suitable for manoeuvring, for artillery, for cavalry-the nature of the roads, of the hills, of the rivers, should all be looked to, and rough contouring and hill-shading is a very important part of the officer's work. He must also get information of defensible spots, of the water supplies, the food supplies, and the resources of the country, and this should be embodied as much as possible on the plan, or on the report attached to it.

Though any degree of elaborateness may be used, any instruments of precision employed, the typical military sketch is made with a sketching case, which is merely an improvised plane table. The main lines or traverses are taken from the bearings of the prismatic compass laid down on the sketch itself. The lines are generally paced or guessed, distant objects fixed by bearings from the station points, and the contouring measured angularly by Abney's level, or sketched by the eye. The shading of the hills shows steepness by the lines used to indicate them being drawn closer or further apart.

CARTOGRAPHER.-The plans and maps having been drawn, and all notes made of information, they reach the cartographer or atlas-maker. His duty is first to compare all new information with what is already known. To eliminate manifest errors, to reduce to scale and to projection uniform with his great maps of the same part of the world, and generally to make everything ship-shape for publication.

ATLAS MAKING. I do not here refer to the Ordnance Survey maps, drawings, and prints, which were described with the utmost detail and precision by Sir Charles Wilson, but to the general atlases, such as Johnston's and Bartholomew's. With the information so gleaned the cartographer is able to make those beautiful orographical maps, which are now so common, showing different levels.

In our diagram I have coloured the island orographically, which is done by drawing the contour lines and washing over the areas so marked with different variations of tint. But I shall not go far into this subject. Mr. Bartholomew must do that some other day. Imagine the map drawn. It may be then engraved, like any other picture or line engraving, on a copper-plate, and either printed from that plate or from lithographic stones, to which an impression of the plate has been transferred.

In the Ordnance Survey printing-office, instead of lithographic stone, the maps are printed from sheets of zinc, which has much the same property of absorbing greasy ink.

By this time we have got into the printing-office, and to describe it in detail would be beyond my province; this part of the subject, though very interesting, really embraces the whole art of the engraver, the lithographer, and the printer. But there is one process I desire to show before closing.

You see daily in books and newspapers, and in our own journal, maps printed in black along with the type. There are numberless processes for their production-one only I shall briefly note. It is the type-process of Messrs. Walker & Boutall, who have kindly sent me a specimen in course of manufacture.

On a brass plate a coating of a waxy composition is laid; the outlines

VOL. VII.

2 H

After

You

of the map are either drawn on this coating or photographically transferred to it. The engraver then scratches through the wax down to the brass with a needle. He next takes suitable types, and stamps in the names also down through the wax to the brass, and completes the matrix with the necessary amount of detail, which may be great or little. verification and correction the matrix is ready for electrotyping. who know the appearance of stereotype moulds will see that this resembles the mould of an ordinary stereotype or electrotype page. The mould is next covered with black-lead, and an electrotype taken from it, when all the punctures that have been made through the wax to the level brass plate come out level-the scratches as lines, and the type as lettering. It is then mounted on wood, and is ready to insert among type and be printed along with it.

I have tried to give you very roughly an outline of how maps are made from the beginning to the end, in almost the same form that actual necessity forced me to learn it for practical use.

REWA RIVER, FIJI.

BY H. H. THiele.

(With a Map.)

It is with the view of contributing some useful information about a small part of Fiji that I take the liberty of laying this paper before the Royal Scottish Geographical Society.

During the last thirty or forty years several books have been published describing these islands and their inhabitants, but, without being exactly unfaithful to the subject on the whole, the authors have evidently written with a view to the sale of their books rather than for the purpose of imparting accurate and useful information. Dr. Seemann's A Mission to Fiji forms an honourable exception, but yet I cannot help thinking that if he had resided five or six years amongst the natives of these islands, the experience he would have gained would have led him to alter his opinion on several subjects relating to them.

The Fiji group consists of two large islands, Viti Levu (Large Viti) and Vanua Levu (Large Land), and upwards of 250 smaller ones, ranging in size from over 200 square miles (Taviuni) down to islets consisting of an acre or less of barren rock. All the islands together make up an area of about 7400 square miles, of which about 4200 square miles belong to Viti Levu, the island dealt with in this paper.

