Shoot density of Miscanthus sinensis populations in different habitats and their maintenance mechanisms in relation to shoot growth

Abstract

How perennial grass populations are maintained in different climates is poorly understood at the level of individual shoots (ramets). During the years 1982–1987 and 1991–1993, measurements of shoot dynamics and growth in populations of a clonal grass, Miscanthus sinensis, were made at two sites in Japan that differed by approximately 5 °C in mean temperature. While annual shoot births were very stable during the period 1982–1987 at both sites, the number of flowering shoots fluctuated cyclically every year. The clonal propagation of shoots was size-independent, whereas the reproduction (flowering) of shoots was size-dependent and negatively affected their own offspring size. Shoot size negatively affected the overwintering of shoots. In the warm climate with a long growing period (9 months), both early-emerging shoots and the subsequent high order tillering shoots developed in large numbers. In the cool climate with a short growing period (6 months), more than half of the annual births occurred in August and September. Nevertheless, average longevity and wintering competency of shoots were not greatly different between the two populations. In response to a warmer climate, tillerings started earlier. This appeared to increase the total number of new shoots that would die within the year; nevertheless, the shoot densities remained much higher because a longer growing season would increase the number of high order tillerings. There was thus a trade-off between the annual survival ratio of new shoots and the number of annual shoot births.

Appendices

Appendix S1

Shoot replacement for maintaining the populations in different climates

At the S and F plots, each of the shoot populations exhibited similar patterns of density changes every year (Fig. 1). The monthly net number of new or old shoots over the 5-year period was expressed as the mean of a single year, and for convenience of comparison it was normalized to the mean number of old shoots in early June: for example, the mean number of old shoots at the S plot in early June, 779 ± 111 (mean ± SD; n = 5), was reckoned at 1, being taken as a base of comparison (Fig. S3). Over a period of 5 months from late May–early June to late October–mid-November, the survival ratio of old shoots (defined as the ratio of the final number to initial number) in an average year was 0.66 ± 0.07 (n = 5) at the S plot and 0.73 ± 0.09 (n = 5) at the F plot, being significantly similar between the two plots (Mann–Whitney test, P = 0.310). At the S plot, most of the old shoots withered until January, except for a tiny minority of shoots in 1982–1984 (0.8% of initial number), whereas at the F plot, all the old shoots died by December.

The mean numbers of overwintering shoots at the S and F plots peaked at 970 ± 128 (n = 5) in December and at 512 ± 50 (n = 5) in May, respectively (Fig. S3). To maintain a stable population size at S plot, new shoots needed to be at least 24% greater than the number of old shoots in early June, whereas at the F plot, the number of new shoots needed to be only about the same as the number of old shoots in early June (Figs. 2, S3).

Appendix S2

Environmental factors for flowering

On the whole, the growing period of a plant falls into two main phases, the vegetative and reproductive phases. For Miscanthus sinensis at the Shizuoka and Fujinomiya sites, which have about the same photoperiods, flower buds began to form inside the apex of shoots in late July at both sites (Kobayashi, pers. obs.). At this time, the mean day length was half an hour shorter than the maximum in June (14.5 h). The mean temperatures differed by several degrees between the two sites, but the reproductive phases began at the same time. Therefore, the onset of reproductive phase appears to be regulated by a physiological signal such as the photoperiod rather than the local temperature, as has been observed in most plants (e.g. Bradley et al. 1999; Yoshie 2007). Vernalization may not act as a trigger for such a reproductive factor because some of early-emerging shoots at the S plot flowered within the year (Fig. 4).

It is known that flowering frequencies were positively correlated with rainfall (Rabinowitz et al. 1989) or temperature (Suvensson et al. 1993; Carlsson and Callaghan 1994). However, for M. sinensis in August-October, flowering was not correlated with rainfall or temperature at either the Shizuoka site (n = 5, P = 0.753 or 0.946) or the Fujinomiya site (n = 6, P = 0.800 or 0.879).

Appendix S3

The minimal growing season

The vegetative growing period before the occurrence of tillerings was called a pre-propagative phase for primary tillerings. The duration of the pre-propagative phase for primary tillerings was 3 months at longest, and the time occupied mainly by primary tillerings would be about 2 months between June and July at the Shizuoka site (Kobayashi 1981) or between August and September at the Fujinomiya site. Consequently, the minimal growing season needed to form primary tillers (‘critical duration’) is about 5 months. This time is needed to maintain the potential minimal level of shoot densities by primary tillerings only in a cooler climate. But in an even cooler climate, where the growing season of M. sinensis is shorter than the critical duration, shoots may be unable to obtain sufficient EAT to produce primary tillerings and cause a shortage of resources for primary tillerings by the undergrown parent shoots. Such an insufficient resource is likely to cause a loss in branching intensity and frequency, which should bring about even lower shoot density, as seen also in C. calceolus genets in shady conditions (Kull 1995).