280 CLINICAL BACTERIOLOGY AND H^MATOLOGY 



infection is certain, whatever be the clinical condition; a cessa- 

 tion of the fall is good also, but to a less extent. (Beware, 

 however, of mistaking a concentration of the blood from 

 diarrhoea or profuse vomiting- for a true rise.) As an example 

 I may quote the case of a patient under Dr. Hayes in King's 

 College Hospital, in whom about a fortnight after parturition 

 there were 1,306,000 reds, 26 per cent, haemoglobin, 13,400 

 leucocytes, of which 87 per cent, were polynuclears. Here 

 2,700,000 reds had been lost in two weeks, indicating a very 

 severe sepsis. Her condition appeared desperate, yet in a 

 week the reds had increased to 2,750,000, in a fortnight more 

 to 3,760,000, when for the first time she showed clinical im- 

 provement. In nine days more they had reached 4,000,000, 

 and the patient was out of danger. Here a good prognosis 

 was given solely on the blood-counts. 



In these cases a great diminution or total disappearance of 

 the eosinophiles is a bad sign, their reappearance a good one. 



2. Suppuration. Where the sepsis is localized and not so 

 severe, so that a limited focus of suppuration occurs, there is 

 a leucocytosis which, in round figures, exceeds 20,000, and the 

 increase is mainly due to a rise in the polynuclears. Various 

 writers give different figures as that above which pus is indi- 

 cated : some take 15,000, which I find to be reached fairly 

 often when there is no pus ; others 25,000, which is frequently 

 not reached when suppuration has occurred. The number 

 20,000 has been taken as the result of a considerable amount 

 of experience of all forms of suppuration, and will prove a 

 correct indication in at least 90 per cent, of all cases. It is 

 especially useful in appendicitis, in which it is an almost cer- 

 tain guide. The glycogen reaction is usually present, and is 

 a valuable confirmatory test. 



The following considerations must be remembered : 



(a) The increase of leucocytes to the figures mentioned 

 above only occurs when the pus is pent up, not when it occurs 

 on a free surface and can escape. I have twice seen a rapid 

 fall of the leucocytes due to rupture of an appendicitic abscess 

 into the intestine. 



(b) Where the spread of the suppuration ceases (due to the 

 death or latency of the organisms it contains) the leucocytosis 

 gradually subsides, and an old, thick-walled collection of pus 

 in the tissues may give a normal count. This is especially 



