24• Plants' spectra provide integrative measures of their chemical, morphological, anatomical, and 25 architectural traits. We posit that the degree to which plants differentiate in n-dimensional 26 spectral space is a measure of niche differentiation and reveals functional complementarity. 27• In both experimentally and naturally assembled communities, we quantified plant niches using 28 hypervolumes delineated by either plant spectra or 10 functional traits. We compared the niche 29 fraction unique to each species in spectral and trait spaces with increasing dimensionality, and 30 investigated the association between the spectral space occupied, plant growth and community 31 productivity. 32• We show that spectral niches differentiated species better than their functional trait niches. The 33 amount of spectral space occupied by individuals and plant communities increased with plant 34 growth and community productivity, respectively. Further, community productivity was better 35 explained by inter-individual spectral complementarity than by productive individuals 36 occupying large spectral niches. 37 • The degree of differentiation in spectral space provides the conceptual basis for identifying 38 plant taxa spectrally. Moreover, our results indicate that the size and position of plant spectral 39 niches reflect ecological strategies that shape community composition and ecosystem function, 40 with the potential to reveal insight in niche partitioning over large areas with spectroscopy. 41 42 Keywords 43 complementarity, functional traits, ecosystem function, Hutchinsonian niche, plant species 44 identification, spectroscopy, spectral space, remote sensing 45 3