(3) Sacramental eating of an animal.—The custom of "hunting the wren," found over the whole Celtic area, is connected with animal worship and may be totemistic in origin. In spite of its small size, the wren was known as the king of birds, and in the Isle of Man it was hunted and killed on Christmas or S. Stephen's day. The bird was carried in procession from door to door, to the accompaniment of a chant, and was then solemnly buried, dirges being sung. In some cases a feather was left at each house and carefully treasured, and there are traces of a custom of boiling and eating the bird.699 In Ireland, the hunt and procession were followed by a feast, the materials of which were collected from house to house, and a similar usage obtained in France, where the youth who killed the bird was called "king."700 In most of these districts it was considered unlucky or dangerous to kill the bird at any other time, yet it might be ceremonially killed once a year, the dead animal conferred luck, and was solemnly eaten or buried with signs of mourning. Similar customs with animals which are actually worshipped are found elsewhere,701 and they lend support to the idea that the Celts regarded the wren as a divine animal, or perhaps a totem animal, that it was necessary to slay it ritually, and to carry it round the houses of the community to obtain its divine influence, to eat it sacramentally or to bury it. Probably like customs were followed in the case of other animals,702 and these may have given rise to such stories as that of the eating of MacDatho's wonderful boar, as well as to myths which regarded certain animals, e.g. the swine, as the immortal food of the gods. Other examples of ritual survivals of such sacramental eating have already been noted, and it is not improbable that the eating of a sacred pastoral animal occurred at Samhain.
(4) Exogamy.—Exogamy and the counting of descent through the mother are closely connected with totemism, and some traces of both are found among the Celts. Among the Picts, who were, perhaps, a Celtic group of the Brythonic stock, these customs survived in the royal house. The kingship passed to a brother of the king by the same mother, or to a sister's son, while the king's father was never king and was frequently a "foreigner." Similar rules of succession prevailed in early Aryan royal houses—Greek and Roman,—and may, as Dr. Stokes thought, have existed at Tara in Ireland, while in a Fian tale of Oisin he marries the daughter of the king of Tír na n-Og, and succeeds him as king partly for that reason, and partly because he had beaten him in the annual race for the kingship.703 Such an athletic contest for the kingship was known in early Greece, and this tale may support the theory of the Celtic priest-kingship, the holder of the office retaining it as long as he was not defeated or slain. Traces of succession through a sister's son are found in the Mabinogion, and Livy describes how the mythic Celtic king Ambicatus sent not his own but his sister's sons to found new kingdoms.704 Irish and Welsh divine and heroic groups are named after the mother, not the father—the children of Danu and of Dôn, and the men of Domnu. Anu is mother of the gods, Buanann of heroes. The eponymous ancestor of the Scots is a woman, Scota, and the earliest colonisers of Ireland are women, not men. In the sagas gods and heroes have frequently a matronymic, and the father's name is omitted—Lug mac Ethnend, Conchobar mac Nessa, Indech, son of De Domnann, Corpre, son of Etain, and others. Perhaps parallel to this is the custom of calling men after their wives—e.g. the son of Fergus is Fer Tlachtga, Tlachtga's husband.705 In the sagas, females (goddesses and heroines) have a high place accorded to them, and frequently choose their own lovers or husbands—customs suggestive of the matriarchate. Thus what was once a general practice was later confined to the royal house or told of divine or heroic personages. Possibly certain cases of incest may really be exaggerated accounts of misunderstood unions once permissible by totemic law. Cæsar speaks of British polyandry, brothers, sons, and fathers sharing a wife in common.706 Strabo speaks of Irish unions with mothers and sisters, perhaps referring not to actual practice but to reports of saga tales of incest.707 Dio Cassius speaks of community of wives among the Caledonians and Meatæ, and Jerome says much the same of the Scoti and Atecotti.708 These notices, with the exception of Cæsar's, are vague, yet they refer to marriage customs different from those known to their reporters. In Irish sagas incest legends circle round the descendants of Etain—fathers unite with daughters, a son with his mother, a woman has a son by her three brothers (just as Ecne was son of Brian, Iuchar, and Iucharba), and is also mother of Crimthan by that son.709 Brother and sister unions occur both in Irish and Welsh story.710
In these cases incest with a mother cannot be explained by totemic usage, but the cases may be distorted reminiscences of what might occur under totemism, namely, a son taking the wives of his father other than his own mother, when those were of a different totem from his own. Under totemism, brothers and sisters by different mothers having different totems, might possibly unite, and such unions are found in many mythologies. Later, when totemism passed away, the unions, regarded with horror, would be supposed to take place between children by the same mother. According to totem law, a father might unite with his daughter, since she was of her mother's totem, but in practice this was frowned upon. Polygamy also may co-exist with totemism, and of course involves the counting of descent through the mother as a rule. If, as is suggested by the "debility" of the Ultonians, and by other evidence, the couvade was a Celtic institution, this would also point to the existence of the matriarchate with the Celts. To explain all this as pre-Aryan, or to say that the classical notices refer to non-Aryan tribes and that the evidence in the Irish sagas only shows that the Celts had been influenced by the customs of aboriginal tribes among whom they lived,711 is to neglect the fact that the customs are closely bound up with Celtic life, while it leaves unexplained the influence of such customs upon a people whose own customs, according to this theory, were so totally different. The evidence, taken as a whole, points to the existence of totemism among the early Celts, or, at all events, of the elements which elsewhere compose it.
* * * * *
Celtic animal worship dates back to the primitive hunting and pastoral period, when men worshipped the animals which they hunted or reared. They may have apologised to the animal hunted and slain—a form of worship, or, where animals were not hunted or were reared and worshipped, one of them may have been slain annually and eaten to obtain its divine power. Care was taken to preserve certain sacred animals which were not hunted, and this led to domestication, the abstinence of earlier generations leading to an increased food supply at a later time, when domesticated animals were freely slain. But the earlier sacramental slaying of such animals survived in the religious aspect of their slaughter at the beginning of winter.712 The cult of animals was also connected with totemic usage, though at a later stage this cult was replaced by that of anthropomorphic divinities, with the older divine animals as their symbols, sacrificial victims, and the like. This evolution now led to the removal of restrictions upon slaying and eating the animals. On the other hand, the more primitive animal cults