Mow the Grass at the Mouse’s Peril: Diversity of Small Mammals in Commercial Fruit Farms

Simple Summary For the first time in the Baltic countries, small mammal communities were evaluated in commercial orchards and berry plantations to test for the influence of crop type and intensity of agricultural practices. Out of ten species registered in the fruit farms, the most dominant were common vole and striped field mouse, confirming the spread of the latter species in the country. Small mammal diversity and abundance were not dependent on crop type but decreased in line with the intensity of agricultural practices, not being found in the most intensively cultivated farms. Unexpectedly, small mammal diversity in apple orchards exceeded the diversities found in most types of forests and was significantly higher than in crop fields. Abstract Small mammals are not only pests but also an important part of agricultural ecosystems. The common vole is a reference species for risk assessment of plant protection products in the European Union, but no data about the suitability of the species in the Baltic countries are present so far. Using the snap-trap line method, we evaluated species composition, abundance, and diversity of small mammal communities in commercial orchards and berry plantations in Lithuania, testing the predictions that (i) compared with other habitats, small mammal diversity in fruit farms is low, and (ii) the common vole is the dominant species. The diversity of small mammals was compared with control habitats and the results of investigations in other habitats. Out of ten small mammal species registered, the most dominant were common vole and striped field mouse. Small mammal diversity and abundance increased in autumn and decreased in line with the intensity of agricultural practices but were not dependent on crop type. In the most intensively cultivated fruit farms, small mammals were not found. The diversity of small mammal communities in fruit farms was significantly higher than in crop fields and exceeded the diversities found in most types of forests except those in rapid succession.


Introduction
In 2017, there were 25,980 ha of commercial fruit farms in Lithuania (including 9851 ha of berry plantations), which yielded 99,215 tons of production [1]. However, despite commercial fruit farms being an important agricultural industry, not all aspects of their ecology have been subjected to scientific research in Lithuania. We present the first results of investigations into small mammals in commercial orchards and berry plantations in the Baltic countries.
Being a part of the food web and providing ecosystem functions in agricultural ecosystems [2], small mammals also are pests of many crops [3,4]. Their negative impact is not limited to crop damage [5] but also includes their role in the dispersal of weed seeds and being a reservoir for various Table 1. Characteristics of study sites. Site numbers as in Figure 1.

Small Mammal Trapping
Small mammals were trapped using 7 × 14 cm wooden snap traps set in lines of 25 traps, each set 5 m apart as in [37]. The number of lines per site depended on the size of the fruit farm, ranging from one to four. Despite the warning of [38,39] about the effects of removal trapping of small mammals on abundance and diversity, our data show no influence of short trap lines on the diversity and abundance of small mammals, even after long−term trapping of up to three times per year [40][41][42][43]. In this study, we trapped two times. Several sites had no small mammals in the first trapping session; thus, the first trapping could not have negatively influenced the second trapping session. Traps were baited with bread crust with raw sunflower oil, set for three days, and checked once per day. According to [37], the bait should be changed after rain or when consumed by other mammals, birds, insects, or slugs. The total trapping effort was 8880 trap days ( Table 2). The relative abundance of small mammals was expressed as standard capture rates to number of animals/100 trap days of the first day of trapping. As there was no rain on the first day of trapping, correction of abundance values for sprung traps was not used; at sites 13 and 15, no small mammals were trapped in summer and autumn. Location of the study sites in Lithuania: 1-Aukštikalniai, 2-Pajiešmeniai, 3-Gaižiūnai, 4-Užpaliai, 5-Kalpokai, 6-Šeimyniškiai, 7-Šedbarai, 8-Šiekštinės, 9-Dembava, 10-Užubaliai, 11-Barčiai, 12-Luksnėnai, 13-Naujasis Obelynas, 14-Gaurė, 15-Kvėdarna. Crop type is indicated by color: green-apple orchards, brown-plum orchards, magenta-raspberry plantations, yellow-currant plantations, and blue-highbush blueberry plantations. Location coordinates are not presented to ensure privacy of the owners.

