The data reported on the radiation yields of the saturated Ci compounds produced by radiolysis of Cn n-alkanes are considered ( n < i < 2 n, where n is the number of C atoms in an alkane molecule). Attention has been paid primarily to the ratios of radiation yields of the "abnormal" Cn+1 group of products, to the yields observed for Cn+2 , Cn+3 , etc. groups. The "abnormal" ratio of radiation yields in the Cn+1 group of products has been accounted for by WIDMER and GäUMANN ( Helv. chim. Acta 1964, 46, 944, 2766 and 2780), who suggested that the formation in irradiated n-hexane of the C 7 linear product is due not only to the combination of CH 3 and C 6H 13 radicals, but, to a considerable extent, to the combination of other linear radicals produced from n-hexane. The quantitative kinetic treatment of the data on radiation yields of Cn+1 products measured by PANCHINIet al. ( Int. J. Radiat. Phys. Chem. 1970, 2, 147) for n-pentane and n-hexane shows that this explanation is not satisfactory. It has been noted that the experimentally found ratio of Cn+1 product yields is very near to the relative abundance of CH bonds of different types in the irradiated alkane molecule. In the present communication a new hypothesis is suggested accounting for the observed ratio of yields of all Cn+1 radiolysis products, linear as well as branched. It has been supposed that in the radiolysis of n-alkans Cn+1 isomers are produced rather by the indiscriminate insertion of methylene, CH 2 (in a singlet state), into CH bonds of alkane, rather than by the combination of CH 3 and parent radicals. Some considerations are given supporting this hypothesis.