1894. ] fiistology of the Pontederiacee. 183 
refers to Meyen for authority for the statement that the 
“membrane of the sac covering the ends of the crystal ceases 
to be apparent, so that the latter seems to protrude freely 
into the air-space.” None in the diaphragms of the P. cor- 
data studied presented this phenomenon. On the contrary, 
it was found that all were in thick-walled sacs; i ita 
probability, the unusual thickness of the walls partly explains 
the long resistance to the action of acids. Concentrated HCl 
applied to the section of the diaphragm slowly dissolves the 
taphides in one to two minutes, while the lance-shaped crys- 
tals were not completely dissolved for 11 hours or more, leav- 
ing the thick wall of the sac but slightly shrunken. In 
Heteranthera limosa, no long crystals were observed. There 
was, however, an abundance of raphides, especially in the 
small peripheral cavities of the petiole. 
There was also present in most of the cross-plates of all the 
five Species, but most abundant in P. cordata, a reddish- 
brown secretion in ellipsoid or dumb-bell shaped cells with 
pad thick walls, scattered in few numbers among the stellate 
. : igs. 5 and 6). Ether acts very slowly on this se- 
cretion, Preliminary experiments led to the belief that it 
Mines... But, in 24-48 hours, the section left in ether 
vay a yellowish-brown color in the cells immediately sur- 
the ce the ellipsoidal cell, intimating the slow dissolution of 
lution Tn Benzine acted much more rapidly in its disso- 
there * Nall probability, the substance is a fatty otl, though 
foom for doubt. 7 i 
P. inden the upper row of long palisade cells of the leaf of 
lowish se @ are many others of similar shape containing a yel- 
betwee aes (fig. 7). These are somewhat regularly placed 
tial ee Bang many stomates, as can be seen in a thin tangen- 
this substar Probably also in P. crassipes and H. graminea, 
rmined nee occurs sparingly. Nothing was positively de- 
n th aS to its composition. 
iarity “tigege stellate tissue of P. cordata, a marked pecul- 
Cells met ees in the walls where the arms of adjoining 
dozen . nder high power frequently were seen one to a 
* More breaks in the thick wall which separates 
Served pin:.\'8 ©) Incross section, small pits were ob- 
n§ a narrow partition. In testing for cellulose 
€ stellat rated sulphuric acid shrunk the protoplasm in 
* cells, in some leaving a fine shrunken thread run- 
