V O L U M E 2 8 , N O . 6, J U N E 1 9 5 6
915
openings are used to indicate whether the third digit is 4 4 (unmarked punch to right of S o . 29) or >, 5 (position 29). OPTICAL DATA
Refractive Indices. Only two refractive indices are given, either omega ( w ) and epsilon ( e ) or alpha ( C Y )and beta ( p ) . The third index, gamma (y), can, of course, be computed from the optic avial angle and the sign of double refraction. They are ~ecordedin the same way as the density and with the same :issumptions in positions 9 to 16 on the right plus the upper of two unmarked positions in the lower right corner ( C Y or W ) and i n positions 1 td 9 ( p or e). Sign of Double Refraction. Two openings are available: one iS o . 10) if Dositive and the other ( S o 11) if negative. Pleochroism. As probably all aniscstropic coibred compounds :ire pleochroic and as all pleochroic ccmpounds are colored, this position (No. 12) will probably be pu:iched for all colored compounds. Optic Axial Angle (2V). This angle must be less than 90": hence two digits are recorded n-ith tvro 7, 4, 2, 1 combinations in positions 13 t o 20 along the top. Extinction Angle. This property i.3 very important, as it is essential to roper correlation of the optical and morphological properties o!h te crystal. The angle thosen for recording (positions 21 to 28) is therefore the smallest angle between c and the lower of the two refractive indices lying in 010. It is then necessary t o specify further whether th: lorn-er index is CY or p (No. 29) and whether i t lies in the acute or obtuse angle. This is done with additional punching; if the extinction position is in acute p, the upper of two unmarked openings in the upper left corner is punched. ~
Dispersion. If dispersion is unusually strong or anal] tic;tllv significant in any way, the lower of the two unmarked openings in the upper left corner is punched. Readily apparent anomalous polarization colors or strong dispersion of the optic axes (greater than, say, about 2" per optic axis) will be sufficient t o justify a punch. Finally, the card has been designed for a special purpose. .Iny other laboratory might well be satisfied u-ith a simpler card designed for its particular purpose. Most analytical laboratories use only the H a n a d t cards, m-hich tabulate x-ray p o ~ d e r diffraction data. This type of data has been omitted almost entirely from the cards described here, because the data are so readily available on the Hanawalt cards. The abundance of such additional analytically useful data as x-ray, optical, and morphological is emphasized by these cards. I t is hoped that this emphasis on cr) stallographic properties 13 ill help to call attention t o this generally neglected field of analysis, and tempt additional analysts to learn the necessary techniques, so that the task of accumulating analytically useful crystallographic data may be made easier. LITERATURE CITED
(1) Kirkpatrick, 9.F., AI.. C H E l r . 20, 847 (1948). (2) McCrone, W. C . , Ibid., 20, 274 (1948). RECEIVED f o r review October 6 , 105Z. Accepted blarch 8. 1956. llirision of Physical and Inorganic Clirtiiistry, 123rd Meeting. ACS, Los .ineeles, Calif., Marcli 1953.
Chromatographiic-Acid Hydrolysis Method for Determination of Triacetin in Double-Base Powder JAMES 0. WATTS and HARRY STALCUP Research and Development Department,
U. S. Naval
Powder Factory, Indian Head,
This investigation was undertaken to develop a direct method for the determination of triacetin in doublebase powder. The method in general use heretofore, which depends on a value obtained by difference, is basically unreliable. The method submitted in this report is based on chromatographic separation of the extracted ester from nitroglycerin and other components, followed by hydrolysis in standard acid and subsequent titration of the liberated acetic acid.
A
S INVESTIGATION n a s undeataken t o develop a rapid
direct method for the determination of triacetin in doublebase powder (so called because it ccntains both nitrocellulose and nitroglycerin). .Ipplication c f the chromatographicgravimetric procedure of Pauling (IO) I S modified by French and 1Iosely ( 3 ) has shown that considerable error results from the presence of impurities in the triacetin fraction. Kight ( ? ) has reported a method for the determinatim of triacetin in a doublebase formulation based on chroma ographic separation and saponification of the triacetin rrsidue with aqueous standard alkali solution. Recent experiments in this laboratory have indicated that a method based on hydrolysis in dilute acid would have application over a wider range of propellant powder formulations. The and Yamasaki ( 1 2 ) on previous xork of Geitel (4,6 ) , 3Ieyer (8), the determination of the velocity constants for triacetin a t successive stages of hydrolysis in dilute acid indicated the feasibilitv of this approach.
