Kalmucks are content to cover the corpse with something red whilst it is awaiting burial.1 And, according to Count Goblet d'Alviella,2 "in certain graves, the earliest of which go back to the reindeer age (those of Mentone, for example), the bones of the dead are painted red with oligist or cinnabar; and in our own day some of the North American tribes, who expose their dead on trees, collect the naked bones and paint them red before finally burying them. An analogous custom has been observed amongst the Mincopies of the Andaman Islands and the Niams of Central Africa." The feeling towards the dead in all these examplesexamples which a learned anthropologist would with ease, I am convinced, have made many times as numerous—is or in all cases may be that prompted by the affection, parental, filial, conjugal, which was even more necessary for the selfpreservation of the human race in the earliest days than it is in civilised times. But it is not here suggested that love was the only feeling ever felt for the deceased. On the contrary, it is admitted that fear of the dead was and is equally widespread, and is equally "natural.” What inference, then, is to be drawn from these two sets of apparently opposed facts, or what explanation is to be given of them? To this question the right answer is given both by savages themselves and by careful observers of savage modes of thought. Kubary, long a resident in the Pelew Islands, says the islanders "are only afraid of ghosts of strangers, as they are safe from the ghosts of their own people because of the good understanding which exists between the family and its own ghosts." So on the Gold Coast, though the spirit of the dead man wanders about, if homeless, doing good or evil according to his disposition, it is to his own family that he does good. "Black people," said a Zulu, "do not worship all Amatongo indifferently, that is, all the dead of their tribe. But their father whom they knew is the head by whom they begin and end in their prayer, for they know him best, and his love for his children;

1 Bastian, Oest. Asien, vi. 607.

2 Hibbert Lecture, 17, referring to Castailhac, La France prehistorique, 292. 3 In Allerlei aus Volks und Menschenkunde, i. 10.

+ Ellis, Ewe-speaking Peoples, 102.

they remember his kindness to them whilst he was living; they compare his treatment of them whilst he was living, support themselves by it, and say, 'He will still treat us in the same way now he is dead. We do not know why he should regard others beside us; he will regard us only.'" In fine, as we might reasonably expect, the man who was loved during his lifetime did not immediately cease to be loved even by savages, when he died, nor was he who was feared in life less feared when dead.

In primitive societies there is no state or central power administering justice between its members and protecting them from external aggression. The only bond which unites the society is the tie of blood. The individual exists only as

a member of a family or clan, and only so far as it supports and protects him. The survival of the race thus depends on the ready and effective aid rendered by the clan to its members. Consequently the individual's only friends are his clansmen, and "stranger" means are philologically the same word.

enemy "guest and hostis Nor does a man cease to

be a member of his clan when or because he dies. On the contrary, his claims on his clansmen may then become more sacred and more exacting than ever, for if he has been murdered they must avenge him at all costs. It is then quite intelligible that strangers, who as strangers were enemies while alive, should continue to be hostile after death; and that clansmen, especially "the father whom they knew," should both show and receive the loving-kindness which during their lifetime marked their relations with their fellow-members.

The object of this chapter was to conjecture what there was in the daily experience of the earliest form of society which may have suggested the possibility of maintaining permanently friendly relations with some of the spirits by which primitive man was surrounded and by which his fortunes were influenced. The conjecture offered is that he was ordinarily and naturally engaged in maintaining such relations with the spirits of his deceased clansmen; that he was necessarily led to such relations by the operation of

1 Callaway, Religious System of Amazulu, part ii., quoted by Tylor, Prim. Cult. ii. 116.

those natural affections which, owing to the prolonged, helpless infancy of the human being, were indispensable to the survival of the human race; and that the relations of the living clansman with the dead offered the type and pattern, in part, though only in part, of the relations to be established with other, more powerful, spirits.

