72 REPORT OF COMMISSIONER OF FISH AND FISHERIES. [70] 



further so long as the bulb is upward, and thus the temperature to be 

 read will not be affected. 



jSTow, whenever the thermometer can be handled it can readily be 

 turned bulb upward for reading the existing temperature. It must be 

 clearly understood that this thermometer is only intended to give the 

 temperature at the time and place where it is turned over ; it is simply 

 a recording thermometer. In its present state it cannot be used as a 

 self- registering maximum and minimum, though, if required, it could 

 be constructed to act as a maximum. 



In order to make the thermometer perfectly satisfactory, it was neces- 

 sary to protect it from pressure as well in shallow as in the deepest 

 seas, for in either case the pressure would cauTse an error of greater or 

 less degree in its indications. Like an ordinary thermomieter it is de- 

 void of air, and so quite different from Sixe's, which, containing com- 

 pressed air, has a certain internal resistance. Hence it would be more 

 affected by pressure than Sixe's thermometer, however thick the glass 

 of the bulb. By the simple expedient of inclosing the thermometer in 

 a glass shield, e, hermetically sealed, the effect of external pressure is 

 entirely eliminated. The shield must of course be strong, but not ex- 

 hausted of air. It will, however, render the inclosed thermometer less 

 readily affected by changes of temperature, making it more sluggish. 



To counteract this tendency mercury is introduced into that portion 

 of the shield surrounding the bulb, and confined there by a partition, rf, 

 cemented in the shield around the neck of the thermometer bulb. This 

 mercury acts as a carrier of heat between the exterior of the shield and 

 the interior of the thermometer; and the efficacy of this arrangement 

 having been experimentally determined, the instrument has been found 

 far superior in sensibility to Sixe's. 



So long as the shield withstands the pressure — that is, does not break — 

 the thermometer will be unaffected by pressure, and there is a,buudaut 

 experience to show that such a shield will stand the pressure of the 

 deepest ocean. Doubtless the shield will be slightly compressed under 

 great pressure, but this can never cause an internal pressure sufficient 

 to have an appreciable effect upon the thermometer. This method of 

 shielding is, therefore, quite efficacious, and deep-sea thermometers so 

 ])rotected do not require to be tested for j^ressure in the hydraulic press. 

 They simply require accurate tests for sensitiveness and for errors of 

 graduation, because they are standard instruments adapted to the de- 

 termination of very small as well as great differences in temperature, 

 some one or two tenths of a degree in shallow water. The tost for sen- 

 sitiveness should determine the time the instrument requires to take 

 up a change of 5°, rise or fall, and the time is found to be from five to 

 ten seconds. 



Thus, ])rovided the turning-over gear is found to answer, this instru 

 ment evidently possesses great advantages. It has no attached .^eale, 

 the figures and graduations being distinctly marked on the stem uself. 



