorex araneus) exploits a variety of terrestrial as well as semi-terrestrial habitats [10]. Though today's meadow management does not aim specifically at improving Common shrew habitats, the species is one of the most common mammals on meadow communities, and may play an important role in the trophic interactions in this biotope. In the present study, we therefore focus on the Common shrew in order to reveal the impact of three different grazing regimes applied in meadow management on the population characteristics of this species.

During the entire trapping period we caught 570 individual Common shrews. Pygmy shrews S. minutus and Water shrews Neomys fodiens were also caught, but only in small numbers.

In all six trapping plots, the number of individual Common shrews caught in each trapping session showed large fluctuations among trapping sessions as well as inter-annually (Figure 1). The autoregressive component was not significant in the model (P > 0.05), whereas trapping session nested within year was significant (P < 0.05). The number of Common shrew individuals varied significantly among meadows (F_1,4.65 = 28.62, P = 0.0038) and among grazing treatments (F_2,4.65 = 18.76, P = 0.0059). Meadow East held more Common shrews than meadow West, and the high grazing intensity treatment (HIGH) held significantly fewer Common shrews than the ungrazed control (NO GRAZING) (P = 0.018) and the low intensity grazing treatment (LOW) (P = 0.006). Common shrew numbers in LOW and NO GRAZING did not differ significantly (P = 0.401).

During the entire trapping period, the overall sex ratio was close to unity (52% males, 48% females; χ² = 0.2523, P = 0.6154). No correlation between female corrected body mass was found on any of the treatments (F_1,9 = 2.66, P = 0.1543), and no statistically significant differences in the number of foetuses or uterine scars among grazing treatments was found (F_2,6 = 1.12, P = 0.3319; Table 1).

The three grazing treatments applied on the two meadows affected the population dynamics of the Common shrews markedly, and despite the overall differences in Common shrew numbers between the two meadows, high intensity grazing always resulted in significantly lower Common shrew numbers compared to both low intensity grazing and no grazing. Areas with low grazing intensity often held more Common shrews than the ungrazed control, though the overall differences in abundance were not statistically significant (Table 1). The pattern in the Common shrew trappings is very similar to that observed for field voles Microtus agrestis on the very same meadows [11]. Thus, in the low intensity grazing treatment, the potentially negative impacts of grazing livestock, such as mechanical disturbance, were apparently fully compensated for by the positive effects of livestock grazing.

The similarity in the response to grazing intensity amongst two ecologically distinct species, a rodent and an insectivore, points to a common environmental parameter as driver of the population dynamics. Schmidt et al. [11] suggested that for field voles the observed pattern was primarily due to livestock grazing creating a more heterogeneous vegetation in the low intensity grazed areas as compared to areas with high grazing intensity or no grazing, which, in case of the Common shrew, fully compensate for the potentially negative effects of grazing, such as mechanical disturbance. As for the field voles [11], vegetation cover, and, thus, risk of predation, is a probable cause of the grazing treatment effects observed on Common shrews in the two meadows [see also [8]]. Additionally, grazing may affect plant species composition, and livestock trampling may create a more heterogeneous micro-topographic environment, which, in turn, may affect the composition and availability of invertebrates. Increasing grazing intensity is generally believed to be accompanied by decreasing invertebrate abundance and species numbers [e.g. [12]]. Roberts & Morton [13], however, reported that Scarabaeidae biomass peaked an intermediate grazing intensity. Also, invertebrate species richness may benefit from low intensity grazing [14]. Shrews generally adapt rapidly to spatial and temporal changes in prey availability [10], and the observed pattern of Common shrew abundance found in this study may, thus, be attributed to indirect effects of livestock grazing affecting the distribution of Common shrew food.

Unlike for the field voles [11], we found no qualitative differences between Common shrew individuals caught in the three grazing treatments. That is, no differences in body mass, reproductive output, or sex ratio between treatments. The only qualitative difference we found was between individuals from the two meadows studied, and individuals from the western meadow were lighter than individuals from the eastern meadow. This difference in body mass points out meadow West as being sub-optimal compared to meadow East. Generally, only few shrews were caught on this meadow, and generally the populations on meadow West fluctuated more irregularly compared to meadow East. Meadow West was generally more water-logged and flooded more often than meadow East, and may therefore be a less suitable and more unpredictable habitat than meadow East. Shrews are rapid colonisers [15], and Common shrew numbers on meadow West was therefore probably more determined by an unstable alternation between immigration and emigration, whereas the shrews on meadow East belonged to more stable populations. Recapture rates were, however, too low the verify this.

The consistent response of the insectivorous Common shrew and the rodent Field vole [11] to grazing intensity across these two, highly different meadows stresses low intensity livestock grazing as a highly suitable means in today's meadow management, at least in the short run. Long-term changes in meadow vegetation composition etc. induced by grazing livestock may either alter or consolidate the response of both rodents and insectivores to livestock grazing reported here.

Livestock grazing intensity had marked effect on Common shrew numbers, and the highest number of Common shrews was found in the low intensity grazed treatments and the ungrazed controls, while the high intensity grazing treatments held the lowest number of Common shrews. Thus, when grazing is needed as part of the meadow management scheme, low intensity grazing seems suitable for Common shrews and small mammals in general [see [11]].