This paper reports the results of a study evaluating the ion temperature partition coefficient &bgr;∥ over a range of F region altitudes. The data have been selected from three EISCAT CP-0 experiments, each of which displayed clear evidence of ion frictional heating. The data were averaged at a time resolution of 30 s. These observations, made in an observing direction parallel to the Earth's magnetic field, have the advantage that the line-of-sight ion thermal velocity distribution can be closely approximated to a Maxwellian, while still giving a line-of-sight ion temperature which can be interpreted on the basis of a well-known energy balance equation. The technique for determining &bgr;∥ depends upon fitting the variation of the field-parallel ion temperature to simplified forms of this balance equation. The method is an extension of the curve-fitting approach previously used by Glatthor and Hern¿ndez (1990) to deduce both ion temperature partition coefficients at a single F region altitude. The results of this procedure are compared to the theoretical predictions for &bgr;∥ obtained from an equation originally due to St.-Maurice and Hanson (1982).

By introducing measured and modeled parameters, it is found that values of &bgr;∥ close to those expected for resonant charge exchange collisions are predicted around the F region peak. At greater heights, however, the increased influence of Coulomb collisions is predicted to give rise to an increase in &bgr;∥ with altitude, corresponding to ion thermal velocity distributions which tend toward isotropy. This height dependence, which has been theoretically predicted in some other recent studies, would be an important factor in calculations of the energy balance in the upper F region. While the experimentally derived values of &bgr;∥ are close to the RCE predictions near the F region peak, the expected increase in &bgr;∥ with altitude is seen in only one of the three selected events. In the remaining cases, the increases are notably smaller. Possible reasons for this discrepancy are discussed. Changes in ion and neutral composition, together with effects such as heat conduction and thermal diffusion can act to bias the curve-fitting technique. The degree of coupling between the ionized and neutral atmospheres is also an important factor, and there is considerable uncertainty in some of the terms used in the theoretical part of the study. Nonetheless, the results suggest that Coulomb collisions play a potentially important role in determining the energy balance of the upper F region. A clear need exists for further studies in this area to establish more fully the contribution of Coulomb processes. ¿ American Geophysical Union 1993