(UuAP. VII. SLEEP OF LEAVES. 325 



close contact ; and from this circumstance it might be thought 

 that the object of the folding was the protection of their lower 

 surfaces. If this had been the case, it would have formed 

 a strongly marked exception to the rule, that when there is any 

 difference in the degree of protection from radiation of the two 

 surfaces of the leaves, it is always the upper surface which is 

 the best protected. But that the folding of the leaflets, and 

 consequent mutual approximation of their lower surfaces, 

 serves merely to allow them to sink down Tcrtically, may bo 



Fig. 127. 



V 



A. B. 



Oxalis acetosella: A, leaf seea from vertically above; B, diagram of leaf 

 asl«ep, also seen from vertically above. 



inferred from the fact that when the leaflets do not radiate 

 from the summit of a common petiole, or, again, when there is 

 plenty of room, from the sub-petioles not being very short, the 

 leaflets sink down without becoming folded. This occurs with 

 the leaflets of 0. sensiHva, Plumierii, and hupleurifolia. 



There is no use in giving a long list of the many species 

 which sleep in the above described manner. This holds good 

 with species having rather fleshy leaves, like those of 0. carnosa, 

 or large leaves like those of 0. Orti-gesii, or four leaflets like 

 those of 0. variabilis. There are, however, some species which 

 show no signs of sleep, viz., 0. pentnphylla, ennenphylta, hirta, 

 and rubella. We will now describe the nature of the movements 

 in some of the species. 



Oxalis acetosella. — The. movement of a leaflet, together with 

 that of the main petiole, are shown in the following dia- 

 gram (Fig. 128), traced between 11 a.m. on October 4th and 

 7.45 A.M. on the 5th. After 5.30 p.m. on the 4th the leaflet sank 

 rapidly, and at 7 p.m. depended vertically. Fw some time 

 before it assumed this latter position, its movemectB could, of 

 course, no longer be traced on the vertical glass, and the 

 broken line in the diagram ought to be extended much farther 



