Introduction
Although silicon (Si) is abundant in the Earth's crust, it may be depleted in soils that have been intensively cultivated or are highly weathered tropical or organic soils (Epstein, Reference Epstein1999; Savant et al., Reference Savant, Korndorfer, Datnoff and Snyder1999). Approximately 60% of the soils within the South African sugarcane industry are light textured (<20% clay), moderately to strongly acidic (Meyer et al., Reference Meyer, Wood and Harding1998), and are typically deficient in plant-available Si (J.H. Meyer, personal communication). Sugarcane is known to be an Si accumulator (Ma et al., Reference Ma, Miyake, Takahashi, Datnoff, Snyder and Korndorfer2001). The beneficial effects of Si are usually obvious in crops that actively accumulate Si in their shoots, especially under various abiotic and biotic stress conditions. Most of these effects are expressed through Si deposited in the leaves and stems. Silicon can contribute to the control of insect herbivores and plant pathogens, and may also alleviate the effects of various abiotic stresses including drought and salt stress, metal toxicity, nutrient imbalance, high temperatures and freezing (Ma et al., Reference Ma, Miyake, Takahashi, Datnoff, Snyder and Korndorfer2001, Reference Ma, Li, Gao and Xin2004). Beneficial responses, which are wide-ranging, have been observed in Si-amended barley, rice, sugarcane, maize and other monocotyledonous crops (see review by Epstein, Reference Epstein1999). In solution-cultured cucumber, melon, strawberry, soybean and tomato, which take up Si passively, the beneficial effects of Si are also observed if the Si concentration in the solution is high (see reviews by Jones & Handreck, Reference Jones and Handreck1967; Savant et al., Reference Savant, Snyder and Datnoff1997; Epstein, Reference Epstein1999; Ma et al., Reference Ma, Miyake, Takahashi, Datnoff, Snyder and Korndorfer2001, Reference Ma, Li, Gao and Xin2004).
Silicate fertilizers have been recommended for use in countries where agricultural soils are Si-deficient, including Australia, Brazil, USA (Florida, Hawaii), and Mauritius, and soils under sugarcane cultivation in South Africa (Savant et al., Reference Savant, Korndorfer, Datnoff and Snyder1999), recognizing that the role of Si in agriculture is increasingly important for sustainable production (Savant et al., Reference Savant, Snyder and Datnoff1997, Reference Savant, Korndorfer, Datnoff and Snyder1999). Applications of Si fertilizers have many agronomic benefits including improved plant growth, increased yield and positive interactions with applied N, P and K fertilizers (Savant et al., Reference Savant, Snyder and Datnoff1997; see also Jones & Handreck, Reference Jones and Handreck1967 and references cited therein). Silicon amendments reduce the severity of plant pathogens in many crop species (Rodrigues et al., Reference Rodrigues, Datnoff, Korndorfer, Seebold and Rush2001; Bélanger et al., Reference Bélanger, Benhamou and Menzies2003; Dannon & Wydra, Reference Dannon and Wydra2004), and enhance resistance to attack by herbivorous insects including borers (Keeping & Meyer, Reference Keeping and Meyer2002), leaf miners (Goussain et al., Reference Goussain, Moraes, Carvalho, Nogueira and Rossi2002) and sap feeders (Moraes et al., Reference Moraes, Goussain, Basagli, Carvalho, Ecole and Sampaio2004).
