THE FRUIT OF TROPIDOCARPUM. Ill 



tion and completely destitute of partition. If the valves are 

 removed it will be seen that at the point where the change in 

 the valves occurs, the two sides of the replum (Fig. 4) sud- 

 denly diverge, so that, instead of being separated only by 

 the thickness of the pod, they are separated by its whole 

 breadth. Both the upper and the lower parts of the pods 

 bear ovules and mature seeds, which are similar so far as 

 observed. The third or rhombic type (Fig. 14) is linear, 

 elongated and 2-valved, but completely unicellular and flat- 

 tened laterally throughout, so that the valves, instead of being 

 conduplicate, are scarcely carinate. The cross-section of this 

 capsule is rhombic as shown in Fig. 7, but is often more 

 strongly flattened. The fourth or capparideum type of fruit 

 (Fig. 5), is shorter and considerably broader. It is not 

 strongly compressed, but is 4-valved and, except at the very 

 base, devoid of septum. There are four placentae, all 

 maturing seeds. 



It will be noted that of the four kinds of capsule^; described, 

 each has a characteristic replum. In the gracile type, after 

 the falling of the valves, the replum (Fig. 3) is seen to con- 

 sist of two closely approximate and essentially parallel col- 

 ums bearing a complete and very narrow septum. In the 

 duhinm type the upper portion of the replum (Fig. 4, a) is 

 exactly as in the gracile type, while in the lower portion 

 (Fig. 4, b) the two columns are farther apart and bear no 

 septum. In the rhombic type the replum (Fig. 15) consists 

 of two parallel columns united merely at the base and sum- 

 mit, and bearing no septum. In the capparideum type the 

 replum (Fig. 8) is formed of four similar columns united at 

 base and summit, and also without septum. It thus happens 

 that even after the dehiscence of the capsules the repla, 

 which are so characteristic, indicate accurately the nature of 

 the fruit. 



As these four types of capsules are so striking, it is only 

 natural that they should have been eagerly taken as furnish- 

 ing excellent specific characters in a group where other dis- 

 tinctions were so poor. Yet it is very important, before 



