THE PORTUGUESE IN SOUTH AFRICA FROM 1505 TO 1700 .

CHAPTER VIII.

ARAB AND PERSIAN SETTLEMENTS IN SOUTH-EASTERN AFRICA.

At some unknown period in the past the territory between the Zambesi and Limpopo rivers was occupied by people more advanced in knowledge than the Bantu who were found living there at the beginning of the sixteenth century. Their nationality is uncertain, and nearly everything conected with them is involved in mystery. They were miners, and were skilled in looking for gold-bearing reefs and in working them when found. It is not impossible, though it is only a conjecture of some writers, that traders from the great commercial city of Tyre on the eastern shore of the Mediterranean sea visited them, and that in holy scripture an account of those visits is given. The conditions mentioned of the fleets that went down the Red sea to Ophir in the time of Solomon are not inapplicable to voyages to the mouths of the Zambesi or to Sofala, and the articles — gold, silver, precious stones, almug trees, ivory, apes, and peacocks — with which they returned are all found in South-Eastern Africa, if by almug trees ebony or some other very hard wood is meant, by precious stones pearls, and by peacocks the bustards that to-day are called wilde pauwen (wild peacocks) by the Dutch colonists.

The name of the bird given in the bible is said, however, to be of Tamil origin, and to be used for the peacock (pavo cristatus) at the present day in Oeylon. This appears to be the greatest impediment to the supposition that the Ophir of scripture is the Rhodesia of to-day, unless there was intercourse between Eastern Africa and Southern India in those early times, in which case an African bird might have received from strangers a Dravidian name.

The object that brought the first traders to South Africa can only be conjectured. They may have come in search of ivory, which from very early times was a valuable article of commerce, and the fact that elephants were plentiful in the country may have been ascertained by hunters pushing their way down from the north. The ivory traders may have found gold, and then mining would have commenced. This, however, is mere surmise, and there are obstacles in the way of accepting it or any other supposition that has yet been made. For instance, it is not likely that either hunters or traders would recognise the presence of the precious metal in quartz reefs in a country where alluvial gold is found only in a few localities and in small quantities. Thus it is not possible to say how, when, or by whom gold was first discovered in the Rhodesia of our day, any more than it is to say how, when, or by whom tin was first discovered in Cornwall.

What is certain is that at some time in the past, of which there is neither written record nor tradition now, mining operations were carried on over an immense tract of country south of the Zambesi. The miners were sufficiently skilful to be able to sink pits and run underground galleries along reefs, but they were obliged to cease operations when water began to flow in, as they had no means except buckets and human labour for keeping the excavations dry. The quantity of a reef that could be removed depended thus entirely upon its position, and where drainage was good considerable depths were reached.

With the appliances at their disposal there was only one way apparently in which this kind of mining could be carried on successfully, for a vast amount of labour was needed in excavating the gold-bearing rock, bringing it to the surface of the ground, there crushing it to powder, and then washing the dust to obtain perhaps a thirty or forty thousandth part of gold from it, though the value of that metal relatively to other articles must then have been very much greater than it is now. With a large number of slaves, the men employed in extracting and crushing rock and the women in raising food, it was possible to make gold mining profitable, and it may be taken for certain that this was the condition of things at that time in the territory called Rhodesia today.

Among various articles manufactured by these people that have been found where they were left or lost are an ingot mould, crucibles, and beads, tacks, and thin plates of beaten gold. The thin plates in little squares of uniform size were intended to overlay wood, perhaps the ceilings and columns of grand buildings in other parts of the world, and the wedge-shaped tacks were for fastening them on. Many of these plates and tacks have been found.

Gold mining thus commenced never afterwards wholly ceased until the beginning of the nineteenth century of our era, though the population of the country may have changed many times. Wave after wave of war must have rolled over the land before Europeans made their first appearance, just as since that date, and on each occasion the greater number of the inhabitants must have perished, but some few would always remain to impart their knowledge to the conquerors. This seems the only way of accounting for people like the Makalanga, who were recent immigrants, being found collecting gold for sale at the beginning of the sixteenth century. At this time the metal was no longer beaten into thin plates, nor were tacks made, these articles having long since ceased to be in demand.

Over the whole country where mining was carried on, ruins of stone buildings and walls are scattered, in some places in astonishing numbers, implying a prodigious amount of labour exerted over a long period of time. The oldest of these ruins may be of much later origin than the commencement of the mining operations, but the articles found in the debris beside them prove that some of them were strongholds and factories of miners and workers in gold. Others, however, seem to have had no connection with mining industry, but to have been simply hill fortifications constructed for purposes of defence by people subject to attack by enemies.

Even in the best of these structures the workmanship was of a low order. The arch was unknown, the walls, though circular or elliptical, were not perfectly regular in form, nor were they absolutely perpendicular or of equal thickness throughout. The men who built them were not sufficiently refined to appreciate mathematical correctness of shape or finish. In the very best the stones, taken from strata or layers of rock which could be broken off like slabs and which were consequently flat on two surfaces and of equal thickness, were trimmed at the sides, so that the courses are regular ; but in far the greater number, though the stones have been selected, they have not been dressed, and the work is rough. In only a very few was cement or mortar used, except for levelling floors.

