82 C. M. CHILD. 



material. The cells of the Castagnea hair attain their full size 

 very near the basal end, so that in that portion of the hair visible 

 in the living plant the cells are all of practically the same di- 

 ameter and length and it is difficult or impossible to determine 

 the position of the growing point from a gradation in cell size. 

 This case shows how readily incorrect conclusions may be drawn 

 where the material is not in perfect condition. These hairs, 

 like those of Fucns, are very sensitive and the gradient undergoes 

 reversal very readily, but there can be no doubt that it is primar- 

 ily acropetal, not basipetal, as stated in my earlier paper. The 

 hairs are so fragile that they are readily broken or injured me- 

 chanically and at least many of the irregularities are due to such 

 injuries. 



BRANCHED MULTICELLULAR HAIRS. 



Among hairs of this kind those of Chondria have served as the 

 chief material, but the hairs of Polysiphonia and Griffithsia have 

 also been examined. The hair of Chondria develops by repeated 

 apical dichotomy (Fig. 4), and since growth is primarily apical 

 a basipetal death gradient is to be expected. In 1915 hairs of 

 this species were examined by first staining deeply with neutral 

 red and then killing in KCN and the gradient from cell to cell 

 was found to be distinctly basipetal while within single cells an 

 acropetal gradient was almost invariably observed (Child' i6a, 

 pp. 105-106). Since these observations were made during the 

 last days of my stay at Woods Hole in 1915 there was no op- 

 portunity to repeat the observations with fresh material until 

 the following year. As regards the acropetal cell gradient the 

 results of 1916 are as follows: When the hairs are killed in a high 

 concentration of neutral red alone (e. g., before precipitation in 

 sea-water) both hair gradient and intracellular gradient are very 

 uniformly basipetal. Similarly when the hairs are killed in some 

 other agent (KCN m/iooo -- w/ioo, HgCl 2 OT/SOOOO) to which a 

 few drops of neutral red have been added to stain the hairs 

 slightly and so make the death-changes clearly visible, both hair- 

 gradient and cell-gradient are basipetal. On the other hand, 

 when the hairs are first deeply stained with neutral red and then 

 killed in KCN or some other agent, the method used in 1915, the 



