HAN Yong-qiang , WEN Ji-hui PENG Zhao-pu ZHANG De-yong HOU Mao-lin,

1 Institute of Plant Protection, Hunan Academy of Agricultural Sciences, Changsha 410125, P.R.China

2 State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing 100193, P.R.China

3 Southern Regional Collaborative Innovation Center for Grain and Oil Crops in China, Changsha 410128, P.R.China

Abstract Rice is one of the most important staple foods for the world population, but it is attacked by a number of destructive pests. While evidence from greenhouse and laboratory tests has shown that silicon (Si) amendment can confer enhanced resistance to pests in rice, few studies have directly demonstrated the Si-mediated protection from pests in a field situation.In this study, field plots with silicon amendments at 0, 75, 150 and 300 kg SiO2 ha–1 in early- and late-season rice were employed to evaluate the effects of silicon amendment on the occurrence of major insect pests and diseases and rice yield.Compared with the control plots without silicon amendment, plant damage by stem borer and leaf folder and population size of planthopper were significantly lower in three to five of the seven monitoring observations in each season in the plots amended with 300 kg SiO2 ha–1. The disease index of rice blast in the early-season rice was lower in the plots amended with Si at 300 kg SiO2 ha–1 than in the control plots, while Si protection from rice blast in the late-season rice and from rice sheath blight in the early-season rice were not apparent. An insignificant increase of rice yield by 16.4% (604 kg ha–1) was observed in the plots amended with 300 kg SiO2 ha–1 over the control plots. Our results indicate that Si amendment at 300 kg SiO2 ha–1 can provide substantial protection from some of the rice pests under field conditions. These findings support the recommendation of silicon amendment as a key component of integrated management of rice pests.

Keywords: silicon, rice, insect pest, disease, yield, plant resistance

1. lntroduction

Rice is a staple food for nearly half of the global population(FAO 2004). However, this crop is subject to attack from a number of devastating insect pests and diseases. Three groups of rice insect pests are generally considered the mostdestructive (Pathak and Khan 1994), i.e., the rice stem borer(RSB, Chilo suppressalis Walker), the rice leaf folder (RLF,Cnaphalocrocis medinalis Guenée), and the planthoppers(white-backed planthopper (WBPH), Sogatella furcifera Horváth, and brown planthopper (BPH), Nilaparvata lugens Stål). The rice crop is also damaged by some destructive diseases like blast (Magnaporthe oryzae) and sheath blight(Rhizoctonia solani Kühn).

For many years, rice pests have been controlled by the indiscriminate application of chemical pesticides, which has caused severe environmental contamination, reduction of natural enemy populations, and resurgence of pests (Wang et al. 2014). Therefore, developing safe and effective alternative methods to control rice pests is essential. Crop resistance and management are currently the dominant tactics for such purposes (Hou and Han 2010). Crop resistance is one of the principal components in integrated pest management (IPM) strategies, which provides a longterm and sustainable crop protection option for subsistence farmers at a low cost. Crop resistance can be achieved through breeding programs and can be regulated through crop management.

Silicon (Si) amendment is a crop management practice that can confer enhanced crop resistance. Evidence in the last decade has found that Si plays an important role in plant defense against biotic stress (Reynolds et al. 2016).Si has been found to improve plant defenses through two primary modes of action. The first is the mechanical defense derived from the deposition of inorganic amorphous silicon oxide (SiO2) phytoliths in the epidermis of the plant tissues(Ma 2004; Hartley et al. 2015), which renders the plant surface less vulnerable to fungal pathogens (Wang et al.2017) and tougher for herbivores to feed on and digest(Massey and Hartley 2009; Reynolds et al. 2016). The second is enhancement of induced defenses through greater mobilization and activation of enzymes such as polyphenol oxidase and trypsin protease inhibitor (Ye et al. 2013; Han et al. 2016; Yang et al. 2017a), higher callose deposition(Yang et al. 2018), or increased release of plant volatiles that attract the natural enemies of the attacking insect herbivores(Kvedaras et al. 2010; Liu et al. 2017). The phytohormone pathways are involved in the foregoing mechanisms of Simediated plant resistance (Ye et al. 2013).

