Ecological Factors Influencing Community Composition of Leaf-tying Caterpillars
By Elisha Sigmon Rubin
B.S. in Biology, May 2008 University of North Carolina at Asheville
A Dissertation submitted to The Faculty of Columbian College of Arts and Sciences of The George Washington University in partial fulfillment of the requirements for the degree of Doctor of Philosophy
August 31, 2013
Dissertation directed by
John T. Lill Associate Professor of Biological Sciences
The Columbian College of Arts and Sciences of The George Washington University certifies that Elisha Sigmon Rubin has passed the Final Examination for the degree of
Doctor of Philosophy as of May 2, 2013. This is the final and approved form of the dissertation.
Ecological Factors Influencing Community Composition of Leaf-tying Caterpillars
By Elisha Sigmon Rubin
Dissertation Research Committee:
John T. Lill, Associate Professor of Biological Sciences, Dissertation
Director
Gina M. Wimp, Assistant Professor of Biology, Georgetown University,
Committee Member
Scott Powell, Assistant Professor of Biological Sciences, Committee
Member
ii
© Copyright 2013 by Elisha Sigmon Rubin All rights reserved
iii
Dedication
To my husband, Dustin Rubin, for never-ending support
iv
Acknowledgements
First and foremost I would like to thank my advisor, Dr. John Lill, for teaching me to be a better scientist, writer, mentor, and person. John’s enthusiasm and constant positive attitude make him a joy to work with. His excitement about my research is always encouraging and his thoughtful insights help me find the meaning within the data.
His belief in me makes me believe in myself.
I’d also like to thank my committee members, Gina Wimp, Martha Weiss, Scott
Powell and Amy Zanne. Gina and Martha have been extremely helpful in both designing experiments and writing this dissertation. A special thanks to Gina for all of the statistics lessons, both in and out of the classroom. And thank you to Adam Smith for serving as the defense moderator.
This work would not have been possible without all the assistance from the Lill lab. Teresa Stoepler and Mariana Abarca Zama are terrific lab-sisters. They are always there for me with research advice, assistance in the field and lab, encyclopedic knowledge of ecology, and a shoulder to lean on. Michelle Sliwinski is the best mentee I could have. She is organized and proactive, which makes my life so much easier. I could never have accomplished all of this research without field and lab assistance from
Shannon Murphy, Lillian Power, Victoria Fiorentino, Kylee Grenis, Megan Eurle, Patrick
Lill, Robert Oppenheimer, Arjun Aswathi, Katie Costantini, and Luke Fey.
I’d like to thank John Brown for identifying Tortricid moths and Gustavo
Hormiga and Ligia Benavidas for identifying spiders. Thanks to Dan Gruner for teaching me R. Many thanks to the members of the DC Plant-Insect Group for helpful comments on manuscripts and the always-stimulating conversations.
v
Finally, I’d like to thank my friends and family for their support and encouragement, even when they have no idea what I’m talking about. My husband,
Dustin, has always supported me and celebrates every step of this process, making even the tiniest accomplishment feel important. My parents, brother, and extended family are extremely supportive and always try to understand my research. Thanks to all my friends, especially my fellow GWU graduate students, for the love, encouragement, and much- needed relaxation.
Financial support for this work was provided by George Washington University
Department of Biological Sciences (Weintraub Fellowship, Harlan Fellowship and the
Mortensen Fund), The Washington Biologists Field Club Research Award, and National
Science Foundation Doctoral Dissertation Improvement Grant (DEB-1210600).
vi
Abstract of Dissertation
Ecological Factors Influencing Community Composition of Leaf-tying Caterpillars
The composition of ecological communities is determined by a variety of factors that span multiple scales, from interactions between individuals to the landscape characteristics. Leaf-tying caterpillars that build shelters between leaves act as physical ecosystem engineers—species that influence local community composition via modifications of the abiotic environment. Other arthropods secondarily inhabit leaf ties, which leads to an increase in arthropod abundance and diversity on the plant. My dissertation examined the composition of the arthropod community within leaf ties at both small and large scales as well as the mechanisms driving compositional differences.
The arthropod community occupying leaf ties was examined throughout a summer season on two host plants, American beech and white oak. Weekly censuses were combined into three time periods: early, middle and late summer. Diversity measures for leaf-tying caterpillars and the entire arthropod community within ties varied little between tree species and time periods. However, density of leaf-tiers and all arthropods was significantly higher on white oak than beech and density increased on both tree species as the season progressed. The composition of the leaf-tying caterpillar community and the arthropod community as a whole differed between host tree species and sampling periods. Though the arthropod communities on American beech and white oak differed, they showed similar patterns of compositional turnover, with distinct communities in early and late summer and a transitional community mid-summer.
vii
The impact of leaf ties on the arthropod community is well-documented, but interactions within the ties have largely been ignored. Using artificial leaf ties composed of one leaf and a piece of transparency paper, I compared the behavior of four species of common oak leaf-tying caterpillars when competing over an existing shelter. The caterpillars were observed to push and hit each other with their head until one retreated from the shelter. The occupant caterpillar, which built the shelter, was more likely to maintain possession of its shelter than lose or share possession, especially when it was larger than the intruding caterpillar. The four species examined differed significantly in their behavior toward other caterpillars with Psilocorsis cryptolechiella and P. reflexella being more aggressive than P. quercicella and Pseudotelphusa quercinigracella. There were also behavioral differences within and between species when acting as an occupant or an intruder. Psilocorsis cryptolechiella was more aggressive and P. quercinigracella was less aggressive when defending a shelter than when attempting to usurp a shelter.
Leaf-tying caterpillars frequently engage in direct, physical competition for shelters that serve as a territory as well as a food source.
The behavioral differences found between species led to speculation that the caterpillar species that initially builds a shelter may influence which arthropods colonize that shelter. Many groups of organisms engineer their habitat in a similar way, but generate different outcomes, which can affect the assembly process of the community associated with an engineer. In the field, I created initially empty control ties and experimental ties initially occupied by one of three species of caterpillar and allowed the ties to be naturally colonized for one or two weeks to examine the initial stages of community assembly when the caterpillar is most likely to interact with potential
viii
colonists. The community changed drastically from one to two weeks, with more arthropods and more species per tie after two weeks. While there were no differences between communities within ties initially built by different species, different communities assembled in ties initially containing a caterpillar compared to initially empty controls. There were also compositional differences in the communities within leaf ties where the initial occupant remained in the tie throughout the study compared to ties that had been abandoned. This indicates that the arthropod community inhabiting leaf ties responds dynamically to the presence of a leaf-tying caterpillar, but not to its species identity.
ix
Table of Contents
Dedication…………………………………………………………………...... iv
Acknowledgments…………………………………………………………………………v
Abstract of Dissertation………………………………………………………...………..vii
Table of Contents……………………………………………………………………….....x
List of Figures………………………………………………………………………….....xi
List of Tables………………………………………………………………………….....xii
General Introduction………….…………………………………………………………...1
Chapter 1: Phenological variation in the composition of a temperate forest leaf tie community..….……………………..……………………………………………………11
Chapter 2: Interspecific variation in aggressive fighting behavior of shelter-building caterpillars………………………………………………………………………………..42
Chapter 3: Caterpillar presence, but not species identity, influences arthropod community assembly in leaf shelters...…………………………………………………..66
General Conclusion..……………………………………………………………………105
Appendices……………………………………………………………………………...111
x
List of Figures
Figure 1.1………………………………………………………………………………...36
Figure 1.2………………………………………………………………………………...37
Figure 1.3………………………………………………………………………………...38
Figure 1.4………………………………………………………………………………...39
Figure 1.5………………………………………………………………………………...40
Figure 1.6………………………………………………………………………………...41
Figure 2.1………………………………………………………………………………...63
Figure 2.2………………………………………………………………………………...64
Figure 2.3………………………………………………………………………………...65
Figure 3.1………………………………………………………………………………...99
Figure 3.2……………………………………………………………………………….100
Figure 3.3……………………………………………………………………………….101
Figure 3.4……………………………………………………………………………….102
Figure 3.5……………………………………………………………………………….103
Figure 3.6……………………………………………………………………………….104
xi
List of Tables
Table 2.1…………………………………………………………………………………61
Table 3.1………………………………………………………………………………....96
xii
GENERAL INTRODUCTION
Understanding the determinants of community composition has been a central aim of ecologists for over a century (Clements 1916, Connell and Slatyer 1977, Connor and
Simberloff 1979, Summerville et al. 2003). Abiotic factors largely determine which species can survive in a particular habitat and biotic factors influence how species interact and thrive (Keddy 1992, Poff 1997, Weiher et al. 2011). Competition, facilitation and trophic interactions shape how members of a community respond to one another and thus how the community is composed. Understanding these interactions and their outcomes can shed light on the processes that determine community composition.
Ecosystem engineers are organisms that modify their abiotic environment in various ways, which in turn affects associated biota as well as their interactions (Jones et al. 1994, 1997). Caterpillars in at least 24 families of Lepidoptera function as ecosystem engineers by tying leaves together with silk to create shelters (Lill and Marquis 2007).
Leaf ties are secondarily inhabited by a variety of arthropods that benefit from altered microclimate (Willmer 1980, Joos et al. 1988, Hunter and Willmer 1989), decreased toxic effects of foliage consumption (Sandberg and Berenbaum 1989, Sagers 1992, Fukui et al.
2002), and protection from a variety of natural enemies (Damman 1987, Atlegrim 1992,
Loeffler 1996, Eubanks et al. 1997, Sipura 1999, Abarca and Boege 2011). Shelter colonists include other herbivores (e.g., shelter-building and non-shelter-building caterpillars, beetles, aphids, hemipterans), predators (e.g., spiders, beetles, centipedes, hemipterans), and scavengers (e.g., thrips, beetles, psocids, collembola). The presence of leaf shelters increases arthropod abundance and species richness 3 to 10-fold when compared to non-shelter leaves, and significantly alters arthropod community structure at
1
the whole plant level (Martinsen et al. 2000, Lill and Marquis 2003, Nakamura and
Ohgushi 2003, Lill and Marquis 2004, Wang et al. 2012). The ecological significance of these engineered resources is enhanced by the fact that leaf shelters tend to remain intact long after the initial builder is gone. The strong facilitative effect of shelter-building caterpillars on the forest arthropod community has been demonstrated repeatedly, but little is known about how the community within leaf ties assembles and how interactions between colonists may modulate the assembly process.
Large-scale processes such as seasonality strongly influence arthropod communities. In a variety of systems the composition of arthropod communities shifts markedly from spring, through summer and into fall (Summerville and Crist 2003,
Wiwatwitaya and Takeda 2005, Hirao et al. 2007, Sanford and Huntly 2010).
Phenological patterns in some insect communities are driven by larval feeding times
(Niemela and Haukioja 1982), so insect larvae that have large impacts on arthropod communities, such as leaf-tying caterpillars, can drive the phenological patterns of the entire community. Beyond these seasonal responses, the composition of arthropod communities is often strongly impacted by host-plant species (Butler and Strazanac 2000,
Summerville et al. 2003, Southwood et al. 2005). Seasonal peaks in arthropod abundance and species richness frequently vary between tree species (Niemela and Haukioja 1982,
Marquis and Passoa 1989). In some communities, host plant diversity has a greater effect on arthropod beta diversity than temporal variation within a season (Sobek et al. 2009).
Additionally, plant quality typically changes seasonally leading to an interaction between host plant use and seasonality that influences arthropod community composition. In
Chapter 1, leaf tie inhabitants on two con-familial, co-occurring host tree species
2
(American beech, Fagus grandifolia, and white oak, Quercus alba) were studied to determine if phenological changes in the leaf tie community can be generalized or if they are host plant species-specific.
Competition is an important structuring force in ecological communities, but has not been thoroughly investigated among leaf-tying caterpillars. Evidence for direct competition among insect herbivores for plant resources is sparse (Kaplan and Denno
2007). Of the few examples in the literature demonstrating direct competition, many are among caterpillars that build shelters, which compete via physically aggressive behaviors or vibratory signals (Berenbaum et al. 1993, Green et al. 1998, Yack et al. 2001, Fletcher et al. 2006, Bowen et al. 2008, Scott et al. 2010, Scott and Yack 2012). However, these studies focus on describing the behaviors exhibited and have not compared behaviors across species or examined competition between species. Spaces to build leaf ties are limited and high occupancy is common, indicating that they are a valuable resource over which herbivores are likely to compete. However, the frequency, intensity, and mode of such competition have not been thoroughly examined. In Chapter 2, I compared the behavior of four species of common oak leaf-tying caterpillars using staged encounters with both conspecifics and heterospecifics.
The interactions that take place within leaf ties are just beginning to be explored and likely contribute to arthropod community composition. When ecosystem engineers create new habitats, the identity of the engineer may have a significant impact on the colonist community due to ecological differences between engineer species. Many groups of organisms engineer their habitat in a similar way, but generate different outcomes
(Machicote et al. 2004, Boots et al. 2012). The effects of early colonists on the
3
community assembly process, termed priority effects, have been examined in many systems (Wilbur and Alford 1985, Drake 1991, Keddy 1992, Lawler and Morin 1993,
Belyea and Lancaster 1999), but are not commonly applied to ecosystem engineers.
Competition is a common driver of priority effects, but facilitation can also affect the community assembly process, and the two are not mutually exclusive (Connell and
Slatyer 1977). In Chapter 3, I examined how the behavioral differences between caterpillar species found in Chapter 2 may influence the compositional patterns described in Chapter 1 by comparing the early stages of community assembly in ties initiated by different caterpillar species. Overall, the goal of the studies is to determine how species interactions influence large-scale patterns of community composition.
References
Abarca, M. and K. Boege. 2011. Fitness costs and benefits of shelter building and leaf
trenching behaviour in a pyralid caterpillar. Ecological Entomology 36:564-573.
Atlegrim, O. 1992. Mechanisms regulating bird predation on a herbivorous larva guild in
boreal coniferous forests. Ecography 15:19-24.
Belyea, L. R. and J. Lancaster. 1999. Assembly rules within a contingent ecology. Oikos
86:402-416.
Berenbaum, M. R., E. S. Green, and A. R. Zangerl. 1993. Web costs and web defense in
the parsnip webworm (Lepidoptera: Oecophoridae). Environmental Entomology
22:791-795.
4
Boots, B., A. M. Keith, R. Niechoj, J. Breen, O. Schmidt, and N. Clipson. 2012. Unique
soil microbial assemblages associated with grassland ant species with different
nesting and foraging strategies. Pedobiologia 55:33-40.
Bowen, J. L., S. J. Mahony, A. C. Mason, and J. E. Yack. 2008. Vibration-mediated
territoriality in the warty birch caterpillar Drepana bilineata. Physiological
Entomology 33:238-250.
Butler, L. and J. Strazanac. 2000. Macrolepidopteran larvae sampled by tree bands in
temperate mesic and xeric forests in eastern United States. Proceedings of the
Entomological Society of Washington 102:188-197.
Clements, F. E. 1916. Plant Succession. Carnegie Institution of Washington, Washington.
Connell, J. H. and R. O. Slatyer. 1977. Mechanisms of succession in natural communities
and their role in community stability and organization. American Naturalist
111:1119-1144.
Connor, E. F. and D. Simberloff. 1979. The assembly of species communities - Chance or
competition. Ecology 60:1132-1140.
Damman, H. 1987. Leaf quality and enemy avoidance by the larvae of a pyralid moth.
Ecology 68:88-97.
Drake, J. A. 1991. Community-assembly mechanics and the structure of an experimental
species ensemble. American Naturalist 137:1-26.
Eubanks, M. D., K. A. Nesci, M. K. Petersen, Z. W. Liu, and H. B. Sanchez. 1997. The
exploitation of an ant-defended host plant by a shelter-building herbivore.
Oecologia 109:454-460.
5
Fletcher, L. E., J. E. Yack, T. D. Fitzgerald, and R. R. Hoy. 2006. Vibrational
communication in the cherry leaf roller caterpillar Caloptilia serotinella
(Gracillarioidea : Gracillariidae). Journal of Insect Behavior 19:1-18.
Fukui, A., M. Murakami, K. Konno, M. Nakamura, and T. Ohgushi. 2002. A leaf-rolling
caterpillar improves leaf quality. Entomological Science 5:263-266.
Green, E. S., A. R. Zangerl, and M. R. Berenbaum. 1998. Reduced aggressive behavior:
A benefit of silk-spinning in the parsnip webworm, Depressaria pastinacella
(Lepidoptera : Oecophoridae). Journal of Insect Behavior 11:761-772.
Hirao, T., M. Murakami, A. Kashizaki, and S. I. Tanabe. 2007. Additive apportioning of
lepidopteran and coleopteran species diversity across spatial and temporal scales
in a cool-temperate deciduous forest in Japan. Ecological Entomology 32:627-
636.
Hunter, M. D. and P. G. Willmer. 1989. The potential for interspecific competition
between 2 abundant defoliators on oak-leaf damage and habitat quality.
Ecological Entomology 14:267-277.
Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers.
Oikos 69:373-386.
Jones, C. G., J. H. Lawton, and M. Shachak. 1997. Positive and negative effects of
organisms as physical ecosystem engineers. Ecology 78:1946-1957.
Joos, B., T. M. Casey, T. D. Fitzgerald, and W. A. Buttemer. 1988. Roles of the tent in
behavioral thermoregulation of eastern tent caterpillars. Ecology 69:2004-2011.
6
Kaplan, I. and R. F. Denno. 2007. Interspecific interactions in phytophagous insects
revisited: a quantitative assessment of competition theory. Ecology Letters
10:977-994.
Keddy, P. A. 1992. Assembly and response rules - 2 goals for predictive community
ecology. Journal of Vegetation Science 3:157-164.
Lawler, S. P. and P. J. Morin. 1993. Temporal overlap, competition, and priority effects
in larval anurans. Ecology 74:174-182.
Lill, J. T. and R. J. Marquis. 2003. Ecosystem engineering by caterpillars increases insect
herbivore diversity on white oak. Ecology 84:682-690.
Lill, J. T. and R. J. Marquis. 2004. Leaf ties as colonization sites for forest arthropods: an
experimental study. Ecological Entomology 29:300-308.
Lill, J. T. and R. J. Marquis. 2007. Microhabitat manipulation: ecosystem engineering by
shelter-building insects. Pages 107-138 in K. Cuddington, J. E. Byers, W. G.
Wilson, and A. Hastings, editor. Ecosystem engineers: plants to protists.
Academic Press, New York.
Loeffler, C. C. 1996. Adaptive trade-offs of leaf folding in Dichomeris caterpillars on
goldenrods. Ecological Entomology 21:34-40.
Machicote, M., L. C. Branch, and D. Villarreal. 2004. Burrowing owls and burrowing
mammals: are ecosystem engineers interchangeable as facilitators? Oikos
106:527-535.
Marquis, R. J. and S. Passoa. 1989. Seasonal diversity and abundance of the herbivore
fauna of striped maple Acer pensylvanicum L. (Aceraceae) in western Virginia.
American Midland Naturalist 122:313-320.
7
Martinsen, G. D., K. D. Floate, A. M. Waltz, G. M. Wimp, and T. G. Whitham. 2000.
Positive interactions between leafrollers and other arthropods enhance
biodiversity on hybrid cottonwoods. Oecologia 123:82-89.
Nakamura, M. and T. Ohgushi. 2003. Positive and negative effects of leaf shelters on
herbivorous insects: linking multiple herbivore species on a willow. Oecologia
136:445-449.
Niemela, P. and E. Haukioja. 1982. Seasonal patterns in species richness of herbivores -
Macrolepidopteran larva on Finnish deciduous trees. Ecological Entomology
7:169-175.
Poff, N. L. 1997. Landscape filters and species traits: Towards mechanistic understanding
and prediction in stream ecology. Journal of the North American Benthological
Society 16:391-409.
Sagers, C. L. 1992. Manipulation of host plant-quality - Herbivores keep leaves in the
dark. Functional Ecology 6:741-743.
