INDUSTRIAL AND ENGINEERING CHEMISTRY
December, 1926
1245
Table 111-Effect
Effect of Varying the Solvent
of Varying t h e S o l v e n t A S OLEIC--Cottonseed Soy bean Linseed oil oil oil 0.036 0.94 2.28 0.044 0.94 2.26 0.032 0.97 2.38 0.032 0.97 2.35 0.044 0.97 2.34 0.032 0.98 2.31 0.036 0.91 2.15
--.%CIDITY
Thgvariable introduced in this series of experiments consisted in the use of petroleum ether, ethyl ether, acetone, benzene, and a mixture of benzene and isopropanol (1:l) as solvents for the oils. Tung, linseed, soy bean, and cottonseed oils served as experimental material. The ethyl alcohol method5 again was used as the control. As before, 28.2gram samples were dissolved in 50 cc. of the solvent. Neutralization was effected with 0.1 N isopropanol potash. The tung oil dissolved in the isopropanol only after warming to about 60" C. All oils gave clear, sharp end points except tung oil, which with petroleum ether, acetone, and ethyl ether gave slight precipitates during titration. These were dispelled, however, on gentle heating. Data are recorded in Table 111 in which acidities have been
Solvents Ethyl alcohol Isopropyl alcohol Petroleum ether Ethyl ether Acetone Benzene Isopropyl alcohol
4- benzene
(1:l)
Tung oil 3.81 4.19 3.92 3.89 3.95 3.89 3.88
expressed in terms of oleic acid. Satisfactory agreement exists in all cases. Experimental data on the recovery of lauric acid when added to the same oils show that the smallest discrepancies exist between theoretical and determined values when benzene or benzene-isopropanol are used as solvents. This suggests that benzene inhibits the saponifying action of the alkali. This point will be investigated at greater length.
Rate of Molecular Weight Increase in the Boiling of Linseed and China Wood Oils'" By J. S. Long and Graham Wentz W r . H. CHANDLER CHEMICAL LABORATORY, LEHIGHUNIVERSITY, BETHLEHEM, PA.
T
BENzENE-The benzene was HE molecular weight of The use of various solvents to study rate of molecular raw and bodied linseed weight increase by the freezing point method gives :$ : s,o,'~t~~:h$~?~~-f~~~ oils has been detercomparable results in the case of linseed and China method suggested by Richards. 6 mined by a number of inwood oils. The freezing point constant of vestigators.3 Determination Extraction by hot solvents permits extension of the this benzene was determined of the molecular weight by freezing point method so as to determine the molecular $ ~ f d ~ ~ :$ :~: ~ ~ n the freezing point method has weights of very thick oils. benzene, and anthracene. The been suggested as a means of The "break" of linseed oil is a factor of the first result of closely agreeing demagnitude in affecting the rate of molecular weight terminations with these subcontrolling the boiling of drystances gave a value of 5065, The rate of rnolecuincrease when linseed oil is boiled. ing compared with 5067 as calcular weight increase of linFurther evidence has been obtained which indicates lated by Raoult's formula. seed oil under varying condithat the reactions which occur during the boiling of LINOLENIC ACIDAND MONOlinseed oil include condensations as one of the major GLYcERIDE-hoknic acid and tions has been observed and linolenic monoglyceride were s u g g e s t e d a s a m e a n s of types of reactions occurring. prepared and purified in this studying the mechanism of laboratory.' the reactions which occur during the boiling pro~:ess.~The SOLVENTS-In addition t o the benzene used in most of this work, solvents of the highest possible purity were obtained and present paper is an extension of this work.
ti:
gL&
Materials LINSEEDOIL-TWO shipments of raw linseed oil were used, both derived from northwest seed. They showed the following constants a t the time used: I Specific gravity a t 15,5°/15.50 C. Iodine number Acid value
0.9336 189 2.05
I1 0.9330 187 2.34
Neither oil showed a,ly break when heated to 300 : in a test tube over a Bunsen flame with interDosed asbestos ,.cauze. I was used in run 48. I1 was used in all &her runs. CHINAWOODOIL-The China wood oil was furnished from a reliable source and had the following constants : Iodine number Acid value Browne heat test
167 5.10 11 minutes 33 secoids
Presented before the Midwest Regional Meeting and the Meeting of the Section of Paint and Varnish Chemistry of the American Chemical Society, hladison, Wis., M a y 27 to 29, 1926. 2 This work was carried out under the New Jersey Zinc Company Fellowship a t Lehigh University, acknowledgment of which is gratefully made. 3 Seaton and Sayer, THIS JOURNAL, 8, 490 (1916); Friend, J . Chem. Soc. ( L o n d o n ! , 111, 162 (1917); Morrell, J . Oil Colour Chem .ASSoC., 7, 146 (1924). ,~ 4 Long and Smull, THISJ O U R N A L , 17, 138 (1925). 6 Long and Wentz, I b i d . , 17, 905 (1923).