Rice is a Si accumulator (Liang et al. 2006; Ma and Yamaji 2006) and harbors Si concentrations above 10% of shoot dry weight (Yamamoto et al. 2012). Si amendment benefits soil health and rice production (Savant et al. 1997),especially in regions where soils are deficient in plantavailable Si. As in China, where half of the paddy fields are Si deficient (Dai et al. 2005), Si deficiency in paddy fields has become a worldwide constraint for sustainable rice production (Liu et al. 2014). Recently, Si amendment has been proven to impart substantial rice plant resistance to pests, including the major insect pests WBPH (Yang et al. 2014), BPH (He et al. 2015; Wu et al. 2017; Yang et al. 2017a, b), RSB (Hou and Han 2010), and RLF (Ye et al. 2013; Han et al. 2015, 2016, 2017); and diseases,such as rice blast (Seebold et al. 2000; Kim et al. 2002;Cai et al. 2008; Sun et al. 2010), sheath blight (Rodrigues et al. 2001), bacterial blight (Song et al. 2016), and brownspot (Ning et al. 2014). However, these results have been obtained exclusively from trials with potted rice plants grown in soil or nutrient solutions under greenhouse conditions.It is essential to demonstrate whether such Si-mediated protection is also in place in field conditions.

In this study, we assessed the effects of Si amendment to rice plants on the occurrence of major insect pests and diseases in field conditions. Our specific objectives were to: 1) evaluate the degree of protection afforded by Si amendment against the principal rice insect pests and diseases in a field design, as indicated by their population sizes and/or damage; and 2) measure rice yield response to different levels of Si amendment.

2. Materials and methods

2.1. Experimental site

The experiment was conducted from April to November 2014 at Guilin Experiment Station for Crop Pests (25°36´00´´N and 110°41´24´´E), Ministry of Agriculture, China. The region has an annual average temperature of 17.8°C and precipitation of 1 842 mm. Paddy fields at the station have loam soil, with pH=5.7, organic C=2.19%, total N=1.58 g kg–1, available P=3.22 mg kg–1, available SiO2=211.31 mg kg–1, and available K=98.16 mg kg–1(Han et al. 2015).

2.2. Experimental plots

Si was soil-applied as calcium silicate (soluble SiO2≥30%,Shanxi Fubang Fertilizer Co., Ltd., China) at 0 (control), 75,150 or 300 kg SiO2ha–1to field plots of 8 m×20 m with three replications, resulting in twelve plots arranged in complete randomized design. The plots were separated by ridges and individually watered and drained as necessary for rice growth.

Fertilization in the plots was as described in Hou and Han (2010). Briefly, all the plots received 270 kg ha–1of N, applied 3 days before transplanting, and at the tillering,heading and milk stages, at a ratio of 4:3:2:1, respectively;60 kg ha–1K2O was supplied 3 days before transplanting and at heading stage at a ratio of 2:1; 60 kg ha–1of P2O5and the above-mentioned calcium silicate were applied 3 days before transplanting. The experiments were conducted on two seasons; for the early-season, thirty-day old seedlings(var. Liangyou Y3218, with duration of 137 days) were transplanted on May 20 and for the late-season, twentyfive-day old seedlings (var. Teyou 808, with duration of 130 days) were transplanted on August 24. After harvest of the early-season rice, the plots were maintained and the Si treatment in each plot for the late-season was the same as that in the early-season. Hill spacing was 30 cm×20 cm in both seasons, resulting in 26 rows by 100 hills in each plot.Pesticides were not used at all throughout the experiments.

2.3. Occurrence of pests and/or damage by rice pests

Monitoring for the occurrence of pests and/or damage by rice pests was initiated at 20 days after transplant (DAT) and subsequently at 10-day intervals for both the early- and the late-season rice. For damage by RSB and RLF, a systematic sampling method was employed. The sampling was started in the row 5 m from the ridge and proceeded in every fifth row across the plot. Two sampling points were randomly selected from each sampling row, which resulted in eight sampling points per plot. Numbers of RSB-damaged plants and RLF-damaged leaves were counted for 5 rice hills in each sampling point. For the calculation of percentages of RSB-damaged plants and RLF-damaged leaves, the numbers of plants and leaves per hill were counted from 25 randomly selected rice hills from each plot on each of the monitoring days immediately before booting stage. They were also used to calculate the percentages of damaged plants after the booting stage. In the mature stage (on 17 August for the early-season rice and 13 November for the late-season rice), 30 randomly selected whitehead rice plants were cut at the stem base. RSB in the stem, if any,was sorted according to developmental stages (Song et al.1958). From the early-season rice, the collected RSB was all at pupal stage while from the late-season rice, the collected RSB was all in the larval stage. The collected insects were individually weighed.

