22Q MISSOURI BOTANICAL GARDEN. 



fond of nectar, and with phytophagous larvae, which is to 

 be regarded as the progenitor of the Pronubas and their 

 close relatives of the genus Prodoxus, the history of which 

 has been so well summarized by Professor Eiley in the 

 Third Garden Report. Whether the first Prodoxids were 

 also pollen feeders, so that their mouth parts became 

 accidentally laden with pollen between their visits to the 

 stigma in search of its secretion, I cannot surmise. As my 

 friend Dr. Lind has suggested, they may have learned to 

 collect this substance and deposit it in the stigma through an 

 instinct such as prompts the mud-dauber to place its prey 

 about the point where its eggs are laid, or the bee to deposit 

 honey and pollen where it can be used by the lame, as 

 though its young were to be directly nourished by the pol- 

 len. At present, however, the act appears to be strictly 

 voluntary, without food compensation, and entirely con- 

 nected with the fertilization of the ovules. The fact that 

 the septal glands of the several species still secrete at least 

 a small amount of nectar may, as Riley believes to be the 

 case,* depend upon the indirect utility of this secretion near 

 the base of the flower in drawing insects other than Pronuba 

 away from the stigma ; or, as I am inclined to think, it may 

 merely indicate that the abortion of the glands is not yet 

 complete. The glands themselves need not be expected to 

 disappear or to be reduced in size to an extent correspond- 

 ing with their loss of activity under this latter supposition, 

 for their persistence does not itself imply any drain on the 

 forces of the plant, since they represent mere clefts in the 

 connate walls of the carpels which have united to make the 

 compound Yucca ovary. It is possible, however, that the 

 rather abundant stigmatic secretion is not, as I have sup- 

 posed, now being reduced from a still more abundant 

 quantity that once served as true nectar, and this 

 assumption of a former nectariferous function of the stigma 

 would be superfluous if Dr. Lind's theory were correct, 



* i.e. no 



