Macroecology of Neotropical Clearwing Butterflies (Nymphalidae

MISSING IMAGE

Material Information

Title:
Macroecology of Neotropical Clearwing Butterflies (Nymphalidae Ithomiini)
Physical Description:
1 online resource (76 p.)
Language:
english
Creator:
Gallice,Geoffrey R
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Entomology and Nematology
Committee Chair:
Willmott, Keith Richard
Committee Members:
Daniels, Jaret

Subjects

Subjects / Keywords:
abundance -- density -- distribution -- ithomiini -- macroecology -- niche -- range
Entomology and Nematology -- Dissertations, Academic -- UF
Genre:
Entomology and Nematology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Macroecology is the study of large scale patterns in species abundance and distribution. While a positive relationship between abundance and distribution is widely documented, most studies to date have focused on vertebrate species in temperate regions, whereas the majority of Earth?s biodiversity is concentrated in the tropics, particularly among invertebrates. To investigate several key macroecological phenomena for the first time in a diverse Neotropical insect group, the clearwing butterflies of the nymphalid tribe Ithomiini, I combined measures of mean local density gathered in the field (five sites in eastern Ecuador) with range size data compiled using ecological niche modeling, as well as measures of body size and dispersal ability measured from museum specimens, to investigate: 1) the relationship between mean local density and body size, 2) the relationship between mean local density and distribution, and 3) the relationship between dispersal ability and distribution. There was an overall negative relationship between density and body size, although inspection of arithmetic plots of density as a function of two estimates of body size, mean forewing length and mean thorax width, revealed rather triangular relationships, with species of intermediate body size displaying the highest densities in both cases. Two measures of mean local density, density at field sites where species were found and density averaged over all visited sites, did not predict range size, and there was no relationship between density and range size within any field site. There was no relationship between mean local density and distribution within genera, and no relationship between density and distribution among closely related, sister species pairs. At a higher taxonomic level, no relationship was found between generic mean local density and generic range size. Finally, three indirect measures of dispersal ability, mean forewing length, mean forewing area, and mean wing load (ration of thorax width to forewing length) failed to predict species range size. These results collectively suggest that macroecological patterns in the tropics, at least for ithomiine butterflies, might differ from those reported for temperate species. Future directions for research, in ithomiine butterflies and other Neotropical nymphalid groups, are discussed.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Geoffrey R Gallice.
Thesis:
Thesis (M.S.)--University of Florida, 2011.
Local:
Adviser: Willmott, Keith Richard.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2011
System ID:
UFE0043383:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 MACROECOLOGY OF NEOTROPICAL CLEARWING BUTTERFLIES (NYMPHALIDAE: ITHOMIINI) By GEOFFREY R. GALLICE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE D EGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Geoffrey R. Gallice

PAGE 3

3 ACKNOWLEDGMENTS First and foremost, I thank Keith Willmott of the Florida Museum of Natural History for providing invaluable guidance and assistance from start to finish, from the development of my project to planning field work in Ecuador, from providing butterfly distributional data essential for modeling ranges to identifying specimens. Without his generous support and insightful discussions, this project would not have been possible. In Ecuador, I thank the Ministerio del Ambiente for granting me permission to conduct field work in Ecuador, and Santiago Villamarn for facilitating the acquisition of the required research/collecting and export permits. The Reser va Biolgica Jatun Sacha, Estacin Cientfica Yasun of the Pontificia Universidad Catlica del Ecuador and Tiputini Biodiversity Station of the Universidad San Francisco de Quito all provided permission to conduct research at their facilities. I am espec ially grateful to Yarina Lodge and Sacha Lodge for providing outstanding accommodations at their tourist lodges. Cristina Maribel Porras was helpful in data collection at Jatun Sacha. The McGuire Center for Lepidoptera and Biodiversity provided financial a ssistance in the form of a research assistantship, and granted me access to their substantial Ithomiini collections. The Dean of the College of Agricultural and Life Sciences at the University of Florida provided additional funding in the form of a teachin g assistantship. The National Science Foundation provided a fellowship to fund my final year of study at the University of Florida. The Tropical Conservation and Development Program at the University of Florida and the Sophie Danforth Conservation Biology Fund at the Roger Williams Park Zoo in Providence, RI, provided generous support in the form of grants for field work.

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 3 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ..................... 8 CHAPTER 1 GENERAL INTRODUCTION ................................ ................................ ...................... 10 Macroecology ................................ ................................ ................................ ......... 10 Study Group ................................ ................................ ................................ ............ 11 Objectives ................................ ................................ ................................ ............... 13 2 MODELING ITHOMIINE BUTTERFLY RANGES ................................ ....................... 14 Introduction ................................ ................................ ................................ ............. 14 Method s ................................ ................................ ................................ .................. 18 Data ................................ ................................ ................................ .................. 18 Modeling Butterfly Ranges ................................ ................................ ............... 19 Results ................................ ................................ ................................ .................... 20 Range Size Frequency Distributions ................................ ................................ 20 Ecological niche modeling ................................ ................................ .......... 20 Minimum convex polygon ................................ ................................ ........... 21 Discussion ................................ ................................ ................................ .............. 22 3 SPECIES RELATIVE DENSITY WITHIN BUTTERFLY COMMUNITIES ................... 26 Introduction ................................ ................................ ................................ ............. 26 Methods ................................ ................................ ................................ .................. 28 Study Sites and Butterfly Sampling ................................ ................................ .. 28 Completeness of Sampling and Species Richness at Sites ............................. 29 The Relationship Between Mean Local Density and Body Size ....................... 30 Results ................................ ................................ ................................ .................... 30 Relative Species Density ................................ ................................ .................. 30 Completeness of Sampling and Species Richness at Sites ............................. 30 The Relationship Between Mean Local Density and Body Size ....................... 31 Discussion ................................ ................................ ................................ .............. 31 4 THE RELATIONSHIP BETWEEN DENSITY AND DISTRIBUTION ........................... 39 Introduction ................................ ................................ ................................ ............. 39

PAGE 5

5 Methods ................................ ................................ ................................ .................. 40 Data ................................ ................................ ................................ .................. 40 Statistical Analysis ................................ ................................ ............................ 41 Results ................................ ................................ ................................ .................... 43 Disc ussion ................................ ................................ ................................ .............. 44 5 DISPERSAL ABILITY AND DISTRIBUTION ................................ .............................. 53 Introduction ................................ ................................ ................................ ............. 53 Overview ................................ ................................ ................................ .......... 53 Defining Dispersal ................................ ................................ ............................ 53 Objectives ................................ ................................ ................................ ......... 54 Methods ................................ ................................ ................................ .................. 55 Dispersal Ability ................................ ................................ ................................ 55 Range Size ................................ ................................ ................................ ....... 56 Statistical Analysis ................................ ................................ ............................ 56 Results ................................ ................................ ................................ .................... 56 Discussion ................................ ................................ ................................ .............. 56 6 GENERAL CONCLUSIONS ................................ ................................ ....................... 61 Macroecology and Study Group ................................ ................................ ............. 61 Species Range Sizes ................................ ................................ .............................. 61 Relative Density Within Butterf ly Communities ................................ ....................... 61 The Relationship Between Density and Distribution ................................ ............... 62 The Relationship Between Dispersal Ability and Distributi on ................................ .. 62 Outlook ................................ ................................ ................................ ................... 62 APPENDIX BUTTERFLIES COLLECTED IN THE FIELD WITH RANGE SIZES ............................. 64 LIST OF REFERENCES ................................ ................................ ............................... 67 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 76

PAGE 6

6 LIST OF TABLES Table page 3 1 Field sites in Ecuador, August December 2010 ................................ ................. 34

PAGE 7

7 LIST OF FIGURES Figure page 2 1 Frequency distributions of butterfly range sizes.. ................................ ................ 24 2 2 Arithmetic plot showing correlation between range size e stimates obtained using two independent methods ................................ ................................ ........ 25 3 1 Field sites in e astern Ecuador.. ................................ ................................ .......... 35 3 2 Species rank abundance curves (Whittaker plots) for ithomiines sampled at all study sites.. ................................ ................................ ................................ .... 36 3 3 Speci es accumulation curves for Ithomiini at 5 eastern Ecuadorian field sites.. ................................ ................................ ................................ .................. 37 3 4 Species mean local density, where found, as a function of two measures of body size for ithomiine butterflies found at the field sites.. ................................ .. 38 4 1 Arithmetic plots of global range size as a function of mean local density for ithomiine butterflies. ................................ ................................ ............................ 49 4 2 Arithmetic plots of global range size as a function of mean local density within field sites.. ................................ ................................ ................................ 50 4 3 Arithmetic plots of global range size as a function of mean local density wi thin genera.. ................................ ................................ ................................ .... 51 4 4 Arithmetic plot of generic global range size as a function of generic mean local density (N=23). ................................ ................................ ........................... 52 5 1 Ari thmetic plots of range size as a function of 3 estimators of dispersal ability for ithomiine butterflies. ................................ ................................ ...................... 60

