Red Cells In Saline Research Articles

Red cell substitutes from hemoglobin—Do we start all over again?

Red cells are the oxygen-carrying components of blood. In modern medical practice, transfusions are given as suspensions of type-matched red cells in saline to replace lost blood, preventing organ damage and allowing for recovery. Since red cells cannot be stored for more than about 40 days and because they can transmit infections, alternative materials for transfusions were developed to replace the oxygenation function of the red cells. One approach involves chemically stabilizing hemoglobin, the oxygen-carrying protein of the red cell, while also adjusting its oxygenation properties to replicate that of the red cell. Evaluation of clinical trials of all products led to the conclusion that none that were tested would be suitable for clinical use [Natanson C, Kern SJ, Lurie P, Banks SM, Wolfe SM: Cell-free hemoglobin-based blood substitutes and risk of myocardial infarction and death: a meta-analysis. J Am Med Assoc 2008, 299:2304-2312]. Most notably, the materials increased blood pressure and some were associated with increased risk of heart attacks. More recently, it was found that materials from covalent addition of polyethylene glycol polymers (PEG) to hemoglobin do not elicit the undesired effects on blood pressure [Vandegriff K, Bellelli A, Samaja M, Malavalli A, Brunori M, Winslow RM: Rates of NO binding to MP4, a non-hypertensive polyethylene glycol-conjugated hemoglobin. FASEB J 2003, 17:A183; Vandegriff KD, Malavalli A, Wooldridge J, Lohman J, Winslow RM: MP4: a new nonvasoactive PEG-Hb conjugate. Transfusion 2003, 43:509-516]. Also, materials with higher oxygen affinity than red cells are able to provide oxygenation at the sites in capillaries that have the most critical need for oxygen [Villela NR, Cabrales P, Tsai AG, Intaglietta M: Microcirculatory effects of changing blood hemoglobin oxygen affinity during hemorrhagic shock resuscitation in an experimental model. Shock 2009, 31:645-652]. It had been considered that the origin of the negative effects of the tested hemoglobin derivatives was because of their scavenging of endogenous nitric oxide (NO), the signal for vasodilation. It has been observed that an increase in the concentration of nitrite in circulation leads to an increase in NO concentration. This is consistent with the well-known reaction of hemoglobin with nitrite that produces NO and oxidized hemoglobin [Cannon RO 3rd, Schechter AN, Panza JA, Ognibene FP, Pease-Fye ME, Waclawiw MA, Shelhamer JH, Gladwin MT: Effects of inhaled nitric oxide on regional blood flow are consistent with intravascular nitric oxide delivery. J Clin Invest 2001, 108:279-287; Cosby K, Partovi KS, Crawford JH, Patel RP, Reiter CD, Martyr S, Yang BK, Waclawiw MA, Zalos G, Xu X, et al.: Nitrite reduction to nitric oxide by deoxyhemoglobin vasodilates the human circulation. Nat Med 2003, 9:1498-1505]. The PEG-hemoglobin and nitrite results are especially interesting as the hemoglobin to which PEG has been conjugated produces NO from nitrite at an enhanced rate [Lui FE, Dong P, Kluger R: Polyethylene glycol conjugation enhances the nitrite reductase activity of native and cross-linked hemoglobin. Biochemistry 2008, 47:10773-10780; Lui FE, Kluger R: Enhancing nitrite reductase activity of modified hemoglobin: bis-tetramers and their PEGylated derivatives. Biochemistry 2009, 48:11912-11919].

SIFT-MS measurement of VOC distribution coefficients in human blood constituents and urine.

The new technique of selected ion flow tube-mass spectrometry (SIFT-MS) has been applied to the measurement of Henry's Law constants for the volatile organic chemicals o-xylene and trichloroethylene that both have low solubility in aqueous solvents. The method is validated by measurements in water at 298 K using the equilibrium partitioning in closed systems (EPICS) methodology in which the equilibrium headspace concentrations for the volatile organic compounds (VOCs) are measured in two sealed bottles containing different liquid volumes of very dilute solutions of the VOC. The range of solvents is then extended to human body fluids at 309 K including urine, saline, whole blood, red cells in saline, and plasma. The dimensionless distribution coefficients for these solvents vary markedly in the different fluids. For o-xylene they range from k(H) = 0.12-0.15 for water, saline, and urine; 0.53 for red cells in saline; 1.9 for whole blood; to 2.4 for plasma. For trichloroethylene the distribution coefficients range from k(H) = 0.070-0.091 for water, saline, and urine; 0.28 for red cells in saline; 0.35 for plasma; to 0.48 in whole blood. The very different solubilities of organic solvents in body fluids influence the uptake of solvents in workers exposed to VOCs. Some implications of these measurements are briefly discussed.