Silicon application decreases lepidopteran borer recovery in sugarcane, as observed in, for example, Diatraea saccharalis (Fabricius) in Florida (Anderson & Sosa, Reference Anderson and Sosa2001), and Chilo infuscatelus Snell (Rao, Reference Rao1967) and Scirpophaga excerptalis Walker (Gupta et al., Reference Gupta, Yazdani, Hameed and Agarwal1992) in India. Pan et al. (Reference Pan, Eow and Ling1979) also noted a reduction in percentage nodes damaged by stem borers, in general, following Si application. In both field and pot trials, the application of Si to sugarcane plants significantly reduced damage and numbers of borers recovered from stalks of the African sugarcane borer, Eldana saccharina Walker (Lepidoptera: Pyralidae) (Keeping & Meyer, Reference Keeping and Meyer2002; Meyer & Keeping, Reference Meyer and Keeping2005). However, the beneficial effects of Si are usually more apparent in Si-accumulating plants, such as sugarcane, when under abiotic stress, such as increased salinity or drought, or biotic stresses such as pathogens or insect pests (Epstein, Reference Epstein1994; Ma, Reference Ma2004). Applied Si can reduce water stress by reducing excessive leaf transpiration in rice (Ma et al., Reference Ma, Miyake, Takahashi, Datnoff, Snyder and Korndorfer2001) and maize (Gao et al., Reference Gao, Zou, Wang and Zhang2004), by reducing water loss via decreased water flow rate in the xylem of maize plants (Gao et al., Reference Gao, Zou, Wang and Zhang2004), or by increasing plant water uptake ability in sorghum (Hattori et al., Reference Hattori, Inanaga, Araki, An, Morita, Luxová and Lux2005). Silicon also acts additively with osmotic stress in enhancing pathogen resistance in barley against barley powdery mildew (Wiese et al., Reference Wiese, Wiese, Schwartz and Schubert2005). The effect of Si on the growth of rice was greater under low humidity (water-stressed) than high humidity conditions (Ma et al., Reference Ma, Miyake, Takahashi, Datnoff, Snyder and Korndorfer2001).
Historically, drought stress in plants is considered to be a major factor underlying outbreaks of herbivorous insects (White, Reference White1984; but see also Huberty & Denno, Reference Huberty and Denno2004), including E. saccharina in both rainfed sugarcane in South Africa (Atkinson & Nuss, Reference Atkinson and Nuss1989) and in maize in the Ivory Coast (Moyal, Reference Moyal1995). Insect herbivore populations frequently achieve higher densities on plants that are intermittently, rather than continuously, water-stressed (Huberty & Denno, Reference Huberty and Denno2004). Reduction of moisture stress is one of several recommendations for combating E. saccharina in the South African sugar industry (Anon, 2005a).
However, a key question, which remains unanswered, is whether Si provides greater protection against insect herbivores when plants are water-stressed, than it does in the absence of water stress. Sugarcane (Saccharum spp. hybrids) is an important crop in South Africa, but since 1970 E. saccharina has been a major pest, particularly in coastal rainfed areas prone to drought (Atkinson et al., Reference Atkinson, Carnegie and Smaill1981). During periods of water stress, susceptibility of sugarcane to E. saccharina is significantly increased, particularly in the presence of excess applied nitrogen (Atkinson & Nuss, Reference Atkinson and Nuss1989). Water stress also increases the available stalk nitrogen content of cane, resulting in increased E. saccharina larval survival and biomass, as well as shorter development times (Nuss et al., Reference Nuss, Bond and Atkinson1986; Atkinson & Nuss, Reference Atkinson and Nuss1989). We tested the primary hypothesis that applied Si affords greater protection against E. saccharina in plants subjected to water stress. Our secondary hypothesis was that the benefit of Si application to water-stressed plants is likely to be greater in susceptible cultivars than in resistant ones, given that susceptible cultivars are generally less drought tolerant (Keeping & Rutherford, Reference Keeping and Rutherford2004). If applied Si is more efficacious in drought stressed crops, then its use may provide an enhanced benefit to growers by suppressing borer infestations in areas where soils are deficient in Si.
