The frequency distribution of concentrations of certain hydrocarbons in crude oils has been investigated. It was found that most hydrocarbons in the gasoline-range of petroleum have either a monomodal or a bimodal lognormal distribution. All of the very water soluble hydrocarbons (C₅ and below) however, are normally distributed. An explanation of the genesis of both the normal and lognormal distributions in terms of the hypothesis that crude oils consist of hydrocarbons that collected from waters containing natural solubilizers is presented.

The distribution of hydrocarbons in petroleum appears to reflect variations in the kind and size of the micelles in which hydrocarbons selectively dissolve. The subsequent release of the solubilized hydrocarbons from aqueous solution results in the formation of crude oil droplets. Such a geochemical process of oil formation explains not only the characteristic proportions of certain homologous and isomeric hydrocarbons present in a particular crude oil, but also accounts for the statistical nature of the variation in their absolute amounts--from one crude oil to another.

The frequency distribution analysis also permits the composition of a "typical" petroleum to be defined. The composition of API Research Project 6 crude oil is compared with that of the "median" petroleum.