SEARCH

SEARCH BY CITATION

Keywords:

  • direct seeding;
  • disease spread;
  • Rhizoctonia;
  • rice cultivation;
  • sheath blight;
  • transplanting

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Establishment methods for rice crops in tropical Asia are very diverse, leading to variation in the structure of rice canopies. Differences in canopy structure can in turn affect the spread of the rice sheath blight pathogen, Rhizoctonia solani. Rice sheath blight epidemics were compared during two seasons in crops established by different methods: direct broadcasting of pregerminated rice seeds, and transplanting of rice seedlings at spacings of 20 × 20 cm, 13 × 25 cm and 25 × 25 cm between hills (i.e. along and between rows, respectively). In both years, the apparent infection rate based on incidence data and the terminal severity of sheath blight were lower in the direct-seeded crops than in any of the transplanted ones, regardless of spacing. The frequency of leaf-to-leaf contacts (CF) between hills (or plants) was highest in direct-seeded rice, and lowest in rice transplanted at a spacing of 25 × 25 cm. Larger CF is known to favour rice sheath blight epidemics. The apparent contradiction between higher incidence and lower CF in the transplanted stands than in the direct-seeded stands is interpreted in terms of accessibility of healthy host tissues to the spread of the pathogen in the canopy, and accounts for within-host (rice hill or plant) and between-host (hill or plant) disease spread. The analysis of incidence-severity relationships indicated a less aggregated distribution of the disease in direct-seeded rice, which was related to the spatial distribution of the tillers. These findings have direct implications for the management of the disease.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Sheath blight, caused by Rhizoctonia solani, has become an important disease of rice in tropical Asia, especially in intensive rice cropping systems ( Teng et al., 1990 ; Savary et al., 1994 ). Cultivars highly resistant to this pathogen ( Xie et al., 1992 ; McClung et al., 1997 ) are not currently available in Asia, and so far control of the disease has relied mainly on the use of fungicides, when affordable by farmers ( Savary & Mew, 1996). However, this option is not considered sustainable because of residue problems and the potential risk of emergence of pathogen populations that are resistant to fungicides. Other options to control sheath blight are necessary, and these include cropping practices that may minimize the disease.

Rice crop establishment methods in Asia are very diverse, and include direct seeding by broadcasting, and transplanting of seedlings at different spacings. These different crop establishment methods lead to differences in the structure of the rice crop canopy, i.e. in crop geometry. In the case of a direct-seeded crop established by broadcasting rice seeds, the crop consists of randomly distributed plants with three to five tillers each. Such a crop can thus be seen as a population of tillers that are slightly aggregated at a scale corresponding to the area occupied by one plant. In the case of transplanted rice, the crop consists of uniformly distributed hills, each having 10–40 tillers. A transplanted rice crop, by contrast, can thus be seen as a population of tillers that are strongly aggregated at a scale corresponding to the area occupied by one hill (which is usually larger than the area occupied by one direct-seeded plant).

During the leaf-borne phase of sheath blight epidemics, the fungus spreads by means of runner hyphae that originate from lesions and progress towards healthy tissues ( Ou, 1987). These healthy tissues may belong to the same tiller, the same ‘host aggregation unit’ (whether a transplanted hill or a direct-seeded plant), or a different unit. Contacts between infected and healthy tissues are therefore a prerequisite for disease spread. The frequency of leaf-to-leaf contacts represents an important epidemiological factor for sheath blight ( Savary et al., 1995 ; Castilla et al., 1996 ). This factor may be affected by crop geometry. Sheath blight epidemics are favoured by close spacing of hills in the case of transplanted rice ( Kannaiyan & Prasad, 1983; Castilla et al., 1996 ), and by high sowing rates in the case of direct-seeded rice ( Mithrasena & Adikari, 1986; Guzman Garcia & Nieto Illidge, 1992). A positive effect of dense canopy on diseases caused by Rhizoctonia solani has also been shown in other crops such as potato ( Firman & Allen, 1995), soybean ( Yang et al., 1990 ) and tall fescue ( Giesler et al., 1996a ). Such canopy–disease relationships reflect the combined effects of microclimate and canopy structure on the spread of this type of disease ( Yang et al., 1990 ; Savary et al., 1995 ; Giesler et al., 1996b ).

The objective of this study was to compare sheath blight epidemics and key epidemiological factors in rice crops with different crop establishment methods and therefore in crop stands that differ in their canopy geometry.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The experiment was conducted in a farmer’s field near Pila, Laguna, Philippines, during the rainy season in 1996 and during the dry season in 1997. IR 72, a semidwarf, high-yielding, short-cycle cultivar, was used in both experiments.

