Root contrast enhancement for measurement with optical scanner-based imageanalysis

Fine roots (diameter <0.30 mm) account for about 80% of total root length i n many species but are often underestimated owing to their small size and n ear transparency. We have assessed a technique for rapid and accurate estim ation of root length when using image analysis. To allow the collection of clear images, several stains were tested to determine when staining is bene ficial. Plants of six species (barley (Hordeum vulgare L.), maize (Zea mays L.), oat (Avena sativa L.), soybean (Glycine max (L.) Merr.), strawberry ( Fragaria x ananassa), and wheat (Triticum aestivum L.)) were grown for eith er 3 or 5 weeks after emergence. Roots of five woody species (honey locust (Gleditsia triacanthos L.), Norway maple (Acer platanoides L.), red ash (Fr axinus pennsylvanica Marsh.), silver maple (Acer saccharinum L), and hackbe rry (Celtis occidentalis L.)) were collected outdoors. Stains were evaluate d over a range of concentrations and for three staining times. Root variabl es were measured with WinRHIZO, an interactive scanner-based image analysis system. Two approaches were used: (i) measuring specific amounts of roots at several stain concentrations and contrasting the results with an unstain ed control and (ii) measuring unstained roots, recovering the roots, staini ng them and measuring them again. The tristimulus coordinates (L*a*b*) of u nstained root colour were measured by spectrophotometer to determine if roo t colour might be used as an indicator as to whether or not staining improv ed the image contrast. Staining improved image contrast and substantially i ncreased measured root length estimates for species with low root contrast (maize, barley, and wheat), gave small improvements for hackberry, Norway m aple, and honey locust, but had no effect on other species. Staining for 10 min was sufficient to maximally enhance the image contrast of species with a high proportion of fine roots. Tristimulus coordinates indicated that di fferences among species correlated with the distribution of root diameter. Establishing whether staining of root samples is required should be accompl ished before any study of root systems with a high proportion of fine roots is undertaken.