each of the 18 containers. Since the test materials were aerated 

 more vigorously in this study, the increased air flow made equalizing 

 the degree of aeration in each container difficult and the agitation 

 of the test materials probably varied more. 



One portion of the experiment (containers 1 through 8) was 

 designed to compare the effects of millipore (HA membrane) and 

 paper (Whatman No, 4.O) filtration on the toxicity of one of the 

 bacteria-free G. brevis cultures. One batch of the sterile culture 

 medium treated in the same manner as the G. brevis culture served as 

 control material. The millipore and paper residues of both the G. brevis 

 culture and the sterile culture medium were each eluted in 1 liter 

 of sterile culture medium to obtain four of the eight test materials 

 used in this phase of the study. The test materials for the remaining 

 portion of this experiment (containers 9 through 18) consisted of 

 duplicate containers of two different bacteria-free G. brevis cultures, 

 sterile culture medium, autoclaved 85% aged sea water, and a unialgal 

 G. brevis culture. The distribution of the test materials and 

 numerical designation of the containers are presented in table 8. 



Samples were taken for bacterial and G. brevis counts 

 immediately after dispensing the test materials . All containers of 

 unfiltered bacteria-free G. brevis culture (9, 10, 11, and 12) and 

 unfiltered culture medi^um ( 13 and I4) in addition to a container of sea 

 water (15) and unialgal G. brevis culture (17) were sampled for 

 bacterial counts. None of the containers of filtrates or residues 

 were sampled because such counts were not needed. These samples were 

 refrigerated 4j to 6-| hours before plating. Samples for G. brevis 

 counts were obtained from the containers with filtrates of G. brevis 

 culture (1 and 2) and bacteria-free G. brevis (9, 10, 11, 12). The 

 G. brevis concentration of the unialgal culture used in containers 

 17 and 18 was ascertained by withdrawing a sample from the culture 

 prior to dispensing. The G. brevis samples were counted between 

 1^ to 2-1/3 hours after collection. 



Each container received four fish: Three small mullet (l to 1-^ 

 inch long) and one large mullet (^l to 5^ inches long). The small 

 mullet were maintained in aerated aquaria overnight and the large 

 mullet were used within a few hours, after collection. The volume of 

 test material (l liter) was somewhat small for some of the large mullet, 

 which thrashed about vigorously. This activity possibly injured some 

 of the smaller test fish. ' 



Excepting the filtrates of G. brevis cultures (containers 1 

 and 2), the bacteria and G. brevis were again enumerated for those 

 containers for which such counts were made initially. The second 

 bacterial samples from the five containers with G. brevis (9, 10, 

 11, 12, and 17) were plated after A— 3/ A to 5^ hours refrigeration 

 whereas those from the containers with culture medium (13 and I4.) 

 and sea water (15) were stored only 2/3 to 1-^ hours. The second 



27 



