A PRECISE METHOD TO MEASURE LEAF RUST (Puccinia triticina Eriksson) SEVERITY IN WHEAT

Abstract

Visual estimation of disease severity in plant organs is subjective and not repeatable, thus accurate methods to measure damaged surfaces are needed. This study proposes a method for measuring severity of leaf rust (Puccinia triticina Eriksson) on wheat (Triticum aestivum L.) cultivars INIA F-66, Jupateco-73R, Morocco, Sonora F-64 and WL-711 by using digital image analysis (AID). The actual percentage of leaf area damaged by the disease (MED, %) was estimated using AID and then compared to a visual assessment method of disease severity (EST, %) performed by three volunteers, on two samples with 10 replicates. Leaf images were scanned, and then AID was performed via automation with software (ImageJ 1.48r). Leaf area (AS, mm²), damaged area (AD, mm²), number (NTL), size (TAM, mm²) and shape of lesions were measured. Number of lesions per cm² (LPC) and MED was calculated. The two methods correlated with each other (r_s = 0.86, P ≤ 0.0001); although EST lacks accuracy. Severity was different among volunteers and AID (K-W ≈ X² = 21.73, P ≤ 0.05). MED and EST were different between cultivars (P ≤ 0.001). Volunteers overestimated EST when AD was less than 19 %, and they underestimated it when it exceeded this level. Morocco had the largest MED (49.4 %). Sonora F-64 and Jupateco-73R had the lowest NTL, TAM and LPC (P ≤ 0.001). Usage of AID has many advantages, among which it allows accurate identification of damaged and healthy leaf area; it requires less than 1 min to determine variables related with severity of leaf rust; and this method is repeatable, reduces experimental errors and subjectivity.