In microdialysis experiments, 'recovery' estimations are required to calculate extracellular concentrations of the compounds determined. Generally, relative recovery (RR) is determined in vitro as: RR = cd/cs, with (cd) being the concentration of a compound in a dialysate fraction and (cs) its known concentration within a sample solution. To determine recoveryin vivo, relative loss (RL) was defined RL = (cp-cd)/cp with (cp-cd) being the loss of a compound from the perfusate and (cp) its perfusate concentration. RL was determined in vitro and in vivo by adding an 'internal reference compound' to the perfusate. Here, 14C-labelled lactate was used as the compound of interest. Comparing RL and RR in vitro, we found both to be similar. In vivo, however, RL was 34% of RL(in) vitro (CSF) and 46% of RL(in) vitro in agar-containing CSF. During ischaemia, RL of lactate even decreased to only 35% of the pre-ischaemic control level. We conclude that RL and RR represent inverse measurements of 'recovery.' Whereas RR can only be determined in vitro, RL can be determined in vivo. We found recoveryin vivo to be different from recoveryin vitro. Moreover, recoveryin vivo decreased during ischaemia. By means of the measured recoveryin vivo extracellular lactate concentrations prior and during ischaemia were calculated. The results, therefore, validate the 'internal reference technique' as a practical method for estimating recoveryin vivo and for controlling dialysis efficacy in vivo even continuously.