Influence of canopy structure on sheath blight epidemics in rice

Authors

  • W. Wu,

    1. College of Agronomy, Northwest A&F University, Yangling, Shaanxi, China
    2. National Key Laboratory of Crop Genetic Improvement, MOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author
  • L. Nie,

    1. National Key Laboratory of Crop Genetic Improvement, MOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author
  • F. Shah,

    1. National Key Laboratory of Crop Genetic Improvement, MOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, China
    2. Department of Agriculture, University College of Science, Abdul Wali Khan University Mardan, Mardan, Pakistan
    Search for more papers by this author
  • Y. Liao,

    1. College of Agronomy, Northwest A&F University, Yangling, Shaanxi, China
    Search for more papers by this author
  • K. Cui,

    1. National Key Laboratory of Crop Genetic Improvement, MOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author
  • D. Jiang,

    1. Provincial Key Laboratory of Plant Pathology of Hubei Province, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author
  • J. Xie,

    1. Provincial Key Laboratory of Plant Pathology of Hubei Province, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author
  • Y. Chen,

    1. National Key Laboratory of Crop Genetic Improvement, MOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author
  • J. Huang

    Corresponding author
    1. National Key Laboratory of Crop Genetic Improvement, MOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, Hubei, China
    Search for more papers by this author

Abstract

Sheath blight (caused by Rhizoctonia solani) is one of the most important constraints in achieving high grain yield in intensive rice production systems. Canopy structure can influence the development of sheath blight epidemics. The objective of this study was to determine the effect of canopy structure parameters such as shoot number, leaf area index, biomass production, contact frequency, light transmittance and plant height on the development of sheath blight epidemics in commercial fields. Field experiments were conducted in both early and late seasons of 2009 and 2010 in Wuxue, Hubei province, China. The effects of nitrogen (N) rate and hill density on structure and production parameters and sheath blight severity were investigated. Sheath blight severity was recorded as a sheath blight index or relative lesion height on inoculated and uninoculated plants in each crop. Lesion length was measured on inoculated plants in 2010. The results showed that the sheath blight index increased with an increase of N rate and hill density in uninoculated plots in each trial. Stepwise multiple regression analysis demonstrated that contact frequency was consistently related to sheath blight. Lesion length on inoculated plants was not correlated with canopy structure. These results indicate that canopy structure influences sheath blight epidemics. A ‘healthy’ canopy resulting from appropriate crop management practices can be used to suppress sheath blight epidemics in rice.

Introduction

Sheath blight, caused by Rhizoctonia solani anastomosis group AG–1 IA, has become a potentially devastating disease of rice (Oryza sativa) in all temperate and tropical rice production regions, especially in intensive production systems (Teng et al., 1990). Recently, the severity of sheath blight in rice growing regions has increased through intensified rice production systems that are characterized by abundant nitrogenous fertilizer application, high planting density and adaptation of high-yielding cultivars (Castilla et al., 1996; Wu et al., 2013). These practices, together with the saprophytic nature and wide host range of the pathogen, have led to the dissemination and persistence of the fungus in all rice producing areas (Lee & Rush, 1983).

Two pathozones for a R. solani epidemic can be distinguished (Savary et al., 1997, 1998). In the first phase, soilborne inoculum usually infects lesions on the lower internodes of the leaf sheaths, which later coalesce and expand to the upper parts of the rice plant. This represents the soilborne phase of the pathogen (Savary et al., 2001). In the second phase, the pathogen spreads to the upper structure of the canopy by means of mycelial strands growing from primary or secondary lesions to healthy tissues of the same or adjacent plants, where these mycelial strands establish new infections. These two phases have been termed horizontal and vertical spread, respectively (Kozaka, 1961). Generally, under favourable conditions, initial symptoms on the rice plant are found at the mid-tillering stage in lowland fields, as a small, ellipsoidal or ovoid, water-soaked shape near the water line. Lesions may coalesce forming bigger lesions with an irregular outline and cause the death of the infected tissue. Then, infection may spread rapidly by runner hyphae to near or upper healthy tissue. Lesions become dry, tan, and eventually grey with brown borders, and then produce sclerotia (Lee & Rush, 1983).

Double cropping of rice is practised on a large proportion of the rice land in the southern provinces of China to meet the demand for rice production (Zhu et al., 2010; Zou, 2011). In these areas, early season rice is generally grown from April to July and late season rice from July to October. Because mostly short stature and early maturing varieties are used in a double rice cropping system, sheath blight incidence is more severe than in a single rice cropping system (Sharma et al., 1995; Wang, 2006).

Most recent studies have focused on varietal or genotypic differences in sheath blight resistance and yield loss resulting from sheath blight (Marchetti & Bollich, 1991; Cu et al., 1996; Zou et al., 2000; Han et al., 2003; Groth & Bond, 2007). Although no gene confers complete resistance to sheath blight in rice (Srinivasachary et al., 2011), a number of genes provide significant resistance. These genes have been mapped using quantitative trait loci (QTL) mapping (Pan et al., 1999; Zou et al., 2000; Liu et al., 2009). However, there is a lack of agreement among the QTLs identified by different groups (Pinson et al., 2005). Canopy architecture and density can vary greatly among rice genotypes, which in turn can affect the spread of the rice sheath blight pathogen (Han et al., 2003; Zhong et al., 2006). Also, various factors are influenced by canopy structure, such as microclimatic conditions (including sunlight, humidity and temperature), nitrogen status, contact frequency, light transmittance and shoot number, which may all contribute to the variation in disease expression and thus make evaluation of host resistance difficult (Savary et al., 1995; Castilla et al., 1996; Zhong et al., 2006; Tang et al., 2007; Srinivasachary et al., 2011).

