Abstract
Woody plants are commonly fertilized in ornamental landscapes, based in part on the rationale that fertilization enhances pest resistance. However, a critical evaluation of evidence finds little to support this claim. Rather, many studies have found that fertilization decreased woody plant resistance to spider mites, sap sucking insects, mandibulate folivores, subcortical feeding insects, and browsing mammals by enhancing the nutritional quality of the plant and/or decreasing secondary metabolite concentrations. The growth/differentiation balance hypothesis (GDBH) postulates a physiological trade-off between growth and secondary metabolism, and predicts a parabolic response of secondary metabolism to variation in nutrient availability. Specifically, fertilization of moderately nutrient-deficient plants is predicted to decrease secondary metabolism if growth is increased but photosynthesis is not affected. However, fertilization of extremely nutrient-limited plants is predicted to increase secondary metabolism if photosynthesis is also increased. A number of studies have found fertilization to increase growth and decrease secondary metabolism. A few studies on extremely nutrient-deficient sites found fertilization to increase foliar secondary metabolism, but insect performance was not affected, possibly because increased foliar nitrogen counteracted effects of secondary metabolites on host quality. These studies, while consistent with the GDBH, do not represent adequate tests because none measured effects of fertilization on photosynthesis as well as growth. Only a few studies have addressed effects of fertilization on the ability of woody plants to tolerate herbivory, and all found fertilization to have no effect. The entrenched paradigm that fertilization enhances the insect resistance of woody plants in ornamental landscapes needs to be reassessed.
Trees and shrubs in ornamental landscapes are commonly fertilized (Beverly et al. 1997, Braman et al. 1998), based in part on the rationale that fertilization enhances pest resistance (Smith 1978, Rathjens and Funk 1986, Neely and Himelick 1987, Caldwell and Funk 1999, Funk 2000, Iles 2000). This assumption, which has emerged in the virtual absence of supporting evidence (Raupp et al. 1992), is no doubt rooted in the long-standing belief that vigorous trees are better able to repel or tolerate herbivory (Houser 1918, Potter 1986, Nielsen 1989). The objective of this forum was to review the substantial evidence that challenges this entrenched paradigm.
Fertilization and Insect Resistance: Evaluating the Evidence
Host plant resistance results from traits that decrease herbivore growth, fecundity, and survival (antibiosis), the attractiveness of plants to herbivores (antixenosis), and/or increase the ability of plants to compensate for herbivory (tolerance) (Painter 1951,1958). Although resistance traits are genetically based, their expression can be influenced dramatically by environmental factors, including fertilization (Painter 1951,1958). Caldwell and Funk (1999) cited the reviews of Stark (1965) and Foster (1967) as evidence that fertilization can increase tree resistance to insects. Both reviews summarized virtually the same set of studies concluding that large-scale fertilization of forests could decrease populations of outbreak insects. However, this limited evidence does not hold up to critical scrutiny.
First, these experiments suffer from limited or no replication of the fertilization treatment (e.g., Oldiges 1959; Schwenke 1960,1962). Therefore, statistically valid conclusions about effects of fertilization cannot be drawn (e.g., Hurlbert 1985), a problem acknowledged by Stark (1965). Furthermore, these studies generally quantified insect population densities rather than direct measures of host plant resistance (which is most relevant to the maintenance of individual plants in landscapes where pest populations transverse property lines). For these reasons, it is not possible to determine whether changes in insect populations were caused by fertilization effects on host plant resistance, indirect effects of fertilization on other factors affecting herbivore density such as natural enemies (e.g., Hartvigsen et al. 1995, Kytö et al. 1996, Forkner and Hunter 2000), or inherent variation among the fertilized and nonfertilized sites.
