Focal animals belonged to three different nest groups that were located in their nests on 24, 16 and 31 days per group, respectively. We were able to observe animals during emergence and return on 26 evenings and 24 mornings. The total time of ad libitum nest behaviour observations before returning and after emergence comprised 41.5 hours.
Group size of the three nest groups was two to four individuals (Fig. 3). Nest Groups 1 and 3 contained only one (sub-adult) female but several mature males. Maturity was assumed as these males had fully developed testes. Nest Group 2 usually consisted of an adult female and her young. Nests were exclusively used by one group only. Only once did we observe another individual entering a nest (see below). A fourth group is shown in Fig. 3, but although animals were captured at the same site it is unclear if they represented a nest group.
The three groups used one, three and three different sleeping trees, respectively. On all days of direct observation the nest group compositions were stable, that is, all group members but no additional animals slept in the nest. Only on three days we observed an unidentified animal sleeping in the nest of Group 2, consisting of a female and her young.
Return time in the morning was between 412h and 547h (mean 528h, SD 17:14 min, n = 42). In the evening the animals emerged from the nests between 1722h and 1749h (mean 1736h, SD 5:52 min, n = 63). When emerging from nests the individuals left the nest site immediately. At their return they often entered the area before sunrise and engaged in grooming and social behaviour such as playing. The latter was only observed between mixed-sex pairs. Allogrooming was observed once between two males.
Group 1 stayed in the same nest during all 24 sampling days over a 44-day period. Group 2 was located in three nests during all 16 sampling days in a 35-day period. This group swapped between two close sleeping trees during a 19-day period (9 sampling days), but after a storm lasting eight days they changed their nest to a new area. We detected Group 3 on 31 sampling days during a 50-day period; they used three close nests during this time. A swap between two of these nests took one week, during which one or two individuals alternately slept in the new nest on different nights until the whole group settled there. The old nest disintegrated quickly during a storm. General return rate (actual individual returns divided by possible individual returns) of all M. zaza was high, with an average of 91.9 % (SD 11.3 %, n = 8). Togetherness in Group 1 was 1 and average sleeping group aggregation size was 4 (n = 13, 4 animals). In Group 2 togetherness was 1.3 (n = 20, 3 animals) and average sleeping group aggregation size 2.31 (n = 26, 3 animals).
Allelic diversity of microsatellites was comparably low (Markolf et al., 2008). A summary of the characteristics of the five microsatellite loci is given in Table 1. Maximum number of detected alleles was five, though length differences of detected alleles ranged up to 22bp. Mean relatedness within sleeping groups ranged between -0.11 and 0.35 (Fig. 3). The microsatellite data showed a relatively low genetic diversity with few alleles and low levels of heterozygosity (Table 1). Genetic diversity was also very low in the mitochondrial sequences, as a total of three haplotypes were found that differed from each other in only 1-3 bp. In one sleeping group (Group 3) mean relatedness was slightly lower than the mean relatedness estimate of all twelve individuals (rmean = -0.08). The composition of sleeping groups appeared to vary in respect to the presence of related individuals. Nevertheless, relatedness of co-sleeping dyads (n = 14, mean r = 0.06) was higher than relatedness within non-co-sleeping dyads (n = 77, mean r = -0.12, unpaired Wilcoxon rank sum test: W = 471, p-value = 0.04).
According to the comparison of individual genotypes, r-values and mitochondrial haplotypes, we found eight dyads that reached a level of relatedness compatible with first degree relatives, one male-male dyad, five male-female and two female-female dyads (Table 5, Fig. 3). If these relationships could be corroborated through further data, this would mean that for Mirza zaza, first-degree relatives could be found within and between sleeping groups. Not all closely-related animals, however, shared the same mitochondrial haplotype, i.e., belonged to the same matriline. Relatedness may therefore partly have paternal origins (e.g. dyad M1/M3). It can be assumed that F2 was the mother of FJ3 (141 g) as they were sleeping in one nest, belonged to the same mt haplotype, had matching genotypes for all five microsatellite loci, were often seen together during nocturnal activities, and F2 had enlarged nipples. Co-sleeping adult males were unrelated except for one dyad (M1/M3) with an r-value of 0.72, which had no mismatches for the microsatellites but mismatching mitochondrial haplotypes.