The Invasive Cactus Moth, Cactoblastis Cactorum

University of South Florida Scholar Commons

Graduate Theses and Dissertations Graduate School

January 2011 The invasive cactus moth, Cactoblastis cactorum: Host plant testing, species interactions, and effects on local Opuntia populations Heather Jezorek University of South Florida, [email protected]

Follow this and additional works at: http://scholarcommons.usf.edu/etd Part of the American Studies Commons, and the Ecology and Evolutionary Biology Commons

Scholar Commons Citation Jezorek, Heather, "The invasive cactus moth, Cactoblastis cactorum: Host plant testing, species interactions, and effects on local Opuntia populations" (2011). Graduate Theses and Dissertations. http://scholarcommons.usf.edu/etd/3731

This Dissertation is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact [email protected].

The invasive cactus moth, Cactoblastis cactorum: Host plant testing, species interactions,

and effects on local Opuntia populations

Heather Jezorek

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Integrative Biology College of Arts and Sciences University of South Florida

Major Professor: Peter Stiling, Ph.D. Gordon Fox, Ph.D. Earl McCoy, Ph.D. Jason Rohr, Ph.D.

Date of Approval: November 18th, 2011

Keywords: cacti, oviposition, larval performance, associational resistance, predation

ACKNOWLEDGEMENTS

I have been fortunate to have a wonderful support system while at USF. I first and foremost thank my major professor, Peter Stiling, for his exceptional guidance and encouragement, not only on my research, but on all aspects of graduate school. I thank my committee members, Dr. Gordon Fox, Dr, Earl McCoy, and Dr. Jason Rohr for their insightful comments on my dissertation, and for freely sharing their time and expertise during the past six years. I would also like to thank Henry Mushinsky and Susan Bell for help in acquiring funding and for creating an outstanding environment for higher education. Many people have helped in the production of this work by assisting with field work, offering advice, and sharing their knowledge. In particular, I would like to thank the following past and present members of the Stiling lab: Laura Altfeld, Lance

Arvidson, Kerry Bohl, Tatiana Cornelissen, Christy Foust, Jennifer Maul, and Keith

Stokes. I thank the Department of Integrative Biology, the Fred L. and Helen M. Tharp

Endowed Scholarship Fund, the Biology Graduate Student Organization and USF

Student Government for providing funding for summer research and travel. I thank the

Florida Department of Environmental Protection Southwest District, in particular Terry

Hingten, and Pinellas County Parks, for allowing me use of their parks for field work. I thank Jim Carpenter, Stephen Hight, and Susan Drawdy for providing live Cactoblastis cactorum specimens and the opportunity to collaborate. Additional thanks to Jim for agreeing to serve as an external chair for my doctoral defense. I thank my parents, John

and Susan Jezorek, for their moral and financial support, as well as their superior grandparenting skills, which permitted me to finish writing this work. They have been generous, encouraging, and patient throughout all of my life decisions, both major and minor. I thank my daughter Chessa for timing her arrival impeccably, and for bringing a new level of love and delight to my life during the final phase of writing. Lastly, I thank my husband PJ for being a truly amazing person and partner. He has been there for me each and every day, through tears and joy. I cannot fully express my gratitude, but am certain that this work would not have been completed without him by my side.

TABLE OF CONTENTS

LIST OF TABLES ...... iv

LIST OF FIGURES ...... v

ABSTRACT ...... viii

CHAPTER 1- INTRODUCTION ...... 1 Invasive Species ...... 1 Cactaceae: Opuntioideae ...... 9 Cactoblastis cactorum ...... 10 Biology ...... 10 History: Australia, South Africa, and the Caribbean ...... 12 History: U.S. and Mexico ...... 15 References cited ...... 17

CHAPTER 2- OVIPOSITION PREFERENCE AND LARVAL PERFORMANCE OF CACTOBLASTIS CACTORUM ON NORTH AMERICAN OPUNTIOID CACTI ...... 26 Synopsis ...... 26 Introduction ...... 27 Materials and methods ...... 32 Oviposition experiment ...... 32 Larval performance experiment ...... 34 Results ...... 36 Oviposition experiment ...... 36 Larval performance experiment ...... 37 Discussion ...... 46 Acknowledgements ...... 50 References cited ...... 50

CHAPTER 3- ASSOCIATIONAL EFFECTS OF CHAMAECRISTA FASCICULATA ON ANT PREDATION OF CACTOBLASTIS CACTORUM AND HERBIVORY OF OPUNTIA...... 56 Synopsis ...... 56 Introduction ...... 57 Study system and methods ...... 61 Insects ...... 61 Plants ...... 62 i

Damage and growth surveys ...... 63 Predation experiment ...... 66 Results ...... 68 Damage and growth surveys ...... 68 Predation experiment ...... 69 Discussion ...... 77 Acknowledgments...... 82 References cited ...... 82

CHAPTER 4- FINDINGS FROM A SIX YEAR STUDY: EFFECTS OF CACTOBLASTIS CACTORUM ON THE SURVIVAL AND GROWTH OF LOCAL OPUNTIA ...... 90 Synopsis ...... 90 Introduction ...... 91 Methods...... 94 Study system ...... 94 Data collection and analysis...... 95 Q1: Are plants attacked by C. cactorum more likely to die than unattacked plants? ...... 97 Q2: How does frequency of attack by C. cactorum affect survival of plants? ...... 98 Q3: For surviving plants, does C. cactorum attack affect growth rate? ...... 100 Results ...... 101 Q1: Are plants attacked by C. cactorum more likely to die than unattacked plants? ...... 101 Q2: How does frequency of attack by C. cactorum affect survival of plants? ...... 103 Q3: For surviving plants, does C. cactorum attack affect growth rate? ...... 104 Discussion ...... 115 Acknowledgments...... 120 References cited ...... 120

CHAPTER 5- LACK OF ASSOCIATIONAL EFFECTS BETWEEN TWO COMMON HOSTS ON ATTACK RATES BY CACTOBLASTIS CACTORUM ...... 125 Synopsis ...... 125 Introduction ...... 126 Methods...... 129 Study system ...... 129 Data collection and analysis...... 131 Results ...... 134 Discussion ...... 143 Acknowledgments...... 146 References cited ...... 146 ii

APPENDIX I: TAXONOMIC INFORMATION FOR THE OPUNTIOIDEAE TRIBES CYLINDROPUNTIEAE AND OPUNTIEAE ...... 154

iii

LIST OF TABLES

Table 2.1: Eggstick and spine data for the fourteen opuntioid taxa...... 40

Table 2.2: Mean values (± SE) for recorded plant characteristics ...... 42

Table 2.3: Results from the no choice larval performance experiment ...... 43

Table 2.4: Adult weights and number of eggs per female (mean ± SE) ...... 44

Table 2.5: Results from regressions of larval performance variables on epidermal toughness of the fourteen tested opuntioids ...... 45

Table 3.1: Native Opuntia herbivores found in central Florida and their characteristic feeding damage ...... 71

Table 3.2: Relative abundance of ant species in a sample (n=99) collected from Opuntia individuals ...... 74

Table 4.1: Summary of attack and survival rates for each survey year ...... 105

Table 4.2: Results from likelihood ratio tests ...... 108

Table 4.3: Results from log linear models ...... 109

Table 4.4: Results from logistic regression models with final survival status as the dependent variable ...... 110

Table 4.5: Results from multiple regressions of ln(proportional cladode change/generation of C. cactorum) on attack frequency and ln(initial cladode number) ...... 112

Table 5.1: The four native Opuntia herbivores examined in the study ...... 137

Table 5.2: Results from a Mann-Whitney test for differences between O. humifusa (N=94) and O. stricta (N=96) ...... 138

LIST OF FIGURES

Figure 3.1: Results from C. cactorum damage surveys ...... 72

Figure 3.2: Results from native herbivore and ant surveys ...... 73

Figure 3.3: Results from a growth survey from April-2009 to Oct-2009 of O. humifusa plants (n=40) near to and far from C. fasciculata ...... 75

Figure 3.4: Results from the predation experiment ...... 76

Figure 4.1: Histogram of C. cactorum attack and survival rates ...... 106

Figure 4.2: The proportion of plants that died each survey year grouped by C. cactorum attack status for that same year ...... 107

Figure 4.3: Mean proportional change in initial cladode number per generation of C. cactorum for all surviving plants ...... 111

Figure 4.4: Partial regression plots for O. humifusa ...... 113

Figure 4.5: Partial regression plots for O. stricta...... 114

Figure 5.1: Total cladodes for the four plant species/neighborhood groups surveyed at HM in 2006 ...... 139

Figure 5.2: Proportion of cladodes with damage from the weevil G. hubbardi for the four plant species/neighborhood groups surveyed at HM in 2006 ...... 140

Figure 5.3: Proportion of cladodes with C. cactorum larvae from the three O. humifusa neighborhoods surveyed at HM in 2009 ...... 141

Figure 5.4: Proportion of cladodes with old C. cactorum damage from three O. humifusa neighborhoods surveyed at HM in 2009 ...... 142

ABSTRACT

The invasive cactus moth, Cactoblastis cactorum, poses a threat to opunitoid cacti species of North America. The following work contains four separate studies investigating C. cactorum host plant preference and performance, predation and parastitism of C. cactorum, effects of C. cactorum on local Opuntia populations, and associational effects of host and non-host plants on C. cactorum and native Opuntia- feeding herbivores. We found that, among southwestern and Mexican opuntioid taxa, moths preferred O. engelmannii var. linguiformis and var. engelmannii for oviposition, while Consolea rubescens and O. streptacantha were superior larval hosts. Oviposition was best predicted by number of cladodes and degree of spininess; epidermal toughness was a significant predictor of most larval fitness parameters. In general, oviposition preference was not correlated with larval performance. A lack of co-evolutionary history between C. cactorum and North American opuntioid species may help explain this disconnect. We placed irradiated C. cactorum eggsticks and pupae on Opuntia plants in the field to test for predation. We found evidence of predation, most likely from ants, on

~16% of eggsticks and ~18% of pupae. Predation rates, ant abundance, and cladode growth were higher, and C. cactorum damage lower, on Opuntia located near the extrafloral nectar-producing legume Chamaecrista fasciculata. We attribute these associational effects to the ability of C. fasciculata to attract ants to its extrafloral nectar.

Over the course of six years, ~78% of nearly 600 tagged Opuntia were attacked by C.

vi cactorum at least once and ~76% of the plants survived. Two separate studies found that

Opuntia stricta was more likely to be attacked by C. cactorum than O. humifusa; we also found that O. stricta was more likely to die following an attack. A plant‟s odds of survival decreased as C. cactorum attack frequency increased, but plants that did survive had positive growth rates, regardless of attack status. We did not find significant evidence of associational effects for O. humifusa and O. stricta, either for C. cactorum or native

Opuntia-feeding herbivores. It could be that present herbivore densities are low enough, and host plants plentiful enough, to avoid mechanisms that usually lead to associational resistance or susceptibility. Overall, our results suggest that the presence and spread of C. cactorum should be taken seriously, especially for rare opuntioids and the Opuntia-rich deserts of North America. However, for more common opuntioid host species, there may be enough resistant or tolerant individuals, and sufficient top down control through ant predation, for populations to persist at current C. cactorum densities. We acknowledge that information on Opuntia reproduction and recruitment rates are needed to confirm this suggestion, and see this as an excellent opportunity for future research.

vii

CHAPTER 1

INTRODUCTION

Invasive species

Inarguably, the spread and establishment of non-indigenous species has had an enormous impact worldwide. Much of this impact has been positive; introduced crops and livestock make up 98% of the US food industry and bring in approximately $800 billion annually

(Pimentel et al. 2005). However, non-indigenous species can also have far-reaching negative impacts. It has been suggested that the introduction and spread of such species ranks second only to habitat loss as a threat to imperiled species in the United States

(Wilcove et al. 1998). Of particular concern is the small percentage of non-indigenous species that rise to the status of „invasive‟. According to Executive Order 13112 (1999), a non-indigenous species is considered invasive when its “introduction does or is likely to cause economic or environmental harm or harm to human health.

It is nearly impossible to accurately quantify the number of non-indigenous species that quickly perish after arriving in a new habitat, the number that survive for a brief period before going extinct, and the number that successfully establish (Mack et al.

2000). Based upon past studies of the British Isles, Williamson and Fitter (1996a) suggest the “tens rule”: 1 out of 10 imported species (accidentally or deliberately) will appear in the wild, or naturalize; 1 out of those 10 will become established; 1 out of those

10 will become a pest (have a negative economic affect). Of course, this rule is an

1 approximation and is subject to taxonomic variation and exceptions. For example, biological control insects are deliberately released into the wild and tend to follow more of a “threes rule” for the proportion that establish and the proportion that lead to control of the target species (analogous to reaching pest status, as both have an economic effect)

(Williamson and Fitter 1996a). In addition, the rule does not consider those species that may have negative ecological effects without inflicting any obvious economic effects.

When trying to predict the number of biological invasions that will occur out of a given number of non-indigenous arrivals, it is only natural to attempt to determine if there are certain biological attributes shared by successful invaders. Similarly, there may be biotic and/or abiotic factors of communities that make them more prone to invasion.

This topic has been the subject of many studies (Ehrlich 1989, Noble 1989, Simberloff

1989, Williamson and Fitter 1996a, Williamson and Fitter 1996b, Williamson 1999,

Kolar and Lodge 2001, Colautti et al. 2006) and several theories have been developed to explain the success of invaders (Tsutsui et al. 2000, Keane and Crawley 2002, Colautti et al. 2004) and the invasibility of a community (Lonsdale 1999, Davis et al. 2000,

Kennedy et al. 2002). In plants, short juvenile period, short time between large seed crops and small seed mass were found to be significantly, positively correlated with invasion success (Rejmanek and Richardson 1996). Ehrlich (1989) suggests that for vertebrates, most of the characteristics of successful invaders can be grouped under

“broad ecological amplitude”; this includes qualities like large native range, generalist diet, and ability to tolerate a large range of physical conditions. Today, most scientists agree that multiple factors act simultaneously on the probability of an introduced species establishing and becoming invasive. The complexity of the interactions of these factors,

2 coupled with the bias of differential opportunities to receive non-indigenous species, may make it impossible for us to ever form a predictive theory of invasion (Williamson 1999,

Mack et al. 2000).

Population-, community-, and ecosystem-level effects have been attributed to the arrival of invasive species. Although there is much more information available for population- and community-level effects, in the last two decades there has been an increase in the number of studies investigating the alteration of ecosystem processes by exotics (Vitousek et al. 1987, Vitousek 1990, Heath et al. 1995, Ehrenfeld et al. 2001,

Schwindt et al. 2004). Effects at any level can be direct, such as the introduction of a virulent pathogen, or indirect, such as reduced pollination of a plant species following the introduction of an organism that preys upon one of the plant‟s important pollinators.

Some of the most devastating consequences of invasive species are brought about by the diseases they carry. A well-known example of this type of direct effect is the elimination of the American chestnut, Castanea dentata, in the first half of the century by the Asian chestnut blight, Cryphonectria parasitica, a fungus thought to have arrived via imported

Japanese chestnuts (Anagnostakis 2001, Davelos and Jarosz 2004). Introduction of avian malaria and avian poxvirus by the mosquito species Culex pipiens quinquefasciatus has caused substantial declines in Hawaii‟s native avian populations and has contributed to the extinctions of at least twenty species of honeycreepers and the Hawaiian crow

(extinct in the wild) (Freed et al. 2005). A similar situation could develop on the

Galapagos Islands; Gottdenker et al. (2005) found 13 pathogens in domestic poultry stocks that are known to adversely affect wild birds. Prevalence of bird pox in Darwin‟s ground finches rose 33% between 2000 and 2004, and male finches that carried the virus

3 were less likely to find a mate (Kleindorfer and Dudaniec 2006). The rise in disease could be due to recent increases in the number of domestic chickens on inhabited islands

(Gottdenker et al. 2005), to increasing populations of introduced arthropod vectors, such as C. pipiens quinquefasciatus (Kleindorfer and Dudaniec 2006), or to a combination of factors. Declines in some of Galapagos‟ endemic avian species have already been observed (Wikelski et al. 2004).

Invasive species can also cause the loss of native species through predation or herbivory. The brown tree snake, Boiga irregularis, was introduced to Guam in 1949 and has since been responsible for the extirpation of the majority of the island‟s native bird and bat species, as well as several species of lizards. The loss of these vertebrates will inevitably lead to indirect ecological effects due to decreased seed dispersal (Fritts and

Rodda 1998, Rodda and Savidge 2007). Boiga irregularis has also caused severe economic losses and social disturbances in the form of electrical outages, death of pets and poultry, infant bites, and a decrease in tourism (Rodda and Savidge 2007).

Introduced red deer locally eliminated the understory species Aristotelia chilensis in

Chilean rainforests and drastically reduced its numbers in Patagonian mesic forests

(Veblen et al. 1989). Areas of Nahuel Huapi National Park in Argentina that have dense red deer and cattle populations have significantly less seedlings and saplings of native canopy trees, suggesting that browsing and grazing are hindering recruitment (Veblen et al. 1992). Hybridization of native and introduced congeners, post-introduction evolution, and competition (to be discussed in more detail later) are still other ways that invasive species can have detrimental outcomes on native flora and fauna (Mack et al. 2000,

Manchester and Bullock 2000).

Most of the above examples deal with population or community level effects.

However, in some instances, invasive species can cause ecosystem-level effects, leading to even more severe economic and environmental consequences. Alteration by invasive species of processes like nutrient and energy cycling, water flow, and fire regime have been found to occur (Vitousek 1990, Mack et al. 2000, Manchester and Bullock 2000,

Crooks 2002). In Hawaii, the nitrogen fixer Myrica faya has been shown to alter primary succession after volcanic ash fall by adding significant amounts of available fixed nitrogen to the soil (Vitousek et al 1987, Vitousek 1990). The invasive paper bark tree,

Melaleuca quinquenervia, has totally changed the structure of some of Florida‟s native sawgrass prairies and cypress swamps. The large scale transformations are thought to be due to the tree‟s tendency to form dense stands that exclude native plants and provide poor habitat for native animals, its large water intake, and its ability to increase fire intensity and frequency while being extremely fire-tolerant itself (Myers 1983, Gordon

1998, Mack et al. 2000, Crooks 2002). Species considered to be “ecosystem engineers” can have ecosystem-level effects by altering the physical structure of a habitat, which can, in turn, alter the processes mentioned above by directly or indirectly changing the availability of resources to other species (Jones et al. 1994). The reef-building polychaete

Ficopomatus enigmaticus is an ecosystem engineer that is considered invasive or exotic in most of the world‟s temperate, brackish waters. It has clogged intake pipes in power plants and has also been found to alter sediment deposition and transport and relative water flow in an Argentinean coastal lagoon (Schwindt et al. 2004).

Many pathways exist for the arrival of new species. Some of these pathways can be considered natural; air, ocean, and river currents can carry seeds, spores and plant

5 propagules to a new location (Mack 2003); insects can disperse aerially over long distances (Kiritani and Yamamura 2003). However, in modern times, natural introductions have a negligible effect compared to human-assisted introductions, both purposeful and accidental. Some of the most common accidental pathways include the arrival of non-indigenous organisms via ballast water or soil, encrusted ship hulls, and cargo. In 1986, Dreissena polymorpha, the zebra mussel, was unintentionally released into the Great Lakes from ballast water from the Caspian Sea. It has spread throughout the eastern US and into parts of Canada and now threatens many native invertebrates, as well as some deep water fish (Vanderploeg et al. 2002, Stiling 2004). “Stowaway” species are commonly found in shipments of the agriculture and nursery industries, but packing materials (straw, wood crates) of any industry can provide a pathway for the introduction of a new plant, animal, or pathogen.

Intentional introductions occur most often through the pet and nursery industries and often result in unintentional escapes of non-indigenous species (Frank and McCoy

1995, Kraus 2003, Mack 2003). The deliberate release of pets into natural systems can cause large scale invasions. In some areas of southern Florida, Black Spiny tail iguanas have become invasive due to a single release event of three individuals by a pet owner

(Krysko et al. 2003). Their omnivorous diet means they not only pose an ecological threat via consumption of hatchling sea turtles and bird eggs, but are also a nuisance to residents, eating the flowers and fruits in their gardens. They may possibly be competing with gopher tortoises for burrowing space and native vegetation (Krysko et al. 2003).

Vertebrates, particularly fish, birds, and mammals, have often been deliberately introduced to increase hunting or sport fishing opportunities. Micropterus salmoides, the

6 largemouth bass, and Oncorhynchus mykiss, the rainbow trout, have been widely introduced for recreational fishing; both now claim spots on the „One Hundred of the

World‟s Worst Invasive Alien Species‟ list compiled by the Global Invasive Species

Database (Cambray 2003).

Organisms of all taxa have been released intentionally in the hopes of providing biological control of other invasive species. Although there are many successful cases of biological control, the risks to non-target species are difficult to fully assess, and have prompted much debate (Simberloff and Stiling 1996b, Louda et al. 2003, Louda and

Stiling 2004). Despite the existing potential, there are relatively few documented examples of adverse effects on native or non-target species by biological control agents, but this is very likely due to lack of monitoring for these effects (Simberloff and Stiling

1996a, Frank and McCoy 2007). Cactoblastis cactorum, the primary study organism of this dissertation, is one of ten documented cases of non-target impacts reviewed by

Louda et al (2003). Cactoblastis cactorum has been used as a successful biological control agent of Opuntia (prickly pear) species in Australia, South Africa, and the

Caribbean. It eventually arrived in Florida where it currently threatens three endemic opuntioids in the Keys as well as native, non-pest Opuntia populations throughout the southeastern US.

The rate at which exotic species have been introduced over the past several hundred years reflects our increasing global trade and travel. Insects, in particular, provide a good example of this. The number of exotic insects introduced annually into the U.S. increased from 1640 to 1980 with the rate of increase becoming exponential around 1920 (Sailer1983, in Mooney and Drake 1989). A study by Frank and McCoy

(1992) found that the rate of insect immigration to the US, and perhaps specifically to the state of Florida, increased markedly in the 1980s, although the increase may partially be a product of better detection methods. Levine and D‟Antonio (2003) predicted an increase of 3-60% in the number of non-indigenous insect species present in the U.S. from 2000 to 2020. The large range in the percent increase demonstrates the difficulty in making accurate, quantitative, long-term predictions. It is not only our travel and trade- related activities that lead to establishment of exotics. Humans continue to pollute and alter the atmosphere, climate, waterways, and terrestrial landscapes, leading to even greater opportunities for invasive species to establish (Mooney and Drake 1989).

The state of Florida, where this doctoral research took place, is particularly prone to the establishment of non-indigenous species because the subtropical portion of the peninsula is bordered by water on three sides and by frost to the north, making it a

“habitat island”. Similar to oceanic islands, it has relatively poor native species richness

(Simberloff 1997). In addition, Florida‟s many miles of coastline, lakes, and rivers ease the spread of introduced species, as do the nearly 50 million tourists that visit annually

(Invasive Species Working Group 2003). The large ornamental plant business contributes significantly to the problem; the Port of Miami receives 85% of the live nonindigenous plant shipments that arrive each year in the US (Gordon 1998). The Florida Exotic Pest

Plant Council lists 147 species of invasive plants on its website, 73 of which are altering native plant communities (FLEPPC 2009). Florida has the largest number of established non-indigenous herpetofaunal species in the entire world (Krysko et al. 2011) and is second only to Hawaii in number of breeding exotic vertebrates (Forys and Allen 1999).

While certainly bad news for native flora and fauna, the statements above make Florida an ideal place for undertaking research of invasive species.

Cactaceae: Opuntioideae

The subfamily Opuntioideae in the family Cactaceae contains five tribes and between fifteen and eighteen genera (Griffith and Porter 2009). The taxa in this work come from two tribes, Cylindropuntieae and Opuntieae, and four genera, Consolea, Cylindropuntia,

Nopalea, and Opuntia (see Appendix I).

All opuntioid species are native to the New World. Members of Opuntieae can be found as far north as British Columbia, in nearly all the contiguous United States, the

Caribbean, and south to the Strait of Magellan. Cylindropuntieae has a narrower distribution, mainly Mexico and the southwestern US (Britton and Rose 1937, USDA-

NRCS 2010). Florida has six native opuntioids and three species which are considered naturalized (Wunderlin and Hansen 2008). Taxa in both Opuntieae and Cylindropuntieae have jointed stems called cladodes. In Opuntieae the cladodes are generally flattened while in Cylindropuntieae they are cylindrical or globular. Both tribes have species that are low-growing and shrub-like and species that are tree-like. The cladodes often bear larger, fixed spines as well as small, barbed, deciduous bristles called glochids.

Opuntioids are both ecologically and economically important. Songbirds, rodents, snakes, and lizards often take shelter in their spiny cladodes in order to receive shade and protection (Soberon et al. 2001). The plants also provide nesting structures for birds such as the Northern bobwhite, reddish egrets, cactus wrens and curve-billed thrashers

(Chavez-Ramirez et al. 1997, Slater et al. 2001, Hernandez et al. 2003, NatureServe

2011). The fruits, seeds, and pads are eaten by a wide variety of animals, including deer, javelinas, lagomorphs, rodents, birds, and tortoises (Macdonald and Mushinsky 1988,

Hoffman et al. 1993, Chavez-Ramirez et al. 1997, Soberon et al. 2001, Birkhead et al.

2005). Many arthropods feed upon the cladodes and make use of their complex structure for oviposition sites or web construction, while the flowers provide nectar and pollen to bees, beetles, and butterflies. Prickly pears also prevent soil erosion and desertification, particularly in arid regions (Soberon et al. 2001).

In Mexico, approximately 250,000 ha of Opuntia are cultivated, bringing in over

80 million dollars annually, and another 3 million ha of wild populations are used for fodder (Soberon et al. 2001). The fruit is used in jams and juices and the cladodes are eaten as a vegetable (nopales), used as fodder, burned as an energy source, and used in manufacturing soap, cosmetics, fertilizer, and adhesives (Soberon et al. 2001, Vigueras and Portillo 2001). Clearly, there are ample ecological and economical reasons to protect opuntioid diversity.

Cactoblastis cactorum

Biology

Cactoblastis cactorum is a cactus-feeding moth belonging to the Lepidopteran family

Pyralidae and the subfamily Phycitinae. It is native to northern Argentina, Paraguay,

Uruguay, and southern Brazil, where it is distributed over a wide range of climates.

Larvae are oligophagous, feeding on at least 30 different hosts in the genera Opuntia,

Consolea, and Nopalea (Dodd 1940, Mann 1969, Zimmermann et al. 2004).

Adult females are pale brown and grey and have a wingspan of 27-40mm; males are lighter in color and smaller (23-32mm). The sex ratio is normally 1:1 (Legaspi et al

2009, J. Carpenter, personal communication), but shifts to favor males when conditions for larvae are suboptimal (Dodd 1940, Zimmermann et al 2004) . The average adult lives about 9 days and does not feed. Moths emerge from pupae at dusk and mate at dawn on the first or second day after emergence; females begin laying eggs the next night.

Females lay their eggs one on top of the other to form an eggstick that is attached to the tip of a cactus spine or directly to the cladode. An average eggstick has 70-90 eggs and a female may lay several sticks for a total of 200-300 eggs. All eggs from a particular eggstick hatch at the same time, usually within three or four hours. The neonates work together to chew a communal entrance hole into the cladode, allowing them to better overcome the mucilage produced by the cactus when its epidermis is penetrated. They tunnel inside the cladode, feeding as a group until only the cuticle and vascular bundles remain. At this point they may either tunnel internally into an adjacent segment or exit the damaged cladode to crawl over the plant‟s surface to a fresh one (Dodd 1940, Mann

1969). The eaten cladodes are left with a hollowed out, rotting appearance and are more susceptible to secondary microbial infections (Starmer et al. 1987). Larvae go through five instars. Early instars are pinkish-cream with dark red dots on the back of each segment. Later instars become bright orange, and the dots expand and join to become a dark band across each segment. The caterpillars reach a length of about 1.5 cm. By the time the larvae are full-grown, the colony has split up and each caterpillar separately exits the plant, drops to the ground below, and pupates in the soil or leaf litter (Dodd

1940, Zimmermann et al. 2004).

In both its native range in and temperate areas of introduction C. cactorum has two generations per year; the shorter summer generation has an average lifecycle of 100-

120 days while the longer winter generation averages 235-265 days. In warmer areas, such as Florida, central Queensland, and the Caribbean, there are three generations per year with a shorter average lifecycle length of about 90 days (Zimmermann et al. 2004,

Hight and Carpenter 2009). The host plants of C. cactorum in Florida include all six native opuntioid species (Consolea coralicolla, O. stricta, O.humifusa, O. pusilla, O. triacantha, and O. cubensis) and several naturalized species (O. ficus-indica, O. cochenillifera, and O. leucotricha). There is concern for all the native species, but especially for C. coralicolla, O. cubensis, and O. triacantha because in the U.S. they are found only in the Florida Keys. Consolea coralicolla is an endangered, endemic species that is restricted to two populations. The Little Torch Key population contains just seven mature stems (Stiling et al. 2000), while a larger population on Swan Key in Biscayne

National Park contains 586 individuals that have not thus far been attacked by C. cactorum (Cariaga et al. 2005).

History: Australia, South Africa, and the Caribbean

C. cactorum has been used in several countries to control Opuntia species. The most famous instance of its use as a biological control agent took place in Australia, specifically in Queensland and New South Wales (NSW). The first two introductions of

C. cactorum failed to establish, but the third attempt in 1925 was immensely successful due to intensive mass rearing and distribution efforts by the Commonwealth Prickly Pear

Board. The intial 1925 shipment of 2,750 eggs was increased to 2,540,000 after two

12 generations of cage-rearing. Mass rearing, collection from field sites, and distribution throughout Opuntia-infested areas continued until 1930, with an estimated total of 3 billion eggs released (Dodd 1940). By 1933, C. cactorum had eradicated 90% of the 25 million hectares infested with nonnative O. inermis and O. stricta, allowing the land to be reclaimed for agricultural purposes. However, persistent, dense stands of Opuntia still exist in some coastal regions; water and nutrient stress and dispersal of females to nearby islands have been suggested as explanations for the lack of control in these areas (Monro

1975, White 1980, Hosking et al. 1994).

