An objective scheme based on empirical orthogonal function analysis to detect patterns in a single or multivariate data set is developed and applied to rawinsonde observations taken in the near-tornado enviornment. If only temperature data are considered, two distinct categories, differentiated cheifly by the tropopause height, are found. When moisture observations are included with the temperatures, the separation between categories becomes less distinct. However, it is noted that within the near-tornado environment there is an inverse relationship between the degree of observed convective and conditional instability. Analysis of only the winds shows that a low-level veering with height is the rule. However, the strength of the veering can vary considerably. When the temperature, moisture, and winds are treated in concert, two categories again appear. One group occurs with strong winds and a low tropopause, while the other group features weak winds and a high tropopause. These groups correspond to ''springtime'' and ''summertime'' synoptic situations, respectively. Comparisons of the various analyses indicate that the near-tornado environment typically features a balance between the strength of the veering of the winds and the amount of conditional instability present. Summer tornadoes feature strong conditional instability and weak winds, while springtime tornadoes occur with stronger veering and convective instability. The strongest tornadoes occur with springtime-type conditions. ¿ American Geophysical Union 1988