mys mechowiclade and the Cryptomys hottentotusclade
(fig. 36.2) or even two genera (Coetomysand Cryptomys;
Ingram et al. 2004). While the latter is distributed through-
out South Africa, extending into part of Mozambique and
Zimbabwe, species in the Cryptomys mechowiclade are
found in southern, central, and western Africa, but are ab-
sent from the horn of Africa, tropical rainforests of Central
and West Africa, and the Sahara (fig. 36.1). As with Hete-
rocephalus,the ranges of some of these social cryptomids
extend into areas of very low rainfall, which is sporadic and
unpredictable (sometimes 200 mm per annum). How-
ever, some species also occur in mesic areas, like the com-
mon mole-rat, Cryptomys hottentotus hottentotus. The re-
lationships between rainfall patterns, food distribution, and
sociality have received much discussion (e.g., Jarvis et al.
1994; Faulkes et al. 1997; Bennett and Faulkes 2000; Burda
et al. 2000), and will be reviewed in the following.
The bathyergid fossil record
Using a molecular clock approach inferred from DNA se-
quence analysis, an ancient (Eocene) origin for the Bathyer-
gidae has been consistently suggested (see the following).
Unfortunately, fossil-bearing strata of the Eocene /Oligocene
are extremely rare in Africa (Lavocat 1978), so if these mo-
lecular timings are correct, from a palaeontological point of
view, this earliest stage of bathyergid evolution may remain
shrouded in mystery. Fossil bathyergids first appear in the
more common early Miocene deposits, and three extinct
genera have been identified from this period (around 20 mil-
lion years ago) in both East Africa and Namibia (Lavocat
1973, 1978). Of the extant genera, Heterocephalusis the
first to appear in the fossil record. The earliest (but un-
named) fossil resembling H. glaberwas discovered in Early
Miocene deposits at Napak in Uganda, together with fos-
sil Bathyergoidea, a sister taxon to the Bathergidae (Bishop
1962). Bishop et al. (1969) subsequently dated these strata
to a minimum age of 17.8 million years (myr). The tempo-
ral association of Heterocephalusfossils with extinct ba-
thyergid ancestors supports the early divergence of Hetero-
cephaluswithin the family that is suggested by molecular
phylogenies.Phylogeography and patterns of speciation
The availability of robust molecular phylogenies and the
application of molecular clocks to estimate divergence times
now enable us to infer the potential phylogeographic influ-
ences that underlie the adaptive radiation of the Bathyergi-
dae and patterns of social evolution. Molecular phylogenies
of the Bathyergidae are firmly rooted in East Africa (Hete-
rocephalusand Heliophobiusforming the basal lineages;
fig. 36.2). It has been suggested (Honeycutt et al. 1991) that
a possible route for the spread of the Bathyergidae was via
a corridor of fluctuating aridity linking east and south West
Africa (Bakker 1967). This arid corridor has been impli-
cated in the distribution of both modern and Early Miocene
fossil faunas (Van Couvering and Van Couvering 1976).
Thus arid- or mesic-adapted ancestral bathyergids could
have exploited this corridor, according to the prevailing cli-
matic conditions. At the same time, large-scale physical and
climatic changes were also occurring in this part of Africa
as a result of continental drift, leading to rifting and in-
creased volcanism. Apart from its possible role as a physi-
cal barrier, in the form of volcanic uplands and deep valleys
(some eventually forming the great lakes of Africa), it is
also likely that the climatic and vegetative changes that in-
directly resulted from the rifting process have been of im-
portance in the distribution and social evolution of the
Bathyergidae.
Molecular clock-based timing of the major divergences
within the family suggest that the initial cladogenesis of
the Bathyergidae was sufficiently early to be independent of
rifting, and that a general radiation occurred from East Af-
rica into southern and south-central Africa. Divergence of428 Chapter Thirty-Six
Figure 36.1 Map showing the approximate ranges of extant genera, sepa-
rating the genus Cryptomysinto the two subclades suggested by molecular phy-
logenies (adapted from Bennett and Faulkes 2000).