After seeing the impressive results produced in Australia, several other countries imported C. cactorum eggsticks in an effort to control prickly pears. In 1932, South

Africa received 112,600 eggs from Australia, mainly to use against O. ficus-indica.

Although nearly 580 million eggsticks were distributed in the eastern provinces, the moth did not afford the same level of control as in Queensland (Pettey 1948). The difference in success is thought to be due to several factors, including the less succulent, woodier texture of O. ficus-indica, ant predation, and South Africa‟s colder temperatures that inhibit egg-laying and prolong development times (Pettey 1948, Robertson 1989,

Robertson and Hoffmann 1989). Although it has not completely cleared areas of O. ficus- indica, C. cactorum has helped to thin out dense stands and reduce its spread by killing small plants and destroying terminal segments of larger plants. Less than 100,000 hectares of the original 900,000 remain, but much of this success has been attributed to the scale insect Dactylopius opuntiae, which is considered the primary biological control agent (Zimmermann and Moran 1991). The species O. stricta and O. aurantiaca are also considered pest weeds in South Africa. Opuntia stricta is particularly problematic in

Kruger National Park, where C. cactorum has decreased the average size of plants, but increased the density of plants through fragmentation, resulting in no net reduction in fruiting (Moran and Zimmermann 1991, Hoffmann et al. 1998, Foxcroft et al. 2007).

In 1957, C. cactorum was released on the Caribbean island of Nevis to control several Opuntia species that had reached abnormal densities due to overgrazing. Despite their status as weeds, the targeted cacti were all native species. Non-target effects and the possible dispersal of the moth to neighboring islands were apparently not enough of concern for the release to be contested. Following the very successful control of the

Opuntia species, especially of O. triacantha, on Nevis, C. cactorum was purposely introduced on Montserrat and Antigua in 1960 and Grand Cayman in 1970. It also spread to St. Kitts, the US Virgin Islands, Puerto Rico, Haiti, the Dominican Republic,

Hispaniola, the Bahamas, and Cuba. Whether C. cactorum spread among these islands via natural dispersal, unintentional human activity, or intentional, but illegal transport by humans is not known for sure, but it is likely that all three factors played a role.

Although many of the weedy, target species of Opuntia were brought under control in the Caribbean, non-target species, such as O. repens in Puerto Rico, were also affected (Tuduri et al. 1971, Bennett and Habeck 1995, Zimmermann et al. 2005). The status of C. cactorum and Opuntia in the Caribbean was the subject of a 2005 joint report by the FAO and the IAEA (Zimmermann et al. 2005). Perhaps the most surprising piece of information from this report was the fact that, except for Puerto Rico, none of the personnel from quarantine or plant health agencies on the surveyed islands knew of the presence and effects of C. cactorum on their islands prior to this survey. The authors attribute this ignorance to a general lack of interest in Opuntia species in the Caribbean,

14 where they are not actively used or cultivated and now mostly exist only as minor pests.

On many of the islands surveyed in 2005 (Montserrat, Antigua, St. Kitts, Grand Cayman,

Guadeloupe, Dominique, Grenada, St. Lucia and Martinique) Opuntia populations were sparse and a high proportion of the plants that were located had been attacked by C. cactorum. The first confirmed record of the moth on Jamaica was made during the 2005 survey. Cactoblastis cactorum damage was found on five species and botanists described a recent decline in several Opuntia species, including the already rare O. sanguinea. The moth is apparently also abundant in Cuba and the Dominican Republic, although time and political constraints prevented a thorough survey of these islands. A 2002 survey of

Nevis (Pemberton and Liu 2007), the original site of C. cactorum introduction in the

Caribbean, found the moth on two native (O. stricta (dillenii) and O. triacantha) and one naturalized species (N. cochenillifera), although less than 30% of plants were attacked.

As has been found in previous studies, larger plants were more likely to be attacked, but smaller plants had a greater proportion of attacked cladodes, indicating that they may suffer a higher mortality risk (Pemberton and Liu 2007).

Other locales where C. cactorum has been successfully introduced for control of non-native Opuntia include Mauritius, Hawaii, St. Helena Island, and Ascension Island.

It has been introduced, but has not established in Pakistan, Kenya, and Israel

(Zimmermann et al. 2004).

History: U.S. and Mexico

In 1989, after being warned of C. cactorum‟s presence in Cuba (still unconfirmed at that point), scientists from the Fairchild Tropical Gardens in Miami visited the Florida

Keys to check for the moth. It was found on O. stricta on Big Pine Key and, less than one year later, in Key Biscayne State Park near Miami. The large distance (200 km) between these two locations, coupled with records of C. cactorum interceptions in Miami between 1981 and 1986, suggest that the moth may have arrived via imported Opuntia from the Dominican Republic (Pemberton 1995, Stiling 2002). Regardless of whether the moth‟s arrival was human-mediated, or it flew the 145 km from Cuba to the Keys, C. cactorum has since spread up the east coast to Bull Island, SC, which has remained its northern leading edge since 2004. The farthest western detection occurred in 2009 in

Terre Bonne Parrish, south of New Orleans, LA (USDA-APHIS-PPQ 2009b).

Ongoing efforts in the southeastern states are being made to carefully monitor the spread of C. cactorum and to push back the western leading edge of its distribution.

Mechanical control (removal of eggsticks by hand) is not practical and biological and chemical control methods could have detrimental non-target effects on native lepidopterans (Stiling 2002). Therefore, genetic control via the release of sterile males is the best option for halting the spread of the moth. This technique is currently used on barrier islands along the AL and MS coast (Carpenter et al. 2001, Hight et al. 2005).

Pheromone traps are also being used to monitor the moth‟s current range and to survey for new outbreaks (USDA-APHIS-PPQ 2010).

In 2006 C. cactorum was discovered on Isla Mujeres in the Mexican state of

Quintana Roo; this was followed by detection in 2007 on nearby Isla Contoy. The most

16 likely arrival path is thought to be from Cuba, Haiti, or the Dominican Republic

(Zimmermann et al. 2004). An eradication program on Isla Mujeres, focusing on host removal, pheromone trapping, and outreach, was quickly put into place. Subsequently, the eradication program was extended to Isla Contoy, where infested Opuntia cladodes were removed and sterile moths from the USDA mass rearing program in Tifton, GA were released. Traps were also set up on the Mexican mainland in the Cancun area

(Rodriguez 2007). Both islands have since been declared eradicated and as of this writing, no other Mexican infestations have been found (USDA-APHIS-PPQ 2008,

2009a)

References cited

Anagnostakis, S. L. 2001. American chestnut sprout survival with biological control of the chestnut-blight fungus population. Forest Ecology and Management 152:225-233.

Bennett, F. D. and D. H. Habeck. 1995. Cactoblastis cactorum: A successful weed control agent in the Caribbean, now a pest in Florida? Pages 21-26 in E. S. Delfosse and R. R. Scott, editors. Biological control of weeds. Proceedings of the 8 International Symposium on Biological Control of Weeds. Lincoln University, Canterbury, New Zealand, 2-7 February 1992.

Birkhead, R. D., C. Guyer, S. M. Hermann, and W. K. Michener. 2005. Patterns of folivory and seed ingestion by gopher tortoises (Gopherus polyphemus) in a southeastern pine savanna. American Midland Naturalist 154:143-151.

Britton, N. L. and J. N. Rose. 1937. The Cactaceae: Descriptions and Illustrations of Plants of the Cactus Family. Dover Publications, New York.

Cambray, J. A. 2003. Impact on indigenous species biodiversity caused by the globalisation of alien recreational freshwater fisheries. Hydrobiologia 500:217- 230.

Cariaga, K. A., C. E. Lewis, J. Maschinski, S. J. Wright, and J. Francisco-Ortega. 2005. Patterns of genetic diversity in the critically endangered Florida key endemic

Consolea corallicola Small (Cactaceae): Evidence from inter-simple sequence repeat (ISSRs) DNA polymorphisms. Caribbean Journal of Science 41:225-233.

Carpenter, J. E., K. A. Bloem, and S. Bloem. 2001. Applications of F-1 sterility for research and management of Cactoblastis cactorum (Lepidoptera : Pyralidae). Florida Entomologist 84:531-536. Chavez-Ramirez, F., X. G. Wang, K. Jones, D. Hewitt, and P. Felker. 1997. Ecological characterization of Opuntia clones in South Texas: Implications for wildlife herbivory and frugivory. Journal of the Professional Association for Cactus Development 2:9-19. Colautti, R. I., I. A. Grigorovich, and H. J. MacIsaac. 2006. Propagule pressure: A null model for biological invasions. Biological Invasions 8:1023-1037. Colautti, R. I., A. Ricciardi, I. A. Grigorovich, and H. J. MacIsaac. 2004. Is invasion success explained by the enemy release hypothesis? Ecology Letters 7:721-733. Crooks, J. A. 2002. Characterizing ecosystem-level consequences of biological invasions: The role of ecosystem engineers. Oikos 97:153-166. Davelos, A. L. and A. M. Jarosz. 2004. Demography of American chestnut populations: Effects of a pathogen and a hyperparasite. Journal of Ecology 92:675-685. Davis, M. A., J. P. Grime, and K. Thompson. 2000. Fluctuating resources in plant communities: A general theory of invasibility. Journal of Ecology 88:528-534. Dodd, A. P. 1940. The biological campaign against prickly pear. Commonwealth Prickly Pear Board Bulletin:1-177. Brisbane, Australia. Ehrenfeld, J. G., P. Kourtev, and W. Z. Huang. 2001. Changes in soil functions following invasions of exotic understory plants in deciduous forests. Ecological Applications 11:1287-1300. Ehrlich, P. R. 1989. Attributes of invaders and the invading process: Vertebrates. Pages 315-328 in J. A. Drake et al, editors. Biological Invasions: A Global Perspective. Wiley and Sons, Chichester, U.K. Executive Order No. 13112. 1999. Federal Registry 64:6183 FLEPPC. 2009. List of Invasive Plant Species. Wildland Weeds 12:13-16. Forys, E. A. and C. R. Allen. 1999. Biological invasions and deletions: Community change in south Florida. Biological Conservation 87:341-347. Foxcroft, L. C., J. H. Hoffmann, J. J. Viljoen, and J. J. Kotze. 2007. Factors influencing the distribution of Cactoblastis cactorum, a biological control agent of Opuntia stricta in Kruger National Park, South Africa. South African Journal of Botany 73:113-117.

Frank, J. H. and E. D. McCoy. 1995. Introduction to insect behavioral ecology: The good, the bad, and the beautiful: Nonindigenous species in Florida. Invasive adventive insects and other organisms in Florida. Florida Entomologist 78:1-15. Frank, J. H. and E. D. McCoy. 2007. The risk of classical biological control in Florida. Biological Control 41:151-174. Freed, L. A., R. L. Cann, M. L. Goff, W. A. Kuntz, and G. R. Bodner. 2005. Increase in avian malaria at upper elevation in Hawai'i. Condor 107:753-764. Fritts, T. H. and G. H. Rodda. 1998. The role of introduced species in the degradation of island ecosystems: A case history of Guam. Annual Review of Ecology and Systematics 29:113-140. Gordon, D. R. 1998. Effects of invasive, non-indigenous plant species on ecosystem processes: Lessons from Florida. Ecological Applications 8:975-989. Gottdenker, N. L., T. Walsh, H. Vargas, J. Merkel, G. U. Jimenez, R. E. Miller, M. Dailey, and P. G. Parker. 2005. Assessing the risks of introduced chickens and their pathogens to native birds in the Galapagos Archipelago. Biological Conservation 126:429-439. Griffith, M. P. and J. M. Porter. 2009. Phylogeny of Opuntioideae (Cactaceae). International Journal of Plant Sciences 170:107-116. Heath, R. T., G. L. Fahnenstiel, W. S. Gardner, J. F. Cavaletto, and S. J. Hwang. 1995. Ecosystem-level effects of zebra mussels (Dreissena polymorpha): An enclosure experiment in Saginaw Bay, Lake Huron. Journal of Great Lakes Research 21:501-516. Hernandez, F., S. E. Henke, N. J. Silvy, and D. Rollins. 2003. The use of prickly pear cactus as nesting cover by northern bobwhites. Journal of Wildlife Management 67:417-423. Hight, S. D. and J. E. Carpenter. 2009. Flight phenology of male Cactoblastis cactorum (Lepidoptera: Pyralidae) at different latitudes in the southeastern United States. Florida Entomologist 92:208-216. Hight, S. D., J. E. Carpenter, S. Bloem, and K. A. Bloem. 2005. Developing a sterile insect release program for Cactoblastis cactorum (Berg) (Lepidoptera : Pyralidae): Effective overflooding ratios and release-recapture field studies. Environmental Entomology 34:850-856. Hoffman, M. T., C. D. James, G. I. H. Kerley, and W. G. Whitford. 1993. Rabbit herbivory and its effect on cladode, flower, and fruit production of Opuntia violacea var. macrocentra (Cactaceae) in the Northern Chihuahuan Desert, New- Mexico. Southwestern Naturalist 38:309-315.

Hoffmann, J. H., V. C. Moran, and D. A. Zeller. 1998. Evaluation of Cactoblastis cactorum (Lepidoptera : Phycitidae) as a biological control agent of Opuntia stricta (Cactaceae) in the Kruger National Park, South Africa. Biological Control 12:20-24. Hosking, J. R., P. R. Sullivan, and S. M. Welsby. 1994. Biological control of Opuntia stricta (Haw) Haw var. stricta using Dactylopius opuntiae (Cockerell) in an area of New South Wales, Australia, where Cactoblastis cactorum (Berg) is not a successful biological control agent. Agriculture Ecosystems & Environment 48:241-255. Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69:373-386. Keane, R. M. and M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis. Trends in Ecology & Evolution 17:164-170. Kennedy, T. A., S. Naeem, K. M. Howe, J. M. H. Knops, D. Tilman, and P. Reich. 2002. Biodiversity as a barrier to ecological invasion. Nature 417:636-638. Kiritani, K. and K. Yamamura. 2003. Exotic insects and their pathways for invasion. Pages 44-67 in G. M. Ruiz and J. T. Carlton, editors. Invasive Species: Vectors and Management Strategies. Island Press, Washington D.C. Kleindorfer, S. and R. Y. Dudaniec. 2006. Increasing prevalence of avian poxvirus in Darwin's finches and its effect on male pairing success. Journal of Avian Biology 37:69-76. Kolar, C. S. and D. M. Lodge. 2001. Progress in invasion biology: Predicting invaders. Trends in Ecology & Evolution 16:199-204. Kraus, F. 2003. Invasion pathways for terrestrial vertebrates. Pages 68-92 in G. M. Ruiz and J. T. Carlton, editors. Invasive Species: Vectors and Management Strategies. Island Press, Washington D.C. Krysko, K. L., J. P. Burgess, M. R. Rochford, C. R. Gillette, D. Cueva, K. M. Enge, L. A. Somma, J. L. Stabile, D. C. Smith, J. A. Wasilewski, G. N. Kieckhefer, M. C. Granatosky, and S. V. Nielsen. 2011. Verified non-indigenous amphibians and reptiles in Florida from 1863 through 2010: Outlining the invasion process and identifying invasion pathways and stages. Zootaxa Issue 3028 (1-64). Krysko, K. L., F. W. King, K. M. Enge, and A. T. Reppas. 2003. Distribution of the introduced black spiny-tailed iguana (Ctenosaura similis) on the southwestern coast of Florida. Biological Sciences 66:74-79. Legaspi, J.C., I. Baez, B. C. Legaspi. 2009. Phenology and egg production of the cactus moth (Lepidoptera: Pyralidae): comparison of field census data and life stage development in the field. Journal of Entomological Science 44: 341-352.

Lonsdale, W. M. 1999. Global patterns of plant invasions and the concept of invasibility. Ecology 80:1522-1536. Louda, S. M., R. W. Pemberton, M. T. Johnson, and P. A. Follett. 2003. Nontarget effects - The Achilles' Heel of biological control? Retrospective analyses to reduce risk associated with biocontrol introductions. Annual Review of Entomology 48:365-396. Louda, S. M. and P. Stiling. 2004. The double-edged sword of biological control in conservation and restoration. Conservation Biology 18:50-53. Macdonald, L. A. and H. R. Mushinsky. 1988. Foraging ecology of the gopher tortoise, Gopherus polyphemus, in a sandhill habitat. Herpetologica 44:345-353. Mack, R. N. 2003. Global plant dispersal, naturalization, and invasion: Pathways, models, and circumstances. Pages 3-30 in G. M. Ruiz and J. T. Carlton, editors. Invasive Species: Vectors and Management Strategies. Island Press, Washington, D.C. Mack, R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. A. Bazzaz. 2000. Biotic invasions: Causes, epidemiology, global consequences, and control. Ecological Applications 10:689-710. Manchester, S. J. and J. M. Bullock. 2000. The impacts of non-native species on UK biodiversity and the effectiveness of control. Journal of Applied Ecology 37:845- 864. Mann, J. 1969. Cactus-feeding Insects and Mites. United States National Museum Bulletin 256. Monro, J. M. 1975. Environmental variation and the efficiency of biological control - Cactoblastis in the southern hemisphere. Pages 204-212 in Kikkawa, J. and H.A. Nix (Eds.). Proceedings of the Ecological Society of Australia, Vol. 9. Managing Terrestrial Ecosystems. Symposium. Brisbane, Australia., May 15-16, 1975. Mooney, H. A. and J. A. Drake. 1989. Biological Invasions: A SCOPE program overview. Pages 491-503 in J. A. Drake et al, editors. Biological invasions: A global perspective. Wiley and Sons, Chichester, U.K. Moran, V. C. and H. G. Zimmermann. 1991. Biological control of jointed cactus, Opuntia aurantiaca (Cactaceae), in South Africa. Agriculture Ecosystems & Environment 37:5-27. Myers, R. L. 1983. Site susceptibility to invasion by the exotic tree Melaleuca- quinquenervia in Southern Florida. Journal of Applied Ecology 20:645-658. NatureServe. 2011. NatureServe Explorer: An online encyclopedia of life, Version 7.1 NatureServe.

Noble, I. R. 1989. Attributes of invaders and the invading process: Terrestrial and vascular plants. Pages 301-313 in J. A. Drake et al, editors. Biological invasions: A global perspective. Wiley and Sons, Chichester, UK. Pemberton, R. W. 1995. Cactoblastis cactorum (Lepidoptera: Pyralidae) in the United States: An immigrant biological control agent or an introduction of the nursery industry? American Entomologist 41:230-232. Pemberton, R. W. and H. Liu. 2007. Control and persistence of native Opuntia on Nevis and St. Kitts 50 years after the introduction of Cactoblastis cactorum. Biological Control 41:272-282. Pettey, F. W. 1948. The biological control of prickly pear in South Africa. Science Bulletin. Department of Agriculture of the Union of South Africa. Pimentel, D., R. Zuniga, and D. Morrison. 2005. Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecological Economics 52:273-288. Rejmanek, M. and D. M. Richardson. 1996. What attributes make some plant species more invasive? Ecology 77:1655-1661. Robertson, H. G. 1989. Seasonal temperature effects on fecundity of Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae): Differences between South Africa and Australia. Journal of the Entomological Society of Southern Africa 52:71-80. Robertson, H. G. and J. H. Hoffmann. 1989. Mortality and life-tables of Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae) compared on two host-plant species. Bulletin of Entomological Research 79:7-17. Rodda, G. H. and J. A. Savidge. 2007. Biology and impacts of pacific island invasive species. 2. Boiga irregularis, the Brown Tree Snake (Reptilia: Colubiridae). Pacific Science 61:307-324. Rodriguez, R. Z. 2007. Detection and eradication campaign on Isla Mujeres, Quintana Roo, Mexico. International Cactoblastis cactorum Conference, Phoenix, AZ. Schwindt, E., O. O. Iribarne, and F. I. Isla. 2004. Physical effects of an invading reef- building polychaete on an Argentinean estuarine environment. Estuarine Coastal and Shelf Science 59:109-120. Simberloff, D. 1989. Which insect introductions succeed and which fail? Pages 61-75 in J. A. Drake et al, editors. Biological invasions: A global perspective. Wiley and Sons, Chichester, UK. Simberloff, D. 1997. The biology of invasions. Pages 3-20 in D. Simberloff, D. C. Schmitz, and T. C. Brown, editors. Stranger in Paradise: Impact and Management of Nonindigenous Species in Florida. Island Press, Washington, D.C.

Simberloff, D. and P. Stiling. 1996a. How risky is biological control? Ecology 77:1965- 1974. Simberloff, D. and P. Stiling. 1996b. Risks of species introduced for biological control. Biological Conservation 78:185-192. Slater, S. C., D. Rollins, R. C. Dowler, and C. B. Scott. 2001. Opuntia: A "prickly paradigm" for quail management in west-central Texas. Wildlife Society Bulletin 29:713-719. Soberon, J., J. Golubov, and J. Sarukhan. 2001. The importance of Opuntia in Mexico and routes of invasion and impact of Cactoblastis cactorum (Lepidoptera : Pyralidae). Florida Entomologist 84:486-492. Starmer, W. T., V. Aberdeen, and M. Lachance. 1987. The yeast community associated with decaying Opuntia stricta (Haworth) in Florida with regard to the moth. Cactoblastis cactorum (Berg). Florida Scientist 51:1-6. Stiling, P. 2002. Potential non-target effects of a biological control agent, prickly pear moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae), in North America, and possible management actions. Biological Invasions 4:273-281. Stiling, P. 2004. Ecology: theories and applications. 4th edition. Prentice Hall India, New Delhi. Stiling, P., A. Rossi, and D. Gordon. 2000. The difficulties of single factor thinking in restoration: replanting a rare cactus in the Florida Keys. Biological Conservation 94:327-333. Tsutsui, N. D., A. V. Suarez, D. A. Holway, and T. J. Case. 2000. Reduced genetic variation and the success of an invasive species. Proceedings of the National Academy of Sciences of the United States of America 97:5948-5953. Tuduri, J. C. G., L. F. Martorell, and S. M. Gaud. 1971. Geographical distribution and host plants of the cactus moth, Cactoblastis cactorum (Berg) in Puerto Rico and the US Virgin Islands. Journal of the Agricultural University Puerto Rico 58:130- 134. USDA-APHIS-PPQ. 2008. Cactus moth (Cactoblastis cactorum) program update. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine. USDA-APHIS-PPQ. 2009a. Cactus moth (Cactoblastis cactorum) program update. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine. USDA-APHIS-PPQ. 2009b. Cactus moth program annual report, Eastern region. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine. 23

USDA-APHIS-PPQ. 2010. Cactus moth program annual report, Eastern region. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine. USDA-NRCS. 2011. The PLANTS Database (http://plants.usda.gov), 1 October 2011. National Plant Data Team, Greensboro, NC 27401-4901 USA. Vanderploeg, H. A., T. F. Nalepa, D. J. Jude, E. L. Mills, K. T. Holeck, J. R. Liebig, I. A. Grigorovich, and H. Ojaveer. 2002. Dispersal and emerging ecological impacts of Ponto-Caspian species in the Laurentian Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 59:1209-1228. Veblen, T. T., M. Mermoz, C. Martin, and T. Kitzberger. 1992. Ecological impacts of introduced animals in Nahuel-Huapi-National-Park, Argentina. Conservation Biology 6:71-83. Veblen, T. T., M. Mermoz, C. Martin, and E. Ramilo. 1989. Effects of exotic deer on forest regeneration and composition in northern Patagonia. Journal of Applied Ecology 26:711-724. Vigueras, A. L. and L. Portillo. 2001. Uses of Opuntia species and the potential impact of Cactoblastis cactorum (Lepidoptera : Pyralidae) in Mexico. Florida Entomologist 84:493-498. Vitousek, P. M. 1990. Biological invasions and ecosystem processes - towards an integration of population biology and ecosystem studies. Oikos 57:7-13. Vitousek, P. M., L. R. Walker, L. D. Whiteaker, D. Muellerdombois, and P. A. Matson. 1987. Biological Invasion by Myrica faya alters ecosystem development in Hawai‟i. Science 238:802-804. White, G. C. 1980. Current status of prickly pear control by Cactoblastis cactorum in Queensland. Pages 609-616 in Proceedings of the Vth International Symposium on Biological Control of Weeds, Brisbane, Australia. Wikelski, M., J. Foufopoulos, H. Vargas, and H. Snell. 2004. Galapagos birds and diseases: Invasive pathogens as threats for island species. Ecology and Society 9. Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying threats to imperiled species in the United States. Bioscience 48:607-615. Williamson, M. 1999. Invasions. Ecography 22:5-12. Williamson, M. and A. Fitter. 1996a. The varying success of invaders. Ecology 77:1661- 1666. Williamson, M. H. and A. Fitter. 1996b. The characters of successful invaders. Biological Conservation 78:163-170.

Wunderlin, R. P. and B. F. Hansen. 2008. Atlas of Florida Vascular Plants. Institute for Systematic Botany, University of South Florida, Tampa. Zimmermann, H. G., S. Bloem, and H. Klein. 2004. Biology, history, threat, surveillance and control of the cactus moth, Cactoblastis cactorum. International Atomic Energy Agency/Food and Agriculture Organization, Vienna, Austria. Zimmermann, H. G., M. P. Cuen, and A. Rivera. 2005. The status of Cactoblastis cactorum (Lepidoptera: Pyralidae) in the Caribbean and the likelihood of its spread to Mexico. IAEA, Vienna. Zimmermann, H. G. and V. C. Moran. 1991. Biological control of prickly pear, Opuntia ficus-indica (Cactaceae), in South Africa. Agriculture Ecosystems & Environment 37:29-35.

CHAPTER 2

OVIPOSITION PREFERENCE AND LARVAL PERFORMANCE OF

CACTOBLASTIS CACTORUM ON NORTH AMERICAN OPUNTIOID CACTI

Synopsis

Cactoblastis cactorum Berg (Lepidoptera: Pyralidae), the cactus moth, is a well-known biological control agent of prickly pear cactus (Cactaceae: Opuntia Miller). The arrival of the moth in Florida and its subsequent spread through the southeastern United States poses a threat to opuntioid diversity in North America. Of particular concern are the ecological and economic impacts the moth could have in the southwestern United States and Mexico, where both native and cultivated Opuntia species are important resources. It is unknown which species would best support larval development if the moth were to spread further westward in North America. This study aimed to determine if ovipositing females demonstrate preferences for any of fourteen common opuntioids native to or naturalized in Mexico and the southwestern United States; which of these opuntioids best support larval development; and if oviposition preference correlates with larval performance, as predicted by simple adaptive models. Results from a field experiment showed that female moths preferred O. engelmannii Salm-Dyck ex Engelmann var. linguiformis (Griffiths) Parfitt and Pinkava and O. engelmannii var. engelmannii for oviposition. A generalized linear model showed number of cladodes and degree of spininess to be significant predictors of oviposition activity. Results from a no-choice

26 larval survival experiment found Consolea rubescens (Salm-Dyck ex de Candolle.)

Lemaire and O. streptacantha Lemaire to be the best hosts among the taxa tested.

Epidermal toughness was a significant predictor of most larval fitness parameters. In general, oviposition preference was not correlated with larval performance. A lack of co- evolutionary history between C. cactorum and North American opuntioid species may help explain this disconnect.

Introduction

The cactus moth, Cactoblastis cactorum Berg (Lepidoptera: Pyralidae), is well known for its role in the biological control of Opuntia Miller species, including the highly successful campaign in Australia during the 1920s and 1930s, the less successful campaign begun in the late 1930s in South Africa, and the campaign against native cacti in the Caribbean begun in the 1950s (Dodd 1940, Pettey 1948, Bennett et al. 1995,

Zimmermann et al. 2004, Zimmermann et al. 2005). However, with the 1989 discovery of C. cactorum in the Florida Keys (Habeck and Bennett 1990, Dickel 1991), the moth has become a potential threat to Opuntia biodiversity in continental North America

(Bennett et al. 1995, Zimmermann et al. 2000, Stiling 2002). Of particular concern are:

1) populations of rare opuntioids (members of the subfamily Opuntioideae: Cactaceae), for example Consolea corallicola Small and Opuntia triacantha (Willdenow) Sweet in the Florida Keys (Stiling et al. 2004, Stiling 2010) and O. sanguinea Proctor and O. jamaicensis Britton in Jamaica (Zimmermann et al. 2005); 2) the possible adverse effects on wild opuntioid species diversity found in the southwestern United States and Mexico

(Perez-Sandi 2001, Soberon et al. 2001, Vigueras and Portillo 2001, Zimmermann et al.

2004); and 3) the economic devastation to Mexico, where cultivated Opuntia is a significant commodity (Soberon et al. 2001, Vigueras and Portillo 2001, Simonsen et al.

2005).

The Opuntioideae are both ecologically and economically important. In the desert regions of the southwestern US and Mexico, wild Opuntia and Cylindropuntia

(Engelmann) Kreuzinger species provide nesting sites for birds, reptiles, and rodents, cladodes and fruit for a variety of wildlife, and contribute to soil stability (Chavez-

Ramirez et al. 1997, Hervert et al. 2005). In Mexico, approximately 250,000 ha of

Opuntia and Nopalea Salm-Dyck species are cultivated for food, livestock fodder, fuel, and industrial manufacturing; another 3 million ha of wild populations are used for fodder (Soberon et al. 2001, Vigueras and Portillo 2001). The estimated annual value of opuntioid production in Mexico is over 80 million dollars (Soberon et al. 2001). If C. cactorum is able to establish populations in the southwestern United States or in Mexico, the impacts on opuntioid diversity and agriculture could be severe.