In their comprehensive review, Waring and Cobb (1992) concluded that fertilization enhanced arthropod growth, fecundity, survival, and/or density in the vast majority of studies with herbaceous and broadleaf trees, but decreased performance in eight of 15 experiments with conifers. However, none of these eight experiments provide firm evidence that standard fertilizer applications can enhance tree resistance to insects. In four cases, statistically-valid conclusions cannot be reached because the large-scale forest fertilization treatment was not replicated (Smirnoff and Bernier 1973, Larsson and Tenow 1984), or replication was extremely limited, and no statistical tests or measures of variation for measures of insect performance were reported (the two experiments reported by Goyer and Benjamin 1972). In two other studies (Carrow and Graham 1968, Carrow and Betts 1972), bi-weekly foliar applications of nitrogen-based fertilizers decreased numbers of balsam wooly adelgid [Adelges piceae (Ratzeburg)] on fir (Abies amabilis Dougl. ex J. Forbes). However, potential insecticidal effects of repeated foliar applications were not evaluated, nor were standard soil fertilizer applications. The other two studies (Xydias and Leaf 1964, Waring and Pitman 1985) also do not provide convincing evidence that fertilization enhances insect resistance of trees. Over the course of their 3-yr study, Waring and Pitman (1985) found that fertilization had no effect on bark beetle-induced mortality of lodgepole pine (Pinus contorta Dougl.), although fertilization did impact the dynamics of mortality, which was lower in fertilized than in control plots in the third year of study. Xydias and Leaf (1964) found that fertilization actually increased incidence of herbivory by white pine weevil [Pissodes strobi (Peck)] on white pine (Pinus strobus L.).
On the other hand, a large number of properly controlled and replicated studies strongly support the generalization that fertilization decreases tree resistance to insects. In a comprehensive review, Kytö et al. (1996) concluded that "where a statistically significant effect of (nitrogen) fertilization on insect performance (body size, development time, survival) was recorded, it was nearly always beneficial for the insects." For example, fertilization of woody plants has been shown to favor piercing-sucking arthropods, including aphids and adelgids (Mitchell and Paul 1974; McClure 1991,1992; Kainulainen et al. 1996; Wainhouse et al. 1998), scales (McClure 1977,1980; Berlinger et al. 1999), psyllids (Pfeiffer and Burts 1983), plant bugs (Holopainen et al. 1995), lace bugs (Casey and Raupp 1999), and spider mites (Garman and Kennedy 1949, Rodriguez 1958, Storms 1969, Jackson and Hunter 1983, Wermelinger et al. 1985, Orians and Floyd 1997). Fertilization has also been shown to decrease tree resistance to caterpillars [Shaw et al. 1978, Brewer et al. 1985 (although, insect performance declined at extremely high rates), Bryant et al. 1987a, Thomas and Hodkinson 1991, Mason et al. 1992, Hättenschwiler and Schafellner 1999, Mutikainen et al. 2000], sawflies (Popp et al. 1986, Mopper and Whitham 1992, Lyytikäinenen 1993, Orians and Floyd 1997, Wainhouse et al. 1998), leaf beetles (Athey and Connor 1989, Lawler et al. 1997), leafminers (Marino et al. 1993, Møller 1995), browsing mammals (Bryant et al. 1987b, Nams et al. 1996, Hartley et al. 1997, Ball et al. 2000), and subcortical-feeding insects, including lesser peach tree borer [Synanthedon exitiosa (Say)] (Smith and Harris 1952), white pine weevil (Xydias and Leaf 1964), Nantucket pine tip moth [Rhyacionia frustrana (Comstock)] (Ross and Berisford 1990), and the xylem-feeding stem-miner Phytobia betulae Kangas of European white birch (Betula pendula Roth) (Ylioja and Rousi 2001). Although, Nowak and Berisford (2000) found that fertilization had no effect on incidence of Nantucket tip moth infestations in a loblolly pine (Pinus taeda L.) grove, they did find that fertilization substantially increased damage from southern pine coneworm [Dioryctria amatella (Hulst.)].
Not all studies found fertilization of woody plants to enhance insect performance; some found little or no effect. For example, fertilization of jack pine (P. banksiana Lamb) (McCullough and Kulman 1991) and Scots pine (Pinus sylvestris L.) (Björkman et al. 1991) growing in extremely nutrient deficient forests increased tree growth without affecting the performance of sawflies. Hare et al. (1989) found that fertilizing orange [Citrus sinensis (L.) Osbeck] orchards increased damage caused by citrus thrips [Scirtothrips citri (Moulton)], and decreased populations of citrus red mite [Panonychus citri (McGregor)]. However, effects on both species were small, and Hare et al. (1989) concluded that fertilization would not have important effects on pest problems one way or the other.
