Male reproductive tactics in house mice: consistent individual differences, intrinsic factors and density effects

Alternative reproductive tactics (ARTs) describe fixed or flexible alternative strategies to secure fertilization within species. For example, while some males defend territories to attract females, others invade them to attempt sneaky matings. Often, male ARTs are considered to be status-dependent, explained by mass or competitive differences. Here, we used 244 male mice, Mus musculus domesticus, from semi-natural populations to address those caveats and describe ARTs in Mus musculus for the first time. We followed males throughout their life and categorized them as territorials or roamers over multiple monthly intervals, after validating our method of assigning a tactic with detailed spatial data. We explored if tactic choice is consistent, whether multiple social and/or intrinsic factors predict tactic choice, and tested for fitness consequences and physiological differences between ARTs. Tactic choice was consistent and associated with mass, age, the operational sex ratio, and population size. We also found that territorials had a higher probability of reproduction but a lower gonadosomatic index. Our results reveal a personality component of ARTs, confirm equal fitness among tactics, and show ARTs as multifaceted traits that probably are under various selective pressures.

Methods
Mice (Mus musculus domesticus) descended from wild populations sampled in the Cologne/Bonn region of Germany (N = 18 original breeding pairs) (50°45′N–51°N, 6°45′E–7°E). Founding mice (N = 160, 80 males, 80 females) for our semi-natural populations were distributed to four replicate enclosures (19.6 m2 each), with 20 males and 20 females founding one population.    

For house mice, the natural environment is a barn or a human shelter, and, to mimic such an environment as closely as possible, each semi-natural enclosure was equipped with various nesting materials and 12 nest-boxes/houses (hereafter we use the word territory to describe that a male occupies one nest-box/house—the territory holder—; however, in many species “territory holder” refers to a male monopolizing multiple houses—which is not the case here). Food and water were provided ad libitum and distributed uniformly across the room at nine feeding stations. Indeed, such conditions are similar to those found in natural European populations of M. m. domesticus. After releasing the founders, we monitored population developments every 4–5 weeks. Population development and densities in semi-natural enclosures mimic mice populations that are under long-term observation with the possibility for dispersal. During monitoring, we caught all mice within an enclosure, measured body mass, checked for fresh bite marks, and took a tissue sample of new individuals (weighing >10 g) for parentage assignment. New animals received an RFID pit tag (Planet ID, 1.4 × 9 mm) for permanent recognition. Whenever a given population reached at least 80 chipped offspring, we removed the older generation to keep the density below the carrying capacity. Some of the removed animals were euthanized in order to obtain physiological characteristics, e.g., conduct sperm counts. 

We classified reproductive tactics according to males’ location during our monitoring. If a male was located inside a nest-box/house it was identified as the territory holder and if roaming outside nest-boxes as roamer. We excluded all observations from months when more than one potentially mature (> 4 weeks) male was found at a particular house, as it was unclear which male was the territory holder. 

We used a stepwise approach to determine our final dataset: from an initial of 1699 observations from 336 males, we excluded (1) the first two observations of monitoring for each male since they were still immature (males only reproduce vary rarely within the first two months; only 1 male did from Nmales = 336). Moreover, we rejected (2) those males that were not assigned as, or were not possible to categorise them as, territorials or roamers (explained below). In total, we ended up with 727 observations from 244 males.

We used seventeen microsatellite markers to determine parentage and assign fitness using the procedure adapted from Linnenbrink et al. (2013). In summary, DNA was extracted from ear clips, amplified using a Multiplex PCR kit (QIAGEN), and the samples were run on an ABI 3730 Sequencer (Applied Biosystems). We used GeneMarker (V2.6.4) to identify alleles, and Colony [©COLONY | Zoological Society of London (ZSL)] to assign the parentages based on the maximum likelihood of each potential parental pair. We conditioned our parentage assignment on the following assumptions: sexual reproduction, polygamous mating system, possible inbreeding, and all animals being present the month before sampling (i.e., when juveniles were born) being possible parents. Overall, our fitness estimates quantify the number of offspring that survived to at least 10g. 

We used fitness to explore differences between ARTs in reproductive success (Nmales = 104), we quantified fitness in four different ways: (1) We first used a binary variable describing if a male reproduced (1) or not (0) during its lifetime; (2) we also added the total number of offspring each male produced as a measure of absolute fitness. We then used two different measures of relative fitness, (3) the number of offspring produced by each male during its life divided by the total yield of offspring of all males that belonged to the same generation, and (4) the total number of offspring of each male divided by its lifespan. Measure 3 allowed us to identify the relative fitness of each individual at the population level while the latter (i.e., 4) “stratifies” individuals based on both their offspring yield and their survival rate. We used individuals that reproduced successfully, along with those that did not, as omitting the latter would inevitably increase average estimates of fitness for each tactic.

Physiological traits: Sperm quantity was assessed using a modified version of the procedure described in Wang Y. (2002). Epididymal sperm count. Curr. Protoc. Toxicol. 14 (1), 16–16. doi: 10.1002/0471140856.tx1606s14. The epididymis was sliced in a Phosphate buffer solution (PBS). The sperm suspension was incubated in a thermomixer for 10 min at 40°C to kill the sperms. In a Burker chamber, 10 μl of 1:20 or 1:40 diluted sperm suspension was used accordingly to get a sperm count of around 3–10 per square. Sperms were counted twice under a microscope at 40× with a PH2 filter in 25 squares.
Randomly chosen adult males out of all males present in the enclosure were dissected to assess testes’ mass and sperm concentration. Males were killed using carbon dioxide, followed by cervical dislocation. The whole body and testes were weighed, and the testes/body mass percentage was calculated.

Instrument- or software-specific information needed to interpret the data: All statistical analyses were performed using R (version 4.2.1 2022-06-23). Repeatability (R) was calculated using the rpt function (rptR package). For the modeling part, we used the packages stats, lme4, and pscl.