All gas exchange parameters were recorded using a Li-Cor 6400XT portable gas exchange system (Li-Cor, Lincoln, NE, United States) in the open air on sunny days. A 2 cm × 3 cm chamber was used, and a LED 6400-02B lamp served as the light source. Throughout the measurement, the reference infrared gas analyzer (IRGA) CO₂ concentration was maintained at 400 μmol mol^–1; the IRGA flow rate was set at 500 μmol s^–1; the chamber temperature was kept at 28°C; and the leaf-to-air vapor-pressure deficit (VPD_leaf–air) was maintained at 1.3 ± 0.2. All measurements were conducted using the youngest fully expanded leaf from 9:00 a.m. to 16:00 p.m. in April 2018.

During the measurements, the leaf was first equilibrated at a photosynthetically active photon flux density (PPFD) of 100 μmol m^–2 s^–1 until the photosynthetic rate (A) and stomatal conductance (g_s) reached a steady state. Then, the PPFD was increased to 1,500 μmol m^–2 s^–1 until 600 s of light induction for A and g_s. The data were recorded every 4 s. To estimate the response of stomatal opening to a stepwise increase in light intensity, P_50g and P_90g were calculated, which represent the time required for g_s to reach 50 and 90% of the difference between the initial and final stomatal conductance (g_si and g_sf) after switching to high light conditions. Similarly, the response rate of photosynthesis to a stepwise increase in light intensity was obtained by calculating the time required for A to reach 50 and 90% (P_50A and P_90A) of the difference between the initial and final photosynthetic rate (A_i and A_f) (Figure 1).

Carbon gain (C_g) and Carbon loss (C_l) represent the integrated amount of CO₂ uptake and loss during light induction, which were calculated as follows:

where A_t represents the transient photosynthetic rate during light induction, and A_f is the final rate at the end of light induction.

The epidermal impressions were collected from the abaxial surface of the youngest fully expanded leaves. The middle segment of leaf samples was fixed in formalin-acetic acid-alcohol (FAA) solution. The abaxial surface of the leaf was obtained by gently scratching the adaxial surface using blades. Afterward, the transparent epidermis was kept in clean water. The epidermal impressions were placed on the microscope slides and analyzed using an optical microscope equipped with a digital camera. In each treatment, five slides from three plants were used for determination. Stomatal density and number were determined using ImageJ. Five stomata were randomly selected from each slide to determine the stomatal size, which was calculated by multiplying stomatal length and width.

Once all gas exchange and stomatal morphology measurements were accomplished, one plant was sampled per pot to estimate the plant growth. The plants were then separated into leaves and stems. A leaf area meter (LI-3000, LI-COR Inc., Lincoln, NE, United States) was used to determine the total leaf area. Besides, the number of stems per plant was recorded. Finally, to determine the total dry weight of the aboveground part, leaves and stems were oven-dried at 80°C to constant weight.

One-way ANOVA and the least-significant difference (LSD) test were used to assess the measured parameters using SPSS 21.0 (SPSS for Windows, Chicago, IL, United States). Linear regression was analyzed to test the correlations between the measured parameters using SigmaPlot 12.5 (Systat Software Inc., San Jose, CA, United States).

The response of stomatal conductance to a stepwise increase in light intensity was determined by calculating the P_50g and P_90g among twelve rice genotypes (Figure 1). P_50g and P_90g showed significant variations among rice genotypes, ranging from 91 to 200.7 s and 254.5 to 469.8 s, respectively (Figures 2A,B). To estimate the effect of stomatal kinetics on the photosynthetic rate under fluctuating light, we also calculated P_50A and P_90A. Similarly, significant variations were observed among different rice genotypes in P_50A and P_90A, which ranged from 39.5 to 99.3 s and from 227.5 to 358.2 s, respectively (Figures 2C,D). Interestingly, tetraploid rice showed a significantly slower response of stomatal conductance and photosynthetic rate to fluctuating light than diploid rice, since P_90g and P_90A were higher for WH (4x) and YD6 (4x) than for WH (2x), and YD6 (2x), and it was the same case for P_50g and P_50A. Leaf carbon gain (C_g) and carbon loss (C_l) significantly varied among twelve rice genotypes during 600 s of light induction (Figures 2E,F). A significant positive correlation was observed between P_50g and P_50A, as well as between P_90g and P_90A (Figure 3).