The reader will already have noticed if not, his attention is now drawn to the fact that hitherto, with the exception of the last quotation (that referring to the Zulus) no mention has been made of ancestor-worship. The reason is not merely that ancestor-worship may be and is explained -erroneously, in the opinion of the writer of these lines-as due in its origin solely to fear, like all worship; but that ancestor - worship implies a belief on the part of the worshipper that the spirit worshipped is a supernatural spirit. Now, according to the thesis set forth in the previous chapters, not all spirits are necessarily supernatural spirits; the man who believes the bowing tree or the leaping flame to be a living thing like himself, does not therefore believe it to be a supernatural being-rather, so far as it is like himself, it, like himself, is not supernatural, for we have seen reason to reject the conjecture that man began by thinking he himself possessed supernatural powers. With this distinction between spirits and supernatural spirits, it has not been necessary for the purpose of this chapter to assume that the spirits of the dead possessed in the earliest form of society that power of thwarting man's best-grounded anticipations, which is of the essence of supernatural power. There may indeed be no à priori reason why man, when casting round for the source of this mysterious, supernatural interference with natural laws, should not have found it in the action of the spirits of the dead as well as in that of any other class of spirits. And, as a matter of fact, in some religious systems the spirits of the dead are credited with supernatural powers, though, it must be remarked, their powers are not by any means so great as those of the national or local gods, and the general feeling is that it is the dead who are dependent on the living for their comfort and even for their continued existence, rather than vice versa; in Egypt the ka was annihilated, if the survivors did not

embalm the body of the deceased and make images of the dead man; in Greece and Aryan India the main motive for marriage was, and in China is, anxiety to provide descendants competent to continue the rites on which the post-mortem welfare of the deceased depends; and amongst savages generally the belief is that the dead stand in actual need of the food that is offered to them. But, as a matter of fact, there are grounds for believing that it was to another quarter altogether than ancestral spirits that man looked in his attempts to locate the supernatural in the external world. This point will be fully discussed in a later chapter.

In the next place, if, as is here argued, man's communion with the spirits of his dead suggested the possibility of communication with other and supernatural spirits, then it is intelligible that, if ever the ritual for approaching both classes of spirits came to be the same, the similarity would eventually react to the advantage and increased honour of the spirits of the dead. The acts which constituted worship in the case of the supernatural spirit would not differ from those in which affection for a deceased father found its natural expression; and consequently, not differing, would come to be worship in the case of the deceased ancestor also. Thus, on this guess, ancestor-worship is secondary on and a by-product of the act of worship in the proper sense (i.e. the worship of a god).

To restate the argument: (1) The family-feast held immediately after the death of the deceased and repeated at intervals afterwards, and the other offerings of food to the deceased, are not originally acts of worship; (2) the same sort of offerings and festivals come to be employed in the case of supernatural spirits and to constitute the (external) worship of those spirits; (3) the offerings to the spirits of the dead then become ancestor-worship. This argument

depends for its validity largely on the identity, here alleged, of the ritual for approaching spirits of the dead and supernatural spirits. The identity cannot be exhibited fully until the act of "worship" in the proper sense has been-as in a later chapter it will be-fully set forth; and the reader is accordingly requested to suspend his final judgment on the question till the full evidence is before him. There are, however, some outstanding points to consider before we can

proceed to consider this evidence. For instance, it will have struck some readers as a serious omission that no reference has been made in this discussion to the "uncleanness" which is very generally, if not universally, considered to attach to a corpse and to all who come in contact with it-an omission all the more serious because this "taboo" has been explained as due to fear lest the spirit of the deceased should lodge on the person who touches the dead body.1 The omission, however, has been intentional, and the reasons for it are twofold. First, whatever the theory of this taboo, in practice the taboo may and does coexist with love for and confidence in the spirit of the deceased. Thus amongst the Pelew Islanders, who, as has been said already, have no fear of the ghosts of their own people, "because of the good understanding which exists between the family and its own ghosts," the relatives of the deceased are "unclean" for several days.2 In Samoa, where the natural affection for the deceased finds touching expression, 'those who attended the deceased were most careful not to handle food, and for days were fed by others as if they were helpless infants . . . fasting was common at such times, and they who did so ate nothing during the day, but had a meal at night; reminding us," says the Rev. G. Turner, "of what David said when mourning the death of Abner: So do God to me and more also, if I taste bread or ought else till the sun be down.' The fifth day was a day of purification.' They bathed the face and hands with hot water, and then they were clean,' and resumed the usual time and mode of eating." On the Gold Coast, where the wives of the deceased try to tempt his soul to return by offering him his favourite dish, "those persons who have touched the corpse are considered unclean; and after the interment, they go in procession to the nearest well or brook, and sprinkle themselves with water, which is the ordinary native mode of purification." In ancient Greece, also, where ancestors were worshipped, the relatives were tabooed.5 In 1I have not been able to see the paper in which this explanation is put forth; but cf. Frazer, Golden Bough, i. 154.

2 Kubary in Allerlei, i. 6.

Ellis, Tshi-speaking Peoples, 241.

5 See my paper,

4

3 Nineteen Years in Polynesia, 228.

"Funeral Laws and Folk-Lore in Greece," in the Classical

Review for June 1895, for instances.