Both Yaos and Anyanja trace descent through the mother, and cannot marry within their own clan, which is, of course, the mother’s. The Yao clans are still clearly known and named. Mr. R. S. Hynde says: ‘The Yaos are divided among themselves into sub-tribes, stocks, or totemistic clans,[30] each with its own distinctive name, e.g. the Amwale, the Asomba, the Apiri clan. If you question them on the subject, they will usually be able to tell you this clan name, unless the person questioned be a slave, who, from various causes, may not know it.’ Somba means ‘a fish,’ and mwale ‘a girl’; piri in Nyanja is ‘a mountain,’ though I have been unable to discover its significance as a Yao word. Thus, if Mwepeta, of the Somba clan, marries Ndiagani, of the Mwale, their children will be Mwale, and none of them can marry a Mwale. Their nearest relation and natural guardian will be their mother’s brother, who (if the grandfather is dead) will be head of the family. But his children will be no relatives of theirs, as they belong to their mother’s clan, and she (by the rule) cannot be a Mwale. Any of Mwepeta’s sons or daughters will therefore be free to marry any of these cousins; but they could not marry the children of Ndiagani’s sisters, who would be Mwale.

Some tribes of Anyanja east of the Lake have, in addition to this, a system of agnatic descent, through the father, called chilawa. This may be borrowed from the Zulus (who always count descent in this way, though the importance assigned to the maternal uncle is probably a survival of a former state of things), and the more southern tribes call it ‘the Angoni system.’ These last do not appear to have surnames; but those who recognise chilawa do; these names descend in the male line, and show at once to what family a man belongs on his father’s side. ‘Although a person’s surname is not generally known to those who are not his near relations or intimate acquaintances, because, so to speak, he does not make personal use of it, and is not called by it, yet every one knows his own surname, and is ready to give it at once, if asked for it.’ The late Bishop Maples, who put the above facts on record, was of opinion that, in spite of the close relationship existing between the sister’s son and the mother’s brother, the father is really the head of the Nyanja family, and arrived at the conclusion that a distinction must be made between ‘kin’ and ‘blood’: ‘the mother preserves to her offspring the tie of kinship, the father that of blood.’

The Yao chief, Kapeni, belonged to the Abanda clan, and was succeeded on his death, not by any of his sons (who, of course, were not Abanda), but by the son of his sister, born of the same mother. Had any younger sons of his mother survived, they would have had the preference; but half-brothers or sisters (children of the same father, but not of the same mother), are not counted as relatives—except that they cannot marry. According to the Yao system of descent, a man should be able to marry his father’s sisters, but this is seldom done, and is, in fact, considered very wrong; but he may marry their daughters. Where chilawa prevails, however, these too are forbidden—they are really reckoned as sisters.

Native terms of relationship are often very puzzling. Mbale is a word which may be applied to a brother, sister, cousin, or relative of almost any sort—sometimes even to a friend. There is no word for ‘sister’ or ‘brother’; but there are words meaning ‘elder brother (or sister)’ and ‘younger brother (or sister)’; and these are never used apart from their possessive pronouns. There is a word which means ‘sister’ when used by a brother, and ‘brother’ when used by a sister, but is never applied to one of the same sex as the speaker. A man will call all his father’s brothers ‘father,’ and all his mother’s sisters ‘mother’; and the term ‘grandparents’ may include all the great-uncles and great-aunts.

Mr. Duff Macdonald well shows the process by which a family may grow into a small state. A man wishing to found a new village asks permission of his chief—which in most cases is readily granted—and moves out into the bush with his wives and children. Temporary shelters are built, and then the man cuts down the trees, while his wives hoe up the ground for gardens; and, when these are ready, and planted, more permanent dwellings are erected. If there are daughters old enough to marry, the village is soon enlarged by the sons-in-law who come and build their huts there. The new chief may be accompanied by his younger brothers, or by friends who call themselves by that name, and place themselves under his authority. As the new settlement grows in power and importance, it will be joined by others, and may grow wealthy by trading.

