72 THE WATERLILIES. 



The lower epidermis (Fig. 29, a) has been frequently referred to in 

 previous pages. It is a perfectly continuous layer, o.oi to o.oiS mm. 

 thick, quite smooth on the outer surface, and consisting for the most part 

 of polygonal cells, many of which are about twice as long as broad (0.018 X 

 0.036 mm.). In these is located the pigment which colors the under side 

 of the leaf. More or less plentifully interspersed with them are thickened 

 circular cells (suberized, according to Schilling, 1894) representing the 

 bases of hairs. Frequently the other epidermal cells are arranged 

 around these in a radial manner. This caused Barthelemy (1874) to 

 speak of them as pseudostomata, and he thought he could see in the 

 center of some of them a small perforation. His opinion lacks confirmation. 



The circular cells are about 0.02 mm. 

 l" across in N.flava, mexicana, amazonum, 

 }^J^ odorata, o. rosca, and tuberosa ; 0.02 to 

 0.028 mm. in N. zanzibariensis X and alba 

 candidissima, and 0.025 to 0.029 mm. in 

 N. tetragona. Their relative number is 

 shown by the following figures, represent- 

 FIO. ^.-Development of mucilage hah- otjv. i n g the number visible in a single field 



alba; 1-5, successive stages. (After Schilling.) . ... 



of my microscope (Leitz, oc. 2, obj. 7) : 



N. /lava 32, mexicana 19-20, amazonum 52, tetragona 21, odorata 36, 

 o. rosea 34, tuberosa 28, rubra 46. In the last-named species permanent 

 hairs stand on many of these basal cells. Hofmeister in 1868 mentioned 

 N. alba in company with Fagus, Salix, Ouercus, &c., as having hairy leaves 

 only in the bud. 



From what has already been said concerning the occurrence of basal 

 cells of hairs in the epidermis of the petioles and lower leaf-surface, it will 

 be seen that hair structures always occur on these parts at some time in 

 their history ; the main ribs of the leaves and their upper surface are alone 

 totally devoid of them. On a leaf rudiment of N. flava a trifle over 

 o.i 6 cm. long, some of these hairs are already complete, but others continue 

 to develop, so that on a leaf three times as long all stages may still be 

 found. A young epidermal cell swells outward beyond its neighbors and 

 is cut in half by a transverse wall ; the outer cell continues to swell, and is 

 soon divided by a wall which is strongly concave outward, making a termi- 

 nal spherical cell (Fig. 32). A second concave wall may follow the first in 

 the apical cell. Now, according to Schilling (1894), who investigated ./V. 

 alba and other aquatics, the cellulose walls of this cell become mucilagin- 

 ized, the cuticle is ruptured, and the outer cell breaks down into a mass of 



