42 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Correction: + 150•o brown. Result: Sample now reads L=73.2, a= +9.3, b= + 12.9. Visual assessment: A good match, but sample showing slight geometrical metamerism. In this case the correction was less easy since the sample started out, much too dark (too much brown and yellow had been added to it). Never- theless, a reasonable match was achieved with three corrections for a sample so far from the standard, this would have been no mean feat for the unaided colour matcher. For less complicated cosmetics, Fig. 11(a)-ll(d) can be replaced by 'skew axis' graphs of the sort illustrated in Fig. 12. This relates to a pearly powder eyeshadow product composed of two blue pigments and a base. A=3% Bluel, 25% Blue2 A peorlized blue powder eyeshodow A b I Blue ', ",, 'x '- ..... x x ' Ix 3 ' x •-• • • "• ' x 'x '• 'x x .... .•-, '• x 'x 'x .... ' ' 'x x 'x 'x :x ............ •'-q, x. x 'x x • .... -•'x •'•'• ', • 'x x x 'x. .'N ....... • "' 'x -Z x'c6 .-'• .... X. L4 x..-3 'x-2 •- •oø/•"J ...• ............ '• 2 x'•3 • 4 '• 5 x 6 _• z - •., 'x ', x ...'•."x'X ? •o% 1.. 'x. x., x. _._.• ..... • • ,. ....... ,.•o½-, ', '. '• .... •.-'-' 'l ......... '. " ", ...... q'- ". ", 'x •. • -' .... x 'x x, -,• x. . •, •. •- • '., x. x x. x x, . • ,• _ •. m 'x ', '• ...... c'- x. x. ',.1-5 ,. 2 ...... •'-' , ' x. C' ' ' '. -' • ...... '5 '1 , '., , '..• -.- -. •', , 'x x x .... • ...... x ß /'c9 • _ -.•'..-'- - ' ' 'x ., '. '• 2• ..... •'- ', '• x / 'x • ...... •' - x x. ', _k x..•._,•._.- x. x, 'x x. x. xl._ ,• .... x 'x '". x. x ..... 'x x. x 'x •.•.•.'•'-' x, 'x x x. 'x-.• - •'•' x e •/o x. x Blue l (purple) Figure 12. 'Skew' axis colour difference graph (Aa/Ab). Such a graph is prepared by plotting the differences in each dimension from the standard resulting from the addition of known quantifies of each pigment to the standard batch. The 'skew' axes are then drawn through each line of points and a grid of parallel lines constructed. A batch sample having differential co-ordinates situated at A, for example, would require approximately 3•o of blue 1 and 25•o of blue 2 to get to the same 'a' and 'b' co-ordinates as the standard.

INSTRUMENTAL COLOUR ASSESSMENT 43 FURTHER EXAMPLES OF THE APPLICATION OF COLOUR MEASUREMENTS IN A COSMETICS LABORATORY We have found that tristimulus colour measurements are applicable to problems other than colour matching within our laboratory, especially those concerned with the storage testing of raw materials and products. Table V, for example, relates to three alternative blue foam bath formulae. Table V. Appearance changes produced during 24 h fade test on three bath foams with varying quantities of UV absorber Foam Bath 1 2 3 Change AL Aa Ab AL Aa Ab AL Aa Ab No UVabsorber +4.9 +15.3 +5.9 -1.4 +10.6 +0.6 -4.0 +16.8 -1.6 0.1% UVabsorber +3.9 +10.5 +4.6 -0.3 +6.7 +0.9 -2.4 +11.1 -0.2 0.2%UVabsorber +4.9 +10.7 +4.7 +1.6 +4.0 +1.4 -2.9 +8.9 -0.9 Since actual numbers can be meaningfully applied to these appearance changes, it becomes muclx easier to rate the relative performance of each formulation in the fade test. In addition, these data also indicate an un- expected relationship between the amount of UV-absorber added and the subsequent change in appearance of the product--a clue worth following up. Table VI relates to the storage testing of a certain shade of soap made from alternative soap bases. Once again, the value of these data lies in the acquisition of meaningful figures concerning subjective appearance differ- ences by which means the relative performance of each base can be assessed in relation, for example, to its cost. Table VI. Changes in the appearance of finished soap made from two soap bases during laboratory storage test procedures of 1 month duration Base 1 2 Change AL Aa Ab At, Aa Ab 4øC -4.1 +3.8 -4.2 -0.8 -10.7 -0.4 Ambient -4.8 +4.4 -5.0 -1.2 +1.2 -0.5 Cyde* -3.4 +4.1 -2.9 -0.8 +1.3 +0.1 32øC -4.7 +4.4 -4.3 -0.8 +1.2 0.0 37øC -3.6 +4.0 -3.6 -1.5 +1.3 -0.3 46øC -3.6 +3.8 -1.8 0.0 +1.4 +1.0 * 4øC - 46øC twice in 24 h.