Clin. lab. Haemat.

1986, 8, 169-179

REVIEW

Reticulocytes

J.F. K O E P K E M D & J.A. K O E P K E M D

Department of Hospital Laboratories, North Carolina Memorial
Hospital, Chapel Hill, NC 27514 and Department of Hospital
Laboratories, Duke University Medical Center, Durham, NC
27710

The determination of erythropoietic activity is important in the evaluation of a

patient with anaemia. Anaemia can result from a great variety of causes. They
can be classified into two main groups. On the one hand, anaemia may be due to
decreased production of erythrocytes by the bone marrow, while on the other
hand, anaemia may be caused by increased erythrocyte destruction (as is the case
in haemolytic anaemia). In these latter cases the marrow may be
erythropoietically hyperactive.
The enumeration of newly-formed erythrocytes (i.e. reticulocytes) can be
useful for the estimation of the functional integrity of the erythropoietic marrow.
This review will recount the anatomy and physiology of reticulocytes, the
identification and counting of these cells and an interpretation of these results in
clinical medicine.

Definition of the reticulocyte

In the 1860s Erb noted granules in red blood cells of humans and animals made
anaemic by venesection. He used picric or acetic acid to demonstrate these
granules (Erb 1865). These granules may have been ‘reticulum’ or perhaps
denatured haemoglobin similar to Heinz bodies. Ehrlich was probably the first
investigator to describe the cells now regarded as reticulocytes. He used
methylene blue to stain the reticulum. He thought that the reticulum was a
‘progressive coagulation necrosis of the stroma’ (Ehrlich 1891). In 1891 Smith
described supravitally stained erythrocytes containing reticulum in pernicious
anaemia of cattle, a disease caused by the tick-borne protozoan Boophilus
annulatus. He felt that these cells represented not degenerative forms, but rather
‘embryonic corpuscles, sent into the circulation before their time to make good
the losses going on’ (Smith 1891). In the early 1930s Heilmeyer published the
Correspondence: John A. Koepke, MD, P.O.Box 2929, Duke University Medical Center, Durham,
N C 27710.

169
170 J.F. Koepke and J.A. Koepke

classic descriptions of reticulocyte morphology at different stages of maturation

as well as the relative frequencies of these stages in the peripheral blood.
Heilmeyer divided the reticulocytes into four groups, plus a group 0 for
normoblasts that contain a nucleus as well as a dense perinuclear reticulum. In
group I, the reticulum appears in the form of a dense clump; in group!II, in a
wreath; in group 111, the wreath has disintegrated; and in group IV, only a few
scattered granules of the reticulum remain (Heilmeyer & Westhauser 1932) (see
Figure 1). In normal individuals, about 61% of the circulating reticulocytes are
group IV, about 32% are group 111, about 7% are group 11, and only about 0.1%
are group I (Seip 1953).
Noting the difficulty in precisely differentiating reticulocytes from mature
erythrocytes in interlaboratory surveys, Gilmer & Koepke (1976) proposed a
definition for this cell. A reticulocyte is therefore defined in the NCCLS Standard
for Reticulocyte Counting (Finch, Marshall & Trobaugh 1984) as any non-
nucleated red cell which contains two or more particles of blue-stained material
corresponding to ribosomal RNA. Other authorities require that a minimal
network remain to be a stage IV cell; still others require three dots of reticulin.
Non-nucleated cells containing larger amounts of blue-staining material will
obviously be included in the reticulocyte category.

