OCEANÜGRAPHICAL RESULTS, MICHAEL SARS, 1900. 137 



good quality, so that they easily change their zero pointy and 

 further their scales are usually so roughly made that accurate 

 readings cannot be taken. In order to avoid at least some of 

 these difficulties I had several new thermometers made by Ne- 

 GRETTi and Zambra, with a much longer scale than usual. The 

 degrees Centigrade of the scale were 1 cm. long and were divi- 

 ded into 0*1 °C. In order to reduce the friction between the 

 mercury and the sides of the tubes, I had ordered the tubes to 

 be cylindrical instead of eliptical as they generally are ; but 

 this was not done, and consequently caused trouble, as some of 

 the thermometers sometimes refused to register, the friction in 

 the tube probably being too great. I had also ordered the ther- 

 mometers to be made of Jena Glass No. 16 HI. Only one ther- 

 mometer, however, could be made of this glass, and this was 

 by far the best instrument. This thermometer worked on the 

 whole very satisfactorily, and as a rule gave the deep-sea tempe- 

 ratures with an accuracy of about + 0'01° C. when its corrections 

 were daily checked by the insulated water-bottle. But even this 

 thermometer changed its zero-point from one day to another. I 

 do not know whether this was due to contaminations of the 

 mercury and the inside of the tube, or to changes in the glass. 

 The following table will give some idea of these changes : 



Zero- Corrections of Negretti and Zambra Reversing 

 Thermometer No. 96018. 



1900 



Correction 



1900 



Correction 



July 23 



„ 24 

 „ 25 



— 0-05°C. 

 -0-06 „ 



- 0-08 „ 



July 28 

 Aug. 3 



- 0-03° C. 

 -0-16 „ 



