, 98:863 | Cite as

Food preferences and mound-building behaviour of the mound-building mice Mus spicilegus

  • Michaela Hölzl
  • Ján Krištofík
  • Alžbeta Darolová
  • Herbert Hoi
Original Paper


Optimal foraging strategies and food choice are influenced by various factors, e.g. availability, size and caloric content of the food type and predation risk. However, food choice criteria may change when food is not eaten immediately but has to be carried to a storage site for later use. For example, handling time in terms of harvesting and transport time should be optimized, particularly when the risk of predation is high. Thus, it is not clear whether food selected by hoarding animals reflects their food preference due to intrinsic features of the food type, e.g. size, caloric or lipid content, or whether the food type selected is a compromise that also considers the handling time required for harvesting and transport. We investigate this question in relation to food hoarding behaviour in mound-building mice. In autumn, mound-building mice Mus spicilegus collect seeds and other plant material and cover it with soil. Such above-ground storage is quite unusual for rodents. Here, we investigated whether there is a relationship between the seed species preferred as building materials and those preferred for food. We conducted a seed preference test using three most collected weed species for mound building. Controlling factors like food availability or predation risk, mice prefer Setaria spp. as food, although Amaranthus spp. and Chenopodium spp. were preferentially harvested and stored. By including the availability of the three species, our experimental results were confirmed, namely, a clear preference for Setaria spp. Also, handling time and seed size revealed to influence plant choice.


Mus spicilegus Seed preference Food storage Optimal foraging 


Optimal foraging strategies and food choice are influenced by various factors such as the availability, size and caloric content of the food (Brewer 2001; Bradford et al. 2008; Lobo et al. 2009), predation risk (Jackson 2001) and even learning processes (Muñoz and Bonal 2008a, b).

The importance of these factors may change, and other factors may also be important when food is not eaten immediately but has to be carried to a storage site for later use. Handling time in terms of harvesting and transport time should be optimized, particularly when the risk of predation is high (Jackson 2001; Chang et al. 2010). In fact, the handling time for a food item has a significant influence on the hoarding strategies of hoarding animals, e.g. food items with a longer handling time may be more likely to be hoarded (Chang et al. 2010). The question is whether food selected by hoarding animals reflects their food preference due to intrinsic features of the food type, e.g. food size, caloric or lipid content, or whether the food type selected reflects a compromise that takes into account the handling time for harvesting and transport. We investigated this question in relation to the food hoarding behaviour of mound-building mice. Food hoarding behaviour is quite common among rodents and is a response to temporally variable and unpredictable food supplies (Vander Wall 1990). Rodents such as the common vole Microtus arvalis (Stein 1958) and the European hamster Cricetus cricetus (Grulich 1981) usually hoard food in chambers or tunnels below ground. Only a few species, such as mound-building mice and the banner-tailed kangaroo rat Dipdomys spectabilis, hoard seeds above ground (Festetics 1961; Reichman et al. 1985; Best et al. 1988; Unterholzner et al. 2000).

The mound-building mouse Mus spicilegus is closely related to the house mouse Mus musculus. These two species often occur sympatrically. In summer, they can be found in the same habitat, but differences in habitat use become quite obvious during winter. In winter, house mice usually live in or near human buildings (Bronson 1979; Carlsen 1993), whereas mound-building mice create their own homes, usually in or near agricultural fields. In autumn, they begin to construct mounds, which can contain up to 700 l of stored plant material (Hölzl et al. 2009) and are covered with soil (Muriaru 1981; Garza et al. 1997; Unterholzner et al. 2000; Poteaux et al. 2008). When the mound is half covered, they dig tunnels underneath it in which they overwinter (Unterholzner et al. 2000). Adult individuals do not usually survive the winter, and juveniles, who are born late in summer, spend the cold period in the nest chambers underneath the mound in large groups without reproducing (Gouat et al. 2003a). In spring, the mice leave the mounds and disperse (Muriaru 1981; Gouat et al. 2003b; Simeonovska-Nikolova 2007). Earlier studies assumed that the mounds constructed by mound-building mice were built for food storage (Festetics 1961; Unterholzner et al. 2000).

Studies have since revealed that mice use an area of up to 140 m2 around the mound when searching for food (Sokolov et al. 1998). However, the area from which mound-building mice collect plant material varies with the abundance of desirable seeds. When the necessary plants are available, the near surroundings of the mound are usually completely harvested, suggesting that they try to minimize travel costs. This behaviour further requires above-ground activity, which is usually performed at dawn and throughout the night according to our video monitoring of mound-building individuals in enclosures (unpublished results). Thus, mound building does not only involve the energetic costs of harvesting and transport but also a high risk of predation.

