1902.] 



of the Blood by Hammer 'schlag's Method. 



175 



merit that each instrument possessed an intrinsic error which tended 

 to minimise the error of reading. 



A hydrometer and its containing vessel were carefully cleansed with 

 benzol and protected from contamination with greasy matter, and a 

 sample of water taken, which was the cleanest readily obtainable, i.e., 

 tap-water which had been allowed to run through the pipe for some 

 fifteen minutes. On immersing the hydrometer in this water, the mark 

 1 -000 rested a slight distance below the surface. The water in which 

 the hydrometer was standardised must have been contaminated, and 

 hence possessed a considerably lower surface tension than that of the 

 comparatively clean water in which my experiment was performed. 



All the four hydrometers of my tables I found to possess this 

 intrinsic error, which I estimated somewhat roughly, and have 

 expressed in scale units. When these innate errors are added to the 

 errors of Table I, the totals more closely approximate to the calculated 

 errors. (See Table TIT.) 



Table III. 



Hydrometer. 



Error due to 

 difference between 

 surface tensions of 

 impure water and a 

 chloroform-benzol 

 mixture. 



Error due to 

 difference of 



surface tensions 

 of impure 



and clean water. 



Total error. 



Calculated 

 error. 



1 



0-002 



0-0014 



0034 



-0035 



2 



-003 



0-002 



0-005 



-0056 



3 



-0095 



0-002 



-0115 



-0123 



4 



o-oio 



0-003 



0-013 



0-0146 



There is thus sufficient agreement between the values of the observed 

 and calculated errors to demonstrate that the disturbing influence of 

 surface-tension is sufficient to cause the whole of the error in the hydro- 

 meter reading and to account for the inaccuracy of Hammerschlag's 

 method. Taking into consideration the varying value of the surface 

 tension of water, and the fact that no accurate determinations were 

 made by me of each individual specimen of water or of chloroform and 

 benzol, very exact calculations are precluded. Had this been done 

 doubtless a more exact agreement between observations and calcula- 

 tions would have resulted. Furthermore, observations of this nature 

 .are replete with difficulties which can not he touched on here, but 

 which may be gathered from a paper upon an elaborate investigation 

 into a similar subject by Fridtjof Nansen.* 



* " Scientific Kesults of the Norwegian North Polar Expedition," vol. 3, 

 Part 10 