In other cases, fertilization had an effect on a species in one study but not in another. For example, nutrient availability had no effect on gypsy moth [Lymantria dispar (L.)] feeding on quaking aspen (Populus tremuloides Michx.), as larvae performed poorly on plants receiving both high and low nutrient treatments (Hemming and Lindroth 1999). In another study, however, gypsy moth growth rate was increased on quaking aspen by the high nutrient treatment (Osier and Lindroth 2001). Similarly, Potter (1992) found that fertilization did not affect holly leafminer (Phytomyza ilicicola Loew) populations on American holly (Ilex opaca Ait.). However, in another study with holly leafminer, fertilization did increase insect growth, survival, feeding damage, and density (Marino et al. 1993).
Traits conferring plants with resistance to herbivores can be expressed constitutively (preexisting), or they can be induced in response to herbivory (Painter 1951, Karban and Baldwin 1997). Defoliation has been shown to enhance the resistance of woody plants to herbivores in a number of species (Schultz 1988, Haukioja 1990). These herbivore-induced responses have been classified as rapid or delayed (Haukioja 1990), with rapid induced-resistance (RIR) occurring within hours or days of damage, and delayed induced-resistance (DIR) expressed in the year(s) following defoliation.
All of the studies previously cited in this paper have addressed effects of fertilization on the expression of constitutive resistance. Only a few studies have addressed effects of fertilization on the expression of induced resistance of woody plants to folivores, with conflicting results. Hunter and Schultz (1995) concluded that fertilization suppressed the expression of RIR in chestnut and red oak (Quercus prinus L. and Q. rubra L., respectively), while Mutikainen et al. (2000) only observed RIR in European white birch trees that had been fertilized. Effects of nutrient availability on the expression of DIR of trees have also been inconsistent. Fertilization suppressed the expression of DIR in defoliated Alaska paper birch (Betula resinifera Britton), while defoliation enhanced the resistance of nonfertilized trees (Bryant et al. 1993). However, in another study, DIR was expressed in both fertilized and unfertilized mountain birch trees (Betula pubescens Ehrh.) (Ruohomäki et al. 1996).
Caution is required when extrapolating results of these studies to conclusions regarding effects of fertilization on the amount of herbivory experienced by ornamental plants under actual landscape conditions. I am not aware of any studies conducted in ornamental landscapes that have examined the effects of fertilization on the amount of insect damage experienced by plants. However, a number of studies have found fertilization to increase the amount of herbivory experienced by plants in field plots and forests (e.g., Xydias and Leaf 1964, Pfeiffer and Burts 1983, Popp et al. 1986, Lightfoot and Whitford 1987, Glyphis and Puttick 1989, Marino et al. 1993, Nams et al. 1996, Orians and Floyd 1997, Berlinger et al. 1999, Ball et al. 2000), which certainly suggests the potential of fertilization to generate similar patterns in ornamental landscapes and urban forests.
Fertilization and Insect Resistance: Potential Mechanisms
Fertilization effects on insect resistance have been attributed to phenotypic variation in plant nutrient content and secondary metabolite concentrations (Mattson 1980, Bryant et al. 1983, Herms and Mattson 1992, Kytö et al. 1996). The nutritional quality of plants plays a central role in their resistance to insects. Nitrogen, an essential component of protein, is a limiting nutrient for most plants and animals (White 1993). The growth and reproduction of phytophagous insects is frequently limited by the nutritional quality of their hosts, and generally increases as the nitrogen content of the plant increases (Mattson 1980, Scriber and Slansky 1981, Mattson and Scriber 1987, but see Fischer and Fiedler 2000). Increased foliar nitrogen concentration is a near universal response of woody plants to nitrogen fertilization (Koricheva et al. 1998). Hence, fertilization frequently benefits insects by increasing the nutritional quality of their host plant (Mattson 1980).
Fertilization also frequently decreases secondary metabolite concentrations in plants (Mattson 1980, Bryant et al. 1983, Waterman and Mole 1989, Kytö et al. 1996, Koricheva et al. 1998). The growth/differentiaton balance hypothesis (GDBH) attributes this response to a resource-based physiological trade-off between primary and secondary metabolic pathways (reviewed in Herms and Mattson 1992). When fertilization stimulates growth, it may divert plant resources from other processes, including secondary metabolism. Trees and other plants have limited resources to support their physiological processes, and all requirements cannot be met simultaneously. Hence, trade-offs occur among growth, storage, reproduction and defense (Bazzaz et al. 1987, Chapin 1991, Kozlowski 1992, Lambers and Poorter 1992). The trade-off between growth and reproduction is perhaps best documented (e.g., Caesar and Macdonald 1983, Antos and Allen 1990, Dick et al. 1990, Newell 1991, El-Kassaby and Barclay 1992). However, evidence increasingly supports the existence of a similar resource-based trade-off between growth and secondary metabolism (e.g., Coley 1986, Han and Lincoln 1994, Sagers and Coley 1995, Hwang and Lindroth 1997, Zangerl et al. 1997, Baldwin 1998).