Materials and methods
A potted sugarcane trial (96 pots) was established in a ‘shade house’ (14×25×3.3 m) with transparent fibreglass roofing and walls of 40% green shade cloth, at the South African Sugarcane Research Institute (SASRI), Mount Edgecombe, KwaZulu-Natal. Sugarcane transplants were produced from single budded setts, cut from mature stalks of four commercial cultivars, resistant (N21, N33) and susceptible (N11, N26) to E. saccharina attack. One-month-old transplants of each cultivar were planted into 25 l PVC pots containing 31 kg (dry weight) of clean, sieved and thoroughly leached river sand, which allowed control of the nutrient supply. Pots were randomly arranged (using the random function in Microsoft® Office Excel, 2003) in a split plot design, where whole plot was ‘cultivar’ and sub-plot was ‘silica’. There were six replicates for each cultivar, each pot containing four transplants of one cultivar. Guard pots (N33) were placed at the end of each row to reduce differences in growth between inner and outer pots, which probably derive from prevailing winds. Before planting, half the pots were treated (Si+) with 124 g (equivalent to 10 t ha−1; 4000 ppm) of wollastonite (i.e. calcium silicate, CaSiO3; 7.9% Si; 60% plant available Si), and the other half left untreated (Si−). The calcium silicate was mixed thoroughly with the sand, dampened and the filled pots left to stand for one week before planting.
All pots were treated monthly with 4:1:1 (44) N:P:K fertilizer or ammonium sulphate (250 g per 25 l water) at 500 ml per pot and Hygrotech® Hydroponic Nutrient Mixture for seedlings (25 g per 25 l water) at 500 ml per pot at planting and every two months thereafter, to provide micronutrients (excluding Si). Fertilizer application ceased to all treatments when the plants were water-stressed. This was done as the rate of crop growth under stress declines dramatically; and, therefore, the demand for nutrients declines accordingly (Marschner, Reference Marschner1986). Continuous application of fertilizer results in a build up of nutrients and unused fertilizer, which in turn may result in increased salinity and further stress, as well as a nutrient imbalance (Marschner, Reference Marschner1986). Plants were drip irrigated using tap water (3 ppm Si), at 0.20 to 2.0 l water per pot per day, depending upon the stage of plant growth and stress treatment. Insecticide spray was applied monthly (chlorpyriphos 2 ml/l−1 water or alphacypermethrin 1 ml/l−1 water) to prevent feral infestations of E. saccharina and other pests. Spraying was halted two months before inoculation with E. saccharina to ensure no pesticide residue remained on the plants.
At 8.5 months, plants in 48 pots (half the trial) were water-stressed intermittently through a staged reduction in their water supply, such that at the end of the stress periods the susceptible cultivars, N11 and N26, had a mean of four green leaves and the resistant cultivars had an average of five green leaves. With the onset of stress, leaf senescence and reduced new leaf appearance combine to reduce green leaf number per stalk compared to non-stressed plants, a recognized indicator for measuring plant water stress in sugarcane (Inman-Bamber & De Jager, Reference Inman-Bamber and De Jager1986; Inman-Bamber, Reference Inman-Bamber2004). Typically, at the time of larval inoculation non-stressed cultivars held an average of 12 green leaves per stalk. Watering was increased slightly after 11 weeks, after which the plants were again water-stressed for three weeks according to a different schedule. This regime was followed because it emulates the variable water availability prevailing in rain-fed sugarcane in South Africa (K.J. Nuss, personal communication). The watering schedules for the two stress periods were as follows. Stress period one: week one, 1.0 l per pot per day; week two, 0.7 l per pot per day; week three, 0.5 l per pot per day; week four, 0.3 l per pot per day; weeks five to 11, 0.2 l per pot per day. Stress period two: week one, 0.5 l per pot per day; week two, 0.3 l per pot per day; week three, 0.2 l per pot per day. The final irrigation rate was maintained until harvest. Manual control of irrigation was sometimes necessary, especially during very hot periods when the plants could be killed.
