It is remarkable that he attributes to the dragon the same desire and capacity to attack the elephant as is attributed to the Pa snake in Western China, and by the old Arabian voyagers to serpents in Borneo. The _Shan-hai-king_, a Chinese work of extreme antiquity, of which special mention will be made hereafter, says: "The Pa snake swallows elephants, after three years it ejects the bones; well-to-do people, eating it, are cured of consumption." Diodorus Siculus, in speaking of the region of the Nile in Libya, says that, according to report, very large serpents are produced there and in great numbers, and that these attack elephants when they gather around the watering places, involve them in their folds till they fall exhausted, and then devour them. Diodorus, in another passage referring to the crocodiles and hippopotami of Egypt, speaking of Ethiopia and Libya, mentions a variety of serpents as well as of other wild beasts, including dragons of unusual size and ferocity. While El Edrisi says: "On peut encore citer le serpent de Zaledj dont parlent Ben Khordadébe, l'auteur du Livre des Merveilles, et divers autres écrivains qui s'accordent à dire qu'il existe dans les montagnes de l'ile de Zaledj une espèce de serpent qui attaque l'elephant et le buffle, et qui ne les abandonnent qu'après les avoir vaincu."[155] Artemidorus, also, according to Strabo,[156] "mentions serpents of thirty cubits in length, which can master elephants and bulls. In this he does not exaggerate; but the Indian and African serpents are of a more fabulous size, and are said to have grass growing on their backs." Iphicrates, according to Bryant, "related that in Mauritania there were dragons of such extent that grass grew upon their backs." It is doubtful whether large serpents, or real dragons, are referred to by Pliny in the following interesting passages which I give at length: the surprise which he expresses at Juba's believing that they had crests, leads me to suspect that there was possibly some confusion of species involved; that Juba might have been perfectly accurate so far as the crests are concerned, and that the beasts in question, in place of being pythons of magnitude, were rather some gigantic lizard-like creature, of great length and little bulk, corresponding with the Chinese idea of the dragon, and, therefore, naturally bearing horny crests, similar to those with which the monster is usually represented by the latter people. It must be noticed here, that if we postulate the existence of the dragon, we are not bound to limit ourselves to a single species, or even two, as the same causes which effected the gradual destruction of one would be exceedingly likely to effect that of another; we must not, therefore, be too critical in comparing descriptions of different authors in different countries and epochs, since they may refer only to allied, but not identical, animals.