The persistence of protein synthesis-dependent long-term potentiation (late-LTP) is thought to require heterosynaptic activation of both glutamate and neuromodulatory receptors in the hippocampus. The present series of experiments contrasts two alternative accounts of heterosynaptic activation. The original version of the synaptic-tag hypothesis of the variable persistence of LTP implied that neuromodulatory and glutamatergic activation could occur independently, albeit within a critical time-window; an alternative view is that there needs to be simultaneous co-activation of both receptors to trigger the up-regulation of relevant protein synthesis (Neuron 34 (2002) 235). Our findings include a replication, over 6 h post-LTP-induction, of earlier findings showing heterosynaptic influences on LTP persistence. Specifically, 'strong' tetanisation with multiple trains of stimulation of one input pathway in a conventional hippocampal slice preparation induces a D1/D5 receptor-dependent form of late-LTP that enables 'weak' tetanic stimulation to induce late-LTP on an independent pathway. However, we also observed that when the first pathway was tetanised in the presence of AP5, not only was no LTP observed on that pathway, but there was also no rescue of late-LTP on the second pathway. Thus, it appears that DA receptors must be co-activated with NMDA receptors in a common pool of neurons to enable LTP persistence, although late-LTP can still be induced by selective activation of glutamatergic synapses if this occurs at time periods shortly before or shortly after this essential coactivation.