Adult Xenopus laevis, rendered anaemic by phenylhydrazine injection, have been studied during the recovery from such anaemia. Electron microscopy of liver and spleen sections indicates that both of these organs are active in the phagocytosis and destruction of the old damaged red blood cells. May-Grunwald and Giemsa staining of liver and spleen cells following anaemia has been used to show that erythropoiesis also occurs in both liver and spleen, and this has been confirmed by electron-microscope studies of these organs.

Cell counting and radiolabelling of the new population of circulating erythroid cells in the period following phenylhydrazine injection suggests that a sudden release of basophilic erythroblasts from liver and spleen is followed by mitosis of this new cell population in circulation, and that no further release of erythroid cells from these organs is likely until complete recovery has occurred.

Adults of Xenopus laevis can be rendered anaemic by injection with phenylhydrazine (Thomas & Maclean, 1975). The induced anaemia is sufficiently severe to involve entire elimination of the existing mature erythrocytes from circulation. Before their disappearance, the red cells become somewhat mis-shapen, and their rapid elimination from circulation clearly implies very efficient removal by the appropriate tissues.

The anaemic animal responds to its condition by releasing relatively early erythroid cells into the blood from the erythropoietic tissues, and these cells pass through a number of divisions in the blood as they differentiate to the mature erythrocyte stage.

We have already published evidence of the biochemical events which are involved in this differentiation (Maclean & Jurd, 1971; Thomas & Maclean, 1974, 1975; Hilder, Thomas & Maclean, 1975) and of the regulation of red cell numbers which it involves (Aleporou & Maclean, 1978). In this paper we present our findings relating to the sites of erythropoiesis during anaemia, the divisions of the new erythroid cells in circulation, and the process by which the old red cells are removed from circulation and destroyed.

Animals

Mature adult female Xenopus laevis were obtained from Harris Biological Supplies, Weston-Super-Mare, and maintained as previously described (Maclean & Jurd, 1971). The temperature of the aquarium which housed the animals was not allowed to fall below 15 °C.

Chemical reagents

All reagents were obtained from British Drug Houses Ltd, Poole, England, except where otherwise indicated. The radioactive reagents were obtained from the Radiochemical Centre, Amersham, England.

Induction of anaemia

Mature, female Xenopus, of body weight approximately too g, were made anaemic by phenylhydrazine injections as previously described (Thomas & Maclean, 1975).

Collection of blood

The anaemic animals, at various times during recovery, were anaesthetized with MS 222 (Sandoz Products Ltd, London) and blood was collected by cardiac puncture. An aliquot of the blood collected was used for determination of the blood cell count (Improved Neubauer haemocytometer). The remainder was washed twice in 20 vol. of Rugh Ringer’s solution containing 8 mg/ml bovine serum albumin and finally resuspended in 1 ml Ringer’s solution. The cells were lysed with 1 mg saponin, the haemolysate was centrifuged at 1000 g to pellet the cell debris, and the clear red supernatant was collected for haemoglobin estimations.

Haemoglobin estimations

The method of Singer, Chernoff & Singer (1951) as described by Wallet & Robinson (1968) was used.

Preparation of tissue touch imprints

Liver and spleen of normal and anaemic Xenopus were removed from thoroughly bled animals. Cut surfaces of these organs were touched onto clean glass slides after being blotted on filter paper to remove excess blood. Air-dried liver and spleen touch imprints were stained in May-Grunwald solution (G. T. Gurr Ltd, London) for 15 min. The cells were then counterstained with Giemsa stain Improved R66 (G. T. Gurr Ltd) diluted 1:20 in 0 ·01 M phosphate buffer, pH 7 · 0 for 30 min. The tissue touch imprints were then washed in running tap water, air-dried and mounted in Euparal (GBI Labs Ltd, Manchester, England).

Autoradiography

At stated times after the induction of anaemia with phenylhydrazine, adult Xenopus were injected with 0·1 ml of sterile solution containing 0·1mlof tritiated thymidine. Immediately after injection and on each day following, a drop of blood was obtained by nicking the foot web, smeared on a clean slide, and used for the preparation of an autoradiograph. Slides were air dried, fixed in methanol and brought to water through a graded series of ethanol concentiations. Following exposure to 5 % trichloroacetic acid at 20 °C for 2 h, slides were washed, dried, and coated with K5 emulsion (Ilford Ltd.) diluted 1/1 with distilled water. After drying, the smears were exposed at 4 °C for 4 days, then developed in Kodak D19 developer (Kodak Ltd) and fixed in Kodafix (Kodak Ltd).

