Trace gas mixing ratios were measured in the upper troposphere and lower stratosphere during the second mission of the Cryogenic Infrared Spectrometers and Telescopes (CRISTA) experiment in August 1997. Water vapor data at high vertical and horizontal resolution are presented here. Respective data for CFC 11 (CFCl3) are given in a future paper by V. Kuell et al. (Tropopause region temperatures and CFC 11 mixing ratios from CRISTA 2, submitted to Journal of Geophysical Research, 2004). Some ozone data are also available. Zonal mean mixing ratios of water vapor are compared to several other measurements, and good agreement is obtained. Data variance of all three gases is high and is mostly attributable to atmospheric fluctuations rather than instrument noise. Diffusive (eddy) fluxes $overline{left(mu^{prime} cdot upsilon ^{prime}right)}$ were derived from these fluctuations for all three gases. They were converted to meridional eddy coefficients Kyy by means of respective mixing ratio gradients. Kyy values obtained from CFC 11 and H2O show a close correlation (r = 0.62). This suggests that water vapor may be a usable tracer at medium latitudes. (This does not include the tropics, where our Kyy values appear to be unreliable.) Our Kyy values are substantially higher than those frequently used in the literature. Another difference is that our summer values are larger on average than the winter ones (by a factor of 2). A similar result, however, is seen in the effective diffusivities Keff of Haynes and Shuckburgh (2000a, 2000b) (hereinafter referred to as HS a,b). There are other agreements of the two data sets Kyy and Keff: HS a,b discuss a number of transport barriers (equator and tropopause). These are found in a very similar way in our data. The latitude structures of the Kyy and Keff coefficients can be compared at three altitude levels. These patterns are also found to be quite similar.