Potential applications of air pocket type tensiometers in measuring hydraulic head profiles in deep vadose zones are discussed. Advantages of this method include (i) the ability to obtain tensiometer measurements far beyond the approximately 9 mdepth often associated with the limit of conventional tensiometry, (ii) ease of regular gauge calibration, and (iii) low cost. Advantages relative to buried, dedicated pressure transducer tensiometers are gained at the expense of substantial losses in gauge sensitivity, S*. In view of this compromise, an analysis was performed to determine the optimal fractional waterfilled length, F, for air pocket tensiometers. It is shown that the critical ratio governing the nature of S*‐optimization is approximated by (II*— IIo)z*, where II*represents the absolute matric head, IIois the vapor pressure of water expressed in head units, and z*is the depth of the tensiometer tip. When (II*— IIo)/z*> 1, S*is optimized when F→1. However, when (II*— IIo)/z*< 1, S*is optimized as F→0. The central role of (II*— IIo)/z*arises from the fact that S*= Pa/Vg, where Parefers to the absolute pressure of all tensiometer headspace gasesexcludingwater vapor, and Vgrefers to the volume of the gas phase within the tensiometer headspace. When (II*— IIo) is less than z*, S*goes to zero because the absolute pressure in the tensiometer headspace approaches the vapor pressure of the tensiometer water (Po) when attempts are made to fill the tensiometer column with liquid water. In the more familiar case of II*— IIobeing larger than z*, the dominance of Paover Poassures that S*increases as the instrument is filled. To test the predicted nature of S*, laboratory experiments were performed on 1.11‐6.36‐, and 11.91‐m long tensiometers over a range of values of (II*— IIo) and F sufficient to provide three orders of magnitude variation in S*. Measured S*agreed well with predicted values, and supports the conclusion that response times are minimized with F→0, in situations where (II*— IIo)/z*< 1.