The pathogen spreads by means of runner hyphae that originate from lesions and progress towards healthy tissues within a hill (a group of seedlings planted in the same place and treated as a single plant) or between hills (Ou, 1987). Contact between infected and healthy tissues lead to infection of healthy tissues. The contact frequency is the number of leaf-to-leaf and leaf-to-sheath contacts between adjacent plants (no. m−2) and it reveals an important epidemiological factor for sheath blight (Savary et al., 1995; Castilla et al., 1996; Willocquet et al., 2012). Similarly, canopy structure has an important effect on Septoria wheat disease (Septoria nodorum and Septoria tritici) epidemics. Variation in canopy structure alters the leaf area index, plant height and rate of shoot extension, which can result in a different rate of vertical disease progress for a splash-dispersed pathogen when rain drops hit the infected leaves (Shaw, 1987; Lovell et al., 1997; Pielaat et al., 2002; Robert et al., 2008).

Little direct information is available regarding the effects of canopy structure on sheath blight epidemic development under field conditions. Manipulation of the canopy structure offers promising possibilities for keeping pathogens under control (Tivoli et al., 2013). The present research examines the relationship of canopy structure parameters with disease epidemic development on uninoculated and inoculated plants. It is hypothesized that contact frequency is a key epidemiological factor related to canopy structure for sheath blight, and that other canopy structure parameters are also related to epidemic development through their relationships with contact frequency.

It has been documented that planting density and N rate affect rice canopy structure parameters such as plant height, leaf area index, biomass production, leaf angle and light extinction coefficient, which are highly related to grain yield and solar radiation use efficiency (Duy et al., 2004; San-oh et al., 2004). However, the role of these factors in sheath blight epidemic development has not been sufficiently studied. A range of N rates and planting densities were used in this study to produce canopy structures that differed in architecture and density. Leaf area index and biomass production were recorded because they represent canopy density well (Zhong et al., 2006) and are also essential traits related to yield performance (Ying et al., 1998); Shoot number and light transmittance were regarded as indicators of canopy connectedness and plant architecture (San-oh et al., 2006). Similarly, contact frequency among leaves and sheaths is an important epidemiological factor for rice sheath blight (Savary et al., 1995). Plant height affects sheath blight development by constraining the vertical progress of the fungus, and alters microclimate and light transmission inside the canopy (Han et al., 2003). Soil–plant analysis development (SPAD) values represent nitrogen status, and high availability of N is conducive to rice sheath blight epidemics. N source is not only required for fungal growth, but also influences shoot numbers and leaf expansion, which together determine canopy size (Walters & Bingham, 2007). Dead leaves and sheaths, mainly caused by sheath blight infection, are good indicators of sheath blight severity (Lee & Rush, 1983). The objectives of this study were to: (i) determine the influence of N rate and hill density on canopy structure parameters and sheath blight development, (ii) identify essential epidemiological factors related to canopy structure that can affect sheath blight epidemics, and (iii) examine the relationship between canopy structure and sheath blight severity.

Materials and methods

Experimental design and cultural practices

Field experiments were conducted in 2009 and 2010 in two adjacent fields at Dajin, Wuxue, Hubei province, China (29°59′N 115°36′E). The soil properties of the experimental site in each trial are described in Wu et al. (2013). Climatic condition data were obtained from the Meteorological Bureau of Wuxue County and are presented in Table 1.

Table 1. Mean air temperature, total rainfall and mean relative humidity from transplanting (TR) to mid–tillering (MT), MT to panicle initiation (PI), PI to flowering (FL) and FL to physiological maturity (PM) in rice during the early season (ES) and late season (LS) cropping periods in 2009 and 2010
SeasonsTR to MTMT to PIPI to FLFL to PM
Mean air temperature (ºC)
2009 ES22·122·828·229·1
2009 LS28·124·725·921·5
2010 ES21·322·725·928·2
2010 LS29·728·625·920·9
Total rainfall (mm)
2009 ES61874120
2009 LS113523222
2010 ES1936614370
2010 LS6615456235
Mean relative humidity (%)
2009 ES75·482·177·981·7
2009 LS88·191·481·978·4
2010 ES82·678·885·389·0
2010 LS84·487·289·086·7

The experiment was conducted each year in the early season (April–July) and repeated in the late season (July-October) in the same field. Treatments were arranged in a split-plot design with nitrogen (N) rate as the main plot and hill density as the subplot. There were four replications per treatment. In 2009, the main plot treatments were two N rates: a zero-N control and a high N rate (180 and 195 kg N ha−1 in the early and late season, respectively). Subplot treatments were three hill densities: 19, 28 and 38 hills m−2, representing hill spacing of 13 × 40, 13 × 27 and 13 × 20 cm, respectively. Subplot size was 30 m2. In 2010, three N rates (120, 180 and 240 kg N ha−1 in both seasons) were allotted to main plots, while the three hill densities (same as in 2009) were arranged in subplots. The sub-subplot size was 45 m2.

