In recent years, the increased use of proton beams for clinical purposes has enhanced the demand for accurate absolute dosimetry for protons. As calorimetry is the most direct way to establish the absorbed dose and because water has recently been accepted as standard material for this type of beam, the importance of water calorimetry is obvious. In this work we report water calorimeter operation in an 85-MeV proton beam and a comparison of the absorbed dose to water measured by ionometry with the dose resulting from water calorimetric measurements. To ensure a proper understanding of the heat defect for defined impurities in water for this type of radiation, a relative response study was first done in comparison with theoretical calculations of the heat defect. The results showed that pure hypoxic water and hydrogen-saturated water yielded the same response with practically zero heat defect, in agreement with the model calculations. The absorbed dose inferred from these measurements was then compared with the dose derived from ionometry by applying the European Charged Heavy Particle Dosimetry (ECHED) protocol. Restricting the comparison to chambers recommended in the protocol, the calorimeter dose was found to be 2.6% +/- 0.9% lower than the average ionometry dose. In order to estimate the significance of chamber-dependent effects in this deviation, measurements were performed using a set of ten ionization chambers of five different types. The maximum internal deviation in the ionometry results amounted to 1.1%. We detected no systematic chamber volume dependence, but observed a small but systematic effect of the chamber wall thickness. The observed deviation between calorimetry and ionometry can be attributed to a combination of the value of (Wair/e)p for protons, adopted in the ECHED protocol, the mass stopping power ratios of water to air for protons, and possibly small ionization chamber wall effects.