In general, the Bantu have everywhere much the same system of government: the same features can nearly always be traced, even when modified by local circumstances. The Anyanja, when they first became known to Europeans, lived in small villages (as they do now), each under the control of its own head-man. A district, containing a large number of villages, was ruled by a sub-chief: such were Chinsunzi and Kankomba, in the Shiré Highlands, in 1861; and over the whole country was the Paramount Chief, or Rundo (Lundu), who at the same period was Mankokwe.[31] Mankokwe’s dominions appear to have extended from Lake Chilwa to the Shiré, and down the latter river as far as the Ruo; below the Ruo was another paramount chief, Tingani. Above the confluence of the Shiré and the Zambezi, between Kebrabasa and Zumbo, were two other independent Nyanja chiefs, Sandia and Mpende. All these chiefs seem at one time to have been ‘united under the government of their great chief, Undi, whose empire extended from Lake Shirwa to the River Loangwa; but after Undi’s death it fell to pieces, and a large portion of it on the Zambezi was absorbed by their powerful southern neighbours, the Banyai.’[32] This process has been repeated over and over again in the history of Africa. The Anyanja, being an agricultural (and on the whole a peaceable) people, kept up no national life outside their little village communities, but tended more and more to what German historians call Particularism. Consequently, they were unable to withstand the shock of an invasion; and their organisation, such as it was, went to pieces before the onslaught of Yaos, Makololo, and Angoni.

Women chiefs are mentioned several times by Livingstone as ruling in various parts of this region: Chikandakadzi, near Morambala (her position with regard to Tingani is not stated); Nyango, who seems to have ranked as Rundo in part of the Upper Shiré Valley, and Mamburuma, near Zumbo on the Zambezi; also Manenko and Nyamosana in the Lunda country. More recently, we find Nalolo, a sister of Liwanika, occupying the position of a chief in the Barotse country. The present Kazembe appears to be a woman. Sebituane, the Makololo chief, appointed his daughter as his successor, ‘probably,’ says Livingstone, ‘in imitation of some of the negro tribes with whom he had come into contact.’ She, however, soon resigned what proved a distasteful position; for her father, unwilling that she should transfer her power to a husband, directed her not to marry, but to contract any number of temporary alliances. It may be, however, that this feature of the situation was due, not so much to Sebituane’s Bechuana ideas of the husband necessarily being ‘the woman’s lord,’ as to some lingering Rotse and Lunda traditions of polyandry; and a writer in the Livingstonia Missionary Magazine characterises the present Kazembe as ‘a thoroughly bad woman—a woman of Samaria over again,’ which may be due to a misunderstanding of a very peculiar institution.

It is impossible not to connect these scattered indications with those afforded by the Yao system of kinship, and marriage customs, as to the state of things in an earlier period of which we have no record.

The Yao tribal organisation is in itself much the same as that of the Anyanja, but it was more closely knit, owing to the exigencies of war; and the relations of the conquerors to the conquered tribes must be distinguished from those which obtained among themselves. But it must be remembered that the Yaos were not an aggressive tribe, organised for conquest, under a chief like Tshaka or Mziligazi. In their own country—between Lake Chilwa and the Upper Rovuma—they seem to have been both a pastoral and an industrial people. ‘Yao-land proper,’ says Archdeacon Johnson, ‘had plenty of smelting-furnaces, cattle, and peas and beans, plenty for man and beast.’ They cultivated ‘down both sides of the Lujenda, till the valleys were full of Indian corn, and settlement extended its fields to those of the next settlement.’ The Machinga, who occupied this country, were dislodged by the Alolo (Makua) from the south-east, who themselves expelled from the north by the (Zulu) Magwangwara, drove them into the country of the Mangoche, forcing the latter into the Shiré Highlands. This was the so-called invasion of 1861.

There are five branches of the Yao nation: the Makale, near the sources of the Rovuma; the Namataka (or Mwembe people), on the hills west of the Lujenda; the Masaninga, Mangoche, and Machinga. The last three were the tribes who entered the Shiré Highlands. Their chiefs seem to have been quite independent of one another; Kapeni of Sochi was perhaps the most powerful.

The chieftainship is hereditary, and passes, as already stated, to the deceased’s younger brothers in succession, or, failing those, to the eldest son of his sister. The new chief takes, at the same time, his predecessor’s official name, so that there is always a Kapeni, or Malemya, or Mponda, as the case may be. The Angoni chiefs, however, observe the Zulu rule of inheritance, and are succeeded by the eldest son of the principal (or ‘official’) wife, who is the one married after accession—earlier ones do not count.