Nitrogen in the form of urea was split-applied at basal (BS), mid-tillering (MT), panicle initiation (PI) and flowering (FL) stages. In 2009, the N-splitting patterns for the early and late seasons were 35% (BS) + 20% (MT) + 30% (PI) + 15% (FL) and 41% (BS) + 24% (MT) + 35% (PI), respectively. In 2010, the N-splitting pattern for both seasons was 50% (BS) + 30% (MT) + 20% (PI). Phosphorus at 40 kg P ha−1, potassium at 50 kg K ha−1 and zinc at 5 kg Zn ha−1 were applied as a basal application in all plots. The basal fertilizers were broadcast and incorporated 1 day before transplanting. Additional potassium at 50 kg K ha−1 was topdressed at PI in all plots. In the 2009 early and late season trials, fungicide (100 g propiconazole ha−1) was applied to each plot using a calibrated backpack sprayer at 8 days after the FL stage. No fungicide was applied to any plot in 2010.

Two widely adapted high-yielding F1 hybrid commercial cultivars were used. Liangyou-287 was used in the early season and T-you207 in the late season for both years. Early season rice was sown on 24 March 2009 and 23 March 2010, and transplanted on 30 April 2009 and 2 May 2010. Late season rice was sown on 19 June 2009 and 18 June 2010, and transplanted on 28 July 2009 and 20 July 2010. Two seedlings were transplanted per hill in each experiment. Each main plot was surrounded by a 35 cm wide ridge covered with plastic film. The plastic film was installed to a depth of 20 cm below the soil surface 3 days before transplanting. Crop management followed the standard cultural practices for the region. The plots were flooded 3 days after transplanting and a 4–10 cm water depth was maintained until 7 days before maturity, when the field was drained. Insects and pests were controlled using insecticides to avoid biomass and yield loss. Herbicide (48 g acetochlor and 12 g bensulfuron-methyl ha−1) was used to control weeds. Diseases other than sheath blight were not observed in any trial.

Pathogen inoculation

An R. solani anastomosis group AG–1A isolate (WH–1 strain), which was isolated from typical sheath blight lesions in a rice field at Huazhong Agricultural University, Wuhan, China (Xie et al., 2008), was used to inoculate plots in each trial. The method of inoculation was adopted from Zou et al. (2000). Inoculation using toothpicks is an effective and accurate method for the evaluation of variation in resistance to R. solani (Prasad & Eizenga, 2008; Srinivasachary et al., 2011). Briefly, autoclaved 10 cm long wooden toothpicks were incubated with the WH–1 strain on potato dextrose agar (PDA) at 28°C for 4 days and then placed behind the leaf collar of the third sheath, counting from the top of the rice plant. In each plot, 24 hills (4 × 6) were selected as microplots for inoculation at the PI stage, i.e. 32 days after transplanting (d.a.t.), because inoculation efficiency is highest when infection starts at the PI stage (Rodrigues et al., 2003). Three main shoots were selected for inoculation in each hill.

Measurements

The plants from 12 hills were randomly sampled from each plot at the PI (32 d.a.t.) and FL (50 d.a.t.) stage in each trial. Plants were washed to remove soil and the roots were separated from the rest of the plant. The shoots from each hill were counted to determine the shoot number per m2. The above-ground parts were divided into leaves, leaf sheaths plus shoots and panicles. The leaf blades that retained some green portion were collected, and the sections that were not green were removed. Leaf area was then determined with a LI-COR area meter (LI-COR Model 3100, LI-COR Inc.), and LAI (leaf area index) was calculated as total leaf area per m2 of ground area at PI and FL. Above-ground total biomass was determined after oven drying at 80°C to a constant weight. In both trials in 2010, dead leaves and sheaths were separated from 12 hills at FL and 10 days after FL to calculate the weight of dead leaves and sheaths per m2.

Some canopy-related parameters (contact frequency, plant height, light transmittance and SPAD) were measured in the present study. Contact frequency was recorded at 7 days after inoculation (d.a.i.) in each trial. Two hills were randomly selected within the central area of each plot. The total number of leaf-to-leaf and leaf-to-sheath contacts between each sampled hill and all of the eight hills surrounding it were counted. Contact frequency per m2 was calculated according to Willocquet et al. (2000). Plant height of the main shoot was measured at flowering. Light transmittance (%) was also recorded at 7 d.a.i. in the 2009 late season and the 2010 early season. The light intensity above the canopy (I0) and at the surface of the soil under the canopy (Ib) was measured simultaneously by the Sunscan canopy analysis system (Delta-T Devices Ltd). These values were then used to calculate light transmittance as: light transmittance (%) = Ib/I0 × 100. SPAD was measured using a chlorophyll meter (SPAD-502, Minolta Camera Co.) at 10 days after the FL stage in both trials in 2010. Twelve leaves were randomly selected on the third internode from the top of plants to measure SPAD.

Sheath blight index (ShBI, %) was measured in uninoculated plants in each trial; in inoculated plants, ShBI was measured in both seasons of 2010 and relative lesion height (RLH, %) was measured in both seasons of 2009. The methods of recording ShBI and RLH are described elsewhere (Wu et al., 2012). Total lesion length (cm per plant) were measured at 7 d.a.i. based on the 12 inoculated shoots per plot from four adjacent hills in both seasons of 2010. To minimize the variability among ratings, sheath blight severity (including ShBI, RLH and lesion length) was recorded by one person in each trial. Lodging index (%) was investigated at 30 days after flowering using the method of Wu et al. (2012). Grain yield was determined from a 5 m2 area in each plot at maturity stage and adjusted to the standard moisture content of 14%.