Sandberg, S. L. and M. R. Berenbaum. 1989. Leaf-tying by tortricid larvae as an
adaptation for feeding on phototoxic Hypercium perforatum. Journal of Chemical
Ecology 15:875-885.
Sanford, M. P. and N. J. Huntly. 2010. Seasonal patterns of arthropod diversity and
abundance on big sage brush, Artemisia tridentata. Western North American
Naturalist 70:67-76.
Scott, J. L., S. M. Matheson, and J. E. Yack. 2010. Variation on a theme: Vibrational
signaling in caterpillars of the rose hook-tip moth, Oreta rosea. Journal of Insect
Science 10:15.
8
Scott, J. L. and J. E. Yack. 2012. Vibratory territorial signals in caterpillars of the poplar
lutestring, Tethea or (Lepidoptera: Drepanidae). European Journal of Entomology
109:411-417.
Sipura, M. 1999. Tritrophic interactions: willows, herbivorous insects and insectivorous
birds. Oecologia 121:537-545.
Sobek, S., I. Steffan-Dewenter, C. Scherber, and T. Tscharntke. 2009. Spatiotemporal
changes of beetle communities across a tree diversity gradient. Diversity and
Distributions 15:660-670.
Southwood, T. R. E., G. R. W. Wint, C. E. J. Kennedy, and S. R. Greenwood. 2005. The
composition of the arthropod fauna of the canopies of some species of oak
(Quercus). European Journal of Entomology 102:65-72.
Summerville, K. S. and T. O. Crist. 2003. Determinants of lepidopteran community
composition and species diversity in eastern deciduous forests: roles of season,
eco-region and patch size. Oikos 100:134-148.
Summerville, K. S., T. O. Crist, J. K. Kahn, and J. C. Gering. 2003. Community structure
of arboreal caterpillars within and among four tree species of the eastern
deciduous forest. Ecological Entomology 28:747-757.
Wang, H. G., R. J. Marquis, and C. S. Baer. 2012. Both host plant and ecosystem
engineer identity influence leaf-tie impacts on the arthropod community of
Quercus. Ecology 93:2186-2197.
Weiher, E., D. Freund, T. Bunton, A. Stefanski, T. Lee, and S. Bentivenga. 2011.
Advances, challenges and a developing synthesis of ecological community
9
assembly theory. Philosophical Transactions of the Royal Society B-Biological
Sciences 366:2403-2413.
Wilbur, H. M. and R. A. Alford. 1985. Priority effects in experimental pond communities
- Responses of Hyla to Bufo and Rana. Ecology 66:1106-1114.
Willmer, P. G. 1980. The effects of fluctuating environment on the water relations of
larval Lepidoptera. Ecological Entomology 5:271-292.
Wiwatwitaya, D. and H. Takeda. 2005. Seasonal changes in soil arthropod abundance in
the dry evergreen forest of north-east Thailand, with special reference to
collembolan communities. Ecological Research 20:59-70.
Yack, J. E., M. L. Smith, and P. J. Weatherhead. 2001. Caterpillar talk: Acoustically
mediated territoriality in larval Lepidoptera. Proceedings of the National
Academy of Sciences of the United States of America 98:11371-11375.
10
CHAPTER 1: PHENOLOGICAL VARIATION IN THE COMPOSITION OF A
TEMPERATE FOREST LEAF TIE COMMUNITY
*This chapter is published as: Sigmon, E. and J. T. Lill. 2013. “Phenological variation in the composition of a temperate forest leaf tie community.” Environmental Entomology
42(1):29-37.
Arthropod communities in an array of temperate ecosystems follow similar phenological patterns of distinct compositional turnovers during the course of a season. The arthropod community inhabiting leaf ties are no exception. Many caterpillars build leaf ties, shelters between overlapping leaves attached together with silk, which are secondarily colonized by a variety of arthropods. We created experimental leaf ties by clipping overlapping leaves together with metal clips. We censused the arthropod community within experimental ties on two host plants, American beech, Fagus grandifolia Ehrhart, and white oak, Quercus alba L., weekly for ten weeks during the summer of 2009. Diversity measures for leaf-tying caterpillars and the entire arthropod community within ties varied little between tree species and sampling periods, but caterpillar and arthropod density per tie was significantly higher on white oak than beech and abundance increased on both tree species as the season progressed. The composition (i.e., species presence and abundance) of the leaf-tying caterpillar community and the arthropod community as a whole differed between host tree species and sampling periods. Though the arthropod communities on American beech and white oak differed, they showed similar patterns of
11
compositional turnover, with distinct communities in early and late summer and a transitional community mid-summer.
Introduction
Lepidoptera larvae (caterpillars) comprise one of the most significant components of temperate forest arthropod communities and are typically the largest and most diverse group of herbivores in these communities (Schowalter et al. 1986). Among caterpillars, leaf-tiers are a dominant component of temperate, broad-leaved forest communities in the summer (Forkner et al. 2008). Caterpillars in at least 24 families of Lepidoptera function as ecosystem engineers by tying leaves together with silk to create shelters (Lill and
Marquis 2007). Leaf ties are secondarily inhabited by a variety of arthropods that benefit from altered microclimate (Willmer 1980, Joos et al. 1988, Hunter and Willmer 1989), decreased toxic effects of foliage consumption (Sandberg and Berenbaum 1989, Sagers
1992, Fukui et al. 2002), and protection from a variety of natural enemies (Damman
1987, Atlegrim 1992, Loeffler 1996, Eubanks et al. 1997, Sipura 1999, Abarca and
Boege 2011). In addition to building a shelter that other arthropods use, leaf-tiers are consumed by a variety of predators and parasitoids and create food for scavengers in the form of frass and exuviae. Leaf ties have been shown to increase arthropod abundance
(Martinsen et al. 2000, Fukui et al. 2002, Lill and Marquis 2004), species richness
(Martinsen et al. 2000, Lill and Marquis 2003), and alter community structure at the whole tree level (Lill and Marquis 2004). Leaf ties persist throughout the season beyond the life span of the initial builder and thus can be used by temporally separated
12
arthropods. However, the general phenology of the leaf tie community is poorly understood.
Arthropod phenology is becoming an increasingly popular area of study, in particular because the phenology of many plants and arthropods is being altered by climate change (Walther et al. 2002, Parmesan 2006, Miller-Rushing and Primack 2008,
Yang and Rudolf 2010). It is important to understand general characteristics of arthropod community phenology in order to detect when patterns are altered. Turnover in
Lepidoptera and Coleoptera community composition has been reported to occur both monthly and seasonally (Summerville and Crist 2003, Hirao et al. 2007). In a variety of systems the composition of arthropod communities shifts markedly from spring, through summer and into fall (Summerville and Crist 2003, Wiwatwitaya and Takeda 2005, Hirao et al. 2007, Sanford and Huntly 2010). Factors that determine arthropod phenology include emergence time, development time, voltinism, generation length, and adult lifespan (Hirao et al. 2007). Phenological changes in some insect communities have been shown to be driven by larval feeding times (Niemela and Haukioja 1982). Insect larvae that have large impacts on arthropod communities, such as leaf-tying caterpillars, can thus drive the phenological patterns of the entire community.
Beyond these seasonal responses, the composition of arthropod communities is often strongly impacted by host-plant species (Butler and Strazanac 2000, Summerville et al. 2003, Southwood et al. 2005). Seasonal peaks in arthropod abundance and species richness frequently vary between tree species (Niemela and Haukioja 1982, Marquis and
Passoa 1989). In some communities, host plant diversity has a greater effect on arthropod beta diversity than temporal variation within a season (Sobek et al. 2009). Host plant use
13
by arthropods may also vary phenologically, with some herbivores switching between hosts both within and between generations (Moran 1992, Moir et al. 2011). Furthermore, plant quality typically changes seasonally leading to an interaction between host plant use and seasonality that influences arthropod community composition.
In this study, leaf tie inhabitants on two con-familial, co-occurring host tree species [American beech, Fagus grandifolia Ehrhart, and white oak, Quercus alba L.
(Fagaceae)] were studied to determine if phenological differences in the leaf tie community can be generalized or if they are host plant species-specific. Experimental leaf ties were created on trees at the beginning of the summer and sub-sampled weekly for ten weeks to determine the abundance and identity of arthropod inhabitants.
Specifically, this study examined the following question: What are the effects of time of year (early, middle and late season) and host tree species on the diversity, abundance, and composition of (1) leaf-tying caterpillars and (2) the entire arthropod community that secondarily inhabits leaf ties?
Materials and Methods
Study System. This study took place during the summer of 2009 at Little Bennett
Regional Park in Montgomery Co., MD (39.26°N and 77.28°W), a site within the piedmont region of Maryland. The park contains a 13.4 km2 secondary growth forest dominated by oak, hickory, and beech.
The arthropod community inhabiting leaf ties was examined on two co-occurring canopy dominants, white oak, Quercus alba, and American beech, Fagus grandifolia.
These host plants were chosen because they are taxonomically related but differ in quality as a food source for herbivores. On average, American beech has lower water
14
content, less nitrogen, and a lower C:N ratio than white oak and suffers less damage over the course of the season (J.T.L. unpublished data).
Procedure. To survey the leaf tie inhabitants throughout a summer season, experimental ties were created on white oak and American beech. Experimental ties were made by clipping two adjacent similar sized leaves together using metal double prong curler clips to completely overlap the two leaves. Experimental ties are colonized quickly and the presence of the clip does not affect arthropod abundance (Martinsen et al. 2000,
Lill and Marquis 2004). Experimentally created ties provide a controlled yet realistic proxy for natural ties that can be systematically sampled because colonization of experimental ties is positively correlated with the number of natural ties on an individual tree (Marquis and Lill 2010). Accessible understory foliage (< 3 m) from a range of tree sizes on ten trees each of white oak and beech were used for this study. All trees or tree branches censused had over 1,000 leaves so were large enough that the addition of experimental ties did not dramatically increase the proportion of tied leaves on the tree, which is ≈10-15% at this study site (E.S. pers. obs.). Fifty-five experimental leaf ties were created on each tree on 11 June 2009 for a total of 550 ties per tree species. Sub- samples of five ties were haphazardly collected each week for ten weeks from 23 June to
25 August 2009 for a total of 500 ties per tree species. The additional five experimental ties per tree allowed for the loss of some ties without altering the sample size in the final week. One beech tree and one white oak were excluded from the last three weeks of collection because other trees had fallen on them during a storm, damaging the trees and making the ties inaccessible. Due to the nature of the sampling method, ties collected later in the season were available for colonization for a longer period of time. However,
15
this mimics the natural patterns seen of ties built at the beginning of the season being re- colonized by arthropods later in the season. Due to the facts that caterpillars can only build ties between overlapping leaves and white oak and beech trees only produce new leaves in the spring, the spaces naturally available for leaf tie construction are set once leaves are fully expanded and can not change throughout the summer. Thus, experimental ties were created prior to the first generation of any leaf tying caterpillars, rather than adding ties throughout the summer, to mimic natural leaf tie availability. We do not believe that the different amount of time ties were available for colonization biased occupancy of the ties because ties were heavily colonized by the first sampling date and the proportion of colonized ties remained relatively stable.
Each week, the collected ties were gently removed from the tree and immediately placed in individual plastic zipper bags. In the laboratory, ties were opened and carefully inspected with a hand lens to count and identify all arthropods (excluding mites) inhabiting the ties. Further identification of arthropods was aided by use of a dissecting microscope when necessary. All caterpillars were subsequently reared in the lab and were identified as caterpillars or from reared adult moths. All non-lepidopteran arthropods were placed in individual glass vials containing alcohol and identified to species or morphospecies. Morphospecies classification is frequently used for large community data sets and richness estimates using morphospecies correspond closely to those using expert taxonomic identification (Oliver and Beattie 1996, Longino and Colwell 1997). The leaf tie community contains many juvenile arthropods, which are not easily identified to species. This most likely lead to some species being lumped together as the same morphospecies in the analyses. However, because this would tend to lead to an
16
underestimation of diversity and community-wide differences, our results are likely to be conservative. It is also possible that juveniles and adults of the same species were classified as different morphospecies, but we believe this was rarely done due to our extensive work with this community including observations of the development of common nymphs.
Data Analysis. For analysis, weekly samples were grouped into early-, mid- and late-season periods. Samples from the first three weeks (23 June – 6 July) comprised the
‘early’ period, the middle four weeks (14 July - 4 August) the ‘middle’ period, and the last three weeks (11 – 25 August) the ‘late’ period. These groupings were based on observed generation times and natural peaks of occupancy of common arthropods sampled. For most bivoltine species, which included the most common leaf tiers, this placed generational peaks in the early and late periods with the between-generation valley in the middle period. Most univoltine species were restricted to one or two of the periods.
Arthropods were combined across all sampled ties on each individual tree during each time period (early: 15 ties/tree × 10 trees = 150 ties/species; middle: 20 ties/tree × 10 trees = 200 ties/species; late: 15 ties/tree × 9 trees = 135 ties/species) and the tree was used as the sampling unit.
To determine if occupancy of experimental ties varied between tree species and time periods, the proportion of ties occupied per tree during each time period was calculated. Proportions were arcsine square root transformed and compared between tree species using a repeated measures analysis of variance (ANOVA) because the same trees were sampled in each time period.
17
Diversity of both the total arthropod community and a subset consisting of only leaf-tying caterpillars was estimated for each individual tree during each time period using EstimateS (Cowell 2006). Diversity was quantified using the Exponential Shannon
Diversity Index because it is effective for relatively small samples and is expressed in units of equivalent, equally abundant species so the values are easily comparable across studies (Magurran 2004, Jost 2006, 2007). Density of arthropods per tie was calculated for each tree as the abundance of arthropods in the sampling period divided by the number of ties sampled to account for differences in sampling effort between time periods. Diversity and density measures were transformed as needed to meet assumptions of parametric tests. Diversity and density were compared between host species using a repeated measures ANOVA. All analysis of variance tests were performed in JMP (SAS
Institute Inc 2007).
It is important in community surveys to examine differences in community composition as well as diversity because these two qualities explain different aspects of changing communities. For example, two samples that have an equal number of species may be composed of entirely different sets of species. To visually compare arthropod community composition between the two host tree species and three time periods we used non-metric multi-dimensional scaling (NMDS; Kruskal 1964). This technique displays compositional differences between samples in n-dimensional space. The NMDS performed compared the communities on each tree during each time period to one another using a dissimilarity matrix derived from species abundance using the Bray-
Curtis distance measure. Separate NMDS analyses were performed for the entire arthropod community and for the leaf-tying caterpillar community alone. For the latter
18
analysis, some beech trees were omitted due to an absence of leaf tying caterpillars in a particular sample resulting in slightly reduced sample sizes (N = 7, 9, and 8 trees for the early, middle and late sampling periods, respectively). We also performed an analysis of dissimilarity (ADONIS; Anderson 2001) on each dissimilarity matrix to determine whether compositional differences between tree species and time periods were greater than differences within groups. All NMDS and ADONIS analyses were performed in R
(R Core Development Team 2011) using the ‘vegan’ package (Oksanen et al. 2011).
Results
Over the ten weeks of sampling, 931 and 1,761 arthropods were collected from the 10 American beech trees and the 10 white oak trees, respectively, for a total of 2,692 arthropods. There were 45 species and morphospecies found exclusively on American beech and 58 solely on white oak, with an additional 35 occurring on both tree species for a total of 138 species (Appendix 1.1). Leaf tying caterpillars constituted 13.1% (122) of the arthropods on American beech and 34.3% (604) of the arthropods on white oak. The most common caterpillar was Psilocorsis quercicella Clemens (Oecophoridae), which was restricted to white oak, with 441 individuals collected (25% of all arthropods collected on white oak). Psilocorsis cryptolechiella Chambers, P. reflexella Clemens
(Oecophoridae) and Pseudotelphusa quercinigracella Chambers (Gelechiidae) were abundant on both beech and white oak (Fig.1). Abundant non-lepidopteran arthropods included aphids (Hemiptera: Aphididae), wooly aphids (Hemiptera: Eriosomatidae),
Ostearius melanopygius Cambridge (Araneae: Linyphiidae), Elaver sp. Cambridge
(Araneae: Clubionidae), the Asiatic oak weevil Cyrtepistomus castaneus Roelofs
(Coleoptera: Curculionidae), rove beetles (Coleoptera: Staphylinidae), and thrips
19
(Thysanoptera), a collembolan (Unknown sp 48), and a predaceous beetle larva
(Coleoptera: Unknown sp 38; Appendix 1.1).
Across all censuses, the proportion of ties occupied by arthropods was 10-20% higher on white oak than on American beech (F1, 16 = 15.38, P = 0.0012; Fig. 2). Tie occupancy increased significantly as the season progressed (F2, 15 = 12.58, P = 0.0006;
Fig. 2).
Fifteen species of leaf tying caterpillars were recorded, ten of which were found on both beech and white oak. Four species were unique to white oak and only one was unique to beech (Appendix 1.1). Leaf-tying caterpillar diversity did not differ between tree species or sampling periods (all P > 0.38; Fig. 3A). Density of leaf tying caterpillars per tie was higher on white oak than on American beech (F1, 16 = 35.25, P < 0.0001) and differed among sampling periods (F2, 15 = 6.61, P = 0.0087), peaking in late summer (Fig.
3B). However, the most abundant leaf-tier, P. quercicella, was only found on white oak and when it was removed from the data set the model was no longer significant (all P >
0.08). The NMDS of the leaf tier community reached a stable solution using 2 dimensions (Fig. 4) with a minimum stress value of 14.87, which is considered low enough to represent real community differences (Clarke 1993). Despite similar diversities, the composition of the leaf-tying caterpillar community differed significantly between host tree species (F1, 47 = 15.19, P < 0.001) and among time periods (F2, 47 = 2.58,
P = 0.004).
Total arthropod diversity did not differ between host tree species (F1, 16 = 1.74, P =
0.2060). However, there was a significant interaction between host tree species and time
(F2, 15 = 3.85, P = 0.0447); during the mid-season, diversity was highest on white oak and
20
lowest on beech (Fig. 5A). Arthropod density per tie was significantly higher on white oak than beech (F1, 16 = 14.65, P = 0.0015) and differed between time periods (F2, 15 =
52.05, P < 0.0001). There was also a significant interaction between tree species and time due mainly to a difference among tree species in density during mid-season (F2, 15 = 9.51,
P = 0.0021; Fig. 5B). The NMDS of the entire arthropod community reached a stable solution using 3 dimensions (Fig. 6) with a minimum stress value of 14.69. The composition of the arthropod community differed significantly between host tree species
(F1, 52 = 11.66, P < 0.001) and time periods (F2, 52 = 3.60, P < 0.001). There was also a marginally significant interaction between tree species and time periods (F2, 52 = 1.60, P =
0.046). The differences in the arthropod community were not merely due to differences in the leaf tier community as the same qualitative results persisted when leaf tiers were excluded from the data set (tree species: F1, 52 = 3.89, P < 0.001; time: F1, 52 = 3.93, P <
0.001; interaction: F1, 52 = 1.66, P = 0.028).
Discussion
We have shown that the arthropod community inhabiting leaf ties on American beech and white oak changes significantly throughout the summer season. Though the communities on American beech and white oak differed in composition and density, they exhibited similar phenological patterns. Different communities were present in early and late summer with a turnover occurring in July.
A wide variety of arthropods were found inhabiting the experimental leaf ties.
The dominant herbivores in leaf ties are leaf-tying caterpillars, which are almost exclusively microlepidopterans that are never found outside of leaf ties. Lill and Marquis
(2004) found that caterpillar migration and oviposition by adult moths contributed
21
equally to the colonization of experimental leaf ties. Many of the leaf-tiers in this study are bivoltine so are present in the community throughout much of the summer. Moths in the late summer frequently oviposit within existing leaf ties, even when much of the foliage has already been consumed by previous occupants (pers. obs.). This constant re- use of the same tie demonstrates the considerable value of these microhabitats. Aside from caterpillars, many other juvenile arthropods were found inhabiting leaf ties.
