up in most unexpected places. Mr. Frazer, Mannhardt's disciple, is very
severe on solar theories of Osiris, and connects that god with the corn-
spirit. But Mannhardt did not go so far. Mannhardt thought that the
myth of Osiris was solar. To my thinking, these resolutions of myths
into this or that original source--solar, nocturnal, vegetable, or what
not--are often very perilous. A myth so extremely composite as that of
Osiris must be a stream flowing from many springs, and, as in the case of
certain rivers, it is difficult or impossible to say which is the real
fountain-head.

One would respectfully recommend to young mythologists great reserve in
their hypotheses of origins. All this, of course, is the familiar
thought of writers like Mr. Frazer and Mr. Farnell, but a tendency to
seek for exclusively vegetable origins of gods is to be observed in some
of the most recent speculations. I well know that I myself am apt to
press a theory of totems too far, and in the following pages I suggest
reserves, limitations, and alternative hypotheses. Il y a serpent et
serpent; a snake tribe may be a local tribe named from the Snake River,
not a totem kindred. The history of mythology is the history of rash,
premature, and exclusive theories. We are only beginning to learn
caution. Even the prevalent anthropological theory of the ghost-origin
of religion might, I think, be advanced with caution (as Mr. Jevons
argues on other grounds) till we know a little more about ghosts and a
great deal more about psychology. We are too apt to argue as if the
psychical condition of the earliest men were exactly like our own; while
we are just beginning to learn, from Prof. William James, that about even
our own psychical condition we are only now realising our exhaustive
ignorance. How often we men have thought certain problems settled for
good! How often we have been compelled humbly to return to our studies!
Philological comparative mythology seemed securely seated for a