The most imposing of the ruins is that known as Great Zimbabwe, in latitude 20° 16' 30" south, longitude 31° 10' 10" east of Greenwich, 22-4 kilometres or fourteen miles from the present township of Victoria. The building there that has attracted most attention is elliptical in form, two hundred and ninety -two feet or eighty -nine metres in its greatest length by sixty-seven metres in greatest breadth, and was built of blocks of stone flat above and below when taken from the rock and trimmed to about double the width of ordinary bricks, but of varying lengths. The greatest height of the wall still standing is thirty-five feet or 1068 metres, and its thickness varies from sixteen feet two inches or 4'9 metres to five feet or a metre and a half. The only ornamentation consists of two courses of stone laid in oblique positions in contrary directions along a fourth part of the wall near the top, but in some other structures courses of outer stones were laid about five centimetres or two inches apart for the same purpose. A solid conical stone tower, seventeen metres in circumference at the base and still about nine metres high, stands within the wall of this building. Among the debris numerous stone phalli have been found.

The labour required for the erection of such a building, or of another of great size on a hill close by, would be enormous in amount at the present day ; what then must it have been when mechanical appliances such as are now in common use were unknown? Yet even this is trifling when compared with the vast amount of industry expended in the erection of lower walls at other places, such as the Niekerk ruins described by Dr. D. Randall-MacIver in his Mediaeval Rhodesia.

The existence of ruins of such extent and magnitude implies a dense population, although many centuries may have passed away between the laying of the first and the last stone in them. That population was not content to leave its supply of food entirely to the chances of the seasons. Droughts were guarded against by a system of irrigation pronounced by competent authorities from its remains to have been almost as perfect as could be devised at the present day, so that abundance of grain could be relied upon, for here, as everywhere else in South Africa, only water was needed to make the soil as productive as any in the world. At first sight it might seem that to conserve it nothing more was necessary than to construct dams across the courses of streams, but so violent were the floods in the rainy season that unless the dams were immensely strong they would certainly be swept away. Under such circumstances artificial reservoirs were requisite, into which water could be led when the streams were full, and from which it could be drawn into furrows for irrigating purposes when dry weather set in. Such reservoirs required skill and much labour to construct and afterwards to preserve in order. This part of Africa must therefore have presented a scene of industry in mining, building, and cultivation of the soil that is not easy to picture by those who know it at the beginning of the twentieth century of the Christian era. It is possible, however, that the whole of the vast territory from the Zambesi to the Limpopo was not occupied at the same time, but portions of it successively.

Travels of el Maçoudi.

When or why the massive buildings and lines of fortifications were abandoned remains as much a mystery as when or by whom they were constructed. A historian, being without means to determine this question, must leave it for archaeological research to decide. There is no mention of any of them being occupied at or after the beginning of the sixteenth century, and if those farthest east had been, the fact could hardly have escaped the knowledge of Portuguese residents at Sofala and the Zambesi stations. Of those south and west of their trading posts they might never have heard. At some unknown time then they were abandoned, probably as the result of war and at different periods, and afterwards materials accumulated within the walls, in which at length great trees sprang up and helped to complete their ruin. With few exceptions the pits by which the mines were reached also became filled, and the irrigation works were all but completely obliterated. Bantu tribes with only the ordinary amount of intelligence and energy possessed by these people now occupied the territory where once so much industry had been expended. Still a little gold was obtained from a few of the shallowest of the old workings down to the days of Tshaka.

In the tenth century of our era we come upon something like firm footing with regard to a knowledge of the inhabitants of the coast of South-Eastern Africa, for which we are indebted to an Arab writer. Abou'l Haçan Ali el Macoudi was born at Bagdad towards the close of the third century after the hegira. In early manhood he was fond of travel, and visited India and many parts of the East, as well as the island of Kanbalou or Madagascar and the African coast somewhere north of the equator. In later life he resided for a time at Antioch and at Bassorah, and died at Old Cairo in the year of our lord 956. In 943 he completed a great work, which has been translated by Messrs. C. Barbier de Meynard and Pavet de Courtaille and pub- lished in Paris for the Asiatic Society of France. The translation is entitled Les Prairies d'Or, and was issued, with the original Arabic text, in nine octavo volumes, in the years 1861 to 1877.

In Maçoudi's time the Indian sea was frequented by Arab and Persian vessels, the monsoons were taken advantage of to cross over from Hindostan and Ceylon to Madagascar and the African coast, and up and down between the Red sea and Persian gulf on the north and Sofala and Madagascar on the south sailed ships manned by Arabs of Oman and the tribe of Azd or Persians from the ports farther east. From the way it is spoken of, this commerce was no new thing, it was as thoroughly established as the Portuguese found it five centuries later, though as yet there were no Asiatic settlements on the African shore.

Macoudi describes the Bantu, or Zendj as he terms them, as inhabiting the country as far south as Sofala, where they bordered upon the Wakwak or Bushmen. Unfortunately he speaks of Sofala as a territory or country, not as a town or river, so that the exact boundary between the two races cannot be ascertained. The Bantu of Sofala were under a ruler with the dynastic title of Waklimi, who was paramount over all the other tribes to the north and could put three hundred thousand men in the field.

