Replicating lithospheric diamonds experimentally helps reveal their histories at depth, which may be complicated by successive growth and dissolution episodes. Here, we present diamond growth and dissolution experiments to constrain the conditions of diamond formation and residence in the Earth’s lithosphere before their transport to the surface. Experiments were performed on mixtures of carbonates, natural lherzolite or MORB, water, diamond seeds, and with or without graphite using multi-anvil presses over a few hours to more than a day at conditions relevant to the lithosphere (7 GPa, 1300–1670 °C). We observed growth within a volatile-rich melt resulting from the miscibility of silicate-carbonate melt and aqueous fluid via the reaction: Carbonate minerals + Silicate minerals/glasses + H2O = [volatile-rich melt] ± minerals + Cdiamond + O2. In the absence of graphite, we observed diamond dissolution within the same hydrous-carbonate-silicate-bearing melts and under similar pressure and temperature conditions used to grow diamond, indicating that diamond formation requires an oxygen sink (graphite, metallic and/or sulphide melts). In dissolution experiments, we also observed resorption features similar to those described in lithospheric monocrystalline diamonds, which we thus attribute to mantle fluids and not to kimberlite-induced resorption during magma ascent. We show that dissolution and growth may alternate in the mantle, and that protracted periods are not necessary to explain natural diamond histories. During experimental growth, inclusions were trapped in diamonds as they are in nature. We observed both syngenetic (formed during growth, representing the parental fluid) and protogenetic inclusions (here particles from the capsule), indicating that both kinds of inclusions can be trapped within a single growth event. Our experiments confirm that inclusions trapped in natural diamonds do not necessarily represent remnants of their parental fluids and are not necessarily contemporaneous: two inclusions near each other within a single monocrystalline diamond may have different histories, and inclusions must be shown to have achieved chemical and isotopic equilibrium before being considered synchronous.