60 FRIDTJOF NANSEN. M.-N. KI. 
ratures and salinities given, heavier water should frequently have been 
placed on top of much lighter strata; but it is often impossible to decide 
whether the errors are chiefly due to errors in the temperature or in the 
salinity!. The values of the latter have been determined from the 
amount of chlorine per litre, and have to be reduced by about 0:07 °/o9? 
to be comparable to the salinities found from Knudsen’s Tables?. The 
highest salinity obtained at Arrhenius’s Stations was 35'22 °/,, (originally 
35°29 09), at 400 metres (Stat. IV), but this is evidently eronous, if the 
temperature of 2'46° C. given for the same depth be correct. For the 
density of that water-stratum would then have been 28:14, and much 
heavier than all underlying waters. If the temperature be correct the 
salinity must have been less than 35°17 /o9, but if the salinity be cor- 
rect, which is improbable, the temperature must have been above 3°C. 
The Nathorst Expedition to East Greenland took, in June and 
July, 1899, a few Stations (NW VI—N IX, PI. V), in the region be- 
tween Amundsen’s Stations (20 and 19), and the Greenland coast. 
The results are described by Mr. Filip Akerblom4, the oceano- 
grapher of the expedition. The temperatures were taken usually by a 
Pettersson Insulated Water-Bottle of the old form. The thermometer 
was inserted after the bottle came on deck and cannot therefore be 
expected to give perfectly accurate temperatures; but as the depths were 
not great these temperatures may be expected to be very satisfactory. 
For depths greater than 500 metres reversing thermometers were used, 
the accuracy of which were inside the limits of o‘1° C., according to 
repeated experiments. The salinities given by Åkerblom appear, however, 
to be less trustworthy. Åkerblom gives in his tables the amount of 
chlorine per litre at 15° C., but he says that salinities are computed 
The latter errors may have been due to evaporation of water through the cork-stoppers 
of the sample bottles on the way home. 
> Mr. B. Helland-Hansen informs the writer that values of salinity computed from the 
amount of Chlorine by means of the tables formerly used in O. Pettersson’s labora- 
tory, are about 0'08 or o’09 °/,, higher than those obtained by Knudsen’s Tables. If 
however the salinity be computed from the permillage of Chlorine (per 1000 grams 
seawater) by means of the factor 1'809 the value obtained will be nearly 0.05 °/ oo 
higher than that obtained by Knudsen’s Tables. A reduction of 0'07 °/,, as a mean 
between the two, has here been employed. 
Pettersson himself does not seem to be aware of this fact as he has recently com- 
w 
pared the old values of salinities from Arrhenius’s Stations with those found in the 
North Polar Basin. He reduced the latter salinities but does not reduce the former 
(see Geograph. Journal., London, vol. XXIV, 1904, p. 316). 
* Filip Akerblom, Recherches Océanographiques. Expedition de M. A. G. Nathorst 
in 1899. Upsala Universitets Ärsskrift 1903. Matematik och Naturvetenskap, II, No. 1, 
Upsala, 1904. 
