A COMPARATIVE STUDY 
OF 
THE SO-CALLED POLYCHROMATOPHILOUS 
DEGENERATION OF RED BLOOD 
CORPUSCLES 
ERNEST LINWOOD WALKER 
(From the Laboratory of Comparative Pathology of the Harvard 
Medical School) 
Reprinted from 
The Journal of the Boston Society of Medical Sciences 
November, 1899 
BOSTON 
MASSACHUSETTS 
U.S.A. A COMPARATIVE STUDY 
OF 
THE SO-CALLED POLYCHROMATOPHILOUS DEGEN- 
ERATION OF RED BLOOD CORPUSCLES. 
Ernest Linwood Walker. 
( From the Laboratory of Comparative Pathology of the Harvard Medical School.) 
It is well known that during certain pathological conditions 
in which anaemia is a primary or a secondary factor there 
appear in the circulating blood red corpuscles showing 
characteristic morphological and microchemical differences 
from the normal circulating corpuscles. These “ anaemic 
forms” include corpuscles abnormal in size (microcytes and 
macrocytes) and in shape (poikilocytes), corpuscles deficient 
in haemoglobin, corpuscles having abnormal staining affinities 
(polychromatophilocytes), nucleated corpuscles (normo- 
blasts, microblasts, and megaloblasts), and corpuscles show- 
ing two or more of these changes simultaneously. All of 
these, with the exception of the normoblasts, are generally 
believed to be corpuscles that have undergone changes of a 
degenerative character, either direct necrobiotic changes 
(polychromatophilia, poikilocytosis, etc.), or else changes 
resulting from degenerative processes in the haematopoietic 
organ (the megaloblast degeneration of the marrow of 
Ehrlich), whereby there is a return to a so-called foetal type 
of haemocytogenesis. The normoblasts, being generally 
recognized as embryonic red corpuscles, are looked upon as 
representing regenerative changes in the blood. Certain 
investigators have, however, seen reasons for believing that 
other of these abnormal corpuscles might belong, with the 
normoblasts, to the regenerative processes following anaemia 
rather than to the degenerative changes of anaemia. Their 2 
diagnostic value in addition to their purely biological signifi- 
cance makes a knowledge of the origin and nature of these 
changes of considerable importance. In this paper there 
will be considered only those corpuscles that show the so- 
called polychromatophilous staining, one of the most con- 
stant of these changes associated with anaemia. 
The discoplasm of the living red corpuscle is said to be 
achromatophilous ; that is, it has no affinity for any of the 
ordinary histological staining reagents. When, however, the 
corpuscles are killed and fixed they become chromatophilous ; 
but the stroma of the corpuscles normal to the human blood 
have an affinity only for those dyes, the staining properties 
of which reside in the acid part of the staining compound, 
and are, therefore, spoken of as acidophilous (oxyphilous). 
In anaemic blood, however, the discoplasm of certain of the 
corpuscles shows a selective affinity for dyes, the staining 
properties of which reside in the base of the compound, and 
may consequently be called basophilons. These are the 
polychromatophilous corpuscles of Gabritschewsky, who 
thus designated them because when stained with the tricolor 
stain of Ehrlich, or when double stained with eosin and 
haematoxylin or methylene blue, they sometimes stain with a 
mixture of two or more colors, either as a diffuse mixture or 
irregularly, some parts of the corpuscle taking color differ- 
ently from others. 
Such diffuse acid stains as eosin do stain all red corpuscles to a certain 
degree, although the basophilons more feebly than the acidophilous 
variety. If methylene blue or haematoxylin, used as a contrast stain, be 
applied for only an instant the basophilons corpuscles will present a mixed 
or polychromatophilous staining due to the fact that the basic stain has 
not had sufficient time to completely substitute itself for the diffuse acid 
stain. If allowed to stain several minutes it will entirely supplant the eosin. 
Methylene blue or haematoxylin used alone stain only the basophilons 
corpuscles, the normal corpuscles remaining entirely unaffected. 1 The 
1 In preparations of dried blood kept several months the stroma of all the red cor- 
puscles may take up alkaline methylene blue faintly and diffusely. This diffuse ab- 
sorption is, however, distinguishable from the chemical affinity of the basophilons cor- 
puscles in the same preparation. If the preparation be lightly decolorized with 
absolute alcohol or very dilute acetic acid (o.i $) the acidophilous corpuscles are 
decolorized, leaving the basophilous corpuscles still stained. other definite basic stains, like methyl green, vesuvin, etc., have a weaker 
affinity for this variety of cytoplasm. Definite basic stains already 
mentioned may be used directly or progressively, but certain diffuse basic 
stains like safranin and fuchsin have to be employed indirectly or regres- 
sively for the differentiation of the basophilous corpuscles, and in general 
give less satisfactory results. The tricolor stain (orange G., acid fuchsin, 
and methyl green) is unsuitable Ehrlich to the contrary for studying 
these chemical changes in the cytoplasm of red blood corpuscles. Methyl 
green, the only basic stain in the mixture, having but a feeble affinity for 
the basophilo-plasm, is submerged by the diffuse acid stains. The more 
intensely basophilous corpuscles are either more feebly stained than the 
acidophilous corpuscles or else show a polychromatophilous staining. 
This selective affinity for basic stains is more commonly 
uniform throughout the stroma of the corpuscle, the corpus- 
cle staining a homogeneous diffuse color varying in intensity 
from the faintest perceptible tint to a moderately deep color; 
but it sometimes appears limited to certain granules or frag- 
ments in the acidophilous stroma, the corpuscle appearing 
punctate with small stained dots. These differences in stain- 
ing affinity are microchemical reactions that indicate the dif- 
ferent chemical composition of the discoplasm of the two 
groups of corpuscles, the nature of which will be fully con- 
sidered in another section. Their staining affinities furnish 
the easiest means of identifying these corpuscles. 
Basophilous corpuscles show slight morphological varia- 
tions from the acidophilous variety. They are usually larger 
—in severe anaemias often much larger than the normal 
corpuscles. The biconcavity of the disc is generally less 
marked. And in the much enlarged forms the corpuscle 
usually has a delicate, attenuated appearance, sometimes as- 
sociated with irregularity of shape (poikilocytosis). 
A review of the literature on the subject shows a diversity 
of opinion concerning the interpretation of these corpuscles. 
