A majority of male rodents will vigorously attack a male conspecific introduced in its home cage, as observed in the resident-intruder (R/I) paradigm. This behaviour, once manifest, remains stable over time. Intriguingly, however, aggression phenotype is not distinguishable upon the first intruder encounter, and requires multiple (typically three) exposures to emerge. unfamiliar mice to settle at a level that is then maintained through adult life . This “aggression career” suggests that learning and plasticity may be key features of innate social behaviour, yet the neural mechanisms of this phenomenon remain obscure. Murine intermale aggression has been shown to depend critically on a population of dopamine transporter -expressing neurons of the ventral premammillary nucleus of the hypothalamus (PMvDAT cells; Stagkourakis et al., 2018). Notably, the excitability of these cells differs significantly between aggressive and non-aggressive mice. Here, we asked if PMvDAT activity is necessary and/or sufficient as the animal assumes a stable aggression phenotype over repeated intruder encounters. Repeated R/I tests were performed in initially naïve male C57BL/6J mice while PMvDAT .Using a combination of opto and chemo -genetic approaches we manipulated PMvDAT neuronal activity prior to or during the first experiences of a resident mouse with an intruder. We first inhibited PMvDAT cells during three consecutive RI trials followed by three RI trials with no neuronal modulation. In another group we stimulated PMvDAT cells while the animal was temporarily alone in its home cage across repeated stimulation sessions, followed by three days of RI testing with no neuronal modulation. To access the contribution of PMvDAT neurons during the establishment of an aggressor phenotype, we bilaterally expressed cre-dependent viruses coding for either inhibitory halorhodopsin (eNpHR 3.0) or DREADDs (hM4Di). Silencing of PMvDAT neurons during the first three R/I sessions prevented the establishment of an aggressor phenotype. In subsequent sessions, where no cell manipulation was performed, initially absent aggression gradually intensified finally reaching a stable plateau state . We next examined the behavioral consequence of bilateral activation of PMvDAT neurons in adult male mice which had not yet experienced an intruder. For this purpose, excitatory channelrhodopsin-2 (ChR2) or DREADDs (hM3Dq) were used. Unexpectedly, rather than simulating RI experience, repeated stimulation of PMvDAT cells in an identical environmental context but in the absence of an intruder, prevented the later development of a typical R/I aggressor phenotype. Together, these findings suggest that PMvDAT activity during the peer encounter is required for development of an aggressor phenotype , and that the activation of this circuit must be coupled with a specific social context or associated other neuronal or physiological cues to enable the development of the relevant behavioural phenotype.