Solid water is often the phantom material of choice for dosimetry procedures in radiotherapy high-energy X-ray and electron beam radiation calibration and quality assurance. This note investigates variation in heat conduction that can occur for a common commercially available solid water stack phantom when a temperature differential occurs between the phantom and ambient temperature. These variations in temperature can then affect radiation measurements and thus the accuracy of radiation dosimetry. In this manuscript, we aim to investigate the variations in temperature which can occur in radiation measurement incorporated (RMI) solid water phantoms, their thermal properties and the effects on radiation dosimetry which can occur because of temperature differentials. Results have shown that the rate of temperature change at a phantom center is a complex function but appears relatively proportional to the surface area of the phantom in normal clinical usage. It is also dependent on the thermal conductivity of any material in contact with the phantom; and the nature of the phantom construction, i.e., the number and thickness of slices within the phantom. A thermal time constant of approximately 20 min was measured for a 2-cm solid water phantom slice when located on a steel workbench in comparison to 60 min when located on a wooden workbench (linac couch insert). It is found that for larger solid water stack phantoms, a transient (within 1°C) thermal equilibrium exists at the center for up to 2 h, before the temperature begins to change. This is assumed to be due to the insulating properties of multiple slices within the stack, whereby very small air spaces are introduced inhibiting the heat conduction through the phantom material. It is therefore recommended that the solid water/phantom material is kept within the treatment room for closest thermal accuracy conditions or at least placed within the room approximately 10 h before dosimetry measurements. If these options are not available, a standard linear interpolation method for calculation of temperature should be used to minimize uncertainty of temperature measurements.