Presently, C. cactorum is established in the southeastern United States and has been found as far north as Bull Island, South Carolina and as far west as southeastern

Louisiana (Simonsen et al. 2005, USDA-APHIS-PPQ 2009). Two populations of the moth were also detected on islands off the Yucatan peninsula, the first on Isla Mujeres

(detected April 2006) and the second on Isla Contoy (detected May 2007) (USDA-

APHIS-PPQ 2009). A combination of tactics including the sterile insect technique (SIT), in particular F1 or inherited sterility (Carpenter et al. 2001b, Carpenter et al. 2001a,

Hight et al. 2005, Tate et al. 2007), sanitation activities (removal of infested plant parts), and host plant removal has been used in western coastal Alabama since March 2005 and

28 at Isla Mujeres and Isla Contory in 2008. Both sites in Mexico have since been declared eradicated and are under surveillance by the Mexican government (USDA-APHIS-PPQ

2009). These tactics have been successful in reducing moth populations along the U.S.

Gulf Coast, but recent detections of C. cactorum in Mississippi (in January and April

2008 and August 2009) and Louisiana (in June and July 2009) (USDA-APHIS-PPQ

2009) attest to the continued threat of this pest to Opuntia-rich areas of North America.

The long-term effects of C. cactorum on cacti populations in the United States and

Mexico are uncertain, but the moth is known to kill small plants and decrease biomass of larger plants, leading to the expectation of extinctions and reduced population sizes

(Dodd 1940, Johnson and Stiling 1996, Hoffmann et al. 1998, Zimmermann et al. 2000,

Baker and Stiling 2009).

Determining the potential effects of a C. cactorum invasion involves verifying its potential host range by testing possible host species from the threatened region for oviposition preference and larval performance. Among the five species of Cactoblastis

Ragonot, all of which are native to South America, C. cactorum has the widest geographic and host range. It is distributed through southern Brazil, Uruguay, Paraguay, northern Argentina and has at least 30 known host species from three genera within the subfamily Opuntioideae (Opuntia, Consolea Lemaire, and Nopalea Salm-Dyck), many of which occur outside its native range (Mann 1969, Zimmermann et al. 2004). Opuntia and Nopalea spp. can be tree- or shrub-like, with determinate, segmented trunks and/or branches and flattened cladodes. Consolea spp. are endemic to the Caribbean Basin and are usually tree-like, with an indeterminate, unsegmented trunk from which flattened cladodes branch off in an asymmetrical fashion (Anderson 2001). Members of Consolea

29 have a unique breeding system in that they are cryptically dioecious/subdioecious, with functionally staminate flowers that occasionally produce fruit (Strittmatter et al. 2006). It is unclear how risk would vary among those hosts native to Mexico and the western

United States. Cactoblastis cactorum has been shown to attack two economically important species native to Mexico, O. ficus-indica (L.) Miller and O. streptacantha

Lemaire, but is not thought to attack Cylindropuntia species, although low levels of damage are seen when such species grow in close proximity to preferred hosts

(Zimmermann et al. 2004). In the wild, the moth would encounter many opuntioid species for the first time, and risks to these “new association” species are largely unknown. Host range testing in the southwestern United States or Mexico is not practical because of concerns about unintentional releases or escapes. To minimize the risk of escape, tests can be done in laboratories under strict quarantine, but the host range determined from laboratory tests can sometimes be a poor indicator of the host range in the field (Hajek et al. 1996, Louda et al. 1997, Louda et al. 2003). For potential biological control agents, an emerging idea is to perform field studies in the invaded region with sterilized insects; this is particularly promising for lepidopteran studies because of the process of F1 sterility (Tate et al. 2009). In C. cactorum, it has been shown that dispersal, mating frequency, and host preference are not significantly different between untreated moths and moths treated with 200 Gy of gamma radiation, the most effective dose for F1 sterility in C. cactorum (Hight et al. 2005, Tate et al.

2007, Marti and Carpenter 2009, Tate et al. 2009). An alternative option is to assess risks to target host species by moving them from the threatened region into regions where the insect is native or well-established.

The first part of this study tests C. cactorum oviposition preference among fourteen common opuntioid taxa, most of which are native to the southwestern United

States and/or Mexico, but which were moved to Florida for the experiment. Our assumption is that C. cactorum females, if they were to establish in the desert southwest, would use the same oviposition cues as they do in Florida, and would thereby exhibit similar preferences to those found in this study. We included plant height, cladode number, epidermal toughness, and spininess in a generalized linear model to determine which, if any, of these characteristics could be useful predictors of oviposition preference. The second part of this study tested larval performance in a laboratory setting, using the same cactus taxa from the oviposition experiment. Here we assume that larval performance in the laboratory will be similar to larval performance in the desert southwest, although this assumption is made cautiously. Finally, we compare larval performance with oviposition preference, as it is often thought that these two attributes should be correlated among various hosts, especially in cases where the larvae are relatively immobile (Levins and Macarthur 1969, Jaenike 1978, 1990).

Opuntioid taxonomy is not fully agreed upon and remains highly confusing, in part because of the tendency for hybridization and polymorphism among its species

(Chavez-Ramirez et al. 1997). Here, the classifications of Anderson (2001) and the

USDA-ARS Germplasm Information Resources Network (USDA-ARS 2009) were followed when classifying the fourteen potential hosts used in the study. The taxa vary in their native and introduced ranges and conservation status (see Supplemental taxa list for detailed species information).

Methods

Oviposition experiment

The oviposition study took place at Fort DeSoto Park (hereafter FDS), Pinellas County,

Florida. Three distinct flight periods of C. cactorum occur at FDS: a spring flight from

February to May, a summer flight from June to August, and a fall flight from September to November (Hight and Carpenter 2009 and pers. obsv.). Although specific densities of

C. cactorum at FDS are unknown, Baker and Stiling (2009) found that approximately

44% of Opuntia plants and 6% of cladodes at three west Florida coastal sites were damaged by C. cactorum. No effort was made to either increase or control the naturally occurring population of C. cactorum at FDS. Within the park, the native species Opuntia stricta Haworth and O. humifusa Rafinesque are found in sand dunes, scrub, and pinelands. In August of 2006, potted plants of fourteen common opuntioids were sunk into the ground, with six individuals of each taxa placed in a randomized block design.

After the cacti had been in the field for approximately two weeks, censusing was begun.

During each census, all plants were inspected for eggsticks, which were then removed.

The number of eggsticks per plant and the number of eggs per eggstick were recorded.

Censusing occurred biweekly until August 2007, with the exception of December 2006 and January 2007 when plants were censused once per month. Plant height and cladode number were recorded at the beginning and end of the experiment, and the mean values were used in analyses. Epidermal toughness, measured as grams of pressure per mm2 needed to puncture the epidermal layer, was recorded at the end of the experiment using an Effegi penetrometer (Model FT-011, International Ripening Company, Norfolk, VA).

Neither primary cladodes, i.e., highly lignified, “woody” cladodes, nor tertiary cladodes,

32 i.e., young cladodes with ephemeral true leaves, were used for toughness measurements, as they were not available on all species and, in the case of primary cladodes, are not typically eaten by C. cactorum larvae. Three secondary cladodes, i.e., cladodes that are mature, but not woody, were randomly selected from each plant and three penetrometer readings were taken per cladode. Starting from the point of attachment to the plant (or the

“joint” of the cladode), a proximal, medial, and distal reading were taken for a mean toughness reading for that cladode. As C. cactorum females often oviposit on the tips of cactus spines or on glochids at areoles where spines might appear, each opuntioid was assigned a level for a categorical variable representing density of glochids and spines

(termed spininess). Levels for spininess were found by giving each taxa separate scores for glochid and spine density (0=none to very sparse, 1=sparse, 2=moderate, 3=dense,

4=very dense), and then summing the two scores. Scores in each category were based upon personal observation and on descriptions in Anderson (2001).

Statistical analyses were performed in STATISTICA 5.5 (StatSoft 1984-2000).

There were no significant effects of block, so data were pooled among blocks. Data for proportions of total eggsticks (the sum of eggsticks laid on a plant over the course of the year divided by the sum of eggsticks laid on all plants over the course of the year) were highly non-normal, violated the homogeneity of variance assumption, and did not respond to transformations. These data were therefore compared among taxa and levels of spininess using a Kruskal-Wallis ANOVA by ranks (Kruskal and Wallis 1952).

Dunn‟s test for significant differences of mean rank (as described in Zar 1999) was used for post hoc comparisons. Numbers of eggs per eggstick met assumptions of normality and homogeneity of variance and were compared among host taxa and levels of spininess

33 using a single factor ANOVA; only those taxa that received two or more eggsticks were used in these analyses. Epidermal toughness, plant height and number of cladodes were log transformed to meet assumptions of homogeneity of variance. These variables were then used in single factor ANOVAs, grouped by taxa, and in regressions of numbers of eggs per eggstick. Means were separated by Bonferroni post-hoc tests. A generalized linear model (StatSoft 2010) was used to determine if the plant characteristics of toughness, cladode number, height, or the categorical variable spininess, could be used as predictors of oviposition preference. Number of eggsticks was used as the Poisson distributed response variable and a log link was used. Continuous variables used in the generalized linear model were examined for tolerance and found to be non-redundant (all tolerance values ≥0.83). A Type III log likelihood test was performed on the full model, and all subsets of models were compared using Akaike‟s Information Criterion (AIC)

(Akaike 1974).

Larval performance experiment

The plants used in the oviposition experiment were removed from Fort DeSoto in

August 2007 and brought to University of South Florida Botanical Garden. On

September 5, 2007, freshly laid C. cactorum eggsticks were obtained from the mass rearing facility located at the USDA-ARS Crop Protection and Management Research

Unit in Tifton, Georgia. Eggsticks containing between 40 and 60 eggs were placed upon cladodes of each of the fourteen opuntioids used in the oviposition experiment and kept in acrylic cages with ventilated lids (22x18x16cm). This experiment was designed to be

34 of the “no-choice” type, therefore only one eggstick and one type of cactus were placed in each cage, with five individuals of each taxa arranged in a randomized block design.

Larvae were kept at ~25°C and a 14L:10D photoperiod. The cages were checked daily and cladodes were replaced as needed. At the final instar, newspaper was placed in cages for pupation media. Pupae were removed after completion of a cocoon and moved to individual vials. Vials were checked daily for adult emergence and adults were placed in a freezer overnight. Variables recorded were: initial number of eggs that hatched, penetration success (into the cladode), number of penetration attempts, pupation rate, adult emergence rate, adult sex ratio (proportion of males), adult weight, number of eggs per female, and development times. Adults were sexed so that weight could be compared separately for males and females.

Analyses were performed in Statistica 5.9 (StatSoft 1984-2000). Initial numbers of larvae were not correlated with any fitness variables, and were therefore not included as co-variates. To meet assumptions of normality and homogeneity of variance, pupation and emergence rates were arcsine square root transformed, and weights and number of eggs per female were either log- or square root-transformed. Durations of larval and pupal stage were log transformed to meet assumptions of homogeneity of variance.

Pupation rate, emergence rate, weights, number of eggs per female, and development times were compared among host taxa using a single factor ANOVA. Means were separated by Bonferroni post-hoc tests. For development times and adult variables, only host taxa in which two or more replicates produced adults were used. Proportions of females were highly non-normal and did not respond to transformations; they were compared among species using a Kruskal-Wallis ANOVA by ranks (Kruskal and Wallis

1952). To determine if epidermal toughness could predict larval fitness variables, least squares linear regressions of pupation rate, emergence rate, adult weights, number of eggs per female, and development times were performed.

Results

Oviposition experiment

Ranks of proportions of eggsticks were significantly different among all taxa (Kruskal-

Wallis H = 39.59, df = 13, p = 0.0002) and when taxa receiving zero eggsticks were excluded (Kruskal-Wallis H = 20.42, df = 9, p = 0.0155). Ranks of proportions of eggsticks were significantly different among all levels of spininess (Kruskal-Wallis H =

30.35, df = 7, p = 0.0001) and when levels of spininess that received zero eggsticks were excluded (Kruskal-Wallis H = 11.34, df = 5, p = 0.0450). When comparing ranks of all levels of spininess, level 0 was significantly different from levels 6 and 7 under Dunn‟s test (Q = 3.69, Q0.05, 8 = 3.12). No other pair wise differences in ranks were found to be significant. Opuntia engelmannii Salm-Dyck ex Engelmann var. linguiformis (Griffiths)

Parfitt and Pinkava received the highest proportion of eggsticks, followed by O. engelmannii var. engelmannii Engelmann (Table 2.1). Four species [O. leucotricha de

Candolle, O. streptacantha, Cylindropuntia acanthocarpa (Engelmann and Bigelow)

F.M. Knuth, and C. spinosior (Engelmann) F.M. Knuth] received no eggsticks over the course of the year. Opuntioids given a score of 5 for spininess received the most eggsticks, followed by those with a score of 3. Those given scores of 6 or 7 did not receive any eggsticks over the course of the year (Table 2.1).

Numbers of eggs per eggsticks were not significantly different between taxa (F7,

57 = 0.94, p = 0.49) or between levels of spininess (F5, 62 = 1.28, p = 0.29), although unequal sample sizes (n = 2-17) reduced the power of the analysis. Numbers of eggs per eggstick were highest for O. macrocentra Engelmann, followed by O. engelmannii var. engelmannii, and lowest for Consolea rubescens (Salm-Dyck ex de Candolle) Lemaire

(Table 2.1). Host taxa differed significantly in epidermal toughness (Table 2.2; F13, 70 =

19.07, p < 0.0001), cladode number (Table 2.2; F13, 70 = 9.93, p < 0.0001), and height

(Table 2.2; F13, 70 = 17.10, p < 0.0001). Consolea rubescens was significantly different in epidermal toughness from all other taxa except Nopalea cochenillifera (L.) Salm-Dyck.

Regression of number of eggs per eggsticks on plant characteristics did not show an effect (r2 = 0.0008, 0.0041, and 0.0036; p = 0.82, 0.60, and 0.62, for toughness, cladode number and height, respectively).

Dispersion statistics from the generalized linear model indicated that the model was a good fit for the data (deviance/df = 0.9743, Pearson Χ2/df = 0.8809, log likelihood

= −76.6720). The Type III log likelihood test showed that number of cladodes and

2 spininess were significant predictors of number of eggsticks (Χ 1 = 9.3699, p = 0.0022

2 and Χ 7 = 71.8413, p < 0.0001, respectively), while height and epidermal toughness were not. This was confirmed by using AIC to compare between all subsets of models. AIC values ranged from 172.14 to 233.65; the model that included only the effects of cladode number and spininess had the lowest AIC value (df = 8, p < 0.0001). The predicted mean

(±SE) number of eggsticks from this model was highest for a spininess score of 5, followed by a score of 3 (1.097 ± 0.24 and 0.359 ± 0.22, respectively).

Larval performance experiment

Overall mean (±SE) pupation rates and adult emergence rates were low (0.22 ± 0.03 and

0.11 ± 0.02, respectively) and were significantly different between taxa (F13, 56 = 3.99, p

< 0.001 and F13, 56 = 3.97, p < 0.001, respectively). The highest pupation rate was found on C. rubescens, followed by O. streptacantha (Table 2.3). Consolea rubescens also had the highest adult emergence rate, followed by O. dillenii (Ker Gawler) Haworth (Table

2.3). Opuntia macrocentra Engelmann had zero percent pupation and O. engelmannii var. lindheimeri (Engelmann) Parfitt and Pinkava and C. spinosior had emergence rates of less than one percent (Table 2.3). Opuntia macrocentra had significantly different pupation and emergence rates from C. rubescens, O. ficus-indica, and O. dillenii.

Consolea rubescens had significantly different emergence rates from C. spinosior, O. engelmannii var. lindheimeri, O. microdasys, and O. santa-rita.

Overall mean (±SE) development time (hatch to adult emergence) was 75.9 ± 1.3 days. Overall mean (±SE) durations of the larval and pupal stages were 54.7 ± 1.3 days and 21.6 ± 1.0 days. Total development time and duration of the larval stage were significantly different among taxa (F10, 26 = 2.35, p = 0.039 and F10, 26 = 9.165, p < 0.001, respectively). Total development time was least on O. microdasys (Lehmann) Pfeiffer

(although only 2% of larvae survived to adult emergence on this species), followed by C. rubescens, and greatest on O. engelmannii var. lindheimeri (Table 2.3). Opuntia engelmanni var. lindheimeri was significantly different from O. dillenii and C. rubescens for total development time, and was significantly different from all taxa except the other two O. engelmanni varieties and C. acanthocarpa for duration of the larval stage.

The overall mean (±SE) proportion of females was low (0.28 ± 0.04) and was not significantly different among taxa (Kruskal-Wallis H=13.45, df = 9, p = 0.143). Opuntia microdasys, O. engelmannii var. lindheimeri, and C. acanthocarpa produced only males

(Table 2.3). Male and female weights were significantly different among taxa (F12, 214 =

5.78, p < 0.0001 and F9, 150 = 2.42, p = 0.014, respectively), but number of eggs per female was not different among taxa (F9, 150 = 1.64, p = 0.109). Number of eggs was significantly correlated with female weight (rS = 0.52, p < 0.0001), but was not significantly correlated with development times. Consolea rubescens had the heaviest females and the highest numbers of eggs per female, followed by O. streptacantha

(Table 4). Opuntia santa-rita (Griffiths and Hare) Rose had both the lightest females and the lightest males (Table 2.4).

Regressions of pupation rate, emergence rate, number of eggs per female, duration of larval stage, and total development time on epidermal toughness were significant, but r2 values were low overall, indicating that untested factors have considerable influence on the response variables (Table 2.5).

Table 2.1. Eggstick and spine data for the fourteen opuntioid taxa. Columns show proportion of eggsticks, mean number of eggsticks (± SE), mean number of eggs per eggstick (± SE), glochid and spine scores and summed spininess score

Proportion Number of Number of eggs per Glochid Spine Spininess Taxa of eggsticksa eggstick (n)b score score scorec eggsticks

Consolea rubescens 0.0294 0.33 ± 0.21 36.0 ± 3.0 (2) 1 0 1

Cylindropuntia 0 0 -- 3 4 7 acanthocarpa

Cylindropuntia spinosior 0 0 -- 3 3 6

Nopalea cochenillifera 0.0735 0.83 ± 0.31 49.4 ± 6.8 (5) 0 0 0

Opuntia dillenii 0.1471 1.67 ± 0.33 47.7 ± 4.5 (10) 3 0 3

Opuntia engelmannii var. 0.1618 1.83 ± 0.65 57.5 ± 7.0 (11) 2 2 4 engelmannii

Opuntia engelmannii var. 0.0147 0.17 ± 0.17 -- 2 2 4 lindheimeri

Opuntia engelmannii var. 0.2500 2.83 ± 0.79 48.0 ± 3.3 (17) 2 3 5 linguiformis

Opuntia ficus-indica 0.0588 0.67 ± 0.49 56.0 ± 10.7 (4) 2 0 2

Opuntia leucotricha 0 0 -- 3 4 7

Opuntia macrocentra 0.1176 1.33 ± 0.49 60.5 ± 7.4 (8) 2 1 3

Proportion Number of Number of eggs per Glochid Spine Spininess Taxa of eggsticksa eggstick (n)b score score scorec eggsticks

Opuntia microdasys 0.0147 0.17 ± 0.17 -- 4 0 4

Opuntia santa-rita 0.1324 1.5 ± 0.50 53.9 ± 4.7 (9) 3 1 4

Opuntia streptacantha 0 0 -- 2 4 6

a n = 6 for all taxa. b Mean number of eggs per eggstick is shown only for those taxa receiving ≥2 eggsticks. c The categorical variable spininess, used in a generalized linear model to predict oviposition by C. cactorum, is the sum of a particular taxa‟s glochid density score and its spine density score. Scores were assigned prior to the experiment and were based upon descriptions in Anderson (2001) and personal observation.

Table 2.2 Mean values (± SE) for recorded plant characteristics

Height Number of Epidermal toughness Species/variety (cm)a cladodesa (g/mm2)a Consolea rubescens 46.8 ± 3.6 7.2 ± 1.7 330.40 ± 16.41 Cylindropuntia acanthocarpa 59.6 ± 2.1 17.3 ± 2.6 583.53 ± 40.49 Cylindropuntia spinosior 69.8 ± 8.2 15.4 ± 2.4 675.33 ± 65.09 Nopalea cochenillifera 34.4 ± 3.7 9.8 ± 1.6 417.60 ± 12.52 Opuntia dillenii 32.1 ± 2.2 9.1 ± 0.9 536.40 ± 29.27 Opuntia engelmannii var. 49.7 ± 2.8 10.1 ± 1.1 787.2 ± 32.29 engelmannii Opuntia engelmannii var. 40.0 ± 1.4 3.6 ± 0.6 979.07 ± 38.57 lindheimeri Opuntia engelmannii var. 71.3 ± 5.3 8.5 ±1.1 732.33 ± 36.26 linguiformis Opuntia ficus-indica 59.3 ± 4.7 6.3 ± 0.7 709.8 ± 15.17 Opuntia leucotricha 52.0 ± 4.5 8.5 ± 0.9 596.73 ± 25.87 Opuntia macrocentra 25.7 ± 1.4 2.9 ± 0.3 959.80 ± 72.31 Opuntia microdasys 24.0 ± 2.1 6.8 ± 1.7 535.73 ± 23.12 Opuntia santa-rita 40.2 ± 6.1 10.6 ± 1.6 775.40 ± 97.19 Opuntia streptacantha 86.0 ± 3.7 7.3 ± 0.6 685.27 ± 58.44 a n = 6 for all taxa

Table 2.3 Results from the no choice larval performance experiment. Columns show pupation and emergence rates (mean ± SE), development times (mean number of days ± SE), and proportion of females (mean ± SE).

Emergence Hatch to Hatch to Proportion of Species/variety Pupation ratea ratea pupation (n)b emergence (n)b females (n)b Consolea rubescens 0.471 ± 0.158 0.353 ± 0.126 48.7 ± 1.2 (5) 70.6 ± 1.8 (5) 0.43 ± 0.15 (5) Cylindropuntia acanthocarpa 0.066 ± 0.031 0.038 ± 0.017 58.9 ± 2.4 (3) 75.3 ± 0.9 (3) 0 (3) Cylindropuntia spinosior 0.068 ± 0.062 0.009 ± 0.008 ------Nopalea cochenillifera 0.342 ± 0.089 0.132 ± 0.082 56.7 ± 2.5 (5) 75.7 ± 2.6 (5) 0.18 ± 0.12 (5) Opuntia dillenii 0.445 ± 0.090 0.284 ± 0.082 43.6 ± 1.2 (5) 71.4 ± 6.5 (5) 0.36 ± 0.06 (5) Opuntia engelmannii var. 0.412 ± 0.156 0.220 ± 0.083 61.3 ± 1.8 (3) 81.3 ± 1.4 (3) 0.46 ± 0.15 (3) engelmannii Opuntia engelmannii var. 0.025 ± 0.014 0.008 ± 0.004 76.7 ± 0.3 (2) 93.7 ± 3.7 (2) 0 (2) lindheimeri Opuntia engelmannii var. 0.131 ± 0.076 0.068 ± 0.051 57.5 ± 0.6 (2) 83.8 ± 4.8 (2) 0.27 ± 0.38 (2) linguiformis Opuntia ficus-indica 0.460 ± 0.100 0.271 ± 0.086 53.1 ± 2.0 (5) 74.8 ± 2.3 (5) 0.29 ± 0.11 (5) Opuntia leucotricha 0.081 ± 0.014 0.025 ± 0.011 55.6 ± 4.6 (5) 79.8 ± 3.5 (3) 0.33 ± 0.17 (3) Opuntia macrocentra 0 0 ------Opuntia microdasys 0.074 ± 0.042 0.021 ± 0.013 49.2 ± 0.8 (2) 70.1 ± 1.9 (2) 0 (2) Opuntia santa-rita 0.049 ± 0.044 0.039 ± 0.035 ------Opuntia streptacantha 0.464 ± 0.175 0.128 ± 0.072 53.3 ± 0.2 (3) 75.3 ± 0.7 (2) 0.64 ± 0.17 (2) a n = 5 for all taxa. b Development times and proportion of females are only shown for taxa that had adults emerge from ≥2 replicates.

Table 2.4 Adult weights and number of eggs per female (mean ± SE)

Male weights Females weights Number of eggs per Species/variety (mg) (n)a (mg) (n)a female (n)a Consolea rubescens 19.58 ± 1.11 (45) 46.1 ± 3.1 (41) 60.8 ± 3.9 (41) Cylindropuntia acanthocarpa 10.09 ± 1.38 (9) -- -- Cylindropuntia spinosior 14.00 ± 4.12 (5) 34.2 ± 11.3 (3) 30.0 ± 10.6 (3) Nopalea cochenillifera 15.06 ± 2.61 (14) 32.3 ± 4.4 (13) 43.7 ± 5.3 (13) Opuntia dillenii 19.27 ± 1.24 (39) 43.5 ± 3.0 (28) 51.7 ± 4.2 (28) Opuntia engelmannii var. 16.18 ± 1.09 (30) 37.2 ± 3.9 (23) 44.7 ± 4.4 (23) engelmannii Opuntia engelmannii var. 22.00 ± 5.70 (2) -- -- lindheimeri Opuntia engelmannii var. 32.48 ± 5.64 (9) 28.6 ± 5.7 (8) 42.6 ± 6.3 (8) linguiformis Opuntia ficus-indica 15.12 ± 0.81 (42) 40.9 ± 4.3 (17) 54.1 ± 7.6 (17) Opuntia leucotricha 30.72 ± 11.81 (6) 44.9 ± 11.8 (3) 43.3 ± 8.7 (3) Opuntia macrocentra ------Opuntia microdasys 16.85 ± 1.83 (6) -- -- Opuntia santa-rita 10.06 ± 1.40 (8) 21.9 ± 2.5 (3) 43.3 ± 1.3 (3) Opuntia streptacantha 20.08 ± 2.69 (12) 21.9 ± 2.5 (21) 59.6 ± 6.8 (21) a Data were pooled among the six replicates. Means are only shown for those taxa that produced ≥2 adults of the respective gender.

Table 2.5 Results from regressions of larval performance variables on epidermal toughness of the fourteen tested opuntioids

Variable Coefficient r2 F

Pupation rate -0.0008 ± 0.0002 0.1665 13.58*** Emergence rate -0.0005 ± 0.0001 0.1379 10.87** Duration of larval stage 0.0194 ± 0.0064 0.1983 9.15** Duration of pupal stage 0.0024 ± 0.0055 0.0051 0.19ns Total development time 0.0218 ± 0.0062 0.2519 12.45** Number of eggs -0.0023 ± 0.0007 0.0608 10.22** Female weights -0.0009 ± 0.0006 0.0142 2.27ns Male weights -0.0002 ± 0.0002 0.0094 2.12ns

** Significant at p < 0.01. *** Significant at p < 0.001 ns Non-significant

Discussion Plant structure, both in the form of spine density and cladode number, had a clear effect on oviposition preference of female C. cactorum moths on the North American opuntioids tested here. Taxa with intermediate spininess ratings (3-5 on a scale of 0-7) received the most eggsticks. Females may prefer to “camouflage” their eggsticks on the tips of spines, rather than lay them directly on the cladode. At the same time, too many spines and glochids interfere with the process of landing, probing the surface of the cladode, and oviposition (Stange 1997). Plants with moderately high numbers of cladodes received more eggsticks than those with fewer cladodes or very high numbers of cladodes. However, cladode number was not separated from true biomass, or controlled for within species, so our conclusions about the effect of plant size on oviposition behavior are tentative. Past studies have shown an inconsistent response to plant size by C. cactorum females (Myers et al. 1981, Robertson 1987, Johnson and

Stiling 1998, Tate et al. 2009). In South Africa, females have been shown to prefer oviposition sites that are sheltered from wind (Robertson 1987), so some of the cladode effect found in our study could be due to the fact that higher numbers of cladodes offer sites that are better sheltered from wind. However, we would also therefore expect taxa with the most cladodes to have the highest number of eggsticks, which we did not find.

This inconsistency was most likely driven by the two cylindropuntia species, which had more cladodes than the platyopuntia species used in the study. Clearly, oviposition preferences cannot be explained by external plant characteristics alone. It is likely that volatile organic compounds (VOCs) play a large role in determining interspecific oviposition preferences of C. cactorum, as female moths have been shown to respond to several VOCs, including nonanal, an alkyl aldehyde produced by O. stricta (Pophof et 46 al. 2005). More work is needed to identify the exact compounds to which C. cactorum females are attracted.

Epidermal toughness was a good predictor of most larval fitness parameters, and it is interesting to note that the two toughest hosts, O. engelmannii var. lindheimeri and

O. macrocentra, were also the two species with the lowest pupation and emergence rates.

Although pre-penetration mortality was not quantified, larvae on these two species often failed to penetrate the cladode, and perished after getting stuck in the mucilage secreted from their attempts. It seems likely that the apparently thicker mucilage, along with epidermal toughness, contributed to the low pupation rates found on O. engelmannii var. lindheimeri and O. macrocentra.

This study confirms the finding of past studies that members of the genus

Cylindropuntia are likely to be at less risk of C. cactorum attack than other opuntioid species (Dodd 1940, Zimmermann et al. 2004). Neither C. acanthocarpa nor C. spinosior received any eggsticks, and both proved to be submarginal hosts. However, overall, our findings did not support the idea that ovipositing females prefer host species that maximize larval fitness. The two taxa most preferred by females, O. engelmannii vars. linguiformis and engelmannii, were not superior hosts for larvae. Out of the fourteen opuntioids, they ranked seventh and fourth, respectively, for percent emergence and had the second and third longest development times. Additionally, O. streptacantha received no eggsticks over the course of a year in the field, yet ranked second for pupation rate, female weights, and number of eggs, and had the highest proportion of emerging females in the laboratory. Of the opuntioids tested, the best overall host for larvae appeared to be C. rubescens, although it received less than half of the expected

47 proportion of eggsticks laid in the field (3.1 ± 0.02% observed versus 7.14% expected if eggsticks were distributed evenly among the tested species).

Cactoblastis cactorum neonates generally enter the cladode upon which the eggstick was laid, and therefore larval host choice in a natural setting is dictated by ovipositional behavior. However, in no choice experiments, larvae have been shown to develop normally on some species that are not selected by females (Dodd 1940), and females have been shown to prefer species that do not maximize fitness. For example,

Johnson and Stiling (1996) found that when given a choice of Florida cactus species, larval preference seemed to reflect ovipositional preference; both groups preferred

Consolea corallicola over three other native Florida species. However, survivorship was significantly lower on C. corallicola than on the less preferred hosts O. stricta and O. humifusa. Mafokoane et al. (2007) found a similar disparity between preference and performance; C. cactorum larvae on O. imbricata had the lowest survivorship, yet the second highest number of eggsticks out of the six species tested.