They used the ox as their beast of burden, there being neither camels nor horses nor mules in their country, and they were even ignorant of the existence of these animals. The climate was hot, and snow and hail were to them unknown. Among them were some whose teeth were sharpened, and who were cannibals. Wild elephants were numerous in the country, but a single tame one was not to be found. The country of Sofala produced gold in abundance, panther skins which the inhabitants used to clothe themselves and to sell to the Mohamedans, who covered their saddles with them, tortoise shells such as could be applied instead of horn to make combs, and ivory. The inhabitants killed elephants with darts (assagais) and sold the ivory, which was taken to Arabia, and thence sent to India and China, where it was in demand.

These people used iron for personal ornaments instead of gold or silver. They employed oxen in war and for riding purposes as if they were camels or horses, and the animals ran with as great speed. They expressed themselves with elegance, and there were even orators among them. They termed God Maklandjalou, meaning supreme master. They had no religious creed, but their chiefs observed certain customs, and followed political rules. Each one worshipped what pleased him, a plant, an animal, or a mineral. The staple food was millet and a plant called kalari, which was drawn out of the ground like a truffle or the root of an alder. They also ate honey and the flesh of animals.

This was the description of the Bantu of Sofala given by Maçoudi, and its general accuracy leaves the reader in no doubt as to the condition of the country and its commerce in his time.

At the beginning of the sixteenth century, when South-Eastern Africa was first visited by Europeans, small quantities of gold — chiefly alluvial — were collected, but no traditions were extant of either the first working of the mines or of the erection of the great stone buildings. The Bantu who occupied the country had not been there for many generations. Asiatics were in possession of all the trade, but not of the soil, or of dominion over the inhabitants. From the Moors, as they termed these people, the Portuguese learned of the existence of extensive ruins inland, which they do not appear at any time to have visited themselves, for the descriptions given by their writers are very far from being correct. Thus the temple at Great Zimbabwe, according to their accounts, was a square building, not circular as it really is, and they stated that there was an inscription over one of its doors which no Arabic scholar could decipher, whereas not only is there no such inscription now, but no indication of a stone having been removed on which one could have been displayed at any time.

The Asiatics who were found trading and occupying various stations along the coast were Arabs and Persians, and as they possessed a literature and preserved records of their original settlements and subsequent transactions, the Portuguese writers into whose hands these records came were able to give a very clear account, not only of their condition in the early years of the sixteenth century, but of their previous history and dealings with the Bantu inhabitants. That history was as follows : —

A certain man named Zaide, great-grandson of Ali, nephew and son-in-law of Mohamed, maintained religious opinions that were not in accordance with the koran as interpreted by the Arabian teachers, and was therefore banished from his home. With his adherents, who from him were termed the Emozaidi, he passed over to the African coast, and formed some temporary settlements of no great importance along it. These people were of a roving disposition, and gradually moved southward, avoiding conflicts with the blacks but incorporating many of them, until in course of time they became hardly distinguishable from Africans except by the profession of a form of the Mohamedan creed and a somewhat higher way of living. The trading instinct of the Arabs led them, however, to carry on a petty commerce in gold and probably in other productions of the country. How far south the Emozaidi eventually wandered cannot be ascertained with precision, but some of them appear to have reached the equator before the nest stream of immigration set in.

Foundation of Kilwa.

This was from Central Arabia, and consisted of a number of families driven out by the oppression of a neighbouring sheik. In three vessels they crossed over to the African coast, and founded first the town of Magadosho, and subsequently that of Brava, both not far north of the equator. In time Magadosho became a place of importance, various subordinate settlements were made to the southward, and its trade grew to large proportions. The Emozaidi, who were regarded as heretics by these later immigrants, would not submit to their authority, and were driven inland and forced into still closer connection than before with the blacks of Africa. They became the wandering traders of the interior, the people who collected the products of the country and conveyed them to the coast for sale.

A vessel belonging to Magadosho, having been driven from her course by a storm, put into the port of Sofala, where her crew learned that gold was to be obtained in trade. This led to a small settlement of Arabs at that place, and to a knowledge of the coast as far as Cape Correntes.

Rather more than seventy years elapsed after the founding of Magadosho and Brava when, towards the close of the fourth century of the Mohamedan era, or about fifty years before the Norman conquest of England, another band of strangers settled on the East African seaboard. A ruler of Shiraz in Persia died, leaving seven sons, one of whom, named Ali, was despised by his brothers on account of his mother having been an Abyssinian slave. He was a man of energy and ability, however, so to avoid insult and wrong he resolved to remove to some distant land. With his family and a few followers he embarked in two vessels at the island of Ormuz, and sailed to Magadosho. The Persians and the Arabs were alike followers of the creed of Mohamed, and professed to hold the koran as their guide, but they formed rival sects, and at that time regarded each other with great bitterness. Ali could not settle at or near Magadosho therefore, so he steered down the coast in search of a place where he could build a town of his own, free of the control of everyone else.

