Seasonal Reproduction Shift among Three Murine Rodents in a Mediterranean Area of North-Western Africa

Samira Merabet1, Nora Khammes-El Homsi1,*, Lydia Aftisse1 and Stéphane Aulagnier2 1Laboratoire Ecologie et Biologie des Organismes Terrestres, Université Mouloud Mammeri, Bastos Tizi-Ouzou 15000, Algérie 2Comportement et Ecologie de la Faune sauvage, INRAE, Université de Toulouse, CS 52627, 31326 Castanet-Tolosan cedex, France Article Information Received 14 November 2020 Revised 12 April 2021 Accepted 10 June 2021 Available online 07 December 2021 (early access)


INTRODUCTION
A ll latitudes find some mammals reproducing seasonally, even in the deep tropics (Bronson, 2009). In the temperate regions the reproductive process of wild rodents is adaptively timed with the season contrary to synanthropic species which can have litters along the year (Bronson and Perrigo, 1987). Winter is a period of high energetic demand and many rodents show a break in reproductive activity (Gockel and Ruf, 2001), under the main influence of poorer foraging conditions and shorter photoperiod. Ovulation more than spermatogenesis is commonly sensitive to both low temperature and food restriction in small mammals (Bronson, 2009). Hence the end of winter usually initiates the reproductive period in response to lengthening of days, rising of temperature and vegetation growth (e.g. Martinet and Spitz, 1971). Within the reproductive period extending from spring to the beginning of autumn many species show a peak of reproductive activity depending mostly on their feeding ecology, herbivores being earlier than granivores, as they face an enormous energetic drain during lactation (Bronson and Perrigo, 1987). However, this reproductive pattern can be softened by winter reproduction for some members of the population (e.g. Gockel and Ruf, 2001) or even be inverted for the whole population. So, for example, populations of wood mouse (Apodemus sylvaticus) breed from March to November in Britanny (north-western France), and from November to April in Doñana (southern Spain), Pyrénées-Orientales (southern France) and Corsica (Moreno and Kufner, 1988;Fons and Saint Girons, 1993).
In North Africa, rodent females usually give birth during winter and spring, after autumn rainfalls, either in Dipodidae, Gerbillinae, Murinae or Ctenodactylidae (Bernard, 1969;Osborn and Helmy, 1980). For some species, the reproductive period may vary geographically, according to the wet season in the most arid zones, to the temperature in mountains. So, for the lesser Egyptian gerbil (Gerbillus gerbillus), it extends from January to May in Egypt (Osborn and Helmy, 1980) and from July to September in Mauritania (Klein et al., 1975). Pregnant wood mouse females have been recorded from September to February or April in Tunisia, Algeria and Moroccan lowlands (Bernard, 1969;Kowalski, 1985;Harich and Benazzou, 1990;Khidas, 1993;Hamdine and Poitevin, 1994), while the reproductive period extends to May in Moroccan high mountains where there is also a winter break (Saint Girons, 1972). The Algerian mouse (Mus spretus) exhibits a similar last pattern with an approximately three-month winter rest period (Bernard, 1969;Palomo et al., 1985;Kowalski and Rzebik-Kowalska, 1991), like in southern Spain (Vargas et al., 1984(Vargas et al., , 1991Antuñez et al., 1990). However, Orsini et al. (1982) reported an absence of reproduction in summer in southern France. The Barbary striped grass mouse (Lemniscomys barbarus) is the third non-commensal murine species widely distributed in north-western Africa (Happold, 2013). Zaime (1985) trapped juveniles mainly during winter and spring periods in central Morocco, suggesting births from September-October or December-January depending on the year. However, pregnant females were reported in May, June and September in Tunisia (Bernard, 1969), in spring/summer months in northern Morocco (Lahmam et al., 2008). Investigating such variations was the starting point of a study in a Mediterranean area where the three smallsized murine species live sympatrically. We aimed to identify the specific responses to the same physical and ecological conditions in the wild. We hypothesized that all three species may adapt similarly to photoperiod and temperature variations for adjusting their reproductive activity to vegetation growth.