It is often shown that explicit tests of the preference-performance hypothesis do not support its expectation (Thompson 1988a, Mayhew 1998, Forister 2004, Forister et al. 2009), and a variety of reasons have been proposed to explain this lack of support.

These include mechanistic and phylogenetic constraints (Courtney and Kibota 1990,

Jaenike 1990, Mayhew 1997), the exclusion of the effects of adult behavior and natural enemies from the models (Thompson 1988b, Scheirs et al. 2000, Ballabeni et al. 2001a,

Forister et al. 2009), a lack of distinction between host plants that provide protection versus those that provide nutrition (Ballabeni et al. 2001b), and lack of sufficient time for adaptation to have taken place (Mayhew 1997, Mayhew 2001, Forister et al. 2009).

Cactoblastis cactorum adults do not feed, and the larvae feed internally, therefore adult behavior and natural enemies of larvae are not predicted to be influential factors of oviposition behavior. A more likely explanation for the discrepancies between preference and performance is the lack of a co-evolutionary history between C. cactorum and its ever-growing assemblage of “new association” hosts. It seems that when encountering novel hosts, C. cactorum mothers do not always “know best” (Courtney and Kibota

1990). In the case of C. cactorum’s potential invasion of the southwestern United States and Mexico, we can only draw tentative conclusions about the opuntioids that will be at highest risk, namely O. streptacantha, in the event that larval fitness drives oviposition preference, and O. engelmannii vars. engelmannii and linguiformis, in the event that oviposition preference drives larval adaptation. The best overall host from our study, C. rubescens, is native to the Caribbean and West Indies, where C. cactorum is known to kill seedlings and small plants (Zimmermann et al 2005). It is not native to the southwestern United States or Mexico, but is naturalized in some areas, including

Arizona, Nevada, and California. Although clearly not a conservation priority, naturalized populations of C. rubescens in the southwest could help sustain C. cactorum populations in areas with other, native hosts. The second best host from our study, O. streptacantha, is native to northern and central Mexico and is widely cultivated for its fruits (tunas) and cladodes (nopales) (Zimmermann et al 2001). Although ignored by ovipositing females in this study, O. streptacantha is a known host of C. cactorum and did receive C. cactorum eggsticks in recent caged choice trials of three different groups of cacti (Tate et al. 2009). Additionally, the superior performance of C. cactorum larvae on O. streptacantha raises concerns about the moth becoming an agricultural pest in

Mexico. The opuntioid most preferred by females, O. engelmannii var. linguiformis, is reportedly native to Texas (Anderson 2001) and is very rare in the wild. It is cultivated in the United States for use in landscaping because of its uniquely shaped pads. There are therefore both conservation and economic risks posed to this species. Although O. engelmannii var. linguiformis had very slow development of larvae and a low proportion of females, the fact that there are so few individuals in the wild suggests they could be particularly at risk. For now, it is our hope that efforts to control C. cactorum will be successful, preventing both predicted and unforeseen consequences of its range expansion to the opuntioid-rich west.

Acknowledgements

We gratefully thank Susan Drawdy (USDA-ARS, Crop Protection and Management

Unit, Tifton, GA) for help in acquiring eggsticks for the larval performance experiment,

Amanda Baker and Mountain States Wholesale Nursery, Litchfield Park, AZ, for providing cactus plants, and Rachael Vistein and Leah Pope for their assistance during the larval experiment.

References Cited

Akaike, H. 1974. New look at statistical model identification. IEEE Trans Automat Contr. AC19: 716-723.

Anderson, E. F. 2001. The cactus family. Timber Press, Portland, OR.

Baker, A. J., and P. Stiling. 2009. Comparing the effects of the exotic cactus-feeding moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae) and the native cactus-feeding moth, Melitara prodenialis (Walker) (Lepidoptera: Pyralidae) on two species of Florida Opuntia. Biol. Invasions.11: 619-624.

Ballabeni, P., M. Wlodarczyk, and M. Rahier. 2001a. Does enemy-free space for eggs contribute to a leaf beetle's oviposition preference for a nutritionally inferior host plant? Funct. Ecol. 15: 318-324.

Ballabeni, P., D. Conconi, S. Gateff, and M. Rahier. 2001b. Spatial proximity between two host plant species influences oviposition and larval distribution in a leaf beetle. Oikos 92: 225-234.

Bennett, F. D., D. H. Habeck, E. S. Delfosse, and R. R. Scott. 1995. Cactoblastis cactorum: A successful weed control agent in the Caribbean, now a pest in Florida? Biological control of weeds, pp. 21-26. In E.S. Delfosse and R.R. Scott (eds.), Proceedings, VIII International Symposium on Biological Control of Weeds, 2-7 February 1992, Lincoln University, Canterbury, New Zealand. CSIRO Publications, East Melbourne, Victoria, Australia.

Carpenter, J. E., S. Bloem, and K. A. Bloem. 2001a. Inherited sterility in Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla. Entomol. 84: 537-542.

Carpenter, J. E., K. A. Bloem, and S. Bloem. 2001b. Applications of F-1 sterility for research and management of Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla. Entomol. 84: 531-536.

Chavez-Ramirez, F., X. G. Wang, K. Jones, D. Hewitt, and P. Felker. 1997. Ecological characterization of Opuntia clones in South Texas: implications for wildlife herbivory and frugivory. J.Prof. Assoc. Cactus. 2: 9-19.

Courtney, S. P., and T. T. Kibota. 1990. Mother doesn't know best: selection of hosts by ovipositing insects, pp. 161-188. In E. A. Bernays (ed.), Insect-plant interactions, vol. 2. CRC Press, Inc., Boca Raton, FL.

Dickel, T. S. 1991. Cactoblastis cactorum in Florida. (Lepidoptera: Pyralidae: Phycitinae). Trop. Lepid. 2: 117-118.

Dodd, A. P. 1940. The biological campaign against prickly pear. Commonwealth Prickly Pear Board Bulletin:1-177. Brisbane, Australia.

Forister, M. L. 2004. Oviposition preference and larval performance within a diverging lineage of lycaenid butterflies. Ecol. Entomol. 29: 264-272.

Forister, M. L., C. C. Nice, J. A. Fordyce, and Z. Gompert. 2009. Host range evolution is not driven by the optimization of larval performance: the case of Lycaeides melissa (Lepidoptera: Lycaenidae) and the colonization of alfalfa. Oecologia. 160: 551-561.

Habeck, D. H., and F. D. Bennett. 1990. Cactoblastis cactorum Berg (Lepidoptera: Pyralidae), a phycitine new to Florida. Florida Department of Agriculture and 51

Consumer Services, Division of Plant Industry Entomology Circular, Gainesville, FL.

Hajek, A. E., L. Butler, S. R. A. Walsh, J. C. Silver, F. P. Hain, F. L. Hastings, T. M. Odell, and D. R. Smitley. 1996. Host range of the gypsy moth (Lepidoptera: Lymantriidae) pathogen Entomophaga maimaiga (Zygomycetes: Entomophthorales) in the field versus laboratory. Environ. Entomol. 25: 709-721.

Hervert, J. J., J. L. Bright, R. S. Henry, L. A. Piest, and M. T. Brown. 2005. Home-range and habitat-use patterns of Sonoran pronghorn in Arizona. Wildl. Soc. Bull. 33: 1.

Hight, S. D., and J. E. Carpenter. 2009. Flight phenology of male Cactoblastis cactorum (Lepidoptera: Pyralidae) at different latitudes in the southeastern United States. Fla. Entomol. 92: 208-216.

Hight, S. D., J. E. Carpenter, S. Bloem, and K. A. Bloem. 2005. Developing a sterile insect release program for Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae): effective overflooding ratios and release-recapture field studies. Environ. Entomol. 34: 850-856.

Jaenike, J. 1978. Optimal oviposition behavior in phytophagous insects. Theor. Popul. Biol. 14: 350-356.

Jaenike, J. 1990. Host specialization in phytophagous insects. Annu. Rev. Ecol. Syst. 21: 243-273.

Johnson, D. M., and P. D. Stiling. 1996. Host specificity of Cactoblastis cactorum (Lepidoptera: Pyralidae), an exotic Opuntia-feeding moth, in Florida. Environ. Entomol. 25: 743-748.

Johnson, D. M., and P. D. Stiling. 1998. Distribution and dispersal of Cactoblastis cactorum (Lepidoptera: Pyralidae), an exotic Opuntia-feeding moth, in Florida. Fla. Entomol. 81: 12-22.

Kruskal, W. H., and W. A. Wallis. 1952. Use of ranks in one-criterion variance analysis. J. Am. Stat. Assoc. 47: 583-621.

Levins, R., and R. Macarthur. 1969. An hypothesis to explain the incidence of monophagy. Ecology. 50: 910-911.

Louda, S. M., D. Kendall, J. Connor, and D. Simberloff. 1997. Ecological effects of an insect introduced for the biological control of weeds. Science 277: 1088-1090.

Louda, S. M., R. W. Pemberton, M. T. Johnson, and P. A. Follett. 2003. Nontarget effects - The Achilles' Heel of biological control? Retrospective analyses to reduce risk associated with biocontrol introductions. Annu. Rev. Entomol. 48: 365-396.

Mafokoane, L. D., H. G. Zimmermann, and M. P. Hill. 2007. Development of Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae) on six North American Opuntia species. Afr. Entomol. 15: 295-299.

Mann, J. 1969. Cactus-feeding insects and mites. Smithsonian Institution, Washington, D.C.

Marti, O. G., and J. E. Carpenter. 2009. Effect of irradiation on the incidence of mating in Cactoblastis cactorum. Fla. Entomol. 92: 159-160.

Mayhew, P. J. 1997. Adaptive patterns of host-plant selection by phytophagous insects. Oikos 79: 417-428.

Mayhew, P. J. 1998. Testing the preference-performance hypothesis in phytophagous insects: lessons from chrysanthemum leafminer (Diptera: Agromyzidae). Environ. Entomol. 27: 45-52.

Mayhew, P. J. 2001. Herbivore host choice and optimal bad motherhood. Trends Ecol. Evol. 16: 165-167.

Myers, J. H., J. Monro, and N. Murray. 1981. Egg clumping, host plant selection and population regulation in Cactoblastis cactorum (Lepidoptera). Oecologia 51: 7- 13.

Perez-Sandi, M. 2001. Addressing the threat of Cactoblastis cactorum (Lepidoptera: Pyralidae), to Opuntia in Mexico. Fla. Entomol. 84: 499-502.

Pettey, F. W. 1948. The biological control of prickly-pear in South Africa. Sci. Bull. Dept. Agric. Union South Africa. 271: 1-163.

Pophof, B., G. Stange, and L. Abrell. 2005. Volatile organic compounds as signals in a plant-herbivore system: electrophysiological responses in olfactory sensilla of the moth Cactoblastis cactorum. Chem. Senses 30: 51-68.

Robertson, H. G. 1987. Oviposition site selection in Cactoblastis cactorum (Lepidoptera): constraints and compromises. Oecologia. 73: 601-608.

Scheirs, J., L. De Bruyn, and R. Verhagen. 2000. Optimization of adult performance determines host choice in a grass miner. Proc. R. Soc. Lond., Ser. B: Biol. Sci. 267: 2065-2069.

Simonson, S. E., T. J. Stohlgren, L. Tyler, W. Gregg, R. Muir, and L. Garrett. 2005. Preliminary assessment of the potential impacts and risks of the invasive cactus moth, Cactoblastis cactorum Berg, in the U.S. and Mexico. Final Report to the International Atomic Energy Agency, Vienna, Austria.

Soberon, J., J. Golubov, and J. Sarukhan. 2001. The importance of Opuntia in Mexico and routes of invasion and impact of Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla. Entomol. 84: 486-492.

Stange, G. 1997. Effects of changes in atmospheric carbon dioxide on the location of hosts by the moth, Cactoblastis cactorum. Oecologia. 110: 539-545.

StatSoft, Inc. 1995. STATISTICA for Windows. StatSoft Incorportated, Tulsa, OK.

StatSoft, Inc. 2010. Electronic Statistics Textbook. StatSoft Incorporated, Tulsa, OK.

Stiling, P. 2002. Potential non-target effects of a biological control agent, prickly pear moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae), in North America, and possible management actions. Biol. Invasions. 4: 273-281.

Stiling, P. 2010. Death and decline of a rare cactus in Florida. Castanea.75: 190-197.

Stiling, P., D. Moon, and D. Gordon. 2004. Endangered cactus restoration: mitigating the non-target effects of a biological control agent (Cactoblastis cactorum) in Florida. Restor. Ecol. 12: 605-610.

Strittmatter, L. I., V. Negron-Ortiz, and R. J. Hickey. 2006. Comparative microsporangium development in male-fertile and male-sterile flowers of Consolea (Cactaceae): when and how does pollen abortion occur? Grana 45: 81- 100.

Tate, C. D., J. E. Carpenter, and S. Bloem. 2007. Influence of radiation dose on the level of F-1 sterility in the cactus moth, Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla. Entomol. 90: 537-544.

Tate, C. D., S. D. Hight, and J. E. Carpenter. 2009. Oviposition preference of Cactoblastis cactorum (Lepidoptera: Pyralidae) in caged choice experiments and the influence on risk assessment of F-1 sterility. Biocontrol Sci. Technol. 19: 317- 333.

Thompson, J. N. 1988a. Evolutionary ecology of the relationship between oviposition preference and performance of offspring in phytophagous insects. Entomol. Exp. Appl. 47: 3-14.

Thompson, J. N. 1988b. Coevolution and alternative hypotheses on insect plant interactions. Ecology. 69: 893-895.

USDA-APHIS-PPQ. 2007. Cactoblastis cactorum activities report for Jan-Feb-Mar 2007. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine.

USDA-APHIS-PPQ. 2009. Cactus moth program annual report, Eastern region. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine.

USDA-ARS, National Genetic Resources Program. Germplasm Resources Information Network - (GRIN) [Online Database].National Germplasm Resources Laboratory, Beltsville, Maryland.URL: http://www.ars-grin.gov/cgi-bin/npgs/html/taxweed.pl (01 October 2011)

Vigueras, A. L., and L. Portillo. 2001. Uses of Opuntia species and the potential impact of Cactoblastis cactorum (Lepidoptera: Pyralidae) in Mexico. Fla. Entomol. 84: 493-498.

Zar, J. H. 1999. Biostatistical analysis. Prentice Hall, Upper Saddle River, N.J.

Zimmermann, H. G., V. C. Moran, and J. H. Hoffmann. 2000. The renowned cactus moth, Cactoblastis cactorum (Lepidoptera: Pyralidae): Its natural history and threat to native Opuntia floras in Mexico and the United States of America. Fla. Entomol. 84: 543-551.

Zimmermann, H. G., S. Bloem, and H. Klein. 2004. Biology, history, threat, surveillance and control of the cactus moth, Cactoblastis cactorum. Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture. International Atomic Energy Agency, Vienna, Austria.

Zimmermann, H. G., M. P. Cuen, and A. Rivera. 2005. The status of Cactoblastis cactorum (Lepidoptera: Pyralidae) in the Caribbean and the likelihood of its spread to Mexico. Report to the International Atomic Agency, Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture and the Plant Health General Directorate, Mexico (DGSVB/SAGARPA) as part of the TC Project MEX/5/029. International Atomic Energy Agency, Vienna, Austria.

CHAPTER 3

ASSOCIATIONAL EFFECTS OF CHAMAECRISTA FASCICULATA ON ANT

PREDATION OF CACTOBLASTIS CACTORUM AND HERBIVORY OF OPUNTIA

Synopsis

The legume Chamaecrista fasciculata attracts ants to its extrafloral nectar (EFN) which can lead to reduced herbivory and increased fecundity for the plant. In Florida, Opuntia stricta and O. humifusa, hosts of the invasive moth Cactoblastis cactorum, are often found growing in close association with C. fasciculata. We tested the hypothesis that O. stricta and O. humifusa individuals have higher ant abundance, lower levels of herbivore damage, and increased growth when growing in close association with C. fasciculata compared with individuals not growing near the plant. We also experimentally placed C. cactorum eggsticks and pupae on Opuntia individuals to see if ant predation of these stages occurred, and if so, whether predation rates were higher on individuals growing close to C. fasciculata. Opuntia plants near C. fasciculata were less likely to be attacked by C. cactorum and had higher ant abundance than plants far from C. fasciculata. Field surveys also showed that Opuntia plants near C. fasciculata had a lower proportion of cladodes with C. cactorum damage of any type. Proportions of cladodes with damage from five native herbivores were not significantly different between treatments. In addition, Opuntia individuals growing near C. fasciculata added proportionately more

56 cladodes during the growing season. We found evidence of ant predation on 15.9% of C. cactorum eggsticks and 17.6% of pupae. In August and October of 2008, there was significantly more evidence of predation on eggs and pupae placed on Opuntia individuals near C. fasciculata. No effect of distance to C. fasciculata was seen in

November of 2008, potentially because plants were no longer producing EFN at this time. Our finding that Opuntia plants close to C. fasciculata show reduced herbivory from invasive C. cactorum, but not from the native herbivores examined, suggests that patterns of associational resistance may be influenced by the co-evolutionary history of the organisms in question.

Introduction

The susceptibility of a plant to herbivory can be affected by the characteristics of its plant neighbors. Neighboring plants can increase or decrease the probability of a target plant being detected and consumed by an herbivore. The former phenomenon is termed associational susceptibility (hereafter, AS) (Brown and Ewel 1987; Wahl and Hay 1995), the latter associational resistance (hereafter, AR) or associational defense (Atsatt and

Odowd 1976; Hay 1986; Tahvanainen and Root 1972). AS and AR have been demonstrated in both terrestrial (Brown and Ewel 1987; Finch et al. 2003; Hamback et al. 2000; Karban 1997; Russell et al. 2007; Stenberg et al. 2007; Thomas 1986; White and Whitham 2000) and marine (Gagnon et al. 2003; Hay 1986; Pfister and Hay 1988;

Wahl and Hay 1995) habitats. Several mechanisms have been invoked to explain these plant-plant interactions, including the repellant-masking plant hypothesis (Atsatt and

Odowd 1976; Tahvanainen and Root 1972), the attractant-decoy plant hypothesis (Atsatt

57 and Odowd 1976), the resource concentration hypothesis (Root 1973), and the enemies hypothesis (Root 1973). The concepts of AR and AS have also been shown among herbivore species (sometimes termed apparent competition), where they are often mediated by shared natural enemies (Barbosa and Caldas 2007; Hamback et al. 2006;

Holt and Lawton 1994; Settle and Wilson 1990; Stiling et al. 2003). Oviposition preference (Shiojiri et al. 2002), differences in plant chemistry (Redman and Scriber

2000; Shiojiri et al. 2001), and habitat modification (White and Andow 2006) have also been shown to drive AR/AS among herbivore species. Various AR/AS mechanisms, and their effects on plant and herbivore populations, have been the subject of several reviews

(Agrawal et al. 2006; Andow 1991; Barbosa et al. 2009; Hamback and Beckerman 2003;

Milchunas and Noy-Meir 2002; Russell 1989; Sheehan 1986) and therefore only the enemies hypothesis, which the present study tests, will be detailed here.

The enemies hypothesis predicts that a higher availability of alternative energy sources (e.g., nectar, pollen, or alternative prey) can lead to higher densities of predators and parasites in polycultures, as compared with monocultures (Root 1973). For these

“insectary plants” (Atsatt and Odowd 1976) to actually provide AR, their alternative energy sources must increase predator and/or parasitoid efficiency, longevity, or herbivore encounter rate, leading to increased herbivore mortality and a subsequent decrease in damage on the target plant. Support for the enemies hypothesis comes from both agricultural (Harmon et al. 2000; Mathews et al. 2007; Spellman et al. 2006) and natural systems (Stenberg et al. 2007; Stiling et al. 2003). Harmon et al (2000) found that density of common dandelion, Taraxacum officinale, in alfalfa fields was negatively correlated with density of pea aphid, Acyrthosiphon pisum, and positively correlated with

58 density of the lady bird beetle Coleomegilla maculata. Caged laboratory experiments showed that C. maculata predation of A. pisum was twice as high in alfalfa-dandelion patches compared with alfalfa only patches. The authors posit that dandelion pollen provides an alternative energy source for C. maculata, increasing tenure time in alfalfa patches with higher dandelion densities and potentially increasing control of pea aphids across generations. Such studies show that understanding the effects of plant neighbors on natural enemies can lead to more effective control of pest herbivore species.

Extra-floral nectar (hereafter, EFN) is an important alternative energy source for natural enemies of herbivores, particularly ants. Numerous studies have shown a protective benefit of EFN-tending ants to plants in natural systems (Barton 1986; Bentley

1977; DelClaro et al. 1996; Heil and McKey 2003; Janzen 1966; Koptur 1984; Pickett and Clark 1979; Tilman 1978). In addition, conservation biological control studies have demonstrated that interplanting crops with species or cultivars that produce EFN can increase the performance of biological control agents (Lewis et al. 1998; Limburg and

Rosenheim 2001; Mathews et al. 2007). EFN-producing plants have also been studied in the context of invasive ant-plant interactions, which have the potential to either benefit or harm the plant (Fleet and Young 2000; Lach 2003; Ness 2003; Stiles and Jones 2001).

For example, it has been suggested that EFN-producing plants can “fuel” the growth of invasive ant populations, thereby reducing local native ant diversity (Ness and Bronstein

2004; Savage et al. 2009). However, to our knowledge, there have been no studies documenting EFN-mediated AR to an invasive herbivore in a natural setting.

Cactoblastis cactorum, the cactus moth, is an invasive pest in coastal regions of the southeast US. Ant predation of C. cactorum has been documented in its native South

America (Lobos and de Cornelli 1997) and in areas where it was released for biological control of Opuntia spp. (e.g. Australia: Dodd 1940; Hawaii: Fullaway 1954; South

Africa: Pettey 1947; Robertson and Hoffman 1989). However, there has been almost no investigation into the potential for ant predation of C. cactorum in the southeast US (but see Bennett and Habeck 1996 and Miller et al. 2010). In this region the C. cactorum hosts Opuntia stricta and O. humifusa frequently co-occur with the legume

Chamaecrista fasciculata, which attracts ants to its EFN. Such a system provided us the opportunity to answer the following questions: (1) In Florida, are any ant species preying upon C. cactorum eggs or pupae? (2) If so, what are the predation rates? (3) Do Opuntia spp. receive AR from C. fasciculata via ants collecting its EFN? We predicted that evidence of ant predation would be found for both egg and pupal stages (the larval stage was not examined because C. cactorum feed internally as larvae). We also predicted that

C. fasciculata would provide AR to Opuntia spp. growing in close association, and tested this prediction with damage and growth surveys of Opuntia individuals and manipulative ant predation experiments.

Study system and methods

Insects

Cactoblastis cactorum is a cactus-feeding, pyralid moth native to northern Argentina,

Paraguay, Uruguay, and southern Brazil. It has been released intentionally in a number of countries, most notably Australia, to control pest Opuntia species, and was discovered in

Florida in 1989. Cactoblastis cactorum has since spread along the Atlantic and Gulf coasts of the southeastern United States. Females lay about 40-100 eggs one on top of the other to form an eggstick that is attached to the tip of a cactus spine or directly to a cladode. The larvae feed internally on a number of opuntioid species, during which time they relatively protected from predation. Larval damage results in hollowed out cladodes and can lead to secondary infections and in some cases, the death of entire plants

(Zimmermann et al 2004). After completing six instars, they drop to the ground to pupate in the soil or hollowed out cladodes. In central Florida, there are three non-overlapping flight periods per year: a spring flight beginning in mid- February, a summer flight beginning in early June, and a fall flight beginning in late August (Hight and Carpenter

2009).

Five native species of Opuntia-feeding insects were included in damage surveys

(Table 3.1). All of these species are specialists on opuntioids, with the exception of

Diaspis echinocacti, which feeds of a variety of cactus species (Hunter 1912; Mann

1969). Dactylopius confusus and D. echinocacti are primarily sessile and are therefore easily quantified. The remaining three species leave characteristic signs of damage

(Table 3.1), allowing damage levels to be assessed even when insects are not directly observed. The ant species Monomorium minimum has been reported as an egg predator of

Chelinidea vittiger (Hamlin 1924). Pickett and Clark (1979) found that C. vittiger individuals spent less time on Opuntia acanthocarpa plants occupied by Crematogaster opuntiae; they also observed C. opuntiae attack and kill C. vittiger nymphs. The remaining four insect species included in damage surveys have no accounts in the literature of interactions with ants (either through predation or tending behavior), nor were any such interactions observed during the course of the study.

Plants

Two Opuntia species native to Florida were used in this study, O. stricta and O. humifusa. Both are common hosts of C. cactorum, although O. stricta is attacked more often than O. humifusa (Baker and Stiling 2009). They exhibit slightly different morphologies; O. stricta has a more erect growth form and more spines, while O. humifusa has a low, spreading growth form and fewer spines. Both species produce EFN at the areoles of newly developing cladodes and flower buds, which are mainly present in

April and May in the Tampa Bay area (H. Jezorek pers. obsv.). EFN is produced both day and night (Oliveira et al. 1999); data on the volume and composition of EFN produced by O. stricta and O. humifusa were not found in the literature and were not collected in the present study. Studies from the native range of O. stricta recorded 16 ant species, mainly belonging to the subfamilies Myrmicinae and Formicinae, associated with its EFN (Oliveira et al. 1999; Miller at al. 2010; Robbins and Miller 2009). To the best of our knowledge there are no studies of ant-EFN associations on O. humifusa.

Chamaecrista fasciculata is a native legume that is common throughout the mid- western and eastern regions of North America. It co-occurs with O. humifusa throughout 62 most of its range and co-occurs with O. stricta in coastal regions of the southeast and

Gulf coast states. Chamaecrista fasciculata produces EFN at the base of each leaf petiole and EFN is produced both day and night throughout the growing season (Kelly 1986).

Plants can have over 200 active nectaries and secrete up to 3 µl of EFN per day (Rutter and Rausher 2004). Ants collect EFN from two weeks after plant emergence to pod senescence (Barton 1986). In central Florida, this period lasts from mid-February to late

October or early November (H. Jezorek pers. obsv.). At least 28 ant species have been recorded collecting EFN from C. fasciculata (Barton 1986; Boecklen 1984; Kelly 1986;

Rios et al. 2008; Stiles and Jones 2001); the vast majority of these species are found in the subfamilies Myrmicinae (14 species) or Formicinae (10 species). Two studies performed in northern Florida found the most common visiting ants to be Forelius (=

Iridomyrmex) pruinosus, Camponotus floridanus, and Crematogaster ashmeadi (Barton

1986; Boecklen 1984).

Damage and growth surveys

Surveys were conducted at Honeymoon Island (hereafter, HI) and Caladesi Island

(hereafter, CI) State Parks (Pinellas County, FL). Both parks lie within Florida‟s west- central barrier island chain and are separated by approximately 4.5 km. Opuntia stricta,

O. humifusa, and C. fasciculata are common in the sand dune and coastal scrub habitats of these islands.

Four surveys of naturally occurring Opuntia were conducted: Aug-2008 at HI

(n=42 plants) and Nov-2008, May-2009, and Sept-2009 at CI (n=28, 40, and 42, respectively). Surveys were conducted between 0900 and 1200. For each survey, Opuntia

63 plants were chosen haphazardly such that half were ≥3 m from any C. fasciculata individual (“far from” category) and half were ≤ 0.25 m from a C. fasciculata individual

(“near to” category). Distances were measured from the center of each plant, with the result being that the “near to” Opuntia were usually touching C. fasciculata leaves, i.e. vegetation bridges were present. Each survey represented a new sample of plants; distances between the surveyed plants varied, but were ≥5 m. The HI survey included both O. stricta (n=16) and O. humifusa (n=26) while the CI surveys included only O. humifusa. The following parameters were measured for each Opuntia individual during each survey: number of primary, secondary, and tertiary cladodes, number of cladodes with old C. cactorum damage, number of cladodes with active, i.e. feeding, C. cactorum larvae, number of cladodes with C. cactorum eggsticks, number of cladodes with native insect damage (for each of five species in Table 3.1), and ant activity. Ant activity was measured by counting the number of ants on a plant during a one minute period and then dividing by the total number of cladodes. Most plants had 1-2 ant species present during the observation period, therefore relative abundance of ant species was not directly quantified during surveys. Instead, when a survey plant had ants present, one ant per observed species was aspirated immediately after the observation period and transferred to ethanol for off-site species identification (total ants collected=99). Any plant with a C. cactorum eggstick was checked four weeks later for larval entrance holes or active larvae. For the second CI sample (n=40), number of cladodes and height (cm) were measured in April-2009 and again in Oct-2009 in order to estimate growth. Growth was not quantified for the other three samples, as they were not surveyed during the Opuntia growing season.

We conducted separate analyses for each survey, as each represented a unique combination of plants, site, and date, and our questions were not aimed at comparing between surveys, but rather between plant categories within surveys. Data for insect response variables included many zeros and could not be transformed to meet the assumptions of parametric analysis. Therefore, proportions of pads with C. cactorum damage and native insect damage were compared between “near to” and “far from” categories with Mann-Whitney tests for independent samples. For the HI survey, data from O. stricta and O. humifusa were pooled, as Kolmogorov-Smirnov tests found no significant differences in the distribution of insect response variables (P-values ranged from 0.289 to 1.00). Fisher‟s exact tests were used to compare the frequency of plants in each category with 1) any type of C. cactorum damage and 2) active larvae one month after discovery of an eggstick. Growth variables (proportional change in cladode number and change in height from April-2009 to Oct-2009) were compared between “near to” and “far from” categories using single factor ANOVA, as these data met normality and homogeneity of variance assumptions. All analyses were conducted in STATISTICA 5.5

(StatSoft 2000).