The relationships between canopy structure parameters, RLH or ShBI were evaluated using correlation analyses (Statistix, 2003). Means among treatments were compared using the least significant difference test (LSD) at a probability level of 0·05. Stepwise regression using sas v. 9.1.2 (SAS Institute, 2003) was performed to assess which canopy structure parameters were related to sheath blight across the four trials. Contact frequency (CF) was converted to ln(CF) prior to stepwise regression (Willocquet et al., 2000).

Results

The mean air temperature was higher in the late season than the early season during the period from TR to MT (Table 1). Subsequently, the difference in air temperature from PI to FL was lower. Conversely, air temperature was higher during the FL to PM phases of the early season in 2009 and 2010 than in the late seasons. Total rainfall from TR to MT was highest in 2010 early season, followed by the 2009 late season; from MT to FL, it was highest in the 2010 late season, and from FL to PM it was highest in the 2010 early season, followed by the 2010 late season. The mean relative humidity (%) was higher in the late season than in the early season from MT to FL in both years, while the early season had a higher mean relative humidity (%) from the FL to PM stage.

No sheath blight lesion was observed on uninoculated plants in the zero-N control treatment in either trial in 2009. N application increased ShBI and RLH significantly (Table 2). ShBI also increased with an increase in hill density, but the difference was significant only in the late season of 2009. In both trials in 2010, plants which received 240 kg N ha−1 had significantly higher ShBI than those with 120 kg N ha−1 (Tables 3 & 4). ShBI increased with increasing hill density in the 2010 early season regardless of N treatment. In the late season of 2010, high hill density increased ShBI significantly only at 120 and 240 kg N ha−1.

Table 2. Sheath blight index (ShBI), contact frequency (CF), plant height, light transmittance (LT), lodging index and grain yield of uninoculated plots, and relative lesion height (RLH) of inoculated and uninoculated plots under selected N rates and hill densities in 2009
N rate (kg ha−1)Hill density (no. m−2)ShBI uninoc.RLH uninoc.RLH inoc.CF (no. m–2)Plant height (cm)LT (%)Lodging index (%)Grain yield (t ha−1)
  1. Within a column for each season, means followed by the same letter are not significantly different according to LSD (0·05). Lowercase and uppercase letters indicate comparisons among the three hill densities and between the two N treatments, respectively.

Early season
018·70016·39 a74·0 a132 a2·30 a
28·10016·35 a68·4 b121 a2·17 a
37·50016·319 a65·8 b140 a2·18 a
Mean0 B0 B16·3 B11 B69·4 B131 B2·22 B
18018·727·4 a25·7 b53·4 a181 c86·7 a223 b4·93 b
28·127·7 a31·3 ab59·3 a323 b84·4 a261 ab5·94 a
37·535·4 a35·5 a62·0 a694 a79·9 b317 a5·63 a
Mean30·2 A30·8 A58·2 A399 A83·7 A267 A5·50 A
Late season
018·70026·1153 b111·9 a23·0 a273 a4·93 b
28·10026·1206 b110·3 a23·3 a248 ab5·45 a
37·50026·1425 a109·5 a17·1 a237 b5·72 a
Mean0 B0 B26·1 B262 B110·6 B21·1 A253 B5·37 B
19518·722·4 b26·0 a43·6 a506 c119·4 a5·3 a351 a7·26 a
28·129·5 ab27·5 a48·8 a942 b118·1 a2·3 b371 a7·19 ab
37·542·3 a34·1 a50·2 a1219 a116·2 a2·8 ab404 a6·70 b
Mean31·4 A29·2 A47·5 A889 A117·9 A3·5 B375 A7·05 A
Table 3. Sheath blight index (ShBI) of inoculated and uninoculated plots, lesion length, contact frequency (CF), plant height, light transmittance (LT), SPAD values, dry weight of dead leaves and sheaths (DLS) at flowering (FL) and at 10 days after flowering, lodging index and grain yield under selected N rates and hill densities in the 2010 early season
N rate (kg ha−1)Hill density (no. m−2)ShBI uninoc. (%)ShBI inoc. (%)Lesion length (cm)CF (no. m−2)Plant height (cm)LT (%)SPAD valuesDLS at FL (g m−2)DLS at 10 d after FL (g m−2)Lodging index (%)Grain yield (t ha−1)
  1. Within a column, means followed by the same letter are not significantly different according to LSD (0·05). Lowercase and uppercase letters indicate comparisons among three hill densities and between three N treatments, respectively.

