16 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS tion spectrum of medicinal liquid paraffin, the B.P. Acid Test appears in a most unsatisfactory light. The ultra-violet absorption spectrum is a measure of the quality and stability of medicinal liquid paraffin, and a combination of ultra-violet and infra-red data is a powerful tool to unravel the constitution of the "impuri- ties" left in the refined product. If in the Acid Test the reaction of the sulphuric acid were only with these minor "impurities," the spectrophotometric and the Acid Test data would directly correlate. No such agreement is found, and any such relation is impossible because of the dehydrogenating action on the naphthene molecules of the acid used [or the test. CONCLUSION The ultra-violet absorption spectrum of medicinal liquid paraffin is a quantitative measure of the unsaturated hydrocarbons, esters, and sulphon- ates of which very small amounts are present. These "impurities" fall into two groups, "A" and "B." The former consists of alkylated symmetrical octahydrophenanthrenes, and the latter of alkylated asymmetrical octa- hydrophenanthrenes, substituted tetrahydrophenanthrene, esters, and sulphonates. There is a significant difference between the evanescent odour of a liquid paraffin which is kept in a warm place and the persistent pungent odour of an unstable material that is noticeable even though the material may be stored in a cool place. Responsible for the development of this persistent odour of unstable materials is their comparatively high content of "Group B" compounds. Owing to the characteristic differences of the spectra of the "Group A and B" compounds, a stability criterion is based on the shape of the ultra- violet absorption spectrum of medicinal liquid paraffin. This relates the absorption intensity at the peak of the band at 2710 2• to the ratio of the absorption intensities at the maximum and minimum and to the wavelength position of the minimum. The latter consideration is particularly important in those cases of so highly purified materials that the spectra show no band structure because the level of the absorption intensity is very low. Comparison of spectrophotometric data with B.P. acid numbers shows that the latter bea• no direct relation to the quality of the material. This is corroborated by the results of daylight exposure tests. In fact, an extremely pure liquid paraffin may be failed by the B.P. Acid Test while on the other hand many unstable materials will easily pass this test, whose results depend on many operational variables and on the hazard that the acid used for the test will also dehydrogenate naphthene molecules and sulphonate the thus created reaction products.

SPECTROPHOTOMETRY AND MEDICINAL LIQUID PARAFFIN 17 In concluding this account I wish to thank those of my present and past colleagues who have taken active and devoted part in this work. A great number of the experimental data are taken from the above-mentioned M.Sc. thesis of Miss P.M. Martin. Mr. E. Kendrick, M.Sc., and Mr. R. Mayoh, B.Sc., Mr. F. Smith, M.Sc., and Mr. W. F. Maddams, M.Sc., have contributed to the experimental evidence. DISCUSSION Dr. Mitchell asked to be informed of the concentration of liquid paraffin, and the solvent used, in the solution used for spectrophotometric examina- tion. Results obtained in his laboratories showed many samples, apparently of good quality and stability, to give ratios ranging from 20 upwards. While such samples literally complied with it, the standard (not less than 2) did not appear to be entirely satisfactory or should one suspect such samples-- or the results ? Most analysts disliked the sulphuric acid test as troublesome, and would welcome a spectrophotometric test instead. However, we were now told that the two tests measured different things. Analysts certainly would not want both tests. Further, they would want assurance that reliable and consistent results would be obtained, so that different analysts could expect to get reasonable agreement between their figures. Dr. Mitchell suggested that results recently obtained in his own, and other, laboratories had failed to give this assurance, and he did not consider that the test, as now defined, was satisfactory. In particular, he could not support its adoption as an official test. Author: The solvent used for the preparation of solutions of liquid paraffin must be transparent in the wavelength region between 2400 • and 2800 •. Most commercially obtainable solvents such as hexane, cyclo- hexane, heptane, and iso-octane must be purified before they are used for spectrophotometric work. The concentration of the liquid paraffin solution must be adjusted to suit the absorption level of any particular sample of liquid paraffin. The greatest accuracy is obtained when, with a photographic plate instrument both the maximum at 2710 2• and the minimum at 2490 2• are recorded at an optical density of 1.00, and with a photo-electric instrument at an optical density of about 0.4. Whenever the spectrum of a solution is measured, the cell in the reference beam should be filled with the corresponding solvent. When the absorption level of a liquid paraffin sample is sufficiently low that the material can be measured "neat," either in a 1 cm. cell or in an adjustable cell of smaller path length, a similar cell should be placed in the reference beam and a check should be made by filling this cell with freshly distilled water to ensure that reflections from the inside of the cell windows are not significant.