5. After a blank distillation of the hydriodic acid (10 c.c.) yielding 0.0098 AgI and the subsequent addition of 5 c.c. of acetic anhydride:

04127 gave 0.1548 AgI below 180°. and 0.0688 AgI above 290°.

OMe 4.9.

==

NMe = 2.1.

6. After a blank distillation of the hydriodic acid (10 c.c.) yielding 00218 AgI, and the subsequent addition of 5 c.c. of acetic anhydride:

0.4130 gave 0.1401 AgI below 180°.

and 0.0764 AgI above 290°.

OMe = 4.5.

NMe = 2.3.

7. Blank distillation with 10 c.c. of hydriodic acid for one hour gave 0.0027 AgI, after addition of 5 c.c. of acetic anhydride, and a double distillation for two hours gave 0·0183 AgI, after a further double distillation for two and one-third hours gave 0.0025 AgI: 0.4966 gave 0.1501 AgI below 180°.

and 0.0713 AgI above 290°. CHON, requires OMe = 51;

Ergotoxine Phosphate.

OMe = 4.0.

NMG = 1.7.
NMe = 4.8 per cent.

Specimen No. 1 was prepared from ergotinine, the rest from ergot. Nos. 4-6 were dried at 100° and analysed with the addition of acetic anhydride. The melting point was 187-188°. 01482, dried at 100° (mixed with CuO), gave 0.3120 CO2 and 0.0900 H2O. C=57·4; H=6·8.

C35H41O6N5,H3PO4 requires C=57·9; H=6·1; OMe=4·3;

NMe 40 per cent.

On receipt of the above data we re-examined the action of alcoholic phosphoric acid on ergotinine, on a larger scale than was originally possible. We were able to transform the salt, first obtained as needles, into plates by recrystallisation from alcohol.

One gram of ergotinine was suspended in 15 c.c. of absolute alcohol, and rather more than one molecular proportion of syrupy phosphoric acid was added. On heating under reflux on the water

bath the ergotinine dissolved, but after a total heating of thirty to forty-five minutes the solution became almost solid with a mass of prismatic needles, which were collected.

The yield was 0.95 gram, or 80 per cent. of the theoretical; the melting point, 189-190°. After drying at 100°:

0.5720 gave 0.0246 AgI (two hours at 140-150°). OMe=0.56. On recrystallisation from alcohol, plates were formed, melting at 190°, which were analysed:

0.4216 gave 0.0412 AgI (two hours at 150°). OMe=1.3.

From these experiments we conclude that the phosphate was that of ergotoxine, and that no ethyl ester grouping was present, for it would have been readily removed under the analytical conditions employed.

We next examined the action of hydriodic acid on ergotinine, using in each case 10 c.c. of acid (D 1·7) and 5 c.c. of acetic anhydride, the bath being heated for two hours at the temperature intervals indicated.

A. Ergotinine, older (less pure) specimen:

0.4127 gave 0·1259 AgI at 150-1950 and 0.0389 AgI at 340-350°. Total AgI=0·1648. OMe=5.3 or NMe=4.9.

B. Ergotinine, purer specimen :

0.4129 gave 0·0105 AgI at 140-145°, 0·1333 AgI at 180—200°, and 0.0303 AgI at 300-370°. Total AgI=0·1741. OMe 5'6 or NMe=5*2.

C35H39O5N5 requires OMe=5·1; NMe=4.8 per cent.

We have calculated our results from the total silver iodide formed, for we believe that only one methyl group is present, and think that this methyl group is attached to nitrogen. Messrs. Carr and Pyman's six analyses, in which heating was continued to 290°, when calculated as NMe only, give the mean 64 per cent., which is one-third more than the theoretical. In our opinion the amount of silver iodide is, however, insufficient to account for a methoxyas well as a methylimino-group. There is abundant evidence that many methylimino-groups give off part of their methyl as methyl iodide on merely boiling with hydriodic acid, particularly if carbonyl groups are adjacent (Busch, Ber., 1902, 35, 1565; Goldschmiedt and Hönigschmid, Ber., 1903, 36, 1850; Monatsh., 1906, 27, 849; Kirpal, Ber., 1908, 41, 819). Decker (Ber., 1903, 36, 2895) has pointed out that in this way N-alkyl may be mistaken for O-alkyl, and Herzig (Monatsh., 1908, 29, 295) states that with nitrogenous substances, unless the formation of silver iodide begins soon after the hydriodic acid boils, and proceeds rapidly to an end, it is unsafe to conclude that a methoxyl group is present. In

our own experiments the evolution of methyl iodide from the boiling acid was slow and incomplete, whence our conclusion, that ergotinine and ergotoxine contain one N-methyl group, but no methoxy-group.

We desire to thank Messrs. Carr and Pyman for having directed our attention to the error in our former paper, and for having placed their analytical results at our disposal.

LISTER INSTITUTE OF PREVENTIVE MEDICINE,
CHELSEA GARDENS, LONDON, S.W.

[Received, March 16th, 1918.]

XXVIII.--Interaction of Formaldehyde and Carbamide. By AUGUSTUS EDWARD DIXON.

IN a paper by Dixon and Taylor on the interaction of aldehydes and thiocarbamides (T., 1916, 109, 1244), it is mentioned in passing that a molecular mixture of carbamide and formaldehyde, when condensed by means of hydrochloric acid, yielded a substance, apparently identical with the methylenecarbamide obtained by Hemmelmayr (Monatsh., 1891, 12, 94), from carbamide and chloromethyl alcohol. Subsequently-but not until most of the present investigation had been completed-it was learned that several chemists have already studied the interaction between carbamide and formaldehyde (see, for example, Hoelzer, Ber., 1884, 17, 659; Tollens, ibid., 1896, 29, 275; Goldschmidt, ibid., 2439; Chem. Zeit., 1897, 21, 460, 586; Einhorn and Hamburger, Ber., 1908, 41, 24). In one case only is it necessary to comment on the results.

According to Goldschmidt, carbamide yields, with "excess" of formaldehyde, a compound, CH10O3N4 (that is, 2CH ON2+ 3CH2O-2H2O), which is formed alike in the presence or in the absence of acids *; and as the yield is theoretical, he concludes that carbamide may thus be determined quantitatively. With these statements, the author's experience does not quite tally; for, in an acidified mixture, the precipitate may not be formed quantitatively; it may have a different composition, or it may not be formed at all. Moreover, in a neutral mixture of the components, even after six months' keeping, no sign of condensation was observed (Expt. 1). From a slightly acidified solution of carbamide (1 mol.), the condensate, with 0.75 mol. of formaldehyde, is methylenecarbamide;

Presumably, this means in a neutral solution

with 1 mol. of formaldehyde it is slightly contaminated with Goldschmidt's compound. With 1 mols. of formaldehyde, and thenceforward up to well beyond 2 mols., the latter compound is the sole product. Further increase of the formaldehyde ratio diminishes the yield, which, at 4 mols., is small, the product being a substance, CH12O4N4; whilst, with 11 mols., condensation ceases.

The precipitates obtained by "catalysing" solutions of carbamide containing variable proportions of formaldehyde, although differing, it may be, largely in composition, exhibit a close general resemblance. They occur in minute, white granules, apparently amorphous, each of which, in plane-polarised light, usually displays a black cross, like that of certain starches. Practically insoluble in all the common solvents, and in cold, dilute mineral acids, they are resolved by heating with the latter into formaldehyde and carbamide; infusible, they decompose at a high temperature, considerably variable with the duration of the preliminary heating (somewhere in the neighbourhood of 240-250°, uncorr.).

From these facts it is concluded that the so-called methylenecarbamide cannot properly be represented by the simple formula, CONHCH2, but is a dimeride, containing, like Goldschmidt's compound, two carbamide residues in the molecule. Its configura

NH

tion may be represented as CONHCH, NH H>CO (I), Gold

*NH•CH, NH schmidt's compound being the methylol derivative,

CO<N(CH,OH)-CHNH>00 (II),

whilst the substance, CH12O4N4, mentioned above, is probably NH——-CH, ---NH

similar, that is, CONHCH, OH HO:CH, NH CO (III). The

formula, CO(NMe-CO-NH2)2, proposed by Goldschmidt (loc. pr. cit.) for his compound does not seem to be in harmony with its decomposition into carbamide and formaldehyde by dilute acids; it is slowly attacked, too, by hot alkalis, with the evolution of ammonia, but no methylamine was detected.

Ready-formed "methylenecarbamide," when kept in contact with formalin, did not combine with it, to yield the compound II above, nor did this, under like conditions, give the possible derivative, CON(CH2OH).CH, NH>CO (IV); the last compound, in fact, *NH•CH,(CH,OH)N

2

has not been isolated, even in circumstances apparently favourable.

From a neutral solution of carbamide (1 mol.) in formalin (1 mol.), methylolcarbamide, together with a little dimethylolcarbamide, is deposited on concentration in a vacuum over sulphuric acid. The former compound, it would seem, is a near precursor of

methylenecarbamide"; for, such a solution, if kept for one-anda-half hours, so as to give the components a little time to combine before treatment with acid (they do not quickly unite in quantitative amount), began to condense nearly twice as soon as a mixture, otherwise similar, but freshly prepared. Moreover, when readyformed methylolcarbamide, in like circumstances of concentration and of temperature, was acidified to the same extent, condensation started in one-fifteenth of the time required by a freshly-made mixture of carbamide and formaldehyde (Expt. 15). Dimethylolcarbamide, on the other hand, is relatively slow to condense—a fact already noted by Einhorn and Hamburger (loc. cit.). The reason, no doubt, is that the dimethylol derivative of methylenecarbamide' (IV, above) resists formation in the presence of acids; so that, until time enough has passed for the elimination of a certain amount of formaldehyde, condensation is barred. No doubt, too, the complete stoppage of the carbamide-formaldehyde condensation (see above) by a very large excess of formaldehyde is due to the maintenance of the carbamide in the state of its dimethylol derivative; if the uncondensable mixture is added to a concentrated aqueous solution of carbamide, a precipitate separates forthwith (Expt. 9).

Having regard to the above facts, one may reasonably judge that "methylenecarbamide" results from the decomposition of some compound, generated by the action of, say, hydrochloric acid on methylolcarbamide. Of the stages that must occur, the first is probably the change of NH, CO NH CH2 OH into

NH,CO NH•CH,C1,

a compound, likely to be unstable, owing to the loss (at least, if water is present) of hydrogen chloride. Through the elimination of hydrogen chloride there originates the residue, NH.CO.NH·CH2',

whence, by polymerisation, the compound CO< NH•CH,NH >co NH•CH,NH 2

(" methylenecarbamide ") could occur. An essentially similar change, in fact, has been observed by Chattaway (T., 1909, 95, 236), chlorocarbamide, by the loss of hydrogen chloride, polymerising thus:

NHCI
CONH,

Really, however, the carbamide-formaldehyde process is somewhat more complex than is here indicated, for, as previously mentioned, molecular proportions of carbamide and formaldehyde do not give pure "methylenecarbamide"; moreover, the condensate of a tolerably pure specimen of methylolcarbamide was markedly contaminated with a substance poorer in nitrogen. Further inquiry