The roots of the vast majority of terrestrial plant species are colonized by mycorrhizal fungi. The most common mycorrhiza is the vesicular–arbuscular type which is characteristic of most herbaceous species and a great many woody species (Harley & Smith 1983). It seems likely that the vesicular–arbuscular mycorrhiza is common at least partly because the symbiosis can increase the fitness of the host plant.
On 30 May 1991 five blocks each containing four 4 m2 plots of mycorrhizal (M, Glomus intraradices Schenck & Smith) and four plots of non-mycorrhizal (NM) Abutilon theophrasti Medic. were established in a field site at the Russell E. Larson Agricultural Research Center of the Pennsylvania State University in Centre County, PA, USA (details published in Stanley et al. 1993). The soil is Hagerstown silty clay loam which has a bicarbonate-extractable phosphorus concentration of ≈ 10 μg P g–1. None of these plants received phosphate fertilizer. On 19 July 1991 the mean fractional root length infection by mycorrhizal fungi was 62% and 0·2% in M and NM plots, respectively. Seeds were collected from M and NM plots on 16 September 1991 and stored for future use at a constant temperature (22 ºC).
The seeds used in this study were chosen randomly from the stored seeds. Seed nitrogen (Nessler method, Jensen 1962) and seed phosphorus (molybdo-phosphate method, Watanabe & Olsen 1965) concentrations were determined for 36 individual seeds randomly selected from each maternal mycorrhizal treatment following digestion in a mixture of H2SO4 and H2O2 at 400 ºC.
Eight boxes (representing eight replicates) were constructed of untreated wood with interior dimensions of 50 cm × 50 cm × 32 cm high. Each box was filled with ≈ 80 l of Hagerstown silty clay loam soil collected from the site described above in May 1994 following fumigation with methyl-bromide/chloropicrin (2:1) at a rate of 560 kg ha–1. The top 7–8 cm of soil within each box was sieved through a 1 cm mesh to create a uniform seed bed. Individual square planting cells within each box were delineated by a grid of nylon monofilament laid out across the soil surface of each box. Each cell was 1·5 cm on a side. The boxes were placed in a greenhouse under ambient light in June 1994. The greenhouse was ventilated to maintain near-ambient temperatures.
Seeds from M and NM plants were randomly chosen for planting. They were first acid-scarified in concentrated H2SO4 for 10 min, rinsed in distilled water, then placed on moist paper towels. Within 12–24 h imbibed seeds were approximately twice as large as unimbibed seeds and a much lighter shade of brown. Imbibed seeds from M and NM plants were planted (1–3 June) in alternate cells in the centre 30 cm × 30 cm ‘observation area’ of each box. Each observation area contained 200 seeds from both M and NM plants. In order to reduce edge effects, the remaining 10 cm-wide perimeter around the observation area in each box was planted with seeds from the same batch of seed originally used to create the maternal plants (Stanley et al. 1993). All boxes were watered daily to field capacity with tap water but received no additional nutrients. Safer soap (Impede) was applied on 17 July and Dycarb was applied on 23 July to control White Flies.
The offspring seedlings were monitored daily for the first 2 weeks and emergence dates were recorded for all seedlings in the 30 cm × 30 cm observation areas. Two partial harvests from within the observation areas were performed during the course of the experiment, the first at 20 days after planting and the second at 47 days after planting. These times coincided with a period before and immediately after extensive self-thinning. For the harvests at 20 and 47 days offspring were taken from a randomly located 6 × 6 cell square from each box, each square containing potentially 18 offspring of both M and NM plants. Care was taken at 47 days to avoid sampling near the edges created by the 20 day harvest. When individuals were harvested, the shoots were cut at the soil surface. Shoot material was dried in a 65 °C oven for 48 h for both harvests. Dried shoot material was analysed for N and P concentration as above. For the offspring plants from the 20 day harvest, survival, shoot height, shoot dry mass, the number of leaves, total leaf area, and shoot N and P concentrations were recorded. Leaf area was measured using a leaf area meter (Delta-T, Cambridge, UK). For the offspring plants from the 47 day harvest, survival, shoot height, shoot dry mass, the number of leaves, and shoot N and P concentrations were recorded. At the 47 day harvest, samples of roots from each harvest area within each box were taken with a cork borer (size 15) and assessed for mycorrhizal infection (Koide & Mooney 1987).
