SO MINNESOTA BOTANICAL STUDIES. 



in the peripheral part. The former are irregularly quad- 

 rangular in outline in both vertical and cross-section. Their 

 diameters in cross-section are equal (Fig* 6). They are 

 more regularly disposed in rows than the cells in the mar- 

 ginal parts. The cells of the holdfast have walls of good 

 thickness, composed in part of the usual mucilaginous sub- 

 stance. 



Taking the central part of the holdfast (Fig. 6), the lowest 

 cells are dead and empty and partially disintegrated into mu- 

 cilage. Clefts arise among the still living cells from this disin- 

 tegration. Gaps are also caused in this way in the body of the 

 holdfast. Here and there individual cells at the bottom have a 

 disc-like lower surface as if they had a holdfast of their own. 

 The decay of the cells near the central part of the holdfast ex- 

 tends not more than one or two cells deep. The next few rows 

 of cells are slightly flattened parallel with the base of the hold- 

 fast. The succeeding rows of cells become gradually elongated 

 in a vertical direction till, at the tenth or eleventh row, a rapid dif- 

 ferentiation begins with cells elongated in the vertical direction 

 to two or four times their horizontal diameter. Evidently the 

 stipe begins at this zone. 



The peripheral portion (Fig. 7), as was stated above, is com- 

 posed of rows of cells descending obliquely from the axis of the 

 stipe. A vertical section through this part shows that these rows 

 of cells branch dichotomously in the horizontal plane as they go 

 down and outward. The cells decrease in length as one follows 

 the dividing branches, till a zone is reached in what corresponds 

 to the cortex of the stipe. Here the ultimate dichotomous divi- 

 sions of the main strands form a meristem of small cells. The 

 cells of this meristem run in straight rows perpendicularly to 

 the surface of the holdfast. They are in active division. This 

 meristematic layer enables the holdfast to grow in thickness and 

 also to form the rhizoid-flaps on its edge. It is about eight to 

 ten cells deep. In the specimen examined the basal cells, three 

 to five deep in this part of the holdfast, showed advanced disin- 

 tegration. The epidermal layer near the lower edge also was 

 in a similar condition. But the cells of this layer are more re- 

 sisting and persist alive after the two rows beneath are already 

 dead. Probably the mucilage derived from the disintegration 

 of these basal cells is useful in attaching the plant to the sub- 

 stratum. 



