All the plant experiments were carried out in accordance with relevant guidelines. The plant materials used in this study were identified by Dr. Satoshi Kakishima in National Museum of Nature and Sciences.
Study sites. The study sites were located in the subtropical forests on Okinawa Island on the Ryukyu Islands of Japan. The population of Strobilanthes flexicaulis Hayata (26° 38' N, 127° 55' E, 280 m a.s.l.) and the population of Strobilanthes tashiroi Hayata (26° 48' N, 128° 16' E, 365 m a.s.l.) were located on Mt. Yae-dake and Mt. Nishimei-dake, respectively. The S. tashiroi population was located at approximately 39 km northeast of the S. flexicaulis population. S. flexicaulis is a monocarpic, perennial shrub and has the periodic mass flowering of every six years, whereas S. tashiroi is a polycarpic perennial herb and has a non-masting phenology [18]. Almost all S. flexicaulis plants flowered in autumn and winter of 2015, and the seed germination and seedling growth started in spring 2016. The seedling density under the parent plants was 780 ± 324 m-2 (mean ± 1SD) in June 2016, showing high intraspecific competition following germination in the monocarpic S. flexicaulis. Their seeds do not have intrinsic dormancy after dropping to the ground. In S. flexicaulis, all plants examined in summer in 2019 and in winter in 2020 were thus 3 years old.
Plant materials. Phylogenetic analysis based on the chloroplast and nuclear DNA sequences of the small Strobilanthes group within the Ryukyu and Taiwan Islands, has been published. According to the results, S. flexicaulis and S. tashiroi are closely related species. Furthermore, their evolutionary processes have been estimated as follows: polycarpic perennial phenology via monocarpic perennial phenology, to monocarpic mass/synchronous flowering phenology within the local area19. Therefore, it is predicted that the periodical phenology of S. flexicaulis has locally evolved from the polycarpic phenology of S. tashiroi19.
Climatological measurements. The air temperature at approximately 1 m high was measured at every 1 hr. with a thermistor sensor (TER-51i, T&D Co Ltd, Nagano, Japan) at just near the S. flexicaulis population from 2017 to 2020. From 2017 to 2020, the mean annual precipitation was 2415 mm at the closest climatological station in Nago (observation by the Japan Meteorological Agency). No snow accumulation was not found in the understory.
Photosynthetic measurements. The photosynthetic light-response curves and water-vapor stomatal conductance were measured in the fully expanded young leaves on Sep. 23 and Sep. 24 (summer) in 2019 and on Jan. 24 and 25 (winter) in 2020 with a portable open gas exchange system (LI-6400; LI-COR Inc., Lincoln, NE, USA) in the field. The measurements were conducted with 400 µmol mol− 1 CO2 in the inlet gas stream. The exposed photon flux density (PFD) was stepwise decreased from 1500, 1000, 700, 400, 300, 200, 100, 70, 50, 30, 20, 10, 7, 4 and 0 µmol m− 2 s− 1 with red-blue light emitting diodes. The relative humidity (RH) in the outlet gas stream was adjusted to the ambient air RH. All leaf gas exchange measurements were conducted before noon to avoid the midday stomatal closure. In each season, the fully expanded young leaves of 7–8 individual plants from each species were used for the measurements.
The initial slope (Φ), the light compensation point (LCP) and dark respiration rates (Rd) were calculated for each leaf from the linear regression between low PFDs (10, 7, 4 and 0 µmol m− 2 s− 1) and the corresponding net assimilation rates (An) (r2 = 0.999 − 0.858). Light-response curves in net assimilation rates (An) were calculated by using the following equation:
$${A}{n}=\frac{\varPhi I+{A}{max}-\sqrt{{\left(\varPhi I+{A}{max}\right)}^{2}-4 \theta \varPhi I {A}{max}}}{2 \theta }-{R}_{d}$$
where Amax are the maximum gross assimilation rates, I is the PFD, and θ is the curvature. When θ = 0, the fitting curve is converted into rectangular hyperbola equations; when θ = 1, it is converted into Blackman equations. The values of θ in S. flexicaulis were 0.805 and 0.678 in winter and summer, respectively; those in S. tashiroi were 0.669 and 0.701 in winter and summer, respectively.
Dark respiration measurements. After the photosynthetic measurements, we collected seven individual plants for which the top canopies were not directly covered by the other plants, in each season or each species. Their stem heights were 113.5 ± 9.0 and 54.3 ± 16.9 cm (mean ± 1 SD) for S. flexicaulis and S. tashiroi, respectively. Their stem diameters near the ground were 9.0 ± 2.0 and 4.8 ± 1.1 mm for the monocarpic S. flexicaulis and the polycarpic S. tashiroi, respectively.
We separated the individual plants into three parts (leaves, stems, and roots) with scissors, and immediately measured their dark respiration rates under the field conditions. The plant parts were put into a closed plastic box with a small fan (1.3 L, 5 L or 14.5 L in volume; the size of the plastic box was selected according to the volume or length of the samples), and then the box was covered with black cloth. To measure the air temperature, copper-constantan thermocouples were placed in the box together with the plant parts. Here, we assumed that the plant temperature was in equilibrium with the air temperature in the box. We measured the increasing rates in air CO2 concentration in the box for approximately 3–5 min, using a nondispersive infrared CO2 gas analyzer (GMP343, Vaisala Inc., Helsinki, Finland) connected with a data logger (GL240, Graphtec Co. Ltd., Yokohama, Japan). From the CO2 increasing rates and the box volume, we calculated the dark respiration rates. After the field measurements, the plant samples were dried at 60°C for at least three days and then the dry mass was weighed. Dry mass-based respiration rates were calculated. Furthermore, to evaluate metabolic activity in cells at a given temperature, the respiration rates were standardized at 19°C and 28°C in both seasons by using Q10 = 23, 15. The chosen values of 19°C and 28°C were the mean box temperatures in winter and summer, respectively.
Allocations of dry matter, whole-plant respiration and N in plants. To evaluate the interspecific variations in plant form, the plant samples used for respiration measurements were used to calculate dry matter allocation. Leaf areas of individual plants were calculated from the leaf dry mass in individual plants and the mean values of leaf mass per area (LMA) in each species. The standardized whole-plant respiration rates were calculated from the mean dry matter allocation in each species and the mean dark respiration rates of each tissue in each season.
The N concentrations in the plant tissues were also examined. Each tissue in individual plants was milled into powder with a cutter mill (ABSOLUTE 3, Osaka Chemical Co. Ltd., Osaka, Japan). The N concentrations were measured with an NC analyzer (Vario Max CN, Elementar Analysensysteme GmbH Co. Ltd., Hanau, Germany). The standardized N and respiration allocations among tissues in the plants were calculated from the mean dry matter allocation in each species and the mean values of N and respiration in each tissue and each season.
Statistics. Statistical analyses were conducted with R Ver. 3.0.2 (R Development Core Team, R Foundation for Statistical Computing, Vienna, Austria). To evaluate seasonal acclimation, the significant differences between winter and summer were examined with ANOVA at P < 0.05 (electronic supplementary material, table S1). The significant differences between species were also examined with ANOVA at P < 0.05 (electronic supplementary material, table S2). The mean (1 SD) values and all statistical results are shown in the supplementary materials.