But there are some chiefs who have not succeeded to their position by right of birth, but attained it by superior cleverness and energy. Such a man may even be a slave, as was said to have been the case with Chibisa, who, in the early sixties, had much more real power than his neighbour Mankokwe, the Rundo of the Shiré Highlands. This man began by representing himself as possessed by the spirit of Chibisa, a deceased prophetess of note among the Nyungwe tribe, near Tete, whose name he assumed. The Nyungwe believed him, and he gradually obtained a complete ascendency over their chief, Kapichi, finally inducing part of the latter’s people to secede with him and settle at the foot of the Murchison Cataracts. His history may be read in the Rev. H. Rowley’s Story of the Universities’ Mission, where it is related how, at last, he fell in battle, fighting Terere, though without mention of the sand-bullet which killed him—the only thing against which he had no charm.

The unexpected rise of a man like this has often been the agency in breaking up Bantu ‘empires’ like that of Undi. But the new power is seldom permanent, as it does not often happen that such parvenu chiefs leave behind them successors of equal ability; while, having no backing but their own immediate followers, they lack that support of custom and tradition which in normal times will keep a mere average ruler in his place, so long as he does not forfeit it by any act of his own.

The customary order of succession is sometimes set aside, not so much by the tribe collectively (though it, too, being represented by the head-men, has a share in deciding the question), as by the household of the late chief. As both wives and slaves have a personal interest in the appointment of the successor, it is but just that they should have a chance to express their objections, if any. When Malemya of Zomba died, in 1878, his slaves, and many of his head-men, disliked the obvious heir, his younger brother Kumtaja, while the widows openly preferred a nephew, Kasabola. The head-men announced that, if Kumtaja were appointed, the people of the chief’s village would all leave and go to live elsewhere. Kasabola, accordingly, was installed, and took the name of Malemya, while Kumtaja left, taking with him such head-men as would go, and founded a new village not far off. Malemya, finding him an inconvenient neighbour, called in the Angoni, who came and raided Kumtaja’s village in August 1884. He fled first to Lake Chilwa, and then to the Upper Shiré, where he died some years ago.

When the new chief is appointed, some little time is allowed to pass before he is formally inducted. The day is then fixed for him to assume his official title (literally, ‘to enter the name’), after which his old name is never heard again. He is lectured on his duties to his people—which are held to consist chiefly in exercising hospitality, and not beating them too much; and, if he is a Yao of certain families, he is invested with the lisanda, a white head-band with hanging ends. Some Yao chiefs are not entitled to wear the lisanda; while, on the other hand, the right is enjoyed by some minor head-men who belong to the privileged families. It is henceforth worn on all solemn occasions—and sometimes at beer-drinkings—and the chief’s first appearance in it is hailed with songs of rejoicing. The proceedings, as might be expected, end with feasting.

On the Lake a special oblong house, with one side open, is built for the chief’s investiture. The insignia of royalty are here, a red blanket, and a red fez, called chisoti cha zindi—both probably imported.

The chief’s powers are not despotic;[33] he is not supposed to act without consulting his head-men, who represent the general views of the tribe; and he seldom disregards their opinion to any serious extent. Should he persist in doing so, his career would either come to a sudden and violent end, or his people would leave him to seek some more congenial ruler, and he would find himself lord of deserted villages. This is a recognised and constitutional remedy for grievances, and no chief refuses an asylum to such refugees; indeed, it is to his interest to welcome them. Fugitive slaves, on the other hand, are often returned to their masters.

The village head-man settles all local matters, usually with the assistance of the elders or heads of families, who are called his ‘younger brothers.’ He consults them before engaging in war, or undertaking any public work, such as constructing a stockade round the village; but he cannot summon them to work on his own private account, nor exact tribute from them. He settles any disputes among them, but if they are not satisfied with his decision, they can appeal to a higher head-man, or sub-chief, or to the chief himself.

Graver matters are reported by the head-man to the sub-chief, and by him, if necessary, to the chief. The latter holds the head-man responsible for any wrong-doing of his people which may come to his ears, just as the chief in his turn will be held responsible for any aggression of his head-men against outsiders. So far is this principle carried that, when a man has been injured by an inhabitant of a certain village, he and his friends are quite satisfied if they can catch any other man belonging to the same village, whom they will either put to death or hold to ransom till reparation is made.

