After achieving the required age, samples for the nonevaporable water content (WN ) determination were ground to a powder using a mortar and pestle and washed with methanol, using a porous ceramic filter and a vacuum, to stop the hydration. The resultant powder was divided approximately in half and placed in two crucibles of known mass and left overnight (on the order of 20 h) in an oven at 105ºC. When they were removed from the oven, the masses of the crucibles and samples were redetermined before placing them in a furnace at 950ºC for a minimum of 4 h, after which their masses once again were determined. The nonevaporable water content was calculated as the difference between the 950º and 105 ºC mass measurements divided by the mass of cement remaining at 950ºC, corrected for the loss on ignition of the cement powder itself, which was assessed in a separate crucible experiment. Based on the expected uncertainty of the mass measurements (0.001 g), a typical expanded uncertainty 27 in the calculated WN, is estimated to be 0.001 g of H2O/(g of cement), assuming a coverage factor 27 of 2. The nonevaporable water contents were converted to degrees of hydration by normalizing by the values measured for fully hydrated samples of the two cements, 14 0.226 and 0.235 g of H2O/(g of cement) for Cements 115 and 116, respectively.