Saharan Africa. The first fossil bathyergids are known from
this period (Early Miocene) in both East Africa and Na-
mibia, confirming that southern Africa was also being col-
onized at this time. During the later Early and Middle Mio-
cene the Cryptomysgenus diverged into its two subclades
(at 12 –17 myr; fig. 36.2, node 4; fig. 36.3d, 36.3e). Again,
this coincides with a critical period when rifting was pro-
gressing in the Kenya rift and also just beginning in the West-
ern rift. While the C. hottentotussubclade appears to have
speciated almost exclusively in South Africa, the C. me-
chowisubclade underwent a more extensive radiation, par-
ticularly in Zambia and Central Africa, resulting in a wide
diversity of genetically divergent chromosomal forms. All
extant species in these clades are social, cooperative breed-
ers. Interestingly, the initial radiation of Cryptomysis coin-
cident with the onset of volcanism in the adjacent Western
rift, which presumably resulted in considerable environ-
mental challenges as climate and vegetation changed. It is
possible that in these circumstances, social elaboration
could have been adaptive. This period has previously been
reported as a time of faunal turnover (Van Couvering and
Van Couvering 1976). The increasing volcanism and for-
mation of the East Africa rift during the Miocene appears
to have almost completely isolated the populations of Het-
erocephalusand Heliophobiusto the east, and restricted
Cryptomysto the west of the Rift. Exceptions to this are a
few populations of Cryptomys whyteiin western Tanzania,
and Heliophobiusfound in Malawi and Zambia. The latter
appear to have diverged before local rifting restricted their
movement (fig. 36.2, node 5; fig. 36.3f; Faulkes et al. 2004).
Social Diversity in the Bathyergidae
Despite many early species descriptions, current interest in
the Bathyergidae began to increase in the 1980s, after the
eusocial system of the naked mole-rat was first reported
(Jarvis 1981). While it is now clear that there is wide so-
cial diversity between genera across the family (solitary-
dwelling genera, Heliophobius, Georychus,and Bathyer-
gus,versus social genera, Heterocephalusand Cryptomys),
there may also be considerable variation within the social
genusCryptomys. Robust molecular phylogenies now make
comparative analyses possible, enabling the evolution of so-
cial traits to be correlated with environmental factors (e.g.,
Faulkes et al. 1997). However, fundamental difficulties re-
main in quantifying both social behavior and ecological
parameters.
Definitions of sociality
Ironically, a major problem with the quantitative study
of social evolution is defining what is meant by sociality.
The term “eusocial” was first used to describe groups of
insects living in close-knit communities but which have a
reproductive division of labor. By definition, only a small
number of individuals are actually involved in direct repro-
duction (reproductive skew). The remainder of the social
group is composed of overlapping generations of function-
ally or irreversibly sterile nonbreeding helpers that coop-
erate in the rearing of offspring (Michener 1969; Wilson
1971).
Rigid, categorical definitions, such as those of Michener
(1969) and Wilson (1971), which have been used to de-
scribe eusocial species, are less useful for quantitative com-
parative studies than a continuous measure of social com-
plexity. In comparative studies of mole-rats, Faulkes et al.
(1997) used maximum group size as an indication of soci-
ality and constraint on dispersal. The use of group size as
a criterion was criticized by Burda et al. 2000, who argued
that eusocial mole-rats should be defined by “permanent
philopatry.” However, while all social species are philopat-
ric to some degree, not even the most social of the mole-rats
are permanently so, e.g., H. glaber(Braude 2000) and C.
damarensis(Hazell et al. 2000; Burland et al. 2002, 2004).
It is generally accepted that some kind of continuous mea-
sure of the type proposed by Sherman et al. (1995) and Kel-
ler and Perrin (1995) will prove to be the most useful esti-
mate of the degree of sociality. These indexes of eusociality
and reproductive skew are difficult to quantify empirically
because estimates of relatedness and lifetime reproductive
success are needed. However, studies of group kin-structure,
parentage, and dispersal using molecular genetic markers
are now beginning to gather these kinds of data, and show
great promise for the future (e.g., Bishop et al. 2004; Bur-
land et al. 2002, 2004; Griffin et al. 2003).The common ancestor of mole-rats:
Was it solitary or social?
The question of the social status of the ancestral bathyergid
is an interesting and important one, but difficult to answer.
Because few other subterranean mammals are social (Nevo
1979), Jarvis and Bennett (1990) have suggested that the
first bathyergids were probably solitary. The earliest known
fossil Bathyergoidea (a sister taxon to the Bathyergidae)
were large animals (Lavocat 1978) and the largest living
bathyergids are solitary (Jarvis and Bennett 1990). By in-
ference this might suggest that if these large fossil forms
were completely subterranean they were also solitary. Re-
cent studies suggest an arid, savanna-type habitat at the fos-
sil sites for naked mole-rats, e.g., Acacia savanna at Lae-
toli (Denys 1987), and assemblages similar to Rushinga and
Songhor (i.e., savanna) at Napak I (Bishop 1962). Con-
versely, the presence of Heterocephalus jaegeriat Olduvai
is used to infer aridity of the habitat (Denys 1989). How-African Mole-Rats: Social and Ecological Diversity 431