The peculiar variable V838 Mon

V838 Mon underwent, after a first nova-like outburst in January and a usual decline, a second outburst after one month, and a third weak one again a month later. Moreover, a very small increase of the temperature at the beginning of April gives us a hint about a physical process with a period of one month. We obtained a BV R IC time-sequence and modelled the photometric behaviour of the object. This leads us to the conclusion that the interstellar foreground extinction has to be 0.6 mag E(B − V ) 0.8 mag and that the quasi-photosphere had persistently unusually low temperatures for nova-like systems. The photometry was used to follow the dramatic changes of the expansion. While the appearing 10-µm excess can be well described by the heating of material ejected during this event, the IRAS emission near the location of the progenitor originates most likely from dust, which were formed during the previous evolution of the object. Assuming that the light echoes are coming from circumstellar material, the distance is 640 to 680 pc ‐ smaller than the 790 pc given in Munari et al. (2002). In our opinion V838 Mon and V4332 Sgr are manifestations of a new class of eruptive variables. We do not count M31RV to be in this class.