Encoding of memories relies on modifications in the strength of the synapses connecting the different cells within a neuronal network. The selective increases in synaptic weight are thought to be biologically implemented by long-term potentiation (LTP). Classically, in area CA1 of hippocampal slices a single train at 100-Hz (1s) triggers a short-lasting LTP, which lasts 1–2 h and is independent of new protein synthesis, whereas multiple trains induce a long-lasting LTP, which lasts several hours and can be blocked by protein synthesis inhibitors. However, it has been repeatedly shown that the threshold and the features of these LTP depend on the history of the neurons, a phenomenon known as metaplasticity. We already demonstrated that slices recovery conditions modified the characteristics of LTP (Capron et al. 2006). By maintaining the slices in submersion during recovery, it was possible to induce a long-lasting LTP with a single train. Further investigations demonstrated that long-lasting LTP (more than 10 hours) can be induced by a very short stimulation (15 pulses) and that increasing the number of pulses increased just the level of potentiation. Moreover, this kind of LTP became nearly independent of new protein synthesis. In this case, synaptic tagging could not be observed anymore. The only way to impair this long-lasting potentiation was to increase the frequency of basal stimulation: stimulating the slices every 10 sec prevent the stabilization of LTP induced by one or 4 trains. But in this case again, synaptic tagging could not be observed. So, in vitro, de novo protein-synthesis dependence of LTP and synaptic tagging can only be observed in very specific conditions depending on the set-up used, the experimenter, the conditions of slices recovery, the composition of ACSF....This dramatically increases the difficulty for comparing the different results obtained in different labs and prevents rapid progress in the understanding of this phenomena.