Predation Experiment

Cactoblastis cactorum eggsticks and pupae, obtained from the USDA-ARS Crop

Protection and Management Research Unit in Tifton, GA, were irradiated at 200 Gy with a Cobalt 60 Gamma cell 220 irradiator. Irradiation at this level results in sterility of any

F1 progeny (Carpenter et al. 2001; Tate et al. 2007) and was taken as a precaution in case of early hatch, early emergence, or unrecovered eggsticks or pupae.

In Aug-2008 and Oct-2008, eggsticks containing an average of 56.1(standard error ±1.2) eggs were placed on Opuntia plants (n=42) at HI, with 3-6 eggsticks per plant. These months correspond to the summer and fall flights, respectively, of C. cactorum at similar latitude in Florida (Hight and Carpenter 2009). The same plants (16

O. stricta and 26 O. humifusa) and categories (“near to” and “far from” C. fasciculata) from the Aug-2008 damage and growth survey were used, but the survey was completed prior to placement of experimental eggsticks or pupae. We were unable to replicate the usual oviposition behavior of C. cactorum females, i.e. eggsticks laid on the tip of cactus spines. Instead, eggsticks were placed on the margins of secondary Opuntia cladodes by inserting their tips into small holes created with a dissecting probe. This method resulted in the first 3-5 eggs of the eggstick being inaccessible to predators, as mucilage from the created hole dried around the end of the eggstick to hold it in place. The location of eggsticks was marked to enable recovery after 14 days in the field. In Sept-2008 and

Nov-2008, when moths from the summer generation are naturally pupating, five pupae were placed around the same Opuntia plants. In order to mimic natural pupation behavior, pupae were covered with sand and/or leaf litter, or placed in dead, hollowed out cladodes at the base of the plant. The location of each pupa was marked to enable

66 recovery after 10-14 days in the field. We recognize that our methods of placing eggsticks and pupae do not perfectly imitate the natural behavior of C. cactorum.

However, using experimentally placed, sterile eggsticks and pupae allowed us to obtain adequate sample sizes; this would be extremely difficult with natural eggsticks and pupae.

Recovered eggsticks and pupae were placed in individual, ventilated vials, taken to the lab, and examined under a dissecting microscope for evidence of predation. They were then returned to ventilated vials and kept at ~25°C and a 14L:10D photoperiod in order to rear out any parasitoids. Eggsticks were placed into one of four groups:

“missing” (not recovered from HI), “intact” (no breaks or rips), “evidence of predation”

(distinct jagged break or eggs ripped open), or “undetermined” (clean break or break not sufficiently jagged to infer predation). Eggsticks placed in the “undetermined” category were excluded from the analysis. Pupae were also placed into one of four groups:

“missing” (not recovered from HI), “intact” (no holes in cocoon and pharate present),

“evidence of predation” (cocoon appeared ripped open; the placement and size of hole in cocoon was atypical for C. cactorum emergence and the pupal case was absent), or

“emerged” (the placement and size of hole in cocoon was typical for C. cactorum emergence and the pupal case usually present).

For each replication of the experiment, proportions of eggsticks or pupae in each group for each plant were compared among categories and species using Mann-Whitney tests for independent samples, as residuals were again highly non-normal. A Fisher‟s exact test was used to compare the frequency of eggsticks or pupae showing evidence of

67 predation between the “near to” and “far from” C. fasciculata treatments. All analyses were conducted in STATISTICA 5.5 (StatSoft 2000).

Results

Damage and growth surveys

The proportion of cladodes with any stage of C. cactorum damage (eggsticks, active larvae, or old damage) was higher on “far from” plants than “near to” plants in all surveys except Nov-2008 (Fig. 3.1a). The difference was significant in Aug-2008 (Mann-

Whitney U=291.0, p=0.035) and Sept-2009 (U= 282.0, P=0.024). The proportion of cladodes with eggsticks was not significantly different between plant categories for any surveys (Fig. 3.1b). The proportion of cladodes with active larvae was higher on “far from” plants than “near to” plants in all surveys except Nov-2008 (Fig. 3.1c), but the difference was only significant for the Sept-2009 survey (U=273.0, P=0.019) when no active larvae were found on “near to” plants. The proportion of cladodes with old C. cactorum damage was higher on “far from” plants than “near to” plants in all surveys

(Fig. 3.1d), with significant differences in Aug-2008 (U=296.0, P=0.008) and May-2009

(U=248, P=0.050). Fisher‟s exact tests revealed that “far from” Opuntia plants had C. cactorum damage significantly more often than expected in Sept-2009 (one-tailed

P=0.022), with marginally significant results from the Aug-2008 and Nov-2008 surveys

(one-tailed P=0.090 and 0.063, respectively). Additionally, “near to” Opuntia plants were significantly less likely than expected to contain active larvae one month after discovery of an eggstick (one-tailed P=0.0048).

Proportions of cladodes damaged by the five native insect species were not significantly different between “near to” and “far from” plants, although a slight trend for higher damage on the “far from” plants was observed (Fig. 3.2a-e). For all surveys, more ants per cladode were found on plants in the “near to” group (Fig. 2f); this result was significant in Aug-2008 (U=130.0, P=0.015), May-2009 (U=125.0, P=0.032), and Sept-

2009 (U=123.5, P=0.012). Six species of ants were identified from the sample collected from survey plants: Monomorium viride, Forelius pruinosus, Solenopsis invicta,

Dorymyrmex bureni, Pseudomyrmex gracilis, and Camponotus planatus (Table 3.2). M. viride, F. pruinosus, and D. bureni are native to Florida, the remaining three species are non-native (Deyrup et al 2000).

Opuntia individuals in the “near to” category added significantly more cladodes than those in the “far from” category (F=4.61, P=0.038); the difference for growth by height was only marginally significant (F=3.28, P=0.078) (Fig. 3.3).

Predation experiment

In the field, ants were frequently observed searching the tips of spines of O. stricta and

O. humifusa plants (H. Jezorek pers. obsv.), but predation of eggsticks was not directly observed. Solenopsis invicta were observed investigating C. cactorum cocoons nearly immediately upon placement at the base of Opuntia plants, prior to concealment with sand or leaf litter, and two S. invicta individuals were found stuck in the silk of a C. cactorum cocoon that was ripped open and emptied.

In Aug-2008, “near to” and “far from” plants differed significantly in the proportion of eggsticks displaying evidence of predation (U= 145, P=0.030) (Fig. 3.4a).

In Oct-2008 this difference was marginal (U=153.5, P=0.062) and the difference in the proportion of intact eggsticks was significant (U=130.0, P=0.020) (Fig. 3.4b). There were no significant differences between O. stricta and O. humifusa in either month.

Results from the Fisher‟s exact test showed that eggsticks placed on “near to” plants showed evidence of predation significantly more often than expected in both Aug-2008 and Oct-2008 (one-tailed P=0.017 and 0.023, respectively). Of the intact eggs recovered,

92.1% hatched and no parasitoids were found.

In Sept-2008, all categories of pupae were significantly different between “near to” and “far from” plants (Intact, U=325, P=0.006; Missing, U=116, P=0.006; Emerged,

U=315, P=0.010; Evidence of predation, U=134, P=0.021) (Fig. 3.4c). No significant differences were found in Nov-2008 (Fig. 3.4d) or between O. stricta and O. humifusa for either month. The Fisher‟s exact test revealed that pupae placed on “near to” plants showed evidence of predation significantly more often than expected in Sept-2008, but not in Nov-2008 (one-tailed P=0.017 and 0.531, respectively).

Five chalcid wasps emerged from pupae placed on “near to” plants and another five emerged from pupae placed on “far from” plants. Three of the wasps were recovered from the Sept-2008 experiment and seven from the Nov-2008 experiment. Nine of the individuals were identified as Brachymeria pschye; the remaining individual was identified as Conura side. For recovered, intact pupae the parasitism rate was 5.29%; for all pupae placed in the field the rate was 2.38%.

Table 3.1: Native Opuntia herbivores found in central Florida and their characteristic feeding damage

Common Insect species Order: Family Description of feeding/damage name Chelinidea vittiger McAtee Hemiptera: Coridae cactus bug mobile sap sucker; leaves pale half or full circles on surface of cladode; high levels of infestation lead to desiccation of cladodes Dactylopius confusus Hemiptera: Cochineal sessile sap sucker; exudes fluffy white Cockerell Dactylopiidae substance on surface of cladodes Diaspis echinocacti Bouché Hemiptera: cactus scale sessile sap sucker; circular, slightly convex, Diaspididae whitish to tan scale; high levels of infestation lead to chlorotic patches and desiccation of cladodes Gerstaeckeria hubbardi Coleoptera: cactus weevil internal feeder; leaves light brown circular LeConte Curculionidae cell; cladode appears pierced by tiny “bullet hole” Marmara opuntiella Busck Lepidoptera: cactus stem internal feeder; pale green to white serpentine Gracillariidae miner mines on cladodes

71 a. Any damage b. Eggsticks

0.16 0.016 ** 0.12 0.012 **

0.08 0.008

0.04 0.004 Proportion of of Proportioncladodes 0 0

c. Larvae d. Old damage 0.12

0.05 ** *** 0.04 0.10 0.08 0.03 ** 0.06 0.02 0.04

0.01 0.02 Proportionof cladodes 0 0.00 Aug. '08 Nov. '08 May '09 Sept. '09 Aug. '08 Nov. '08 May '09 Sept. '09

Figure 3.1: Results from C. cactorum damage surveys. Proportion of cladodes with (a) any type of Cactoblastis cactorum damage, (b) C. cactorum eggsticks, (c) C. cactorum larvae, and (d) old C. cactorum damage for surveyed Opuntia plants near to C. fasciculata (unfilled bars) and far from C. fasciculata (filled bars). For ease of interpretation, bars show mean (+SE), but asterisks represent significant results from a non-parametric Mann-Whitney test (**, P<0.05, ***, P<0.01).

a. C. v. aequoris b. D. confusus

0.12 0.10

0.10 0.08 0.08 0.06 0.06 0.04 0.04

0.02 0.02 Proportion of of Proportioncladodes 0.00 0.00 c. D. echinocacti d. G. hubbardi

0.08 0.25

0.06 0.20

0.15 0.04 0.10 0.02

Proporiton ofcladodesProporiton 0.05

0 0.00

e. M. opuntiella f. Ants

0.0150 0.50 **

0.40 0.0100 0.30 ** ** 0.20 0.0050

Number perNumber cladode 0.10 Proportion of of Proportioncladodes 0.0000 0.00 Aug. '08 Nov. '08 May '09 Sept. '09 Aug. '08 Nov. '08 May '09 Sept. '09

Figure 3.2: Results from native herbivore and ant surveys. Proportion of cladodes with damage from (a) Chelinidea vittiger, (b) Dactylopius confusus, (c) Diaspis echinocacti, (d) Gerstaeckeria hubbardi, and (e) Marmara opuntiella, and (f) number of ants per cladode counted during one minute observation periods for surveyed Opuntia plants near to Chamaecrista fasciculata (unfilled bars) and far from C. fasciculata (filled bars). For ease of interpretation, bars show mean (+SE), but asterisks represent significant results from a non-parametric Mann-Whitney test (**, P<0.05)

Table 3.2: Relative abundance of ant species in a sample (n=99) collected from Opuntia individuals. Collections were from 7 O. stricta and 79 O. humifusa. One ant per observed species was aspirated immediately following a 1 minute observation period.

Ant species Relative abundance Native to Florida? Monomorium viride Browna 0.5960 Yes Forelius pruinosus Roger 0.1616 Yes Solenopsis invicta Buren 0.1414 No Dorymyrmex bureni Trager 0.0505 Yes Pseudomyrmex gracilis Fabricius 0.0303 No Camponotus planatus Rogera 0.0202 No a Species are congeners of known C. cactorum predators in South Africa (Robertson 1988)

b. a. 0.25 7 p=0.038 6 p=0.078 0.2 5 0.15

4 0.1

3 added 0.05 (cm) Growth 2

0 1 Proportion of of Proportioncladodes -0.05 0 Near C.f. Far from C.f. Near C.f. Far from C.f.

Figure 3.3: Results from a growth survey from April-2009 to Oct-2009 of O. humifusa plants (n=40) near to and far from C. fasciculata. a) Mean (+SE) proportional change in cladodes; b) Mean (+SE) change in height.

a. August 2008 b. October 2008

eggsticks

Proportion of of Proportion

a a

c. September 2008 d. November 2008

Proportion of pupae of Proportion

b b b a

Figure 3.4: Results from predation experiment. Proportion of eggsticks (a-b) or pupae (c- d) in each category after collection from Opuntia plants far from Chamaecrista fasciculata (n=21, unfilled boxes) and near to C. fasciculata (n=21, patterned boxes). I= intact, M= missing, E= emerged, PR= evidence of predation. Boxes indicate interquartile ranges, bold lines indicate medians, whiskers indicate inner fences, points indicate outliers, and asterisks indicate extreme values. a = P<0.05, b = P<0.01

Discussion

We found evidence to support our prediction that ants prey upon C. cactorum eggsticks and pupae in Florida. Approximately 16% of our experimental eggsticks and 17.5% of our experimental pupae displayed breaks and/or rips that are consistent with descriptions of ant predation (Dodd 1940; Pettey 1947; Robertson and Hoffmann 1989). The rates of pupal predation we found are within the range reported from the Florida Keys (0- 36%,

Bennett and Habeck 1996) and South Africa (13-34%, Robertson and Hoffmann 1989).

However, substantially higher rates of egg mortality from predation have been reported from both Argentina (30%) (Lobos and de Cornelli 1997) and South Africa (55-78%)

(Robertson and Hoffmann 1989). It should be noted that predation rates from the present study have the potential to be underestimates, for two reasons. First, eggsticks were collected after 14 days and pupae after 10-14 days, although the average duration of these life stages is 30 and 20 days, respectively (USDA-APHIS 2007). They were therefore not vulnerable to predation for the full development period. Second, we were conservative when determining whether specimens showed evidence of predation and it is conceivable that missing eggsticks and/or pupae could have been carried away by ants. Pettey (1947) reported that in some South Africa C. cactorum colonies, up to 59% of eggsticks were removed by ants. Future studies that include experimental manipulation of ants and extended observation periods could reveal that actual predation rates are higher than what we have reported. The lower rates found in our study could also be due in part to the newer association of C. cactorum and ants in Florida than in the moth‟s native range or in

South Africa or Hawaii, where the moth has been present since 1933 and 1950, respectively (Zimmermann et al. 2004).

Of the six ant species identified from our sample, half were non-native species, accounting for about 20% of individuals (Table 3.2). The two most abundant species, M. viride and F. pruinosus, are native to Florida and accounted for >75% of individuals.

Two congeners of M. viride (M. albopilosum and M. minutum) are known egg predators of C. cactorum in South Africa (Robertson and Hoffmann 1989). More extensive surveys and experiments are needed before statements can be made about the relative importance of various ant species to C. cactorum predation in Florida. For example, S. invicta accounted for only 14% of ants in our sample, yet its aggressive nature (Deyrup et al. 2000; Ness 2003; Porter and Savignano 1990) could mean that it accounts for a greater proportion of predation.

The absence of egg parasitism in our recovered eggsticks is consistent with previous findings in Florida, where the only recorded instance comes from two eggsticks parasitized by Trichogramma sp. (Bennett and Habeck 1996). Accounts of egg parasitism from other regions vary widely (Pemberton and Cordo 2001). Approximately 5% of our recovered pupae were parasitized; this agrees with what Bennett and Habeck (1996) found during the same times of year in Florida (10% in September and 6% in November), although their rates for other months were considerably higher (e.g. 30% parasitism in

July). Although both of the parasitoid species we found represent new records for C. cactorum, they have relatively broad host ranges and are therefore not good candidates for inundative biological control.

We found partial support for our prediction that Opuntia plants receive AR when closely associated with C. fasciculata. Results from our Aug-2008, May-2009, and Sept-

2009 surveys provide evidence that Opuntia plants growing close to C. fasciculata may

78 experience less C. cactorum damage than plants far from C. fasciculata (Fig. 3.1) and that this benefit could potentially be attributed to higher ant activity (Fig. 3.2f). However, damage by five native insect species was not affected by distance to C. fasciculata, indicating that AR benefits may not extend to all herbivores. The lack of an effect may be explained by the life histories of the insects surveyed. The two scale insects, D. echinocacti and D. confusus, may be vulnerable to predation during their crawler phases, but nymphs in both species settle and begin producing waxy or wooly coverings, respectively, within 24 hours (Mann 1969; Oetting 1984). Egg, larval, and pupal stages of

Gerstaeckeria hubbardi are passed inside cladodes and adults drop to the ground and feign death if disturbed (Mann 1969; Woodruff 2009). Egg and larval stages of Marmara opuntiella occur beneath the epidermis of a cladode (Hunter et al. 1912). In contrast to these native insects, C. cactorum eggsticks and pupae are accessible to predators for a relatively long duration. The absence of a significant difference in damage by C. vittiger is more difficult to explain, as eggs of this species are known suffer from ant predation

(Hamlin 1924), and at least the early nymph stages would seem to be vulnerable to predation. Adults may be protected from ant predation because of their relatively large size of up to 13mm long and 4.5mm wide (Hamlin 1924).

Miller et al (2010) found that excluding ants from Opuntia spp. did not deter oviposition by C. cactorum or a native cactus-feeding moth, Melitara prodenialis. Low numbers of eggsticks found in our surveys make it difficult to draw definite conclusions, but we also found no evidence that higher ant activity on the “near to” plants affected C. cactorum oviposition. It is interesting to note that eggsticks found on “near to” Opuntia were far less likely to result in larval damage than eggsticks found on “far from” Opuntia.

Again, this effect could be mediated by differences in ant activity, whereby higher ant- eggstick encounter rates occur on Opuntia growing in close to C. fasciculata.

“Near to” Opuntia plants added more cladodes, but not height, than “far from” plants (Fig. 3.3). However, only one set of plants was surveyed for growth and all individuals were O. humifusa. This species has a low, sprawling growth form and proportionate change in cladode number may be a more informative measure of growth than height. The fact that O. humifusa plants far from C. fasciculata actually lost cladodes over the growth season (Fig. 3a) could be due to increased C. cactorum damage, as cladodes hollowed out by larvae often fall off the plant. Vegetative reproduction is extremely common in Opuntia (Anderson 2001). Gimeno and Vila (2002) found that vegetative recruitment accounted for close to 50% of O. stricta juveniles in Spanish olive groves. However, the relative importance of vegetative versus sexual recruitment in

Opuntia is known to vary by habitat (Del Carmen Mandujano et al 1998), so studies that examine both methods are necessary to determine the full effects of plant neighbors and predation on Opuntia fitness. For example, Oliveira et al (1999) found that excluding ants from O. stricta branches translated into a 50% decrease in fruit set. Local environmental factors should also be examined. C. fasciculata fixes nitrogen (Naisbitt et al 1992), so it is possible that differences in soil chemistry contribute to its observed effect on Opuntia growth and insect damage levels.

Our results from the predation experiments provided additional support for our prediction of AR, in that more eggsticks from Opuntia close to C. fasciculata showed evidence of predation. The fact that eggsticks from these same plants had fewer intact eggsticks is also of interest. C. cactorum neonates work together to enter a cladode, often

80 making several attempts to overcome mucilage before successfully penetrating (Jezorek et al. 2010; Robertson and Hoffmann 1989; Zimmermann et al. 2004). As a result of this communal behavior, the loss of even a few eggs of an eggstick could end up being fatal to the entire cohort if they are unable to successfully penetrate the cladode. The minimum number of neonates needed for successful penetration is unknown, but is a potential topic for further study. Results from our Sept-2008 pupal experiment were consistent with the

AR prediction, but distance to C. fasciculata did not have an effect in our Nov-2008 experiment. This could be attributed to the fact that C. fasciculata has generally stopped producing EFN by this point in the year (H. Jezorek, pers. obsv.) meaning that ants are equally likely to encounter C. cactorum on Opuntia plants near to or far from C. fasciculata.

In summary, C. cactorum is very likely subject to ant predation of its egg and pupal stages in Florida, although it is not yet known which species function as the main predator(s). Overall, our results support the enemies hypothesis (Root 1973).

Chamaecrista fasciculata may be acting as an “insectary plant” (Atsatt and Odowd 1976) for nearby Opuntia individuals by “sharing” the anti-herbivore effects of its EFN production and its well-documented attractiveness to predacious ants (Abdala-Roberts and Marquis 2007; Barton 1986; Kelly 1986; Rios et al. 2008; Rutter and Rausher 2004).

What is not clear from our study is whether this AR is due solely to differences in ant abundance, or also to differences in ant species assemblages, which we did not measure.

Although the effects of AR detailed here may only last as long as EFN is produced, i.e. mid-February to October in our study sites, this period overlaps with all three C. cactorum flight periods (Hight and Carpenter 2009), and most of the three pupation

81 periods. Land managers and others concerned with protecting Opuntia from C. cactorum may want to consider adopting practices, such as frequent prescribed burning, that encourage the growth of C. fasciculata (Galloway and Fenster 2000) or other EFN- producing plants.

Acknowledgements

We gratefully thank Susan Drawdy (USDA-ARS, Crop Protection and Management

Unit) for help in acquiring C. cactorum eggsticks and pupae, Mark Deyrup (Archbold

Biological Station) for ant identification, Michael Gates (USDA-ARS, Systematic

Entomology Laboratory) for parasitoid identification, Terry Hingtgen (Florida

Department of Environmental Protection, Southwest Division) for his assistance in acquiring the necessary permits for this work, and University of South Florida for providing funding.

References cited

Abdala-Roberts L, Marquis RJ (2007) Test of local adaptation to biotic interactions and soil abiotic conditions in the ant-tended Chamaecrista fasciculata (Fabaceae). Oecologia 154:315-326

Agrawal AA, Lau JA, Hamback PA (2006) Community heterogeneity and the evolution of interactions between plants and insect herbivores. Q Rev Biol 81:349-376

Anderson EF (2001) The Cactus Family. Timber Press, Portland, OR

Andow DA (1991) Vegetational diversity and arthropod population response. Annu Rev Entomol 36:561-586

Atsatt PR, Odowd DJ (1976) Plant Defense Guilds. Science 193:24-29

Baker AJ, Stiling P (2009) Comparing the effects of the exotic cactus-feeding moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae) and the native cactus- 82

feeding moth, Melitara prodenialis (Walker) (Lepidoptera: Pyralidae) on two species of Florida Opuntia. Biol Invasions 11:619-624

Barbosa P, Caldas A (2007) Do larvae of species in macrolepidopteran assemblages share traits that influence susceptibility to parasitism? Env Entomol 36:329-336

Barbosa P, Hines J, Kaplan J, Martinson H, Szczepaniec A, Szendrei Z (2009) Associational resistance and associational susceptibility: Having right or wrong neighbors. Annu Rev Ecol Evol S 40:1-20

Barton AM (1986) Spatial variation in the effect of ants on an extrafloral nectary plant. Ecology 67:495-504

Bennett FD, Habeck DH (1996) Cactoblastis cactorum: A successful weed control agent in the Caribbean, now a pest in Florida? In: Delfosse ES, Scott RR, (eds) Biological control of weeds: Proceedings of the VIII international symposium on biological control of weeds. CSIRO Publishing, Melbourne pp 21-26

Bentley BL (1977) Extrafloral nectaries and protection by pugnacious bodyguards. J Ecol 65:27-38

Boecklen WJ (1984) The role of extrafloral nectaries in the herbivore defense of Cassia fasciculata. Ecol Entomol 9:243-249

Brown BJ, Ewel JJ (1987) Herbivory in complex and simple tropical successional ecosystems. Ecology 68:108-116

Carpenter JE, Bloem S, Bloem KA (2001) Inherited sterility in Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla Entomol 84:537-542

Del Carmen Mandujano M, Montana C, Mendez I, Golubov J (1998) The relative contribution of sexual reproduction and clonal propagation in Opuntia rastrera from two habitats in the Chihuahuan desert. J Ecol 86: 911-921

DelClaro K, Berto V, Reu W (1996) Effect of herbivore deterrence by ants on the fruit set of an extrafloral nectary plant, Qualea multiflora (Vochysiaceae). J Trop Ecol 12:887-892

Deyrup M, Davis L, Cover S (2000) Exotic ants in Florida. T Am Entomol Soc 126:293- 326

Dodd AP (1940) The biological campaign against prickly pear. Commonwealth Prickly Pear Board Bulletin: 1-177. Brisbane, Australia.

Finch S, Billiald H, Collier RH (2003) Companion planting - do aromatic plants disrupt host-plant finding by the cabbage root fly and the onion fly more effectively than non-aromatic plants? Entomol Exp Appl 109:183-195

Fleet RR, Young BL (2000) Facultative mutualism between imported fire ants (Solenopsis invicta) and a legume (Senna occidentalis). Southwest Nat 45:289- 298

Fullaway DT (1954) Biological control of cactus in Hawaii. J Econ Entomol 47:696-700

Gagnon P, Himmelman JH, Johnson LE (2003) Algal colonization in urchin barrens: defense by association during recruitment of the brown alga Agarum cribrosum. J Exp Mar Biol Ecol 290:179-196

Galloway LF, Fenster CB (2000) Population differentiation in an annual legume: local adaptation. Evolution 54:1173-1181

Gimeno I, Vila M (2002) Recruitment of two Opuntia species invading abandoned olive groves. Acta Oecologica 23: 239-246

Hamback PA, Agren J, Ericson L (2000) Associational resistance: Insect damage to purple loosestrife reduced in thickets of sweet gale. Ecology 81:1784-1794

Hamback PA, Beckerman AP (2003) Herbivory and plant resource competition: A review of two interacting interactions. Oikos 101:26-37

Hamback PA, Stenberg JA, Ericson L (2006) Asymmetric indirect interactions mediated by a shared parasitoid: Connecting species traits and local distribution patterns for two chrysomelid beetles. Oecologia 148:475-481

Hamlin JC (1924) A review of the genus Chelinidea (Hemiptera-Heteroptera) with biological data. Ann Entomol Soc Am 17:193-208

Harmon JP, Ives AR, Losey JE, Olson AC, Rauwald KS (2000) Colemegilla maculata (Coleoptera: Coccinellidae) predation on pea aphids promoted by proximity to dandelions. Oecologia 125:543-548

Hay ME (1986) Associational plant defenses and the maintenance of species-diversity - turning competitors into accomplices. Am Nat 128:617-641

Heil M, McKey D (2003) Protective ant-plant interactions as model systems in ecological and evolutionary research. Annu Rev Ecol Evol S 34:425-453

Hight SD, Carpenter JE (2009) Flight phenology of male Cactoblastis cactorum (Lepidoptera: Pyralidae) at different latitudes in the southeastern United States. Fla Entomol 92:208-216 84

Hight SD, Carpenter JE, Bloem S, Bloem KA (2005) Developing a sterile insect release program for Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae): effective overflooding ratios and release-recapture field studies. Environ Entomol 34:850- 856.

Holt RD, Lawton JH (1994) The ecological consequences of shared natural enemies. Annu Review of Ecol Syst 25:495-520

Hunter WD, Pratt FC, Mitchell JD (1912) The principle cactus insects of the United States. United States Department of Agriculture, Bureau of Entomology, Washington DC

Janzen DH (1966) Coevolution of mutualism between ants and acacias in Central America. Evolution 20:249-275

Jezorek HA, Stiling PD, Carpenter JE (2010) Targets of an invasive species: oviposition preference and larval performance of Cactoblastis cactorum (Lepidoptera: Pyralidae) on 14 North American opuntioid cacti. Environ Entomol 39:1884-1892

Karban R (1997) Neighbourhood affects a plant's risk of herbivory and subsequent success. Ecol Entomol 22:433-439

Kelly CA (1986) Extrafloral nectaries: ants, herbivores and fecundity in Cassia fasciculata. Oecologia 69:600-605.

Koptur S (1984) Experimental evidence for defense of Inga (Mimosoideae) saplings by ants. Ecology 65:1787-1793

Lach L (2003) Invasive ants: unwanted partners in ant-plant interactions? Ann Mo Bot Gard 90:91-108

Lewis WJ, Stapel JO, Cortesero AM, Takasu K (1998). Understanding how parasitoids balance food and host needs: importance to biological control. Biol Control 11:175-183

Limburg DD, Rosenheim JA (2001) Extrafloral nectar consumption and its influence on survival and development of an omnivorous predator, larval Chrysoperla plorabunda (Neuroptera: Chrysopidae). Environ Entomol 30:595-604

Lobos E, de Cornelli JO (1997) Observations on Cactoblastis cactorum (Berg) as a pest of cactus pear (Opuntia ficus-indica) in Argentina with suggestions on possible control methods. J Prof Assoc Cactus 2:97-102

Mann J (1969) Cactus-feeding insects and mites. Smithsonian Institution Bulletin 256, Washington, D.C.

Mathews CR, Brown MW, Bottrell DG (2007) Leaf extrafloral nectaries enhance biological control of a key economic pest, Grapholita molesta (Lepidoptera: Tortricidae), in peach (Rosales: Rosaceae). Environ Entomol 36:383-389

Milchunas DG, Noy-Meir I (2002) Grazing refuges, external avoidance of herbivory and plant diversity. Oikos 99:113-130

Miller TEX, Legaspi JC, Legaspi B (2010) Experimental test of biotic resistance to an invasive herbivore provided by potential plant mutualists. Biol Invasions 12:3563- 3577

Naisbitt T, James EK, Sprent JI (1992) The evolutionary significance of the legume genus Chamaecrista, as determined by nodule structure. New Phytol 122:487-492

Ness JH (2003) Contrasting exotic Solenopsis invicta and native Forelius pruinosus ants as mutualists with Catalpa bignonioides, a native plant. Ecol Entomol 28:257- 251.