Juveniles are generally more vulnerable than adults to desiccation and predation so may benefit more from the protection of leaf ties. Aphids were also extremely abundant within the leaf ties in this study. Aphids are common occupants of many types of leaf shelters
(Hajek and Dahlsten 1986, Martinsen et al. 2000, Nakamura and Ohgushi 2003) suggesting that there may be some aspect of their ecology that predisposes them to using shelters. Herbivores use the leaf ties as shelter and a food source, so are likely to remain in the same tie for the majority of their development. Conversely, predators and scavengers may find food within shelters, but may need to move between shelters for a continuous supply of food. Spiders are believed to rest, hunt, or molt within shelters then move on leading to a highly variable subset of the local spider community being found in leaf ties. One of the most common spiders, Elaver sp., ties leaves as well. It creates a small webbed sphere between leaves completely enclosing itself. These shelters were frequently found containing a molt so it is our belief that this spider creates a shelter for protection during vulnerable molting periods. Notably, many predators and herbivores were found in the same tie even after it had been in a plastic bag for up to two days, suggesting that predators may be using leaf ties for benefits other than prey items (e.g. microclimate and protection from their natural enemies; Lill and Marquis 2007).
22
Leaf-tying caterpillar density was higher overall on oak than on beech, though this was driven primarily by one species, Psilocorsis quercicella. The density of leaf tiers in experimental ties increased as the season progressed, which may lead to an increase in the proportion of leaves tied on a tree in the late summer. As leaf-tying caterpillars grow they become capable of pulling together leaves that small caterpillars cannot (Caroll
1977). This leads to an increase in the number of ties as well as the number of leaves involved in a single tie later in the summer. An increase in the density of naturally occurring ties may enhance the ecosystem engineering effects on the arthropod community as a whole in late summer.
Despite similarities in diversity, the leaf-tier community composition differed between host tree species and time periods. This discrepancy highlights the importance of comparing patterns of compositional differences in addition to diversity or richness measures. Some species of leaf tiers, particularly univoltine tortricids, were restricted to just a few weeks of the summer, contributing to the observed compositional differences.
The results of the non-metric multi-dimensional scaling (NMDS) indicate that the leaf- tying caterpillar community on beech shows a greater amount of phenological variation than that on white oak. This may be due to the smaller community on beech (N = 10 spp) being less predictable than the larger community found on white oak (N = 14 spp).
Diversity of the total arthropod communities on American beech and white oak showed a curious interaction with time. Arthropod diversity was highest on oak in mid- season and lowest on beech in mid-season. By contrast, early and late summer diversity was similar on beech and oak. Arthropod density showed a much more predictable trend than diversity. Density increased as the summer progressed on both American beech and
23
white oak. However, the increase began earlier on beech than on white oak. The mid- season, in July, seems to be an important period of community restructuring on both beech and white oak. Despite similarities in diversity, arthropod community composition differed phenologically and between the two host plants. By examining the results of the
NMDS it can be seen that the arthropod community gradually changed from early to late summer with mid-summer being somewhat of an intermediate. This differs from the phenological variation seen in diversity, where early and late season diversity are similar.
This indicates that there is a compositional shift in the arthropod communities on both white oak and beech over the summer despite similarities in diversity across the season.
Further research by the authors in the same forest as the present study as well as other forests in Montgomery Co., MD and DC have demonstrated that the patterns described here are typical of the oak and beech leaf tie communities. Peaks in arthropod abundance and diversity are easily predictable based on this study and tend to only shift slightly across years due to warmer or cooler springs (E.S. unpublished data).
The arthropod community within leaf ties changes seasonally like many other arthropod communities (Summerville and Crist 2003, Wiwatwitaya and Takeda 2005,
Hirao et al. 2007, Sanford and Huntly 2010). Changes in leaf chemistry from spring to summer are frequently cited as the driving mechanism behind differences in arthropod communities between these seasons (Feeny 1970, Forkner et al. 2004). As the summer progresses, oaks decrease in water and nitrogen content and increase in toughness and tannin concentrations, which can affect herbivores (Feeny 1970, Forkner et al. 2004).
Changes in the herbivore community could cascade up to higher trophic levels. Abiotic factors such as temperature and humidity may limit some arthropods to certain times of
24
the season. Leaf ties alter these abiotic factors (Willmer 1980, Joos et al. 1988, Hunter and Willmer 1989), so may be used differentially as heat stress and desiccation risk change seasonally. Space to build a leaf tie appears to be a limiting resource, thus species could be separating phenologically to divide the niche space and minimize competition.
The arthropod communities present in leaf ties on American beech and white oak differed substantially despite the host plants being closely related. Leaf chemistry has been shown to be an important factor in host plant use by leaf-tying caterpillars among closely related tree species (Marquis and Lill 2010). American beech is a lower quality host plant than white oak, which may explain the observed differences in leaf tier density.
Plant architecture, measured as the number of leaves touching, has also been shown to affect leaf tie density within a host plant species (Marquis et al. 2002) and may account for some of the observed differences. Beech leaves are positioned alternately on long branches with even spacing between leaves whereas white oak has a more complex branching pattern with whorls of leaves. This causes more leaves to overlap on white oak than on beech trees (E.S. and J.L. pers. obs.), which may contribute to the higher density of natural leaf ties found on white oak. Arthropods frequently move between ties and readily colonize new ties (Lill and Marquis 2004). High density of natural ties on white oak may explain the higher occupancy of experimental ties and higher arthropod density within ties. Unfortunately, we did not estimate natural tie density as a part of this study so no direct correlations can be made.
Evolutionary history may also play a role in the differences between host plant fauna. While American beech and white oak are in the same family (Fagaceae),
American beech is the only representative of its genus while there are over 300 species of
25
oaks in North America (Manos and Stanford 2001). This allows for a larger, more diverse fauna to develop on oaks leading to more monophages and oligophages feeding on white oak (Wagner 2005). Oaks generally have a larger associated arthropod fauna than beech, which may lead to a proportional increase in their response to leaf ties. Similar comparisons of the herbivore faunas associated with pairs of related species-rich vs. species-poor genera are needed to further substantiate this pattern.
Many plant-based arthropod communities have been shown to be affected by host plant species and seasonality, but few have been shown to be significantly influenced by the interaction between these two factors. The quality of these two host plants may exhibit seasonal declines at different rates, which would lead to similar patterns of host plant use that are offset in time. Additionally, American beech is a lower quality host plant than white oak, which may slow the development of herbivores feeding on beech compared to those on white oak. If this is true, the early season herbivores may take longer to develop on beech than on white oak leading to a shorter gap between the early and late season herbivore communities.
Though the communities found in leaf ties on American beech and white oak differ, they exhibit similar phenological pattern of changes in community composition but not diversity from early to late summer. Future research should focus on probing the mechanisms underlying these observed patterns. In particular, because arthropods may inhabit shelters for food, protection, or alleviation of climatic stress, determining which of these factors most influences leaf tie inhabitants and how these factors may differ among host plants may lead to greater understanding of the ecosystem engineering effect of leaf tiers.
26
Acknowledgments
Thanks to S. M. Murphy and L. Power for assistance in field and laboratory work, G. M. Wimp for advice and assistance with statistical analysis, J. W. Brown (Lepidoptera: Tortricidae) and G.
Hormiga (Araneae) for identification of specimens, and the members of DC Plant Insect Group
(DC-PIG) and the Lill lab for comments on early drafts. Funding provided by George
Washington University Mortensen Fund to E. Sigmon.
27
References Cited
Abarca, M., and K. Boege. 2011. Fitness costs and benefits of shelter building and leaf trenching
behaviour in a pyralid caterpillar. Ecological Entomology 36: 564-573.
Anderson, M. J. 2001. A new method for non-parametric multivariate analysis of variance.
Austral Ecology 26: 32-46.
Atlegrim, O. 1992. Mechanisms regulating bird predation on a herbivorous larva guild in boreal
coniferous forests. Ecography 15: 19-24.
Butler, L., and J. Strazanac. 2000. Macrolepidopteran larvae sampled by tree bands in temperate
mesic and xeric forests in eastern United States. Proceedings of the Entomological
Society of Washington 102: 188-197.
Caroll, M. R. 1977. Observations on microlepidopterous leaft tiers (Lepidoptera: Gelechiodea) in
central Missouri. Masters, University of Missouri-Columbia.
Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure.
Australian Journal of Ecology 18: 117-143.
Cowell, R. K. 2006. EstimateS: Statistical estimation of species richness and shared species from
samples. Version 8.2.
Damman, H. 1987. Leaf quality and enemy avoidance by the larvae of a pyralid moth. Ecology
68: 88-97.
Eubanks, M. D., K. A. Nesci, M. K. Petersen, Z. W. Liu, and H. B. Sanchez. 1997. The
exploitation of an ant-defended host plant by a shelter-building herbivore. Oecologia 109:
454-460.
Feeny, P. 1970. Seasonal Changes in Oak Leaf Tannins and Nutrients as a Cause of Spring
Feeding by Winter Moth Caterpillars. Ecology 51: 565–581.
28
Forkner, R. E., R. J. Marquis, and J. T. Lill. 2004. Feeny revisited: condensed tannins as anti-
herbivore defences in leaf-chewing herbivore communities of Quercus. Ecological
Entomology 29: 174-187.
Forkner, R. E., R. J. Marquis, J. T. Lill, and J. Le Corff. 2008. Timing is everything?
Phenological synchrony and population variability in leaf-chewing herbivores of
Quercus. Ecological Entomology 33: 276-285.
Fukui, A., M. Murakami, K. Konno, M. Nakamura, and T. Ohgushi. 2002. A leaf-rolling
caterpillar improves leaf quality. Entomological Science 5: 263-266.
Hajek, A. E., and D. L. Dahlsten. 1986. Coexistence of 3 species of leaf-feeding aphids
(Homoptera) on Betula pendula. Oecologia 68: 380-386.
Hirao, T., M. Murakami, A. Kashizaki, and S. I. Tanabe. 2007. Additive apportioning of
lepidopteran and coleopteran species diversity across spatial and temporal scales in a
cool-temperate deciduous forest in Japan. Ecological Entomology 32: 627-636.
Hunter, M. D., and P. G. Willmer. 1989. The potential for interspecific competition between 2
abundant defoliators on oak-leaf damage and habitat quality. Ecological Entomology 14:
267-277.
Joos, B., T. M. Casey, T. D. Fitzgerald, and W. A. Buttemer. 1988. Roles of the tent in
behavioral thermoregulation of eastern tent caterpillars. Ecology 69: 2004-2011.
Jost, L. 2006. Entropy and diversity. Oikos 113: 363-375.
Jost, L. 2007. Partitioning diversity into independent alpha and beta components. Ecology 88:
2427-2439.
Kruskal, J. B. 1964. Nonmetric multidimensional scaling: a numerical method. Psychometrika
29: 115-129.
29
Lill, J. T., and R. J. Marquis. 2003. Ecosystem engineering by caterpillars increases insect
herbivore diversity on white oak. Ecology 84: 682-690.
Lill, J. T., and R. J. Marquis. 2004. Leaf ties as colonization sites for forest arthropods: an
experimental study. Ecological Entomology 29: 300-308.
Lill, J. T., and R. J. Marquis. 2007. Microhabitat manipulation: ecosystem engineering by
shelter-building insects, pp. 107-138. In K. Cuddington, J. E. Byers, W. G. Wilson, and
A. Hastings (ed.), Ecosystem engineers: plants to protists. Academic Press, New York.
Loeffler, C. C. 1996. Adaptive trade-offs of leaf folding in Dichomeris caterpillars on
goldenrods. Ecological Entomology 21: 34-40.
Longino, J. T., and R. K. Colwell. 1997. Biodiversity assessment using structured inventory:
Capturing the ant fauna of a tropical rain forest. Ecological Applications 7: 1263-1277.
Magurran, A. 2004. Measuring biological diversity, Blackwell Publishing, Malden, MA.
Manos, P. S., and A. M. Stanford. 2001. The historical biogeography of Fagaceae: Tracking the
tertiary history of temperate and subtropical forests of the Northern Hemisphere.
International Journal of Plant Sciences 162: S77-S93.
Marquis, R. J., and S. Passoa. 1989. Seasonal diversity and abundance of the herbivore fauna of
striped maple Acer pensylvanicum L. (Aceraceae) in western Virginia. American Midland
Naturalist 122: 313-320.
Marquis, R. J., and J. T. Lill. 2010. Impact of plant architecture versus leaf quality on attack by
leaf-tying caterpillars on five oak species. Oecologia 163: 203-213.
Marquis, R. J., J. T. Lill, and A. Piccinni. 2002. Effect of plant architecture on colonization and
damage by leaftying caterpillars of Quercus alba. Oikos 99: 531-537.
30
Martinsen, G. D., K. D. Floate, A. M. Waltz, G. M. Wimp, and T. G. Whitham. 2000. Positive
interactions between leafrollers and other arthropods enhance biodiversity on hybrid
cottonwoods. Oecologia 123: 82-89.
Miller-Rushing, A. J., and R. B. Primack. 2008. Global warming and flowering times in
Thoreau's concord: A community perspective. Ecology 89: 332-341.
Moir, M. L., K. E. C. Brennan, M. J. Fletcher, J. D. Majer, and J. M. Koch. 2011. Multi-scale
patterns in the host specificity of plant-dwelling arthropods: the influence of host plant
and temporal variation on species richness and assemblage composition of true bugs
(Hemiptera). Journal of Natural History 45: 2577-2604.
Moran, N. A. 1992. The evolution of aphid life-cycles. Annual Review of Entomology 37: 321-
348.
Nakamura, M., and T. Ohgushi. 2003. Positive and negative effects of leaf shelters on
herbivorous insects: linking multiple herbivore species on a willow. Oecologia 136: 445-
449.
Niemela, P., and E. Haukioja. 1982. Seasonal patterns in species richness of herbivores -
Macrolepidopteran larva on Finnish deciduous trees. Ecological Entomology 7: 169-175.
Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O'Hara, G. L. Simpson,
P. Solymos, M. H. H. Stevens, and H. Wagner 2011. vegan: Community Ecology
Package computer program, version R package version 2.0-1 By Oksanen, J., F. G.
Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O'Hara, G. L. Simpson, P.
Solymos, M. H. H. Stevens, and H. Wagner.
Oliver, I., and A. J. Beattie. 1996. Invertebrate morphospecies as surrogates for species: A case
study. Conservation Biology 10: 99-109.
31
Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual
Review of Ecology Evolution and Systematics 37: 637-669.
R Core Development Team. 2011. R: A language and environment for statistical computing. R
Foundation for Statistical Computing,Vienna, Austria.
Sagers, C. L. 1992. Manipulation of host plant-quality - Herbivores keep leaves in the dark.
Functional Ecology 6: 741-743.
Sandberg, S. L., and M. R. Berenbaum. 1989. Leaf-tying by tortricid larvae as an adaptation for
feeding on phototoxic Hypercium perforatum. Journal of Chemical Ecology 15: 875-885.
Sanford, M. P., and N. J. Huntly. 2010. Seasonal patterns of arthropod diversity and abundance
on big sage brush, Artemisia tridentata. Western North American Naturalist 70: 67-76.
SAS Institute Inc. 2007. JMP User’s Guide. Cary, NC: SAS Institute Inc.
Schowalter, T. D., W. W. Hargrove, and D. A. Crossley. 1986. Herbivory in forested
ecosystems. Annual Review of Entomology 31: 177-196.
Sipura, M. 1999. Tritrophic interactions: willows, herbivorous insects and insectivorous birds.
Oecologia 121: 537-545.
Sobek, S., I. Steffan-Dewenter, C. Scherber, and T. Tscharntke. 2009. Spatiotemporal changes of
beetle communities across a tree diversity gradient. Diversity and Distributions 15: 660-
670.
Southwood, T. R. E., G. R. W. Wint, C. E. J. Kennedy, and S. R. Greenwood. 2005. The
composition of the arthropod fauna of the canopies of some species of oak (Quercus).
European Journal of Entomology 102: 65-72.
32
Summerville, K. S., and T. O. Crist. 2003. Determinants of lepidopteran community composition
and species diversity in eastern deciduous forests: roles of season, eco-region and patch
size. Oikos 100: 134-148.
Wagner, D. L. 2005. Caterpillars of Eastern North America, Princeton University Press,
Princeton, NJ.
Walther, G. R., E. Post, P. Convey, A. Menzel, C. Parmesan, T. J. C. Beebee, J. M. Fromentin,
O. Hoegh-Guldberg, and F. Bairlein. 2002. Ecological responses to recent climate
change. Nature 416: 389-395.
Willmer, P. G. 1980. The effects of fluctuating environment on the water relations of larval
Lepidoptera. Ecological Entomology 5: 271-292.
Wiwatwitaya, D., and H. Takeda. 2005. Seasonal changes in soil arthropod abundance in the dry
evergreen forest of north-east Thailand, with special reference to collembolan
communities. Ecological Research 20: 59-70.
Yang, L. H., and V. H. W. Rudolf. 2010. Phenology, ontogeny and the effects of climate change
on the timing of species interactions. Ecology Letters 13: 1-10.
33
Figure Legends
Fig 1.1. Common leaf-tying caterpillars found on white oak and beech. From top to bottom:
Psilocorsis quercicella (Oecophoridae), P. cryptolechiella (pre-pupa with wasp parasitoid), P. reflexella (photos by Keegan Morrison), and Pseudotelphusa quercinigracella (Gelechiidae; photo by E. Sigmon).
Fig 1.2. Proportion (mean ± S.E.) of experimental leaf ties occupied by arthropods on American beech and white oak in early (white), middle (gray) and late (black) summer. Proportions were arcsine square root transformed for analyses. Occupancy differed between tree species (F1, 16 =
15.38, P = 0.0012) and time periods (F2, 15 = 12.58, P = 0.0006).
Fig 1.3. (A) Exponential Shannon diversity index value (mean ± S.E.) and (B) density of caterpillars per leaf tie (mean ± S.E.) of the leaf-tying caterpillar community inhabiting leaf ties on two host tree species, American beech and white oak, during three time periods during a summer season—early (white), middle (gray) and late (black). Diversity and density were determined for communities on individual trees.
Fig 1.4. Results of non-metric multidimensional scaling (NMDS) analysis of leaf-tying caterpillar community inhabiting leaf ties on American beech and white oak in early, middle and late summer. Species abundance scores were used for individual trees in each time period
(n=54). Analysis of dissimilarity (ADONIS) resulted in significant differences in community composition both between tree species (F1, 47 = 15.19, P < 0.001) and time period (F2, 47 = 2.58, P
= 0.004).
34
Fig 1.5. (A) Exponential Shannon diversity index value (mean ± S.E.) and (B) density of arthropods per tie (mean ± S.E.) for arthropod community inhabiting leaf ties on two host tree species, American beech and white oak, during three time periods during a summer season— early (white), middle (gray) and late (black). Diversity and density were determined for communities on individual trees.
Fig 1.6. Results of non-metric multidimensional scaling (NMDS) analysis of arthropod community inhabiting leaf ties on American beech and white oak in early, middle and late summer. Species abundance scores were used for individual trees in each time period (n=58).
Analysis of dissimilarity (ADONIS) resulted in significant differences in community composition both between tree species (F1, 52 = 11.66, P < 0.001) and between time periods (F2, 52
= 3.60, P < 0.001) and there was a marginally significant interaction (F2, 52 = 1.60, P = 0.046).
