High variability among individuals is often encountered when hemocyte characteristics are measured in bivalves. Such variability is suspected to result partly from genetic factors. In this study, hemocyte characteristics of six families of Crassostrea gigas were compared by flow cytometry at one sampling date in October 2001. These families were obtained from a nested, half-sibling cross design, and reared from July to October 2001 at three sites distributed along the French Atlantic coast from north to south: Baie des Veys (Normandy), Riviere d'Auray (Brittany) and Ronce (Marennes-oleron Basin, Poitou Charentes). Among the 15 measured hemocyte characteristics, production of reactive oxygen species (ROS) of untreated hemocytes (maintained in filtered sterile seawater) and treated hemocytes (zymosan at 20 particles per hemocyte, and with vibrio sp. S322 at 50 bacteria per hemocyte) was the most notable differences between families. This supports the existence of a genetic basis, at least partly, for the hemocyte characteristics of oysters, and especially for ROS production. Among the six families analyzed, three have shown high survival during summer (named as "resistant", mean mortality 5.2%) and three experienced high mortality during summer (named as "susceptible", 30.6% mean mortality). Families showing high or low survival to summer mortality had similar hemocyte characteristics, regardless of the environmental conditions or reproductive state. Resistant families were observed to have higher total hemocyte counts and lower production of ROS than susceptible families. Moreover, ROS production of hemocytes from susceptible families was diminished significantly more by pathogenic vibrio than that of resistant families. However, this study demonstrates also that rearing site strongly affected the hemocyte characteristics of all families of oysters, most notably hemocyte concentration and morphology (size and granularity), production of reactive oxygen species (ROS), and susceptibility to the cytotoxic activity of the pathogenic vibrio sp. S322 (50 bacteria/ hemocyte). Food availability and reproductive state are the most probable explanations for the site differences observed. Finally, it appeared difficult to link oyster survival during summer mortality to hemocyte profiles evaluated at one sampling date; other relevant indicators would probably help explaining oyster survival during summer mortality events. (c) 2007 Elsevier B.V. All rights reserved.