36 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
with the former, as the readings are sufficiently accurate for the pur- 
pose, and they are much more convenient in actual use. Two of 
them were used, their rating being so close that there was no appreci- 
able difference between them. 
We carried four Negretti and Zambra reversing deep-sea thermom- 
eters, unfortunately without auxiliary thermometers for taking the 
temperatures of the detached thread at the moment of reading, such 
as are provided in their latest pattern and in the Richter thermometer. 
Two of these were rated in the U. S. Bureau of Standards at Washing- 
ton, two in the Chemical Laboratory at Harvard University, with the 
following results :— 
Negretti and Zambra Thermometer, U.S. B. F., No. 7,277. 
Reading, °F Correction, °F when 7’. of detached thread is 
3 60° 90° 
32° —.3° —.5° — 6 
60° —.6° —.9° —1.1° 
90° 0 +.3° + .6° 
Negretti and Zambra Thermometer, U.S. B. F., No. 7,259. 
32° —.5° — .8° —1.1° 
60° —.2° —.5° — ,9° 
90° +.4° 0 — .5° 
It is fortunate that the changes in reading consequent on change 
of temperature of the detached thread are so small, for without the 
use of a water-bath, which was not available, the temperature of the 
detached thread could be obtained only by allowing the instrument 
to come to the temperature of the air before reading. 
The corrections for Nos. 84,036 and 49,648 were noted with the 
temperature of the detached thread the same as that of the readings; 
i. e., the freezing point reading was taken at an air temperature of 
32°, the 68.5° reading at an air temperature of 68.5°. They are as 
follows :— 
84,036, 32°, correction —.2°; at 68.5°, —.55°; at 77.13°, —.82°. 
49,648, 32°, correction —.5°; at 68.5°, —.16°; at 77.13°, —.37°. 
With both these thermometers the requisite correction at readings 
between 40° and 50° is about —.3°: and though this is not exact, varia- 
tions from it, within this range, are less than the probable error of 