The effect of hemolysis on acoustic scattering from blood

In an attempt to develop a direct method for measuring the extent of red cell damage in vitro, the effect of the degree of hemolysis on ultrasonic scattering from blood was investigated. Starting with a suspension of 30% hematocrit, a series of suspensions containing different relative concentrations of healthy and damaged red cells in saline were prepared, with the total number of cells present in any one suspension being constant. For each sample, a suspension of equal concentration of healthy cells, but no lyzed cells, was also produced. Using a specially designed container, all samples were exposed to 15 MHz ultrasound in pulse‐echo mode and measurements of backscattering were obtained. At high hematocrits, the samples containing damaged cells were found to scatter substantially more than the suspensions containing exclusively healthy cells. This indicates that damaged cells contribute significantly to the overall backscattered intensity. Below a concentration of 13% per volume of healthy cells, scattering levels from healthy and hemolyzed suspensions were comparable. A theoretical model, which treats healthy cells as weak‐scattering spheres and damaged cells as hard thin disks, is proposed to interpret the observed scattering behavior.

Commentary on the safety of red cells preserved in extended‐storage media for neonatal transfusions

Red cells preserved in extended-storage media are the standard product dispensed by many regional blood centers. When the red cells are intended for neonatal transfusion, concern exists about the safety of the relatively high quantities of additives present in these media. Definitive studies to address these concerns are not available. Therefore, to estimate the effects of additives and to delineate circumstances in which they might be harmful, the quantities transfused in defined clinical settings were calculated, and the following recommendations are offered for transfusing infants less than 4 months of age. First, red cells preserved in extended-storage media should present no substantive risks when used for small-volume (approximately 10 mL/kg) transfusions of premature infants and can be used without additional processing. Second, the risks of the most premature neonatal patients or those with severe renal and/or hepatic insufficiency cannot be defined clearly, and, because data are not available to ensure safety for these infants, removal of the additive medium and resuspension of the red cells in saline or albumin solution immediately before transfusion are recommended. Third, following a similar rationale, it seems prudent to avoid using entire units of red cells preserved in extended-storage media in massive transfusion settings (e.g., exchange transfusion, cardiac surgery, and extracorporeal membrane oxygenation). In these settings, the preservative medium should be removed and the red cells resuspended in the fluid that is most appropriate for the procedure that is planned. It must be emphasized that these recommendations are based on calculations and hypothetical settings, not actual data. Accordingly, they are tentative and should be altered as definitive information becomes available.

Supernatant K+ and in vitro haemolysis of washed red blood cells stored in saline for 3 days.

Washed Red Blood Cells (WR8C) is allowed a shelf-life of 24 hours. This study is undertaken to determine whether 1) the K⁺ level in the supernatant at 24 hours is acceptably-low for neonatal transfusion, and 2) the preservation of red cells in saline, as measured by in vitro haemolysis can be extended to beyond 24 hours. 22 units of donor blood of various storage age washed manually, and 4 units washea by Haemonetics were studied. Samples taken from each were divided into aliquots and tested for K⁺ level (mmol/1) and % loss of haemoglobin (Hb) in the supernatant at 6, 24, 48 and 72 hours. Results were as follows:red cellshours of storageage (n)6244872Manual wash4-7 daysK0.74.89.213.1(7)Hg1.41.72.12.514 daysK0.84.27.610.6(7)Hg2.43.34.04.321 daysK0.94.06.58.0(8)Hg2.64.04.65.0Haemonetics6-12 daysK0.20.83.9/(4)Hg1.32.63.8/ Although the supernatant K⁺ was acceptably low, in vitro haemolysis was excessive compared to blood stored in standard preservatives. We propose in vivo survival study for WRBC after 24 hours of storage in saline; alternative measures such as additives to WRBC immediate postwash for better preservation.

