1911.] Serum Reactions in Pulmonary Tuberculosis. 375 



antigens. With antigens such as alcoholic extracts of normal and pathologic 

 organs, specificity has been maintained. Extracts of other micro-organisms 

 have not as yet been tried. From the relative specificity of a serum, obtained 

 after tubercle immunisation, to different acid-fast cultures (as shown by 

 Gengou, Calmette, and others by different serological reactions), it is likely 

 that comparative results would be obtained with the inhibitive and fixation 

 tests between the serum of the tuberculous and other acid-fast bacteria. 



The routine* for some time, as far as this part of the serological work is- 

 concerned, has been to determine the reactions of the serum in two strengths 

 (0"3 c.c. and 01 c.c.) to the four dilutions of antigen, and subsequently to 

 attempt to show the cause of the various irregularities which may have 

 appeared. For these purposes it is essential that the same samples'of sera 

 be used, and that the succeeding reactions be performed immediately, as- 

 marked variations have been shown sometimes to take place rapidly in stored 

 sera. To meet these requirements blood is usually drawn from 10 to 20 

 patients. As a great number of patients have been followed for a consider- 

 able period, it is frequently possible to include a number of similarly 

 reacting sera. From these it is often possible to determine conditions 

 connected with the irregularities. The results of these observations and 

 inferences drawn from them can be considered most conveniently under their 

 various headings. 



(1) General Preparation. — Despite the utmost care and similarity of 

 procedure, I have found that every now and then any single tube may vary 

 remarkably, even to 100 per cent, error, as can readily be proved by a 

 repetition of the test. Further, the results of an entire series have been 

 thought doubtful because of the unsatisfactory condition of some of the 

 controls which, with the usual technique, are not generally made. Because of 

 these inherent characteristics of the test it seems advisable to outline the 

 general technique which has been adopted. 



Glassware. — These articles are finally cleaned in saline solution, then 

 rinsed in distilled water and allowed to drain and dry. Test tubes, as similar 

 in size as obtainable, are not dried on racks, but packed in wire baskets and 

 inverted on clean filter paper. All articles are sterilised for use. 



(2) System. — Guinea-pig complement and anti-sheep-corpuscle rabbit 

 hemolysin have been used. It has been the custom to re-standardise 



* For economy of materials, 0"5 c.c. units have been employed. With this unit it has 

 been convenient to designate in Arabic numerals the number of 0'05 c.c. quantities 

 employed, which are in each instance completed to - 5 c.c. with physiological saline 

 solution. By this means the insertion of a decimal point converts the protocol into 

 comparison with 1 c.c. units. This is the method that has been adopted in this report, so 

 that actually but one-half the quantities reported have been used. 