The production of new biomass requires particularly high levels of plant resources (McLaughlin et al. 1980, Margolis et al. 1991, Griffin et al. 1993). The production of secondary metabolites is also costly (Gulmon and Mooney 1986, Gershenzon 1994). Consequently, rapidly growing plants may have lower levels of carbon available to support other processes, including secondary metabolism and accumulation of stored energy reserves (Coley 1986, Bazzaz et al. 1987, Chapin et al. 1990, Herms and Mattson 1992, Kozlowski 1992, Lambers and Poorter 1992, Lerdau and Gershenzon 1997).
The differential investment of photoassimilates into new leaf area is largely responsible for both genotypic and phenotypic differences in growth rates (Potter and Jones 1977, Lambers 1987, Körner 1991, Lambers and Poorter 1992). Increased nutrient availability increases tree growth by increasing total leaf area, but has much less affect on the photosynthetic rate of individual leaves unless nutrient availability is extremely low (Linder and Rook 1984, Waring et al. 1985, Ingestad and Ågren 1991, Luxmoore 1991, Bowman and Conant 1994, Ericsson 1995, Kubiske et al. 1998, Samuelson 1998). Under low nutrient conditions, total leaf area is reduced on a whole plant basis, which concentrates nutrients into a smaller canopy (Jurik 1986, Hirose 1987, Sinclair and Horie 1989, McDonald 1990, Ingestad and Ågren 1991, Anten et al. 1995). Trees and other plants also increase their root:shoot ratios in response to decreased nutrient availability (Chapin 1980, Marschner et al. 1996, Wang et al. 1998, Aerts and Chapin 2000), which proportionally increases nutrient uptake. Furthermore, the potential of roots to acquire nutrients per unit biomass may increase in nutrient-limited environments (Chapin 1991, Aerts and Chapin 2000). Since most foliar nitrogen is used to produce photosynthetic enzymes (Chapin et al. 1987, Evans 1989), these buffering mechanisms act to maintain stable rates of photosynthesis in low nutrient sites (McDonald 1990), and nitrogen deficiency rarely induces chlorosis in trees that have had a chance to acclimate to low nutrient conditions (Ingestad 1982, Harris 1992). However, growth rate is reduced because total leaf area is decreased (Chapin 1980,1991; Luxmoore 1991; Lambers and Poorter 1992; Albaugh et al. 1998).
On a whole-tree basis, the greater total leaf area of fertilized plants allows them to acquire more carbon and produce greater quantities of secondary metabolites. Individual herbivores, however, feed on localized tissues, rather than whole trees. Therefore, understanding how a fast growing tree, with a larger overall carbon budget, will have lower concentrations of secondary metabolites in specific organs and tissues than a slower growing tree, requires understanding how fertilization affects patterns of carbon allocation and partitioning. The distribution of carbon is controlled by interactions between carbon sources and sinks, which themselves are under tight hormonal and ezymatic regulation (Chapin et al. 1993, Geiger et al. 1996, Sturm and Tang 1999). The growth of active meristems, which are strong photosynthetic sinks (Patrick 1988, Wardlaw 1990, Marcelis 1996), is supported by carbohydrates exported from carbon sources, including storage organs and neighboring leaves (Harper 1989, Dickson 1991, Kozlowski 1992). If fertilization has little effect on the photosynthetic rate of existing leaves, it will not increase their pool of available carbon, and production of new leaves can be supported only if they export a greater proportion of photosynthate to growing meristems, leaving less carbon available for the production of secondary metabolites. Hence, secondary metabolism may be substrate and/or energy limited in rapidly growing plants (Margna et al. 1989, Herms and Mattson 1992, Jones and Hartley 1999). However, since moderate nutrient deficiency dramatically slows the rate of new leaf production, but has a proportionally smaller impact on photosynthetic rate of individual leaves, carbohydrates that would otherwise be exported to growing meristems accumulate in the leaf where they are produced (Körner 1991, Luxmoore 1991, Geiger et al. 1996). This increases the pool of carbohydrates available for production of secondary metabolites (Chapin 1980, Bryant et al. 1983, Waterman and Mole 1989, Herms and Mattson 1992, Jones and Hartley 1999).