In an earlier study (Hölzl et al. 2009), we showed that mice usually collected inflorescences of 16 of the 56 different plant species identified in the surrounding mounds (within a radius of 3 m). The three most dominant plant species collected were Amaranthus spp. (37.2%), Chenopodium spp. (27.5%) and Setaria spp. (8.2%).

Whether or not the mice eat the plant material (seeds) stored in the mound has not been demonstrated, and a more recent study even suggested that the seeds have a thermoregulatory function (Bihari 2004). These functions are not mutually exclusive, although we found no strong support for a thermoregulatory function in a recent study (Hölzl et al. 2011). Nevertheless, the choice of the building material could be a trade-off between different functions. In this study, we investigated whether mound-building mice show a preference for a specific food type. If the choice of plant material collected reflects food preferences, we predicted that they would prefer Amaranthus spp. over Chenopodium spp. and Setaria spp. Therefore, we conducted seed preference tests to clarify whether or not a food preference exists and, if so, whether this preference (1) reflects the plant species harvested during the building of the mound, (2) might be due to foraging efficiency in terms of seed processing time and food features, for example, seed size, caloric or lipid content, and (3) reflects the handling time in terms of time required to transport inflorescences to the mound because mice usually carry complete inflorescences rather than individual seeds.

Materials and methods

The mice that were used for this experiment came from a harvested grain field located in Bohelov (Western Slovakia, 47°54′25.72″ N, 17°41′ 58.45″ E). The mice were caught using live traps placed around mounds in September. The mice used in the experiments originated from ten different mounds whereby members of the same mound were used in each experiment. The average body mass of experimental mice measured before the start of each experiment was 9.8 ± 0.45 g, ranging from 7.1 to 15.6 g. No individual died after the experiment. The average body mass of the experimental mice determined later, at the onset of the following breeding season (mid April), was 15.7 ± 0.5 g, ranging from 13.1 to 18.5 g (for comparing body mass with a nearby field population, see Unterholzner et al. 2000). As we did not expect sex-specific seed preferences, gender composition was not determined for the seed preference experiment.

Before the experiments started, mice were fed with a seed mixture containing the same amount of the three most important plant species used for building. The dominant plant material in the mounds consisted of the three annual weeds Amaranthus spp., Cheopodium spp. and Setaria spp.

The experimental design consisted of an experimental box made of glass (50 × 50 × 30 cm) with three mice inside. As they are social animals, we kept them in groups in order to avoid stress and abnormal results. Prior to the start of the experiment, the mice were kept inside the experimental box for 2 h to acclimatize, without food but with water ad libitum, a shelter (cardboard roll) in each corner and commercial mouse bedding covering the floor. The experiment started by simultaneously offering the mice with the same quantity of seeds, namely, 5 g of Amaranthus spp., Cheopodium spp. and Setaria spp. seeds in separate food dishes (glass Petri dishes with a diameter of 6 cm). All three dishes were placed in a triangle at the centre of the box. Opposite each food dish, at equal distances, we offered identical shelters (paper roll) so that the mice had equal access to the different food dishes. To reduce the possible effect of position, we used new shelters for each experiment and, in addition, changed the position of the offered seeds. After each trial, we cleaned the box and the food dishes and used new paper rolls. The experiments were performed in October, coinciding with the main building period in the field (Unterholzner et al. 2000).

Thirty trials were performed in total, each with three different mice originating from the same mound. Prior to the experiment, we performed behavioural observations using video recordings. They revealed that foraging activity occurs in bouts also over the whole daylight period. Based on that information, food preference tests were conducted at the same time each morning between 07:30 to 09:30, and the mice were monitored for 30 min. The observations showed that foraging activity declined after 30 min. Every day, three different, visually separated groups (boxes) were tested under semi-natural conditions, namely, in a large outdoor enclosure (4 × 4 × 2.5 m) sheltered from the rain. The weather conditions were noted because these mice are less active in bad conditions (rain, storm, etc.).

The dependent variable was the amount of seeds eaten from each food dish; therefore, the seeds were weighed before and after the experiment (weighing accuracy 0.01 g), and the number of visits to the food zone and the duration of feeding at the food dishes were observed (based on video recordings). This also allowed us to see whether any seeds were accidently scattered outside the food dishes or carried to other places (e.g. shelters) in the box. On some occasions, food was scattered around the food dishes, and this was collected after each experiment and considered as not consumed. We did not observe mice carrying seeds to other places in the experimental box, and we found no indication that seeds were only half eaten.

