— ee 
1910. | Some Enumerative Studies on Malarial Fever. 161 
computing the proportion of parasites to each corpuscle in that specimen, 
then, by a simple calculation, the total percentage error will be e+e’ + ee’/100. 
(For example, if e = +5 percent. ande’ = +10 per cent., the total percentage 
error will be +15°'5 per cent.) 
Our method consists simply in making a measured quantity of blood into 
a thick film preparation and then counting all the parasites in it. The blood 
is measured by determining the diameter of a fine capillary tube under the 
microscope and then calculating and marking off the length of the tube 
required to hold the given quantity (as suggested by Dr. Wakelin Barratt). 
Thus a uniform tube of 0°180 mm. diameter and 4 cm. length will contain 
1018 cu.mm.—a convenient size. The blood is sucked up into the tube and 
then quickly discharged upon the glass slide in one or several measured 
droplets, first samples being rejected. The dehemoglobinised thick film is 
fixed before staining, in order to avoid loss, and the whole area is scrupulously 
searched with the aid of a sliding stage. 
Mr. M. Greenwood, Junior, informs us that if 7 is the number of parasites 
counted in the unit (say, 1 cu. mm.) of blood, and m is the number of such 
units in the patient’s whole body, then the total number of parasites in him, 
assuming uniform distribution, will be mn +067449m,/n. Thus the probable 
percentage error is 67:-449/,/n, the factor m cancelling out from the ratio. 
This is important, because it shows that the error depends, not on the 
magnitude of the measured sample taken, but upon the number of parasites 
actually counted in it. Hence, in order to obtain results within an assigned 
error, we take a large sample when the parasites are scarce and a small one 
when they are numerous; the number of parasites which we must find and 
count being given by the formula n = 4550/e?, where e is the permissible 
error. 
The greatest error occurs when the parasites are so scarce that it is 
difficult to find enough of them to count up to the required standard; or 
when they are so numerous that it is difficult to measure a volume of blood 
small enough to contain an easily countable number of them. In these 
preliminary researches we have seldom examined at one sitting more than 
1 cu. mm. of blood, or less than } cu. mm. 
When the parasites were very numerous, we have sometimes fallen back 
on estimating their number by comparison with leucocytes, but are now 
elaborating improvements and special instruments for such, and other, 
details. The method is also accurate and rapid for other blood parasites, and 
for leucocytes. Other sources of error are (1) inaccurate measurements of the 
quantity of blood used, and (2) difficulty of seeing or distinguishing badly 
Stained parasites. 
