trast, hydrogen cyanide (HCN) studies show that juvenile fish may be 75-80% 

 more sensitive than the egg in some species. Further, the behavioral 

 inhibitions on spawnig of some species like bluegill and brook trout 

 ( Salvelinus fontinalis ) in respose to H 2 S may be apparent at levels far 

 below those which cause acute mortality in 96-hr or which reduce growth 

 over long periods. These factors make it important that an appropriate 

 life-history stage be selected for testing and that an endpoint be chosen 

 which reflects the true sensitivity of that stage to the toxicant. 



Selection of test fish stocks from widely separated geographical areas 

 or different cultured stock may introduce wide variations in results. In 

 fathead minnows acute sensitivity to cyanide (HCN) and H 2 S may vary as much 

 as 30-40% between stocks. It is also important that the influence of test 

 conditions (temperature, pH, 2 , fish numbers) be taken into consideration 

 when selecting an appropriate endpoint for bioassay and in evaluating the 

 results when they are obtained. 



PREDICTION OF LONG-TERM ADVERSE EFFECTS FROM SHORT-TERM TESTS 



Ideally a toxicant should be administered to a fish or an invertebrate 

 throughout its entire life history, beginning with the egg through repro- 

 duction and into the second generation, to determine concentrations which 

 will not adversely affect the population. However, the large number of 

 known toxicants and unknown mixed toxicants which must be tested will pre- 

 vent definition of safe levels for many materials by long-term tests. It 

 is therefore common practice to make an acute test (96 hr) defining some 

 median tolerance limit (LC50 or TLm) and then to apply a mathematical 

 factor which will reduce the value of this concentration to that considered 

 safe for completion of all life-history stages. This factor is usually 

 called an application factor and is calculated by dividing the safe concen- 

 tration by the 96-hr LC50 (TLm) of the toxicant. Historically, first 

 approximations of this factor were made by comparison of acute median 

 tolerance limit tests and long-term or chronic tests of a limited number of 

 toxicants. The tests were conducted on the same species and under the same 

 conditions. Results of acute tests were also compared to concentration of 

 toxicants in streams which contained normal fish populations. While many 

 materials appear to have application factors of similar magnitude, certain 

 families of materials have factors different by orders of magnitude from 

 established means and factors (Table 1). With a single toxicant the appli- 

 cation factor may vary substantially between species. An example can be 

 drawn from our studies of H 2 S in which the application factor varies widely 

 between species and between life-history stages of the same species (Table 

 2). Also fish species react very differently to materials in similar 

 chemical families (Table 3). These examples indicate that no single appli- 

 cation factor can be used to predict "no-effect" toxicant concentrations 

 from short-term tests. In consequence, uses of application factors with 

 short-term tests have tended to be more restrictive in their results than 

 necessary in some cases and much less restrictive than required in others. 



43 