Smith, Gabritschewsky, and Askanazy, basing their judg- 
ment upon comparative and experimental evidence, look 
upon them as normal phases in the life history of the red 
corpuscle, and specifically as immature stages of the devel- 
oping corpuscle, stages which in man are normally passed in 
the blood-forming tissue of the marrow, and which have 4 
escaped prematurely into the circulation during the active 
regeneration of the blood following the anaemia. On the 
other hand, Ehrlich, Maragliano, and their followers, basing 
their opinion chiefly upon pathological evidence, have been 
led to believe, since basophilous corpuscles are abnormal to 
the human circulation and only appear associated with some 
abnormal, usually pathological condition of the individual, 
that they represent abnormal phases in the life of the red 
corpuscle, and, from their association with disease, that they 
are degenerative changes due to the disease. This is the 
opinion that seems to have become the more widely accepted 
and is now generally quoted in text-books as the explanation 
of them, with the occasional qualification that other and op- 
posing views are held by some. 
Ehrlich (5) in 1885 was first to describe corpuscles staining with the 
basic stains haematoxylin and methylene blue, which can be thoroughly 
identified with the corpuscles that have just been described. He calls them 
anaemic degeneration forms, and looks upon the change as a sort of coagu- 
lation-necrosis in the stroma of the corpuscle. In the same year Favre 
and Celli (8) described them in malarial blood. 
Several observers, including Erb (4) , Lowit (13), Foa and Mondino (9), 
Celli and Guarnieri (3), and Howell (12), some of them prior to Ehrlich, 
have described granules and parts of corpuscles that stain abnormally, or 
which were brought out by various treatments of the blood, both from well 
and from diseased animals and from animals upon which venesection had 
been practised. It is questionable, however, how far any of these are 
identical with the corpuscles under consideration. 
In the history of a case of purpura presented (by Dr. Prentiss) before 
the Association of American Physicians in 1890 Theobald Smith was first 
to diagnose as immature certain red corpuscles staining diffusely with 
basic stains. 
In the same year Gabritschewsky (10) described such corpuscles in 
cases of anaemia and leukaemia, calling them polychromatophilous cor- 
puscles, for reasons just stated. He also looks upon them as immature 
corpuscles, on the grounds that the immature nucleated corpuscles of 
reptiles and birds, and the nucleated corpuscles found in the human cir- 
culating blood in anaemia and leukaemia, have a cytoplasm that stains with 
the basic stain methylene blue. That this is not due to degenerative 
changes he thinks proven by the occurrence of different phases of mitosis 
in their nuclei. 
In 1891 Smith (17) presented before the Association of American 
Physicians a paper giving the results of observations upon the anaemic blood 5 
of cattle made during his well-known investigations of the Texas cattle 
fever. Briefly, the important facts of this paper are as follows : 
I. The normal circulating blood of the ox contains only acidophilous 
normocytes. 
2. When through the anaemia of this disease, which is caused by an 
extensive destruction of red corpuscles by a protozoan endoglobular para- 
site, as is malaria in man, the number of red corpuscles in the circulating 
blood is reduced from 5.5-6.5 millions per cmm., the normal number in 
the healthy animal, to about 3 million, there appear in the circulation a 
certain number of enlarged corpuscles (macrocytes). When the number 
has fallen slightly below 3 million, corpuscles with granules staining with 
basic stains appear. If the destruction still continues, at 1.5-2 million 
diffusely staining basophilous corpuscles, and at 1-2 million corpuscles per 
cmm., nucleated corpuscles, are to be found in the circulation. During re- 
covery the disappearance of these abnormal forms from the circulation 
takes place in inverse order. 
3. Smith was able to produce these same phenomena by artificially low- 
ering the number of red corpuscles in the blood through venesection on 
healthy animals. 
4. He considers these abnormal forms to be primarily embryonic cor- 
puscles that have been drawn into the circulation prematurely to supply 
the increased demand created by the destruction of corpuscles by the 
parasite. 
Askanazy (1) confirms Gabritschewsky that the affinity for basic stains 
is characteristic of the cytoplasm of nucleated red corpuscles found in the 
circulating blood, and he finds that the greater number of the embryonic 
nucleated corpuscles of the human bone marrow, as shown from material 
obtained from the resection of a rib, show the same selective affinity. The 
corpuscles of the embryonal human lives, he says, show this same peculi- 
arity. 
Since his first communication on the subject Ehrlich has made several 
contributions to the controversy, always maintaining his belief in the de- 
generative character of these changes. In his latest work on the blood, in 
collaboration with Lazarus (7), he summarizes his evidence in support of 
this theory: 
“ 1. The appearance of those erythrocytes that show the highest degree 
of polychromatophilia. Through the fragmentation of their edges they 
appear to every eye trained in the estimation of morphological relations 
as if in the act of dissolution, as the most pronounced degeneration 
forms. 
“2. The fact that one can cause them to appear in considerable numbers 
in the blood in animal experiments, for example through inanition, there- 
fore in just the condition when it can be least the question of new build- 
ing of red corpuscles. 
“ 3- The clinical experience that after acute blood losses in man, even 
inside of the first twenty-four hours, one can find this staining anomaly 
in numerous corpuscles, while, according to our very extensive investiga- 6 
tions on this point, embracing many hundred cases and conducted with the 
greatest care, one finds at this time no nucleated red corpuscles. 
“4. Frequently nucleated red corpuscles, especially megaloblasts, show 
the polychromatophilous degeneration. This is a so easily demonstrated 
fact that even an untrained observer cannot easily overlook it, and, as is 
well known, it was Ehrlich who first turned his attention to these relations. 
But important in their interpretation is the circumstance that the type of 
the normal regeneration, the normoblasts, are usually free from polychro- 
matophilous degeneration, likewise are the nucleated red blood corpuscles 
of animals. When Askanazy asserts that the nucleated red blood cor- 
puscles of the bone marrow, which he was able to study in some cases of 
empyema immediately after resection of the rib, all showed polychroma- 
tophilia, this may perhaps be connected with the idiosyncrasies of these 
cases or with the uncertainty of the staining method employed. Eosin- 
methylene blue staining is to be denoted as very untrustworthy, since over- 
staining with the blue is likely to happen. (We expressly advise the use 
of the tri-acid solution or haematoxylin-eosin mixture in the study of anaemic 
degeneration).” 
Ehrlich is supported by Maragliano (14, 15), who studied the changes 
going on in corpuscles of drawn blood under the microscope, where he 
observed degenerative changes, including a modification of the staining 
affinities resembling these changes in the corpuscles in anaemic blood, and 
from analogy looks upon these latter as necrobiotic. 
