Idea Transcript
Complexometric Magnesium
Titration
of Calcium and
by a Semiautomated Procedure
James D. Jones and Warren F. McGuckin
A semiautomatic procedure for the direct estimation of both serum calcium and magnesium is described. The indicator used, Eriochrome blue SE (EBSE), has the advantage of stability and freedom from interference by naturally occurring substances. The procedure requires 3-5 mm. per sample and is very reproducible. The values obtained agree with those obtained by back-titration of excess chelator and are consistently 0.3 mg./100 ml. less than those obtained by a standard oxalate precipitation method.
C
OMPLEXOMETRIC TITRATION s of calcium and magnesium in serum with the aid of metallochromic indicators have been proposed by a number of investigators. The advantages over other methods are increased accuracy, speed, and simplicity. However, the detection of the end point is a potential source of variability and error. The indicators used are frequently unstable and can be masked by proteins and naturally occurring ions. The method for direct titration of calcium in serum with EDTA at pH 12.5 described in this report appears to be free of interference from other substances found in the serum. The indicator is stable, and the time required per sample is 3-5 mm. The results are very reproducible, since the detection of the end point is recorded automatically with respect to time. The distance, on the recorder chart, from start to extrapolated end point is proportional to the concentration of calcium in the sample. The results we have obtained are in general agreement with those in From the Section of Biochemistry, Mayo Clinic and Mayo Foundation, Rochester, Minn. The authors wish to express their appreciation to Philip Burnett and Victor Johnson for competent technical assistance and to Geigy Dyestuffs, Division of Geigy Chemical Corpora. tion, for generous gifts of Erioehrome blue SE and CHEL DE. Received for publication Sept. 3, 1963. 767
768
JONES & MCGUCKIN
other
reports
on coniplexometric
methods
Clinical
(1-4)
Chemistry
(0.3 mg. of calcium
per 100 mg. of serum less than determined by the modified Clark and (‘ollip procedure). The values are constant in samples of serum that are highly pigmelited, contam abnormal proteins, or are moderately lipemic. The determination by this method requires only one quantitative transfer, and results agree well with values obtained by backtitration of excess ethylene glycol i)is_($_amnloethyl ether)-N,N’tetraacetic acid (EG TA). This method has now replaced a modified Clark and Collip determmation for routine use in our laboratories and a preliminary study of serum calcium ill a mixed normal population has been initiated. An analogous procedure for magnesium also is described. Experimental Apparatus A titrator (Model SE*) with a Farrand interference filter (608 m, 15 m half bandwidth, 47% transmission) is coupled with a constantrate buret (Model C, 1 ml./min.*) and a logarithmic recorder (Model RL,* with 2_ill./min. chart speed). The chart drive-motor of the recorder is activated simultalwously with the buret. A 1-ml. syringe pipet for standards and samples and 30-ml. Pyrex beakers are the only glassware neede(l. Reagents All solutions, with the exception of the indicator, which is stored in glass, are stored in tightly closed polyethylene bottles. Deionized water is used for all solutions. Standard calcium solutions For the stock calcium standard, 2.4970 gm. of primary standard CaC0:, dried to constant weight at 1100, is transferred with 100 ml. of 1120, dissolved in 50-60 ml. of N HCi, and diluted to 1 L. with water, which gives a final concentration of i: mg. of calcium per milliliter. Standard magnesium solution Tile stock solution is 5.0678 gm. of MgSO4 7H.,O diluted to 500 ml. with water, which gives a final concentration of 1 mg. of Mg per milliliter. Working standards are prepared by diluting 75, 100, and 125 ml. of stock calcium and 20 ml. of stock magnesium solutions to 1 L. to give solutions containing 7.5, 10, and 12.5 mg. of calcium and 2 mg. of magnesium per 100 ml. Two ad-
brown
.
E. H. Sargent
& Company,
Chicago,
Ill.
vol.