Md.
Winter and Butts (11) have reported a method for triacetin determination, applicable to double-base powders, based on hydrolysis of the ester (after evaporation of the extracting solvent) in phosphoric acid, followed by distillation and subsequent titration of the acetic acid formed. The method appears t o be satisfactory with respect to accurary and precision but requires the efiorts of a skilled technician for best results. The method presented in this report is based on chromatographic separation of the extracted ester from nitroglycerin and the other components, follo\r ed by hydrolysis in standard hydrochloric acid and subsequent alkali titration of the librrated acetic acid. Dioctyl phthalate and fatty acid were found to be unaffected by dilute acid, and their presence in the triacetin residue after chromatographic separation offered no interference with final results. However, the presence of more than a trace of 2-nitrodiphenylamine in the triacetin residue interfered somewhat with the end point during the alkali titration of the acetic acid, which made its removal desirable. A round robin test of the method by 12 laboratories of the Joint Army, Kavy, and Air Force Panel on Analytical Chemistry showed a standard deviation of O.OiYc within individual laboratories and 0 5c5 between all laboratories ( 6 ) . i P P A R A T U S 41-D R I 4 T E R I 4 L S
Apparatus. Chromatographic tube, 35 mm. in diameter and a t least 200 mm. long (or Size KO. I11 chromatographic tube with fritted-glass disk sealed to inner joint connection). Suction-filtration apparatus, see apparatus in Figure 1 Coors filter disk (20 to 22 mm.), perforated porcelain.
ANALYTICAL CHEMISTRY
976 Condenser (GO-cm.), Allihn. Heater (550 watts), coil. Soxhlet unit, all glass with siphoning-cup capacity of 25 to 40 mi. Extraction thimble, paper or glass with fritted-glass disk bottom, coarse porosity. Constant temperature bath. Solvents. Methylene chloride, redistilled (Fisher reagent grade). Ethyl ether, anhydrous, redistilled from sodium hydroxide pellets. Acetone, analytical reagent grade Methanol, technical grade.
-
____ CHROMATOGRAPHIC TUBE No. Ill
i Id-
P RF RAT
P~Rc'~LA~FFLBIsK
:;-\NITRATE
DINITROGLYCERIN
t;' 4
-BALL
& SOCKET JOINT -.TO
VACUUM
cVAC UUM
/\
BLEEDER L~~~~~~ HOSE CONNECTOR
'3
- 2 5 0 m l ERLENMEYER FLASK b 2 4 / 4 0
Figure
1.
Chromatographic
apparatus s h o w i n g relative position of adsorbed m a t e r i a l on silicic acid-Celite a d s o r b e n t a f t e r de-
velopment
Reagents and Solutions. Sodium Hydroxide (0.1N). .4 solution containing 4 grams of sodium hydroxide in 1 liter of distilled water is standardized against potassium acid phthalate, phenolphthalein being used ae an indicator Hydrochloric Acid (0.1S). A solution containing 9 ml. of concentrated hydrochloric acid per liter is standardized against the alkali Solution A. A solution containing 5y0 hydroxylamine hydrochloride in 10% potassium hvdroside is freshly prepared each week. Solution B. A solution containing 2% ferric chloride in 1.0-V hydrochloric acid is prepared as required Adsorbent Test Solution. An adsorbent test solution is prepared by dissolving 0 20 gram of triacetin (c.P.)and 0 033 gram of 2-nitrodiphenylamine (c P . ) in 30 ml. of redistilled methylene chloride. Adsorbent. Two parts by weight of silicic acid, Baker or Mallinckrodt (9% water), are thoroughly mixed with one part of Celite (Johns-Rlanville Hyflo Super-Cel), according to the method reported by French and Mosely (3). Preparation of Adsorbent. A large quantity of the adsorbent is washed with solvents to remove any soluble material, then i t is dried, heated, and stored ready for use. Used adsorbent should be washed in the same manner. A 90 X 90 mm. fritted-glass Buchner funnel assembled to a suction-filtration apparatus is prepared. K i t h the suction applied from a xvater aspirator or an! source of vacuum, the adsorbent is slowly added and packed until the funnel is filled to within 0.25 inch of the top. The surface of the adsorbent is then smoothed and packed with the flat end of a wooden rod prior t o the addition of the solvents K i t h the suction still applied, the following solvents are added in the order listed, each new solvent being added just before the to of the column goes dry: 1-V ethyl ether, 1-V acetone, and 1-$methanol, V being defined as the minimum quantity of liquid required to wet the entire adsorbent. The adsorbent is removed from the funnel and air-dried until practically all solvents are evaporated; it is then heated in an oven for 24 hours a t 150" to 160" C. The adsorbent may be stored in 0.5-gallon Mason jars, containing a few ebbles from 0.5 to I inch in diameter to facilitate complete breaKdown of any lumps. The material is shaken vigor-
