474 



MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



[yol. 9, No. 3 



It is difficult at present to suggest whether the subsidiary bundles, such as 

 occur in corollas of Stenopadus kunhardtii, S. stipitatus, and Stomatochaeta 

 crassifolia, represent features of a primitive venation pattern or are secondary 

 additions. 



In the zygomorphic corolla of Geberinae, union of the abaxial three lobes has 

 involved the reduction from pairs of laterals to single bundles in the united 

 portion. Stages in this simplification are offered by Neblinaea and two species of 

 Gongylolepis. The final simplified form was described by Koch (1930a, b) in 

 Perezia microcephala and Mutisia taraxacifolia, although she suggests no stages 

 in its origin. 



The genus Glossarion appears to represent a special case in which the lobes 

 of an actinomorphic flower have united without any great alteration of the vas- 

 cular system. 



(2) Achene. In the achenes of Stenopadus kunhardtii, S. stipitatus, and 5. 

 cucullatus, adjacent laterals continue downward without fusing (although a 

 vascular plexus often interconnects many of these bundles at the achene top). In 

 most of the taxa considered here, however, the pairs of laterals unite in the 

 achene top and continue downward as five of the exterior achene wall bundles. 

 Ten bundles are very commonly present in flowers showing this condition, and 

 the additional five bundles are accounted for by continuation of median corolla 

 veins into the achene, as in Stenopadus campestris. Very often, however, median 

 bundles are not present in the corolla, or present only as isolated portions in the 

 lobes. The presence of additional bundles in the achene may then be accounted 

 for by preferential retention of bundles in the achene, although their upward ex- 

 tension as median traces has been lost phyletically in the corolla. Koch (1930b) 

 described the achene anatomy for Mutisieae only in an advanced form, Mutisia 

 taraxacifolia, in which the bundle numoer in the achene was reduced to five. The 

 presence of ten bundles in achenes of Mutisieae has been described previously 

 by Giroux (1937) in Dicoma tomentosa. 



The reader will have noted that a difference occurs between the genera 

 Stenopadus, Stomatochaeta, and Chimantaea, and the remaining genera in this 

 study in the configuration of bundles at the achene base. In the three genera 

 named a cylinder of independent bundles enters the receptacle, while in the 

 others all the achene bundles fuse to a single strand. Snow (1945) has asked, 

 quite appropriately, why there should be a single vein supplying the base of the 

 achene in the Compositae she investigated, rather than a floral stele consisting 

 of several bundles, such as is found in flowers of other families. This question 

 can be answered by saying that consistent reduction in the vascularization of the 

 composite flower has resulted in this simplified situation, but the presence of an 

 actual cylinder of bundles in the taxa named shows that such a condition is 

 indeed present in primitive Compositae. 



Apparently the pappus of the Mutisieae investigated here is advanced com- 

 pared with that of primitive Heliantheae, and although many of the pappus bristles 

 are quite thick and contain some degree of differentiation among cells in their 

 interior, there are no vascular elements present in the pappus. However, the fact 

 that phloem strands may be seen to lead out to the pappus base in several of the 

 genera (particularly Glossarion and Gongylolepis) could be used in support of the 

 idea that the pappus was primitively a calycine structure more nearly like that 

 shown by certain primitive Heliantheae. 



(3) Style and carpels. Because in flowers of Mutisieae or other Compositae 

 described heretofore only a single circle of bundles has been found in the achene, 