In the second experiment, the weed species that was least preferred in the first experiment was replaced by proso millet Panicum miliaceum, a novel unfamiliar food for our mice since it did not occur at the field site from which they were collected. Furthermore, proso millet has much larger seeds (by ten times), and hence, we predicted that their handling and transport times would be much shorter. In contrast to the three most preferred plant seeds, the calorific content of proso millet seeds is lower. Seeds of all plant species were collected at their original study site except proso millet, which was obtained from a pet shop.

Thirty trials using the same mice as in the first experiment were performed in exactly the same way as described for the first experiment.

According to the optimal foraging theory, animals prefer food items with the highest ratio of energy to handling time (Krebs 1978; McCleery 1978). Therefore, we measured several parameters as an indicator of handling time, defined as the processing time from harvesting to swallowing a particular food item. In mound-building mice, there are two processing components that are temporally separated.

One component of handling time is the gathering time, which includes the time taken to harvest and transport the vegetal material to the mound and may reflect how much the mice can carry in one trip. It is almost impossible to measure this directly, so we measured mound growth as an indirect measure. Piling up the material usually takes about 4 weeks (own unpublished observation). After that, the plant material is covered with a layer of soil (Unterholzner et al. 2000). Even when mounds look finished, mice bring soil parts, sometimes for several more months (own unpublished observations). However, the time to complete mounds depends on mound size and plant availability (Hölzl et al. 2009). Separating the building period into four successive 7-day periods revealed a high repeatability (r = 0.78, df = 3, 29) in mound growth and the largest increase during the second period (unpublished data). Therefore, for this analysis, we used the mound volume estimated 1 week after the start of building and 7 days later and calculated the daily growth rate (cm3 of seed material) as \( \left( {{\text{mound volume 2}} - {\text{mound volume 1}}} \right)/{7} \). Mound measurements were performed in October, the main building period (Unterholzner et al. 2000). To calculate mound volume, we measured the length, width and height (in cm) with a measuring tape. Mound volume was then calculated as a pyramid \( \left[ {{\text{length}}\left( {\text{cm}} \right) \times {\text{width}}\left( {\text{cm}} \right) \times {\text{height}}\left( {\text{cm}} \right)} \right]/{3} \) (Hölzl et al. 2009).

We then used mound growth to examine whether handling time in terms of transport efficiency varies in relation to the different plant species used for building. We also examined the role of plant material in final mound size since differences in handling and transport characteristics could affect this. To determine the mound content, one person (MH) estimated the percentage composition of stored seeds during the mound-building period (approximately 4 weeks after the start of building).

Plant availability in the surroundings is known to affect mound size and consequently may also affect mound growth rate. For this reason, when examining the effect of plant type, we controlled plant availability (Hölzl et al. 2009). Therefore, the percentage coverage of the three plant species was estimated by one person in a radius of 3 m around the mound using the coverage scale based on the methods of Braun-Blanquet (1964). As we did not know the number of inhabitants in any of the mounds and since in an earlier study we showed that the number of inhabitants is independent of mound size (Hölzl et al. 2009), we did not include this variable in the model.

We determined 18 mounds as being made predominantly of Amaranthus spp., with ten Chenopodium spp. and five Setaria spp. The mounds are usually built using a predominant plant species, and we referred to a mound after the predominant plant species when more than 80% of it was composed of one plant species. Plant material composition was determined during the same building period as mound growth was determined (October).

The second component of the handling time is the processing time during eating. Since mound-building mice probably use the plant material as a food store for many months (Unterholzner et al. 2000), the eating process probably takes place long after building the mound. In an attempt to show this, we calculated the seed processing time as part of the handling time from our food preference experiments. We therefore determined (1) the time the mice spent eating at the food dishes based on behavioural observations during the experiment (video recordings) and (2) the amount of seeds eaten over time (mg/min).

Statistical analysis

To determine whether mice showed a significant preference for certain seed types, we used a mixed model analysis. Using this method, we analysed the outcome of the two experiments separately using the box rather than individuals as the sample unit. The amount of seeds eaten per box (three individuals) was entered as the dependent variable. Date was included to examine the effect of time and varying weather conditions on the different experimental days (n = 5). Seed species and date were included as fixed factors and box as a random factor. To examine a possible novelty effect of the unfamiliar seed type (proso millet), we compared the change in the time mice spent foraging in the first and second 5-min intervals between the unfamiliar and familiar and naturally preferred seed types (Setaria spp.) using a paired t-test.

