F. G. HALL. 



of assurance that the mouth of University Creek will be covered 

 with a mat of this duckweed which prevents the processes of 

 photosynthesis from going on below. Thus a region of very 

 low oxygen is produced. 



A wire cage, three feet high, three feet wide, and three feet 

 in length, was sunk in this region. The experimental fishes 

 were placed in this cage, having been previously kept in a cage 

 of similar dimensions in regions of more normal oxygen content. 

 The fish were left in the cage from one to two hours as recorded 

 in Table IV., which shows the results of the experiments. 



Another series of experiments was carried on in the laboratory. 

 A large still of thirty-five gallons capacity, loaned by the Wiscon- 

 sin State Prohibition Commissioner, was used for obtaining 

 oxygen-free water. The water was brought to the boiling point 

 and siphoned through a series of condensers. The water was 

 conveyed directly from the condensers to an aquarium which 

 was kept at the same temperature as the water in which the fish 

 were accustomed to live. Fishes were placed in this aquarium 

 and removed at intervals. The gaseous content of the swimblad- 

 der and the amount of dissolved gases in the water were deter- 

 mined from samples taken simultaneously. The results of 

 these experiments are shown in Table V. Control fishes were 

 kept in aquaria at the same temperature and where other con- 

 ditions were as similar as possible to that in the aquaria where 

 fishes were subjected to low oxygen. The swimbladder analyses 

 of ten controls averaged 16.0 per cent, oxygen and 0.74 per cent, 

 carbon dioxide. 



The results of these experiments show that when perch were 

 subjected to water of low oxygen content in Lake Mendota 

 averaging 2.1 cubic centimeters per liter, the percentage of 

 oxygen decreased, and was 12.1 per cent, as compared with the 

 normal of 19.9 per cent. When perch were subjected to water 

 containing extremely small amounts of oxygen, averaging 0.67 

 cubic centimeters per liter, in aquaria in the laboratory, the per- 

 centage of oxygen in the swimbladder decreased and was 9.78 

 per cent, as compared with normal controls having 16.0 per cent, 

 oxygen. 



These results lead one to speculate concerning to what extent 

 the ability of the fishes to re-absorb oxygen from the swimbladder 