Small Mammal Trapping
Small mammals were trapped using 7 × 14 cm wooden snap traps set in lines of 25 traps, each set 5 m apart as in [37]. The number of lines per site depended on the size of the fruit farm, ranging from one to four. Despite the warning of [38,39] about the effects of removal trapping of small mammals on abundance and diversity, our data show no influence of short trap lines on the diversity and abundance of small mammals, even after long-term trapping of up to three times per year [40][41][42][43]. In this study, we trapped two times. Several sites had no small mammals in the first trapping session; thus, the first trapping could not have negatively influenced the second trapping session. Traps were baited with bread crust with raw sunflower oil, set for three days, and checked once per day. According to [37], the bait should be changed after rain or when consumed by other mammals, birds, insects, or slugs. The total trapping effort was 8880 trap days ( Table 2). The relative abundance of small mammals was expressed as standard capture rates to number of animals/100 trap days of the first day of trapping. As there was no rain on the first day of trapping, correction of abundance values for sprung traps was not used; at sites 13 and 15, no small mammals were trapped in summer and autumn. Table 2. Summary of trapping results in 2018 in the commercial fruit farms and control habitats.
Standard measures (body mass, body length, tail length, hind foot length, and ear length) were taken before dissection. Species were identified morphologically, with specimens of Microtus voles identified by their teeth. Juveniles, subadults, and adults were identified under dissection based on body weight, the status of sex organs, and atrophy of the thymus, the latter of which decreases with animal age [44].
We used snap trapping to obtain material for the analysis of chemical elements in the bodies of the small mammals and to search for various pathogens (data not used in the current publication). The study was conducted in accordance with the principles of Lithuanian legislation for animal welfare and wildlife. Permission to trap wild small mammals was provided according to Regulation No. 6 (2018-02-02) of the Ministry of the Environment of the Republic of Lithuania.

Data Analysis
Species richness was expressed as the number of trapped species, while the Shannon-Wiener diversity index, H, on the base of log 2 [45] was used as a measure of the diversity of the small mammal community and the dominance index, D, as a measure of dominance [19,45,46]. The diversity of the community was compared to other habitats and territories of varying sizes in Lithuania ( [40][41][42][43]47,48]; Appendix B Table A1). Diversity estimates (including bootstrap, yielding 95% confidence interval) were calculated using PAST ver. 2.17c (Ø. Hammer, D.A.T. Harper, Oslo, Norway) [46,49] and EstimateS ver. 9.1.0 (R.K. Colwell, Connecticut, CT, USA) [50,51]. Rarefaction was used according to individual-based data to produce species accumulation curves and compare species richness between fruit farms, berry fields, and control habitats and between fruit farms with different intensities of treatment. Rarefaction eliminates the influence of the sample size on the diversity estimate (number of individuals or trapping effort, thus the effect of the farm size at the same time). Differences of community composition were evaluated using chi-square statistics in PAST.
We applied factorial ANOVA to relative small mammal abundance, diversity, and dominance as dependent variables across the research sites to test the possible cumulative influence of categorical predictors-season, crop type, and intensity of agricultural practices. Hotelling's T 2 was used for multivariate testing. Thereafter, significant factors were used for ANOVA analysis, and Tukey's HSD with unequal N for post hoc comparisons [52]. The confidence level was set as p < 0.05. Calculations were done in Statistica for Windows, ver. 6.0 (StatSoft, Inc., Tulsa, OK, USA).

Results
We trapped 512 individuals of 10 small mammal species (Table 3). Overall, orchards and plantations were inhabited by all recorded species but were dominated by common vole (M. arvalis) (31.0%), striped field mouse (Apodemus agrarius) (27.1%), and yellow-necked mouse (Apodemus flavicollis) (19.5%). Control habitats were characterized by nine species, with pygmy shrew (Sorex minutus) not trapped, and these were dominated by A. agrarius (36.8%), M. arvalis (18.7%), A. flavicollis, and bank vole (Myodes glareolus) (both 13.9%). Species composition significantly differed between the fruit farms and controls (χ 2 = 25.7, df = 9, p < 0.01), the main difference being higher numbers of M. arvalis in the former and higher numbers A. agrarius in the latter habitats. The diversity of small mammals in orchards and plantations was the same as in control habitats, with dominance also not differing ( Figure A5A). No significant differences in diversity indices between the control habitats and commercial fruit farms were found in the separate seasons of summer ( Figure A5B) and autumn ( Figure A5C). In autumn, however, species accumulation curves showed that a higher diversity was reached in control habitats with a smaller number of trapped specimens.

