Environmental exposures affecting human health range from complex mixtures, such as environmental tobacco smoke, ambient particulate matter air pollution and chlorination by products in drinking water, to hazardous chemicals, such as lead, and polycyclic aromatic hydrocarbons, such as benz(a)pyrene. The exposome has been proposed to complement the genome and be the totality of all environmental exposures of an individual over his or her lifetime. However, if measurements of the exposome in biological samples are the sole tool for exposure assessment there are a number of limitations. First, it has limited utility for fully capturing the impact of complex mixtures such environmental tobacco smoke or particulate matter air pollution. Second, a number of relevant environmental exposures such as noise, heat or electromagnetic fields do not have direct correlates as metabolites or protein adducts, but there is important evidence linking them with health effects. Third, functional genomic changes are likely in many instances to be both a susceptibility factor and a marker of internal doses in response to environmental exposures. Fourth, internal dose measurements of environmental exposures might have lost the distinct signature of the relevant sources. This paper emphasises the obligation of environmental epidemiology to provide robust evidence to assist timely and sufficient protection of vulnerable subgroups of populations from environmental hazards. Therefore, in applying the exposome concept to environmental health problems, a strong link with the external environment needs to be maintained.