The mlandu and the ordeal are the two great judicial institutions of Bantu Africa. With the ordeal we have partly dealt elsewhere, but there will be a little more to say about it presently.

Mlandu is a word which may be variously rendered as ‘lawsuit,’ ‘complaint,’ ‘discussion,’ ‘crime,’ and otherwise, according to the context. It is the same thing known as a ‘palaver’ in West Africa, and an indaba by the Zulus. Civil cases are thus settled. The head-man and his ‘younger brothers’ take their seats in the bwalo, and, as a rule, the whole village is assembled, the men sitting on one side, and the women—a little apart—on the other. The accuser speaks first, then the accused, and the various members of the council give their opinions in turn. The speeches are often long and eloquent, and the case may extend over days or even weeks before the head-man gives his decision, or, as the natives say, ‘cuts the case.’ If no decision is come to, or if either party wishes to appeal, the case is transferred to a higher court, and ‘the mlandu spoken’ before the sub-chief or the chief. An important case of this sort is sometimes attended by hundreds of people. The successful party in the suit makes the judge a present out of the damages. Matrimonial cases are settled before a court of this kind—if, for instance, a wife feels herself aggrieved and returns to her relatives. She is represented before the court by her ‘surety.’ The husband may also bring an action for divorce in this way.

Criminal charges, too, in the first instance, are brought before the chief’s or head-man’s court. A man caught stealing may, by native law, be killed, and his death entails no prosecution. He may be caught alive, and would then be put into a slave-stick for safe-keeping, till ransomed by his friends; and killed, or kept as a slave, if no ransom were forthcoming.

When a theft has been committed, without suspicion falling on any particular person, the diviner or the mabisalila is consulted, and the person pointed out by him or her accused before the court. The prosecutors demand restitution of the stolen goods; the defendant pleads not guilty, and offers to drink mwavi to prove it. His friends, if they believe him innocent, will demand the ordeal on his behalf; if they have misgivings, they will be afraid to run the risk, convinced, as they are, that the guilty party invariably dies, and knowing that, in such a case, they will have to pay the full value of whatever was stolen. If guilty, a man will probably confess rather than risk the ordeal—he, or his relations, will have to make restitution and pay a fine besides, the head-man of his village being held responsible. These payments also have to be made for him, if his confidence in the judgment of the ordeal turns out to have been misplaced—his death, in the native view, conclusively proving his guilt. If, on the other hand, he survives, the accusers have to pay over a fine to him, and the sorcerer is assumed to have been mistaken. Some try a second sorcerer, but he must not point out the man just acquitted, as no man can be made to drink mwavi twice on the same charge.

Theft, if brought into court at all, is always punished by a fine; but sometimes the thief is handed over as a slave to the injured party. Other ordeals are sometimes used besides the mwavi—plunging the hand into hot water, or touching red-hot iron—but the principle is the same: injury to the hand proves guilt. It will be noticed that the head-man is held responsible for thefts committed by his villagers, in accordance with the principle already stated. (He may, in fact, be the receiver of the stolen goods.) If he refuses to take the matter up when it is brought to his notice, war may be the result.

A murderer, if caught red-handed, may be killed by the friends of his victim; or they may put him in a slave-stick till slaves have been paid over for his ransom. Some of these are sometimes sacrificed to the manes of the victim. If the actual slayer cannot be caught, a man from the same village may be captured and held to ransom in the same way. In some cases, no distinction is made between accidental homicide and murder; sometimes a gun which has been the cause of an accident is seized instead of the owner, and held till he pays several slaves for it. Sometimes, however, the view is taken that the man or his gun may have been bewitched, and steps are taken to find out and punish the person who has done this. The man who kills his own slave, or even his younger brother, or other ward, is not amenable to justice, but—unless he can protect himself by a charm—he is afraid of the mysterious chirope which overtakes those who shed blood within the tribe. The chief, to whom he goes if he has committed such a murder, procures the charm for him from his own medicine-man, and uses it himself as well, ‘because of the blood that has been shed in his land.’

