Though the palmy days of cinchona cultivation have long since departed, given land and labour cheap and plentiful, there is yet a great possibility in cinchona to the man who can afford to wait.

Sugar might be cultivated to an unlimited extent on the lowlands. It is grown and manufactured at Zomba for local consumption; but as it is practically a twoyear crop in the highlands, profits from it are small. I have never looked upon sugar as an article of export from Nyasaland, though local demand can always be

met.

Tobacco is cultivated by the natives and cured in different ways, and some of this article as manufactured by them is highly prized. From imported seed a finer leafed tobacco has been raised, and a sample of Zomba tobacco was described by a London broker as the finest yet seen from Africa. The native tobacco lies under the disadvantage of being coarse-veined, thick and heavy in every way, totally unsuited for any save the coarsest uses in the home market. Both cut tobacco, cheroots, and cigars, are being manufactured at Zomba for local supply, and, while this is met, planters may look forward to exporting the raw material. A superior tobacco can be cultivated; experience in curing and growing will adapt it to the market.

I feel justified in saying that, commercially, Nyasaland has a bright future before it. We have the backbone of commerce in coffee, cocoa, rubber, tobacco, cotton, cinchona, and it may be, tea, and sugar, and in an already organised trade in oilseeds, capable of unlimited extension, with great probabilities in fibres, grain, hides, beeswax, etc., not to speak of the ivory trade, which will hold its own for several years yet to come.

The development of Nyasaland, considering the circumstances, has been I believe, unique. In June of last year 1891, we entered upon a new régime. It is unnecessary in this place to do more than mention the fact that a Commissioner, in the person of Mr. H. H. Johnston, C.B., sent out by Her Majesty's Government, began to administer the country.

Before concluding, I wish merely to ventilate a railway scheme for British Central Africa which has been in my mind for years. It is simply that we should construct a railway from the Shiré to Lake Nyasa, making Chiromo at the mouth of the Ruo our starting-point, and subsequently another line to connect Nyasa and Tanganyika. For the first project alone a sum of about £500,000 would be necessary. It seems a large sum, but we may as well look the matter in the face at once, for a few years hence the money will have to be forthcoming.

It needs no prophet to foretell that the lion's share of the work of developing Central Africa, and putting down the slave-trade, has fallen upon Britain's shoulders, and as a nation she can no more shirk her responsibility than can an individual his duty. Central Africa calls aloud for development, and I hold, with pride too, that of all nations on the face of the earth, Britain is the most fit for the work.

THE CONSTRUCTION OF A MAP OF THE WORLD ON A SCALE OF 1:1,000,000.

By Professor Dr. A. PENCK.

RECENT years have witnessed the accumulation of an enormous mass of geographical information, laid down on maps of a comparatively large scale. We have maps on a scale of 1: 200,000 (3.15 miles to 1 inch) of nearly all Europe, large tracts of North America, and portions of Africa and the Indian Empire.

More than two-thirds of the whole extent of coast-line of our continents and islands are mapped on a not much smaller scale. Several of the Central and South American States, Australia, and a large portion of Western, Southern, and Eastern Asia, are mapped on scales varying between 1:500,000 (7·89 miles to 1 inch) and 1:1,000,000 (15.78 miles to 1 inch), while itineraries through the wilds of South America and Africa are now frequently laid down on larger scales than that of 1:1,000,000.

Various circumstances tend to considerably reduce the scientific and practical value of this enormous mass of accumulated information. The maps on which it is laid down are not uniform either in scale, projection, or style of execution; they are published at different places all over the world, and are often difficult to obtain. Some are only published in journals, and cannot be had separately, while others, for one reason or another, are not put in the market at all.

The circumstances and interests of our civilised life make good maps almost a necessity. Maps of our own country are absolutely indispensable; commercial interests, missionary undertakings, and colonial enterprise create a demand for maps of foreign countries, while of the maps required for educational purposes and as illustrations of contemporary history, the name is legion.