Electron microscopy

Samples of liver, spleen and blood were taken from anaesthetized or freshly killed anaemic Xenopus. The liver and spleen were cut into small blocks of 1–2 mm3. The tissue pieces as well as the blood cells were fixed in 2 % glutaraldehyde in 0 · 05 M sodium cacodylate containing 0 · 25 M sucrose at pH 7·2–7·4 for 5 h at 4 °C, washed overnight in cacodylate/sucrose buffer then postfixed in 1 % osmium tetroxide in the same buffer for 2 h and finally washed in cacodylate/sucrose buffer (3 changes, 30 min each). Subsequently the tissues were dehydrated in a graded series of ethanol and then passed successively to propylene oxide and through 1:1 mixture of propylene oxide and Spurt resin and finally embedded in Spurr resin overnight. Polymerization took place at 60 °C for 24–48 h. Ultrathin sections approximately grey to gold were cut on an LKB ultratome equipped with glass knives. The sections were collected on 200-mesh uncoated copper grids. These were stained with uranyl acetate followed by lead citrate (Reynolds, 1963) and examined on a Philips EM 300 equipped with a 3o-/tm objective aperture and operated at an accelerating voltage of 80 kV.

Destruction of old mature erythrocytes

Our electron-microscope studies clearly indicate that the old erythrocytes are trapped by both the spleen and the liver. Very active phagocytosis occurs in both spleen (Figs. 1, 2) and liver (Figs. 3, 4), and individual phagocytes containing the remains of four or five erythrocytes are common. Extensive red cell degradation often precedes actual erythrocyte engulfment, both nucleus and cytoplasm of the erythrocytes exhibiting condensation of the normal molecular conformation.

Fig. 1.

Photograph of section of glutaraldehyde-fixed spleen from an adult Xenopus rendered anaemic by phenylhydrazine injection u days previously, viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining. The picture shows a spleen macrophage containing a darkly stained nucleated erythrocyte. Following phenylhydrazine treatment, the spleen is obviously highly active in removal and destruction of the old erythrocytes, × 8000.

Fig. 1.

Photograph of section of glutaraldehyde-fixed spleen from an adult Xenopus rendered anaemic by phenylhydrazine injection u days previously, viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining. The picture shows a spleen macrophage containing a darkly stained nucleated erythrocyte. Following phenylhydrazine treatment, the spleen is obviously highly active in removal and destruction of the old erythrocytes, × 8000.

Fig. 2.

As for Fig. 1, but in this case the remains of many partly degraded erythrocytes are visualized inside a single spleen macrophage, × 6500.

Fig. 2.

As for Fig. 1, but in this case the remains of many partly degraded erythrocytes are visualized inside a single spleen macrophage, × 6500.

Fig. 3.

Photograph of a section of glutaraldehyde-fixed liver from an adult Xenopus rendered anaemic by phenylhydrazine injection 11 days previously, viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining. The picture shows a liver macrophage containing a darkly stained nucleated erythrocyte. × 6900.

Fig. 3.

Photograph of a section of glutaraldehyde-fixed liver from an adult Xenopus rendered anaemic by phenylhydrazine injection 11 days previously, viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining. The picture shows a liver macrophage containing a darkly stained nucleated erythrocyte. × 6900.

Fig. 4.

As for Fig. 3, but in this case the remains of 2 partly degraded erythrocytes are visualized inside a single liver macrophage, × 8000.

Fig. 4.

As for Fig. 3, but in this case the remains of 2 partly degraded erythrocytes are visualized inside a single liver macrophage, × 8000.

Erythropoiesis

Clear evidence about erythropoiesis is not always easy to obtain in anaemic animals because of the possible confusion over trapping of previously circulating erythroblasts. We have attempted to overcome this difficulty by using both light microscopy of tissue-touch imprints and electron microscopy to identify stages of erythroblast differentiation in presumptive erythropoietic tissue, not at that time recoverable from the blood.

