1. The spermatozoa of the following Muridae, which had not previously been described, have been studied in detail, measured, and figured:
Apodemus flavicollis (De Winton's field mouse).
Micromys minutus (harvest mouse).
Rattus rattus (black rat).
Microtus hirtus (short-tailed field mouse).
Evotomys glareolus (bank vole).
Dicrostonyx sp. (Canadian lemming).
Ondatra zibethica (musk rat).
In addition the sperms of the following, described and figured by Retzius, have been measured:
Mus musculus (house mouse).
Apodemus sylvaticus (long-tailed field mouse).
Rattus norvegicus (common rat).
2. Among these, two distinct types of sperm can be distinguished, one with a hooked nucleus associated with the subfamily Murinae and the other with a recessed nucleus associated with the sub-family Microtinae. In both of these, however, the head as a whole is hooked.
3. The sperms of all the genera hitherto described and those described in this paper fall into these two type-groups, with the exception of those of the harvest mouse and the musk rat, the heads of which are simpler and unhooked.
4. It is believed that the simpler type of sperm is the more primitive, since it is similar to those found in more primitive groups of the Rodentia; and that the hooked head has been evolved independently in the two sub-families.
5. Each of the genera examined has a distinctive sperm; but specific differences are not so obvious, the sperms of the two British rats, Rattus norvegicus and Rattus rattus, differing in minor features and in head-length, while those of the two species of Apodemus differ in head-length alone.
6. Within each species, however, differences in mean headlength occur between individuals, and these are sometimes significant. It appears that the mean head-length of its sperms is a characteristic of each male rat or mouse comparable to other characteristics such as the length of its tail. It is presumed that this cytological characteristic, like the grosser ones, varies normally throughout the population.
7. It is shown that a quantitative criterion exists for deciding, from the head-length of their sperms alone, whether two individuals of either the genus Apodemus or the genus Rattus belong to the same species or represent distinct species. Since these were the only two genera in which congeneric species were examined, it is not known whether this criterion may have a more general application.
8. In each of these groups of the Muridae the morphology of the sperm has evolved with the evolution of the animals which bear it. Thus it is possible io recognize any of these species from their sperms alone, and to decide in the same way the genus and (with two exceptions--Micromys and Ondatra) the sub-family to which they belong.
9. Outlines of the nuclei of all the known sperms of the Muridae, taken wherever possible from Feulgen preparations, are embodied in a diagram (Text-fig. 30) which places them in their natural groups and at the same time summarizes the morphological side of this study. Other diagrams (Text-figs. 1 and 2) illustrate a nomenclature for the more specialized features of the Muride sperm.