The compilation of all the existing cartographical material and its condensation into an atlas of the world, would be a work of great practical as well as scientific value, especially for nations having considerable Colonial possessions. A uniform map of the world would be at the same time a uniform map of the British Empire, showing not only the actual territory under British authority, but also the sphere of British commercial activity, and would serve the varied purposes of administration, navigation, and commerce. The desire for maps which shall show a country not merely as a piece of land limited by political boundaries, but as a region in the frame of its natural surroundings, has led to the execution not only of general maps of large portions of Central Europe by the German, French, and Austrian staffs, but also to the production of a large scale map of a whole continent-Africa.

The consideration by the Fifth International Congress (held at Berne, 1891) of the scheme for the execution of a map of the world on the scale of 1:1,000,000 (15.78 miles to 1 inch), was a step in the right direction. On that occasion the author of this paper gave a short sketch of his scheme,* which he had previously treated of in a preliminary note.† De Lannoy de Bissy, the author of the map of Africa on the scale of 1:2,000,000 (31.56 miles to 1 inch), gave his approval,‡ and after receiving the report of a committee, the Congress appointed an International Commission for the investigation of the scheme.§ Dr. Richard Lüddecke of Gotha has since started a discussion on the subject in the pages of Ausland || in which

* A. Penck. Die Herstellung einer einheitlichen Erdkarte im Massstabe vo 1:1,000,000 (Annales du Congrès, Annexe IV.).

† A. Penck. Die Erdkarte im Massstabe von 1:1,000,000 (Beilage zur Allgemeinen Zeitung, Munich, 1891, No. 169: June 20).

A. E. Forster. Über die Herstellung einer Karte im Massstabe von 1:1,000,000 (Das Ausland, No. 31: 1891).

De Lannoy de Bissy. Quelques détails sur la carte de l'Afrique au 2,000,000me à propos de la question de l'élaboration d'une carte de la terre à l'échelle du 1,000,000me (Annales du Congrès, Annexe V.). Also published separately (Imprimerie Fricotel,' Epinal).

§ Première Résolution votée dans la séance de clôture.

|| R. Lüddecke. Zur Erdkarte im Massstabe von 1:1,000,000 (Ausland, No. 46: 1891). Noch einmal zur Erdkarte im Massstabe von 1:1,000,000 (Ausland, No. 11: 1892).

Dr. H. Habenicht (Gotha),* Dr. E. Hammer (Stuttgart),† and the author of this paper have taken part,‡ the main points discussed being the scale and projection of the map. It will be remembered, however, that on both of these points the Congress had already expressed an opinion, the scale of 1 : 1,000,000 being approved and the division of the sheets by meridians and parallels being accepted by preference.

A comparison of the scales usually employed in the construction of general maps of extensive tracts of country will show that they range between 1: 200,000 (3∙15 miles to 1 inch), and 1:1,000,000. Smaller scales are but rarely employed in the construction of general maps of large portions of Europe; on the other hand, maps of European States in the best atlases rarely have larger scales than that of 1:1,500,000 (23.67 miles to 1 inch). In order to fulfil its purpose, so far as Europe is concerned, a map of the kind proposed must be executed on a scale not larger than 1: 1,000,000, nor smaller than 1 : 1,500,000.

It must be confessed that for some parts of the world the scale of 1 : 1,000,000, in view of the present state of our knowledge, is much too large, but these tracts become more and more limited every year, and though twenty years ago a map of Africa on a larger scale than 1 : 5,000,000 would have seemed an impossibility, we now have a map of that continent on a scale of 1 : 2,000,000, and as long ago as 1885 the author of that great work was able to express the opinion that in ten years the scale of 1: 1,000,000 would be the smallest on which our knowledge of Africa could be done justice to. In fact portions of Inner Africa are now mapped on the scale of 1 300,000, large tracts on scales ranging between 1: 750,000 and 1: 1,000,000, while similar scales are frequently used for route maps. It will be seen that in districts at all thickly populated from twenty to fifty place-names will go to the inch.

