included in the Genus Arvicanthis. 69 
separating its tip from the small third lamina (whose large 
axis is transverse) deep, so that the two only become con- 
nected in old age. 
Lemniscomys contains the great mass of the striped and 
spotted rats of Africa. The better-known species like 
barbarus and striatus are profusely striped and spotted, 
while, on the other hand, in the spinalis* group there is only 
a single median dorsal line, and it seemed natural to expect 
some superspecific difference to be present between the two. 
But not only are they identical in all structural characters, 
but even in coloration certain of the rarer species practically 
link up the two extremes. Thus in JL, striatus venustus the 
lateral light spots are less developed than usual, and in 
L. linulus, while present, they are so reduced that the species 
was described as being specially related to L. * dorsalis.” 
And, finally, E. African forms described as subspecies of the 
same animal have a strong suspicion of lateral spotting. 
There is therefore a practically complete intergradation 
from one end of the series to the other. 
RHABDOMYS, gen. nov. 
Rattus, Donovan, 1827 ; nec G. Fisch. 1814; nec Desm. 1822. 
Type. R. pumilio (Mus pumilio, Sparrm.). 
Build lighter than in Arvicanthis. Back distinetly dark- 
lined, but the median line light instead of dark, with two 
dark lines on each side of it, making four in all ; the number 
of dark lines is therefore even instead of odd. Fifth finger 
normal, with a claw. 
Skull comparatively lightly built, somewhat bowed in the 
frontal region. Zygomatic plate not concave anteriorly. 
Molars small, far lighter than in Arvicanthis, their 
structure essentially as in Lemniscomys, the lamine similarly 
bowed, or even alittle more so. Main lamina of m’ strongly 
bent round on itself externally, and, owing to the shallowness 
of the notch behind it, it is very early joined to the posterior 
element of the tooth, which is round or longitudinally oval. 
The pumilio group form a very distinct genus, charac- 
terized as above detailed. Even in coloration, although 
* The well-known animal usually called <Arvicanthis dorsalis un- 
fortunately needs a new name, as Smith’s Mus dorsalis of 1845 is 
invalidated by G. Fischer’s use of the same name for a Svcista in 1814 
(Zoogn. iii. p. 66). I-would propose to replace dorsalis by spinalis, but 
this would only stand asa subspecies of grise/da, Thos., the senior of the 
various subspecific names which have been added to dorsalis. The proper 
name of the South-African subspecies would therefore be Lemniscomys 
griselda spinalis, 