Experimental treatments

Four crop establishment methods were compared: broadcast direct-seeded rice (DS); and transplanted rice, with spacing across and between rows of 20 × 20 cm (TR20×20), 13 × 25 cm (TR13×25) and 25 × 25 cm (TR25×25).

The experimental design was a randomized complete-block design with four replications. Each individual plot was 3 × 3 m in size, and the distance between plots was 1 m. All observations were taken in the 1 × 1-m square located at the centre of the plots, referred to subsequently as the central area. The area between plots was called the buffer area.

Crop management

In 1996, direct sowing was carried out on July 19 and transplanting on July 30. In 1997, these operations were conducted on February 5 and February 11, respectively. In the transplanted crop establishment, 10-day-old seedlings were planted at a rate of 7–8 seedlings per hill. In the direct-seeded crop establishment, seeds were soaked overnight in water, incubated for two days in jute bags and broadcast at 90 kg ha−1. A pre-emergence herbicide (Butachlor) was applied before crop establishment in the direct-seeded plots. The entire experimental area was flooded with 5 cm of water above ground from 20 days after crop establishment until 15 days before harvest, when the field was drained. Time was scaled in DAT, the number of days after transplanting in transplanted plots.

The experimental area was irrigated with water flowing from a fish-pond that also provided most of the nitrogen supply required for crop growth. No nitrogen was applied in 1996, and in 1997 urea was applied at 23 and 42 DAT at a rate of 50 kg N/ha.

Rice seedlings were transplanted into the buffer area on the same date as in the transplanted plots, with 20 × 20-cm spacing.

Establishment of sheath blight in individual plots

Rice seedlings were potted at the same time as transplanting into the field, and were grown in a glasshouse. At the tillering stage, they were inoculated by inserting at their base a rice grain/rice hull mixture colonized by mycelium of an R. solani anastomosis group AG1–1A isolate ( Sharma & Teng, 1990; Savary et al., 1995 ). After one week of incubation, the infected plants were used as disease sources. One heavily infected plant was transplanted to the centre of each plot, 35 DAT in 1996, and 38 DAT in 1997. The infection source was removed once sheath blight lesions were observed on the hills or plants surrounding it, which occurred at 7 days after the transplanting of the infected plant in 1996 and 4 days after in 1997.

Assessment and computation of sheath blight incidence and severity

Assessments of sheath blight were done weekly throughout the growing season on hills or plants located in the central area. A tiller was operationally defined as a tiller having at least two fully expanded leaves. The number of tillers m−2 was calculated from counts of tillers in 10 and 20 host aggregation units per plot in the transplanted and direct-seeded stands, respectively. Incidence at the tiller level was defined as: INC = 100 Fh Fit, where Fh is the fraction of infected host aggregation units (hill or plant), and Fit is the fraction of infected tillers in the diseased host aggregation units. Sheath blight severity was defined as: SEV = Sit Fit, where Sit is the average severity in infected tillers.

In the transplanted plots, the fraction of infected hills, Fh, was calculated from the assessment of disease occurrence in all hills standing in the central area. The fraction of infected tillers in the diseased hills, Fit, was computed on the basis of counts of diseased and healthy tillers in ten infected hills selected at random. If fewer than ten diseased hills were observed in a plot, the fraction of diseased tillers was measured in all diseased hills. The average severity in infected tillers, Sit, was computed on the basis of the assessment of sheath blight severity on five infected hills selected from the ten sampled hills. When fewer than five infected hills were observed, severity was assessed on all infected hills. Severity was assessed on one infected tiller per hill as the percentage of total tiller area (stem plus leaves) covered by sheath blight lesions.

In the direct-seeded plots, the fraction of infected plants, Fh, was calculated from the assessment of disease occurrence in all plants standing in one-eighth of the central area. Within this area, the same sampling scheme and computations as in transplanted plots were followed, except that twenty plants were sampled to assess sheath blight incidence, and one tiller per plant in ten infected plants was used to assess sheath blight severity.

Assessment of contact frequency

Assessments of contact frequency were done in all plots of both experiments at the tillering and panicle initiation stages. Counts were made differently depending on the crop establishment method. Contact frequency between hills was measured in transplanted plots, and contact frequency between plants was measured in direct-seeded plots.