Numerous studies document examples in which fertilization-induced reductions in secondary metabolite concentrations were associated with decreased resistance to herbivores. For example, fertilization of Alaska paper birch and quaking aspen increased tree growth while decreasing concentrations of defensive compounds and resistance to snowshoe hares and leaf-feeding insects, respectively (Bryant et al. 1987a,1987b). Fertilization of loblolly pine increased growth and decreased foliar phenolic concentrations, and decreased resistance to the Nantucket pine tip moth (Ross and Berisford 1990). Glyphis and Puttick (1989) found that fertilization increased growth of oak (Quercus coccifera L.), but decreased tannin levels of leaves, which then received more insect feeding. Lawler et al. (1997) observed the same pattern in Eucalyptus tereticornis Sm.: fertilization increased growth, but decreased concentrations of foliar phenolic compounds and resistance to leaf beetles. Similarly, nitrogen fertilization decreased concentrations of foliar phenolic compounds in Norway spruce [Picea abies (L.) Karst.], which increased the growth of Lymantria monacha L. (Hättenschwiler and Schafellner 1999). Holopainen et al. (1995) observed that increased nitrogen availability decreased total phenolic and resin acid (but not monoterpene) concentrations of Scots pine (P. sylvestris L.), and decreased resistance to tarnished plant bug (Lygus regulipennis Popp.). Kainulainen et al. (1996) ascribed the enhanced growth rate of the aphid Schizolachnus pineti (F.) on fertilized Scots pine to increased foliar nitrogen and decreased resin acid concentrations. Wainhouse et al. (1998), in a study with Sitka spruce [P. sitchensis (Bong.) Carrière], also found that fertilization decreased foliar phenolic and resin acid concentrations, and decreased resistance to an aphid and sawfly. Fertilization also decreased chemical defenses in the foliage of European white birch, resulting in decreased resistance to the outbreak folivore Epirrita autumnata (Borkhausen) (Mutikainen et al. 2000).
Based on the results of their meta-analyses, Haukioja et al. (1998) and Koricheva et al. (1998) concluded that phenolic compounds are generally decreased by fertilization, but terpenes are not responsive to nutrient availability. However, several studies have found fertilization to decrease terpene levels in woody plants (e.g., Bryant et al. 1987a, Johnson and Lincoln 1991, Holopainen et al. 1995, Wilkens et al. 1997, Wainhouse et al. 1998, Warren et al. 1999), while other studies found fertilization to increase terpene concentrations (e.g., Björkman et al. 1991,1998; McCullough and Kulman 1991), suggesting that terpenes can be responsive to nutrient availability. Two studies found that when fertilization of loblolly pine increased growth, resin production in the main stem was dramatically reduced (Wilkens et al. 1997, Warren et al. 1999). Since decreased resin flow has been associated with increased reproductive success of southern pine beetle (Dendroctonus frontalis Zimmerman) (Reeve et al. 1995), these results are inconsistent with suggestions that fertilization can enhance pine resistance to bark beetles in ornamental landscapes (Ham and Hertel 1984). Kytö et al. (1998,1999) found that fertilization did not affect resin flow of very slow-growing Scots pine, but did observe a negative correlation between resin flow and tree vigor index. None of the most vigorous trees produced high levels of resin.
Studies in which fertilization increased terpene concentrations have been deemed inconsistent with the GDBH (Kytö et al. 1996, Haukioja et al. 1998), which they erroneously interpret as predicting that fertilization will always decrease concentrations of carbon-based secondary metabolites. In fact, the GDBH postulates a parabolic response of secondary metabolism to variation in nutrient availability, which explicitly predicts that the resource-based trade-off between growth and secondary metabolism will not always result in a negative correlation between them (Herms and Mattson 1992). It is well documented that resource-based allocation trade-offs do not always result in negative correlations between physiological traits when there is variation in the rate of resource acquisition (e.g., Van Noordwijk and de Jong 1986, Reznick et al. 2000). Accordingly, the GDBH (Herms and Mattson 1992, figure 1, p. 289) predicts that if fertilization of extremely nutrient-limited plants increases photosynthesis, the increased carbon budget of individual leaves will result in increased secondary metabolism and growth, despite competition for a common resource base. Under conditions of moderate nutrient limitation, where net photosynthesis is already close to maximum but fertilization increases growth, the physiological trade-off between growth and secondary metabolism is predicted to manifest itself as a negative correlation between them.
