— ^70 - 



from sheathwearing to sheathless, has returned from sheathless to 

 sheath-wearing. Among the sheath-wearing groups the P. pectinatus- 

 group seems to be related to the P. lucens-group; the leaves of 

 P. jjectinatus are like the P. luceus-leaves, that are reduced to the 

 midrib only. Such leaves are often found on the lower part of the 

 stem and are generally provided with a shorter or longer sheath; 

 such leaves, however, are to be found on the lower part of the stem 

 of other otherwise sheathless species. 



The P. Bobbinsii-gi'oui:) reminds, with exception of the leaf- 

 sheaths, highly of the P. zosterifolius-grou^. The P. ki/bridiis-gvoup 

 is certainly quite near the P. javanicus-gvoup which again seems 

 closely related to the P. lucens-gro{iiJ. 



Fig. 7. r. Juncifolius (22: 1); transverse sections of leaves: Ig, lacunae; k, vas- 

 cular bundles; s, bast bundles. 



Shape of the leaf. While in the before said three instances it 

 seems easy to decide what must be looked upon as the most 

 primary, it is another case with the shape of leaf. With regard to 

 the leaf's shape and quality we have to do with four types: 1) The 

 floating leaf; 2) the thin, broad submerged leaf; 3) the thin, linear 

 submerged leaf; 4) the big, linear submerged leaf. This last form 

 of leaf, found for instance in the P. pectinatus-gvow^ (Fig. 7 and 6 C), 

 I will leave out of consideration as I think it has issued from a 

 broader leaf reduced to the midrib only. 



No doubt the thin, linear submerged leaf is also derivative. I 

 suppose this when considering the fact that all the groups with 

 only this or with partly this kind of leaf are all more or less 

 reduced in the other characters especially regarding the axial cylinder 

 and bundles in the bark. Finally we have the floating leaf and 