Mound growth was analysed for the mounds in which we determined a predominant plant species. In order to examine the variation in mound growth, we used a GLM, with the daily increase in mound volume as the dependent variable, predominant plant species as the factor and plant availability within a distance of 3 m of each mound as the covariate. A stepwise regression model was used to examine the effect of the plant species used on the final mound volume. A one-way ANOVA was used to determine the effect of plant species collected on (1) the amount of seeds eaten per minute, (2) mound growth and (3) final mound size.

Repeatability analyses were used to examine variation between and the consistency of individual mounds in terms of mound-building investment over the main building period (1 month). Repeatability was therefore calculated for four successive 7-day periods following Lessells and Boag (1987).

Since the data were normally distributed, parametric tests were used throughout. To investigate normality of the data, we used Shapiro–Wilk tests.


Seed species had a significant effect on the amount of seeds consumed (mixed model analysis; F2,28 = 7.84, P = 0.002), whereas day (F4,28 = 1.26, P = 0.31) and the interaction between seed species and day (F8,28 = 0.94, P = 0.5) had no effect. We observed no difference between the experimental boxes (Wald Z = 0.71, P = 0.48). Post-hoc pairwise comparisons based on estimated marginal means showed that in the first experiment, the mice mostly consumed Setaria spp. seeds and less frequently Amaranthus spp. seeds; however, this difference was not statistically significant (P = 0.11, see also Fig. 1a). Finally Chenopodium spp. was taken significantly less often than Setaria spp. (P < 0.001) and Amaranthus spp. (P = 0.029).
Fig. 1

The results of a experiment 1 showing the mean number of grams ± SE eaten within 20 min of the three seed species that were mainly preferred for mound building and b experiment 2, in which the least preferred seed of experiment 1 was replaced with a seed species that was not available in the natural habitat of the mice. Results are based on 30 experiments with three mice each

The second experiment again showed that seed species had a significant effect on the amount of seeds consumed (mixed model analysis; F2,28 = 8.83, P = 0.001), whereas day (F4,28 = 0.35, P = 0.84) and the interaction between seed species and day (F8,28 = 0.86, P = 0.56) had no effect. We observed no differences between the experimental boxes (Wald Z = 0.67, P = 0.5). Post-hoc multiple pairwise comparisons based on estimated marginal means showed that in the second experiment, the mice consumed more Panicum miliaceum than Amaranthus spp. (P = 0.001) or Setaria spp. (P = 0.001), whereas we found no difference in the consumption of Setaria spp. and Amaranthus spp. seeds (P > 0.8, Fig. 1b). A possible novelty effect of the new unfamiliar seed type proso millet is not indicated since we found no difference when comparing the change in the time mice spent foraging in the first and second 5-min intervals between the unfamiliar proso millet and the familiar Setaria spp. seeds (paired t-test: t = 0.17, P > 0.8, n = 30).

By examining the amount of seeds that the mice ate per minute, we again found a significant difference between plant species (ANOVA: F = 16.87, df = 2, 95, P < 0.001). The intake rate (grams of seeds eaten per minute) was highest for Setaria spp. seeds and lowest for Chenopodium spp. seeds (Fig. 2). The intake rate of proso millet was similar to that of Setaria spp. (10.83 ± 3.8 mg).
Fig. 2

Mean ± SE intake rate for the three plant species calculated as the amount of seeds eaten per minute (given as mg/min). Results are based on 30 experiments with three mice each

Our results show that the residual increase in mound size (building speed) varied significantly according to the plant species used for building (ANOVA: F = 5.3, df = 2, 31, P = 0.011). Post-hoc multiple comparisons revealed that this difference was mainly due to Setaria spp. (Fig. 3). Using Setaria spp., the daily intake rate was around five times higher than that for Amaranthus spp. (P = 0.005) or Chenopodium spp. (P < 0.004). Our results further indicated that the plant species used also influenced the size of the mound. In a multiple stepwise regression, Setaria spp. and Chenopodium spp. entered the model at a significant level (stepwise regression model: F = 7.8, df = 2, 43, P < 0.001, R2 = 0.3). The partial regression coefficient suggested that the more Setaria spp. the mice used, the bigger the mounds were (rpart = 0.32, P = 0.029). In contrast, when the mice used more Chenopodium spp., the mounds were smaller (rpart = −0.41, P = 0.003).
Fig. 3

