CLIMATE, TROPHIC FACTORS, AND BREEDING PATTERNS OF THE NILE-GRASS RAT(ARVICANTHIS-NILOTICUS-SOLATUS) - A 5-YEAR STUDY IN THE SAHELIAN REGION OF BURKINA-FASO (FORMERLY UPPER-VOLTA)

From 1984 to 1989, five populations of the Nile grass rat, Arvicanthis niloticus solatus, Living in five different habitats of the Sahelian region were studied in northern Burkina Faso (Oursi, 14 degrees N). Th e following were investigated: (i) seasonal variations in the percenta ges of sexually active males and females and the percentage of the pop ulation less than 6 weeks old, and (ii) seasonal variations in testis and seminal vesicle masses and plasma testosterone levels in adults. B oth the large form of A. niloticus occupying habitats with stable trop hic resources and the small form of the species occupying habitats wit h fluctuating resources showed seasonal breeding, mainly in the dry se ason, in contrast with what we have observed elsewhere in other specie s in this region. In dry years with normal rainfall (rains between Jun e and September), breeding occurred from mid-October to mid-July in ha bitats with constant trophic resources and from mid-October to mid-Apr il in habitats with fluctuating resources. In 1986, a year with except ional rainfall (additional rains in January), breeding was continuous in habitats where this climatic disturbance caused a reappearance of p rimary production and there was a population outbreak of A. it. solatu s. Relationships between (i) breeding cycles, (ii) long-day periods, w hich have a gonadoinhibitory effect, (iii) seasonal variations in rela tive humidity and temperature, and (iv) dietary characteristics of A. n. solatus suggest that this subspecies is capable of breeding through out the year. This is due to its adaptable dietary habits, which allow if to take advantage of resources that are either constant (insects, harvest wastes, and the water-rich bark of certain woody plants) or se asonal (rain-dependent herbaceous plants and farm vegetables). The end of breeding appears to be related to the relative timing of the stimu latory effect of certain foods and the inhibitory effects of photoperi od, temperature, and humidity. Under this hypothesis, the gonadoinhibi tory effect of long days was masked in 1986 by an earlier stimulatory effect related to the reappearance of vegetation following the excepti onally heavy rainfall in January.