The French Polynesia Economic Exclusive Zone is located in an important longline fishing ground for albacore (Thunnus alalunga), yellowfin (T. albacares) and bigeye tuna (T. obesus). Longline tuna abundance estimates using commercial catches are particularly biased when hook depth does not coincide with the depths at which tuna prefer to swim. To avoid catchability problems, a direct acoustic estimate of tuna abundance was made in the French Polynesia EEZ using a 38 kHz echo-sounder with a depth range of 500 m. Several biases can influence individual tuna target selection, such as the threshold effect, the risk of multiple target acceptance, the beam width effect and the reduction in target detection at depth. However, they all appeared to be limited in effect. Comparison with experimental longline catches shows that the acoustically selected targets appear to be representative of longline tuna distribution. A density of 1.33 fish per km2, i.e. about 33.8 kg of tuna per km2, was measured. Such a density is slightly greater than the estimate based on tuna catches, as the whole tuna habitat range is not sampled by most professional longlining. (C) 2000 International Council for the Exploration of the Sea.