28 



HYDROGRAFI. 



and emptied into a common tumbler, in which the precipitation was to take place. Next the solution 

 of silver-nitrate was added from a calibrated burette (supplied with a glass-cock), on which tenths of 

 cubic centimetres were marked, so that hundredths might be determined by means of a float As the 

 strength of the solution is kept about the same, the reading of samples by titration of sea-water of which 

 the quantities of halogens differ but little, will be nearly at the same place in the burette , so that eventual 

 caliber-errors in this latter will be of no consequence, in so far as the relative quantity is concerned. 

 Notwithstanding the fact that the vibrations occasioned by the engine and 

 the propeller of the vessel rendered the reading rather difficult, and that the tem- 

 perature in the laboratory varied considerably, it will be seen by the curves representing 

 the salinity, plate IV — XIX, that the titrations have given serviceable results. As the determination 

 of the amount of chlorine was effected by direct comparison with the standard-water, and provision 

 always was made to see that the water-samples had the same temperature as that of the room, and 

 likewise as that of the used standard-water and the solution of the nitrate of silver, a change in the 

 temperature of the laboratory could not have any influence of importance. At all titrations on both 

 voyages the same burette was used, which, in accordance with the Cardano system of suspension — 

 used at the determination of the specific gravity — was fitted on to a steel-bar attached to the table 

 and ceiling. 



The samples of water brought home by the Ingolf, were deposited in 

 bottles closed with parafined corks. This parafination proved, how- 

 ever, to be objectionable, as the parafine might give trouble at the analysis. 

 Thus small particles of parafine may find their way into the pipette during the 

 titration, and even if they be so small that they would expel no appreciable 

 volume of sea-water, it may happen that they attach to the glass, and retain a 

 considerable amount of water round them, when the pipette is emptied, so that 

 this latter had to be rinced with distilled water every time a quantity of water 

 was measured off for titration. At the determination of the strength of the so- 

 lution of the silver-nitrate, the same method of course has to be used, and in 

 this manner a great number of samples have been titrated after the ship arrived' 

 home, and principally samples that had not been titrated an board, but also — to 

 be on the safe side — such as had been titrated on the voyage. 



For the measuring of the sea-water, I had constructed a pipette, in which, 

 as seen by fig. 10, a cock is fitted at the uppermost tube. .Looking at the figure, 

 it can easily be understood that the adjustment to the mark in the uppermost 

 tube may be avoided, as the water, when it is measured off for titration, is sucked 

 up above the cock, which is shut and opened for the second perforation, when 

 the water is to run out. Instead of a tail-cock, a cock with two oblique perfora- 

 tions may be used. I found that this pipette (made by Franz Muller in Bonn) worked properly and 

 conveniently. For those revisions of the determinations which were to be executed at home, 

 I found it advisable to use a system of titration, by which the errors only would be small in compa- 

 rison with those occasioned by titration on board, I employed therefore Volhard's system, namely: 



w 



Fig. 10. 