For investigating population sizes of planthoppers, simple random sampling was employed. One hill of rice was covered with a sheet iron cylinder (diameter 30 cm, height 50 cm before stem elongation and 80 cm after stem elongation); and all the arthropods in the cylinder were collected by a vacuum sucker. Twenty hills were sampled per plot and all the samples from one plot on a given monitoring day were pooled together. The collected arthropods were identified and numbers of planthoppers(BPH and WBPH) were recorded. The population size data were adjusted to number of individuals per 100 hills before data analysis.

The occurrence and severity of rice leaf blast, neck blastand sheath blight were recorded for 10 hills at each of five diagonal points in a plot. The severity of leaf blast and neck blast was evaluated at the maximum tillering stage and the mature grain stage, respectively, and the severity of sheath blight was evaluated at the tillering, booting and dough grain stages, according to the standard evaluation system (Appendices A and B; IRRI 1996). The proportion of diseased plants (PDP) and disease index (DI) for rice leaf and neck blast and rice sheath blight were calculated as:

Where, Niis the number of plants (sheath blight), leaves(leaf blast) or panicles (neck blast) whose damage was rated as score i, Ntis the total number of investigated plants, leaves or panicles, and I is the value of the highestdamage score.

2.4. Rice yield

Ten plants of the early-season rice were collected at each of five diagonal points from each plot at maturity to determine yield components. Panicles were hand-threshed, and filled grains were separated from unfilled grains by submerging them in tap water. Thousand-grain weight of filled grains was determined after oven-drying at 70°C to constant weight.The numbers of total grains and filled grains per panicle were counted, and grain-filling percentages (100×Number of filled grains/Total number of grains) were calculated for each plot.

Another 15 hills were sampled at five diagonal points from each plot at maturity to determine the number of plants per hill and the number of effective panicles. Then the plots were harvested by hand and grain yield per plot was weighed and adjusted to reflect a water moisture content of 0.14 g g–1fresh weight.

2.5. Data analysis

Data in the tables and figures are expressed as mean±SE.All data were analyzed by one-way analysis of variance(ANOVA), followed by Tukey HSD test (P=0.05) for significant differences between treatments. Percentage data for damage were transformed using arcsine square-root,and homogeneity of variance was tested before ANOVA.Age distribution data of the RSB larvae collected from the late-season rice were analyzed using the Chi-square testin Crosstabs. All statistical analyses were performed using SPSS 16.0 (SPSS Inc., USA).

3. Results

3.1. Damage by RSB and RLF

The percentage of plants damaged by RSB in the earlyseason rice was generally high in the control plots compared with the Si-amended plots and decreased with increases in Si amendment (Fig. 1-A). Si addition significantly influenced RSB damage on five of the seven monitoring days, i.e., on 20, 30, 50, 60, and 80 DAT (F≥5.603, df=3, 11, P≤0.023).Compared with the control plots, damage levels were significantly lower in the plots supplied with Si at 150 kg SiO2ha–1on 20 DAT, in the plots with Si addition at 300 and 150 kg SiO2ha–1on 30, 50 and 60 DAT, and in the plots with Si at 300 kg SiO2ha–1on 80 DAT. The patterns of RSB damage in the late-season rice were similar to those in the early-season rice (Fig. 1-B). On four of the seven sampling days, Si amendment significantly influenced RSB damage (F≥9.952, df=3, 11, P≤0.004). Significantly reduced damages compared to control plots were recorded on 40,50 and 60 DAT in the plots with Si amendment at 300 and 150 kg SiO2ha–1. On 70 DAT, damage in the plots with Si at 300 kg SiO2ha–1was significantly lower than that in the plots with Si at 150, 75 and 0 kg SiO2ha–1.

Fig. 1 Percentage of rice plants damaged by the rice stem borer(RSB) in plots with Si amendment at 0, 75, 150 and 300 kg SiO2 ha–1 applied 3 days before transplanting. A, early-season rice. B, late-season rice. The plot sampling was initiated at 20 days after transplanting (DAT) on 9 June for the early-season rice and on 13 September for the late-season rice. ＊ indicates a significant difference between the Si addition rates (Tukey HSD test, P=0.05).

