1. The fall in the refractive index of mineral oils recovered after injection into the peritoneal cavity of mice is usually accompanied by a fall in the density proportional to the fall in the index.

2. With straight distillates from the crude oil the ratio of index fall to density fall is on an average about 0·85 to 1. Wide deviations from these figures were rarely found except when dealing with oils of low toxicity. The ratio tends to be above the mean when the actual index fall is low and consequently below the mean when the index fall is high.

3. The refractivity of the original oil, the index fall of the recovered oil and the density fall of the recovered oil give information as to the carcinogenic potency for mice of the original oil of about equal value, but the two lastmentioned tests are far superior to the first for assessing the probable dermatitic potency of the oil.

4. Other things being equal, the index and density fall are closely dependent upon the index, density and viscosity of the original oil.

5. The index and density fall gradually become larger as the experiment progresses, but within each 24-hour interval the fall gradually becomes less. The ratio of index fall to density fall remains fairly constant for each individual oil during the course of the experiment.

6. Titration of mineral oils in vitro with sulphuric acid and alcohol provides residual oils of lower index and density than the original oil, but the ratio of the fall in index to the fall in density is low, especially when the treatments are not severe.

7. Oxidation of mineral oils tends to raise the density without affecting the index to an appreciable extent. On the other hand reduction tends to lower the index without affecting the density to anything like the same extent.

8. In general terms the animal reacts upon the chemically treated oils in an additive manner, so that if by the chemical treatment we have the ratio of index fall to density fall low, the oil recovered from the animal will have the ratio high, and vice versa. The end product of the severe treatment of a mineral oil with acid and clay is a white oil which for practical purposes may be considered to be fully saturated. The end product of an oil recovered from the animal is probably similar in nature, although the process involved in arriving at this end product is presumably different chemically and physically.

9. Synthetic oils recovered from the animal after inj ection showed a marked rise in density but no corresponding rise in index.

10. The animal reaction to an injected oil was not materially accelerated by emulsification of the oil.

11. Some chemically pure non-carcinogenic compounds were very rapidly changed or eliminated from the peritoneal cavity.

12. All alcohol extracts of mineral oils tested showed an increase in index, density and refractivity, but not always in viscosity compared with the original oil, the residues, of course, having characteristics in an opposite direction. The latter were in every instant less toxic for the animal than the original oil, whether applied to the skin or injected.

13. A mineral oil recovered from animals 7 days after injection, when reinjected into a second series of animals, was altered as regards its physical characters in a normal manner. Thus whether allowed to remain in the peritoneal cavity of an animal of the first series or recovered and reinjected into an animal of the second series, the ultimate index and density fall in a given time were approximately similar.

14. The free oil in the peritoneal cavity becomes, from day to day, less. This is at least partly due to local trapping by the proliferating endothelium. The fall in index of the trapped oil is approximately similar to that of the free oil.

15. The alteration of the physical characteristics of mineral oils by the animal is probably brought about by a chemical reaction followed by removal of the changed product.