No, I did not try to include temperature or composition in this model because my system is similar to a bacterial growth curves -- in these situations you usually fit a logistic/sigmoidal or hill-type model to describe how a particular species grows under a set of different conditions (i.e. temperature, nutrient level, etc.), but then you can compare the non-linear model parameters to determine the fastest growing species and/or worst/best conditions for growth. My thought was to see if there is anyway I could conclude that sample or temperature significantly effect A/B/C/D.