‘412 Dr. J. Mellanby. [June 3, 
the blood it will be ultimately carried into the neighbourhood of any tissue. 
But before it can attack a special tissue it must first pass through the capillary 
walls of the blood-vessels into the lymph bathing the cells of that tissue, — 
and such a passage demands a specialised mechanism of chemiotaxis. 
Further, from the union of toxin and antitoxin im vitro, it is probable that 
the toxin would unite with the antitoxin as soon as it got into the blood 
stream, and the hypothesis of the formation of antitoxin in a distant tissue 
would demand that this toxin-antitoxin compound should be broken up by 
the cell before it could stimulate the production of more side chains—in fact, 
that the formation of antitoxin by a cell should be of no use in protecting it 
from subsequent inoculations of toxin. 
From a general consideration of diphtheria toxin immunisation it is 
probable that the leucocytes are the immediate agents concerned in the pro- 
duction of antitoxin. After a subcutaneous injection of diphtheria toxin intoa 
horse there is usually produced at the site of inoculation a swelling of varying 
size. This swelling is of an cedematous character, and is crowded with finely 
granular oxyphil cells. The size of the swelling gives, to some extent, an 
indication of the degree of antitoxic reaction. The production of antitoxin 
by the leucocytes at the seat of inoculation would bring the mechanism 
involved into line with the facts observed by Metchnikow with pathological 
bacilli and the more recent work on opsonins. The antitoxin need not 
necessarily be produced at the seat of inoculation. After a leucocyte had 
ingested a toxin molecule the secretion of antitoxin by it would take place 
into any fluid in which it was present—the blood or lymph. 
This hypothesis affords a ready explanation of poisoning by excess of toxin. 
If the leucocytes are able to take up the toxin, none of it is carried to distant 
tissues, such as the heart and nervous system, and no pathological effects are 
observed. But if the toxin is injected in too great a quantity to be taken up 
by the leucocytes, then the excess is carried by the blood to the tissues, and 
general toxemia results. Again, Ehrlich’s hypothesis offers no explanation 
why subcutaneous injection of toxin yields much better antitoxic results 
than injection into the blood stream direct. The theory that antitoxin is 
secreted by leucocytes which have ingested toxin molecules affords a ready 
explanation of these results. 
SUMMARY. 
All the properties of diphtheria antitoxic serum indicate that diphtheria 
antitoxin is a protein which possesses characters identical with albumin (a) 
of serum, 2.¢., the protein forming 85 per cent. of the total quantity present. 
The alcohol precipitation curves of antitoxic sera of varying strength 