In the transplanted plots of the 1996 experiment, two hills per plot were randomly selected within the central area. The total number of leaf-to-leaf and leaf-to-sheath contacts between each sampled hill and all eight hills surrounding it were counted ( Savary et al., 1995 ).

In the transplanted plots of the 1997 experiment, four hills per plot were randomly selected in the central area. In the TR20×20 and TR25×25 treatments, leaf-to-leaf and leaf-to-sheath contacts were counted between each selected hill and two of its neighbours (one on the same row and one along the diagonal). In the TR13×25 treatment, leaf-to-leaf and leaf-to-sheath contacts were counted between each selected hill and three of its neighbours (one on the same row, one on the diagonal and one on the adjacent row, in front of the sampled hill).

In the direct-seeded plots of the 1996 and 1997 experiments, four plants per plot were randomly selected within the central area. The total number of leaf-to-leaf and leaf-to-sheath contacts between the sampled plant and one plant surrounding it were counted.

Computation of contact frequency per square metre

The total number of contacts (sheath-to-leaf plus leaf-to-leaf) per m2 (CF) was estimated as follows, using assessments of contact frequency: if NB is the total number of hills m−2, x is the total number of contacts counted between one hill and its neighbours, and n is the number of neighbours for which counts were made, then:

1x/n is the number of contacts between one hill and one of its neighbours;

2 8x/n is the number of contacts between one hill and all its neighbours;

3 8x/(2n) is the average number of contacts between one hill and all its neighbours, because contacts occurring between hill A and hill B are the same as those occurring between hill B and hill A. Thus it follows that CF = 8x NB/(2n), and for 1996 (n = 8), CF = x NB/2, for 1997, TR20×20 and TR25×25 plots (n = 2), CF = 2x NB, and for 1997, TR13x25 plots (n = 3), CF = 1·33x NB, with NB = 25, 16 and 30·8 for TR20×20, TR25×25 and TR13×25, respectively.

In the case of direct-seeded rice, a plant was assumed to be surrounded by six plants. Thus: CF = 6x NB /2, where NB is the number of plants m−2 at each assessment of contact frequency.

For each year, the means of the contact frequency over the four replications were compared within each assessment date, according to the Student–Newman–Keuls multiple range test.

Analyses of disease progress curves

The disease progress curves of sheath blight incidence and sheath blight severity were described with the logistic model dy/dt = ay (1 – y), where y is the disease incidence or the disease severity, and a represents the apparent infection rate ( Van der Plank, 1963).

Linear regressions were done using the equation logit(y= at + b. Separate regressions were done in the two years for the different crop establishment treatments. The parameters estimated from the regressions were compared using a t-test ( Campbell & Madden, 1990).

Regression analyses

Two linear regressions were tested using a stepwise, upward procedure in order to synthesize the relationships between disease incidence, disease severity and contact frequency over the two experiments. The regressions have the following shapes:

  • image
  • image

where CE is a binary variable equal to 0 for direct-seeded stands, and equal to 1 for transplanted stands; CF is the contact frequency at panicle initiation; and INC and SEV are the maximum disease incidence and severity, respectively, observed in each individual plot of both experiments. Thirty-two observations were thus considered in each regression.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Crop establishment method and sheath blight dynamics

1996 experiment

Crop development was faster and the number of tillers was greater in the direct-seeded plots than in transplanted ones. The number of tillers m−2 increased to maximum values of 600, 800, 800 and 1800 for the TR25×25, TR20×20, TR13×25 and DS treatments, respectively; these values were reached at 42, 35, 35 and 21 DAT, respectively. The number of tillers then declined and stabilized at about flowering stage (with 500 and 1100 tillers m−2 for transplanted and direct-seeded treatments, respectively). In direct-seeded plots, the average distance between two neighbouring plants was approximately 5 cm.

Spontaneous sheath blight infections were not observed prior to the establishment of disease sources in the plots. Sheath blight incidence increased rapidly afterwards, and reached 80% (DS treatment) to 100% (transplanted treatments) at the end of the cropping season, with a sigmoidal pattern of disease progress ( Fig. 1A). The sheath blight epidemic in TR25×25 was delayed compared with the two other transplanted stands, and disease incidence in this treatment was the lowest among transplanted stands when disease spread was the fastest, that is, at 49–63 DAT. Disease incidence was significantly (P = 0·05) lower in TR25×25 than in the two other transplanted stands at 49 and 56 DAT. The logistic model described the progress curves of incidence quite well in all crop establishment treatments ( Table 1). The apparent infection rate (parameter a) was significantly (P = 0·05) higher in the TR20×20 treatment than in the TR13×25 and TR25×25 treatments, and was significantly (P = 0·05) lower in the direct-seeded treatment than in all the transplanted treatments.

image

Figure 1. Sheath blight incidence (A) and sheath blight severity (B) in four different rice crop establishment treatments in the rainy season 1996; sheath blight incidence (C) and sheath blight severity (D) in four different rice crop establishment treatments in the dry season 1997.

