• Nishizawa, Tomo
  • Aarssen, Lonnie W.


In most natural plant populations, there is a strong right-skewed distribution of body sizes for reproductive plants—i.e. the vast majority are relatively small, suppressed weaklings that manage not just to survive effects of crowding/competition and other hazards but also to produce offspring. Recent research has shown that because of their relatively large numbers, these relatively small resident plants collectively contribute most of the seed offspring production available for the population in the next generation. However, the success of these offspring will depend in part on their quality, e.g. reflected by seed size and resource content. Accordingly, in the present study, we used material from natural populations of herbaceous species to test the null hypothesis that there is no significant relationship between body size variation in resident plants—resulting from between-site variation in the intensity of crowding/competition—and variation in the mass or N content of their individual seeds.

Using populations of 56 herbaceous species common in eastern Ontario, total above-ground dry plant mass, mean mass per seed and mean nitrogen (N) content per seed were recorded for a sample of the largest resident plants and also for the smallest reproductive plants growing in local neighbourhoods with the most severe crowding/competition from near neighbours.

Mass per seed was numerically smaller from the smallest resident plants for most study species, but with few exceptions, this was not significantly different (P > 0.05) from mass per seed from the largest resident plants. The results therefore showed no general effect of maternal plant body size on individual seed mass, or N content. This suggests that the reproductive output of the smaller half of the resident plant size distribution within these populations is likely to contribute not just most of the seed production available for the next generation but also seed offspring that are just as likely—on a per individual basis—to achieve seedling/juvenile recruitment success as the seed offspring produced by the largest resident plants. This conflicts with the traditional ‘size-advantage’ hypothesis for predicting plant fitness under severe competition, and instead supports the recent ‘reproductive-economy-advantage’ hypothesis, where competitive fitness is promoted by capacity to produce offspring that—despite severe body size suppression imposed by neighbour effects—in turn have capacity to produce grand-offspring.


Sample collection

Natural populations of 56 plant species of annuals, biennials and perennials from 19 different families (Table 1) were sampled from old fields and recently disturbed (e.g. roadside) habitats in the vicinity of Kingston, Ontario (44°17′N, 76°34′W). Study species were selected based on local availability and habitat accessibility but were otherwise chosen without bias. All sampled populations were from sites with conspicuous local variation in neighbourhood density (based on visual estimation), where competition between plants was visually apparent—i.e. with at least some resident plants growing in contact with a high local density of near neighbours and/or with overtopping larger neighbours.

Each population (one per species) was monitored over the growing season (May–September 2010) until seeds were mature, including especially for the smallest plants growing in the most densely crowded neighbourhoods or with intense competition from larger near neighbours (of the same or other species). Once seeds were ready for natural dispersal, 2–5 (mean = 4.05) of the largest plants from the population and 3–11 (mean = 6.96)—depending on population size—of the smallest reproductive plants growing under intense crowding/competition were harvested above ground and stored in open paper bags in the lab for later processing.

Processing and analyses

In the laboratory, all seeds were collected from each of the smallest plants and at least 200 seeds were collected from each of the largest plants. Unfertilized ovules and aborted seeds were easily distinguished by size and colour (and disregarded accordingly), but many species exhibited a small but noticeable range of seed size even on single plants, and these all had equal probability of collection. Collected seeds were air-dried and stored, separately for each plant, at 2°C. The mass of random seed lots from each plant was used to calculate mean individual seed mass—with the number of seeds per lot depending on species seed size, and chosen to ensure a total mass of at least 0.00010g, sufficient for the resolution of the analytical balance (0.00001g) (for the smallest plants: n = 1–400 seeds (mean = 22.0); for the largest plants: n = 1–213 seeds (mean = 28.2)). Harvested plants were dried to constant final mass at 70°C for 48h and individual above-ground biomass was measured.

In order to determine the N content (%) of the samples, 0.1500–0.3100g of dried seeds were placed in a weigh boat and run in a nitrogen analyzer (LECO CNS-2000). Samples were combusted in a furnace to convert elemental N to N2 and NOx; NOx was further reduced to N2 and detected by a thermoconductivity detector. For the smallest plants of each species—because of their severely limited seed production per plant—it was necessary to pool seeds from all replicate plants in order to obtain a mass sufficient for N analysis, and even with pooling, mass was sufficient for only 34 of the 56 study species.