THE SCIENCE AND ART OF PROTEAL THERAPY 213 



proper enrollment to the differently distributed types of leu- 

 cocytes. 



Yet I venture to surmise that nine observers out of ten dis- 

 regard this truism, and ultimately stop their count when they 

 have registered an arbitrary number of corpuscles, say two hun- 

 dred or three hundred, regardless of the fact that the field with 

 which they abandon the count may lie at the middle of the smear. 



If the smear is a short one, so that a good many tours have 

 been required to make up the requisite count of two hundred or 

 three hundred, the miscalculation will not be very notable, but 

 if a long smear is under observation, or if there is a marked 

 leucocytosis, so that only two or three sweeps of the smear are 

 necessary to complete the count, the inaccuracy may be highly 

 significant. Smears are not infrequent, for example, in which 

 practically all small lymphocytes are in the first half and prac- 

 tically all the large monocytes in the second half. If, now, the 

 sweep of the field begins at the beginning of the smear, and two 

 and a half tours of the field are required, the region of the 

 small lymphocytes will have been traversed three times, whereas 

 that of the large monocytes will have been traversed only twice, 

 and the record will show a seeming preponderance of small 

 lymphocytes in excess of the actuality somewhat in the ratio 

 of three to two. But if, contrariwise, three and a half tours 

 of the field were required, then we shall have traversed the region 

 of the large monocytes four times and the regions of the small 

 lymphocytes only three times, and the record will show a ficti- 

 tious preponderance of large monocytes. 



In a word, where the method is employed and it is almost 

 universal of counting a definite number of leucocytes, it can 

 occur only in the exceptional instances where the count happens 

 to be finished at the extremity of the field, that the result is as 

 accurate as it might be and should be. The degree of error will 

 vary with the length of the smear (a short smear being pref- 

 erable) and with the aggregate number counted; the larger the 

 number, naturally, the closer approximation to accuracy. But 

 the method is inherently faulty and, in my opinion, should be 

 abandoned. 



The only correct way to make the differential count, as I see 

 it, is to work clear across the field from right to left and from 

 left to right, several times, preferable at considerable longi- 

 tudinal intervals, so that in the end practically the entire area 

 of the smear has been sectioned, and to stop counting, not 

 when an arbitrary number has been reached, but at the end of 

 a complete tour across the smear, whether that tour be the first 

 or the twentieth. A single or double sweep across the field, with 

 an aggregate count of, let us say 137 leucocytes, would be, in 



