THE MYTHOLOGY OF THE BRAHMAN AS 91
with the demon Vrtra: in one passage (I. vi. 3. 17) the Satapatha Brdhmana divides the dead Vrtra into two parts, one of which goes to make the moon, and the other (the belly) to trouble mankind. The conception is also found in the Maitrdyani Samhitd (II. vii. 8), and it is clear proof that terror of the moon was not unknown to the Indians of the Vedic period. The moon as Candramas often appears with the sun, and the Aitareya Brdhmana (viii. 24) — though in a passage which may be a priestly fiction rather than a genuine belief — states that the moon is born from the sun. A more important conception, which figures largely in the eschatology of the Upanisads, is that the sun is the light of the gods and the moon the light of the fathers, from which it is an easy step to the doctrine that the righteous dead dwell especially in the moon. On the other hand, in its more primitive sense Soma still figures as the heavenly drink in the story of his descent to earth, which is now attributed to the Gayatri metre; and since this metre is used at the morning pressing of the soma and is closely associated with Agni, we thus have a variant of the legend which is seen in the Rgveda (iv. 27) when Soma is brought down by the eagle. The Gayatri is shot at by the archer who guards the soma, and a nail of her left foot, being cut off, becomes a porcupine, while the goat is born of the fat that drips from the wound. The other metres, Jagati and Tristubh, failed in the effort to obtain the soma, being wearied by the long flight to heaven.
Agni does not change his essential features in the later Vedic period, but his character is more fully set out. Thus while the Rgveda mentions only one of the three fires, the Garhapatya, the later texts name also the Ahavaniya and the Daksinagni; and the three are brought into conjunction with the earth, the sky, and the atmosphere respectively, besides being associated with the three categories of men, gods, and fathers, and with Agni, Surya, and Vayu. It is a question how far in these equa- tions we have to see mere priestly schematism: it has been sug-
