BIGELOW: EXPLORATIONS IN THE GULF OF MAINE. 



41 



and this may be considered the extreme. There is one other source 

 of error in any reversing thermometer actuated by a propeller; i. e., 

 uncertainty at what precise level the instrument reverses, with possi- 

 bility of change in reading during its passage upward through the 

 column of water necessary to reverse it. But we so often used two 

 thermometers at each level, and so often repeated the entire series, that 

 I do not believe this possible error is of any practical importance in 

 the present case. On the whole, then, it is better not to claim accu- 

 racy closer than =±=.3°F; i. e., roughly, .15°C. And it is certainly 

 much better to set these limits wide, rather than to claim a higher 

 degree of accuracy than can be relied upon. 



The surface thermometers were extremely reliable, and so far as 

 the instruments themselves are concerned very little error is to be 

 expected. But the readings were taken by various persons, often 

 under difficult conditions, therefore accuracy is not claimed beyond 

 =•= .5°F. 



_ioqjBrt flw , \"' 



Fig. 3. — Bathymetic chart of the Gulf of Maine. 



