|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Interspecific interactions among co mmon insects of the salt myrtle, Baccharis halimifolia L. (Asteraceae) by Laura F. Altfeld A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Biology College of Arts and Sciences University of South Florida Major Professor: Peter Stiling, Ph.D. Susan Bell, Ph.D. Henry Mushinsky, Ph.D. Anthony Rossi, Ph.D. Date of Approval: April 14, 2006 Keywords: Linepithema humile plant-insect interacti ons, myrmecophily, host plant quality Copyright 2006, Laura F. Altfeld
ACKNOWLEDGEMENTS There are many people whose assistan ce, support and encouragement made completion of this dissertation possible. Fi rst and foremost, I want to acknowledge my committee advisor and friend, Peter Stiling; thank you for your confidence, support, and patience. To my committee, Susan Be ll, Anthony Rossi, Gary Huxel and Henry Mushinsky, thank you for your guidance and encouragement and for challenging me to become a better scientist. Thank you to Jim Wilson and Robert Browning for allowing me to conduct research within beautiful Fort De Soto Pa rk. Thank you also to Mark Deyrup, Archbold Biological Station, for assi stance with ant identif ications and Susan Halbert, Florida Department of Agriculture, for help with the aphi ds. A special thank you to Sylvia Lukasiewicz, for endless after noons in the field and in the lab. To my colleagues in the Stiling Lab, Tatiana Cornel issen, Mark Barrett, Amanda Baker, Tere Albarracin, Rebecca Forkner, Kerry Bohl and Heather Jezarek, I was in the best of company; thank you for your camaraderie and insight. Thank you to my family, especially Ma rie Hensey, Christopher Hensey, and Jonathan Altfeld. Your optimism, advice, confidence and humor were essential. To Jonathan, I especially want to thank you for your gift of providing the appropriate perspective, reminding me never to take my self too seriously, and for our two girls, Gillian and Madison, to whom this body of work is dedicated.
i TABLE OF CONTENTS LIST OF TABLES iii LIST OF FIGURES iv ABSTRACT vi CHAPTER 1. THEORY, BACK GROUND AND OBJECTIVES 1 Objectives 4 Study System 6 CHAPTER 2. MYRMECOPHILY ON BACCHARIS HALIMIFOLIA 9 Synopsis 9 Introduction 10 Methods 11 Results 14 Discussion 20 CHAPTER 3. NITROGEN LI MITATION, COMPETITION AND MYRMECOPHILY 23 Synopsis 23 Introduction 23 Methods 25 Results 27 Discussion 33 CHAPTER 4. PHOSPHOR US LIMITATION 35 Synopsis 35 Introduction 35 Methods 36 Results 38 Discussion 42 CHAPTER 5. EXOTIC VERSUS NATIVE ANTS 43 Synopsis 43 Introduction 44 Methods 46 Results 48
ii Discussion 58 CHAPTER 6. CLOSING REMARKS 61 REFERENCES 66 ABOUT THE AUTHOR End Page
iii LIST OF TABLES Table 2.1 Mean SE abundances of myrmecophilous and nonmyrmecophilous insects on th e host plant in each ant attendance level treatment. 15 Table 5.1 Ant species effects on insects of B halimifolia Results from Nested ANOVA with Ant Species as the main factor. 49 Table 5.2 Effects of early-season feeding by T. bacharidis larvae on herbivores of B. halimifolia Results from Nested ANOVA with Larval Herbivory as nested factor within Ant Species. 50 Table 5.3 Bottom-up effects from nitrogen fertilization on insects of B. halimifolia Results from Nested ANOVA with Fertilization nested within Larval Herbivory (nested Within Ant Species). 51 Table 5.4 Top-down effects of ants on insects of B. halimifolia Results from Nested ANOVA with Ant Foragers nested within Larval Herbivory (nested within Ant Species). 52
iv LIST OF FIGURES Figure 2.1 Seasonal mean density of aphids and aphid predators 1 standard error on 25 B. halimifolia branches per tree from May through September 2003 in high (H; n = 6) versus low (L; n = 9) ant attendance level treatments. 16 Figure 2.2 Mean 1 standard error estimated number of aphids in paired ant-tended / untended populations. 18 Figure 2.3 Comparison of the mean 1 standard error number of aphid predators in ant-tended versus untended aphid populations. 19 Figure 3.1 Mean seasonal abunda nces of ant-damaged mines, A. maculosa and L. trifolii in the presence and absence of ant attendants. Bars represent 1 standard error. 30 Figure 3.2 Mean seasonal abundances of aphids, ants and coccinellids on smooth and grooved trees in the presence of ant attendants. Bars represent 1 standard error. 31 Figure 3.3 Mean monthly abunda nce of larval and adult T. bacharidis on control and fertilized trees. Bars represent 1 standard error. 32 Figure 4.1 Mean 1 SE abundances of T. bacharidis on host plants under different fertilization treatments. 39 Figure 4.2 Mean 1 SE abundances of L. humile A. coreopsidis and aphid predators under different fertilization treatments. 40 Figure 4.3 Mean 1SE abundances of A. maculosa and L. trifolii on host plants under different fertilization treatments. 41 Figure 5.1 Comparison of mean 1 SE abundances of (a.) insects involved in the ant-aphid mutualism and (b.) other herbivores not involved in the ant-aphid mutualism, when the ant foragers were either the native C. floridanus or the exotic L. humile 53
v Figure 5.2 Comparison of mean 1 SE abundances of herbivores on host plants with the pres ence and absence of earlyseason herbivory by T. bacharidis larvae. Results are presented for host trees with (a.) native C. floridanus and (b.) exotic L. humile ants. 54 Figure 5.3 Comparison of mean 1 SE abundances of herbivores on control and nitrogen en richment host plants with (a.) native C. floridanus and early-season herbivory, (b.) native C. floridanus but no early-season herbivory, (c.) exotic L. humile and early-season herbivory, and (d.) exotic L. humile but no early-season herbivory. 55 Figure 5.4 Comparison of mean 1 SE abundances of insects involved in the ant-aphid mutualism when ants were present and absent on host plants with (a.) native C. floridanus and early-season herbivory, (b.) native C. floridanus but no early-season herb ivory, (c.) exotic L. humile and early-season herbivory, and (d.) exotic L. humile but no early-season herbivory. 56 Figure 5.5 Comparison of mean 1 SE abundances of herbivores not involved in the ant-aphid mutualism when ants were present or absent on host plants with (a.) native C. floridanus and early-season herbivory, (b.) native C. floridanus but no early-season herb ivory, (c.) exotic L. humile and early-season herbivory, and (d.) exotic L. humile but no early-season herbivory. 57 Figure 6.1. Abundance of ant-damaged l eaf mines as a quadratic function of the abundance of intact leaf mines on the host plant, B. halimifolia 64
vi Interspecific interactions among co mmon insects of the salt myrtle, Baccharis halimifolia L. (Asteraceae) by Laura F. Altfeld ABSTRACT Baccharis halimifolia is host to many species of insect herbivore, including myrmecophilous aphids. P revious studies with B. halimifolia have revealed strong effects of competition by an early-feeding beetle, Trirhabda bacharidis, and nitrogen enrichment. The effects of ant mutualists, however, have not been evaluated for their potential influence on interspecific interactio ns among insects on the host plant. I have employed a series of experiments aimed at answering the followi ng general questions. (1) How does the density of aphid-tending ants affect common insects on the host plant? (2) What are the relative effects of (a) competition from T. bacharidis (b) nutrient enrichment, and (c) ant presence on common in sects of the host plant? (3) How do the effects of exotic ants differ from those of native ants? The density of aphid-tending ants had positive effects on myrmecophilous aphids and aphid predators. However, given a choice between tended and untended aphid populations, aphid predators preferred to forage in the absence of ants. The density of
vii aphid-tending ants also increased preda tion on leaf miners although this did not necessarily translate into reduced densities of leaf miners. Competition by early-season feeding of T. bacharidis negatively affected laterfeeding herbivores but the effects of competition were unaffected by nutrient enrichment. Nutrient enrichment had positive effects on some herbivores, often only in the absence of early season herbivory. Trirhabda bacharidis larvae showed eviden ce of nitrogen and phosphorus limitation and suffered no predati on by aphid-tending ants. Ant presence increased host plant survivorship from stemborer damage in 2004. Ant species identity was an important factor determining the densities of myrmecophiles and non-myrmecophiles on the host plant in addition to affecting the responses of herbivores to increases in host plant quality. Aphids were more abundant in the presence of the exotic Linepithema humile (Hymenoptera: Dolichoderinae) versus the native Camponotus floridanus (Hymenoptera: Formicinae) Aphid predators also had higher densities in the presence of L. humile versus C. floridanus. Only L. humile acted as predator on leaf mines although pred ation did not always translate into reduced seasonal abundances for both sp ecies of leaf miner.
1 CHAPTER 1 THEORY, BACKGROUND AND OBJECTIVES The meta-subject for this dissertation is interspecific inte ractions among common insects that share the host plant, Baccharis halimifolia L. (Asteraceae), in Florida. I chose this general topic because of my interests in how mu ltiple species interact when sharing a resource and what factors influen ce the population densitie s of those species. This dissertation draws upon eco logical theories for compe tition, nutrients (so called bottom-up factors), myrmecophily, and exotic sp ecies as these theories pertain to plantinsect herbivore interactions The inference space for c onclusions drawn from this body of work is small, an island, a single host plan t, relative to the in finite possibilities for interspecific interactions among insects on this host plant, let alone the possibilities for plant-insect interactions in general. That notwithsta nding, the following series of experiments add to increasing evidence in s upport of community-wide effects of exotic species, conditional importance of bottom-up factors, and asymmetric competition while at the same time challenging ecologists to que stion generalities about ant mutualisms and how both the presence of an ant-aphid mutualis m and the identity of the ant mutualist can differentially affect mrymecophilous and nonmrymecophilous insects that share a host plant.