subjected to slight further purification as follows: The nitrobenzene was redistilled under atmospheric pressure and the portion which distilled over between 206 and 206.5 was collected for Use, Small quantities of alcohol were removed from the bromoform by means of a fractionating column under a pressure of 18 mm. The alcohol came over a t 31 C. The resulting bromoform was transferred directly to the freezing point tubes. Para-cresol was redistilled under a pressure of 24 mm. The portion distilling between 107" and 108" C. was used. UMBER-This was a commercial product, brown in color, and having the Per cent 16,80 A1203 6.40 FezOa 3 8 . 4 0
sioz
MnO Ma0 Ca?O
Per cent o,46 0.14 3.10
Per cent 0.22 Carbonaceous matter and Hz0 3 4 . 4 8 (by difference)
so3
Method
1
Essentially the same apparatus and method were used as has been previously d e ~ c r i b e d . ~Five , ~ hundred grams of oil were heated in open 13-em. porcelain casseroles and mechanically stirred a t the rate of 200 r. p. m. I n run 53, 500 grams of oil were heated in a %liter Pyrex flask into m-hich steam \vas passed so as t o maintain an at8
7
J . A m . Chem. Soc., 47, 2285 (1925). Method to be described in a later paper.
INDUSTRIAL AND ENGINEERING CHEMISTRY
1246
mosphere of steam in the flask. Temperature was maintained constant to *2’ C. in all runs. Samples were withdrawn at regular intervals up to heat, after 0.5 hour, 1, 1.5, 2, 2.5, and 3 hours, and their molecular weights were determined by the freezing point method. Benzene was used as solvent in all experiments except the comparison of various solvents. Accurately weighed samples of 1 gram * 1 per cent were used to eliminate any molecular weight variation due to concentration of the solution in benzene. I n run 57, 500 grams of linseed oil which had been dried for 2 hours a t 110’ C. in an atmosphere of carbon dioxide were heated in a 2-liter Pyrex flask connected to a vacuum pump through a train consisting of (1) a 50-cc. glass flask filled with glass beads to catch mechanically entrained oil; (2) a Ytube and stopcocks; (3) two pairs of calcium chloride tubes which were weighed a t regular intervals to determine the weight .of water given off from the oil in that interval; (4) gas washing bottles containing calcium hydroxide and acid sodium sulfite solutions, together with suitable stopcocks and calcium chloride protection tubes which were not weighed; and (5) a mercury manometer. The pressure during the run varied from 14 mm. to 25 mm. of mercury with the oil a t 293’ C. Comparison of Molecular Weights as Determined in Various Solvents
Five hundred grams of linseed oil were heated for 3 hours at 293’ C. with mechanical stirring a t 2@0r. p. m. Samples were taken when the oil was “up to heat” and after 1.5 hours and 3 hours. The molecular weight of these samples was determined by the freezing point method using six solvents. Five hundred grams of China wood oil were heated for 1.5 hours at 192” C. with mechanical stirring. Samples were taken “up to heat” and after 1 hour and 1.5 hours. The molecular weight of these samples was determined by the freezing point method using four solvents. The freezing point constant, K , of the purified solvent was determined in each case using both aromatic and aliphatic substances. Benzene and carbon tetrachloride were used for each solvent. Carbon disulfide and anthracene were used as further standards in the case of several solvents. The constant used is the average of two to five determinations. 1 gram * 1 per cent ofsample was dissolved in 25 cc. of solvent. The weight of solvent as delivered by a 25-cc. pipet was determined as follows: Solvent Benzene Para-cresol Diphenyl ether Bromoform Ethylene bromide Nitrobenzene
Grams 21.8434 25.6244 26.6600 71.5971 54.5025 30.0100
The molecular weight values obtained with the different solvents are given in Table I. T a b l e I-Molecular
W e i g h t s of Linseed and C h i n a W o o d Oils as Obtained w i t h Various Solvents“ -MOLECULAR WSIGHT--
