FISHERY BULLETIN: VOL 79. N0.4 



and 0800 h. Plankton was then siphoned into the 

 tank at a constant rate over a 7-h period from 

 about 0800 to 1500 h. At intervals prior to, during, 

 and for about 20 h after the feeding period, respi- 

 ration rates and voluntary swimming speeds were 

 measured (Durbin et al. 1981), and samples of the 

 tank water were collected for determination of 

 ammonia and dissolved organic nitrogen (DON) 

 excretion rates. Feces were periodically siphoned 

 from the tank during the feeding period and for 41 

 h thereafter for the determination of assimilation 

 efficiency. The tank was briefly flushed with fil- 

 tered seawater at the end of the feeding period to 

 reduce the ammonia concentration. Control mea- 

 surements on the tank and tank water with and 

 without plankton, demonstrated that these did not 

 measurably effect ammonia or dissolved oxygen 

 concentrations during the experiments. 



Seven experiments were carried out using cul- 

 tures of the solitary diatom Ditylum brightwelli, 

 and three using Narragansett Bay zooplankton 

 consisting mainly of adult Acartia tonsa and a 

 small number of unidentified crab zoeas. Details of 

 the culturing and processing of the D. brightwelli 

 during experiments are given in Durbin et al. 

 (1981). Narragansett Bay zooplankton were col- 

 lected with 0.5 m diameter, 300 fxm mesh nets on 

 the day before an experiment and maintained 

 alive in 1.2 m diameter tanks overnight. 



During each experiment the plankton was con- 

 centrated into seven equal batches of about 18 1. 

 Each batch was then siphoned into the tank over a 

 1-h period, providing an approximately constant 

 input of food over the 7-h period without greatly 

 changing the volume of the tank (1,400 1). Each 

 batch of phytoplankton was subsampled for de- 

 termination of C and N (Durbin et al. 1981). Each 

 hourly batch of zooplankton was concentrated 

 onto a 100 /xm mesh and weighed just before use. A 

 subsample of this was removed for determination 

 of the dry weight:wet weight ratio, C and N con- 

 tent (Hewlett-Packard Model 185B CHN Ana- 

 lyzer), ash (combustion at 475° C for 4 h), caloric 

 (Parr adiabatic bomb calorimeter), and chitin 

 (Windell 1966) content. The remaining sample 

 was dispersed in about 18 1 of water and siphoned 

 into the tank. By changing the concentration of 

 plankton in the batches, we obtained different 

 concentrations of food in the tank and different 

 ration sizes. Turbulence produced by the swim- 

 ming of the fish kept the water and plankton in the 

 tank well mixed at all times. 



Feces settled to the bottom of the tank, and were 



collected at 1-6 h intervals depending on the rate 

 at which they were eliminated. Feces were gently 

 siphoned into a plastic bucket and allowed to set- 

 tle; the water was then aspirated off, and the feces 

 were briefly rinsed with a little distilled water and 

 transferred to preweighed aluminum dishes. Fecal 

 pellets formed cohesive, cylindrical rods; the loss 

 of material during collection and concentration 

 was nominal. These dishes were then freeze-dried 

 and reweighed to determine the dry weight of the 

 feces. Each sample of feces was subsequently ana- 

 lyzed for carbon and nitrogen (3 replicates), ash 

 (4-8 replicates), and calories (4 replicates). Each 

 sample of feces from the phytoplankton experi- 

 ments was also analyzed in triplicate for particu- 

 late silicon (Durbin 1977). 



Samples for dissolved nitrogen analysis were 

 siphoned from the tank into a clean, acid washed 

 container, and then filtered through prerinsed 

 glass fiber filters. The sampling interval was 1-4 h, 

 depending on the excretion rate. All determina- 

 tions were carried out in triplicate. 



Ammonia determinations were carried out im- 

 mediately with a modification of the method of 

 Solorzano (1969), using 10 ml samples which were 

 incubated in the dark for 4 h for color development. 

 The range of values for 3 replicates was <0.1 jjlM 

 of ammonia. 



Total dissolved nitrogen was determined using 

 the alkaline persulphate method of D'Elia et al. 

 (1977). Nitrate formed during digestion of the 

 samples was determined using a Technicon Au- 

 toAnalyzer II. At the same time, nitrate, nitrite, 

 and ammonia concentrations were also deter- 

 mined on noncombusted subsamples using the 

 AutoAnalyzer. These concentrations were sub- 

 tracted from the total dissolved nitrogen deter- 

 mined from the persulphate digestion to give a 

 measure of DON. Ammonia concentrations of the 

 samples which had been frozen were always very 

 similar to the concentrations determined from 

 fresh samples immediately after collection. The 

 range of values for three replicates of each mea- 

 surement of total dissolved N was s;0.5 /xM N. 



Excretion rates of the nonfeeding fish were low, 

 and thus changes in total dissolved nitrogen con- 

 centration in the tank water were small. Although 

 the ammonia excretion rate could always be satis- 

 factorily determined, the greater errors inherent 

 in the DON determination made it impossible to 

 accurately measure DON excretion except during 

 and soon after feeding, when the N excretion rates 

 were high. During feeding the increase in NH3 



602 



