http://www.idealist.ws/EPA pu238 data in excel format included on link
The EPA's data includes negative values - which have no physical meaning. In order to trust the positive values put forward by the EPA, we must determine if there is a negative bias, which would drag down positive values. This is an important point because the EPA has a history of maintaining a negative bias in their analysis of samples. 'Oddly enough, in 1982 there were some months in which many cities were reported to have large negative concentrations of radionuclides in milk, which can have no physical meaning.11 It is possible to have slight negative readings because of statistical uncertainty in measuring milk with zero levels of radioactivity, but it is not possible for such uncertainty to reach levels of greater than -5 picocuries per liter.'('The Energy Within,' by Jay Gould, Four Walls Eight Windows, 1996, p. 57, more on pp. 58-60).
To be able to make conclusions on this data, we need to have the EPA's full data that was printed from their instruments - this includes blanks and QC samples. Only using such information - all obtainable from EPA - as the detection limit(s), raw instrument data and 'blank biases' (a blank should always be zero and any positive or negative value will bias the analysis) will we be able to trust the data. Because a very, very small amount of plutonium-238 - if ingested - can exceed an individual's expected lifetime radiation dose, any bias can have a huge impact on calculated dose. MORE AT LINK
![](http://www.idealist.ws/EPAPU238.PNG)