1915-lG.] BOTANICAL SOCIETY OF EDINBURGH 7 



an incidental formation ; on!}' in a very few favoured spots 

 does it occur to the exclusion of shrub growth, and then 

 it is usually Abies, covering only a small area. The 

 greater part of the divide is covered with shrubs, and a 

 thin belt of trees is found at high altitudes. This great 

 difference is largely due to the difference of rainfall on 

 the two divides, but wind is probably just as important, as 

 the following tables suggest. It should be noted that 

 thouHi the chano-e of climate is sufficient to check tree 

 growth on the Mekong- Yangtze divide, and introduce very 

 considerable differences into the composition of the forest 

 and flora generally, it is not sufficiently great to do away 

 with tree growth altogether. Some of the differences in 

 the composition of the flora too must be ascribed more 

 directly to other causes — for instance, the retreat of the 

 glaciers which has plainlj^ modifled the alpine flora, though 

 this is indirectly due to the changed climate, which has 

 caused the retreat. There is, however, good reason to 

 believe, as I hope to show in a future paper, that whatever 

 the differences in the flora of the two mountain chains 

 now, and whatever gulf separates them, they must once 

 have been derived from a common origin. 



The following tables were drawn up after taking a series 

 of observations with a small instrument, called an evapori- 

 meter, devised by Sir Francis Darwin. It consists of a 

 small cylindrical copper vessel fitted with a lid, and an 

 elbow-joint carrying a capillary tube gauge of glass, with 

 a scale of millimetres. A small frame, inserted through a 

 slot in the lid, serves to spread a small T-shaped piece of 

 blotting-paper which dips into the water. It w^as not a 

 very satisfactory instrument, as the evaporating surface 

 was, in humid air, too small to give visible results, while, 

 on the other hand, the capillary tube gauge from which the 

 amount of water evaporated was read off, was apt to get 

 clogged during the fine drizzling rains on the Mekong- 

 Salween divide, and so vitiate the readings. Nevertheless 

 it served to give some indication of the comparative rates 

 of evaporation at the selected stations (see tables), though 

 the paucity of records, partly owing to the above-mentioned 

 defects, and partly to causes bej-ond my control, renders the 

 results only approximate. It must not be forgotten that 