Mature plants with RSB damage were sampled to check the developmental stages of the insects. All the insects collected in the early-season rice had pupated.Si application significantly reduced pupal weight (F=8.62,df=3, 323, P＜0.001) (Fig. 2-A). In the late-season rice, the collected insects were all at the larval stage. Larval weight was also significantly reduced by Si application (F=5.875,df=3, 347, P=0.001). The larvae collected from the plots with Si amendment tended to be earlier instars than the larvae from the control plots (Chi-square test, P＜0.001; Fig. 2-B).In the plots with Si addition at 300 kg SiO2ha–1, the 3rd,4th, 5th, and 6th instars accounted for 3.0, 34.3, 48.5, and 14.3% of the total collected larvae, respectively, in contrast to 0.0, 20.7, 61.4, and 18.0% in the control plots, respectively.

RLF damage peaked on 50 DAT in the early-season rice(Fig. 3-A). A significant influence of Si amendment on RLF damage was observed on four of the seven monitoring days,i.e., 40, 50, 60, and 70 DAT (F≥4.876, df=3, 11, P≤0.033).Damage was significantly lower in the plots with Si addition at 300 kg SiO2ha–1than in the control plots on the four monitoring days. On 50 and 60 DAT, less damage was also recorded in the plots with the highest Si addition than in the plots with Si at 150 and 75 kg SiO2ha–1, respectively.In the late-season rice, the percentage of RLF-damaged rice leaves was generally low in comparison with that in the early-season rice, but the damage increased gradually as the season advanced (Fig. 3-B). The damage was generally high in the control plots compared to Si-amended plots and decreased with increasing Si amendment. On four of the seven monitoring days, i.e., 40, 50, 60, and 70 DAT, significant effects of Si addition on RLF damage were observed (F≥4.7, df=3, 11, P≤0.036). Lower damage was recorded in the plots with Si amendment at 300 kg SiO2ha–1than that in the control plots on all the four monitoring days.

Fig. 2 Influence of Si amendment on development of the ricestem borer (RSB) in the damaged plants collected from plots with Si amendment at 0, 75, 150 and 300 kg SiO2 ha–1. A, pupal(early-season rice) and larval (late-season rice) weight. B, larval age distribution (%) in late-season rice. In panel A, histogram indicates mean and error bar is SE; different letters over the bars within pupae or larvae indicate significant differences(Tukey HSD test, P=0.05).

3.2. Population size of planthoppers

The planthopper populations showed the same temporal dynamic patterns across all the treatments in either the early- or the late-season rice (Fig. 4). The populations were relatively high in the control plots compared to those in the Si-amended plots and decreased with increasing Si amendment. In the early-season rice, the planthopper population sizes in the Si-amended plots (300 and 150 kg SiO2ha–1) were significantly decreased by more than 22.8% (F≥8.388, df=3, 11, P≤0.007) compared to that in the control plots on 40, 60 and 80 DAT (Fig. 4-A). On 70 DAT, planthopper populations in the plots amended with 300 kg SiO2ha–1were significantly lower than those in the plots with 150, 75 and 0 kg SiO2ha–1(F=15.018, df=3, 11,P=0.001) (Fig. 4-A). In the late-season rice, compared to the control plots, Si addition (300 and 150 kg SiO2ha–1)significantly reduced planthopper population sizes by more than 30.9% on 30, 60 and 70 DAT (F≥6.345, df=3, 11,P≤0.016) (Fig. 4-B).

Fig. 3 Percentage of leaves damaged by the rice leaf folder(RLF) in plots with Si amendment at 0, 75, 150 and 300 kg SiO2 ha–1 applied 3 days before transplanting. A, early-season rice. B, late-season rice. The plot sampling was initiated at 20 days after transplanting (DAT) on 9 June for the early-season rice and on 13 September for the late-season rice. ＊ indicates a significant difference between the Si addition rates (Tukey HSD test, P=0.05).

3.3. Occurrence and severity of rice blast and sheath blight

The proportion of diseased plants (PDP) and disease index (DI) of leaf and neck blast decreased with increasing Si amendment in the early-season rice (Table 1). The differences in PDP were not significant between the treatments in the early-season rice (leaf blast: F=3.813,df=3, 11, P=0.058; neck blast: F=1.326, df=3, 11, P=0.332).In contrast, DI of leaf and neck blast in the plots with Si addition at 300 and 150 kg SiO2ha–1were significant lower than that in the control plots in the early-season rice (leaf blast: F=5.385, df=3, 11, P=0.025; neck blast: F=7.842,df=3, 11, P=0.009). However, no significant differences were observed between the treatments in PDP and DI of sheath blight at tillering, booting or dough stages in the early-season rice (Table 2) or of leaf and neck blast in the late-season rice (Table 3).