Download figure to PowerPoint

Table 1.  Linear regression analyses a of sheath blight incidence (logit) by the number of days after crop establishment in the four crop establishment treatments in 1996 and 1997
YearTreatmentd.f. bacGrouping dbcGrouping dFPR2e
  • a

    The equation tested was: logit(INC/100) = at + b, where t is the number of days after transplanting (DAT).

  • b

    Number of degrees of freedom of the regression.

  • c

    Estimates are followed by their standard error.

  • d

    Each year, parameter estimates followed by a different grouping letter are significantly (P = 0·05) different according to the t-test.

  • e

    Coefficient of determination.

1996TR25×2570·174(0·010)B− 10·07(0·54)B 3150·00010·98
TR20×2060·248(0·016)A− 13·66(0·81)A 2420·00010·98
TR13×2580·166(0·016)B − 9·75(0·92)B 1120·00010·93
DS60·107(0·021)C − 6·72(1·39)B  240·00430·80
1997TR25×2540·343(0·018)A− 17·06(0·91)A 3520·00030·99
TR20×2040·376(0·028)A− 18·62(1·41)A 1770·00090·98
TR13×2540·338(0·021)A− 16·58(0·91)A 2550·00050·98
DS50·236(0·003)B− 12·85(0·19)B49270·00010·99

The three transplanted treatments showed similar severity progress curves, whereas the slope of severity increase was lower in the direct-seeded plots ( Fig. 1B). The terminal severities, at 84 DAT, were 15%, 25%, 35% and 35% for the DS, TR13×25, TR20×20 and TR25×25 treatments, respectively ( Fig. 1B). Severities at 77 DAT and 84 DAT were significantly (P = 0·05) lower in the direct-seeded treatment than in the transplanted treatments, according to the Student–Newman–Keuls multiple range test. The logistic model described the progress curves of severity reasonably well ( Table 2). The slope was significantly (P = 0·05) lower in direct-seeded (0·11) than in transplanted plots (0·13–0·18).

Table 2.  Linear regression analyses a of sheath blight severity (logit) by the number of days after crop establishment in the four crop establishment treatments in 1996 and 1997
YearTreatmentd.f. bacGrouping dbcGrouping dFPR2e
  • a

    The equation tested was: logit(SEV/100) = at + b, where t is the number of days after transplanting (DAT).

  • b

    Number of degrees of freedom of the regression.

  • c

    Estimates are followed by their standard error.

  • d

    Each year, parameter estimates followed by a different grouping letter are significantly (P = 0·05) different according to the t-test.

  • e

    Coefficient of determination.

1996TR25×2580·129(0·013)A− 10·60(0·75)A103·00·00010·93
TR20×2080·178(0·028)A− 13·98(1·67)A39·00·00040·83
TR13×2580·136(0·020)A− 11·16(1·19)A45·00·00030·85
DS60·107(0·020)B− 10·12(1·30)A28·00·00320·82
1997TR25×2560·133(0·044)A− 11·30(2·68)A9·00·03010·57
TR20×2060·130(0·045)A− 11·05(2·70)A8·40·03400·55
TR13×2570·158(0·042)A− 13·07(2·42)A14·00·00960·65
DS60·074(0·020)A − 8·04(1·20)A14·00·01370·63
1997 experiment

A few spontaneous sheath blight infections were observed before the establishment of disease sources. At 35 DAT, 0·36, 0·45, 1·38 and 0·01% of the tillers were infected in the TR20×20, TR25×25, TR13×25 and DS treatments, respectively.

The dynamics of the number of tillers were similar to those observed in 1996, but the number of tillers was larger for all four treatments in 1997 than in 1996, especially in the transplanted rice treatments. The number of tillers m−2 increased to maximum values of 850, 900, 980 and 1900 for the TR25×25, TR20×20, TR13×25 and DS treatments, respectively. The number of tillers then declined and stabilized at about flowering stage (with 800 and 1200 tillers m−2 for transplanted and direct-seeded treatments, respectively).