At 12 months, the trial was hand inoculated with 150 E. saccharina eggs per pot (eggs placed on two stalks per pot at 75 eggs per stalk), following the methods of Keeping (Reference Keeping2006). At the time of inoculation, most of the eggs were in the ‘black head’ stage of development and hatched within 24 h, reducing exposure to egg predation. Larvae were allowed to develop for 66 days (520 degree days; t=10°C; Tempest® Degree-day Units; Insect Investigations Ltd, Cardiff, UK) before harvesting.
At harvest, stalk length, rind hardness at the mid-point of the central internode (Durometer, Rex Gauge Company, Glenview, Illinois 60025, US), and total length of borings per stalk were measured. Thereafter, all stalks were dissected, and leaves and leaf sheaths inspected. Pupal numbers were low. The number of surviving larvae+pupae recovered and the length of stalk bored were used as measures of borer numbers and stalk damage, respectively. Stalk samples for Si analysis were taken at the time of harvest from the same pots used for the borer assessment. Stalk Si% was determined using the procedures of Fox et al. (Reference Fox, Silva Younge, Plucknet and Sherman1967).
All the variables analysed were first submitted to the Anderson Darling test for univariate normality and Bartlett's test for homogeneity of variance. Log transformations were employed to stabilize variances and normalize the data where necessary. However, for the sake of clarity, figure axes and means (±1 SEM) show untransformed data. To test for differences in stalk damage, stalk Si content and rind hardness among treatments, these were analysed using a three-way ANOVA with Si, stress and cultivar as main effects. Borer numbers were analysed using a generalized linear mixed model (GLMM), with a Poisson error distribution and a logarithm link function. In a separate analysis of borer numbers and stalk damage, cultivars were grouped and classified as susceptible (N11, N26) and resistant (N21, N33), to reveal differences between susceptible and resistant cultivars. A three-way ANOVA and GLMM (Poisson error distribution and logarithm link function), with Si, stress and ‘combined cultivar’ as main effects, were used to analyse stalk damage and borers recovered from stalks, respectively. For all analyses, where applicable, the least significant difference (LSD) method was used to determine where significant differences lay. Rind hardness and percentage stalk length bored were analysed at a ‘stalk-within-a-pot’ level. The stalks were considered to be subsamples within the experimental unit (pot). This allowed for two sources of error to be identified, i.e. sampling and experimental error, where the differences between the stalks within a pot (observational differences) could be separated from the differences between the experimental units, increasing the precision of the analysis (Steel et al., Reference Steel, Torrie and Dickey1997). For logistical reasons, stalk Si% and borer numbers were analysed at the pot level (i.e. the sum of borers within each pot). All statistical analyses were performed using Genstat 8.0 for Windows (Genstat, 2005).
Results
Silicon content
Silicon treatment had a significant effect on stalk Si content (F1,20=160.83, P<0.001), which was higher in Si+ plants than in Si− plants (fig. 1). Stalk Si differed significantly between cultivars (F3,15=5.29, P=0.011), being lowest in N11 (fig. 1). There was no significant effect of stress on stalk Si (F1,32=2.02, P=0.165), nor were there any significant interactions. As the trial was only stressed after 8.5 months of normal watering and plant growth, the plants had ample time for Si uptake prior to stressing, as illustrated by the uptake in all Si+ cultivars (fig. 1).
Fig. 1. Percentage of stalk Si at age 14 months in four cultivars (N21 and N33 (resistant), and N11 and N26 (susceptible)) of non-stressed (NS) and water-stressed (S) sugarcane, treated (Si+,
) and untreated (Si−, □) with Si. Bars are SE.
