874 
ocytes and lymph glands has well expressed 
this conviction in the conclusion that “we 
must await some new method of attacking the 
problem.” The recent development in the 
technique of tissue culture, with the striking 
results obtained by Harrison, Burrows, Lewis, 
Loeb and others, leads one to expect that sim- 
ilar methods may yield important data con- 
cerning the complicated problem of the life 
history and genetic relationship of the various 
types of blood cells. With this hope a study 
of the behavior of blood corpuscles in plasma 
cultures was begun somewhat over a year ago. 
Some of the results regarding the red blood 
cells appear of sufficient importance to justify 
a preliminary statement. 
The present observations relate to the origin 
of the erythrocytes or non-nucleated red blood 
corpuscles and bear directly upon the long- 
standing question of the elimination of the 
nucleus. The description will be confined to 
certain results reached in a study of the blood 
of the pig embryo. Stating the method 
briefly, the erythroblasts of the embryonic 
blood were taken at various stages of differ- 
entiation and observed in hanging-drop cul- 
tures, employing the technique devised by 
Harrison and Burrows, with such modifieca- 
tions as appeared essential for the present 
purpose. Biased by the prevalent view that 
the nucleus of the erythroblast subsequently 
disappears, either by disintegration or by ex- 
trusion, it was hoped that it might be possible 
to observe the process directly in the living 
cell. It can be readily appreciated then with 
what surprise it was discovered that erythro- 
cytes may arise in a manner fundamentally 
different from either of the above alternatives. 
During the first half day (circumstances 
necessitated beginning the cultures in the 
afternoon) the erythroblasts manifest a va- 
riety of cytoplasmic activities, not inade- 
quately described as amceboid in character. 
Numerous pseudopodia-like projections of va- 
rious types are present. The greater number 
are in the form of tapering points or slender, 
elongated processes, varying from one to sey- 
eral in a given erythroblast. In a smaller 
number of cases the cytoplasm extends out in 
SCIENCE 
[N.S. Vou. XXXV. No. 909 
blunt, rounded, somewhat bud-like projections. 
These processes manifest constant quivering 
and oscillating movements. Other erythro- 
blasts, spherical in form, are more or less 
quiescent. During the second day, of the 
above types of processes, the blunt bud-like 
form has become much more predominant. 
In other respects the cells appear as on the 
preceding day. These bud-like processes may 
involve one third or even more than one half 
of the cytoplasm of the cells. <A striking fea- 
ture is that not only are the buds completely 
filled with hemoglobin, but in many cases 
practically the entire hemoglobin content of 
the erythroblast has become segregated in the 
bud, leaving a more or less clear and hemo- 
globin-free cytoplasmic area surrounding the 
nucleus. If observations are begun upon the 
cell at this stage, the bud will be found in a 
state of ameboid activity manifested in 
changes in the contour varying from slight 
modifications to elongated projections. If the 
cell is at the height of its activity an aston- 
ishing thing may now occur: a constriction 
becomes evident and within a few minutes or 
even seconds this constriction is completed 
and the hemoglobin-containing bud is liber- 
ated from the cell. Typically this liberated 
bud corresponds in size and appearance with 
the adjacent erythrocytes; it may assume a 
bi-coneave dise shape, and in one instance it 
was possible to continue the observations suffi- 
ciently to follow the final transformation of 
the liberated bud beyond this dise shape to a 
typical cup-shaped form. In each ease con- 
trol specimens of the blood originally em- 
ployed for the cultures were fixed in formalin 
vapor and stained. Budding erythroblasts 
were also successfully fixed and stained and 
were found to correspond both in their cyto- 
plasmie and nuclear characteristics with the 
control specimen. LErythroblasts were main- 
tained in normal condition in several experi- 
ments for three or more days. Budding 
activities were continuous throughout this 
time, and under favorable conditions ten or 
more cells in various stages of active budding 
could be counted in a given field of the micro- 
scope. 
