Chemometric analysis of near-infrared (NIR) spectra recorded directly on fresh leaves of barley plants (Hordeum vulgare, L.) enabled the separation of control and Cu deficient samples before any visual deficiency symptoms developed. This demonstrates that the molecular structure of leaves is modified during latent Cu deficiency. Lignin biosynthesis is a primary target of Cu deficiency, but lignin concentrations were unaltered when separation was first possible, indicating that alteration of lignin composition, not concentration, is among the earliest effects of Cu deficiency in plants. Validation of chemometric models using an independent test set found that 92% of samples were correctly classified as control or Cu deficient, respectively. Models were undisturbed by including spectra from plants deficient in P, Mg, B, or Mn, establishing their specificity for Cu deficiency. This study is the first to demonstrate that NIR has the potential to successfully diagnose the deficiency of an essential trace element in plants.