Fig. 4 Population size of planthoppers in plots with Si amendment at 0, 75, 150 and 300 kg SiO2 ha–1 applied 3 days before transplanting. A, early-season rice. B, late-season rice.The plot sampling was initiated at 20 days after transplanting(DAT) on 9 June for the early-season rice and on 13 September for the late-season rice. ＊ indicates a significant difference between the Si addition rates (Tukey HSD test, P=0.05).

3.4. Rice yield

Rice yield was evaluated for the early-season rice (Table 4).The number of effective panicles showed no significant difference between the treatments (F=0.404, df=3, 11,P=0.754). However, a significant difference in number of grains per panicle was recorded between the treatments(F=13.841, df=3, 247, P＜0.001), with the Si-added plots showing 8.3–22.2% increases over the control plots. The grain-filling percentage also differed significantly between the treatments (F=9.527, df=3, 247, P＜0.001), and increased in the plots with Si addition at 300 and 150 kg SiO2ha–1by 10.7 and 10.3% in comparison with that in the control plots,respectively. Compared to the control plots, the thousandgrain weight in the plots with Si addition at 300 kg SiO2ha–1was significantly higher by 6.3% (Tukey HSD test, P=0.047)and in the plots with 75 kg SiO2ha–1by 6.2% (Tukey HSD test, P=0.048). Rice yields increased in the Si-amended plots over that in the control plots by 6.8–16.4%, however the increases were not statistically significant (F=1.895,df=3, 11, P=0.209).

Table 1 Effects of Si amendment on occurrence and severity of rice blast in early-season rice1)

Table 2 Effects of Si amendment on occurrence and severity of rice sheath blight in early-season rice1)

Table 3 Effects of Si amendment on occurrence and severity of rice blast in late-season rice1)

Table 4 Effects of Si amendment on rice yield components and yield in early-season rice

4. Discussion

To ensure sustainable rice production, it is essential to develop ecologically sound alternative methods to achieve successful pest control. Si amendment may be one such potential alternative (Savant et al. 1997). Increasing evidence shows that a high quantity of Si in plants confers resistance or tolerance to various biotic stresses (Ma 2004). Si amendment in soils or hydroponic solutions has previously been shown to enhance rice resistance to stem borers (Hou and Han 2010; Sidhu et al. 2013; Jeer et al.2017), leaf folders (Ye et al. 2013; Han et al. 2015, 2016,2017), and planthoppers (Yang et al. 2014; He et al. 2015;Wu et al. 2017; Yang et al. 2017a, b, 2018). Previous studies have proposed two mechanisms to account for Si-mediated plant defense against insect herbivores: the physical barrier mechanism and the induced defense mechanism. However,few studies have directly assessed the Si-mediated plant resistance to pests in a field situation.

