Despite much research on Euphausia superba, estimates of their total biomass and production are still very uncertain. Recently, circumpolar krill databases, combined with growth models and revisions in acoustics have made it possible to refine previous estimates. Net-based databases of density and length frequency (KRILLBASE) yield a summer distributional range of ~19 x106 km2 and a mean total abundance of 8x1014 postlarvae with biomass of 379 million tonnes (Mt). These values are based on a standardised net sampling methodology but they integrate over the period 1926-2004, during which krill abundance has fluctuated. To estimate krill biomass at the end of last century we combined the KRILLBASE map of relative krill density around Antarctica with the most recent, acoustics-derived, value for the CCAMLR Synoptic Survey of the Scotia Sea area (37.3 Mt). Thus the CCAMLR 2000 survey area contains 28% of the total stock, with total biomass of ~133 Mt in January-February 2000. Gross postlarval production is estimated conservatively at 342-536 Mt y-1, based on three independent methods. These are high values, within the upper range of recent estimates, but consistent with the concept of high energy throughput for a species of this size. The similarity between the three production estimates reflects a broad agreement between the three growth models used, plus the fact that, for a given population size, production is relatively insensitive to the size distribution of krill at the start of the growth season. These production values lie within the envelope of what can be supported from the Southern Ocean primary production system and what is required to support an estimated predator consumption of 128-470 Mt y-1. Given the current debate over acoustic methodology, plus the need for precautionary management of the developing krill fishery, our net-based data help to set a conservative estimate of total krill biomass.