G. brevis counts were completed between 1 and 2 hours after collecting 

 the samples . In addition, samples for pH and salinity determinations 

 were taken from all 18 containers. Bacterial, G. brevis , pH, and 

 salinity samples were taken either shortly after the death of the 

 last fish in the container or at the end of the test period (24 hours) 

 providing one fish in the container survived, except for some pH and 

 salinity samples as noted xmder "remarks" in table 8. The room 

 temperature varied from 22 to 25° C. during the 24-hour test period. 



The fish subjected to bacteria-free cultures as well as filtrates 

 and residues of such cultures died more rapidly for the most part than 

 those exposed to the control materials (Table 8). Some of the control 

 fish, especially small M. cephalus , did not survive well. Fish of 

 this size group died rapidly in one batch of sterile culture medium 

 (containers 13 and 14). The "death times" varying from 25 minutes to 

 2-2/3 hours for the small mullet in these two containers were 

 comparable to the "death times" varying from 15 minutes to 2-| hours 

 for the same sized fish in unfiltered bacteria-free and unialgal 

 G. brevis cultures. However, the large M. cephalus in the two 

 containers of culture meditmi sxirvived considerably longer than those 

 in the G. brevis cultures. One large fish survived the 24-hour test 

 period and the other lived at least 8-^ hours. In contrast, the 

 "death times" for the six large mullet in three different unfiltered 

 G. brevis cultures, two of which were bacteria-free, varied from 10 

 to 41 minutes. All fish in the sea water lived for at least 7^ hours 

 and two of them survived the test period. 



The early deaths of the small mullet in containers 13 and I4 

 were probably due to the abnormally low pH of the culture medium. 

 The pH of the material in container I4 was 5.9 about I4 hours 

 after beginning of experiment and increased to 6.4 after 25 hours. 

 A pH value of 6,6 was obtained for the culture mediiim in container 

 13 at the end of the test period. 



The control fish lived considerably better in the batch of 

 sterile culture medium used in the filtration phase of this experi- 

 ment. All of the large mullet (4) and two- thirds of the small mullet 

 (8) subjected to filtrates and residues of culture medium survived 

 the 24-hour test period. On the contrary, none of the fish (4 large 

 mullet and 12 small mullet) exposed to filtrates and residues of the 

 G. brevis cvlture survived the test period. The difference between 

 the effects 01 the two methods of filtration on the toxicity of 

 the G. brevis culture was marked. The millipore filtrate was much 

 more toxic to the fish than the residues; in the filtrate the "death 

 times" varied from I4 to IO4 minutes in contrast to a variation of 

 5 to 132- hours in the residues. The toxicity of the filter paper 

 residues was greater than the filtrate; the test fish lived from 39 

 to 42 minutes in the residues whereas they survived from 44 minutes 

 to about 8 hours in the filtrate . 



28 