10, No. 9, 1964
SEMIAUTOMATED
769
CA AND MG
ditiollal magnesium standards are prepared by diluting 10 and 30 ml. of stock magnesium solution, each with 100 ml. of stock calcium solution, to 1 L. to give solutions which contain 1 and 3 mg. of magnesium and 10 mg. of calcium per 100 ml., respectively. Other reagents The indicator is Eriochrome blue SE (EBSE) (Geigy Chemical), 0.380 gm./i00 ml. 1120. The titrating reagent is 0.9922 gm. of EDTA, (disodium ethylenediaminetetraacetate dihydrate, Baker Analyzed) per liter of water. Concentrated NaOH solution is diluted to 2.4 N. Buffer, pH 10.1, contains 67.5 gm. of reagentgrade NH4C1 and 570 ml. concentrated NH4OH (c.p.) per liter of 1120. Procedure Calcium
One milliliter of each calcium standard (7.5, 10, and 12.5 mg./100 ml., all with 2 mg. of magnesium per 100 ml.) is delivered to separate beakers, diluted to 15 ml. with deionized water, and mixed with 2 ml. of 2.4 N NaOH and 3 drops of indicator. The prepared sample is placed in position in the titrator (set at “spectro’’), the stirrer is activated, the recorder is set at “stalid by’’ turned to “CK” (check), and the pen is adjusted to 100 on the scale (extreme right) by use of the displacement knob. The recorder is set on ‘‘my’’ (millivolts) and the pen is adjusted to zero (extreme left) liv B
-A
Direct io I of pen
Fig.
use of the variable the button marked passed
(Fig.
range
knob.
“automatic”
1), both
and the analysis
1.
buret
is completed
and
Titratioti
curve.
titration tlmeti is started by turning on the buret. After the end point is
The
stirrer
by drawing
are
shut off, the pen is raised, one line Oil tile slope (A in
770
JONES & MCGUCKIN
Clinical
Chemistry
Fig. 1) and allother on the final line (B in Fig. 1) at the end of the titration. The distance, in centimeters, from the origin to the intercept of these two lines is proportional to the calcium concentration. The titration values, in centimeters, are plotted against the concentrations, in milligrams per 100 ml., of the standards (Fig. 2).
12
s-
s Calcium
o
0
Magnesium
10
0 0
4
/
/
2
/
/
/ / 3
5
4
6
7
8
9
cm. Fig.
2.
Standard
curve.
One-milliliter samples of serum are titrated in exactly the same manner as the standards. The serum calcium concentration is determined by interpolation from the graph prepared from standards (Fig. 2), or t.he concentration of EDTA can be adjusted so that the measured distance in centimeters, when multiplied by a suitable factor (we adjust the solution so that the factor is 2), gives the concentration of calcium in the serum in milligrams per 100 ml. Since a syringe pipet is used for standards and samples, a direct comparison is made; however, if samples containing less than 0.05 mg. of calcium are to be analyzed-for example, 0.5 ml. or less of serum-a standard curve including this concentration is necessary. An alternative procedure is to use an internal calcium standard. Delivery of air in lieu of titrant or failure to eliminate the “slack” in the chart drive mechanism of the recorder are the most common sources of error in the determination.
Vol. 10, No. 9, 1964
SEMIAUTOMATED
CA AND MG
771
Magnesium
One milliliter of each magnesium standard (1.0, 2.0, and 3.0 mg./100 ml., all with 10 mg. of caicium/100 ml.) is delivered to separate beakers, diluted to 15 ml., and mixed with 2 ml. of pH 10.1 buffer and 3 drops of indicator. The titration values obtained at pH 10.1 minus the value for a 10 mg./100 ml. calcium standard titrated at pH >12 are plotted, in centimeters, against the magnesium concentrations. The value for magnesium in a serum sample treated in a similar fashion is obtained by interpolation from the standard curve. Additional Procedures Measurements of pH were made with the Metrion 1)11 meter* with the Hyalk glass electrode. Trichloroacetic acid filtrates were prepared as described by Malmstadt and Hadjiioannou (5) and were titrated with the aid of 650-ms and yellow cutout filters. The back-titration procedure of Yarbo and Golby (6) was modified to allow detection of the end point with the titration assembly described. Titration of excess CHEL BEt (EGTA) at pH >12 was done with standard calcium solution and Calcoi4 as indicator for calcium, and titration of excess EDTA at pH 10.1 was done with standard calcium for total magnesium and calcium with Eriochrome black T as indicator. Serum was ashed by placing 1 ml. of serum and 2 ml. of a mixture of HNO3 and HC1O4 (concentrated acids, 1:1) in a 50-mi. Erlenmeyer flask and evaporating the mixture slowly to dryness on a hot plate. The ash was dissolved in 1 ml. of 0.5 N HC1 and the solution so obtained was used for analysis. In the modified Clark and Collip procedure used in this study, 2 ml. of serum, 2 ml. of 1120, and 1 ml. of 4% (NH4)2 C204 were mixed well and allowed to stand for 3 hr. at room temperature. The resulting precipitate, removed by centrifugation at 1325 RCF for 30 mm., was resuspended in 4 ml. of wash solution (327 ml. of EtOH, 327 ml. of diethyl ether, 327 ml. of 1120, and 27 ml. of cone. NH4OH) and recentrifuged. The entire washing procedure was repeated, the supernate was poured off, andthe precipitate, which was dried for 6 hr. at 55-60#{176} C., was dissolved in 2 ml. of N HOSO4, heated in boiling water for 3 mm., and titrated with 0.01 N KMnO4. Normal values for serum calcium for adults by this procedure are 9.2-10.4 mg./100 ml. (7). *Coleman
Instruments,
Inc., Maywood,
tGeigy Chemical Corporation, Ardsley, J. T. Baker Chemical Co., Phullipsburg,
Ill. N. Y. N. J.
