Coupled influence of precipitation regimes and seedling emergence time on the reproductive strategy in Chloris virgata

Experimental design and treatment

The experiment was conducted at Jilin Jianzhu University, Jilin Province, China in 2016. C. virgata seeds were obtained from the Greenhouse of Jilin Jianzhu University, Jilin Province, China (44°33N, 123°82 31E) in 2015. Twenty seven PVC pipes, 25 cm diameter and 50 cm height, were filled with saline-sodic soils at different sowing dates in 2016. The soil was collected from the surface layer (0–20 cm) of a grassland near the Northeast Normal University field experiment station, in Changling, Jilin Province, China (123°44 0E, 44°44 0N, 167 m above sea level). The soil total nitrogen was 6.9%, organic carbon 0.4%, pH 8.6, electrical conductivity 91 µs cm⁻¹ and field capacity of 200 g kg⁻¹. And the maximum/minimum temperatures were 32/16 °C (day/night) during the treatment period from 15 June to 15 September. The soil-filled pipes were all buried in the soil and the top edge of each pipe and the soil surface were kept at the same level. Twenty seeds were sown on 15 May (T₁), 15 June (T₂) and 15 July (T₃), respectively, to ensure three seedling emergence times (SETs). Most seedlings emerged 5 d after sowing and 10 uniform and robust seedlings were selected per pipe 10 d after sowing. Three treatments were used to simulate local precipitation regimes (PRs), corresponding to 120 (L: low, soil moisture was 6.0%), 200 (T: typical, soil moisture was 11.0%), and 280 (H: high, soil moisture was 19.0%) mm precipitation, and these were created by adding water every 5 d to the pipes from 30 d after seedling emergence (see Wang et al., 2018). The typical precipitation regime was based on the average amount and frequency of the local region’s precipitation (June–October) over the past 21 years in the Songnen grassland, where C. virgata is widely distributed. The low and high precipitation regimes were 40% lower and higher than the typical precipitation regime, based on data from the Songnen grassland in wet and dry years, respectively. For example, the plants were watered 12.5 mm 5 d intervals (The number of precipitation was 16 from 15 June to 15 September because most precipitations occurred in this period) under typical precipitation regime. Then the precipitation was calculated according to the precipitation depth and cross section of the pipe. There were 3 replicates for each precipitation level at each seedling emergence time, giving 30 individuals per treatment for determining tiller traits and seed production. There was a total of 27 pipes across all treatments for the three seedling emergence times and the pipes were arranged outdoors in a replicated randomized block design under a rain shelter made from transparent 0.2 mm polyethylene sheeting on 15th May, 2016.

Data collection

The plants from each treatment at each seedling emergence time were harvested in mid-September, when all C. virgata individuals were fully mature. Some pipes were damaged by locusts during the course of the experiment and were excluded from the analyses, giving 22–30 individuals. The numbers of reproductive and vegetative tillers of C. virgata were recorded for each individual in each pipe. One first-order tiller was selected randomly from each individual in each pipe. Then the numbers of mature seeds and unripe seeds were counted to obtain seed number. The number of left spikes from each plant were recorded and the spikes were kept in individual paper bags. Fifty seeds were selected randomly from each plant and dried at 65 °C for 48 h and weighed to determine mean seed size. The remaining seeds from each plant were dried at 65 °C for 48 h and weighed to obtain seed yield.

Data analysis

All data were tested for the normality and homogeneity and subjected to analyze by SPSS statistical software (version 17.0, SPSS Inc., Chicago, IL). One-way ANOVA was used to analyze variables such as tiller number, proportion of reproductive tiller number to total tiller number, seed number, spike number, seed size and seed yield. Two-way ANOVA was also used to analyze all data under different precipitation regimes over seedling emergence times. Differences under different precipitation regimes and seedling emergence times were analyzed using the least significant difference (l.s.d.) test.

Vegetative and reproductive tiller growth

The ANOVA results showed that seedling emergence time individually had a significant effect on vegetative tiller number, and that precipitation regime and seedling emergence time individually had significant effects on reproductive tiller number. However, their interactions had no effect on vegetative tiller number or reproductive tiller number (Table 1).

Growth and reproduction in the C. virgata plants responded differently to different levels of precipitation and seedling emergence times (Fig. 1). The vegetative tiller number was similar under different precipitation regimes at T₁ and T₂, while significant variations were found among the three precipitation regimes at T₃ (Fig. 1A). The vegetative tiller number varied significantly within increasing precipitation and the vegetative tiller number was significantly higher under high precipitation than under low precipitation at T₃. There was no significant change in reproductive tiller number with increasing precipitation at T₁ (Fig. 1B). At T₂ the reproductive tiller number was similar between low and typical precipitation but was significantly higher under high precipitation. At T₃ the reproductive tiller number was similar under typical and high precipitation treatments and was significantly higher than under low precipitation. Vegetative tiller number decreased significantly from T₁ to T₃ under all the precipitation regimes. The highest reproductive tiller numbers occurred at T₁ under the low precipitation regime. The reproductive tiller numbers were significantly higher at T₁ than at T₂ under typical precipitation. Furthermore, changes in seedling emergence time did not affect reproductive tiller number under the high precipitation regime (Fig. 1). The proportion of reproductive tiller number to total tiller number were similar between T₁ and T₂, while were significantly lower than that at T₃ (Fig. 2).

