In the matter of choice of site, the same general considerations as to soil and configuration of the ground that determine our choice in temperate climates, as a rule, hold good. A gravelly or sandy soil, and gradients favourable to natural drainage, are even greater desiderata in the Tropics than in Europe, and this is especially the case in climates characterised by a heavy rainfall; but ideal sites are rare in all countries, and as a rule, one must be content to make the best of less favourably placed spots. In the countries which we are at present considering, the especial danger against which we have to guard is always that of malaria, and hence, in choosing the site for a house or station, the great point is to select one, which is, as far as possible, free from natural or artificial collections of water, within a radius of a quarter of a mile; or at any rate, including such only as can be easily filled in, drained, or otherwise dealt with. The site should also be sufficiently raised above the level of some natural watercourse to afford an adequate outfall for its surface drainage. For a single house, no better position can be selected than the summit of a mound, whether natural or artificial; and such situations are generally to be preferred to the slope of a hill, even where the latter affords a considerably greater elevation. On the sea coast, and not unfrequently in the neighbourhood of some of the great rivers, sand dunes, where sufficiently clad with vegetation to afford a sufficiently stable foundation, form excellent sites for single houses, good examples of which are to be found in Chittagong, where nearly every European residence has its own little hill, on which it is perched by itself; and it is doubtless to this circumstance that the comparative healthiness of the European population of the town, under otherwise unfavourable surroundings, is mainly due. The neighbourhood of jungle, and even of trees, should be as far as possible avoided, for trees undoubtedly harbour mosquitoes, and their presence is generally equivalent to that of malaria: moreover, the appearance of coolness, associated with trees, is deceptive rather than real. As a rule, even when numerous and thickly set, they throw no actual shade on the walls of the house, which hence receives as fully the power of the sun as it would in an open plain, and added to this they obstruct the breeze and generally impede ventilation; so that a house placed in the midst of a glaring, treeless space is often really far cooler than one surrounded with fine timber. Even a garden is by no means too desirable an adjunct to a tropical residence, for unless there is abundant labour to keep it in a condition of perfect neatness, and constant intelligent supervision to ensure that the cultivation of flowers and vegetables be not associated with the breeding of mosquitoes, it is only too likely to originate fever of a luxuriance at least equalling that of its roses and salads.