Maragliano considers these changes in anaemic blood to be due to a 
toxicity of the plasma. Cabot (2) thinks a lessened resistance to the 
ordinary plasma environment on the part of the red corpuscles would 
explain them. Troja (19) looks upon this change in the staining affinity 
of the protoplasm of the red corpuscle as a sort of karyolysis, whereby 
a diffusion of the chromatine of the nucleus into the cytoplasm has taken 
place. 
Taking the clue from the investigations of Smith, Gabrit- 
schewsky, and Askanazy, my own studies have been made on 
the blood and bone marrow of vertebrates with the purpose 
of finding out more of the origin and nature of this change 
and thereby gaining a better insight into its significance in 
anaemic blood. 
In studying the staining reactions of red corpuscles, Ehrlich dry prepara- 
tions of the blood and of the marrow were made according to the method 
described by Smith (17'), and subsequently by Mannaberg, Heiman, and 
Zettnow. The edge of a square cover-glass held in clamp forceps is 
touched to the drop of blood and then drawn over the surface of a second 
cover-glass at an angle of 30°, leaving a thin layer of blood that dries 
almost instantly, fixing the corpuscles in good condition. The ease and 7 
rapidity with which the preparations can be made, the ability to vary the 
thickness of the preparation to a certain extent independent of the amount 
of blood taken, and the fact that any slight dirt or grease on the cover- 
glass is not necessarily fatal to the preparation, make this method prefer- 
able to the usual method of allowing the blood to flow between the two 
cover-glasses by capillarity and then drawing them apart. 
Cover-glass preparations thus made and dried in the air at ordinary 
temperature were fixed half an hour in a mixture of absolute alcohol and 
ether in equal parts by volume. For the differentiation of basophilous 
cytoplasm, staining with Ldfflef’s solution of alkaline methylene blue, either 
alone or as a contrast stain to eosin, gives most satisfactory results. A 
supplementary use was also made of all of the other appropriate histologi- 
cal stains. 
In order to compare the proportion of basophilous or other abnormal 
corpuscles in the blood or marrow of different species and individuals, 
counts were made in a microscope field of known area with a Zeiss oil 
immersion lense. Percentages were determined by differential counts, the 
average being taken of the counts of one hundred fields. When the pro- 
portion of normal to abnormal corpuscles was very great, as in the circu- 
lating blood of adult animals, the ratio was obtained approximately by 
actually counting the abnormal forms and calculating the total number of 
all forms in a single field. This method gives results of approximate 
accuracy that are sufficient for our purpose. 
Normal Blood of Vertebrates. 
Although the appearance of basophilous corpuscles in the 
circulating blood of man consequent to pathological condi- 
tions is well known, less seems to be known and still less sig- 
nificance given to what is known concerning their normal 
presence in the blood of many species of vertebrates. 
Ehrlich (5), in 1885, in the same communication in which he first 
describes basophilous corpuscles in anaemic human blood, also says he has 
observed them in the blood of kittens. 
W. Erb (4), in 1865, describes granules in red corpuscles of animals, 
both in normal blood and in the blood after repeated venesection, as well 
as in the blood of man after severe haemorrhage or when suffering with 
chronic disease, which were made apparent by treating fresh blood with 
1 % acetic acid or with picric acid. 
Lowit (13) observed granulations in red corpuscles of the rabbit when 
fresh blood from certain vascular territories was heated with a modified 
Picini’s fluid. 
Howell (12) describes granules in the corpuscles from the blood of 8 
kittens that had been severely bled which stain with methyl green, and 
which are considered by him as remains of the nucleus. 
Smith (17) briefly mentions that he has observed corpuscles staining 
with methylene blue in the blood of guinea-pigs, rabbits, mice, and 
pigeons kept in cages. 
I find basophilous corpuscles of the diffusely staining 
variety to be normal elements of the circulating blood of all 
the lower species of vertebrates. They are invariably present 
in small numbers in the blood of snakes, frogs, pigeons, house 
mice, rats, guinea-pigs, rabbits, cats, and dogs. Their number 
varies in different species, but is fairly constant in healthy 
adult animals of the same species. In Table I. are grouped 
the results of counts of the basophilous corpuscles in the 
blood of the different species, the percentages in each case 
being the average of one hundred fields. This table does not 
represent the number of blood examinations made, but only 
those in which counts were made from animals taken without 
selection. The condition of the animals was determined 
either by post-mortem examination, or, in cases when the 
animal was not killed, by daily observation. Many of them 
were animals bred and kept in confinement, or were domesti- 
cated animals. Some of them, as the snakes, frogs, rats, and 
some of the mice, were freshly caught. The pigeons were in 
confinement but a short time before the blood counts were 
made. 
Table I. 
Species. 
Number. 
Percentage of 
basopliilous 
corpuscles 
in the circu- 
lating blood. 
Remarks. 
Snake 
I 
2-73 
Frog 
1 
I4.4 
Pigeon 
I 
I.84 
Healthy adult animal. 
Pigeon 
2 
3-09 
Healthy adult animal. 
Pigeon 
3 
3.°° 
Healthy adult animal. 
House mouse.. 
i 
1.82 
Healthy adult animal. 9 
Species. 
Number. 
Percentage of 
basophilous 
corpuscles 
in the circu- 
lating blood. 
Remarks. 
House mouse.. 
2 
O.64 
Healthy animal; % grown. 
House mouse.. 
3 
O.4O 
Healthy adult animal. 
House mouse.. 
4 
0.62 
Healthy adult animal. 
House mouse.. 
5 
I.II 
Healthy animal; % grown. 
House mouse.. 
6 
I.44 
Rat 
i 
Numerous 
Healthy adult animal. 
Guinea-pig .... 
i 
0.065 
Healthy animal; weight 477 grams. 
Guinea-pig .... 
2 
0.041 
Healthy animal; weight 335 grams. 
Guinea-pig .... 
3 
0.03 
Healthy animal; weight 271 grams. 
Guinea-pig .... 
4 
0.26 
Tuberculous animal. 
Guinea-pig .... 
5 
0.19 
Tuberculous animal. 
Guinea-pig 
6 
0.094 
Pregnant animal. 
Rabbit 
i 
0.22 
Adult female; weight i,8oo grams. 
Rabbit 
2 
0-54 
Adult male; weight 2,025 grams. 
