We used the Very Large Array in 1997 January, February, March, and April to observe comet Hale-Bopp in the 1.35 cm H2O line. No line was detected in any of the four observations, and upper limits to the H2O line emission are presented. These upper limits are typically around 60 mJy, which is much lower than the ~2.5 Jy flux density reported for this line in comet Bradfield (1974b) and jansky-level flux densities sporadically reported for a few other comets. This discrepancy is most likely due to a combination of questionable reliability of past reported detections and the possibility that masing in particular observational geometries is required for detection of the line. We concur with previous studies that indicated that this line could be detected only if it were masing. In the only masing scenario likely to yield a positive result, a line may be seen only if the masing region is observed along a jet against the radio continuum background of the nucleus. We estimate the probabilities of a detection in a random search for two typical positions of the origin of a jet on a rotating nucleus: (1) close to the pole and (2) close to the equator. Because the probabilities are very low (0.01) in either case, our negative result for comet Hale-Bopp (as well as previous negative results for other comets) cannot exclude the possibility of such masing jets. Carefully scheduled observations of future comets whose rotation and jet locations have been determined must be used to evaluate this possibility.