190 MISSOURI BOTANICAL GARDEN. 



ized flowers show conclusively that they thrust this well 

 into the stigmatic canal, — in some cases apparently even 

 into the top of the ovarian cells, which, owing to the short 

 style and the deep stigmatic notches, they can reach easily 

 with their long maxillary tentacles. In ovipositing, they 

 doubtless back down between the upper ends of the 

 stamens, and the ovary is pierced at about its middle. The 

 septa, and the median line of each cell (which is produced 

 into the cell as a false dissepiment, dividing the cell into 

 two ), being covered by the appressed lower part of the fila- 

 ment, the moth is constrained to puncture the wall part 

 way between the true and false partitions, as she commonly 

 does in filamentosa* this being also its thinnest part. At 

 San Diego, where the moths were more abundant than at 

 Banning, as many as three to five punctures in a vertical 

 series were seen several times in one cell of an ovary, the 

 other cells sometimes showing none at all. 



The compound microscope reveals the presence of pollen 

 in the stigmas of all such flowers punctured for oviposition 

 as I have examined, and a hand lens commonly shows it, 

 though not always. The open stylar canal is frequented 

 by large numbers of a white Thrips, which sometimes pene- 

 trate into the cells of the ovary, and doubtless scatter 

 the pollen greatly, perhaps devouring some of it, On dried 

 fruits of the preceding season, the perforations made by 

 escaping larvae are commonly elevated, the uninjured tissue 

 evidently shrinking more in drying than that immediately 

 surrounding the tunnel of the larva. These dried fruits 

 sometimes show constrictions corresponding to the points 

 of oviposition, but they are frequently obliterated in the 

 development of the pulpy exocarp. 



Y. australis, (Engelm.). (PL 4, 5).— A study of 

 material on the Sierra Blanca plateau of southwestern 

 Texas, and a careful review of the literature of the species 

 related to baccata, enables me to fix without further ques- 



* Riley, I c. pi. 36, f. 2. 



