The first report of synthetic diamond involved a high-pressure high-temperature (HPHT) process in which diamond was the thermodynamically stable phase(1). Subsequent attempts to make diamond at less extreme conditions culminated in the rapid growth of diamond films by chemical vapour deposition(2,3). But the question of whether diamond might be grown in hydrothermal conditions mimicking those under which it is formed in the Earth has been long debated(4-6). It seems reasonable to suppose that metals might play a catalytic or solubilizing role in this context, given their role in the HPHT method(1), in a recent low-pressure solid-state synthetic approach(7) and in the recrystallization of diamond in the Ni-NaOH-C system(8). Hydrothermal synthesis of diamond has been explored at some length(9-13), but with equivocal results. Here we report evidence from spectroscopic, diffraction and microscopic techniques which suggest that aggregates, tens of micrometres in size, of small diamond crystals can be grown in a hydrothermal environment from a mixture of carbon, water and metal (usually pure nickel). Crucially, we have been able to distinguish new diamond from the diamond seeds added to nucleate new growth.