772
JONES & MCGUCKIN
Clinical
Chemistry
Results and Discussion Calcium The indicator, 3- (5-chloro-2-hydroxyphenylazo) -4, 5-d i hy dr oxyiiaphthylene-2,7-disulfonic acid (EBSE), used earlier by Flaschka and associates (8) is sensitive to calcium and magnesium, is watersoluble, is stable in alkaline and neutral solutions, and meets the general requirements for an indicator (9). The differences in absorbance for the dye at pH 12.5 and 10.1 in the presence of calcium and magnesiuni with and without au excess of El)TA are shown in Fig. 3. The quantity of indicator used does not affect the value obtained for either calcium or magnesium but does determine the height of the titration curve (total change in absorbance). The specificity of a complexometric titration is determined to a great extent by the pH at which it is performed. Figures 4 and 5 illustrate that it is necessary to add more NaOH thaii is required to attain
S ‘1
0
S b 0 0
Fig. 3. Change in absorbance with and without excess chelator.
S
0
I 0
SI
Wave length,millimicrons
a p11 greater than 12 in the determillation of calcium. Tile 1)reakS in the curves depicting pH occur at a lower concentration of NaOH than do the breaks in the curves showing titration value. This observation is similar hi both instances (standards and serum samples) and maybe
Vol. 10 No. 9, 1964
SEMIAUTOMATED
773
CA AND MG
7.-
-
5r ‘3 -0-
-o
--
-
-0-
-o
---0---
12.5
,0
0
0’
12
I
I
I
I
2
3
4
5
Ml. 2.4W NoOH 7
12.5
o 12
I
I
I
I
2
3
4
5
MI. 2.4N NoOH
Fig. 4. (top).
Effect
of concentratioa
of NaOH
on pH and on result
of titration
of stand-
ard solutions. Changes in pH are shown by open circles and broken line, which correspond to scale at right. Changes in titration values are shown by solid circles and solid line, which correspond to scale at left. Fig. 5. (bottom,). Effect of concentration of NaOH on pH and on result of titration of serum. Changes in pH are shown by open circles and broken line, which correspond to scale at right. Changes in titration values are shown by solid circles and solid line, which correspond to scale at left.
774
JONES & MCGUCKIN
Clinical
Chemistry
the reason for poor reproducibility noted in some other complexometric methods. This is undoubtedly a reflection of a change in apparent stability constant of the calcium-dye complex with varying ionic strength of the solution (10). The titration value obtained in the presence of a great excess of NaOH is not altered, as has been reported with murexide as the indicator (11). increasing the quantity of the buffer does not change the calcium-plus-magnesium value. One criterion that has been used in evaluation of methods is the effect of dilution of sample on the value obtained. A plot of quantity of sample against titration value ideally should give a straight line with the intercept at zero. This was found with our method both in standards and in samples of serum of varying calcium content (Fig. 6). In the procedure described by Malmstadt and Hadjiioannou (5) a correction is needed since the intercept of standards and of serum is not zero. The inability of the procedure described here to measure low concentrations accurately is indicated by the point representing 0.5 ml. of a serum containing 7.4 mg. of calcium per 100 ml. Recoveries of cal28
24
Fig. 6. Effect of dilution of serum on determination of calcium with EDTA. Each line represents dilutions of different serum.