Eight days after the final phenylhydrazine injection, the circulating blood cells substantially comprise a mixture of damaged mature erythrocytes and new basophilic erythroblasts. By day 11 of anaemia, no basophilic erythroblasts are detectable in the blood, only the later stages of polychromatophilic erythroblasts. At that time, however, light microscopy of tissue-touch imprints and electron microscopy of embedded tissue indicate that both liver and spleen contain many cells identifiable as basophilic erythroblasts (see Figs. 5-8). The rate of red cell replacement in the animals used in our present study is indicated by Table 1.

Table 1.

Rate of red cell replacement in the animals used in the study

Rate of red cell replacement in the animals used in the study
Rate of red cell replacement in the animals used in the study
Fig. 5.

Light-microscope photograph of cells from adult Xenopus liver, 12 days after induction of anaemia by phenylhydrazine. Cells were obtained as a tissue-touch imprint, and were fixed and stained by the May-Grunwald/Giemsa technique. A number of the large erythroid cells are identified by this technique as being basophilic erythroblasts (arrowed), × 500.

Fig. 5.

Light-microscope photograph of cells from adult Xenopus liver, 12 days after induction of anaemia by phenylhydrazine. Cells were obtained as a tissue-touch imprint, and were fixed and stained by the May-Grunwald/Giemsa technique. A number of the large erythroid cells are identified by this technique as being basophilic erythroblasts (arrowed), × 500.

Fig. 6.

As for Fig. 5, but this tissue-touch imprint is from spleen. Again some basophilic erythroblasts are in evidence (arrowed), × 500.

Fig. 6.

As for Fig. 5, but this tissue-touch imprint is from spleen. Again some basophilic erythroblasts are in evidence (arrowed), × 500.

We believe that we can reliably distinguish both in the light and electron microscope, between basophilic and polychromatophilic erythroblasts and other cells also present in liver, spleen and blood. The basophilic erythroblast is a large cell as compared to the polychromatophilic erythroblast, both nuclear and cytoplasmic areas being visibly increased in cell sections, and the cytoplasm stains strongly with basophilic dyes. In electron-microscope section, they are identifiable by their abundant clusters of ribosomes and comparatively sparse endoplasmic reticulum (see Fig. 7). It therefore seems that, during recovery from anaemia, Xenopus laevis employs both spleen and liver in erythropoiesis, although the liver seems to be the more active site of the two (Thomas & Maclean, 1975). We have been unable to find evidence of red blood cell production in any other tissues.

Fig. 7.

Photograph of section of glutaraldehyde-fixed liver from adult Xenopus, 11 days after induction of anaemia by phenylhydrazine, viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining. The picture shows a cell which we identify as a basophilic erythroblast, as well as parts of mature erythrocytes. Notice in this cell the relatively uncondensed chromatin of the large nucleus, large mitochondria, and cytoplasm containing many clustered ribosomes but very little endoplasmic reticulum, × 10300.

Fig. 7.

Photograph of section of glutaraldehyde-fixed liver from adult Xenopus, 11 days after induction of anaemia by phenylhydrazine, viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining. The picture shows a cell which we identify as a basophilic erythroblast, as well as parts of mature erythrocytes. Notice in this cell the relatively uncondensed chromatin of the large nucleus, large mitochondria, and cytoplasm containing many clustered ribosomes but very little endoplasmic reticulum, × 10300.

Erythroblast mitosis in circulation

Basophilic erythroblasts are released into circulation by the liver and spleen some 6–10 days after phenylhydrazine treatment. Such immature cells are not detected in the blood at later dates, however, so the replenishment of the erythrocyte pool is not accomplished by a continuous release of such cells. Rather the shedding of erythroblasts appears to be a fairly sudden ‘one-off’ event and the recovery of erythroid cell numbers is largely by division of these new cells in circulation.

Between days 10 and 25 of anaemia the erythroid cell population increases from 1 × 108 to about 5 × 108/ml and matures from basophilic erythroblasts to late polychromatophilic erythroblasts/reticulocyte type cells. (A normal healthy adult has a blood cell concentration close to 7 × 107 cells/ml.) This increase could clearly be accomplished by 2–3 rounds of mitosis by the circulating erythroblasts (see Fig. 9).

