Low-resolution spectrophotometry of Jupiter by Karkoschka (1994, Icarus 111, 174-192) revealed a weak apparent absorption feature in the spectrum of Jupiter at 920-940 nm. This is the wavelength of the strong 2[v.sub.1] + [v.sub.3] band of [H.sub.2]O, and Karkoschka tentatively identified the feature as jovian [H.sub.2]O absorption. If this spectral feature is indeed due to [H.sub.2]O absorption, it would offer the exciting possibility of monitoring the jovian [H.sub.2]O with visible band spectroscopy. This feature has been seen in previous low-resolution spectroscopy of Jupiter such as that of Woodman et al. (1979, Icarus 37, 73-83), but was not identified as jovian in origin because of uncertainties in extinction corrections for the very strong telluric [H.sub.2]O absorption. Cochran et al. (1995, Bull. Am. Astron. Soc. 27, 66) tentatively identified the feature as due to [NH.sub.3], an identification that was supported by further low-resolution spectrophotometry by Karkoschka (1998, Icarus 133, 134-146). Here we present high-resolution spectra (R = 155,000) of Jupiter using the 2dcoude spectrometer on the 2.7-m Harlan Smith Telescope at McDonald Observatory, in order to make a definitive identification of this absorption feature. Observations were obtained when Jupiter was near quadrature in order to Doppler shift potential jovian [H.sub.2]O features away from the corresponding telluric lines. We demonstrate conclusively that the absorption band identified by Karkoschka is not, in fact, due to [H.sub.2]O. We present further evidence that it is most likely to be an absorption band of [NH.sub.3]. We searched our spectra carefully to attempt to identify absorption lines which are actually due to [H.sub.2]O, without success. Key Words: Jupiter, atmosphere; spectroscopy; atmospheres, composition.