Mean ± SE mound growth per day, estimated during 1 week of the main building period. The number of mounds included is given in the bars

Finally, when examining the use of different plant species for building while controlling their availability near the mounds (within 3 m), our results revealed that Setaria spp. and Chenopodium spp. were used more often than expected and Amaranthus spp. less often than expected (Fig. 4).
Fig. 4

Amount of the three plant species used for mound building in relation to the availability of the three species in the study area. The zero line indicates a predicted even distribution of plant availability. The number of mounds included is given in the bars

Comparing seed size, calorific and lipid contents of the plant species used for building and eating (Table 1) suggested that seed size was not related to the calorific value or lipid content of the four species.
Table 1

Weight (mg), size (length in mm), caloric value (kcal) and lipid content (%) of seeds, the percentage used for building out of the total amount of each plant species/genus found within 3 m of the mounds and the percentage of each plant eaten in comparison to the others in the experiment (foraging preference) for Amaranthus spp., Amaranthus viridis, Chenopodium album, Setaria spp., Panicum miliaceum and Chenopodium quinoa (see also Hölzl et al. 2009) according to the preference tests (see Fig. 1a, b). For seed weight and caloric content, see Harrold et al. (1980), and Emmerling-Skala (2005), and for lipid content, see Sridhar and Lakshminarayana (1994), Sena et al. (1998), Jancurová et al. (2009) and Deshpande and Poshadri (2011)

Seed species

Seed weight

Seed size

Calorific content

Lipid content

Percent used (building)

Foraging preference

Amaranthus spp.







Chenopodium spp.







Setaria spp.







Panicum miliaceum








If the three dominant plant species used for mound building reflected the food preferences of mound-building mice, Amaranthus spp. would be the preferred food type. However, our food preference tests revealed a different result, namely, that the mice clearly demonstrated a preference for Setaria spp. over the two other species present in their habitat (Fig. 1a). We found no interaction with the box or day of experiment, which suggests that this difference was stable over time (or changing weather conditions) and did not vary between individuals. As well as being the most preferred species, our results also suggested that the handling time of Setaria spp. was more economic for both of the components analysed. The handling time in our experiment included only the time it takes to manipulate free seeds and not the time it needs to extract the seeds from the plant, e.g. from the inflorescences. Seed extraction very likely takes place in the mound as mice usually carry whole inflorescences to the mound. Furthermore, seed extraction usually occurs after some time, when the plant material is already decomposed. Consequently, this part of the handling time might not be so important as seeds fall off after some time and are probably available without further manipulation.

On the one hand, our results showed that the mounds grew significantly faster when Setaria spp. was used compared to the two other plant species (Fig. 3) and even that the final mound size was affected by the type of plant, with Setaria spp. mounds being the largest. In contrast, the percentage of Chenopodium spp. found in the mound was significantly and negatively related to final mound size. At first glance, this result makes no sense as it is not what we would have predicted based on, for example, seed size (Table 1). As larger seeds offer a higher nutrient content (Kerley and Erasmus 1991), many granivorous rodents prefer large seeds (Brewer 2001), but such preferences may be counterbalanced by the higher costs of handling and transporting such seeds (Jacobs 1992; Muñoz and Bonal 2008a). In addition, it was found that foraging decisions depend simultaneously on the size of both the seeds and the rodents (Muñoz and Bonal 2008a).

However, mice transport not only the seeds but also the seed pods. The inflorescences (spica) of Setaria spp. are about 2 to 8 cm long, the inflorescences of Chenopodium spp. usually consist of small shoots with only a few seed cases (about 2 cm), and the size of the parts of Amaranthus spp. that are collected is somewhere in between (own unpublished results).

Furthermore, Setaria spp. also seem to be the easiest to process during foraging since more Setaria spp. seeds were eaten per unit time than the others (Fig. 2). Finally, the energetic value of the seeds was also highest for Setaria spp. (Table 1). Thus, the optimal foraging theory would clearly predict a preference for Setaria spp. over Amaranthus spp. and Chenopodium spp., and this result was actually confirmed by our food preference experiments (Fig. 1a). The discrepancy between these results and the actual frequency found in mounds may be due to other plant characteristics that were not considered, but it is more likely due to the availability of the three plant species in the study area. Including the availability of the three species in the study area revealed a clear preference for Setaria spp. over Chenopodium spp. and Amaranthus spp., as predicted in theory.