A cursory glance at a map shows that the island of Viti Levu is of an oval shape; it is about 97 miles long from east to west, and 67 miles across from north to south at its broadest part.

Although the island is of no great extent, its climate varies very considerably in different parts; generally speaking, the southern and eastern districts and the greater part of the interior receive plenty of rain, while

the northern and western parts, especially the coast-lands, are considerably drier. The mountain ranges, the general direction of which is from north-east by north to south-west by south, are highest in the interior, where they attain a mean height of about 3000 feet, and they cover a greater area in the western than in the eastern half of the island. The south-east trade-wind prevails for about two-thirds of the year; it comes from the ocean laden with moisture, and, striking the eastern sides of the ranges, sweeps up their slopes till it reaches an altitude where, owing to the fall in temperature, some of its moisture is condensed and falls in rain on the windward side of the mountains. The average yearly rainfall on the south-east corner of the island, at Suva, the capital, is about 105 inches; at Viria, further inland and closer to the mountains, 145 inches; and at Ba, on the north-west coast, probably not much more than half of the latter quantity. No very reliable meteorological observations have so far been taken in that part of the island, but it may be noted that irrigation has to be resorted to there, in order to cultivate sugar-cane successfully.

From the circumstances stated above it will be easily understood that the principal rivers must be found in the southern and eastern parts of the island. The Siga Toka is by far the most important river in the middle and southern parts, while the Rewa holds the same position in the eastern half of the island. The former enters the ocean on the south coast, at about 177° 30′ E. long., and the so-called delta of the latter occupies the south-east corner of the island. Besides these there are two other rivers of minor importance, viz., the Navua, which, rising somewhere to the south-east of the centre of the island, pursues a southerly course till it reaches the ocean about midway between the mouths of the Siga Toka and the Rewa; and the Ba River, the basin of which occupies the north-western corner of the island. These four rivers are the only ones on Viti Levu of any importance to traffic. Numerous creeks are found all along the coast, carrying the water from the slopes of the nearest ranges to the ocean.

In the northern part of the island the watershed between the Siga Toka and the Rewa is formed by the so-called Dividing Range.

The length of the Siga Toka cannot be estimated with any approach to exactness, as it is not decided which of the many head-waters should be regarded as the main stream. It runs for nearly the whole of its course through deep gorges in the most mountainous part of the interior; only near the mouth do extensive fertile plains and sandy hills lie beside. its banks. It is full of rapids, shallows, and deep holes, and a dangerous bar obstructs its mouth; it receives many affluents both from east and west, all more or less of the same character as itself. During the rainy season an immense volume of water comes rushing down this treacherous river, while in the dry season there is in many places not sufficient water to float a laden native canoe. That the scenery on this river is magnificent can easily be imagined: in the north the banks are mostly formed by steep and high walls of barren rocks, and the narrow river-bed is strewn with large and small fragments of conglomerate, thrown or rolled into it during volcanic disturbances, between and over which the

water rushes along with great speed, forming whirlpool after whirlpool and an endless succession of foaming cascades. Any one who has seen the "Elvs" of Old Norway cannot help being struck by the great resemblance, especially if the hour happens to be early morning or late afternoon, when a gloomy twilight increases the wildness of the scenery, and seems to add to the darkness of the water in the deep pools and bring into greater contrast the silvery whiteness of the countless little waterfalls. A turn of the river and, may be, the whole scenery is changed from a Norwegian to a thoroughly tropical one. You pass into a narrow valley, where tree-ferns, wild figs, and smaller tropical plants occupy every inch of ground, and the more sloping sides of the ranges on either side are covered with dense forest from foot to crown: you might fancy yourself in the interior of Guadeloupe, or some other West Indian island. Where fertile flats of sufficient extent to produce the necessary food for the inhabitants occur, native villages stand on the banks. The Fijians here, and on the mountains generally, are called "Koi Colos"; they were the last of all the tribes to submit to British rule, and even now the greater part of the interior is included in the so-called "Disturbed District," through which a white man can travel only by special permission from the Government.

Leaving the Siga Toka, I will now proceed to the description of the Rewa, which is the principal subject of this paper, as the river happens to run through the district of Viti Levu with which I am best acquainted.