2 Connell (1983) defined competition among species as the term used when each species of a pair has a negative effect on the other. The extent to which species compete with each other is a topic of conti nued interest, particularly in terms of the influence competition may have in structuring communities. Contemporary ecological theory is divided in support of the role competition plays in determining the abundances and identities of species within a community. Up until the late 20th century, competition was considered the primary structuring force in communities. Ecological theory had long since been under the strong influence of proce sses that were born out of studies done on plant communities, such as the role of competition among plants during succession. Hutchinsons (1959) classical competition model suggested that competition is the primary structuring force in communities. Later, this model was amended by Hairston et al. (1960) who established a t op-down trophic scheme stressing that, while competition was important among plants and predators, herbivores do not compete because of the top-down pressure from predators. Janzen (1973) considered competition among herbivores to be important on an evolut ionary time scale but suggested that, on an ecological time scale, the importance of compe tition, particularly for insect herbivores, is dependent upon the predilection of herbivore species for part icular plant parts and the host plants response to herbivory. Weins (1977) also challenged the Hutchinson model by suggesting that because reso urces are not always in shor t supply, competition is not a primary structuring force for herbivore comm unities, i.e. competition is only important during the times when resources are in shor t supply relative to demand. In 1983, two independent literature review s (one by Connell and one by Schoener) were published that concluded a less pervasive role for in terspecific competition among herbivores,
3 advocating the more frequently observed and relevant role of intraspecific competition in affecting population densities. Sih et al.s (1985) review conclude d that competition was no more important than predation in structur ing communities. The theory of competition had become much maligned as the primary factor structuring communities owing to a significant lack of evidence and confusion a bout what validly qualif ied as evidence of true competition among species. Competition is currently considered to be one of several potentially important factors shaping co mmunities. Among insect herbivores, the presence of competition among species has b een documented in all types of feeding guilds (Denno et al. 1995) and ha s been shown to be important for host plants, herbivores and predators in several systems (Hudson and Stiling 1997, Denno et al. 2000, Moon and Stiling 2002). The community of insect herbivores on Baccharis halimifolia (Asteraceae) provides an example of a system that can be substantially influenced by competition. The salt myrtle has a broad range in the U. S. and has been reported to support 133 phytophagous insect species, 11 of which ar e specialists (Palmer 1987). Among these is an early-season foliage-feeding beetle, Trirhabda bacharidis (Coleoptera: Chrysomelidae), that reduces the densities of leaf miners and a stem galling fly which occur on the plant later in the year (Hudson a nd Stiling 1997). The negative effects of T. bacharidis were described by Hudson and Stilin g as an example of exploitative competition of leaves and growing tips by the specialist beetle. Several insect herbivores on B. halimifolia have been shown to be positively affected by the bottom-up effect of increases in host plant quality (Moon and Stiling 2004). In fact, it is common for insect he rbivores to respond positively, in terms of
4 growth and/or fecundity, to increases in host plant quality because herbivores are typically limited by the nitrog en content of their food re lative to their demand for nitrogen (Waring and Cobb 1992, Kyto et al. 1996). In addition to the importance of nitrogen, phosphorus is also a limiting nutrient for some herb ivores, particularly those with rapid growth rates (Waring and C obb 1992, Sterner and Elser 2002). The relative importance of such bottom-up factors for herbivores may change depending on heterogeneity in other variables such as e nvironmental stress, plant chemistry, omnivory, competition or predation (Menge and Suther land 1976, Hunter and Price 1992, Polis and Strong 1996, Forkner and Hunter 2000, Moon and Stiling 2002). Insect herbivores on the host plant B. halimifolia contend with competition by an early-season feeding herbivore, nitrogen lim itation and ant predation by exotic aphidtending ants (Hudson and Stili ng 1997, Moon and Stiling 2004, A ltfeld and Stiling 2006). The goals of this research are to examine th e relative roles of these factors when they occur in concert. Objectives The field experiments that will be desc ribed herein have three major objectives. First, because B. halimifolia commonly hosts an ant-aphid mutualism, and ant attendants may affect host use by other herbivores on the ho st plant, I evaluated the effects of aphidtending ants on aphids, their predators and ot her herbivorous insect s on the host plant. This was done by manipulating the level of ant attendance, creating naturally high and low densities of ant foragers and subsequent ly measuring the densit ies of aphids, their predators, two leaf miners, a stem borer a nd a gall fly over the course of a season. In
5 addition, because short-term trends may help explain seasonal dynamics, the short term effects of ant removal on aphids and aphidopha ges were also inves tigated. Second, as follow-up to Hudsons and Stili ngs work, I evaluated the eff ects of nutrient enrichment, specifically nitrogen and phosphorus, on exploita tive competition by the specialist beetle T. bacharidis A factorial experiment that combined N-fertilization, T. bacharidis larval herbivory, and ant attendance was used to teas e apart the effects of each of these three important factors in affecti ng the densities of herbivores on the host plant. In a complementary experiment, phosphorus was eval uated for its potential as an important limiting nutrient in this system by compari ng P-fertilization with NP-fertilization and ambient (non-fertilized) conditions. Lastly, I compared the effects of an exotic aphidtending ant ( Linepithema humile ) to those of a native aphid-tending ant ( Camponotus floridanus ) on plant-insect and interspecific in teractions among common insects of the host plant. The Argentine ant, Linepithema humile (Formicidae: Dolichoderinae), is the most commonly encountered ant attendant of the Baccharis aphid, Aphis coreopsidis (Hemiptera: Aphidinae), on Mullet Key, the site used for the experiments described above. However, some aphid populations on Mullet Key are tended by the native ant Camponotus floridanus (Formicidae: Formicinae). The final experiment in this dissertation describes a nested experimental design conducted to compare the effects of ant species identity, early-season herbivory a nd the relative effects of bottom-up and topdown factors on common insects of the host plant. Study System Baccharis halimifolia also known as the groundsel tree or salt myrtle, is a native North American species that spans the east co ast of the United States from Florida to
6 Massachusetts and along the Gulf of Mexi co coast from Florida to Texas. Baccharis halimifolia is a dioecious perennial sh rub that grows to a height of 5 m. Trees flower in late autumn at which time male and female trees can be distinguished. An excellent colonizer of disturbe d or denuded habitats, B. halimifolia typically produces large numbers of air-dispersed seeds (P almer 1987). Allelochemicals from B. halimifolia include the triterpenoid, baccharis oxide, a cardiac glycoside, and an acetone soluble resin (Anthonse et al. 1970, Kraft and Denno 1982, Krischik & Denno 1990, Hudson 1995). In addition, leaves of B. halimifolia increase in toughness and resin content while also decreasing in both nitrog en and water content throughout the year (Kraft & Denno 1982). Baccharis halimifolia is associated with a large number of phyt ophagous insects (Palmer 1987). The most important of these is the specialist beetle Trirhabda bacharidis Weber (Coleoptera: Chrysomelidae). The univoltine Trirhabda bacharidis consumes B. halimifolia both in the larval stage and as an adu lt. Adult female beetles lay their eggs within deep grooves of B. halimifolia tree bark and in the late winter, stimulated by warmer temperatures and coinciding with newly budding leaves, larvae emerge to feed on foliage. Larvae pass through three instars in 35 40 days before pupating in the soil for an additional 20 26 days. Adults then em erge from the soil and return to the host to feed on foliar re-growth, remaining on the hos t for a further six weeks (Boldt 1989). Trirhabda bacharidis larvae are capable of complete ly defoliating a tree and their presence on the host plant has negative effects on the densities of other herbivores on the host plant later in the year (Hudson and Stiling 1997).
7 The field site selected for all experiment s described in this dissertation was Fort De Soto Park on Mullet Key, St. Petersburg, Flor ida. Mullet Key is an island that lies between the mouth of Tampa Bay and th e Gulf of Mexico. In addition to T. bacharidis five other herbivorous insects are common on B. halimifolia These are the monophagous gall fly Neolasioptera lathami Gagne (Diptera: Cecidomyiidae), two leaf miners: the polyphagous serpentine miner Liriomyza trifolii Burgess (Diptera: Agromyzidae) and the oligogophagous blotch miner Amauromyza maculosa Malloch (Diptera: Agromyzidae); the monophagous stem boring plume moth Oidaematophorus balanotes Meyrick (Lepidoptera: Pterophor idae); and the myrmecophilous Aphis coreopsidis (Hemiptera: Aphidinae). The a phids are preyed upon by a suite of coccinellids, both larv ae and adults, and syrphid larvae. The ant attendants on the host plant were either the native Camponotus floridanus or the exotic Linepithema humile Ant attendance on hemipterans well docum ented. The degree of association between hemipterans and ants can vary fr om untended hemipterans, which may be preyed upon by ants, to facultative species th at may be tended by multiple species of ants, to obligate myrmecophiles that are mutualists with particular species of ants (Buckley 1987). Myrmecophilous hemipterans provide the ant foragers with honeydew in exchange for protection from predators and parasites (Buckley 1987). Of the myrmecophilous aphids, the subfamily Aphidi nae contains the largest proportion of myrmecophiles and these are mostly facultative in nature with a few species that exhibit an obligate mutualism with ants (Stadler et al. 2003). Some or al l of the following are considered to be important benefits to an t-tended hemipterans: (1) protection against parasites and predators; (2) reduction of fungal infections that can result from an
8 accumulation of honeydew; (3) increases in f eeding rates and fecundity (Addicott 1979, Buckley 1987); and (4) reduction of potential competition from other nonmyrmecophilous herbivores on the host plan t (Messina 1981). Potential costs of attracting ant attendant are (1) the potentia l for predation by ants, (2) restriction to habitats where ant attendants are pres ent (Stadler 2004), and (3) production of competitively superior honeydew (Addicott 1979, Sakata and Hashimoto 2000, Engel et al. 2001, Woodring et al. 2004). Ant-aphid mutualisms are conditional (Cushman and Whitham 1989) and may vary in time and space and by species of host plant, aphid and ant (Messina 1981, Fritz 1983, Mooney and Tillb erg 2005). Ant-tended aphids may be less sensitive, in terms of altering their grow th rates, to changes in host plant quality relative to untended species, a consequence which may work as both a cost and benefit (Stadler et al. 2002). Further, ant attendant s can be both a limited and limiting resource for myrmecophilous aphids (Cus hman and Addicott 1989).
9 CHAPTER 2 MYRMECOPHILY ON BACCHARIS HALIMIFOLIA Synopsis The main objective of this experiment wa s to evaluate the effects of aphid-tending ant density myrmecophilous and non -myrmecophilous herbivores of Baccharis halimifolia, via field manipulations of ant attendan ce levels. High and low levels of ant attendance were established for a five month evaluation of the effects of ant densities on Aphis coreopsidis aphidophages, and th e two leaf miners, Amauromyza maculosa and Liriomyza trifolii High levels of ant attendance resu lted in significantly higher densities of both aphids and their predators over the course of the five month experiment. The densities of the two leaf miners were redu ced on high attendance host plants although this trend was not significant despite a consequent significant incr ease in ant-damaged mines. In a complementary short-term (20-day) e xperiment, aphid populati ons in ant-attended colonies increased in size and persisted longe r than those in non-an t attended colonies; and aphid predators remained more abundant on non-ant attended aphid colonies until the colonies went locally extinct. The results of this experiment sugge st that ant tending by L. humile on B. halimifolia more strongly affects insect s normally involved in the antaphid mutualism.
10 Introduction Often in ant-hemipteran mutualisms ther e is a direct relati onship between aphid colony size and number of ants tending the colo ny with the strength of the relationship varying according to the species of ant and/ or aphid involved in the mutualism (Cushman and Addicott 1989, Engel et al. 2001, Fischer et al. 2001). Generally, ants increase their level of attendance on larger ap hid colonies; colonies that re ceive attendance by a greater number of ants, i.e. a greater ant-aphid ratio, receive greater fitness benefits such that the tending level of ants is an important determ inant of aphid fitness and colony persistence (Cushman and Addicott 1989, Breton and A ddicott 1992). Previous studies have primarily relied on ant exclusion methods to elucidate patterns and mechanisms in anthemipteran mutualisms. This experiment ma nipulated ant densities in the field by both increasing and decreasing tending ant dens ities on open honeydew-producing hemipteran populations in order to evaluate the respons es of hemipterans, their predators and nonmyrmecophilous herbivores of a host plant. The design of this experiment is unique and provides several benefits. Manipulating ant dens ities on host plants in the field allows for a more biologically relevant interpretation of the dynamics of ant-hemipteran-natural enemy interactions as compared with greenhouse studies in which environmental heterogeneity is not incorporated. The durati on of this study also encompassed the time period during which hemipterans at this s ite are most abundant, namely spring and summer, and the effects of level of ant tend ing are evaluated acro ss time in search of generalizations about the broad consequences of ant tending for a community of insects on a host plant.