35
Figure 1.1
36
Figure 1.2
37
Figure 1.3
38
Figure 1.4
39
Figure 1.5
40
Figure 1.6
41
CHAPTER 2: INTERSPECIFIC VARIATION IN AGGRESSIVE FIGHTING
BEHAVIOR OF SHELTER-BUILDING CATERPILLARS
Caterpillars that build leaf shelters have frequently been shown to have facilitative effects on arthropod communities. However, the role of competition in mediating these effects has largely been ignored. The objective of this study was to characterize the behaviors of four common, co-occurring caterpillar species that tie together overlapping leaves on white oak – Psilocorsis quercicella, P. cryptolechiella, P. reflexella, and
Pseudotelphusa quercinigracella. Artificial leaf ties were created by clipping a piece of transparency paper to a white oak leaf. An ‘occupant’ caterpillar was allowed to build a shelter between the transparency paper and leaf, after which an ‘intruder’ caterpillar was introduced into the leaf tie. Caterpillars were observed pushing and hitting one another to gain or maintain access to the shelter. Outcomes included the occupant ‘winning’
(maintaining possession of the shelter; 52% of interactions), the intruder ‘winning’
(usurping the shelter from the occupant; 24% of interactions), or caterpillars sharing the shelter (24% of interactions). The four species examined differed significantly in their behavior toward other caterpillars, with P. cryptolechiella and P. reflexella being more aggressive than P. quercicella and P. quercinigracella. There were also behavioral differences within and between species when acting as an occupant or an intruder.
Psilocorsis cryptolechiella was more aggressive and P. quercinigracella was less aggressive when defending a shelter than when attempting to usurp a shelter. Despite facilitative effects on the arthropod community, this facilitation may be mediated by frequent direct competition for shelters that serve as a territory as well as a food source.
42
Introduction
Caterpillars in at least 24 families of Lepidoptera function as ecosystem engineers by tying leaves together with silk to make shelters (Lill and Marquis, 2007). Following the construction of leaf shelters, a wide array of other arthropods secondarily occupies the shelters benefitting from a combination of decreased temperature (Hensen, 1958), increased humidity (Willmer, 1980), decreased toxic effects of foliage consumption
(Fukui et al., 2002; Sagers, 1992), and/or protection from a variety of predators (Abarca and Boege, 2011; Damman, 1987; Loeffler, 1996; Sipura, 1999). The presence of leaf shelters has been shown to strongly affect arthropod abundance, species richness, and community structure (Lill and Marquis, 2003, 2004; Martinsen et al., 2000; Nakamura and Ohgushi, 2003; Wang et al., 2012). Despite the abundance of evidence for facilitative effects of leaf shelters on the arthropod community, the role of competition in mediating these interactions has largely been ignored.
In a recent review, Kaplan and Denno (2007) found that indirect interactions accounted for more than 65% of the documented cases of competition among herbivorous insects. By comparison, evidence for direct competition among insect herbivores for plant resources is sparse. Of the few examples in the literature demonstrating direct competition, many are among caterpillars that build shelters (Berenbaum et al., 1993;
Scott et al., 2010a; Yack et al., 2001). For example, the parsnip webworm
(Oecophoridae) has been shown to aggressively defend its leaf webs from conspecifics
(Berenbaum et al., 1993; Green et al., 1998). A few shelter-building drepanid caterpillar species have also been shown to use vibrational signals in competitive interactions with conspecifics (Bowen et al., 2008; Fletcher et al., 2006; Scott et al., 2010b; Scott and
43
Yack, 2012; Yack et al., 2001). However, these studies focus on describing the behaviors exhibited and have not compared behaviors across species or examined competition between multiple species. Many species of shelter-building caterpillars can be found co- occurring in the same shelter and are therefore likely to compete over the limited space and food resources contained therein.
Natural History
For the present study, I examined four caterpillar species that build shelters on white oak (Fageacea: Quercus alba) in Maryland forests: Psilocorsis reflexella, P. cryptolechiella, P. quercicella (Oecophoridae), and Pseudotelphusa quercinigracella
(Gelechiidae). These species were chosen because they are the most abundant shelter builders locally, they can be found in the field cohabitating within the same shelter, and they construct leaf shelters in a similar manner. These caterpillars build a type of leaf shelter known as a leaf tie, which is a pair of overlapping leaves attached together with silk to form a sandwich-type shelter. They build ‘walls’ within the overlapping leaves using silk combined with frass to fasten together the upper and lower leaf (Figure 1). For the purpose of this paper the ‘shelter’ will be defined as the area between walls of frass and silk created by the caterpillar and the ‘leaf tie’ will be defined as the larger area encompassing the entire area of leaf overlap. Thus, it is possible for more than one shelter to exist within one tie, but each individual caterpillar generally builds and resides in its own shelter.
Leaf-tying caterpillars are restricted to building shelters in areas of natural overlap between leaves, as they do not fold leaves and are incapable of pulling together leaves that do not touch (Carroll and Kearby, 1978). The number of leaves touching one another
44
on a tree has been shown to be strongly correlated with the density of leaf ties formed on that tree (Marquis et al., 2002), suggesting that areas to build leaf ties are a limiting resource. Leaf ties routinely experience very high occupancy rates as well as high rates of herbivory compared with non-tied leaves. For example, Lill (2004) found that 95% of experimental leaf ties created on white oak were colonized by arthropods within two weeks of their creation and were highly skeletonized (epidermis of leaf consumed) by the end of the season. In the eastern U.S., more than a dozen leaf-tying caterpillar species feed on white oak trees and frequently co-occur within the same leaf tie (Sigmon and Lill,
2013). Recent surveys of Maryland forests have found an average of 3 caterpillars and
1.5 species per leaf tie and up to 23 caterpillars and 7 species within a single leaf tie
(unpublished data). Beyond noting the frequent co-occurrences of multiple individuals of the same or different herbivore species (Lill and Marquis, 2004; Sigmon and Lill, 2013), no studies to date have examined whether disputes over the occupation of a leaf tie occur or if it is passively shared.
Spaces to build leaf ties are limited and high occupancy is common, indicating that they are a valuable resource over which herbivores are likely to compete. In the present study, focal species were observed in both interspecific and intraspecific interactions in the lab using artificial leaf ties, which allowed observations of the interactions within the ties without disturbing the caterpillars. I compared each species’ behavior using staged encounters with both conspecifics and heterospecifics. The objectives were to (1) describe the behaviors displayed in competitive interactions between shelter builders; (2) compare behaviors between species; (3) compare the
45
behavior of caterpillars defending a shelter to those trying to usurp a shelter; and (4) compare behavior towards conspecifics and heterospecifics.
Methods
Individual caterpillars of Psilocorsis reflexella, P. cryptolechiella, P. quercicella, and Pseudotelphusa quercinigracella were lab-reared or field-collected in Montgomery
Co., MD for use in competition trials. The caterpillars were reared in the laboratory in individual deli containers and fed white oak (Quercus alba) foliage that was replaced weekly. Penultimate and ultimate instar caterpillars were used in the trials, which were performed in the laboratory using artificial leaf ties.
Artificial leaf ties were constructed using one white oak leaf with the stem in an aquapic to which a piece of transparency paper was clipped to the upper surface using a spring-loaded metal curler clip (Marianna Industries, Omaha, NE). Large leaves with
<5% feeding damage were used. Prior to each trial, an individual caterpillar was placed in an artificial leaf tie and allowed to build a shelter for at least two days. During each trial, the aquapic was clamped to a ring-stand so that the leaf was held with the petiole down and leaf surface perpendicular to the floor. A second caterpillar, the ‘intruder,’ was introduced into the leaf tie containing the original ‘occupant,’ which was residing within the shelter (Figure 1). The intruder was placed 3-4 cm away from the shelter using a small paintbrush to handle the caterpillar. The length of both the occupant and the intruder were measured to the nearest 0.1 mm using dial calipers prior to each trial. The length was converted to mass using a regression equation for each species derived from length and mass measurements of individual caterpillars (Appendix 2.1). Pairs of
46
caterpillars had a range of size differences, from similar size up to a 5-fold difference in mass. The pair of caterpillars was then observed through the transparency paper for 30 min and was video recorded using a Panasonic video camera (Model PV-GS90P-S).
Individual caterpillars were used in only one trial each.
For each trial, I quantified how many times each caterpillar moved towards/away from the other caterpillar, pushed or hit the other caterpillar, and the final outcome of the interaction. The caterpillar that had possession of the shelter at the end of the 30 min trial was considered the ‘winner’ and the other caterpillar that either lost possession or failed to gain possession was considered the ‘loser.’ In some trials, both caterpillars remained in the shelter so the competitive interaction resulted in sharing the shelter. Each caterpillar was assigned an aggression score based on its series of behaviors during the encounter.
The aggression score was a total of points for each of the behaviors observed multiplied by the number of times the caterpillar performed that behavior. Caterpillars were given positive points for aggressive behaviors such as moving towards, pushing, hitting, and usurping a shelter and negative points for submissive behaviors such as moving away and sharing a shelter (Table 1). All points were summed for each individual caterpillar to create an overall aggression score. I staged 127 trials, of which the caterpillars interacted in 75 trials, leading to 150 caterpillars with aggression scores.
The difference in mass of each pair of competing caterpillars was expressed by a ratio of the caterpillar’s mass to that of its competitor (hereafter mass difference ratio).
Logistic regression was used to determine the effect of the mass difference ratio on whether an occupant caterpillar won or lost, excluding those that shared shelters (N=57).
The aggression score of all caterpillars that interacted (N = 150) was used to compare
47
behaviors within and among species. Aggression scores were cube-root-transformed to improve normality and homogenize variances. An analysis of variance (ANOVA) using type III sum of squares was performed using the transformed aggression score for each individual caterpillar as the response variable and species, occupancy (caterpillar acted as occupant or intruder) and competition type (interspecific or intraspecific) as fixed effects.
Two-way interactions between species and occupancy and between species and competition type were also included in the model. The aggression score of the opponent and the mass difference ratio were included as covariates. When effects were significant, post hoc tests were performed using Tukey HSD tests to determine differences among groups. Analyses were performed in R 2.14.2 (R Core Development Team, 2011). The
ANOVA was performed using the ‘car’ package (Fox and Weisberg, 2011). Adjusted means of aggression scores for figures were calculated according to the above model using the ‘effects’ package (Fox, 2003).
Results
Caterpillars were frequently observed pushing and hitting one another in or near a shelter until one caterpillar retreated. A typical interaction began with the intruder approaching the shelter. Usually the intruder would enter the shelter and initiate contact by pushing the occupant. Rarely, the occupant would move towards the intruder to push it away first (12% of interactions). The caterpillars would continue pushing and hitting one another until one caterpillar backed away from the shelter. In some trials (25% of interactions) this sequence occurred multiple times with one caterpillar retreating and later returning to the shelter to engage the occupant. In some instances (16% of
48
interactions), the caterpillar in control of the shelter would leave it to push or hit the other caterpillar without that caterpillar having approached the shelter. Caterpillars pushed by placing their head against the other caterpillar and contracting body segments to walk forward. They hit each other by rapidly moving the head laterally. Pushing was usually done inside the shelter and hitting was usually done outside of the shelter, likely because open space is required for lateral movement. Occasionally (< 20% of interactions) a caterpillar would tap the leaf surface with its head repeatedly creating an audible sound, but this could not be positively associated with any other behaviors (i.e., before an interaction, after being attacked or after a specific outcome). Interactions resulted in either the occupant maintaining possession of the shelter, the intruder usurping the shelter, or the two caterpillars sharing the shelter.
Caterpillars came into physical contact with one another in 59% of the trials (75 trials out of 127). In the remainder of the trials the caterpillars either remained stationary or the intruder never approached the shelter; these trials were not considered in further analyses. Of the 75 trials in which the caterpillars interacted, the occupant maintained possession of the shelter in 52% of the interactions. The intruder gained access to the shelter and the occupant left the shelter in 24% of the interactions and the two caterpillars shared the shelter in the remaining 24% of the interactions.
The occupant caterpillar was more likely to win a contest if it was larger than the intruding caterpillar (Wald’s χ2 = 7.19, P = 0.0073). The full model including species, occupancy, and competition type (with the aggression score of the competitor and mass difference ratio as covariates) significantly predicted the aggression score of individual caterpillars (F13,136 = 14.79, P < 0.0001). A caterpillar’s aggression score was positively
49
associated with both its competitor’s score (F1,136 = 86.26, P < 0.0001) and the mass difference ratio (F1,136 = 5.43, P = 0.0213). Caterpillars were more aggressive when their competitor was aggressive and when the competitors were similar sizes. The four caterpillar species examined differed significantly in aggression scores (F3,136 = 3.56, P =
0.0160; Figure 3). On average, Psilocorsis cryptolechiella and P. reflexella had 35% higher aggression scores than P. quercicella and P. quercinigracella. Intruder caterpillars had 10% higher aggression scores than occupant caterpillars (F1,136 = 7.75, P = 0.0062).
There was also a significant interaction between caterpillar species and occupancy (F3,136
= 7.69, P < 0.0001; Figure 3); Psilocorsis cryptolechiella was more aggressive and P. quercinigracella was less aggressive when defending a shelter than when attempting to usurp a shelter. The caterpillars behaved similarly when engaged in interspecific compared to intraspecific interactions (F1,136 = 1.49, P = 0.2243), regardless of caterpillar species (i.e., competition type x species interaction was not significant; F3,136 = 0.87, P =
0.4588).
Discussion
It is clear that direct competition over shelter resources frequently occurs among the common shelter-building caterpillars occupying white oak leaf ties. Psilocorsis reflexella, P. cryptolechiella, P. quercicella, and Pseudotelphusa quercinigracella often engaged in physically aggressive competitive ‘fights’ to gain or maintain access to existing leaf shelters. During these fights, caterpillars pushed and hit each other with their heads within or near the shelter until one caterpillar retreated. The behavioral interactions
50
observed in this study result from direct competition for a structure, territory and shelter as well as for a food source.
The original builder of the shelter, or ‘occupant,’ succeeded in defending its shelter in more than half of the trials. This may reflect differential costs incurred by the occupant and intruder associated with the shelter they compete over. The occupant has invested time and energetic costs in the construction of the shelter, whereas the intruder invests only as much energy as the physical competition requires. Also, if the intruder does not successfully usurp the shelter, then it is no worse off than before the competitive interaction (i.e., still without a shelter). Therefore the occupant likely incurs a greater cost in losing the shelter than the intruder does in not winning. Despite winning less frequently, the intruders were more aggressive than the occupants overall. This likely reflects the occupant’s advantage of already being in the shelter, which would require the intruder to push and hit more frequently to force the occupant to move. The intruder would thus need to be more aggressive in order to successfully usurp the shelter.
The consequences of leaving a shelter for any caterpillar are as yet unknown. In most trials the loser built a shelter in another portion of the same leaf tie and rarely crawled out of the tie. Previous studies in this system and others have shown that there is no detectable fitness cost in rebuilding a shelter in terms of growth rate, development time, or pupal mass (Abarca and Boege, 2011; Lill et al., 2007; Weiss, 2003). However, these studies were performed in laboratories and so do not reflect possible ecological costs, such as increased exposure to predators and parasitoids. When searching for a new site to build a shelter, caterpillars are highly vulnerable to predation and frequently fall off the tree (unpublished data). Some caterpillars require 1-3 days to build a complete
51
shelter (pers. obs.), so either usurping a pre-existing shelter or cohabitating with another caterpillar may alleviate this time cost, during which they may be particularly vulnerable to predation. It seems likely that predation risk is much higher outside of a shelter than within. However, although leaf shelters have been shown to provide protection from some predators (Abarca and Boege, 2011; Damman, 1987; Loeffler, 1996; Sipura, 1999), they seem to increase predation rates by others (Jones et al., 2002; Weiss et al., 2004), and parasitism rates can be substantial (Lill, 2001). Given the small size of these caterpillars, however, exposure to either natural enemies or the elements is likely to be deadly, placing a premium on quickly finding and occupying a shelter.
If sites to build shelters are rare, caterpillars may be forced to share leaf ties, in which case they may be more likely to become food-limited. Competition among cohabitants of a tie will increase as food resources decrease, as they only feed within leaf ties. It is unclear to what extent the observed competition is driven by need for the protection of a shelter, food resources or both. Trials similar to those performed here could be conducted with smaller or herbivore-damaged leaves to determine if caterpillars are more aggressive when food-limited. Further studies should also examine the amount of foliage required for each caterpillar species to complete development and compare that to the average size of natural leaf ties to determine to what extent caterpillars become food-limited when sharing leaf ties in nature.
The four species examined in this study differed significantly in the average level of aggression they exhibited in these staged competitive encounters. Psilocorsis reflexella and P. cryptolechiella were the most aggressive while P. quercicella and P. quercinigracella were the least aggressive. Many factors may contribute to these
52
interspecific differences in aggression. Each species constructs its shelters in a characteristic shape by depositing silk and frass in different patterns (Figure 1). The more aggressive P. reflexella and P. cryptolechiella each create a narrow tunnel with only two walls. This shape may be easier to defend because it has only two entrances. This type of shelter is also much more difficult to enter without coming into direct contact with the occupant, so shelter-sharing is often precluded. Psilocorsis quercicella builds one arched wall that it rests behind. Pseudotelphusa quercinigracella fills a large area with many strands of silk and frass that attach the leaves but rarely builds well-defined walls. These two larger, more open shapes are likely to be more difficult to defend because they can be entered from many points. They can also be more easily shared without the caterpillars coming into frequent contact.
Another notable trend is that caterpillar aggression increased with maximum size.
Final instar P. reflexella is twice as large as P. quercinigracella and the other two species are intermediate. Larger caterpillars require more food resources to complete development so may be more aggressive in defense of their territory to avoid food- limitation. A caterpillar was more likely to win a competitive interaction if it was larger than it’s opponent. This may be due to larger caterpillars being relatively stronger and pushing harder. Also, small caterpillars could be injured by larger caterpillars. I was unable to determine if the caterpillars in this study use their mandibles to bite as other caterpillars are reported to do (Okuda, 1989; Scott et al., 2010a), but a bite to the soft tissues of the thorax or abdomen could be fatal. However, it appears that P. quercinigracella overcomes this size disadvantage through different phenological patterns from Psilocorsis. All of these species are bivoltine and are abundant mid to late
53
summer, but P. quercinigracella emerges slightly before Psilocorsis, allowing for it to be the same size or larger than co-occurring caterpillars. Thus, the phenology of P. quercinigracella likely promotes coexistence between these species due to the size advantage it confers in competitive interactions.
Surprisingly, the caterpillars examined here did not differ in aggression when competing with conspecifics compared to heterospecifics. It is commonly believed that intraspecific competition must be stronger than interspecific competition for two species to coexist (Chesson, 2000; Tilman, 1982). However, aggression scores of caterpillars did not differ for trials involving interspecific vs. intraspecific pairs. This suggests that leaf tie resources are perceived as equally valuable for each species. These common, co- occurring species coexist at high densities so may be partitioning resources in some other way.
Each of the species examined in this study was observed to use vibratory signals that created an audible noise, but these behaviors were rare. The vibratory signals observed, hitting the leaf surface with the head, did not seem to be ritualized as has been described in other caterpillars (Scott et al., 2010b; Yack et al., 2001). Scott and colleagues (2010a) hypothesized that ritualized vibratory signals evolved from physical aggression behaviors to minimize the risk of injury. However, their study only compared caterpillars in one family, Drepanidae. The minimal use of vibratory signaling by the caterpillars in this study may represent an early evolutionary step towards the use of ritualized vibratory signals in competitive encounters. Use of both vibratory signaling and physical aggression has been described in one other shelter-building caterpillar, Tethea or
(Drepanidae), so may be more common than previously thought (Scott and Yack, 2012).
54
Clearly, direct competition between shelter builders is widespread so future studies of aggressive signaling behavior evolution would benefit from the inclusion of caterpillars in a range of families with diverse types of behaviors.
Only late-instar caterpillars were used in this study, but there may be ontogenetic changes in aggressive behavior. Early instar caterpillars of the same and different species have been observed building shelters very close to one another, possibly facilitating each other in pulling the leaves together (pers. obs., Carroll and Kearby, 1978). Because larger caterpillars consume more leaf tissue and are thus more likely to experience food- limitation, tolerant or facilitative behaviors would be expected to decrease as the caterpillars grow. In a similar manner, leaf miners have been shown to facilitate siblings during mine initiation (when they first hatch), but compete for unmined leaves when more developed (Damman, 1994). Future studies should compare the behaviors exhibited by leaf-tying caterpillars during each instar to determine how behaviors toward other caterpillars change ontogenetically. Additionally, Psilocorsis and Pseudotelphusa moths oviposit in clusters (Carroll and Kearby, 1978) so siblings generally share leaf ties, especially in the first few instars. Further studies should compare aggression towards siblings and non-siblings to determine if there is sibling recognition and differential behavior towards siblings and non-siblings, which may change ontogenetically.
