The scope, mechanism and synthetic value of the acid-catalysed thermal decomposition of N,N'-diphenyl-1,3-propanadiamine has been investigated. This diamine has been decomposed under a variety of acid conditions and the products have been analysed by gas chromatography. Using the optimum conditions found in these experiments, a number of other N;N'-diaryl-1,3-propanediaminas have been decomposed with hydrogen bromide at elevated temperatures. The volatile products have been analysed by gas chromatography and have bean separated chemically to give the expected anilines and 1,2,3,4-tetrahydroquinolines and also significant quantities of julolidines in many cases where the formation of such tricyclic products is allowed. As well as these major components, a number of N-alkylanilines have also been isolated and some N-alkyltetrahydroquinolines have been provisionally identified. N,N’-Diphenyl-cis-1,3-cyclohexanediamine and sym-dianilino-dimethylether have been prepared and decomposed; examination of the reaction products has shown that 1,3-endoimino cyclohexane is probably produced in the first case while aniline and an intractable tar are obtained in the second case. A number of N-(3-bromopropyl)-arylarmines have been prepared and decomposed by heat. A comparison of the product ratios with those of the corresponding diamine decompositions has been made and the possibility that these compounds are intermediates in the diamine decomposition has been discussed. Comments have also been made of other mechanistic aspects of the decomposition, in the light of the experimental data. During the course of this work, it was noted that no reports could be found in the literature, of the application of the Hofmann-Martius reaction to N-alkylnaphthylamines. Consequently, N-methylnaphthylamine hydrohalides have been prepared and decomposed. With the hydrochlorides only dibenzacridines were isolated, but with the hydrobromides, normal Hofmann-Martius products were obtained as well. The influence of the halogen on the course of such break-down has been discussed.