Seed production

The ANOVA results showed that precipitation regime and seedling emergence time individually had significant effects on seed number, seed size, spike number and seed yield (Table 1). Their interactions had an effect on seed size and seed yield. However, their interactions had no effect on seed number and spike number (Table 1).

The seed number and seed size of the C. virgata plants were significantly different under the various precipitation regimes and seedling emergence times (Fig. 3). Seed number remained unchanged among the three precipitation regimes at T₂ and T₃, whereas seed number was significantly lower under low precipitation than under the typical and high precipitation regimes at T₁. Within the same precipitation treatment seed numbers were similar at T₁ and T₂ and were significantly higher than at T₃ (Fig. 3A). Seed size was unchanged from T₁ to T₃ under typical and high precipitation and was higher than under low precipitation at all sowing times. Within the same precipitation treatment seed size remained unchanged at T₁ and T₃ and was higher than at T₂ (Fig. 3B).

There was a significant difference in spike number among different levels of precipitation at different seedling emergence times (Table 2). Spike number increased from the low to typical precipitation regimes but under high precipitation there was no further significant change at T₁. There was a significant difference between low and high precipitation at T₂. Spike numbers increased significantly from low to high precipitation at T₃. Spike numbers remained unchanged from T₁ to T₃ under low and high precipitation. While the spike numbers under the typical precipitation regime decreased substantially from T₁ to T₂. There was no significant difference between T₁ and T₃, or T₂ and T₃ under typical precipitation.

Seed yield was significantly higher under the typical precipitation regime than under low precipitation at T₁ and T₃, but there was no difference between typical and high precipitation for these two emergence times (Fig. 4). At T₂ there was no significant difference in seed yield between low and typical precipitation but there was a marked increase under high precipitation. The highest yield values occurred at T₁ under all precipitation regimes.

Vegetative and reproductive tiller growth

Plants select different reproductive strategies according to their growing environment, and the reproductive strategy selected in a given environment should maximize fitness (Zhang et al., 2018). However, resources are the primary factor determining how plants establish their reproductive machinery (Weiner et al., 2009). Reproductive tiller number is an important characteristic influencing seed yield in many plants (Peng et al., 2015). Whether a bud develops into a reproductive or a vegetative tiller is based on the environment and the evolutionary history of the population (Warringa & Kreuzer, 1996). Our results showed that the trends in tiller number in C. virgata under different precipitation regimes were similar at T₁ and T₂ but were different at T₃. The trends were similar with the results in 2018 when sowed at the early date. It can be concluded that water might be one of the limiting resource for tiller number at T₃. Furthermore, the proportional allocation of reproductive tiller number to total tiller number was significantly higher at T₃ than at T₁ and T₂ (Fig. 2). High temperature improved the proportion of reproductive tiller number to total tiller number and the results were consistent with those from previous study (Escalada & Plucknett, 1975). Plant morphology was affected by temperature and daylength, while many plants did not react to critical periods of photoperiod unless their thermal requirements were met. C. virgata produced tiller 10 days later after sowing. The temperature was higher when C. virgata produced tillers at T₃ than that at T₁ and T₂. Delayed sowing and low precipitation reduced C. virgata growth (total tiller number), while maximized it’s reproduction and restoration potential. This strategy has also evolved in monocarpic species and according to optimal allocation theory, such plants should allocate more resources to reproductive parts to improve seed production in unfavorable environments (Ellner, 1987).

Seed production

Seed production are related to plant adaptations to their growing environment. For example, the maternal nutrient environment affects seed size, seed yield and the competitive ability of progeny (Gómez, 2004). In our study, seed size was smaller at T₂ compared to the other two emergence times. The larger seed size at T₁ and T₃ may indicate higher seed quality because large seed size has been associated with higher seed germination ability and seedling survival as reported by other authors (Cowell & Doyle, 1993) and in our earlier work (Wang et al., 2018). This also suggests a management method to produce high quality seeds from C. virgata.

Moreover, our results showed that when C. virgata was delayed in emergence beyond 30 d or 60 d it produced fewer seeds. It has been demonstrated that sowing date was one of the critical factors for productivity of winter wheat (Triticum aestivum L.) due to different amounts of soil water storage at different sowing times (Ren et al., 2019). Further, Awan & Chauhan (2016) have reported that a delay in emergence would result in fewer seeds returning to the seed bank, which has important implications for vegetation succession and restoration (Shang et al., 2016). However, seed yields were similar among the three water supply treatments and the first-order spike number was one of the most important factors in determining seed yield in our former work (Wang et al., 2018). Seed yields in the current work were higher in typical and high precipitation than under low precipitation. The pattern of seed yield was not consistent with our former work in 2018, and the reason for this might be the different plant densities in the two experiments. However, the first-order spike number (reproductive tiller number) was higher under low precipitation than under typical and high precipitation in these earlier experiments under a density of five plants per pot, whereas the first-order spike number was significantly higher under typical and high precipitation compared to low precipitation with a density of ten plants per pot in the current work (Fig. 1). This indicated the existence of different reproductive strategies under different plant densities. Here, reproductive tiller number was demonstrated to be an important component for C. virgata seed yield, and future work should explore the effect of plant densities on growth and reproductive characteristics.