METHODS OF OBSERVING BACTERIA. 153 



fact, seldom longer than six to eight weeks, sometimes not 

 more than two or three ; therefore it is best to prepare it 

 in very small quantity by pouring about 1 c.cm. of pure 

 anilin into a test-tube, rilling the tube about one-half 

 with distilled water, shaking the mixture well, then fil- 

 tering as much as is desired into a small dish. To this 

 the saturated alcoholic solution of the dye is added until 

 the surface becomes distinctly metallic in appearance. 



Friedlander recommends that the section remain from 

 fifteen to thirty minutes in warm stain, and in many cases 

 the prolonged process gives better results. 



From the stain the section is given a rather hasty wash- 

 ing in water, and then immersed from two to three min- 

 utes in Grain's solution (a dilute Lugol's solution) : 



Iodin crystals, 1 ; 



Potassium iodid, 2 ; 



Water, 300. 



While the specimen is in the Gram's solution it 

 appears to turn a dark blackish-brown color. When 

 removed from the solution it is carefully washed in 95 

 per cent, alcohol until no more color is given off and 

 the tissue assumes a grayish color. If it is simply 

 desired to find the bacteria, the section is dehydrated 

 in absolute alcohol for a moment, cleared up in xylol, 

 and mounted in Canada balsam. If it is necessary to 

 study the relation between the bacteria and the tissue- 

 elements, a nuclear stain, such as alum carmin or Bis- 

 marck brown, may be subsequently used. Should a 

 nuclear stain requiring acid for its differentiation be 

 desirable, the process of staining must precede the Gram 

 method altogether, so that the acid shall not act upon 

 the stained bacteria. 



The success of Gram's method rests upon the fact that 

 the combination of mycoprotein, anilin dye, and the iodids 

 forms a compound insoluble in alcohol. 



The process described may be summed up as follows : 