Steady-state gas exchange is considered as an important factor affecting the dynamic response to fluctuating light. We observed significant differences in initial and final stomatal conductance (g_si and g_sf) and photosynthetic rate (A_i and A_f) among rice genotypes (Figures 4A–D and Supplementary Figure 2). Moreover, the tetraploid rice of WH (4x) and YD6 (4x) showed significantly lower g_sf and A_f than WH (2x) and YD6 (2x) (Figures 4B,D). The initial and final water use efficiency (W_i and W_f) also significantly varied among rice genotypes (Figures 4E,F). Significantly lower W_i and W_f were observed in diploid rice than in tetraploid rice. No significant correlation was observed between g_si and P_50g, as well as between g_sf and P_90g (Figures 5A,B). However, A_i was negatively correlated with P_50A, and no correlation between A_f and P_90A was observed (Figures 5C,D). Leaf _iWUE was mainly determined by stomatal conductance under fluctuating light (Supplementary Figure 3).

The rice genotypes varied significantly in stomatal size and density, ranging from 224 to 491 μm² and from 252 to 730 mm^–2, respectively (Figure 6). Moreover, significant differences were observed in stomatal size and density between diploid and tetraploid rice (Figure 6). Compared with WH (2x) and YD6 (2x), WH (4x) and YD6 (4x) exhibited significantly larger stomatal size and lower stomatal density. The stomatal density was found to have significant negative correlations with both P_50g and P_90g (Figures 7A,C). Inversely, the stomatal size was significantly positively correlated with both P_50g and P_90g (Figures 7B,D). Also, a higher density of smaller stomata could contribute to a faster photosynthetic induction and higher carbon gain (C_g) (Figures 7E,F).

The natural variations of plant growth were further explored within twelve rice genotypes, including the number of tillers, total leaf area, leaf mass per area (LMA), and biomass (Table 2). Similarly, significant variations of the abovementioned indices were observed across different rice genotypes. Moreover, compared with diploid rice, tetraploid rice showed significantly fewer tillers, a smaller total leaf area and, therefore, lower biomass. The stomatal kinetics and photosynthetic induction showed negative correlations with the plant biomass (Figure 8).

Stomatal conductance (g_s) is known to be determined by stomatal morphology and aperture. Generally, there are significant negative correlations between stomatal size and density across or within species (Franks et al., 2009; Fanourakis et al., 2015), which is also supported by our results (Figure 6). Numerous studies have reported that stomatal morphology has strong correlations with the g_s and photosynthetic rate (A) under constant light conditions (Franks and Beerling, 2009; Xiong et al., 2017; Xiong et al., 2018). Recently, some studies have reported that stomatal movement in response to environmental fluctuations is often affected by stomatal morphology (Lawson et al., 2014; Fanourakis et al., 2020); however, other studies have suggested that there is no correlation between stomatal kinetics and stomatal morphology (Eyland et al., 2021). In this study, significant variations were observed in stomatal morphology and stomatal response rate to fluctuating light among different rice genotypes (Figures 2A,B, 6). Interestingly, the rate of stomatal response to fluctuating light was significantly positively correlated with the stomatal size, while negatively correlated with the stomatal density (Figures 7A,B), which is in line with the findings of Drake et al. (2013). It is worth noting that P_50g has a stronger correlation with the stomatal morphology than P_90g, suggesting that the initial phase of stomatal response might be more likely affected by stomatal morphology (Figures 7A–D). Moreover, the tetraploid rice WH (4x) and YD6 (4x) showed a larger size and lower density of stomata and correspondingly slower stomatal response than WH (2x) and YD6 (2x), which again indicates that stomatal morphology plays an important role in regulating stomatal kinetics under fluctuating light (Figures 2A,B, 6).

Many studies have been focused on the coordination between g_s and A under fluctuating light conditions (Adachi et al., 2019; Kimura et al., 2020; Sakoda et al., 2021). After shifting to illumination from a shading environment, stomatal opening often shows a typical delay response relative to photosynthetic induction, which will result in a stomatal limitation to A (McAusland et al., 2016). Several studies have investigated the key limiting factors during photosynthetic induction, among which g_s is the main factor that limits A during light induction, and the biochemical processes only have a very short-term limiting effect at the initial phase (Kaiser et al., 2016; Adachi et al., 2019; De Souza et al., 2020; Eyland et al., 2021). However, Acevedo-Siaca et al. (2020a,b) proposed that photosynthesis is primarily limited by biochemistry, especially the activation of RuBisCo under fluctuating light. In this study, we observed significant differences in photosynthetic induction across twelve rice genotypes (Figures 2C,D). Also, we found a significant contribution of stomatal opening to photosynthetic response under a stepwise increase in light intensity, since P_50g and P_90g were positively correlated with P_50A and P_90A, respectively (Figure 3). Therefore, a higher density of smaller stomata may contribute to faster stomatal kinetics and photosynthetic induction under fluctuating light.

