105] NORTH AMERICAN AND WEST INDIAN CUSCUTAYUNCKER 15 



dilatations of the lower (attached) part of the filaments, perhaps of the 

 character of stipules, as Prof. A. Braun suggests; or they are a sort of 

 stamineal crown, attached at base to the corolla, but not a duplication of 

 the same." 



Miss Cunningham's idea regarding the origin of the scales does not 

 agree with Engelmann's. She states (1898): "it was noticed that in some 

 species the filament of the stamen extends under the apex of the scale, in 

 others the base of the filament can be traced nearly to the base of the corolla, 

 while the scale forms two lateral wings, one on either side of the filament. 

 For this work specimens from each of the three groups were examined. 

 Longitudinal sections were made through the corolla with its attached 

 stamen and scale and a careful study showed that the scales have their 

 origin from the corolla. The stamens also originate from the corolla, but 

 at a diflferent level from the scale so that they cannot possibly be attached 

 to each other. However, in the third section a few species showed some 

 connection between the scale and the filament; but, while there may have 

 been a slight attachment of these parts in individual specimens, yet the 

 examination of other sections fully demonstrated the fact that the origin 

 of the scale is unquestionably from the corolla, and the base of the stamen 

 is slightly above that of the scale. The results of these examinations, so 

 far as made, confirm us in the belief that the scales are not epistamineal, 

 and do not form a stamineal crown, but are petaloid and are in the nature 

 of a duplication of the petals." 



It would appear inconsistent with the morphology of a typical flower 

 to consider the scales as outgrowths of the corolla and at the same time 

 originating opposite the stamens. After examination of a great many 

 specimens of different species and of sections made of many of them, it is 

 quite evident to me that the scales are outgrowths of the filaments, as 

 Engelmann believed. Inspection of various species in other genera of the 

 Convolvulaceae shows that there are numerous filamentous appendages 

 frequently present towards the base of the filaments. In Convolvulus 

 septum, for example, the filaments are found to be flattened laterally 

 forming slight wings which bear numerous filamentous projections. It 

 does not seem difficult to imagine these wings uniting at the base between 

 two filaments forming the "inverted arch" or bridge, as I shall call this 

 region and also uniting above and becoming free as the filament becomes 

 more firmly attached to the corolla. I believe that the epistaminal 

 scales of Cuscuta have originated in this manner. The scales are con- 

 sidered in the descriptions as continuing to the base of the corolla in 

 all cases and the description of the height of the bridge is in relation 

 to their total length. The bridge is of different heights and fairly char- 

 acteristic for the different species. The shape, size and texture of the 

 scales and the length and abundance of their processes is very different in 