PAGE 8

8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Ful fillment of the Requirements for the Degree of Master of Science MACROECOLOGY OF NEOTROPICAL CLEARWING BUTTERFLIES (NYMPHALIDAE: ITHOMIINI) By Geoffrey R. Gallice August 2011 Chair: Keith R. Willmott Major: Entomology Macr oecology is the study of large scale patterns in species abundance and distribution. While a positive relationship between abundance and distribution is widely documented, most studies to date have focused on vertebrate species in temperate regions, wherea particularly among invertebrates. To investigate several key macroecological phenomena for the first time in a diverse Neotropical insect group, the clearwing butterflies of the nymphal id tribe Ithomiini, I combined measures of mean local density gathered in the field (five sites in eastern Ecuador) with range size data compiled using ecological niche modeling, as well as measures of body size and dispersal ability measured from museum s pecimens, to investigate: 1) the relationship between mean local density and body size, 2) the relationship between mean local density and distribution, and 3) the relationship between dispersal ability and distribution. There was an overall negative relat ionship between density and body size, although inspection of arithmetic plots of density as a function of two estimates of body size, mean forewing length and mean thorax width, revealed rather triangular relationships, with species of intermediate body s ize displaying the highest densities in both cases. Two measures of

PAGE 9

9 mean local density, density at field sites where species were found and density averaged over all visited sites, did not predict range size, and there was no relationship between density and range size within any field site. There was no relationship between mean local density and distribution within genera, and no relationship between density and distribution among closely related, sister species pairs. At a higher taxonomic level, no rel ationship was found between generic mean local density and generic range size. Finally, three indirect measures of dispersal ability, mean forewing length, mean forewing area, and mean wing load (ration of thorax width to forewing length) failed to predict species range size. These results collectively suggest that macroecological patterns in the tropics, at least for ithomiine butterflies, might differ from those reported for temperate species. Future directions for research, in ithomiine butterflies and o ther Neotropical nymphalid groups, are discussed.

PAGE 10

10 CHAPTER 1 GENERAL INTRODUCTION Macroecology Macroecology is the study of patterns in species abundance and distribution at large spatial and temporal scales (Gaston & Blackburn 1999). As opposed to the up approach to understanding biological systems, in which a system is reduced to components that are analyzed independently, macroecologists search for emergent properties of entire systems. For example, in a pioneering macro ecological work, Brown (1984) showed that there was typically a positive relationship between abundance and geographic distribution among species comprising large, continental scale biotas, such as North American birds, mammals, and plants. The number of s tudies aiming to document such large scale patterns in species abundance and distribution has since proliferated (for a review see Blackburn et al 2006). While studies of the interactions between species and their local environments have certainly gone a long way in helping to uncover processes underlying the structuring of local communities, the search for broader trends has the unique potential to help reveal the processes that shape community composition and patterns in species richness at larger scales (Brown et al 2002; Blackburn & Gaston 2002b; Brown & Maurer 1989). The implications for fields ranging from ecological and historical biogeography (e.g. Blackburn & Gaston 2002a) to evolutionary biology (e.g. Verbruggen et al 2009) to conservation biolo gy (e.g. Kerr et al 2007) are diverse.

PAGE 11

11 Study Group There are several key considerations in choosing a model system for macroecological study. First, and perhaps most importantly, is the availability of the data required to evaluate the relationship betwee n abundance, distribution, and ecological attributes of species in the system. Second, since species typically represent the units of analysis in macroecology, the system should be sufficiently speciose and widespread to investigate large scale patterns. F inally, practical considerations dictate that the system be amenable to study using often limited resources. Currently, relatively few groups fulfill each of these criteria, and thus most macroecological work to date has focused on select vertebrates of te mperate regions (e.g. birds, mammals), for which species are not temperate vertebrates, but tropical invertebrates (Myers et al 2000), macroecologists must seek g roups for study that more accurately represent global biodiversity. Neotropical butterflies are emerging as an attractive model system for macroecological study, having been collected and studied extensively since the earliest European explorers visited th e region. More distributional, taxonomic (e.g. Lamas 2004), and ecological (e.g. Beccaloni et al. 2008) information is currently available for many species of butterflies than for most other Neotropical groups. Furthermore, butterflies are extremely diverse and found commonly in most terrestrial ecosystems throughout the Neotropics, and can be easily sampled using simple techniques and inexpensive equipment. The nymphalid butterfly tribe Ithomiini (ithomiines) is a diverse clade of diurnal Lepidoptera (Papil ionoidea) found exclusively in the Neotropics, consisting of approximately 370 species (Lamas 2004) that inhabit moist forests from sea level to

PAGE 12

12 approximately 3000 meters elevation. They range from Mexico through Central America, south to southern Brazil, northern Argentina and Paraguay, and across three Caribbean islands (Cuba, Hispaola, Jamaica) (Willmott & Freitas 2006). The tribe is diagnosed by the presence in males of a patch of elongated, erectile androconial scales along the anterior edge of the do rsal hindwing (Brower et al 2006). These scales, often called hair pencils, are used by males to disseminate volatile pheromones used in courtship, the precursors of which are acquired in most species by feeding on flowers of Asteraceae and dried Boragina ceae plants (Schultz et al 2004). These precursors, which comprise various dehydropyrrolizidine alkaloids, also make adults of all species unpalatable (Trigo & Brown 1990; Brown 1985). As a result of this unpalatability, ithomiine species form the basis o f a diverse array of Batesian and Mllerian mimicry complexes (Beccaloni 1997) that have stimulated a wealth of study on natural selection, beginning notably with the theory of mimicry by Bates (1862) and M ller (1879). Due in part to their peculiar biolo gy and natural histories, and also to their overall abundance, charisma, and relative ease of sampling, distributional data in the form of georeferenced museum specimens is more readily available for ithomiines than for many other Neotropical insect groups As such, they are an attractive choice for studies in biogeography and macroecology that employ recently developed tools that require extensive data sets, for example distribution or niche modeling applications. Furthermore, species level phylogenetic hy potheses have been proposed for several genera and communities (de Silva et al 2010; Elias et al 2009; Elias et al 2008; Mallarino et al 2005), allowing for the use of phylogenetic comparative methods, such as phylogenetically independent contrasts (Fe lsenstein 1985).

PAGE 13

13 For all of these reasons, I have chosen the Ithomiini as a model group for investigating large scale patterns in abundance and distribution, as well as for exploring some potential underlying factors such as dispersal and body size. The ta xonomy that I follow is that of Lamas (2004), with the exception of the genus Napeogenes for which I follow the revised nomenclature proposed by Elias et al (2009). Objectives Throughout the following chapters, I explore several fundamental macroecologic al phenomena, using data on ithomiine butterfly density, distribution, and morphology gathered from various sources, including museum specimens, databases, and from the field. In Chapter 2, I describe a method for estimating geographic range sizes of ithom iine species found in the western Amazonian lowlands of eastern Ecuador, from where I also collected mean local density data. An analysis of these density data is presented in Chapter 3, where I also test for a relationship between density and body size. I n Chapter 4, I explore whether the commonly observed relationship between density and range size applies to ithomiine butterflies. Finally, in Chapter 5, I investigate whether there exists a relationship between dispersal ability and range size. This is th e first time such a macroecological study has investigated these phenomena in a Neotropical insect clade, and among the first for tropical invertebrates anywhere.

PAGE 14

14 CHAPTER 2 MODELING ITHOMIINE B UTTERFLY RANGES Introduction The concept of species range siz e is central to a variety of fields in theoretical and applied ecology and their related disciplines. Although more attention has historically been paid to species traits such as abundance, a growing number of biogeographic studies have recently begun to i nvestigate patterns in range size among species, driven in part by the increasing availability of distributional data sets and a growing recognition that an improved understanding of large scale patterns such as those among range sizes can lead to a better knowledge of the processes acting upon local assemblages (Gaston 1996 a ). Patterns in species richness are also increasingly being 0). With the popularization of the field of macroecology by studies such as Brown (1984) and Brown and Maurer (1989), a growing number of macroecological studies have attempted to compare such patterns in species distributions with other species attributes including abundance and ( Blackburn & Gaston 2002b; Brown 1984 ). In conservation, range size is often considered an important component of rarity, and is also often used as a criterion by which a species may be considered to be in danger of extinction For example, according the International Union for the Conservation of Nature (IUCN) Red List Categories and Criteria (IUCN 2001), a widely recognized system for the establishme nt of species threat status, a species may be considered threatened if it is range restricted (e.g. if its range size is < 20,000 km 2 ).

