THE FRUIT OF TEOPIDOCARPUM. 113 



find two or more of them associated upon the same individual. 

 Noteworthy examples of this variation upon the same indi- 

 vidual may be seen in figures 9 and 12. In figure 9, pod a is 

 of the gracile type ; pods b and g are of the duhium type ; 

 pod d, of the rhombic type ; while pods e and / are 3-valved 

 (a fact not very easily shown in the drawing); and pod i is 

 of the 4-valved or capparideiim type. In figure 12 we have 

 a very similar inflorescence, with the persisting repla after 

 the falling of the valres. Of these repla, a is of the gracile 

 type, 6, c and d of the dubium type, while e and / are par- 

 tially 3-valved, and g partially 4-valved, thus closely approach- 

 ing the capparideum type. 



As the different types occur thus upon the same individuals 

 it is quite impossible to ascribe much value to them in the 

 separation of species. It is certainly true that specimens of 

 the so called T. capparideum^ Greene, with mostly short thick- 

 ish pods, look strikingly different from any of the other pro- 

 posed species, yet close scrutiny shows in almost all plants, 

 examined by the writer, that there are one or more 2-valved 

 and rather frequent 3-valved pods associated with the 4-valved. 

 There are of course many individuals in which the capsules 

 are prevailingly of the dubium, the rhombic, or the cap- 

 parideum type. The question naturally arising is, if these are 

 not species, what is their status ? Are they merely marked 

 forms or teratological states of T. gracile, Hook. ? The chief 

 evidences against this view are that they occur in great num- 

 bers and appear to bear perfect seeds, which may very well 

 "come true" and produce like individuals, although this 

 remains to be proved. On the other hand there are several 

 things to suggest monstrosity. First, there is the inherent 

 variability in all forms of the capsules except the gracile 

 type, a variability not only in different plants but shown so 

 notably in the successive fruits of the same individual. Then 

 it is possible, as we have seen, by the comparison of the differ- 

 ent kinds of capsule, to derive capsules of the dubium, 

 rhombic and capparideum types by a gradual series of modi- 

 fications or deformations of a capsule of the gracile type, 



