Various theoretical studies have shown that highly diluted plutonium solutions could have a positive temperature effect, as a result of the expansion effect in the solution, the Doppler effect on the cross sections and the temperature effect on the neutron thermalisation in the plutonium solution. This paper shows that the application of an S(α,β) table at the appropriate temperature is required to correctly characterize the temperature effect in the studied diluted plutonium solutions. Using a slightly different temperature for the S(α,β) table can even result in a fundamentally different behaviour. A negative effect can for instance be obtained instead of the observed positive effect if the temperature of the S(α,β) table is off by only a few degrees. From the point of view of criticality safety and the phenomenology of a potential criticality accident, this makes a big difference. In order to solve this issue, it would be necessary to produce S(α,β) tables at any given temperature. A number of interpolation techniques on the basic S(α,β) tables and the processed data has been investigated, unfortunately without any success. Other techniques to remedy the problem are under further investigation.