IX Leaves 173 



water, and are relatively shrivelled, their walls are applied 

 to one another and the stoma is closed. The condition 

 determining the action of these ventilators is explained 

 then, in the first place, by reference to the state of water- 

 tension within the guard-cells. The state of water-tension 

 in the plant depends primarily upon the relation between 

 the water within the plant and the external conditions of 

 drought or humidity. Thus on dry days, when the plant is 

 necessarily losing much water to the air, when the leaves in 

 consequence tend to droop, the water-tension of the guard- 

 cells is necessarily lessened, and the stomata automatically 

 close, or partially close, thus beneficially retarding the drying 

 up of the plant. This explanation was for a long time alone 

 current : its insufficiency gradually became clear as obser- 

 vation showed that stomata in sunshine tend to open, in 

 darkness to close ; that light in fact is of more importance 

 to their movements than is the humidity of the atmosphere. 

 But how does the light act ? It does not usually affect the 

 opening or closing of ventilators : what can be the rationale 

 of this strange phenomenon ? This Schwendener has 

 explained by help of their chlorophyll already mentioned ; 

 they form starch and grape-sugar (glucose) as mere epider- 

 mis cells cannot do. Hence when assimilating vigorously 

 their glucose draws water from the neighbouring cells, they 

 become turgescent, and so open ; while, when light fails 

 and their activity and stores diminish or disappear they 

 close again. It seems, moreover (and thirdly), that the 

 light, apart from its influence on the guard-cells through the 

 formation of starch and consequently sugar, has a direct 

 influence on the state of water-tension in the cells. We 

 have already noticed a similar influence in the case of those 

 plants which bend towards the light. The rationale of 

 this has next to be sought for, and although neither so ob- 

 vious or certain as in the preceding case, need not be 