Physiology of reticulocytes

Reticulocytes are immature red blood cells (RBC) in the final stages of
differentiation contemporary with the metarubricyte’s release into the peripheral
circulation. A reticulocyte is formed when a metarubricyte (orthochromatic
erythroblast) ejects its nucleus. The erythroblast divides unequally so that the
smaller portion contains the nucleus and a thin rim of cytoplasm while the larger
portion becomes the reticulocyte. The expelled nucleus is quickly ingested by a
neighbouring macrophage and is destroyed. The reticulocyte must then traverse
the wall of the marrow sinusoid, a specialized marrow vessel, in order to gain
access to the peripheral, blood circulation. The exact mechanism by which the
reticulocyte crosses this barrier is not known. Some investigators feel that the
diapedesis occurs at a junction between two endothelial cells similar to leukocytes
(Bessis 1973) while others feel that the mature reticulocyte passes through the
cytoplasm of a single endothelial cell (emperiopolesis) rather than between two
cells (Tavassoli & Takahashi 1982). In vitro measurements of membrane and
whole cell deformability indicate that erythroid maturation is accompanied by a
progressive increase in cellular deformability (Leblond, LaCelle & Weed 1971).
Thus, irrespective of the mechanism, more mature erythroid cells may more easily
transverse the marrow sinus wall.
The early reticulocyte contains mitochondria, a small number of ribosomes,
the centriole, and remnants of Golgi bodies. It is the precipitation and
aggregation of these organelles (particularly ribonucleoprotein) by supravital dye
techniques that give reticulocytes their name. The exact appearance of the
 Reticulocytes 17 1

0. 1. It.

111. IV.
Figure 1. Reticulocyte maturation according to the five stages proposed by Heilmeyer (1932). 0,
nucleated red cell within marrow space; I, dense reticulum in the most immature non-nucleated cell;
11, more extensive but looser reticulum; 111, residual reticulum network; IV, scattered granules of
reticulum. This stage proves to be the most difficult to consistently identify. Reprinted with
permission Friedman (1984).

precipitate is extremely variable. When stained it may take the form of a delicate
network, a collection of fragments and rods, a few scattered dots, or diffuse fine
granules. This precipitate has been variously termed reticulum, granulo:
filamentous substance and reticulofilamentous substance (Marshall 1978). Hence
the cell is called a reticulocyte.
Maturation of the reticulocyte, determined by radioiron kinetic studies,
normally requires about three and one half to four days, and usually only the last
24 hours ,,or so is spent in the peripheral blood (Labardini et al. 1973,
Papayannopoulou & Finch 1975, Hillman & Finch 1985). Numerous
morphologic changes occur during the maturation of the reticulocyte (Bessis &
Bricka 1952). Using phase microscopy it has been shown that the youngest
reticulocytes have a multilobular form and are motile while more mature
reticulocytes are cup-shaped and non-motile (Coulombel, Tchernia & Mohandes
1979). Ultrastructural studies have shown that as the reticulocyte matures, the
mitochondria decrease in size and number and show degenerative changes such
as vacuolization and loss of cristae. The reticulocyte ejects degenerating
organelles by autophagy, the process in which small vacuoles encircle organelles,
fuse into a large single autophagosome, and then ejects the contents into the
extracellular space (Simpson & King 1968). Maturation of the reticulocyte is
concurrent with synthesis of the remaining one-third of the red cell haemoglobin
content (Papayannopoulou & Finch 1975).
172 J.F. Koepke and J.A. Koepke

Manual reticulocyte counting procedure

In the United States the most popular supravital staining technique for
reticulocytes is the new methylene blue (NMB) method, first described in the late
1940s (Brecher 1949). t

The manual method for counting reticulocytes is simple. A few drops of a

1.0% (w/v) saline (NaCl, 0.152 mol/l) solution of New Methylene Blue (NMB) or
Brilliant Cresyl Blue are mixed with an equal volume of peripheral blood
collected in EDTA (1.5 mg/dl). The solution is mixed and using a pipette or other
transfer device an evenly thin film of stained blood is made on a 2.5 x 7.5 cm
microscope slide (Finch, Marshall and Trobaugh, 1984). Staining is rapid and
need not be done for more than 1 or 2 minutes.
After the film has dried, an area of the film where the cells are undistorted and
the reticulocytes appear to be well-stained is chosen for examination. Routinely
1000 erythrocytes are examined and the proportion of erythrocytes containing
reticulin is determined (Finch, Marshall & Trobaugh 1984). In some laboratories,
special devices which allow for more efficient counting are used, although their
effect on improving precision is questionable (Savage, Skoog & Rabinovitch 1985).

Reference ranges

Reticulocyte counts are commonly expressed as a percentage4.e. the number of

reticulocytes per 100 erythrocytes. The reference range for the reticulocyte count,
when 1000 cells are counted, is usually given as 1.00% k0.5 (meank2 SD). The
reference range is significantly higher in women than in men; 2.45%*0.82 vs
1.65%&0.44, respectively, in one study (Deiss & Kurth 1970). This probably
reflects a response to the need to replace blood lost during menstruation. The
reticulocyte count is higher in newborn infants with the normal range usually
given as 2 to 6%, but this drops to normal adult levels in 1 to 2 weeks
(Lowenstein 1959). Higher reticulocyte counts have also been reported in normal
subjects living at altitudes above 1850 m.