Such a place he found at Kilwa, the Quiloa of the Portuguese. The island was occupied by blacks, but they were willing to sell their right to it, which Ali purchased for a quantity of cloth, when they removed to the mainland. He then formed a settlement, and constructed fortifications sufficiently strong for defence against the African blacks and the Arabs higher up the coast who were unfriendly towards him. Whether the island had a name before is not known : he called it Kilwa. Admirably situated for commerce, the settlement attracted immigrants and grew rapidly, so that even in Ali's lifetime it was able to send out a colony to occupy the island of Mafia not far to the northward. Successively different settlements were formed or those founded by the Arabs were conquered, until in course of time Kilwa, notwithstanding various civil wars, became not only the most important commercial station, but the ruling town on the East African coast.

At first the houses were built of wood and clay, but these were afterwards replaced by others of stone and mortar, with flat roofs or terraces which could be used for the same purposes as stoeps in the Cape Colony in our day. The streets between the rows of houses were very narrow, mere alleys in fact, but in the outskirts were large gardens planted with various kinds of vegetables, in which grew also palms and different trees of the orange species. In front of the town, close to the harbour, was the residence of the ruler, which was built to serve also as a fortress, and was ornamented with towers and turrets. The mosques were adorned with minarets, so that, as looked upon from the sea, Kilwa presented the appearance of a beautiful and stately eastern town.

About the year 1330 of our era it was visited by a learned Mohamedan named Abu Abdallah Mohamed Ibn Abdallah el LawSti, a native of Tanjier, usually known as Ibn Batuta. This great traveller left his home in the year of the hegira 724, and did not return until 754, having visited in that time Egypt, the Soudan, Syria, Arabia, the East African coast, Persia, Hindostan, Java, Sumatra, China, and other countries, of which he wrote an account. Makdashu, or Magadosho as now termed by Europeans, he describes as a large city. From it he proceeded down the coast to Mombasa, and thence to Kilwa, whose ruler when he was there was Abu el Mozaffir Hasan. According to the Chronicle of the Rulers of Kilwa, the sultan at this time, the nineteenth in succession from Ali, was named Hacen, who is described therein as a very gallant man. Ibn Batuta relates that he gained great victories over the infidel Zendj, or Bantu, so that the one account corroborates the other. He speaks of the abundance of ivory, and mentions gold, hut only to say that the people of Kilwa did not give much of it in charity. The houses were still built mainly of wood.

There were now three distinct communities of Asiatic origin on the East African coast : the Emozaidi, deemed by both the others to be heretics, the orthodox Arabs, holding one form of the Mohamedan faith, and the Persians, holding another. They were all at variance, and strife between them was constant. This is the key to their easy conquest by the Portuguese in later times. They termed the Bantu inhabitants of the mainland Kaffirs, that is infidels, an epithet adopted by modern Europeans and still in use. None of them, however, scrupled to take women of that race into their harems, and thus at all their settlements the number of mixed breeds was large. At the commencement of the sixteenth century the majority of those who called themselves Arabs, including the descendants of the Persian immigrants, were undistinguishable in colour and features from the ordinary Bantu. It followed that while those in whom the Asiatic blood was predominant were strict Mohamedans, the others were almost indifferent in matters concerning that religion.

Sofala was wrested from Magadosho by the people of Kilwa in the time of Soleiman, ninth successor of Ali, and with it a trade in ivory and gold was secured which greatly enriched the conquerors and enabled them to extend their power. In the zenith of its prosperity Kilwa was mistress of Melinde and Sofala on the mainland, the islands of Mombasa, Pemba, Zanzibar, Mafia, Comoro, Mozambique, and many others of less note, various stations on the coast of Madagascar, and numerous small trading posts along the African shore as far south as Cape Correntes, beyond which no vessel in those times ever passed. But owing to internal strife and perpetual feuds among the different communities, all of these places except Mozambique were lost before the beginning of the sixteenth century, and each of the others had become a petty but sovereign state.

The forty-third ruler of Kilwa after Ali was named Abraham, and it was he who held the government when the Portuguese arrived on the coast. He did not rule, however, by right of descent, but had seized the supreme authority under pretence of keeping it in trust for an absent heir. On this account he was conceded no higher title than that of Emir. When he thus usurped the administration of Kilwa a man named Isuf t was governor of Sofala, having received that appointment many years before. This Isuf was held in high esteem for ability and valour, and as he did not choose to acknowledge Emir Abraham as a superior, he made himself independent and opened his port to the trade of Melinde and other towns on the coast.

Commerce of East Africa.

The Asiatic communities on the African seaboard existed almost entirely by commerce. Except at Pemba, Zanzibar, and one or two other places they did not carry on agriculture to any large extent, though they introduced various fruit-trees and the cultivation of rice and probably a few foreign vegetables among the Bantu. The small islands were not adapted for the growth of grain, and the supplies of food needed by the inhabitants of such towns as Kilwa and Mombasa could be obtained without difficulty in exchange for such wares as they had to barter. One product of the ground, however, they paid particular attention to. That was the cocoa-palm, without which they could not have existed as they did. From its fruit they obtained not only an agreeable article of diet, but a fibre of the greatest utility ; from its leaves material for mats and thatching ; and from its trunk timber for the habitations of the poorer classes, masts and spars for their vessels, and wood for a great variety of other purposes. There was no part of this valuable tree of which some use could not be made.