Adultery is theoretically a capital crime with most Bantu tribes; that is, the man may be (and frequently is) shot or speared by the husband; the wife is frequently let off with a warning the first time, but for a second offence either killed or divorced and sent back to her relatives, who in such a case must return whatever present was made at the marriage. Sometimes she drinks mwavi, and is, of course, accounted guilty if she dies. But in practice, the matter is often arranged by paying damages, or the guilty man may be sold into slavery. Still there can be no question that (where they have not been corrupted by outside influences, or their customs and institutions disorganised by war, etc.) they look on it as a very serious affair.

Slave wives are more summarily dealt with, and are often either killed or sold. I have seen one ‘Angoni’ woman who had had her nose and ears cut off, but seemed to be living on in her husband’s (or master’s) family, as before,—though evidently doing most of the heavy work. But such cases are not common, except perhaps among the coast Arabs, who have large slave harems and rule them by terror.

From what has already been said, it will be seen that in ordinary procedure formal executions are not common; if the criminal has not been killed red-handed, or if he does not undergo the ordeal, the trial usually ends in a fine, or in his being handed over as a slave to his accusers. But where the chief orders a man to be executed, he is usually stabbed or has his throat cut. Sometimes a wizard is shot at once on being detected by the Mabisalila, and sometimes, when convicted by the ordeal, the crowd fall upon him and lynch him without waiting for the poison to do its work: an outbreak of panic ferocity which has its parallel, in a more deliberate form, in the records of English and Scottish witch-trials.

Witches were in former times sometimes burned alive by the Yaos; but Mr. Macdonald says that in his time this was only done if they refused the mwavi test, which was not likely to happen. It was done also by the Anyanja at Likoma, and the stake where these executions took place stood on the site of the present Cathedral. But by far the greater number of cases were left to be decided by the issue of the poison.

At one time, the Yaos used to torture the person pointed out by the witch-detective by squeezing his head between two pieces of wood, till he pointed out where the horns were buried; but in later times the Mabisalila found the horns (as already described) as well as naming the witch. But I do not think that this, or judicial torture to extort confession of ordinary crime, is common.

An act of sacrilege held to be penal is when a free man sets fire to grass or reeds near a lake supposed to be the abode of a tutelary spirit, in which case he would be thrown into the lake. This perhaps would be rather a sacrifice than an execution.

Imprisonment as a punishment is scarcely known, and indeed scarcely possible, though, as has been said, men are sometimes detained in the slave-stick till ransomed or otherwise disposed of. It is a small log with a fork at one end, long enough for the other end to rest on the ground when a man’s neck is inserted into the fork and secured with an iron pin. Slaves are confined in these sticks on the march (as in the familiar picture in Livingstone’s Zambezi Expedition), or when they are likely to run away,—or sometimes as a punishment. Debtors, adulterers, and thieves may be put in the slave-stick till their debts or fines are paid up. Slaves, when thus confined, sometimes have the other end of the stick fastened up to a tree, so that they can do their usual work of pounding corn, or the like. There is also a form of stocks called in Yao ugwalata, consisting of a hole in the verandah-part of a hut, through which a man’s arm or leg is passed, and secured so that he cannot draw it back. Slaves are sometimes severely beaten.

The Makololo chiefs approach more nearly the idea of an irresponsible despot than any others in this part of Africa; but this is owing to a special set of circumstances, and they cannot be taken as typical. It is the more necessary to bear this in mind, because the chapter of horrors which Mr. Macdonald gives under the heading, ‘A Slave Government,’ may seem to contradict some of the statements we have made. These chiefs, then, were placed in a very exceptional position. They were a small minority of warriors in the midst of an unwarlike population whom they regarded with contempt, but who were strong enough to make them think they must secure their position by ruthless severity. Some of them had actually been slaves themselves, though the only Makololo among them, Ramakukane, was of good family. Cruelty is not a Makololo trait, though Sebituane and Sekeletu could act with firmness and even harshness when the occasion seemed to demand it. But some unusually barbarous punishments seem to have been used in the Barotse valley where the Makololo had settled, and to which some of these chiefs belonged by birth. They may have brought some of these customs with them; and their becoming possessed of virtually unlimited power, while at the same time their footing was but a precarious one, did the rest. None of them was subject to the others; they were far enough apart to be quite independent; but they acted together in face of a common enemy. It is a pleasing fiction that their despotism was on the whole of a benevolent character, and voluntarily submitted to by the Mang’anja, who welcomed them as protectors against the Yaos; but this illusion is dispelled by a closer acquaintance with the facts. Even before the departure of the Livingstone expedition, they had begun to tyrannise over the Shiré population; but it was only after that event that their power became fully established. Several of them were undoubtedly men of fine qualities; but a careful examination of their careers before and after (roughly speaking) 1861, leads to the conclusion that some at least must have degenerated sadly.

