368 



NATURE 



[August 14, 1879 



are an indefinite number of positions these marked mole- 

 cules (representing the portion of introduced gas) could 

 take up in the vessel, consistent with equilibrium, and 

 there would be (practically) an infinite number of chances 

 against the portion of introduced gas arranging itself as 

 in a vacuum : for to do this, the marked molecules (com- 

 posing the portion of gas) would require to arrange 

 themselves in such a way that their mean distance is 

 everywhere the same throughout the vessel, a contingency 

 almost infinitely unlikely. What applies to marked mole- 

 cules applies to chemically different molecules of equal 

 mass, or which are dynamically similar. Hence it would 

 follow that portions of gas of the same kind, or portions 

 of chemically different gases of equal molecular mass 

 could not be said " to behave to each other as vacua," in 

 regard to arrangement. On the other hand, where the 

 gases have unequal molecular masses there is (as we have 

 seen) a forcible dynamical tendency for the gases to 

 diffuse themselves symmetrically through each other, so 

 that each gas behaves to the other as a vacuum, each gas 

 becoming uniformly diffused through the vessel, as if it 

 existed alone in a vacuum. The successive introduction 

 into a vessel of portions of gas of the same kind (or of 

 portions of chemically different gases of equal molecular 

 mass) may be compared to the introduction into any 

 closed space of successive sets of equal differently coloured 

 perfectly elastic balls (the balls being supposed left in free 

 motion among each other in analogy with the molecules 

 of a gas), when evidently no one arrangement of the 

 different coloured balls in the closed space (at any given 

 instant) could be said to be more probable than another, 

 and it would be extremely unlikely that the sets of 

 coloured balls should " behave to each other as vacua," 

 in the sense of each set diffusing itself symmetrically 

 through the closed space, as it would do in a vacuum. 

 But if the sets of balls were of unequal masses [in analogy 

 to gases of unequal molecular masses], then no doubt the 

 different sets would behave to each other as vacua, or 

 each set would forcibly tend to arrange itself according to 

 strict dynamical principles, so as to pervade uniformly 

 the entire closed space, precisely as it would do a vacuum. 



S. ToLVER Preston 



OBSERVATIONS ON THE PHYSICAL GEO- 

 GRAPHY AND GEOLOGY OF MADAGASCAR 



A LTHOUGH Madagascar is known to be the third 

 •**• largest island in the world, its actual size and ex- 

 tent is not very generally understood. It is easy to see 

 how misconception on this point arises, for in maps the 

 island is usually seen only in connection with Africa, and 

 that gfreat continent is so large that it dwarfs by com- 

 parison with itself everything in its near neighbourhood, 

 so that the really large island sheltering under its south- 

 eastern side appears but an inconsiderable appendage to 

 its vast neighbour. If, however, we take a good-sized 

 map of Madagascar, and put by its side the outline, to the 

 same scale, of another country with whose dimensions we are 

 familiar, such, for instance, as England, we begin to realise 

 how important an island it is as regards size, being nearly 

 1,000 miles long' by about 250 in average breadth, so 

 that it is nearly four times as large as England and 

 Wales. 



During the last ten years much light has been thrown 

 upon the physical geography of Madagascar, principally 

 through the researches of M. Alfred Grandidier, and the 

 numerous exploratory journeys made in various parts of 

 the country by missionaries and others. Until a very 

 recent period there was no reliable map of the island ; a 

 number of mountain ranges were shown in positions 

 where no such geographical features are to be found, and 

 the physical geography was completely misunderstood. 

 But it is now quite clear that instead of a " central 

 ' More exactly, 975 miles. 



mountain chain," as described in most histories and 



gazetteers, there is an elevated mountainous region, which, 

 however, does not occupy the centre of the island, but is 

 more to the east and north, leaving a considerable extent 

 of country to the west, and all beyond the 23rd parallel 

 of south latitude, at a much lower level above the sea. 

 Broadly speaking, therefore, Madagascar consists of two 

 great divisions, viz., (i) an elevated interior region raised 

 from 3,000 to 5,000 feet above the sea-level ; and (2) a 

 comparatively level country surrounding it, not much 

 exceeding 400 or 500 feet in elevation, and most extensive 

 in the west and south. 



The elevated region is largely composed of primary and 

 crystalline rocks. Lines of hills traverse it in all direc- 

 tions, but they do not rise to a very great height, the 

 highest points in the countrj', the peaks of the Ankiratra 

 group of mountains, being a little under 9,000 feet above the 

 sea-level. A very large extent of this portion of Mada- 

 gascar is covered with bright red clay, through which the 

 granite and basaltic rocks protnide. But there are also 

 extensive rice-plains, especially near the capital cities of the 

 two chief provinces, where there is a rich black allu\"ial soil ; 

 and it can hardly be doubted that some at least of these 

 plains, from their perfect level, out of which the red clay 

 hills rise like islands, have foi-merly been the beds of 

 extensive lakes, subsequently drained, possibly by slight 

 changes in the level through subterranean action. 



A good deal of this portion of Madagascar is bare and 

 somewhat dreary-looking country. The long roUing 

 moor-like hills are only covered witli a coarse grass, which 

 becomes very brown and dry towards the close of the 

 seven months' rainless season ; but the hollows and river- 

 valleys are often filled with a luxuriant tropical vegetation, 

 and, wherever there is population, with the bright green 

 of the rice-fields. There is, nevertheless, an element of 

 grandeur in the landscape, from the great extent of country 

 visible from many points in the clear, pure atmosphere, 

 which renders veiy distant objects wonderfully sharp and 

 distinct. And many portions of the central region possess 

 still greater claims to admiration from its picturesque 

 mountain scener)'. 



In the southern Betsildo country, the grand and varied 

 forms of the mountains filled me with an exultant kind 

 of delight. To the south was a crowd of mountain-tops, 

 peak beyond peak, with the greatest variety of outline : 

 one had the appearance of a colossal truncated spire, 

 another had a jagged saw-like ridge, another was like a 

 pyramid with successive steps, and another an enormous 

 dome. Their summits were never long free from clouds, 

 and many of the peaks must be at least 3,000 feet above 

 the plain. 



Sections taken by the aneroid across this elevated 

 region from east to west at the latitude of the capital 

 show that it has a depression in the centre, the edges on 

 either side being considerably higher than the country 

 between them. At some points this height of 4,000 to 

 5,000 feet is gained by a series of steps from the maritime 

 plains, each range of hills rising higher and higher, while 

 at other points it descends almost at one steep slope for 

 nearly 3,000 feet. The water-shed is not in the centre 

 of the island, but is much nearer the eastern side. 

 Through the eastern wall many of the rivers cut their 

 way by magnificent gorges, amidst dense forest, finding 

 their way to the sea by a succession of rapids and cata- 

 racts, and occasionally by stupendous falls, as in the case 

 of the M^titinana river, which descends at one plunge 500 

 or 600 feet. Some of the western rivers, also, are said 

 to form grand waterfalls, particularly that of the Jtlania, 

 whose sound is reported to be heard at a distance of two 

 days' journey, i.e., about forty to fifty miles. 



The lower region of Madagascar consists of extensive 

 plains only a few hundred feet above the sea-level, but 

 there are at least three prominent chains of hills travers- 

 ing it from north to south, one of which appears nearly 



