TUBERCLE BACILLI IN SPUTUM. 259 



to run over the surface, and then, after dashing off the excess of 

 alcohol, to ignite that remaining on the film in the flame and allow to 

 burn out. 



A mark with a grease pencil, about one-half inch from the end, 

 gives a convenient surface to hold with the forceps and also prevents 

 the stain subsequently used from running over the entire surface. 

 Sputum should as a routine measure be stained by the Ziehl-Neelson 

 method and by Gram's method. 



In examining for tubercle bacilli it may be necessary to employ 

 some method for concentrating the bacterial content of the sputum 

 prior to making the smear. A very satisfactory method is that of 

 Muhlhauser-Czaplewski. Shake up the sputum with four to eight 

 times its volume of 1/4% solution of sodium hydrate in a stoppered 

 bottle. When the mixture has become a smooth, mucilaginous -looking 

 fluid, add a few drops of phenolphthalein solution and bring the pink 

 mixture to a boil. 



Then add drop by drop a 2% solution of acetic acid, stirring con- 

 stantly, until the pink color is just discharged. If the least excess of 

 acid is added over that just sufficient to cause the pink color to disap- 

 pear, mucin w r ill be precipitated. Now pour this mixture into a centri- 

 fuge tube and smear the sediment on a slide and stain for tubercle 

 bacilli. 



Sputum smears stained by some Romanowsky method or by the 

 haematoxylin eosin stain are best adapted for the study of various 

 cells, and in particular of the eosinophile cells so characteristic of 

 bronchial asthma. In sputum from cancer of the lungs the large 

 vacuolated cells may be found. 



When examining the sputum of the bronchopneumonia of in- 

 fluenza the formol fuchsin gives the best results. The influenza 

 bacilli are found in little masses, frequently grouped about small 

 collections of M. tetragenus. The cocci stain a rich purplish-red, 

 while the small influenza bacilli take on a light pink color. 



Red cells show up well in specimens stained by the Romanowsky 

 method; if rouleaux formation is marked, it may indicate pulmonary 

 infarction. 



In culturing sputum a mucopurulent mass should be washed in 



