350 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Methods A modification of the wet-dry bulb thermometer apparatus described by List and coworkers (14) was used for the in vitro evaluations (Fig. 1). This consisted of an air-tight chamber, made of styrofoam, with a glass front, a plexiglass top, and partially filled with silica gel to act as a desic- cant. The chamber was fitted with a thermometer-rubber stopper system on the left, which constitutes the dry bulb, and a thermometer-metal stopper system on the right, which constitutes the wet bulb. The prep- arations were introduced directly into the metal cylinder from the oint- ment tube by means of a specialized attachment. The thermometer bulb was then placed into the preparation, rotated to cover the bulb com- pletely, and then suspended at the same level as the dry bulb with the cork stopper. The difference in the readings of the dry and wet bulbs, At, was read each 1/2 min as a function of time. Silica Gel Figure 1. Wet-dry bulb thermometer for in vitro testing of cooling effect The in vivo experiments were conducted on students using a tele- thermometer (Model 43TA) * and skin probe (Model 408).* A plastic stencil 2 mm thick with a circular opening of 4 cm •' was placed on a smooth area of the lower arm the opening in the stencil was then filled with the preparation to be tested, and smoothed over with a spatula to make sure a constant volume was applied for each test. The skin probe was placed into the preparation, moved to the skin surface, and the ß Yellow Springs Instrument Go., Yellow Springs, Ohio 45387.

COOLING EFFECT OF COLD CREAMS 351 difference between the normal skin temperature and the temperature of the treated area, /xt', was read each minute as a function of time. No fluctuations in temperature of the control area of the skin were observed. This procedure was selected over several others which have appeared in the literature (15). The other procedures were tried, but found to be inadequate. To determine if the heating of skin (--/xt') exhibited by some prepara- tions was due to an oil-based occlusive film depositted by the oil of the preparation which did not permit normal skin respiration, the following test was performed. Preparations 1-5, 8, and 10 were placed on the skin. The cooling effect was allowed to rise and fall back to zero, at which time a plastic prescription vial containing a dried strip of filter impregnated with cobalt chloride solution was placed over the prepara- tion. If normal respiration ensued, the cobalt paper would turn blue to pink if respiration was blocked, the paper would not change color. The vials were kept capped in an analytical oven until the time of use. RESULTS Each preparation was tested three to seven times in vitro. The data obtained for each preparation were then averaged and plotted against time on numeric graph paper. Figure 2 shows an example of one of these graphs. Figure 3 shows the mean curves of all the preparations tested. The axes and scales are the same as in the previous graph, but the range lines have been omitted. The results of the in vitro tests show that the 20 250 40 50 60 MINUTES Figure 2. Example of a typical in vitro cooling curve (formula No. 6 in Table I). Vertical bars indicate standard deviations in 7 runs