The issue of whether the mechanism of infection is independent or co-operative for low doses of transmissible spongiform encephalopathy (TSE) agent is critical for risk assessment. The susceptibility (and hence ID50) of individuals with the same prion protein (PrP) genotype may vary considerably with a small proportion being very susceptible. Assuming independent action, the incubation period (IP) would continue to increase until the dose is below the ID50 of the most susceptible individuals in the experiment, at which point it would become constant. This may explain the observed increase in IP with decreasing dose below the apparent ID50 in experiments with untreated TSE agent. In contrast, IPs for autoclaved or NaOH-treated TSE agent increase greatly at doses < ID50 suggesting strong co-operative action, or even a threshold. It is proposed here that the unit of infectivity for prion disease is a nucleation seed comprised of PrP and host phospholipid (PL). PL would play a structural role through mediating protein/lipid interactions with PrP. Heating or alkali treatment destroys the PL breaking up the nucleation seeds, which require long IPs to reform at low doses. Replenishing those inactivated PLs with host PL would explain how the phenotypic effect of long IP at low dose is lost on subpassage. It is proposed here that strain thermostability is controlled by the nature and strength of the PrP/PL interactions, which are independent of the host PrP genotype. Although repeated oral exposure to low doses of scrapie is less harmful than a single large exposure, this effect may reflect interference by competition rather than diminished risks due to a co-operative effect, and is of little importance for 'one-off' low-dose environmental exposures.