ill. K. E V ~ R F TA T N D FAYSHEPPARD
1792
water, alcohol arid hydrochloric acid, and practically insoluble in ether, ethyl acetatc, etc. They arc unusually difficult to analyze. Their proprrties are described ill Table 111.
Summary A series of hydroxy- and alkoxyphenylethyldi-
[CONTRIBUTION FRO51 THE BIOCHEMISTRP
Vol. BO
a of ~lydroxy-and alkoxyphenylethyltrimethylamine chlorides (18 compounds) have been prepared, all by the same sequence of reactions, which appears to be general. The preparation of the intermediates also is described. TYCKAFIOF, XEWYORK
LABORATORY OF THE
RECEIVED MAY20, 1938
UXIVERSITY OF OKLAHOMA
MEDICAL SCHOOL]
Relations of cis-trans Isomerism to Asymmetric Oxidation of Sugars* BY
M. R. EVERETT AND
Richtmyer and Hudson2 receiitly reported the interesting phenomenon of unequal quantitative reduction of alkaline copper reagents by d- and Z-forms of aldoses, when d- or 1-tartaric acids were used in Shaffer-Hartman-Somogyi reagents. IVe have now completed similar experiments with d - , 1- and meso-tartrate Folin-Wu reagents3 The I-tartaric acid employed was prepared by the method described by Richtmyer and Hudson, and had thecorrect [ a I 2 k-I-&*. Thed- and mesotartaric acids were Central Scientific Co. and Eastman Kodak Co. products with [ c t l l Z o ~+ I 4 and +0.05', respectively. The sugar solutions and special reagents4 were freshly prepared, immediately before analysis. Several concentrations of each sugar were investigated and compared with suitable d-glucose standards. Determinations were made under strictly comparable conditions. The results reported in Table I are the reducing equivalents relative to d-glucose and d-tartrate reagent taken as 1.00 and are averages of four or more determinations. Corrections for moisture content of sugars and proportionality of the analytical method have been applied. The reagents employed by us appear to be somewhat morc sensitive to spacial configuration than the I) Aided by a grant from the Rewarch rund of the lintvercity of Okldioma Medical School. The authors a h wish to acknowledge the following gifts. wd-mannohcptose and d-mannoketohepto% from Dr C. S Hudson and d-galacturonic acta from Dr Karl Ltnk (2) Richtmyer and Hudson, THISJOURNAL, 98, 2540 (19%) (3) Polin and Wu, J Brol Chenz , 67, 357 (l'I20) (4) For each 20 cc of aikaline tartrate reagent we used 0 1696 g of tartaric acid ( d - , 1- or meso-), 3 cc of distilled water, 2 cc of 1 125 3 sodium hydroxide and 15 cc of buffer solution (containing 4 667% sodium carhonste and 1 4 6 7 % sodium bicarbonate). 18 cc of this solution %as then m r e d with 2 cc of ,5y0 cupric sulfate solution The iinie of heatiug u a s uniformly scven and one-half nunutes on the boiling water-bath 'The I'olirii acid redgetit w a s uced for color drvelopment fey) Foliii 7 Bro2 Cirpirz , 82, 97 (l'i2Ci)
FAY
SHEPPARD
copper reagents used by Richtmyer and Hudson, but both investigations show that d-forms of arabinose, fructose and mannose and 1-forms of fucose and rhamnose select the I-tartrate reagent ; d-forms of galactose, lactose and mannoketoheptose and I-arabinose select the d-tartrate reagent; and d-forms of glucose and xylose show little selectivity. For sugars not investigated by Richtmyer and Hudson, we find that d-forms of maltose, melibiose and ribose select the I-tartrate reagent ; d-forms of galacturonic and glycuronic acids, mannoheptose and sorbose select the dreagent ; d-forms of cellobiose, gentiobiose, glucosamine and lyxose and 1-forms of ascorbic acid, sorbose and xylose are not very selective. Our