Root growth patterns respond to small-scale resource heterogeneity and the presence of roots of neighboring plants, but how a plant integrates its responses to these cues is not well understood. In the presence of neighbors, plants may shift allocation to roots as a consequence of plant size and root:shoot allometry, as a response to resource depletion by neighbors, or through a direct response to neighbor presence. The same response pathways also have the potential to alter proliferation in resource-rich patches in soil.

Four species of grassland plants were grown in the greenhouse as single plants, monocultures, and mixtures. Root length allocation as a function of shoot mass was examined for background soil and fertilized patches. Plants grown with same-species neighbors followed the same allometric trajectory as single plants for root length in background soil, so any change in root allocation was due only to reduced plant size. Root proliferation in patches declined with neighbors, consistent with a response to resource depletion. Mixtures overproduced roots in both background soil and in patches, relative to plants of the same size in monocultures.

Four co-occuring grassland species were included in the experiment: two grasses, Dactylis glomerata L. and Poa pratensis L., and two forbs, Centaurea jacea L. and Plantago lanceolata L.

Each pot (15 cm square x 30 cm deep) was filled with a background soil containing a mix of 50% (by volume) commercial topsoil, 30% sand, 10% milled peat, and 10% perlite. Two resource-rich patches were created in each pot by removing a soil core (1 cm diameter x 15 cm deep), adding 0.5 g of Osmocote slow-release fertilizer (19-6-12 NPK; Scotts-Sierra Horticultural Products, Marysville, Ohio, USA) mixed with about 15 ml background soil, and refilling the rest of the core with background soil. Patches were placed in opposite corners of the pot, equidistant from the pot edge and center. Control cores containing only background soil with no fertilizer were removed and refilled in the two remaining corners.

Seeds were germinated in background soil and transplanted into experimental pots after four weeks. Each pot was planted with a single plant, a monoculture (four individuals of the same species), or a mixture (one individual of each species). Each plant was positioned between a patch and a control core. Specific positions of single individuals and of species in mixtures were determined randomly. Pots were maintained in the greenhouse with natural light and watered daily.

Plants were harvested after 60 days of growth in pots. Shoots were clipped at the soil surface, dried at 80 C for 48 hours and weighed. Shoots of all four individuals in monocultures were weighed together. Soil cores (2 cm diameter x 20 cm deep) were removed at the positions of resource-rich patches and control patches. Cores from the two patches of each kind were combined for each pot. The soil was air-dried for at least 48 hours and the roots were removed from the soil. Each root sample was scanned for determination of total root length using a WinRhizo root image analysis system (Regent Instruments, Quebec, Canada). Roots in mixtures could not be separated by species.