Species-specific seed preferences have been observed in several rodent species (Muñoz and Bonal 2008a; Lobo et al. 2009). In mammals, the selection of the appropriate food types from those available in the environment is one of the main problems that offspring have to solve when first becoming independent. Investigation of the mother's mouth by the offspring may be the best way to obtain this information (Suárez and Kravetz 1998). Thus, mothers may play an active role in the tradition and transmission of food selectivity to their offspring. Food (seed) choice in rodents can also be affected by learning without a tutor (mother), and learning by doing may be advantageous over inherited behaviours, particularly in variable and unpredictable environments like human-managed agricultural fields (Muñoz and Bonal 2008b). Our results indeed support these findings since our mice were flexible in their response to the seed plants offered and immediately adjusted their choice when a new food type was offered (Fig. 1b); indeed, our mound-building mice inhabit environments with high plant species diversity (Unterholzner et al. 2000; Hölzl et al. 2009).

In mammals (Brewer 2001; Muñoz and Bonal 2008a) as well as in birds (Bradford et al. 2008), food choice is very much influenced by seed size. Our second experiment also revealed that food choice is influenced by seed size (Muñoz and Bonal 2008a). When there was free choice, we found a clear preference for the novel food type, namely, proso millet, which had the biggest seeds by far (Table 1). This is in line with the highest overall preference found in the experiments. Setaria spp. showed the highest calorific value, and they have the second largest seeds in terms of length which may influence handling time when extracting them from the plant. These two facts may explain the high preference found in nature for this plant species over the two other species, which have more or less the same seed sizes. The differences in calorific and lipid contents between the proso millet and the three other species are probably outbalanced by the enormous size differences. A novelty affect per se such as a change in behaviour due to the new food type is unlikely since we found no change (increase or decrease) in the time spent at food dishes containing proso millet in comparison to Setaria spp. (see results). Furthermore, proso millet is known to be a regular food in other populations (own unpublished observations).

Lipid content is not obviously related to seed size (Table 1), but proso millet has not only the biggest seeds but also rather high lipid content. Therefore, since it is known that, for example, rats and mice prefer lipid-rich foods and since recent observations suggested that, besides olfactory cues, the sense of taste is also involved in fat preference (Passilly-Degrace et al. 2008), the lipid content may also play a role in food choice.

The plant seeds mice collect and eat may not necessarily reflect food preference. Our results showed that the plant seeds stored in the mounds were not necessarily those that were the most preferred. This may be due to the availability of the different plant species, the mobility of the mice and the predation risk during harvesting events. Since food gathering takes place above ground, the mice face the risk of predation by different predators such as the red fox Vulpes vulpes, European polecat Mustela putorius, steppe polecat Mustela eversmanni, stout Mustela erminea, weasel Mustela nivalis, common buzzard Buteo buteo and European kestrel Falco tinnunculus and, during the night, the long-eared owl Asio otus, the barn owl Tyto alba and the little owl Athene noctua. To reduce the risk of predation, the mice usually gather plant material when it is dark. In this way, they can avoid diurnally active predators but may still encounter many nocturnal hunters. Thus, optimal foraging may be extremely influenced by predation risk (Jackson 2001). One may predict that the role of food processing time during ingestion could be of lesser importance, but that handling time in terms of harvesting and transport time should be optimized (Jackson 2001; Chang et al. 2010). Chang et al. (2010) found that handling time for a given food item has a fundamental impact on the hoarding strategies of hoarding animals. Food items with a longer handling time may be more likely to be hoarded due to the increased risk of predation if the mice consume such food above ground (outside their burrows). Little is known about how the burrows used by scatter-hoarding animals influence their foraging behaviours, but it could be that above-ground food storage evolved as a response to predation risk and clumped and sporadic food distribution. Whenever mice encounter a rich food source, they may pile the food up and hide it on site instead of transporting it to distant burrows. This patchy food distribution may also be responsible for the rather untypical social behaviour of the Muridae family. Therefore, future studies should try to include the effect of predation risk on food choices for mound-building mice as well as their habitat choice because selection of an appropriate mound site may be particularly important in this respect.


