TEOPICAL FLORAS 81 



perfectly known countries seems to ine to offer so many ad- 

 vantages that the adoption of it in Java by the Dutch botanists 

 must, I think, be looked upon as an important discovery. It 

 has the great advantage of being at once economical and ef- 

 fective; it brings about the maximum of scientific result with 

 the minimum of cost, of time and of labour. It has proved 

 that the careful and systematic study of very small areas is 

 calculated to extend our knowledge of the vast world of plant- 

 life more than any other that has hitherto been adopted. The 

 plan is to have, in any extensive country or island, a suit- 

 able number of what may be termed " botanical reserves " 

 (but which wdll also serve as zoological reserv^es, especially for 

 bird and insect life) ; these to be of small size, say one square 

 mile each, to be kept absolutely in a state of nature, except 

 the provision of numerous paths giving access to at least one 

 specimen of every species of tree the reserve contains. Ex- 

 perience in Java seems to show that one man, or two if 

 necessary, can keep the paths open, watch for the flowering 

 and fruiting of trees, gather and send specimens to the head 

 of the department, and also, I presume, serve as guide to any 

 botanical visitors to the reserve. But when the trees had been 

 all found, numbered, and named, the same superintendent 

 or keeper would have time and opportunity for the collec- 

 tion of specimens of all the shrubs, climbers, epiphytes, and 

 herbs that grew in the reserve, identifying the place of all 

 the rarer species by direction and distance from the nearest 

 named tree, the epiphytes, orchids, ferns, mosses, etc., being 

 identified by the tree they grew upon being numbered, and 

 made accessible by a path. Of course this area of 3 square 

 kilometres, or about a square mile, may not be in all 

 cases sufficient, but it seems likely to be the most suitable for 

 luxuriant tropical forests. In more open country, as at 

 Campo Santo, a space of from 10 to 50 square miles might 

 be advisable, because the trees on such an area might be as 

 easilv found as in a mile of unbroken forest, and w^ould not 

 be much more numerous. In any new tropical country of 