Ness JH, Bronstein IL (2004) The effects of invasive ants on prospective ant mutualists. Biol Invasions 6:445-461

Oliveira PS, Rico-Gray V, Diaz-Castelazo C, Castillo-Guevara C (1999) Interaction between ants, extrafloral nectaries and insect herbivores in Neotropical coastal sand dunes: herbivore deterrence by visiting ants increases fruit set in Opuntia stricta (Cactaceae). Funct Ecol 13:623-631

Pemberton RW, Cordo HA (2001) Potential and risks of biological control of Cactoblastis cactorum (Lepidoptera: Pyralidae) in North America. Fla Entomol 84:513-526

Pettey FW (1947) The biological control of prickly pears in South Africa. Department of Agriculture of the Union of South Africa

Pfister CA, Hay ME (1988) Associational plant refuges: Convergent patterns in marine and terrestrial communities result from differing mechanisms. Oecologia 77:118- 129

Pickett CH, Clark WD (1979) Function of extrafloral nectaries in Opuntia acanthocarpa (Cactaceae). Am J Bot 66:618-625

Porter SD, Savignano DA (1990) Invasion of polygyne fire ants decimates native ants and disrupts arthropod community. Ecology 71:2095-2106

Redman AM, Scriber JM (2000) Competition between the gypsy moth, Lymantria dispar, and the northern tiger swallowtail, Papilio canadensis: interactions mediated by host plant chemistry, pathogens, and parasitoids. Oecologia 125:218-228 86

Rios RS, Marquis RJ, Flunker JC (2008) Population variation in plant traits associated with ant attraction and herbivory in Chamaecrista fasciculata (Fabaceae). Oecologia 156:577-588

Robbins M, Miller TEX (2009) Patterns of ant activity on Opuntia stricta (Cactaceae), a native host-plant of the invasive cactus moth, Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla Entomol 92:391-393

Robertson HG, Hoffmann JH (1989) Mortality and life-tables of Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae) compared on two host-plant species. B Entomol Res 79:7-17

Root RB (1973) Organization of a plant-arthropod association in simple and diverse habitats: fauna of collards (Brasscia oleracea). Ecol Monogr 43:95-120

Russell EP (1989) Enemies hypothesis: A review of the effect of vegetational diversity on predatory insects and parasitoids. Environ Entomol 18:590-599

Russell FL, Louda SM, Rand TA, Kachman SD (2007) Variation in herbivore-mediated indirect effects of an invasive plant on a native plant. Ecology 88:413-423

Rutter MT, Rausher MD (2004) Natural selection on extrafloral nectar production in Chamaecrista fasciculata: The costs and benefits of a mutualism trait. Evolution 58:2657-2668

Savage AM, Rudgers JA, Whitney KD (2009) Elevated dominance of extrafloral nectary- bearing plants is associated with increased abundances of an invasive ant and reduced native ant richness. Divers Distrib 15:751-761

Settle WH, Wilson LT (1990) Invasion by the variegated leafhopper and biotic interactions: Parasitism, competition, and apparent competition. Ecology 71:1461- 1470

Sheehan W (1986) Response by specialist and generalist natural enemies to agroecosystem diversification: A selective review. Environ Entomol 15:456-461

Shiojiri K, Takabayashi J,Yano S, Takafuji A (2001) Infochemically mediated tritrophic interaction webs on cabbage plants. Popul Ecol 43:23-29

Shiojiri K, Takabayashi J, Yano S, Takafuji A (2002). Oviposition preferences of herbivores are affected by tritrophic interaction webs. Ecol Lett 5:186-192

Spellman B, Brown MW, Mathews CR (2006) Effect of floral and extrafloral resources on predation of Aphis spiraecola by Harmonia axyridis on apple. Biocontrol 51:715-724

StatSoft, Incorporated (2000) STATISTICA for Windows, Version 5.5. StatSoft, Incorporated, Tulsa, OK

Stenberg JA, Heijari J, Holopainen JK, Ericson L (2007) Presence of Lythrum salicaria enhances the bodyguard effects of the parasitoid Asecodes mento for Filipendula ulmaria. Oikos 116:482-490

Stiles JH, Jones RH (2001) Top-down control by the red imported fire ant (Solenopsis invicta). Am Midl Nat 146:171-185

Stiling P (2002) Potential nontarget effects of a biological control agent, prickly pear moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae), in North America, and possible management actions. Biol Invasions 4:273-281

Stiling P, Rossi AM, Cattell MV (2003) Associational resistance mediated by natural enemies. Ecol Entomol 28:587-592

Tahvanainen JO, Root RB (1972) Influence of vegetational diversity on population ecology of a specialized herbivore, Phyllotreta cruciferae (Coleoptera: Chrysomelidae). Oecologia 10:321-346

Tate CD, Carpenter JE, Bloem S (2007). Influence of radiation dose on the level of F-1 sterility in the cactus moth, Cactoblastis cactorum (Lepidoptera: Pyralidae). Fla Entomol 90:537-544

Thomas CD (1986) Butterfly larvae reduce host plant survival in vicinity of alternative host species. Oecologia 70:113-117

Tilman D (1978) Cherries, ants and tent caterpillars: Timing of nectar production in relation to susceptibility of caterpillars to ant predation. Ecology 59:686-692

United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine (2007) Life history of the cactus moth. http://www.aphis.usda.gov/plant_health/plant_pest_info/cactoblastis. Accessed 25 March 2011

Wahl M, Hay ME (1995) Associational resistance and shared doom: effects of epibiosis on herbivory. Oecologia 102:329-340

White JA, Andow DA (2006) Habitat modification contributes to associational resistance between herbivores. Oecologia 148:482-490

White JA, Whitham TG (2000) Associational susceptibility of cottonwood to a box elder herbivore. Ecology 81:1795-1803

Woodruff RE (2009) Cactus Weevils, Gerstaeckeria hubbardi (LeConte) and Gerstaeckeria fasciata Pierce (Insecta: Coleoptera: Curculionidae). University of Florida Institute of Food and Agricultural Sciences, Department of Entomology and Nematology EENY-82

Zimmermann HG, Bloem S, Klein H (2004) Biology, history, threat, surveillance and control of the cactus moth, Cactoblastis cactorum. International Atomic Energy Agency/Food and Agriculture Organization, Vienna

CHAPTER 4

FINDINGS FROM A SIX YEAR STUDY: EFFECTS OF CACTOBLASTIS

CACTORUM ON THE SURVIVAL AND GROWTH OF LOCAL OPUNTIA

Synopsis

Cactoblastis cactorum is known for being both a biological control agent and an invasive pest of opuntioid cacti. The spread of C. cactorum in the southeastern United States may threaten the biological and physical integrity of desert, scrub, and coastal habitats.

However, the effects of invasive species are known to vary spatially and temporally, and

C. cactorum‟s efficacy as a biological control agent varies considerably from region to region. Therefore, the long term effects of C. cactorum within its U.S. range are still uncertain. We surveyed tagged Opuntia stricta (n=253) and O. humifusa (n=327) plants along the west coast of Florida for six years to determine the effects of C. cactorum attack on survival and growth rate of plants, and to examine host species differences and the effects of plant size. 78.11% of the Opuntia plants were attacked at least once by C. cactorum during the six year study and overall survival rate was 75.8%. Plants attacked by C. cactorum were more likely to die than unattacked plants and a plant‟s odds of surviving the six year period decreased as C. cactorum attack frequency increased.

However, plants that survived the six year period showed, on average, positive growth and there was no significant difference in growth rates between surviving attacked and unattacked plants. O. stricta plants were more likely to be attacked, were attacked more

90 frequently, and were more likely to die after being attacked than were O. humifusa plants.

Plant size did not predict plant survival, but larger surviving plants lost proportionally more pads over the six years than smaller surviving plants. Although C. cactorum should still be considered a threat, particularly for rare opuntioids, overall survival at our sites is currently high and plants that are able to survive C. cactorum attack are not being reduced in size, possibly because they possess traits that render them more tolerant of C. cactorum damage. Our findings suggest that an assumption of severe negative effects of an invasive species, based on its effects in other regions or over short periods of time, may not always be justified.

Introduction

Invasive species are often responsible for negative effects on native biodiversity which can result from a variety of direct and indirect mechanisms (Mack et al. 2000, Snyder and

Evans 2006, Kenis et al. 2009). The severity of these negative effects certainly varies from species to species, but is also expected to vary spatially and temporally for any particular invasive species, leading to variation in its effects on the community (Parker et al. 1999, Strayer et al. 2006). Despite the expectation of temporal variation, most studies of invasive species take place over relatively short time scales. A review of studies of invasive species from 2001-2005 found that approximately 80% were of durations of 3 years or less (Strayer et al. 2006). Studies of invaders over longer durations or over multiple points in time post-invasion can reveal patterns that would otherwise be missed.

Severe initial impacts during early exponential growth of an invasive species, i.e., the log phase (Mack et al. 2000), may be followed by a decline in its abundance, allowing for at

91 least partial recovery of resilient species and communities. Surveys from central Texas early in the invasion stage of Solenopsis invicta showed that infested sites had dramatically lower native ant and non-ant arthropod diversity (Porter and Savignano

1990). Twelve years later, however, ant and arthropod diversity at the same sites had returned to pre-invasion levels and S. invicta abundance had declined (Morrison 2002).

Long term studies of invasive species can also reveal an increase in the severity of an invader‟s effects. Wilson et al (2004) analyzed 19 years of data from a Wisconsin lake undergoing dispersal by the invasive Orconectes rusticus (rusty crayfish). They found that macroinvertebrates, macrophytes, and native crayfish experienced dramatic negative effects very quickly after exposure to O. rusticus invasion, while negative impacts on more mobile, longer-lived species took several more years to become evident and large piscivorous fish such as bass had not yet suffered declines (Wilson et al. 2004). In both of these cases, data from the first few years post-invasion leads to different conclusions, and therefore different control and management decisions, than data from a decade post- invasion. Continuing research for multiple decades post-invasion could lead to yet different conclusions regarding the severity of an invader‟s effects, although logistical and monetary constraints make this difficult in more cases than not.

Classical biological control provides some of the best evidence for spatial and temporal variation in the impacts of non-native species, as the efficacy of a biological control agent often varies across the invasive range of the target species (Monro 1975,

Phillips and Baird 2001, Denslow and D'Antonio 2005, Romeis et al. 2005, Shea et al.

2005). The pyralid moth Cactoblastis cactorum was a phenomenally successful biological control agent of Opuntia stricta (erect prickly pear) in Australia (Dodd 1940,

Zimmermann et al. 2000). However, in South Africa, C. cactorum has not achieved the same level of control of Opuntia spp. (Hoffmann et al. 1999, Zimmermann et al. 2004) and in particular, has failed to control O. stricta in Kruger National Park (Hoffmann et al.

1998). Even within regions where C. cactorum is considered successful, levels of control are variable, with persistent populations of target Opuntia spp. remaining decades after the moth‟s introduction (Monro 1975, White 1980, Pemberton and Liu 2007).

In North America, C. cactorum is an invasive pest which currently ranges from the Florida Keys north to Bull Island, SC and has been detected as far west as Terrebonne

Parish, LA (Rose et al. 2011). Although it is known to feed on nearly all opuntioid species in its current U.S. range (Johnson and Stiling 1996, Jezorek et al. 2010, Rose et al. 2011), the long term effects of C. cactorum are as yet unknown. Certainly for rare species such as Consolea corallicola, which is endemic to the Florida Keys, the risk of extinction is grave (Stiling 2010). For more common opuntioid species, quantitative data on mortality rates within C. cactorum‟s U.S. range are lacking. Johnson and Stiling

(1998) found a mortality rate of 15% for O. stricta, but their data were from a small sample size and gathered from 1991-1993, soon after the first U.S. detection of C. cactorum in 1989. A second study found even lower mortality rates (~3.5%) for plants damaged by C. cactorum (Baker and Stiling 2009); this study used data gathered more recently and from a much larger sample size, but was of relatively short duration (2 years,

2003-2005). Such studies suggest that although C. cactorum is capable of high levels of damage, an assumption that this damage will result in extirpation of common species of

Opuntia may not be warranted. Based upon the known variation in the efficacy of C. cactorum as a biological control agent, and of the general tendency for impacts of

93 invasive species to vary over time, we wanted to document mortality and growth of local

Opuntia populations over a longer duration. We used data collected over a six year period to determine: (1) Are plants attacked by C. cactorum more likely to die than unattacked plants? (2) How does rate of C. cactorum attack affect survival of plants? (3) For surviving plants, does C. cactorum attack affect growth rate? We also examined differences between two host species, O. humifusa and O. stricta, and the relationship of plant size to C. cactorum attack, survival, and growth rate.

Methods

Study system

Cactoblastis cactorum is a cactus-feeding, pyralid moth native to South America. During the twentieth century, it was intentionally released to control pest species of opuntioid cacti in Australia, South Africa, Mauritius, Hawaii, and the Caribbean. The moth was first discovered in Florida in 1989 and has since spread along the Atlantic and Gulf coasts of the southeastern United States. Females oviposit ~50-100 eggs one on top of the other to form an eggstick that is attached to the tip of a cactus spine or directly to a cladode.

The gregarious larvae feed internally, completing five instars, and then drop to the ground to pupate in the soil or hollowed out cladodes. In central Florida, there are three non-overlapping flight periods per year: a spring flight beginning in mid- February, a summer flight beginning in early June, and a fall flight beginning in late August (Hight and Carpenter 2009). Cactoblastis cactorum has a wide host range within the genera

Opuntia, Consolea, and Nopalea. At our study sites, it feeds on the species O. stricta and

O. humifusa. Surveys were conducted at Fort deSoto Park (Pinellas County; hereafter

FDS), Honeymoon Island State Park (Pinellas County; hereafter, HI), and Lido Key

(Sarasota Country; hereafter LK). All three sites are part of Florida‟s west-central barrier island chain and share similar plant communities; O. stricta and O. humifusa are common in the dune and coastal scrub areas found on these islands.

Data collection and analysis

In March of 2003, Opuntia plants were marked at the three sites. Each plant was flagged and had a numbered metal tag driven into the ground at its base. Sample sizes by site and species were as follows: 140 plants (92 O. humifusa and 48 O. stricta) at FDS, 253 plants

(176 O. humifusa and 77 O. stricta) at HI, and 200 plants (100 O. humifusa and 100 O. stricta) on LK. In April of 2006, an additional 43 O. stricta were marked at HI. Fifty-six plants that were marked in 2003 were lost in prescribed burns and were removed from analyses, leaving a total sample size of 580 (327 O. humifusa and 253 O. stricta).

Plants were surveyed monthly from April 2003 until July 2005 and tri-annually from November 2005 until March 2009. Tri-annual surveys took place in March/April,

July, and October/November in order to correspond with peak larval activity. During a survey, the number of live primary (old, woody), secondary (green, mature), and tertiary

(new growth with true leaves) cladodes and the number of cladodes with active C. cactorum damage and eggsticks were counted for each plant. Damage was considered active if cladodes were not yet fully hollowed out and had signs of larval infestation including entrance holes, dripping frass, and a yellow or black “mushy” appearance.

When in doubt, cladodes were cut open to check for live larvae. Larvae and eggsticks of the native cactus boring moth, Melitara prodenialis, are easily distinguishable from C.

95 cactorum, and were never observed at our sites. Plants were scored as dead when they had zero live cladodes for a period of a year (i.e., three consecutive tri-annual surveys).

We tested for differences between sites using log likelihood tests for categorical variables and ANOVA for continuous variables. No significant differences between sites were found, therefore, data from the three sites were pooled for all analyses.

As some plants were marked in 2003 and other in 2006, data were analyzed according to „survey year‟, rather than calendar year. For example, „survey year 2‟ is calendar year 2004 for plants marked in 2003 and 2007 for plants marked in 2006. The

2006 sample consisted of O. stricta plants only and was therefore tested for heterogeneity against O. stricta from the 2003 sample. We used log likelihood tests for categorical variables and Student‟s t- tests for unequal sample sizes for continuous and scaled variables. For categorical variables, no significant differences were found for survey years 1 and 3; no plants from the 2006 sample died in survey year 2 so this comparison could not be made. For continuous and scaled variables, there were no significant differences between the two samples, although a Levene‟s test revealed unequal variances for the variable attack frequency. A Welch‟s t-test, which does not assume equal variance or sample size, was then performed and was non-significant. Therefore, for survey years 1-3 data were pooled between the two samples.

All analyses were performed in SPSS 19 (SPSS Inc. 2010) unless otherwise indicated. Prior to performing ANCOVA, assumptions of linearity of the covariate and homogeneity of variance and slopes were checked. The variable attack frequency was arcsine square root transformed to meet assumptions. Prior to performing linear regression, assumptions of linearity, homogeneity of variance, normality of residuals, and

96 collinearity of predictors were checked. The variables proportional change in cladodes/generation and initial cladode number were natural log-transformed to meet these assumptions.

Q1:Are plants attacked by C. cactorum more likely to die than unattacked plants?

For each survey year, we used a likelihood ratio test (G2 test) of the null hypothesis that attack status, survival status, and species were independent. We defined attack status for any survey year as positive if a plant was attacked during at least one of the three C. cactorum generations of that year and negative if a plant was not attacked at all that year.

We defined survival status for any survey year as positive for plants that were alive at the end of the year, negative for plants that died that year, and undetermined for plants whose fate was unknown at the end of that year. Only plants with known fates at the end of each year were included in the test. We tested for homogeneity of odds ratios between species with Tarone‟s X2 test. If odds ratios were found to be homogeneous, we used a Mantel-

Haenszel test for conditional independence of attack and survival status while controlling for species, and found the Mantel-Haenszel common odds ratio estimate, which in this case represents the ratio of the odds of dying for attacked versus unattacked plants, taking into account species differences.

We also used a likelihood ratio test of the null hypothesis that final attack status, final survival status, and species were independent, and determined relative risk and odds ratios. We defined final attack status as positive if a plant was attacked by C. cactorum at least once over the duration of the study and negative if a plant was never attacked by C. cactorum. Final survival status was positive for plants that survived the duration of the

97 study, negative for those that did not survive, and undetermined for plants whose fate was unknown at the end of the study. Only plants with known fates at the end of the study were included in the test (n=532). We included the same Tarone‟s and Mantel-Haenszel tests as described above. In order to test significance and find confidence intervals of odds ratios, we performed backward elimination model selection with the same plants and factors. Model selection indicated that all two way interactions and lower order effects were significant (p=0.000-0.021), but the three way interaction was non- significant and was not included in the model. Based on the results of the first likelihood ratio test, and because nearly all O. stricta plants were positive for final attack status, we performed a second likelihood ratio test and used a second log linear model to analyze just those plants with both a positive final attack status and a known final survival status

(n=415). The purpose of these analyses was to determine if there were species differences in survival specifically after C. cactorum attack.

Q2: How does frequency of attack by C. cactorum affect survival of plants?

Attack frequency for each plant over the duration of the study was defined as follows:

Number of C. cactorum generations for which active larvae were recorded Total number of C. cactorum generations surveyed

For example, a plant surveyed from March 2006 to March 2009 was surveyed for ten C. cactorum generations (3 per year for 2006-2008, plus one in 2009). If this plant had active larvae during six of those generations, its attack frequency would be 6/10=0.600.

To determine if attack frequency differed between host species, we performed an

ANCOVA on all plants with species as a fixed factor and initial cladode number as a covariate. This covariate was included because plants with more cladodes clearly have an advantage in that they can “last longer”, that is, they provide a greater number of opportunities for attack before suffering 100% cladode loss and therefore, death. Initial cladode number refers to the number of live secondary and tertiary cladodes present on a plant during its first survey; primary cladodes are generally not used by C. cactorum and therefore were not included in the variable.

To determine the effects of attack frequency on plant survival, we used backward stepwise selection to compare between logistic regression models. Inclusion in the model was based upon a significant (p≤0.05) change in deviance (-2*log likelihood) when a predictor term was removed from the model. The dependent variable was final survival status and potential predictor variables were: attack frequency, host species, initial cladode number, attack frequency*host species, and attack frequency*initial cladode number. We performed the procedure with two sets of plants. First, we used all plants with known fates at the end of the study (n=532); second, we used plants with known fates and an attack frequency greater than zero (positive attack status, n=415). Model fit was checked with a Hosmer-Lemeshow test.

Q3: For surviving plants, does C. cactorum attack affect growth rate?

Growth rate over the duration of the study was measured as proportional change in live cladodes as follows:

Final number of cladodes – initial number of cladodes Initial number of cladodes

This proportional change in cladodes was then divided by the number of C. cactorum generations for which the plant was followed. For example, a plant with an initial cladode number of 200 in April 2003 and a final cladode number of 100 in March of 2009 would have a proportional cladode change of -0.500 over 19 C. cactorum generations (3 per year for 2003-2008, plus a spring generation of 2009), which would give a proportional cladode change per generation (hereafter PC/G) of -0.026. Surviving O. humifusa and O. stricta were found to have equal variances for PC/G (Levene‟s F=0.019, p=0.890), so a

Student‟s t-test for independent samples was used to test for differences in mean PC/G between host species. Attacked and unattacked survivors had unequal variance (Levene‟s

F=7.13, p=0.008) and unequal sample size (attacked, n=331; unattacked, n=109), so a

Welch‟s t-test was used to test for differences in mean PC/G between plants with positive and negative final attack status. To determine if attack frequency or starting size of a plant predicted PC/G, we performed multiple linear regression of ln(PC/G) on attack frequency and ln(initial cladode number), both for all surviving plants and for each host species separately.

100

Results

Attack rates for each survey year ranged from 27% to 41.5%, but survival rates were high in all survey years (Table 4.1). Of the 580 plants used in analyses, 78.10% were attacked by C. cactorum at least once during the study and 75.87% survived until March of 2009.

A higher percentage of O. stricta were attacked and died over the course of the study than of O. humifusa (Fig.4.1).

Q1: Are plants attacked by C. cactorum more likely to die than unattacked plants?

For survey years 3-6, the proportion of plants that were attacked in the year they died was lower than the proportion that were unattacked in the year they died. Survey years 1 and

2 showed the opposite trend (Fig. 4.2). We could not reject the null hypothesis of independence of attack and survival status for survey years 1-4 and 6 (p-values for all G2 and Mantel-Haenszel X2 statistics >0.100). For survey year 5, we rejected the null hypothesis for the total sample (both species pooled, G2=6.31, p=0.012) and for O. humifusa (G2=6.28, p=0.012), but the relationships were relatively weak (Theil‟s

U=0.044 and 0.085, respectively) and we could not reject the null hypothesis for O. stricta (G2=3.80, p=0.052). When controlling for species, our rejection of the null held

(Mantel-Haenszel X2=6.09, p=0.014), and the common odds ratio was 0.136, indicating that the odds of a plant dying during survey year 5 were lower if it experienced at least one C. cactorum attack than if it did not experienced an attack. For all survey years, the odds ratio of dying for attacked versus unattacked plants was homogeneous between species (p-values for all Tarone‟s X2 statistics >0.250).

101

We rejected the null hypothesis that final attack and final survival status were independent for the total sample (both species pooled) or for O. humifusa. However, the associations were weak, as evidenced by low values of Theil‟s U (Table 4.2). We could not reject the null hypothesis for O. stricta. This non-intuitive result is due to the fact that

92.6% of O. stricta plants in the sample were attacked. The very low frequency of unattacked O. stricta made comparisons between attacked and unattacked categories somewhat problematic, and led us to perform a second test examining species differences in survival after C. cactorum attack (see below and Table 4.2 and 4.3). When controlling for species, our rejection of the null held (Mantel-Haenszel X2 statistic = 4.14, df=1, p=0.043). For both species, the relative risk of dying was higher for attacked plants than unattacked plants; the increase in risk was higher for O. humifusa but the difference in risk between species was not significant (Table 4.2c). The Mantel-Haenszel common odds ratio was 2.33 (p=0.033), indicating that the odds of a plant dying during the study if it experienced at least one C. cactorum attack was 133% higher than if it experienced no attacks.

The log linear model using all two-way interactions and lower order effects of final attack status, final survival status, and species adequately fit the data (G2=1.92, df=1, p=0.166). Plants were significantly more likely to survive than die (Z=4.29, p=0.000) and significantly more likely to be have a positive attack status than a negative one (Z=8.753, p=0.000). Interaction parameters confirmed that plants with a positive final attack status had a significantly higher risk of death over the course of the study, and revealed that O. humifusa plants had lower risk of death, regardless of attack status, and a lower risk of being attacked by C. cactorum (Table 4.2).

102

The second likelihood ratio test revealed that species and survival after attack by

C. cactorum are not independent, although their association was weak (Table 4.2). The relative risk of dying after an attack was lower for O. humifusa than O. stricta (Table

4.2); the parameter estimate from the second log linear model showed that this was significant (Table 4.3).

Q2: How does frequency of attack by C. cactorum affect survival of plants?

The mean attack frequency for all plants of both species was 16.67%. The ANCOVA interaction term between host species and initial cladode number was non-significant

(F=1.52, p=0.219), showing homogeneity of the coefficient for cladode number across species. Opuntia stricta had a higher mean attack frequency (24.13%) than O. humifusa

(10.90%) and ANCOVA indicated that this difference was significant (F=150.24, p=0.000) with a relatively large effect size (η2=0.207). Initial cladode number also accounted for a significant proportion of variation in attack frequency (F=80.82, p=0.000), but the effect size was smaller than that of host species (η2=0.123).

For both sets of plants analyzed, the variables attack frequency, host species and attack frequency*host species were retained in the logistic regression model as significant predictors of final survival status (Table 4.4). Initial cladode number and attack frequency*initial cladode number were not found to be significant predictors of final survival status. Partial coefficients (β, Table 4.4) indicated that the predicted log odds of survival decrease as attack frequency increases, and that this effect is much greater for O. stricta than for O. humifusa.

103

Q3:For surviving plants, does C. cactorum attack affect growth rate?

The mean PC/G for all plants of both species was -0.015 ± 0.004, meaning plants lost on average 1.5% of their initial cladodes per generation of C. cactorum. For surviving plants of both species, the mean PC/G was 0.016 ± 0.005, meaning surviving plants added about

1.6% of their initial cladodes per generation of C. cactorum. Surviving O. humifusa had a higher (more positive) mean PC/G than surviving O. stricta (Fig. 4.3), but this difference was not significant (t=1.01, p=0.311). Surviving attacked plants had a higher (more positive) mean PC/G than unattacked plants (Fig. 4.3), but again, this difference was not significant (Welch‟s t=1.80, p=0.074). Multiple regression models showed that both ln(initial cladode number) and attack frequency were significant predictors of ln(PC/G) for surviving plants (Table 4.5); results were similar for O. humifusa (Fig.4.4) and O. stricta (Fig.4.5). Coefficients indicated that proportional growth rate of survivors (as defined by PC/G), decreased as initial cladode number increased and increased as attack frequency increased. However, standardized coefficients, partial correlations, and partial regression plots revealed that initial cladode number was the stronger predictor of the two

(Table 4.5, Figs. 4.4 and 4.5).

104

Table 4.1 Summary of attack and survival rates for each survey year.

Survey year Start N Number attacked (%) Number Surviving (%) 1 580 180 (31.0) 573 (98.8) 2 573 238 (41.5) 564 (98.4)

3 564 184 (32.6) 496 (87.9)

4 459 a 124 (27.0) 442 (96.3)

5 442 170 (38.5) 424 (95.9)

6 424 169 (39.9) 403 (95.0)

Finalb 580 453 (78.1) 440 (75.9)

a The lower „Start N‟ for survey year 4 reflects that fact that some plants were only surveyed for three years (see Methods) b„Final‟ row displays the total number of plants that were attacked at least once by C. cactorum during the study and the total number of plants that survived for all years they were surveyed

105

Figure 4.1: Histogram of C. cactorum attack and survival rates. Data are from those plants whose fate was known at the end of the study (n=532)

106

1.00

0.90 Attacked 0.80 Unattacked 0.70 0.60 0.50 0.40

0.30 Proportion of dead plantsof dead Proportion 0.20 0.10 0.00 1 2 3 4 5 6 Survey year

Figure 4.2: The proportion of plants that died each survey year grouped by C. cactorum attack status for that same year

107

Table 4.2 Results from likelihood ratio tests

Sample Variables tested df Species G2, p Theil‟s U a Relative risk b 1: All plants with known Final attack status, O. humifusa 7.19, 0.007 0.035 3.28 c final survival status final survival status, 1 O. stricta 0.068, 0.795 0.000 1.12c (N=532) species Total 13.52, 0.000 0.028 2.96 2: Only plants with positive final attack status Final survival status 1 --- 14.57, 0.000 0.035 0.465 and known final survival and species status (N=415)

a Theil‟s U, also called the uncertainty coefficient, is a directional measure representing a percent reduction in uncertainty in predicting one variable when the value of a second variable is known. For Test 1 it represents the reduction in error in predicting final survival status if final attack status is known. For Test 2 it represents the reduction in error in predicting final survival status for an attacked plant if species is known b Relative risk is a ratio of the probabilities of an event. For Test 1 it represents the ratio of the probability of death for an attacked plant to an unattacked plant. For Test 2, it represents the ratio of the probability of death after a C. cactorum attack for O. humifusa to O. stricta c A test for homogeneity of the relative risk between species was non-significant (Tarone‟s X2=2.04, p=0.153)

108

Table 4.3 Results from log linear models

Sample Predictor Parameter estimate a SE Z p 95% CI Survival*Attack 0.857 b 0.400 2.14 0.032 0.073 − 1.64 1: All plants with known Species*Survival -1.07 b 0.248 -4.29 0.000 -1.55 − -0.578 final survival status (N=532) Species*Attack -1.66 b 0.290 -5.74 0.000 -2.30 − -1.09 2: Only plants with positive final attack status and known Species*Survival -0.945 c 0.255 -3.71 0.000 -1.45 − -0.445 final survival status (N=415)

a Only coefficients for the interaction terms are presented, as lower order coefficients are conditional upon reference categories and therefore, more difficult to interpret. The interaction coefficient is an estimate of a conditional log odds ratio. b Interpretation of parameter estimates for Model 1: Survival*Attack = the log of the ratio of odds of dying for an attacked versus an unattacked plant; Species*Survival = the log of the ratio of odds of dying for O. humifusa versus O. stricta; Species*Attack = the log of the ratio of odds of experiencing at least one attack for O. humifusa versus O. stricta c Interpretation of parameter estimate for Model 2: Species*Survival = the log of the ratio of odds of dying for an attacked O. humifusa versus an attacked O. stricta

109

Table 4.4 Results from logistic regression models with final survival status as the dependent variable

Hosmer-Lemeshow Sample goodness of fit test Nagelkerke R2b Predictor β SE pc Exp(β) Χ2 df pa

Attack frequency -3.79 0.962 0.000 0.023 1: All plants with known final survival status 2.29 7 0.942 0.138 Host species -2.19 0.575 0.000 0.113 Attack frequency*host (N=532) 3.15 1.20 0.009 23.34 species 2: Only plants with Attack frequency -4.54 1.39 0.001 0.011 positive final attack status 10.35 8 0.241 0.097 Host species -2.45 0.836 0.003 0.087 and known final survival Attack frequency*host status (N=415) 3.70 1.66 0.025 40.55 species

a A Hosmer-Lemeshow statistic with p≤0.05 indicates a poor fit for the data b The Nagelkerke R2 is a pseudo-R2 statistic which attempts to estimate a coefficient of determination. It has a scale of 0 (the predictor variables do not explain any of the variation in the dependent variable) to 1 (the predictor variables perfectly explain variation in dependent variable) c Significance of the predictors is based upon the Wald statistic, which is the partial coefficient (β) divided by its SE and follows a Χ2 distribution

110

Figure 4.3 Mean proportional change in initial cladode number per generation of C. cactorum for all surviving plants. n=440, error bars represent ± 1 SE.