  1. Best TL, Intress C, Shull KD (1988) Mound structure in three taxa of Mexican kangaroo rats (Dipodomys spectabilis cratodon, D. s. zygomaticus and D. nelsoni). Am Midl Nat 119:216–220CrossRefGoogle Scholar
  2. Bihari Z (2004) A güzüegér (Mus spicilegus) életmódjának sajátságai és mezögazdasági jelentösége. Növényvédelem 40:245–250Google Scholar
  3. Bradford MG, Dennis AJ, Westcott DA (2008) Diet and dietary preferences of the southern cassowary (Casuarius casuarius) in North Queensland, Australia. Biotropica 40:338–343. doi:10.1111/j.1744-7429.2007.00372.x CrossRefGoogle Scholar
  4. Braun-Blanquet J (1964) Pflanzensoziologie, Grundzüge der Vegetationskunde. Springer, WienGoogle Scholar
  5. Brewer SW (2001) Predation and dispersal of large and small seeds of a tropical palm. Oikos 92:245–255CrossRefGoogle Scholar
  6. Bronson FH (1979) The reproduction ecology of the house mouse. Quart Rev Biol 54:265–299PubMedCrossRefGoogle Scholar
  7. Carlsen M (1993) Migration of Mus musculus musculus in Danish farmland. Z Säugetierk 58:172–180Google Scholar
  8. Chang G, Xiao Z, Zhang Z (2010) Effects of burrow condition and seed handling time on hoarding strategies of Edward's long-tailed rat (Leopoldamys edwardsi). Behav Process 85:163–166CrossRefGoogle Scholar
  9. Deshpande HW, Poshadri A (2011) Physical and sensory characteristics of extruded snacks prepared from foxtail millet based composite flours. Inter Food R J 18:730–735Google Scholar
  10. Emmerling-Skala A (2005) Sultan der Gemüsegärten? - der Weiße Gänsefuß (Chenopodium album L.) als Nahrungspflanze, VEN Verein zur Erhaltung der Nutzpflanzenvielfalt e. V., CremlingenGoogle Scholar
  11. Festetics A (1961) Ährenmaushügel in Österreich. Z Säugetierk 26:112–125Google Scholar
  12. Garza JC, Dallas J, Duryadi D, Gerasimov S, Croset H, Boursot P (1997) Social structure of the mound building mouse Mus spicilegus revealed by genetic analysis with microsatellites. Mol Ecol 6:1009–1017. doi:10.1046/j.1365-294X.1997.00278.x PubMedCrossRefGoogle Scholar
  13. Gouat P, Féron C, Demouron S (2003a) Seasonal reproduction and delayed sexual maturity in mound-building mice Mus spicilegus. Reprod Fertil Dev 15:187–195PubMedCrossRefGoogle Scholar
  14. Gouat P, Katona K, Poteaux C (2003b) Is the socio-spatial distribution of mound-building mice, Mus spicilegus, compatible with a monogamous mating system? Mammalia 67:15–24CrossRefGoogle Scholar
  15. Grulich I (1981) Die Baue des Hamsters (Cricetus cricetus, Rodentia, Mammalia). Folia Zool 30:99–116Google Scholar
  16. Harrold RL, Craig DL, Nalewaja JD, North BB (1980) Nutritional value of green or yellow foxtail, wild oats, wild buckwheat or redroot pigweed seeds as determined with the rat. J Anim Sci 51:127–131PubMedGoogle Scholar
  17. Hölzl M, Hoi H, Darolová A, Krištofík J, Penn DJ (2009) Why do the mounds of Mus spicilegus vary so much in size and composition? Mamm Biol 74:308–314. doi:10.1016/j.mambio.2009.02.004 CrossRefGoogle Scholar
  18. Hölzl M, Hoi H, Darolová A, Krištofík J (2011) Insulation capacity of litter mounds built by Mus spicilegus: physical and thermal characteristics of building material and the role of mound size. Ethol Ecol Evol 231:49–59. doi:10.1080/03949370.2010.529827 CrossRefGoogle Scholar
  19. Jackson TP (2001) Factors influencing food collection behaviour of Brants' whistling rat (Parotomys brantsii): a central place forager. J Zool 255:15–23CrossRefGoogle Scholar
  20. Jacobs LF (1992) The effect of handling time on the decision to cache by gray squirrels. Anim Behav 43:522–524CrossRefGoogle Scholar
  21. Jancurová M, Minarovičová L, Dandár A (2009) Quinoa—a review. Czech J Food Scien 27:71–79Google Scholar
  22. Kerley GIH, Erasmus T (1991) What do mice select for in seeds? Oecologia 86:261–267. doi:10.1007/BF00317539 CrossRefGoogle Scholar