i;lieso-tartrate reagent is reduced more than either of the trans reagents by d-forms of cellobiose, gentiobiose, glycuronic acid, lactose, maltose, Qmannoheptose, mannoketoheptose and mannose. It is evident from these results that asymmetric oxidation of sugars bears no simple relation to ordinary optical or planar isomerism, and is influenced by extraplanar molecular relations (group substitution, cis-trans relations, etc.). The reducing values of Table I are definitely related to the cis-trans classification of sugars suggested by the a ~ t h o r s . ~ This , ~ arrangement of sugars in eight cis -trans groups, corresponding to the eight aldopentoses, is reproduced in Table 11. Oxidation in alkaline solution is complicated by mutarotation and epimerization, and recently IsbelP has shown that mutarotation is correlated with cis-trans configuration. The behavior of (0) Everett and Sheppard
Pvoc
ORla Acad
S c i , Xovember,
3'93fl
( 7 ) Ererett and Sheppard
'The Oxidation of Carbohydrates in
4cirl 5olution ' l'iii~ir.ity of Oklahoma Medical School Monograph 1 ' 4 X t i l 4 r ~ l l .I R i ~ i o i r l r L ~ i lR,ti V n ~ d m d \ 18, 505 f ? O ? i )
Aug., 1938
cis-truns
TABLE I RELATIVE REDUCINO EQUIVALENTS
ISOMERISM AND
ASYMMETRIC OXIDATION
OF SUGARS (d-GLUCOSE
= 1.00) Sugars d-Xylose d-Xylose 1-Xylose 2-Xylose d-Lyxose d-Lyxose d-Arabinose d-Arabinose d-Arabinose 1-Arabinose 2- Arabinose 1-Arabinose &Ribose &Ribose d-Ribose d-Glucose d-Glucose d-Mannose d-Mannose d-Galactose d-Galactose d-Galactose d-Glucosamine.HC1 d-Glucosamine. HC1 d-Glycuronic acid d-Glycuronic acid d-Glycuronic acid d-Galacturonic acid. HtO d-Galacturonic acid, Hz0 d-Sorbose d-Sorbose 2-Sorbose I-Sorbose d-Fructose d-Fructose I-RhamnoseHzO I-Rhamnose.Hz0 I-Fucose 2-Fucose a-d-Mannoheptose a-d-Mannoheptose d-Mannoketoheptose d-Mannoketoheptose 1-Ascorbic acid 1-Ascorbic acid Gentiobiose Gentiobiose Cellobiose Cellobiose Maltose.Ha0 Maltose.Ht0 Melibiose.2HzO Melibiose.2HzO Lactose HtO Lactose.Hz0
.1
.2 .4
0.97 .985. .9G5 .99 .96 .96 .69
1.02 0.91 .995 .94 1.05 ,955 1.05 .955 0.94 .855 ,935 -88 .92 .76 .88 .78
1.14 1.18 1.16 1.13
.1
.73 .69 .84 .83
.765 .77
1.14
1.08 .70 .73 1.00 1.00 0.70 .G8
.2 .4 .1 .2 .2 .4 .1 .2 .4 .1 .2 *1 .2 .4 .2
.62 .655 1.00 1 .oo 0.45 .57 .78 .79
.655
,135 ,715 -45
.4
.65
.46
.56B
.1 .2 .1 .2 .1 .2 .3 .6 .B .G .2 .4 .2 .4
.7G
.70 ' .71 ,795 .83 ,965 .975 .40 .41 .53 .48 ,515 .51
.745 ,785
.99
.785
.95
.77 .81 .71 .SO
.80 .80 ,85 .89 .93 .31.5 ,376 .45 .42 .60 .63 .68 .70 .55 .925 .54 .535 .535 .52 .415 .425 .495 .49 .475 .475
.98 .975 .56
.63
.755 .72 .73 .755 .95 ,945 .58
.83 .94 .98 .385 .44 .475 .45 .67 .66 .75 .73 .55 ,945
d-Xylose group
1-Xylose group
d-Isorhamnose d-Sorbose d-Glucose 1-Idose
1-Isorhamnose 1-Sorbose 1-Glucose d-Idose
d-Lyxose group
I-Lyxose group
d-Rhamnose d-Tagatose d-Mannose Z-Gulose
2-Rhamnose 1-Tagatose 2-Mannose d-Gulose
1.00 0.97 .97
.64 .63 .79 .80
FORMS tram Forms
2,3-cis Forms
.685
.86 .87
1793
TABLE TI CLASSIFICATION OF PENTOSES, METHYLPENTOSES ANI) HEXOSES ACCORDINGTO cis-trans ISOMERISM OF PYRANOID
Sumner Folin-Wu tartrate reagents reagent 2meso (0.5 mg.:cc.) mg./cc. d0.1 .2 .1 .2 .1 .2 .1 .2 .4
OF S U G A R S
.33' .91
.87
3 , 4 - ~ kForms d-Arabinose group
1-Arabinose group
1-Fucose d-Fructose d-Altrose Z-Galactose
&Fucose I-Fructose Zdltrose d-Galactose 2 , 3 , 4 - ~Forms i~
d-Ribose group
I-Ribose group
Epifucose d-Psicose d-Allose I-Talose
Epirhodeose 1-Psicose 1-Allose d-Talose
Type formula (1) C/l0\C
.99
1
I
(5)
.98
1.00 0.83
Carbon 1 is the characteristic potential carbonyl.