12018·721·9 b71·8 a3·89 a352 c96·0 a16·3 a40·5 a10·0 b 9·8 b322 a4·94 a
28·132·8 ab69·4 a3·67 a730 b93·1 ab10·2 b35·8 a13·5 b14·5 ab293 a5·14 a
37·543·2 a66·0 a4·04 a1346 a92·6 b8·1 c34·4 a25·2 a21·1 a334 a4·89 a
Mean32·6 B69·1 A3·87 A809 C93·9 A11·5 A36·9 A16·2 B15·1 B316 A4·99 A
18018·727·9 b63·3 b3·87 a462 c95·8 a12·6 a40·9 a9·7 b12·4 b310 a5·08 a
28·135·6 ab68·4 ab4·08 a826 b95·4 a6·3 b33·3 ab20·9 ab21·5 ab300 a4·41 a
37·551·4 a76·0 a4·54 a1507 a91·1 b5·9 b29·4 b22·8 a28·9 a348 a4·30 a
Mean38·3 AB69·2 A4·16 A932 B94·1 A8·3 B34·5 AB17·8 B20·9 AB319 A4·60 AB
24018·731·9 b72·7 a4·29 a441 c97·8 a10·9 a32·2 a15·5 b17·6 b329 b4·59 a
28·149·5 ab75·1 a4·44 a912 b95·4 ab5·7 b34·2 a25·4 ab26·4 ab360 ab4·22 ab
37·555·2 a83·2 a4·38 a1601 a92·9 b3·8 b29·7 a32·9 a38·1 a371 a3·91 b
Mean45·5 A77·0 A4·37 A985 A95·4 A6·8 C32·0 B24·6 A27·3 A354 A4·24 B
Table 4. Sheath blight index (ShBI) of inoculated and uninoculated plots, lesion length, contact frequency (CF), plant height, SPAD values, dry weight of dead leaves and sheaths (DLS) at flowering (FL) and at 10 days after flowering, lodging index and grain yield under selected nitrogen rates and hill densities in the 2010 late season
N rate (kg ha−1)Hill density (no. m−2)ShBI uninoc. (%)ShBI inoc. (%)Lesion length (cm)CF (no. m−2)Plant height (cm)SPAD valuesDLS at FL (g m−2)DLS at 10 d after FL (g m−2)Lodging index (%)Grain yield (t ha−1)
  1. Within a column, means followed by the same letter are not significantly different according to LSD (0·05).

  2. Lowercase and uppercase letters indicate comparisons among three hill densities and between three N treatments, respectively.

12018·725·7 b56·8 b4·64 a408 c126·6 a43·6 a43·6 a56·9 b451 b7·10 a
28·151·9 a64·6 ab4·93 a752 b123·6 a43·7 a48·6 a81·1 a468 ab6·93 a
37·550·6 a66·7 a5·10 a1158 a126·9 a42·8 a64·1 a77·5 a500 a6·93 a
Mean42·7 B62·7 A4·89 A773 C125·7 A43·3 A52·1 A71·8 AB473 A6·99 A
18018·742·8 a67·2 a4·67 a401 c124·1 a45·7 a29·6 b62·0 b445 a7·22 a
28·147·3 a65·1 a4·31 a907 b126·7 a42·7 a49·9 a57·7 b444 a7·32 a
37·554·2 a58·9 a4·78 a1303 a127·2 a42·4 a62·1 a89·7 a476 a7·04 a
Mean48·1 B63·7 A4·59 A870 B126·0 A43·6 A47·2 A69·8 B455 A7·19 A
24018·747·7 b58·3 b4·33 a462 c123·5 a47·1 a33·8 b61·3 b466 a7·20 a
28·159·8 a66·1 a4·36 a970 b124·4 a44·2 ab57·7 a84·4 a480 a6·38 a
37·559·9 a67·7 a4·56 a1636 a123·4 a39·5 b58·6 a93·1 a493 a6·83 a
Mean55·8 A64·1 A4·42 A1023 A123·8 A43·6 A50·1 A79·6 A480 A6·81 A

In inoculated plants in each trial, sheath blight development infected by R. solani was nearly uniform and lesions were observed at 48 h after inoculation, resulting in 100% infection rate because high humidity was retained by the sheath covering. RLH in 2009 was lower in the zero-N control than at 180 or 195 kg N ha−1, but there was no significant difference among hill densities (Table 2). In 2010, ShBI was not significantly different among the N rates. ShBI was higher under high hill density in the case of 180 kg N ha−1 in the 2010 early season, and 120 and 240 kg N ha−1 in the 2010 late season. N rate and hill density had no effect on lesion length in 2010 (Tables 3 & 4).

High N treatment significantly increased shoot numbers, LAI and biomass compared with the zero-N control at the PI and FL stages in both trials in 2009 (Table 5). Moreover, at the PI stage, shoot numbers, LAI and biomass consistently increased with increasing hill density across the two seasons. At the FL stage, high hill density increased shoot numbers significantly across both seasons. In contrast, hill density had no consistent effect on LAI or biomass at the FL stage in 2009. In 2010, increasing N rate did not affect shoot number, LAI or biomass in both trials (Table 6). Increasing hill density increased shoot numbers, LAI and biomass consistently in the 2010 early season. In the 2010 late season, increasing hill density increased LAI and biomass significantly at the PI stage, but it had no consistent effect on shoot number at the PI and FL stages or on LAI and biomass at the FL stage.

Table 5. Canopy structure parameters under selected N rates and hill densities at panicle initiation and flowering in both early and late season trials in 2009
TreatmentPanicle initiation stageFlowering
N rate (kg ha−1)Hill density (no. m−2)Shoot number (no. m−2)Leaf area indexBiomass (g m−2)Shoot number (no. m−2)Leaf area indexBiomass (g m−2)
  1. Within a column in each season, means followed by the same letter are not significantly different according to LSD (0·05). Lowercase and uppercase letters indicate comparisons among three hill densities and between two N treatments, respectively.