Small Mammal Diversity and Abundance in Relation to Crop Type
Small mammal species richness was highest in the apple orchards and their control habitats, then declined in the following order: raspberry plantations and their controls, currant plantations, currant controls, and plum orchards and their control habitats. No small mammals were trapped at all in the highbush blueberry plantation (Table 3). Thus, our first prediction was not confirmed.
Small mammal diversity (Table 3) in the apple orchards was higher than in plum orchards, currant plantations, and raspberry plantations. Plum orchards had the same small mammal diversity as currant and raspberry plantations, while diversity in raspberry plantations exceeded that in currant plantations. Species accumulation curves ( Figure 2A) fully confirmed the diversity differences. Differences in the small mammal diversity between crops and respective control habitats were not found. Animals 2019, 9 6 of 17 autumn ( Figure A5C). In autumn, however, species accumulation curves showed that a higher diversity was reached in control habitats with a smaller number of trapped specimens. Testing the cumulative influence of season, crop type, and intensity of agricultural practices on relative small mammal abundance, diversity, and dominance showed a significant dependence on season (Hotelling's T 2 = 0.353, df = 3, p = 0.001) and intensity (T 2 = 0.316, df = 6, p = 0.02) but not habitat type (T 2 = 0.207, df = 12, p = 0.59). Univariate tests showed a significant influence of season (ANOVA, F = 17.01, df = 1, p < 0.001) and intensity of agricultural practices (F = 4.09, df = 2, p < 0.05) on the relative abundance of small mammals, as well as the influence of season (F = 9.65, df = 1, p < 0.01) on diversity and the influence of intensity of agricultural practices (F = 3.78, df = 2, p < 0.05) on dominance in small mammal communities.

Small Mammal Diversity and Abundance in Relation to Crop Type
Small mammal species richness was highest in the apple orchards and their control habitats, then declined in the following order: raspberry plantations and their controls, currant plantations, currant controls, and plum orchards and their control habitats. No small mammals were trapped at all in the highbush blueberry plantation (Table 3). Thus, our first prediction was not confirmed.
Small mammal diversity (Table 3) in the apple orchards was higher than in plum orchards, currant plantations, and raspberry plantations. Plum orchards had the same small mammal diversity as currant and raspberry plantations, while diversity in raspberry plantations exceeded that in currant plantations. Species accumulation curves ( Figure 2A) fully confirmed the diversity differences. Differences in the small mammal diversity between crops and respective control habitats were not found. Dominance in the small mammal community was most notably expressed in the currant plantations (Table 3), where M. arvalis constituted nearly 70% of all trapped individuals (64% in the control habitats). This species also dominated in plum orchards (46%), whereas A. agrarius dominated in raspberry plantations (48%). Control habitats of the raspberry plantations were dominated by A. flavicollis (33%), A. agrarius (27%), and M. glareolus (24%). Four small mammal species were represented almost equally in apple orchards ( Figure 2B). Control habitats of the apple and plum orchards were dominated by A. agrarius (over 40% of all individuals in both controls). Other small mammal species were not trapped in large numbers. Thus, our second prediction about the dominance of M. arvalis was confirmed but not for all habitats. Dominance in the small mammal community was most notably expressed in the currant plantations (Table 3), where M. arvalis constituted nearly 70% of all trapped individuals (64% in the control habitats). This species also dominated in plum orchards (46%), whereas A. agrarius dominated in raspberry plantations (48%). Control habitats of the raspberry plantations were dominated by A. flavicollis (33%), A. agrarius (27%), and M. glareolus (24%). Four small mammal species were represented almost equally in apple orchards ( Figure 2B). Control habitats of the apple and plum orchards were dominated by A. agrarius (over 40% of all individuals in both controls). Other small mammal species were not trapped in large numbers. Thus, our second prediction about the dominance of M. arvalis was confirmed but not for all habitats. The relative abundance of small mammals varied across the study sites and ranged between 0 and 12 individuals per 100 traps/day in control habitats and commercial fruit farms in summer. In June, no small mammals were trapped in fruit farms at seven sites (Nos. 1, 2, 5, 11, and 13-15, representing all crops). In the autumn, the relative abundances were 0-36 in commercial fruit farms and 0-30.7 individuals per 100 traps/day in control habitats. Fruit farms at three sites (Nos. 5, 10, and 13) were not inhabited by small mammals in the autumn, two of which were unoccupied in both seasons. The averages of the relative abundance of small mammals between the respective controls and fruit farms in summer and autumn did not differ (Table 3).

