Two methods for the analysis of colorimetric components applied to plant stress monitoring

Several authors have shown the possibility of early detection of many stress conditions inplants by analysing leaf colour. The availability of portable instruments, increasinglyaccurate and reliable, leads to the belief that, in the near future, this technique could be usedfor routine analysis in the field. Nevertheless, due to the variability of the leaf colourparameter and consequently to the large amount of data to be collected to obtain arepresentative sample, direct analysis of chromaticity coordinate values, as generated by aninstrument, could give results that are hard to interpret, especially when the colour variationto be detected is subtle, such as in leaves at the first stages of stress. The availability ofstandard methodologies to analyse colorimetric data would allow them to be interpretedcorrectly and, in a short time, to widen their fields of application. In this work two methodsof elaboration, with related software (PCA and COLORE) that give an objective evaluationof colorimetric variations of stressed plants by means of specific indexes of alteration, areillustrated. PCA, based on principal components analysis, allows two 'indexes', Tij and D, tobe calculated, that can describe the differences between two or more treatments. It alsoallows the location of the different treatments on the chromaticity diagram to be graphicallyrepresented. COLORE calculates the values of the dominant wavelength and purity of thecolour. From these, it computes the CI index that can be related to degree of stress. The twosoftwares were applied to data from different studies of thermal and water stress in Sorghumbicolor ((L.) Moench) and Citrus macrophylla (L). The results showed a clear correlationbetween the colorimetric indexes and the ecophysiological parameters chosen as indicators ofstress.