183 



which were significajitly more toxic than physically- 

 dispersed oil. Concentrate dispersants usually produced a 

 dispersion which appeared to have a toxicity greater than 

 the oil alone but only in five cases was this increase in toxi- 

 city statistically significant (Student's t-test; P = <0.05). 



3.2 Tests under non-standard conditions 



In order to validate the test procedure used for licensing 

 purposes, experiments were conducted to determine the 

 sensitivity of the test to variations in test conditions. These 

 are described in Sections 3.2.1 - 3.2.8. 



3.2.1 Effectiveness of physical dispersion 



Although every effort was made to form a repro- 

 ducible dispersion of oil by physical means, the 

 degree to which the oil was mechanically dispersed 

 by the propeller action tended to vary between the 

 test tanks. In Table 2 a subjective assessment based 

 on a visual observation of the degree of dispersion 

 is compared with the mean mortality (%) in a series 

 of 284 test tanks. These data suggest a strong link 

 between the subjective assessment' of the degree of 

 dispersion and its toxicity. The 'poor' dispersions 

 were characterised by coarse oil droplets and thus a 

 low on/water surface area: the 'good' dispersions 

 by fine droplets and a larger oil/water surface area. 

 The suggested association between the degree of dis- 

 persion and toxicity may be due inter alia either to 

 the differences in oil droplet size (perhaps affecting 

 its availability to the organism) or to enhanced dis- 

 solution of soluble oU fractions across the increased 

 oil/water interface. 



tank was fUtered through a Whatman No. 1 fJter 

 paper and the filtrate analysed for 'dissolved' oil by 

 UV fluoresctnte spectrophotometry. The results 

 (Table 3) showed that the fuier dispersions using 

 Method 1 were accompanied by an increase in the 

 concentration of oil in solution. No such clear rela- 

 tion was obtained with Method 2, mainly because 

 no 'poor' or "very good' dispersions were observed. 

 The concentiarson of the water-soluble fraction of 

 oil required to cause mortality of a range of species 

 during a 96 h exposure period was found by Rice 

 et a/.(1976a, i) to be in the range 1.8-10.8 /j1 1''. 

 In the sea tea animals were exposed only for 100 

 min, and it appears unlikely that the concentration 

 of the water-soluble fraction was responsible for the 

 difference in toxicity observed; it seems likely that 

 droplet size was a more important factor. 



Regardless of the cause of the change in toxicity, it 

 is clear that variability in the degree of dispersion of 

 the physicallyJispersed oU in the cont:ol can mark- 

 edly influence mortality. In order to minimise this 

 possibility and to ensure consistency of standard 

 tes". conditions, results of tests in which the oil 

 dispersion arrears to be exceptionally poor or 

 variable are discarded. 



3.2.2 The rsiationship between dispersion charac- 

 teristia and toxicity 



To investigate further the relationship between the 

 droplet size of the dispersed oil and the toxicity of 



Table 2 - The relationship between extent of dispersion and toxicity of physicaUy-aaispersed oil in the sea test 



A simple experiment was conducted to determine 

 whether finer dispersions were associated with in- 

 creased amounts of oil in solution. For this purpose 

 12 tanks were set up under standard test conditions; 

 oil was added to six tanks already subject to agitation 

 (as in Method 1 in Section 2.2) and to another six 

 tanks before agitation was started (Method 2, in 

 Section 2.2). The degree of dispersion was varied 

 within each set of six replicates and assessed at the 

 end of the 100 min exposure period. At the end of 

 the experiment a sample of the oily water from each 



the dispfsma, a limited number of tests were carried 

 out in »indi some of the characteristics of the oil 

 dispersions poduced by both methods of dispersant 

 application ■«ere investigated. The standard test con- 

 ditions WBE ased for these experiments with analyses 

 of water 5aE5»les taken from the centre of the tanks 

 during l3i£ !30 min exposure period. The size distri- 

 bution of As oil droplets in suspension in the oil and 

 dispersant inixtures was measured using a Coulter 

 counter »i!i 200 Mm aperture. Although the Coulter 

 counter dateminations were carried out immediately 



