Until now, significant progress was achieved in understanding the mechanisms of long-term memory formation and stabilization. However, processes of the reverse amnestic direction, such as memory loss, forgetting, amnesia, despite their undoubted theoretical, social, and medical significance, are far less understood. Discovery of memory packing phenomena provided a significant contribution to elucidation about amnesia mechanisms. It was demonstrated that reactivation of long-term memory, induced in particular by retrieval, leads to destabilization of memory trace and its transition into a labile phase. Nonetheless, if in these behavioral conditions, memory trace cannot be behaviorally expressed, especially for a long time after amnesia induction, it is suggested to be the result of memory trace erasure. However, it should be noted that at present there is no persuasive experimental evidence to distinguish these two scenarios. Even in cases in which amnesia persisted for a long time after memory impairment, it can be supposed that the memory is preserved. However, it cannot be expressed at the behavioral level due to impairment of the retrieval mechanisms. In contrast, in cases in which memory impairment is transient and can be recovered spontaneously or by presentation of different reminding stimuli, it cannot be excluded that memory trace impairment could be compensated by the passage of time or is a result of recovering procedures. Another approach to studying amnesia mechanisms is retraining of amnestic animals. In some cases, this approach can be informative in determining the degree of preservation of impaired memory trace and possibility of its retrieval. It is logical to suggest that there is a direct correlation between the stability of amnesia and the number of paired conditioned and unconditioned stimuli presented during second training, which are necessary for memory formation. Neurotransmitter systems play an important role in memory and amnesia mechanisms, involving glutamate, adrenergic, cholinergic receptors, etc. In addition, in most studies, memory trace preservation was evaluated only at limited time points after its impairment. A systematic study of amnesia dynamics and its features due to different time intervals passed from memory impairment has never been conducted. In earlier experiments on snails, a behavioral model was employed of certain food rejection to study the mechanisms. Memory formation features in experiments with repeated training in the context of training and in a new context were studied as well. To study amnesia dynamics, retraining procedures were conducted at different phases following amnesia induction at 1, 3, 10, or 30 day time points.  Injections of serotonin receptor antagonists or NMDA glutamate receptor antagonists combined with a reminder 2 days after training caused memory impairment and development of amnesia, which persisted for more than 30 days Snails were injected with vehicle (n=8), methiothepin (n=8), or MK-801 (n=8) 2 days after training and a reminder was presented as described above. Two other experimental groups were injected with methiotepin (n=7) or МК-801 (n=8) without following with a reminder. A test 30 days after substance/reminder administration demonstrated that latencies of consummatory reactions to CS were lower than those among control training animals that were injected with a vehicle prior to the reminder (Mann-Whitney rank sum test, for methiothepin: -z = 3.4 p < 0.001; for MK-801: -z =3.4 p < 0.001). In addition, it did not differ from the latencies of consummatory reactions to the DS (Wilcoxon signed-rank test, for methiothepin: z = 1.2, p=0.25; for MK-801: z = 1.1, p=0.27) or from the reactions to CS before training (Wilcoxon, for methiothepin: z = 0.9, p=0.35; for MK-801: z = 0.7, p=0.48) (fig. 1, T3). The test 30 days after methiothepin or MK-801 injections without following with a reminder showed that latencies of consummatory reactions to CS did not differ from those in control training animals (Mann-Whitney, for methiothepin: z = 0.21, p = 0.8; for MK-801: z = 0.1, p = 0.9), and they were higher than the latencies of consummatory responses to DS (Wilcoxon, for both groups: z = 2.5, p < 0.05). Thus, the methiothepin or MK-801 injection paired with the reminder 2 days after training caused impairment of memory retrieval, and it caused amnesia development. These persisted for more than 30 days without spontaneous memory recovery. Antagonist injections without reminders did not impair memory.