10 



J. O. HAGSTROM, CRJTJLCAL RESEARCHES ON r THE POTAMOGETONS. 



to prevent the leaves to be pulled to pieces. 

 pally consist in anatomical facts: 



The arrangements of the stem princi- 



D 



1) Adaption of the epidermis- 

 cells in being very narrow with 

 maintaining of the usual length or 

 being very short almost cubic with 

 keeping of the width. In general 

 the epidermis cells are 3 — 6 times 

 the width. 



2) Generation of a pseudo- 

 hypoderma. In some species it ap- 

 pears one-layered (P. perfoliatus, 

 P. pulcher etc.), in others two-layer- 

 ed (P. prcelongus, P. ochreatus etc), 

 in one species even three- (2 — 4) 

 layered: P. linguatus. 



3) Developement of particular 

 lists in the angles where the par- 

 tition-walls join the epidermis or 

 hypoderma. A usual fact. 



4) Developement of particular 

 thin or strong subepidermal or subhypodermal (P. prcelongus) bast 

 bundles. A very important distinctive mark. 



5) Evolution of a system of mechanical strands and tissuesboth 

 in the bark and the central cylinder, the endodermis included, too. 



6) Evolution of a plurality of air-channels by decreasing of 

 their lumen; occasionally in non-sclerenchymatous species. In 

 this case a multitude of longitudinal walls in the bark afford 

 the same strength to a plant as the mechanic strands give to 

 another. 



The groups thus established have been distributed on five 

 sections. When, however, such a significance as has hitherto been 

 ascribed to the floating leaves cannot be attributed to them, I have 

 been obliged to disapprove of the old division HeterophyUi, Ho- 

 mophylli etc, and the modifications of it made by låter authors. 

 Eere the behaviour of the leaf-sheaths and ligules offers itself 

 as a convenient princi})le of division. On the peculiarity of /'. 

 jig. 2. Behematic deiioea- densus, see under Latcrales! 



tion »i tln' different trava ut TT . r j. ,. ,, , .,■ . ,1 , 



Bpike-prodnotion.Beethetezti Vcrnation forms a distinctive mark within tlie exten- 



sivc" seotion IV Azillares, 

 The subsection Pusilli may oonveniently be divided into two series: Connati 

 and Convolwti according as the ligules are connate or open, whioh fact has hitherto 

 not been observed at all. 