After After K Up t o heat 1.5 hours 3 hours Linseed Oil Benzene 5065 789 1067 1868 Nitrobenzene 6261 732 997 1612 Ethylene bromide 10860 804 1082 1900 Bromoform 13360 707 1012 1437 Diphenyl ether 7370 957 1433 1926 Para-cresol 602 793 1541 6154 China Wood Oil After After 1 hour 1.5 hours Benzene 5065 815 1503 1935 Ethylene bromide 11040 828 1500 1940 Bromoform 14034 773 1264 1796 Nitrobenzene 6261 853 1639 2134 a Data determined b y P. Wetterau, Callender-Carnell fellow a t Lehigh University. Solvent
Vol. 18, No. 12
Modification of M e t h o d f o r High Molecular Weights
Thickened linseed oils with molecular weights of 2000 or upwards, as determined by the freezing point method using benzene or similar solvents, dissolve but slowly in the cold solvent. With dark-colored oils or enamels such as those made with relatively high proportions of iron driers, it is very difficult to tell when the sample is completely dissolved, although filtration through dry paper can be used as a test after the molecular weight determination has been completed. Jelly-like lumps left on the paper would indicate incomplete solution and consequent molecular weight values which are too high. I n order to facilitate solution and insure complete solution, the oil was subjected to extraction in a n alundum cup in a Soxhlet extractor. The method in detail is as follows: Three and five-tenth to 4.5 grams of the thick oil, weighed only approximately, are introduced into a n alundum cup. The weight of oil taken should be great enough t h a t after extraction the solution obtained contains more than 1 gram of oil for each 25 cc. (21.8434 grams) of benzene. The oil is extracted by means of benzene in a Soxhlet extractor. The benzene should siphon over eight to ten times for thick oils or until no undissolved oil is left in the cup. A residue consisting of “skins” or of mineral driers may remain. A 10- or 25-cc. portion of the solution is introduced into a. weighed porcelain or silica evaporating dish, the benzene evaporated on a steam or water bath, and the residue of oil promptly weighed. The specific gravity of t h e benzene and oil must be known or determined. For many purposes the variation in specific gravity of the oil introduces such a small correction that it need not be determined afreshfor each sample but may be taken from previous data for the product under examination.
From the weight of residue and the specific gravities of the oil and benzene, the weight of benzene in a 25-cc. portion of the solution may be calculated. The solution obtained in the Soxhlet extractor is then diluted so that 1 gram * 1 per cent of oil is contained in 25 cc. (21.8434 grams) of benzene. The freezing point of this solution is then determined and the molecular weight of the oil calculated from the freezing point depression. Typical results obtained by this method are given in Table 11. T a b l e I1 -RUN Sample 1 2 3 4
‘2-1-7 Mol. wt. 1678 1738 1841 1997
--RUN Sample 1 2 3
C .-2
Mol wt. 2408 2683 2858
The samples were taken a t half-hour intervals in latter stages of runs in which 1 gram of lead oxide and 4 grams of umber were present as “driers” in 500 grams of linseed oil. The following is a sample calculation: Assume t h a t evaporation of the benzene from a 10-cc. sample of the solution furnished a residue of 0.4850 gram and t h a t t h e densitv of the oil is 0.97. The volume of oil = 0.485 gram = 0.5 cc. 0.970 gram/cc. The volume of benzene in 10 cc. of solution is 10 - 0.5 = 9.5 cc. Since the specific gravity of benzene is 0.8737, the weight of benzene in 10 cc. is 9.5 X 0.8737 = 8.3002 grams. In 50 cc. of solution there are 50/10 X 8.3002 = 41.5010 grams benzene and 50/10 X 0.4850 = 2.4250 grams of oil. We want for 1 gram oil, 21.8434 grams benzene, or for 2.425 grams of oil we want 2.425 X 21.8434 = 52.9702 grams benzene. Benzene t o add = 52.9702 grams 41.5010 grams 11.4692 grams = = 13.1 CC. 0.8737
If, therefore, 13.1 cc. of pure benzene are added to 50 cc. of the solution from the Soxhlet extraction, the resulting solution will contain 1 gram of oil for each 25 cc. (21.8434 grams) of benzene.
