1914] Sumner, et al.: Physical Conditions in San Francisco Bay 13 
(3) The surface temperature was taken by means of a Negretti- 
Zambra reversing thermometer, immersed to a depth of a foot or two 
beneath the water surface. The thermometer was operated by a mes- 
senger, which was released after an interval of ten minutes. This 
was believed to be a more reliable method than that of dipping an 
ordinary thermometer into a pail of water on deck, for the temper- 
ature of such water might change appreciably before the reading could 
be made. 
(4) A water sample from the bottom was taken by means of an 
Ekman reversing water bottle. This instrument has been fully de- 
seribed by its inventor (Ekman, 1905a), and no further account of it 
is required here. Two of these instruments were made for the ‘‘ Alba- 
tross’’ by Mr. Valdemar Arntzen, expert mechanic of the University 
of California. 
Both surface and bottom waters were preserved in snap-top pint 
bottles, having a porcelain stopper and rubber ring. The kind used 
were those known to the trade as ‘‘citrate of magnesia bottles.’’ They 
were thoroughly washed with sea water before use. 
(5) The bottom temperature was determined by means of two 
Negretti-Zambra thermometers, mounted in the frame of an Ekman 
water bottle, and reversing with this. The mean of the two readings 
has been employed by us in our computations. The instruments were 
left at the bottom for ten minutes. 
The thermometers used were all standardized, and the proper 
corrections have been applied to the readings. Although the scale, 
in our instruments, was comparatively short, our readings are be- 
lieved to have been aecurate to within one-tenth of one degree Fahren- 
heit.2. Amid differences of such magnitude as were found in San 
Francisco Bay, errors of smaller extent are believed to be negligible. 
(6) Air temperatures were not taken by the scientific staff, since 
they were recorded at hourly intervals on the ‘‘ Albatross”’ 
to custom, by the ship’s quartermasters. The thermometers used for 
this purpose have not been carefully tested, but they are believed to 
be accurate to within a half degree F. A greater source of error is 
, according 
undoubtedly the heat from the ship’s furnaces, which must, at times, 
have a considerable effect upon the recorded air temperature. For 
this and other reasons little use has been made by us of these figures. 
2 Unfortunately the instruments all bore the Fahrenheit scale, but the centi- 
grade equivalents are given in the tables. 
