Many forms of neurodegeneration are ascribed to excessive cellular Ca2+ loading (Ca2+ hypothesis). We examined quantitatively whether factors other than Ca2+ loading were determinants of excitotoxic neurodegeneration. Cell survival, morphology, free intracellular Ca2+ concentration ([Ca2+]i), and 45Ca2+ accumulation were measured in cultured cortical neurons loaded with known quantities of Ca2+ through distinct transmembrane pathways triggered by excitatory amino acids, cell membrane depolarization, or Ca2+ ionophores. Contrary to the Ca2+ hypothesis, the relationships between Ca2+ load and cell survival, free [Ca2+]i, and Ca2+-induced morphological alterations depended primarily on the route of Ca2+ influx, not the Ca2+ load. Notably, Ca2+ loading via NMDA receptor channels was toxic, whereas identical Ca2+ loads incurred through voltage-sensitive Ca2+ channels were completely innocuous. Furthermore, accounting quantitatively for Ca2+ loading via NMDA receptors uncovered a previously unreported component of L-glutamate neurotoxicity apparently not mediated by ionotropic or metabotropic glutamate receptors. It was synergistic with toxicity attributable to glutamate-evoked Ca2+ loading, and correlated with enhanced cellular ATP depletion. This previously unrecognized toxic action of glutamate constituted a chief excitotoxic mechanism under conditions producing submaximal Ca2+ loading. We conclude that (a) Ca2+ neurotoxicity is a function of the Ca2+ influx pathway, not Ca2+ load, and (b) glutamate toxicity may not be restricted to its actions on glutamate receptors.