52 A sinus in Tegulis.
come back again. ^ The logical Greek followed out this hypothesis to the natural conclusion ; being a ghost, the man could proceed to be re-born, swaddled, suckled, and then allowed to grow up as soon as he liked ; the common- sense Roman fetched him in by the recognised path of spirits, and then presumably had a good look at him, decided ,that he was too solid to be an ordinary spook, and after a reasonable number of piacida, admitted him once more to embodied human society.
Frazer has no difficulty whatsoever in showing that ghosts are provided with an exit of their own, nor that their path is very often a hole in the roof. His examples range from Norway to South Africa, and from Russia via Green- land to China, and the islands of the Pacific. I have nothing of importance to add here, but wish to point out that ghosts are merely one of the uncanny things whose way is on or through the roof ; and that some at least of the others are not reasonably to be connected with the fear of ghosts, or indeed of spirits of any kind.
[a) Inmiiinate Things and the Roof. It is well known, that if a bound man got into the house of the flamen Dialis at Rome, not only must he be unbound, but his fetters must be pulled out through the impliiuium on to the tiles, and then flung down into the street (Gelhus x. 15, 8). Here there is no ghost and no spirit in question. The fiamen must avoid anything suggestive of tying or binding, — hence he might not wear an ordinary ring, nor have a knot in his clothes, nor pass under a vine-trellis, — so, if the evil magic of binding actually got into his house it must be thrust forth by the regular road of all uncanny things. A similar idea, perhaps, attaches to the removal of the fig-tree from the roof of the temple of the Dea Dia by the Arval Brothers (Henzen, Acta fratrum arualium, p. clxxxvi., cf. 141) ; the fig was generally speaking lucky, but this
^ Le Conte C. N. de Cardi in J. R.A.I, xxix. (1899), p, 53.