Nonautocatalytic methemoglobin formation by sodium nitrite under aerobic and anaerobic conditions

Aerobic exposure of human hemoglobin in 1:1 molar ratio to sodium nitrite induced an immediate and progressive formation of methemoglobin. Neither lag nor autocatalytic phases were observed, contrary to what was seen when diluted hemoglobin in buffered solutions was used. Hemolyzed whole blood, washed red cells with their plasma replaced by saline, and whole blood were exposed to identical sodium nitrite concentrations. In the presence of oxygen, the speed of methemoglobin formation was three to four times faster in the hemolyzed and plasma-free samples than in whole blood. The maximum rate was reached within the first minute of the reaction. Restoring plasma to washed red cells reduced the rate to that of whole blood; restoring plasma to hemolyzed red cells had no effect on rate. We advance the hypothesis that sodium nitrite or its derivatives form a complex with some element of the blood plasma which slows the passage of nitrite through the red cell membrane without hindering the oxidative capacity of the nitrite. With no free oxygen present, whole blood, washed red cells in saline, and hemolyzed blood all showed a similar rate of methemoglobin formation when exposed to a 1:1 molar ratio of sodium nitrite. The presence of nitrosylhemoglobin formed in the anaerobic reaction seems to be a for the lack of changes seen in the reaction.

An Autoantibody with Activity Dependent on Red Cell Age in the Serum of a Patient with Autoimmune Haemolytic Anaemia and a Negative Direct Antiglobulin Test

A red cell IgM autoantibody with anti-e specificity was identified in the serum of a rhesus-negative (rr) patient presenting with haemolytic anaemia and a negative direct antiglobulin test. The autoantibody strongly agglutinated normal allogeneic rhesus-negative (rr) red cells in saline at 37°C but had only weak activity for autologous red cells. Incubation of the patient’s serum with subpopulations of normal allogeneic red cells obtained by density fractionation, demonstrated that the strong agglutinating activity of the autoantibody was for red cells with density greater than 1.090 g/ml. Young red cell subpopulations of lower density gave weak reactions and low titration scores equivalent to those obtained with autologous red cells during the haemolytic episode. The patient’s red cells during remission however, when the patient’s haemoglobin level had returned to normal, were strongly agglutinated by serum samples taken during the haemolytic episode; as was the case with normal allogeneic red cells, the strong activity was for patient red cells with density greater than 1.090 g/ml, red cell populations of lower density giving low titration scores. The findings in this case indicate that the patient’s red cells which had survived haemolysis during the haemolytic episode were young red cells (density <1.090 g/ml), the weak sensitization of these cells being insufficient to cause their destruction by macrophages. Furthermore, these findings, together with observations that IgG autoantibodies may also bind less strongly to young red cells [Gray et al., Br. J. Haemat., 55: 335-345, 1983; Branch et al., Blood 63: 177-180, 1984], imply that quantitative assays of cell-bound antibody during haemolytic episodes may provide an erroneous measure of the red cell antibody load required to effect in vivo haemolysis.

Role of heparin in red cell aggregation

One of the chief functions of hyaluronic acid (HUA) and soluble protein~chrondroitinkeratan,sulfate (PCKS) in the animal organism is to restrict dispersion and concentration of cells, their organelles, and certain other tissue components in the necessary volume [3, 5, 8, 11-14]. Investigations using suspensions of rabbit red cells in saline as the model of isolated cells have shown that heparin (the total preparation) and its two individual fractions, one of which contains three (HP-3) the other four (HP-4) sulfate groups for each re. peated dissacharide unit of the macromolecules, weaken or totally suppress this action of HUA and PCKS in a concentration of I0 mg.ml-* [3].

The accuracy of Doppler ultrasound measurement of pressure gradients across irregular, dual, and tunnellike obstructions to blood flow.

The accuracy of Doppler-estimated pressure gradients in the setting of irregular, multiple, and tunnellike stenoses was investigated. An in vitro model of the left ventricular outflow tract was designed to allow pulsatile flow of red cells in saline across valve orifices from 0.01 to 2.5 cm2. Simultaneous pressure gradients were estimated by both Doppler and direct-pressure manometer techniques. Gradients obtained by the two methods correlated well for valve areas in the range of clinical stenoses at pressure gradients of 10 to 150 mm Hg (r = .97 to .99). Model valves were constructed with a large orifice (0.75 to 1.25 cm2) placed beside a small orifice (0.02 to 0.25 cm2) in the same outflow tract. A distinct jet was recorded when the Doppler transducer was aligned with each orifice. Doppler-estimated gradients for each pair of large and small orifices were identical and correlated well with those measured by manometer (r = .97 to .99). Irregularly shaped orifices also provided good correlation between the two methods (r = .98 to .99). Pulsatile flow was generated through long tunnellike obstructions with cross-sectional areas varying from 0.06 to 1.25 cm2. Tunnel length varied from 0.1 to 4 cm. Tunnel areas above 0.25 cm2 gave good Doppler-to-manometer correspondence at all tunnel lengths. Doppler underestimated manometer-determined values in the 0.25 cm2 tunnel by 8% at 3 cm and by 15% at 4 cm. In the 0.06 cm2 tunnel, Doppler underestimated manometer gradients by 12%, 15%, 32%, and 42% at lengths of 1, 2, 3, and 4 cm, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

