body that has considerable rigidity will push up 

 through the substance of the nucleus when the cell 

 is flattened, even though it is formed in the cyto- 

 some. With the nuclear substance displaced, 

 the vacuole appears to be within the nucleus. An 

 additional basis for suggesting that these granules 

 are artifacts is the fact that they tend to occur 

 most abundantly in cells that are most distorted 

 {A, 13, 15, and B, 23). 



The second type of vacuole is a highly refrac- 

 tory one. In the drawings each such vacuole is 

 indicated by a blue ring (figs. 224 and 225). 

 The number within a cell varies widely from none 

 to as many as are seen in figure 225, 23, but three 

 seems to be an average number. Only a few are 

 present in the cells of figure 224 A, and by 69 

 hours incubation (fig. 225) they have practically 

 disappeared; but these vacuoles are typical for 

 the cells of figure 224 B. The nature of the sub- 

 stance they contain, their function, and their 

 specificity are unknown. Microchemical stud- 

 ies, or the tracing of their course of appearance 

 and disappearance with the phase microscope in 

 living cells, might be revealing. 



The third type of vacuole is small and round, 

 is not refractile, has a tinge of color, and tends 

 to form clusters. At the stage represented by fig- 

 ure 224 A they are best shown in cell i, but in B 

 numerous cells show these bodies (4, 5, 8, and 

 12-14). When they first appear, they are scat- 

 tered luit as they increase in number they form a 

 group of spheres. By 65 hours incubation the 

 individual bodies of these clusters have coalesced 

 and are responsible for the light area in the cyto- 

 plasm adjacent to the nucleus. Dantschakoff 

 (1908b and 1909a) observed in her sectioned 

 material the presence of an orange-stained area 

 beside the nucleus of the early erythrocyte found 

 in the yolk sac. Similar orange-staining spheres 

 were illustrated and described by Maximow 

 (1909) in his study of early formation of blood 

 cells in the mammalian embryo. The whole 

 process of differentiation of primary erythro- 

 blasts from mesenchyme has been described by 

 Murray (1932). He used tissue culture prep- 

 arations of the primitive streak of the chick 

 embryo. It was obsei'ved that the immature 

 primitive erythrocytes were rather small cells 

 with pseudopodia and with nuclei which con- 

 tained one or two nucleoli. The cytoplasm 

 was strongly basophilic and contained one 

 or more eosinophilic masses adjacent to the nu- 



cleus. From such cells as these arose the later 

 stages in the differentiation process. The 

 orange-stained area may well have consisted of 

 vacuoles of this type. If additional study should 

 prove that the vacuoles seen beside the nucleus 

 both in sectioned material and in dried smears 

 are the same, this would offer an excellent cell 

 organelle that could be used as a common basis 

 for identifying the same cell under two different 

 technics. Dantschakoff (1908b) observed that 

 the light-staining spheres contained centrosomes. 



None of the three types of vacuoles or spheres 

 observed in the cytosome of the primary genera- 

 tion of erythroblasts were observed in the cyto- 

 some of later generations. Tliis difference be- 

 tween the first and subsequent generations is in 

 agreement with observations ]jy Dantschakoff 

 (1908b) on sectioned material. 



Since none of these types of spheres have been 

 observed in later generations of erythroblasts in 

 smears, it might be possible to establish whether 

 later generations are similar to the cells (megalo- 

 blasts) * of the primary generation and to the 

 megaloblasts of anemia — a question that has been 

 ably discussed by Jones (1943). This author 

 has pointed out that there are differences in nu- 

 clear cytology among primary and normal de- 

 finitive megaloblasts and megaloblasts of anemia. 



Studies by the smear method show that such 

 differences certainly do exist in the primary and 

 normal definitive erythroblasts of the chicken. 

 As an example of this difference, the nucleus of 

 the primary erythroblast generally contains two 

 nucleoli. No other blood cell included in this 

 study has been obseiTcd to have this numljer of 

 nucleoli. The chromatin reticular pattern is 

 somewhat coarser in primary erythroblasts than 

 in the cells of a corresponding degree of differen- 

 tiation in later embryonic generations. The nu- 

 cleoli seem to be more prominent in the first gen- 

 eration than in later ones but this is probably due 

 entirely to physical causes associated with the 

 coarseness of the screen which permits a better 

 view of the interior of the nucleus than can be 

 obtained when looking through the fine screen 

 that is characteristic of later generations of 

 erythroblasts. 



Differences in size among primary erythro- 

 blasts have not been specifically mentioned thus 

 far. The differences may be seen by glancing 



* "Erythroblasts" and "megaloblasts" are used as synony- 

 mous terms. 



116 



