Although PAH are expected to be significant contributors to the toxicity observed in these experiments, the actual contribution of each PAH compound in conjunction with PAH metabolites and other potential additional stressors identified as additional potential confounding factors to the different lethal and sublethal endpoints measured remains to be defined. The oiled-gravel columns produced effluents containing different concentrations and compositions of TPAH and total alkanes, proportional to the initial loading of oil on the columns. However, the initial relative concentrations of different PAH were not the same for the different treatments in the LWO and

MWO experiments and the compositions changed in different Pexidartinib in vitro ways during the two 16-day experiments because of different rates of depletion of PAH in the oil-on-gravel by dissolution, dispersion, and biodegradation. Therefore, it is not possible to determine the contribution of different PAH, alkanes, microbial degradation products MDV3100 manufacturer and microbial fouling that led to the different lethal and sublethal endpoints observed. In addition,

it is likely that the oiled gravel columns produced a mixture of dissolved and non-dissolved PAH (Page et al., 2012 and Redman et al., 2012), further complicating the definition of aqueous exposure concentrations. Neither potency nor causation were determined by Carls et al. (1999) nor can they be determined based on the available data from this study. The issues of causality and confounding factors identified here for the Carls et al. (1999) study have also been described for a similar study of pink salmon embryos and larvae (Landrum et al., 2012 and Page et al., 2012). Given the rapidly declining aqueous PAH concentrations and variable aqueous PAH compositions produced by oiled gravel columns, it is not possible to define an aqueous PAH concentration causally associated with an observed

effect (Landrum et al., 2013). This is particularly true for embryo toxicity tests, where embryos undergo rapid biochemical and morphogenic changes at the same time the exposure concentrations are declining and composition is changing most rapidly. This raises the question of whether the next use of oiled-gravel columns to generate hydrocarbon-contaminated exposure media for toxicity studies can yield reliable and reproducible results that can be extrapolated to the field. Toxicity studies need to demonstrate clear and convincing monotonic dose–response relationships between suspected toxicants and observed biological effects. The presence of two or more concentration–response relationships in multiple treatment studies is a strong indication of the presence of multiple stressors and/or mechanisms of toxicity.