RAN ALES 417 



The perianth shows stages in the development of distinct calyx and 

 corolla. In some genera, especially in Nijmphaea, there are many transi- 

 tional forms between stamens and petals, with evidence that the corolla 

 in this family represents sterile stamens, morphologically (Fig. 54). 

 Other, less primitive genera have no staminodia; the distinction between 

 stamen and petal is obvious. But in the polymerous taxa, a line is often 

 not distinct; the nature of the sepals is not clear. In Nuphar, specializa- 

 tion in the perianth has progressed further than in the other Nymphaea- 

 ceae; the sepals have become petaloid, and the petals reduced and 

 somewhat bractlike. In the Cabombaceae, calyx and corolla are sharply 

 distinct; the differentiation of a corolla has accompanied the establish- 

 ment of cyclic arrangement of organs. 



The stamens in these two families show many stages in evolutionary 

 modification (Figs. 54 and 55). The most primitive type is similar to 

 that of the more primitive members of the Magnoliaceae — flat and 

 broad stamens, without distinction of anther and filament (Fig. 55B to 

 D); wall-less, elongate, median sporangia, deeply sunken in the lamina 

 and remote from the vascular supply; and prominent distal appendage. 

 This primitive type is characteristic of Victoria and some species of 

 Nijmphaea. Other taxa show forms transitional to the stamens of 

 Cahomha, Brasenia (Figs. 54 and 55), and species of Euryale and 

 Barclm/a (Fig. 49), which have well-differentiated anther and filament, 

 no sterile, distal appendage, and protuberant sporangia. The Nymphaea- 

 ceae and Brasenia are alike in having introrse dehiscence; Caboinba 

 has extrorse dehiscence. The significance of this difference is discussed 

 in Chap. 4. The water lilies, like the magnolias, demonstrate the evolu- 

 tionary advance of the stamen in two independent lines of the Ranales. 



The morphology of the gynoecium in these families has been much 

 discussed. Connation and adnation, complicated by spiral arrangement 

 of the numerous appendages, has made the perigyny and epigyny of 

 the Nymphaeaceae difficult of anatomical interpretation. NeJumbo, 

 with its isolated carpels, "sunken in the receptacle," increases the ana- 

 tomical complexity. No broad and critical study of the gynoecial struc- 

 ture of these families has been made. 



The morphology of the perigyny and epigyny in the Nymphaeaceae 

 has been interpreted in two ways: as the result of the adnation of the 

 outer, lower organs to the gynoecium; and as a result of the envelop- 

 ment of the bases of the appendages by the receptacle. The ontogenetic 

 growth of receptacular tissue about the free carpels in Nelumbo has 

 been considered evidence that perigyny and epigyny in the other genera 

 have come about in the same way. But ontogeny does not support this 

 interpretation. This family, like the Rosaceae, shows all stages in the 

 development of epigyny — here complicated by spiral arrangement of 