An advantage of the scale proposed is that it corresponds almost exactly to that of the Indian Government's 16 miles to the inch map of the Indian Empire and neighbouring countries, so that one-seventh of Asia may be said to be already mapped on this scale. Further it differs but little from the Russian military map of Russian and Turkish Asia (1:840,000 and 1:1,680,000). Moreover, the French and Dutch have given us maps of Further India and the East Indian Islands respectively, on the scale of 1:1,000,000, which is also used for the maps of the Transvaal and of certain Central and South American States.

It may further be mentioned that 10 per cent. of all the African maps in Petermann's Mitteilungen are on this scale, as also many maps published in the Proceedings of the Royal Geographical Society. If De Lannoy de Bissy already considers the scale of 1:2,000,000 too small for Africa, we may safely assume that the same is true with regard to the other continents, and that the proposed scale of 1:1,000,000 is in every respect the most suitable. Lüddecke has denied this, and declared that for certain parts of the earth such scales as 1:3,000,000 (47.34 miles to an inch) or even 1:4,000,000 (63.13 miles to 1 inch) would meet the requirements. The latter scale is most certainly too small. It will not allow of an adequate representation of the whole of our topographical and orographical knowledge of the earth's surface even in those districts which as yet are not fully surveyed, while in regions which are completely mapped such a scale would limit

* H. Habenicht. (Ausland, Nos. 1 and 19: 1892).

E. Hammer. Zur Projektion der Erdkarte 1: 1,000,000 (Ausland, No. 40: 1892).

‡ A. Penck. Zur Erdkarte im Massstabe von 1:1,000,000 (Ausland, No. 52: 1891; No. 19: 1892).

the amount of detail which could be inserted to such an extent as to materially prejudice the practical utility of the map.

A moment's consideration of existing maps of the kind will show the truth of this statement. Stieler's atlas has maps on the scale of 1:3,700,000 (58.38 miles to 1 inch) of Europe and the United States. South and Central Europe are also given on the scale of 1:1,500,000 (23.67 miles to 1 inch). The maps of countries not so favoured are very evidently too crowded. The scale of 1: 3,700,000 must therefore be looked upon as distinctly too small for a map such as that proposed. Still, though for some parts of the earth even the scale of 1: 2,000,000 may be too large, it must be borne in mind that the object of this scheme is to condense all our geographical knowledge, and place it upon a map which shall give a representation of the earth's surface uniform both in scale and style of execution. It appears from a paper by Mr. J. G. Bartholemew* that 56 per cent. of the land surface of the globe has already been fully surveyed, while only 12 per cent. can be described as unexplored. It is therefore apparent that only for one-eighth of the land surface of the globe could the scale of 1:1,000,000 be considered too large, while for one-half it might rather be considered too small.

During the many years that will be occupied in the execution of the map, the unexplored areas will be greatly reduced, and it would hardly seem worth while on their account to choose a smaller scale than is otherwise suitable.

Equally absurd would be the adoption of a smaller scale for these regions, as such a measure would make the map useless for purposes of measurement and comparison. In addition to the general advantages hitherto claimed for the scale of 1:1,000,000, there is the special advantage that any metre-measure can be used as a scale for a map constructed on this scale, which, moreover, allowing for paper shrinkage is for all practical purposes identical with that of 16 miles to the inch (1 : 1,013,760) and 25 versts to the inch (1 : 1,050,000). It must be borne in mind that the sheets of this map are not intended to be joined together so as to form a map of the whole world or even of a single continent. Asia on this scale would cover 30 feet square. For such a purpose maps on a smaller scale or in a different style of execution would be needed. It might rather be described as an “Atlas of the World," the term being used in the sense in which it is applied to the largescale maps of certain countries, the separate sheets of which no one would attempt to piece together into a map of the whole country.

