JOURNAL OF COSMETIC SCIENCE 140 Table XXIV. Foundation Summary (n = 10) Results (mg/kg) As Cd Cr Co Pb Hg Ni Maximum 0.40 TR 5.0 2.5 1.2 TR 5.8 Minimum TR NF 0.46 0.16 0.16 NF 0.56 Mean 0.18 0.0072 2.3 0.77 0.64 NF 1.8 Median 0.14 NF 2.0 0.47 0.69 NF 1.2 NF: Not found, or less than detection limit TR: Trace, or greater than detection limit, but less than quanti- tation level. Table XXV. Body Powder Summary (n = 5) Results (mg/kg) As Cd Cr Co Pb Hg Ni Maximum 0.80 NF 3.9 1.6 1.5 NF 4.6 Minimum NF NF TR NF NF NF TR Mean 0.32 NF 1.6 0.82 0.48 NF 2.1 Median 0.29 NF 1.7 1.1 0.21 NF 1.6 NF: Not found, or less than detection limit TR: Trace, or greater than detection limit, but less than quanti- tation level. REFERENCE MATERIAL STUDY As a measure of accuracy, four reference materials were analyzed. Recovery limits were selected as 70%–130% for all certifi ed values less than 10 times the LOQ. For certifi ed values above 10 times the LOQ, recovery limits of 90%–110% were used. As a further test of ruggedness, recoveries were tested for three different weights of reference material: 0.25g, 0.50g, and 0.75g. There was a trend toward higher recoveries with lower weights for the custom oil standard for arsenic, cadmium, and possibly cobalt and nickel. These trends, however, were not deemed to be signifi cant with respect to variation in results. The recovery of mercury from its custom oil standard was also lower than for the other reference materials for an unknown reason. The recoveries for mercury ranged from 71% to 79% for 0.25g and 0.75g custom oil, respectively. The variation in recovery for differ- ent weights is not deemed to be signifi cant. Certifi ed values for the analytes in each refer- ence material and mean recoveries (from values obtained by analyzing the reference materials with each sample batch) are shown in Tables III and IV. The recoveries demon- strate good accuracy of the methods for all analytes. SUMMARY OF METHOD PERFORMANCE AND COSMETICS SURVEY There were no evident systematic errors. All calibration curves (y = ax + blank) had cor- relation coeffi cients of at least 0.997. All initial and CCV standards were within the control limits of 90%–110% recovery. All RPD for duplicates were less than 20% with

SURVEY OF COSMETICS FOR SEVEN INORGANIC ELEMENTS 141 Table XXVI. Compact Powder Summary (n = 5) Results (mg/kg) As Cd Cr Co Pb Hg Ni Maximum 0.65 0.07 16 14 7.8 0.004 28 Minimum 0.28 NF 1.4 0.82 1.6 NF 3.9 Mean 0.50 0.014 7.6 4.8 4.6 0.0010 12 Median 0.59 NF 4.8 3.0 4.4 NF 9.1 NF: Not found, or less than detection limit TR: Trace, or greater than detection limit, but less than quanti- tation level. Table XXVII. Shaving Cream Summary (n = 5) Results (mg/kg) As Cd Cr Co Pb Hg Ni Maximum 0.15 NF 0.50 NF TR NF 0.13 Minimum NF NF NF NF NF NF NF Mean NF NF 0.21 NF NF NF NF Median NF NF 0.16 NF NF NF NF NF: Not found, or less than detection limit TR: Trace, or greater than detection limit, but less than quanti- tation level. the exception of some analytes that were near or below the estimated LOQ, suggesting good sample homogeneity. SURVEY OF COSMETIC SAMPLES After method performance was found to be acceptable, the method was used to survey the 150 cosmetic samples. Results for individual cosmetic types are presented in Tables V—XVI. Capital letters designate manufacturers and small letters differentiate brands within a manufacturer. Summary tables for each cosmetic type and for all cosmetics are presented in Tables XVII—XXIX. Maximum and minimum results and numbers of samples (n) analyzed are included in the summary tables. Calculations of means and medians were made with values less than the LOD set to zero, but included numerical values for “trace” (TR) lev- els, which are results between the LOD and LOQ. Values shown in parentheses were calculated without three signifi cantly high results for cadmium, cobalt, and nickel that were identifi ed as outliers (see Discussion section). DISCUSSION The elements determined in this study range from less than the respective detection lim- its of each element to 1.7 mg/kg for arsenic in an eye shadow, 0.36 mg/kg for cadmium in an eye shadow, 22,000 mg/kg for chromium in an eye shadow, 13 mg/kg for cobalt in