Relative frequency histograms were constructed for each replication using 12 size classes for shoot masses at 20 and 47 days that included offspring of both M and NM plants. A similar 12-size class histogram was constructed for seed masses using four random samples of 36 seeds each from both M and NM plants for comparison.
Seeds produced by the offspring were collected from individual plants within the 30 cm × 30 cm observation area as they matured. A final harvest of all the remaining plants occurred 94–98 days after planting. Dry masses of stems only were recorded, because many of the leaves had abscised. A tally of the apparently viable (not aborted) seeds remaining on the plants was also made. Shoot material was dried in a 65 °C oven for 72 h.
Data (except for survival) were analysed using the single factor (maternal mycorrhizal treatment) analysis of variance procedure (STSC 1991). An effect was considered significant when P ≤ 0·05. The difference in survival between offspring of different maternal mycorrhizal treatments was compared with a χ2 analysis of the numbers of individuals following the procedure for replicated χ2 tests pooled over all replicates (Zar 1984). The numbers of individuals alive at the 20 and 47 day harvests relative to the number of individuals that were alive at 10 days were used to calculate survival.
The distributions of the masses of seeds used in this study (produced by either M or NM plants) were essentially normal (Fig. 1a). The mean P concentration of seeds produced by M plants (0·67%, SE 0·02) was 37% greater (P < 0·05) than that of seeds produced by NM plants (0·49%, SE 0·01). The mean nitrogen concentration of seeds was not significantly affected by maternal plant mycorrhizal infection (M plants: 3·4%, SE 0·03; NM plants: 3·2%, SE 0·03). The mean mass of seeds produced by M plants (10·3 mg seed–1, SE 0·1) was 6% greater (P < 0·05) than the mean mass of seeds produced by NM plants (9·7 mg, SE 0·1).
Emergence by 10 days was very high and not significantly different between offspring of M and NM plants (means were 98·6% and 97·3%, respectively). All seedlings that emerged were still alive at 20 days (Fig. 1b, inset). At 20 days there were fewer offspring of NM plants represented in larger size classes and more offspring of NM plants represented in the smallest size classes compared with offspring of M plants (Fig. 1b). Moreover, offspring of M plants had significantly greater mean shoot height, shoot dry mass, number of leaves and leaf areas than offspring of NM plants (Table 1). Offspring of NM plants had significantly greater mean shoot N and P concentrations than offspring of M plants but the reverse was true for N and P contents owing to the significantly greater mass of offspring of M plants (Table 1).
Table 1. . Mean (SE) of growth and nutrient uptake of offspring of mycorrhizal (M) and non-mycorrhizal (NM) plants at harvest 1 (20 days after planting): n = 8
As a result of extensive mortality, only 41·6% of all seedlings that had emerged survived to 47 days. Mortality was not random with respect to maternal treatment; however, as the χ2 analysis showed that offspring of M plants had significantly greater survival than offspring of NM plants (Fig. 1c, inset). This harvest occurred when flower buds began to appear. Of all the plants harvested at 47 days only 16 plants possessed flower buds and 15 of these were offspring of M plants. Offspring of M plants accounted for nearly all individuals in the larger size classes as well as the majority in the smallest size class (Fig. 1c), suggesting that many of the smallest offspring of NM plants died. Offspring of M plants had significantly greater mean shoot height, shoot dry mass and number of leaves than offspring of NM plants (Table 2). Again, shoot N and P concentrations were significantly higher in offspring of NM plants, but the reverse was true for shoot N and P contents (Table 2). None of the roots sampled at the 47 day harvest were colonized by mycorrhizal fungi.
Table 2. . Mean (SE) of growth and nutrient uptake of offspring of mycorrhizal (M) and non-mycorrhizal (NM) plants at harvest 2 (47 days after planting): n = 8
Mortality continued between the 47 day harvest and final harvest but, again, cumulative mortality was not random with respect to maternal mycorrhizal treatment. The χ2 analysis showed that significantly more offspring of M plants survived than offspring of NM plants (Table 3). Offspring of M plants therefore constituted 68·6% of the total survivors while offspring of NM plants constituted 31·4%.
Table 3. . Mean (SE) of survival and reproduction of offspring of mycorrhizal (M) and non-mycorrhizal (NM) plants by the final harvest (94–98 days after planting): n = 8
Surviving offspring of M plants had significantly greater mean stem masses than offspring of NM plants [mean (SE) for offspring of M plants: 501 (42) mg; for offspring of NM plants: 277 (34) mg]. Total shoot masses are not reported because there was substantial leaf loss by the final harvest.
