ioo THE WATERLILIES. 



on the authority of Caspary and Braun (in personal letters), reluctantly 

 accepts the belief that the flowers are axillary, but the bract and bracte- 

 oles are carried up on the summit of the peduncle as anterior and lateral 

 sepals. The argument for this has been based largely upon the uncertain 

 ground of teratology. Planchon (1850, e] figured a flower of N. cacrulea 

 "whose peduncle showed accidentally [at the lower end] three spatulate 

 bracts, with petioles wavy here and there on the margins as the sepals are 

 at base." * Caspary (fide Eichler) found in the same species a flower 

 with the anterior sepal at the base of the peduncle as a ribbon-shaped 

 bract, the lateral sepals also pushed down, and the posterior sepal in its 

 proper place. This was followed by four sepals in a whorl, placed diago- 

 nally like the second whorl of a normal flower. Caspary found similar 

 cases in N. alba, gigantea, rubra, and hybrids of N. caerulea and capensis. 

 My own observations on the development of the flower of N. lotus fully 

 substantiate Caspary's view, and, I may say, completely turned me from 

 an opponent to an adherent of his theory. The low insertion of the 

 rudiment of the outer sepal and its very early appearance, followed at once 

 by the two lateral sepals, are convincing. This also explains the peculiar 

 relations of the sepals in the bud ; on any other theory, except a blank 

 displacement, the anterior sepal should be, like the posterior, covered by 

 the lateral sepals. 



It remains, however, to explain the presence of a single posterior sepal 

 and the oblique position of the next four floral leaves. Caspary and Braun 

 looked upon the single sepal as an " Erganzungsblatt. " But if the floral 

 diagram of Nuphar as given by Baillon ( j : 82) can be depended on, in 

 which the outermost of the five sepals is posterior and the petals " are 

 inserted along a spiral," we have the most perfect starting point for the 

 development of the flower of Nymphaea. The outermost sepal of Nuphar 

 homologizes with the posterior sepal of Nymphaea, and a very slight 

 shifting of the remaining four sepals away from the median anterior line 

 gives them the oblique position of the four outer petals of Nymphaea. 

 The strongly sepaloid character of these petals, especially in Brachyceras, 

 indicates such an origin. As to the inner petals, from the acycly of 

 Nuphar, Nymphaea shows a continually increasing tendency to cyclic 

 symmetry. In N. lotus the rudiments of all of the floral leaves (excepting 

 the three outer sepals) arise in series, one after another, but Payer's 

 account Q{ N. alba places the four outer petals in a whorl, and the inner 



dont le pedoncule offre accidentellement trois bractecs spathulees, a petiole ondule 



ca-et-la sur les bords, comme les sepales le sont a Icur base. 



