332 



Dr. S. Judd Lewis. 



a view to building up the resultant curve from its parts. This problem 

 is one of great importance, for only in such manner can one hope to 

 determine what constituents of the serum are affected in any given 

 disease. It has already been solved in general terms and its precise study 

 is in progress. It has been found that the chief central band is due almost 

 entirely to proteins. The constancy of the form of section £ points also to the 

 same constituents. This has been ascertained in the following manner : — 

 A portion of serum is weighed and mixed with alcohol, whereby the proteins 

 are precipitated, leaving the non-proteins in solution. The operations are 

 conducted so that after dilution 100 volumes of solution contain the non- 

 proteins from 1 volume of serum. A portion of the solution is examined 

 in a 2 cm. cell ; this layer corresponds with the non-proteins in a film 

 of serum 0*2 mm. thick, so that the absorption curve requires but little 

 correction to make it comparable with that of the film of serum usually 

 employed. The effect of the solvent is eliminated in the manner already 

 indicated. The distribution and proportion of that part of the absorption 

 by serum due to non-proteins may then be readily appreciated. It is 

 found to consist essentially of a general absorption which is slight from 

 the visible to a wave-length of about 2100, and then to increase rapidly. 



The proteins are washed with 90-per-cent. alcohol, separated in the 

 centrifuge and dissolved in such a quantity of water that 100 volumes 

 of the solution contain the proteins from one volume of serum. The manner 

 of examination is similar to that for the solution of the non-proteins. The 

 form of the resulting curve approximates that for the original serum, 

 showing that the proteins account for most of the absorption observed with 

 serum. 



Horse serum (three specimens) has also been the subject of inquiry. Its 

 absorption curve is very similar to that of human serum, but (a) the 

 depression is at wave-length 2510 instead of 2540, (b) the amplitude from 

 this point to the head of the curve at 2800 is rather greater, and (c) the 

 first step-like prominence in section 7 is somewhat lower and decidedly 

 more pronounced, so that the curve for the protein band is in general of 

 rather larger dimensions than is the case with human serum. 



Egg white (one specimen) was dealt with in similar manner for com- 

 parison. The curve of the protein band is more symmetrical in form. The 

 sections e and £ meet at 2540 as with human serum. 



Pathological. — A considerable number of pathological specimens have been 

 examined, with some significant results. As would be expected, abnormality 

 is confined to certain diseases ; and again, as must be anticipated, the 

 magnitude of the disturbance is usually small. One cannot look for severe 