Rabbit 
3 
1.047 
Tuberculous animal. 
Rabbit 
4 
0.838 
Tuberculous animal. 
Cat 
I 
0-399 
Healthy animal; 6 weeks old; 4 nucleated 
corpuscles in 100 microscopic fields. 
Dog 
i 
0.163 
Healthy adult animal. 
Dog 
2 
0.091 
Healthy adult bull-dog. 
Dog 
3 
0.104 
Healthy adult pug-dog. 
Dog 
4 
O 
4* 
00 
Pregnant greyhound. 
Ox 
i 
O j 
TT 
{ 
Basophilous corpuscles never present in the 
I 
circulating blood of healthy adult animals. 
Man 
— 
0 ) 
Table I. Continued. 
The number of basophilous corpuscles is never large in 
the normal blood of adult animals, amounting to 2-14% of 10 
all red corpuscles in the blood of reptiles and birds, and 
rarely reaching I % in the blood of the mammalia. Basophi- 
lous corpuscles have not been found in the normal blood of 
the sheep, ox, horse, or man. 
The normal presence of basophilous corpuscles in the 
blood of so many species of vertebrates must mean that they 
are normal stages in the life history of the red corpuscles in 
these species, and consequently make us hesitate to believe 
that they represent degenerative changes due to disease when 
occurring in the blood of other species. And their normal 
absence from the circulating blood of some species makes it 
improbable that they are corpuscles showing degenerative 
changes due to old age and natural death of the individual 
corpuscles. 
Opportunity was offered for examining the blood of the 
foetuses of a guinea-pig and a dog. Post-mortem examina- 
tion showed both mother and foetuses to be healthy. Non- 
nucleated basophilous corpuscles were present in both cases 
in numbers vastly in excess of those in the blood of the 
mother. Normoblasts, megaloblasts, and, in the foetuses of 
the dog, corpuscles with basophilous granules, were also pres- 
ent in considerable numbers. Table 11. gives the percentages 
of these different basophilous corpuscles in the foetal blood, 
as obtained by differential counts, and the percentages in the 
blood of the mother, for comparison. In both cases the blood 
of the mother contained basophilous corpuscles above the 
average for the species, owing to pregnancy; yet, while these 
corpuscles in the blood of the mother did not average one in a 
microscopic field, the blood of the foetuses showed them so 
numerous and so prominent because of their large size and 
deep staining that they appeared at first glance to make up half 
or two-thirds of all the red corpuscles. Comparison of per- 
centages shows that the blood of the foetuses of the guinea- 
pig contained ninety times as many basophilous non-nucle- 
ated corpuscles as the blood of the mother, and the blood of 
the foetuses of the dog over seventy-two times as many as the 
blood of the mother. 
Fcetal Blood. 11 
Smith (17, 18) examined the blood of a three months’ 
foetus of a cow that had died suddenly of Texas fever. The 
mother had succumbed so early in the disease that basophi- 
lous corpuscles had not made their appearance in her blood, 
and the blood of the foetus was not infected with parasites, 
nor appeared to be diseased in any way. Corpuscles stain- 
ing diffusely with basic stains were numerous. 
Table 11. 
Mother. 
Fcetus. 
Species. 
Number. 
Percentage of 
non-nucleated 
basophilous 
corpuscles in 
the circulating 
blood. 
Percentage of 
non-nucleated 
basophilous 
corpuscles in 
the circulating 
blood. 
Percentage of 
nucleated cor- 
puscles in the 
circulating 
blood. 
Remarks. 
Guinea-pig . 
6 
O.094 
8.48 
O.IO9 
- 
Of the basophi- 
lous corpuscles in 
the foetal blood 
Dog 
4 
0.478 
34-75 
0.7 ■ 
■ 
4.89 % contained 
granular basophi- 
lous fragments. 
Bone Marrow. 
Even more significant are the staining reactions of the 
corpuscles in the red bone marrow of healthy adult animals. 
The relative number of non-nucleated basophilous corpuscles 
in the bone marrow varies in different individuals and in dif- 
ferent parts of the same individual, and in general depends 
upon the activity of the haemocytogenesis, being greatest in 
young animals and in those parts of the marrow where the 
haematopoietic function is most active. The selection of 
preparations of the marrow for making differential counts 
was therefore guided by the activity of haemocytogenesis as 
disclosed by the presence and number of nucleated cor- 
puscles. Differential counts were made of the normal 
acidophilous, of the non-nucleated basophilous, and of the 
nucleated corpuscles, and the percentage compared with the 12 
circulating blood of the same individuals. The results of 
these counts are given in Table 111. 
Table 111 
Species. 
Number. 
Percentage of 
basophilous cor- 
puscles in the cir- 
culating blood. 
Percentage of 
basophilous cor- 
puscles in the 
bone marrow. 
Percentage of 
nucleated cor- 
puscles in the 
bone marrow. 
Pigeon 
3 
3-oo 
62.05 
62.05 
House mouse 
4 
0.62 
36.06 
1.28 
House mouse 
5 
I.II 
20.21 
I.25 
House mouse 
6 
I.28 
53-47 
IO.56 
Guinea-pig 
5 
O.ip 
19-34 
9-39 
Guinea-pig 
6 
O.09 
26.34 
6.21 
Rabbit 
4 
O.84 
23.16 
24.26 
Cat 
r 
0.40 
34.00 
23-9 
Dog 
4 
O.48 
34.08 
15.06 
Dog 
5 
0.09 
27.87 
7.08 
Dog 
6 
0.10 
30.42 
I5-72 
Ox 
i 
0.00 
12.46 
8.85 
Horse 
i 
0.00 
Present 
Present 
Not only are the large percentages of basophilous cor- 
puscles in the marrow striking, but even more so is a compar- 
ison of these percentages with those obtained from the 
circulating blood of the same individuals. Thus in the 
pigeon the percentage in the marrow is 20 times as great, in 
the house mouse 18 to 58 times as great, in the guinea-pig 
163 times as great, in the rabbit 27 times as great, in the cat 
85 times as great, and in the dog 71 to 306 times as great 
as in the circulating blood of the respective animals. 
The blood of the ox, horse, and man, it will be remem- 
bered, do not normally contain basophilous corpuscles. The 
preparations from the marrow of the ox of the one animal 13 
examined contained 12.5 per cent, non-nucleated basophilous 
corpuscles and nearly 9 per cent, nucleated corpuscles. The 
circulating blood of this animal did not contain them. The 
only opportunity to study the marrow of the horse was from 
an old animal that showed extremely little haemocytogenesis, 
at least in that part from which my preparations were made. 