carbon a t pyranoid positions 1 or 5 (Table 11) limits mobility of the cyclic oxygen, allowing .BO opposite planar arrangements of cyclic oxygen .1 .ti5 0.44 .90 .4 in aldohexoses of each cis-tram group and the .2 .54 .57 .GO resultant alteration of cis-trans influences at .535 .57 .4 .2 .53 ,666 .83 pyranoid position 5. .4 .51 ,645 In our experiments cis sugars with opposite .2 .46 .51 .16 .4 .44 .51 extraplanar relationships showed opposing selec.52 .50 .2 .63 tivity for asymmetric oxidizing agents. d-Man.4 .50 .50 .45 .535 .2 .77 nose and 1-rhamnose, with 2,3-cis hydroxyls trans .445 .53 .4 to cyclic oxygen, selected I-tartrate reagent, and Used 1 mg./cc. d-galactose, an extraplanar opposite as far as pentoses with alkaline reagents (Table I) follows positions 2, 3 and 4 are concerned, selected the epimeric relations ; thus, xylose and lyxose (epi- d-reagent (Table I). The oxidation of sugars is meric trans types) react almost equally with d- thus influenced by spacial configuration in three and 1-tartrate reagents, whereas ribose and arabi- dimensions. nose (epimeric cis types) select opposite enantioIn Table I11 we have listed molar reducing morphs of reagents. .Six and seven carbon sugars equivalents of sugars and their percentage deviabehave qualitatively like related epimeric pentose tions from the d-glucose standard. The molar types, but deviate quantitatively. Additional equivalents have been calculated as percentages ,605
1.oo
M. R. EVERETT AND FAYSHEPPARD
1794
Vnl. 60
rIr MOLARREDUCING EQUIVALEN~ s AND PERCENTAGE DCVIATION~ OF SUGARS RELATIVETO d-GLucosE TABI E
5ugari
&Xylose &Xylose 1-Xylose 1-Xylose d-Glucose d-Glucose d-Glucosarninc d-Glucosaminc d-Glycuronic acid d-Glycuronic acid d-Glycuronic acid d - Sorbow d Yorbosc 1 Sorbosc l-,Sorhow
d-lyosc d-Lyxose d-hlannose d -Mannose Z-Rhamnose l-Rhamno5c d-blannoketoheptoie ci-Mannoketoheptorc d- Arabinosc d-Arabinose d-Arabinow 1-Arabinose L-Arabinose /-Arabinose d-Galactose d-Galactose d-Galactose &Galacturonic acid &Galacturonic acid &Fructose d-Fruciosc 1-Fucose &Fucose a-d-Mannoheptose a-d-hfannoheptosc
d-Ribose &Ribose &Ribose
Fohn-WU'u equivalents-----d Reap i Reag
m
Re14
81 82 805 825 1 00 I 00 0 c)L 97
0 8,; 83
-
70;
io
7ii 7% 795 795 1O(I I 00 0 87.5 903 1 025 1 113
4
Xf I
77 70
0 7 1-3
0 1
>
I
I 2 1
2 1 2 1 )
1
..,
I I1
X75 875 98 97.5
1)
783 7%-
S,"!