Fig. 8.

Photograph of section of glutaraldehyde-fixed spleen from adult Xenopus, 11 days after induction of anaemia by phenylhydrazine viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining The picture shows a group of cells, 2 of which we identify as basophilic erythroblasts (arrowed). Parts of mature erythrocytes are also visible, darkly stained, as well as other types of spleen cells As in the liver section, basophilic erythroblasts are recognizable by their large nucleus with rather uncondensed chromatin, and a cytoplasm densely filled with clusters of ribosomes but with little endoplasmic reticulum Some of the vesicles present in the cytoplasm of these cells may be the remains of mitochondria, since there is a dramatic reduction in the mitochondrial population at this stage of erythroid cell differentiation, × 3900.

Fig. 8.

Photograph of section of glutaraldehyde-fixed spleen from adult Xenopus, 11 days after induction of anaemia by phenylhydrazine viewed in the electron microscope after Spurr embedding and uranyl acetate/lead citrate staining The picture shows a group of cells, 2 of which we identify as basophilic erythroblasts (arrowed). Parts of mature erythrocytes are also visible, darkly stained, as well as other types of spleen cells As in the liver section, basophilic erythroblasts are recognizable by their large nucleus with rather uncondensed chromatin, and a cytoplasm densely filled with clusters of ribosomes but with little endoplasmic reticulum Some of the vesicles present in the cytoplasm of these cells may be the remains of mitochondria, since there is a dramatic reduction in the mitochondrial population at this stage of erythroid cell differentiation, × 3900.

Fig. 9.

Electron-microscope photograph of blood cells of adult Xenopus 15 days after the induction of anaemia by phenylhydrazine. Most of the cells are at the ‘reticulocyte’ stage of maturation to the erythrocyte. The central cell is in mitosis and sections of paired chromatids are clearly visible. This probably represents the last mitosis of this cell, since its cytoplasm is already indicative of a late erythroblast stage. × 3500.

Fig. 9.

Electron-microscope photograph of blood cells of adult Xenopus 15 days after the induction of anaemia by phenylhydrazine. Most of the cells are at the ‘reticulocyte’ stage of maturation to the erythrocyte. The central cell is in mitosis and sections of paired chromatids are clearly visible. This probably represents the last mitosis of this cell, since its cytoplasm is already indicative of a late erythroblast stage. × 3500.

That division is occurring in circulating cells is shown by the incorporation of [3H]thymidine into DNA following in vivo administration or in vitro culture. Injection of [3H]thymidine into anaemic frogs (at day 13) results in over 40% of the circulating cells becoming labelled in autoradiographs. See Fig. 10. [3H]thymidine is rapidly cleared from circulation, certainly within 24 h (Thomas, 1975), so this is, in fact, a type of ‘pulse-chase’ experiment. The fate of the labelled cells was followed by bleeding from the footweb on subsequent days. From the distribution of grains/cell it was concluded that: (1) there was no further recruitment of labelled or unlabelled cells from the erythropoietic tissues after this time; but (2) there was a shift towards the lower grain count, presumably as a result of division of the labelled cells (see Fig. 11). The approximate rate of division of the cells over this period was estimated from the rate of loss of label from highly labelled cells. In 3 separate experiments the results obtained were:

Fig. 10.

Autoradiograph of Xenopus blood cells, harvested from the foot web 24 h after 0·1 mCi of tritiated thymidine had been injected into the dorsal lymph sac. When, injected, the animal was anaemic, having been administered with phenylhydrazine 13 days previously. Cells have been prepared as described in Materials and methods and photographed under phase contrast. Over 40 % of the cells are invariably labelled in this procedure. In this photograph an unusually high percentage are labelled, × 650.

Fig. 10.

Autoradiograph of Xenopus blood cells, harvested from the foot web 24 h after 0·1 mCi of tritiated thymidine had been injected into the dorsal lymph sac. When, injected, the animal was anaemic, having been administered with phenylhydrazine 13 days previously. Cells have been prepared as described in Materials and methods and photographed under phase contrast. Over 40 % of the cells are invariably labelled in this procedure. In this photograph an unusually high percentage are labelled, × 650.

Fig. 11.