In all treatments, there was an increase in sheath blight incidence from 0 to 100% within four to five weeks, with a sigmoidal pattern of disease incidence progress ( Fig. 1C). The curves for the three transplanted treatments were nearly identical. The epidemic was delayed by one week in the direct-seeded plots, compared with transplanted plots. As in 1996, the apparent infection rate was significantly (P = 0·05) higher in transplanted treatments (0·34–0·38) than in the direct-seeded treatment (0·24) ( Table 1).

Sheath blight severity increased until 63 DAT and then remained stable in the transplanted plots ( Fig. 1D). The increase in sheath blight severity was slower in the direct-seeded plots than in the transplanted ones. This was reflected by higher apparent infection rates (based on severity data) in transplanted plots ( Table 2). No significant difference between the apparent rates of infection in the different treatments was found, due to the relatively large standard errors of the parameter estimates ( Table 2). In all assessments between 49 DAT and 84 DAT, severity was significantly (P = 0·05) lower in the DS treatment than in any of the TR treatments, according to the Student–Newman–Keuls multiple range test.

Repeated-measure analyses of variance

Longitudinal analyses ( Zadoks, 1972) were performed on disease incidence and severity progress data for both years, using a repeated-measure anova design ( Madden, 1986) ( Table 3). While the crop establishment method had no effect on disease incidence in 1996, a significant (P < 0·05) effect on incidence was found for the 1997 epidemics. Significant (P < 0·01) effects of crop establishment methods on the overall variation in severity were found in both years, as well as significant interactions between crop establishment method and observation date (1996: P = 0·05; 1997: P = 0·01), indicating increasing differences among crop establishment methods over time ( Fig. 1). When the same analyses were performed after removal of data pertaining to the direct-seeded treatment, none of the above effects were significant, indicating that the contrast between the DS treatment and all TR treatments accounted for most of the effects detected in the complete anova.

Table 3.  Repeated-measures analyses of variance for sheath blight incidence and sheath blight severity progress data on four crop establishments taken in 11 assessments in 1996 and 10 assessments in 1997
 19961997
  IncidenceSeverity IncidenceSeverity
Source of variationd.f. aMSFPMSFPdf aMSFPMSFP
  • a

    Degrees of freedom.

  • b

    Fischer ratios compared with conservative Fischer values [with d.f.(A) = 1; Madden, 1986].

Crop estab. (C)  30·841  2·110·170·3318·280·006  30·4864·160·040·14214·60·0008
Replication (R)  31·301  3·270·070·1774·420·036  30·0690·590·640·0121·260·35
Error a  90·398  0·04    90·117  0·010  
Assessment (A) 1023·773280·00< 0·01 b2·971253·00< 0·01 b  930·818720·00< 0·01 b1·880518·00< 0·01 b
A*C 300·151  1·78NS b0·0393·350·05 b 270·099230·00NS b0·0308·210·01 b
Error b1200·085  0·012  1080·043  0·004  

Contact frequency

In both years, contact frequency was significantly (P = 0·05) lower in the TR25×25 treatment than in the other treatments, and was higher in the DS treatment than in the other treatments at the two development stages considered ( Fig. 2). Contact frequency in the TR20×20 and TR13×25 treatments did not differ significantly, and was intermediate (P = 0·05) between that of the DS and TR25x25 treatments. The contact frequency in direct-seeded plots ranged from 10 to 100 times that observed in transplanted plots. Contact frequency increased between tillering and panicle initiation stages and was between 1·7- and 3·8-fold higher in 1997 than in 1996 in all the treatments, reflecting the profuse growth of the canopy in the dry season (1997 experiment) compared with that in the rainy season (1996 experiment).

image

Figure 2. Numbers of leaf-to-leaf and leaf-to-sheath contacts per square metre between hills (transplanted plots) or plants (direct-seeded plots)(logarithmic scale), in 1996 (A) and 1997 (B). For each development stage, bars labelled with different letters are significantly (P = 0·05) different according to Student–Newman–Keuls multiple range test done on log-transformed values.

