FISHERY BULLETIN: VOL. 78, NO. 3 



yolk sac was absorbed. The criterion for embryo 

 mortality was the appearance of opacity of the 

 embryo, and the criterion for larval mortality was 

 failure to respond to a prod with a polished glass 

 rod. Cumulative mortality with time, percentage 

 hatching, and the stage of development at mortal- 

 ity were taken to be indices of the toxic effect of 

 copper. 



Total copper concentrations were measured 

 every other day during all tests. Labile copper 

 concentrations were measured every other day 

 during embryonic test III and larval tests I and II. 

 Total copper in samples containing >200 ixg Cu/1 

 was determined by direct aspiration of seawater 

 into the flame of a Model 303 Perkin Elmeri" 

 atomic absorption spectrophotometer (AAS) with 

 a deuterium background corrector; total copper in 

 samples containing <200 and >10 ^ig Cu/1 was 

 determined by direct injection of a sample aliquot 

 into an HGA 2100 model graphite furnace after 1:1 

 dilution of the sample with ultrapure 2 N HNO3. 

 Labile copper, defined operationally as that frac- 

 tion passing through a 0.45 ^tm filter and retained 

 by NHj-Chelex resin, was determined by the 

 method of Riley and Taylor (1968). Eluants from 

 the columns were analyzed directly in the flame or 

 in the graphite furnace of the AAS. 



The mean total copper concentrations measured 

 during each test are given in Tables 1 and 2. The 

 percentage of the total copper in the labile form for 

 all concentrations averaged 96% (SD = ±2.60). 

 The mean pH of the exposure seawater for all tests 

 was 8.06 (SD = ±0.05). 



The primary measure of toxicity for this study 

 was the copper concentration resulting in 50% 

 mortality over a given time (median lethal con- 

 centration, LCgp). This toxicity measure was de- 

 termined by performing weighted least squares 

 estimates and maximum likelihood estimates for 

 the parameters cc and f3 in the logit model: 



Pix) 



g^ + fix 



1 + e X + /iv 



The linear transform of the logistic function is 

 logit P = lnP(x)/l-P(x) = oc + /3x; thus if logit P 

 is plotted against x, the points should fall on a 

 straight line with cc as the intercept and /3 as the 

 slope (Berkson 1953). The weighted least squares 



estimates for « and (i were found first and then 

 used as the initial estimates for the maximum 

 likelihood estimates (Koshiver and Moore 1979). 



In our calculation of LC5o,P(x) is the proportion 

 responding at dose x. Our method followed that 

 outlined by the American Public Health Associa- 

 tion (1976) except that the logit analysis was used 

 in place of a probit analysis. For each observation 

 time, an estimated LC50 value was determined. 

 The series of LC50 values obtained were used to 

 construct a toxicity curve that was used to esti- 

 mate the incipient lethal concentration (lethal 

 threshold concentration, ILCg^; Sprague 1969). 



RESULTS 



Northern anchovy embryos continuously ex- 

 posed to copper showed high mortality during the 

 first 8-10 h of exposure (Table 1). After 10 h, the 

 mortality rate was relatively constant until hatch- 

 ing (Figures 2-4). The embryos took two different 

 forms at mortality. The first form (Type I) was 

 observed predominantly during the initial 8-10 h 

 of exposure and accounted for varying proportions 

 of the total mortality, depending on the copper 

 exposure concentration (Table 3). These embryos 

 appeared to have had epiboly disrupted; the yolk 

 was naked and a deformed opaque mass of proto- 

 plasm was found at the animal pole. The second 

 form (Type II) appeared similar to normally devel- 

 oping embryos (the embryo encircling the yolk 

 sac), except for an opacity of the embryo. In em- 



100 - 



10 



20 30 40 



Time — h 



50 



'"Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA, or the University of 

 California. 



Figure 2. — Percentage cumulative mortality of northern an- 

 chovy embryos continuously exposed to copper during test I: 

 numbers next to curves are the exposure concentrations in ju.g 

 Cu/1. 



678 