In conclusion, direct competition among oak leaf-tying caterpillars is common and varies in intensity among species. The high frequency of physically aggressive encounters between shelter builders provides evidence that direct competition among herbivorous insects may be more common than previously thought. In spite of these aggressive interactions that routinely occur within shelters, co-occurrence of leaf tiers is
55
common and shelter-builders have strong net facilitative effects on the larger arthropod community (Lill and Marquis, 2003, 2004; Martinsen et al., 2000; Wang et al., 2012).
Leaf ties are a valuable resource around which arthropods communities are structured via competition and facilitation.
Acknowledgements
Funding provided by GWU Mortensen Fund and GWU Harlan Trust. Thanks to members of the Lill lab for assistance in caterpillar collection and the DC Plant-Insect Group (DC
PIG) for comments on early drafts.
References
Abarca M, Boege K, 2011. Fitness costs and benefits of shelter building and leaf
trenching behaviour in a pyralid caterpillar. Ecological Entomology 36:564-573.
doi: 10.1111/j.1365-2311.2011.01299.x.
Berenbaum MR, Green ES, Zangerl AR, 1993. Web costs and web defense in the parsnip
webworm (Lepidoptera: Oecophoridae). Environmental Entomology 22:791-795.
Bowen JL, Mahony SJ, Mason AC, Yack JE, 2008. Vibration-mediated territoriality in
the warty birch caterpillar Drepana bilineata. Physiological Entomology 33:238-
250. doi: 10.1111/j.1365-3032.2008.00627.x.
Carroll MR, Kearby WH, 1978. Microlepidopterous oak leaftiers
(Lepidoptera:Gelechioidea) in central Missouri. Journal of Kansas Entomological
Society 51:457-471.
56
Chesson P, 2000. Mechanisms of maintenance of species diversity. Annual Review of
Ecology and Systematics 31:343-+. doi: 10.1146/annurev.ecolsys.31.1.343.
Damman H, 1987. Leaf quality and enemy avoidance by the larvae of a pyralid moth.
Ecology 68:88-97.
Damman H, 1994. Defense and development in a gregarious leaf-mining beetle.
Ecological Entomology 19:335-343. doi: 10.1111/j.1365-2311.1994.tb00250.x.
Fletcher LE, Yack JE, Fitzgerald TD, Hoy RR, 2006. Vibrational communication in the
cherry leaf roller caterpillar Caloptilia serotinella (Gracillarioidea :
Gracillariidae). Journal of Insect Behavior 19:1-18. doi: 10.1007/s10905-005-
9007-y.
Fox J, 2003. Effect displays in R for generalised linear models. Journal of Statistical
Software 8:1-27.
Fox J, Weisberg S, 2011. An {R} companion to applied regression. Second ed Thousand
Oaks, CA: Sage.
Fukui A, Murakami M, Konno K, Nakamura M, Ohgushi T, 2002. A leaf-rolling
caterpillar improves leaf quality. Entomological Science 5:263-266.
Green ES, Zangerl AR, Berenbaum MR, 1998. Reduced aggressive behavior: A benefit
of silk-spinning in the parsnip webworm, Depressaria pastinacella (Lepidoptera :
Oecophoridae). Journal of Insect Behavior 11:761-772.
Hensen WR, 1958. The effects of radiation on the habitat temperatures of some poplar-
inhabiting insects. Canadian Journal of Zoology 36:463-478.
57
Jones MT, Castellanos I, Weiss MR, 2002. Do leaf shelters always protect caterpillars
from invertebrate predators? Ecological Entomology 27:753-757. doi:
10.1046/j.1365-2311.2002.00465.x.
Kaplan I, Denno RF, 2007. Interspecific interactions in phytophagous insects revisited: a
quantitative assessment of competition theory. Ecology Letters 10:977-994. doi:
10.1111/j.1461-0248.2007.01093.x.
Lill JT, 2001. Selection on herbivore life-history traits by the first and third trophic
levels: The devil and the deep blue sea revisited. Evolution 55:2236-2247.
Lill JT, Marquis RJ, 2003. Ecosystem engineering by caterpillars increases insect
herbivore diversity on white oak. Ecology 84:682-690.
Lill JT, Marquis RJ, 2004. Leaf ties as colonization sites for forest arthropods: an
experimental study. Ecological Entomology 29:300-308.
Lill JT, Marquis RJ, 2007. Microhabitat manipulation: ecosystem engineering by shelter-
building insects. In: Cuddington K, J. E. Byers, W. G. Wilson, and A. Hastings,
editor. Ecosystem engineers: plants to protists New York: Academic Press. p.
107-138.
Lill JT, Marquis RJ, Walker MA, Peterson L, 2007. Ecological consequences of shelter
sharing by leaf-tying caterpillars. Entomologia Experimentalis Et Applicata
124:45-53. doi: 10.1111/j.1570-7458.2007.00546.x.
Loeffler CC, 1996. Adaptive trade-offs of leaf folding in Dichomeris caterpillars on
goldenrods. Ecological Entomology 21:34-40.
Marquis RJ, Lill JT, Piccinni A, 2002. Effect of plant architecture on colonization and
damage by leaftying caterpillars of Quercus alba. Oikos 99:531-537.
58
Martinsen GD, Floate KD, Waltz AM, Wimp GM, Whitham TG, 2000. Positive
interactions between leafrollers and other arthropods enhance biodiversity on
hybrid cottonwoods. Oecologia 123:82-89.
Nakamura M, Ohgushi T, 2003. Positive and negative effects of leaf shelters on
herbivorous insects: linking multiple herbivore species on a willow. Oecologia
136:445-449. doi: 10.1007/s00442-003-1285-5.
Okuda T, 1989. Aggressive characteristics of diapausing larvae of a stem borer, Busseola
fusca Fuller (Lepidoptera: Noctuidae) in artificially crowded conditions. Applied
Entomology and Zoology 24:238-239.
R Core Development Team. 2011. R: A language and environment for statistical
computing Vienna, Austria R Foundation for Statistical Computing.
Sagers CL, 1992. Manipulation of host plant-quality - Herbivores keep leaves in the dark.
Functional Ecology 6:741-743.
Scott JL, Kawahara AY, Skevington JH, Yen SH, Sami A, Smith ML, Yack JE, 2010a.
The evolutionary origins of ritualized acoustic signals in caterpillars. Nature
Communications 1:9. doi: 410.1038/ncomms1002.
Scott JL, Matheson SM, Yack JE, 2010b. Variation on a theme: Vibrational signaling in
caterpillars of the rose hook-tip moth, Oreta rosea. Journal of Insect Science
10:15. doi: 54.
Scott JL, Yack JE, 2012. Vibratory territorial signals in caterpillars of the poplar
lutestring, Tethea or (Lepidoptera: Drepanidae). European Journal of Entomology
109:411-417.
59
Sigmon E, Lill JT, 2013. Phenological variation in the composition of a temperate forest
leaf tie community. Environmental Entomology 42:1-9.
Sipura M, 1999. Tritrophic interactions: willows, herbivorous insects and insectivorous
birds. Oecologia 121:537-545.
Tilman D, 1982. Resource competition and community structure. Princeton: Princeton
University Press.
Wang HG, Marquis RJ, Baer CS, 2012. Both host plant and ecosystem engineer identity
influence leaf-tie impacts on the arthropod community of Quercus. Ecology
93:2186-2197.
Weiss MR, 2003. Good housekeeping: why do shelter-dwelling caterpillars fling their
frass? Ecology Letters 6:361-370. doi: 10.1046/j.1461-0248.2003.00442.x.
Weiss MR, Wilson EE, Castellanos I, 2004. Predatory wasps learn to overcome the
shelter defences of their larval prey. Animal Behaviour 68:45-54. doi:
10.1016/j.anbehav.2003.07.010.
Willmer PG, 1980. The effects of fluctuating environment on the water relations of larval
Lepidoptera. Ecological Entomology 5:271-292.
Yack JE, Smith ML, Weatherhead PJ, 2001. Caterpillar talk: Acoustically mediated
territoriality in larval Lepidoptera. Proceedings of the National Academy of
Sciences of the United States of America 98:11371-11375.
60
Table 2.1. Descriptions of the behaviors observed and their point values used to calculate an aggression score for each individual caterpillar in each trial. The points were multiplied by the number of times a caterpillar was observed performing each action and summed for each caterpillar.
Behavior Description of Behavior Point value Move Toward Crawl toward other caterpillar, each crawling 0.5 movement between stops in movements are scored as 1 move Move Away Crawl away from other caterpillar, each -0.5 crawling movement between stops in movements are scored as 1 move Push Place their head against the other caterpillar 1 and contract body segments to walk forward, each contraction of body segments is 1 push Hit Rapid movement of head laterally that results 1 in head coming into contact with other caterpillar Enter Shelter Crawling into the boundaries of the shelter 2 Attack Away from Resident caterpillar exits shelter to push or hit 2 Shelter the other caterpillar without that caterpillar having approached the shelter Usurp/Defend Shelter The result of a competitive interaction when 5 one caterpillar retreats from the shelter and the other remains within the shelter Share Shelter The result of a competitive interaction when -5 both caterpillars remain within the shelter
61
Figure Legends
Figure 2.1. Characteristic shelter shapes of A) Psilocorsis cryptolechiella, P. reflexella,
B) P. quercicella and C) Pseudotelphusa quercinigracella.
Figure 2.2. Diagram of artificial leaf tie used in competitive trials. (A) White oak leaf clipped to (B) transparency paper comprises the (C) leaf tie. The (D) shelter is built within the leaf tie by the (E) occupant caterpillar and the (F) intruder caterpillar is placed in the leaf tie at the beginning of the trial.
Figure 2.3. Mean aggression score (and 95% confidence interval) for all caterpillars of
Psilocorsis cryptolechiella, P. reflexella, P. quercicella, and Pseudotelphusa quercinigracella acting as an intruder (white bars) or an occupant (grey bars) adjusted according to the model. The scores are based on the number of times each caterpillar performed certain behaviors (Table 1). Letters (species by occupancy interaction) and lines (species main effect) above bars indicate significant differences according to Tukey
HSD test and numbers at the bottom of the bars denote the sample size.
62
Figure 2.1
63
Figure 2.2
64
Figure 2.3
65
CHAPTER 3: CATERPILLAR PRESENCE, BUT NOT SPECIES IDENTITY,
INFLUENCES ARTHROPOD COMMUNITY ASSEMBLY IN LEAF SHELTERS
Abstract
Physical ecosystem engineers, organisms that alter their abiotic environment and thereby affect associated biota, influence local community composition both directly and indirectly via their constructs. Leaf-tying caterpillars act as ecosystem engineers by building shelters between leaves that are subsequently colonized by other arthropods. The arthropod community varies widely between shelters, which may be influenced by ecological and behavioral competitive interactions as well as the sequence of colonizers.
To examine the effects of leaf-tying caterpillar species identity on the assembly of the arthropod community within leaf ties, we created experimental ties that were initially empty or occupied by one of three species of caterpillar. We allowed the ties to be naturally colonized for 1-2 weeks to examine the initial stages of community assembly when the caterpillar is most likely to interact with potential colonists. There were 50% more arthropods and 30% more species that colonized leaf ties after two weeks compared to one week. In addition, arthropod abundance and species richness were 20% higher and different communities assembled in ties initially containing a caterpillar compared to initially empty controls. Herbivores and predators were significantly more abundant in leaf ties that initially contained a caterpillar relative to initially empty control ties. There were no measurable differences in the communities associated with different species of original occupant caterpillars, but arthropod community composition differed between leaf ties where the initial occupant remained in the tie throughout the study compared to
66
ties that had been abandoned. This indicates that the arthropod community inhabiting leaf ties responds to the presence of a leaf-tying caterpillar, but not to its species identity.
Introduction
Ecosystem engineers are organisms that modify their abiotic environment in various ways, which in turn affects associated biota as well as their interactions (Jones et al. 1994, 1997). This widespread phenomenon includes a variety of organisms from beavers creating ponds behind their dams, to coral and oyster reefs providing a habitat for other organisms, to cyanobacterial mats that alter desert hydrology (Naiman et al. 1986,
Eldridge and Greene 1994, Ruesink et al. 2005, Buhl-Mortensen et al. 2010). However, not all engineers are created equal. Many groups of organisms engineer their habitat in a similar way, but generate different outcomes. For example, burrows made by vizcachas are more suitable for use by nesting owls than those made by armadillos, because vizcachas thin vegetation around burrows (Machicote et al. 2004). Ant nests alter soil nutrient cycling which affects vegetation composition. However, ant-species-specific microbial communities mediate this effect and alter soil nutrients in different ways (Boots et al. 2012). While trophic interactions do not constitute ecosystem engineering on their own, they can be an important structuring force in communities associated with engineers. For example, earthworms act as ecosystem engineers and facilitate salamanders by creating tunnels. Earthworms are also prey for the salamanders and indirectly compete with and reduce other salamander prey (Ransom 2012). The effects of ecosystem engineers on a community can be complex, involving facilitation via engineering, competition and trophic interactions.
67
When ecosystem engineers create new habitats, the identity of the engineer may have a significant impact on the community that colonizes that habitat. The effects of early colonists on the community assembly process, termed priority effects, have been examined in many systems (Wilbur and Alford 1985, Drake 1991, Keddy 1992, Lawler and Morin 1993, Belyea and Lancaster 1999). However, most experimental studies of priority effects involve only a few species in mesocosms and the priority effects are driven by the first colonist consuming the majority of a shared resource (Wilbur and
Alford 1985, Robinson and Dickerson 1987, Lawler and Morin 1993, Fincke 1999,
Collinge and Ray 2009). Competition is a common driver of priority effects, but facilitation can also affect the community assembly process, and the two are not mutually exclusive (Connell and Slatyer 1977). The mode of interaction among colonists and the traits of individual species strongly influence the composition of the resulting assemblage. When multiple species of engineers create the same type of habitat, the identity of the engineer can lead to priority effects. For example, different species of bark- and wood-boring insects in decaying trees can either facilitate or inhibit fungal species and later colonizing wood-borers (Weslien et al. 2011). Priority effects among ecosystem engineers are likely to be strong, but have rarely been studied.
Caterpillars that use silk to build shelters out of leaves comprise a widespread and varied guild of ecosystem engineers (Fukui 2001, Lill and Marquis 2007). Following the construction of leaf shelters, a wide array of other arthropods occupies the shelters both concurrently with the caterpillar and after the caterpillar has abandoned the shelter.
Shelter colonists include other herbivores (e.g., shelter-building and non-shelter-building caterpillars, beetles, aphids, hemipterans), predators (e.g., spiders, beetles, centipedes,
68
hemipterans), and scavengers (e.g., thrips, beetles, psocids, collembolans; Sigmon and
Lill 2013). Occupants of leaf shelters can profit from an array of benefits, including decreased temperature (Hensen 1958), increased humidity (Wilmer 1980), decreased toxic effects of foliage consumption (Sagers 1992, Fukui et al. 2002), and/or protection from a variety of predators (Damman 1987, Loeffler 1996, Sipura 1999, Abarca and
Boege 2011). The presence of leaf shelters increases arboreal arthropod abundance and species richness 3 to 10-fold compared to leaves without shelters, and significantly alters arthropod community structure on the whole tree (Martinsen et al. 2000, Lill and Marquis
2003, Nakamura and Ohgushi 2003, Lill and Marquis 2004, Wang et al. 2012). The ecological significance of these engineered resources is enhanced by the fact that leaf shelters tend to remain intact long after the initial builder is gone. While many benefits of shelter occupancy are known, they do not fully account for the impact of shelters on the arthropod community. Some benefits of shelters, perhaps relating to the shelter-building caterpillar itself, remain unexplored. The strong facilitative effect of shelter-building caterpillars on the forest arthropod community has been demonstrated repeatedly, but the mechanism behind this effect is largely unknown.
Despite the discovery of aggressive behaviors by many shelter-building caterpillars, the role of behavioral interactions in structuring the community within leaf shelters has not been explored. Several caterpillar species have been shown to compete over shelters by pushing and hitting each other and by using vibratory signals
(Berenbaum et al. 1993, Yack et al. 2001, Fletcher et al. 2006, Scott et al. 2010, Guedes et al. 2012, Scott and Yack 2012, Sigmon in prep). Competition has been shown to be a strong driving force in the assembly of many communities (Shulman et al. 1983, Lawler
69
and Morin 1993, Blaustein and Margalit 1996, Kennedy et al. 2009), but has not yet been examined as a structuring force in leaf shelter communities. Caterpillars that build leaf ties (i.e., a shelter between overlapping leaves) on white oak trees have recently been shown to push and hit each other in competition over a shelter, with some species being more aggressive than others (Sigmon in prep). Within leaf ties, the ‘shelter’ is a small area where silk is used to attach leaves together and the ‘leaf tie’ encompasses the entire area between the overlapping leaves that can be inhabited by other arthropods. Thus, there can be multiple shelters within one leaf tie. Oak leaf-tying caterpillars can potentially interact with over 100 other species of arthropods within leaf ties, including a dozen species of leaf-tying caterpillars (Sigmon and Lill 2013). In staged encounters between leaf-tiers, behavioral differences between caterpillar species affected the outcome of their interactions (Sigmon in prep), so we predicted that these behaviors influence the assembly of the communities within leaf ties.
This study aims to answer the following question: does the species identity of the original leaf-tier influence the assembly process of the leaf tie arthropod community? To address this question, we established experimental leaf ties in a Maryland forest and allowed them to be naturally colonized by the arthropod community. Experimental ties included initially empty controls and ties that contained one of three of the most common oak leaf-tiers, which vary in aggression towards other caterpillars: Psilocorsis cryptolechiella (Oecophoridae), P. quercicella, and Pseudotelphusa quercinigracella
(Gelechiidae). Psilocorsis cryptolechiella is the most aggressive, P. quercicella is moderately aggressive, and P. quercinigracella is the least aggressive when defending its own shelter against intruders (Sigmon in prep). The larval lifespan of these caterpillars
70
averages 4 weeks so we chose to examine the arthropod community within the first two weeks of leaf tie colonization, when the original caterpillar is most likely to be present in the leaf tie. We examined the arthropod community that assembled one and two weeks after the tie was initially built to determine how interactions between the caterpillar and subsequent colonists influence the initial assembly of the leaf tie community.
Methods
Experimental Design
We created experimental leaf ties by clipping together adjacent leaves with metal hair curler clips on small, understory white oaks, Quercus alba (Fagaceae), in Little
Bennett Regional Park, Montgomery Co., Maryland in the summers of 2011 and 2012.
This procedure for creating ties has been used successfully in numerous prior experiments (Lill and Marquis 2003, 2004, Wang et al. 2012, Sigmon and Lill 2013).
Experimental ties are colonized quickly with up to 97% of ties colonized within two weeks (E. Sigmon unpublished data). Also, the presence of the clip does not affect arthropod abundance (Martinsen et al. 2000, Lill and Marquis 2004). Five experimental leaf ties were made on a single branch, which served as an experimental replicate. Two such replicates were placed on opposite sides of a single white oak tree. The tree was treated as a block to account for between-tree differences in arthropod communities. Each replicate contained one of each of the following ties: 1) control without an occupant or cues (Control 1); 2) control without an occupant but with occupancy-related cues (i.e., feeding damage, silk and frass; Control 2); 3) occupied by Psilocorsis cryptolechiella; 4) occupied by P. quercicella; and 5) occupied by Pseudotelphusa quercinigracella (Table
71
1). Each species was used as an initial colonist in one tie per replicate (Treatments in
Table 1). The Control 2 tie was created by placing a P. cryptolechiella caterpillar in the tie and removing it after 2 days. Psilocorsis cryptolechiella was used for this control because it was the most abundant species in the laboratory colony. Insects often rely on chemical cues to find food so many arthropods are likely to be drawn to leaf ties due to the volatile cues from frass, silk, and feeding damage (Mattiacci and Dicke 1995,
Pettersson et al. 2005). By using two controls, one with cues and one without caterpillar cues, we were able to determine which colonists responded to occupancy-related cues and how those responses were altered by the presence of a caterpillar.
