462 



BOTANICAL GAZETTE 



[DECEMBER 



The tanks, which were specially constructed for the experi- 

 ments, were 76 cm. in diameter and 81 cm. in depth, inside measure- 

 ments (fig. 1). Both cylinder and bottom were constructed of 

 heavy galvanized iron, all joints being well riveted and soldered. 



A heavy iron band was riveted to the rim, through which were 

 bored four holes, allowing the tank to be supported by spikes 

 driven into the framework of the pontoons. Lighter iron bands 

 were placed at intervals along the length of the tank, to which 



were riveted internal iron 

 braces, one of which is 

 shown at a> fig. 1. The 

 lower edge of the tanks 

 projected about 10 cm. 

 below the bottom. Pieces 

 of wood 5 cm. by 10 cm. 

 in cross-section were fitted 

 into this space and spiked 

 to the projecting sides, as 

 shown in fig. 2. This con- 

 struction made a tank of 

 great stiffness, which was 

 not easily deformed by 

 rough handling or the tilt- 

 ing caused by wave action, 

 and which produced a solid 



surface to which the meas- 

 uring device was secured. 



Fie. 1 



deformation 



The foregoing features were 



the experiments would cause a corresponding change in the volume 

 of the tank and a similar error in the reading. 



measuring device was a modification 



Hooke 



- 



made 



mm 



diameter 



The lower end of the tube was threaded and screwed into a socket 

 which was riveted and soldered to the bottom of the tank, while 



