Predicting crown damage to Eucalyptus grandis by Paropsis atomaria with direct and indirect measures of leaf composition

A genetic basis for variation in resistance to defoliating insects within Eucalyptus species has been identified in many studies. This variation has frequently been ascribed to variation in secondary metabolites but studies investigating variation in resistance to defoliation by paropsine chrysomelids have failed to correlate foliar chemistry with resistance. We found that the extent of crown damage due to defoliation by Paropsis atomaria (Chrysomelidae: Colepoptera) in two matched progeny trials of Eucalyptus grandis was a heritable trait that exhibited a strong correlation with provenance latitude. Despite this, neither foliar nitrogen or concentrations of a recently discovered group of compounds, formylated phloroglucinol compounds, could account for significant variation in defoliation. We also investigated whether defoliation in the field could be predicted from foliar near-infrared reflectance spectra. Such an approach takes into account all compositional variation simultaneously rather than relying on a restricted number of measured traits. Modified partial least squares regression models performed poorly in predicting variation in crown defoliation between trees within a site primarily due to the high level of variability and coarseness of the calibration data. Discriminant analyses however demonstrated a consistent difference between spectra from trees in families suffering low level defoliation from families more susceptible to defoliation by P. atomaria suggesting chemical differences between the two groups are important in determining resistance to this insect.