Mid-instar caterpillars from multiple families were obtained from established laboratory colonies and supplemented with field-collected individuals to randomize genetic relatedness. We measured the length of each caterpillar to the nearest 0.1 mm prior to the experiment using dial calipers. This allowed us to estimate the size of the original occupant upon collection based on previously determined growth curves, which assisted in determining if the original occupant remained in the tie throughout the experiment. To create experimental leaf ties, leaves were overlapped and fastened together such that the area within all ties was approximately equal (around 40 cm2). Only relatively undamaged leaves (< 5% damage) were used, in order to control for potential effects of prior damage and to provide an adequate food source throughout the experiment. Before the experiment began, all arthropods were removed from the leaves.
After caterpillars were placed in ties, all ties were bagged using netting for two days. It takes up to two days for these species to build the walls of silk and frass that define the boundaries of their shelter (E.S. pers. obs.). Bagging the ties allowed the caterpillars time
72
to complete this task without being disturbed by intruders and allowed us to verify that the initial caterpillar created a shelter in the tie they were placed in. When the bags were removed after two days, the ‘extra’ P. cryptolechiella caterpillar was removed from the
Control 2 tie. The leaves of this tie were pulled apart and then clipped back together so that any subsequent occupants were required to spin silk to fasten the leaves together, as in Control 1.
One or two weeks after bag removal, the ties were sampled as described below (in
Community Sampling Protocol). In 2011, all ties were sampled after one week. In 2012, one replicate per tree (a full set of experimental ties from one side of the tree) was sampled after one week and the remaining replicates were sampled after two weeks
(different ties sampled each week) for a fully-crossed factorial design. Communities were sampled at one or two weeks because the larval lifespan of these caterpillars is approximately four weeks, and we wanted to conclude the experiment before the initial colonist completed development. Two sampling periods totaling 27 replicates were conducted in 2011. In 2012, five sampling periods were conducted with a total of 110 replicates sampled (57 after one week and 53 after two weeks). However, due to limited caterpillar availability, 24 blocks in 2 of the sampling periods in 2012 (9 replicates in one sampling period and 15 in another) did not include a tie containing P. quercinigracella
(13 collected after one week and 11 after two weeks) leading to an unbalanced design.
However, other replicates in these sampling periods did contain a P. quercinigracella treatment so P. quercinigracella was present in all sampling periods.
73
Community Sampling Protocol
Arthropods colonize a leaf tie by one of two means: migration and oviposition.
After one or two weeks of natural colonization, we opened the experimental ties in the field and identified all arthropod migrants, except mites and other microscopic animals.
This was done in the field rather than bringing unopened ties to the laboratory for inspection to ensure that all ties were inspected on the same day and that no further predation occurred. We noted if the caterpillars originally placed in the experimental ties were present at the time of collection, hereafter ‘engineer presence.’ All caterpillars were returned to the laboratory and reared until identification was possible. Any arthropods that could not be identified in the field were stored in alcohol and later identified with the aid of a dissecting microscope. All arthropods were identified to species or morphospecies based on distinguishing characteristics. Morphospecies classification is frequently used for large community data sets and corresponds closely to expert taxonomic identification (Oliver and Beattie 1996, Longino and Colwell 1997).
To determine whether the presence of a caterpillar influenced oviposition choice of adult arthropods, the leaves used for each tie treatment were removed from the tree and brought into the laboratory after being cleared of migrant arthropods. Leaves were stored in plastic deli containers with moist filter paper and checked for hatching arthropods every 2-3 days for one week after being removed from the tree as in Lill et al. (2007).
Analysis
The original occupants that were placed in the treatment ties, if still present, were not included in any analyses We compared the number of ties that were colonized/not colonized by arthropods among treatments and weeks, blocked by tree, using a logit
74
model. We also used a logit model to determine if engineer presence differed among treatments and weeks, blocked by tree.
We examined the effects of treatment tie type and weeks available for colonization on species richness, arthropod abundance, and community composition using a blocked factorial design. We compared arthropod richness and abundance per tie among treatments, weeks and treatment by week interaction, blocked by tree and sampling period, using a generalized linear model (GLM). Species richness followed a
Poisson distribution and abundance data followed a negative binomial distribution. When main effects were significant, post hoc tests were performed using Tukey’s HSD test. The composition of the arthropod communities in each tie was compared among treatments, weeks and the treatment-by-week interaction using a blocked permutational multivariate analysis of variance (PERMANOVA) on Bray-Curtis dissimilarities derived from species abundance data. Because the dissimilarity of empty samples cannot be calculated, a dummy species column with a value of 1 was added to all samples prior to analysis (as in: Wang et al. 2012). To visualize treatment and week effects on arthropod community composition we used non-metric multi-dimensional scaling (NMDS; Kruskal 1964). This technique displays compositional differences between samples in 2-dimensional space.
We summed the arthropods within each species on a single tree for each treatment (i.e. summed 2 ties for each treatment and 5 ties for each week). We performed the NMDS using Bray-Curtis dissimilarity measure and added dispersion ellipses for each treatment.
The stress values all of the NMDS ordinations performed were rather high (.26-.29), and this is likely due to the very high number of 0 values in the data set. To determine which species contributed to treatment differences, we performed a similarity percentage
75
analysis (SIMPER) on the community matrix comparing among treatments and weeks.
This analysis identifies which species contribute the most to between-group dissimilarity up to a cumulative contribution of 70% of the dissimilarity.
We also examined differences between only the treatments that initially contained a caterpillar to examine the effects of engineer presence/absence. A GLM on abundance and species richness and a PERMANOVA, NMDS and SIMPER on community composition were performed as above on a subset of the data including ties only initially containing a caterpillar (i.e., excluding control ties) and the additional predictor of engineer presence (present/absent). We also explored how behavioral differences between caterpillar species may have influenced leaf-tying caterpillars that migrated into the ties. Caterpillar size is a significant predictor of its ability to usurp a shelter (Sigmon in prep), so we expected larger caterpillars to colonize ties made by the most aggressive leaf-tier, P. cryptolechiella. We compared the size of colonist leaf-tying caterpillars
(N=36), relative to the size of the initial occupant (larger, same or smaller) between species of initial occupants using Fisher’s exact test. We also compared engineer presence/absence among caterpillar species in ties that were colonized by leaf-tying caterpillars using Fisher’s exact test.
To determine if trophic groups responded differently to the treatments we classified each species as herbivore, predator, scavenger, adult parasitoid, or unknown juvenile using information from published literature and personal observations. There were too few adult parasitoids and unknown juveniles to perform analyses so we only examined herbivore, predator, and scavenger abundance. The total number of arthropods per tie from each trophic group was compared between treatments and weeks blocked by
76
tree and sampling period using GLM with negative binomial distribution. We also performed correlations among abundance of each pair of trophic groups and adjusted the p-values using a Bonferroni correction. All analyses were performed in R version 2.15.2
(R Development Core Team 2012). The NMDS, PERMANOVA, and SIMPER were performed in the ‘vegan’ package (Oksanen et al. 2012). Tukey’s HSD test was performed using the ‘agricolae’ package (de Mendiburu 2012). Adjusted means and confidence intervals for figures were computed using the ‘effects’ package (Fox 2003).
Results
In total, we collected 2067 arthropods from 137 morphospecies from experimental white oak leaf ties. Of these, 597 arthropods hatched from the ties (29%) but consisted of only 15 species (Appendix 3.1). Experimental ties were colonized quickly, though more ties were colonized after two weeks than one (89% and 80% occupied respectively; χ2 = 15.43, df = 1, p < 0.0001). Ties initially containing a caterpillar had 8% higher arthropod occupancy than initially empty control ties, but this difference was not significant (χ2 = 8.022, df = 4, p = 0.0908).
The arthropod community within experimental leaf ties that initially contained a caterpillar differed from that in controls that were initially empty. Colonists that migrated into the leaf ties and those that colonized via oviposition were combined for analyses, though when analyzed separately they produce qualitatively the same results (data not shown). Ties that initially contained a caterpillar were colonized by 20% more species (χ2
= 12.92, df = 4, p = 0.0117; Figure 1A) and 20% more arthropods per tie (χ2 = 11.74, df
= 4, p = 0.0033; Figure 1B) relative to Control 2 ties (with occupancy related cues).
77
However, arthropod abundance and species richness were similar among the Control 1 ties (without cues) and ties that initially contained a caterpillar (Figure 1). Leaf tie colonization increased over time, with 30% more species (χ2 = 37.28, df = 1, p < 0.0001;
Figure 1A) and 50% more arthropods per tie (χ2 = 70.78, df = 1, p < 0.0001; Figure 1B) colonizing ties after two weeks than after one week. Furthermore, the arthropod community differed in composition among tie treatments (F4, 580 = 2.03, p <0.001; Figure
2) and between weeks (F1, 580 = 23.32, p < 0.001; Figure 3). Thus, not only did more arthropods and species colonized caterpillar-initiated ties, but different species colonized these ties. There were no significant interactions between treatment and week on species richness (χ2 = 0.22, df = 4, p = 0.9935), arthropod abundance (χ2 = 1.07, df = 4, p =
0.8989), or community composition (F4, 580 = 0.75, p = 0.873) indicating that differences between treatments were consistent through time.
According to the SIMPER analysis, the psocid Polypsocus corruptus (Psocoptera:
Amphipsocidae), the common leaf-tying caterpillar Psilocorsis quercicella, an unidentified thrips (Thysanoptera), and an unidentified collembolan (Hypogastruridae) contributed to nearly 50% of the dissimilarity among treatments and weeks. Other arthropods that contributed to up to 70% of the treatment and week differences included the leaf beetle Metachroma sp. (Coleoptera: Chrysomelidae), the Asiatic Oak Weevil
Cyrtepistomus castaneus (Coleoptera: Curculionidae), the rove beetle Stictocranius sp.
(Coleoptera: Staphylinidae), the aphid Myzocallis punctata (Hemiptera:
Drepanosiphidae), an unidentified aphid, an unidentified predaceous beetle larva and another unidentified thrips (Thysanoptera). In addition, the sac spider Elaver sp.
(Araneae: Clubionidae) and an unidentified wooly aphid (Hemiptera: Eriosomatidae)
78
contributed to the differences between the empty control ties and the ties initially containing a caterpillar. In general, ties containing caterpillars contained more predators, thrips, and leaf-tying caterpillars while empty controls contained more thrips and aphids.
Juvenile and oviposited insects increased after two weeks and herbivorous beetles decreased.
The sampling period was significantly associated with species richness (χ2 =
13.77, df = 1, p = 0.0002), arthropod abundance (χ2 = 18.70, df = 1, p < 0.0001), and community composition (F1, 580 = 7.31, p < 0.001). However, post hoc tests showed that sampling periods in different years were similar to each other and there were no obvious seasonal trends, indicating that the patterns shown are consistent across months and years.
Effect of engineer presence
The original caterpillar that was placed in the tie was half as likely to be present in the tie after two weeks than after one week (χ2 = 37.79, df = 1, p < 0.0001; Figure 4).
Irrespective of week, the most aggressive caterpillar, P. cryptolechiella, was most likely to remain in the tie, followed by P. quercicella; P. quercinigracella, the least aggressive caterpillar, was least likely to remain in the tie (χ2 = 26.21, df = 2, p < 0.0001; Figure 4).
When comparing only ties that initially contained caterpillars, the treatment (i.e., species identity) was no longer a significant predictor of arthropod abundance (χ2 = 1.50, df = 2, p = 0.4709), species richness (χ2 = 0.19, df = 2, p = 0.9084) or community composition (F2, 309 = 0.95, p = 0.501), indicating that species identity did not alter the effects of leaf-tiers on the community. Arthropod abundance was marginally higher in abandoned ties (χ2 = 2.917, df = 1, p = 0.0877), but species richness was not affected by
79
engineer presence (χ2 = 1.41, df = 1, p = 0.2359). However, engineer presence did significantly affect the composition of the arthropod community (F1, 308 = 6.43, p < 0.001;
Figure 5). The compositional differences in the arthropod community between ties where engineers were present or absent were primarily attributed to the same four species that influenced the differences between treatments and weeks: the psocid Polypsocus corruptus (Psocoptera: Amphipsocidae), Psilocorsis quercicella, an unidentified thrips
(Thysanoptera), and an unidentified collembolan (Hypogastruridae). The dissimilarity between engineer presence/absence was also attributed to the leaf beetle Metachroma sp.
(Coleoptera: Chrysomelidae), the Asiatic Oak Weevil Cyrtepistomus castaneus
(Coleoptera: Curculionidae), the rove beetle Stictocranius sp. (Coleoptera:
Staphylinidae), the sac spider Elaver sp. (Araneae: Clubionidae), an unidentified predaceous beetle larva, another unidentified thrips (Thysanoptera), and an unidentified wooly aphid (Hemiptera: Eriosomatidae). These species were all more abundant in abandoned ties. Week was still a strong predictor of arthropod abundance (χ2 = 45.50, df
= 1, p < 0.0001), species richness (χ2 = 17.45, df = 1, p < 0.0001), and community composition (F1, 308 = 15.05, p < 0.001) in the reduced data set, but there was no interaction between week and engineer presence in any of these measures.
The relative size of leaf-tying caterpillar colonists did not differ between initial colonist species (p = 0.7567). However, in ties subsequently colonized by leaf-tiers, P. cryptolechiella was more likely to be present in the tie whereas P. quercicella and P. quercinigracella were more likely to have abandoned the tie at the time of collection (p =
0.0406).
80
Arthropod trophic groups
We collected 55 species of herbivores, 48 species of predators, 16 species of scavengers, 12 species of adult parasitoids, and 6 different nymphs that could not be identified below the order level. Herbivore and predator abundance per tie was correlated
(r = 0.14, p = 0.0002), but no other pairs of trophic groups were significantly correlated.
Herbivores (χ2 = 17.99, df = 4, p = 0.0012) and predators (χ2 = 35.67, df = 4, p < 0.0001) were up to twice as abundant in leaf ties that initially contained a caterpillar compared to controls (Figure 6). However, scavengers had similar abundances in all treatments (χ2 =
6.79, df = 4, p = 0.1476; Figure 5). There were twice as many herbivores (χ2 = 34.73, df
= 1, p < 0.0001), 1.5 times as many predators (χ2 = 9.19, df = 1, p = 0.0024), and twice as many scavengers (χ2 = 26.19, df = 1, p < 0.0001) per tie after two weeks of colonization than one week (Figure 5).
Discussion
We found strong differences between the arthropod community that colonizes leaf ties occupied by a caterpillar and the community that colonizes empty leaf ties. The presence of the ecosystem engineer attracts more arthropods, more species, and a compositionally different community than the altered habitat of the leaf tie alone. This indicates that the presence of a leaf-tier enhances the ecosystem engineering effect of the shelters. However, the species identity of the leaf-tier does not affect the arthropod community composition, richness or abundance, indicating that these ecosystem engineers are interchangeable. Leaf-tying caterpillar species vary widely in their relative abundance between forests on a local scale (E. S. unpublished data) and between states
81
(Carroll and Kearby 1978, Lill and Marquis 2003, Sigmon and Lill 2013). Despite differences in the leaf-tying caterpillar community composition, their engineering effect will likely be consistent across sites and their effects as a guild may be considered equivalent to the effects of each individual species.
The benefit of the presence of a leaf-tier to the community may simply be that the leaf-tier maintains the shelter, preventing it from coming apart. Over time, the silk in unoccupied shelters degrades and is broken by arthropods moving in and out. Leaf-tying caterpillars constantly add silk to their shelter so that a leaf tie containing a caterpillar is more sturdy and dependable than a tie that has been abandoned. Older shelters that are not maintained by caterpillars may not provide the same protection against abiotic stress and predation. Many of the occupants of leaf ties are juvenile arthropods that are prone to desiccation. The increased humidity in a well-maintained leaf tie would be particularly beneficial in the hot summer months, which is when leaf ties have the highest density of occupants (Sigmon and Lill 2013). Additionally, dislodgement from the tree by wind or rain is less likely in a leaf tie (Loeffler 1996). Leaf-tying caterpillars are rarely found in beat sampling (i.e., sampling done by beating a tree branch with a stick and catching falling caterpillars; per. obs.), suggesting that residing within a well-maintained tie lessens the likelihood of being knocked off a leaf. For small arthropods, falling off a tree is most likely deadly, as it would be difficult to find and climb a tree and there are many predators in the leaf litter.
Another possible benefit of caterpillar presence may be that small arthropods suffer decreased predation by associating with larger caterpillars. Birds, wasps, ants, predaceous beetle larvae, and spiders are common predators of leaf-tying caterpillars.
82
These predators, particularly the visually-oriented ones, may be more likely to consume the caterpillar in a leaf tie rather than the other much smaller arthropods. While this theory has yet to be experimentally tested, the increase in both herbivore and predator abundance in ties containing a caterpillar is consistent with this scenario. Also, many of the predators found were juveniles and likely to be prey items for birds and larger arthropods as well. The only predation events we have witnessed in a tie are a beetle larva eating caterpillars and aphids, and a spider eating an ant. It is surprising that herbivores are more abundant in ties containing a caterpillar since they are likely to compete with it for food resources. Because leaf ties are continually colonized, even when most of the leaf epidermis has been consumed (pers. obs.), the benefits of microclimate and protection from larger predators may outweigh the costs of being near meso-predators and food limitation.
A few other studies have compared occupied to unoccupied leaf ties with mixed results. Our findings are similar to those of Lill and Marquis (2004), who found that experimental ties initially inhabited by a leaf-tier, Psilocorsis sp., contained a higher density of arthropods and supported a different trophic group structure than initially unoccupied ties after two weeks. This study confirms that in the initial stages of community assembly the presence of a leaf-tier has significant effects on the arthropod community. However, studies that have examined occupied and unoccupied ties months after construction have mixed results. Lill and Marquis (2003) found no difference in the species richness and community composition of arthropods in experimental ties initially inhabited by P. quercinigracella compared to initially unoccupied ties after one, two, and three months. A more recent study found the same results when comparing initially
83
empty ties to those built by P. quercinigracella after one and two months. However, communities in ties built by P. cryptolechiella differed compositionally from communities in empty or P. quercinigracella ties (Wang et al. 2012). In these last two studies, sampling was conducted beyond the larval lifespan of the initial colonist so direct engineer-colonizer interactions were unlikely to be detected. This time lag allowed initially unoccupied ties to be colonized by other shelter-building caterpillars, potentially obscuring effects of the initial occupancy status.
There was a very strong effect of the time ties were available for colonization on all community metrics, indicating that the arthropod community colonizes new leaf ties quickly and changes rapidly in the initial stages of community assembly. Eighty percent of experimental ties were colonized within a week and even more were colonized after two weeks with an increase in arthropod abundance and species richness. This reiterates that leaf ties are a highly valued resource for the arthropod community. Additionally, the compositional change between the communities present one and two weeks after a tie is built indicates that there are not just more arthropods colonizing leaf ties, but different arthropods. Compositional turnover in the leaf tie community has been demonstrated seasonally (Sigmon and Lill 2013), but this is the first study to show a weekly change in the community. Adult arthropod oviposition largely contributes to this weekly change.
Adults were more common in the first week and their offspring were more common in the second week of colonization. Within group differences were also larger among communities in ties collected after one week than after two, shown by the larger spread of points in the NMDS for one-week samples. After two weeks communities within ties
84
became more similar to each other, which may indicate the community moving towards a stable state.
Only 22% of the initial occupants remained in the ties for the entire two weeks, so the compositional differences between weeks and between engineer presence/absence are likely interwoven. As time progresses and leaf-tiers move out of shelters, the arthropod community changes. Leaf ties that the engineer abandoned contained more scavengers and more arthropods that hatched in the tie. Similar effects of an engineer leaving have been shown in other systems. For example, in a beaver meadow, species richness is significantly influenced by the time since the beaver abandoned the site. (Wright et al.