ously to fluff it and transferred to another jar just prior to use in the column. Reclaiming Used Adsorbent. The above procedure may also be used for reclaiming used adsorbent. After six consecutive determinations, the chromatographic column is extruded and the upper 0.5 inch is discarded. When a sufficient quantity of the used material has been accumulated, it is treated exactly as described above. No change in adsorptive properties of the used adsorbent sufficient to require a modification of the procedure has been observed as a result of this treatment. Preparation and Adsorptive Test of Chromatographic Column. A chromatographic tube about 35 mm. in diameter and between 200 and 250 mm. long is assembled to a suction-filtration apparatus (Figure 1). If the tube is not provided with a coarse porous disk, a plug of glass wool is inserted into the bottom. The adsorbent is added to the tube until a column about 12 cm. high is formed. Suction from a vacuum line (about 10 mm. of mercury) is then turned on and the top of the column leveled. If the column height goes below 10 cm., additional silicic acid is added until the column height is 10 cm. A porcelain porous disk is placed on top of the column. Suction is again applied, and the following solvents and solutions are added t o the column in the order listed, each new solvent or solution being added just before the previous one is sucked into the column: 60 ml. of petroleum ether, 30 nil. of the adsorbent test solution, and 170 ml. of redistilled methylene chloride. The column should be free from 2-nitrodiphenylamine, as indicated by the absence of yellow coloration on the lover portion. Suction is maintained until the liquid begins to drip slowly from the tube, then it is released. K i t h the aid of a rod, the column is extruded onto a paper towel or sheet of paper. The position of the triacetin on the column is determined as follows (9). A contiguous series of drops of Solution A is placed along the extruded adsorbent, from top to bottom, followed by a superimposed streaking with Solution B. If the purple color produced by the triacetin extends lover than 3 cm. from the bottom, a new column should be prepared containing more adsorbent. If the 2-nitrodiphenylamine is not washed completely from the column, more developer should be used or the column made shorter. The column may be used for six consecutive sample runs before repacking with fresh adsorbent. PROCEDURE
Extraction of Sample. -4sample consisting of approximately 2 grams of the powder (ground in a Kiley mill to pass 20 mesh) is accurately weighed into a tared balance pan. The sample is transferred t o a small paper extraction thimble and the balance pan is reweighed. The thimble is plugged with a thin layer of glass wool and dropped into a small, all-glass Soxhlet extractor having a siphoning cup of 25- to 40-ml. capacity. A 250-ml. extraction flask containing 50 ml. of methylene chloride is attached to the Soxhlet extractor. The extractor is then attached t o a water condenser, the siphoning cycle is adjusted to 1 or 2 minutes, and the sample is extracted over a steam bath for 3 or 4 hours. After the extraction has been completed, the volume in the flask is adjusted to approximately 30 ml. Chromatographic Development. The chromatographic tube containing the adsorbent column packed t o a height of 10 cm. is attached t o the suction-filtration assembly illustrated in Figure 1. Vacuum sufficient to provide a flow rate of approximately 25 ml. per minute is applied and the column is wetted Kith 60 ml. of petroleum ether. The 30-ml. volume of sample solution in the extraction flask is then poured upon the adsorbent column. The sample flask is rinsed with tn-o 5-ml. portions (or less) of methylene chloride