PAGE 15

15 Although a large number of studies have used measures of range size to answer often similar questions, there are major d ivisions between related studies in how it is actually measured. Measurements of range size typically fall into one of two categories: 1) geographic ra which is the area within the most widely spaced lo calities for a species, or 2) occupancy of grid cells drawn over a given area of interest (Gaston 1996b) Importantly, neither measurement scheme produces precise measures of range size, but estimations or predictions, and both methods have their limitatio ns. For instance, the measures of range size derived from the extent of occurrence method are sensitive to bias in distributional or locality data from which they are derived particularly at range edges The method also likely overestimates range size in many cases, since no species occupies every single area within its range (e.g. species are not found in areas of suboptimal habitat). On the other hand, measures of occupancy are sensitive to the scale at which grid cells are considered. Large grid cells m ay overestimate range size if a species is counted in grid cells where its actual distribution is limited. Small grid cells may underestimate range size if species are not sampled but occur in cells. In macroecology, systematic sampling bias may lead to an artifactual relationship between abundance and distribution using the occupancy method, since rare species might not be sampled in all cells in which they occur (Gaston et al 1997). Additionally, it is important to draw a distinction between measures of global range size (i.e. the entire range area of a species, or its complete extent of occurrence) and partial range size (e.g. range size within a study area usually a measure of occupancy ). While studies that consider species range size at these differi ng resolutions might

PAGE 16

16 attempt to explore similar fundamental questions, they may actually address different phenomena entirely and thus might not be comparable (Gaston 1996 a ). The disparity between studies in their description of range size is almost certa inly attributable, at least in part, to the quality of data available for different taxa and from different biogeographic regions Rarely, distributional data representing species presence and absence are sufficient that can be measu red almost directly, such as in the case of North American and western European birds (e.g. Sauer et al can also be measured almost directly, such as in studies where occupancy o f grid cells is known over a given region (e.g. Cowley et al 2001). In all other cases, however, various sources of variable accuracy and precision. Recent technical and methodological advances in physiography (physical geography) have delivered large scale environmental data sets, such as global, high resolution digital elevation models and interpolated climatic parameters (e.g. high resolution interpolated global cli mate surface, Hijmans et al [2005]). Such environmental data, coupled with powerful information systems, have opened the door to computer based distribution modeling, and ecologists have responded with the development of a number of tools for predicting s pecies distributions based upon comparatively limited distributional data sets. Key to this approach is the calculation of a species ecological niche, which is a mathematical representation of a species distribution in environmental space. For example, the bioclimatic predictive system

PAGE 17

17 locations where the species is known to occur. BIOCLIM then em ploys an algorithm to space, or bioclimatic envelope, in a process known as environmental or ecological niche modeling. Similar approaches using different algorithms have also been developed, including the Genetic Algorithm for Rule Set Production ( GARP ; Stockwell & Peters 1999), and maximum entropy methods (e.g. MAXENT; Phillips et al 2006, 2004), among others. Such ecological niche modeling techniques require that s pecies are in equilibrium in their environments (i.e. distributions are stable), and that distributional data represent all relevant environmental gradients (i.e. no systematic bias in sampling, sufficient sampling of all habitats occupied by the species). In addition, and perhaps most importantly, these methods require that the climatic parameters used are, in fact, determinants of species distribution (e.g. Peterson 2001; Mackey & Lindenmayer 2001). relatively modest distributional data sources (Beaumont et al 2005; Pearson & Dawson 2003). Precise range estimates are currently unavailable for most Neo tropical butterfly taxa, and ecological niche modeling methods provide an objective, reliable, and efficient way of compiling such estimates. Here I outline a method for predicting and calculating butterfly range sizes using ecological niche modeling techn iques, and present range size estimates for the ithomiine butterfly community of the Amazonian lowlands of eastern Ecuador. I end with a discussion of the range size frequency distribution for this ithomiine community.

PAGE 18

18 Methods Data A total of 26,540 unique georeferenced locality records were available from throughout the Neotropics for n=73 species of ithomiine butterfly species occurring in the Amazonian lowlands of eastern Ecuador. An average of 363.6 locality records were available per species (s=378.1) with a minimum of 21 records ( Hyalyris [n. sp.] ; recorded from 3 sites, with 19 records from a single site in Ecuador), and a maximum of 2,111 ( Mechanitis lysimnia ; recorded at a total 477 sites throughout the Neotropics). Data were compiled from various sources, including collections in Ecuador, the University of Florida (McGuire Center for Lepidoptera and Biodiversity, Florida Museum of Natural History), and databases compiled for the Butterflies of Ecuador Project (www.butterfliesofecuador.com) and the Tropical Andean Butterfly Diversity Project (Mallet et al. 2007; www .andeanbutterflies.org). Recently collected and databased specimens (i.e. those collected after the introduction of commercially available hand held GPS units) were usually georeferenced rather accurately (e.g. coordinates measured at time of collection, with error usually <10 m). Older records were georeferenced by searching various sources for GPS coordinates, including published maps and online sources such as Google Earth (www.google.c om/earth/index.html). The majority of these records were georeferenced at an accuracy of at least 2 to 3 decimal places (1.11 km and 111 meters at the equator, respectively), with a small number of records being georeferenced at an accuracy of 1 decimal de gree (approximately 111 kilometers at the equator). Georeferencing errors were checked by inspecting maps of species locality records and investigating possible outliers or erroneous records.

PAGE 19

19 Modeling Butterfly Ranges Global range estimates for ithomiine b utterflies were estimated using the ecological niche modeling tool BIOCLIM, implemented in the free GIS software DIVA GIS ( http://www.diva gis.org ). BIOCLIM was used to summarize 19 climatic parameters derived from glo bal, interpolated precipitation and mean, minimum, and maximum temperature data (Hijmans et al 2005) for each species. Histograms for each parameter were inspected, and those parameters for which histograms followed either a normal or highly skewed distri bution were included in the model (mean=6.5 parameters per species, s=2.56), following Beaumont et al (2005). Parameters with even distributions, or otherwise non normal or non skewed distributions, were deemed to have little influence in the determinatio n of distribution, and were therefore excluded from models (see Introduction). BIOCLIM produces an initial predicted range divided into percentile based categories, with the resulting range typically having core areas of high climatic suitability surrounde d by areas of decreasing suitability. All areas deemed as suitable, thus including all areas with climate similar to any location where a species was recorded in its predicted range. A distribution mask was then applied, to exclude areas deemed climatically suitable by BIOCLIM, but where a species was known not to occur due to biogeographic barriers, dispersal limitation, etc. Range estimates were then projected t o a world cylindrical equal area projection in ArcGIS ( ESRI, Redlands, CA, USA ), since BIOCLIM calculates a predicted range without considering curvature of urther towards either pole are smaller and thus should be weighted less in calculations of area). Range sizes were then calculated from these projected range estimates in ArcGIS. In addition, ranges were estimated using the

PAGE 20

20 minimum convex polygon method in DIVA GIS, and range sizes were subsequently calculated from projected minimum convex polygons in ArcGIS (excluding sea area where appropriate), for comparison with modeled ranges. All range sizes were calculated in km 2 Data were sufficient for calculatin g range sizes for 73 species using the minimum convex polygon method; since H. [ n. sp. ] (a very locally distributed species found only in eastern Ecuador) was recorded in so few grid cells of the digital elevation model used by BIOCLIM, data were insuffici ent for modeling using this program, and thus H. [n. sp.] was excluded from all analyses with range sizes derived with BIOCLIM (hence range sizes were calculated for 72 species using BIOCLIM) Results Range Size Frequency Distributions E cological niche modeling Global range sizes for N=72 species of clearwing butterflies estimated using ecological niche modeling in BIOCLIM vary from approximately 4.3 x 10 4 km 2 ( Athesis acrisione found only in the eastern Andean foothills of Ecuador and e xtreme southern Colombia) to 1.5 x 10 7 km 2 ( M. lysimnia an extremely widespread species found throughout the entire Neotropical region), spanning 4 orders of magnitude The frequency distribution for ranges modeled in BIOCLIM (Figure 3 1 a) is strongly rig ht skewed, with a median value of 4.3 x 10 6 km 2 Most species have relatively small ranges (ca. 60% of species have a range size <5 million km 2 ), whereas a small er number have very large ranges (ca. 15% have range size >10 million km 2 ). See Appendix A for a list of species and their range estimates with BIOCLIM.

PAGE 21

21 Minimum convex polygon Global range sizes for N=73 species estimated using the minimum convex polygon method spanned 7 orders of magnitude, ranging from approximately 2.7 x 10 2 km 2 ( H. [ n. sp. ] ve ry narrowly distributed only in Eastern Ecuador) to 1.6 x 10 7 km 2 ( M. lysimnia ). With H. [ n. sp. ] removed from the analysis, range sizes for N=72 species ranged from 3.2 x 10 4 km 2 ( A. acrisione ) to 1.6 x 10 7 km 2 ( M. lysimnia ), spanning 4 orders of magnitud e. The frequency distribution for these data (Figure 3.1b) are strongly right skewed, with a median value of 3.6 x 10 6 km 2 (including H. [ n. sp. ] ). Similar to ranges estimated using BIOCLIM, most species have range sizes that are comparatively small, with relatively few species very widely distributed throughout the Neotropics. Comparing Methods for Estimating Butterfly Range Sizes Global, untransformed range size estimates derived using both methods correlated very well (Figure 3 2 ; slope=0.96, R 2 =0.93, p <0.0001), using range size estimated with the minimum convex polygon method as the predictor variable. H. [n. sp.] was excluded from the analysis due to lack of a modeled (BIOCLIM) range estimate. A power analysis was conducted to estimate type II error pr obability and to ensure that the analysis was conducted using a large enough sample size from the fitted model involve comparatively larger estimates of range size by t he minimum convex polygon method compared to estimates using BIOCLIM; this is unsurprising, considering the biologically irrelevant restriction imposed upon the polygon method, that is, that the polygon must be convex. While serving as a way of introducing objectivity to the range estimation process, this is also likely to overestimate ranges in