Small Mammal Diversity and Abundance in Relation to Intensity of Agricultural Practices
The number of small mammal species, recorded in the sites with medium and high intensity of agricultural practices, was lower than in respective control habitats. The small mammal diversity in the fruit farms with high-intensity practices was the lowest (Table 4), differing from the respective control habitats and fruit farms with medium-and low-intensity practices. Low-and medium-intensity practices had similar effects on small mammal diversity. In fruit farms with medium intensity of agricultural practices, it did not differ from control habitats. Species accumulation curves ( Figure 3A) confirmed that the highest diversity of small mammals was in control habitats and habitats with medium intensity of disturbance. Table 4. Small mammal diversity, numbers trapped, and relative abundance (recalculated to 100 traps per day) in fruit farms according to the intensity of agricultural practices. Control data presented in Table 3.

Species
Intensity of Agricultural  The dominance index was highest in small mammal communities living under the highest pressure of agricultural practices (Table 4). In these fruit farms, three small mammal species-M. arvalis, A. flavicollis, and A. agrarius-constituted over 95% of all trapped individuals, while respective control habitats were dominated by A. agrarius, representing 47% of all trapped individuals ( Figure 3B). Over 85% of all trapped individuals in fruit farms with low pressure were represented by A. agrarius, M. arvalis, and M. glareolus, with the same species dominating respective control habitats. Habitats  (Table 4). The dominance index was highest in small mammal communities living under the highest pressure of agricultural practices (Table 4). In these fruit farms, three small mammal species-M. arvalis, A. flavicollis, and A. agrarius-constituted over 95% of all trapped individuals, while respective control habitats were dominated by A. agrarius, representing 47% of all trapped individuals ( Figure  3B). Over 85% of all trapped individuals in fruit farms with low pressure were represented by A. agrarius, M. arvalis, and M. glareolus, with the same species dominating respective control habitats. Habitats with medium pressure of agricultural activities were dominated by M. arvalis, and their respective control habitats by A. agrarius ( Table 4).
The average relative abundances of small mammals in summer and autumn were highest in fruit farms with low intensities of agricultural practices (Table 4), exceeding those in sites with medium intensity by about 4 times and those in sites with high disturbance by nearly 10 times in summer and 2-4 times in autumn. At the beginning of summer, no small mammals were trapped in four fruit farms with high intensity (Nos. 2, 11, 13, and 15) and three fruit farms with medium intensity (Nos. 1, 5, and 14) of agricultural practices. In autumn, the number of farms with no small mammals was two (Nos. 10 and 13) and one (No. 5), respectively. Small mammals were always present at all sites with a low intensity of measures. Thus, intensity had a negative effect on small mammal presence.
The intensities of agricultural practices were also well reflected by the variation of relative abundance of small mammals across sites. In summer, relative abundances were 0.3-12 ind. in fruit farms with a low intensity of influence, 0-4 ind. in medium intensity, and only 0-1.1 ind. in the highly affected sites. In the autumn, the respective figures were 4-36, 0-9, and 0-28 ind. The averages of the relative abundance of small mammals between the respective controls and farms with different intensity of agricultural practices in summer and autumn did not differ (Table 4).

Discussion
As recognized in the European Common Agricultural Policy Reform [53], one of the dominant threats to biological diversity is intensive agriculture, resulting in a decline of small mammal populations in Europe [54]. Thus, even while recognizing negative small mammal impacts such as crop damage and monetary losses [3,7], the current position is to maintain their communities, even in agricultural ecosystems, as providers of ecosystem services [2]. Therefore, chemical−based pest management methods are being changed to those based on small mammal ecology [55] or biological control [29].
Knowledge regarding the species and their habitats should assist in the development of new rodent pest management strategies [56]. In disturbed habitats with varying agricultural practices, The average relative abundances of small mammals in summer and autumn were highest in fruit farms with low intensities of agricultural practices (Table 4), exceeding those in sites with medium intensity by about 4 times and those in sites with high disturbance by nearly 10 times in summer and 2-4 times in autumn. At the beginning of summer, no small mammals were trapped in four fruit farms with high intensity (Nos. 2, 11, 13, and 15) and three fruit farms with medium intensity (Nos. 1, 5, and 14) of agricultural practices. In autumn, the number of farms with no small mammals was two (Nos. 10 and 13) and one (No. 5), respectively. Small mammals were always present at all sites with a low intensity of measures. Thus, intensity had a negative effect on small mammal presence.
The intensities of agricultural practices were also well reflected by the variation of relative abundance of small mammals across sites. In summer, relative abundances were 0.3-12 ind. in fruit farms with a low intensity of influence, 0-4 ind. in medium intensity, and only 0-1.1 ind. in the highly affected sites. In the autumn, the respective figures were 4-36, 0-9, and 0-28 ind. The averages of the relative abundance of small mammals between the respective controls and farms with different intensity of agricultural practices in summer and autumn did not differ (Table 4).