They built vessels adapted for the navigation of the upper part of the Indian sea, where the monsoons blow regularly at different periods of the year from the east and from the west, though in them they could not venture on such stormy waters as those south of Cape Correntes. In these vessels no iron was used, the planks being fastened to the timbers with wooden treenails, and all the parts sewed or bound together with cord of coir. As they did not use saws, the planks were formed by splitting the trunks of trees down the centre, and then trimming each block with an axe, a tedious and clumsy process, in which much timber was lost. The sails were of close and strong matting, and the standing and running gear alike was made of coir. The largest of these vessels — now called dows — were used for crossing over to the coasts of Arabia, Persia, and Hindostan ; those next in size — which were called pangayos by the first Europeans who saw them — for the most important part of the home trade ; and the smallest — termed zambucos and luzios — for communicating between the settlements, conveying cargoes up and down the mouths of the Zambesi, and other purposes where heavy tonnage was not needed. The zambucos and luzios, indeed, were nothing more than large boats, half decked, and commonly provided with awnings. In shallow places, as in rivers, they were propelled with poles.

The pilots, called malemos, who conducted the vessels to foreign ports, were remarkably expert. Steering across to the coast of Hindostan, for instance, they seldom failed to make the land within a very short distance of the place they were bound to. They determined the latitude by means of measuring the angular altitude of certain stars when on the meridian, for which purpose they used an instrument which they regarded as superior to that by which the first Portuguese navigators in those seas found their way. Of any other method of determining longitudes than by dead reckoning, however, they were as ignorant as all the rest of the world at that time.

The commerce carried on by these people with distant lands was indeed small when compared with that which passed from India either up the Persian gulf and thence by caravans to the shore of the Mediterranean, or up the Red sea, then overland to Cairo, and down the Nile to Alexandria, where the produce of the East was obtained by the Venetians to be distributed over Europe ; but for Africa it was considerable, and it was not subject to much fluctuation.

From India they obtained silks, spices, and other articles of luxury for the use of their own people of pure or nearly pure Asiatic blood, and cotton cloth and beads for trade with the Bantu; from Arabia and Persia rich fabrics, dates, scimitars, large sheathed daggers, and various other kinds of merchandise. Every man, no matter how black, who claimed to be a Mohamedan, wore at least a turban and a loin cloth, and carried a weapon of some kind on his person. The men of rank and wealth, who were of lighter colour, dressed in gorgeous robes of velvet, silk, or cotton, had sandals on their feet, and at their sides ornamented scimitars of finely tempered steel. The women naturally were clothed more or less richly according to the position of their parents and husbands, and they were particularly fond of trinkets. Every article of dress or adornment, all glassware, the best of the furniture of every description, the choicest weapons, and various luxuries of diet were imported from abroad.

With pieces of calico to be used as loin-cloths, beads, and ornaments of trifling value, the traders went among the Bantu on the mainland. Ingratiating themselves with the chiefs by means of presents, they induced those despots to send out men, here to hunt elephants, there to wash the soil for gold, and so forth. Time was to them of less importance than to Europeans, and their mode of living was so nearly like that of the pure Africans that they could reside or travel about without discomfort where white men could hardly have existed. Thus the trade that they carried on was much greater in quantity than that of their Portuguese successors, though its exact amount cannot be ascertained. Upon their wares they obtained enormous profits. They received in exchange gold, ivory, pearls from the oyster beds at the Bazaruta islands, strips of hippopotamus hide, gum, and ambergris washed up on the coast, with which they carried on their foreign commerce ; and millet, rice, cattle, poultry, and honey, which they needed for home consumption.

Commerce was open to any one who chose to engage in it, but practically was confined to the pure Asiatics, who employed the mixed breeds as their agents in conducting the inland barter, working the vessels, and performing the rough labour of every kind. The governments, Arab, Persian, and Bantu alike, derived a revenue from the trade that today seems extortionate. When an elephant was killed, the tusk next the ground belonged to the chief, and when the upper one was sold he took about half the proceeds. On all other articles disposed of by his subjects, his share was about the same proportion, besides which the traders on the other side were obliged to make him large presents before commencing to barter. When Mombasa after the independence of Isuf was able to trade with Sofala, an export duty of rather over fifty per cent was levied on the merchandise for the benefit of the government of that town. At Kilwa any one desiring to trade with Sofala was obliged to pay about seventy per cent of the value of the goods before leaving the port, and on arrival at his destination one-seventh of what was left. Upon his return he paid a duty of five per cent of the gold he had acquired. The duty on ivory brought to Kilwa was very heavy, so that in fact the government obtained a large proportion of the profits on commerce.