11 Specifically, this experiment was designe d to address the follo wing questions: (1) what are the consequences to aphids of vari able levels of ant a ttendance on a host plant; (2) how do differences in aphid abundances affect the distribution of aphid predators among host plants; and (3) what is the effect of variable degrees of myrmecophily on other non-myrmecophilous herbivores of the host plant? Methods Two experiments were conducted in fulfillm ent of the objectives just listed. The first of these, hereby referred to as Experi ment 1, was initiated in April 2003 at which time 24 B. halimifolia trees were selected The trees ranged in height from 1-2 m and trees were a minimum of 1 m apart. Each tree was randomly assigned to one of three treatments. The three treatment levels consisted of the following: (1) ant reduction via the application of Tanglefoot adhesive at the base of experime ntal trees; (2) ant enhancement via the use 4, 0.76 m long (2.5 cm diameter) PVC pipes half-buried vertically in the ground around the base of expe rimental trees; and (3) control, or ambient ant densities. Tanglefoot is a sticky nontoxic substance that immedi ately traps insects and is often used in ant exclusion experiments becau se it effectively restricts foraging ants. Tanglefoot was applied to experimental ant removal trees by applying the substance to the base of the trees and removing foraging an ts on the foliage with a large soft-bristle brush. Surrounding vegetation, such as gra ss blades and branches from neighboring trees, were clipped to prevent ants from circumventing the Tanglefoot and foraging on treatment trees. Ant enhancement was car ried out by inserting 4, 0.76 m long PVC pipes
12 half-buried vertically around the base of e xperimental trees in a front-back-side-side arrangement. The use of PVC pipes, typically for plot demarcation, in other salt marshes around Tampa Bay has been found to increase local ant densities because the pipes become colonized by ants (Altfeld, personal obse rvation). At the end of the experiment all PVC pipes were inspected for ant colonization. Post-treatment sampling was carried out monthly from MaySeptember 2003. On each sampling date, 25 randomly selected br anches per tree were visually inspected for numbers of ants, A. coreopsidis coccinellid adults and larvae, and syrphid larvae. Due to difficulty in counting individuals in the field, aphid abundances were estimated using the following categories: 0, 1-10, 11100, and 101-1000 individu als per branch. Leaf mine densities were estimated each month by counting the number of leaf mines on 200 randomly selected leaves per tree. Leaf miners were identified to species based on the morphology of the leaf mine; A. maculosa creates blotch mines and L. trifolii creates serpentine mines. Ant-damage d mines encountered in the 200-leaf subsample were also included in the data to represent the propor tion of total mines ripped/destroyed by foraging ants. In July 30th, 2004, Experiment 2 was initiated. This experiment represented an evaluation of the short-term dynamics of ant exclusion on aphid and predator abundance. For this experiment, 5 pairs of branches from four B. halimifolia trees that supported anttended aphid colonies were selected. Each br anch consisted of several separate stems all with their own growing tips. One branch fr om each of the 5 pairs received a band of Tanglefoot at the base of the branch where it met with the first division of the tree trunk and all ants on the branch were removed using a soft bristle brush. The second branch of
13 the pair was left un-manipulated. As suc h, each pair of branches represented an untended aphid colony and a tended aphid colony on a single host plant. The number of ants, aphids, syrphid larvae, coccinellid larvae and coccinellid adults were recorded once a day on days 1 4 and 20, post-treatment. Aphid abundance was recorded using abundance categories similar to the method used in Experiment 1 but with a finer scale of estimation that would enable smaller change s in aphid population si ze to be detected. The abundance categories were as follows : 0, 1-25, 26-50, 51-100, 100-200, +200. As in Experiment 1, the midpoint of each abundance category was used in the analysis of aphid data. Statistical analysis was carried out using SPSS version 12 for MS Windows. All variables were subject to Kolmogorov-Smirnov normality testing and Levenes homogeneity of variance test. Transformations were used when variables did not readily meet assumptions of normality and homosced asticity for ANOVA testing. The relevant comparisons in Experiment 1 were between hi gh and low levels of ant attendance. Prior to analysis, the categorical data on ap hid abundance was transformed for use as continuous data by taking the median of each category and transforming it using a fourth root transformation ((a + 0.5)1/4) (for similar method see Sloggett and Majerus 2000). Aphid predator densities, however, were unable to be transformed to meet assumptions of homoscedasticity and normality and were analyzed using the non parametric MannWhitney U. In Experiment 2, time was an im portant factor in the analysis of response by aphids and their predators to ant removal. As such, aphid and predator abundances in the second experiment were analyzed usi ng repeated-measures ANOVA. For both experiments, = 0.05 was designated as an appropriate level of statistical significance.
14 Results Experiment 1 The experiment was initially designed to provide eight replic ate host plant trees within each ant attendance level. However, owing to heterogeneity in the distribution and behavior of ants across the field site, no t all trees with PVC pipes were colonized and one control tree failed to maintain foraging ants. The following numbers of replicates were used in the analysis of the data : high n = 6, control n = 9, and low n = 9. Ant abundances were effectively alte red on experimental trees via field manipulations. During the experimental seas on, mean ant density in the high attendance level treatment was 3x greater than in the c ontrol (natural) attendance treatment (Table 2.1), 2.5x fewer in the low attendance level tr eatment and the difference between the high and low treatments was significant (F1,13 = 27.726, P < 0.001). Ant attendance level significantly affected the seasonal abundance of A. coreopsidis and their predators (Figure 2.1). Ap hids were more abundant in the high attendance level treatment compared to the low attendance level treatment (Table 2.1, F1,13 = 7.781, P = 0.015). Aphid predators consis ted of 3.6% larval coccinellids, 44.4% adult coccinellids and 52% larval syrphids (Table 2.1). Pred ator densities were greater on trees with higher densities of their prey, na mely trees in the high ant density treatment (U = 48.5, P = 0.010). The densities of the two sp ecies of leaf miners were not significantly affected by level of ant attendance (Table 2.1, A. maculosa F1,13 = 2.418; P = 0.144; L. trifolii F1,13 = 0.04; P = 0.845), despite a significantly greater number of ant-damaged mines on trees in the high ant density treatment (F1,13 = 13.787; P = 0.003).
15 Table 2.1. Mean SE abundances of myrm ecophilous and non-myrmecophilous insects on the host plant in each ant attendance level treatment. High Control Low Ants 770.0 150.4 249.7 71.8 100.9 32.8 A. coreopsidis 2986.7 1278.8 1161.7 502.4 194.4 89.4 Predators 14.7 6.2 10.0 2.0 4.4 0.5 Coccinellids 6.3 2.3 7.0 0.78 3.4 0.69 Syrphids 14.8 7.3 5.0 2.7 1.0 0.29 Other Herbivores Intact Mines 103.2 19.4 84.2 7.8 74.9 4.7 A. maculosa 81.0 17.8 65.2 6.2 51.8 3.4 L. trifolii 22.2 3.4 19.0 3.9 23.1 3.1 Ant Damaged Mines 65.0 11.3 48.6 9.4 24.8 4.8
16 Ant Density Treatment HLEstimated Aphid Density 0 1000 2000 3000 4000 5000 Predator Density 0 10 20 30 40 50 Aphids Predators Figure 2.1. Seasonal mean density of aphids a nd aphid predators 1 standard error on 25 B. halimifolia branches per tree from May through September 2003 in the high (H; n = 6) versus low (L; n = 9) ant density treatments.
17 Experiment 2 The results from the short-term antexclusion experiment showed that the application of Tanglefoot on an t-exclusion branches was fully effective. Ants were completely excluded on branches where the sticky substance was applied. The abundance of aphids on branches that excluded ants showed steady decline (Figure 2.2) and by day 20 the estimated number of aphi ds in the ant-exclusion treatment was significantly lower than in corresponding tended colonies. A repeated measures ANOVA provided significant time and time-treatmen t effects (P=0.003 and P=0.028 respectively). The treatment effect alone, however, was not significant (P=0.095) because it wasnt until day 20 of ant exclusion that untended a phid colonies were sufficiently reduced to show a significant difference from the tende d colonies. The abundance of aphidophages, namely coccinellid adults, larvae and syr phid larvae, also responded to ant exclusion (Figure 2.3). Repeated measures ANOVA for predators produced significant time (P=0.035), time-treatment (P=0.028), and tr eatment effects (P=0.001). Predator abundance responded to ant exclusion by day 2 of the experiment with predators remaining more abundant in untended aphi d colonies until the aphids disappeared.
18 Sampling Day 123420Estimated No. Individuals 0 100 200 300 400 Ants Excluded Ants Present Figure 2.2. Mean 1 standard error estimat ed number of aphids in paired anttended/untended populations.
19 Sampling Day 123420No. Individuals 0 1 2 3 4 5 6 Ants Absent Ants Present Figure 2.3. Comparison of the mean 1 sta ndard error number of aphid predators in anttended versus untended aphid populations.
20 Discussion Using Tanglefoot to exclude ants in the field was challenging because of the physical disturbances caused by weather and tides that were fr equent at the field site and because the ants were persistent in their determination to forage in the canopy of host trees. However, I was able to create a treatment which resulted in 59.6% fewer foraging ants as compared to control trees over the c ourse of the experiment. The reduction in ant abundance was biologically relevant to aphids and aphidophages, whic h had statistically lower abundances in this treatment. Restri cting ant foragers from reaching the aphids they tended prevented aphid pe rsistence on the host plant ove r the course of the season, i.e. late spring and summer. When aphid popul ations were left unte nded they were more efficiently preyed upon by coccinellids and sy rphid larvae as was shown in the rapid demise of untended aphid populations concurrent with the increase in predator abundance in Experiment 2. In tended aphid populati ons, however, the efficiency of predators was either inhibited by ant attendants, or mort ality by predators was compensated for by population growth enhancement due to ant tending, or both. Examples of both negative and positive density dependence have been shown to exist in ant-hemipteran mutualisms. Addicott (1979) suggests that ants seem to have the greatest positive effect on re latively small aphid populations via predator exclusion and increases to aphid feeding rates while the posi tive effects of ants on relatively large aphid populations diminish due to low ant-aphid ratios. Breton and A ddicott (1992) showed that small aphid populations (<30 individuals) of Aphis varians on fireweed benefited more from ant attendance than did large aphi d populations (>30 indivi duals) in terms of increased growth of local populations that was attributed to the per capita increase of ant
21 attendants in small aphid colonies. However, large colonies of the ant tended membracid, Publilia modesta, received increased benefits of 36 46% relative to small colonies (Cushman and Addicott 1991). On B. halimifolia, A. coreopsidis populations persist on host plants longer when tended by L. humile than when left untended, even when aphid populations are large (> 30 individuals). When evaluating the accumulated effect s of ant exclusion on the abundances of aphids and their predators, an understanding of the relevant ti me frame of the system is important. Aphid predators in this system do seem to prefer foraging on untended aphid colonies. However, results from the short term experiment suggest that colonies of aphids are quickly consumed to extinction. These results concur with other studies that show the rapid extinction of aphid colonies by coccinellid and syr phid predators in the absence of ant attendants (Addicott 1979, Fl att and Weisser 2000, Shingleton and Foster 2000, Yao et al. 2000). However, results from Experiment 1 also showed that in the relative absence of unt ended aphid colonies the predat ors are more abundant on trees where their prey more consistently persists in ant-tended colonies. The results of this study are similar to those of Sloggett and Maje rus (2000) in which a greater proportion of predators were found on aphid colonies tended by Formica rufa during times of untended aphid scarcity and the persistence of tended a phid colonies provided an alternative source of food, a situation which they called aphid-me diated coexistence of ladybird beetles and ants. An additional aim of the current study was to look at potential predation on leaf miners by aphid-tending ants. The two common leaf miners, A. maculosa and L. trifolii, were unaffected by in creased density of L. humile on the host plant despite increases in
22 the number of ant-damaged mines. This suggest s that either ant pred ation is not critical or that compensatory mortality is occurring at some later stage in the life cycle. Clearly, the presence of ant attendant s can have dramatic effects on those organisms which are involved in the mutualism. But in terms of effects of the mutualism on other herbivores the patterns are less clear. Studies of ant guards on ant plants have produced equivocal results. He rbivory on plants with extrafloral nectaries has been shown to be reduced (Dyer and Letourneau 1999) or unaffected (Mody and Linsenmair 2004) in the presence of ant guards. On plan ts that host ant-hemipteran mutualisms, the evidence supports a reduction in herbivory in th e presence of the mutualism, particularly when the herbivores are foliage feeding b eetles and their larvae (Dansa and Rocha 1992, Floate and Whitham 1994). This effect of reduced herbivory, however, depends in part on whether or not the tending ants ar e aggressive (Messina 1981).
23 CHAPTER 3 NITROGEN LIMITATION, COMP ETITION AND MYRMECOPHILY Synopsis Here I test the effects of (1) ni trogen enrichment, (2) competition by Trirhabda bacharidis (Coleoptera: Chrysomelidae) and (3) aphid-tending by Linepithema humile (Hymenoptera: Dolichoderinae) on insects of B. halimifolia using a fully factorial design. The presence of ants had a greater eff ect on herbivore densities than either T. bacharidis larvae presence or nitrogen enrichment: aphi ds were more abundant but leaf miners and stemborers were less abundant. Ant presence increased tree survival because stemborers frequently killed host plants. Competition by T. bacharidis affected aphids and their predators which were more abundant on trees with reduced herbivory. Nitrogen fertilization had no effect on most herbivores although T. bacharidis larvae achieved higher densities and earlier pupati on on fertilized trees. These results indicate that aphid tending by L. humile affects other insects on B. halimifolia more so than herbivory by an exploitative competitor such as T. bacharidis or nitrogen limitation. Introduction Fertilizing host plants with nitrogen ofte n results in increased herbivore densities because of the frequency of nitrogen limita tion for insect herbivores (Mattson 1980,
24 Waring and Cobb 1989, Awmack and Leather 2002) However, the absence of increased herbivore densities on fertiliz ed host plants can indicate th at other factors, such as competition or predation, may be preventi ng the expression of positive effects from nitrogen fertilization (Kyt et al 1996, Strengbom et al. 2005, Moe et al. 2005). Predator manipulations, unlike manipulations of host plan t quality, often result in either speciesspecific effects on herbivore densities becau se of the unique interactions between herbivores and their natural enemies or unpredicted responses in herbivore densities caused by environmental heterogeneity (Moon and Stiling 2002). In addition, the outcomes of interspecific interactions among co-occurring herbivores can be affected by both host plant quality and predators (Strauss 1987, Inbar et al. 1995, Denno et al. 2000, Vestergrd et al. 2004). No one theory of community dynamics predicts the outcomes of species interactions among multiple species of herbivores, mutualists and predators that share a host plant nor how host plant quality can affect those species interactions. Here I evaluated how host plant quality affects competition among herbivores, the intensity of a mutualism between aphids and ants, and predation by ants. The community of common insects on the groundsel tree, Baccharis halimifolia Linnaeus (Asteraceae) provides a model for evaluation of the relative effects of nitrogen limitation, competition between herbivores, and predation by ants be cause each of these factors have previously been found to be important. Common insects of the host plant were nitrogen limited (Moon and Stiling 2004), the pr esence of aphid-tending ants led to predation on leaf miners (Altfeld and Stiling 2006), and the presence of an early-season herbivore, Trirhabda bacharidis (Coleoptera: Chrysomelidae), ne gatively affected late-season herbivores (Hudson and Stiling 1997).