20 cI
0 #{188}
16
12
8
4
0
MI. serum
cium added to serum are 100 ± 0.5%. Error in the analysis of samples containing a low concentration of calcium, less than 0.05 mg., is eliminated either by adding an internal standard or by preparing a new standard curve in this range of calcium concentrations.
Vol. 10, No. 9, 1964
SEMIAUTOMATED
775
CA AND MG
The precision of the method is excellent. Titration values, in centimeters, are presented in Table 1 ; these values multiplied by a factor of 2 represent milligrams of calcium per 100 ml. of the respective samples. Reproducible values were obtained when the calcium concentration of the pooled serum sample used in Table 1 was determined repeatedly over a period of 15 days (N = 39, = 9.59 mg./100 ml., S. B. = ± 0.044). Phosphate does not appear to have any significant effect on the analysis as performed on serum (Table 2). This also is indicated by the reproducibility of the shape of the titration curve obtained, in which no lag (indicative of a poor end point) was noted with these samples. This is in contrast to a direct titration of calcium in urine in which the end point, however slowly it is approached, is not readily detected regardless of the indicator used. Cyanide, fluoride, and triethanolamine, commonly used masking agents, had no effect on values obtained, even in severely hemolyzed samples; this indicates a lack of masking effect on the indicator by other ions (9). A comparison of the modified Clark and Cohlip method and the diTable
1.
PRECISION
OF TITRATION
OF
CAI.cluht Titration
Sample*
(mg./100
ml.)
Standard, 7.5 mg./l00 Standard, 10.0 mg./100 Standard, 12.5 mg./100 Serum (pooled sample)
BY
EDTA
,alues
(em.) Mean
Repiwates
3.76, 4.99, 6.20, 4.80, 4.80,
ml. ml. ml.
3.74, 5.00, 6.23, 4.79, 4.80,
375
3.74 4,98, 5.00, 4.97 6.20 4.80, 4.80, 4.81, 4.81, 4.81, 4.78, 4.80, 4.85
4.99 6.21 4.80
5One milliliter.
Table 2.
EFFECr
OF ADDED
PHOSPHORUS
Phosphorus water
added* (mp.
x
I Normal
Poole
d serumt
(mg.
)< 102)
as KH,P02. serum
range,
3.2-4.3
X
ay
10’)
)(
10.0 10.0 10.0 10.0 9.9 9.8
10 15 20 50
OF CALCIUM
Calcium standar4
(mg.
10-2)
0 5
*Added
ON DETERMINATION
10” mg. in this size sample.
9.6 9.6 9.6 9.5 9.8 9.8
EDTA
776
JONES & MCGUCKIN
Clinical
Chemistry
rect titration method in over 200 samples indicated that the new method gave values that were consistently 3% lower. To allow a direct comparison, the values shown ill Table 3 are corrected by this factor as are the values obtained by the complexometric procedure applied to TCA filtrates (Table 4). The procedure with TCA filtrates necessitated inclusion of TCA in the standard solutions in preparation of the standard curve, and the values obtained appeared somewhat more variable, as reported by Sadek and Reilley (12). The new method also was tested by the comparison of two complexometric titrations: direct with EDTA and back-titration of excess chelator with standard calcium (Table 5). The values obtained agree reasonably well and again illustrate that a lower value is obtained by this method than by the modified Clark and Collip procedure. Table
3.
RESULTS
BY CLARK
AND
C0LLIP
METHOD
AND
EDTA* (lark (mg./100
10.02 9.71 10.10 9.96 9.14 9.45 9.83 9.64 9.83 9.96
Table
4.
TITRATION
seru,n)
Calco,,.
9.88 9.67 10.05 9.80 9.01 9.54 9.92 9.63 9.92 9.97
10.14 10.01 10.30 9.92 9.30 9.49 10.00 9.62 10.11 10.10
values.
RESULTS
BY CLARK
aad Collip ml. seru,n)
AND
COLLIP
METHOD
AND
Erioel, blue
rome SE
9.75 10.15 9.96 9.29 9.75 9.71 9.74 9.89
9.71 10.08 9.96 9.14 9.83 9.64 9.83 9.96 5Corrected
EDTA
blue SE
BY
EPTA* (lark (my./100
ml
Eriochrorne
a,Id (olliI. ml. serum)
Correetetl
BY DIRECT (rny./100
values.
EDTA (mg./l00
TITRATION ml.
OF
TCA
FILTRATES
serum)
Call-on
9.91 10.22 9.91 9.32 10.22 -
9.96 9.85
Vol. 10, No. 9,
Table
5.