Many studies have been focused on the underlying mechanisms of light-induced stomatal movement, which may be triggered by the products of the photosynthetic process in guard cells or mesophyll cells, but the exact signals remain unclear (Lawson, 2009; Lawson et al., 2014; Santelia and Lawson, 2016). Drake et al. (2013) and Zhang et al. (2019) reported that higher initial and final g_s contribute to a faster g_s response to fluctuating light. A higher initial g_s at low light may reduce the initial lag (λ) and promote the initial response of stomatal conductance and photosynthetic rate to fluctuating light (Adachi et al., 2019). Differently, in this study, g_si and g_sf showed no correlation with P_50g and P_90g, respectively (Figures 5A,B). It has been widely accepted that stomatal morphology determines the potential maximum g_s under a steady state (Franks and Beerling, 2009; Franks et al., 2009). However, stomatal morphology showed no significant effect on g_sf in this study (Supplementary Figure 4), possibly because the measurement of g_sf during light induction cannot accurately reflect the potential maximum g_s.

Light is one of the most dynamic factors under natural conditions, which often results in fluctuations of gaseous exchange on the leaf surface (Durand et al., 2019). Notably, the steady-state measurement generally cannot accurately indicate leaf photosynthetic efficiency in the natural environment when considering leaf carbon uptake. Currently, several studies have reported the low correlation between steady-state and non-steady-state photosynthesis (Acevedo-Siaca et al., 2020a,b). Consistently, less correlation was observed between steady-state and non-steady-state photosynthesis in this study (Figures 5C,D). One possible reason is the trade-off between photosynthetic proteins inside leaves, including RuBisCo and RuBisCo activase content, which determines the difference between steady-state and non-steady-state photosynthesis (Acevedo-Siaca et al., 2020b). Thus, further evidence is still needed to evaluate the relationship between the steady-state and non-steady-state gaseous exchange, which will facilitate the improvement of leaf photosynthetic efficiency under natural conditions in the future.

The improvement of photosynthesis has always been a major target to increase crop yield to meet the demand of the increasing global population. However, the relationship between leaf photosynthesis and plant growth is not always predictable, since there are various confounding factors arising from plant growth, developmental dynamics, and complex growing environments (Wu et al., 2019). Fluctuating light is a common factor affecting plant carbon uptake under natural conditions (Durand et al., 2019). In this study, we estimated the differences in leaf gas exchange in response to fluctuating light across twelve rice genotypes and the influence on plant biomass. As a result, stomatal kinetics and photosynthetic induction showed negative correlations with the plant biomass (Figure 8). Faster stomatal kinetics contributes to a higher photosynthetic rate under fluctuating light and, thereafter, higher biomass accumulation, which is in line with the results reported by Kimura et al. (2020). These results again suggest that stomatal morphology plays an important role in regulating leaf photosynthetic induction and plant biomass accumulation under dynamic environmental conditions.

Stomata are micropores composed of pairs of guard cells on the epidermis of leaves, which control the balance of CO₂ uptake for photosynthesis and water loss via transpiration. Low g_s to water vapor can conserve water by limiting CO₂ uptake for A. With a stepwise increase in light intensity, the g_s and A displayed asynchronous responses, as stomatal kinetics are often a magnitude slower than photosynthetic induction, which will result in at least a 20% decrease in _iWUE (Lawson et al., 2014). This asynchronous response was also observed in this study. During the initial phase of induction, photosynthesis was mainly limited by biochemistry, and the g_s was higher than needed, resulting in a higher C_i and lower _iWUE (Supplementary Figures 2C,D). Also, we observed a significant negative correlation between steady-state water use efficiency (W_i and W_f) and stomatal conductance (g_si and g_sf), which again indicates that excessively higher g_s will decrease the _iWUE (Supplementary Figures 3A,C).