They took advantage of the famine of 1862-3 to enslave the Mang’anja, and ‘their power increased every day till they could claim all on the Lower Shiré for their subjects.’ They had no council of head-men, and though each village had a head-man, he was not a responsible local ruler, but a mere taskmaster appointed by the chief. Forced labour and oppressive tribute were exacted. No woman had a ‘surety,’ as with the Yaos, but the chief disposed at will of his subjects’ daughters—assigned them to husbands of his own choosing, or took them into his harem, as he felt disposed. Wholesale mwavi-drinkings took place, at which no one was allowed to refuse the cup; and judicial torture was frequent.

The ordinary tribute paid to chiefs varies in different tribes, but is not in general excessive. In some parts, when an elephant is killed, the chief claims ‘the ground tusk’—i.e. the one which touches the ground when it falls; elsewhere this is not insisted on. Presents are usually expected from strangers passing through the country, but they get something in return; and a chief (though in practice he may fall short of the ideal) is always supposed to be generous. Yao local head-men send their chief a percentage of the ivory when they kill elephants, and (if they live near enough) a haunch of any large animal (such as an eland) which they may shoot. It is also the custom for them to invite him to a beer-drinking at least once in the year. Sometimes the chief sends for the village head-men, or orders them to find men, to do some work for him at his village—hoeing, or building huts. This was frequently done by the Angoni chiefs, who also (as has been said before) made periodical levies of their subjects’ sons to herd their cattle, and of their daughters for the harem.

There is no regular priestly class. The professional diviners and medicine-men to a certain extent occupy the same position, and a Yao chief sometimes appoints a ‘sacrificer,’ whose duties are of a somewhat miscellaneous character. Besides taking the omens before a battle, he has to carry the banner and lead the army—the chief himself, like David in later life, not going into action. (He stays behind to ‘supply powder and deal with deserters.’) The ‘sacrificer’ tastes the beer offered to the chief’s guests, to show that it is not poisoned, and beats one of the drums at witch-dances, where he represents the chief, if the latter is unable to be present. Whether he is the same as the chief’s medicine-man is not clear.

But the strictly religious functions of a priest, as we have seen, are performed by the chief on behalf of his tribe, by the head-man for the village, by the father for the family, and (in private matters) by the individual for himself.

We have seen how the chief presides over, or at least takes part in, public prayers for rain; but the Yaos and Anyanja do not at present seem to have anything corresponding to the ‘feast of first-fruits’ among the southern Bantu, where the chief ceremonially ‘tastes’ the first of the new crops before the people are allowed to gather them. There are traces of such a rite among the Yaos, and I am inclined to think that the Angoni keep up something of the kind, or did a few years back, because I was informed at one of the Ntumbi kraals, about the beginning of the harvest season, that the father of the family was away at Chekusi’s, ‘eating maize,’—an expression of which I did not at the time grasp the probable bearing. The Zulus keep the ukutshwama with great solemnity, and the Angoni would have brought the custom with them from the south, though I do not know how they observe it in detail.

The chief is supposed to be the owner of all the land, but in strictness he cannot alienate it without the consent of the tribe. It seems, however, as if, apart from European or Arab influence, the idea of permanent property in land scarcely existed. No one is supposed to own land except so long as he actually cultivates it; and, owing to the method of agriculture, it is abandoned every few years. Any member of the tribe can make a fresh garden where he likes, provided no one else has bespoken the ground; but a stranger would require the chief’s permission to settle. The chief’s land is well defined, and has recognised boundaries, but there seem to be no definite limits to the territory occupied by a tribe.

The Mang’anja used to recognise certain animals as nyama ya lundu, ‘king’s meat,’ not allowed to be eaten by the people in general. Among these were the nkaka, or scaly ant-eater, whereof the Rev. D. C. Scott was on one occasion invited to partake by Ramakukane, and a certain kind of large frog or toad called tesi, said to be very delicate eating.