  23. Krebs JR (1978) Optimal foraging: decision rules for predators. In: Krebs JR, Davies NB (eds) Behavioural ecology: an evolutionary approach. Blackwell Scientific Publications, Oxford, pp 23–63Google Scholar
  24. Lessells CM, Boag PT (1987) Unrepeatable repeatabilities: a common mistake. Auk 104:116–121Google Scholar
  25. Lobo N, Duong M, Millar JS (2009) Conifer-seed preferences of small mammals. Can J Zool 87:773–780. doi:10.1139/Z09-070 CrossRefGoogle Scholar
  26. McCleery RH (1978) Optimal behaviour sequences and decision making. In: Krebs JR, Davies NB (eds) Behavioural ecology: an evolutionary approach. Blackwell Scientific Publications, Oxford, pp 377–410Google Scholar
  27. Muñoz A, Bonal R (2008a) Are you strong enough to carry that seed? Seed size/body size ratios influence seed choices by rodents. Anim Behav 76:709–715. doi:10.1016/j.anbehav.2008.03.017 CrossRefGoogle Scholar
  28. Muñoz A, Bonal R (2008b) Seed choice by rodents: learning or inheritance? Behav Ecol Sociobiol 62:913–922. doi:10.1007/s00265-007-0515-y CrossRefGoogle Scholar
  29. Muriaru D (1981) La presence de Mus musculus spicilegus Petenyi, 1882 dans le delta du Danube accompagne de son "parasite" Apodemus agrarius (Pall., 1771). Trav Mus Hist Nat "Grigore Antipa" 23:297–304Google Scholar
  30. Passilly-Degrace P, Gaillard D, Besnard P (2008) Perception gustative des lipides alimentaires. Cahiers Nutr Diet 43:273–281CrossRefGoogle Scholar
  31. Poteaux C, Busquet N, Gouat P, Katona K, Baudoin C (2008) Socio-genetic structure of mound-building mice, Mus spicilegus, in autumn and early spring. Biol J Linn Soc 93:689–699. doi:10.1111/j.1095-8312.2007.00944.x CrossRefGoogle Scholar
  32. Reichman OJ, Wicklow DT, Rebar C (1985) Ecological and morphological characteristics of caches in the mounds of Dipodomys spectabilis. J Mammal 66:643–651CrossRefGoogle Scholar
  33. Sena LP, Vanderjagt DJ, Rivera C, Tsin ATC, Muhamadu I, Mahamadou O, Milleson M, Pastuszyn A, Glew RH (1998) Analysis of nutritional components of eight famine foods of the Republic of Niger. Plant Foods Hum Nutr 52:17–30. doi:10.1023/A:1008010009170 PubMedCrossRefGoogle Scholar
  34. Simeonovska-Nikolova DM (2007) Spatial organisation of the mound-building mouse Mus spicilegus in the region of northern Bulgaria. Acta Zool Sinica 53:22–28Google Scholar
  35. Sokolov VE, Kotenkova EV, Michailenko AG (1998) Mus spicilegus. Mammal Spec 592:1–6CrossRefGoogle Scholar
  36. Sridhar R, Lakshminarayana G (1994) Contents of total lipids and lipid classes and composition of fatty acids in small millets: foxtail (Setaria italica), proso (Panicum miliaceum), and finger (Eleusine coracana). Cereal Chem 71:355–359Google Scholar
  37. Stein GHW (1958) Die Feldmaus (Microtus arvalis Pallas). Die Neue Brehm-Bücherei, A. Ziemsen-Verlag, Wittenberg-LutherstadtGoogle Scholar
  38. Suárez OV, Kravetz FO (1998) Transmission of food selectivity from mothers to offspring in Akodon azarae (Rodentia, Muridae). Behaviour 135:251–259Google Scholar
  39. Unterholzner K, Willenig R, Bauer K (2000) Beiträge zur Kenntnis der Ährenmaus Mus spicilegus Petenyi, 1882. Österreichische Akademie der Wissenschaften, ViennaGoogle Scholar
  40. Vander Wall SB (1990) Food hoarding in animals. The University of Chicago Press, ChicagoGoogle Scholar

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  • Michaela Hölzl
    • 1
  • Ján Krištofík
    • 2
  • Alžbeta Darolová
    • 2
  • Herbert Hoi
    • 1
  1. 1.Konrad Lorenz Institute of Ethology, Department of Integrative Biology and EvolutionUniversity of Veterinary MedicineViennaAustria
  2. 2.Institute of ZoologySlovak Academy of SciencesBratislavaSlovakia

Personalised recommendations