Effects of fertilization (18:5:9, N:P:K applied at a rate of 225 kg N/ha/yr) on the ability of 5-yr-old paper birch (Betula papyrifera Marsh.) and 7-yr-old sugar maple (Acer saccharum Marsh.) to tolerate severe (75-100%) defoliation by forest tent caterpillar (Malacosoma disstria Hilbner). Trunk growth (± SEM) was measured as the increase in trunk diameter at 50 cm above ground. The statistical interaction between the fertilizer and defoliation treatments was not significant for paper birch (F = 0.37; df = 1,34; P = 0.55) or sugar maple (F= 0.03; df = 1,34; P = 0.87), indicating that there is no evidence for either species that the slope of the line describing the effects of defoliation on growth differed for fertilized and nonfertilized trees. Thus, there is no evidence that fertilization had any effect on the ability of either species to tolerate defoliation (modified from Dankert 1996).
The studies in which fertilization increased growth and foliar terpene concentrations of pine trees were conducted on extremely nutrient deficient sites where tree growth was quite slow (Björkman et al. 1991,1998; McCullough and Kulman 1991). Hence, these responses may be consistent with predictions of the GDBH. The photosynthetic rates of nutrient-deficient conifers have been shown to increase in response to fertilization (Brix 1981, Linder and Rook 1984, Brown et al. 1996). However, Björkman et al. (1991,1998) and McCullough and Kulman (1991) did not measure photosynthesis, so it is not possible to judge if these studies are truly consistent with the model. Lerdau et al. (1995) found that fertilization did increase photosynthesis and monoterpene concentration of Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco]. A parabolic response of secondary metabolism to nutrient availability would suggest the potential for fertilization to increase insect resistance on extremely nutrient-deficient sites. However, McCullough and Kulman (1991) and Björkman et al. (1991) found that fertilization did not affect sawfly performance, perhaps because positive effects of increased foliar nitrogen counteracted negative effects of increased terpene concentrations on host quality.
Fertilization and Tolerance of Woody Plants to Defoliation
Relative to studies of fertilization effects on antibiosis and antixenosis, very few studies have addressed the effects of fertilization on the ability of woody plants to tolerate herbivory. Fertilization of defoliated trees has been recommended based on the assumption that defoliation will have less impact on fertilized than unfertilized trees (e.g., Wargo 1978). Studies have shown that fertilization can increase the growth of defoliated trees (e.g., Wickman et al. 1992), as it does the growth of nondefoliated trees. However, testing the hypothesis that fertilization increases the tolerance of trees to defoliation requires a direct comparison of fertilization effects on both defoliated and nondefoliated control trees. Increased tolerance is indicated by a shallower slope in the growth response of the fertilized trees to defoliation (Strauss and Agrawal 1999, Stowe et al. 2000). I am aware of only a few studies with woody plants that have tested this hypothesis, and all found that fertilization had no effect on the ability of woody plants to tolerate defoliation (Parker 1978, McGraw et al. 1990, Dankert 1996, Houle and Simard 1996).