71 -!I,? x, 1
is
->
Y( I h0 45
7% 3ri
71 'iJ io
4
r) (
t),$
(:8
7
80
1
so
a
Y
I
x
I
??E
( 3
_
-
4 - I - 7 - 5 + 2 2 3 - 3
+ +1 + 0 r, + 9 + (i +9 + 0 tj +2 +2 +2
Deviation %-
Ti
Z
-15 - 17 -12 5 -12 5 0 0 - 5 - 4
-24 -22 -20 5 -20 5 0 0 -12 5 - 9 5 + 2 5 2
-23 6 - 14 -24 - 20 -20 - 15
+
lo.-)
2 k
70 8
1 1 ,? 71
V5
+X
71)
ST,
+l"
I 2 4 1 2 3 1 2 4
--.),
I ,
i;:i
71
1i.i
-15 5
-23 -27
-37 - 35
+lI 5
-28 -27 6
-36
(i4
01
755
+13
-21
-24 5
72 68 fiC6 94
$24 +22 5 - 7 5 -4,5 - 7 - 0 +10
-28 -32 -33 6 - 6 - 2 - 57 -59 -22 -23 -42
_.)
)
--
- 85 - 3 6
72 -'I
-
1-1)
04 t i1
,-
1 2 3
89 Y3 11 38 70
97.5 48 44 00
98 43 41 7s
73.5
393
c-
it
+14
1 2
52
70 69
58
-17
-30 -31
I-
.i4
I-Ascorbic acid 1 I-Ascorbic acid 4 Gentiobiose 2 Gentiobiose 4 Cellobiose 2 Cellobiose 4 Maltose 2 4 Maltose 2 Melibiose 4 Melibiose Lactose 2 4 Lactose " Concentrations as in Tshk
I_
iI I
b?
bJ
56 013 1 03 1 02 1 02 0 99 2"3 8.5 104 103 0 95 95
GI 56 92 1 03 1 02 1 01 ri 97 92 88 109 105 0 90 80
56 965 1 08 1 08 1.24 1 23 1 02 1 02 105 105 1 07 1
06
-
2 0
-60.5
-2 -1 -5 - 3.5
97
-
99
-1
1 16
+16
U.94
-6
.95
-5
.97
-3
3
-36
-22
*
2 t
95
575 7s 79
-23 -23 5 - 3 5 - 2 3 - 52 -A. -30 -20 5
I?
98
.)I
63 53 34 965
2
0.395
.965
:IS
0
- 5
1 00
-29 -27 -30 -32 -61 -55 5 -12 5 - 15
.12
-11
-
98
.99
( 8
_-
095
-25 5 -21 5 -21 3 - 17
-20 -20 -44 -37 -59 5 -5s 5 -21 - 18
7 I
Sumner Deviation, equivalents C:
.91
+ 2.5 -1 -9
.97
-3
.90
-10
.45
-55
+ 8
1 25
+25
+ 8 4-24
1 57
+57
1 52
f52
1 32
+32
1%
+54
1 025 0.99
-39
0 + 2 5 0 0 1 2 - 9
+ + -
3
- 5 - 2 5
+
-+ G
-44 - 5 5 + 3 + 2 $ 2 - 1 - 8 - 12 + 9
+ 5 - 5
-
5
-44 - 4 5
$23 + + + f
2 2 5 5
7, t-6 f
Aug., 1938
CiS-tYUnS
ISOMERISM AND
ASYMMETRIC OXIDATlON O F SUGARS
of the reduction of d-tartrate reagent by dglucose. The deviations are defined as follows: X, representing planar influences for trans reagents, is the molar equivalent for the d-reagent minus that for the 1-reagent. Y, representing extraplanar influences for trans reagents, is the equivalent for that truns reagent most redured by any particular sugar minus the equivalent of the trans sugar, d-glucose, for the d-reagent. 2, representing the total influence upon the meso reagent, is the equivalent for this reagent minus that of d-glucose for the d-reagent. The deviations of monosaccharides with trans tartrate reagents reveal that epimerized d-glucose, d-glucosamine, d-xylose, d-lyxose and 1-sorbose are most symmetrical in a planar sense, since their X deviations approach zero. However, these trans sugars have average Y deviations of 0, -4.5, -16, -20 and -17.5, respectively. The presence of the CHzOH group a t position 5 in d-glucose therefore increases the trans characteristics of the pyranoid molecule. In the three-dimensional strainless model of d-glucopyranose, viewed from the plane of the ring, the additional hydroxyl group of carbon 6 balances the other oxygens in both a planar (right and left) and an extraplanar (polar) sense. In 1sorbose and d-xylose models the oxygens are well balanced in a planar sense but not in an extraplanar sense, since there is no CHzOH group at pyranoid position 5 . Lengthening the carbon chain therefore results in appreciable alteration of extraplanar influences, which are interrelated with planar influences. The interrelation is demonstrated by the fact that d- and 1-isomerism can be represented alternately as a polar phenomenon relative to the pyranoid cyclic oxygen. Inversion of strainless pyranoid models of 1-sugars, whose ring systems are thus viewed as identical with those of d-sugars, reveals polar interchanges between carbons 1 and 5 and between carbons 2 and 4. Relating sugar antipodes in this fashion facilitates comparisons which are difficult with the usual right and left disymmetric formulas. Since planar and extraplanar influences have a common space coordinate, the substitution of CHzOH or COOH groups a t pyranoid positions 1 or 5 affects both cis-trans (Y) and planar (X) influences, the latter causing a “shift to the right” in reagent selection by ketoses and alduronic acids. (Compare these with related aldohexoses in Table 111.)
