172 Dr. A. G. Levy. Estimation of Specific Gravity [Nov. 25, 



before a recent meeting of the Physiological Society, in which he 

 recorded estimations by Hammerschlag's method which exceeded by 

 0*012 the estimation of the specific gravity of the blood by the 

 picnometer method. Dr. Baumann also mentioned similar, but less 

 considerable, excessive readings noted by certain other observers. I 

 had nryself, in the course of a series of experiments upon the blood 

 of dogs,* occasion to remark upon the consistently high results 

 yielded. by Hammerschlag's method, the excess being, in my cases,, 

 from 0-007 to 0-008. 



"With the purpose of investigating the source of this error, I 

 prepared two mixtures, the one (A) of chloroform and benzol, and 

 the other (B) of glycerine and water, and, in each case, adjusted 

 the relative proportions until identical readings were obtained on 

 the scale of the same hydrometer. On immersing a drop of mixture 

 (B) in a vessel of mixture (A), it rapidly sank to the bottom, thus 

 indicating, in the absence of interfluid exchanges, an actually higher 

 specific gravity of (B). This indeed could be demonstrated by other 

 methods of finding the densities of the two liquids, i.e., by the weigh- 

 ing bottle or picnometer, or by Westphal's specific gravity balance. 

 The occurrence of a gross fault in the hydrometer method being 

 established, it remained to investigate its extent and origin. 



I proceeded by immersing four dissimilar hydrometers, all graduated 

 from 1*000 to 1*060 to the test of immersion in a mixture of chloro- 

 form and benzol, which had been prepared, with the aid of Westphal's 

 balance, of a specific gravity = 1*000, and the following table shows 

 the point at which the lowest level of the fluid meniscus intersected 

 each scale, i.e., the apparent specific gravity for each hydrometer. The 

 measurements of each instrument are included in the table for future 

 reference. 



The error, as set forth above, is explicable by a consideration of the 

 differences between the values of the surface tensions of water and of 

 chloroform and benzol. The exact experimental proof of this would, 

 I find, involve an extended investigation into a complicated subject. 

 This I have not attempted — it suffices to show how far the error,, 

 calculated from certain known conditions, agrees with my comparatively 

 rough observations. 



The action of surface tension of a fluitLupon a floating hydrometer 

 is evidenced in a downward pull upon the stem, so that a hydrometer 

 becomes immersed, not only until it has displaced a weight of water 

 equal to its own weight, but is still further immersed by the action of 

 surface tension upon its stem until the additional weight of water 

 displaced balances this surface tension pull. It is this point of ultimate 



* The Changes in the Blood of Dogs after Thyroidectomy," ' Journ. Path, and 

 Baot. October 1898 p. 317. 