and the rinsings are added to the column. The column is then developed with additional methylene chloride, a total of 1'70 ml. being used for both rinsing and developing. The column may be used for as many as six consecutive determinations r i t h o u t repacking. Removal of Eluting Solvent. When the column has almost stopped dripping, the vacuum is released and the flask containing the filtrate is replaced with another Erlenmeyer flask. Vacuum is applied and the column is eluted with 150 ml. of anhydrous ethyl ether. The solvent in the Erlenmeyer flask is removed by evaporation under a reduced pressure (approximately 500 mm. of mercury) with the flask partially immersed in a constant temperature bath maintained between 40" and 42" C. The flask is removed from the vacuum and water bath immediately after all the solvent has evaporated. Hydrolysis and Titration. To the residue in the Erlenmeyer
V O L U M E 2 8 , NO. 6, J U N E 1 9 5 6 flask are added 25 ml. of standard (approximately 0 . l N ) hydrochloric acid and about 20 small glass beads. The flask is then attached to a 60-cm. Allihn condenser and, by means of a hot plate or coil heater, refluxed a t boiling temperature for i 5 minutes, the time required for the complete hydrolysis of the triacetin. T h e flask is cooled to room temperature while still attached to the condenser. T h e condenser is washed with 50 ml. of distilled water, after vihich the flask is disengaged from the condenser, and the combined hydrochloric and liberated acetic acids are titrated with standard 0 . 1 s sodium hydroxide, phenolphthalein being used as the indicator. A blank is run on 25 nil. of the standard acid in the same manner as the sample. The per cent triacetin is calculated by means of the following equation :
N X 0.02 x
Pel s.ciit triacetin = ( A - B ) X
100
TV
11 = milliliters of standard alkali required to titrate the sample 15 = milliliters of standard alkali required to titrate the blank -3- = normality of the standard alkali T I - = grams of yon-der sample 0,073 = grams of triacetin (glycerol triacetate) equivalent to 1 ml. of I.\- alkali The factor 0.073 is derived from the following equations:
AI%0cOc&
Fe+++
I
CH?OCOCIHa Triacetin
HC1
- 1
1
CIlOH
+ 3CH3COOH
Per cent triacetin
4
=
( A - B ) X S X 0.021 X 1.154 X 100 TT'
\J-here
B
milliliters of titanous chloride required to titrate the sample R = milliliters of titanous chloride required to titrate the ferric iron in the blank = normality of the titanous chloride TV = grams of the powder sample 0 021 = grams of nitric arid equivalent to 1 nil. of 1 S titanous chloride 1 154 = conversion factor for nitric acid to triacetin =
Table I. Effect of Nitroglycerin Impurities (in Eluate) on Triacetin Recovery from Synthetic Double-Base Powder Samples Replica
T A Added, Gram
T.4 Found, Gram
0.1807 0.1759
+ 1120
CH,ONOZ CHOH
Av. 1 2 3
0.2039 0.2368 0.1967
0.2040 0.23fi0 0.1946
(>H?OKO2
kH20H
Dinitroglycerol Because the nitric acid interfered with triacetin results, i t became necessary either to remove the compound or to find a means of correcting for its presence. It was desired to keep the chromatographic phase of the procedure as simple as possible with respect t o the developing solvent; therefore no effort was made to remove the dinitrogl-cero1 from the column. It rras found that the standard procedure for the determination of nit,roglycerin \vas applicable to the determination of dinitroglycerol in the presence of triacetin. The correction was determined as follows.
.i 2-gram powder sample was extracted R-ith methylene chloride for 3 hours. T h e extract was chromatographed and the column was eluted as described in the procedure. T h e eluate was then transferred to a standard titanous chloride titration flask and evaporated to dryness, and the nitrates were determined by the ferrous iron-titanous chloride method of Becker ( 1 ) used for nitroglycerin determination in accordance with the following equations:
99.07 100.03 99.65 98.94
hv.
99.54
Table 11. Effect of Evaporation of Eluting Solvent on Triacetin Recovery, Dry Basis TA Added, T A Found, Recovery, Replica Gram Gram % T A Assayed Directly by Hydrolysis with 0.liV Hydrochloric Acid 1 2 3
0,1820 0.1540 0.1806
100,00 100.06 99.95
0,1820 0.1341 0,1805
4v.