Download figure to PowerPoint

Linear regressions

In the first regression, the three parameters were significantly different from zero, a being negative, and b and c being positive. The model gave a fairly good description of incidence variation (R2 = 0·59). In the second regression, ln (CF) was not retained, the constant (d) was negative, and the two other parameters (f and g) were positive. This second regression gave a good description of severity variation (R2 = 0·69). The first regression describes disease incidence, i.e. exodemic, between–tiller disease increase, as a function of increasing contact frequency for a given crop establishment method. Disease incidence is lower in direct-seeded stands than in transplanted ones. The second regression describes disease severity, i.e. esodemic, within–tiller disease increase, as a function of increasing incidence in a given crop establishment, irrespective of contact frequency. Disease severity is lower in direct-seeded stands than in transplanted ones.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Comparison of direct-seeded and transplanted rice

In these experiments, direct-seeded and transplanted stands differed in their canopy structure: the direct-seeded stand was characterized by (a) more tillers m−2 than transplanted stands; (b) a shorter distance between neighbouring plants (around 5 cm) than between neighbouring hills (at least 13 cm); (c) in consequence, more frequent contacts between plants than between hills; and (d) fewer tillers per host aggregation unit than transplanted stands. Direct seeding of the crop thus leads to a more shady and probably more humid microclimate in the canopy, and to more contacts between plants. These factors are favourable to sheath blight epidemics ( Savary et al., 1998 ). In spite of this, the different analyses done on sheath blight incidence and sheath blight severity based on the data from this study indicate that direct-seeded rice was less conducive to sheath blight epidemics than transplanted rice. Other epidemiological factors therefore need to be considered in order to explain this apparent contradiction. The fungus spreads from an infected tiller to a healthy one by growth of hyphae on areas where the two tillers are in close contact ( Ou, 1987; Savary et al., 1995 ). The spatial distribution of tillers should thus be expected to greatly affect disease spread. The spatial distribution of tillers in a direct-seeded crop is slightly aggregated at the plant scale, but it is strongly aggregated within each hill in the case of transplanted rice. The spread of the disease in transplanted rice involves two stages: spread between tillers in the same hill, and spread between tillers belonging to different hills. The first stage (within-hill spread) presumably is much faster than the second (between-hill spread), which in turn is analogous to spread from plant to plant in a direct-seeded crop. When incidence is assessed, the two stages are combined in their speeds, and the faster spread in a transplanted crop is predicted. The spatial distribution of tillers may thus explain why direct-seeded stands are less conducive to sheath blight than transplanted ones. It may be also hypothesized that the greater aggregation of tillers within a (transplanted) hill than within a (sown) plant would lead to a microclimate more favourable to sheath blight.

The difference observed in these experiments between the direct-seeded and the transplanted crop establishments is in agreement with results obtained at a regional scale. A survey conducted in 456 farmers’ fields in different countries of Asia (China, India, Philippines and Vietnam) allows analyses of relationships between cropping practices and rice diseases ( Savary et al., 2000 ). More specifically, relationships between sheath blight and crop establishment can be studied. A contingency table of crop establishment by sheath blight levels was constructed ( Table 4). The chi-square value obtained (37·8, d.f. = 2, P = 0·0005) indicates that crop establishment and sheath blight levels are strongly associated. The comparison of observed and expected values ( Table 4) indicates that, while crop establishment does not affect the occurrence of sheath blight in a field (in SHB0 (no sheath blight) , the expected values are close to the observed ones), sheath blight incidence is lower in direct-seeded fields than in transplanted ones.

Table 4.  Contingency table of the number of fields recorded for their levels of sheath blight infection under different methods of crop establishment
SheathCrop establishments
blight levels aDirect-seeded bTransplanted b
  • a Sheath blight levels were derived from assessments in each field at tillering, booting, early dough and maturity (Savary et al. 2000) . The maximum sheath blight incidence (% infected tillers) over the four assessments was computed. SHB0, no sheath blight; SHB1, moderate sheath blight incidence > 0% and ≤ 5%; SHB2, high sheath blight incidence, > 5%.

  • b

    Observed values are followed by the expected values.

SHB055(52·1)70(72·9)
SHB166(41·7)34(58·3)
SHB269(96·3)162(134·8)

