MOVEMENTS OF HERRING AND OTHER MARINE FISHES. 329 



almost to the center while the central region was essentially- 

 neutral. Consulting graph 5, Chart I., we note that the fish 

 moved the entire length of the tank for two minutes and then 

 began to turn back before the highest salinity was reached. 

 After a few such turnings it went the entire length of the tank 

 for a short period with one exception. Between the 7th and 14th 

 minutes the excursions into the salt water were gradually 

 shortened. In other words after a few brief entrances into the 

 salt water the fish gradually shortened its invasions of the salt 

 water until it was turning rather regularly just on the alkaline 

 side of neutrality, which continued to the end of the observation. 

 It will also be noted that the fish turned back twice from the 

 fresh water end, which is significant because in other cases the 

 fishes selected this region. This was true in four other experi- 

 ments with the incline and in six out of eleven performed without 

 the incline. It appears that the herring select either brackish 

 or slightly alkaline water. The control, graph 6, is symmetrical. 



In some of the experiments performed with the incline there 

 was a slight difference in temperature between the two ends, 

 the fresh water being a little higher. To check this source of 

 error, the experiment was performed with the incline but with 

 the difference in temperature reversed, and the fishes selected 

 the opposite end of the tank, showing that this was not only a 

 reaction to solutes but that the solutes inhibited any reac- 

 tion to temperature that might otherwise have taken place 

 (graph 7). In the temperature experiments the fishes selected 

 the higher temperature when the stock was fresh and the lower 

 temperature near the close of the work showing that the fishes 

 had undergone some slight physiological change during their 

 stay in the float-tank. 



The tendency to come to rest in the region on the alkaline 

 side of neutrality was very clearly shown in all the experiments 

 except one. The salmon oriented with their heads toward the 

 fresh water end, drifting very slowly back, probably floating 

 in a current and then swimming up again to the same point. 

 This was very striking and constituted an unmistakable difference 

 between the experiment and control. Chart II., graph 13 and 

 14, show such an experiment and control. The swimming up 

 occurred notably in the 13th and i8th minutes. 