25 Specifically I tested whether nitrogen fe rtilization alleviated or enhanced (1) competition between T. bacharidis and other herbivores a nd (2) predation on other herbivores by aphid-tending ants. My hypot heses were that, (1) if fertilized trees maintained higher densities of T. bacharidis larvae, then the densitie s of at least some of the other herbivores would be lower on those trees because of greater exploitative use of the host plant; and (2) ant presence would have positive effects on aphids and their predators but negative effects on leaf miners because of mine predation by the ants. Thus, nitrogen fertilization might not elicit increases in the densities of all insects utilizing the host plant because of interactiv e effects of predation and competition. Methods All experimental trees were between 1-2 m in height and were located within a 3.2 km long stretch of coastline shoreward of a mangrove fringe along the east coast of Mullet Key. In December 2003, a total of 40 B. halimifolia trees, 20 trees with primarily smooth bark and 20 trees with primarily groov ed bark, were selected, based on size and the presence of aphids, for random assignment to experimental treatments in a 2 x 2 x 2 factorial design with five repl icates. The three factors were nitrogen fertilization (added versus control), T. bacharidis larvae abundance (grooved vers us smooth bark), and ant attendants (present versus absent). The experiment was conducted for 9 months from January through August and October 2004. The passage of several tropical storms in September precluded data collection.
26 Nitrogen fertilization was accomplished vi a monthly applications of 40g granular nitrogen fertilizer (46-0-0 urea fertilizer, PCS Sales Skokie, IL) around the base of experimental trees. The density of T. bacharidis larvae was manipulated by arbitrarily selecting host trees with eith er primarily smooth or grooved tree bark. The selection of one of two bark phonologies provided a natural treatment for T. bacharidis larvae presence / absence because female beetles te nd to deposit eggs in th e grooves of host tree bark (Boldt 1989, Hudson and Stiling 1997) su ch that host trees with primarily smooth bark have few to no larvae. Ant attendan ce was manipulated by excluding ant foragers on experimental trees using a band of Tangl efoot (The Tanglefoot Company, Grand Rapids, MI) around the trunk of experiment al trees and surrounding vegetation was clipped to prevent ants from bypassing the Tanglefoot and foraging in the trees. Insect abundances were estimated at th e beginning of each month from January through October using three census methods. First, 25 branches, each a minimum of 10 cm in length, were arbitrarily c hosen and the numbers of individual T. bacharidis larvae and adults, ants, aphids, coccinellids and sy rphid larvae were determined by visual inspection. Aphid abundance on each branch wa s assigned to the following categories: 0, 1-25, 26-50, 51-100, 101-200, 201+ individuals. Second, 200 leaves were inspected for leaf mines and were categorized by leaf miner species and as intact or ant-damaged. Third, 200 different branches, also 10 cm minimum length, were examined for the number of galls. Finally, a single count of recent stemborer emergence holes on host plant branches was made by visually inspec ting all branches on each host tree in August during the peak in stemborer activity. Recen t emergence holes were identified by the presence of frass around the holes.
27 In order to quantify differences in foliar nitrogen between fertilized and unfertilized trees, foliar nitr ogen was analyzed monthly fo r 10 randomly collected leaves per tree. Leaves were oven dried at 50o C for 3 days, ground in a Wiley mill, and analyzed for total nitrogen as percent dry mass (CN 2100 Soil Analyze, CE Elantech, Inc. Lakewood, NJ). Treatment effects were tested on mont hly abundances of insects using fully factorial, repeated measures ANOVA. Prior to analysis, the categorical data on aphid abundance were transformed for use as cont inuous data by taking the median of each category to the fourth root ((a + 0.5)1/4, for similar method see Sloggett and Majerus 2000). The abundance of stemborer holes was tested using a univari ate ANOVA. Main factors for ANOVA tests were nitrogen fertilization (+/-), T. bacharidis larvae (+/-) and ants (+/-). The effect of the presence of ants on stembore r-induced mortality among host trees was tested by chi-square with the Yate s correction for continuity (Zar 1999). Prior to analysis all variables were subjecte d to Kolmogorov-Smirnov normality testing and Levenes homogeneity of vari ance test. All stat istical tests were carried out using SPSS Version 13 for Windows (SPSS, Inc. 2004). Results The addition of nitrogen fertilizer si gnificantly increased foliar nitrogen in B. halimifolia from 1.67% 0.05 (standard error) to 1.84% 0.05 (F1,32=6.364, P=0.017). Abundance of T. bacharidis larvae on host trees with smooth bark was significantly lower, 33 7 SE (F1,32=4.211, P=0.048,) than on host trees with grooved bark, 55 14 SE. Tanglefoot effectively reduced the number of foraging ants, 16 5 SE, on
28 exclusion trees (F1,32=39.323, P<0.001) compared to 108 21 ants on control trees. In addition, Tanglefoot also captured some T. bacharidis larvae such that there was a significant Tanglefoot by tree t ype interaction with the lo west densities of larvae on smooth trees banded with Tanglefoot. The presence of L. humile attendants influenced abundances of aphids, both species of leaf miners, ant-damaged mines, and stemborers. Aphids were present in significantly higher densities on trees with ant attendants (F1,32=9.557, P=0.004). The presence of ants significan tly reduced the abundances of both species of leaf miners, A. maculosa (F1,32=4.996, P=0.033) and L. trifolii (F=1,324.885, P=0.034) (Fig. 3.1), and the stemborer, O. balanotes (F1,32=25.835, P<.001). The reduction in the abundances of intact leaf mines coincided with significan tly higher numbers of ant-damaged mines on trees with ant foragers (F1,32=20.308, P<0.001) (Fig. 3.1). The abundance of N. lathami galls was unaffected by the presence of ant attendants. The abundance of T. bacharidis larvae on B. halimifolia had no effects on any of the insects in this study, apart from T. bacharidis adults, where adults were more abundant on grooved trees. However, the a bundances of ants, aphids and coccinellids were highest on trees with reduced numbers of T. bacharidis larvae when the tree had ant attendants (ants: F1,32=4.497, P=0.042; aphids: F1,32=4.648, P=0.03; coccinellids: F1,32=4.477, P=0.042) (Fig. 3.2). In addition, Trirhabda bacharidis larvae were more abundant and left their hosts earlier on fe rtilized trees (Fig. 3.3; significant time x nitrogen interaction: F2, 64 = 3.093, P = 0.052).
29 Finally, there was 47.5% mortality of experimental trees during September and October of 2004 from stemborers. Of the tr ees that died from stemborer damage, significantly more (15 out of 19) were ant-exclusion trees (2 = 5.2632, P<0.025).
30 Ants AbsentAnts PresentAbundance 0 10 20 30 40 50 60 Ant Damaged Mines A. maculosa L. Trifolii Figure 3.1. Mean seasonal abundances of ant-damaged mines, A maculosa and L. trifolii in the presence and absence of ant attendants. Bars represent 1 standard error.
31 Tree Bark Type SmoothGroovedNo. Coccinellids 0 2 4 6 8 10 No. Aphids and Ants 0 200 400 600 800 Coccinellids A. Coreopsidis Ants Figure 3.2. Mean seasonal abundances of aphids, ants and coccinellids on smooth and grooved trees in the presence of ant attendant s. Bars represent 1 standard error.
32 JanFebMarAprMayJunNo. Individuals 0 10 20 30 40 Nitrogen Addition Control T. bacharidis larvae T. bacharidis adults Figure 3.3. Mean monthly a bundance of larval and adult T. bacharidis, on control and fertilized trees. Bars repr esent 1 standard error.
33 Discussion The ant attendants, L. humile, on myrmecophilous aphids of B. halimifolia, had significant positive effects on aphids and cocci nellids but negative effects on leaf miners and stemborers. Nitrogen did not affect the densities of ant atte ndants nor affect the outcome of competition from T. bacharidis larvae. In fact, only Trirhabda larvae responded to nitrogen addition inferring that T. bacharidis larvae may be nitrogen limited and experience decreased development time an d/or increased survi vorship on fertilized trees. Invasion of non-native habitats by L. humile often causes significant decreases in native arthropod divers ity (Holway et al. 2002, Sanders et al. 2003). However, for honeydew producing myrmecophiles, L. humile can have positive effects. Linepithema humile makes an exceptional partner because th ey recruit to honeydew sources quickly, tend 24 hours a day and have a modified prove ntriculus that increases honeydew uptake rates and carrying capacity, and assists in food sharing among nestmates (Holway et al. 2002, Davidson et al. 2004, Ness and Bronstein 2004). Aphid-tending ants, whether native or non-native, have significant effects on aphid colony size via predator exclusion in a variety of systems (Addicott 1979, Flatt and Weisser 2000, Shingleton and Fost er 2000, Yao et al. 2000). On B. halimifolia, L. humile has a positive effect on aphid density but di fferentially affects co ccinellids depending on the spatial scale used for observation. On individual host trees, the presence of L. humile reduces coccinellid densities; among a sta nd of host trees, coccinellid densities are positively affected by the presence of L. humile by providing alternative resources when untended aphid populations are ra re (Altfeld and Stil ing 2006). The positive effects of
34 ant attendants on coccinellids is facilitated by the relatively higher densities of aphids on trees with ant attendants (Sloggett and Maje rus 2000) and a positive relationship between aphid density and oviposition preference by fe male coccinellids (Evans and Dixon 1986). The mechanism by which ants benefit pl ants by removing herbivores is complex and often strongly related to the species of ant and th e presence of myrmecophilous aphids (Skinner and Whittaker 1981, Del-Clar o and Oliveira 2000, Su zuki et al. 2004). Two other Trirhabda spp. on Solidago were differentially affected by the presence of ant attendants, depending on the size and aggressiveness of the species of ant present. The larger and more aggressive Formica sp. had significant negative effects on beetles but the smaller ant species, Myrmica sp.and Prenolopis sp., had no effects (Messina 1981). Even herbivores that feed within protective stru ctures such as mines are not necessarily protected from predation by ants (Faeth 1980, Fowler and MacGarvin 1985, Pezzolesi and Hager 1994, Fagundes et al. 2005). Camponotus sp. and Crematogaster sp. of ant attendants have significant negative effects on a psyllid gall fly of Baccharis dracunculifolia (Fagundes et al. 2005). Linepithema humile, in particular, can have strong negative effects on he rbivory levels experienced by a host plant via herbivore exclusion or predation (Way 1963, Way et al 1992, Way et al. 1999) especially because they can reach high densities where they te nd honeydew-producing in sects (Holway et al. 2002). On B. halimifolia, the presence of L. humile was responsible for a trophic cascade because the presence of ants significan tly reduced stemborer-induced tree death.