In the present study, we measured pest occurrence and damage and rice yield in response to Si amendment in field plots. Significantly reduced plant damage by RSB and RLF,compared to control plots, was documented on four or five out of seven monitoring occasions in either the early- or lateseason rice amended with Si at 300 kg SiO2ha–1(Figs. 1 and 3). Nakano et al. (1961) documented heavy infestations of rice stem borers in fields deficient in available Si, while application of calcium silicate in those fields decreased both the insect damage and populations. Savant and Sawant (1995) observed a decrease in the incidence of dead hearts (stem borer damage) after transplanting when the rice seedlings were fertilized with black-gray ash of rice hulls (a type of Si source) at the nursery stage. For RLF,reduced damage in potted plants with Si amendment was also reported (Han et al. 2015, 2017). In the current study,the population size of planthoppers was generally low in the plots with Si addition and significantly reduced in the plots amended with Si at 300 kg SiO2ha–1when compared to the control plots on three and four out of the seven monitoring days in the late- and early-season, respectively (Fig. 4).These findings confirm a positive relationship between rice Si content and resistance to BPH (Sujatha et al. 1987).Another noteworthy feature of the planthopper populations was that they showed similar dynamic patterns in all the plots: decreasing gradually from a peak on 40 DAT to a low level on 60 DAT and then abruptly increasing to a higher peak on 70 DAT. Since no pesticides were sprayed in the plots during the whole investigation period, the decrease might have been the result of emigration while the rapid increase might have resulted from immigration.At the population peaks, a reduced average planthopper population size was recorded in the plots with Si addition at 300 kg ha–1than in the control plots, which could be the result of antixenosis afforded by Si amendment (Yang et al.2017b, 2018). The disease index of both leaf and neck blastin the early-season rice was lower in the plots amended with 300 kg SiO2ha–1than in the control plots. Previous studies also reported that Si application resulted in a decrease in severity of rice blast (Seebold et al. 2000; Kim et al. 2002;Cai et al. 2008). However, no influence of Si amendment on PDP or DI of sheath blight in early-season rice or on blast in late-season rice was detected in the present fieldstudy. These observations contrast to previous reports for the effects of Si amendment on rice sheath blight (Rodrigues et al. 2001) and rice blast (Seebold et al. 2000). While the reasons for this discrepancy are uncertain, one possibility is that Si amendment might have been insufficient to function against the high level of rice blast that occurred in the lateseason rice in this study. Another reason may be that Si was applied at lower levels (75, 150 and 300 kg SiO2ha–1)in the current study than the levels in previous studies, e.g.,(equal to 1 185 kg SiO2ha–1) in the report by Rodrigues et al.(2001) and (equal to 1 056 and 2 111 kg SiO2ha–1) in thestudy by Seebold et al. (2000). However, Si amendment at such high levels can work against its practical use in rice production due to the associated high cost. Even with the modest Si amendment levels in this study, our results suggest that Si amendment at 300 kg SiO2ha–1can provide affordable, substantial protection from the major rice pests.Although not measured in the present study, Si content was found to increase in Si-amended potted plants compared to control plants in our previous studies (Hou and Han 2010; Han et al. 2015, 2016, 2017; Yang et al. 2017a, b,2018), where the plants received similar Si addition rates.Therefore, it is reasonable to assume that in the presentstudy, Si content in the rice plants with Si amendment also increased relative to the control plants, which would add to the enhanced protection from rice pests in Si-added plots.

The reduced occurrence of pests or damage by the pests in the current field tests could be the result of antibiosis mediated through Si amendment. We recorded significantly reduced RSB larval and pupal weights and retarded larval development in the plots with Si application(Fig. 2), indicating potential antibiosis effects. Similar results of negative effects of Si application on stem borer larval performance have been reported in studies using potted plants. Sidhu et al. (2013) reported that Si incorporation into soil led to reduced performance and relative growth rates in Diatraea saccharalis (Fabricius) larvae. Si amendment has been observed to decrease RSB larval weight gain and stem damage, and to prolong penetration duration and larval development (Hou and Han 2010). Si amendment also reduced damage (percent dead heart and white ear)by the yellow stem borer, Scirpophaga incertulas (Walker)as compared with the control (Voleti et al. 2008; Jeer et al.2017). The fact that rice varietal resistance to RSB has been linked with Si content (Hao et al. 2008) also indicates Si-mediated antibiosis. In RLF, Ye et al. (2013) and Han et al. (2015, 2017) observed reduced larval survival,fecundity and population growth rates in assays using excised-leaves from rice plants grown in media amended with Si. For planthoppers, He et al. (2015) and Yang et al.(2017b) demonstrated that Si addition to rice plants reduced fecundity, nymph survival rate, and population growth rate.These studies collectively show that Si-mediated antibiosis functions against a variety of herbivorous insects with different feeding habits.

Several mechanisms contribute to the Si-mediated antibiosis. First, Si content is increased in Si-amended plants (Han et al. 2015; Yang et al. 2018), and the increased deposition of amorphous phytoliths in plant tissues reduces their digestibility to herbivores (Massey and Hartley 2009;Reynolds et al. 2016). This is evident from the fact that Si amendment significantly increased the C:N ratio of rice leaves (Han et al. 2017) and reduced the efficiency of conversion of ingested and digested food in the third instars of RLF (Han et al. 2015). Second, the enhanced induction of defenses through greater mobilization and activities of an array of defensive enzymes in Si-amended and pest-attacked plants has been shown (Ye et al. 2013; Han et al.2016; Yang et al. 2017a), which can ultimately stress the development of the herbivores. Third, for the sucking insect pests, higher callose deposition in sieve tubes (Yang et al.2018) and increased soluble silicic acid in rice leaf sheaths(Yoshihara and Sögawa 1979) in Si-added plants serve to reduce herbivore feeding and thus contribute to antibiosis.Recent results show that phytohormones are involved in the induction of the defense responses associated with Simediated plant resistance (Ye et al. 2013).

