It must be noticed that, though the members of a gens were of no recognizable kin to each other in one sense, yet they showed a certain solidarité-putting on mourning when one of the kin was in disgrace (Livy, vi. 20), sharing common religious rights peculiar to themselves, and at one time having a right to inherit property. All these things point to consciousness of distant blood-relationship. Still one feature of the ghotra seems absent. It is hardly proved that there was a time when Romans might not marry within the gens. Indications of the past existence of the rule are found in the fact that Roman genealogies do not, as it is said, show us examples of marriages between persons of the same gens. More to the point is Plutarch's statement (Hepi aitíwv 'Pwμaïkŵv), "In times past it was unlawful for Romans to marry women of their own kin (συγγενίδας); nay, they did not wed ladies in any degree connected with them by blood, just as now they do not take sisters or aunts, and it was long before they ventured to take cousins to wife." It seems then that, just as in the case of the contemporary ghotra of the Hindus, an ancient and wide prohibition to marry in the gens was thrown off by the Romans. Here it must be noted that the ghotra of the Hindu law of inheritance is not identical with the ghotra in which marriage is prohibited by custom. It is rather a body composed of all the cognates within certain large limitations.

In the example of the Greek yévos we again find the common name a patronymic, generally thought to be derived from a hero. We find that all who bore the name shared certain religious rights, and before Solon's date were co-heirs to property, and took up the blood feud if one of the yévos were slain. Yet the yevra are often defined as not akin in blood, so entirely did the old sense of relationship dwindle, in Greece as in Rome. The lexicographers supposed that the yen were constituted by legislative enactment, vóμw Tivi exovtes Koivwvíav. (See Meier, De Gentibus Atticis; Philippi, Der Areopag und die Epheten, Berlin, 1874, p. 68; Schoemann, Griechische Alterthümer, Berlin, 1861, vol. i. p. 329, with Schoemann's theory of the growth of the yévos; F. Haase, Die Athenische Stammverfassung; also Grote's History of Greece, iii. 53.) Now, hard as it is to ascertain the exact nature of the yévos, and of its relation to the tribe, it seems, on the whole, more analogous to the totem-kin than to the caste or joint family of the modern Hindus. (See Sir Henry Maine, "South Slavonians and Rajpoots," Nineteenth Century, December, 1877.) A common name, co-heirship, the duty of avenging a member, all point to the idea of kinship. As to exogamy, a Greek could certainly marry in his own yévos, for the common name went by the father's side, and a Greek might marry his father's though not his mother's daughter. It has been argued that the prohibition to marry a uterine sister, though kinship in historic Greece went by the male line, indicates a past when the maternal strict,-when, in fact, a man who married his uterine sister married within the yévos, and a man who took his half-sister by the father's side married outside the yévos. Here it may be observed that Aristotle (Pol., 1, 2, 5, 6) gives as very ancient synonyms of yevra the terms duoyáλakres (nourished on the same milk), dμooivo (eating from the same vessel), oμoкávot (warmed by the same fire). These terms speak of a time when motherhood or fosterage, when community of shelter, not blood kinship, were the bonds that kept members of the same kin together. The words may be compared with Gaelic teadhloch and coediche, "Gaelic names for family, signifying, the first, having a common residence, the second those who eat together" (M'Lennan, Prim. Mar., p. 154).

It has been usual, almost universal, to explain the Greek yévos and Roman gens by simply saying, like Mr Freeman

(Comparative Politics, Macmillan, 1873, p. 111), "The family grew into the clan, the clan grew into the tribe." Mr Freeman says we can trace this process best "among men of our own blood." But when we examine the early associations of the English (Kemble's Saxons in England, vol. i. p. 458), we find, just as in America, just as in Australia, groups of kindred of the same name, take Billing, by way of example,-scattered from north to south through all the local tribes. We have seen how this happens in America and Australia, we have seen that there the family, in Mr Freeman's sense, does not grow into the clan. Did it do so in Attica and Italy, and, if so, how did a tribe, which was ex hypothesi but a swollen clan, contain so many stocks which claimed distinct origin and distinct mythical ancestors? How did these stocks come to be scattered through local tribes, not grouped in one? The growth of savage tribes is not a development of the family; tribes singularly like those of savages are found in early civilizations. Had the two kinds of kindred different origins?

