That there was such an Italian deity is placed almost beyond doubt by the name itself, which all authorities agree in connecting with cerus = genius, and with the _cerfus_ and _cerfia_ of the great inscription of Iguvium[225]. The verbal form seems clearly to be _creare_[226]; and thus, strange to say, we actually get some definite aid from etymology, and can safely see in the earliest Ceres, if we recollect her identification with the Greek goddess of the earth and its fruits, a deity presiding over or representing the generative powers of nature. We cannot, however, feel sure whether this deity was originally feminine only, or masculine also, as Arnobius seems to suggest[227]. Judging from the occurrence of forms such as those quoted above, it is quite likely, as in the case of Pales, Liber, and others, that this numen was of both sexes, or of undetermined sex. So anxious were the primitive Italians to catch the ear of their deities by making no mistake in the ritual of addressing them, that there was a distinct tendency to avoid marking their sex too distinctly; and phrases such as ‘sive mas sive femina,’ ‘si deus si dea,’ are familiar to all students of the Roman religion[228]. We may be satisfied, then, that the oldest Ceres was not simply an importation from Greece. It is curious however, that Ceres is not found exactly where we should expect to find her, viz. in the ritual of the Fratres Arvales[229]. Yet this very fact may throw further light on the primitive nature of Ceres. The central figure of the Arval ritual was the nameless Dea Dia; and in a ritual entirely relating to the fruits of the earth we can fairly account for the absence of Ceres by supposing that she is there represented by the Dea Dia—in fact, that the two are identical[230]. No one at all acquainted with Italian ideas of the gods will be surprised at this. It is surely a more reasonable hypothesis than that of Birt, who thinks that an old name for seed and bread (i. e. Ceres) was transferred to the Greek deity who dispensed seed and bread when she was introduced in Rome[231]. It is, in fact, only the name Ceres that is wanting in the Arval ritual, not the numen itself; and this is less surprising if we assume that the names given by the earliest Romans to supernatural powers were not fixed but variable, representing no distinctly conceived personalities; in other words, that their religion was pandaemonic rather than polytheistic, though with a tendency to lend itself easily to the influence of polytheism. We may agree, then, with Preller[232], that Ceres, with Tellus, and perhaps Ops and Acca Larentia, are different names for, and aspects of, the numen whom the Arval brothers called Dea Dia. At the same time we cannot entirely explain why the name Ceres was picked out from among these to represent the Greek Demeter. Some light may, however, be thrown on this point by studying the early history of the Ceres-cult.