Oxygen consumption of individual larvae of the Antarctic sea-star Odontaster validus was measured during the 50-day period following fertilisation. Values ranged from 0.76 pmol O-2 h(-1) for one specimen at the coeloblastula stage to 77.6 pmol O-2 h(-1) for one bipinnaria larva. At 0degreesC the mean oxygen consumption rate of an individual larva increased from 10.9 pmol O-2 h(-1) (standard error of the mean, SEM, 0.13) for a gastrula larva, 13 days post-fertilisation, to 25.4 pmol O-2 h(-1) (SEM 3.5) at the bipinnaria stage (50 days post-fertilisation). Gastrulae reared at -0.5degreesC did not have significantly different oxygen consumption rates between dabs 13 and 45 post-fertilisation (mean=11.4 pmol O-2 h(-1)). Individual metabolic rates were highly variable, covering more than a 40-fold range. At 2degreesC gastrula oxygen consumption was on average 45% higher (17.35 pmol O-2 h(-1)) giving a Q(10) temperature effect of 4.4. For bipinnaria, mean oxygen consumption in 2degreesC larvae (31.4 pmol O-2 h(-1)) was not significantly different from that in larvae at -0.5degreesC, suggesting bipinnaria metabolism may be less sensitive to temperature change than earlier stages. At 2degreesC the bipinnaria stage was,reached at 30-35 days compared with 45-50 days at 10degreesC, giving a Q(10) of 4.5 for temperature effects on development. The method here used a new, highly sensitive micro-respirometry method that is inexpensive and straightforward in design. Individual larvae of O. validus were held in 35- to 50-mul respirometers. These larvae :have very low metabolic rates, and published work on,such organisms have utilised at least 25 individuals per chamber. The oxygen content of the respirometers was measured using a 25-mul sample injected into a couloximeter. Oxygen consumption rates down to -1 pmol h(-1) can be detected. Under optimum conditions oxygen consumption of a single larva of -4 pmol O-2 h(-1) was measured with an accuracy of +/-20%. Values of similar to15 pmol h(-1) could routinely be measured with this accuracy. This method would allow oxygen consumption to be evaluated in individual field-caught larvae of most marine ectotherms.