These animal experiments corroborate the findings in the blood of anilin and nitro-benzine workers and enable us to follow the sequence of events, which appear to be the same in cases of poisoning by either of these bodies, though they are much more marked in the case of the latter.

These bodies may gain access to the body either by (1) inhalation of their vapours; (2) absorption through the skin; (3) absorption from the alimentary canal after being swallowed.

Probably the first method occurs most commonly, since cases of poisoning are much more frequent in hot close weather than during the colder seasons of the year. No matter how they are absorbed these bodies appear to have the same action on the blood, which is as follows:

They first convert the oxy-haemoglobin of the red corpuscles into met-haemoglobin, this is followed by haemolysis, degeneration of cytoplasm and escape of haemoglobin into the plasma.

If a considerable amount of met-haemoglobin is present in the blood, the respiratory capacity of the red corpuscles is diminished, the patient becomes cyanosed and the colour of the blood is changed to a chocolate brown. This fact renders the estimation of haemoglobin by the ordinary colour standards impossible, as they are based on the colour of pure oxy-haemoglobin, which is quite different from that of met-haemoglobin.

In the less severe cases of poisoning in which met-haemoglobin is not present in sufficient amount to be detected spectroscopically, the changes which may be found in the blood are sufficiently characteristic to enable a diagnosis of anilin or nitro-benzine poisoning to be made.

These consist in:

(1) A decrease in the percentage of haemoglobin as estimated from the specific gravity of the blood of from 5% to 50%.

(2) A decrease in the number of red corpuscles if the amount of poison absorbed is considerable; if the dose is very small this decrease is not found, blood formation apparently keeping pace with blood destruction.

(3) Degeneration and imperfect development of the red corpuscles as shown by the occurrence of basophil granulations, polychromasia, poikilocytosis, and variations in size.

(4) The presence of nucleated red corpuscles in severe cases.

(5) A decrease in leucocytes rapidly followed by an increase; the increase being principally due to the number of lymphocytes.

The simultaneous occurrence of all or several of these signs in the blood enables a diagnosis of poisoning by anilin or nitro-benzine to be made in quite mild cases.

Cessation from work for a short time enables the blood to recover rapidly from the effects of poisoning.

The best treatment in any severe case of poisoning by these bodies appears to be inhalation of oxygen.

Haldane, Makgill, and Mavrogordato found that mice could be kept alive in oxygen at a pressure of two atmospheres even when 90% of their blood had been converted into met-haemoglobin by injection of sodium nitrite. They also showed that animals showing severe symptoms of di-nitro-benzol poisoning recovered rapidly when placed in oxygen, and that as other substances besides met-haemoglobin are produced by the decomposition of the haemoglobin, the absence of a met-haemoglobin band is no evidence that the symptoms are not due to oxygen asphyxia.

Permission to publish this paper has kindly been given by Mr H. L. Samuel, Chairman of the Home Office Committee on Industrial Diseases.