Our results were consistent with our first prediction of the differences in effects of nutrient distribution pattern on plant size between ranks. Nutrient distribution pattern significantly affected the sizes of higher-ranked plants but not those of the smallest plants, i.e., the sixth-ranked. The growth of the smallest plants would be more limited rather than their neighbors independently of the nutrient distribution pattern. This suggests that the effects of nutrient distribution pattern on the size of I. tricolor are likely to be small, or even negligible, if the plant growth is restricted. In contrast, different growth rates could cause large differences in the sizes of higher-ranked plants between heterogeneous and homogeneous environments because a size hierarchy develops not only by competition but also by differences in size-dependent growth rate (Hutchings 1997). As far as we know, this is the first study to show that the effect of nutrient distribution pattern on the size of higher-ranked plants is more prominent than on lower-ranked plants under the control of aboveground competition.
Moreover, given that the limited plant growth in our study would result from a shortage of nutrient acquisition as well as the negative effect of aboveground competition, differences in plant growth between ranks could occur in response to differences in root response within a nutrient-rich patch. In our experiment, the plants in the pots were equally located on the border of the nutrient-rich patch and their roots were able to reach the patch immediately after they had been transplanted. Thus, differences in plant growth between ranks may be caused by differences in the plastic responses of the roots after arrival at the patch (Hodge 2004), rather than by differences in the exploring of the patch, such as differences in the distance from the plant to the nutrient-rich patch (e.g., Hutchings et al. 2003) or by variation in the growth or size of roots among plants (e.g., Weiner
1990; Campbell et al. 1991; Wijesinghe et al. 2001).
Our results were consistent with our first prediction of differences in plant sizes between heterogeneous and homogeneous treatments (i.e., the 75:25 and the 50:50 treatment). However, mean plant size in the 100:0 treatment (one of the heterogeneous treatments) was the smallest among nutrient distribution patterns (Fig. 3-1). Significant selective root placement to a nutrient-rich patch in the 100:0 treatment (Fig. 3-2) might have caused severe competition for nutrients, thus decreasing the rate of nutrient intake per plant under this treatment. This phenomenon, “the tragedy of the commons”, means that increasing root placement into the nutrient-rich patch of competing individuals would decrease the nutrient availability to all (Gersani et al. 2001). In contrast, the roots within a patch could have been relatively dispersed in the 75:25 treatment, since nutrients were provided both inside and outside the patch. Although the root dispersal could be costly, no obvious costs were detected in our study. Therefore, efficient nutrient acquisition and growth increment of plants could have been achieved. It should be noted, however, that we still do not know when this scenario is applicable, and thus the studies are necessary that pay attention to the presence or absence of belowground competition (e.g., Hodge et al. 1999; Robinson et al. 1999; Kembel and Cahill 2005).
Aboveground competition negatively affected the size of plants of all ranks. Thus, our second prediction—that the effects of aboveground competition on plant growth would be significant in lower-ranked plants—was supported. However, a significant reduction in the sizes of higher-ranked plants in the presence of aboveground competition was also observed. The leaves of higher-ranked plants are generally able to explore the highest positions of the canopy to capture strong light, whereas those of lower-ranked plants have to grow in the shade of their taller neighbors (Ford 1975; Anten and Hirose 1998). This advantage of higher-ranked plants may depend on the arrangement of leaves along the stem of the plant. Most leaves of I. tricolor remained during the experimental period. Mutual shading of leaves at the lowest parts of the stem would have had negative effects, even on the size of the first-ranked plants, though lower-ranked plants would be
likely to experience more severe aboveground competition than higher-ranked plants (e.g., Anten and Hirose 1998).
The significant interaction between the effects of nutrient distribution pattern and aboveground competition in MANOVA was consistent with the third prediction, i.e., that the effects of nutrient distribution and aboveground competition on plant size would interact with each other.
Low light availability by shading might negatively affect selective root placement under heterogeneous conditions (Bilbrough and Caldwell 1995) and reduce rates of nutrient uptake (Jackson and Caldwell 1992). As a result of these changes in responses, the small amounts of nutrients acquired would not cause any differences in size of the lowest-ranked plants between nutrient distribution patterns. Nutrient distribution pattern did not affect the sizes of the lowest-ranked plants but did affect those of the higher-ranked plants, a finding that is consistent with that of Maestre and Reynolds (2006a). They suggested that size differences between heterogeneous and homogeneous conditions could be reduced among lower-ranked plants because of aboveground competition. The same scenario was applicable to the study by Nagashima et al. (2003). However, these experiments did not control for the effects of competition or for the separate effects of above- and belowground competition. Our experiment, in contrast, by eliminating aboveground competition, allowed plants to compete only with their belowground parts. Thus, our results showed that the effects of nutrient distribution pattern and aboveground competition on plant size differed according to size rank.
In conclusion, the spatial patterns of nutrient distribution affected the size of especially higher-ranked plants in a population but not the size of the lowest-ranked plants at late growth stages because the plant sizes and their hierarchy is the results of aboveground competition and the size-dependency of growth rate throughout the growth dynamics. Nutrient distribution pattern would thus interact with competition, and, consequently, affect size structure in a population.
Table 3-1 Multivariate analysis of variance (a) and the corresponding univariate analyses of variance (b) of the effects of nutrient distribution pattern and aboveground competition on plant size in each rank.
(a)
Source Wilk’s lambda df F P
Nutrient distribution pattern 0.711 12,120 1.857 0.047 Aboveground competition 0.779 6,60 2.844 0.017
Interaction 0.699 12,120 1.963 0.034
(b)
Nutrient distribution pattern Aboveground competition
Rank N df MS F P df MS F P
First 71 2 3.620 5.263 0.008 1 3.851 5.598 0.021
Second 71 2 1.608 3.521 0.035 1 5.524 12.094 <0.001
Third 71 2 2.097 5.423 0.007 1 5.056 13.078 <0.001
ln(Fourth) 71 2 0.428 3.463 0.037 1 0.773 6.252 0.015
Fifth 71 2 1.243 3.471 0.037 1 2.236 6.246 0.015
Sixth 71 2 0.748 2.057 0.137 1 2.354 6.473 0.014
The first-ranked plant was the largest plant and the sixth-ranked plant was the smallest. Because of heteroscedasticity, data on the fourth-ranked plants were log-transformed (ln) before analysis. A few missing values for sixth-ranked plants were estimated by the method of Zar (1999); for details see Materials and methods. Wilk’s lambda, number of individuals (N), degrees of freedom (df), mean squares (MS), F-statistic (F), and significance (P) are reported.
Figure 3-1 Mean (+ SE) sizes of the first- to sixth-ranked plants of Ipomoea tricolor in heterogeneous (100:0 and 75:25) and homogeneous (50:50) fertilizer treatments. Six individual plants in a pot were ranked according to their aboveground biomass. The same letter in each figure indicates no significant difference in plant size between treatments at P < 0.05, following Holm multiple-means comparison tests.
Figure 3-2 Mean (± SE) of ratio of root biomass inside a patch to root biomass outside the patch.
Plants were grown with neighbors in the presence of both above- and belowground competition or belowground competition alone and under three different nutrient distribution patterns: 100:0, 75:25, or 50:50. The same letter indicates no significant difference in ratio between treatments at P < 0.05, following Holm multiple-means comparison tests.