This is a common problem in experiments that require you to weight crucible before and after roasting. After roasting crucible is usually dry, but before it may be wet - and if it is, its mass will be not that of dry crucible, but of crucible plus some amount of water. If you try to find mass of the substance left after roasting subtracting initial mass form the final mass, your result wil be off by the mass of water that was not removed before the roasting.