was used in addition to these two bottles of material, which were 

 <4-| months old at this time. These three materials were tested in 

 duplicate (containers 1 through 6). The test materials in all of 

 these containers, except one of sea water (2), were sampled for 

 bacterial counts just before the fish were added. These samples 

 were refrigerated 20 minutes to 2-^ hours before pouring the plates. 



Five containers of the test material (containers 7 through 11) 

 were used to test the effects of heating and filtration on the 

 toxicity of unialgal G. brevis cultures. Bacterial counts were not 

 made for these materials because such information was not needed. 

 A filtrate which was prepared by passing G. brevis culture through 

 filter paper (Whatman No. 4.2), was tested in duplicate. A single 

 container of another test material consisted of the residues retained 

 by the two filter paper discs eluted in 2 liters of sea water. Two 

 liters of G. brevis culture were passed through each disc. Other 

 test materials included single containers of G. brevis cultures 

 which had been heated to 35 and 45° C. 



Each of the 11 containers (4.-liter beakers) received two common 

 killifish ( Fundulus grandis ), 3 to 3^ inches long. The test 

 materials, 2 liters in each container, were not aerated. 



Only one of the four fish placed in each of the two control 

 materials, sea water and Prorocentrum culture (bottle 2), failed 

 to survive the 4-hour test period (Table 4). On the contrary, 

 none of the four fish subjected to the G. brevis culture (bottle l) 

 were alive at the end of the test period. The "death times" were 

 9, 16, 100, and 13O minutes. Again, however, the G. brevis culture 

 (bottle 1) had a greater bacterial concentration than the material 

 in bottle 2, in spite of the addition of Prorocentrum a month earlier. 

 The count for the former was 2.2 to 2.4 million bacteria per ml in 

 contrast to 0.19 to 0.20 million per ml for the latter. The latter 

 counts are quite similar to those obtained for bottle 2 in experiment 5 

 (Table 3). 



The fish lived 20 to 100 minutes in the G. brevis culture heated 

 to 35° C. In the culture heated to 45° C. the "death times" were 

 only 13 and 18 minutes. Three of the four fish exposed to filtrates 

 of a G. brevis culture survived the experimental period. The two 

 fish subjected to the materials eluted from filter paper through which 

 G. brevis culture had passed died in 23 and 13O minutes. Filtration 

 appears to reduce the toxicity of G. brevis cultures. However, other 

 filtering methods must be tested before this effect can be established 

 as a characteristic of filtration. 



14 