Early season
018·7109 b0·44 b58 b130 b0·70 a201 a
28·1118 b0·51 ab77 a152 b0·68 a204 a
37·5137 a0·55 a85 a188 a0·73 a236 a
Mean121 B0·50 B73 B157 B0·70 B200 B
18018·7245 b1·73 b188 b245 c2·20 a506 b
28·1364 a2·55 a281 a332 b2·64 a580 ab
37·5406 a2·32 ab252 ab384 a2·80 a660 a
Mean338 A2·20 A240 A320 A2·53 A575 A
Late season
018·7187 a2·21 b366 b164 b1·82 a584 b
28·1193 a2·33 ab424 a189 a1·94 a686 a
37·5216 a2·41 a443 a193 a1·86 a702 a
Mean199 B2·32 B411 B182 B1·87 B657 B
19518·7332 b4·77 a469 b282 b4·29 a760 a
28·1442 a5·96 a600 a357 a5·15 a815 a
37·5367 ab5·22 a600 a354 a5·00 a799 a
Mean380 A5·32 A556 A331 A4·81 A792 A
Table 6. Canopy structure parameters under selected nitrogen rates and hill densities at panicle initiation and flowering in both early and late season trials in 2010
TreatmentPanicle initiation stageFlowering
N rate (kg ha−1)Hill density (no. m−2)Shoot number (no. m−2)Leaf area indexBiomass (g m−2)Shoot number (no. m−2)Leaf area indexBiomass (g m−2)
  1. Within a column for each season, means followed by the same letter are not significantly different according to LSD (0·05). Lowercase and uppercase letters indicate comparisons among three hill densities and between three N treatments, respectively.

Early season
12018·7267 c2·56 b216 b286 c3·79 b536 b
28·1440 b4·43 a378 a349 b4·47 ab616 a
37·5549 a5·12 a452 a409 a5·04 a625 a
Mean418 A4·04 A349 A348 B4·43 A592 A
18018·7310 b2·75 b242 b284 c3·83 b508 b
28·1390 b3·53 b301 b367 b4·58 ab566 b
37·5540 a5·02 a440 a414 a5·41 a665 a
Mean413 A3·77 A328 A355 AB4·61 A579 A
24018·7288 c2·55 c234 c314 c3·99 b554 a
28·1402 b3·89 b328 b377 b4·71 ab576 a
37·5534 a4·95 a439 a431 a5·21 a614 a
Mean408 A3·80 A333 A374 A4·64 A581 A
Late season
12018·7361 b5·98 b556 b294 b5·04 b682 b
28·1405 ab7·55 a650 a325 ab5·91 ab805 ab
37·5445 a7·80 a707 a360 a6·81 a922 a
Mean404 A7·11 B638 A326 A5·92 A803 A
18018·7390 a6·89 b557 b314 a5·94 a747 b
28·1423 a8·02 a667 a323 a5·84 a830 a
37·5433 a7·55 ab656 a339 a5·80 a822 a
Mean415 A7·49 AB627 A325 A5·86 A799 A
24018·7400 a6·95 b569 b304 b6·12 a795 a
28·1424 a8·18 a644 ab338 a5·72 a802 a
37·5457 a8·01 a669 a358 a6·53 a856 a
Mean427A7·71 A627 A333 A6·12 A818 A

Contact frequency (no. m−2, leaf-to-leaf and leaf-to-sheath contacts) increased significantly with increasing N rate in each trial (Tables 2, 3 & 4). In the 2009 early season, increasing hill density increased contact frequency significantly only in the N application treatment. In each trial except the 2009 early season, high plant density increased the contact frequency significantly across different N treatments. In the 2009 late season and the 2010 early season, light transmittance decreased significantly with the increasing N rate (Tables 2 & 3). In the 2009 late season, increasing hill density decreased light transmittance significantly only under the N application treatment. In the 2010 early season, increasing hill density decreased light transmittance significantly regardless of N treatment.

N treatment had a significant effect on the SPAD value only in the 2010 early season (Tables 3 & 4). Hill density has no consistent effect on the SPAD value. Dead leaves and sheaths increased significantly with the increasing N rate in the 2010 early season. In the 2010 late season, N treatment had a significant effect on dead leaves and sheaths only at 10 days after heading. In general, increasing hill density increased dead leaves and sheaths consistently across N rates and cropping seasons.

N application increased lodging index and grain yield in 2009 (Table 2), compared to zero-N. The effects of hill density on lodging index and grain yield were not consistent across the two N treatments and seasons. In 2010, excessive application of N (240 kg ha−1) decreased rice yield significantly, although it did not affect the lodging index in the two seasons (Tables 3 & 4). Hill density had no consistent effect on lodging index and rice yield across N treatments and seasons.

Canopy structure parameters including LAI, shoot number and biomass at PI and FL were positively correlated with ShBI on uninoculated plants in each trial, although correlation coefficients between LAI and ShBI were not significant at FL stage in the 2010 late season (Table 7). Furthermore, ShBI on uninoculated plants was positively correlated with contact frequency in each trial, and with dead leaves and sheaths at each stage in both trials in 2010 except at FL in 2010 late season (Table 8). ShBI was negatively correlated with light transmittance in the 2009 late season and the 2010 early season trials. In the early season of 2010, SPAD values had a negative relationship with ShBI on uninoculated plants. Plant height also showed negative correlations with ShBI in each trial, although only significant in the 2010 early season. Stepwise multiple regression analysis indicated that only contact frequency had significant effects on ShBI in each trial. Shoot number and plant height showed significant effects on sheath blight in the 2010 late season (Table 9). However, canopy structure parameters including shoot number, LAI and biomass at the two considered stages and contact frequency had no significant relationships with lesion length or ShBI on inoculated plants during both seasons of 2010 (data not shown). Figure 1 shows that ShBI increased in a linear fashion as contact frequency increased, whilst lodging index and grain yield levelled off after a contact frequency of 400.