Correction methods

In certain cases the percentage reticulocyte count may be misleading, e.g. in

patients with anaemia or in periods of intense erythropoietic stimulation. In these
cases the percentage reticulocyte count may be corrected with a packed cell
volume correction and/or a reticulocyte maturation time (or so-called ‘shift’)
correction.

Packed cell volume (PCV) correction

The percentage reticulocyte count can increase either because there are more
 Reticulocytes 173
reticulocytes in the circulation, or because there are fewer mature red blood cells
circulating, as in anaemia. For this reason, the observed reticulocyte count is
corrected to a normal PCV, i.e. 0.45. A reticulocyte count corrected for anaemia
is termed the reticulocyte index. It is calculated using the following formula:
patient’s PCV
Reticulocyte index = observed reticulocyte count (in %) x
0.45
An alternate way of correcting for anaemia is to obtain an absolute reticulocyte
count. This is obtained by multiplying the reticulocyte percentage by the red
blood cell count. In absolute numbers the mean normal value is around 5 x 10”
reticulocytes/l.

Shift correction
The reticulocyte count in the peripheral blood is a result of three variables: (1) the
rate of release of reticulocytes from the bone marrow; (2) the degree of
immaturity of the freshly released reticulocytes; and (3) the rate of disappearance
of reticulum (Baldini & Pannacclulli 1960). The last variable is usually not
appreciated, Under most conditions the relationship between the absolute
reticulocyte count and the number of erythrocytes produced in the bone marrow
is linear. However, in periods of intense erythropoietic stimulation younger
reticulocytes (so-called ‘shift cells’) are released into the peripheral blood. These
‘shift’ cells are basophilic macroreticulocytes with a cell diameter about 25%
greater than normal cells (Perrotta & Finch 1972). Some investigators feel that
under the effects of an erythropoietic stress, a generation is skipped in the
maturation sequence and larger cells are released prematurely into the circulation
(Brecher & Stohlman 1961, 1962). Other investigators do not postulate a skipped
generation but rather feel that these larger cells are not significantly different
from normal marrow reticulocytes but are merely released prematurely. Once in
the circulation they lose both haemoglobin and water and become normal
erythrocytes (Ganzoni, Hillman & Finch 1969, Papayannopoulou & Finch 1975).
In periods of intense erythropoietic stress, the reticulocyte maturation time in the
marrow may be shortened to less than 1 day, while the reticulocyte maturation
time in the peripheral blood correspondingly lengthens. Thus the number of
reticulocytes may be markedly increased in the blood without any increase in the
erythropoietic activity in the marrow. In theory only the last day of a
reticulocyte’s life should be considered when evaluating erythrocyte production.
There is no practical way, based on RNA content, to accurately determine the
age of a reticulocyte. However, if one sees an abnormally’high percentage
(greater than 5%) of ‘shift’ cells an empirical correction for RBC maturation
based on the packed cell volume may be made. A reticulocyte count corrected
both for PCV and maturation time is termed the reticulocyte production index.
Use of the reticulocyte production index as compared to an uncorrected
174 J.F. Koepke und J.A. Koepke

Marrow
Normoblasts and Blood
Hematocrit Reticulocytes Reticulocytes
t

45 3.5

35 3.O
25 2.5

15 1.5
I
Figure 2. The correlation of the packed cell volume (haematocrit) with the blood reticulocyte
maturation time. With increasing anaemia (and erythropoietin production) the maturation time of
the erythroid marrow normoblasts and reticulocytes progressively shortens from a normal 3.5 days
to 1.5 days or less. Conversely the reticulocytes in the peripheral blood persist for a longer time
when the patient is anaemic. These facts should be taken into account when calculating the
reticulocyte production index (reprinted with permission from Hillman & Finch 1985).

reticulocyte count clearly improves the diagnostic correlation. The reticulocyte

production index is calculated using the following formula:

Observed reticulocyte Y Patient's PCV

count (in %) 0.45 "

Reticulocyte production index =

Maturation time in peripheral blood

The maturation time of the reticulocyte in the peripheral blood is taken as 1 day
when the PCV is 0.45k0.5, 1.5 days when the PCV is 0.35k0.05,2 days when
the PCV is 0.25 k0.05,and 3 days when the PCV is 0.15k0.05 (Hillman & Finch
1967). Thus if a patient has an observed reticulocyte count of 9% and a PCV of
0.25, the reticulocyte production index is 2.5 (9% x 25/45+2). These
relationships are shown graphically in Figure 2. The normal range for the
reticulocyte production index is the same as for the uncorrected reticulocyte
count since the normal maturation time in the peripheral blood is 1 day.
There are certain pathological states, e.g., renal failure, where the
erythropoietin output is not consistent with the degree of anaemia present, and
the expected prolongation of reticulocyte maturation time does not occur. For
the shift correction to be valid, there must be a normal relationship between the
degree of anaemia and the increased erythropoietin concentration which
produces the shift. This is validated by finding a high percentage of basophilic
macroreticulocytes, or 'shift' cells, in the peripheral blood film.
 Reticulocytes 175

Clinical requirements of reticulocyte counting

Clinicians frequently request a reticulocyte count to help delineate the cause of

anaemia or to assess a patient’s response to therapeutic intervention. The
reticulocyte count can discriminate between bone marrow failure (decreased
reticulocyte count) and increased bone marrow output (increased reticulocyte
count). If it is within the normal range, it is of relatively little help in diagnosis of
an anaemic patient, although one might presume neither haemolysis nor marrow
failure are responsible for the anaemia.
In addition to diagnostic use, the reticulocyte count is also useful in assessing
the cause for an elevated MCV (reticulocytes have an increased MCV) and in
evaluating response to nutritional supplements in iron, folate, or vitamin B,,
deficiency states.
The recent introduction of the concept of reticulocyte production indices and
methods of correcting ‘raw’ reticulocyte counts for PCV and maturation time can
better help the clinician assess specific causes of anaemia. Such derived
information provides valuable data in the diagnosis of anaemia. Elevated
reticulocyte production indices ( > 3) are associated with chronic haemolysis,
recent haemorrhage, or response to therapy. Depressed indices (<2) are generally
associated with either marrow failure or ineffective erythropoiesis, as in the
megaloblastic anaemias due to vitamin BI2 or folate deficiencies (Hillman &
Finch 1985, Crosby 1981). Thus, the reticulocyte has progressed in importance
from a morphological and staining curiosity to a widely used method of assessing
erythrocyte production.

Assessment of reticulocyte counting

There are two major factors which may influence the reliability of the reticulocyte
count; one is physiological and the other is related to laboratory technique
(Friedman 1984).
The physiological factor relates to the physiological variation of reticulocyte
production. Variations in the reticulocyte count have been claimed to occur
diurnally, daily or seasonally, but these deviations are not felt to be significant
(Lowenstein 1959).
Inter- and intralaboratory imprecision and inaccuracy affect reticulocyte
counting. The initial studies in this area were done in the 1970s by the College of
American Pathologists Hematology Resource Committee. They demonstrated an
extremely high interlaboratory coefficient of variation (from 25 to 48%) on three
stained reticulocyte smears distributed in the 1971, 1972 and 1974 Comprehensive
Hematology Surveys (Gilmer & Koepke 1976). Subsequent studies indicate four
different analytical variables which may contribute to the technical variation in
reticulocyte counts. They are: (1) interobserver variation in reticulocyte
definition; (2) the size of the sample of red cells evaluated; (3) the type of blood
176 J.F. Koepke and J.A. Koepke