Evaluation of the pulmonary immune response by analysis of bronchoalveolar fluids obtained by serial lung lavage.

Lavage fluids were obtained from Beagle dogs by serial bronchoalveolar lavage from immunized and control lung lobes 3-21 days after local deposition of sheep red blood cells. Results showed that immunized lobes contained more total protein, albumin, IgG and antigen-specific antibody than did unimmunized control lobes or lobes that received autologous red cells in saline. Antigen-specific antibody was present in the lavage fluid from 7 through 21 days after immunization. Mercaptoethanol treatment of lavage fluid had little effect on the antibody titer, suggesting that IgG, and possibly IgA, were the predominant immunoglobulin classes in the antigen-specific antibody pool throughout the course of the serial lavage study. The data suggest that most of the antigen-specific antibody was made locally by cells in the lung.

Fish erythrocytes are bicarbonate permeable: problems with determining carbonic anhydrase activity using the modified boat technique.

Effects of foaming agents (blood plasma, bovine serum albumin, polyvinyl-pyrrolidinone) and defoaming agents (octanol, n-foam) on manometric determination of carbonic anhydrase activity in intact erythrocytes of fish were examined. Foaming agents abolished the activity of red cells. Defoaming agents increased the activity and negated any differences in carbonic anhydrase activity between whole blood (red cells in plasma) and "Cortland" blood (red cells in saline). It is concluded that effects of plasma on the carbonic anhydrase activity of intact erythrocytes of fish measured using the modified boat technique are largely nonspecific and related to reagent foaming during mixing, rather than to the action of a specific enzyme inhibitor.

EFFECT OF PLASMA PROTEINS AND TEMPERATURE ON ECHOGENICITY OF BLOOD

We have previously described echogenicity of unclotted blood during high resolution B-mode ultrasonic scanning. This occured during stasis and was less evident after agitation or in flowing blood. To determine more precisely the factors associated with echogenicity, a series of in-vitro experiments were performed using blood from normal subjects and patients with multiple myeloma. Whole blood and combinations of red cells with fibrinogen, gelatin, saline and serum were examined with real-time B-mode and A-mode ultrasound using 7.5 or 10 MHz transducers at temperatures of 0°, 25° and 37°C. RESULTS: Normal whole blood was echogenic but blood from patients with myeloma was more echogenic than normal blood. Myeloma blood showed increased rouleaux formation on smear. When fibrinogen and washed red cells were used, echogenicity increased with greater concentrations of fibrinogen. The action of fibrinogen was not specific. Gelatin added to red cells also produced an echogenic suspension while red cells in saline or serum were not echogenic. These experiments were done at room temperature (22° C). Compared to echogenicity at room temperature, whole blood was markedly less echogenic at 0° C and more echogenic at 37° C. CONCLUSIONS: Echogenicity of unclotted blood is probably due to reflection of ultrasound from red cell aggregates. The most prevalent form of aggregation is rouleaux, the formation of which increases in the presence of asymmetric molecules such as fibrinogen, gamma globulin or gelatin. This probably explains the greater than normal echogenicity in myeloma blood and the direct relation between concentrations of fibrinogen and gelatin and the amount of echogenicity. The effect of temperature on echogenicity may be related to the interaction between plasma proteins and red cells in the production of red cell aggregation.

Erythrocyte damage caused by the Haemotherm microwave blood warmer.