Table 7. Correlation coefficients between sheath blight index measured in uninoculated plots and canopy structure parameters in each trial in 2009 and 2010
SeasonsPanicle initiation stageFlowering
Shoot numberLeaf area indexBiomassShoot numberLeaf area indexBiomass
  1. Levels of significance indicated: ns = not significant, *significant at  0·05, **significant at  0·01, = 6 and 9 in 2009 and 2010, respectively.

2009 Early season0·95**0·96**0·95**0·93**0·99**0·99**
2009 Late season0·90*0·93**0·91*0·97**0·96**0·84*
2010 Early season0·83**0·80**0·81**0·93**0·92**0·73*
2010 Late season0·85**0·90**0·67*0·78*0·62 ns0·72*
Table 8. Relationship of sheath blight index measured from uninoculated plants with contact frequency, light transmittance, SPAD values, plant height and dry weight of dead leaves and sheaths (DLS) in each trial in 2009 and 2010
Traits2009 ESa2009 LS2010 ES2010 LS
  1. Levels of significance indicated: ns = not significant, *significant at  0·05, **significant at  0·01, = 6 and 9 in 2009 and 2010, respectively.

  2. a

    ES and LS represent early and late season, respectively.

  3. b

    Data were not investigated.

  4. c

    Correlation coefficients between sheath blight index and plant height was calculated using data only from nitrogen application treatment.

  5. d

    FL, flowering.

Contact frequency 0·88* 0·94**0·91** 0·73*
Light transmittanceb–0·93**–0·92**
SPAD values–0·81**–0·35 ns
Plant heightc–0·96 ns–0·99**–0·63 ns–0·34 ns
DLS at FLd0·93** 0·55 ns
DLS at 10 d after FL0·95** 0·80*
Table 9. Stepwise regression analyses of the effects of shoot number, leaf area index, biomass production at panicle initiation (PI) and flowering, SPAD values, dry weight of dead leaves and sheaths (DLS), contact frequency (CF), plant height and light transmittance on sheath blight index (ShBI) in uninoculated plants in each trial in 2009 and 2010
SeasonsRegression model R 2 P for ln(CF)
  1. **Significant at ≤ 0·01.

  2. a

    ES and LS represent early and late season, respectively.

2009 ESaShBI = −16·44 + 7·85 × ln(CF)0·95**0·001
2009 LSShBI = −105·60 + 19·94 × ln(CF)0·80**0·017
2010 ESShBI = −195·24 + 42·59 ×  ln(CF)−0·12 × (Shoot no. PI)0·93**0·004
2010 LSShBI = 335·98 + 16·70 ×  ln(CF)−3·19 × (plant height)0·83**0·002
Figure 1.

Relationships of contact frequency with sheath blight index, lodging index and grain yield across each trial in 2009 and 2010. ** indicates significant at  0·01. ES and LS represent the early and late season trials, respectively.

Discussion

In uninoculated plants, ShBI in each trial and RLH in 2009 increased significantly with increasing N rate. Also ShBI and RLH significantly increased for most cases under N application with increasing hill density. These results are consistent with previous reports that sheath blight severity increased in rice grown under high N input and dense plant populations (Savary et al., 1995; Willocquet et al., 2000). Similar to ShBI in uninoculated plants, canopy structure parameters such as shoot number, LAI and biomass increased with the increase of N rate and hill density. Canopy structure in this study was different under various N rates and hill densities as a result of the differences in shoot number, LAI and biomass. Canopies with more shoots, higher LAI, and larger biomass usually have dense canopy structure, which in turn lead to severe sheath blight incidence (Zhong et al., 2006). This result can be attributed to the strong correlation of ShBI with shoot number, LAI, biomass, light transmittance, contact frequency, plant height, SPAD value, dead leaves and sheaths.

One needs to be cautious while establishing the cause–effect relationships between ShBI and canopy structure parameters, because these traits are themselves correlated with each other or closely correlated with other variables. Some of these parameters would promote an increase in disease severity, while others are probably the result of a severe disease. Stepwise multiple regression analysis indicated that contact frequency was a significant factor in the regression model in each trial. This indicates that contact frequency is a key factor for sheath blight related to canopy structure. This is consistent with previous reports that contact frequency between host tissues provides a physical bridge for the running hyphae strands to progress (Savary et al., 1995; Castilla et al., 1996). On the other hand, shoot number and plant height were significant in the regression model in 2010 early and late seasons, respectively, which agree with the previous findings that plant height and shoot number also play a vital role in sheath blight development (Han et al., 2003; Zhong et al., 2006). It seems logical to hypothesize that short internodes combined with excessive shoot number result in more leaves per unit space. These two factors in turn increase the possibility of leaf-to-leaf contact, which increases the opportunities for the running hyphae to progress upward. Moreover, SPAD values decreased and the amount of dead leaves and sheaths increased under severe disease pressure associated with dense canopy structure.