Study plan
Rodents were trapped along a 150 m linear transect, a low disturbing but efficient method for sampling rodent populations in the temperate zone (Spitz et al., 1974). Three-day captures were conducted each month from January to December 2017. Baited traps with bread and pilchard were 3-meter spaced, giving a total of 1620 night-trappings. Animals were euthanized, sexed, aged, measured and any sexual activity was recorded. According to Kowalski (1985), males were considered sexually active when the diameter of testes was over 10 mm and seminal vesicles were developed. For females, reproductive state was estimated by appearance of open vagina, developed mammary glands and uterus (including embryos). They were considered sexually active when gestating and/or lactating (Birkan, 1968).
Age classes were identified according to upper molar wear after cleaning the skulls by boiling heads and then soaking them in bleaching water during 5-10 min. For Apodemus sylvaticus, we followed Saint-Girons (1972) who considered three age classes: juveniles, sub-adults and adults. For Mus spretus, we identified the same three age classes according to Palomo et al. (1983). For Lemniscomys barbarus, we adapted the table published by Van der Straeten (1980) for L. linulus to our trapped specimens for identifying the same three age classes.

RESULTS
The 1620

O n l i n e F i r s t A r t i c l e
along the year with larger numbers in winter (December to February) and lower numbers in summer (June to September) for A. sylvaticus and M. spretus, which no specimen was trapped in June, July and August (Fig. 1A).
The monthly low number of L. barbarus (maximum 6 specimens in November) was quite similar along the year. The gross sex-ratio was balanced for A. sylvaticus (33:32) and in favour of males for M. spretus (32:15) and L. barbarus (28:10), including deep monthly variations for the two first species and not for the third one, males being always more numerous than females (Fig. 2). No female of L. barbarus was trapped from March to May, in August and October.  Few juveniles were trapped for any species except for M. spretus with 11 juveniles over 26 specimens. Juveniles of L. barbarus were trapped in June and July (Fig. 1B) and sub-adults mainly from August to December (Fig. 1C), suggesting a winter break of reproduction. At the opposite, juveniles of M. spretus were trapped from September to May, and sub-adults from November to May. A quite similar pattern was observed for A. sylvaticus with juveniles from October to May and sub-adults from December to March. None of sub-adults was sexually active as well as none adult male of A. sylvaticus trapped from June to September, the first gestating female being recorded in December. Adult males of L. barbarus were sexually active from April to October.

DISCUSSION
The vegetation of our study site is relatively dense, so it is not surprising that Apodemus sylvaticus was the most trapped species among small mammals. This species which prefers forest or forest edges is found in a wide variety of habitats (Denys, 2017a), including mountain grasslands, shrubs and undergrowth cover (lower woody vegetation) in Kabylia (Hamdine and Poitevin, 1994;Khidas et al., 2002). Mus spretus is associated with Mediterranean scrub, bush, grasslands and cultivated fields (Denys, 2017b), and was found syntopic with wood mouse where high woody vegetation is sparse (Khidas et al., 2002) despite some competition for food (Fons et al., 1988). Lemniscomys barbarus prefers bushes and grasses habitats with dense ground cover (Happold, 2013) whereas M. spretus includes a high percentage of bare ground in its home range in Kabylia (Khidas et al., 2002). Such occurrence of the three murine species has been rarely reported, for example in cultivated fields of Esperada (Morocco), among 17 sampled sites, and Lansarine region (Tunisia) with a low number of A. sylvaticus each time Ben Ibrahim et al., 2019). In Kabylia, these three species were previously trapped in four sites: Bouberak (Khidas, 1993), Azazga (Khammes, 1998), Cap Djinet andBoukhalfa (Amrouche-Larabi et al., 2015), with a lower number of L. barbarus each time, like in our study site.
Sex-ratio was balanced for A. sylvaticus, contrary to Hamdine and Poitevin (1994) who reported a larger percentage of males, as we recorded for M. spretus and L.
barbarus. An unbalanced sex-ratio is most often observed in trapped small murine species, particularly during the breeding period for A. sylvaticus (Butet and Paillat, 1997) and after the breeding period for M. spretus (Vargas et al., 1984;Cassaing and Croset, 1985). In L. barbarus males are more often trapped than females after the age of five months when the later become more sedentary (Zaime, 1985). Juvenile dispersal and exploratory behaviour are widely reported in small mammals to explain sex ratios skewed in favour of males in trapping sessions (Stenseth and Lidicker, 1992).
Despite the small sample size our results clearly suggest a different reproductive period for A. sylvaticus and M. spretus with a summer break (June to August), and L. barbarus with a winter break (November to January).

