1Danish Institute of Agricultural Sciences, Research Centre Flakkebjerg, 4200 Slagelse, Denmark; 2Ubon Rice Research Centre, Ubon Ratchathani 34000, Thailand; 3Mariano Marcos State University, Batac, The Philippines 2906; 4University of Agriculture and Forestry, Ho Chi Minh City, Vietnam; 5International Rice Research Institute, 1099 Manila, The Philippines

Background and objectives
Epidemiological and crop loss research depend on data indicating population sizes of biotic stress agents and/or crop injury levels caused by these. While quick yet accurate appraisals of pest problems require biologically meaningful parameters reflecting these, the procedures for collecting and processing data should be simple. A widely accepted parameter to characterize the whole-season pest burden on vegetative plant organs is the area under the disease or crop injury progress curve (AUDPC) while the maximum severity value (MSV) or the last severity value (LSV) observed during the growing season or just before harvesting, respectively, are suitable parameters to characterize pest problems on whole tillers and/or reproductive plant organs [1,2]. AUDPC is the most costly to obtain parameter, followed by MSV, LSV, and parameters representing crop injury levels at defined phases of crop development. The objective of this work was to identify the least costly to obtain and simplest yet biologically meaningful parameters for characterizing severity levels of biotic stresses in rice to facilitate the field assessment of pest problems.

Materials and methods
The severity levels of 24 types of biotic crop injury were assessed during the seedling, tillering, vegetative lag, early generative, and ripening phases as well as at harvesting in a total number of 377 plots in farmers' rice-fields in the Philippines, Thailand, and Vietnam in 1992 and 1993. General linear models were fit to the data to explain AUDPC, MSV, and LSV by covariates representing simpler parameters of crop injury levels and using country and year as dummy variables.

Results and conclusions
R2 values were >0.93 for most of the injury types when AUDPC was the dependent variable (y) and MSV the covariate (x). Any other constellation of y and x led to considerably lower R2 values in most cases, especially for some injury types typically occurring early in the season on vegetative plant organs, such as injury caused by hopper burn, leaf folders, thrips, and whorl maggot. For all types of panicle injury, such as dirty panicle, false smut, panicle blast, sucking insect damage and stem borer white heads as well as for some types of injury occurring late in the season on vegetative plant organs, such as bacterial leaf blight, brown spot, leaf scald, narrow brown spot, sheath blight, and sheath rot, other constellations were also associated with high R2 values, especially when the severity value observed during the ripening phase (RSV) served as x. The effects of country and year were usually significant, but their relative contribution to explaining the total variance of y was much lower than that of x in most cases. It was concluded that AUDPC, MSV, and LSV may often be estimated based on variables that are easier to obtain than each one of these. MSV may substitute for AUDPC in most cases. RSV may substitute for AUDPC and/or MSV and LSV for most biotic injuries on panicles and for injuries on vegetative plant organs caused by pest problems occurring late in the season. This may considerably facilitate field assessments of pest infestations and crop losses because MSV and RSV can be obtained from only one observation per growing season which, in the case of RSV, can conveniently be scheduled in advance. However, the relationships between various parameters of biotic crop injury may vary across countries and years and may have to be quantified based on observations in a subset of field plots.

1. Pinnschmidt HO, dela Pela F, Suriya-Arunroj D, Don LD, Teng PS, 1996. Journal of Plant Diseases and Protection 103, 620-643.
2. Savary S, Elazegui FA, Moody K, Litsinger JA, Teng PS, 1994. Agricultural Systems 46, 385-408.