100.00
T h Dissolved in 150 h l l . ,.4nhvdrous E t h e r , Evaporated and Assayed b y Hydrolysis wiih 0.1.V Hydrochloric Acid 1 0.1850 0.1838 99.33 2 0.1990 0.1970 99.00 3 0,2050 0.2038 99.40
CHzOH
+ 2H20 HCI d H O H + 2HSO3 +
98.96 99.18
Samples with 0.50 Gram of Sitroplycerin
DISCUSSION AIVD RESULTS
The presence of any dioctyl phthalate with the eluted triacetin does not interfere n-ith results. 2-Nitrodiphenylaniine in more than trace amounts has a masking effect on the end point tliiring the alkali titration, although no chemical change was observed after the hydrolysis step. Chromatographic investigations on synthetic mixtures simulating the methylene chloride estract from a double-base powder revealed the presence of a small quantity of impurity, identified by French and Uosely (3) as dinitroglycerol, on the column as a small narrow band \\-ithin the same area as t h a t occupied by the triacetin. During the acid hydrolysis step of the procedure this compound formed nitric arid as s h o r n i n the follon.ing equation:
%
0.1788 0.1745
Acetic acid
C'II,3COOH f S a O H -+ CH&OO?r'a
Recovery,
Samples x i t h o u t Xitroglycerin 1 2
CHZOH Glycerol
+ TiC18
+ 3Fef++ + 2Hz0 F e + + + TiCl,
SO
The nitrates were then calculated in ternis of per cent triacetin by means of the following equation:
CHZOH
+ 3H.O
-
A\-
where
CH?OC()C&
+ 3 F e + + + 3H'
"03
977
Av.
99.24
TA l s s a y e d by Chromatographic-Acid Hydrolysis Procedure 1 2 3
0.1781 0.1518 0.2189
0.1761 0.1502 0,2173
99.07 98.95 99.26
Av.
99.09
When the proposed procedure was applied to synthetic samples containing triacetin but no nitroglycerin, a recovery of 99.1% (dry basis) of the actual amount of triacetin originally added was obtained (Table I). This loss, most of which was found to occur during the eluate evaporation stage (Table 11),is equivalent to an error of 0.09% in the total analysis of the double-base powder sample which nominally contains 9.507, triacetin. However, when the procedure was applied to synthetic double-base pori-der samples containing nitroglycerin, almost full recovery was obtained (Table I). This recovery could not be fully accounted for by the increase in acid due to the dinitroglycerol hydrolysis. Synthetic 2-gram samples were likewise analyzed for dinitroglycerol by the same procedure. Table 111 shows a comparison of results obtained for the synthetic and the standard double-base porrder samples and indicates that the dinitroglycerol
A N A L Y T I C A L CHEMISTRY
978
present in the standard double-base p o der ~ raises the percentage obtained for triacetin in the total powder analysis by O.O-iCc. An effort was then made to determine the amount of acidic impurities other than dinitroglycerol. Synthetic double-base powder samples with and without nitroglycerin (and xithout triacetin in both cases) were piepared and tested by the proposed procedure.
Table 111. Dinitroglycerol Content of Synthetic and Standard Double-Base Powder Samples (Ferrous sulfate-titanous chloride method for nitroglycerin) Per Cent DSG (Calcd. as Triacetin) Replica Synthetic Standard 0.04 0 03 0 05 -41.. 0 . 0 4
0.03 0.02 0.06 0.04
The results in Table IV indicate that the dinitroglycerol in the nitroglycerin accounts for only one half of the total acid impurities. The total acid impurities were found to add 0.087, to the value obtained for triacetin in the total powder analysis, which appears t o compensate for the O.OSC/, loss of triacetin during the eluate evaporation stage of the procedure. On the basis of this investigation neither a correction factor for acidic impurities nor a recovery factor for triacetin \vas used in the proposed procedure.
Table IV. Acidic Impurities in Synthetic Double-Base Powder without Triacetin by Chromatographic-Acid Hydrolysis Method yo .4cidity (Calcd. as TA) in Total ilnalysis of Powder With Without 0.50 gram of Replica nitroglycerin nitroglycerin 1 2 3
0.03 0.03 0.03 0.04
4
0.05 Av.
0.04
0.09
0.07 0.07
0.07 0.08
0.08
The increase in the value obtained for TA due t o the presence of dinitroglycerol and other acid impurities adds 0.08% t o the total analysis percentage.