PAGE 22

22 Andean foothills. For this reason, and s ince range size estimates obtained using the se two methods were otherwise generally very consistent, all subsequent analyses (see following chapters) were performed using range size estimates obtained using the BIOCLIM method. Discussion Ithomiine range sizes exhibited a right skewed frequency distr ibution (Fig. 3 1 ), which is perhaps the most common distribution for range sizes among animals (Blackburn et al 2004; Gaston 1998). Some authors have suggested evolutionary causes for such frequency distributions, for example in island regions where isol ation and limited dispersal have promoted the evolution of geographically restricted species on islands (e.g. Beck et al 2006). Historical biogeographic processes might also play a limited role. For example, small, secondary peaks in range sizes are appar ent towards the right tails of ithomiine range size frequency distributions, which are comprised entirely of species that have distributions spanning both sides of the Andes, including into Central America and Mexico in all cases. These peaks might have ar isen in one of two ways: 1) species comprising the peaks have breached a major biogeographical barrier (i.e. the Andes) after its formation, or 2) the formation of the barrier within t not all taxa, with newly formed sister species on either side of the barrier having smaller ranges than unspeciated taxa. The relative roles of these competing hypotheses remain to be explored. However, such skewed frequency distributions can also result from purely ecological processes (Gaston & He 2002), a hypothesis that I argue is the most well supported based upon my results for ithomiine butterflies. For example, processes

PAGE 23

23 acting upon island faunas are not likely to have significant influence over a diverse, widespread, continental fauna. In addition, while a major geographic barrier seems to influence a small, secondary peak in species distributions, most species are distributed on only one side of the barrier (in South America) and thus other proce sses must be acting to determine their distributions. The next step is to search for underlying, ecological causes of such a distribution in ithomiine butterflies and to test them using carefully designed field and observational studies. To begin to explor e this question, in subsequent chapters I examine further species attributes, such as local density and dispersal ability, in order to identify which attributes correlate with, and might help to explain, the distribution of range sizes that I observed.

PAGE 24

24 Figure 2 1. Frequency distribution s of butterfly range sizes E stimated using (a) ecological niche modeling (BIOCLIM) (N=72 spp.); ( b ) mini mum convex polygon method (N=73 spp .). (a) (b)

PAGE 25

25 Figure 2 2 Arithmetic plot showing correlation between range size e stimates obtained using two independent methods. X=minimum convex polygon method (MCP ); Y= ecological niche modeling using BIOCLIM (MODEL ). MODEL=4.84x10 5 + 0.96MCP; R 2 =0.93; p<0.0001; N=72.

PAGE 26

26 CHAPTER 3 SPECIES RELATIVE DEN SITY WITHIN BUTTERFL Y COMMUNITIES I ntroduction Relative species density is a major topic of interest in ecology and many ecologists work to understand its relationship with species distribution and the underlying causes. One universal feature of any biological community is that abundance i s variable a mong the species comprising the community; some species are common, whereas others, and usually most others, are relatively rare (Magurran 2005). This common observation has led to the formation of a number of mathematical and ecological models of community structure (e.g. geometric series, Motomura 1932; log series, Fisher 1943; log normal, Preston 1948; unified neutral theory, Hubbel 2001). While the number of models for describing patterns in abundance within communities has grown, so too ha s the number of applications, in fields ranging from community ecology to macroecology to conservation biology. B utterflies are an ideal system for investigating patterns in abundance for several reasons. First of all, they are extremely diverse and widesp read, being found in most terrestrial ecosystems. Secondly, they are relatively easy to sample using simple, inexpensive techniques and equipment. Thirdly due at least in part to their broad appeal among biologists and popularity with collectors throughou t recent history, more distributional data is available for this group than for many other groups in the Neotropics, allowing the application of recently developed tools for modeling comprehensive ranges for many species. My main goal here is to compile da ta on species mean local density for ithomiine butterflies of the Amazonian lowlands of eastern Ecuador, an important first step towards a better understanding of key

PAGE 27

27 macroecological phenomena, such as the relationship between species density and distribut ion. Several authors have argued that body size in animals is negatively related to local density (Currie et al. 1993). For example, Damuth (1981) found a very tight, negative relationship between body size (mass) and local density for mammalian primary co and Wassenberg (1983) found similar relationships in a number of widely distributed animal groups, including herbivorous and carnivorous mammals and birds, vertebrate poikilotherms, and aquatic and terrestrial invertebrates. Interestingly, however, studies that investigate the relationship between body size and local density within local or natural communities often fail to find this relationship, finding instead a relationship that is rather triangular, or no relationship at all. In a study of various guilds of birds and beetles in Borneo and North America, Blackburn et al. (1990) found no relationship between density and body size in any case. According to the authors, studies that find a tight negative relationship between density and body size might actually be showing the hypotenuse of a triangle since these studies use data collected in areas where species are most dense and thus worth studying. As a result, many species that are generally found at lower densities in natural communities are excluded from the plot of density versus body size, hiding a truly triangular relationship. In order to investigate this relationship in a diverse community of tropical insects, the clearwing b utterflies of the nymphalid tribe Ithomiini, I used measures of body size compiled from morphological measurements and measures of mean local density gathered in the field.

PAGE 28

28 Methods Study Sites and Butterfly Sampling Species density data for ithomiine butte rflies were collected at 5 sites in the Amazonian lowlands of eastern Ecuador during August December, 2010 (see Table 3 1 Figure 3 1 for a description of sites). Since population synchrony has been demonstrated at distances of up to approximately 1 km f or butterflies in the study area (Engen et al. 2002), sites were located an average of 51 km apart, to avoid issues of spatial autocorrelation and to ensure that sites were independent samples of butterfly communities (i.e. not pseudoreplicates). A ll sites were within relatively close proximity of the Napo River (mean=7.4 km), a large Amazonian tributary, and the largest in eastern Ecuador. Habitat at each site comprised mostly terra firme forest that is never flooded, with patches of varzea or flooded forest, that is partially or fully inundated at least periodically. Ithomiine density data were collected along e stablished trails, referred to here as tr ansects. Under the reasonable assumption that trails are created rather haphazardly and therefore rep resent habitat types in roughly the proportion in which they occur at a site, measures of ithomiine density gathered from transects should be representative of overall local density patterns. Total trail length within sites varied from at least 5 km (Jatun Sacha) to >20 km (Yasun Scientific Research Station). Transects were patrolled daily, weather permitting, and ithomiines found within 2.5 meters of either side of the transect were netted with a hand net following a modified form of the technique outline d by Pollard (1977, 1979). Each specimen was then given a unique voucher number and stored for later identification. No butterflies were collected during specimen processing in the field to avoid biasing collection towards preferred microhabitats, lekking sites (pockets) etc. Transect start and stop times were recorded.

PAGE 29

29 Species mean local density was the number of in dividuals collected divided by the time spent collecting at sites where the species was found I define density in the same manner throughout this chapter, and throughout the remainder of this thesis unless otherwise noted The s pecies rank density distribution was tested for goodness of fit to a standard distribution model, namely the lognormal distribution, using the non parametric Kolmogorov Smirnov test. Completeness of Sampling and Species Richness at Sites Species accumulation curves were used to determine the completeness of sampling of the ithomiine community at the study sites. These curves plot the total, accumulated number of species S by the total number of samples (hand net days) n. Since the order with which samples are added can affect the form of the curve (Colwell & Coddington 1994), curves were produced using original sampling data and original data with 50 randomizations for com parison. In order to evaluate the total species richness of the ithomiine community at the study sites, total species richness was computed using the non parametric estimator proposed by Chao (1984). This estimator takes the form: max = S obs + ( a 2 /2b) (equation 3 1 ) where a is the number of species represented by 1 individual (singletons), b is the number of species represented by 2 individuals (doubletons), S obs is the total number of species collected, and max is the estimated total species richness.

