The Nardoo Plant 143 



from the anterior half ring, which, however, retained, as 

 vestiges of their points of affixion, a corresponding number 

 of cicatrices. The sori, by this breaking off, were simul- 

 taneously opened. After the lapse of about an hour the 

 whole cord had protruded as a closed ring. In this state 

 the ring remained for three days in water unaltered, main- 

 taining thus a normal form. The perhaps more frequent 

 appearance of a torn and then vermicular cord is easily 

 accounted for by its vulnerability at the points of construc- 

 tion. A portion of the ring may perhaps also be destroyed 

 when the water only penetrates by gradual decomposition of 

 the valves, whilst the artificially promoted ingress of the 

 water is favourable for the development of the process in its 

 perfection. 



Observing the volumen of the swollen sori-cord in com- 

 parison to its original size in the dry fruit, it will be patent 

 that an at least 200 fold enlargement took place. It was 

 therefore of interest to analyse this mechanism. 



The fruit-valves, as already in particular sketched by 

 Mettenius, consist of layers of cellular tissue of very different 

 structure. The exterior one consists in a ripe state of the 

 remnants of a cutaneous layer formed by roundish thick- 

 walled cells, which contain a granular substance. Beneath 

 this follow two layers of prismatic cells, which stand per- 

 pendicular to the surface-stratum and possess very thickened 

 and lignescent walls, so that their lumina, which contain 

 yet a little granular matter, are only observable at both 

 extremities, but are obliterated in the middle. The exterior 

 one of these layers is formed again of two substrata of small 

 cells, whilst the interior one is formed of an almost single 

 stratum of larger cells. The former is of pale yellow, the 

 latter of brown yellow colour. These strata of cells seem to 

 prevent the ingress of water into the unvulnerated fruit. 

 On these strata again is deposited at the inner side a tissue 

 of parenchymatic thick-walled cells, of which the first row, 

 in consequence of their variable size and form, leaves many 

 vacant places between them. The following ones become 

 gradually more thinly walled and more ample. On the 

 inner limits of this stratum cease the anastomosing vascular 

 bundles of the fruit-valves, as already pointed out by 

 Mettenius. This inner parenchym, in fruits not perfectly 

 ripe, is densely filled with starch, of which even the similar 

 cells of the surface-layer participate. 



In fruits perfectly ripe the starch is no longer extant, 