2003) Also, more ant species colonize abandoned arboreal termite nests than occupied nests (Santos et al. 2010). It is possible that whatever factor causes the initial caterpillar to abandon the leaf tie may also be responsible for the change in community structure between occupied and abandoned leaf ties. The caterpillars leave their ties to pupate, but they also move between ties even when food resources have not been depleted and are more likely to do so when other caterpillars are present (E.S. unpublished data). In this study, some of the initial caterpillar occupants were found in other treatment ties or in ties they built elsewhere on the tree. A leaf tie may become unsuitable, or the presence of certain arthropods may prompt emigration.
Adult arthropods may make oviposition decisions based on caterpillar presence.
Moths of P. quercicella preferentially oviposited in ties that initially contained a caterpillar, particularly a conspecific, but the caterpillar had abandoned the tie. These moths usually oviposit on the leaf margin, but we found some eggs within the tie near the shelter and feeding damage made by the leaf-tier showing that moths also enter the tie to
85
oviposit. These moths may be assessing the resident leaf-tier’s presence and size. Late instar caterpillars will continue to maintain the shelter, which could protect the eggs, but are likely to leave the tie to pupate in the week that eggs take to hatch; thus the hatchlings will not be directly competing for food with a large caterpillar. Moths preferentially ovipositing in leaf ties containing a caterpillar will cause the ties to be inhabited by caterpillars for longer than the lifespan of the original builder of the shelter. This pattern of consecutive caterpillar occupancy would prolong the effect of engineer presence seen in this study. This may explain why Wang and colleagues (2012) found differences between initially occupied and unoccupied ties up to 2 months after the tie was built.
In addition to oviposited arthropods, empty ties also had a higher abundance of some scavengers. Scavengers are a common inhabitant of leaf ties (Sigmon and Lill
2013) and we assume that they consume silk, exuvia, and mold that commonly grows on the frass and silk in shelters. Mold on the silk of the shelter would be more accessible when the caterpillar is absent, which may lead to an increase in scavengers in abandoned ties. Most of the scavengers are extremely small, such as psocids and collembola, so may be drawn to the protected space within a leaf tie regardless of the presence of food resources. There was no difference in total scavenger abundance between the types of tie treatments, but many scavengers contributed to the compositional differences between treatments. Psocids were most abundant in Control 1 ties (without cues), and rove beetles were most abundant in ties initially containing caterpillars. Trends such as these in opposite directions may have led to the total abundance of scavengers being similar across treatments while individual species contribute to compositional differences.
86
The leaf-tying caterpillars used in this study differ in their behavioral aggression, maximum size, and shelter size and shape (see Sigmon in prep). Despite differences between the species, we did not find evidence for priority effects. This indicates that the facilitative effects of the shelter on the arthropod community outweigh any competitive or behavioral interactions that takes place within the leaf ties. It is possible that expanding this study to include caterpillars that build ties in different ways, such as web- makers or leaf-folders, or caterpillars with longer larval life spans may find evidence for priority effects. However, the caterpillars used in this study are the most common in this community and appear to be interchangeable. Despite overall similarities in the communities associated with different engineers, there was some evidence for behavioral differences shaping the community. The most aggressive caterpillar, P. cryptolechiella, was the most likely to remain in its shelter throughout the study and the least aggressive caterpillar, P. quercinigracella, was the most likely to abandon its shelter. Additionally, in ties that were colonized by other leaf-tiers, P. cryptolechiella was more likely to be present at the time of collection and P. quercicella and P. quercinigracella were more likely to have abandoned the shelter. This is possibly due to P. cryptolechiella defending its shelter from intruding caterpillars, leading to the caterpillars building a separate shelter within the tie. Psilocorsis quercicella and P. quercinigracella were likely forced out of the leaf tie by other caterpillars that usurped their shelter.
Arthropod abundance and species richness was lowest in the control ties with occupancy-related cues (Control 2) relative to any other treatment. This could be due to either colonists not responding to volatile cues or the volatiles dissipating long before we collected the ties. If colonists did use volatile cues to find leaf ties, then the presence of a
87
caterpillar strongly influenced their decision to remain in the leaf tie. However, it is likely that any feeding-related volatiles were no longer present when we collected ties one and two weeks after the cues were created and thus we were unable to detect any arthropods attracted to the volatiles. The communities in the two types of controls were quite similar suggesting that the volatiles were most likely not effective on this time scale. The initially empty Control 1 ties had higher arthropod abundance than Control 2 ties, but this was mainly due to an increase of psocids, particularly those that colonized via oviposition. On average, there were .5 more psocids per tie in Control 1 compared to the other treatments, which accounts for most of the increase in abundance in Control 1 ties.
In conclusion, leaf-tying caterpillar presence has a facilitative effect on the arthropod community beyond that solely of the habitat they create. Leaf ties containing, or at least initially containing, a caterpillar contained more species, more arthropods, and a different community than attached, overlapping leaves. However, the identity of the leaf-tier had no effect on the community as a whole, despite differences in aggressive behavior, indicating that these ecosystem engineers are interchangeable.
Acknowledgements
Thanks to Michelle Sliwinski, Mariana Abarca, Katherine Costantini, Arjun Aswathi, and
Luke Fey for assistance in data collection. Thanks to the DC Plant-Insect Group, Gina
Wimp, and Martha Weiss for comments on earlier drafts. Thanks to Ligia Benavides
Silva for assistance with spider specimen identification. This project was funded by NSF
Doctoral Dissertation Improvement Grant (DEB-1210600).
88
References
Abarca, M. and K. Boege. 2011. Fitness costs and benefits of shelter building and leaf
trenching behaviour in a pyralid caterpillar. Ecological Entomology 36:564-573.
Belyea, L. R. and J. Lancaster. 1999. Assembly rules within a contingent ecology. Oikos
86:402-416.
Berenbaum, M. R., E. S. Green, and A. R. Zangerl. 1993. Web costs and web defense in
the parsnip webworm (Lepidoptera: Oecophoridae). Environmental Entomology
22:791-795.
Blaustein, L. and J. Margalit. 1996. Priority effects in temporary pools: Nature and
outcome of mosquito larva toad tadpole interactions depend on order of entrance.
Journal of Animal Ecology 65:77-84.
Boots, B., A. M. Keith, R. Niechoj, J. Breen, O. Schmidt, and N. Clipson. 2012. Unique
soil microbial assemblages associated with grassland ant species with different
nesting and foraging strategies. Pedobiologia 55:33-40.
Buhl-Mortensen, L., A. Vanreusel, A. J. Gooday, L. A. Levin, I. G. Priede, P. Buhl-
Mortensen, H. Gheerardyn, N. J. King, and M. Raes. 2010. Biological structures
as a source of habitat heterogeneity and biodiversity on the deep ocean margins.
Marine Ecology-an Evolutionary Perspective 31:21-50.
Carroll, M. R. and W. H. Kearby. 1978. Microlepidopterous oak leaftiers
(Lepidoptera:Gelechioidea) in central Missouri. Journal of Kansas Entomological
Society 51:457-471.
Collinge, S. K. and C. Ray. 2009. Transient patterns in the assembly of vernal pool plant
communities. Ecology 90:3313-3323.
89
Connell, J. H. and R. O. Slatyer. 1977. Mechanisms of succession in natural communities
and their role in community stability and organization. American Naturalist
111:1119-1144.
Damman, H. 1987. Leaf quality and enemy avoidance by the larvae of a pyralid moth.
Ecology 68:88-97. de Mendiburu, F. 2012. agricolae: Statistical procedures for agricultural research.
Drake, J. A. 1991. Community-assembly mechanics and the structure of an experimental
species ensemble. American Naturalist 137:1-26.
Eldridge, D. J. and R. S. B. Greene. 1994. Microbiotic soil crusts - A review of their roles
in soil and ecological processes in the rangelands of Australia. Australian Journal
of Soil Research 32:389-415.
Fincke, O. M. 1999. Organization of predator assemblages in Neotropical tree holes:
effects of abiotic factors and priority. Ecological Entomology 24:13-23.
Fletcher, L. E., J. E. Yack, T. D. Fitzgerald, and R. R. Hoy. 2006. Vibrational
communication in the cherry leaf roller caterpillar Caloptilia serotinella
(Gracillarioidea : Gracillariidae). Journal of Insect Behavior 19:1-18.
Fox, J. 2003. Effect displays in R for generalised linear models. Journal of Statistical
Software 8:1-27.
Fukui, A. 2001. Indirect interactions mediated by leaf shelters in animal-plant
communities. Population Ecology 43:31-40.
Fukui, A., M. Murakami, K. Konno, M. Nakamura, and T. Ohgushi. 2002. A leaf-rolling
caterpillar improves leaf quality. Entomological Science 5:263-266.
90
Guedes, R. N. C., S. M. Matheson, B. Frei, M. L. Smith, and J. E. Yack. 2012. Vibration
detection and discrimination in the masked birch caterpillar (Drepana arcuata).
Journal of Comparative Physiology a-Neuroethology Sensory Neural and
Behavioral Physiology 198:325-335.
Hensen, W. R. 1958. The effects of radiation on the habitat temperatures of some poplar-
inhabiting insects. Canadian Journal of Zoology 36:463-478.
Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers.
Oikos 69:373-386.
Jones, C. G., J. H. Lawton, and M. Shachak. 1997. Positive and negative effects of
organisms as physical ecosystem engineers. Ecology 78:1946-1957.
Keddy, P. A. 1992. Assembly and response rules - 2 goals for predictive community
ecology. Journal of Vegetation Science 3:157-164.
Kennedy, P. G., K. G. Peay, and T. D. Bruns. 2009. Root tip competition among
ectomycorrhizal fungi: Are priority effects a rule or an exception? Ecology
90:2098-2107.
Kruskal, J. B. 1964. Nonmetric multidimensional scaling: a numerical method.
Psychometrika 29:115-129.
Lawler, S. P. and P. J. Morin. 1993. Temporal overlap, competition, and priority effects
in larval anurans. Ecology 74:174-182.
Lill, J. T. and R. J. Marquis. 2003. Ecosystem engineering by caterpillars increases insect
herbivore diversity on white oak. Ecology 84:682-690.
Lill, J. T. and R. J. Marquis. 2004. Leaf ties as colonization sites for forest arthropods: an
experimental study. Ecological Entomology 29:300-308.
91
Lill, J. T. and R. J. Marquis. 2007. Microhabitat manipulation: ecosystem engineering by
shelter-building insects. Pages 107-138 in K. Cuddington, J. E. Byers, W. G.
Wilson, and A. Hastings, editor. Ecosystem engineers: plants to protists.
Academic Press, New York.
Lill, J. T., R. J. Marquis, M. A. Walker, and L. Peterson. 2007. Ecological consequences
of shelter sharing by leaf-tying caterpillars. Entomologia Experimentalis Et
Applicata 124:45-53.
Loeffler, C. C. 1996. Adaptive trade-offs of leaf folding in Dichomeris caterpillars on
goldenrods. Ecological Entomology 21:34-40.
Longino, J. T. and R. K. Colwell. 1997. Biodiversity assessment using structured
inventory: Capturing the ant fauna of a tropical rain forest. Ecological
Applications 7:1263-1277.
Machicote, M., L. C. Branch, and D. Villarreal. 2004. Burrowing owls and burrowing
mammals: are ecosystem engineers interchangeable as facilitators? Oikos
106:527-535.
Martinsen, G. D., K. D. Floate, A. M. Waltz, G. M. Wimp, and T. G. Whitham. 2000.
Positive interactions between leafrollers and other arthropods enhance
biodiversity on hybrid cottonwoods. Oecologia 123:82-89.
Mattiacci, L. and M. Dicke. 1995. The parasitoid Cotesia glomerata (Hymenoptera,
Braconidae) discriminates between frist and 5th larval instars of its host Pieris
brassicae, on the basis of contact cues from frass, silk, and herbivore-damaged
leaf tissue. Journal of Insect Behavior 8:485-498.
92
Naiman, R. J., J. M. Melillo, and J. E. Hobbie. 1986. Ecosystem alteration of boreal
forest streams by beaver (Castor canadensis). Ecology 67:1254-1269.
Nakamura, M. and T. Ohgushi. 2003. Positive and negative effects of leaf shelters on
herbivorous insects: linking multiple herbivore species on a willow. Oecologia
136:445-449.
Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O'Hara, G. L.
Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner. 2012. vegan:
Community Ecology Package.
Oliver, I. and A. J. Beattie. 1996. Invertebrate morphospecies as surrogates for species: A
case study. Conservation Biology 10:99-109.
Pettersson, J., V. Ninkovic, R. Glinwood, M. A. Birkett, and J. A. Pickett. 2005. Foraging
in a complex environment - semiochemicals support searching behaviour of the
seven spot ladybird. European Journal of Entomology 102:365-370.
R Development Core Team. 2012. R: A language and environment for statistical
computing R Foundation for Statistical Computing, Vienna, Austria
Ransom, T. S. 2012. Comparison of direct, indirect, and ecosystem engineering effects of
an earthworm on the red-backed salamander. Ecology 93:2198-2207.
Robinson, J. V. and J. E. Dickerson. 1987. Does invasion sequence affect community
structure. Ecology 68:587-595.
Ruesink, J. L., H. S. Lenihan, A. C. Trimble, K. W. Heiman, F. Micheli, J. E. Byers, and
M. C. Kay. 2005. Introduction of non-native oysters: Ecosystem effects and
restoration implications. Pages 643-689 Annual Review of Ecology Evolution
and Systematics. Annual Reviews, Palo Alto.
93
Sagers, C. L. 1992. Manipulation of host plant-quality - Herbivores keep leaves in the
dark. Functional Ecology 6:741-743.
Santos, P. P., A. Vasconcellos, B. Jahyny, and J. H. C. Delabie. 2010. Ant fauna
(Hymenoptera, Formicidae) associated to arboreal nests of Nasutitermes spp.
(Isoptera, Termitidae) in a cacao plantation in southeastern Bahia, Brazil. Revista
Brasileira De Entomologia 54:450-454.
Scott, J. L., A. Y. Kawahara, J. H. Skevington, S. H. Yen, A. Sami, M. L. Smith, and J.
E. Yack. 2010. The evolutionary origins of ritualized acoustic signals in
caterpillars. Nature Communications 1:9.
Scott, J. L. and J. E. Yack. 2012. Vibratory territorial signals in caterpillars of the poplar
lutestring, Tethea or (Lepidoptera: Drepanidae). European Journal of Entomology
109:411-417.
Shulman, M. J., J. C. Ogden, J. P. Ebersole, W. N. McFarland, S. L. Miller, and N. G.
Wolf. 1983. Priority effects in the recruitment of juvenile coral-reef fishes.
Ecology 64:1508-1513.
Sigmon, E. in prep. Interspecific variation in aggressive fighting behavior of shelter-
building caterpillars. Behavioral Ecology. *Chapter 2
Sigmon, E. and J. T. Lill. 2013. Phenological variation in the composition of a temperate
forest leaf tie community. Environmental Entomology 42:29-37.
Sipura, M. 1999. Tritrophic interactions: willows, herbivorous insects and insectivorous
birds. Oecologia 121:537-545.
94
Wang, H. G., R. J. Marquis, and C. S. Baer. 2012. Both host plant and ecosystem
engineer identity influence leaf-tie impacts on the arthropod community of
Quercus. Ecology 93:2186-2197.
Weslien, J., L. B. Djupstrom, M. Schroeder, and O. Widenfalk. 2011. Long-term priority
effects among insects and fungi colonizing decaying wood. Journal of Animal
Ecology 80:1155-1162.
Wilbur, H. M. and R. A. Alford. 1985. Priority effects in experimental pond communities
- Responses of Hyla to Bufo and Rana. Ecology 66:1106-1114.
Wilmer, P. G. 1980. The effects of fluctuating environment on the water relations of
larval Lepidoptera. Ecological Entomology 5:271-292.
Wright, J. P., A. S. Flecker, and C. G. Jones. 2003. Local vs. landscape controls on plant
species richness in beaver meadows. Ecology 84:3162-3173.
Yack, J. E., M. L. Smith, and P. J. Weatherhead. 2001. Caterpillar talk: Acoustically
mediated territoriality in larval Lepidoptera. Proceedings of the National
Academy of Sciences of the United States of America 98:11371-11375.
95
Table 3.1. Description of experimental ties used in the study. One of each tie type was placed on a single branch of white oak (2 branches/tree). Pairs of leaves were fastened together using a hair clip at the petiole using spring-loaded hair curler clips. Cues from a caterpillar include frass and silk used to build a shelter between the clipped leaves.
Control Control Treatment 1 Treatment 2 Treatment 3 1 2 Leaves clipped Cues from caterpillar Leaf tying Psilocorsis Psilocorsis Pseudotelphusa caterpillar cryptolechiella quercicella quercinigracella occupant
96
Figure Legends
Figure 3.1. A) Species richness (mean ± 95% confidence interval) and B) abundance
(mean ± 95% confidence interval) per tie of arthropods colonizing experimental leaf ties given one or two weeks to colonize the ties (white bars). Treatments (gray bars) included control tie without occupant or cues (Control 1), control without an occupant with cues
(Control 2), and ties initially containing a caterpillar of either Psilocorsis cryptolechiella,
P. quercicella, or Pseudotelphusa quercinigracella. Letters above treatment bars indicate significant differences according to Tukey’s HSD test.
Figure 3.2. Results of non-metric multidimensional scaling ordination (Stress = 0.2665) displaying community composition in each treatment tie. Treatments were ties initially containing a caterpillar of either Psilocorsis cryptolechiella (Pc), P. quercicella (Pq), or
Pseudotelphusa quercinigracella (Ps), control tie without occupant or cues (C1), and control without an occupant with cues (C2). Dispersion ellipses represent the standard error of average scores with the label placed at the centroid. Arthropod communities differed significantly between treatments (F4, 580 = 2.03, p <0.001).
Figure 3.3. Results of non-metric multidimensional scaling ordination (Stress = 0.2775) displaying community composition in ties collected after 1 or 2 weeks. Dispersion ellipses represent the standard error of average scores with the label placed at the centroid. Arthropod communities differed significantly between weeks (F1, 580 = 23.32, p
< 0.001).
97
Figure 3.4. Percent of ties with original leaf-tier present at the time of collection. Ties in treatments (ties initially containing a caterpillar of either Psilocorsis cryptolechiella, P. quercicella, or Pseudotelphusa quercinigracella; grey bars) were combined across weeks and ties in weeks (white bars) were combined across treatments.
Figure 3.5. Results of non-metric multidimensional scaling ordination (Stress = 0.2967) of arthropods colonizing leaf ties initially containing a caterpillar when the caterpillar was either present or absent at the time the tie was collected. Dispersion ellipses represent the standard error of average scores with the label placed at the centroid. Arthropod community composition significantly differed between engineer presence/absence (F1, 308
= 6.43, p < 0.001).
Figure 3.6. Mean abundance per tie of scavengers, herbivores, and predators collected after one or two weeks from experimental leaf ties, which included control tie without occupant or cues (Control 1), and control without an occupant with cues (Control 2), and ties initially containing a caterpillar of either Psilocorsis cryptolechiella, P. quercicella, or Pseudotelphusa quercinigracella. Herbivore (χ2 = 34.73, df = 1, p < 0.0001), predator
(χ2 = 9.19, df = 1, p = 0.0024), and scavenger abundance (χ2 = 26.19, df = 1, p < 0.0001) differed between weeks. Abundance of herbivores (χ2 = 17.99, df = 4, p = 0.0012) and predators (χ2 = 35.67, df = 4, p < 0.0001) differed between treatments. Letters within treatment bars indicate significant differences according to Tukey’s HSD test.