On the islands the governments of the Asiatics were not only independent, but all other authority was excluded, and on some of them fortifications were erected, as well as mosques and houses of stone. But on the mainland south of Kilwa, it was different. Here the mixed breeds were permitted by Bantu chiefs to reside for purposes of trade, but they were by no means lords of the country. The sheiks ruled their own people, but no others, like Bantu clans which are often found intermingled, whose idea of government is tribal rather than territorial. They were obliged to make the Bantu rulers large presents every year for the privilege of living and trading in the country, which presents may be regarded rather as rent for the ground and license fees than as tribute. Under these circumstances they did not construct any buildings of stone.

Description of Arab Settlements

The pure Asiatic settlers on the African coast were grave and dignified, though courteous in demeanour. They were as hospitable as any people in the world, but they were attached to their ancestral customs, and keenly resented anything like an affront. They were enterprising, though so conservative in their ideas that they were incapable of making what Europeans would term rapid progress in civilisation. As superstitious as their Bantu neighbours, they especially regarded dreams as figuratively foreshowing events, and he was regarded as wise who pretended to be able to interpret them. The tombs of men celebrated for piety were places of ordinary pilgrimage, but every one endeavoured when in the prime of life to visit the city of Mecca in Arabia, thereby to obtain the highly honoured title of hadji.

The mixed breeds, who formed the great bulk of the nominally Mohamedan population, had all the superstitions of both the races from which they were descended. They would not venture to sea on a coasting voyage if one among them had an adverse dream, or without making an offering, if only of a shred of calico or a piece of coir cord, at the tomb of some holy man. They believed that the winds could be charmed to rise or fall, that the pangayos were subject to bewitchment, that even the creatures of the sea could be laid under spells. They lived in short in the atmosphere of the Arabian Nights, darkened by the gloom of Bantu fear of malignant sorcery.

Coming down the eastern coast of Africa in the year 1500, the principal Mohamedan settlements and trading stations were in geographical order as follows : —

Magadosho, in latitude 2° 2.' north of the equator. The town was on the coast of the mainland, partly built upon an eminence rising to a height of about 12-19 metres above a sandy plain. It contained several mosques and many stone houses with flat roofs. In front, at no great distance from the shore and parallel with it, was a coral reef seven or eight kilometres in length, which protected the channel within from the fury of the sea. At low spring tides the water in the channel was only two fathoms in depth, but that was sufficient for the dows used in the Indian trade. There was no other port.

Brava, in latitude 1° 7' north, was also built on the coast of the mainland. It stood on an eminence about thirty metres above the beach, and was enclosed with a wall. The town was well built, and was governed as an aristocratic republic, the only one of the kind on the coast. The port somewhat resembled that of Magadosho, being a channel along the shore partly protected by islets and. reefs, but was more exposed to heavy rollers from the sea.

Melinde, in latitude 3° 15' south of the equator, situated on the coast of the mainland, was also a well-built town. Adjoining it was an extensive and fertile plain, covered with beautiful gardens and groves, in which flourished fruit trees of various kinds, principally orange and lemon. To gain this advantage the town was built some distance from the nearest anchorage, which itself was far from safe, being a roadstead protected to some extent by a reef, but made dangerous by numerous shoals. It possessed, however, in a narrow rocky peninsula extending into the sea an excellent natural pier for landing cargo from boats.

Mombasa, on a coral island about five kilometres long by three broad, was situated in the estuary of the Barrette river, in latitude 4° 4' south. The island was like a huge fortress, standing from twelve to eighteen metres out of the water and presenting steep cliffs of madrepore on the seaward side. It possessed one of the best natural harbours in the world, easily accessible at all times. On each side the passage between the island and the banks of the estuary was broad and deep, though winding, and when in them or in the fine sheet of water to which they led a vessel was perfectly sheltered. This sheet of water could only be reached by large vessels through the northern strait, because a submerged reef stretched across the inner end of the other, and at low tide formed a ford to the mainland. The town was built along the steep shore of the northern passage, not far from the sea, and was next to Kilwa the most celebrated on the coast. The houses were of stone, so well constructed that the first Europeans who saw them compared them favourably with residences in Spain. Mombasa, owing to its excellent site and to the prevalence of sea breezes, was less troubled with fever than any other settlement on that part of the coast.

Pemba, a coral island, rising in the highest part to ninety-two metres above the level of the sea, was sixty-one kilometres in extreme length by twenty-one in width. It was about twenty-nine kilometres from the mainland, with a clear passage for ships inside, though coral reefs abounded near the shore. The island was fertile, and produced large quantities of provisions, particularly rice, for exportation. The principal Arab settlement on it was in latitude 5° 25' south.

Zanzibar, not far south of Pemba, was an island similar in every respect, though larger, being sevehty-six kilometres in extreme length by thirty-two in breadth. It rose to a height of one hundred and thirty-four metres above the level of the sea. The principal Arab town, from which the island took its name, was on the western side, in latitude 6° 3' south. The anchorage in front of it was good and capacious, and there were many secure harbours among the islets and reefs in the channel between it and the mainland. Here were built the greater number of the vessels used in the Indian and the coasting trade, and from the island considerable quantities of provisions were exported.

Mafia, (Monfia or Monfeca) a coral island rising abruptly from a great depth of water, lay about fourteen kilometres from the mainland. This island was about forty-three kilometres in length by fourteen in extreme breadth, between 7° 38' and 8° south latitude. It was of much less importance than either Zanzibar or Pemba.

