| Available from UMI in association with The British Library.; A major concern for forensic scientists is the assessment of the evidential value of fibre comparisons. To this end, a fibre data collection has been established, for which a colour notation system was required. Modifications to the CIE tristimulus colorimetry system, giving complementary chromaticity coordinates, are shown to be acceptable for this purpose.; A method of describing errors in replicate colour measurements from single fibres, based on elliptical error regions, has been developed. Such ellipses are shown to be caused mainly by the inherent substrate fibre colour, with minor contributions from instrumental errors, delustrant level and fibre shape. The use of error ellipses, and some less computationally demanding alternatives, for colour-matching is demonstrated.; Microspectrophotometry is also shown to provide solutions to two problems in fibre characterisation; the measurement of birefringence and the discrimination of polyolefin fibres. For the latter, infra-red spectroscopy, melting-point determination and thin-layer chromatography are also evaluated.; For visually colourless fibres, examination by microspectrophotometry is inappropriate. Modifications made to the microspectrophotometer, converting it into a microspectrofluorimeter, enable fibres treated with fluorescent brightening agents to be discriminated. In addition, microspectrofluorimetry is shown to provide further discrimination for some conventionally dyed fibres.; In a similar manner to microspectrophotometry, the CIE system may be modified to provide colour definition for self-luminous sources. This has been evaluated for a range of fluorescent materials, but limitations in the method of excitation are shown to impose restrictions on the chromaticity coordinates obtained.; Upon prolonged excitation, many fluorescent species exhibit a marked decay in fluorescence emission intensity. The kinetics of this process, and its effects on colour definition have been assessed. Decay parameters are shown to be of little use as a method of discrimination. |