Under uninoculated conditions, a dense canopy structure facilitates the horizontal spread of the disease by increasing contact frequency in a leafborne phase of the pathogen and can also speed up the vertical spread of the fungus as a result of increased humidity in the plant canopy (Savary et al., 1995; Castilla et al., 1996; Willocquet et al., 2000; Han et al., 2003; Tang et al., 2007). Also, severe sheath blight under a dense canopy structure decreases the leaf nitrogen status of each plant, leading to more dead leaves and sheaths in the lower internodes. This results in a decrease in photosynthetic ability and carbohydrate production, which can cause a large yield loss from sheath blight (Marchetti & Bollich, 1991; Wu et al., 2013). The associated yield reduction is further aggravated by the additional losses resulting from increased lodging (Wu et al., 2012). This phenomenon was observed in uninoculated plots, where movement of the pathogen between shoots belonging to different or the same hills plays a substantial role in the development of sheath blight. However, in inoculated plots, sheath blight severity was not necessarily restricted by infection between different hills because enough hyphae for progress were available in each infected shoot, so severity was mainly determined by vertical spread of the pathogen within each hill. It can be argued that, although the inoculated area varied between different hill densities, the sheath blight development could not be influenced by this variable, because each inoculated hill has the same fungal biomass provided by the toothpick inoculum. Willocquet et al. (2000) suggested that the spatial distribution of shoots was strongly aggregated within each hill in the case of transplanted rice, and this greater aggregation of shoots than in direct-seeded rice would lead to microclimate conditions within a hill that are more favourable for sheath blight development regardless of variation in plant density. The present study concludes that differences in canopy structure resulting from planting densities and N rates have a smaller effect on sheath blight severity in inoculated plots than in uninoculated plots.

Generally, disease resistance of the host is genetically inherited (Pan et al., 1999; Prasad & Eizenga, 2008), and should not be affected by changes in the canopy structure for a genotype. Similarly, in the present study, there was no difference in sheath blight resistance as reflected by lesion length among various canopy structures. It can be concluded that plant traits affected sheath blight epidemic in the case of naturally infected field conditions by changing canopy structure, as reflected by contact frequency, rather than changing sheath blight susceptibility of individual plants.

The ShBI in 2009 was lower than in 2010 at an equivalent N rate which mainly resulted from different canopy structure parameters, where the contact frequency and LAI were lower in 2009 than in 2010. Moreover, the total rainfall and RH were higher in 2010 than 2009 and therefore conditions were more conducive for sheath blight (Savary et al., 1995; Castilla et al., 1996; Zhong et al., 2006). Lesion length and ShBI of the early season in 2010 were comparable with those of late season in 2010 on inoculated plants, although the climatic conditions and canopy structure were different between these two seasons. In addition to the above-mentioned climatic factors and canopy structure parameters, Savary et al. (1998) proposed that the quantity of soilborne inoculum can be an important factor when trying to explain the difference between seasons and years.

Mean ShBI in the late season was 26% higher than in the early season across N rates and hill densities in uninoculated plants in 2010. Average air temperature and mean relative humidity (RH) were consistently higher in the late season than the early season, especially from TR to PI. Furthermore, canopy structure parameters such as shoot number, LAI and biomass of the late season plants were higher than those of the early season. These two aspects can favour disease development in the late season (Savary et al., 1995; Castilla et al., 1996; Zhong et al., 2006). However, in 2009 there was no consistent difference in ShBI between the early and late season trials, which could be the result of shorter plants in the early season; a trait conducive for the running hyphae to progress towards the top three leaves (Han et al., 2003).

In order to get a further understanding of R. solani epidemiology, additional research is needed to address the spatiotemporal distribution and infection response of this fungus to changes in dynamic canopy structure. This topic will become an important field of research because of the advances in experimental and modelling activities which enable interdisciplinary work when tackling complex systems, and the societal pressure to develop environmentally friendly disease control strategies (Tivoli et al., 2013).

Contact frequency is a key epidemiological factor for sheath blight, and can be modified by altering canopy parameters to help with management of the disease. Such an approach could complement the QTL identification for genetic and physiological resistance because the QTL mapping can be confused by canopy related traits (Srinivasachary et al., 2011). It can be regarded as an important indicator for recommending preventive use of a fungicide because greater contact frequency increases the probability of disease spread upwards. Lower sheath blight severity can be achieved in a sparse canopy structure, where contact frequency and risk of lodging are reduced (Fig. 1a,b). However, a sparse canopy would limit the biomass production and subsequently suppress grain yield (Fig. 1c) because the harvest index is quite steady around 0·5 (Ying et al., 1998). Hence, a consideration of the balance between disease pressure, lodging resistance and grain yield is important. For instance, when contact frequency ranged from 500 to 1000, grain yield was not restricted substantially by the relatively sparse canopy structure, but the disease pressure and risk of lodging were kept controlled to some extent (Fig. 1). An increase in the contact frequency beyond 1000 will have no positive effect on yield but lodging index, and especially disease pressure, will markedly increase.

Resistance to R. solani is a complex trait that is affected by heritability of resistance or other plant traits (Han et al., 2003). Breeding for genetic improvement is the most promising strategy in managing sheath blight (Pan et al., 1999; Srinivasachary et al., 2011). However, completely resistant genotypes to this pathogen have not been found (Prasad & Eizenga, 2008; Willocquet et al., 2012). Importantly, appropriate crop management practices, including environmentally friendly strategies such as rational use of fertilizers and optimum planting density resulting in a ‘healthy’ canopy structure, should be given priority for managing sheath blight in rice.

Acknowledgements

The authors are grateful to N. Paveley, C. Robert, D. Fernando and the anonymous reviewers for their valuable comments and suggestions which helped to improve this paper. They also thank the National Basic Program of China (No. 2009CB118605), the National Key Technology Program R&D of China (Project No. 2012BAD04B12), and Hubei provincial special fund for the Innovation Team (2010CDA080) for financial support.

Ancillary