HYDROGRAPHY. 



93 



that we can easily, and quickly, take out a lot of samples for gas-analysis as soon as the water has 

 come up with the water bottle, and thus have made as many analyses as desired. 



Petterssori 's method has afterwards been used at nearly all determinations of the gases contained 

 in the water, but on account of Doctor Rerdam being in doubt whether these determinations could 

 be relied xipon, I did not employ the said method in 1896. As the irregularities in the distribution of 

 the amount of oxygen, on which Dr. Rordam founded his doubt, may be attributed to other causes than 

 the consumption of the oxygen dtiring the storing of the sample, I resolved to deposit water for some 

 time in glass-balloons void of air and remaining from the voyage in 1895, while another sample of 

 the same water was analysed at once. To perform the analysis of the samples brought home in glass- 

 balloons, I was obliged to have the boiling-flask of the gas-analysis-apparatus made larger, as the 

 capacity of a balloon was a little more than 300 cubic centimetres. The balloon was connected with 

 the boiling-flask almost in the same manner as in Professor Pettersson's apparatus. As soon as the air 

 was expelled from the boiling-flask, the point projecting into the tube that connects the balloon with 

 the flask was broken off. The other point was led down into mercury, and broken off under the 

 surface of the mercury. The mercury was then sucked up into the balloon, driving the water before 

 it into the boiling-flask, in which boiling out and analysis were effected in the usual manner. The 

 volume of the water was determined by weighing of the balloon before and after it was emptied. 