film evaluated (i.e., variability in cell distribution); and (4) the use or lack of use
of an insert in the ocular of the microscope as a counting aid for standardization
of area reduction (Savage & Skoog 1985).
The best documented cause of imprecision in reticulocyte counting ,is
interobserver variation in the definition of a reticulocyte. The identification of the
Heilmeyer group IV cell is the most variable measurement due to interlaboratory
variation in the exact definition of a group IV cell and it was suggested that two
or more clumps or granules of reticulin be visible (Gilmer & Koepke 1976).
However, other authorities require an identifiable reticulum network remain
while others require three dots of reticulum. The inaccuracy of counting group IV
cells is further magnified by the fact that cytoplasmic particles unrelated to
ribosomal RNA may appear as granules on a supravital stain. These include
siderotic granules, nuclear remnants, Heinz bodies, precipitated stain and other
debris. One can readily appreciate that the difficulty in positivity identifying
group IV cells combined with the fact that about 60% of the reticulocytes in the
circulation are group IV cells can easily lead to markedly increased imprecision in
reticulocyte counts.
A statistical appraisal of manual reticulocyte enumeration was investigated by
Peebles, Hochberg & Clark (1981) who concluded that technologist-to-
technologist variation was the major source of inaccuracy at all reticulocyte
levels. The source for this large variation was different criteria for the
morphological identification and enumeration of reticulocytes. Although the
results might be clinically useful, they felt it was extremely difficult to obtain
manual results with sufficient accuracy to serve as a ‘reference’ reticulocyte
method (Peebles et al. 1981).
The second major analytical variable contributing to imprecision in
reticulocyte counting is the size of the sample evaluated. The precision of
reticulocyte counting follows a binomial distribution. Since the normal
percentage of reticulocytes is low, statistical variation can be great. For example,
the 95% confidence limits for a 1% reticulocyte count when 100 red blood cells
are counted is 0.03 to 5.45%; when 1000 red blood cells are counted the 95%
confidence limits drops to 0.48 to 1.84%. Thus it is appropriate to count at least
1000 red blood cells to maximize analytic precision in manual reticulocyte
counting (Greenberg & Beck 1984).
Variation due to the type of smear stained for examination is the third major
analytical variable which may contribute to the imprecision in reticulocyte
counting. One study reports that the distribution of reticulocytes and
erythrocytes is more uniform on spun slides as compared with wedge films (May
& Sage 1976). Further testing is needed to evaluate this potential source of
reticulocyte count imprecision.
Finally, the use or the lack of use, of a standardized area reduction device as a
counting aid in the microscope may contribute to imprecision in reticulocyte
counting. Early studies indicated that the use of the Miller disk for reticulocyte
counting diminished the observed standard error of the reticulocyte count
particularly at low reticulocyte levels (Brecher & Schneidermann 1950). However,
 Reticulocytes 177

most laboratories have chosen not to use such an aid. This too is an area which
requires further testing to evaluate its contribution to the imprecision of
reticulocyte counting.

Other reticulocyte counting methods

The present manual New Methylene Blue method for performing reticulocyte
counts is time consuming since 1000 red blood cells must be counted. However,
even when counting this number of cells this method is still relatively imprecise.
Because of this, methods for automating this laboratory test are being developed.
One technique recently developed uses automatic scanning of a NMB-stained
blood film with a pattern recognition device such as a Geometric Data Hematrak
(Perel, Herrman & Watson 1980), or a Coulter diff 3 (Kaplow & Crouch 1982).
The data obtained with these methods appears to be highly reproducible and in
good agreement with manual counting techniques. However, one major
drawback of the Hematrak system is that the operator must manually review all
reticulocytes found by the Hematrak. With time this leads to a decreasing
counting accuracy due to operator fatigue.
In 1952, a method for counting reticulocytes using acridine orange
fluorescence was first reported (Kozenow & Mai 1952). Certain fluorescent dyes
(e.g. acridine orange, pyronin Y and thioflavin T) combine with ribosomal RNA
in reticulocytes to produce fluorescence in ultraviolet light. The amount of
fluorescence in a particular cell is proportional to the amount of RNA present,
which in turn relates to the maturity of the reticulocyte. There is satisfactory
agreement between the microscopic fluorometric results and those obtained by
counting separately the four Heilmeyer reticulocyte maturation groups (Vander,
Harris & Ellis 1963; Thaer & Becker 1975). These fluorescent staining techniques
combined with flow cytometry have led to new automated systems for
reticulocyte counting (Tanke et al. 1980, 1983; Sage, O’Connell & Mercolino
1983). These flow cytometry methods have the potential to overcome the two
major deficiencies of the currently used manual reticulocyte counting methods.
First, since large numbers of cells are counted, the statistical reliability of low
reticulocyte counts as well as small changes in the reticulocyte count are much
improved. Second, as the amount of fluorescence in a cell is proportional to the
amount of RNA in the reticulocyte, these systems give information on the degree
of maturation of the reticulocyte population. This provides a more accurate
means for assessing the effective rate of red cell production by accounting for
shortened marrow transit time.
Reticulocytes have transferrin binding sites on their surface that are lacking on
mature red cells. This fact has led to an adaptation of the automated system
described above in which an immunofluorescent assay with antitransferrin
receptor antibody (which identifies the transferrin receptor on reticulocytes) is
combined with flow cytometry. This system also appears to give reproducible
reticulocyte counts (Seligrnan et al. 1983).
178 J.F. Koepke and J.A. Koepke