Borer numbers
There was a significant effect of Si on numbers of borers (F1,78=6.87, P=0.009) with fewer borers recovered from Si+ cane than Si− cane (fig. 2). There was also a significant effect of cultivar (F3,78=24.84, P<0.001), with fewer borers found in the resistant cultivars N21 and N33 than in susceptible cultivars N11 and N26 (fig. 2). There was no significant effect of stress on borers recovered from stalks (F1,78=3.1, P=0.078); however, there was a significant Si×stress interaction (F1,78=7.52, P=0.006), with significantly less borers in stressed Si+ cane compared with either stressed Si− or non-stressed Si− and Si+ cane. There were no other significant interactions. Similarly, when results from the susceptible cultivars (N11+N26) were combined and compared with combined resistant cultivars (N21+N33), there were significantly lower borer numbers on combined resistant cultivars compared with combined susceptible cultivars (F1,86=78.54, P<0.001) and significantly fewer borers recovered from stalks on Si+ cane compared with Si− cane (F1,86=7.02, P=0.008). There was no significant effect of stress on borer recovery from stalks (F1,86=3.07, P=0.08). However, a significant Si×stress interaction (F1,86=7.92, P=0.005) showed that significantly fewer borers were recovered from stalks in stressed Si+ cane compared with stressed Si− cane, non-stressed Si− and non-stressed Si+ cane (LSD, P<0.05). There was a significant negative correlation between stalk Si% and borer numbers when all cultivars were combined in the analysis (Spearman rank order correlation; r=−0.30; P<0.05; N=88).
Fig. 2. Mean number of Eldana saccharina larvae and pupae recovered from Si-treated (Si+,
) and untreated (Si−, □), resistant (N21, N33) and susceptible (N11, N26) sugarcane cultivars under non-stressed (NS) and water-stressed (S) conditions. Bars are SE.
Stalk damage
There was a significant effect of Si on percentage stalk length bored (F1,20=34.87, P<0.001), with the least damage recorded in Si+ cane; and a significant effect of cultivar (F3,15=31.82, P<0.001), with the least stalk damage recorded for N33 (fig. 3). However, there was a significant Si×cultivar interaction for percentage stalk length bored (F3,20=5.77, P=0.005; fig. 3), with N21, N11 and N26 exhibiting significantly less damage on Si+ cane compared with Si− cane (LSD, P<0.05, fig. 3). There was no significant effect of stress (F1,316=1.26, P=0.263), nor were there any other significant interactions. When combined results from the susceptible (N11+N26) and resistant (N21+N33) cultivars were compared, there was a significant effect of Si (F1,360=27.29, P<0.001), with less damage recorded on Si+ cane and ‘combined cultivar’ (F1,360=153.54, P<0.001), with the least damage recorded for combined resistant cultivars. There was no significant effect of stress (F1,360=0.96, P=0.328), but a significant ‘combined cultivar’×Si interaction (F1,360=8.28, P=0.004) and a significant Siבcombined cultivar’×stress interaction was obtained (F1,360=6.08, P=0.014), with significantly less damage in Si+ combined susceptible cultivars than in Si− combined susceptible cultivars under non-stressed (LSD, P<0.05) and stressed conditions (LSD, P<0.001). There was a significant negative correlation between stalk Si and percentage stalk length bored when data from all cultivars was combined and analysed (Spearman rank order correlation; r=−0.29; P<0.05; N=88).
Fig. 3. Percentage stalk length bored by Eldana saccharina in Si-treated (Si+, ) and untreated (Si−, □), resistant (N21, N33) and susceptible (N11, N26) sugarcane cultivars under non-stressed (NS) and water-stressed (S) conditions. Bars are SE.
Rind hardness
Treatment with Si significantly increased rind hardness (F1,20=28.21, P<0.001; fig. 4). There was also a significant effect of cultivar on rind hardness (F3,15=14.09, P<0.001), being hardest in N33 followed by in order of hardest to softest, N11, N21 and N26. There was no significant effect of stress on rind hardness (F1,256=3.75, P=0.054); however, there was a significant cultivar×stress interaction (F3,256=3.7, P=0.012) and although the trend for all cultivars, except N21, was for a harder rind when non-stressed compared with stressed, this was only significant for N11 (LSD, P<0.05).
Fig. 4. Internode rind hardness of four sugarcane cultivars, N21, N33, N11 and N26, Si-treated (Si+,) and untreated (Si−, □), under non-stressed (NS) and water-stressed (S) conditions. Bars are SE.