Histogram plotting the effect of time on autoradiographic grain count in circulating erythroid cells of anaemic Xenopus following in vivo [3H]thymidine pulse label. The data show the shift in labelling intensity from day 1 (continuous line) to day 5 (interrupted line) after administration of label. Notice that the scale on the vertical axis is not continuous between 100 and 200 labelled cells.

Fig. 11.

Histogram plotting the effect of time on autoradiographic grain count in circulating erythroid cells of anaemic Xenopus following in vivo [3H]thymidine pulse label. The data show the shift in labelling intensity from day 1 (continuous line) to day 5 (interrupted line) after administration of label. Notice that the scale on the vertical axis is not continuous between 100 and 200 labelled cells.

formula

This fits very well with 2–3 divisions between days 13 and 25.

We had hoped to derive evidence from our autoradiographs about the synchrony of DNA synthesis or mitosis in the circulating cells, but the data fit best to a wave of cells released over a few days, and we cannot detect a non-random distribution of synthesis.

Cell number and haemoglobin content

As shown in Figs. 12, 13, we have scored recovering animals for blood cell number and blood cell haemoglobin concentration between days 12 and 150. Prior to day 12 the old erythrocytes interfere with the cell count and haemoglobin determinations. These figures illustrate that animals kept at a higher temperature range, 16-20 °C, recover more rapidly than animals kept between 10 and 14 °C, It is also interesting to notice that, at the higher temperatures, recovering animals at day 50 have overshot their normal erythroid cell number, but the cells are severely deficient in haemoglobin. These observations underline the fact that the population of erythroid cells released in response to severe induced anaemia are abnormal and that a return to correct levels is not recorded until 140 days after phenylhydrazine injection, a time at which fresh mature erythrocytes are probably being released into circulation and the cohort of anaemia-released cells withdrawn.

Fig. 12.

Peripheral blood cell count during recovery from phenylhydrazine-induced anaemia. Points are drawn through the mean counts from at least four animals, and standard deviations from the means are represented by vertical bars. Recovering animals were maintained within 2 separate temperature ranges (A, 16–20 °C, B, 10—14 °C), although at the lower temperatures animals were scored only up to 100 days. It will be seen that animals retained at a higher temperature recover more rapidly from anaemia, and also that there is some evidence of an overshoot in cell numbers between days 50 and 100 in these animals. The counts at 140 days represent the normal distribution for a healthy Xenopus.

Fig. 12.

Peripheral blood cell count during recovery from phenylhydrazine-induced anaemia. Points are drawn through the mean counts from at least four animals, and standard deviations from the means are represented by vertical bars. Recovering animals were maintained within 2 separate temperature ranges (A, 16–20 °C, B, 10—14 °C), although at the lower temperatures animals were scored only up to 100 days. It will be seen that animals retained at a higher temperature recover more rapidly from anaemia, and also that there is some evidence of an overshoot in cell numbers between days 50 and 100 in these animals. The counts at 140 days represent the normal distribution for a healthy Xenopus.

Both light and electron microscopy of spleen and liver tissue indicate that both organs are active in blood cell formation following anaemia, and presumably both were originally involved in releasing basophilic erythroblasts into circulation in order to restore the blood cell population. We cannot, of course, be certain that both organs are active in erythropoiesis in the normal healthy adult. Normal blood cell production is very slow in Xenopus and therefore evidence of erythropoietic activity is never dramatic. The liver is certainly much more active than the spleen in erythropoiesis following anaemia. Evidence for erythropoiesis in other amphibians compares interestingly with Xenopus, since it is believed to be purely a splenic and circulatory phenomenon in the newt Triturus cristatus (Grasso, 1973), but to occur in liver and bone marrow in the bullfrog Rana catesbiana (Maniatis & Ingram, 1971). We have never found evidence of erythropoietic activity in Xenopus bone marrow.

As with blood cell production, the destruction of such cells is much easier to follow in the experimentally anaemic animal and here both spleen and liver play a vigorous role. We have not so far been able to record sites of red cell destruction in normal healthy Xenopus.