PAGE 30

30 The Relationship Between Mean Local Density and Body Size In order to investigate a potential correlate of density, body size, regression analyses were performed using two estimates of body size: 1) forewing length, from wing base to apex and 2) thorax width at the base of the forewings (for a thorough description of these measurements see Chap. 5). These metrics of body size were the predictor variables in all analyses. All variables were log 10 transformed in both a nalyses to normalize residual distributions. Results Relative Species Density A total of N=55 species were collected at the 5 field sites during the sampling period (Appendix A). density fit the lognormal distribution wel l (D=0.0998, p > 0.150 ), when rank density data were included for all species sampled at the sites, including H. sarepta (an extremely abundant outlier, found very locally but abundantly at 1 site). When H. sarepta was excluded from the rank abundance calc ulations, the resulting rank abundance distribution still fit the lognormal distribution (D= 0.1094, p=0.104). Species rank abundance curves are shown in Fig. 3 2 Completeness of Sampling and Species Richness at Sites Species accumulation curves (Fig. 3 3 ) stabilized after approximately 2 0 hand net days, suggesting that the ithomiine community at the study sites was adequately sampled. While both curves begin to approach a similar asymptote after about 20 hand net days, the curve constructed using actual, n on randomized data climbs more slowly and monotonically compared to the curve produced by a randomization process, which might be due to temporal turnover in species at least at the first site visited (Jatun

PAGE 31

31 Sacha), a possibility that will be explored in a subsequent publication. Species accumulation curves constructed for each site individually (not shown) failed to reach asymptotes, due to the limite d sampling period at each site. Using the so estimator (Colwell & Coddington 1994) of speci es richness, total species richness at the max a value comparable to the apparent asymptotes of both species accumulation curves. The Relationship Between Mean Local Density and Body Size Forewing length, a surro gate for body size, predicted mean local density, when an extremely dense outlier ( H. sarepta ) was both included (abundance=1.3187 + 1.1235forewing length; R 2 =0.1008; p=0.0193) and excluded (abundance=1.1138 + 1.0718forewing length; R 2 =0.1009; p=0.0205) from the analysis. The remaining surrogate for body size, thorax width, also predicted mean local density, when an extremely dense outlier (again, H. sarepta ) was included (abundance= 1.7220 + 1.0200thorax width; R 2 =0.1077; p=0.0154) and excluded (abundan ce= 1.7931 + 0.9627thorax width; R 2 =0.1054; p=0.0177) from the analysis. All variables were log 10 transformed to normalize residual distributions. See Figure 3 4 for double log and arithmetic plots of density as a function of body size. Discussion Data f or species relative densities fit well with a lognormal distribution, a hollow curve model that has been documented almost ubiquitously in biological communities (Magurran 2005; May 1975). This suggests that the community of ithomiine butterflies at the st udy sites contained a relatively small proportion of very abundant species, with the remainder and majority of the community comprising species that are relatively rare.

PAGE 32

32 Two surrogates of body size were overall, negatively correlated with mean local densi ty for ithomiine butterflies found at the study sites, and double log plots (Fig. 3 4 ) of each measure of body size with mean local density show a negative relationship, however with relatively little variation in density explained by body size in either c ase. Inspection of the arithmetic plots of mean local density as a function of mean forewing length and mean thorax width (Fig. 3 4 ) suggests that this negative relationship is heavily influenced by several comparatively large bodied but rare species (e.g. forewing length: Athesis acrisione, Methona curvifascia, M. grandior ; thorax width: A. acrisione, M. confusa, M. curvifascia, M. grandior ). These arithmetic plots also suggest that the relationship between mean local density and body size may be somewhat triangular, with species of interme diate body size having the high est mean local densities A similar, triangular relationship was reported for herbivorous beetles in Borneo by Blackburn et al. (1990), in which density increased with body size to some inte rmediate value and then subsequently decreased. For ithomiine butterflies, however, it is unclear the extent to which this relationship is truly triangular, with lower limits to body size (the smallest ithomiine species sampled at the Ecuadorian field site s was Scada zibia, with a mean forewing length of 17.61mm and a mean thorax width of 1.01mm) potentially obscuring the fully triangular form of plots of density as a function of body size. Notably, studies of the relationship between body size and density at larger scales (e.g. global density measures compiled from the literature for a diversity of taxa ) often find a much stronger negative relationship that is not triangular (e.g. mammals, Dam uth 1981; birds, mammals, vertebrate poikilotherms, invertebrate s, Peters & Wassenberg 1983). Blackburn et al (1990) suggested this pattern results from, at least

PAGE 33

33 in part, the exclusion of rare species in analyses, since data for these species is often difficult to gather. Second, they suggest such a relationship betw een density and body size might be an artifact of density data collected for species in areas of maximum density, since areas where species are common and thus easier to study are more likely to be visited. As a result, a tight, negative relationship might actually reflect the hypotenuse of a triangle, with species maximum density, but not necessarily overall density, being highly correlated with body size. Interestingly, the relationship between body size and density in ithomiine butterflies in eastern Ec uador appears to be, at once, both triangular and yet generally negative. It remains to be seen what form this relationship will take with the inclusion of taxa in analyses that differ more widely in their ecology. For example, density and body size data a re currently becoming available for other guilds of Neotropical butterflies, and in the future I plan to present results of tests of the relationship between density and body size among a wider variety of butterfly species, including bait trapped butterfli es of the nymphalid subfamilies Charaxinae, Biblidinae, and the genus Adelpha In addition to allowing comparisons with ithomiine butterflies, data for these groups will permit analysis of the relationship between density and body size in a more diverse Ne otropical butterfly community.

PAGE 34

34 Table 3 1. Field sites in Ecuador, August December 2010 Site name Latitude Longitude Elevation (m) Jatun Sacha Biological Reserve (JS) S 0104.2' W 07737.0' 450 Yarina Lodge (YL) S 0028.1' W 07650.6' 250 Sacha Lodge ( SL) S 0028.0' W 07628.0' 235 Yasun Scientific Research Station ( YN ) S 0040.6' W 07623.7' 230 Tiputini Biodiversity Station (TBS) S 0038.2' W 07609.0' 230

PAGE 35

35 Figure 3 1. Field sites in eastern Ecuador Sites were visited between Aug. and Dec., 2 010 ( JS=Jatun Sacha Biological Reserve; YL=Yarina Lodge; SL=Sacha Lodge; YN=Yasun Scientific Research Station; TB S=Tiputini Biodiversity Station)

PAGE 36

36 Figure 3 2 Species rank abundanc e curves (Whittaker plots) for i thomiines sampled at all study sit es (a) I ncluding H. sarepta (an extreme outlier); (b) excluding H. sarepta. Species are ranked by abundance, i.e. the most abundant species have higher ranks (lower numbers). Note the strong right skew of both curves, indicating that ithomiine communities are comprised primarily of rare species. (a) (b)

PAGE 37

37 Figure 3 3 Species accumulation curves for Ithomiini at 5 eastern Ecuadorian field sites E mpty circles represent actual species accumulation curve; solid circles represent curve based on 50 randomizations calc ulated using EstiamteS v 8.2 (Colwell 2009)

PAGE 38

38 Figure 3 4 Species m ean local density, where found as a function of two measures of body size for ithomiine butterflies found at the field sites. (a) A rithmetic and (b) double log plots of density as a function of forewing length; (c) arithmetic, and (d) double log plots of density as a function of thorax width. (a) (b) (c) (d)

PAGE 39

39 CHAPTER 4 THE RELATIONSHIP BET WEEN DENSITY AND DIS TRIBUTION Who can explain why one species ranges widely and is very numerous, and why another allied species C. Darwin (1859) Introduction A positive relationship between species abundance and distribution is very well documented, and it is often considered one of the few general rules in ecology (Boc k & Ricklefs 1983; Brown 1984; Gaston & Lawton 1988). In a recent review and meta analysis of 279 instances of a relationship between abundance and distribution, Blackburn et al (2006) found a significant positive relationship in groups as varied as North American birds (Bock and Ricklefs 1983), British mammals (Blackburn et al 1997), grassland, desert, and forest plants (Collins and Glenn 1990; Guo et al 2000; Leite and Lopes 2001), British marine invertebrates (Foggo et al 2003), and 444 species of mo ths, aphids, carabid beetles, and bracken feeding insects in Europe (Gaston & Lawton 1988). However, w hile these studies span a variety of realms and taxonomic groups, their geographical scope is relatively limited, with most studies being conducted in tem perate regions ( Gaston & Blackburn 1999). Indeed, of the 253 studies included in the meta analysis that could be assigned to broadly defined regions (i.e. Old/New World tropical/temperate areas), only 47 (ca. 19%) were from the tropics

PAGE 40

40 (Neotropics, n=15; A frotropics, n=28; Australasia, n=4) with the remaining 206 (ca 81%) from temperate regions (Palearctic, n=174; Nearctic, n=32) In contrast, the s living in tropical regions (M yers et al 2000; Rea ka Kudla et al 1997; Erwin 1982) While disproportionately such a focus on temperate species might have also largely failed to address potentially important issues associated with climatic differences between temperate versus tropical regions. For instance, temperate regions have experienced significant recent effects of glaciation (Hawkins & Porter 2003), and after the last glacial maximum. In addition, temporary colonization by cold intolerant yet highly dispersive species during warm summer months may contribute further confounding effects in these regions (e.g. Hawkins & DeVries 2009). Clearly, more work is needed in order to justify the acceptance interspecific abundance distribution relationship, with further studies needed especially for tropical invertebrates. Here I investigate the relationship between mean local density (a measure related to, but not equal to abundance; see Chap. 3) and geographic range size, a measure of global distribution ( Chap. 2), for the first time in a diverse Neotropical insect clade, the clearwing butterflies of the tribe Ithomiini. Methods Data Measures of species mean local density were available for N=54 species of ithomiine butterflies occurring at 5 sites in the Amazonian lowlands of eastern Ecuador (see Chap. 3). Mean local density was the number of individuals collected, divided by