In addition to the biochemical mechanisms, the reduced pest severity in Si-amended plots in the current field tests can also result from the behavioral changes of insect herbivores.Hou and Han (2010) and Sidhu et al. (2013) revealed that boring success of RSB and D. saccharalis larvae into ricestems decreased in plants supplied with Si. In preference tests, RLF larval settlement and egg deposition on rice plants amended with Si were significantly reduced (Han et al. 2017).Using electric penetration graphing, Yang et al. (2017b)recorded impaired-sucking behaviors that led to reduced feeding in BPH. BPH female adults showed non-preference for plants with Si addition (Yang et al. 2017b). These results suggest that Si amendment functions to increase herbivorous antixenosis and deter feeding behavior.

Further, natural enemies and other indirect lethal factors in the field can also contribute to the reduced occurrence and damage of insect pests in Si-amended fields. When attacked by insect herbivores, Si-amended plants increase the release of plant volatiles that attract the natural enemies of the insect herbivores (Kvedaras et al. 2010; Liu et al.2017), which can help reduce the population size or damage of the herbivores. Additionally, poor/delayed development of insect herbivores or prolonged penetration (as in the case of stem borers) on Si treated rice plants can increase their exposure to adverse environmental factors (e.g., rain and high temperature) and predation/parasitism (Hou and Han 2010). These factors are believed to constitute a major component of Si-mediated plant resistance to insect herbivores (Massey at al. 2006; Kvedaras and Keeping 2007; Hou and Han 2010).

In addition to enhancing plant resistance to pests, Si has been recognized as a beneficial element to improve rice grain yields (Tamai and Ma 2008). In the present study,we observed significant increases in the number of grains per panicle, grain-filling percentage, and thousand-grain weight in the plots with 300 kg SiO2ha–1compared to the control plots (Table 4). However, the increase of rice yield of 16.4% (604.5 kg ha–1) in the plots with 300 kg SiO2ha–1over the control plots was insignificant; this is due to the lack of a difference in number of effective panicles (Table 4).Pati et al. (2016) observed a significant increase in rice yield in field plots amended with ＞480 kg SiO2ha–1; and in an economic analysis of Si application, Alvarez and Datno(2001) assumed a yield increase of 500 kg ha–1from Si amendment. The current results generally confirm previous reports that exogenous Si supply enhances straw biomass,number of spikelets per panicle, and particularly the percentage of filled spikelets (Tamai and Ma 2008; Detmann et al. 2012) due to its beneficial effects on water use efficiency and cell elongation (Hossain et al. 2002; Isa et al.2010). Although the mechanisms for the Si-mediated rice yield increase are complex and largely unclear (Detmann et al. 2012), it is certain that the reduced pest occurrence and severity has played a role.

From the present study and previous reports, it can be generalized that Si amendment is not a ‘cure’ like pesticides but may act as a ‘regulator’ of pests. Instead of directly killing pests, Si functions indirectly (i.e., mediated through hostplant) to regulate the behavior, development and physiology of pests in ways which make them less destructive. In that way, Si amendment, especially applied to soils deficient in available Si, can help keep pest outbreaks in check.

5. Conclusion

Our study demonstrates in a field design that Si amendment at 300 kg SiO2ha–1contributes substantially to the reduction of population size and damage by rice insect pests in both seasons and the occurrence of rice blast in the early season. The addition of Si also marginally enhances rice yield. These findings provide a basis to recommend the incorporation of Si amendment as a key component of integrated management of rice pests.

Acknowledgements

This study was financially supported by the National Natural Science Foundation of China (31371951), the National Key Technology R&D Program for Grain Crops, Ministry of Science and Technology of China (2016YFD0300701) and the Science and Technology Innovation Project of Hunan Academy of Agricultural Sciences, China (2017JC41). We thank internship students Qin Guobiao, Zhong Mingtai and Cai Zhiyang from Hunan Agricultural University for technical assistance with the experiments.

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm