the benzene ring of the halogen atom to be replaced and the complexity of the carbon chain of the alkoxide or aryloxide determined the procedure used in each synthesis. Variations in the general procedure involved reaction temperature, exposure time and amount of water present.

No abstract

March, 1951 POLYAMIDES FROM FORMALDEHYDE AND DINITRILES 1031 cake was slurried successively with 600.ml. each of water and ethyl alcohol. The combined filtrate was evaporated a t 15 mm. Presswe by warming on a water-bath (not above 65'). A white solid remained weighing 90.5 g. (86%) after drying in a vacuum desiccator. The crude methylene-bisacrylamide was recrystallized from a mixture of 300 cc. of acetone and enough water (about 75 cc.) to dissolve the product in the hot solution. The white solid that separated from the cooled solution weighed 71 g. (67.6% yield of purified material) , m.p. 185' with polymerization.

Olefins can be converted by amylsodium to alkenylsodium reagents. This reaction, just observed, is important in a preparative way because the products, when carbonated, yield j3,y-unsaturated acids which, in most cases, have not been prepared hitherto. The work furnishes also a basis for a comparison of the relative hydrocarbo acidities' of the olefins and of the relative acidities a t various positions in an olefin, and shows that an alkyl group reduces this acidity even as it does in aromatic compounds. Finally, this acidity appears to be closely related to the general reactivity of the olefins in reactions where a labile hydrogen atom is involved, because the compounds primarily produced by polymerization of olefins in acid solution can be predicted from the data of relative acidities, as determined by the alkaline reagents.Metalation of Olefins.-The formation of alkeriylsodium compounds takes place so easily that its non-observance in our previous studies now appears surprising. Any interchange that did occur in the past has, however, been extremely slight, since the quantity of olefin present has been very low, the effectiveness of stirring has been much less than a t present and the reaction time has been short. Attention was first directed to this new reaction by a preparation of amylpotassium, analogous to that of amylsodium. The product, after carbonation, contained a trace of unsaturated acids which boiled within the same range as caproic acid. The new products were derived from the reaction shown below between CH?=CHCH (IC) ClHo and KCH=CHCjH7 amylpotassium and 1-pentene, an olefin which happened to be present in the reaction mixture because of a disproportionation in the reaction of amyl chloride with amylpotassium which occurred during the preparation of the latter by addition of aniyl chloride to metallic potassium. Amylsodium was then found to react with the olefin as did amylpotassium provided a sufficient quantity of the hydrocarbon was present and the reaction was given time for completion.Isobutene is readily attacked by amylsodium. The product, after carbonation, is the hitherto unreported P-methylvinylacetic acid, as shown by analysis of the p-bromophenacyl ester derivative, and by isomerization of the acid under the influence of alkali to the known @,@-dimethylacrylic acid. Evidence that a dicarboxylic acid, @-CjE€iiI( + GHie + (1) Morton, Chem. Rm., 35, 1 (1944) H2C=C(-CHoCOzNamcthyleneglutaric acid, H2C=C (CK2COOH)a, the result of dimetalation, is present in a higher boiling fraction, is given by ozone cleavage to formaldehyde stid acetonedicarboxylic acid. By a similar series of reactions propene gives vinylacetic acid, though in smaller yield than the niethylvinylacetic acid. This metalation occurs readily also but a cornparison of. the relative activities of propene and isobutene is difficult to make because propene is so volatile. Propene shows a smaller amount of interchange with benzylsodium than with amylsodium, and this fact accords with the view that the latter is a salt of a weak...

aspect of-the problem. Considerable disubstitution also occurs in some instances.The work on the olefins suggests that rubber also should be a hydrocarbo acid and susceptible to metalation by amylsodium or other organoalkali metal reagent. This expectation is realized but the rubber undergoes also considerable crosslinking. The details of this work have an important bearing on the problem of gel formation of rubber during sodium polymerization and will be discussed at length in a paper on that subject.The activity relationships discussed in this paper are, fundamentally, those in acidity, not merely in reactivity to a specific reagent. Therefore, a similar interchange could be a factor in the case of any reagent that is at the tip of the growing system. In this particular instance the