S. Merabet et al.
In the Mediterranean region, the reproductive period of A. sylvaticus was linked to the availability of food, usually reduced in summer as a consequence of drought, and abundant in autumn (fruits and berries) and winter after vegetation growth following autumn rainfalls (Soriguer and Amat, 1979;Torre et al., 2002;Díaz and Alonso, 2003;Rosário and Mathias, 2004). Similarly, the reproductive period of M. spretus is linked to water availability and vegetation growth, mainly mast production and grass seeds in southern France (Orsini et al., 1982). Similarly, according to Zaime (1985) food availability synchronized the reproductive period of L. barbarus and the two syntopic gerbilline species Meriones grandis and Gerbillus campestris. But this period is spread over winter and spring in this sub-arid study site of central Morocco, before the spring and summer period estimated from our data which are supported by previous studies in Tunisia (Bernard, 1969) and northern Morocco (Lahmam et al., 2008), where the climate is sub-humid. These results confirm that, in the Mediterranean area, small mammals are irresponsive to variation in photoperiod (Bronson, 2009). For a species which reproduces along the whole year in captivity (Lenkiewicz and Saint-Girons, 1964), the reproduction shift agrees with the hypothesis of an adjustment of the breeding activity to vegetation growth. Moreover, Lenkiewicz and Saint-Girons (1964) showed that activity declined in cold weather, which can contributes to delay the reproductive period to hotter days in some parts of its range.
The influence of temperature on the shift of reproductive period among the three murine species is hardly supported by our data. First, the mean January temperature is lower in central Morocco (Zaime, 1985) than in Boudjima. Second, L. barbarus, a diurnal species, should be less affected by winter cold nights than A. sylvaticus and M. spretus, which are mainly nocturnal. Then, feeding ecology remains the main likely cause of this seasonal reproduction shift. A. sylvaticus is omnivorous, but predominant food is seeds and acorns, and M. spretus eats mainly fruits, seeds and green parts of plants (Denys, 2017b). In Kabylia, both species are primarily granivorous (Khammes, 1998;Khammes and Aulagnier, 2007). L. barbarus is "probably herbivorous" (Happold, 2013), like most Lemniscomys species (Taylor, 2017) and its reproductive activity seems to respond to herbaceous vegetation growth starting at the end of winter, whereas the two granivorous species get an energetic source from seed and acorn production in autumn and winter.

CONCLUSION
The occurrence of the three small-sized non commensal murine species in one area is a quite rare event that is worth to be reported, mainly when populations seem to be almost balanced, contrary to previous trapping sites reported in the literature. As some competition between Apodemus sylvaticus and Mus spretus was recorded by Fons and Saint Girons, (1988), our study site provides a good opportunity for investigating their interactions with a third phylogenetically related species that could influence their population dynamics.
Our main result is the seasonal reproduction shift between A. sylvaticus and M. spretus on one side, and Lemniscomys barbarus on the other. Roughly, the summer break of the formers stands in opposition to the winter break of the later. This pattern cannot be related to physical factors widely reported to influence reproductive period, such as photoperiod or temperature. Ecological factors are also poorly relevant since, even if they exhibit different daily activity pattern, the three species are living in syntopy. However, a more comprehensive study could reveal some different use of the study area. More likely, the reproduction shift among the three species is linked to their feeding ecology that should be investigated, our study site becoming a higher research spot.

Statement of conflict of interest
Authors have declared no conflict of interest.