35 CHAPTER 4 PHOSPHORUS LIMITATION Synopsis This experiment was designed to eval uate the limiting value of phosphorus for insect herbivores on B. halimifolia. Experimental trees were fertilized with phosphorus and nitrogen + phosphorus and insect densities compared on experimental trees relative to untreated control pl ants. As predicted, T. bacharidis larvae responded positively to phosphorus fertilization. However, adult be etles showed no effect s of fertilization. Aphids and the two species of leaf mine rs responded negatively to phosphorus fertilization although this result is considered to be an artifac t of host trees not re-flushing with leaves after T. bacharidis larval herbivory. Results suggest that T. bacharidis is phosphorus limited, likely due to the rapid deve lopment of this univoltine insect, but the limiting value of phosphorus of the othe r insect herbivores, especially A. coreopsidis, bears further investigation. Introduction Phosphorus is an important limiting nutrient for herbivores. Phosphorus represents only about 0.1 to 0.8% of plant tissue by mass but is an essential element in the construction of nucleic ac ids, amino acids and energy transferring molecules such as
36 ATP and NADP (Sterner and Elser 1998). While the positive effects of nitrogen fertilization have frequently been reported for insect herbiv ores, the effects of phosphorus fertilization on insect herbi vores are much less well documented. However, due to the evidence of phosphorus limitation of organisms such as Daphnia and the theory that rapidly growing organisms are more likely to be phosphorus limited due to the high phosphorus demand of RNA produc tion during growth, more st udies that evaluate the role of phosphorus as a limiting nutrient are being conducted. The insects on B. halimifolia that may be most sensitive to phosphor us limitation due to characteristic rapid growth are the specialist beetle T. bacharidis, which completes its entire life cycle within 6 months, and aphids. Phloem feeders in pa rticular have been suggested to experience extreme stoichiometric imbalances, a condi tion which may be exa ggerated in aphids. The objectives of this experiment were to evaluate (1) whether herbivores on B. halimifolia exhibit phosphorus limitati on and (2) if nitrogen an d phosphorus fertilization would have additive effects on herbivore densities. Methods In February, 2004 12 B. halimifolia host trees, ranging in height from 1 2 m and separated by a minimum distan ce of 2 m, were assigned to one of two fertilization treatments (n = 6). The treatments were (1) phosphorus fertiliza tion or (2) nitrogen + phosphorus fertilization. All tr ees were fertilized with a one -time application of 74 g of dry calcium phosphate fertiliz er (total phosphoric acid 16%, Loncola Inc., High Springs,
37 FL). Trees in the nitrogen + phosphorus treatment were also fertilized with an additional 50 g of nitrogen fertilizer (46-0-0 urea fertilizer, PCS Sales Skokie, IL). The densities of T. bacharidis larvae and adults, A. coreopsidis, ants, aphid predators, A. maculosa, L. trifolii and N. lathami were estimated monthly from March through October. Insect densities were meas ured using 3 census methods. First, 25 branches, each a minimum of 10 cm in length, were arbitrarily chosen and the numbers of individual T. bacharidis larvae and adults, ants, aphids, coccinellids and syrphid larvae were determined by visual in spection. Aphid abundance on e ach branch was assigned to the following categories: 0, 1-25, 26-50, 51-100, 101-200, 201+ individuals. Second, 200 leaves were inspected for leaf mines and were categorized by leaf miner species. Third, 200 different branches, also 10 cm minimum length, were examined for the number of galls. Foliar nitrogen and phosphorus content were not measured for host trees in this experiment because most trees did not re-flu sh with leaves after larval herbivory by T. bacharidis. The densities of insects on the host plant were compared among the following treatments: control, phosphor us fertilization and nitrogen + phosphorus fertilization. Data for insect abundances on control, unf ertilized, host plants came from 6 randomly selected host trees that were also being used in a concurrent experi ment (see Chapter 3). Treatment differences were evaluate d with ANOVA. The densities of A. coreopsidis were fourth root (x + 0.5) transformed and L. humile square root (x + 0.5) transformed to meet assumptions of normality and homoscedasticity.
38 Results The densities of T. bacharidis larvae were higher on hos t plants fertilized with phosphorus and nitrogen + phosphor us relative to control (F2,15 = 4.745, P = 0.025). There were no additive effects for the combination of nitrogen and phosphorus fertilization (Figure 4.1). Trirhabda bacharidis adults exhibited no significant effects of fertilization by either phosphor us or nitrogen + phosphorus (Figure 4.1). The densities of A. coreopsidis and L. humile were negatively affected by phosphorus and nitrogen + phosphorus fertilization relativ e to control host trees (F2,15 = 4.044; P = 0.055 and F2,15 = 4.437; P = 0.031, respectively) but aphid pred ators were unaffected (Figure 4.2). Leaf miners were also negatively affected by phosphorus and nitrogen + phosphorus fertilization treatments (Figure 4.3. A. maculosa: F2,15 = 7.249; P = 0.006 and L. trifolii: F2,15 = 4.046; P = 0.039). The gall fly, N. lathami, was unaffected by phosphorus and nitrogen + phosphorus fertilization treatments.
39 Fertilization Treatment CPNPNo. Individuals 0 10 20 30 40 50 60 Larvae Adults a b ab A A A Figure 4.1. Mean 1 SE abundances of T. bacharidis on host plants under different fertilization treatments.
40 Fertilization Treatment CPNPNo. Individuals 0 100 200 300 400 500 No. Individuals 1 2 3 4 5 6 7 A. coreopsidis Ants Aphid Predators a ab b A B AB Figure 4.2. Mean 1 SE abundances of L. humile, A. coreopsidis and aphid predators on host plants under different fe rtilization treatments.
41 Fertilization Treatment CPNPAbundance 0 10 20 30 40 A. maculosa L. trifolii a b b A B B Figure 4.3. Mean 1 SE abundances of A. maculosa and L. trifolii on host plants under different fertilization treatments.
42 Discussion Trirhabda bacharidis was the only herbivore positively affected by phosphorus fertilization and only during the larval stag e. This insect was predicted to respond positively to phosphorus enrichment of the host pl ant because this beetle is univoltine and has rapid growth during its larval stage. Adult beetles, however, showed no preference for fertilized trees. The majority of host plan ts in this experiment did not recover from T. bacharidis larval herbivory by re-flushing with new leaves. This lack of recovery from larval herbivory is more likely to ha ve caused the observed negative effects of phosphorus fertilization in A. coreopsidis, ants, A. maculosa and L. trifolii than changes in host plant quality.
43 CHAPTER 5 EXOTIC VERSUS NATIVE ANTS Synopsis I compared the effects of native ants, Camponotus floridanus, and exotic ants, Linepithema humile, on plant-insect interactions on B. halimifolia. I used a nested design to determine whether (1) the identity of aphi d-tending ant matters to co-occurring insects on the host plant, (2) the speci es of ant changes the effect s of early-season herbivory on herbivores that occur on the host plant later in the year, a nd (3) if the strength of bottomup and top-down factors changes in the presence of an exotic species of ant, compared to the presence of a native ant. Densities of all insects except the stem-galling fly Neolasioptera lathami were increased on host plants with exotic ants. The negative effects of early-season herbivory on later feeding leaf miners and a gall fl y were more pronounced on host plants with native ants than on host plants with exotic ants On plants with exotic ants, only the gall fly was significantly reduced by early-season herbivory. In ad dition, bottom-up effects of foliar nitrogen increases were more likely to be translated into increased densities of leaf miners on host plants with native ants. Top-down effects of an t predation were only important for herbivores when exotic ants we re present on the host plant. These results
44 suggest that community-level processes can be significantly affected by identity of ants and early-season herbivory on a host plant. Introduction Ants are among the most common omnivores in terrestrial sy stems (Holldobler and Wilson 1990). Ant species vary in th eir abilities to act as attendants to myrmecophilous insects or as predators on other insects (Messina 1981, Buckley 1987, Freitas and Oliveira 1996, Mody and Lins enmair 2004). In its native range, the Argentine ant, Linepithema humile (Hymenoptera: Dolichod erinae), establishes mutualisms with honeydew producing hemipter ans in their native areas (Holway et al. 2002) and makes an excellent mutualist because it recruits to honeydew sources quickly, tends 24 hours a day and has a modified pr oventriculus that in creases honeydew uptake rates and carrying capacity (H olway et al. 2002, Davidson et al. 2004, Ness and Bronstein 2004). Outside its native range, however, th is ant reduces native ant and arthropod diversity and has been associated with outbreaks of honeydew-s ecreting hemipterans (Way 1963, Ness and Bronstein 2004). In the U.S., L. humile has spread widely and has the potential to impact many native insect communities. The direct and indirect effects of an ts on insects that co-occur on a host plant differ depending on the feeding guild of the inse ct and its association with ants (Mahdi and Whittaker 1993). Myrmecophilous insects benefit from the presence of ants because ants provide protection from natural enemie s, reduce incidences of fungal infections, increase feeding rates and fecundity (Addicott 1979, Buckley 1987); and reduce competition from other non-myrmecophilous herbivores on the host plant (Messina
45 1981). In Florida, the presence of L. humile on the host plant, Baccharis halimifolia (Asteraceae), strongly affect s the myrmecophilous aphid Aphis coreopsidis and its predators and, potentially, other members of the insect community (Altfeld and Stiling 2006). There are, however, many other selectiv e pressures which may generate in this system. To begin with, not all host plants are patrolled by exotic Argentine ants. Many support only patrols of the native Camponotus floridanus, a much larger species. Host plants tend to have only one species of ant, not both. In addition, the community can also be negatively affected by early-season herbi vory from larvae of the specialist beetle, Trirhabda bacharidis which defoliates plants and re duces subsequent densities of herbivores which appear after the leaves re-flush (Hudson and Stiling 1997). Most herbivores are also affected by increases in foliar nitrogen (Moon and Stiling 2004). Here I evaluated the extent to which th e responses of the insect community on B. halimifolia to both host plant quality and predators are changed in the presence of native or exotic ants. Specifically I examined how the species of ant on a host plant affects interspecific interactions among herbivores under differing conditions of early-season herbivory (present / absent), hos t plant quality (increased / c ontrol) and predation by ants (present / absent). I manipulated ant species, early-seaso n herbivory, foliar nitrogen and ant predation using a nested design and made the following predic tions. (1) The presence of an exotic ant on the host plant will affect th e densities of all insects utilizing the host plant. The direction of the effects of the exotic ant will differ depending on whether the insects are or are not involved in the antaphid mutualism. (2) Early-season herbivory will negatively affect all herbivores of the hos t plant, regardless of the species of ant on
46 the host plant. (3) Bottom-up factors will be important for all insects of the host plant but will increase in importance in the abse nce of early-season herbivory and for nonmyrmecophilous herbivores. (4) Top-down eff ects of ants will be species-specific, affecting only those insects either closely a ssociated with ants or more vulnerable to predation. Methods All experimental trees were located with in Fort De Soto Park on Mullet Key, St. Petersburg, Florida, USA. Host plants sel ected for inclusion in this experiment were between 1 2 m in height, a minimum of 2 m apart and had either native C. floridanus or exotic L. humile ants tending A. coreopsidis. In January 2005, 80 B. halimifolia trees were assigned to experimental groups based on a 3-tier hierarchical design. The main factor (1st tier) was aphid-tending ant species, C. floridanus or L. humile. Nested within ant species was the presence of ab sence of early-season herbivory by T. bacharidis larvae (2nd tier). Nested within ant species and ear ly-season herbivory were foliar nitrogen (control and enriched) and ant foragers (absen t and present) applied in a 2 x 2 design. Data was collected for 4 months from April to July 2005, when insect densities were highest on the host plant. Ant attendance and early-season herbivory by T. bacharidis larvae were manipulated by applying either a temporar y or permanent band of Tanglefoot (The Tanglefoot Company, Grand Ra pids, MI) around the base of experimental trees. To prevent early-season herbivory and exclude ants, a permanent band of Tanglefoot was placed around the base of the trunks of e xperimental trees, but above where the T.
