The body size approach was used to deal with the diversity of species and developmental stages of copepod assemblages. Zooplankton samples for gut pigments, taxonomic composition and abundance were collected by vertical tows from bottom (or from 200m where the depth was greater than 200m) to the surface with a 80cm diameter conical net of 500 μm mesh size for the large (> 1000 μm) and medium (500 - 1000 μm) size groups, a modified Bogorov net of 200 μm mesh size was used for the small (< 500 μm) size group. Only the cod end content was immediately poured into soda / seawater solution (1:5. V / V) to anaesthetize; the animals were then sieved through 1000 μm, 500 μm and 200 μm meshes, and the resultant fractions were filtered onto GF/C filters. The filters were frozen and kept at -20°C for subsequent analysis. Samples for evacuation experiments were taken at 0, l0, 20, 30, 60, 120min. during the evacuation experiments using the same method for gut pigment analysis, but anaesthetic was not used before experiments. Gut pigments were measured at 9 stations in 1993 cruise, and at 25 stations in 1994 cruise. Evacuation experiments were undertaken on board at two time-series stations, where the diurnal feeding rhythms of copepod assemblages (by gut pigments measurement), primary production and new production were also observed. The results show that mean level of gut pigments increases with increasing size, but is relatively low. The same distribution pattern of gut pigment contents was found in all the three groups, although the absolute value are quite different, which implies that grazing of the 3 size groups were affected by the same environmental factors (food availability, temperature, etc.). The large size group of copepod assemblages (dominated by Calanus sinicus) performed clear diurnal feeding rhythms with maximum level at midnight and with minimum level at noon time The medium size group (dominated by Paracalanus parvus and early copepodid stages of Calanus sinicus) showed the same rhythms, but with a smaller amplitude. Feeding cycles were not apparent in the small size group (mainly Oilhona spp.). The fact that the small size group is associated with relatively weak feeding rhythms is probably related to the diurnal vertical migration. Large copepods, like Calanus sinicus, feed in the upper layer during the night and migrated to the deep layer during the day time. However this is not the case for small copepods. In spring the fraction of primary production consumed daily by the copepod assemblages was relatively small 6.75% at St. 410 and 4.86% at St. 111(Tab.4). This is probably due to the high primary production and low biomass of copepods. Daily consumption of phytoplankton was mainly caused by grazing of the large size group (58% - 76%). Grazing by the medium size group accounted for only 16%-17%. Grazing by the small size group was 6% - 25% (Tab. 4). The daily fecal pellets production carbon estimated at the two stations were 4.56 and 7.44 mg/(m²*d), respectively (Tab. 4). Among the three size groups the large size group plays a more important role in phytoplankton consumption and in fecal pellets production.

There might be some sources of underestimation in the present results. Pigment destruction in gut by digestion may occur. Evacuation may occur before anaethetization caused by disturbance during the sampling. The underestimation may also be caused by the decrease of evacuation rate under starved condition.