Kilwa, a low coral island, rather over six kilometres in length by three in breadth, rising on the northern side to fourteen metres above the sea level, was set like an arrow in a drawn bow in the estuary of the Mavudyi river. It lay in latitude 8° 57' south. With the sea in front, a strait on each side, and a sheet of water extending sixteen or twenty kilometres beyond its inner extremity, it was a very strong position. As at Mombasa, the southern strait was crossed at its far end by a reef, along which access to the mainland could be had at low water. This strait was interspersed with islets, and made a capacious harbour, admirably adapted for shipping, but that on the northern side of the island was difficult to navigate on account of its containing numerous reefs and sand banks.

Passing south of Cape Delgado, in latitude 10° 40', a chain of coral islets and reefs parallel to the coast at a distance of thirteen to twenty-one kilometres, and extending one hundred and eighty-eight kilometres along it, was to be seen. The principal islet was termed Kerimba, or Querimba, and from it the whole group was named. Next in importance was Ibo. Most of the others were uninhabited, being mere rocks rising from the sea. Along the strait within were numerous harbours for ships.

The northern extremity of the Mozambique channel has now been reached, and halfway across it lay the Comoro islands, all of volcanic origin. The principal of these were named Comoro, Johanna, Mohilla, and Mayotta, but there were many smaller in size. These islands were also possessed by the Arabs, who made use of them as convenient stopping places on their way to the great Island of the Moon, which we term Madagascar.

Keeping down the African coast, an inlet about nine kilometres across and ten in depth was reached, in latitude 15° south. Into its inner end ran three streamlets, but of inconsiderable size. Lying across the centre of the mouth of the inlet, within a line joining its two outer points, was a low coral island, about two kilometres and a half in length and three hundred and sixty-six metres in breadth, named Mozambique. About five kilometres farther out in the sea were two others, similar in formation, then uninhabited, one of which is now called Saint George and the other Saint Jago. Behind Mozambique was a spacious harbour, easily accessible and perfectly sheltered. At long intervals indeed a furious cyclone would sweep over it and cause great destruction, but the same could be said of any part of that coast and sea. Such a position as the island of Mozambique could not escape the observation of the Mohamedans, though it had not the advantages of Kilwa or Mombasa. The island itself produced nothing, not even drinking water. On the northern shore of the inlet, since termed Cabaceira, the ground was fertile, but it was exposed to irruptions of the Bantu inhabitants, who were generally hostile. So Mozambique never rose to be more than a dependency of Kilwa, a mere halfway station for vessels bound up or down the coast, Its Mohamedan occupants had their gardens and cocoa nut groves on the mainland, but could not always depend upon gathering their produce.

The Angosha (Angoxa, Angozha, or Angoche) islands lay off the mouth of the Angosha river, between latitude 16° and 16° 40' south. The river was five kilometres wide at the bar, and could be ascended by boats nearly two hundred and forty kilometres, which circumstance gave to the six coral islets off its entrance a value they would not have had in another position. There was a good roadstead between the bar of the river and the island Mafamede, which was a mere crown of sand on a coral reef above sea level.

The Primeiras islands were nothing more than a row of coral hummocks extending northward from latitude 17° 18' in a line parallel with the coast. In the channel between them and the mainland there were places where a pangayo could find shelter to refit, or during the prevalence of contrary winds.

At Mozambique the direction of the coast line had changed from nearly north and south to north-east and south-west, and the aspect of the land had altered also. Thence to Cape Correntes as far as the eye could reach nothing was visible but a low flat tract, bordered along the sea by sand hills from fifteen to one hundred and eighty metres high, with here and there a dark-coloured rock. In latitude 18° south the mouth of the Kilimane, or Quilimane, river was reached. This was the northernmost of the several outlets of the great river Zambesi, which therefore bounded the delta on that side. The other large outlets were the Luabo and the Kuama, but there were many smaller ones, a distance of a hundred and sixty kilometres separating the extreme southern from the extreme northern mouth, while the inland extremity of the delta, where the river began to fork, was over eighty kilometres in a straight line from the sea. In later years this whole tract of land and water was termed by the Portuguese the Rivers of Kuama, the largest of the islands in the delta bearing that name.

If an accurate survey of the delta and its streams had been made in any one year, in the next it would have been imperfect, and in a decade misleading, for two causes were constantly operating to alter the features of land and water. In the rainy season the Zambesi, which stretched nearly across the continent, poured down a flood bearing sand, soil, and gravel, which spread over great areas, blocked up old channels, tore away huge fragments of islands, and opened new passages in every direction. When the flood subsided, former landmarks were gone, and where vessels had sailed the year before sand flats alone were seen. The Kilimane arm in the year 1500 was the best entrance into the Zambesi during six months of the year, in 1900 its upper course is much higher than the bed of the great river farther inland, of which it is no longer regarded as an outlet. The other cause of change was the mangrove. This tree, with its gloomy dark-green foliage, grew only on the confines of land and water, where it spread out its roots like gigantic snakes, intertwining and retaining in their folds the ooze and slime that would otherwise have been borne away. Sand was blown up by the wind or deposited when the currents were gentle, vegetable mould accumulated, the inner line of the swamp became soil on which grass and herbs could grow, and the mangrove spread farther out to reclaim ever more and more land from the shallow water. So the floods washed away and reformed, and the mangrove bound together and extended, in the ever varying scene.