47bacharidis eggs were laid. To prevent early-seas on herbivory but allow ants, a temporary band of Tanglefoot was placed on trees onl y for the three months, January through March, when T. bacharidis larvae were actively hatching an d feeding on host trees. After the larvae left host trees at the end of March, the tem porary Tanglefoot was removed from trees to allow the ants access to foliage. To allow early-season herbivory and exclude ants, a band of Tanglefoot was placed on trees after the larvae left host trees in late March. Host trees with early-season herbivory and ants present were free of Tanglefoot manipulations. The temporary bands of Tanglefoot cons isted of a band of tape wrapped around the trunk with Tanglefoot placed over the tape The tape was easily cut and peeled off host trees for removal when necessary. Bands were made permanent by applying the Tanglefoot directly to host trees and reapplying when necessary. In addition, vegetation surrounding ant-excl usion trees was clipped to prevent ants from bypassing the Tanglefoot. Foliar nitrogen enrichment was achieved via monthly applications of 60g granular nitrogen-based fertilizer (460-0 urea fertilizer, PCS Sales Skokie, IL) around the base of experimental trees. In order to quantify diffe rences in foliar nitrogen between fertilized and unfertilized trees, foliar nitrogen was an alyzed monthly for 10 randomly collected leaves from all trees at both site s. Leaves were oven dried at 50o C for 3 days, ground in a Wiley mill, and analyzed for total n itrogen as percent dry mass (CN 2100 Soil Analyzer, CE Elantech, Inc. Lakewood, NJ). Beginning in April, insect abundances at each site were estimated monthly using three census methods. First, 25 branches, each a minimum of 10 cm in length, were
48 arbitrarily selected and the numbers of individual T. bacharidis adults, ants, aphids, coccinellids and syrphid larvae were reco rded. Aphid abundance on each branch was assigned to the following cat egories: 0, 1-25, 26-50, 51100, 101-200, 201+ individuals. Second, 200 leaves were inspected for leaf mi nes, which were categ orized by species and presence or absence of ant damage. Third, 200 different stems, each a minimum of 10 cm in length, were examined for galls. In addition, a single count of recent stemborer emergence holes on host plant branches was ma de by visually inspecting all branches on each host tree in August during peak stemborer activity. Recent emergence holes were identified by the presence of frass around th e holes. In September, host tree mortality due to stemborer damage was evaluated. Data were analyzed using a nested A NOVA design with ant species as the main factor, early-season herbivory by T. bacharidis larvae nested within ant species, and foliar nitrogen (control / enriched) and ant fo ragers (present / absent) fully crossed and nested within larval herbivory. The abunda nces of aphids and aphid predators were square root (x + 0.5) transformed to meet the assumptions of normality and homogeneity of variance. Prior to analysis all va riables were subjected to Kolmogorov-Smirnov normality testing and Levenes homogeneity of variance test. All statistical tests were carried out in SPSS Version 13 for Windows (SPSS, Inc. 2004). Results The densities of all insects, except the specialist gall fly, N. lathami, were significantly affected by the species of ants foraging on the host plant (Table 5.1). The densities of ants, A. coreopsidis and aphid predators were hi gher on host plants with the
49 exotic L. humile as compared to insect densities on host plants with the native ant C. floridanus (Figure 5.1a). The densities of herb ivores not involved in the ant-aphid mutualism, namely the leaf miners, gall fl y, and stem borer, were also higher on host plants with L. humile as compared to host plants with C. floridanus (Figure 5.1b.). Table 5.1. Ant species effects on insects of B. halimifolia. Results from Nested ANOVA with ant species as the main factor. Insect F1,64P Ants 46.908< 0.001 A. coreopsidis 60.118< 0.001 Aphid Predators 30.822< 0.001 T. bacharidis Adults 52.519< 0.001 A. maculosa 123.737< 0.001 L. trifolii 357.328< 0.001 N. lathami 0.1970.659 O. balanotes 43.888< 0.001 Ant-Damaged Mines 8.897< 0.001 Early-season feeding by T. bacharidis larvae affected the densities of A. coreopsidis, A. maculosa, L. trifolii, and N. lathami (Table 5.2). On host plants with native C. floridanus, the densities of A. maculosa, L. trifolii and N. lathami were increased and the density of A. coreopsidis decreased when early-season herbivory by T. bacharidis larvae was prevented (Figure 5.2). When exotic L. humile was present on host plants, however, only the specialist gall fly, N. lathami, was negatively affected by earlyseason herbivory (Figure 5.2).
50 Table 5.2. Effects of early-season feeding by T. bacharidis larvae on herbivores of B. halimifolia. Results from Nested ANOVA with Larval Herbivory as nested factor within Ant Species. Insect F2,64P Ants 1.4420.244 A. coreopsidis 3.2480.045 Aphid Predators 2.7460.072 T. bacharidis Adults 0.2450.784 A. maculosa 7.4410.001 L. trifolii 4.3380.017 N. lathami 17.5880.001 O. balanotes 0.4310.652 Foliar nitrogen increased on fertilized trees by an average of 17.25%. This increase in foliar nitrogen affected the densities of A. maculosa, L. trifolii, N. lathami and aphid predators (Table 5.3). However, speci es of ant and the presence or absence of early-season herbivory by T. bacharidis larvae influenced the extent to which herbivore densities were affected by increased folia r nitrogen (Figure 5.3). Aphid predators increased in response to nitrogen fert ilization only on host plants with native C. floridanus. However, on host plants with exotic L. humile, aphid predator densities were unaffected by fertilization while A. coreopsidis densities decreased. Amauromyza maculosa, L. trifolii, and N. lathami increased in response to n itrogen fertilization only in the absence of early -season herbivory by T. bacharidis larvae.
51 Table 5.3. Bottom-up effects from nitrogen fertilization on insects of B. halimifolia. Results from Nested ANOVA with Fertilizatio n nested within Larval Herbivory (nested within Ant Species). Insect F4,64P Ants 0.2160.929 A. coreopsidis 1.7890.142 Aphid Predators 2.7140.037 T. bacharidis Adults 0.7310.574 A. maculosa 5.5040.001 L. trifolii 7.035< 0.001 N. lathami 16.151< 0.001 O. balanotes 1.3930.246 The presence of ants affected A. coreopsidis, aphid predators, L. trifolii and antdamaged mines (Table 5.4). The effects, however, were specie s-specific. Aphid predators were less abundant in the presence of C. floridanus but more abundant in the presence of L. humile (Figure 5.4). Herbivores not i nvolved in the ant-aphid mutualism were unaffected by the presence of ants when the ant was the native C. floridanus but L. trifolii was less abundant in the presence of ants on host plants with exotic L. humile because the ants preyed on leaf mines (F igure 5.5). There were no significant interactions observed between bottom-up and top-down mani pulations for any insects except L. trifolii, which showed the higher densities on nitrogen enriched, ant-exclusion host plants compared to host plants that were either nitrogen enriched or ant-exclusion.
52 Table 5.4. Top-down effect s of ants on insects of B. halimifolia. Results from Nested ANOVA with Ant Foragers nested within Larval Herbivory (nested within Ant Species). Insect F4,64P A. coreopsidis 39.299< 0.001 Aphid Predators 6.546< 0.001 T. bacharidis Adults 2.6290.042 A. maculosa 2.0830.093 L. trifolii 28.987< 0.001 N. lathami 1.6810.165 O. balanotes 1.8590.129 Ant-Damaged Mines 8.897< 0.001
53 C. floridanusL. humileAbundance 0 100 200 300 400 500 Ants A. coreopsidis Aphid Predators *a. b. C. floridanusL. humileAbundance 0 20 40 60 80 100 120 140 160 T. bacharidis adults A. maculosa L. trifolii N. lathami O. balanotes Ant-Damaged Mines Figure 5.1. Comparison of mean 1 SE abundances of (a.) insects involved in the antaphid mutualism and (b.) other herbivores not involved in the ant-aphid mutualism, when the ant foragers were either the native C. floridanus or the exotic L. humile.
54 Invasive L. humile PresentAbsentAbundance 0 100 200 300 400 500 600 A. coreopsidis T. bacharidis adults A. maculosa L. trifolii N. lathami O. balanotes Native C. floridanus PresentAbsentAbundance 0 20 40 60 80 100 A. coreopsidis T. bacharidis adults A. maculosa L. trifolii N. lathami O. balanotes a. b. Figure 5.2. Comparison of mean 1 SE abunda nces of herbivores on host plants with the presence and absence of early-season herbivory by T. bacharidis larvae. Results are presented for host trees with (a.) native C. floridanus and (b) exotic L. humile ants.
55 ControlEnrichedAbundance 0 200 400 600 800 ControlEnrichedAbundance 0 200 400 600 800 ControlEnrichedAbundance 0 20 40 60 80 ControlEnrichedAbundance 0 20 40 60 80 100 120 140 A. coreopsidis T. bacharidis adults A. maculosa L. trifolii N. lathami O. balanotes a. b. c. d. Figure 5.3. Comparison of mean 1 SE abundances of herbivores on control and nitrogen enriched host plants with (a.) native C. floridanus and early-seas on herbivory, (b.) native C. floridanus but no early-season herbivory, (c.) exotic L. humile and earlyseason herbivory, and (d.) exotic L. humile but no early-seaso n herbivory.
56 AbsentPresentAbundance 0 20 40 60 80 100 AbsentPresentAbundance 0 200 400 600 800 1000 AbsentPresentAbundance 0 20 40 60 80 100 120 140 160 180 Ants A. coreopsidis Aphid Predators AbsentPresentAbundance 0 200 400 600 800 1000 1200 a. b. c. d. Figure 5.4. Comparison of mean 1 SE abunda nces of insects involved in the ant-aphid mutualism when ants were present or absent on host plants with (a.) native C. floridanus and early-season herbivory, (b.) native C. floridanus but no early-season herbivory, (c.) exotic L. humile and early-season herbivory, and (d.) exotic L. humile but no early-season herbivory.
57 AbsentPresentAbundance 0 50 100 150 200 250 AbsentPresentAbundance 0 50 100 150 200 250 AbsentPresentAbundance 0 10 20 30 40 50 60 AbsentPresentAbundance 0 5 10 15 20 25 30 A. maculosa L. trifolii N. lathami O. balanotes Ant-Damaged Mines a. b. c. d. Figure 5.5. Comparison of mean 1 SE a bundances of herbivores not involved in the ant-aphid mutualism when ants were present or absent on host plants with (a.) native C. floridanus and early-season herbivory, (b.) native C. floridanus but no early-season herbivory, (c.) exotic L. humile and early-season herbivory, and (d.) exotic L. humile but no early-season herbivory.
58 Discussion The presence of the exotic L. humile on B. halimifolia affected all of the insects, except the specialist gall fly N. lathami, on the host plant. By comparison with trees patrolled by the native ant C. floridanus, trees containing the exotic L. humile maintained greater densities of all observ ed insects. Aphids were 7 times more abundant in the presence of L. humile. The exotic L. humile makes an exceptional partner for honeydewproducing hemipterans, enabling them to have very high densities in the ants presence (Altfeld and Stiling 2006, Davidson et al. 2004, Ness and Bronstein 2004, Murdoch et al. 1995). The higher relative densities of aphid predators, especially coccinellids, were likely due to female ladybird beetles using honeydew and aphid presence as cues for oviposition to increase larval densities where aphids are abundant (Evans and Dixon 1986). Coccinellids feeding on aphids tended by Formica rufa also showed higher densities in the presence of ant attendants when untended aphid populations were rare (Sloggett and Majerus 2000). Hi gher relative densities of leaf miners, despite leaf mine predation, suggest compensatory mortality when L. humile is on the host plant, as may be produced if parasitism is reduced by th is exotic ant (Inouye and Agrawal 2004). Negative indirect effects of early-season herbivory by T. bacharidis larvae were observed for the leaf miners and gall fly on host trees with the native ant, C. floridanus, but, when L. humile was on the host plant, were obser ved only for the gall fly. The mechanism by which T. bacharidis early-season herbivory aff ects later occurring insects on B. halimifolia is indirect and thought to be chemical in nature because host trees fully re-flush with leaves after the larvae f eed on the plants (Hudson and Stiling 1997).
59 Bottom-up effects of host plant quality were more important to herbivores in the absence of early-season herbivory. Howeve r, for aphids the results were variable depending on whether they were tended by native or exotic ants. In addition, increases in foliar nitrogen trickled up to the third trophic level in the native insect community. Regardless of the species of ant on the host plant, the specialist gall fly, N. lathami, was positively affected by increases in foliar nitrogen and unaffected by the removal of ants, suggesting that this herbivore s success on its host is tightly linked to host plant quality and the fly larvae are susceptible only to specialized parasitoids not generalist predators such as ants. Top-down effects of ant predation we re unaltered by early-s eason herbivory but were strongly affected by the species of ant on the host plant. Nativ e ants had negative effects on aphid predator abundance and no effect s on other herbivores of the host plant. L. humile, however, dramatically increased aphid de nsities, aphid pred ator densities and preyed on leaf mines, reducing th e densities of the leaf miner L. trifolii. Leaf mine predation by ants is not uncommon. Liriomyza commelinae, a congener of L. trifolii, is preyed upon by Crematogaster ants with mortality rates of 21.2% (Freeman and Smith 1990). The delicate mines of Eriocraniella sp., which mine young oak leaves, suffer significant mort ality by arboreal ants; however, Cameraria sp., which mine mature oak leaves, are better prot ected within thicker mines and do not suffer mortality by ants (Faeth 1980). Leaf mining beetles on bamboo suffer predation by Pseudomyrmex gracilis in Costa Rica (Memmott et al. 1993), but leaf miners on bracken fern, a plant with extra-flor al nectaries, suffer no attacks by ants (Heads and Lawton 1985, Rashbrook et al. 1992). Coleophora serratella miners on birch are also unaffected
60 by ants(Fowler and MacGarvin 1985). Furthe rmore, larvae of the leaf mining beetle Odontota have higher survivorship in the pres ence of membracid-te nding ants because the ants deter an important predator of the beetle larvae (Fritz 1983). Community-level processes can be str ongly influenced by the presence of an exotic species, such as L. humile (Way and Khoo 1992). On B. halimifolia, nutrient limitation, predation and competition were all changed by the presence of this ant. Linepithema humile had direct and indirect effect s on most herbivores. Predicted negative effects of ear ly-season herbivory were less pronoun ced on host trees with exotic ants because the generalist herbivores we re more abundant on those host trees and bottom-up factors were more important in th e absence of early-season herbivory. Only the specialist N. lathami responded as predicted to early -season herbivor y regardless of the identity of ant on B. halimifolia. Top-down effects of ant predation were only important for leaf miners in the presence of exotic ants. Clearly, the effects of this exotic ant are persistent, significant, direct and indirect, and permeate all levels of the community. The long-term consequences of L. humile presence on B. halimifolia are unknown but may be to alter adap tive behaviors of herbivores that co-occur on this host plant.