Need for standardization .

Given the technical problem with reticulocyte counting, it is probable that
standardization of this common test would yield significant clinical rewards. The
imprecision is probably of little concern when the reticulocyte count is elevate'd.
However, the converse, the accurate assessment of a decrease or absence of
reticulocytes, promises to significantly improve our ability to evaluate bone
marrow failure. Standardization efforts should, in our view, be directed toward
improvement in the precision in low reticulocyte counts. Normal, and especially
high reticulocyte counts, done manually are usually adequate if done according
to published standards.
The second standardization need is to provide a reference method for the
several automated systems now available for clinical use. We are again faced with
the prospect of using a less precise reference for evaluation of the considerably
more precise instrumental methods of reticulocyte count. And here the normal
and high counts are more important for the calibration of instrumental methods.

References

BALDINI M. & PANNACCLULLI I. (1960) The maturation of reticulocytes. Blood 15,614 .

BESSISM. (1973) Living blood cells and their ultrastructure. p. 140 Springer-Verlag, Berlin
B~ssrsM. & BRICKA M. (1952) Aspect dynamique des cellules du sang. Son etude par la microcint-
matographie en contraste de phase. Rev. Hkmatol. 7,407
BRECHER G. (1949) New methylene blue as a reticulocyte stain. Am. J. Clin. Pathol. 19, 895
BRECHER G. & SCHNEIDERMAN M. (1950) A time saving device for counting reticulocytes. Am. J.
Clin. Pathol. 20, 1079
BRECHER G. & STOHLMAN F. (1961) Reticulocyte size and erythropoietic stimulation. Proc. SOC.
Exp. Biol. Med. 107, 887
BRECHER G. & STOHLMAN F. (1962) The macrocytic response to erythropoietin stimulation. In
Erythropoiesis. Eds L. Jacobson & M. Doyle. p. 216 Grune and Stratton, New York
COULOMBEL L., TCHERNIA G. & MOHANDES N. (1979) Human reticulocyte maturation and its
relevance to erythropoietic stress. J. Lab. Clin. Med. 94,467
CROSBY W . (1981) Reticulocyte counts. Arch. Intern. Med. 141, 1747
DEIS A. & KURTHD. (1970) Circulating reticulocytes in normal adults a s determined by the new
methylene blue method. Am. J. Clin. Pathol. 53,481
EHRLICH P. (1891) Uber einige Beobachtungen am anamischen. Blut. Berlin Klin. Wochenschr. 18,
43
ERBW.(1865) Zur Enturichelungseschichte der rothen Blutkorperchen. Virchows Arch. Path. Anat.
Physiol. 34, 138
FINCH C.A., MARSHALL P.N. & TROBAUCH F.E. (1984) Method for reticulocyte counting. NCCLS
Proposed Standard H16-P National Committee for Clinical Laboratory Standards, Villanova.
Pennsylvania
FRIEDMAN E. (1984) Reticulocyte counts: How to use them, what they mean. In Diagnostic
Medicine, June. pp. 33
GANZONI A., HILLMAN R. & FINCH C. (1969) Maturation of the macroreticulocyte. Brit. J. Haemat.
16. 119
GILMER P. & KOEPKE J. (1976) The reticulocyte: An approach to definition. Am. J. Clin. Pathol. 66,
262
GREENBERG E. & BECKJ. (1984) The effects of sample size on reticulocyte counting and stool
examination. Arch. Pathol. Lab. Med. 108, 396
 Reticulocytes 179