111

Table 4.5 Results from multiple regressions of ln(proportional cladode change/generation of C. cactorum) on attack frequency and ln(initial cladode number)

Partial Sample R2 F, p Tolerancea Coefficients Bb SE βc p correlation Constant 0.130 0.012 0.000 All surviving plants 55.88, ln(Initial cladode 0.204 0.881 - 0.040 d 0.004 - 0.481 0.000 - 0.452 (n=440) 0.000 number) Attack frequency 0.073 e 0.024 0.140 0.002 0.146 Constant 0.131 0.015 .000 Surviving O. humifusa 37.12, ln(Initial cladode 0.210 0.863 - 0.040 d 0.005 - 0.493 0.000 - 0.458 (n=284) 0.000 number) Attack frequency 0.120 e 0.037 0.183 0.002 0.187 Constant 0.136 0.020 0.000 Surviving O. stricta 25.42, ln(Initial cladode 0.250 0.780 - 0.050 d 0.007 - 0.566 0.000 - 0.500 (n=156) 0.000 number) Attack frequency 0.122 e 0.040 0.244 0.003 0.215 a Tolerance is an indication of the proportion of variance in a predictor that cannot be explained by other predictors; values <0.20 indicate problematic collinearity b Unstandardized coefficients c Standardized coefficients d Due to ln transformations, B coefficients are interpreted as the % change in the dependent variable for a 1% increase in the predictor variable e Due to ln transformation of the dependent variable, B coefficients are interpreted as the B*100% change in the dependent variable for a one unit change in the predictor

112

a) b)

Figure 4.4: Partial regression plots for O. humifusa. Plots are of a multiple regression of ln(proportional cladode change/generation of C. cactorum) on (a) attack frequency and (b) ln(initial cladode number) and indicate that initial cladode number is a stronger predictor of proportional growth rate than attack frequency.

113

a) b)

Figure 4.5: Partial regression plots for O. stricta. Plots are of a multiple regression of ln(proportional cladode change/generation of C. cactorum) on (a) attack frequency and (b) ln(initial cladode number) and indicate that initial cladode number is a stronger predictor of proportional growth rate than attack frequency.

114

Discussion

Among scientists, a common argument is that we should adapt a “guilty until proven innocent” stance on non-native species (Ruesink et al. 1995, Mack et al. 2000, Van

Driesche and Van Driesche 2001, Simberloff 2003). However, others argue that the assumption that alien status implies harmful effects is misplaced, and that we need to be more objective in our terminology, assessment of, and policies concerning non-natives

(Brown and Sax 2004, Colautti and MacIsaac 2004, Schlaepfer et al. 2011). Lively debate over how we should define and regulate non-native and invasive species, and whether they generally have harmful effects, has ensued over recent years (Brown and

Sax 2005, Cassey et al. 2005, Sagoff 2005, Simberloff 2005, Brown et al. 2007, Larson

2007). Many invasion biologists are now pushing for a system in which evaluation of species is based more upon documented species‟ impacts, and less on native versus non- native status (Davis et al. 2011). Such a principle seems to be particularly relevant in light of the rapid changes to ecosystems caused by everything from local land use change to global climate change. However, policies and programs based on species‟ impacts in the past may not be appropriate for the current era, and those based on impacts in one region may not work well in another region, especially if conclusions about said impacts are drawn mainly from qualitative observations or inconclusive data (Hager and McCoy

1998). Variation in the establishment and efficacy of biological control agents clearly demonstrates this notion (McFadyen 1998).

Habitat differences are an evident explanation for the varying success of non- native species, biological control agent or otherwise. Both abiotic factors, for example, distance to urban areas (Niemela and Spence 1991, Holway et al. 2002), soil moisture

115

(Holway et al. 2002), and stream discharge (Pintor and Sih 2011), and biotic factors, for example, species diversity (Morrison and Porter 2003) and prey biomass (Pintor and Sih

2011), have been found to correlate with variation in the density of invasive species. The lack of effective control of Opuntia by C. cactorum in South Africa can be at least partially explained by higher levels of predation than in Australia, both by ants

(Robertson and Hoffmann 1989) and baboons (Pettey 1948, Hoffmann et al. 1998).

Variation in the success of non-native species can also result from differing source populations, which can display variable fecundity (Phillips and Baird 2001) or abiotic tolerances (Romeis et al. 2005), among other differences. Often, explanations will be more complex. Shea et al (2005) found that population growth of invasive Carduus nutans (musk thistle) in New Zealand is driven by early life stage fecundity, while in

Australia it is driven by survival of later life stages, even though the plant spread to

Australia from New Zealand. Differences in life history may explain the lack of success of Rhinocyllus conicus (seed head weevil) in Australia as compared to North America, and lead to the prediction of more effective thistle control by Urophora solstitialis (a seed-reducing agent) in New Zealand and by Trichosirocalus horridus (a rosette feeder) in Australia (Shea et al. 2005).

We found that C. cactorum attack and Opuntia survival were statistically independent when each survey year was analyzed separately, but as the detrimental effects of attack may be delayed, examining attack and survival status in terms of plants‟ final status after six years probably gives a better indication of their relationship. When doing so, we found that C. cactorum-attacked plants had a higher risk of death, and this risk increased with increasing frequency of attack. Encouragingly, plants were still more

116 likely to survive than die over the six years, and when they did survive, most showed positive cladode growth, in agreement with findings by Johnson and Stiling (1998). This suggests that there may be substantial variation in the Opuntia population for tolerance to

C. cactorum attack.

Our finding that O. stricta is more likely to be attacked and to experience a higher attack frequency than O. humifusa is not unique, but our larger sample size and longer study duration provide superior support of the notion that O. stricta is the „preferred‟ of these two hosts of C. cactorum. More novel is the fact that even when an O. humifusa was attacked, it was more likely to survive than an attacked O. stricta. Together, these findings indicate that host species differ not only in resistance to C. cactorum attack, but also in tolerance to attack. Therefore, local and regional variation in the composition and density of various host communities may lead to dramatic differences in the effects of C. cactorum in North America.

We also found interesting relationships between plant size, C. cactorum attack, and plant survival. Plants with a larger initial cladode number had a higher attack frequency; this result is expected and agrees with findings of Pemberton and Liu (2007) and Johnson and Stiling (1998). However, we also found a negative relationship between cladode number and growth rate of surviving plants, such that larger plants lost proportionally more cladodes than smaller plants. This is in contrast to Pemberton and

Liu (2007) who found that larger plants have proportionally fewer cladodes attacked by

C. cactorum. Despite this, initial cladode number was not found to predict survival of our plants.

117

Perhaps most surprising were our findings concerning Opuntia growth rate.

Surviving plants that were attacked by C. cactorum added proportionally more cladodes than those that were unattacked, although this difference was not significant, and there was a significant positive relationship between attack frequency and proportional cladode change. One could argue that this is a by-product of the positive correlation between plant size and attack frequency; C. cactorum may selectively attack larger plants because they are growing more vigorously (Strauss 1988), leading to an apparent higher growth rate for attacked plants. However, we found the correlation between size and attack frequency to be relatively weak. Additionally, as stated above, larger surviving plants actually lost proportionally more cladodes when controlling for attack frequency (Table

4.5), leading us to conclude that the positive relationship between attack frequency and growth rate is not driven by initial plant size. This raises the intriguing possibility of compensatory growth as a topic for further study.

Given the widespread destruction of O. stricta in Australia and the Caribbean, our finding of high survival rates (87% for O. humifusa, 62% for O. stricta) in conjunction with high C. cactorum attack rates may seem unexpected. However, a closer look at the situation in these regions decades after the release of C. cactorum reveals that our results may be less surprising than at first glance Although C. cactorum provided effective control of Opuntia over an enormous area of Queensland, it proved ineffective in the southwest and coastal areas (Monro 1975, White 1980) and in New South Wales

(Hosking et al. 1994). White (1980) proposed that the dense stands of Opuntia in these areas are largely resistant to C. cactorum because they are under water and nutrient stress, resulting in lower plant quality and higher levels of mucilage than unstressed

118 plants. It has also been suggested that loss of female moths through dispersal could provide an explanation for the lack of control on small coastal islands off of Australia and Mauritius (Monro 1975), although to our knowledge this has not been further explored.

A survey of the Caribbean islands of Nevis and St. Kitts nearly 50 years after the initial release of C. cactorum exposed a similar situation (Pemberton and Liu 2007).

Opuntia triacantha and O. stricta populations have been greatly reduced, but not eradicated, even though C. cactorum remains established on the islands. A previous survey (Bennett and Simmonds 1966) found plants to be very scarce and small, but

Pemberton and Liu (2007) found that plants were “abundant enough to be easily located” and documented plants with up to 100 cladodes. The authors state that C. cactorum populations appear lower today than they were soon after its introduction, allowing some

Opuntia to escape attack and recover, which again emphasizes the importance of assessing data from multiple points in time (or from longer term studies) before drawing definitive conclusions on species‟ effects.

Our objective is not to trivialize the threat of C. cactorum. It is obvious that it could extirpate rare or less resistant opuntioid species, and even low mortality rates of more common species could have complex, negative effects on the community. We acknowledge that without data on plant reproduction and recruitment, definite statements on the long term viability of these Opuntia populations cannot be made. We also acknowledge that as conclusions about an invasive species‟ effects often change through space or time, our results may not hold true for other North American regions, or for studies done in west Florida ten years from now. In light of this we adamantly support

119 control and regulation policies of C. cactorum. Still, we feel our study provides some support of Pemberton and Liu‟s (2007) hopeful prediction that the effects of C. cactorum in North America will not necessarily be catastrophic.

Acknowledgements

We would like to thank Debi Tharp and Kerry Bohl for their enormous contribution to data collection, Terry Hingtgen (Florida Department of Environmental Protection,

Southwest Division) for his assistance in acquiring the necessary permits for field work at Honeymoon Island State Park, and the employees of Fort deSoto County Park and

Lido Key beaches for their support during the course of the study.

References cited

Baker, A. J. and P. Stiling. 2009. Comparing the effects of the exotic cactus-feeding moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae), and the native cactus-feeding moth, Melitara prodenialis (Walker) (Lepidoptera: Pyralidae), on two species of Florida Opuntia. Biological Invasions 11:619-624.

Bennett, F. D. and F. J. Simmonds. 1966. Biological control of Opuntia spp. by Cactoblastis cactorum in the Leeward Islands (West Indies). BioControl 11:183- 189.

Brown, J. H. and D. F. Sax. 2004. An essay on some topics concerning invasive species. Austral Ecology 29:530-536.

Brown, J. H. and D. F. Sax. 2005. Biological invasions and scientific objectivity: Reply to Cassey et al. (2005). Austral Ecology 30:481-483.

Brown, J. H., D. F. Sax, D. Simberloff, and M. Sagoff. 2007. Aliens among us. Conservation Magazine 8:14-21.

Cassey, P., T. M. Blackburn, R. P. Duncan, and S. L. Chown. 2005. Concerning invasive species: reply to Brown and Sax. Austral Ecology 30:475-480.

120

Colautti, R. I. and H. J. MacIsaac. 2004. A neutral terminology to define 'invasive' species. Diversity and Distributions 10:135-141.

Davis, M., M. K. Chew, R. J. Hobbs, A. E. Lugo, J. J. Ewel, G. J. Vermeij, J. H. Brown, M. L. Rosenzweig, M. R. Gardener, S. P. Carroll, K. Thompson, S. T. A. Pickett, J. C. Stromberg, P. Del Tredici, K. N. Suding, J. G. Ehrenfeld, J. P. Grime, J. Mascaro, and J. C. Briggs. 2011. Don't judge species on their origins. Nature 474:153-154.

Denslow, J. S. and C. M. D'Antonio. 2005. After biocontrol: assessing indirect effects of insect releases. Biological Control 35:307-318.

Dodd, A. P. 1940. The biological campaign against prickly pear. Commonwealth Prickly Pear Board Bulletin 1-177. Brisbane, Australia.

Hager, H. A. and K. D. McCoy. 1998. The implications of accepting untested hypotheses: a review of the effects of purple loosestrife (Lythrum salicaria) in North America. Biodiversity and Conservation 7:1069-1079.

Hight, S. D. and J. E. Carpenter. 2009. Flight phenology of male Cactoblastis cactorum (Lepidoptera: Pyralidae) at different latitudes in the southeastern United States. Florida Entomologist 92:208-216.

Hoffmann, J. H., V. C. Moran, and H. G. Zimmermann. 1999. Integrated management of Opuntia stricta (Haworth) Haworth (Cactaceae) in South Africa: an enhanced role for two renowned insect agents. African Entomology:15-20.

Holway, D. A., A. V. Suarez, and T. J. Case. 2002. Role of abiotic factors in governing susceptibility to invasion: A test with argentine ants. Ecology 83:1610-1619.

Hosking, J. R., P. R. Sullivan, and S. M. Welsby. 1994. Biological control of Opuntia stricta (Haw) Haw var. stricta using Dactylopius opuntiae (Cockerell) in an area of New South Wales, Australia, where Cactoblastis cactorum (Berg) is not a successful biological control agent. Agriculture Ecosystems & Environment 48:241-255.

Jezorek, H. A., P. D. Stiling, and J. E. Carpenter. 2010. Targets of an invasive species: oviposition preference and larval performance of Cactoblastis cactorum (Lepidoptera: Pyralidae) on 14 North American opuntioid cacti. Environmental Entomology 39:1884-1892.

121

Johnson, D. M. and P. D. Stiling. 1996. Host specificity of Cactoblastis cactorum (Lepidoptera: Pyralidae), an exotic Opuntia-feeding moth, in Florida. Environmental Entomology 25:743-748.

Johnson, D. M. and P. D. Stiling. 1998. Distribution and dispersal of Cactoblastis cactorum (Lepidoptera : Pyralidae), an exotic Opuntia-feeding moth, in Florida. Florida Entomologist 81:12-22.

Kenis, M., M. A. Auger-Rozenberg, A. Roques, L. Timms, C. Pere, M. Cock, J. Settele, S. Augustin, and C. Lopez-Vaamonde. 2009. Ecological effects of invasive alien insects. Biological Invasions 11:21-45.

Larson, B. M. H. 2007. An alien approach to invasive species: Objectivity and society in invasion biology. Biological Invasions 9:947-956.

Mack, R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. A. Bazzaz. 2000. Biotic invasions: Causes, epidemiology, global consequences, and control. Ecological Applications 10:689-710.

McFadyen, R. E. C. 1998. Biological control of weeds. Annual Review of Entomology 43:369-393.

Monro, J. M. 1975. Environmental variation and the efficiency of biological control - Cactoblastis in the southern hemisphere. Pages 204-212 in Kikkawa, J. and H.A. Nix (Eds.). Proceedings of the Ecological Society of Australia, Vol. 9. Managing Terrestrial Ecosystems. Symposium. Brisbane, Australia., May 15-16, 1975.

Morrison, L. W. 2002. Long-term impacts of an arthropod-community invasion by the imported fire ant, Solenopsis invicta. Ecology 83:2337-2345.

Morrison, L. W. and S. D. Porter. 2003. Positive association between densities of the red imported fire ant, Solenopsis invicta (Hymenoptera : Formicidae), and generalized ant and arthropod diversity. Environmental Entomology 32:548-554.

Niemela, J. and J. R. Spence. 1991. Distribution and abundance of an exotic ground beetle (Carabidae) - a test of community impact. Oikos 62:351-359.

Parker, I. M., D. Simberloff, W. M. Lonsdale, K. Goodell, M. Wonham, P. M. Kareiva, M. H. Williamson, B. Von Holle, P. B. Moyle, J. E. Byers, and L. Goldwasser. 1999. Impact: toward a framework for understanding the ecological effects of invaders. Biological Invasions 1:3-19.

Pemberton, R. W. and H. Liu. 2007. Control and persistence of native Opuntia on Nevis and St. Kitts 50 years after the introduction of Cactoblastis cactorum. Biological Control 41:272-282.

122

Pettey, F. W. 1948. The biological control of prickly pear in South Africa. Science Bulletin. Department of Agriculture of the Union of South Africa.

Phillips, C. B. and D. B. Baird. 2001. Geographic variation in egg load of Microctonus hyperodae (Hymenoptera : Braconidae) and its implications for biological control success. Biocontrol Science and Technology 11:371-380.

Pintor, L. M. and A. Sih. 2011. Scale dependent effects of native prey diversity, prey biomass and natural disturbance on the invasion success of an exotic predator. Biological Invasions 13:1357-1366.

Porter, S. D. and D. A. Savignano. 1990. Invasion of polygene fire ants decimates native ants and disrupts arthropod community. Ecology 71:2095-2106.

Robertson, H. G. and J. H. Hoffmann. 1989. Mortality and life-tables of Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae) compared on 2 host-plant species. Bulletin of Entomological Research 79:7-17.

Romeis, J., D. Babendreier, F. L. Wackers, and T. G. Shanower. 2005. Habitat and plant specificity of Trichogramma egg parasitoids: Underlying mechanisms and implications. Basic and Applied Ecology 6:215-236.

Rose, R., R. Weeks, and S. Usnick. 2011. Cactus moth, Cactoblastis cactorum, 2011 survey plan for PPQ and state cooperators. United States Department of Agriculture, Animal and Plant Health Inspection Service.

Ruesink, J. L., I. M. Parker, M. J. Groom, and P. M. Kareiva. 1995. Reducing the risks of nonindigenous species introductions - guilty until proven innocent. Bioscience 45:465-477.

Sagoff, M. 2005. Do non-native species threaten the natural environment? Journal of Agricultural & Environmental Ethics 18:215-236.

Schlaepfer, M. A., M. Pascal, and M. A. Davis. 2011. How might science misdirect policy? Insights into the threats and consequences of invasive species. Journal of Consumer Protection and Food Safety 6:27-31.

Shea, K., D. Kelly, A. W. Sheppard, and T. L. Woodburn. 2005. Context-dependent biological control of an invasive thistle. Ecology 86:3174-3181.

Simberloff, D. 2003. Confronting introduced species: A form of xenophobia? Biological Invasions 5:179-192.

Simberloff, D. 2005. Non-native species do threaten the natural environment! Journal of Agricultural & Environmental Ethics 18:595-607.

Snyder, W. E. and E. W. Evans. 2006. Ecological effects of invasive arthropod generalist predators. Annual Review of Ecology Evolution and Systematics 37:95-122. 123

Stiling, P. 2010. Death and decline of a rare cactus in Florida. Castanea 75:190-197.

Strauss, S. Y. 1988. Determining the effects of herbivory using naturally damaged plants. Ecology 69:1628-1630.

Strayer, D. L., V. T. Eviner, J. M. Jeschke, and M. L. Pace. 2006. Understanding the long-term effects of species invasions. Trends in Ecology & Evolution 21:645- 651.

Van Driesche, J. and R. Van Driesche. 2001. Guilty until proven innocent: Preventing nonnative species invasions. Conservation in Practice 2:8-18.

White, G. C. 1980. Current status of prickly pear control by Cactoblastis cactorum in Queensland. Pages 609-616 in Proceedings of the Vth International Symposium on Biological Control of Weeds, Brisbane, Australia.

Wilson, K. A., J. J. Magnuson, D. M. Lodge, A. M. Hill, T. K. Kratz, W. L. Perry, and T. V. Willis. 2004. A long-term rusty crayfish (Orconectes rusticus) invasion: Dispersal patterns and community change in a north temperate lake. Canadian Journal of Fisheries and Aquatic Sciences 61:2255-2266.

Zimmermann, H. G., S. Bloem, and H. Klein. 2004. Biology, history, threat, surveillance and control of the cactus moth, Cactoblastis cactorum. International Atomic Energy Agency/Food and Agriculture Organization, Vienna, Austria.

Zimmermann, H. G., V. C. Moran, and J. H. Hoffmann. 2000. The renowned cactus moth, Cactoblastis cactorum (Lepidoptera : Pyralidae): its natural history and threat to native Opuntia floras in Mexico and the United States of America. Florida Entomologist 84:543-551.

124

CHAPTER 5

LACK OF ASSOCIATIONAL EFFECTS BETWEEN TWO COMMON HOSTS ON

ATTACK RATES BY CACTOBLASTIS CACTORUM

Synopsis

Associational susceptibility occurs when herbivory on a target plant is increased due to the influence of plant neighbors. It has been suggested that associational susceptibility often arises when a plant neighbor is more palatable to a particular herbivore, resulting in attraction and spill over of the herbivore onto the target plant. We tested this prediction on a two common hosts of Cactoblastis cactorum in Florida: Opuntia humifusa, generally considered less preferred by C. cactorum, and O. stricta, considered to be the preferred host. We used a combination of observational surveys and experimental plantings to test for differences in C. cactorum and native herbivore damage and cladode growth on isolated O. humifusa plants, O. humifusa in interspecific pairs, and O. humifusa in conspecific pairs. We found that O.stricta had significantly more cladodes with C. cactorum larvae, eggsticks, and old damage than O. humifusa, lending support to our assumption that it is preferred by C. cactorum over O. humifusa. Three of the four native herbivores tested were found at higher levels on O.humifusa than O.stricta, suggesting that O.humifusa may experience decreased damage from these herbivores, that is, associational resistance, when paired with O. stricta. However, we found little support for associational resistance or susceptibility. Opuntia humifusa plants in

125 interspecific pairs did have more C. cactorum larvae and old damage than plants in conspecific pairs, but did not have more than isolated O. humifusa plants, meaning this result could be driven by density, as opposed to diversity, of host plants. Plants from the site used for the experimental plantings had little to no C. cactorum damage, regardless of plant neighbors, and revealed no associational effects for native herbivores. We believe that the lack of support for associational susceptibility may be because a spillover effect requires higher densities of herbivores, and lower densities of their preferred host species, than were present at our sites.

Introduction

The risk of herbivory for a plant is influenced by the diversity and spatial distribution of its surrounding plant community. A host plant may gain protection from its herbivores when found in a diverse, versus simple, community or when growing in close association with non-host or less palatable species. This phenomenon is termed „associational resistance‟ (hereafter AR) and was first used to describe reduced herbivory on crops grown in polycultures versus those grown in monocultures (Tahvanainen and Root 1972,

Root 1973). AR was later extended to include reduced herbivory, predation, or parasitism when a focal host/prey is associated with a more resistant or less preferred host/prey

(Wahl and Hay 1995, Hjalten and Price 1997, Ostfeld and Keesing 2000, LoGiudice et al. 2003, Stiling et al. 2003, Stenberg et al. 2007, Raffel et al. 2008). Conversely, if a focal host/prey experiences increased herbivory, predation, or parasitism due to association with a less resistant or more preferred host/prey, the phenomenon is termed

126

„associational susceptibility‟ (hereafter AS) (Thomas 1986, Brown and Ewel 1987, Wahl and Hay 1995, White and Whitham 2000). Multiple mechanisms and hypotheses have been proposed to explain AR and AS, including the resource concentration hypothesis

(Root 1973), enemies hypothesis (Root 1973), plant defense guilds (Tahvanainen and

Root 1972, Atsatt and Odowd 1976), plant „eavesdropping‟ (Karban 2001, 2007,

Himanen et al. 2010), shared natural enemies (or apparent competition) (Holt and

Lawton 1994), and biotic or abiotic habitat modification (White and Andow 2006,

Barbosa et al. 2009); these mechanisms and hypotheses have been reviewed by Russell

(1989), Andow (1991), Agrawal et al (2006), Barbosa et al (2009), and Letourneau et al

(2011).

Many of the mechanisms listed above could, in theory, lead to either AR or AS.

For example, a plant neighbor that serves as an attractant could reduce damage on the focal plant by drawing away shared herbivores (AR). This is the conceptual basis for trap cropping in agricultural systems, which involves mixing attractive, but less economically valuable, plant species with crops in order to lure herbivores away from the crops.

However, if an herbivore species increases in abundance due to feeding on the attractive neighbor(s), spill over onto the focal plant could occur, thereby increasing damage (AS).

In general, AR is more likely when a neighbor is less palatable relative to the focal plant (Pfister and Hay 1988, Hjalten and Price 1997, Barbosa et al. 2009), while AS is more likely when a neighbor is more palatable (Brown and Ewel 1987, White and

Whitham 2000). For example, White and Whitman (2000) found significantly higher cankerworm densities and percentage defoliation on potted cottonwoods placed under the worm‟s preferred host, box elder, then on isolated cottonwoods or cottonwoods placed

127 under conspecifics. Agrawal (2006) refines this prediction for specialist herbivores, by observing that for specialists, AS should become more likely as plant neighbors offer increasingly essential resources. However, actual occurrence of AR or AS will clearly be linked to other factors including herbivore density and motility (Brown and Ewel 1987,

White and Whitham 2000, Russell et al. 2007), presence of other herbivores and natural enemies (Stiling et al. 2003, Hamback et al. 2006, White and Andow 2006) and distance between focal and neighbor plants (White and Whitham 2000, Stiling and Moon 2001,

Stiling et al. 2004, Russell et al. 2007).

The primary objective of this study was to determine whether Opuntia humifusa, a host of the invasive cactus moth, Cactoblastis cactorum Berg (Lepidoptera: Pyralidae), experiences any associational effects when growing near to O. stricta, a second host that is generally considered more preferred than O. humifusa (Baker and Stiling 2009). For insect herbivores, AS has been more commonly observed than AR (Barbosa et al. 2009), and as described above, is generally expected when a neighbor is more preferred than a focal plant. Cactoblastis cactorum females normally disperse very little and tend to oviposit very near to their emergence site, leading to a clumped distribution of eggsticks

(Dodd 1940, Myers et al. 1981, Robertson 1987b, Zimmermann et al. 2004). The general observations and expectations for AS, coupled with the specific behaviors of C. cactorum females, suggest that spill over from O. stricta is likely to lead to AS for O. humifusa growing in close proximity. We therefore predicted that O. humifusa would have higher levels of C. cactorum damage when growing near O. stricta, as compared to damage levels when growing near conspecifics or isolated. Our secondary objective was to determine if there were differences in damage levels or abundance of native herbivores

128 that could be attributed to associational effects of the two Opuntia species. We examined four native herbivores: Chelinidea vittiger, Dactylopius confusus, Diaspis echinocacti, and Gerstaeckeria hubbardi (Table 5.1). Preferences for O. humifusa versus O. stricta have not been documented for any of these species, therefore specific predictions concerning the occurrence of AR or AS were not made. In order to address our two objectives, we conducted both observational surveys and experimental plantings.

Methods

Study system

Cactoblastis cactorum Berg (Lepidoptera: Pyralidae), is well known for its role in the biological control of Opuntia species, including the highly successful campaign in

Australia during the 1920s and 1930s, the less successful campaign begun in the late

1930s in South Africa, and the campaign against native cacti in the Caribbean begun in the 1950s (Dodd 1940, Pettey 1948, Bennett and Simmonds 1966, Zimmermann and

Moran 1991). However, since its discovery in the Florida Keys in 1989, the moth has become a potential threat to North American opuntioid biodiversity (Bennett and Habeck

1995, Zimmermann et al. 2000, Stiling 2002). Presently, C. cactorum is established in the southeastern United States as far north as Bull Island, South Carolina and as far west as southeastern Louisiana (Simonsen et al. 2008, USDA-APHIS-PPQ 2009). Female moths lay about 40-100 eggs one on top of the other to form an eggstick that is attached to the tip of a cactus spine or directly to a cladode. The larvae feed internally and gregariously on a number of opuntioid species. Larval damage results in hollowed out cladodes and can lead to secondary infections and in some cases, the death of entire

129 plants (Zimmermann et al. 2004). After completing six instars, larvae drop to the ground to pupate in the soil or in hollowed out cladodes. In central Florida, there are three non- overlapping flight periods per year: a spring flight beginning in mid- February, a summer flight beginning in early June, and a fall flight beginning in late August (Hight and

Carpenter 2009).

Four native species of Opuntia-feeding insects were included in damage surveys:

C.vittiger, D. confusus, D. echinocacti, and G. hubbardi. Diaspis echinocacti feeds on a variety of cactus species, while the remaining three species are specialists on opuntioid cacti (Hunter et al. 1912, Mann 1969). Dactylopius confusus and D. echinocacti are primarily sessile and live adult stages are therefore easily quantified. Chelinidea vittiger nymphs and adults are motile and G. hubbardi larvae feed internally, making accurate, direct counts of these two species more difficult. However, both leave characteristic signs of feeding (Table 5.1) which allow for quantification of damage even if insects are not directly observed.

Two Opuntia species native to Florida were used in this study, O. stricta and O. humifusa. Both are common hosts of C. cactorum, although O. stricta is attacked more often than O. humifusa (Baker and Stiling 2009). They exhibit slightly different morphologies; O. stricta has a more erect growth form and more spines, while O. humifusa has a low, spreading growth form and fewer spines. However, both species can range in size from a few cladodes to several hundred cladodes.