Of the sixteen preparations made fifteen of them showed 
no embryonic corpuscles. A single preparation showed a 
slight activity of the haematopoietic function, a few nucleated 
and about an equal number of non-nucleated basophilous 
corpuscles being present. The circulating blood of the animal 
was apparently free from them. Material is not at this 
time available for the study of the human bone marrow. 
Askanazy (1), however, says that the greater number of the 
nucleated red corpuscles in the human marrow, as shown by 
material obtained from a resected rib, have a polychromato- 
philous cytoplasm. He found the same to be true of the 
nucleated corpuscles of the embryonal human liver. But he 
makes no mention of the staining affinities of the non- 
nucleated corpuscles of the human marrow. 
The universal presence of basophilous corpuscles as elements 
of the embryonic red corpiLScle tissue in the healthy vertebrate 
marrow furnishes unqualified disproof that they are corpuscles 
showing changes due to pathological conditions. 
The Relation of the Affinity for Basic Stains to the 
Development of Hcemoglobin. 
In the haematopoietically active bone marrow of mammals 
there are, beside the marrow cells proper and the leucocytes, 
four types of cells morphologically distinguishable that will 
be generally accepted as representing different stages in the 
development of the red blood corpuscle. There are (a) 
normal non-nucleated corpuscles like that of the circulating 
blood ; (b) large non-nucleated corpuscles (macrocytes) ; (c) 
what may be designated as mature nucleated corpuscles, i.e., 
corpuscles morphologically like the non-nucleated red cor- 
puscles, except for a medium-sized spherical nucleus in the 14 
otherwise homogeneous cytoplasm ; and (d) young nucleated 
corpuscles in different stages of development. 
The younger erythroblasts are comparatively large cells, 
with a somewhat granular cytoplasm, having a large nucleus 
with a loose reticulated structure occupying the larger portion 
of the cell. Transitional forms, showing every intermediate 
stage between these cells and those designated as the mature 
nucleated corpuscles, are present. In a single microscopic 
preparation, sometimes in a single microscopic field, there 
may be seen nucleated corpuscles showing every possible 
gradation, so that the relation of the different stages is un- 
questionable. The progressive morphological changes ac- 
companying the development of these embryonic corpuscles 
are a decrease in size of the whole cell, an increase in the 
homeogeneity of its cytoplasm, a decrease in size of the 
nucleus both relatively and absolutely, the nucleus coming 
to occupy less of the cell in the mature nucleated corpuscle, 
and an increase in the density of the nuclear substance. 
The mature nucleated corpuscle is acknowledged to give rise 
to the non-nucleated red corpuscle through the loss of its 
nucleus, either by extrusion or by absorption. One often 
sees mature nucleated corpuscles with partially or wholly 
extruded nuclei, whose cytoplasm is morphologically indis- 
tinguishable from the non-nucleated red corpuscle. 
Accompanying these morphological changes attending the 
development of the red corpuscle are parallel chemical 
changes. The nucleus, with its increasing density of struct- 
ure, stains more and more uniformly and intensely with the 
chromatin stains. The cytoplasm of the younger nucleated 
corpuscles is an unmodified protoplasm containing no haemo- 
globin. With the development of this embryonic cell there 
goes on a gradual transformation of the embryonic cytoplasm 
into the haemoglobiniferous discoplasm of the mature red 
corpuscle. This is manifested by the appearance and gradual 
intensification of haemoglobin color in the cytoplasm of the 
more and more mature corpuscles. But while morphological 
maturity is practically attained with the loss of the nucleus, 
complete haemoglobin development is not usually reached 15 
until some time afterwards. An examination of an unstained 
preparation of the haematopoietically active mammalian 
marrow will show the greater number of the large non- 
nucleated corpuscles to be deficient in haemoglobin color. 
These macrocytes show all degrees of haemoglobin develop- 
ment, from the faintest perceptible tint to the deep color of 
the normal circulating corpuscle. In general, it may be said 
that these non-nucleated macrocytes take up the degree of 
haemoglobin color attained by the mature nucleated cor- 
puscles, and extend it through all distinguishable gradations 
to the full color of the mature corpuscle. These macrocytes 
would thus seem to be corpuscles slightly immature, some of 
them not having attained complete haemoglobin development, 
and others, although chemically mature, are still morphologi- 
cally immature. Occasionally corpuscles of normal size show 
incomplete haemoglobin development. And rarely haemo- 
globin development is completed before the loss of the 
nucleus. 
This tendency for the haemoglobin development of the cor- 
puscle to lag behind the morphological development seems 
to be connected with the activity of the haematopoietic 
function. Haemocytogenesis is more extensive and active 
in foetal than in extra-uterine life, in young than in old 
animals, in animals that have suffered an abnormal loss of red 
corpuscles than in normal animals, and, apparently, in the 
lower than in the higher species of vertebrates. In all of 
these cases the chemical development of the corpuscle shows 
a corresponding greater tendency to fall behind the mor- 
phological development; as if the time necessary for the 
complete transformation of the embryonic cytoplasm into the 
hsemoglobiniferous stroma was nearly constant, while any 
acceleration of the haematopoietic function hastened the mor- 
phological development of the corpuscle. 
If, with the microscopic field of an unstained marrow 
preparation in focus, a drop of Loffler’s alkaline methylene 
blue solution be allowed to flow between the cover-glass and 
the slide the cytoplasm of all the corpuscles deficient in 
haemoglobin will instantly take on the diffuse bluish color 16 
characteristic of basophilous corpuscles stained with this 
solution, while the mature, deeply haemoglobin colored cor- 
puscles remain unaffected, standing out in the blue back- 
ground in strong contrast as bright yellow discs. If careful 
notice be taken previous to the staining of the depth of 
haemoglobin color of the different corpuscles in the micro- 
scopic field it will be seen that the intensity with which the 
cytoplasm of the different corpuscles stain with the basic 
stain varies inversely as their haemoglobin development. 