There remains a point to notice. The thoroughly savage totem-kindreds revere the animal, plant, or other object from which they take their name and claim descent, and they use it as a badge. For Greek and Roman survivals of this usage see Plutarch, Theseus; M'Leunan, "The worship of plants and animals," in the Fortnightly Review, 1869, 1870; and the Antiquities of Heraldry, by W. S. Ellis, 1869. If the ordinary theory, that the tribe and clan are overgrown families, be rejected, the converse theory may be stated thus:-The totem-kindreds of savages grow up through exogamy and female kin. The change to male kinship (a change which is demonstrably taking place in America and Australia) produced something like the Chinese circle of relationship. The substitution of the name of a fictitious ancestor for that of the sacred plant, animal, or natural object produced a circle of affinity like the Hindu ghotra of customary religion.1 The decay of the prohibition to marry within the kin united by the family name, like the growing laxity of rule in the ghotra, produced something like the Greek yévos and the Roman gens. Nothing remained but joint religious rites, a common place of burial, a common name, a vague feeling of connexion, traditions of the prohibition to marry within the gens, the duty of taking up the blood-feud, and vestiges of the jointheirship. In process of time the intenser affections of the family caused the old gentile ties to disappear, and gentile law became an empty memory.

It has been shown that arrangements ruder than the modern family exist among contemporary savages, and have existed among ancient peoples. It has been shown that these rude institutions produce large associations of men, tribes and totem-kindred, among savages, and that, by a series of changes, every one of which is exemplified in experience, the Greek and Roman gentes, the units of early political society, may have been developed out of barbarous groups. There are next certain customs to be examined, which tend, as far as they go, to show that civilized society passed through savage stages. The chief of these customs are the ceremony of capture and bridal etiquette. As to the ceremony of capture it is superfluous to say much, as the subject has been handled, with complete originality and copious illustrations, in M'Lennan's Primitive Marriage. The classic example of the ceremony of capture is thus stated by C. O. Müller, (History and Antiquities of the Doric Race, English translation, Oxford, 1830, vol. ii. p. 298): "Two things were requisite as an introduction and preparation to

1 We have examples in Zulu-land of the declining belief in animal

ancestry (Callaway's Religion of the Amazulu), and in Greek history we have frequent instances of the fictitious adoption of eponymous heroic ancestors.

marriage at Sparta: first, betrothing on the part of the father; secondly, the seizure of the bride. The latter was clearly an ancient national custom." Müller then describes the clandestine intercourse, which lasted for some time, before the man "brought his bride, and frequently her mother, into his house." The intercourse of bride and groom among the Iroquois of Lafitau's time was likewise clandestine. For the practice in Crete Müller quotes Strabo, x. 482, d. A similar custom prevailed in Rome (Apuleius, De As. Aur. iv.; Festus, s. v. Rapi"), and was supposed to be derived from the time of the rape of the Sabines. Mr M'Lennan finds the practice necessary to the constitution of the relations of husband and wife among the Calmucks, the Tunguzians, the Khonds, the Fuegians, the Welsh, the Arabs, the Irish, and various other races. He explains its existence by the institutions of exogamy (i.e., the rule prohibiting marriage between people of the same blood), and by the prevalence of hostility between the tribes of rude times. Suppose the rule to exist that a man may not marry a woman of his own community, and suppose that, by an exhaustive division, all other communities without exception are hostile, he must steal a wife if he is to marry at all. The fiction of capture, as men grow more polite, will endure as part of the marriage ceremony when the need of the reality is passed. It is to be noticed that the theft of the woman is, in the fictitious capture, generally the work of more than one man, as it well might be, if the early marriages were polyandrous. If it be granted that the prohibition to marry within the community is as early as it is widely prevalent, this explanation of the form of capture will seem sufficient. The origin of the early prohibition will be discussed later. Thus, on the evidence of a sportive feature in the marriage ceremony of civilized peoples, a vestige is revealed of customs connected with a very early form of the family.

A strange piece of barbarous etiquette may hint that the kindred of the bride and groom were once hostile groups. The daughter-in-law, among many races, is forbidden to speak to her father-in-law; the mother-in-law must hide when she sees her son-in-law. The wives treat their husbands with what may be a survival of hostility, and never name them by their names. Examples are collected in Sir John Lubbock's Origin of Civilization, pp. 11, 12. The practices are found among races on the border of the Polar Sea, in the Rocky Mountains, in Southern Africa, among the Caribs, Mongols, and Calmucks, in China, in Siberia, and in Australia. To these instances adduced by Sir John Lubbock we may add Bulgaria (Dozon, Chants Populaires Bulgares).

Herodotus says (i. 146) that the wives of the early Ionians would not call their husbands by their names nor sit at meat with them, and instructed their daughters to practise the same reserve. The reason assigned is that the women were originally Carians, whose parents the Ionians had slain. It may be allowed that this world-wide practice, too, testifies to a time when men married out of their own group, and all groups were hostile each to the other. Perhaps the English local custom, which forbids the parents of bride and bridegroom to be present at the marriage ceremony, holds the same antiquity.