Surveys were conducted at Honeymoon Island State Park, Pinellas County

(hereafter, HI) and experimental plantings were conducted at Fort deSoto Park, Pinellas

County (hereafter, FDS). Both sites are part of Florida‟s west-central barrier island chain

130 and have similar plant and herbivore communities. O. stricta and O. humifusa are common in the dune and coastal scrub areas and C. cactorum is well-established.

Data collection and analysis

In September of 2006, 190 Opuntia plants were marked at HI as follows: Neighborhood type 1 = 50 O. humifusa alone; Neighborhood type 2 = 52 O. stricta alone;

Neighborhood type 3 = 44 interspecific pairs. Plants in “alone” neighborhoods were >5m from any other Opuntia individual, while plants in “paired” neighborhoods were ≤1m apart, but >5m from any other Opuntia individuals, with distances measured between the two closest cladodes of any two plants at the time of marking. All plants were surveyed in September 2006 and December 2006. At each survey, the following variables were recorded: number of live, non-woody cladodes, proportion of cladodes with active C. cactorum larvae, proportion of cladodes with old C. cactorum damage, number of C. cactorum eggsticks per cladode, and proportion of cladodes with D. confusus, D. echinocacti (live insects present), C. vittiger, and G. hubbardi ( feeding marks present).

Data were highly non-normal so to check for differences in herbivore damage between O. stricta and O. humifusa, a Mann-Whitney test was performed with species as the grouping factor. To check for associational effects on herbivore damage, a non- parametric ANOVA by ranks was used to compare between plant species/neighborhood combinations. All analyses were conducted using IBM SPSS Statistics 19 and where appropriate, Dunn‟s test for multiple contrasts was used for pair wise comparisons.

The design described above allows for testing of O. humifusa versus O. stricta for cladode number and herbivore damage levels, but does not allow for a comparison to

131 tease out any associational effects due to density of Opuntia, regardless of host species, versus associational effects due explicitly to the species of Opuntia. For example, if O. humifusa in interspecific pairs were found to have significantly more C. cactorum larvae than O. humifusa alone, the effect could be due to a higher density of hosts in the paired neighborhood or to a specific effect of O. stricta. Therefore a second set of surveys were conducted at HI in 2009 with 100 Opuntia plants marked as follows: Neighborhood type

1 =20 O. humifusa alone, Neighborhood type 3 = 20 interspecific pairs, and

Neighborhood type 4 =20 conspecific pairs. Distances for “alone” and “paired” neighborhoods were as described for 2006 surveys. All plants were surveyed in May,

July, and November of 2009 with the same variables as in 2006 recorded, plus maximum height. Data were again highly non-normal, so similar methods were used for analysis.

Herbivore-related variables were averaged over the three months, but as this survey encompassed the latter part of the Opuntia growth season, the change in height and proportional change in cladode number from May to November were used instead of the average. A non-parametric ANOVA by ranks was used to compare O. humifusa plants with plant neighborhood as the grouping factor.

Analysis of data in the above manner results in a “snapshot” assessment of associational effects. However, the time since establishment in a neighborhood was unknown for each plant and initial damage levels were not controlled for, so changes over time were not of primary interest for the observational surveys. In order to control for these factors we used Opuntia plants raised in a greenhouse at the University of South

Florida Botanical Garden for an out planting experiment. Cladodes from FDS were collected in April of 2007 and a single cladode was planted in a six inch pot using a mix

132 of sand and potting soil. Plants were fertilized twice a year, and watered as needed. By

April of 2009 the experimental (greenhouse raised) plants had multiple cladodes and were used to create two plant neighborhoods at FDS: Neighborhood type 3 = 18 interspecific groups, each consisting of two experimental O. stricta planted ≤0.5m from one naturally occurring O. humifusa, and Neighborhood type 4 = 16 conspecific pairs, each consisting of one experimental O. humifusa planted ≤0.5m from one naturally occurring O. humifusa. We also marked plants for Neighborhood type 1 = 15 naturally occurring O. humifusa alone. Experimental plants were of similar size (2-6 cladodes) and were initially free of damage from C. cactorum, D. confusus, C. vittiger, and G. hubbardi, but due to an outbreak in the greenhouse, all experimental plants had D. echinocacti. Naturally occurring plants were initially free of C. cactorum larvae and had little or no old C. cactorum damage (maximum=2 cladodes with old damage); initial damage levels of the four native herbivores varied as finding plants free of all types of native herbivores proved impossible. All plants were surveyed monthly from April 2009 to November 2009, and the same variables as in the 2009 observational survey were recorded. Although changes over time were of inherent interest, data could not be transformed to meet assumptions of repeated measures ANOVA. Instead, we calculated the difference in height, the proportional difference in cladode number, and the difference in each herbivore-related variable between the April and November surveys. A non-parametric ANOVA by ranks was used to compare O. humifusa plants with plant neighborhood as the grouping factor.

133

Results

Mann-Whitney tests of all O. stricta versus all O. humifusa from the 2006 survey showed that O. stricta plants had significantly higher proportions of cladodes with larvae and old damage, and more eggsticks per cladode. Opuntia humifusa had higher proportions of cladodes with C. vittiger, D. confusus, and G. hubbardi, as well as more total cladodes

(Table 5.2). ANOVA by ranks for plant species/neighborhood combinations were significant for all variables except proportion of cladodes with D. echinocacti, however, pair wise comparisons revealed that this was driven almost entirely by interspecific differences. Opuntia stricta alone and O. stricta in pairs were not significantly different for any variables. The sole intraspecific differences between Opuntia humifusa alone and

O. humifusa in pairs were for total cladodes (p=0.005) and proportion of cladodes with

G. hubbardi (p=0.000). In both cases O. humifusa alone had the highest value of the four groups and was significantly different than all other groups (Figs 5.1 and 5.2).

The 2009 data were used to test among O. humifusa plants alone, in conspecifc pairs, and in interspecific pairs. Plants that were in interspecific pairs had the greatest proportion of cladodes with C. cactorum larvae and old damage (Figs 5.3 and 5.4). The difference between neighborhoods was significant for old damage (p=0.036), but was marginal for larvae (p=0.080). Pair wise comparisons for old damage showed that O. humifusa from interspecific pairs had significantly more old damage than O. humifusa in conspecific pairs (adjusted p=0.030), but neither neighborhood was significantly different than O. humifusa alone (Fig 5.4). Opuntia humifusa plants alone added significantly more height than O. humifusa in interspecific or conspecific pairs (adjusted p=0.004 for both pair wise comparisons), yet lost proportionally more cladodes than the

134 paired plants, although this result was not significant. No other variables were significantly different among the three plant neighborhoods.

Cactoblastis cactorum damage levels were extremely low for the FDS site. Over the seven months, only one plant had active larvae and only one eggstick was found.

Eight plants showed old damage, but half of these were plants with 1-2 cladodes of pre- existing old damage. Therefore, C. cactorum variables were not included in the analysis.

Results from the ANOVA by ranks indicated significant differences between plant neighborhoods for change in height, proportional change in cladodes, and change in D. echinocacti damage (p=0.016, 0.023, and 0.004, respectively). Further examination revealed that results for the latter two variables were driven solely by differences between experimentally planted O. humifusa and naturally occurring O. humifusa. The experimentally planted O. humifusa had a significantly higher proportional change in cladodes when compared to naturally occurring O. humifusa alone, in conspecific pairs, and in interspecific pairs (adjusted p=0.000 for all three pair wise comparisons).

Experimental O. humifusa also significantly decreased the proportion of cladodes with D. echinocati as compared to all three groups of naturally occurring plants (adjusted p=0.000 for all three pair wise comparisons). Experimental O. humifusa had D. echinocacti colonies present at the time of planting and the large decrease in cladodes with D. echinocacti reflects the insects‟ tendency to reach much higher densities in greenhouse or plantation settings than in naturally occurring Opuntia populations

(Hunter et al. 1912, H. Jezorek personal observation).

135

Table 5.1 The four native Opuntia herbivores examined in the study

Insect species Order: Family Common Description of feeding/damage Name

Chelinidea Hemiptera: Cactus bug Mobile sap sucker; leaves pale half vittiger McAtee Coridae or full circles on surface of cladode; severe damage can lead to desiccation of cladode

Dactylopius Hemiptera: Cochineal Sessile sap sucker; exudes fluffy confusus Dactylopiidae white substance on surface of Cockerell cladodes

Diaspis Hemiptera: Cactus Sessile sap sucker; circular, slightly echinocacti Diaspididae scale convex, whitish to tan scale; severe Bouché damage causes chlorotic patches and desiccation of cladode

Gerstaeckeria Coleoptera: Cactus Internal feeder; leaves light brown hubbardi Curculionidae weevil circular cell; cladode appears pierced LeConte by tiny “bullet hole”

136

Table 5.2 Results from a Mann-Whitney test for differences between O. humifusa (N=94) and O. stricta (N=96). Plants were surveyed in September and December of 2006 and values for all variables were averaged across the two surveys prior to analysis. Negative values of the standardized test statistic indicate that the variable was greater for O. humifusa, positive values indicate that the variable was higher for O. stricta.

Variable O. humifusa rank O. stricta rank Standardized test p statistic

Total cladodes 104.27 86.91 -2.176 0.030

C. cactorum larvae a 82.45 108.28 4.54 0.000

C. cactorum old damage a 74.80 115.79 5.62 0.000

Eggsticks/cladode 92.50 98.44 2.46 0.014

D. confusus a 107.63 83.62 -5.04 0.000

G. hubbardi a 121.26 70.28 -7.19 0.000

C.vittiger a 121.30 70.23 -6.56 0.000

D. echinocacti a 97.85 93.20 -0.85 0.394

a Values for these variables were recorded as the proportion of cladodes on a plant containing the herbivore or its characteristic damage

137

a b b b

(50) (52) (44) (44)

Figure 5.1: Total cladodes for the four plant species/neighborhood groups surveyed at HM in 2006. Boxes indicate interquartile ranges, bold lines indicate medians, whiskers indicate inner fences, points indicate outliers, and asterisks indicate extreme values. Numbers in parentheses represent sample sizes. Different lower case letters to the left of boxes represent significant pair wise comparisons according to Dunn‟s test.

138

a b c bc

(50) (52) (44) (44)

Figure 5.2: Proportion of cladodes with damage from the weevil G. hubbardi for the four plant species/neighborhood groups surveyed at HM in 2006. Boxes indicate interquartile ranges, bold lines indicate medians, whiskers indicate inner fences, points indicate outliers, and asterisks indicate extreme values. Numbers in parentheses represent sample sizes. Different lower case letters to the left of boxes represent significant pair wise comparisons according to Dunn‟s test.

139

(20) (40) (20)

Figure 5.3: Proportion of cladodes with C. cactorum larvae from the three O. humifusa neighborhoods surveyed at HM in 2009. Boxes indicate interquartile ranges, bold lines indicate medians, whiskers indicate inner fences, points indicate outliers, and asterisks indicate extreme values. Numbers in parentheses represent sample sizes.

140

a ab b

(20) (40) (20)

Figure 5.4: Proportion of cladodes with old C. cactorum damage from the three O. humifusa neighborhoods surveyed at HM in 2009. Boxes indicate interquartile ranges, bold lines indicate medians, whiskers indicate inner fences, points indicate outliers, and asterisks indicate extreme values. Numbers in parentheses represent sample sizes. Different lower case letters to the left of boxes represent significant pair wise comparisons according to Dunn‟s test.

141

Discussion

Our findings agreed with previous reports that O. stricta is preferred over O. humifusa by

C. cactorum, as O. stricta plants had a higher proportion of damaged cladodes and cladodes with larvae present. In cases where a neighboring species is preferred by an herbivore, the expectation is of AS for the focal species, provided that focal species is an acceptable host. However, we found very little evidence of AS to C. cactorum for O. humifusa growing in close association with O. stricta. The 2006 survey revealed no intraspecific differences for C. cactorum-related variables, while the 2009 survey produced limited support for our prediction of AS. From the 2009 survey, we found that

O. humifusa plants paired with O. stricta had higher proportions of cladodes with C. cactorum larvae and old damage than did O. humifusa plants alone or paired with conspecifics. Although the result for larvae was marginal, we interpreted these findings as partial support for our prediction of AS because observation of active larvae is infrequent compared to that of old damage, leading to lower power for the larval variable. The significant difference for old damage between plants in interspecific and conspecifc pairs is of particular interest because the comparison between these two neighborhoods controls for density of host plants, making it more likely that the observed difference was due to diversity of hosts (that is, to the specific effect of O. stricta).

However, a significant difference between the interspecific plants and the alone plants, which would have strengthened the support for AS, was not observed.

Our study found significantly higher levels of the native insects C.vittiger, D. confusus, and G. hubbardi on O. humifusa than on O. stricta. Although the reason for C. cactorum‟s preference for O. stricta over O. humifusa is unknown, native herbivores may 142 play a role. Plants that have actively feeding herbivores, or eggs of herbivores, on them can produce feeding- or oviposition-induced surface compounds and volatiles that serve to repel gravid females from oviposition (Blaakmeer et al. 1994, Harmon et al. 2003,

Hilker and Meiners 2011). Cactoblastis cactorum females are known to respond to slight variations in the CO2 concentration at the boundary layer of plants, a behavior hypothesized to enable them to choose more actively growing, healthier plants for their offspring (Stange et al. 1995, Stange 1997). It is also possible that they respond to induced surface compounds or volatiles to avoid Opuntia individuals that are relatively more damaged and seek out individuals that are relatively less damaged. Given that all four of the native insects we examined were found in higher levels on O. humifusa than

O. stricta (with three of the four being significantly higher), such behavior in our system would result in C. cactorum females choosing an O. stricta individual more often than an

O. humifusa individual.

If, based on our finding for G. hubbardi damage, we assume it prefers O. humifusa over O. stricta, we would expect to see AR for O. humifusa when paired with

O. stricta (Barbosa et al. 2009). Indeed, the 2006 data showed that O. humifusa plants alone had a higher proportion of cladodes with G. hubbardi damage than did O. humifusa in interspecific pairs. In theory, this could result from O. stricta acting as a repellent/masking plant (Atsatt and Odowd 1976) by interfering with the ability of G. hubbardi to locate its preferred host , or from spill over from O. humifusa onto O. stricta

(Agrawal et al. 2006, Barbosa et al. 2009). If the mechanism at work were repellency/masking, we would not necessarily expect a difference between O. stricta alone and in interspecific pairs. However, if spill over were occurring, we would expect

143 higher levels of G. hubbardi on O. stricta in interspecific pairs than O. stricta alone, as it would be receiving AS from being paired with O. humifusa. Our actual results cannot distinguish the mechanism for three reasons. First, although O. stricta in pairs did have higher G. hubbardi damage than O. stricta alone, the difference was not significant

(p=0.126). Second, we did not mark conspecific pairs of O. humifusa in 2006, so we cannot rule out a density effect, whereby G. hubbardi damage is diluted or spread out when host plants are more densely arrayed, regardless of their species. Third, we did not find any evidence for associational effects of G. hubbardi from our 2009 survey or from the experimental plantings.

Although our study did not set out to document direct competition between the two Opuntia species, the finding from the 2006 survey that O. humifusa plants alone had significantly more cladodes than O. humifusa plants in interspecific pairs raises this possibility by suggesting that plants growing alone are released from this competition and can grow larger. However, the same effect was not seen for O. stricta plants alone versus those in interspecific pairs. Further, the 2009 survey produced a contradictory result in that O. humifusa plants alone actually lost proportionally more cladodes than those in conspecifc or interspecific pairs. Exploration of direct and indirect competition between the two Opuntia species is a topic for future studies dedicated expressively to that purpose.

There are several possible explanations for the lack of evidence for AS/AR, one of which is our choice of distances when determining plant neighborhoods. Russell et al

(2007) found that egg loads of Rhinocyllus conicus on the native thistle Cirsium undulatum decreased significantly as distance from Carduus nutans, the invasive musk

144 thistle, increased. Similarly, White and Whitham (2002) found that both cankerworm densities and defoliation rates on cottonwoods decreased as a function of distance from box elder, the worm‟s preferred host. We based our distances on the behavior of C. cactorum, as documenting AS to this herbivore was our primary objective. Larvae are able to disperse very short distances, but only do so when no cladodes remain on their original host (Dodd 1940). Females are reluctant to fly and seldom disperse from their emergence site prior to first oviposition (Myers et al. 1981, Robertson 1987a). They will disperse over larger distances, but only as hosts become sparser; the longest flight recorded is 24km (Zimmermann et al. 2004). We reasoned that plants >5m from another

Opuntia would be sufficiently isolated, but now consider that this distance may not be large enough to observe differences between alone plants and paired plants. Stiling et al

(2004) found no significant difference in C. cactorum attack rates for O. corallicola plants ≤5m from an O. stricta and those >20m from an O. stricta. However, they found that O. corallicola plants that were ≥500m from an O. stricta suffered no attacks by C. cactorum. Perhaps the distance needed for a host plant to be truly isolated in terms of normal C. cactorum dispersal is more on the order of tens or hundreds of meters. At our sites, there are few to no plants that meet this criterion, but it is possible that associational effects for C. cactorum could be documented at other, larger sites.

The occurrence of a spillover effect depends on the herbivore in question reaching a high enough density to actually begin using the less preferred host. For example, White and Whitham (2000) note that AS of cottonwoods, due to spillover from the attractant “sink” plant, box elder, was only seen in areas with high cankerworm densities. In areas with low densities, cankerworm completed their full life cycle on box

145 elder. Although we observed C. cactorum and native insect damage on both Opuntia species, the density of these herbivores may simply be too low to result in a true spillover effect onto their less preferred hosts. In the future, if the density of any of the herbivores increases, or the density of one of their host species decreases, associational effects could be more readily documented. Findings from the previous chapter show that mortality of

O. stricta is higher than that of O. humifusa, and that O. humifusa plants are more likely to survive a C. cactorum attack. If O. stricta plants are dying faster than they are being recruited, they may become scarce enough to consistently suffer complete “defoliation” by C. cactorum within a single generation, a phenomenon which is currently rarely seen.

If this were the case, O. humifusa plants close to O. stricta could end up as the most likely oviposition sites for the next generation of moths.

The phenomenon of AR and AS for mixtures of host plants with shared herbivores has been fairly well-documented. Despite this, we found, at best, weak support for our prediction of AS to C. cactorum for O. humifusa plants near O. stricta and virtually no support for AR from or AS to four native opuntioid-feeding insects.

While the lack of support could be attributed to an imperfect experimental design, we feel it is more likely due to ecological factors, particularly such as host plant and herbivore density. Future studies that explicitly measure densities of Opuntia plants and their associated herbivores are needed to confirm this, but we can cautiously conclude that at our sites, the herbivore species are sparse enough, and both Opuntia species are dense enough, for associational effects to be non-existent or undetectable.

146

Acknowledgements

We would like to thank Natalie and Brandon O‟Brien and Diane Harshbarger for their help with field work, Terry Hingtgen (Florida Department of Environmental Protection,

Southwest Division) for his assistance in acquiring the necessary permits for field work at Honeymoon Island State Park, and the employees of Fort deSoto County Park for their support during the course of the study.

References cited

Agrawal, A. A., J. A. Lau, and P. A. Hamback. 2006. Community heterogeneity and the evolution of interactions between plants and insect herbivores. Quarterly Review of Biology 81:349-376.

Andow, D. A. 1991. Vegetational diversity and arthropod population response. Annual Review of Entomology 36:561-586.

Atsatt, P. R. and D. J. Odowd. 1976. Plant defense guilds. Science 193:24-29.

Barbosa, P., J. Hines, I. Kaplan, H. Martinson, A. Szczepaniec, and Z. Szendrei. 2009. Associational resistance and associational susceptibility: Having right or wrong neighbors. Annual Review of Ecology Evolution and Systematics 40:1-20.

Bennett, F. D. and F. J. Simmonds. 1966. Biological control of Opuntia spp. by Cactoblastis cactorum in the Leeward Islands (West Indies). BioControl 11:183- 189.

147

Blaakmeer, A., D. Hagenbeek, T. A. Vanbeek, A. E. Degroot, L. M. Schoonhoven, and J. J. A. Vanloon. 1994. Plant response to eggs vs. host marking pheromone as factors inhibiting oviposition by Pieris brassicae. Journal of Chemical Ecology 20:1657-1665.

Brown, B. J. and J. J. Ewel. 1987. Herbivory in complex and simple tropical successional ecosystems. Ecology 68:108-116.

Dodd, A. P. 1940. The biological campaign against prickly pear. Commonwealth Prickly Pear Board Bulletin:1-177. Brisbane, Australia.

Hamback, P. A., J. A. Stenberg, and L. Ericson. 2006. Asymmetric indirect interactions mediated by a shared parasitoid: connecting species traits and local distribution patterns for two chrysomelid beetles. Oecologia 148:475-481.

Harmon, J. P., J. A. White, and D. A. Andow. 2003. Oviposition behavior of Ostrinia nubilalis (Lepidoptera : Crambidae) in response to potential intra- and interspecific interactions. Environmental Entomology 32:334-339.

Hilker, M. and T. Meiners. 2011. Plants and insect eggs: How do they affect each other? Phytochemistry 72:1612-1623.

Himanen, S. J., J. D. Blande, T. Klemola, J. Pulkkinen, J. Heijari, and J. K. Holopainen. 2010. Birch (Betula spp.) leaves adsorb and re-release volatiles specific to neighbouring plants - a mechanism for associational herbivore resistance? New Phytologist 186:722-732.

Hjalten, J. and P. W. Price. 1997. Can plants gain protection from herbivory by association with unpalatable neighbours? A field experiment in a willow-sawfly system. Oikos 78:317-322.

Holt, R. D. and J. H. Lawton. 1994. The ecological consequences of shared natural enemies. Annual Review of Ecology and Systematics 25:495-520.

Hunter, W. D., F. C. Pratt, and J. D. Mitchell. 1912. The principle cactus insects of the United States. USDA, Bureau of Entomology.

Karban, R. 2001. Communication between sagebrush and wild tobacco in the field. Biochemical Systematics and Ecology 29:995-1005.

148

Karban, R. 2007. Associational resistance for mule's ears with sagebrush neighbors. Plant Ecology 191:295-303.

Letourneau, D. K., I. Armbrecht, B. Salguero Rivera, J. Montoya Lerma, E. Jimenez Carmona, M. Constanza Daza, S. Escobar, V. Galindo, C. Gutierrez, S. Duque Lopez, J. Lopez Mejia, A. M. Acosta Rangel, J. Herrera Rangel, L. Rivera, C. Arturo Saavedra, A. Marina Torres, and A. Reyes Trujillo. 2011. Does plant diversity benefit agroecosystems? A synthetic review. Ecological Applications 21:9-21.

LoGiudice, K., R. S. Ostfeld, K. A. Schmidt, and F. Keesing. 2003. The ecology of infectious disease: Effects of host diversity and community composition on Lyme disease risk. Proceedings of the National Academy of Sciences of the United States of America 100:567-571.

Mann, J. 1969. Cactus-feeding Insects and Mites. Smithsonian Institution Bulletin 256, Washington, D.C.

Myers, J. H., J. Monro, and N. Murray. 1981. Egg clumping, host plant selection and population regulation in Cactoblastis cactorum (Lepidoptera). Oecologia 51:7-13.

Ostfeld, R. S. and F. Keesing. 2000. Biodiversity and disease risk: The case of Lyme disease. Conservation Biology 14:722-728.

Pettey, F. W. 1948. The biological control of prickly pear in South Africa. Science Bulletin. Department of Agriculture of the Union of South Africa.

Pfister, C. A. and M. E. Hay. 1988. Associational plant refuges - Convergent patterns in marine and terrestrial communities result from differing mechanisms. Oecologia 77:118-129.

Raffel, T. R., L. B. Martin, and J. R. Rohr. 2008. Parasites as predators: Unifying natural enemy ecology. Trends in Ecology & Evolution 23:610-618.

Robertson, H. G. 1987. Oviposition site selection in Cactoblastis cactorum (Lepidoptera): Constraints and compromises. Oecologia 73:601-608.

Root, R. B. 1973. Organization of a plant-arthropod association in simple and diverse habitats – Fauna of collards (Brassica-Oleracea). Ecological Monographs 43:95- 120.

Russell, E. P. 1989. Enemies hypothesis – A review of the effect of vegetational diversity on predatory insects and parasitoids. Environmental Entomology 18:590-599.

149

Russell, F. L., S. M. Louda, T. A. Rand, and S. D. Kachman. 2007. Variation in herbivore-mediated indirect effects of an invasive plant on a native plant. Ecology 88:413-423.

Simonsen, T. J., R. L. Brown, and F. A. H. Sperling. 2008. Tracing an invasion: Phylogeography of Cactoblastis cactorum (Lepidoptera : Pyralidae) in the United States based on mitochondrial DNA. Annals of the Entomological Society of America 101:899-905.

Stange, G. 1997. Effects of changes in atmospheric carbon dioxide on the location of hosts by the moth, Cactoblastis cactorum. Oecologia 110:539-545.

Stange, G., J. Monro, S. Stowe, and C. B. Osmond. 1995. The CO2 sense of the moth Cactoblastis cactorum and its probable role in the biological control of the CAM plant Opuntia stricta. Oecologia 102:341-352.

Stenberg, J. A., J. Heijari, J. K. Holopainen, and L. Ericson. 2007. Presence of Lythrum salicaria enhances the bodyguard effects of the parasitoid Asecodes mento for Filipendula ulmaria. Oikos 116:482-490.

Stiling, P. 2002. Potential nontarget effects of a biological control agent, prickly pear moth, Cactoblastis cactorum (Berg) (Lepidoptera: Pyralidae), in North America, and possible management actions. Biological Invasions 4:273-281.

Stiling, P., D. Moon, and D. Gordon. 2004. Endangered cactus restoration: Mitigating the non-target effects of a biological control agent (Cactoblastis cactorum) in Florida. Restoration Ecology 12:605-610.

Stiling, P. and D. C. Moon. 2001. Protecting rare Florida cacti from attack by the exotic cactus moth, Cactoblastis cactorum (Lepidoptera : Pyralidae). Florida Entomologist 84:506-509.

Stiling, P., A. M. Rossi, and M. V. Cattell. 2003. Associational resistance mediated by natural enemies. Ecological Entomology 28:587-592.

Tahvanainennen, J. O. and R. B. Root. 1972. Influence of vegetational diversity on population ecology of a specialized herbivore, Phyllotreta cruciferae (Coleoptera-Chrysomelidae). Oecologia 10:321-&.

Thomas, C. D. 1986. Butterfly larvae reduce host plant survival in vicinity of alternative host species. Oecologia 70:113-117.

150

USDA-APHIS-PPQ. 2009. Cactus moth program annual report, Eastern region. United States Department of Agriculture, Animal and Plant Health Inspection Service, Plant Protection and Quarantine.

Wahl, M. and M. E. Hay. 1995. Associational resistance and shared doom – Effects of epibiosis on herbivory. Oecologia 102:329-340.

White, J. A. and D. A. Andow. 2006. Habitat modification contributes to associational resistance between herbivores. Oecologia 148:482-490.

White, J. A. and T. G. Whitham. 2000. Associational susceptibility of cottonwood to a box elder herbivore. Ecology 81:1795-1803.

Zimmermann, H. G. and V. C. Moran. 1991. Biological control of prickly pear, Opuntia ficus-indica (Cactaceae), in South Africa. Agriculture Ecosystems & Environment 37:29-35.

151

APPENDIX I

TAXONOMIC INFORMATION FOR OPUNTIOIDEAE TRIBES

CYLINDROPUNTIEAE AND OPUNTIEAE

Family: Cactaceae

Subfamily: Opuntioideae

Tribe: Cylindropuntieae

Cylindropuntia acanthocarpa (Engelmann and J.M. Bigelow) F.M. Knuth native to US (AZ, CA, NM, NV, UT) and Mexico

*Cylindropuntia spinosior (Engelmann) F.M. Knuth native to US (AZ, NM) and Mexico

Tribe Opuntieae

Consolea rubescens (Salm-Dyck ex de Candolle) Lemaire native to West Indies and Caribbean; said to be naturalized in AZ, CA, NV

Nopalea cochenillifera (L.) Salm-Dyck introduced in US (FL); cultivated in West Indies, Caribbean, Mexico, Central America; probable origin is Mexico or Central America (Panama)

*Opuntia dillenii (Ker Gawler) Haworth native to US (FL, SC,), West Indies, Central America, Mexico; introduced Australia, S. Africa, Madagascar

152

APPENDIX I (continued)

Opuntia engelmannii Salm-Dyck ex Engelmann var. engelmannii native to US (AZ, CA, NV, NM, T, UT) and Mexico

*Opuntia engelmannii Salm-Dyck ex Engelmann var. linguiformis (Griffiths)

Parfitt and Pinkava reportedly native in Bexar County, TX, but possibly extirpated; cultivated and occasionally naturalized in AZ, TX, Mexico

Opuntia engelmannii Salm-Dyck ex Engelmann var. lindheimeri (Engelmann)

Parfitt and Pinkava native to US (LA, MO, MS, NM, OK, TX) and Mexico; introduced S. Africa

Opuntia ficus-indica (L.) Miller native to Mexico; introduced in US (reports of naturalized populations in AL, AZ, CA,

FL, GA, HI, NM, TX); cultivated worldwide

Opuntia leucotricha de Candolle native to Mexico; naturalized in FL

*Opuntia macrocentra Engelmann native to US (AZ, NM, TX) and Mexico

153

Opuntia microdasys (Lehmann) Pfeiffer native to Mexico; introduced in US (cultivated and naturalized in AZ)

Opuntia santa-rita (Griffiths and Hare) Rose native to US (AZ , NM, and TX) and Mexico

Opuntia streptacantha Lemaire native to northern and central Mexico; widely cultivated in Mexico

* These species are of conservation concern in their native range

Taxonomic, distribution, and conservation information obtained from:

Anderson, E.F. 2001. The Cactus Family. Timber Press, Portland, OR. http://plants.usda.gov http://www.ars-grin.gov http://www.natureserve.org

154