In the north-eastern part of the mountain ranges lie the sources of numerous water-courses, which, with the exception of a few comparatively unimportant ones, make direct for the north coast, and join one of the two following rivers: the Wai-ni-Buka, the most northerly, and the Wai-niMala. These, after very tortuous courses of some 30 miles, meet to form the Rewa. The length of this river is about 42 nautical miles,1 and its general direction is towards the south-east. It receives numerous affluents both from the north and south, but only two are of any importance. These are the Wai-Dina and the Wai-Manu, both coming from the west. The former flows into the Rewa about 28 miles, and the latter about 10 miles, from its mouth.

It may fairly be reckoned that the Rewa drains one quarter of the entire island of Viti Levu, or about 1000 square miles. Taking the average rainfull in this basin at 120 inches per annum, and allowing 25 per cent. for evaporation and absorption, this river carries yearly about 5800 million tons of water to the ocean. Evaporation must necessarily be small, where rainy and cloudy days are so very frequent and dry winds never occur; besides, probably half of the total quantity stated above is carried down during floods or heavy freshets-which last from three to six days at a time, and occur on an average about six times yearly—and soon reaches the sea. Nor can the quantity of water absorbed be very large, owing to the impermeable nature of the river-bed, for the hills, where they are not of hard rock, are principally formed of soapstone,

1 When miles are mentioned subsequently, they are "nautical" miles.

which, once saturated, hardly ever gets the chance of becoming dry again. The same to a great extent is also the case with the clayey soil on the flats, and the peaty soil has quite a remarkable power of retaining moisture, owing to its resting on layers of clay and being covered with a thick vegetation of rank grass and reeds.

The effects of the force of this considerable mass of water making its way towards the sea shows itself in several ways. The river-bed gets deepened in some places, while shallows are created in others, banks are undermined and carried away, and the materials formed into islands elsewhere. These changes are continually going on, slowly certainly under ordinary circumstances, but with all the greater violence during heavy floods. In the five or six years I have been here the channel of deep water in the lower part of the river-from Dreke-ni-kelo to the seahas altered to such an extent that in one or two places islands now exist where previously the deepest water was to be found. Immediately after a heavy flood or freshet the river on the whole gets shallower; in course of time, however, and especially as far as the tides take effect, a channel is formed of about the usual depth.

From Laucala bay to the Navoloa mouth of the river the coast-line is formed by a succession of mud-flats, sandy stretches, and some fringing reefs on the whole it is rather low, though a portion of it between Nasoata and Nasalai rises to an elevation of from 30 to 40 feet, and Kaba island is more than 100 feet over high-water mark. Mud-flats of varying breadth, more or less overgrown with mangrove bushes, line the bays and the banks of the lower portion of the river. The coast-line proper is in many places difficult to define, as the tides cause it to vary very considerably; in the accompanying chart it has been laid down on the line of ordinary high-water mark.

The difference of tides at the mouths of the river is about 3 feet under ordinary circumstances, and upwards of 6 feet during spring-tides. At Nansori, about 10 miles from the coast, the difference is 3 feet under ordinary circumstances, and 4 feet at spring-tides. At Naitasiri, 20 miles from the mouth, the usual tidal difference is 1 foot, with foot added at spring-tides. No tides are felt above the place where the WaiDina joins the river. Strong easterly or south-easterly storms increase the flood-tides to some extent. Seeing that the Rewa has six large mouths, it may be easily imagined that the tidal currents are very complicated in its lower portion, a circumstance which, combined with the effects of strong winds, considerably increases or retards the outflow of water. At the Nukulao, or principal, mouth of the river a sand bar has been formed with only about 4 to 5 feet of water on it at ordinary lowtide. A vessel drawing 3 feet requires flood-tide to go up river as far as Bau Levu. It is hardly conceivable with such tides, combined with the usual south-easterly wind blowing fairly on to this part of the coast, how a delta could ever have been formed. The depth of the sea outside the reefs is considerable, and also in the passages leading through the reefs; where the water is deep, mud is always found at the bottom, an indication that the sediment of the river ultimately settles there. The existing mud- and sand-flats were no doubt formed by the sediment of the