172 CARROLL— ON DEVELOPMENT OF 



L. These four roots form a whorl, one pair falling in the median plane 

 of the cotyledons and one at right angles. In cross section these appear 

 as the arms of a cross. On germination these appear soon after the prim- 

 ary roots as little hooks and soon press into the ground. Heinricher's 

 figures show the condition in I. glanduligera and balsamina. Heinricher 

 states on Klebs' authority that in Impatiens noli-tangere at germination 

 the four side roots surpass the main root. Similar behavior of side 

 roots is seen in fulva. The side roots were not found, however, already 

 developed in the seed. 



The reserve food stored in the cells of the cotyledons has been inves- 

 tigated in several species of Impatiens, particularly in I. balsamina by 

 Heinricher. In I. balsamina, capensis and others reserve food is stored 

 in the form of cell- wall thickenings. These thickenings Heinricher (21) 

 determined were not cellulose but of amyloid nature. The thickenings 

 disappear on germination, and starch grains appear in the cells. Sub- 

 sequently the cells become green and function as assimilative cells. 

 The cells of the cotyledons of other species of Impatiens contain small 

 protein grains and a large quantity of fatty oil (21, p. 167) with no 

 carbohydrate. In /. fulva, the cells of the cotyledons are thin- walled, 

 and densely packed with large grains of reserve food. This material 

 proved not to be starch, but its nature was not investigated. It may 

 be the fatty oil characteristic of other species of the genus. 



The history of the endosperm presents some interesting points. 

 The primary endosperm nucleus divides quickly and divisions follow 

 in rapid succession. The nuclei move to the periphery of the embryo 

 sac. Some push around above the embryo and even more into the 

 micropyle (Fig. 27). The nuclei in the micropylar end of the sac become 

 cut off by cell walls (Figs. 28 and 29). These cells soon begin to show 

 signs of degeneration, and in later stages appear as small crushed masses 

 in the micropyle and in the micropylar end of the embryo sac (Fig. 30). 

 The nuclei on each side of the embryo a little later are cut oflf as cells. 

 Thus when the cotyledons begin to appear three rather irregular layers 

 of endosperm cells are to be seen at the sides of the embryo (Fig. 32). 

 As the embryo increases in size the peripheral layer of endosperm nuclei 

 elsewhere in the sac become cut off into cells. At the micropylar end 

 of the sac the cells become much elongated, forming a row around the 

 hypocotylar end of the embryo as if functioning as a suspensor to force 

 the embryo down toward the center of the sac. There seems to be 

 quite a tendency for cells to force their way to the micropylar end of 

 the sac or even into the micropyle. The synergids when cut off become 



