Study system
For this study, we captured rodents of three different rodent species, Apodemus sylvaticus (Linnaeus, 1758) (wood mouse), Mus spretus (Lataste, 1883) (Algerian mouse), and Microtus arvalis (Pallas, 1778) (common vole). The captured specimens of the three rodent species were placed in terrariums for reproduction. Nine descendants of each rodent species were selected for the experiment to eliminate any previous acorn consumption experience.
They were placed in a fenced enclosure of 100 m2. Inside we built 9 plots of 9 m2 (3*3 m side). Each of them was isolated by placing on each side of the square a 3 m wide sheet of metal buried in the ground at a depth of 50 cm and a height from the ground of 1 m to prevent rodents from escaping by jumping. Under these conditions, the mice behaved as if they were in semi-wild conditions: building burrows, creating shallow storages, galleries, pantries, and caches to hide the acorns, which they moved around the plot. The 9 specimens of each species were placed in these plots successively, starting with M. spretus and ending with M. arvalis.
The wood mouse has inhabited the Iberian Peninsula since ancient times (Michaux et al. 2003). Acorns of various Quercus species are among its food sources and are abundant in the area it inhabits (Pulido and Diaz 2005; Muñoz et al. 2009). The Algerian mouse is of North African origin but has been present in parts of the Iberian Peninsula for a long time (Palomo et al. 2008). It lives practically in the same habitat as the wood mouse except in the northern mountains. It also consumes acorns from the Quercus species (Muñoz and Bonal 2007). These two species partially consume acorns, opening them at the basal end and preserving the embryo. For this reason, we include them in the mutualism group. A third rodent species, Microtus arvalis, has recently arrived in the region (Luque-Larena et al. 2013; Paz et al. 2012) and does not use acorns for its diet, but consumes them (Del Arco et al. 2018). Common vole is abundant in central Europe. Until recently, its distribution area in the Iberian Peninsula has been confined to the northern mountains, where it can access fresh food such as soft green herbaceous plants. Due to the increase in the distribution area of irrigated crops, the common vole has expanded to the center of the Iberian Peninsula, where it experiences phases of overpopulation, and it is considered a recurrent crop pest (Luque-Larena et al. 2013; Paz et al. 2012). There are no references to this species feeding on acorns despite that the mountain pastures where it originally lives are surrounded by Q. pyrenaica and Q. petraea forest formations, but the humidity of the area means that fresh vegetables are continuously available as food. This species is a risk for the oaks in the center of the Iberian Peninsula because due to phases of overpopulation or by the increasingly pronounced drought can exhaust their traditional food and seek new sources of resources in the acorns and this species is clearly a predator of these seeds (Del Arco et al. 2018).
To feed the rodents and check their food preferences, we have selected six species of acorns from the environment where the rodents live because they may constitute their source of resources and maintain some kind of relationship with them.
Acorns from the following Quercus species were collected: Quercus ilex subsp. ballota (Desf.) Samp. (holm oak), Q. faginea Lam. (Portuguese oak) at the site where the rodents were caught, Q. pyrenaica Willd. (Pyreneen oak), Q. petraea (Matt.) Liebl. (sessile oak), in the northern mountains of the Iberian Peninsula where the common vole originally lived, Q. suber L. (cork oak) in Salamanca (Spain), and Q. rubra L (American oak) (Bieberich et al. 2016; Merceron et al. 2017) in experimental reforestations in Galicia (Spain). The only non-native species is Quercus. Rubra. We have chosen this species to test its integration in this system of relationships between rodents and oaks if it can adapt to the existing dispersal mechanisms in the area (Bieberich et al. 2016; Merceron et al. 2017).
Experimental Procedures and Design
Nine specimens of the descendants of each rodent species were placed in isolated plots. Every day, for five days each specimen had six acorns available to them from each of the Quercus species used. During the experiment, each specimen was offered 180 acorns, 30 from each of the six Quercus species.
Before being given to the rodents, each acorn was weighed and labeled to estimate the individual mass consumed of each acorn. The remains of the acorns from the previous day were removed and weighed to estimate their mass after consumption. To estimate the daily mass consumed of each acorn species, we calculated the sum of the quantities consumed of each acorn by each rodent species specimen for one day. We also estimated the mean mass consumed per acorn of each plant species by each rodent species.
Some acorns from each species were analyzed for their nutrient content (Table 1). We analyzed the chemical composition of the acorns to see whether it influenced selection during consumption (Table 1). A high lignin content is related to a higher tannin content (De Blas et al. 2003), which, in turn, could have deterred rodents from eating acorns (Zhang et al. 2013). On the contrary, the higher fat, protein, and sugar contents are usually the characteristics that most attract rodents to ingest acorns (Table 1).
To study the similarity in acorn composition, we grouped the species with the most similar composition together (Figure 1) according to the Czekanowski C = (2W/A+B)*100 index.
Where W = the sum of the minimum values of each compound in the side-by-side comparison of the two species
A = the sum of the values of each component in the first species in the side-by-side comparison
B = the sum of the values of each component in the second species in the side-by-side comparison.
Another group of acorns was set aside to study their shell thickness. A Mitutoyo caliper was used to measure thickness at three different points: the circular scar where the acorn is attached to the cupule, the periphery of this circle located at the base of the acorn, and the apical end of the acorn, where the shell covers the embryo.
Data Analyses
The possible effects of rodent species (wood mouse, Algerian mouse, and common vole), day (5 levels), Quercus species (Q. petraea, Q. ilex, Q. pyrenaica, Q. faginea, Q. suber, Q. rubra), and their interactions on the number and mass of acorns eaten per specimen, were analyzed using Linear Mixed Models (LMM) with the Restricted Maximum Likelihood method (REML). The specimens were treated as a random factor and the day as a repeated factor. Finally, working on the model matrix, contrasts were carried out to test differences between fixed factor levels (Pinheiro and Bates 2000). Consequently, the Bonferroni correction was used to adjust for the significance level for each t-test (Sokal and Rohlf 1995). The statistical calculations were implemented in the R software environment (version 2.15.3; Core Team R 2013), using the nlme package for LMM (Pinheiro et al. 2013).