Incidence–severity relationships

Incidence-severity relationships can provide useful information on disease spread and spatial structure ( Seem, 1984). In both seasons, disease incidence increased with disease severity, reaching a plateau with incidence approaching 100% ( Fig. 3). The relationships between sheath blight severity and sheath blight incidence were analysed using a hyperbolic tangent regression model ( Fig. 3). It provided a good fit of the estimated incidences to the observed ones ( Fig. 3). The estimated values of the regression parameter α were 1·3 and 3·2 for the transplanted stands in 1996 and 1997, and 5·8 and 12·0 for direct-seeded stands in 1996 and 1997, respectively. In both seasons, incidence increased with severity more rapidly in the direct-seeded plots than in the transplanted plots. This was reflected by a significantly (P = 0·05) higher value of the regression parameter α in the direct-seeded treatment, according to the t-test ( Campbell & Madden, 1990). The relationships between incidence and severity indicate that at a given level of disease incidence, disease severity was lower in direct-seeded plots than in the transplanted ones, i.e. an increment in incidence corresponds to a faster increase in severity in the transplanted than in the direct-seeded stands, when incidence is low. This suggests that transplanted rice is more conducive than direct-seeded rice to disease intensification within a tiller. It may also suggest that the increase in severity in a direct-seeded stand results to a large extent from alloinfection, whereas autoinfection plays a more important role in the increase of severity in transplanted stands ( Seem, 1984). The relationships between incidence and severity ( Fig. 3) suggest that the disease was more aggregated in transplanted rice ( McRoberts et al., 1996 ). This can be explained by the more aggregated spatial distribution of the tillers in transplanted crops; given that spread of the fungus requires contact between tillers, it may be assumed that disease spatial structure is strongly linked with tiller spatial structure. Disease was less aggregated in 1997 than in 1996. This may be due to the presence of more abundant, randomly located, natural infections.

image

Figure 3. Estimated and observed sheath blight incidence in 1996 (A) and 1997 (B). Sheath blight incidence was estimated using the regression equation INC = tanh(α SEV) = (exp(α SEV) − exp(– α. SEV)) / (exp (α SEV) + exp(– α SEV)). INC and SEV are disease incidence and disease severity, respectively, expressed as proportions. EIDS: estimated incidence for direct-seeded plots; EIDR: estimated incidence for transplanted plots.

Download figure to PowerPoint

Comparison of the transplanted stands

The sheath blight epidemic in the TR25×25 treatment was slightly delayed compared with the TR20×20 and TR13×25 treatments in 1996. This may be explained by the lower density of tillers and the lower contact frequency in TR25×25 plots than the others. These results agree with other reports that showed that close spacing ( Kannaiyan & Prasad, 1983) and high contact frequency ( Savary et al., 1995 ; Castilla et al., 1996 ) were favourable to sheath blight in transplanted rice. In 1997, in spite of a lower tiller density and lower contact frequency in the TR25×25 treatment, the three spacing treatments had similar epidemics. Crop growth was faster in 1997 than in 1996. These results suggest that, in 1997, crop growth was such that near-optimum conditions were reached for sheath blight epidemics in the three spacing treatments, leading to similar epidemics in them. These findings agree with reports showing that sheath blight is favoured in intensive cropping practices (e.g. Savary et al., 1994 ). The faster growth observed in 1997 may be due to more solar radiation and larger nitrogen inputs.

Comparison of the two experiments

Sheath blight epidemics were faster in the 1997 dry season than in the 1996 rainy season. This may be partly attributed to faster rice growth and to the presence of natural inoculum in the plots that caused primary infections in 1997. This is consistent with monitoring of the soil-borne inoculum in the same field in 1995 and 1996. The quantity of soil-borne inoculum was much greater at the beginning of the dry season than at the beginning of the rainy season ( Savary et al., 1998 ).

Implications for disease management

The epidemics observed in transplanted rice suggest that, under intensive cropping practices that allow good plant growth, the spacing between hills, within the range that was studied here, does not affect the spread of sheath blight. In less intensive cropping systems, spacing may affect, and limit, sheath blight growth, as observed in the 1996 experiment and in other reports ( Kannaiyan & Prasad, 1983; Castilla et al., 1996 ).

Direct-seeded rice seems to be less conducive to sheath blight than transplanted rice. This strongly suggests that management options for this disease should differ depending on crop establishment methods.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors are very thankful to Mr E. Camangon, Mr E. Carandang, Mr T. Llaneta, Mr A. Magbanua and Mr L. Matundan for their technical assistance in the field observations.