1795
Of the monosaccharidesstudied, d-glucose gives rise to the most symmetrical three-dimensional epimeric system. In a planar sense, epimerized d-glucosamine, d-xylose, 1-sorbose and d-lyxose are just as symmetrical, while d-galacturonic acid, d-ribose, d-arabinose and d-galactose (3,4-cis types) are most unsymmetrical. In an extraplanar sense, d-glucose, d-glucosamine and d-fructose are most symmetrical, the latter evidently going to a balanced epimeric system by a Lobry de Bruyn rearrangement; methylpentoses are most unsymmetrical and reduce less copper than other sugars. The trans arrangement of epimerized d-glucose and its relatives, d-fructose and d-glycuronic acid, is also the most favorable influence for reduction of meso-tartrate reagent (Table 111). With aldopentoses, extraplanar deviations (Y) for trans reagents are uniformly less than the total deviations (2) for the meso reagent, and the latter is not always reduced equally by planar antipodes. (See xylose and sorbose in Table 111.) Analogous smaller deviations occur even with the nonasymmetric Sumner reagent. The fact that d-glucose reduces meso reagent 23% more than does d-xylose, indicates again that mutarotation and epimerization of the trans pentose are insufficient to balance its molecule in all dimensions and that terminal CHzOH groups are active. Certain sugars which are not optical antipodes reduce the meso reagent equally (dxylose and d-sorbose, also d-mannose and &galacturonic acid) ; others (d-glucose, I-ascorbic acid, the methylpentoses and ketohexoses) have approximately equal Y and 2 deviations. More detailed correlation is problematical because the concentration of certain sugars affects their oxidation. Notable instances are d-mannose and d-glycuronic acid with trans tartrate reagents (Table 111), but the most remarkable example is a-glucosan with d -tartrat e. The P-glucose-glucosides, cellobiose and gentiobiose, resemble the parent monosaccharide with trans reagents, but differ in their effects upon. meso reagent. Cellobiose presents a space arrangement for meso oxidation markedly superior to that of any sugar studied (Table 111). The marked reduction of Sumner’s reagent by disaccharides is significant. We have called attention previously to the usefulness of Sumner/Folin-Wu ratios of glucose equivalents for differentiating sugars in s d u t i ~ n . ~ * ~ (9) Everett, Edwards and Sheppard, J . Biol. Chcm., 104111 (1934).
BRUCEB. ALLENWITH HENRYR. WENZE
1796
The present study provides a more scientific basis for these ratios and correlates them with sugar structure.
Summary 1. Quantitative studies have been made of the
oxidation of twenty-five sugars by d-, 1- and mesotartrate modifications of the Folin-Wu alkaline copper reagent and by Sumner’s dinitrosalicylate reagent. 2. ‘The behavior of reducing sugars is determined by the entire sugar molecule, but intra-
[CON1RIBU~TONNO.
VOl. BO
molecular influences can be conveniently classified as planar and extraplanar, &-trans relations being important extraplanar influences. 3. Epimeric sugars have related behaviors, the glucose epimeric system being most symmetrical in all dimensions. CiS-sUgarS are most UnSymmetrical in a planar sense and methylpentoses in an extraplanar sense. 4. meso-Tartrate reagent is susceptible to both planar and extraplanar inflUellCeS. OKLAHOMA CITY, OKL.4.