Discussion
Silicon application to sugarcane resulted in its increased uptake by four sugarcane cultivars (N21, N33, N11 and N26), regardless of whether the plants were water-stressed at the end of their growth period or not (fig. 1). Treatment with Si was associated with a reduction in E. saccharina borers recovered from stalks and stalk damage, especially in susceptible sugarcane cultivars under water stress. The increase in resistance of Si+ water-stressed susceptible cultivars to E. saccharina was such that borer recovery from stalks and damage in these plants approached and, in many instances, was not significantly different from that of resistant cultivars (irrespective of whether the latter were treated with Si and/or water-stressed) (figs 2 and 3). These observations support our primary hypothesis that Si enhances resistance of sugarcane to E. saccharina when plants are water-stressed and our secondary hypothesis that the benefit of Si application to water-stressed plants (compared with non-stressed plants) is greater in susceptible cultivars than in resistant ones. The very marked effect of Si in reducing borer numbers and damage in N26 (even when non-stressed for stalk damage), we believe was due to this cultivar's known sensitivity to water stress and poor growth in sandy soils (Anon., 2005b). Inevitably, root binding in pot trials will lead to some degree of stress, even when well watered.
Although the mechanism(s) of Si-mediated resistance has yet to be elucidated, the present results show that its efficacy is enhanced by the simultaneous imposition of water stress and that its action is, therefore, likely to be complex. Silicon application has previously been shown to enhance resistance of sugarcane to E. saccharina, especially in susceptible cultivars (Keeping & Meyer, Reference Keeping and Meyer2002, Reference Keeping2006; Kvedaras et al., Reference Kvedaras, Keeping, Goebel and Byrne2005; Meyer & Keeping, Reference Meyer and Keeping2005), but the effect of water stress was not investigated. Comparison of the Si− controls shows that water stress alone neither significantly nor consistently increased borer recovery (fig. 2) or borer damage (with the exception of N11) (fig. 3). Only N11 showed an increase for both variables and N21 for % stalk length bored in response to stress. This is contrary to Atkinson & Nuss (Reference Atkinson and Nuss1989), who reported increased E. saccharina performance in drought-stressed cane.
All sugarcane cultivars in the Si+ treatments doubled, or almost doubled, their stalk Si content compared with their respective controls (fig. 1). Keeping & Meyer (Reference Keeping and Meyer2006) demonstrated that cultivars differed in Si assimilation and found that Si− resistant cultivars (N21, N33) had higher stalk Si content (i.e. concentration) than Si− susceptible cultivars (N26, N30). American researchers also reported significant differences in Si accumulation between sugarcane cultivars (Deren et al., Reference Deren, Glaz and Snyder1993; Savant et al., Reference Savant, Korndorfer, Datnoff and Snyder1999). However, Keeping & Meyer (Reference Keeping and Meyer2002) found that plant Si content of Si+ resistant cultivars was not consistently higher than that of Si+ susceptible cultivars, as was noted in the current study (fig. 1). Furthermore, in the present study Si content did not differ in any consistent way between Si+ cultivars or stress treatments; in particular, a much larger decrease in borer recovery and stalk damage was obtained in Si+ stressed susceptible cultivars (figs 2 and 3) than would be expected based on the corresponding increases in stalk Si content (fig. 1) in these treatments. While there was a significant negative correlation between stalk Si and percentage stalk length bored and between stalk Si and borer numbers when cultivars were combined, this does not provide an adequate explanation for why Si supplementation increased plant resistance to a greater extent in water-stressed susceptible cultivars than in non-stressed susceptible cultivars.
