pointing to the centje of the flower, from which circum- 

 stance it is clear that the slits of dehiscence are not per- 

 fectly latéral but turn a little inwards. 



The abnormal flowers are ail a little shorter, not ex- 

 ceeding 60 c.M., moreover the sepals of the same flower 

 differ more in length as well in width. Besides the corolla 

 may show varions degrees of dialysis and the insertion 

 of the stamens irregularities. 



The pistil always proved undisturbed, both in structure 

 and in relative length. 



The greatest disturbance is, however, shown by the 

 anthers: so e.g. in flg. II the anther bears a curved ap- 

 pendage at top with a stigma-like end. When we trace 

 the origin of this appendage it proves to be in connection 

 with both thecae. Fig. III represents a stamen, from 

 which one of the thecae has been taken away enabling 

 us in this way to see the side of dehiscence as well as 

 the opposed one. The top bears a threadlike appendage 

 without stigma. Although this theca opened at the time 

 of its examination, only a small quantity of pollen was 

 produced; in the split, however, a hard, brown, grainy 

 little body (o) was seen at the base, and even at the edges 

 on the opposite surface. This brown corpuscle stands out 

 more clearly in flg. IV where it coïncides with indubitable 

 distinctness with the style and stigma-like appendage al- 

 ready described. The thecae hâve the aspect of dehiscence 

 while the brown corpuscles coalesce. 



As the brown corpuscles in the other flgures are also 

 marked with o, their further indication will be superfluous. 

 It is worth mentioning that the stronger this corpuscle is, 

 the less is the quantity of pollen in the same anther. In 

 flgures V and VI two other dérangements incidental to 

 the said phenomenon are represented. In fig. V it is a 

 strong twisting, evidently in conséquence of a différence 

 in rate of growth between the filament and an additional 