PAGE 41

41 the time spent pa trolling transects at all sites where the species was found. This measure of density, mean density at sites where a species is found, is often used in macroecological studies since it avoids the generation of an artifactual relationship between density and distribution via the inclusion of increasing numbers of zeroes in calculations of mean density for species that are found at fewer and fewer sites (Gaston 1996b). However, since species range maps predicted all species to occur at or near each of the 5 fi eld sites, a second estimate of mean local density was calculated for comparison with density estimates at sites where species were found during the study period: mean density over all sites. The calculation of this measure of density addresses the possibi lity that all species found in the study region might occur at least rarely and/or occasionally at all sites even though they were not sampled during the study period, and might eventually be recorded with sufficient sampling. Global range size estimates w ere also available for the same N=54 species (see Chap. 2). F or within site analyses (see below) density was the number of individuals collected at that site alone divided by the total amount of time spent patrolling transects at that site All r ange size s were measured in km 2 Statistical Analysis The effects of mean local density ( at sites where found) and mean local density (across all sites) on global range size were tested for all species (N=54) using simple linear regression. In addition, the effects of mean density (within sites) on global range size were tested for species occurring at individual sites (JS, N=20; YL, N=44; SL, N=37; Y N N=37; T BS N=24 ) conducted with linear regression produced no n normal residual distributions. To minimize the effects of sampling locally abundant but not generally abundant species

PAGE 42

42 (e.g. H. sarepta, see results/discussion), linear regression analyses were performed using 2 subsets of the data: 1) excluding species found at only 1 site, and 2) excluding species found at only 2 sites. Potential effects related to phylogeny were controlled for using two additional analyses of the relationship between density and distribution: 1) ecies within certain genera (e.g. Hypothyris N=6; Napeogenes N=6; Oleria N=7 ) over all sites; 2) Wilcoxon signed rank test to test for differences among species in sister species pairs. In this analysis, N=13 sister species pairs were identified based o n a community phylogeny for Ithomiini species found at the study sites from Elias et al. (2008), and species in each pair were ordered so that the first species in each pair had higher mean density than the second. A Wilcoxon signed rank test was then perf ormed for the resulting paired values of range size for each sister species pair. This test examined whether range sizes for the sister species in the first group (i.e. sister species with larger mean densities) were larger than those for the second group (i.e. sister species with smaller mean densities). This procedure was repeated, removing the 6 species pairs with an interspecific difference in mean density < 0.1 (leaving N=7 pairs), since species with only marginally different densities might not be exp ected to also differ substantially in range size (assuming a positive relationship between density and distribution). These tests were performed to investigate the relationship between density and distribution at comparatively fine scales (i.e. within gene ra and among sister species pairs), since the existence of a relationship at this scale might be obscured by the inclusion of multiple, ecologically divergent genera in analyses. Typically, however, so called phylogenetic comparative methods, such as phylo genetic independent contrasts (Felsenstein 1985), are used to account for

PAGE 43

43 niche conservatism in macroecological studies; since some species traits, such as range size, tend to show a phylogenetic signal (e.g. Beck 2007), species may not always constitute i ndependent data points, possibly leading to an artifactual relationship between density and distribution. In order to homogenize the scales of the variables in regression analyses, global range sizes were divided by 10 7 Where necessary, measures of mean local density or range size were transformed (log 10 or square root) to produce normal residual distributions for parametric (i.e. ordinary least squares regression) tests. In all analyses, species mean local density was used as the predictor variable. Sinc e different relationships between density and range size have been reported at different taxonomic levels (e.g. Harcourt et al 2005), the effects of taxonomic level were investigated by simple linear regression of mea sures of generic range and density for N=23 genera. Generic range size was the outline of all overlapping congeneric species ranges found at the field sites. No range data were available for H. mamercus so that species was excluded from generic range estimation for the genus Hypothyris Value s for generic range size were divided by 10 7 to homogenize the scales of the two variables for analysis. Generic density was mean local density where found, averaged over all congeneric species. Variables were otherwise untransformed, and generic density was the predictor variable. Results Species mean local density ( at sites where found; log transformed) did not predict range size (square root transformed) (range size = 0.6548 + 0.1291 density; R 2 =0.0029; p=0.7) when all species (N=54) were included in a linear regression analysis (Fig. 4 1 ) Species mean density (averaged over all sites) did not predict range size in a linear

PAGE 44

44 regression analysis (R 2 =0.0370; p= 0.1632 ; in all subsequent analyses, mean local density is density where found) (Fig. 4 1 ). Spearm did not show a relationship between mean local density and range size for species 0.0612, d.f.=35, p=0.7190; YSRS 0.2179, d.f.=34, p=0.2016; transformed) density as a predictor of range size (log transformed) produced no significant results using the data set with species occurring at only 1 site r emoved (range size = 1.0017 + 0.0702 density; R 2 =0.0039; p=0.6920). Similarly, no significant relationship was found using the data set (double log transformed) with species found at only 2 sites removed (range size = 1.1856 + 0.2191 density; R 2 =0.0458 tests conducted for species within genera showed no relationship between mean local density and range size in any case ( Hypothyris 0.0286, d.f.=4, p=0.9572; Napeogenes Oleria from a Wilcoxon signed rank test for N=13 sister species pairs (all available pairs) showed that species in pairs that had higher mean densi ties did not tend to have larger range sizes (p=0.7354). Results were similar when pairs whose constituent species differed in density by < 0.1 were removed from the analysis, with species in pairs that had higher mean density not tending towards larger ra nge size (p=0.8125). At the generic level, congeneric mean density did not predict generic range size (range size = 0.74886 + 0.39182 density; R 2 =0.0096; p=0.6558) (Fig. 4 4 ) Discussion Several a nalyses clearly show no relationship between two measures o f mean local density and range size for ithomiine butterflies, both across the 5 study sites and

PAGE 45

45 within each site (Figs. 4 1 & 4 2 respectively) Furthermore, no relationship was apparent when species occurring at only 1 and 2 sites were removed from the analyses to minimize the influence of very locally abundant but not generally abundant species The relationship between mean local density and range size was investigated among closely related, congeneric species, in order to minimize the potential confou nding effects of including more distantly related taxa in analyses. No relationship was found. Finally, two analyses conducted for sister species pairs in the ithomiine community of the study area similarly showed no relationship. Taken together, these res ults provide very strong evidence against a relationship between mean local density and range size for ithomiine butterflies at the field sites in eastern Ecuador. While I argue that my results rather strongly suggest that there is no relationship between density and distribution for ithomiine butterflies, I acknowledge two potential confounding factors in my analyses. First is the issue of scale, and whether density data collected on a local scale (5 sites in eastern Ecuador) is comparable to distribution al Indeed, in a meta analysis of 71 instances of a relationship between abundance and distribution, Blackburn et al (2006) were unable to reject their hypothesis tha t studies which employed different measurement scales for density and distribution had an overall lower effect size (i.e. weaker positive relationship) than those for which the scales were roughly equivalent. However, the mean effect size of 0.13 for n=17 studies employing different scales was still significantly different from 0, suggesting that studies that employ differing scales should produce similar results to those employing equal scales, even if they are less strong. The second potential argument fo r error in the

PAGE 46

46 analyses is related to local measures of species density at the field sites, and whether these estimates represent overall patterns in species density. Local density estimates might be argued to differ from overall densities in two ways: spa tially or temporally. Spatial disagreement between local versus broader density patterns might arise from so et al 1997a). For example, since many species included in the analyses were found near the edge of their g eographic or elevational ranges at the base of the Andes in Ecuador (i.e. most were lowland Amazonian species near the western edge of their ranges in the Amazon basin), spurious patterns in density might lead to an erroneous relationship between density a nd distribution if species abundance tends to decline towards the edge of the range (Grinnel 1922). However, I argue that this is unlikely a source of error for two reasons. l positive relationship (Gaston et al. 1997a), not a lack of relationship. Furthermore, the Andes are a rather abrupt, geographic boundary, and there is little reason to believe that species density should systematically decrease as one approaches the boun dary, before its effects (i.e. changes in temperature, precipitation) are felt by actually increasing in elevation. The field sites, with the exception of perhaps Jatun Sacha, were all firmly within the Amazonian lowlands of eastern Ecuador (Jatun Sacha, 4 50 m; all others, ca. 230 m elevation), despite relative proximity to the Andes (as opposed to, say, a site in central Brazil), and climatic factors such as temperature at these sites were essentially unaffected by proximity to the Andes (i.e. climate at t he field sites is strictly lowland Amazonian). On the other hand, disagreement between my local density estimates versus overall estimates might also arise due to temporal effects (i.e. limited sampling

PAGE 47

47 period). For example, several studies of other nympha lid butterflies in the study area have demonstrated strong temporal or seasonal turnover in community composition and abundance patterns (e.g. Checa et al 2009; DeVries et al 1997), and it is unclear whether sampling during only part of a single year may reliably represent general patterns in density over time. However, the a bundance ranking of ithomiine species found at the field sites 211 days spent in the field collecting butterflies, sp anning 14 years and 87 sites in the study region (K. Willmott, pers. comm. ). Furthermore, my sampling scheme thoroughly measured density of an actual community (Fig. 3 3 often cited positive relationship between dens ity and distribution are robust and broadly applicable, then they should apply to communities as they exist at any randomly selected point in space or time. While my results are in contrast to many studies of the relationship between density and distributi on, particularly in temperate regions, they agree with a small but growing body of evidence that suggests the relationship may be weaker or even non existent among tropical taxa. F or example, in Blackburn et al analysis, the mean effect size for n=15 studies from the Neotropics was not statistically significant from zero, indicating that there was no overall relationship between density and distribution for Neotropical taxa. However, these studies were conducted exclusively on mammals, and in particular, primates. More work is needed, especially for other taxa such as butterflies and other invertebrates that are more representative of tropical biodiversity in order to explore the relationship between density and distribution in diverse tropic al regions, and to confirm whether the form of this relationship is generally

PAGE 48

48 different for tropical versus temperate faunas. If a different relationship truly exists among tropical taxa, then this will provide a unique opportunity to test mechanisms and s earch for species attributes that drive a relationship in temperate regions but not in the tropics.