134 FROM THE DEPARTMENT O F CHEMISTRY, THE
RFCEIVFDJANUARY 7, 1938
UNIVERSITY OF TEXAS]
Hydantoins Derived from the Analogs of 1,3-Dichloroisoprog~xyexyethyl Methyl Ketone’ BY BRUCEB. ALLENWITH HENRYR. HENZE As a continuation of attempts to prepare substances possessing soporific properties, attention has now been directed to the synthesis of 5,3-disubstituted hydantoins in which one of the substituents is of the halogenoalkoxyalkyl type. In this Laboratory, the application of Bucherer’s2 hydantoin synthesis has been extended to the preparation of disubstituted heterocyclic compounds from such bifunctional carbonyl substances as alkoxymethyl alkyl (or aryl) ketones3 and phenoxymethyl alkyl (or aryl) ketone^,^ and has been shown to be advantageous over other accepted methods as regards yield, ease of execution, and degree of purity of product. In the present investigation, then, successful use of the Bucherer method in hydantoin formation from dichloroisopropoxyethyl alkyl or aryl ketones resulted in further extension of the application of the method. The ketones used were, with one exception, of the type (CH&l)&3-I-O-CH(CH,)COR where R represents alkyl and includes those groups which have been shown most active physiologically ; the preparation and characterization of these ketones have been described by us pre. v i ~ u s l y . ~Since the hydantoin derived from the phenyl analog of this series of ketones represents a halogenoalkoxyl derivative of Nirvanol (5ethyl-5-phenylhydantoin), a substance shown to (1) From a portion of a dissertation presented by Bruce B Allen to the Faculty of the Graduate School of the University of Texas in partial fulfilment of the requirements for the degree of Doctor of Philosophy, June, 1938 (2) B u c k e r and Lieb, J p m k t Chem , [2] 141, 5 (1934). (3) Rigler with Henze, THIS JOURNAL, MI, 474 (1936) (4) Whiiney with Henze, ibrd , 60, 1148 (1938). ( 6 ) A11m w i t h fTancr,. ibid , 59, 640 (1937).
possess useful medicinal value,6 it was deemed pertinent to effect its preparation and subsequent conversion into the corresponding hydantoin.
Experimental Preparation of
ol-1,3-Dichio~propoxyethyl Alkyl
Ketwes.--The eight ketones of this series which were employed were prepared from a-l,3-dichloroisopropoxypropionitrile by means of the Grignard reaction, these preparations having been described previously.6 Prepatation of ol-i,3-Di~MoroieopropoxyethylPhenyl Ketaae.-The general method as described for the preparation of the alkyl analogs was employed in the synthesis of this ketone. The Grignard reagent from 48.1 g. (0.31 mole) of phenyl bromide, 7.3 g. (0.3 atom) of magnesium in the form of turnings, and 250 cc. of anhydrous ether was treated with the solution of 45.5 g. (0.25 mole) of CY1,3-dichlwoisopropoxypropionitrilein a n equal volume of ether. Although vigorous reaction occurred throughout the nitrile addition, completion was ensured by refluxing for two hours over a steamcone. The addition product, which separated as a gray-colored solid, was decomposed in the usual manner with chilled, dilute hydrochloric acid and crushed ice; the resulting ether layer was separated, washed with dilute sodium carbonate solution and then water, and finally dried over anhydrous calcium chloride. The crude material was freed of ether by evaporation and then purified by two fractionations under 4 mm. pressure; the pure ketone is a colorless, almost odorless, quite viscous liquid which develops a deep red coloration upon standing, yield, 45.4 g. (69.6%); b. p. 169“ (4 mm.); dz0r 1.2356; n% 1.5398; MR calcd., 66.18;’ M R found, 65.31; y20 35.53 dynes/cm.; P calcd. (Sugden’s atomic constants), 538.4; P found, 516.0. ( 6 ) De Rudder, Cbrrn. Zcntr., 33, I , 26248 (1926); Poynton and Schlesinger, Lancet, 11, 267 (1929); Richer and Gerstenberger, A n . J Diseases Children, 40, 1239 (1930), Jones and Jacobs, J .4m Med Assoc., 33, 18 (1932). (7) This summation includes the exdtatiou value due to the G H s CO- grouping which has been determined for acetophenone by A i i w r r s and Biseolohr, J . prekf Chem., [2lU4, 20 (1911).