Dankert (1996) tested the effects of fertilization on the ability of 5-yr-old paper birch (Betula papyrifera Marsh.) and 7-yr-old sugar maple (Acer saccharum Marsh.) to tolerate severe (75-100%) defoliation by forest tent caterpillar (Malacosoma disstria Hübner). The rate and timing of the fertilization treatment (18:5:9, N:P:K applied at a rate of 225 kg N/ha/yr in split applications in May and October) are consistent with recommendations for woody ornamental plants (Smith 1978, Rathjens and Funk 1986, Neely and Himelick 1987). Defoliation dramatically decreased the growth of both paper birch and sugar maple, while fertilization increased the growth of paper birch but had no effect on sugar maple growth (Fig. 1). However, the statistical interaction between the fertilizer and defoliation treatments was not significant for either paper birch (F = 0.37; df = 1, 34; P = 0.55) or sugar maple (F = 0.03; df = 1, 34; P = 0.87). For both species, this indicates that there is no evidence that the slope of the line describing the effects of defoliation on growth was different for fertilized and nonfertilized trees. Thus, there is no evidence that fertilization had any effect on the ability of either species to tolerate defoliation (Fig. 1). The same was true in each of the 2 yr following defoliation. In other words, fertilizer provided no special growth-enhancing benefits to the defoliated trees, either during the year of defoliation, or in following years. Parker (1978) and McGraw et al. (1990) also found that fertilization had no effect on the ability of red oak seedlings to tolerate defoliation, as did Houle and Simard (1996) with Salix planifolia Pursh.
The lack of fertilization effect on the ability of trees to tolerate defoliation may be because the growth of defoliated trees is limited more by carbon than nutrients (e.g., Jonasson 1995). Deciduous trees that are severely defoliated often refoliate during the same growing season, which draws substantially on their carbohydrate reserves (Wargo et al. 1972, Herms et al. 1987), and high levels of storage reserves have been associated with a greater ability of trees to recover from defoliation (Gross 1991). Accumulation of carbohydrate reserves may compete with growth for allocation of resources (Chapin et al. 1990), and fertilization has been shown to decrease levels of storage carbohydrates (Etter 1972, Waring et al. 1985, Larsson et al. 1986, McDonald et al. 1986, Green et al. 1994, Ericsson 1995, Mooney et al. 1995, Von Fircks and Sennerby-Forsse 1998, but see Bollmark et al. 1999). Hence, lower carbohydrate reserves of fertilized trees may decrease their capacity to tolerate herbivory.
Summary and Conclusions
A critical review of the evidence finds little to support the view widely held within the tree care industry that fertilization enhances insect resistance of woody plants. In fact, substantial evidence demonstrates that fertilization frequently enhances insect performance by increasing the nutritional quality of their hosts, and/or by decreasing concentrations of secondary metabolites. Furthermore, I am not aware of any convincing evidence that shows fertilization to enhance tree resistance to insects or mites. I am also not aware of any evidence that fertilization increases the tolerance of woody plants to defoliation.
It has been suggested that it might be possible to manipulate fertilizer formulation to avoid detrimental effects on pest resistance of trees (Caldwell and Funk 1999). Although most studies have used fertilizers formulated with multiple elements (e.g., nitrogen, phosphorus, and potassium), the emphasis of this review has been on tree responses to nitrogen, which is the nutrient that most limits tree growth in managed landscapes (Harris 1992). Studies have found that fertilizing with P and/or K, in the absence of nitrogen, had no effect on tree growth (Neely et al. 1970, Harris 1992). A recent review found that phosphorus fertilization of woody plants also had no effect on secondary metabolite concentrations (Koricheva et al. 1998). Furthermore, field trials have shown that the rate of nitrogen applied is the key factor affecting tree growth, with form of nitrogen or method of application having little effect (Neely et al. 1970, Smith 1978).
It should be emphasized that the general pattern that fertilization decreases tree resistance to sucking and chewing arthropods and browsing mammals has emerged in spite of the great variation among studies in their experimental conditions. These sources of variation include field versus container studies, and varying levels of background fertility, as well as variation in the formulation, rate, amount applied, and timing of fertilization treatments. Similar patterns have been observed across natural gradients of soil fertility (e.g., Muller et al. 1987, Glyphis and Puttick 1989). This suggests that fertilization influences insect resistance of trees via general responses of trees to increased nutrient uptake, rather than through specific effects of particular formulations.
When applied as part of a comprehensive plant health care program (e.g., Lloyd 1997), prescription fertilization programs can have clear nutritional benefits for woody plants, especially in high maintenance landscapes. However, fertilization programs should be implemented with an understanding of the potential consequences for pest resistance. It is time to reassess the entrenched paradigm that fertilization can enhance the insect resistance of woody plants in ornamental landscapes.
Discussions with John Lloyd, David G. Nielsen, and Michael Raupp were motivating forces behind this paper. John Lloyd, David G. Nielsen, Kimberly Wallin, and two referees provided constructive commentary on earlier drafts of the manuscript.
References