Among the survivors, a significantly greater proportion of offspring of M plants were reproductive compared with offspring of NM plants (Table 3). Furthermore, of those reproducing, offspring of M plants produced significantly more capsules than offspring of NM plants and the capsules produced by offspring of M plants contained significantly more seeds than those produced by offspring of NM plants (Table 3). This resulted in almost four times as much cumulative seed production by offspring of M plants as offspring of NM plants (Table 3). The relationships between plant size and seed output were virtually indistinguishable for offspring of M and offspring of NM plants (Fig. 2). The mass, N or P contents of these seeds were not significantly affected by maternal treatment [mean (SE) for offspring of M plants: 9·1 (0·2) mg, 889 (21) μg N seed–1, 291 (8) μg P seed–1; for offspring of NM plants: 8·8 (0·2) mg, 838 (26) μg N seed–1, 289 (9) μg P seed–1].
When all harvesting was complete, boxes were broken open to examine root systems. In all cases the depth of the boxes appeared to be sufficient as only ≈ 10% of all roots reached the bottom.
Previously we showed that for A. theophrasti grown in the field, mycorrhizal infection resulted in an approximate twofold increase in the number of offspring produced (Stanley et al. 1993; Koide et al. 1994). We had earlier suggested that the quality of the offspring might also be significantly affected by mycorrhizal infection of the maternal generation (Lu & Koide 1991). The current study demonstrates two important consequences of this variation in offspring quality.
First, in these self-thinning populations survival among offspring of M plants was more than twice that of offspring of NM plants. In previous studies mortality was shown to occur predominantly among smaller individuals (Black 1958; White & Harper 1970; Schmitt et al. 1987). Our study is consistent with these others because the average offspring of M plants was significantly larger than the average offspring of NM plants. In fact, in our study differences between offspring of M and NM plants in mass actually increased over time. At 20 days, offspring of M plants were about 98% heavier than offspring of NM plants. At 47 days, they were about 240% heavier. In some cases increasing size inequality leading to self-thinning from among suppressed individuals occurs as a consequence of asymmetric competition. Because all plants were well-watered and because N and P concentrations were actually higher in offspring of NM plants, we conclude that light was the most likely resource for which there could have been asymmetric competition.
Second, offspring of M plants were more reproductive than offspring of NM plants. Compared with offspring of NM plants, a greater proportion of offspring of M plants that survived actually reproduced. Moreover, on average each reproductive offspring of M plants also produced more capsules and more seeds per capsule than did each reproductive offspring of NM plants. This appeared to be a simple consequence of the larger average size of offspring of M plants because the relationships between size and reproductive output were not significantly different for the two offspring types. Overall there were nearly four times as many seeds produced per box by offspring of M plants as produced by offspring of NM plants.
The magnitude of the effect of maternal plant mycorrhizal infection on offspring performance under intraspecific competition is likely to be a function of density. At lower densities when competitive interactions are less intense, smaller differences between offspring of M and NM plants might develop. Indeed, at very low densities at which there are no competitive interactions differential mortality may not occur at all. However, even in that circumstance we might expect there to be differential reproduction between offspring of M and NM plants as a consequence of differential offspring vigour (Koide & Lu 1992).
In order to demonstrate the potential effect of mycorrhizal infection of maternal plants on the competitive ability of offspring we maximized the interaction between offspring types by planting seeds of M and NM plants in alternate positions in a regular grid. We certainly do not mean to imply that this arrangement could ever occur in nature. The results do suggest, however, that if an Abutilon plant were unable to develop mycorrhizas for reasons such as mutation, its offspring could be at a distinct competitive disadvantage. Parrish & Bazzaz (1985) demonstrated that an increase in the nutrient content of seeds of A. theophrasti imparted a significant interspecific competitive advantage. Offspring vigour as influenced by mycorrhizal infection of maternal plants might thus be important in determining the survival and reproduction of offspring in plant communities comprising a variety of plant species. Taken as a whole our research suggests that in addition to increasing the fecundity of the infected plant, mycorrhizal infection may also increase the survival of offspring and their subsequent reproduction, especially in competitive environments.
We acknowledge financial support from the U.S. National Science Foundation and the A.W. Mellon Foundation. We thank Tanya Bibikova, Edward Boswell, Cheryl Krazowski, Michael O’Connell, Peter Skogland and Julie Whitbeck for assistance in planting.