341 



does not caiTjf down air, for instance Sigsbee's, must be used 

 in combination with a deep-sea-thermometer. As soon as the 

 water-sample is on deck it should be transferred to a glass- 

 bottle. iNo perceptible loss of carbonic acid need be feared, 

 when this operation is performed by a rubber-tube of suitable 

 bore conducted to the bottom of the bottle. 



On the preservation of water-samples for a subsequent 

 determination of the tension of carbonic acid 1 have not much 

 to say, the more so, as perfectly reliable methods have not, so 

 far, been discovered. 



No perceptible increase of the alkalinity takes place at 

 ordinary temperatures in resistance-bottles; it appears, on the 

 contrary, from some of my determinations that the alkalinity 

 may decrease, at least when the water is kept for some months. 



Carbonic acid will sometimes be produced in the samples 

 (presumably by bacteria) and consequently the tension Aviil rise. 

 In 18 samples from the Davis- Strait and the North- Atlantic, 

 taken in new and clean bottles, this happened with 4, while 

 the others remained unaltered. Most of these indeed became 

 infected during the tension-determinations, and the tension in 

 some of them rose considerably. 



If 1 gr. of sublimate is added to each sample of 1 1. this 

 formation of carbonic acid is completely avoided, but the pre- 

 sence of Hg CI2 causes the tension of normal sea-water to 

 rise 0.35 (0.0035 "/0). 



I have not tried any other antiseptic, but 1 consider it 

 extremely probable that, among the existing multitude of these 

 substances, one or more may be found, capable of preventing 

 the growth of carbonic-acid-producing organisms but free from 

 any chemical action on the mineral constituents of seawater. 



When the tension has to be determined the bottle is pro- 

 vided with a rubber-stopper pierced by two glass-tubes. One 