We have now to note the widespread existence of a system of nomenclature, which can hardly have arisen in times when the monogamous family was the unit of society. Mr Lewis Morgan of New York was the discoverer of a custom very important in its bearing on the history of society. In about two-thirds of the globe - persons in addressing a kinsman do not discriminate between grades of relationship. All these grades are merged in large categories. Thus, in what Mr Morgan calls the "Malayan system,' ," "all consanguinei, near or far, fall within one of

these relationships-grand-parent, parent, brother, sister, child, and grandchild." No other blood-relationships are recognized (Ancient Society, p. 385). This at once reminds us of the Platonic Republic. "We devised means that no one should ever be able to know his own child, but that all should imagine themselves to be of one family, and should regard as brothers and sisters those who were within a certain limit of age; and those who were of an elder generation they were to regard as parents and grand-parents, and those who were of a younger generation as children and grandchildren (Timæus, 18, Jowett's translation, first edition, vol. ii., 1871). This system prevails in the Polynesian groups, and in New Zealand. Next comes what Mr Morgan chooses to call the Turanian system. "It was universal among the North American aborigines," whom Mr Morgan styles Ganowanians. "Traces of it have been found in parts of Africa" (Ancient Society, p. 386), and "it still prevails in South India among the Hindus, who speak the Dravidian language," and also in North India, among other Hindus. The system, as Mr Morgan says, "is simply stupendous." It is not exactly the same among all his miscellaneous "Turanians," but, on the whole, assumes the following shapes. Suppose the speaker to be a male, he will style his nephew and niece in the male line, his brother's children, "son" and "daughter," and his grand-nephews and grand-nieces in the male line, "grandson" and "granddaughter." Here the Turanian and the Malayan systems agree. But change the sex; let the male speaker address his nephews and nieces in the female line,-the children of his sister,-he salutes them as "nephew " and "niece," and they hail him as "uncle." Now, in the Malay system, nephews and nieces on both sides, brother's children or sisters, are alike named "children" of the uncle. If the speaker be a female, using the Turanian style, these terms are reversed. Her sister's sons and daughters are saluted by her as "son" and daughter," her brother's children she calls "nephew" and "niece." Yet the children of the persons thus styled "nephew" and "niece" are not recognized in conversation as "grand-nephew" and "grandniece," but as "grandson" and "grand-daughter." It is impossible here to do more than indicate these features of the classificatory nomenclature, from which the others may be inferred. The reader is referred for particulars to Mr Morgan's great work, Systems of Consanguinity and Affinity of the Human Race (Washington, 1871).

The existence of the classificatory system is not an entirely novel discovery. Nicolaus Damascenus, one of the inquirers into early society, who lived in the first century of our era, noticed this mode of address among the Galactophagi. Lafitau found it among the Iroquois. To Mr Morgan's perception of the importance of the facts, and to his energetic collection of reports, we owe our knowledge of the wide prevalence of the system. From an examination of the degrees of kindred which seem to be indicated by the "Malayan" and "Turanian" modes of address, Mr Morgan has worked out a theory of the evolution of the modern family. A brief comparison of this with other modern theories will close our account of the family. The main points of the theory are shortly stated in Systems of Consanguinity, &c., pp. 487, 493, and in Ancient Society, p. 384. From the latter work we quote the following description of the five different and successive forms of the family:

"I. The Consanguine Family. It was founded upon the intermarriage of brothers and sisters, own and collateral, in a group. "II. The Punaluan Family. It was founded upon the intermarriage of several sisters, own and collateral, with each others' husbands, in a group, -the joint husbands not being necessarily kinsmen of each other; also, on the intermarriage of several brothers, own and collateral, with each others' wives in a group,these wives not being necessarily of kin to each other, although

looked upon with such contempt that they kill all the native females and buy wives from their neighbours." Now, in each group, by the system of capture, are members of alien groups, namely, the women and their children, who, as we have seen, are recognized as connexions of the mothers, not the fathers. Let these practices be formed into customary law, refuse a man permission to marry a woman of his own stock-name (marked by the totem), and you have exogamy, or what Mr Morgan calls "the gentile organization." Within the groups are several families of the earliest type, the female and her offspring. Next, conceive of the sets of mother's sons, as feeling a stronger bond of union between themselves and the other members of the group, and as living with their mother. They cannot marry their sisters (who are of the same name and totem as themselves), but they regard their sisters' children as their heirs. To their own putative children, they can only make presents inter vivos, and the sisters are wedded each to a set of men in the manner of the Nairs. But, as property was amassed, the brothers would prefer to keep their property in the hands of their putative children, and "there would be a disposition in favour of a system of marriage which would allow of the property passing to the brother's own children" (Prim. Mar., 242). This form of marriage would be the one prevailing in Thibet. The elder brother, the first to marry, would have some of the attributes of the paterfamilias. Thus the idea of fatherhood attained something like maturity. Chiefs, moreover, would secure one or more wives to themselves, and their example imitated would produce monandry. The old state of things would leave its trace in the levirate, the duty of "raising up seed to a brother." Even before these changes, the custom of marrying out of the group would have introduced so many strangers of various names and totems, that the members of a local tribe could intermarry with one another and yet not violate the law of exogamy. Such a local tribe, flushed with success in war, might refuse to marry beyond its limits, and become, so to say, a caste divided into ghotras. Let this caste feign itself to be descended from a common ancestor (a process of which Sir John Lubbock gives many examples), and you have a caste believed to be of common blood, yet refusing to marry outside the blood,—that is, an endogamous tribe. Within this tribe (as it were by a reaction from the old kinship through females) grows kinship through males only, the agnatic system of Rome. The wife and children are the husband's property; agnates only can be a man's heirs and, failing these, gentiles, -i.e., members of the kin still denoted by the common