The mechanical barrier hypothesis has traditionally been advanced as an explanation for Si-mediated resistance to disease, wherein polymerized Si is deposited in epidermal cells and forms a barrier to pathogenic penetration (Ishiguro, Reference Ishiguro, Datnoff, Snyder and Korndörfer2001), in much the same way that it may hinder feeding by herbivorous insects (Djamin & Pathak, Reference Djamin and Pathak1967; Moraes et al., Reference Moraes, Goussain, Basagli, Carvalho, Ecole and Sampaio2004). Internode rind hardness among South African sugarcane cultivars is significantly positively correlated with resistance to E. saccharina, making it a likely contributor to mechanical resistance against stalk penetration by early instar larvae (Keeping & Rutherford, Reference Keeping and Rutherford2004). However, increased rind hardness due to Si deposition in the stalk epidermis does not provide a complete explanation for our findings, as this increase (fig. 4) was not associated with comparable decreases in borer numbers or stalk damage (figs 2 and 3), especially in the water-stressed susceptible cultivars. Therefore, the mechanism by which Si-mediated resistance to E. saccharina acts cannot be explained by the mechanical barrier hypothesis alone. Several studies comparing total Si content of insect resistant and susceptible crop cultivars have also indicated that the arrangement and site of silica deposition in pest-targeted tissues is important (Miller et al., Reference Miller, Robinson, Johnson, Jones and Ponnaiya1960; Hanifa et al., Reference Hanifa, Subramaniam and Ponnaiya1974; Moore, Reference Moore1984). Similarly, the imposition of water stress may change the arrangement, form or concentration of Si in the stalk tissue in ways that increase its effectiveness as a barrier against larval stalk penetration, without necessarily increasing tissue hardness.
To our knowledge, this is the first report of an interaction between Si-mediated resistance to an insect herbivore and an abiotic stress factor. Our results parallel those of other studies, where the beneficial effects of Si were greater under conditions of biotic or abiotic stress (Ma, Reference Ma2004; Gong et al., Reference Gong, Zhu, Chen, Wang and Zhang2005). Wiese et al. (Reference Wiese, Wiese, Schwartz and Schubert2005) showed that Si-mediated resistance of barley to barley powdery mildew could be enhanced by the imposition of osmotic stress (which independently can induce resistance to the fungus (Wiese et al., Reference Wiese, Kranz and Schubert2004)) and that the effects of Si and osmotic stress were additive. Such an amplification of Si-mediated resistance to a biotic stressor (plant pathogen) by an abiotic stressor (osmotic stress) bears a strong resemblance to the Si×water stress×E. saccharina interaction in the present study. In rice the effect of Si on growth is more pronounced under conditions of water stress than non-stressed conditions (Ma et al., Reference Ma, Miyake, Takahashi, Datnoff, Snyder and Korndorfer2001). Similar effects may be at work in sugarcane, which also belongs to the Poaceae. Application of Si under conditions of water stress may have an indirect effect (in addition to any direct effects) of reducing borer numbers and stalk damage by enhancing drought tolerance, especially in susceptible drought-intolerant cultivars with low plant Si content.
Silicon has been implicated in metabolic activities in higher plants under drought (Gong et al., Reference Gong, Zhu, Chen, Wang and Zhang2005) and may also play a role in activating the plant's natural chemical defences against insect herbivores (Gomes et al., Reference Gomes, Moraes, Santos and Goussain2005). Specifically in sugarcane, analysis of genes involved in secondary metabolism suggests that most of the expressed compounds may be acting as defensive barriers to insect attack (Falco et al., Reference Falco, Marbach, Pompermayer, Lopes and Silva Filho2001). Plants that have experienced some form of abiotic stress tend to contain higher concentrations of defence compounds (e.g. Isman & Duffey, Reference Isman and Duffey1982; Inbar et al., Reference Inbar, Doostdar and Mayer2001); indeed, the production of these compounds may represent a general response to stress, and herbivory is merely one form (Myers & Bazely, Reference Myers, Bazely, Tallamy and Raup1991; Nicholson & Hammerschmidt, Reference Nicholson and Hammerschmidt1992). Possibly, under various forms of abiotic stress, including water deficiency, Si augments the production of these defensive compounds, making the plant more resistant to insect attack. Evidence of a role for Si as an activator of plant chemical defences against fungal and bacterial pathogens has been reviewed in recent papers by Ghanmi et al. (Reference Ghanmi, McNally, Benhamou, Menzies and Bélanger2004), Rémus-Borel et al. (Reference Rémus-Borel, Menzies and Bélanger2005), Rodrigues et al. (Reference Rodrigues, Jurick, Datnoff, Jones and Rollins2005), and an editorial by Hammerschmidt (Reference Hammerschmidt2005). Fauteux et al. (Reference Fauteux, Rémus-Borel, Menzies and Bélanger2005) considered that the results from monocotyledons and dicotyledons indicate that the role of Si as an activator of plant defences against pathogens could probably be generalized to the plant kingdom as a whole.
