Question: Monocot and dicot perennial herbaceous species can be considered two a priori guilds, which interact in regularly mown grasslands. Complementary resource use in space is important for diversity and productivity of perennial herbaceous communities, yet small-scale spatial patterns of guilds and niche complementarity have rarely been linked explicitly. We ask: 1. Do guild interactions in our community generate small-scale patterns such as guild proportionality? 2. Do these patterns coincide with overyielding of guilds, that is, can small-scale spatial patterns of guilds be experimentally linked to niche complementarity?

Location: Georgia, central Caucasian mountains.

Methods: We sampled randomly placed small quadrats and used presence-absence data to measure small-scale spatial patterns within plots. We experimentally tested overyielding of monocots and dicots grown separately and in mixtures.

Results: We found measurable small-scale pattern of guilds in our community: monocots and dicots showed significant guild proportionality. However, variance deficit in the number of species per small quadrat was stronger in monocots than in dicots. Experiments found considerable overyielding, which was unaffected by water stress.

Conclusions: Coincidence of measurable small-scale arrangements of guilds with robust overyielding indicates that spatial arrangement is an important manifestation of niche complementarity. Stronger variance deficit of monocot species at small scale suggests that monocots compete with each other more strongly and, as a consequence, are more regularly dispersed at small spatial scale. Conversely, morphologically more diverse and more plastic dicots may more easily fit within the neighbourhood patchwork created by monocots.