PAGE 49

49 Figure 4 1 Arithmetic plots of global range size as a function of mean local density for ithomiine butterflies. (a) W here found; (b) across all sites for ithomiines at 5 sites in eastern Ecuador (N=54). Outlier in (a) is H. sarepta Data shown are untransformed.

PAGE 50

50 Figure 4 2 Arithmetic plots of global range size as a function of mean local density within field sites (a) JS (Jatun Sacha), outlier is H. euclea (N=20); (b) YL (Yarina Lodge), outlier is H. euclea (N=44); (c) SL (Sacha Lodge) (N=37); (d) YN (Yasun Scientific Research Station), outlier is H. sarepta (N=37); (e) TBS (Tiputini Biodiversity Station) (N=24). All data shown are untr ansformed. Density (at sites where found)

PAGE 51

51 Figure 4 3 Arithmetic plots of global range size as a function of mean local density within genera (a) Hypothyris (N=6); (b) Napeogenes (N=6); (c) Oleria (N=7). Data shown are untransformed.

PAGE 52

52 Figure 4 4 Arithmetic plot of generic gl obal range size as a function of generic mean local density (N=23). Data shown are untransformed.

PAGE 53

53 CHAPTER 5 DISPERSAL ABILITY AN D DISTRIBUTION Introduction Overview or shift its range (Rundle et al 2002), and also to its ability to sustain a meta population structure in fragmented landscape (Baguette 2003; Conradt et al. 2000; Hill et al 1996). In the context of increasing global anthropogenic disturbance, including habitat fragmentation and climate change, understanding dispersal and its role in shaping patterns in species abundance and distribution will be important if we are to successfully manage natural populations and mitigate biodiversity loss (Stevens et al 2010). There is some evidence that rarity is linked to poor dispersal ability (Gaston 1994; Kunin & Gaston 1993), and that species that are better dispersers are also able to attain wider distributions, for example in Iberian carabid beetles (Gutirrez & Menndez 1997), Swedish mayflies (Malmqvist 1999), and British butterflies (Cowley et al 2001; Dennis et al 2000). To date, however, very few studies have investigated the influence of dispersal on range size in tropical taxa Importantly, present and hi storical climatic conditions in the tropics have exerted different influences on populations there as opposed to temperate areas (Hawkins & Porter 2003; Stenseth et al 2002), and it remains to be seen how the dynamics of dispersal and distribution differ in these areas as a result of different climatic effects Defining Dispersal Many definitions of dispersal exist, with the definition chosen in a particular study being dependent upon its context, and the motivation underlying the dispersal behavior

PAGE 54

54 (Bowle r & Benton 2005; Clobert et al 2001). For example, in macroecology, which is concerned with large scale might be the process by which one or more individuals reach a new site and establish a popul ation there (e.g. Beck 2007). In evolutionary biology, where interest is focused on the evolution of life history traits and speciation, dispersal might be any movement that drives gene flow (Ronce 2002). Alternatively, dispersal might describe any movemen t of an individual, regardless of purpose, whether it be movement by one individual away from another, for example during territorial disputes, or movement of an individual away from its natal site (Johnson & Gaines 1990). However, all of these definitions of dispersal entail movement of individuals, and both the propensity for and capacity for estimated, allowing for investigation of its influence on other species attr ibutes. mobility, that makes it, in part, a successful or unsuccessful disperser. Thus, throughout this chapter, when I mention dispersal, I mean mobility, or a species Objectives Here I investigate the relationship between dispersal and distribution for the first time in a diverse Neotropical insect group, the clearwing but terflies of the tribe Ithomiini. Information regarding butterfly dispersal wa s gathered from museum specimens and from specimens collected in the field. Estimates of distribution were compiled for butterfly species in the form of geographic range sizes, estimated using ecological niche modeling techniques

PAGE 55

55 Methods Dispersal Ability Instead of direct measures of dispersal ability such as those obtained using mark release recapture techniques (e.g. Hill et al 1996), species dispersal ability was estimated indirectly using morphological data (e.g. Beck 2007; Hill et al 1999). Species included in the analyses were those predicted to occur in the western Amazonian lowlands of eastern Ecuador based on range maps created using niche modeling techniques (see Chap. 2), and for which both range sizes and pinned specimens or photographs were available (for a list of species included in analyses see Appendix A). Three morphological characters related to flight morphology were measured for N=570 mean=8.03, s=3.51) representing 69 species Specimens w ere fairly evenly distributed by sex, with N=236 females, N=329 males, and N=5 undetermined. Measurements were : 1) forewing length (measured from the base of the forewing to its apex), 2) area of the forewing (measured from the dorsal surface of the wing), and 3) width of the thorax (at the base of the forewing). These metrics were then used to calculate 3 surrogates for mobility/dispersal ability in each species: 1) mean forewing length, a common estimator of body size in Lepidoptera (Loder et al. 1998) an d a correlate of flight ability in butterflies (e.g. Chai & Srygley 1990); 2) mean forewing area, a second measure of body size /flight ability ; and 3) mean wing load (ratio of thorax width: forewing length), a measure of flight strength (Beck, 2007; Dudley & Srygley 1994; Chai & Srygley 1990) in which larger values denote greater flight strength All measurements were taken from photographs of pinned specimens using the digital imaging software Photoshop CS4 (Adobe Systems, In corporated, San Jose, CA, USA).

PAGE 56

56 Range Size Range size estimates (in km 2 ) for N=72 species were available, including 69 for which dispersal estimators were tabulated, based on range estimates calculated using ecological niche modeling implemented in DIVA GIS using the BIOCLIM modeling s o ftware (for details see Chap. 2 ; Appendix A ). Statistical Analysis The independent effects of each dispersal estimator on range size were tested using ordinary least squares regression models for N=69 ithomiine species with range size as the dependent var iable in all cases. Where necessary, variables were transformed (square root, log 10 ) to normalize residual distributions. Results No measure of dispersal ability was useful in describing variation in range size for ithomiine butterflies: untransformed mean forewing length did not predict untran s formed range size (range size= 3.0 x 10 6 + 7.2 x 10 4 forewing length; R 2 =0.0183; p=0.2679); untransformed mean forewing area did not predict untransformed range size (range size = 4.0 x 10 6 + 3.7 x 10 3 forewing area; R 2 =0.0148; p=0.3201); square root transformed mean wing load did not predict square root transformed range size (range size = 3.5 x 10 3 + 5.6 x 10 3 wing load; R 2 =0.0 063; p=0.5162) (Fig. 5 1 ). Discussion None of the estimators of dispersal ability (forewi ng length, forewing area, wing load) predicted range size for ithomiine butterflies If the metrics that I chose to represent dispersal ability do, in fact, serve as valid surrogates, then my study provides strong evidence against a general relationship be tween dispersal ability and distribution in this group. While support for the notion that dispersal ability can be reliably estimated

PAGE 57

57 using morphological characters is mixed, there is some direct evidence for butterflies that suggests that it can be (e.g. Berwaerts et al 2002; Chai & Srygley 1990). In the only related study of a tropical lepidopteran taxon that I am aware of, Beck (2007) found little evidence of a relationship between dispersal ability and range size in southeast Asian moths of the family Sphingidae, using similar measures of dispersal ability. In contrast, some evidence supports a positive relationship between dispersal ability and distribution in temperate butterflies (e.g. Cowley et al 2001; Dennis et al 2000 ). While this topic has, by no means, been exhaustively examined, if these studies do represent trends in temperate and tropical Lepidoptera, then it begs the immediate question of why this pattern is observed in temperate but not tropical taxa. With a lack of relationship in ithomi ine butterflies further supported (see below), I would proceed to test hypotheses related to differences in current and historical climate between temperate versus tropical regions (e.g. Hawkins & Porter 2003) that might help to explain differences between patterns of dispersal and distribution observed in taxa from these two regions. On the other hand, it is of course possible that dispersal ability cannot be reliably estimated using morphological characters. In gathering estimates of butterfly mobility o r dispersal ability, several authors have employed rather different methods than those I used here, including the circulation of questionnaires among experienced lepidopterists (e.g. Komonen et al 2004; Cowley et al 2001), and the more direct measurement of dispersal through mark release recapture methods (e.g. Pollard & Yates 1993) and the analysis of vagrancy in grids (e.g. Cook et al 2001). Without direct tests of the usefulness of morphological characters as surrogates for dispersal ability in ithomi ine

