DR THOMAS R. FRASER ON STROPHANTHUS HISPIDTJS. 991 



water. The effect of immersion is a gradual transverse rounding of the pericarp, with a 

 corresponding approximation of the placental or ventral edges and a consequent narrow- 

 ing of the placental or ventral surface of the follicle. An originally flat entire pericarp, 

 one inch and a quarter wide, may in a few hours become so greatly rounded that its 

 edges approach to within three-quarters of an inch from each other. If the pericarp be 

 then allowed to dry, by merely exposing it to the air at an ordinary temperature, it 

 gradually resumes its original flat shape, its edges gradually separating from each other 

 to their original distance. If the endocarp alone be immersed in water and then removed 

 and allowed to dry, it assumes the same changes of form as the entire pericarp, but 

 they are more rapidly accomplished. If the endocarp be removed from the pericarp, 

 the conjoined meso- and epicarp also react in the same way as the entire pericarp, or as 

 the endocarp alone, under the influence of moisture and dryness. 



The above experiments seem to show that the changes in form are not dependent on 

 the anatomical structure of any one part of the pericarp. As the changes occur, however, 

 most rapidly and completely in the detached endocarp, the presence in it of elongated 

 indurated cells, arranged in circumferential and longitudinal directions (see Plate V. 

 figs, be and 5d), may confer, as Blondel * has supposed, on this portion of the pericarp 

 a special facility of movement during the change from the condition of moistness to 

 that of dryness. It cannot, however, be overlooked that the existence of indurated cells 

 in the endocarp would strengthen the entire pericarp, and, by enabling it more effectually 

 to resist any bursting force operating in the interior of the follicle, would prevent splitting 

 of the follicle elsewhere than at the feebly resisting placental or ventral surface. It will 

 afterwards be pointed out that, for the satisfactory extrusion of the seeds, it is of import- 

 ance that dehiscence should occur at the ventral surface. 



When the separation of the inverted carpellary edges and the resulting expansion of 

 the ventral or placental surface of the follicle has advanced to a certain stage, the latter- 

 opens either at the middle line alone, by complete separation of the edges of the two 

 previously united carpellary margins, or both at the middle line and at other parts of the 

 dried and brittle ventral surface, by several longitudinal splittings. 



The dragging of the inverted carpellary edges from the interior to the surface of the 

 follicle induces a change in the position and form of the placenta, which is attached to 

 these edges. By this movement it is brought nearer to the ventral surface of the follicle, 

 and, as it is being drawn from its original position, its spiral is unfolded (Plate VII. 

 figs. 3, 4, 5, and 6). The seeds, imbedded at maturity in the elastic hairs which surround 

 them, and fixed in position, also, by their comose appendages, are unable to accompany 

 the placenta in its changes of position, and they thus become detached from it by rupture 

 of the now dry and brittle funiculi. 



The actual extrusion of the seeds appears to be produced by the pressure exerted 

 upon them by the hairs contained in the follicle, and especially by the long basal seed- 

 hairs, which separate the seeds from the endocarp and from each other. These hairs, in 



* Bulletin Ge~n<fral de TMrapeutique, 1888, pp. 100-103. 



