Uncertainty exists over the importance of Antarctic toothfish (Dissostichus mawsoni) as prey of top predators in the Ross Sea. In this paper we assess relative weight given to direct, observational evidence of prey taken, as opposed to indirect evidence from scat and biochemical analysis, and conclude that toothfish are important to Weddell seals (Leptonychotes weddellii). The seals eat only the flesh of large toothfish and therefore they are not detected in scat or stomach samples; biochemical samples have been taken from seal sub-populations where toothfish seldom occur. Using direct observations of non-breeding seals away from breeding haulouts in McMurdo Sound, 0.8–1.3 toothfish were taken per day. Based on these and other data, the non-breeding portion of the McMurdo Sound seal population, during spring and summer, consume about 52 tonnes of toothfish. Too many unknowns exist to estimate the non-trivial amount consumed by breeders. We discuss why reduced toothfish availability to Weddell seals, for energetic reasons, cannot be compensated by a switch to silverfish (Pleuragramma antarcticum) or squid. The Ross Sea toothfish fishery should be reduced including greater spatial management, with monitoring of Weddell seal populations by CCAMLR. Otherwise, probable cascades will lead to dramatic changes in the populations of charismatic megafauna.