Blood units (59) were warmed with the Haemotherm microwave blood warmer and seven units were warmed in a water bath for comparison. The influence of the final blood temperature, the size and the hematocrit of the units on the erythrocyte lesion was studied. Extracellular hemoglobin and potassium, hematocrit, osmotic fragility and mean cellular volume were used as indicators of red cell damage. Warming in the water bath caused no erythrocyte damage at temperatures below +46.8 degrees C. Higher temperatures caused progressive morphological changes and hemolysis. Units of red cells in saline warmed to temperatures above +46.3 degrees C with the Haemotherm showed intense hemolysis and changes in all parameters used. Below this temperature no erythrocyte damage occurred. When the amount of blood warmed was less than 300 g or when the hematocrit exceeded 0.70, the blood mixing mechanism became insufficient, leading to local overheating and hemolysis. Therefore, erythrocyte concentrates in a Fenwal blood bag should never be warmed by the Haemotherm. The mean plasma hemoglobin increase of the ten whole blood units warmed by the Haemotherm to +36.0-36.8 degrees C was 123 mg/l. The results of the present study indicate that microwaves per se are not harmful to erythrocytes but that poor penetrance of microwaves, together with insufficient blood mixing during warming, are the critical factors leading to hemolysis.

Conversion of vitamin B6 compounds to active forms in the red blood cell

In studies with pyridoxine and other B(6) compounds in blood, the active forms pyridoxal and pyridoxal phosphate were measured by differential assays using Lactobacillus casei. Red cell uptake of tritiated pyridoxine was also measured. A new metabolic pathway for conversion of pyridoxine to active forms was demonstrated in red cells. In vivo studies in normal subjects suggested that pyridoxine was taken up by red cells where it was converted to pyridoxal phosphate and then pyridoxal, followed by gradual release of a proportion of pyridoxal into plasma. In vitro incubation of pyridoxine with blood confirmed this observation. Increasing amounts of pyridoxine were taken up and converted as the amount added to blood was increased, and only very small numbers of red cells were needed to convert appreciable amounts. Conversion was markedly inhibited at temperatures lower than 37 degrees C, and stopped altogether at - 20 degrees C.Release of pyridoxal into plasma was always directly proportional to the amount of pyridoxal formed and to the volume of plasma present. That pyridoxal phosphate was not released into plasma was demonstrated in stored blood, for pyridoxine was converted mainly only as far as pyridoxal phosphate, probably due to inactivation of the phosphatase. Pyridoxal phosphate remained in the red cells. Pyridoxine was converted when incubated with washed red cells in saline or phosphate buffer suspension (0.08 M). In saline suspension, pyridoxal formed but was not released in the absence of plasma. In phosphate buffer suspension, pyridoxal phosphate was formed but was not changed to pyridoxal, probably due to inactivation of phosphatase by excess phosphate. Pyridoxamine was converted to active forms in red cells less efficiently. Pyridoxal entered red cells rapidly, equilibrating between plasma and cells within 1 min in the same ratio as pyridoxal formed inside red cells. Pyridoxal phosphate did not enter red cells in whole blood but did so readily in washed cells in saline.

Study of Erythrocyte Aggregation by Blood Viscometry at Low Shear Rates Using a Balance Method

Abalance method was used to investigate the viscometric behavior and aggregation properties of blood at very low shear rates. The system was essentially a parallel-plate viscometer, consisting of a thin brass plate suspended by a fine wire from one end of an analytical beam balance. The yield shear stress of a sample was obtained directly by recording stress relaxation curves. Results were obtained for homogeneous Newtonian and non-Newtonian fluids and for blood and various red cell suspensions. It was found that human and canine blood behave as fluids exhibiting a yield stress. Suspensions of red cells in saline had much smaller values of yield stress, which is consistent with greatly diminished erythrocyte aggregation when fibrinogen is lacking. Measured values of the yield stress were in the range of 0.001 to 0.1 dynes/cm 2 .

Use of Formalin-Treated Red Cells for the Study of Influenza A Virus Hemagglutinating Activity

Human erythrocytes, treated with formalin, the writer has found, retain their ability to react with influenza A virus in a manner quite similar to that of untreated human erythrocytes. For certain types of experiments the use of such formalized red cells possesses distinct advantages over the use of normal red cells.To prepare formalized red cells, equal volumes of a 50% (by volume) suspension of washed red cells in saline and formalin containing sodium chloride in a concentration of 0.85%, were mixed and allowed to stand for 2 to 3 days. If the formalin concentration of the mixture was between 10 and 50%, the cells became dark brown gradually over a period of 24 hours or so. The cells also became cohesive and very resistant to hemolytic agents. The stability of the cells thus treated was marked. Samples of formalized cells stored at 4°C have retained their activity undiminished for influenza A virus for almost two years now. Furthermore, the cells have retained the microscopic appearance and shape of fre...

