r Good Dragon, for it controlled the waters of
irrigation and gave fertility to the soil (_op. cit._, p. 36). The
_fravashi_ was also identified with the third member of the primitive
Trinity, the Warrior Sun-god, not merely in the general sense as the
adversary of the powers of evil, but also in the more definite form of
the Winged Disk (_op. cit._, pp. 67 and 68).

In all these respects the _fravashi_ is brought into close association
with the dragon, so that in addition to being "the divine and immortal
element" (_op. cit._, p. 51), it became the genius or spirit that
possesses a man and shapes his conduct and regulates his behaviour. It
was in fact the expression of a crude attempt on the part of the early
psychologists of Iran to explain the working of the instinct of
self-preservation.

In the text of Chapters I and III I have referred to the Greek,
Babylonian, Chinese, and Melanesian variants of essentially the same
conception. Soederblom refers to an interesting parallel among the
Karens, whose _kelah_ corresponds to the Iranian _fravashi_ (p. 54, Note
2: compare also A. E. Crawley, "The Idea of the Soul," 1909).

In the development of the dragon-myth astronomical factors played a very
obtrusive part: but I have deliberately refrained from entering into a
detailed discussion of them, because they were not primarily the real
causal agents in the origin of the myth. When the conception of a
sky-world or a heaven became drawn into the dragon story it came to
play so prominent a part as to convince most writers that the myth was
primarily and essentially astronomical. But it is clear that originally
the myth was concerned solely with the regulation of irrigation systems
and the search upon earth for an elixir of life.

When I put forward the suggestion that the annual inundation of the Nile
provided the information for the first measurement of the year, I was
not aware of the fact that Sir Norman Lockyer ("The Dawn of Astronomy,"
1894, p. 209), had already made the same claim and substantiated it by
much fuller evidence than I have brought together here.

In preparing these lectures I have received help from so large a number
of correspondents that it is difficult to enumerate all of them. But I
am under a special debt of gratitude to Dr. Alan Gardiner for calling my
attention to the fact that the common rendering of the Egyptian word
_didi_ as "mandrake" was unjustifiable, and to Mr. F. Ll. Griffith for
explai