I N D U S T R I A L A N D ENGINEERING CHEMISTRY
December, 1926
of Molecular Weight Increase of Oil w i t h a n d w i t h o u t Additional S u b s t a n c e s c MOLECULAR WEIGHT After After After After Temp. Up to 0.5 1.0 1.5 2.0 Nature of run C. heat hour hour hours hours 250 g. Sinolenic acid in open 13-cm. casserole 422 443 442 441 446 232 300 g. linseed oil 30 g. linolenic acid in open 13-cm. casserole 293 700 836 1016 1118 1377 300 g. linseed oil in open 13-cm. casserole 775 818 293 960 1038 1116 300 g. linseed oil 30 g. glycerol in open 13-cm. casserole 843 996 1115 293 781 1229 928 1104 300 g. linseed oil 30 g. F e salts of free fatty acid in open 13-cm. casserole 293 1451 1712 500 g. of linseed oil 15 g. sulfur in open 13-cm. casserole 1161 293 1010 1222 500 g. linseed oil in 3-liter flask in atmosphere of steam 293 800 860 988 500 g. linseed oil with break developed 1.0 g. PbO (I,) in open 13-cm. casserole 293 952 1458 1254 2117 500 g. linseed oil with break removed 1.0 g. PbO (I,) in an open 13-cm. casserole 293 947 300 g. linseed oil 30 g. linolenic monoglyceride in an open 13-cm. casserole 293 726 800 895 1057 1221
1247
Table 111-Rate
Run 47 48 49 50 51 52 53 54 55 56
+
+ ++
+
++
Further Evidence for Condensation Reactions
Table I11 shows the rate of molecular weight increase for a number of runs of oil alone and with admixture of other substances which it was thought might throw some light on the mechanism of reactions occurring during the boiling process. It was thought that glycerol might play some part in the reactions, but comparison of runs 49 and 50 shows that the rate of molecular weight increase was less for linseed oil containing 10 per cent of free glycerol than for linseed oil alone. Slight variations in the rate of molecular weight increase on different clays with otherwise identical or duplicate runs had led to the conclusion that the relative humidit.y of the air had a slight effect on the boiling. There was also previous evidence that the boiling process included some reactions which were of the nature of condensations in which water, acrolein, and other substances were eliminated from the oil. To test the effect of humidity in a direct manner and at the same time furnish some evidence bearing on condensation re:tctions, 500 grams of oil were heat'ed in a flask into which steam was constantly passed for 3 hours (run 53). It was thought that this steam would hinder elimination of water from the oil and thus retard condensation reactions if any equilibria were involved. As shown by comparison of runs 49 and 53, the rate of molecular weight increase in an atmosphere of steam is very much retarded, thus showing that the humidity of the air is a factor affecting the boiling of linseed oil. The retarded action in this run seems to furnish evidence that water is being evolved from the oil and that the chemical actions occurring are dependent on the removal of this water. Equilibria are indicated. Table IV-Data on R u n 57 Time of heating Pressure Water evolved Sample Hours Min. Mm. Gram Mol. wt. 770 1 Start 25 0.1037 2 1 20 25 0.4355 860 0.3424 988 3 2 50 22 4 4 5 25 0.2136 1161 0.3595 1255 5 5 35 22 6 7 5 14 0.2846 1500 0.0830 1609 7 8 20 22 8 9 50 17 0.1809 1678 0.0480 1752 9 10 35 18 Total water evolved 1.9474
-
Iodine number 160.7 163.1 136.0
125.5
More direct evidence showing condensation reactions is furnished by run 57, results for which are given in Table IV. I n this run, the apparatus for which is described under Method, the oil was heated a t 180-190' C. under 14 to 25 mm. pressure until calcium chloride tubes between the oil and the pump showed no appreciable gain in weight over two half-hour periods. This required 5 hours. The oil did not change much in this preliminary treatment, as is evidenced by the fact that the molecular weight of the preliminary oil after 5 hours' heating is 770. The temperature was then raised to 293' C. and heating continued until the oil was very thick (mol. wt. 1712). Water was steadily evolved. The calcium chloride in that leg of the first calcium chloride tube nearest to the oil became quite wet. I n a previous preliminary run anhydrous copper sulfate between the hot oil and the vacuum pump became greenish blue.
After After2.5 3.0 hours hours. 443 449 1672 2126 1459 1736 1379 1560' 1860 1252 1132 2111 1660
2323.
