FISHERY BULLETIN: VOL. 76. NO. 2 



to four times during each bioassay to determine 

 the actual concentrations to which the organisms 

 were exposed. Water samples were collected in 

 acid-washed polyethylene jars and acidified to pH 

 2 with concentrated HCl. Total copper was 

 analysed by the APDC-DDDC-MIBK extraction 

 method described by Kinrade and VanLoon 

 (1974). The copper concentration in extracted 

 MIBK solutions was determined with a model 303 

 Perkin Elmer atomic absorption spectrophotom- 

 eter, using an HGA-2100 graphite furnace with a 

 deuterium background corrector. 



Since the chemistry of copper in seawater is 

 complex, more than one form of copper may be 

 present in the bioassay water. To examine the 

 form of the copper in the bioassay system water, 

 out-flow samples were collected from the bioassay 

 system before organisms were introduced to de- 

 termine the particulate-bound fractions 

 (>0.45^im), ionic fraction (bound by Chelex-100 

 resin (Riley and Taylor 1968)), and complexed 

 fraction (not bound by Chelex-100 resin). The 

 analysis scheme is summarized in Figure 2. To 

 monitor the partitioning of copper into each of 

 these fractions, copper-64 was equilibrated with 

 water samples after they were withdrawn from 

 the bioassay system. The partitioning of stable 

 copper in the seawater of the bioassay system was 

 indicated by the percentage of the initial activity 

 recovered in each of the described fractions. 



Statistical Analysis 



The measures of toxicity determined in this 

 study were the time to 50^f mortality at each con-'^ 

 centration of copper tested (median lethal time, 

 LT50) and the concentration of copper resulting in 

 50% mortality over a given time (median lethal 

 concentration, LC50). These toxicity measures 

 were deterimed by performing a weighted linear 

 regression analysis on the sets of cumulative mor- 

 tality data using the logistic function. The 

 straight line transform of the logistic function is: 

 logit P = In PIQ =a + (3x, so that if logit P is plot- 

 ted against X, the points will fall on a straight line 

 with a as the intercept and /3 as the slope (Berkson 

 1953). In our calculations of LT50, x represented 

 the time from the onset of the reaction period in 

 the case of continuous embryo exposures, from 

 hatching in the case of continuous larval expo- 

 sures, and from the termination of a given pulse in 

 the case of pulsed embryo exposures. In our calcu- 

 lations of LC50, X represented concentration, 



350 



Seawater 



APDC-DDDC-MIBK 

 extraction 



0.45 u filter 



Total 

 fraction 



Chelex 100 



ion exchange 



resin 



APDC-DDDC-MIBK 



extraction 



on waste 



water 



Particulate 



bound 



fraction 



Ionic 

 fraction 



_ Complexed 

 fraction 



Figure 2. — Analysis scheme for the separation of copper frac- 

 tions recovered from the bioEissay system used to expose Pacific 

 herring embryos and larvae to copper. 



and our method followed that outlined by the 

 American Public Health Association (1976) with 

 logit analysis used in place of probit analysis. 



A computer was used to calculate the LC50 and 

 LT50 values, and for each fitted line the program 

 determined: the LT50 or LC50, the 959^ confidence 

 limits associated with the LT50 or LC50, Pearson's 

 rho (p), the slope (/3) and the intercept (a), and the 

 mean square error (EMS); no assumptions of 

 homogeneity were made and the EMS was calcu- 

 lated in every case, rather than assuming an EMS 

 of 1 for homogeneous data (Finney 1964). 



In the case of embryos that were exposed con- 

 tinuously or exposed to pulses of copper, deaths 

 prior to the delayed reaction period or the onset of 

 the pulsed exposure, respectively, were not used in 

 the data analysis. In no case were mortalities dur- 

 ing these periods greater than 6%. 



The relationship between time to 50% mortality 

 during continuous exposure of both embryos and 

 larvae and concentration was determined follow- 

 ing the method outlined in the American Public 

 Health Association (1976). The resulting toxicity 

 curve was used to estimate the lethal threshold 

 concentration (incipient LCgo) (Sprague 1969). 



RESULTS 



Physical Parameters of the Bioassay System 



Mean copper concentrations measured during 

 each test are reported in Table 1. The partitioning 



