Box.— Vol. I.] CAMPBELL— XAIAS AND ZANNICHELLLA. 43 



it to the single flower of Naias and considers the envelope 

 as the homologue of the carpellary envelope in the latter. 

 Morong (1893, p. 56) in his description of the plant, how- 

 ever, evidently considers each carpel as a distinct flower, 

 a view whose correctness is borne out by the study of the 

 development. The number of carpels in a cluster varies 

 from two to about eight. In the form studied by the writer 

 there were usually five. 



Each flower consists of a single flask-shaped carpel (fig. 

 104), having the ovary almost globular in form, and the style 

 expanded above into a peltate stigma, which is much more 

 strongly developed on the outer side, and whose edges are 

 more or less evidently lobed . A very evident canal traverses 

 the style. Schumann, following Irmisch, considers the 

 inflorescence (or " flower ") as the termination of the main 

 axis, which is then replaced by a lateral shoot; but we have 

 already seen that this view is incorrect, and the inflorescence 

 is the result of the dichotomy of the main shoot, whose other 

 member continues the growth of the axis. 



Very soon after the stem-apex is divided, the young inflor- 

 escence becomes recognizable by its broadened form, the 

 other member of the dichotomy remaining more pointed 

 (fig. 85 ) . It is soon evident that the broadening of the floral 

 branch is due to a second dichotomy in it. The branch at 

 this time closely resembles the young staminal one, but is 

 somewhat larger. The arrangement of the tissues is very 

 plain. The epidermis is separated from the axial plerome 

 by a layer of periblem, which is more strongly developed on 

 the lower side of the branch. At the time of the first dichot- 

 omy, the plerome consists of four rows of cells (seen in 

 longitudinal section), but at the top it forks, one branch 

 going to each of the branches of the inflorescence (fig. 92). 

 In case there are but two carpels, there is no further division ; 

 but in all the specimens examined by me the dichotomy was 

 repeated in each branch. The fifth carpel presumably is 

 formed by a third dichotomy in one of the secondary 

 branches, and where there are eight, it is natural to suppose 

 that all of the branches have divided. It is plain from this 



