HEMP-SEED AND OLIVE OILS 239 anatomic similarity between pig and human skin makes the results obtained with this technique predictive for human skin absorption. The pig skin was prepared: A fresh pig's ear was thoroughly washed in physiological solution and the bristles removed. The skin was then fixed onto a cork mat and the external surface of the ear cut away. To make the conditions reproducible, skin strips of equal thickness were cut, consisting of horny layer, epidermis, and a part of the dermis there is little or no subcutaneous fat in the edge of the ear. Disks of the same diameter as the Franz cells (2.5 cm2 ) were cut from these strips. The emulsion was applied to the skin disk held in place between the donor and receptor compartments of the Franz-static diffusion cell the epidermis was in contact with the donor compartment and the dermal layer in contact with the receptor compartment, consisting of physiological solution (0.9% NaCl) under continuous magnetic stirring. To quantify the amount of Kathon CG® that had permeated through the skin, samples were taken every 60 minutes. After 24 hours, the receptor compartment was emptied and replaced with fresh physiological solution the solution removed was subjected to HPLC analysis, under the above conditions. At the end of 24 hours, the skin was cleaned, cut into small pieces, and placed in a beaker containing the eluent employed for the HPLC analyses, in order to determine any Kathon CG® trapped in the tissues. Each sample was injected into the HPLC apparatus following the procedures reported below. The experiment was repeated twice. The concentration of Kathon CG® in the receptor solution was found directly from the standard calibration graph obtained by injecting standard solutions of Kathon CG® at different concentrations, between 1 mg/1 and 6 mg/1, into the HPLC apparatus. The analytical conditions were as follows: • Column: Cromasil C 18 • Eluent: methanol/water: 30/70 • Flow: 0.7 ml/min • Detector: UV (A. = 275 nm) • Retention time: 7 .4 min The calibration curve showed a good linear correlation between peak areas and sample concentration (R2 = 0.9978). The results of the study of permeation showed that only small amounts of the preser­ vative penetrated the dermis: after 24 hours the concentration of Kathon CG® in the receptor phase was negligible. RESULTS AND DISCUSSION QUALITY INDICES AND PHYSICOCHEMICAL CHARACTERISTICS The quality indices and physicochemical characteristics of the hemp-seed oils (1998 and 1999) and olive oil are in Table II. ACIDITY The acidity indices and percentages of oleic acid are listed in Table III. Both batches of

240 JOURNAL OF COSMETIC SCIENCE Table II Physicochemical Characteristics of the Vegetable Oils Considered Hemp-seed Hemp-seed Parameters oil (1998) oil (1999) Olive oil Relative density (d204) 0.925 0.924 0.916 Rotational viscosity at 18.6 s- 1 (mPa s) 40 37.5 60 Capillary viscosity (mPa s) 40.420 Refractive index (RI) 1.4735 1.4760 1.4665 Table III Acidity Indices and Percentage of Oleic Acid in the Vegetable Oils Considered IA Oleic acid(%) Hemp-seed oil (1998) 3.85 (±0.46) 1.93 (±0.18) Hemp-seed oil (1999) 14.10(±1.31) 7.13 (±0.95) Values in parentheses are standard deviations, determined on four samples. Extra-virgin olive oil 55 1.4665 Olive oil 0.44 (±0.02) 0.22 (±0.01) IA = aliquot (in mg) of KOH O. lM necessary to neutralize the fatty acids that are free in 1 gram of fat mass. hemp-seed oil (1998 and 1999) had free acidity above 1 %, the upper limit for the "extra-virgin" category (EEC Regulations), while the acidity of olive oil was below the limit. Hemp-seed oil (1999) had a significantly higher free-fatty-acid content, although it was younger, probably due to the different method (Baglioni) employed for crushing the seeds. The acidity index of the commercial olive oil was in agreement with that declared on the label. PEROXIDE NUMBER Peroxide numbers, expressed as milliequivalents of 02 per 1000 grams of fat mass, were found to be 5 .15 for the 1998 hemp-seed oil and 1. 5 7 for the 1999 hemp-seed oil. Both kinds of oil had peroxide values below the limit of 10 established for "extra virgin" oil. Moreover, the peroxide numbers of both hemp oils were lower than 20, the threshold limit for edible oils. CONJUGATED DIENES AND TRIENES Results of UV characterization are reported in Table IV. Absorptivity differences at 232 Table IV K232 and 11K Values of the Oils Considered, Dissolved in 0.1 % 1-Butanol Hemp-seed Hemp-seed oil (1998) Hemp-seed Extra-virgin oil (1998) (after 24 h at 40°C) oil (1999) Olive oil olive oil K232 2.58 (±0.23) 2.66 (±0.11) 2.12 (±0.12) 2.72 (±0.08) 2.20 (±0.09) 11K 0.025 (±0.003) 0.030 (±0.001) 0.025 (±0.001) 0.075 (±0.002) 0.0 Values in parentheses are standard deviations, determined on four samples.