97 



were tested, such as earthenA\'arc filter diaphragms, asbestos 

 cloth, pumice stone, bolting silk, muslin, &c., &c., and it 

 was found that the best result was obtained by using three layers 

 of bolting silk, which broke up the air in a highly satisfactory 

 manner when the pressure was properly regulated. It was decided, 

 however, not to use this at first, as it was considered that it would 

 not completely withstand the action of sea-water, and in the end 

 cane was chosen, though it is by no means a thoroughly satisfactory 

 medium owing to the great variation in its porosity. Experiments 

 are being made with other materials, such as elm, which already 

 has been found to give promising results. 



The final nozzles were made about IJ inches in diameter, and 

 of sufficient length to take a piece of cane about 1| inches long, with 

 suitable arrangements for replacing the cane by discs of various 

 material in carrying out further experiments later. 



These special nozzles were used only in the case of the deep 

 tanks in the main aquarium, as they are not suitable for the shallow 

 tanks in the workers' aquarium, or the various dishes in the private 

 laboratories. In the two latter cases the best results so far have 

 been obtained by sealing one end of a rubber tube which has been 

 punctured by a needle at several points for a length of about 4 

 inches. Fine bubbles in large quantities and over a considerable 

 extent are thereby produced, effectively aerating a shallow tank. 



With reference to the compressor, it was decided to compress 

 the air to the precise degree required, and run the installation 

 continuously day and night with a view to the greatest economy. 

 A special form of rotary compressor was adopted, driven by an 

 electric motor, automatically increasing in speed to meet a greater 

 demand for air. By this means the apparatus is comparatively 

 self regulating, though a bye-pass relief valve was fitted so that any 

 excess of air should be released and returned to the suction side of 

 the compressor. Besides this automatic regulation and bye -pass 

 valve, a suitable regulating, switch was installed to bring the speed 

 of the blower down to the minimum requirements in order that, 

 with only a few tanks at work, both consumption of current and 

 wear and tear on the compressor might be reduced. The maximum 

 pressure required in any of the tanks was found to be 2 lbs., due to 

 the head of water against which the air had to be forced. It was 

 therefore decided to make an allowance for back pressure on the 