The maximum stain is seen in the most immature nucleated 
corpuscles with unmodified cytoplasm. It decreases with the 
increasing haemoglobin color, and, since complete transmu- 
tation of the cytoplasm of the erythroblast into the haemo- 
globiniferous stroma of the circulating corpuscle is not ordi- 
narily completed until some time after the loss of the nucleus, 
it extends over to these non-nucleated corpuscles. Just as 
every possible gradation in haemoglobin color, from the 
faintest tint to the deep color of the mature corpuscle, is to 
be seen among these non-nucleated corpuscles of the marrow, 
so do they display in inverse order every possible gradation in 
intensity of staining with basic stains. It is consequently 
possible, in an unstained microscopic preparation of the 
marrow, to recognize from the depth of haemoglobin color 
not only those corpuscles that are basophilous, but also the 
degree of their affinity for basic stains; and, conversely, in 
a stained preparation, from their affinity for basic stains are 
we able to tell the degree of haemoglobin development the 
corpuscle has attained. 
There may, then, be distinguished four type phases in the 
development of the mammalian red blood corpuscle : 
a. Young nucleated corpuscles. 
b. Mature nucleated corpuscles.1 
Deficient in haemoglobin; 
basophilous. 
XT , , , 
c. JN on-nucleated macrocytes. 
Basophilous or acidophi- 
lous. 
, tvt , , , , 
d. Non-nucleated normocytes, 
Haemoglobin fully devel- 
oped; acidophilous. 
1 Occasionally acidophilous, haemoglobin development being completed before the 
loss of the nucleus. 17 
The transitional forms connecting these different stages are 
so numerous and the gradations so continuous that the rela- 
tion of the different stages to each other and to the degree 
of development is unmistakable. 
This is even more clearly brought out by a study of the 
marrow of the lower vertebrates, as of the pigeon, where the 
red corpuscle is nucleated throughout its life. The embry- 
onic cells of the marrow in the pigeon, from which the red 
corpuscles are developed, are large cells that are spherical, or 
polyhedral from mutual contact, having a somewhat granu- 
lar cytoplasm and a large spherical nucleus with a reticu- 
lated structure occupying the greater part of the cell. The 
cytoplasm of these cells is entirely free from haemoglobin, 
and is strongly basophilous. With development the cell 
decreases in size and becomes oval, almost lenticular, and its 
cytoplasm shows an increasing homogeneity. The nucleus 
becomes smaller, both relatively and absolutely, takes on an 
oval and finally narrow lenticular outline, and shows an in- 
creasing density in structure and intensity in staining. With 
these morphological changes go the parallel chemical changes 
in the cytoplasm: a gradual transformation of the baso- 
philous protoplasm of the embryonic cell into the acidophi- 
lous, haemoglobiniferous stroma of the mature corpuscle, 
manifested by an increasing depth of haemoglobin color and 
a decreasing intensity in its affinity for basic stains. 
The shape of the nucleus especially enables us to recog- 
nize the finer differences in the degree of development in the 
nearly mature avian corpuscle, the spherical nucleus of the 
embryonic corpuscle becoming progressively more and more 
elongated and pointed as the corpuscle approaches maturity, 
becoming narrow lenticular in the mature corpuscle. Any 
corpuscle with a nucleus showing any departure from the 
narrow lenticular outline towards the oval, be it ever so 
slight, will be accompanied by a corresponding degree of 
chemical difference in its cytoplasm, a lessened degree of 
haemoglobin color, and an increased affinity for basic stains. 
Turning to the circulating blood of the pigeon, we have a 
morphological as well as a chemical (haemoglobin) index to 18 
the maturity of the corpuscle. Judged by this morphologi- 
cal criterion, as well as by the haemoglobin development, the 
basophilous corpuscles in the circulating blood of the pigeon 
are immature corpuscles whose cytoplasm is incompletely 
transformed into acidophilous haemoglobin. This relation 
of basophilous staining to the morphological development of 
the corpuscle in reptiles and birds is well brought out in the 
colored plates illustrating Gabritschewsky’s first article on 
this subject in 1891 (10). 
Both Ehrlich (5) and Maragliano (14, 15) call attention to 
the deficiency in haemoglobin of the basophilous corpuscles in 
anaemic human blood. But these authors have looked upon 
the phenomenon as a degenerative change in the discoplasm 
of the mature corpuscle whereby the acidophilous haemoglo- 
bin is converted into, or replaced by, a basophilous substance, 
instead of a process in which embryonic basophilous cyto- 
plasm is being transformed into acidophilous haemoglobin in 
the development of the corpuscle. 
The Conditions of the Appearance of Embryonic Red 
Corpuscles in the Circulating Blood. 
In all but a few of the higher mammalia immature red cor- 
puscles regularly enter the circulation during the physiologi- 
cal regeneration of the blood. Basophilous non-nucleated 
corpuscles are constantly present in the circulating blood of 
all of the lower species of vertebrates examined. Newman 
(16) states that nucleated red corpuscles are always present 
in small numbers in the blood of the pig. Howell (12) says 
that the same is true in the blood of the opossum. Why red 
corpuscles should enter the circulation before they are capa- 
ble of performing their proper function, since they are yet at 
most only incompletely supplied with haemoglobin, and thus 
load the blood to a certain extent with more or less useless 
corpuscles, is not apparent. Whether all new red corpuscles 
undergo their final stages of haemoglobin development in 
the circulation of these species, or whether only a part acci- 
dentally escape with the other mature corpuscles into it, and, 
if accidentally present, whether they complete their develop- 19 
ment after entering the circulation, or whether they never 
attain it, but live out their lives as useless corpuscles, are 
questions that cannot be answered with certainty. It would 
seem possible, however, that in the lower vertebrates the 
red corpuscle completed its final stages of haemoglobin de- 
velopment, at least in some of the corpuscles, after entering 
the circulating blood, instead of in the marrow, as is the case 
with the higher mammalia. Indeed, these are indications that 
as we descend the scale of development there is a* tendency 
for more and more of the final stages of the development of 
the red corpuscle —that is normally completed in the haema- 
topoietic organ in the higher mammalia —to be carried on 
after the corpuscle has entered the circulation. Basophilous 
corpuscles with incomplete haemoglobin development, absent 
in the higher mammals, are more and more prevalent as we 
descend the scale of classification. Nucleated corpuscles 
are said to constantly enter the circulating blood of some of 
the lower mammalia. And among the lower vertebrates, the 
aves and reptilia, the corpuscles are nucleated throughout 
their life. Embryological evidence points in the same direc- 
tion. The red corpuscles of the very young mammalian foetus 
are all nucleated. In the nearly developed foetus nucleated 
corpuscles are still present, but comparatively few in number. 
