THE PROBLEM 45 



all joined by their stalks in a single ring, while 

 in the St. John's Wort they are united in groups. 

 The carpels are more often united than free, 

 anong themselves; in the Orchids, stamen and 

 pistil are fused to form the column. The pistil 

 is always in the middle of the flower, but in many 

 families that part of it, the ovary, which contains 

 the ovules, is sunk below the level of the calyx; 

 the ovary is then called inferior, as in the Hem- 

 lock family, the Campanulas and many more. 



It is not my purpose to go any further into 

 the modifications of the flower; my object is to 

 point out that the possession of a flower, in the 

 sense explained, is characteristic of Angiosperms, 

 and that we may naturally expect to find at least 

 the beginnings of a flower in any group of plants 

 from which the Angiosperms are likely to have 

 been derived. 



Before leaving the subject it must be recalled 

 to mind that the flower in many Angiosperms 

 is much simpler than hi the typical cases dealt 

 with above. Often there is no corolla (Knot- 

 grass; Nettle); often there is no perianth at all 

 (Spurge; Willow). What is more important, 

 the flower may be uni-sexual, only the stamens 

 or only the pistil being present, as is the case in 

 all our catkin-bearing trees, and many of the 

 Arum family. Where, however, the individual 

 flowers are very simple, we almost always find a 



