The anaerobic reactor was initially charged with 300 mL of distilled water, 500 mL of the inoculum and 200 mL of a nutrient-trace element solution. The composition of this nutrient — trace element solution is given in detail elsewhere (Borja et al., 2001).

The start-up of the reactor involved stepped increases in COD loading using an influent substrate concentration of 17.2 g COD/L. During this period the organic loading rate (OLR) was gradually increased from 0.25 to 0.50 g COD/ (L d) between 1 and 15 d, 0.75 g COD/ (L d) between 16 and 30 d, 1.00 g COD/(L d) between 31 and 45 d and finally 1.25 g COD/(L d) between 46 and 60 d.

After the preliminary step, the reactor was fed in series of semicontinuous experiments using OLRs of 0.9, 1.2, 1.4, 1.7, 2.1, 2.8, 3.5, 4.1 L COD/(L d) for the OWSW1, which correspond to hydraulic retention times (HRTs) of 40.0, 28.6, 25.0, 20.0, 16.6, 12.5, 10.0 and 8.3 d, respectively. After these experiments with OMSW 1 five different OLRs were assessed for the OMSW 2, 3.0, 6.0, 9.05, 12.0 and 15.0 g COD/ (L d), these OLRs corresponded to HRTs of 50.0, 25.0, 16.6, 12.5 and 10.0 d, respectively.

Once steady-state conditions were achieved at each feed flow-rate, the daily volume of methane produced, and total and soluble COD, pH, total volatile fatty acids (TVFA) and volatile solids (VS) of the different effluents were determined. The samples were collected and analysed for at least 5 consecutive days. The steady-state value of a given parameter was taken as the average of these consecutive measurements for that parameter when the deviations between the observed values were less than 3% in all cases. Each experiment had a duration of 2-3 times the corresponding HRT.

The organic loadings applied in this work were increased in a stepwise fashion in order to minimise the transient impact on the reactor that might be induced by a sudden increase in loadings.