Not only was water evolved but it came off steadily as the boiling progressed a t the constant temperature (293' C.). As shown by Table IV, the weight of water evolved was, measured during a molecular weight change from 770 to 1752. One unexpected result is rather significant. The larger amounts of water were given off in 1.5-hour periods, in which the molecular weight change was small; the smaller amounts were given off in periods in which the molecular weight change was relatively greater. This suggests consecutive or concurrent reactions, at least one of which eliminates water. The gases evolved were bubbled through aqueous calcium hydroxide acid and sodium sulfite solutions enroute to the pump. Tests of the acid sodium sulfite solution with phenylhydrazine a t half-hour intervals showed that acrolein was also evolved during the entire boil. The rate of evolution of acrolein was not measured in this run. The uniform rate of molecular weight increase with coincident steady evolution of water, acrolein, and other substances observed in this experiment indicates rather conclusively that the reactions occurring during the boiling of linseed oil include those of the type generally classified as condensations. The evolution of condensation products is not, however, the only evidence of reaction involved. The iodine number diminished from 160.7 to 125.5-a decrease of 35.2. Any theory of the mechanism of the reactions must take account of this. Among others, the theory put forward by Salways suggested that it might be due to coupling of molecules of fatty acid at the double bond. Fahrion regards the double bond in the 15 position as the most reactive one. If Salway's premise is correct, then rate of molecular weight increase should be speeded up by having free fatty acid in the oil. This had previously been found true.6 The presence of 10 per cent of pure linolenic acid caused a. great rise in the rate of molecular weight increase, as can be seen by comparing runs 48 and 49. However, when pure linolenic acid was heated alone at 232" C. for 3 hours, no appreciable increase in molecular weight took place. A lower temperature than 293' C. was used to prevent rapid vaporization of the linolenic acid (b. p. 229' C. a t 16 mm.). Previous results with linseed oil alone and linseed oil with fatty acids had shown that the change a t 232' C.; though slower, was considerable in 3 hours. Impure fatty acids alone when heated at 232' C. showed a rapid molecular weight increase, but the pure linolenic acid alone did not appreciably gain in molecular weight. This suggests that the decrease in iodine number might be due to coupling of other bodies. The mono- or diglycerides or other bodies present in the oil during condensations might conceivably form addition products as well as condense. An attempt was therefore made to prepare a mono- or diglyceride of linolenic acid. The product finally obtained was a glyceride; the ultimate analysis indicated i t to be the monoglyceride with a small amount of diglyceride. Its molecular weight as determined by the freezing point method in benzene is 465. That of the monoglyceride is 352, of the diglyceride, 8
J SOC.Chcm. I n & , 39, 324T (1920).
I N D U S T R I A L A N D E,VGINEERING CHE,MISTRY
1248
612. It is inferred from similar work that some association of solvent molecules takes place. If 1 mol of benzene associates with 1 mol of monoglyceride, the molecular weight as 78 = 430. When 10 per cent determined would be 352 of this glyceride was added to linseed oil and the mixture heated a t 293” C. (run 56), the rate of molecular weight increase was greater than that observed in any other run with this shipment of oil. It will be noticed from Table I1 that for a molecular weight change of 942 the iodine number decreased from 160.7 to 125.5-i. e., by 35.2. If the reaction taking place consisted entirely in addition of molecules of linolenic acid at a double bond, as indicated by the reaction
+
CHzOCO(CH3 I C H ~ H C H ~ C H = C H C H Z C H - C H C H Z C H ~ I CHOCO-R HOCOR = I CHzOCO-R CHzOCO(CH2) CH=CHCHzCH-CHCHzCHzCHCHnCHI
+
I I
CHOCO-R
I
-COR
CHiOC-R
then the molecular weight change corresponding to the decrease in iodine value could be calculated as follows: 2 I = 1 CI~H~QCOOH 2 X 126.93 = 278.24 278 24 100 grams of oil absorb 35.2 grams of iodine = 35.2 X -=
253.86 38.58 grams of linolenic acid. The molecular weight corresponding to the iodine number of 160.7 is 770. 1 mol of glyceride of mean molecular weight 770 770 would therefore add - X 38.58 = 297 gram of linolenic acid. 100 The molecular weight of the body produced would be 770 297 = 1067.