This being the case there is no need for a projection such as that described by Sir Henry James (R.G.S. Journal, Vol. XXX., p. 106), which allows of the representation of a large extent of the earth's surface on one plane; the projection known as the polyhedric, employed for the ordnance maps of Germany, the AustroHungarian Empire, the United States and Japan, and the Indian Transfrontier maps, would meet the requirements. There are two methods of carrying out this projection; it can either be done separately for each section or on truncated cones corresponding to successive zones of the map. The first method produces rectilinear trapezes, and adjoining sheets of the same column or zone can be fitted together with mathematical accuracy. There is, however, a certain amount of distortion towards the edges, and the parallels of latitude do not appear as continuous curves when two sheets are fitted together. The second method gives us non-rectilinear trapezes, which can be accurately fitted together within the same zone only. The distortion is, however, considerably less, and the sheets even of different zones will fit together at least as well as paper-shrinkage allows the sheets of a map projected on to a plane to fit together.

*Scott. Geog. Mag., 1890, pp. 293, 575; 1891, pp. 124, 586.

It appears that the discrepancy between the size of such a sheet and the area of the Earth's surface which it represents increases approximately with the square of its size, so that the smaller the sections, the more accurate the map. On the other hand, the larger the sections the smaller the cost of printing the whole map.

Practical considerations make a division of the map into 3° or 5° zones advisable, these being the only divisors of 90° that will give sheets of a convenient size on the scale of 1: 1,000,000. Even the inaccuracy in 5° sections is not as great as that caused by paper-shrinkage. For all practical purposes then we may look upon the projection of the map on 3° or 5° sections as mathematically accurate. There is no need for an equivalent projection of each separate section. This would produce curved meridians, and render the joining together of sheets within the same zone impossible.

Towards the poles the longitudinal extent of the sections is, of course, reduced. On the parallel of 60° it is but half of what it was on the Equator. It would seem advisable then to divide the zones between the poles and 60° into half as many columns as between 60° and the Equator. This would make it possible to have the sheets of the map as nearly as possible of equal size, without making a change in their division, which would complicate the numbering of the sheets. Several considerations point to a division of the map into 5° sections as in every respect the most suitable. Both the single-column sheets between the Equator and 60° and the double-column sheets between 60° and the pole will be found to have a handy size, while for purposes of numeration a division which adapts itself to the decimal system is certainly the best.

Dr. Lüddecke mentions as one of the greatest drawbacks of the polyhedric projection the impossibility of fitting any number of sheets together. We have already seen that it allows adjoining sheets of the same zone to be fitted together with mathematical accuracy. Owing to the difference between the radii of the parallels of latitude it will be seen that if sheets of different zones are grouped symmetrically on either side of a meridian, there is a gap between adjoining zones, which increases by geometrical progression with every column's distance from the central meridian. These inaccuracies, however, are for adjoining sheets not greater than those caused by paper-shrinkage; and when we consider that it cannot possibly serve any practical purpose to fit more than nine sheets together, we may safely disregard this objection of Dr. Lüddecke's. Nine sheets of the map would cover a surface of nearly 6 feet square, about as large an extent as can be conveniently overlooked, and even in this case the greatest extent of the gap between the sheets would not be more than of an inch. This refers to sections near the equator; nearer the poles the gaps are even smaller. The widespread use of the Greenwich meridian makes its adoption in the construction of this map seem advisable. De Lannoy de Bissy recommends the subsidiary insertion of longitudes counted from other meridians.

Sea, rivers, and lakes should be shown in blue; the map should contain as much information as possible regarding tides, variations in the extent of lakes, sea-depths, swamps, &c.

In order to give a really satisfactory representation of the relief of the earth's surface, the map should show both the absolute and the relative elevation. This can really only be done satisfactorily by a combination of contours with colouring, either as hypsographical surface colouring varying with the elevation, or as shading to show steepness. Either of these methods has some disadvantages; and the best plan would, therefore, seem to be a combination of both: the relative elevation being shown by hachures, and the absolute elevation by the insertion of a limited number of contours in conjunction with hypsographical surface colouring.

Four contours would seem to suffice, drawn at 100 (328 feet), 300 (984 feet), No. III.-MARCH, 1893.]

S