It remains to speak of slavery, which has always been, in varying proportions, a feature of Bantu society. The outside slave trade does not so much concern us here, as (in this part of Africa, at any rate) it is entirely an exotic thing, introduced and fostered by the Portuguese on one side and the Arabs on the other. And though this has been largely, if not wholly, done away with (it is certain that there was some smuggling going on, twelve years ago), yet domestic slavery, which is very difficult for governments to interfere with, still continues in fact, if not in name, and will only die out gradually. The proportion of slaves to free people is probably not large—unless all the Anyanja subject to the Angoni are counted as slaves—which is not, strictly speaking, correct; they are rather in the position of serfs or villeins.

Slavery may be a matter of birth; the children of slaves, or of a slave mother and free father, are slaves also. Some slaves are persons taken prisoners in war, or sold (probably to pay a debt) by father, grandfather, or elder brother. Others may have been condemned to slavery as criminals, or bewitchers, or possessed of ‘the evil eye,’ and these are sold to some one at a distance—to get rid of them. Or they may be seized on account of a debt they cannot pay; or, lastly, they may voluntarily become slaves, in time of famine, in order to get food.

The owner has the power of life and death over his slaves, but subject to the moral restraint already mentioned. Slaves may be beaten—sometimes cruelly—or confined in gori-sticks, at the will of their masters, but as a rule they are kindly treated, and, in fact, to an outsider, are often indistinguishable from the family. In speaking of or to them, the master says mwana (‘child’), or mnyamata (‘boy’), rather than kapolo (‘slave’). Some of the families at Nziza and Ntumbi had Yao slaves who must have been captured in the raids across the Shiré a few years before, and who seemed quite contented with their lot.

Slaves are employed about the usual work of a house and garden: the women are generally the master’s junior wives, and share the household labours among them; the men sometimes relieve them of part of the heavy work, such as pounding corn, or fetching wood and water, but are also engaged in more strictly masculine pursuits. They are supplied with guns and go out hunting; they spin, weave, sew, make baskets, etc.; and sometimes they are sent to carry loads for a trading party, or accompany their master to war. A man may even send a confidential slave to the coast to trade on his account. A chief often gives considerable authority to his principal slave, who may attain a position of great importance, and cases are not unknown where such a slave has become a chief.

People kidnapped from another tribe may be, and sometimes are, ransomed by their friends. After a fight it is common to send word that such and such prisoners have been taken, so that a ransom may be sent. It does not seem to be possible, in practice, for a slave to redeem himself; but once free, there are no special disabilities attaching to his position. A slave who runs away places himself under the protection of another master, if he can find one to shelter him; but if he can escape being caught, he may achieve freedom for himself, as, apparently, Chibisa did. But in general a masterless slave does not find the highroad a safe place, and hastens to put himself under some one’s protection.

Slaves are not distinguished by any special mark, badge, or dress. They may possess property (such as cloth, guns, and ivory), as their owner frequently allows them to keep part of what they earn. They may even, in some cases, own other slaves. The master gives the slave a wife—usually a slave woman, but occasionally he may let him marry his daughter. The case of a free woman marrying a slave husband is, however, rare; and he is likely to be superseded at any time.

On the death of a slave-owner, such of his slaves as are not chosen to accompany him (and this, as we have seen, is by no means universal) pass into the possession of his heir. If a slave dies possessed of property, it all goes to his master.

The Machinga, at Mponda’s on the Upper Shiré, made a raid on Ntumbi and Nziza, in May 1894, for the purpose of capturing women and children, but the men of the place frustrated this attempt, and took two prisoners, who were sent up to Chekusi’s, but released (I believe) after their guns had been taken from them. There is reason to believe that they had been more successful on previous occasions, not so very long before, and that the women in question had been smuggled across the Shiré, and, as there was no safe opportunity of sending them down to the coast, bought by various Yaos in the Shiré Highlands, who set them to work in their gardens, and, if inquiry was made, passed them off as their wives. In Livingstone’s time even the Anyanja, who have themselves suffered so much from the slave-trade, at times kidnapped people and sold them to the Portuguese. There is a special word for this (fwamba), and though practised it seems to have been always more or less reprobated—or at any rate felt to be wrong.