98
Figure 3.1
99
Figure 3.2
100
Figure 3.3
101
Figure 3.4
102
Figure 3.5
103
Figure 3.6
104
GENERAL CONCLUSIONS
Leaf ties have a strong facilitative effect on the arthropod community despite the fact that competitive interactions within the ties are frequent and there is rapid turnover in the arthropod community. Interactions with the leaf-tying caterpillar figure more prominently into community dynamics than previously suspected. The research presented here has implications for the fields of community ecology, animal behavior, and insect ecology. It adds to existing knowledge on the theories of ecosystem engineers, community phenology, host plant use, direct competition, territoriality, and community assembly.
Like most arthropod communities, the leaf tie community exhibits predictable phenological patterns throughout the summer season. These patterns are consistent across host plants, despite differences in the communities, demonstrating the generality of the patterns. At mid-latitudes in North America, many leaf-tying caterpillars are bivoltine, which leads to peaks in abundance in early and late summer (e.g., June and late August) that have been consistent across years. Seasonal changes in the leaf tie community reflect changes seen in the forest arthropod community, with different species present at different times of the summer, reflecting their realized temporal niches. Leaf ties are predominately occupied by juveniles so development times are a likely driver of this pattern, as has been shown in other arthropod communities (Niemela and Haukioja 1982).
Documentation of this seasonal pattern adds to the existing literature on community phenological trends. In the face of global climate change it is important to document such patterns so that we can recognize when and how they change.
105
Prior to this research the interactions within leaf ties were a complete mystery. I show that the common leaf-tying caterpillars on white oak frequently compete over existing shelters and vary in their aggression in these interactions. Such behavior has been shown in other shelter-building caterpillars (Berenbaum et al. 1993, Yack et al.
2001, Fletcher et al. 2006, Scott et al. 2010, Guedes et al. 2012, Scott and Yack 2012), suggesting that competition over shelters is common and highlighting the value of these resources to the arboreal arthropod community. The oak leaf tiers compete mainly via physically aggressive behaviors, but also appear to use vibratory signaling suggesting that they may represent an early evolutionary step toward the sole use of vibratory signals
(Scott et al. 2010). Herbivorous insects generally compete only indirectly over plant resources via induced plant defenses or shared predators (Kaplan and Denno 2007).
However when the food resource is also a territory, as is the case for leaf shelters, direct competition is more common (also see: Rathcke 1976, Stiling and Strong 1984, Tack et al. 2009). Competition theory generally only considers one resource at a time, but the results of my research highlight the value of explicitly considering the multiple possible benefits of such resources (i.e., recognizing that a leaf shelter may simultaneously provide food, territory, and protection to occupants).
Despite the plethora of studies on the benefits of residing in a shelter and the effects of shelters on the arthropod community, how and why shelters are used by such a wide variety of arthropods is still unclear. We show that the presence, but not species identity, of a leaf-tier in a tie affects the assembling community more so than the tie alone. Engineer presence has also been shown to be an important predictor of diversity in communities associated with beavers and arboreal termites (Wright et al. 2003, Santos et
106
al. 2010), suggesting that abrupt community changes following engineer abandonment may be a common effect of ecosystem engineers. We speculate that colonizing arthropods (via oviposition or migration) may prefer inhabited shelters because the caterpillar maintains the tie, which provides reliable protection from predators and harsh climatic conditions; further testing of this hypothesis is necessary. There is also the possibility that unknown behavioral interactions could be causing this pattern. We had suspected that aggressive caterpillars would prevent some arthropods from inhabiting the tie, but our understanding of how leaf-tying caterpillars and other arthropods interact is incomplete. We also noticed a strong correlation between herbivore and predator abundance within leaf ties. This seemingly contradictory pattern exemplifies the complexity of leaf tie community assembly. By closely examining the leaf tie community, which represents a diverse yet manageable set of players, we may continue to gain new insights into large-scale patterns of structuring ecological communities.
Future Directions
This research highlights the intricacy of interactions that take place within leaf ties. Future studies similar to those in Chapter 2 should be performed using siblings, different age caterpillars, and small or damaged leaves to explore how ecological context influences expression of aggressive behaviors. Additionally, interactions between other common leaf tie occupants (e.g., non-caterpillars) should be explored. We have generally assumed that leaf-tying caterpillars would defend their shelter from all arthropods, but this has not been tested with behavioral studies. It would also be very interesting to see
107
how herbivores and predators interact within leaf ties. It is perplexing that both herbivores and predators are abundant in leaf ties and frequently co-occur. Observations of their interactions may reveal that they ignore each other or that only a few predators consume other arthropods within ties. Determining the trophic links in the leaf tie food webs would perhaps help resolve some of this uncertainty and could be aided by using stable isotopes (as in: Wimp et al. 2013).
The transparency paper artificial tie technique could also be adapted for use in the field. Transparency paper could be clipped to the underside of leaves allowing arthropods to colonize ties naturally. We have observed caterpillars, crickets and spiders build ties using plastic flagging in the field, so they are likely to accept transparency paper as a suitable substrate. Artificial ties could be observed frequently without having to disrupt the occupants so that rapid changes in the community and interactions within the ties may be seen. This could uncover the interactions between herbivores and predators within ties and the role that scavengers play.
References
Berenbaum, M. R., E. S. Green, and A. R. Zangerl. 1993. Web costs and web defense in
the parsnip webworm (Lepidoptera: Oecophoridae). Environmental Entomology
22:791-795.
Fletcher, L. E., J. E. Yack, T. D. Fitzgerald, and R. R. Hoy. 2006. Vibrational
communication in the cherry leaf roller caterpillar Caloptilia serotinella
(Gracillarioidea : Gracillariidae). Journal of Insect Behavior 19:1-18.
108
Guedes, R. N. C., S. M. Matheson, B. Frei, M. L. Smith, and J. E. Yack. 2012. Vibration
detection and discrimination in the masked birch caterpillar (Drepana arcuata).
Journal of Comparative Physiology a-Neuroethology Sensory Neural and
Behavioral Physiology 198:325-335.
Kaplan, I. and R. F. Denno. 2007. Interspecific interactions in phytophagous insects
revisited: a quantitative assessment of competition theory. Ecology Letters
10:977-994.
Niemela, P. and E. Haukioja. 1982. Seasonal patterns in species richness of herbivores -
Macrolepidopteran larva on Finnish deciduous trees. Ecological Entomology
7:169-175.
Rathcke, B. J. 1976. Competition and coexistence within a guild of herbivorous insects.
Ecology 57:76-87.
Santos, P. P., A. Vasconcellos, B. Jahyny, and J. H. C. Delabie. 2010. Ant fauna
(Hymenoptera, Formicidae) associated to arboreal nests of Nasutitermes spp.
(Isoptera, Termitidae) in a cacao plantation in southeastern Bahia, Brazil. Revista
Brasileira De Entomologia 54:450-454.
Scott, J. L., A. Y. Kawahara, J. H. Skevington, S. H. Yen, A. Sami, M. L. Smith, and J.
E. Yack. 2010. The evolutionary origins of ritualized acoustic signals in
caterpillars. Nature Communications 1:9.
109
Scott, J. L. and J. E. Yack. 2012. Vibratory territorial signals in caterpillars of the poplar
lutestring, Tethea or (Lepidoptera: Drepanidae). European Journal of Entomology
109:411-417.
Stiling, P. D. and D. R. Strong. 1984. Experimental density manipulation of stem-boring
insects - some evidence for interspecific competition. Ecology 65:1683-1685.
Tack, A. J. M., O. Ovaskainen, P. J. Harrison, and T. Roslin. 2009. Competition as a
structuring force in leaf miner communities. Oikos 118:809-818.
Wimp, G. M., S. M. Murphy, D. Lewis, M. R. Douglas, R. Ambikapathi, L. Van-Tull, C.
Gratton, and R. F. Denno. 2013. Predator hunting mode influences patterns of
prey use from grazing and epigeic food webs. Oecologia 171:505-515.
Wright, J. P., A. S. Flecker, and C. G. Jones. 2003. Local vs. landscape controls on plant
species richness in beaver meadows. Ecology 84:3162-3173.
Yack, J. E., M. L. Smith, and P. J. Weatherhead. 2001. Caterpillar talk: Acoustically
mediated territoriality in larval Lepidoptera. Proceedings of the National
Academy of Sciences of the United States of America 98:11371-11375.
110
APPENDICES
Appendix 1.1. List of all arthropods collected within experimental leaf ties, which are leaves clipped together with a metal clip, on American beech, Fagus grandifolia, and white oak, Quercus alba, during the summer of 2009.
American White Order Family Species Beech Oak
Lepidoptera Gelechiidae Arogalea cristifasciella* X
Lepidoptera Gelechiidae Chionodes fuscomaculella* X X
Lepidoptera Gelechiidae Pseudotelphusa quercinigracella* X X
Lepidoptera Gelechiidae Unknown sp 1 X
Lepidoptera Geometridae Unknown sp 2-3 X
Lepidoptera Lasiocampidae Macrocampa marthesia X
Lepidoptera Limicodidae Lithacodes fasciola X
Lepidoptera Limicodidae Prolimacodes badia X
Lepidoptera Lymantriidae Dasychira obliquata X
Lepidoptera Noctuidae Acronicta sp. X X
Lepidoptera Noctuidae Morrisonia confusa* X X
Lepidoptera Noctuidae Zanclognatha sp. X
Lepidoptera Nolidae Meganola phylla X
Lepidoptera Notodontidae Dasychira obliquata X
Lepidoptera Notodontidae Lochmaeus manteo X
Lepidoptera Notodontidae Macrocampa marthesia X
Lepidoptera Oecophoridae Machimia tentoriferella X
Lepidoptera Oecophoridae Antaeotricha osscella* X X
Lepidoptera Oecophoridae Antaeotricha schlaegeri* X
Lepidoptera Oecophoridae Psilocorsis cryptolechiella* X X
Lepidoptera Oecophoridae Psilocorsis quercicella* X
111
American White Order Family Species Beech Oak
Lepidoptera Oecophoridae Psilocorsis reflexella* X X
Lepidoptera Oecophoridae Setiostoma xanthobasis* X
Lepidoptera Pyralidae Pococera expandens* X
Lepidoptera Tortricidae Acleris subnivana/veriana* X X
Lepidoptera Tortricidae Amorbia humerosana* X X
Lepidoptera Tortricidae Pandemis limitata* X X
Lepidoptera Tortricidae Unknown sp 4* X
Lepidoptera Unknown sp 5-6 X
Lepidoptera Unknown sp 7 X
Araneae Clubionidae Elaver sp. X X
Araneae Linyphiidae Ostearius melanopygius X X
Araneae Unknown sp 8-11 X X
Araneae Unknown sp 12-18 X
Araneae Unknown sp 19-21 X
Centipoda Unknown sp 22 X X
Coleoptera Alleculidae Unknown sp 23 X X
Coleoptera Cantharidae Rhagonycha imbecillis X
Coleoptera Chrysomelidae Odontota dorsalis X X
Coleoptera Chrysomelidae Unknown sp 24 X
Coleoptera Curculionidae Cyrtepistomus castaneus X X
Coleoptera Curculionidae Unknown sp 25 X X
Coleoptera Curculionidae Unknown sp 26 X
Coleoptera Staphylinidae Unknown sp 27 X X
Coleoptera Unknown sp 28-37 X
Coleoptera Unknown sp 38-45 X
Collembola Unknown sp 46-47 X
Collembola Unknown sp 48 X X
Diptera Unknown sp 49-50 X
112
American White Order Family Species Beech Oak
Diptera Unknown sp 51 X
Embiidina Unknown sp 52 X
Hemiptera Aphididae Unknown sp 53-54 X X
Hemiptera Cicadellidae Unknown sp 55 X X
Hemiptera Eriosomatidae Unknown sp 56 X X
Hemiptera Pentatomidae Apateticus cynicus X
Hemiptera Reduviidae Zelus luridus X X
Hemiptera Tingidae Corythuca sp. X X
Hemiptera Unknown sp 57 X X
Hemiptera Unknown sp 58 X
Hemiptera Unknown sp 59 X
Hymenoptera Braconidae Apanteles epinotiace** X
Hymenoptera Eulophidae Unknown sp 60** X
Hymenoptera Formicidae Unknown sp 61 X
Hymenoptera Ichneumonidae Diadegma psilocorse** X
Hymenoptera Tenthredinidae Caliroa sp. X
Hymenoptera Unknown sp 62 X X
Hymenoptera Unknown sp 63-66 X
Hymenoptera Unknown sp 67-71 X
Millipoda Unknown sp 72 X
Orthoptera Unknown sp 73 X
Psocoptera Unknown sp 74-75 X
Psocoptera Unknown sp 76 X
Thysanoptera Unknown sp 77 X X
*Species that build leaf ties
**Parasitoids that emerged from reared caterpillars
113
Appendix 2.1 Curves used to calculate mass from length of each caterpillar used in behavior trials. Individual caterpillars from laboratory colony and field collection, not necessarily those used in behavior trials, were measured to the nearest 0.1 mm and weighed to the nearest 0.1 mg. Psilocorsis cryptolechiella and P. reflexella are indistinguishable from each other until the last instar when P. reflexella reaches a larger size, so data for those species are combined in a single curve.
Figure A1. Length and mass for Psilocorsis cryptolechiella (filled circles) and P. reflexella (open circles) caterpillars. Fitted curve is y = 0.0083x3.0070 (R2 = 0.9511).
114
Figure A2. Length and mass for Psilocorsis quercicella caterpillars. Fitted curve is y =
0.0068x3.2266 (R2 = 0.9412).
115
Figure A3. Length and mass of Pseudotelphusa quercinigracella caterpillars. Fitted curve is y = 0.0071x2.9476 (R2 = 0.8172).
116
Appendix 3.1 Species and morphospecies found inhabiting experimental leaf ties on white oak. ‘Oviposition’ indicates arthropods found hatching from leaf ties.
Order Family Species Oviposition Araneae Anyphaenidae Unknown sp #1 Araneae Clubionidae Clubiona Araneae Clubionidae Elaver Araneae Linyphiidae Ostearius melanopygius Araneae Salticidae Unknown sp #2 Araneae Salticidae Unknown sp #3 Araneae Salticidae Unknown sp #4 Araneae Salticidae Unknown sp #5 Araneae Salticidae Unknown sp #6 Araneae Salticidae Unknown sp #7 Araneae Salticidae Unknown sp #8 Araneae Salticidae Unknown sp #9 Araneae Salticidae Unknown sp #10 Araneae Salticidae Unknown sp #11 Araneae Salticidae Unknown sp #12 Araneae Salticidae Unknown sp #13 Araneae Theridiidae Unknown sp #14 Araneae Theridiidae Unknown sp #15 Araneae Thomisidae Misumenini Araneae Thomisidae Unknown sp #16 Araneae Thomisidae Unknown sp #17 Araneae Thomisidae Unknown sp #18 Araneae Thomisidae Unknown sp #19 Araneae Thomisidae Unknown sp #20 Araneae Unknown sp #21 Araneae Unknown sp #22 Araneae Unknown sp #23 Araneae Unknown sp #24 Coleoptera Alleculidae Unknown sp #25 Coleoptera Cantharidae Rhagonycha imbecillis Coleoptera Chrysomelidae Acalymma vittata Coleoptera Chrysomelidae Odontota dorsalis Chrysomelidae: Coleoptera Alticinae Unknown sp #26 Chrysomelidae: Coleoptera Eumolpinae Metachroma Chrysomelidae: Coleoptera Eumolpinae Unknown sp #27 Chrysomelidae:Chryso Coleoptera melinae Unknown sp #28 Coleoptera Coccinellidae Harmonia axyridis
117
Coleoptera Coccinellidae Unknown sp #29 Coleoptera Curculionidae Cyrtepistomus castaneus Curculionidae: Coleoptera Otiorhynchinae Unknown sp #30 Coleoptera Elateridae Unknown sp #31 Coleoptera Elateridae Unknown sp #32 Coleoptera Eucnemidae Unknown sp #33 Coleoptera Latridiidae Unknown sp #34 Coleoptera Mordellidae Unknown sp #35 Coleoptera Mordellidae Unknown sp #36 Coleoptera Phalacridae Unknown sp #37 Coleoptera Pselaphidae Unknown sp #38 Coleoptera Pselaphidae Unknown sp #39 Staphylinidae: Coleoptera Euaesthetinae Stictocranius Coleoptera Unknown sp #40 X Coleoptera Unknown sp #41 Coleoptera Unknown sp #42 Coleoptera Unknown sp #43 Coleoptera Unknown sp #44 Coleoptera Unknown sp #45 Collembola Unknown sp #46 Collembola Unknown sp #47 Collembolla Hypogastruridae Unknown sp #48 X Diptera Cecidomyiidae Lestodiplosis Diptera Serphid Unknown sp #49 Diptera Unknown sp #50 Diptera Unknown sp #51 Diptera Unknown sp #52 Hemiptera Unknown sp #53 X Hemiptera Cercopidae Unknown sp #54 Hemiptera Cicadellidae Unknown sp #55 Hemiptera Cicadellidae Unknown sp #56 Hemiptera Cicadellidae Unknown sp #57 Hemiptera Cicadellidae Unknown sp #58 Hemiptera Drepanosiphidae Myzocallis punctata X Hemiptera Eriosomatidae Unknown sp #59 X Hemiptera Miridae Hyaliodes harti Hemiptera Miridae Unknown sp #60 Hemiptera Miridae Unknown sp #61 Hemiptera Pentatomidae Cosmopepla bimaculata Hemiptera Reduviidae Unknown sp #62 Hemiptera Tingidae Corthuca Hemiptera Unknown sp #63 Hemiptera Unknown sp #64 Hemiptera Unknown sp #65
118
Hemiptera Unknown sp #66 Hemiptera Unknown sp #67 Hemiptera Unknown sp #68 Hemiptera Unknown sp #69 Hemiptera Unknown sp #70 Hemiptera Unknown sp #71 Hemiptera Unknown sp #72 Hemiptera Unknown sp #73 Hymenoptera Eulophidae Unknown sp #74 Hymenoptera Formicidae Unknown sp #75 Hymenoptera Formicidae Unknown sp #76 Hymenoptera Tenthredinidae Caliroa Hymenoptera Unknown sp #77 Hymenoptera Unknown sp #78 Hymenoptera Unknown sp #79 Lepidoptera Gelechiidae Unknown sp #80* Lepidoptera Gelechiidae Arogalea cristifasciella* Lepidoptera Gelechiidae Chionodes fuscomaculella* Pseudotelphusa Lepidoptera Gelechiidae quercinigracella* Lepidoptera Geometridae Anacamptodes defectaria Lepidoptera Geometridae Unknown sp #81 Lepidoptera Geometridae Unknown sp #82 Lepidoptera Limacodidae Euclea delphinii Lepidoptera Lymantriidae Dasychira obliquata Lepidoptera Noctuidae Acronicta haesitata Lepidoptera Noctuidae Morrisonia confusa* Lepidoptera Notodontidae Schizura unicornis Lepidoptera Oecophoridae Antaeotricha schlaegeri* Lepidoptera Oecophoridae Machimia tentoriferella* Lepidoptera Oecophoridae Psilocorsis cryptolechiella* X Lepidoptera Oecophoridae Psilocorsis quercicella* X Lepidoptera Oecophoridae Psilocorsis reflexella* X Lepidoptera Oecophoridae Setiostoma xanthobasis* X Lepidoptera Pyralidae Pococera expandes* X Lepidoptera Tortricidae * Lepidoptera Tortricidae * X Lepidoptera Tortricidae Anclis divisiana* Lepidoptera Tortricidae Pandemis limitata* X Millipoda Unknown sp #83 Neuroptera Unknown sp #84 Neuroptera Unknown sp #85 Gryllidae: Orthoptera Oecanthinae Unknown sp #86 Psocoptera Amphipsocidae Polypsocus corruptus X Psocoptera Unknown sp #87 X
119
Psocoptera Unknown sp #88 Thysanoptera Unknown sp #89 X Thysanoptera Unknown sp #90 Thysanoptera Unknown sp #91 Unknown sp #92 Unknown sp #93 Unknown sp #94 *Leaf-tying caterpillars
120