This research was supported under a scientific agreement between the International Rice Research Institute (IRRI) and the Institut de Recherche pour le Développement (IRD).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Campbell CL & Madden LV, eds., 1990. Introduction to Plant Disease Epidemiology. New York, NY: John Wiley & Sons.
  • 2
    Castilla NP, Leaño RM, Elazegui FA, Teng PS, Savary S, 1996. Effects of plant contacts, inoculation pattern, leaf wetness regime, and nitrogen supply on inoculum efficiency in rice sheath blight. Journal of Phytopathology 144, 187 92.
  • 3
    Firman DM & Allen EJ, 1995. Effects of seed size, planting density and planting pattern on the severity of silver scurf (Helminthosporium solani) and black scurf (Rhizoctonia solani) diseases of potatoes . Annals of Applied Biology 127, 73 85.
  • 4
    Giesler LJ, Yuen GY, Horst GL, 1996a. Tall fescue canopy density effects on brown patch disease. Plant Disease 80, 384 8.
  • 5
    Giesler LJ, Yuen GY, Horst GL, 1996b. The microclimate in tall fescue turf as affected by canopy density and its influence on brown patch disease. Plant Disease 80, 389 94.
  • 6
    Guzman Garcia P & Nieto Illidge L, 1992. Nuevas densidades y fertilizacion para disminuir Rhizoctonia solani. Arroz 41, 10 4.
  • 7
    Kannaiyan S & Prasad NN, 1983. Effect of spacing on the spread of sheath blight disease of rice. Madras Agricultural Journal 70, 135 6.
  • 8
    Madden LV, 1986. Statistical analysis and comparison of disease progress curves. In: LeonardKJ, FryWE, eds. Plant Disease Epidemiology, vol. 1, Population Dynamics and Management. New York, NY: Macmillan, 55 84.
  • 9
    McClung AM, Marchetti MA, Webb BD, Bollich CN, 1997. Registration of ‘Jefferson’ rice. Crop Science 37, 629 30.
  • 10
    McRoberts N, Hughes G, Madden LV, 1996. Incorporating spatial variability into simple disease progress models for crop pathogens. Aspects of Applied Biology 46, 75 82.
  • 11
    Mithrasena YJPK & Adikari WP, 1986. Effect of density on sheath blight (ShB) incidence. International Rice Research Notes 11, 20 1.
  • 12
    Ou SH, ed., 1987. Rice Diseases, 2nd edn. Slough, UK: CAB International.
  • 13
    Savary S, Elazegui FA, Moody K, Litsinger JA, Teng PS, 1994. Characterization of rice cropping practices and multiple pest systems in the Philippines. Agricultural Systems 46, 385 408.
  • 14
    Savary S, Castilla NP, Elazegui FA, McLaren CG, Ynalvez MA, Teng PS, 1995. Direct and indirect effects of nitrogen supply and disease source structure on rice sheath blight spread. Phytopathology 85, 959 65.
  • 15
    Savary S & Mew TW, 1996. Analyzing crop losses due to Rhizoctonia solani: rice sheath blight, a case study . In: SnehB, Jabaji-HareS, NeateS, DijstG, eds. Rhizoctonia Species: Taxonomy, Ecology, Pathology and Disease Control. Dordrecht, Netherlands: Kluwer Academic Publishers, 237 45.
  • 16
    Savary S, Castilla NP, Willocquet L, 1998. Effects of some environmental factors on rice sheath blight epidemiology. In: Proceedings of the 7th International Congress of Plant Pathology, Edinburgh, 9–16 August 1998, Invited Papers, 1, BSPP, ISPP, reference 2.4.3S.
  • 17
    Savary S, Willocquet L, Elazegui FA, Teng PS, Du PV, Zhu D, Tang Q, Huang S, Lin X, Singh HM, Srivastsava RK, 2000. Rice pest constraints in tropical Asia: characterisation of injury profiles in relation to production situations. Plant Disease 84, 341 356.
  • 18
    Seem RC, 1984. Disease incidence and severity relationships. Annual Review of Phytopathology 22, 113 50.
  • 19
    Sharma NR & Teng PS, 1990. Effect of inoculum source on sheath blight (ShB) development. International Rice Research Newsletter 15, 18 9.
  • 20
    Teng PS, Torres CQ, Nuque FL, Calvero SB, 1990. Current knowledge on crop losses in tropical rice. In: Crop Loss Assessment in Rice. Los Baños, Philippines: IRRI, 39 54.
  • 21
    Van der Plank JE, 1963. Plant Diseases: Epidemics and Control. New York, ny:. Academic Press.
  • 22
    Xie QJ, Linscombe SD, Rush MC, Jodari-Karimi F, 1992. Registration of LSBR-33 and LSBR-5 sheath blight resistant germplasm lines of rice. Crop Science 32, 507.
  • 23
    Yang XB, Snow JP, Berggren GT, 1990. Analysis of epidemics of Rhizoctonia aerial blight of soybean in Louisiana. Phytopathology 80, 386 92.
  • 24
    Zadoks JC, 1972. Methodology of epidemiological research. Annual Review of Phytopathology 10, 253 76.