THE ACTION OF LIGHT ON COLOURING MATTERS 253 Occasionally the products of photodegradation are coloured - yellows, in particular, often turn brown, but most other observed changes in hue arise because the original material contained two or more colourinõ matters which faded at different rates. Fadinõ is confined to organic colourinõ matters inorganic pigments used as colouring matters are in general absolutely fast to light with the exception of lead chromate which darkens. As all organic colourinõ matters fade, the technologically important characteristic is the rate of fading and the most important generalization, which was first discovered by Bancroft in 1814, is that the fading rate is markedly affected by the substrate. Cases are known where a change of substrate, e.õ. from cotton to nylon, causes the dye to fade as much as a hundred times faster. Hence we never refer (except loosely) to the light fastness of a dye or pigment but to the light fastness of a dyed sub- strate or of a pigment in a certain medium. With a given dye/fibre combination, however, all the available evidence suggests that the amount of photodegradation is a product of intensity and time-assuming constant spectral enerõy distribution, but, unfortunately, the relationship between the two is rarely simple. In part this is un- doubtedly due to the fact that the amount of degradation a given dose of radiation will produce depends on two other factors- the temperature and moisture content which are themselves related (19). Dyed textiles, the fading rates of which are unaffected by moisture content, usually fade more rapidly at higher temperatures and therefore if the level of light intensity is doubled, the temperature of the textile will inevitably rise and more than twice as much fading will occur for the same dosage. With dyed textiles whose fading rates are moisture-sensitive, however, the con- verse occurs-the temperature increase reduces the moisture content which reduces the amount of degradation for a given dosage so that a doubling of intensity may not even increase the fading rate. This produces the striking phenomenon that a curtain in a north-facing room may actually fade faster than another made from the same material and hung facing south. DETERMINATION OF LIGHT FASTNESS As all organic colouring matters fade, but at widely differing rates, determination of the fading rate of a dyed substrate is of considerable technological importance. The most obvious way of determining the fading rate of a coloured substance is to expose it to a standard light source and determine the time required to cause a certain amount of fading. This method has been widely employed in the U.S.A., using both

254 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS daylight and artificial light as radiation sources. The daylight testing method was based on the assumption that there is a threshold level of intensity below which no fading occurs which Gebhard (20) claimed in 1909. This may not be surprising to cosmetic chemists because there is such a threshold in the case of the action of uv light on skin- the pro- duction of an erythema. To a textile chemist, however, the idea is pre- posterous as he is well aware that coloured fabrics fade in north-facing windows or under fluorescent lighting. However, for many years, com- mercial fading stations in the U.S.A. were reporting only the hours of bright sunshine and ignoring the effect of lower levels of radiation until the threshold theory was completely disproved by extensive experiments (21). The current method is to measure total radiation falling on the specimens by means of a pyrheliometer, an instrument widely used in meteorology. It can only be reliable, however, if the ratio of actinic to total radiation is always constant this may well be true in those sunny parts of the U.S.A. where commercial fading stations have been established, but it is extremely doubtful if it will hold where the skies are frequently overcast the uv content is by no means a constant fraction of the total. The artificial light source which has become a well-established standard in the U.S.A. is the enclosed carbon arc, and if lamps known as Fade- Ometers are correctly maintained and operated, they will emit a reasonably constant (both in amount and composition) dosage in a given time. The Standard Fade-Ometer Hour has thus become a widely used parameter of light fastness in the U.S.A. Its greatest weakness, however, lies in the fact that the composition of the light emitted is markedly different from daylight, the near uv being in considerable excess (22). For this reason it is slowly being displaced by the xenon arc whose spectral quality is, to all intents and purposes, identical with daylight because the intensity decreases as the arc ages, however, the standard fading hour parameter cannot readily be used and is not officially permitted. The method (23) which has become universal throughout the world avoids the necessity of having a standard source by employing a series of coloured wool patterns to measure, in effect, the actinic radiation falling on the specimen which was basically the method used by Dufay in 1729- such standards are termed "actinometers." The current light fastness standards are the direct descendants of work started in Germany in 1911 and augmented by work sponsored by the Society of Dyers and Colourists in 1927. These standards are pieces of wool dyed with different blue dyes, number 1 being the most fugitive, fading perceptibly after 1 hr exposure