Using systems for long-term maintenance of Antarctic krill Euphausia superba in aquaria, positive growth was measured in a large number of specimens under controlled conditions. The first moult always showed the greatest growth increment. The high laboratory growth rates agreed with previous field investigations. Often-reported 'abnormal' slow and predominantly negative growth was probably due to inadequate maintenance procedures. Moulting was partly synchronous, and moult frequency was temperature dependent. Growth of Antarctic krill did not match the slow rates expected of a polar species. Krill was very sensitive to experimental variation in feeding regime: (1) moult frequency (14 to 29 d) and (2) growth increment at moult (-15 to 21 %) varied together. Ecdysis and the staging and timing of the moult cycle were studied. The stage definitions of Buchholz (1982; Mar. Biol. 66: 301-305) were expanded and are now applicable to frozen krill and for field investigations, and include a simple classification key. Regular histological changes in the integument and consistent changes in the titre of the moult hormone could be attributed to distinct stages of the moult cycle. A clear temporal correlation confirmed the physiological relevance of the staging system. A possible 'resting' stage between post- and pre-ecdysis was not found. The apparent plasticity of growth parameters indicates that, in the wild, krill is able to adapt to strong regional and seasonal changes in feeding conditions, typical for Antarctic waters. Krill grows fast, and in spite of being predominantly planktonic, attains a large size. This enables fast swimming which facilitates location of rich food sources. Favourable feeding conditions immediately initiate growth. Essentially unhindered motility during actual moult and extremely fast ecdysis were interpreted as an adaptation to a pelagic way of life.