1906. No. 3. AMUNDSEN'S OCEANOGRAPHIC OBSERVATIONS IN 1901. 5 
The zero corrections determined by Mr. Jakob Schetelig at the 
laboratory in Christiania before and after the voyage were taken with a 
reading microscope, and were consequently fairly accurate. In March, 
1901, at the same time as Schetelig, Amundsen also determined the 
zero-correction of the same thermometers with one of the ordinary lenses, 
such as he used on the voyage, and it than appeared that it was very 
easy to get a somewhat too great correction, as he everal times obtained 
— 0'20 for No. 72620, whilst it should have been —o'16. This is evi- 
dently due to the very thick and coarse division-marks of the thermo- 
meter scale.! 
It is a striking fact that the temperatures taken during the voyage with 
these thermometers are nearly always lower than those taken simultaneously 
with the Richter Reversing Thermometer No. 113. The difference is strangely 
enough very uniform, about o‘10° C. (cf. the Tables of Observations). This 
proves that Amundsen has evidently read of his thermometers with great care, 
and by a very uniform method. It does not seem probable that these old 
Negretti and Zambra thermometers should have altered their error so 
“much as to indicate during the voyage on the average about o0'1" C. lower 
than they did both in March, before the voyage, and in September, after the 
voyage; and besides, this would not be in accord with the zero-corrections 
taken during the voyage. But it does not seem more probable that the Richter 
thermometer, made of Jena glass No. 16111, should suddenly have altered its 
error during the voyage, as much as to indicate 0'1" C. higher than it did both 
before and after. It seems more probable that some fairly uniform error 
has been made during the reading off of the Negretti and Zambra thermo- 
meters, perhaps some error of parallax e. 2. the axis of the reading lens 
may have beeen placed not perfectly perpendicular on the stem of the 
thermometers; and as the stem is very thick, only a slight error of paral- 
lax is sufficient to amount to o0'1" C. in the reading. 
In order to determine the temperature of the broken off mercury, at 
the moment the thermometers were read, they were always placed in a 
waterbath for some time before the reading was taken. The temperature 
of the water-bath was taken simultaneously, and recorded in the journal. 
The reading were subsequently corrected accordingly. 
Nansen Deep-Sea Thermometers. Richter Nos. 109 and 110. 
These thermometers were of the same type as now generally made 
by Richter for the Pettersson-Nansen Water-Bottle. They were made of 
! The determinations of the zero-corrections made on board the Gjøa on September 5, 
1901, in Tromsø, are therefore less trustworthy than those made at Christiania. 
