IN VEGETABLE BIOLOGY. 601 



tinuous rifts in the wall may sometimes be observed in thin 

 sections — one of these is shown at tig. 14 ; whether such result 

 from splitting of the tissue during section or are congenital or 

 consequent on absorption I do not know. Figure 15 shows a 

 small portion of the wall penetrated by holes of varying calibre. 

 Investing the seed are several layers of thin-walled cells forming 

 a somewhat spongy testa, immediately beneath which the outer- 

 most endosperm-cells are smaller and thinner than the under- 

 lying ones ; nevertheless continuity can be traced across their 

 walls. In all essentials D. melanoxylon resembles the species just 

 described. 



One point of great interest remains. Hitherto, as far as I am 

 aware, no one has succeeded in tracing the connection between 

 the intramural threads and the cell-protoplasm. Tangl* says 

 that employment of ordinary reagents causes total plasmolysis, 

 and subsequent authors seem, by their silence, to acquiesce in 

 the truth of the statement. So much attention has recently 

 been directed to this matter f that I have made a special study 

 of the plasmolysis of the cells of Strychnos, and find that Tangl 

 is incorrect. It is true that in the case of many cells the ex- 

 hibition of strong reagents is followed by a plasmolysis which is 

 total, but this is by no means always the case. The best way of 

 observing plasmolytic threads is to place sections in a drop of 

 solution of iodine in alcohol on a slide, a minute or so afterwards 

 placing a cover- slip upon them and examining either in this state, 

 or after addition of a small quantity of water. Figures 16, 17, 

 and 18 were drawn from sections treated in the latter way, the 

 first of them from 8. Nux-vomica, the second from S. Ignatia, the 

 last from S. potatorum. Although easily overlooked, the threads 

 can readily be made out with care ; in many instances they may 

 be seen to run into the intramural threads. Figure 18 repre- 

 sents an unusual condition; although plasmolysis has been violent, 

 yet some of the threads extend unbroken across nearly half of 

 the cell's lumen. It will be observed that in the cases of partial 

 plasmolysis here figured the protoplasm remains for a part of its 

 course in close apposition to the wall of the cell, and it would 

 seem at least possible that the persistence of the threads is due 

 to this fact, the close relation between wall and plasma acting as 



* L. c. 



t H. de Tries, " Untersuch. lib. d. mechanise*. Ursachen d. Zellstreekung;" 

 Gardiner, Phil. Trans, memoir, and Proc. Roy. Soc. 1882; Bower, Quart. 

 Journ.Micr. Sc. 1883. 



