Abstract
Apple proliferation, associated with infection by 'Ca. Phytoplasma mali', is an economically important disease of apple, particularly in Europe. Because there is no curative treatment, the management of the disease lies primarily within management of the insect vectors of 'Ca. P. mali' and the eradication of infected trees, which must be diagnosed by costly and time-consuming molecular analyses or according to the presentation of symptoms that often requires expert knowledge. To complicate matters further, infected trees do not always present symptoms, and other stresses, especially mineral nutrient deficiencies such as those in N and P, cause similar symptoms. Preliminary work has shown that on site spectroscopy has the potential to overcome these obstacles for the diagnosis of 'Ca. P. mali' infection and serves as the basis for other methods that could potentially be applied easily and economically within practical settings. In this presentation, further results elaborating on the diagnosis of 'Ca. P. mali' infection based on an in-field spectroscopic method, as well as results of a potted trial relating to the performance of the same method for diagnosis when other trees are experiencing moderate and severe N and P deficiencies, are presented and compared. As previously observed, predictive ability using data collected in-field, from late spring through summer, was negligible but increased drastically as the growing season progressed. Results from April were inconclusive. Multiclass PLS-DA models from the potted trial were highly successful and also increased as the growing season progressed. These models were more complex than the binary " in-field " models and required wavelengths distributed widely across the spectrum.