61 CHAPTER 6 CLOSING REMARKS The presence of invasive species in communities can have profound consequences for native organisms in the communities where the exotic species invade. Otherwise adaptive strategies of native organisms can become disadvantageous when invasive species enter communities because the inva sive species alter community dynamics producing what Schlaepfer et al. (2002) called an evoluti onary trap. The reverse may also be true whereby the outcomes of intera ctions between native and invasive species benefit either or both organisms, a resu lt which is known as e volutionary release (Schlaepfer et al. 2005). Inte ractions among native species and Linepithema humile have been both positive and negative. Whereas native arthropod diversity is typically reduced, especially ant diversity, other insects, such as scale insects on citrus benefit from reduced parasitism in the presence of L. humile (Holway et al. 2002, Sa nders et al. 2003). On B. halimifolia, L. humile presence dramatically increases the mean annual abundance of A. coreopsidis, a facultative myrmecophile, and subsequently alters aphid predator foraging by providing large populations of prey. The invasive ant also preys on leaf miners, sometimes reducing their densi ties and sometimes not, and is capable of a trophic cascade enabling B. halimifolia to survive periods of intense da mage by the specialist stem borer
62O. balanotes. Temporally variable effects of ant predation are, however, common in many systems (Del-Claro and Oliviera 2000, Mooney and Tillberg 2005). Despite the abovementioned eff ects, the specialist herbivores, T. bacharidis and N. lathami, are unaffected by the presence of this exotic ant. Larvae of N. lathami are well protected within its galls and adults are, evidently, undeterred from ovipositing on host trees with ants. Larvae of T. bacharidis are likely unpalatable, secreting haemolymph when disturbed and adult beetle s are able to fly off host plants when disturbed by ants (Altfeld personal obse rvation). Host use by herbivores of B. halimifolia, then, may only change in response to L. humile presence when the herbivores are myrmecophilous or are vulnerable to preda tion. The evolutionary consequences of L. humile A. coreopsidis mutualism on B. halimifolia may be facilitation of each species by the other. Linepithema humile is widespread on Mullet Ke y and its spread may have been facilitated by honeydew collect ion from myrmecophiles such as A. coreopsidis. In other areas, such as Australia, where both B. halimifolia and L. humile are invasive, the ant may interfere with biologica l control of the plant because O. balanotes is used as a biological control agent. Both scenarios repr esent examples of evolutionary release. However, for generalist leaf miners, such as L. trifolii, B. halimifolia host use can be reduced because of predation by L. humile, an example of an evolutionary trap. The temporal variability in leaf mine predation observed for L. humile on A. maculosa and L. trifolii may have been, at least in pa rt, a result of density dependence because the densities of leaf miners were hi gher on host plants with L. humile than on host plants with C. floridanus. In order to evaluate the de nsity dependent nature of leaf mine predation by ants as a function of overa ll leaf mine abundance, I used the densities
63 of intact and damaged leaf mines from cont rol trees in all experiments from 2003 2005 at the field site that contained L. humile. I plotted that data and then tested the functional relationship against the following models: lin ear, S, quadratic and exponential. The quadratic model best fit the da ta (Figure 6.1). The nature of the quadratic relationship may help explain the temporal variation in leaf mine predat ion on leaf miner abundance. Leaf mine predation was greatest when mine rs were at a density between 50 100 mines per 200 leaves. Below and above this range leaf mine predation was decreased either because leaf mines were rarely encoun tered or overly abundant on host plants.
64 300.00 250.00 200.00 150.00 100.00 50.00 0.00 intact 120.00 100.00 80.00 60.00 40.00 20.00 0.00 damaged Quadratic Observed Figure 6.1. Abundance of ant-damaged leaf mines as a quadratic function of the abundance of intact leaf mines on the host plant, B. halimifolia.
65 Interspecific interactions among insects on B. halimifolia varied only for some insects, usually polyphagous, a nd depended on whether or not L. humile was present. Early-season herbivory by T. bacharidis was clearly an important factor for herbivores on the host plant. This side-to-side interaction affected the relative importance of host plant quality but host plant quality did not affect the relative importance of competition for later feeding herbivores. The effects of early-season herbivory were most pronounced when native C. floridanus were present but, even when L. humile was present, N. lathami predictably decreased in density due to T. bacharidis larval herbivory. Top-down effects of ant predation were unaffect ed by host plant quality but host plant quality was affected by the presence of predation by exotic ants. The leaf miner, L. trifolii, was particularly vulnerable to decreases in density when L. humile was present. The previous series of experiments involving B. halimifolia have shown that the presence of an ant-aphid mutualism has significant effects on community -level dynamics. However, the effects that the mutualism has depend on what sp ecies of ant is tending the aphids.
66 REFERENCES Addicott, J.F. 1979. A multispecies aphi d-ant association: density dependence and species-specific effects. Canadi an Journal of Zoology 57: 558-569. Altfeld, L. and P. Stiling. 2006. Argentine ants strongly affect some but not all common insects on Baccharis halimifolia. Environmental Entomology 35: 31-36. Anthonse, T., T. Bruun, E. Hemmer, D. Holme, A. Lamvik, E. Sunde and N.A. Sorensen. 1970. Baccharis oxide, a new triterpenoid from Baccharis halimifolia. Acta Chemica Scandinavica 24: 2479. Awmack, C.S. and S. R. Leather. 2002. Host plant quality and fecundity in herbivorous insects. Annual Review of Entomology 47: 817-844. Blthgen, N. and K. Fiedler. 2004. Comp etition for composition: lessons from nectarfeeding ant communities. Ecology 85: 1479-1485. Boldt, P.E. 1989. Biology and host specificity of Trirhabda bacharidis (Coleoptera: Chrysomelidae) on Baccharis (Asteraceae: Asteraceae). Environmental Entomology 18: 78-84 Breton, L.M. and J.F. Addicott. 1992. De nsity-dependent mutualism in an aphid-ant interaction. Ecology 73: 2175-2180. Buckley, R.C. 1987. Interactions involving plants, homoptera, and ants. Annual Review of Ecology and Systematics 18: 111-135.
67 Cushman, J.H. and J.F. Addicott. 1989. Intraand interspecific competition for mutualists: ants as a limited and limiting resource for aphids. Oecologia 79: 315321. Cushman, J.H. and J.F. Addicott. 1991. C onditional interactions in ant-plant-herbivore mutualisms. Pages 92-103 in C.R. Huxley and D.F. Cutler, editors. Ant-Plant Interactions. Oxford Univ. Press, Oxford, UK. Connell, J.H. 1983. On the prevalence and relative importance of interspecific competition: evidence from field experime nts. American Naturalist 122: 661696. Dansa, C.L.V.A. and C.F.D. Rocha. 1992. An ant-membracid-plant interaction in a cerrado area of Brazil. J Trop Ecol 8: 339-348. Davidson, D.W., S.C. Cook and R.R. Snelling. 2004. Liqui d-feeding performances of ants (Formicidae): ecologi cal and evolutionary impli cations. Oecologia 139: 255266. Del-Claro, K. and P.S. Oliveira. 2000. Conditional outcomes in a neotropical treehopper-ant association: temporal and species-specific variation in ant protection and homopteran fec undity. Oecologia 124: 156-165. Denno, R.F., M.S. McClure and J.R. Ott. 1995. Interspecific interactions in phytophagous insects: competition revisite d and resurrected. Annual Review of Entomology 40: 297-331/ Denno, R.F., M.A. Peterson, C. Gratton, J. Cheng, G.L. Langellotto, A.F. Huberty and D.L. Finke. 2000. Feeding-induced change s in plant quality me diate interspecific competition between sap-feeding he rbivores. Ecology 81: 1814-1827.
68 Dyer, L.A. and D.K. Letourneau. 1999. Relative strengths of top-down and bottom-up forces in a tropical forest co mmunity. Oecologi a 119: 265-274. Engel, V., M.K. Fischer, F.L. Wckers a nd W. Vlkl. 2001. Interactions between extrafloral nectaries, aphids and ants: are there competition effects between plant and homopteran sugar sources : Oecologia 29: 577-584. Evans, E.W. and A.F.G. Dixon. 1986. Cues for oviposition by ladybird beetles (Coccinellidae): response to aphids. Journal of Animal Ecology 55: 1027-1034. Faeth, S.H. 1980. Invertebrate predation of leaf-miners at low densities. Ecological Entomology 5: 111-114. Fagundes, M., F.S. Neves and G.W. Fernandes. 2005. Direct and indirect interactions involving ants, insect herbivores parasitoids, and the host plant Baccharis dracunculifolia (Asteraceae). Ecological Entomology 30: 28-35. Fischer, M.K., K.H. Hoffman a nd W. Vlkl. 2001. Competition for mutualists in an anthomopteran interaction me diated by hierarchies of an t attendance. Oikos 92: 531-541. Flatt, T. And W.W. Weisser. 2000. The eff ects of mutualistic ants on aphid life history traits. Ecology 81: 3522-3529. Floate, K.D. and T.G. Whitham. 1994. A phid-ant interaction reduces chrysomelid herbivory in a cottonwood hybrid zone. Oecologia 97: 215-221. Forkner, R.E. and M.D. Hunter. 2000. What goes up must come down? Nutrient addition and predation pressure on oa k herbivores. Ecology 81: 1588-1600.
