The material in 12 of the I4 containers received gentle aeration 

 continously from a small aerator; the main air line from the 

 aerator was equipped with a non-absorbent cotton filter. To 

 preclude possible excessive oxygen demand by Go brevls , light was 

 provided continously with two fluorescent lamps equipped with 

 two 18- inch, 15-watt daylight tubes, A photograph of the experi- 

 mental setup is presented in figure 1, 



Six different bacteria-free cultures (containers '}, 4, 6, 1, 

 9t 10, 12, and 13) and four batches of sterile control material 

 (containers 5, 8, 11, and I4.), all of which had incubated for a 

 month, were employed in this study. The control material consisted 

 of uninoculated culture medium otherwise subjected to the same con- 

 ditions as the inoculated medium. The sterility of the G. brevis 

 cultures and control materials was established by inoculating routine 

 sterility-test media with samples withdrawn from the culture vessels 

 shortly before these materials were dispensed into the experimental 

 containers. Duplicate containers of two of the six bacteria-free 

 cultures (3 and 4-j 6 and 7) were set up to compare the survival 

 of test fish in aerated (containers 3 and 6) and non-aerated (containers 

 4. and 7) cultures, A year-old unialgal G, brevis culture (container l), 

 which was replenished with fresh medium about 4. days previously, was 

 used to compare the effects of unialgal and bacteria-free cultures. 

 Aged sea water (container 2) served as control material for the 

 entire experiment. The volume of test material placed in each 

 container varied from 750 to 1000 ml. 



Before adding the fish, samples were taken from four containers 

 of bacteria-free G„ brevis culture (3,6,9, and 12) and two con- 

 tainers of uninoculated culture medium (5 and- 11) for bacterial 

 counts. Also at this time the nine containers with G, brevis 

 were sampled for enumeration of this organism. The samples taken 

 for the initial bacterial counts were refrigerated from 1 to nearly 

 3 hours before preparing the pour-plates. Most of the G, brevis 

 counts were completed within a few minutes to an hour after with- 

 drawing the samples, A few samples, however, stood for a maximum 

 of approximately 3 hours. 



Each of the I4 containers received two M. cephalus and two 

 C, variegatus except the container of sea water (2), which received 

 three M, cephalus . Each fish was removed from the container shortly 

 after it died so that test materials would not become excessively 

 fouled by decomposing fish. 



After commencing the experiment, samples were again taken for 

 Go brevis and bacterial counts. Seven of the containers with G, 

 brevis were sampled for counts of this organism immediately after 

 the death of the last fish in the container. Two other containers 

 of Go brevis (12 and 13), in which the last fish in the container 



21 