PROTEIN METABOLISM AND EXCHANGE AS INFLUENCED BY CONSTRICTION OF THE VENA CAVA

Constriction of inferior vena cava above the diaphragm is used to produce experimental ascites in the dog. This type of experimental ascites drains the body protein reserves, reduces the level of circulating plasma proteins, and in effect is an internal plasmapheresis. As the ascitic fluid is withdrawn and the proteins measured, we observe a production of ascitic protein (80-90 gm. per week) comparable to that removed by plasmapheresis (bleeding and replacement of red cells in saline). High protein diet tends to decrease the ascites but the protein content of the ascitic fluid may increase. Sodium chloride increases notably the volume of the ascites which accumulates and the total ascitic protein output increases. Sodium-free salt mixtures have a negative influence. High protein diet low in sodium salts gives minimal ascitic accumulation under these conditions. The question of circulation of the ascitic fluid is raised-how rapid is the absorption and the related accumulation?

The Spherical form of the Mammalian Erythrocyte

ABSTRACT The most suggestive observation which we have regarding the remarkable disk-sphere transformation which occurs when mammalian red cells in saline are enclosed between slide and coverslip is that the velocity and extent of the shape change depends on the distance between the surfaces, and upon whether or not both surfaces are wetted by the suspension medium (Ponder, 1929). Thus the transformation occurs rapidly and completely between two glass surfaces 50μ. apart, but slowly and incompletely between two glass surfaces 1 mm. apart, and paraffining of both surfaces, or the substitution of surfaces unwetted by saline, prevents the shape change completely. That the transformation is due to electrostatic charges on the surfaces, pressure,2 or diffusion of alkali from the glass, has been disproved experimentally (Ponder, 1929, 1936), and the only suggestion which still remains is that the surfaces have an orienting effect, of an obscure nature, on a “liquidcrystal” structure in the red-cell envelope.

On the Spherical Form of the Mammalian Erythrocyte

ABSTRACT In 1895 Hamburger described a remarkable phenomenon which can be observed in suspensions of mammalian red cells in isotonic, hypotonic, or hypertonic saline and in sugar solutions, but not in serum or plasma. The normally biconcave discoidal cells become perfect spheres, without undergoing any apparent alteration in volume; the addition of serum to the suspension of spherical cells, however, converts the spheres back again into the typical discoidal form. Hamburger has no explanation to offer for the occurrence of this phenomenon, but suggests that it may be due to changes in surface tension. In 1920 Brinkman and van Dam published a study of the phenomenon, in which they record the changes undergone by the red cells of man and of the rabbit when placed in a haemocytometer chamber. They correct Hamburger’s statement in one important respect, for they observe that red cells in saline are not spherical but discoidal, and that they become spherical only when placed in the haemocyto-meter chamber. In becoming spheres they pass through an intermediate form, “the crab-apple form,” which is spherical but finely crenated. Brinkman and van Dam believe that the cause of the phenomenon is that the cells receive an electrostatic charge from the glass when they come into contact with the floor of the chamber. They confirm Hamburger’s statement that the change from disc to sphere is prevented by the presence of serum, and claim that it is the cholesterol contained in the serum which inhibits the change of form. Unaware of these two investigations, Gough in 1924 described the same phenomenon once gain, but fell into Hamburger’s error of believing that the red cells are spherical when suspended in a volume of saline. This error is very easy to make, and both Millar and I have made it also; indeed, as will be seen below, it is impossible to avoid it if an oil immersion objective is used for examining the cells. Gough adds two interesting observations to the previous descriptions: that ammonium oxalate and ammonium chloride act like serum in preventing the cells from becoming spherical in saline, and that the spherical form becomes crenated after a few hours, passing spontaneously into the discoidal form in about 24 hours. He does not give any detailed explanation of the phenomenon. Both Millar and myself (Millar 1925, Ponder 1925) made the same mistake as Gough and Hamburger, in believing that the spherical form is due to the immersion of the cells in saline and that cells are spherical when freely suspended in this medium. On such an assumption, I carried out many measurements of sedimentation velocity without receiving any indication that the assumption was wrong1. It appears, however, on a more detailed examination, that the essential condition under which the spherical form occurs has escaped all the observers who have worked on the subject, and that the phenomenon is much more complex than has hitherto been believed.