COUGH IRRITATION BY DEODORANT SPRAYS 521 In our evaluations, a dispersion of values was noted in every test. In addition, the frequency distribution was not normal, thereby making it impossible to compare two series of measurements by t-test. In an evalua- tion like this, the most precise way is to use the (T) sign test in order to recognize significant differences. In interpreting the data obtained with the consumer test, we preferred, however, to use the Q-test of Wilcoxon. Here each segment of the data in a series is compared with the data obtained from another series. This yields a difference score, called Q-value, allowing the one-side overrun proba- bility to be determined by means of tables. Using the Q-test instead of the T-test has the advantage of allowing comparison between series of different length. Experience has shown that the average rating, as determined with the consumer test method, has a maximum value of 3 to 3.5. Not only when assessing cough stimulation due to deodorants, but in other tests, too, there appears to be some personal dislike for giving the maximum mark for a given effect. This finding is in accordance with the findings of several other investigators (7, 8, 9). The conclusions which can be drawn for the sources of cough stimula- tion by deodorant sprays are that this irritation is due to the type of bac- tericide used and the percentage of bactericide included in the formula. Other factors, like the spray pattern and spray rate, the concentrations of solvents and perfumes, appear to play a secondary role. That respiratory tract irritation appears not to depend on spray rate is quite logical if it is realized that a test subject feels irritation from an aerosol cloud no matter how this cloud has been formed. This statement is supported by the fact that the test procedure has no influence on the actual irritation felt by the subjects. Wide differences appear to exist between the cough-stimulating effect of different bactericides. No correlation could be deduced between the bac- tericidal and the irritant properties for the different compounds studied, so we have only reported our actual findings. (Received: 21st February 1975) REFERENCES (1) Kinkel, H. J. Inhalation and toxicity studies. J. Soc. Cosmet. Chem. 15 395 (1964). (2) Kiibler, H. Physiological properties of propellants. J. Soc. Cosmet. Chem. 14 341 (1963). (3) Gloxhuber, C. Zur Methodik der toxikologischen Prti lung von Kosmetika. J. Soc. Cosmet. Chem. 21 313 (1970).

522 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS (4) O'Connor Ward, C. Current perspectives on aerosol toxicity. J. Soc. Cosmet. Chem. 25 271 (1974). (5) Troy, W. R. Testing for inhalation toxicity. J. Soc. Cosmet. Chem. 25 283 (1974). (6) Joyner, B. D. and Holmes, C. L. Manufacturing Chemist p. 8l May 1970. (7) Amerine, M. A., Pangborn, R. M. and Roessler, E. B. Principles of Sensory Evah•ation of Food (1965) (Academic Press, New York). (8) Guilford, J.P. Psychometric Methods (1954) (MacGraw-Hill, New York). (9) Ough, C. S. and Baker, G. A. Small panel sensory evaluations of wines by scoring. ttilgardia 30 587 (1961).