I spent quite a bit of time doing analytical chemistry. When you...

I spent quite a bit of time doing analytical chemistry. When you do trace analysis the standard deviation is often multiples of the mean due to so many variables. So you often get 'crazy' results like 5 +/-15 mg/Kg.

I suspect if the Edencrete dosage was adjusted in 0.1 gallon/yd3 increments from 0 to 4 gallons/yd3 and sampled 3x for each level (eg a proper experiment) the correlation would be modest above 2 gallons/yd3 and probably non-existent below 1 gallon/yd3.