888 REPORT OF COMMISSIONER OF FISH AND FISHERIES. 



Ko. 8. The same, no number. Bought from James Greene, New York, 

 and marked J. G. 



Kos. 1, 2, 3, and 4 agreed closely when in equilibrium. No. 1, as the 

 most easily read, on account of its larger scale, was selected as the 

 standard by which to record the actual temperature. Its action was so 

 rapid that no attempt was made to record it during its rise or fall. 



No. 6 was defective in that it had an index error averaging 2^° to be 

 added. It was, however, sensitive, and had fair action. 



No. 7 recorded correctly, or nearly so, when in equilibrium. 



No. 8 recorded more slowly, and had an index error of about .5 of a 

 degree subtractive. 



Our first experiments were made in a tub, containing about ten inches 

 of water, on the surface of which about three inches of broken ice 

 floated; the instruments were laid flat and read through the water. 

 The necessity of constantly agitating the water caused the broken ice 

 to frequently interfere with the range, and produced delay and inaccu- 

 racy, and if the water was not thoroughly mixed by stirring, no accu- 

 rate results could be obtained, the difference in temperature between a 

 point one-half inch below the ice and at the bottom, sis inches deep, 

 amounting to six degrees, when the water was tranquil, and so great 

 was this cause of error that between No. 1 lying flat and No. 4 mounted 

 on a metal back, which raised it over one-half inch above No. 1, there 

 was a difference, due to position, of nearly 1°. A bath-tub was there- 

 fore devised, which enabled us to record correctl3\ This consisted of 

 an ordinary aquarium, twenty inches by twelve and fourteen inches 

 deep. This was divided longitudinally by a screen of wire gauze into 

 two compartments, in one of which, nine inches wide, a bath of sea- 

 water, well filled with broken ice, was put from a frame bar lengthwise, 

 and six inches above the narrow section the thermometers were hung, 

 with their bulbs on the same level in the iced water, but free from all 

 contact with the ice. 



The apparatus being placed so that the surface of the mercury was in 

 line with the eye (an important precaution where extreme accuracy is 

 desired), the record was read through the glass side and half-inch of 

 water, which, slight enough to present no obstacle to a clear view, was 

 all-sufficient to protect the delicate instruments from external influ- 

 ences. 



METHOD OF EXPERIMENTS. 



The standard No. 1 was suspended in the water, which was drawn 

 from over the side of the vessel, and cooled down by ice to the de- 

 sired temperature. An assistant, with a bucket of hot water, a bucket 

 of finely-broken ice, and a table-spoon, stood ready, after the water had 

 settled to the required temperature, to add at each reading a little hot 

 or cold, as the standard's record, read aloud, indicated two or three 

 tenths below or above the bath we wished, while at the same time, with 