name.

66

Many criticisms have been made, especially by Sir John Lubbock and Mr Herbert Spencer (Origin of Civilization, third edition; Principles of Sociology, vol. i.), on the scheme here too briefly sketched. Sir John Lubbock holds that exogamy springs from marriage by capture (by which alone a member of a group could get a wife to himself), rather than marriage by capture from exogamy. Mr Spencer advocates the action of various conspiring causes,' "the stealing of a wife might become the required proof of fitness to have one (op. cit., pp. 652, 653). The origin of exogamy lies so far behind us in the past that it inay remain for ever obscure. It is probable that every variety of union of the sexes has existed, while it seems possible that a few have been passed through, as necessary stages, by all advancing races. In this notice we have said little of the custom by which a man is a member of several clubs of men, each with one wife in common.

No hard and fast theory is likely to be accepted as more than provisional in the present state of knowledge, when science has only for a few years been busily occupied in this investigation. (A. L.)

FAMINES. War, pestilence, and famine are regarded by many as the natural enemies of the human race; but in truth these are all more or less associated with the circumstances of civilization. In the highest state of civilized society there ought to be no war; there need be no pestilence; and famine alone would stand as being beyond the range of human prevention-subject to some conditions to be afterwards spoken of. The advancement in the social scale to a state of dependence upon cereal crops, while the facilities of intercommunication between different countries, or even parts of the same country, remained imperfect, led almost necessarily to the periodical recurrence of scarcity. Cereal crops are especially dependent upon conditions of climate for their regular production; and here at least are circumstances beyond human control.

In a matter of such practical importance as the failure of the regular supply of the food of the people, it is not desirable to rely upon merely theoretical surmises; nor is it necessary to do so. A table has recently been prepared1 of over 350 famines which have occurred in the history of the world, beginning with those spoken of in the Scriptures as having been in Palestine and in the neighbouring nations in the time of Abraham (Gen. xii. 10), and again in the time of Isaac (Gen. xxvi. 1); passing on to the seven years' famine in Egypt down through those which afflicted ancient Rome; enumerating in their order those which have visited the three divisions of the United Kingdom (by far the most numerous in the table-the records being more available), as also those devastating Europe in the Middle Ages; reviewing in special detail the 34 famines which have visited India, including, as the first recorded of this group, that of 1769-70 (above 20 have been on a large scale); and concluding with that terrible calamity which is now ravaging the populations in North China. It is not pretended that this table is entirely exhaustive. It is known that many famines have occurred in the Chinese empire of which no details have been found available; and it is supposed that many have desolated Persia and other portions of Asia of which exact particulars are not available. But as this is believed to be the only existing table of the kind, and as great pains have been taken to make it complete, it may for our present purpose be regarded at least as representative. We proceed then to an analysis of it, in view of ascertaining what have been the causes of famines,-a point of the first importance when we come to a consideration of the problem which will naturally force itself into prominence can anything be done to avert these national calamities?

The analysis discloses the following causes, or we may perhaps more accurately say attributed causes-for in this matter we have to follow the authority of the original chronicle, or of such records as have reached us :-1, rain; 2, frost; 3, drought; 4, other meteorological phenomena; 5, insects and vermin; 6, war; 7, defective agriculture; 8, defective transport; 9, legislative interference; 10, currency restrictions; 11, speculation; 12, misapplication of grain. These causes are named, as far as may be, in the order of their importance. It is immediately noticeable that they form themselves into the two distinct groups of natural and artificial causes.

We proceed to consider the first group-natural causes of famines.

1. Rain.-By excess of rain floods are produced, the soil becomes saturated, and decomposition of the seed is occa sioned. In hilly countries the seed is not unfrequently washed entirely out of the ground, and so is destroyed. This cause of famine applies in a marked manner to

1 See Statistical Journal, vol. xli., paper by Mr Cornelius Walford, F.I.A., F.S.S., &c., " On the Famines of the World, past and present."