Summary
- We investigated patterns in the dispersion (i.e. spread and spacing) of plant species traits that are frequently associated with competitive ability, in an old‐field plant community. In contrast with previous studies, we found no evidence for significant over‐ or under‐dispersion of maximum plant height, maximum plant biomass, or seed mass.
- These findings hold across three plot sizes (10 × 10 cm, 30 × 30 cm, 50 × 50 cm), and when plot size is measured in terms of number of ramets (50 ramets, 250 ramets or 500 ramets) rather than per unit area. Plot size, however, significantly affected the direction of dispersion observed across test statistics, supporting previous studies that have observed that trait dispersion patterns are sensitive to plot size.
- While no significant dispersion was detected, dispersion direction, measured as the tendency for a two‐tailed test to indicate trends for a given trait, was more frequently observed when analyses were weighted by abundance. Abundance weighted analyses had significantly different dispersion directions compared with presence/absence analyses for one of three traits considered. These findings suggest that abundance weighted analyses may yield more consistent trait dispersion patterns.
- Synthesis. Our results may be interpreted as evidence for limited, if any, functional niche partitioning between co‐occurring species via size‐mediated differences in rooting depth or physical space niches (i.e. limiting similarity). Alternatively, some species with large differences in competitive traits may avoid competition through niche separation, while others may simply compete asymmetrically, leading to overall patterns of dispersion that cannot be distinguished from random. This interpretation supports recent perspectives that niche and neutral theoretical concepts are not mutually exclusive; both, working together, can be applied to the interpretation of plant community assembly and structure.
Methodology
Study Site
This research was conducted from June to September 2003 at the Queen's University Biology Station, Chaffey's Locks, Ontario, Canada (44°34′ N, 76°20′ W), in the Lane Sargeant Experimental Field. This field is on abandoned farmland and has been managed under a strict disturbance regime with some sections hayed each year, while others are tilled and/or ploughed on different yearly programs. The portion of the field used in this research had previously been undisturbed for a period of 31 years (neither hayed nor tilled nor ploughed) and is 100 × 27 m in size. The plant species living in the field are largely perennial in habit and are typical of what would be found in old‐field vegetation in this region. The five most common species within this plant community (measured in terms of ramet number) were Poa compressa, Carex cryptolepis, Phleum pratense, Elymus repens and Daucus carota .
Data Collection
Vegetation plots were located within the field using a random number table and each was assigned to one of the three plot sizes: 10 × 10 cm, 30 × 30 cm, and 50 × 50 cm. Within plots, each ramet was enumerated and identified to species for fixed increments of 10 × 10 cm until the desired plot size was reached. At that point, individual ramets were enumerated and identified as advocated by Zobel & Liira (1997) until a fixed ramet number was reached. For example, fixed plots of 10 × 10 cm were extended to 40 ramets (if that number had not already been reached), 30 × 30 cm to 250 ramets, and 50 × 50 cm plots to 500 ramets. Because the number of ramets per plot for flexible plots (i.e. flexible in terms of area, but fixed in terms of ramet number) was adjusted after collection to match more closely the average number of ramets for the fixed sized plots used, the number of plots defined by area and those defined by ramet number is not equal. The number of plots in each size class was determined such that overall area sampled per plot size was relatively equal, and overall species richness was similar across plot sizes. Species richness per plot varied both within and among the different plot sizes (Fig. 1).