69 Fowler, S.V. and M. MacGarvin. 1985. The impact of hairy wood ants, Formica lugubris, on the guild structure of he rbivorous insects on birch, Betula pubescens. Journal of Animal Ecology 54: 847-855. Freeman, B.E. and D.C. Smith. 1990. Variatio n of density-dependen ce with spatial scale in the leaf-mining fly Liriomyza commelinae (Diptera: Agromyzidae). Ecological Entomology 15: 265-274. Freitas, A.V.L. and P.S. Oliveira. 1996. An ts as selective agents on herbivore biology: effects on the behaviour of a non-myrmeco philous butterfly. Journal of Animal Ecology 65: 205-210. Fritz, R.S. 1983. Ant protection of a host plants defoliator: c onsequence of an antmembracid mutualism. Ecology 64: 789-797. Hairston, N.G., F.E. Smith and L.B. Slobodki n. 1960. Community structure, population control, and competition. American Naturalist 94: 421-425. Heads, P.A. and J.H Lawton. 1985. Bracken, ants sand extrafloral nectaries. III. How insect herbivores avoid ant predatio n. Ecological Entomology 10: 29-42. Hlldobler, B. and E.O. Wilson. 1990. The Ants. Harvard University Press. Cambridge, MA. Holway, D.A. 1998. Effect of Argentine ant invasions on ground-dwelling arthropods in northern California riparian w oodlands. Oecologia 116: 252-258. Holway, D.A., L. Lach, A.V. Suarez, N.D. Tsutsui and T.J. Case. 2002. The causes and consequences of ant invasions. Annual Review of Ecology and Systematics 33: 181-233. Hudson, E.E. 1995. Exploitative competiti on in the phytophagous insect community
70 associated with Baccharis halimifolia. M.S. Thesis. University of South Florida. Tampa, FL. Hudson, E. E. and P. Stiling. 1997. Exploitative competition strongly affects the herbivorous insect community on Baccharis halimifolia. Oikos 79: 521-528. Hunter, M.D. and P.W. Price. 1992. Playi ng chutes and ladders: heterogeneity and he relative roles of bottom-up and top-down forces in natural communities. Ecology 73: 724-732. Hutchinson, G.E. 1959. Homa ge to Santa-Rosalia or why are there so many kinds of animals. American Naturalist 93: 145-159. Inbar, M., A. Eshel and D. Wool. 1995. Interspecific competition among phloemfeeding insects mediated by induced host-plant sinks. Ecology 76: 1506-1515. Inouye, B.D. and A.A. Agrawal. 2004. Ant mutualists alter the composition and attack rate of the parasitoid community for th e gall wasp Disholcaspis eldoradensis (Cynipidae). Ecological En tomology 29: 692-696. Janzen, D.H. 1973. Host plants as isla nds. 2. Competition in evolutionary and contemporary time. American Naturalist 107: 786-790. Kaplan, I. and M.D. Eubanks. 2005. Aphids alter the community-w ide impact of fire ants. Ecology 86: 1640-1649. Kraft, S.K. and R.F. Denno. 1982. Feeding responses of adapted and non-adapted insects to the defensive properties of Baccharis halimifolia L. (Compositae). Oecologia 52: 156-163. Krischik, V.A. and R.F. Denno. 1990. Patte rns of growth, reproduc tion, defense, and
71 herbivory in the dioecious shrub, Baccharis halimifolia (Compositae). Oecologia 83: 176-181. Kunert, G., S. Otto, U.S.R. Rose, J. Gers henzon and W.W. Weisse r. 2005. Alarm pheromone mediates production of winged dispersal morphs in aphids. Ecology Letters 8: 596-603. Kyt, M., P. Niemela and S. Larsson. 1996. Insects on trees: popul ation and individual response to fertilization. Oikos 75: 148-159. Mahdi, T. and J.B. Whittaker. 1993. Do bi rch trees (Betula pendula) grow better if foraged by wood ants? Journal of Animal Ecology 62: 101-116. Mattson, W.J. 1980. Herbivory in relation to plant nitrogen-content. Annual Review of Ecology and Systematics 11: 119-161. Memmott, J., H.C.J. Godfray and B. Bolton. 1993. Predation and parasitism in a tropical herbivore community. Ecological Entomology 18: 348-352. Menge B.A. and J.P. Sutherland. 1976. Sp ecies diversity gradient s: synthesis of the roles of predation, competition and tem poral heterogeneity. American Naturalist 110: 351-369. Messina, F.J. 1981. Plant protection as a c onsequence of an ant-membracid mutualism: interactions on goldenrod (Solidago sp.). Ecology 62: 1433-1440. Mody, K. and K.E. Linsenmair. 2004. Plantattracted ants affect arthropod community structure but not necessarily herbivor y. Ecological Entomology 29: 217-225. Moe, S.J., R.S. Stelzer, M.R. Forman, W.S. Harpole, T. Daufresne and T. Yoshida. 2005. Recent advances in ecological st oichiometry: insights for population and community ecology. Oikos 109: 29-39.
72 Moon, D. C. and P. Stiling. 2002. Top-dow n, bottom-up, or side to side? Whithintrophic-level interactions modify trophi c dynamics of a salt marsh herbivore. Oikos 98: 480-490. Moon, D.C. and P. Stiling. 2004. The infl uence of a salinity and nutrient gradient on coastal vs. upland tritrophic complexes. Ecology 85: 2709-2716. Mooney, K.A. and C.V. Tillberg. 2005. Te mporal and spatial variation to ant omnivory in pine forests. Ecology 86: 1225-1235. Murdoch, W.W., R.F. Luck, S.L. Swarbrick, S. Walde, D.S. Yu and J.D. Reeve. 1995. Regulation of an insect population under biological control. Ecology 76: 206217. Ness, J.H. and J.L. Bronstein. 2004. The effects of exotic an ts on prospective ant mutualists. Biologica l Invasions 6: 445-461. Palmer, W.A. 1987. The phytophagous insect fauna associated with Baccharis halimifolia L. and Baccharis neglecta Britton in Texas, Louisiana, and Northern Mexico. Proceedings of the Entomologi cal Society of Washington 89: 185-199. Pezzolesi, L.S.W. and B.J. Hager. 1994. Ant predation on two sp ecies of birch leafmining sawflies. American Mi dland Naturalist 131: 156-168. Polis, G.A. and D.R. Strong. 1996. Food web complexity and community dynamics. American Naturalist 147: 813-846. Rashbrook, V.K., S.G. Compton, and J.H. La wton. 1992. Ant-herbivore interactions: reasons for the absence of benefits to a fe rn with foliar nectaries. Ecology 73: 2167-2174.
73 Sakata, H. and Y. Hashimoto. 2000. Should a phids attract or repel ants? Effect of rival aphids and extrafloral nectaries on an t-aphid interactions. Population Ecology 42: 171-178. Sanders, N.J., N.J. Gotelli, N.E. Heller and D.M. Gordon. 2003. Community disassembly by an exotic species. Pr oceedings of the National Academy of Science 100: 2474-2477. Schlaepfer, M.A., M.C. Runge and P.W. Sher man. 2002. Ecological and evolutionary traps. Trends in Ecology and Evolution 17: 474-480. Schlaepfer, M.A., P.W. Sherman, B. Blossey and M.C. Runge. 2005. Introduced species as evolutionary traps. Ecology Letters 8: 241-246. Schoener, T.W. 1983. Field experiments on interspecific competition. American Naturalist 122: 240-285. Shingleton, A.W. and W.A. Fo ster. 2000. Ant tending influe nces soldier production in a social aphid. Proceedings of the Royal Society of London 267: 1863-1868. Sih, A., P. Crowley and M. McPeek. 1985. Predation, competition, and prey communities a review of field experi ments. Annual Review of Ecology and Systematics 16: 269-311. Skinner, G.J.and J.B. Whittaker. 1981. An experimental investigation of interrelationships be tween the wood ant (Formica rufa) and some tree-canopy herbivores. Journal of Animal Ecology 50: 313-326. Sloggett, J.J. and M.E.N. Majerus. 2000. Aphid-mediated coexistence of ladybirds (Coleoptera: Coccine llidae) and the wood ant Formica rufa: seasonal effects,
74 interspecific variability and the evolution of a coccinellid myrmecophile. Oikos 89: 345-359. SPSS Inc. 2004. Users manual, versions 12.0 and 13.0 for Windows. SPSS Inc. Chicago, IL. Stadler, B., A.F.G. Dixon and P. Kindlmann. 20 02. Relative fitness of aphids: effects of plant quality and ants. Ecology Letters 5: 216-222. Stadler B., P. Kindlmann, P. Smilauer, and K. Fiedler. 2003. A compar ative analysis of morphological and ecological characters of European aphids and lycaenids in relation to ant attendance. Oecologia 135: 422-430. Stadler, B. 2004. Wedged between bottom-up and top-down processes: aphids on tansy. Ecological Entomology 29: 106-116. Sterner, R.W. and J.J. Elser. 2002. Ecologi cal stoichiometry: the biology of elements from molecules to the biosphere. Princet on University Press. Princeton, N.J. Strauss, S.Y. 1987. Direct and indirect e ffects of host-plant fer tilization on an insect community. Ecology 68: 1670-1678. Strengbom, J., J. Witzell, A. Nordin and L. Ericson. 2005. Do multitrophic interactions override N fertilization effects on Operophtera larvae? Oecol ogia 143: 241-250. Suzuki, N., K. Ogura and N. Katayama. 2004. Efficiency of herbivore exclusion by ants attracted to aphids on the vetch Vicia angustifolia L. (Leguminosae). Ecological Research 19: 275-282. Vestergrd, M., L. Bjrnlund a nd S. Christensen. 2004. A phid effects on rhizosphere microorganisms and microfauna depend more on barley growth phase than on soil fertilization. O ecologia 141: 84-93.
75 Waring, G.L. and N.S. Cobb. 1989. The im pact of plant stress on herbivore population dynamics. Insect-Plant Interactions. Vol. 4. Elizabeth Bernays, ed. CRC Press. Boca Raton, FL. Pp. 167-187. Way, M.J. 1963. Mutualism between ants and honeydew-producing homoptera. Annual Review of Entomology 8: 307-344. Way, M.J., M.E. Cammell, and M.R. Paiva. 1992. Studies on egg predation by ants (Hymenoptera: Formicidae) especi ally on the eucalyptus borer Phoracantha semipunctata (Coleoptera: Cerambycidae) in Po rtugal. Bulletin of Entomological Research 82: 425-432. Way, M.J. and K.C. Khoo. 1992. Role of ants in Pest Management. Annual Review of Entomology 37: 479-503. Way, M.J., M.R. Paiva and M.E. Cammell. 1999. Natural biological control of the pine processionary moth Thaumetopoea pityocampa (Den. & Schiff.) by the Argentine ant Linepithema humile (Mayr) in Portugal. Agricultural and Forest Entomology 1: 27-31. Weins, J.A. 1977. On competition and variable environments. American Scientist 65: 590-597. Woodring, J., R. Wiedemann, M.K. Fische r, K.H Hoffman and W. Vlkl. 2004. Honeydew amino acids in relation to sugars and their role in the establishment of ant-attendance hierarchy in eight sp ecies of aphids feeding on tansy (Tanacetum vulgare). Physiological Entomology 29: 311-319. Yao, I., H. Shibao and S. Akimoto. 2000. Costs and benefits of ant attendance to the drepanosiphid aphid Tuberculatus quercicola. Oikos 89: 3-10.
76 Zar, J.H. 1999. Biostatistical Analysis. F ourth Ed. Prentice Hall. Upper Saddle River, NJ.
77 ABOUT THE AUTHOR Laura Francine Altfeld, ne Hensey, wa s born to Rose Marie and Daniel Hensey, April 20th, 1970. Raised in Florid a, Laura attended Eckerd College, where she completed a senior thesis on ichthyofa una in created versus natu ral wetlands and earned a Bachelors of Science degree in Marine Scie nce. In keeping with her interests in wetlands, Laura pursued a Masters of Science degree in Biology with Peter Stiling, PhD, at the University of South Florida, wher e she began studying inse ct herbivory on salt marsh plants. In 2003, Laura began work on he r PhD, also under the supervision of Peter Stiling. While at USF, Laura was a gradua te and teaching assistant for the Biology and Biology Honors Labs. Laura is expected to complete her do ctorate in 2006 and continue to pursue research in the area of plant-animal interactions.
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001795700
007 cr mnu|||uuuuu
008 070711s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001578
Altfeld, Laura F.
Interspecific interactions among common insects of the salt myrtle, Baccharis halimifolia L. (Asteraceae)
h [electronic resource]
by Laura F. Altfeld.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: Baccharis halimifolia is host to many species of insect herbivore, including myrmecophilous aphids. Previous studies with B. halimifolia have revealed strong effects of competition by an early-feeding beetle, Trirhabda bacharidis, and nitrogen enrichment. The effects of ant mutualists, however, have not been evaluated for their potential influence on interspecific interactions among insects on the host plant. I have employed a series of experiments aimed at answering the following general questions. (1) How does the density of aphid-tending ants affect common insects on the host plant? (2) What are the relative effects of (a) competition from T. bacharidis, (b) nutrient enrichment, and (c) ant presence on common insects of the host plant? (3) How do the effects of exotic ants differ from those of native ants? The density of aphid-tending ants had positive effects on myrmecophilous aphids and aphid predators. However, given a choice between tended and untended aphid^ populations, aphid predators preferred to forage in the absence of ants. The density of aphid-tending ants also increased predation on leaf miners although this did not necessarily translate into reduced densities of leaf miners. Competition by early-season feeding of T. bacharidis negatively affected later-feeding herbivores but the effects of competition were unaffected by nutrient enrichment. Nutrient enrichment had positive effects on some herbivores, often only in the absence of early season herbivory. Trirhabda bacharidis larvae showed evidence of nitrogen and phosphorus limitation and suffered no predation by aphid-tending ants. Ant presence increased host plant survivorship from stemborer damage in 2004. Ant species identity was an important factor determining the densities of myrmecophiles and non-myrmecophiles on the host plant in addition to affecting the responses of herbivores to increases in host plant quality. Aphids were more abundant in the presence of the exotic^ Linepithema humile (Hymenoptera: Dolichoderinae) versus the native Camponotus floridanus (Hymenoptera: Formicinae). Aphid predators also had higher densities in the presence of L. humile versus C. floridanus. Only L. humile acted as predator on leaf mines although predation did not always translate into reduced seasonal abundances for both species of leaf miner.
Dissertation (Ph.D.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 77 pages.
Adviser: Peter Stiling, Ph.D.
Host plant quality.
t USF Electronic Theses and Dissertations.