Tokamak Fusion Test Reactor (TFTR) profiles from a supershot density-modification experiment [Zarnstorff et al., Plasma Physics and Controlled Nuclear Fusion Research, 1990, Proceedings of the 12th International Conference, Washington (IAEA, Vienna, 1991), Vol. I, p. 109] are analyzed for their local and ballooning stability to toroidal eta(i) modes in order to understand the initially puzzling results showing no increase in chi(i) when a pellet is used to produce an abrupt and large increase in the eta(i) parameter. The local stability analysis assumes that k parallel-to = 1/qR and ignores the effects of shear, but makes no assumption on the magnitude of k parallel-to upsilon(ti)/omega. The ballooning stability analysis determines a self-consistent linear spectrum of k parallel-to 's including the effect of shear and toroidicity, but it expands in k parallel-to upsilon(ti)/omega less-than-or-equal-to 1, which is a marginal assumption for this experiment. Nevertheless, the two approaches agree well and show that the mixing length estimate of the transport rate does not change appreciably during the density modification and has a value close to or less than the observed chi(i), in contrast to most previous theories, which predicted chi(i)'s that were over an order-of-magnitude too large. However, still to be explained is the observed increase of chi(i) (r) with minor radius by adding the effects of (i) the finite-beta drift wave-magnetohydrodynamic (MHD) mode coupling, (ii) the slablike mode, or (iii) the trapped-electron response. The experimental tracking 0.2 < chi(e)/chi(i) < 0.7 suggests that both grad T(i) and trapped-electron driving mechanisms are operating.
