Helioseismology and asteroseismology put high demands on the accuracy and consistent numerical realization of the equation of state (for a review see Däppen, these proceedings). This is explicitly illustrated by the helioseismic determination of the helium abundance of the solar convection zone in a recent investigation by Kosovichev et al. (1992). In that work it was observed that details of the treatment of the heavy elements matter more than is intuitively expected. Naively, one would expect an uncertainty of less than 10−4 in the key thermodynamic quantity, the adiabatic gradient Γ1. This is because in material of solar composition the heavy-element abundance is less than about 1.5 × 10−3 by number, and under solar conditions the dominant nontrivial contributions to the seismically relevant thermodynamic quantities predicted by modern equations of state agree to a few per cent, even for the much more abundant hydrogen and helium. However, Kosovichev et al. (1992) found that uncertainties in the treatment of the heavy elements translate into discrepancies in Γ1 of the order of 10−3, which is enough to disturb the helioseismic helium-abundance determination significantly. We briefly present the reason below. A forthcoming paper will show more detailed results, though some further information can already be found in papers by Kosovichev et al. (1992) and Christensen–Dalsgaard & Däppen (1992).