MODIFICATION OF THE BEHAVIOR OF FISHES 3 



volved re-counting of movements and recalculation of time, etc., 

 which, together with the examination of a large amount of liter- 

 ature has delayed the preparation of this aspect for several months. 



II. CONDITIONS AND METHODS OF STUDY 



The experiments were conducted under conditions as nearly 

 uniform as possible with respect to all factors except the amount 

 of the various solutes used. (For detailed plans and figure of 

 the apparatus, see Shelf ord and Allee, '13, pp. 225-229). Two 

 galvanized iron boxes, 120 cm. long by 14 cm. deep by 20.5 cm 

 wide, with screen partitions 5 cm. from the ends, making the in- 



Control 



:^ + ii +-* 



Experiment 



Figure I 



The upper two quadrangles show the ground plan of the experimental tanks 

 with the positions of the central drain and the lights under which the fishes went 

 when passing to and fro. The positions of the lights are indicated by the crosses 

 and that of the drain by the double broken line near the center. The broken lines 

 near the ends indicate the positions of the screen partitions. The tanks are rep- 

 resented a little less than one-twentieth actual size. The ends A receive tap water 

 and AM treated water. Below the tank used for the experiment the method of 

 recording movements is indicated. A portion of the record sheet is shown about 

 twice natural size, and immediately below the reduced tank so that the marks 

 dividing the tanks into sixths are opposite the corresponding dividing lines of the 

 ruled paper. The movements of two mud-minnows during the first five minutes 

 of expt. 49A are indicated with the time of the principal movements, in minutes 

 and seconds after the beginning, shown in figures. The graph is an enlarged copy 

 of the original . The solid line indicates that the two fishes were moving together; 

 the broken line shows the movement of a single fish. 