Keeping & Meyer (Reference Keeping and Meyer2002) and Correa et al. (Reference Correa, Moraes, Auad and Carvalho2005) proposed that Si may act as an elicitor of allelochemicals or enzymes involved in plant defence against insect herbivores. Gomes et al. (Reference Gomes, Moraes, Santos and Goussain2005) found that Si, alone or together with aphid pre-infestation, negatively affected greenbugs', Schizaphis graminum (Rondani) (Hemiptera: Aphididae), plant preference and population increase rate, and elicited a significant increase in the activities of the defensive enzymes peroxidase, polyphenoloxidase and phenylalanine ammonia-lyase in wheat. We consider that evidence such as the absence of a clear pattern of association between stalk Si content and resistance to E. saccharina, and the observation that rind hardness of susceptible cultivars did not increase with Si treatment to an extent greater than that of resistant ones, provides grounds for arguing in favour of an active role for soluble Si that compliments any passive, amorphous Si-based mechanical barrier.
In this study, both susceptible cultivars are drought-intolerant while both resistant cultivars have good drought tolerance. It is possible that, when subjected to water stress, the borer-susceptible drought-intolerant cultivars experienced a heightened stress response compared with that of the borer-resistant cultivars, which in turn led to a greater effect of Si (soluble and/or amorphous) in enhancing resistance in the borer-susceptible cultivars. Rutherford and co-workers (Rutherford et al., Reference Rutherford, Meyer, Smith and Van Staden1993; Rutherford & Van Staden, Reference Rutherford and Van Staden1996; Rutherford, Reference Rutherford1998; Heinze et al., Reference Heinze, Thokoane, Williams, Barnes and Rutherford2001) have demonstrated that various defensive compounds (tannins, chlorogenates, flavonoids, epicuticular waxes, protease inhibitors) differ in quantity and/or in composition between sugarcane cultivars, susceptible and resistant to E. saccharina. Possibly, in the presence of Si, water-stressed borer-susceptible cultivars may develop a defensive chemistry with a profile similar to that of borer-resistant cultivars.
From an applied point of view, the finding that Si augments the resistance of water-stressed borer-susceptible sugarcane cultivars is particularly relevant for the Si-deficient soils in the cane-growing regions of South Africa. In these regions, Si amendments for susceptible cultivars may provide improved resistance to E. saccharina, but field trials are required to confirm these results.
Acknowledgements
The authors thank Andrew Govender, Sipho Zuma and the Entomology survey teams at SASRI for their assistance in many aspects of the work. Chandani Sewpersad is thanked for the statistical analysis of the data and Rod Harding for the experimental design. Auas Industrial Minerals (Fontaine Bleau, Johannesburg, South Africa), through consultation with Tony Derrick and a local grower, Trevor Thompson, are gratefully acknowledged for the supply of wollastonite. The SASRI Insect Rearing Unit provided the egg batches. Research funding was provided by SASRI, the National Research Foundation of South Africa, the University of the Witwatersrand, and Le Centre de coopération internationale en recherche agronomique pour le développement.