How far up the Zambesi the Mohamedans were accustomed to go cannot be ascertained with precision. They had a small settlement on its southern bank where the Portuguese village of Sena now stands, about two hundred and twenty- five kilometres from the sea, but it is doubtful whether they had any fixed post farther inland, though travelling traders probably penetrated the country to a great distance. About three hundred and seventy-eight kilometres from the sea the great river passed through the Lupata gorge, a narrow cleft in the range that separates the interior plain from the coast belt, where the rapids were so strong that they may not have cared to go beyond them with their boats, though the Portuguese afterwards navigated the stream up to the Kebrabasa rapids, about thirty-two kilometres above Tete, or five hundred and twelve kilometres from the sea.

At the mouth of the Pungwe river, where Beira now stands, there was a very small Mohamedan trading settle- ment, perhaps not a permanent one, and only at best an outpost of Sofala.

Sofala, the most important station south of Kilwa, was in latitude 20° 10'. It was at the mouth of an estuary not quite three kilometres wide from the northern bank to an island named Inyansata, between which and the southern bank there was only a narrow and shallow stream when the tide was low. Across the entrance of the estuary was a shifting bar of sand, which prevented large vessels from crossing, and inside there were so many shoals that navigation was at all times dangerous. The land to a great distance was low and swampy, and the banks of the estuary were fringed with belts of mangrove, so that the place was a hotbed of fever and dysentery. Farther in the interior the stream was of no great size, but it was always bringing down material to add to the deposits of sand and mud above the bar. The sole redeeming feature was a high rise of tide, often nearly six metres at full moon, so that when the wind was fair it was accessible for any vessels then used in the Indian trade. Along the coast was a great shoal or bank like a submerged terrace, extending far into the sea, upon which the waves ran so high at times and the currents were so strong that the locality was greatly dreaded by the mariners of olden days. But all these drawbacks were disregarded in view of the fact that gold was to be obtained here in exchange for merchandise of little value.

At Sofala there were two villages : one close to the sea, on a sand flat forming the north-eastern point, contained about four hundred inhabitants ; the other, about three kilometres higher up the bank of the estuary, also contained about four hundred residents. The sheik lived in the last named. His dwelling house was constructed of poles planted in the ground, between which wattles were woven and then plastered with clay. It was thatched, and contained several apartments, one of considerable size which could be used as a hall of state. The floor, like that of Bantu huts, was made of antheaps moistened and stamped. It was covered with mats, and the room oocupied by the sheik was hung with silk, but was poorly furnished according to modern European ideas. This was the grandest dwelling house in Africa south of the Zambesi, indeed the only one of its size and form, in the first year of the sixteenth century.

The island of Tshiloane (Chiluan, Chilwan, Chuluwan, Kiloane) lay partly in the mouth of the Ingomiamo river, in latitude 20° 37' south. The island was about nine kilometres and a half long by five wide, but a great part of it was a mangrove swamp. The channel into the Ingomiamo on the northern side of the island, now called Port Singune, was used as a harbour by an occasional pangayo or zambuco that put in to trade.

The Bazaruta islands were of much greater importance, for there were the pearl-oyster beds which yielded gems as much coveted by the Arabs and Persians as by the people of Europe and India. There were five islands in this group, stretching over forty-eight kilometres along the coast northward from the cape now called Saint Sebastian, which is in latitude 22° 5' south. The principal island, from which the group takes its name, is twenty-nine kilometres in length.

The last place to the southward frequented by the Mohamedans was the river Nyambana, or Inhambane, the mouth of which is in latitude 23° 45' south. They had a small settlement where the Portuguese village now stands, twenty-two kilometres by the channel, though only thirteen in a direct line, above the bar. The river was easy of access, and formed an excellent harbour. It was navigable for boats about eight kilometres farther up than the settlement, which formed a good centre for collecting ivory, an article always in demand in India. This place was reputed to be the healthiest on the whole coast.

Beyond Cape Correntes, in latitude 24° 4' south, the Arabs and Persians did not venture in their coir-sewn vessels. Here the Mozambique current, from which the cape has its present name, ran southward with great velocity, usually from two to five kilometres an hour, according to the force and direction of the wind, but often much faster. The cape had the reputation also of being a place of storms, where the regular monsoons of the north could no longer be depended upon, and where violent gusts from every quarter would almost surely destroy the mariners who should be so foolhardy as to brave them. The vivid Arab imagination further pictured danger of another kind, for this was the chosen home of those mermaids — believed in also by the Greeks of old — who lured unfortunate men to their doom. So Cape Correntes, with its real and fictitious perils, was the terminus of Mohamedan enterprise to the south, though there were men in Kilwa who sometimes wondered what was beyond it and half made up their minds to go over land and see.

CHAPTER IX. DISCOVERY OF AN OCEAN ROUTE TO INDIA.