Clark
964
RESULTS
and
SEMIAUTOMATED
(No.1100 ML. SERUM) WITH EDTA, AND BY
represent
AND
BACK-TITRATION
COLLIP
METHOD,
OF EXCESS
10.0 9.5 5.3 10.0 9.7 9.7 10.2
9.4 5.4 9.8 9.8 9.5 10.2 9.3 9.7 10.1 9.4 10.0 determinations
done
TITRATION
EGTA
10.0
routine
BY DIRECT
EGTA
EDTA
Collip*
10.4 9.6 5.6 10.2 10.1 9.9 10.3 9.5 9.9 10.5 9.6 10.2 5Values
BY CLARK
777
CA AND MG
9.2 9.6 10.1 9.4 10.0 in the
clinical
laboratory.
Since higher values are ol)tanled by titration of oxalate in the Clark alid Collip procedure, the following experiment was performed. Calcium was removed from serum by precipitation with oxalate for 6-8 hr. The supernate thell was used as the source of oxalate to precipitate the calcium from a water standard, and a blank from each “calciumfree” serum was diluted with an equal amount of water. The final permanganate titration values obtained on the precipitates from the samples of serum and from the standards, minus the blank values and expressed as milligrams of calcium per 100 ml. are shown in Table 6. The calcium standard in water consistently gave correct values (10 mg./100 ml.) OIl permanganate titration when the oxalate was in water. However, when the source of oxalate was a calcium-free serum, a consistently higher value resulted. In similar experiments, these preparations yielded values of 9.7-9.8 nig. of calcium per 100 ml. when measured by direct E1)TA titration of HNO1-HClO4-digested oxalates. Since it has been demonstrated (13, 14) (and also reported to us by M. Power in a personal communication) that the total calcium content found in a sample, when determined by permanganate titration of an oxalate precipitate, is a functiomi of a number of phenomena (including time allowed for precipitation of the oxalate, inclusion of oxalates other than of calcium, amId soluhility of calcium oxalate itself), variation may he expected. The data shown in the preceding tables illustrate that, in a comparison of methods, the reference method may be subject to considerable criticism.
778 Table
JONES & MCGUCKIN 6.
EFFECT
OF SEaUM-OXALATE
PREPARATION FROM
Clinical
ON PRECIPITATION
WATER
ml.)*
(ng./100
Water No.
Standard 1 2 3 4 5 6 7 8 9 10 11 5By permanganate
Original
Exp.
serum
Water ammonium
sol. oxalate
1
9.98 10.35
-
10.16 9.87 10.08 10.21 10.40 7.71 10.24 9.76 7.29 9.89 titration,
corrected
AS OXALATE
STANDARD
Calcium
Sample
OF CALCIUM
Chemistry
standard Exp.
2
9.95 -
10.35 10.35 10.35 10.35 10.28 10.28 10.35 10.40 10.28 10.38 for
blank
(see text).
To evaluate the effect of serum proteins on the method, calcium-free serum was prepared by passing a sample of serum over a chelating resin (is) (Dowex A1*), which removed! all of the measurable calcium. This preparation, when added back to either serum or standards, did not affect the analysis in any detectable manner. Since standards added to serum also are recovered quantitatively, standards in water are used exclusively in the routine determination. The procedure has been applied successfully to serum from patients with jaundice, myeloma, and lipemia. Highly lipemic serum must be extracted with a twofold volume of diethyl ether saturated with water prior to the titration. The values obtained in these instances agree with those obtained on samples ashed with HNO.-HClO4. Magnesium The standard curve obtained in the titration of magnesium is shown in Fig. 2. The curve is not straight and does not have a zero intercept. The curve is not affected by calcium concentration in the range 0-20 X 10 mg. amid maintains a similar shape with serum samples, as determined by analyzing different volumes of a single serum. The use of Eriochrome black T for the combined calcium-plus-magnesium titraThe Dow Chemical
Company,
Midland,
Mich.
