To understand energy passing frorn bacterial production to metazooplankton in an aquatic ecosystem, the relative significance of bacterial and photosynthetic production as carbon sources for metazooplankton was determined in a 14 d coastal mesocosm experiment, using frequent in situ batch incubations with dissolved organic and inorganic '-'C tracers. The contnbution ratio (CR) of bacterial to photosynthetic production as rnetazooplankton food resources fluctuated from 6 to 41 O&, according to successional changes in the plankton community structure in the mesocosm. The lowest CR (6%) occurred when nano-sized autotrophs were highly abundant and appendicularians Oikopleura sp., which are capable of ingesting only pico-to nano-sized particles, increased rapidly. After these small autotrophs disappeared, the CR increased to 30-40%, suggesting that the appendicularians were supported largely by bacterial production at that tlme. When the metazooplankton consisted mainly of copepods. the CR reached about 10% despite the relatively low ratio (ca 15%) of net bacterial community production to photosynthesis; also. the copepods would have been able to exploit bactenal production only through intermedianes such as bactenvorous protists. In this penod with copepods dominant, Gymnodinium mikimotoi was the dominant autotroph and might have been a main source of prey for grazers. However, copepods avoid this dinoflagellate, as often reported. Consequently, not much of the autotrophic production of the dinoflagellate might have been ingested by the copepods. We conducted that nano-sized autotrophs and herbivorous protists were likely to be the main suppliers of photosynthetic production for copepods, as nano-sized bactenvores were the suppliers of bactenal production. These findings suggest that bacterial production and the carbon pathway, 1.e. a microbial loop. are not always the terminus of organic carbon flow in food webs, but can be an integral component of the food pool for metazooplankton in the coastal environment.