Fig. 13. Haemoglobin concentration in the circulating cells. Points are drawn through mean counts from at least four animals, and standard deviation from the means are also represented. Recovering animals were maintained within 2 separate temperature ranges (A, 16–20 °C, B, 10-14 °C), although at the lower temperature animals were scored only up to 100 days. It will be seen that animals retained at a higher temperature have a somewhat higher concentration of haemoglobin in their cells, no doubt resulting from the more rapid erythroid differentiation et this temperature range. The haemoglobin concentration at 140 days represents the normal range for a healthy Xenopus.

The results of blood cell counts and autoradiographs following anaemia confirm the earlier conclusion (Thomas & Maclean, 1974) that the blood is a site of active blood cell production during anaemia. Such production is not, we now believe, from true stem cells, but more probably from basophilic erythroblasts released from liver and spleen after the induction of anaemia. Our results strongly suggest that dramatic anaemia in Xenopus is answered by a ‘panic’ response from the erythropoietic tissues, involving a sudden release of a population of very early erythroid cells in circulation. If an already anaemic Xenopus is given another dose of phenylhydrazine, the treatment is invariably fatal within 10 days (Maclean, unpublished), presumably due to the abolition of the entire erythropoietic potential. Such a release seems to denude the liver and spleen of their short-term erythropoietic potential, so that normal red blood cell numbers are reached by the division of the initial cohort in circulation and not by further release from hepatic or splenic sources. These erythropoietic organs are, at this time, involved in recouping their own erythropoietic potential by erythroblast division and maturation rather than release of erythroid cells. In the non-anaemic adult we find no sign of erythroblasts in circulation, and presume that probably both liver and spleen serve as sites of erythroid cell production and maturation, the red cells being released as early erythrocytes.

We should also point out that the rate of recovery of our anaemic Xenopus is somewhat faster than that previously reported (Thomas & Maclean, 1975). This we believe to be a result of the time of year at whch the experiments were undertaken. Since the Amphibia are poikilothermic, it is certain that the ambient temperature will profoundly affect their rate of recovery from anaemia.

We are grateful to the Medical Research Council for partial support of this work.

Aleporou
,
V.
&
Maclean
,
N.
(
1978
).
Feedback inhibition of erythropoiesis induced in anaemic Xenopus
.
J. Embryol. exp. Morph
.
43
,
221
231
.
Grasso
,
J. A.
(
1973
).
Erythropoiesis in the newt Triturus cristatus. 1. Identification of the erythroid precursor cells
.
J. Cell Sci
.
12
,
463
489
.
Hilder
,
V. A.
,
Thomas
,
N.
&
Maclean
,
N.
(
1975
).
The erythroid cells of anaemic Xenopus laevis. II. Studies on nuclear non-histone proteins
.
J. Cell Sci
.
19
,
521
527
.
Maclean
,
N.
&
Jurd
,
R. D.
(
1971
).
The haemoglobins of healthy and anaemic Xenopus laevis
.
J. Cell Sci
.
9
,
509
528
.
Maniatis
,
G. H.
&
Ingram
,
V. M.
(
1971
).
Erythropoiesis during amphibian metamorphosis. I. Site of maturation of erythrocytes in Rana catesbiana
.
J. Cell Biol
.
49
,
372
379
.
Reynolds
,
E. S.
(
1963
).
The use of lead citrate at high pH as an electron-opaque stain in electron microscopy
.
J. Cell Biol
.
17
,
208
212
.
Singer
,
K.
,
Chernoff
,
A. I.
&
Singer
,
L.
(
1951
).
Specific stains for haemoglobin
.
Blood
6
,
413
435
.
Thomas
,
N.
(
1975
).
Erythropoiesis in Xenopus laevis During Recovery from Phenylhydrazine Induced Anaemia
.
Ph.D. Thesis
,
Southampton University
.
Thomas
,
N.
&
Maclean
,
N.
(
1974
).
The blood as an erythropoietic organ in anaemic Xenopus
.
Experientia
30
,
1083
1085
.
Thomas
,
N.
&
Maclean
,
N.
(
1975
).
The erythroid cells of anaemic Xenopus laevis. 1. Studies on cellular morphology and protein and nucleic acid synthesis during differentiation
.
J. Cell Sci
.
19
,
509
520
.
Wallet
,
L. H. B
&
Robinson
,
J. B.
(
1968
).
Human Haemoglobins and Their Identification
.
London
:
Butterworth
.