+
If instead of molecules of linolenic acid molecules of linolenic glyceride could add a t the double bond, the molecular weight change corresponding to an iodine number change of .35.2 would be 770 E2 X 35.2 = 121 for 100 grams of oil or - X 253.86 100 931.7 for 770 grams of oil,
121 =
which is the mean average molar weight of the glycerides present. This would bring the molecular weight up to a value of 770 931.7 = 1701.7, which compares well with the value found. It might be concluded that this addition was the only type .of reaction involved. But this takes no recognition of the water or other substances liberated or of the fact that the rate of molecular weight increase is much more rapid when the ail contains free fatty acids or linolenic monoglyceride, despite the fact that these have lower molecular weights than the triglycerides. Run 53 is also significant in pointing out that the reactions depend partly on evolution of water. It seems, therefore, that the reactions taking place in the boiling of linseed oil cannot consist entirely of addition at double bonds, but involve t o no small extent condensations between mono- or diglycerides and other compounds present in the oil used or formed during boiling. The molecular weight of the raw linseed oil used is 720. The theoretical value for linolenic monoglyceride is 872.72 and those of the other glycerides present not far different from the value. It is further probable that the values obtained in the various solvents are too high, owing t o association of the solvent. The value obtained for raw oil, therefore, suggests that raw linseed oil may contain mono- or diglycerides formed during pressing or storing. It was noticed in run 53 that when linseed in heated oil was in
+
Vol. 18,No. 12
an atmosphere of steam the oil “broke,” despite the fact that the oil used gave no break at 300” C. when dry. This is of interest in commercial oil boiling in connection with the fact that umber or other materials used contain some water and also that moisture might get in during storage. It was thought that the development of “break” during the boiling might adsorb part of the driers, thus rendering them ineffective, and that this adsorption might not be a constant factor from day to day, and might therefore account for some of the irregularities noticed in practice. The effect of the formation of break is shown in runs 54 and 55, which contain 1 gram of lead monoxide in 500 grams of oil. I n run 54 the break was developed by means of steam while the oil was being raised to 293 and before the lead monoxide was dissolved. The samples of this boiling on standing showed three distinct layers, the bottom layer being a very thin layer of a material of high specific gravity. I n run 55 the break was developed by means of steam and the oil then filtered. One gram of lead monoxide was then introduced into 500 grams of oil and the boiling carried out as in run 54. The samples after the 1-hour sample did not show any deposit. As can be seen from Table 111, the rate of molecular weight increase is very much greater in the oil from which the break had been removed by filtration. Since sulfur, selenium, and tellurium belong to the same family with oxygen, it was thought that sulfur might be taken up by the oil. This idea was interesting in opening up possible avenues of study of the various reactions occurring. Since the heat of formation of hydrogen sulfide is very much less than the heat of formation of water, we felt that condensation reactions involving the formation of hydrogen sulfide from SH bodies would proceed more slowly than the corresponding condensations between OH bodies. The molecular weight would therefore be expected to increase more slowly. Accordingly, 15 grams of flowers of sulfur and 500 grams of raw linseed oil were heated a t 293” C. (run 52). The sulfur dissolved rapidly, furnishing a clear reddish brown liquid, and it is interesting to note that the molecular weight is high (1010) when the oil was “up to heat.” The molecular weight change during the 3 hours’ subsequent heating in air was small, much less than for oil heated in air without addition of sulfur. Work in Progress
The run with 3 per cent of sulfur suggests trying further runs with more sulfur and with selenium and tellurium under various conditions. Since we are also engaged in trying to prepare pure mono- and diglycerides of linolenic and linoleic acids, me will try to prepare and study some of the corresponding sulfur compounds. Further work is in progress in which the gaseous products of the reactions are being determined and the weights of these considered in relation to the molecular weight changes for linseed and China wood oils.
Yale Develops New Process for Making Metal Foils A process for making metal foils less than a millionth of an inch in thickness has been developed by the research workers a t the Sloane Physics Laboratory of Yale University. In confirming the report William F. G. Swann, director of the Sloane Laboratory, said that those foils are perfectly uniform and are almost completely transparent. Joseph E. Henderson is perfecting the technic and is using the foils in an investigation of the relative ease with which electrons of different velocities can pass through metals, and in certain problems pertaining to soft x-rays. More than twenty-five separate investigations are in progress a t the Sloane Laboratory by members of the Yale faculty, professors from other institutions who have chosen t o devote their sabbatical leave to research work a t Yale, research fellows, and graduate students.
