43 



genous matter must be present in them. We might, therefore, 

 for the present, consider this pecuhar substance as a mixture of 

 mucilage and protoplasm, impregnated with tannin. 



As to the question whether the mucilage is produced hi the 

 interior of the secreting cells, or in the outer layer of their walls, 

 we have to refer to the statements of DeBary in his Comparative 

 Anatomy (p. 93, Engl. Ed.) where, in treating of glands, inchid- 

 ing glandular hairs, he says : " The anatomical peculiarity of the 

 glandular parts of the epidermis consists in the appearance in the 

 cell wall of that body, which is termed the secretion of the gland, 

 as a part sharply defined from the cellular layers. The wall 

 grows in thickness at the glandular spot by intercalation of a lay- 

 er between its outer and inner side. The intercalated mass dif- 

 fers in material from the cellulose and cuticular wall and is termed 



a secretion." 



Another passage on the same page reads as follows: — *' More 

 careful investigations are necessary to answer the question as to 

 the appearance and origin of the secretion. But in any case it 

 is incorrect to imagine a "perspiration " in the sense of a passage 

 of large optically determinable masses formed in the interior of 

 the glandular cells through the membrane." And on page gg 

 the author mentions one single exceptional case known to him : 

 '* The bases of the young leaves of OsDiimda are covered with a 

 rich amorphous mucilage. This originates from long septate 

 hairs with large bead-Hke cells, each of which in the stages of 

 development observed, is completly filled with a mass of muci- 

 lage. The origin of the latter remains to be investigated."* 



I venture to suggest that the hairs of Brasenia may form 

 another exception. In the first place we do observe "large opti- 

 cally determinable masses " in the interior of the hair which are 

 similar to the secreted masses, diminishing as the secretion — in 

 the sac — increases, and which finally disappear Moreover we 

 have seen that by destroying the restrictive power of the enclosing 



* Since this paper was read, No. x, Vol. I, of the Annals of Botany, has reached 

 us, in which W. Gardiner and Tokutaro Ito discuss the structure of the secretory cells 

 of Osmunda regaiis^ L., and Bhchnum ocddeniaie, L. According to their investiga- 

 tions, «« the cell-contents usually escape by means of a small localized rupture of the 

 wall" (p. 40), or "the whole wall through disorganization breaks down on all sides 

 and the swollen drops quietly escape " (p. 41). 



