SEPTEMBER 18, 1913] 
chemistry is not singular in this respect: as much 
might be said about controversies not yet settled 
which concern themselves with such every-day pheno- 
mena as the chemistry of the candle-flame or of the 
rusting of iron. 
It is a commonplace that Kekulé, to whom theo- 
retical chemistry owes so many fruitful suggestions, 
was of the opinion that substitution is not a process 
in which what may be called a direct exchange of 
radicals occurs, but is preceded by the temporary 
union of the molecules of carbon compound and 
addendum, followed by disruption into two new 
molecules, the substituted carbon compound being 
one of them. It is clear, then, from the point of view 
of Kekulé’s hypothesis, that some degree of unsatura- 
tion is to be looked for in all carbon compounds and 
in all addenda. Hence, the paraffin hydrocarbons 
which furnish derivatives only by substitution, and, 
under the older, more rigid view of valency pro- 
pounded by Kekulé himself, are typically saturated 
compounds, supply the exceptions to prove the general 
validity of the hypothesis that addition precedes sub- 
stitution. 
Before examining the case of these hydrocarbons, 
however, some advantage may be gained ii the be- 
haviour of other groups of compounds be examined in 
the light of the idea underlying Kekulé’s view. By 
reference to the literature, it is evident that since 
the beginning of this century attention has been con- 
centrated on the phenomena of substitution in the 
important group of carbonyl compounds, particularly 
the ketones and acids, which in many cases yield 
halogen substitution derivatives of one type. Thus 
methyl ethyl ketone when brominated in sunlight 
yields two bromoketobutanes of the constitution shown 
in the following formulz, and propionic acid with 
bromine and red phosphorus under Volhard’s condi- 
tions '° gives a-bromopropionic acid, 
CH,Br.CO.CH,.CH, and CH,.CO.CHBr.CH, » 
CH,.CHBr.CO.OH 
the halogen occupying what is termed the «-position 
with reference to the carbonyl radical. Why is sub- 
stitution easier in the methyl group when it is present 
in the ketone or acid than when it occurs in methane, 
is one question that may be asked. A second will 
inquire whether the carbonyl group has a directing 
influence, and, if so, by what means is it exercised. 
It has been supposed by Werner that the distribu- 
tion of valency is disturbed by the introduction of the 
oxygen atom of the carbonyl group into the molecule 
of the hydrocarbon; that this oxygen atom absorbs 
much of the valency of the carbon atom of the 
carbonyl group, leaving less to bind its neighbour or 
neighbours, which results in their having free valency, 
and thereby attaching substituents to themselves. This 
explanation, if accepted for the bromination of ketones 
and acids, also for the chlorination of ketones, does 
not account for the results recorded by Michael and 
by Montemartini in the case of carboxylic acids. 
Michael has found that the 6-chloro-, not the a-chloro- 
acid is the chief product (60-65 per cent.) when homo- 
logues of acetic acid are chlorinated?2; and Monte- 
martini states that if the radical CH occur in any 
part of the carbon chain the exchange of hydrogen 
for chlorine takes place in that position, however 
distant it may be from the carbonyl group of the 
acid.!* 
10 J. Volhard, Ay ‘ , cexlil, 4 . i 
i L. Van 7 slalantalg ‘Bull, nced. on meen aoe For the 
chloroketobutanes, cf. idem; Kling, Compt. rend. rgos, cxl, 312; Bull. Soc. 
chim., 1905 [iii], xxxiii, 322. 
12 A. Michael, Ber., 1901, xxxiv, 4035, 4045. 
fil, 260 Montemartini, Gazz. chim, ital., 1897, xxvii [ii], 368 ; 1898, xxviii 
NO. 2290, VOL. 92] 
NATURE Tone 
CH,.CH,.CHCI.CH,.CO.OH 
(Michael) 
CHy 
PEO. CH2,CH2.CO.OH 
H; ; 
(Montemaitini) 
At present there seems to be no clue to the reason 
why chlorine and bromine in these reactions behave 
alike towards ketones and not towards acids. 
An alternative explanation of this reaction, which 
has come to be widely accepted, is based on the re- 
markable property called desmotropy or dynamic iso- 
merism, which certain of these carbonyl compounds 
exhibit. A desmotropic compound may exist in two 
or more forms, and its peculiar isomerism is known 
to depend on the mobility of a hydrogen atom in the 
complex .CH,.CO. whereby an equilibrium is set up 
of the type: 
CH, COn—, CH C(O): 
Ketonic torm Enolic form 
Of these two forms, the enolic is the more un- 
saturated, and presumably the more reactive.** Lap- 
worth, making use of this desmotropic relationship, 
supposes that when the ketone reacts with halogen 
in dilute aqueous solution three changes occur which, 
for the case of acetone, may be represented by the 
following expressions :— 
CH CO:.CH. —> CH,.C(OH):CH, 
CH;.C(OH):CH,+Br, —~- CH,.C(OH)Br.CH.Br 
CH;.C(OH)Br.CH,Br —> CH,.CO.CH,Br+ HBr 
the first being one of slow enolisation, accelerated 
catalytically by halogen acid, leading to the produc- 
tion of an unsaturated compound, which then by rapid 
addition of bromine and subsequent elimination of 
hydrogen bromide conforms with Kekulé’s hypothesis. 
The intermediate compounds, it is true, have not been 
isolated, but a study of the dynamics of the reaction 
by Lapworth, and later by Dawson with his collabora- 
tors (using iodine instead of bromine), shows that this 
explanation is in harmony with the data obtained.'* 
When the reaction is applied to carboxylic acids under 
similar conditions, the view that it takes a similar 
course finds support from an investigation of the 
dynamics of the bromination of malonic acid in 
aqueous solution.'® 
Whether evidence drawn from reactions found to 
take place in aqueous solution is relevant when 
bromination is effected by heating a carboxylic acid 
with bromine and red phosphorus may be doubted. 
Certainly it seems to afford no assistance in account- 
ing for the course of chlorination in the acids 
examined by Michael and by Montemartini. Never- 
theless, Aschan employs the keto-enolic hypothesis 17 
to elucidate the results of a recent inquiry into the 
‘““mechanism"’ of the Volhard reaction 8; and it may 
be added that racemisation has been found to occur 
when laevo-valeric acid is brominated by Volhard’s 
method !*—a result which must follow if enolisation 
14 It may be of interest to note that the long controversy respecting the 
composition of ordinary ethyl acetoacetate CHy.CO.CH».CO.OEt, the first 
of these desmotropic compounds to be discovered, has been brought to an 
end by the isolation of each desmotropic form at temperatures sufficiently 
low to inhibit the desmotropic change. From re{ractometric observations 
with mixtures of the pure isomerides, Knorr concludes that this ester at the 
ordinary temperature contains about two per cent. of the enolic form, 
whereas from bromination experiments with the ester itself, which may pos- 
sibly be accompanied by a disturbance of the equilibrium, K. H. Meyer 
infers that the amount may be as much as seven percent. (L. Knorr, O. 
Rothe, and H. Averbeck, Ber., 1911, xliv, 1138; K. H. Meyer, Annalen, 
rgtr, ceclxxx, 222; K. H. Meyer and P. Kappelmeier Ser., 1o1r, xliv, 
2718.) 
15 A, Lapworth, Trans. Chem. Soc., 1904, Ixxxv, 31; H. M. Dawson 
with May S. Leslie, z7d., 1909, xcv, 1860; with R, Wheatley, 7é7¢., 1910, 
xvii, 2048 ; with F. Powis, /ézd., 1912, ci, 1593. 
16 K, H. Meyer, Ber., 1912, xlv, 2867. 
W7 OQ. Aschan, Ber., 1912, xlv, 1913 ; 1913, xlvi, 2162; K. H. Meyer, Ber., 
1912, xlv, 2868. 
8 J. Volhard, /oc. crt. 
19 ©, Schiiiz and W. Marckwald, Bev ., 1896, xxix, 58. 