HEILMEYER L. & WESTHAUSER R. (1932) Reifungsstudien an iiberlebenden Reticulocyten in vitro und

ihre Bedeutung fiir die Schatzung der taglichen Haemoglobinproduktion in vivo. Ztsch. Kline.
Med. 121,361
HILLMAN R. & FINCH C. (1967) Erythropoiesis: Normal and abnormal. Semin. Hemat. 4,327
HILLMAN R. & FINCH C. (1985) Red cell manual, 5th ed. pp. 49, 56-58 FA Davis, Philadelphia
KAPLOW L. & CROUCH J. (1982) Automated reticulocyte counts. Lab. Med. 13, 167
KOZENOW W. & MAIH. (1952) Fluorescence microscopy of blood cells Ztschr. Kinderheilk. 70,552
LABARDINI J., PAPAYANNOPOULOU T., COOKJ. ADAMSON J. WOODSON R., ESCHBACH J., HILLMAN
R. & FINCH C. (1973). Marrow radioiron kinetics. Haematologica 7, 301
LEBLOND P., LACELLE P. & WEEDR. (1971) Cellular deformability: A possible deforminant of the
normal release of maturing erythrocytes from the bone marrow. Blood 37,40.
LOWENSTEIN L. (1959) The mammalian reticulocyte. Int. Rev. Cytol. 8, 135
MARSHALL P. (1978) Reticulation, polychromasia and stippling of erythrocytes. Microscopica
Acta 81, 89
MAYJ. & SAGEB. (1976) Spinner films for reticulocyte counts. Am. J. Med. Technol. 42, 357
PAPAYANNOPOULOU T. & FINCH C. (1975) Radioiron measurements of red cell maturation. Blood
Cells 1, 535
PEEBLES D., HOCHBERG A. & CLARKT. (1981) Analysis of manual reticulocyte counting. Am. J.
Clin.Pathol. 76, 713
PERELI., HERRMANN N. & WATSONL. (1980) Automated differential leukocyte counting by the
Geometric Data Hematrak system: Eighteen months experience in a private pathology
laboratory. Pathology 12,449
PERROTTA A. & FINCH C. (1972) The polychromatophilic erythrocyte Am. J. Clin. Pathol. 57,471
SAGEB., OCONNELLJ. & MERCOLINO T. (1983) A rapid, vital staining procedure for flow
cytometric analysis of human reticulocytes. Cytometry 4, 222
SAVAGE R.A., SKW D.P. & RABINOVITCH A. (1986) Analytic inaccuracy and imprecision in
reticulocyte counting: a preliminary report from the College of American Pathologists
reticulocyte project. Blood Cells 11,97
SEWM. (1953) Reticulocyte studies. Acta Med. Scand. Suppl. 282, 9
SELIGMAN P., ALLENR., KIRCHANSKI S. & NATALE,P. (1983) Automated analysis of reticulocytes
using fluorescent staining with both acridine orange and an immunofluorescence technique.
Am. J. Hemat. 14, 57
SIMPSON C. & KINGJ. (1968) The mechanism of mitochondria1 extrusion from phenylhydrazine-
induced reticulocytes in the circulating blood. J. Cell. Biol. 36, 103
SMITH T. (1891) On changes in the red blood cell corpuscles in the pernicious anemia of Texas cattle
fever. Trans. Assoc. Am. Phys. 6,263
TAVAWLIM. & TAKAHASHI K. (1982) Photo/Essay-Birth of the red cell. JAMA 247,2416
TANICE H., ROTHBARTH P., VOSSEN J., KOPERG. & PLOEM J. (1983) Flow cytometry of reticulocytes
applied to clinical hematology. Blood 61, 1091
TANKEH., NIEUWENHUIS I., KOPERG., SLATSJ. & PLOEMJ. (1980) Flow cytometry of human
reticulocytes based on RNA fluorescence. Cytometry 1, 313
THAER A. & Becker H. (1975) Microscopic fluorometric investigations on the reticulocyte
maturation distribution as diagnostic criterion of disordered erythropoiesis. Blut. 30, 339
VANDER J., HARRIS C., & ELLISS. (1963) Reticulocyte counts by means of fluorescence microscopy.
J. Lab. Clin. Med. 62, 132