But the non-nucleated corpuscles show pronounced embry- 
onal characters : marked enlargement, deficiency in haemoglo- 
bin, and strong basophilous staining. In extra-uterine life 
the nucleated corpuscles disappear and the non-nucleated cor- 
puscles show less and less pronounced embryonal characters, 
and may lose them entirely. 
In certain of the higher mammalia, including man, imma- 
ture red corpuscles are usually restricted to the bone marrow, 
but under certain conditions they do appear in the circulating 
blood. Observation and experiment have determined these 
conditions to be any disturbance of the ratio between supply 
and demand of new red corpuscles. 
In the normal adult physiological regeneration of the blood 
hsemocytogenesis and haemocytolysis balance; new red cor- 
puscles are matured and turned into the blood current as 20 
fast as is required to make good the loss from the natural 
death of old corpuscles. During foetal life there is an exces- 
sive haemocytogenesis accompanying foetal development; red 
corpuscles are formed faster than they are matured. Imma- 
ture red corpuscles are consequently forced into the circula- 
tion and are always present in foetal blood. 
Ehrlich, Muller, Rindfleisch, Engel, etc., while not agree- 
ing as to the details of the changes in the marrow, all look 
upon the pernicious anaemias as due to defective and deficient 
haemocytogenesis. Some pathologists, however, conceive 
these anaemias to be due rather to excessive haemocytolysis. 
The so-called secondary anaemias may result from blood 
losses through haemorrhage, destruction of red corpuscles by 
parasites, or through other excessive haemocytolysis, or they 
may be due to defective haemocytogenesis, resulting from defi- 
cient nutrition, etc. But whether these anaemias be primary 
or secondary, due to deficient haemocytogenesis, to excessive 
haemocytolysis, or to direct blood losses, the demand for 
new red corpuscles exceeds the supply, and incompletely de- 
veloped corpuscles are drawn into the circulation in attempt 
to satisfy this demand. 
This is the explanation advanced by Smith (17) in 1890, 
considering their presence in the blood of cattle consequent 
to the pernicious anaemia of Texas cattle fever: “We shall 
not be far from the truth, I think, in assuming that we have 
here unfinished or embryonic corpuscles sent into the circu- 
lation before their time to make good the losses going on. 
We are led from the enlarged but otherwise normal corpuscles 
through those containing granules and staining diffusely to 
the haematoblast. Each stage is more embryonic than the 
preceding, and called into active service by more and more 
pressing need. We are, however, justified in asking whether 
these stages are those which the corpuscles ordinarily pass 
through, or whether pathological conditions are superadded. 
It is highly probable that there is both an embryological and 
a pathological factor involved.” This last statement is true 
with reference to the aggregation of basophiloplasm and to 
the overgrowth of the corpuscles in extreme anaemia. With 21 
these exceptions we now know these “ anaemia corpuscles ” 
to be normal stages of development of the red corpuscle that 
are always present in the haematopoietically active marrow. 
It is in this regard that Ehrlich (6) arrives at mistaken 
conclusions concerning the appearance of basophilous cor- 
puscles in the blood of starved dogs, and in cases of post- 
haemorrhagic anaemia at a period so early that, as he thinks, 
new corpuscles could not yet have had time to be formed. 
It is, however, not a question of the formation of new 
corpuscles, but of drawing into the circulation immature 
corpuscles that are already present in certain parts of the 
marrow as successive stages in the development of new red 
corpuscles required in the physiological regeneration of the 
blood. It is also probable that under the excitation of an 
increased demand the haematopoietic organ is stimulated to 
an increased activity in the production of new corpuscles. 
The same conditions that cause red corpuscles to escape 
prematurely from the marrow into the circulating blood of 
species that are normally free from them also increases their 
number in the blood of those species that normally contain 
them in small numbers. The effect of anaemias secondary to 
tuberculosis and to pregnancy on the number of basophilous 
corpuscles in the blood of guinea-pigs, rabbits, and dogs is 
shown in Table 1., where a comparison maybe made with the 
number in the blood of normal animals of these species. 
The significance of the basophilous substance in punctate 
or granular fragments in an oxyphilous stroma in certain 
corpuscles of anaemic blood is not understood. In all of the 
embryonic forms in the healthy marrow, in foetuses, and in 
the circulating blood of the lower species of vertebrates the 
transmutation of the basophilous cytoplasm into the oxyphilous 
haemoglobin takes place gradually and uniformly, the only 
exception to this being in the blood of the foetuses of dog 
number 4, in which corpuscles with basophilous granules are 
numerous. On the other hand, in anaemias, both of man and 
of other animals, these punctate forms are characteristic. In 
the blood of cattle in the progressive oligocythaemia of Texas 
fever, and from venesection, they are always the first form of 22 
basophilous corpuscle to appear, followed by the diffusely 
staining variety. In all of these cases with the exception of 
the dog foetuses just mentioned —the presence of basophilous 
corpuscles in the blood is abnormal; and since this punctate 
variety is not found among the developing corpuscles in the 
marrow or in the foetus, at least in healthy animals, they are 
under the suspicion of being the products of abnormal condi- 
tions. Smith (17) suggests that these granules may be derived 
from the diffusely staining cytoplasm through a condensation 
or aggregation of the basophilous substance in the circulation.1 
The practical value of the foregoing is that there is fur- 
nished a simple, quick, and at the same time fairly accurate 
means of microscopic diagnosis of anaemia. By the applica- 
tion of an extremely simple staining method to dried blood 
films, and without the use of the more complicated apparatus 
and technique necessary in making blood-corpuscle counts 
in fresh blood, the presence or absence of anaemia may be 
quickly determined. Exception might be made to possible 
cases in which a decrease of red corpuscles below the normal 
is not followed by an increased demand upon the haemato- 
poietic organ, the organism living, so to speak, upon a lower 
plane of vital activity. lam not aware, however, that such 
conditions have been proven to exist. The degree of the 
anaemia or of the blood loss in haemorrhage is roughly indi- 
cated by the number and phase of development of the em- 
bryonic corpuscles in the circulation, their number and 
1 After this had gone to the printer two papers by Plehn and Litten in recent num- 
bers of the Deutch med. Wochenschrift have come to my notice. 