A considerable quantity of a nitrate compound was found in the methylene chloride extract. Reports received by the authors have identified the nitrate in a similar experiment as nitrocellulose that had been extracted by the methylene chloride and adsorbed a t the top of the chromatographic column. The material remained a t the top of the column during the development and ether-elution operations. After six consecutive determinations had been made, this accumulation reached such proportions as to necessitate repacking the column with fresh adsorbent. Before the adsorbent could be reclaimed, it was necessary to extrude and discard this contaminated upper portion. Figure 1 shows the relative positions of the various materials on the column after development with methylene chloride. T h e use of a Soxhlet tube of 25- to 40-ml. capacity permits a siphoning cycle of 1 to 2 minutes over a steam bath ( 2 ) . Consequently, a 4-hour methylene chloride extraction (instead of the usual 16 hours) is sufficient. Determinations may be made within 2 hours after extraction, and the entire analysis may be completed within 6 hours. I n a determination of the triacetin in a double-base powder sample, the following results were obtained :
Triacetin, OC
Rep1i c n 1
2 3
? Average
Range
Standard deviation
9 9 9 9 9 9 0 0
41 44
45 54 54 48
13 06
The range and standard deviation shown indicate a satisfactoi y degree of precision. The procedure 1%as originally developed for propellant formulations containing dioctyl phthalate, 2-nitrodiphenylamine, and a fatty acid salt in addition to nitroglycerin and nitrocellulose. l l a n y other ingredients commonly found in double-base propellants do not appear to offer interference Compounds specificaally tested (by adding known amounts to the methylene chloride extract of the propellant) and shown to be without effect include ethyl centralite and dibutyl phthalate. Other ingredients such a6 diphenylamine and dinitrotoluene present no problem, as they pass completely through the column with the nitroglycerin component during development. LITERATURE CITED
(1) Becker, W. W.,IND.ENO.CHEM.,ANAL.ED. 5, 152-4 (1933). (2) Corey, R. B., Dekker, A. O., Malmberg, E. W.. LeRosen, 4. L., Schroeder, W. A,, Office of Scientific Research and Development, Division 8, NDRC, OSRD Rept. 1837, 35 (1943). (3) French, J. C., hlosely, B. P., Hercules Powder Co., Kenvil, N. J., “Determination of Diethylphthalate (DEP), Triacetin (TA) and Ethyl Centralite (EC) Plus Diphenylamine (DP-4) in Admixture by Chromatographic Separations-Kenvil Methods,” Smokeless Powder-Analytical AIethods, Rept. Invest. 3112 (Aug. 1, 1952). (4) Geitel, A. C., J . prakt. Chem. 55, 417 (1897). (5) Ibid.,57, 113 (1898). (6) Joint Army-Savy-Air Force Analytical Chemistry Panel for Solid Propellants, hIichelson Laboratory, Naval Ordnance Test Station, China Lake, Calif., Report of Round Robin on Determination of Triacetin, private communication, LIarch 23, 1955. ( i )Kight, W. E., Hercules Powder Co., Allegany Ballistics Laboratory, Cumberland, bId., “Determination of Triacetin in Double-Base Propellant,” private communication, April 15, 1954. (8) Xeyer, J., 2. physik. Chem. 66, 81-125 (1909). (9) Pauling, L., Office of Scientific Research and Development, Division 8, NDRC, OSRD Rept. 5952, 18 (1945). (10) Ibid.,p. 72. (11) Winter, L. S., Butts, P. G., Liberty Powder Defense Corp., Badger Ordnance Works, “Distillation Method for Determination of Triacetin in Double-Base Propellant Powders,” Lab. Dept. Project 6005 (3Iarch 17, 1953). (12) Yamasaki, E., J. A m . Chem. SOC.42, 1455 (1920). R E C E I V Efor D review October 5 , 1955. Accepted March H, 1956. The opinions or asscrtations contained in this article are the private ones of the writers and are not t o be construed as official or reflecting the views of the Navy
Department.
Electron Microscopy-Correct
ion
I n the review on “Electron Microscopy” [Swerdlow, Max, Dalton, A. J., Birks, L. S., ANAL.CHEM.28, 597 (1956)l the Literature Cited should be listed as follows: (200) Rhodin, J., Ezptl. Cvll Research 8 , 572 (1955). (201) Rhodin, J., Monograph, Aktiebolaget Godvil, Stockholm, 1954. (202) Rhodin. J . , Dalhamn, T., Ezptl. Cell Research 9, 371 (1955). On page 604, second column, second paragraph, third line and sixth from last line, on page 605>last line, and on page 607, second column, sixth line, (202) should be changed to (201). On page 607, first column, third and seventh lines of last paragraph, (201) should be changed to (200). Page 606, second column, second paragraph, second line, (200) should be changed to (202).
