COUGH IRRITATION BY DEODORANT SPRAYS 513 Each subject was free to choose how to make the test. Two different methods were used: some subjects (19•o) tended to direct the aerosol spray towards the face from arm's length. The remaining sub- jects sprayed horizontally in front of the body and then remained immersed within the spray cloud. No overall difference in findings appeared between the two groups although, with the second method, the sprayed amounts of fluid were generally greater. Respiratory tract irritation was expressed as an integer, according to the following table: 0 no cough, 1 slight cough, 2 moderate cough, 3 strong cough, 4 very strong cough. After averaging, multiplication by 2.5 and rounding to 0.5, a value was obtained for each spray, giving a measure of the cough-stimulating poten- tial: the cough factor (CF). Cough-stimulating activity was assessed twice for each deodorant spray investigated. Reproducibility appeared to be good. The whole investigation lasted about 2 yr, with a total of 212 irritation tests. EXPERIMENTAL Preliminary studies Since at the beginning of this investigation nothing was known about the possible cough-stimulating potential of the various ingredients, first the cough-stimulating activity of a deodorant spray with a composition in common use was determined

514 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Composition 7o Isopropyl myristate 0.3 Perfume 0.2 Bactericide 0.2 Ethanol 24.3 F11 25.0 F12 50.0 100.0 The fragrance used was a green flowery type. The bactericide was chosen arbitrarily from the six types listed above and was used as the standard bactericide throughout the investigation. The aerosol was equipped with a valve having a stem orifice of 0.45 mm and an actuator orifice of 0.50 mm the spray angle was about 16 ø. All preliminary tests were carried out with this deorodant spray initially the investigation was designed to show whether the consumer test did actually ensure good quantitative reproducibility of the cough-stimulating irritation. This preliminary study indicated that the test method was suitable for the purpose and the spray mentioned above gave a cough factor (CF) of 6.0. This standard spray was used for reference throughout the main investi- gation, a run with it being included in each large test series. The standard spray was tested, overall on 22 occasions by the panel, leading to cough factors as follows: 4 times: CF = 5.5 12 times: CF = 6.0 6 times: CF = 6.5 Between the two extreme values, 5.5 and 6.5, Wilcoxon's Q-test only once showed a significant difference. A number of screening tests were performed, in order to identify a possible main cause for cough stimulation. Several formulae were tested, one ingredient being omitted each time. Table I gives a summary of the formulae used, with corresponding CF values. Clearly the formulation without bactericide was less cough-stimulating than the others. Formulae C, E and G (respectively without perfume, with- out ethanol and without F12) also gave significantly lower CF values than the standard, but not as low as the bactericide-free formula.