Plehn (20) believes the granules and fragments in corpuscles in the blood in tropical 
anaemias that stain with hsematoxylin (in Ehrlich’s acid hasmatoxylin-alum-eosin) and 
with methylene blue to be spore or embryo stages of the malaria parasite. Litten 
(21) hesitates between two, to him possible, origins of the basophilous granules in red 
blood corpuscles in severe anaemias : (1) a karyolydc process (“ kernzerfall ”), and (2) 
a degenerative process in the haemoglobin-holding protoplasm. He finally favors the 
first process. Ehrlich also considers them to be the product of disintegration of the 
nucleus of the megaloblast. 
Smith's studies of the blood in the Texas fever (17, 18) fully explain the relation 
of these granules to the anaemia and to the hasmocytozoan parasite, and refute 
Plehn’s suppositions. That basophilous granules are not the product of disintegra- 
tion of the nucleus is proven by their presence, not only in the non-nucleated corpus- 
cles, but also in normoblasts with intact nuclei and in the immature nucleated corpus- 
cles. Moreover, their microchemical reactions show them to be basophilous but not 
chromatin, and in this identical with the cytoplasm of the embryonic red corpuscle. immaturity increasing with the severity of the anaemia. In 
mild secondary anaemias they are few in number and pre- 
sent simply slight enlargement and incomplete haemoglobin 
development associated with basophilous staining. In palu- 
dal and especially in the primary pernicious anaemias they 
may become very numerous, include nucleated forms, and 
present abnormal phases of development. By a more careful 
comparison of the results of red corpuscle counts with the 
number and phase of development of the embryonic cor- 
puscles in anaemic human blood, as was done by Smith in 
the blood of the ox in Texas fever, one might be enabled to 
estimate the degree of oligocythaemia within narrower limits. 
When considered with the clinical and microscopical evi- 
dence, these embryonic red corpuscles may, if judiciously 
employed, furnish important aid in differential diagnosis. 
They are constantly made use of in this laboratory as a 
supplementary guide in the microscopical diagnosis of ma- 
laria. This disease, being associated with a more or less 
extensive destruction of red corpuscles caused by the specific 
hsemocytozoan parasite, presents a more pronounced oligo- 
cythsemia than is likely to be associated with other febrile 
diseases, or diseases that have febrile symptoms, that might 
be mistaken clinically for malaria. Consequently, the pres- 
ence of immature corpuscles in the blood of suspected cases, 
especially if they are numerous, would lead to the suspicion 
of malarial infection even when the blood be taken for exam- 
ination at a time when the parasites were relatively few in the 
peripheral circulation. In such cases a more careful search 
is demanded, and is sometimes rewarded by the discovery 
of one or more parasites that would otherwise have been 
overlooked. Corpuscles containing punctate or granular 
fragments of embryonic basophiloplasm seem to be very 
characteristic of the severer grades of paludal anaemias. On 
the other hand, the absence of embryonic red corpuscles from 
the blood would help confirm a negative diagnosis. 
For diagnostic purposes the following simple technique is 
recommended: Dried blood films are to be fixed 15 to 30 
minutes in a mixture of equal parts of absolute alcohol and ether, stained 5 to 10 minutes with Ldffler’s alkaline methy- 
lene blue solution, and examined in water. The nuclei of 
the white and the red corpuscles, the basophilous corpuscles, 
and malaria parasites, if present, are stained blue. The 
normal red corpuscles are unstained. 
LITERATURE. 
1. Askanazy, S. Ueber einen interessanten Blutbefund bei rapid letal 
verlaufenden perincidsen Anarnie. Zeitschr. f. klin. Med., Bd. 
21, s. 415, 1892. 
2. Cabot, R. C. A Guide to the Clinical Examination of the Blood. 
[Win. Wood & Co., N.Y., 1897.] 
3. Celli und Guarnieri. Fortschr. der Med., s. 521, 1889. [Quoted 
from Smith (17) and Gabritschewsky (10).] 
4. Erb, W. Arch. f. path. Anat. u. Physiol, u. f. klin. Med. (Vir- 
chow), Bd. 34, s. 138, 1865. 
5. Ehrlich, P. Charite-Annalen, 1885. [Quoted from Smith (17)-] 
6. . Ueber Schwere Aniimische Ziistande. Verhandl. d. Congr. 
f. inn. Med., Leipzig, 1892. 
7. Ehrlich u. Lazarus. Die Anarnie [Special Pathologic und Therapie 
herausgegaben von Hofrath Prof. Dr. Hermann Nothnagel, VIII. 
Band, I. Theil, I. Heft, 1898]. 
8. Favre und Celli. Charitd-Annalen, Bd. x., 1885. [Quoted from 
Smith (17) and Gabritschewsky (10).] 
9. Foa und Mondino. Fortschr. des Med., No. 14, s. 524, ISB9. 
[Quoted from Smith (17) and Gabritschewsky (10).] 
10. Gabritschewsky, G. Klinische hamatologische Notizen. Arch. f. 
expr. Pathol, u. Pharmakol., Bd. 28, No. 5, s. S3, 1890. 
11. . Zeitschr. f. klin. Med., Bd. 27, s. 492, 1895. 
12. Howell. The Life History of the Formed Elements of the Blood. 
Journ. of Morphol., Vol. IV., No. 1, 1891. 
13. Lowit. Sitzungsberichtd. k. Akad. d. Wiss., Bd. XCV., s. 144, ISB7. 
14. Maragliano, E. XI. Congr. f. inn. Med., Leipzig, 1892. 
15. Maragliano, E., und Castellio, P. Zeitschr. f. klin. Med., Bd. 21, 
s. 415, 1892. 
16. Newman. Zeitschr. f. klin. Med., Bd. 3, s. 411, 1881. [Quoted from 
Howell (12).] 
17. Smith, T. On Changes in the Red Blood Corpuscles in the Perni- 
cious Anaemia of Texas Cattle Fever. Transactions of the Asso- 
ciation of American Physicians, September, 1891. 
18. . Investigations into the Nature, Causation, and Prevention 
of Texas or Southern Cattle Fever. U.S. Dept. Agric., Div. of 
Animal Industry, 1893. 
19. Troja. XI. Congr. f. inn. Med., Leipzig. 
20. Plehn, Albert. Tropenanamie und ihre Beziehung zur latenten und 
manifesten Malariainfection. Deutch. med. Wochenschrift., 
XXV. Jahrgang, No. 28-30. 
21. Litten, M. Ueber basophile Kdrnungen in rothem Blutkorpern. 
Deutch. med. Wochenschrift., XXV. Jahrgang, No. 44, s. 717, 
1899. 
Boston Society of Medical Sciences, November 21, 1899.