Vol. 10, No. 9, 1964
SEMIAUTOMATED
CA AND MG
779
tion with either Calcon or EBSE for calcium alone yields a similar curve. The direct titration (with EBSE at pH 10.1) of a solution containing only calcium gives a value 0.15 cm. greater than that found with calcium alone at pH >12. This “initial blank” does not appear to remain constant with increasing concentrations of magnesium since the standard curves do not remain straight. The effect on the end point of increasing the temperature of the titration mixture has not been evaluated (9). Although Lewis and Melnick (14) reported that titrations of calcium at pH 12.5 with Calcon or Cal-Red indicator agree with those at pH 10 with Eriochrome black T, results similar to those observed with EBSE were observed with these indicators. The curve, which breaks at about 5 X 10_2 mg., is obtained by analyzing water solutions of magnesium or magnesium plus calcium or standards to which serum free of calcium and magnesium has been added. Since the curve from 5 X 10-2 mg. to higher concentrations of magnesium would extrapolate to a zero intercept, and since the titration values for magnesium in serums to which varying known amounts of magnesium have been added plot as a straight line, it is necessary to use the standard curve in calculating recovery values of magnesium added to serum. Recovery of magnesium from serum was 97-102% and 101-104% for direct titration and back-titration procedures, respectively. Values obtained on serum by the back-titration procedure consistently were 0.4 mg./100 ml. lower. Manipulations to resolve the apparent discrepancy have failed to yield a satisfactory explanation. Possibly a more accurate value would be obtained by adding an internal magnesium standard (0.05 mg.) but since this would entail a second quantitative transfer, the direct titration as described earlier is preferred for routine use. Thus, the value for serum magnesium obtained by this procedure, although it appears to agree with those obtained by other methods (16-18), may not be the true value. Prior to adopting the method described herein, a number of methods for the estimation of serum calcium were investigated, including highand low-temperature flame photometry, colorimetric-complex formation, complex formation, and other complexometric methods. Precision and agreement with the modified Clark and Collip procedure were not easily realized in the majority of those methods tried, and, while the high-temperature flame photometric procedure was acceptable, it was
780
JONES & MCGUCKIN
Clinical
Chemistry
sufficiently more time-consuming to niake questionable whether it would be possible to process efficiently the large number of samples handled routinely in our clinical laboratory. An extensive study of tile variations in concentration in serum of calcium, magnesium, and phosphorus in normal persons has been initiated; prelimillary data on 40 normal adults indicate calcium to be 9.47 ± 0.26, and magnesium, 2.38 ± 0.14 mg./100 ml. (means ± S.D.).
References 1.
Eldjarn,
L., Nygaurd,
0., and
Sveinsson,
S. L., Seandi,,ar.
,J. Cli,,.
4- Lab.
Invest.
7, 92
(1955). 2. Bowden, C. H., and Patston, Valerie J., J. Cli,t. Pathol. 16, 18 (1963). 3. Watson, D., and Rogers, J. A., Gun. Chint. Aeta. 8, 168 (1963). 4. Lumb, G. A., Gun. Chint. Acta. 8, 33 (1963). 5. Malmstadt, H. V., and Hadjiioannou, T. P., Clip. Chew. 5, 50 (1959). 6. Yarbo, C. L., and Golby, R. L., Anal. Chew.. 30, 504 (1958). 7.
Power,
8. Flaschka, 9. Welcher, 10. 11. 12. 13. 14. 15. 16. 17. 18.
M. H., Toogood,
and Adamson, Lucille, Clin. Chem. 2, 278 (1956). A. A., and Sadek, F., Z. Physiol. Chem. 310, 97 (1958). Uses of Ethylenediawineteira4eeiie Acid. Van Nostrand Co.,
Ruth,
H., Abd El Raheein,
F. J., Analytical Inc., Princeton, 1958. Brush, J. S., Anal. Chew. 33, 798 (1961). Kenny, A. D., and Cohn, V. H., Anal. Chem. 30, 1366 (1958). Spdek, F. S., and Reilley, C. N., J. Lab. 4’ Glii. Med. 54, 621 (1959). Maclntyre, 1., Biochern. J. 67, 164 (1957). Lewis, L. L., and Melnick, L. M., Anal. Chew. 32,38 (1960). Walser, Mackenzie, Anal. Chein. 32, 711 (1960). Hasselman, J. J. F., and Van Kainpen, E. J., Clin. Chin,. Ada. 3, 305 (1958). WaIlacli, Stanley, (‘ahill, L. N., Rogan, F. H.,and .loiies, H. L., J. Lab. 4- Gun. Med. 59, 195 (1962). Stewart, W. K., Hutchi,i.son, Frederiek, and Flemming, Laura W., J. Lab. 4- Clin .Med.
61, 858 (1963).