Discussion
As recognized in the European Common Agricultural Policy Reform [53], one of the dominant threats to biological diversity is intensive agriculture, resulting in a decline of small mammal populations in Europe [54]. Thus, even while recognizing negative small mammal impacts such as crop damage and monetary losses [3,7], the current position is to maintain their communities, even in agricultural ecosystems, as providers of ecosystem services [2]. Therefore, chemical-based pest management methods are being changed to those based on small mammal ecology [55] or biological control [29].
Knowledge regarding the species and their habitats should assist in the development of new rodent pest management strategies [56]. In disturbed habitats with varying agricultural practices, species ecology will be different [53]. Knowledge of the driving factors in small mammal ecology would allow harmonization of conservation needs and economic necessities [18].
We found that small mammal diversity and relative abundance were mainly defined by season and intensity of agricultural practices but not by crop type. In our study, soil scarification of interlines was used in 25% of investigated sites, while rodenticides in approximately 40%, plant protection agents in nearly 75%, and grass mowing at all sites. In the current analysis, usage of rodenticides was treated as any other agricultural activity and not analyzed separately. The application of rodenticides, however, had only a limited effect, as in the sites where rodenticides were used, three species of small mammals were recorded in spring and six species in autumn.
The main agricultural activity in our study was grass mowing. Depending on whether the mowed grass was removed from the fruit farm, used as mulch, or left in place, the impact of mowing on small mammals was different. Small mammal numbers decline rapidly after mowing, but up to 27% of small mammals remain if the cut grass is not moved [57]. Habitats that are not mown every year and ecological compensation areas in agricultural landscape support the highest numbers of small mammals [54]. Rodents respond to decreased vegetation height by reducing their activity and changing their feeding pattern [15]. Unmowed grasslands are characterized by lower giving-up densities and differences in the use of territory; they are less used at night, possibly accounting for owl activity [58]. Using the analogy of understory, we may expect changes in reproduction parameters [59]. Finally, decreased mowing intensity after the crop season in commercial fruit farms may facilitate easier recolonization of the territory from surrounding habitats. In our study, the example is the migration of A. flavicollis into the raspberry plantation at site 11 from the adjacent forest in autumn.
In Lithuania, two dominant species, M. arvalis and A. agrarius, were found in most of the investigated commercial fruit farms. The exception was the old apple orchard at site 7 ( Figure A3d) with high grass cover and almost no agricultural activities, which was dominated by M. glareolus. M. arvalis is one of the most numerous species in Central Europe [11], which may consume nearly 50% of the annual production of alfalfa during outbreaks [60]. An expansion of A. agrarius has been recorded in other European countries [61], and this requires special attention in Lithuania [43].
Based on our long-term small mammal studies [40][41][42][43][44]47,48], we suppose that the number of trapping sites (spatial replications) was enough to represent possible temporal changes in small mammal diversity and abundance (i.e., the results are representative). Earlier, different small mammal densities and number of species were found within the same year in forests, swamps, or meadows located several kilometers from each other [41,42,47,48].
In commercial fruit farms, our results indicate a higher species richness and an equal diversity and dominance in small mammal communities when compared with control habitats but with different species composition and relative abundance.
We compared our results with other agricultural and natural habitats across the country using published data from 125 trapping sessions (Table A1), which represent long-term trapping [41], forest succession [42], flooded and nonflooded habitats [44], various habitats in protected territories [47], and data from small mammal monitoring in the agro-landscape [48].
We found that the small mammal diversity in our control meadows was in line with that in natural, seminatural, and sown meadows evaluated in the long-term investigations, lower than in shrubby or reforesting meadows but much higher than in meadows of the agro-landscape (all mentioned differences significant with p < 0.001). The small mammal diversity that we found in apple orchards was significantly higher than that recorded in various forest types, with the exception of young forest growing in former meadows and regrowing after forest clear-cutting, where the succession between ecosystem stages shapes the small mammal community [42,47]. Small mammal diversity in the plum orchards was in line with most forest types, while small mammal diversity in the berry plantations was significantly lower than in meadows and in line or exceeding those in cropland, forests, and other habitats.

Conclusions
Out of ten small mammal species registered in commercial fruit farms in Lithuania, the most dominant were M. arvalis and A. agrarius. The diversity of small mammal communities in the farms was higher than in crop fields, while the diversity in the fruit orchards exceeded or was in line with that of the forests. The diversity in berry plantations was in line or exceeded cropland and forests but