PAGE 58

58 butterflies, it is difficult to conclude irrefutably that dispersal ability does not influence range size. Data for actual dispersal ability in this group, as measured using mark release recapture techniques, or with more direct surrogates such as expe rimental estimates of flight speed measured in the field/laboratory or perhaps even using radio telemetry techniques (e.g. Wikelski et al 2010 ) would go a long way in corroborating or refuting my findings and confirming the validity of indirect estimators of dispersal ability in Neotropical butterflies If it turns out that, after thorough examination of dispersal ability, dispersal ability really does not account for distribution in ithomiine butterflies, other explanations will have to be explored to acc ount for the incredible variation in range size observed in this group (see Chap. 2). unclear. Nevertheless, other factors almost certainly exert additional influence, and it is example, in addition to competitive interactions, abiotic factors such as climate are key ortant determinants of distribution in animals (Peterson 2001; Mackey & Lindenmayer 2001). Cowley et al ecological niche, correlated significantly with distribution in British butterfl ies. Similarly, Beck (2007) found that larval diet breadth, another measure of ecological niche breadth, was significantly correlated with range size in southeast Asian sphingid moths. I recently visited Ecuador (Aug. Dec. 2010) to collect diverse behavi oral and ecological data for ithomiine butterflies data that I intend to use to estimate ecological niche breadth and to examine how this is related to distribution; I also plan future field work to collect host

PAGE 59

59 plant utilization data for larval ithomiine butterflies, in order to explore the relationship between larval host plant breadth and distribution.

PAGE 60

60 Figure 5 1. Arithmetic plots of range size as a function of 3 estimators of dispersal ability for ithomiine butterflies (a) F orewing length (N=6 9); (b) forewing area (N=69); (c) wing load (ratio of thorax width to forewing length) (N=69). All data shown are untransformed.

PAGE 61

61 CHAPTER 6 GENERAL CONCLUSIONS Macroecology and Study Group In Chapter 1, I introduced the field of macroecology, which is the study of large scale patterns in abundance and distribution. Throughout the remainder of the thesis, I attempted to tackle some fundamental macroecological questions. Are species that are locally abundant also more widespread? Do species with greater disp ersal ability occupy larger geographic ranges? To answer these questions, I chose the Neotropical butterfly tribe Ithomiini. Ithomiines have fascinated biologists since the time of the earliest South American explorers. Indeed, their ubiquity, ease of coll ection, and fascinating biology make them an ideal model system for study in a diversity of fields, including macroecology. Species Range Sizes In Chapter 2, I presented geographic range sizes for ithomiines occurring in the western Amazonian lowlands of e astern Ecuador. Since distributional data are limited for this group, as with most other groups in the Neotropics, I elected to estimate large ranges throughout the Neotropic s, with the majority being more narrowly distributed. Such information is a prerequisite for almost any macroecological analysis, several of which I pursued in subsequent chapters. Relative Density Within Butterfly Communities In Chapter 3, I outlined rece nt field work to collect local density data for ithomiines in Ecuador (Aug. Dec. 2010). I described the species rank abundance distribution from the 5 field sites and compared that to a standard distribution model, the lognormal

PAGE 62

62 model, to which my data f it. Few species were locally very abundant, but the remainder, and majority, were much more rare. I also evaluated completeness of sampling at the field sites using species accumulation curves, and showed that I had a thorough sample of the species occurri ng at the sites. Finally, I investigated the relationship between local density and body size, finding that two separate measures of body size, forewing length and thorax width, both predicted local density; larger bodied species tended to have lower densi ties. The Relationship Between Density and Distribution In Chapter 4, using measures of range size and density compiled in previous chapters, I investigated the relationship between density and distribution (range size) in ithomiines. Using several differe nt methods, including two in which potential phylogenetic non independence was accounted for, I found no relationship. The Relationship Between Dispersal Ability and Distribution Finally, in Chapter 5, I explored the relationship between dispersal ability and distribution (range size) for ithomiines. I measured dispersal ability indirectly using morphological measures associated with flight ability in butterflies. I found no relationship between dispersal and distribution, and discussed future prospects fo r further study on this topic. Outlook Most of my analyses challenge well documented patterns within the animal kingdom, with the unsurprising exceptions of ithomiine rank density curves and range size frequency distributions. Most unexpectedly, I found s trong evidence against a relationship between density and distribution for ithomiine butterflies. There are several potential methodological difficulties worth mentioning, namely my use of measures of

PAGE 63

63 density and distribution compiled at different scales, and of potential sampling bias due to the relatively small temporal and spatial scale of my density sampling effort. In the future, I plan to address these concerns, by adding density and distributional data for additional species, from other locations and times from throughout the Neotropics. results. It is probable that there is at least some (and perhaps much) underreporting of such results, as they are often more difficult to publish (Gaston & Blackburn 1999). If we fail to generate a balanced body of literature, it will be difficult to investigate macroecological patterns as they truly exist in nature (if they exist at all). Indeed, a more well supported lack of relationship between density and distribution in the mega diverse Neotropical ithomiine butterflies might be more interesting, and perhaps more informative, than a confirmation of patterns observed in other groups and regions.

PAGE 64

64 APPENDIX A BUTTER FLIES COLLECTED IN T HE FIELD WITH RANGE SIZES (Above) Total numbers of each species collected by hand net, per field site (JS=Jatun Sacha; YL=Yarina Lodge; SL=Sacha Lodge; YN=Yasun Scientific Research Station; TBS=Tiputini Biodiversity Station), during fi eld work (August December 2010), with species for which ranges were estimated with BIOCLIM. Range sizes are in km 2 Numbered superscripts (1 13) indicate species pairs (each pair with unique number) included in Wilc oxon signed rank test (Chap. 4). (*) indicates no range size data available from range modeling. Superscript ( d ) indicates taxa included in dispersal ability analysis (Chap. 5). Species JS YN SL YL TBS Range size Aeria eurimedia Cramer 1777 11 ,d 0 17 15 4 1 8,328,526 Athesis acrisione Hewitson 1869 d 0 0 0 2 0 43,000 Brevioleria arzalia Hewitson 1876 d 3 3 12 16 2 1,518,976 Callithomia alexirrhoe Bates 1862 5,d 0 0 0 1 0 7,233,941 Callithomia lenea Cramer 1779 5,d 0 3 7 11 1 12,058,406 Ceratin ia tutia Hewitson 1852 3,d 2 2 24 12 0 4,639,863 Episcada sulphurea Haensch 1905 3,d 0 0 0 2 0 1,189,846 Forbestra olivencia Bates 1862 12,d 0 11 23 14 2 4,648,919 Forbestra proceris Weymer 1883 12,d 5 0 0 0 0 2,093,783 Godyris zavaleta Hewitson 1855 d 0 9 34 19 8 3,886,202 Heterosais giulia nephele Hewitson 1855 d 0 0 1 2 0 1,585,144 Hyalyris [n. sp.] Willmott & Lamas* 0 1 0 0 0 no data Hyalyris coeno Doubleday 1847 d 1 0 0 0 0 880,582 Hypoleria lavinia Hewitson 1855 d 1 1 30 3 3 11,363,417 Hypoleria sare pta Hewitson 1852 d 0 47 0 0 6 9,051,825 Hyposcada anchiala Hewitson 1868 2,d 1 8 6 9 5 4,301,553 Hyposcada illinissa Hewitson 1852 d 2 2 27 17 2 4,319,711 Hyposcada kena Hewitson 1872 2,d 0 1 3 2 1 2,365,818 Hypothyris anastasia Bates 1862 d 1 2 0 4 0 3,37 3,879

PAGE 65

65 Hypothyris euclea Godart 1819 7,d 18 1 1 28 3 12,294,947 Hypothyris fluonia Godman 1900 d 0 9 3 8 3 7,087,153 Hypothyris mamercus Hewitson 1869* 7 0 2 6 3 0 no data Hypothyris semifulva Salvin 1869 d 0 0 22 3 1 2,244,316 Ithomia amarilla Haensch 190 3 d 0 22 14 0 7 482,717 Ithomia salapia Hewitson 1853 d 5 4 6 2 6 2,775,514 Mechanitis lysimnia Fabricius 1793 d 0 4 6 2 0 14,879,369 Mechanitis mazaeus Hewitson 1860 d 1 4 10 3 0 7,121,808 Mechanitis messenoides Felder & Felder 1865 d 0 12 14 3 0 1,871,241 Mechanitis polymnia Linnaeus 1758 d 0 0 1 2 0 12,754,819 Melinea menophilus Hewitson 1856 13,d 0 3 6 5 0 4,853,140 Melinea satevis Doubleday 1847 13,d 0 0 0 5 0 5,826,009 Methona confusa Butler 1873 d 4 0 1 2 0 7,314,041 Methona curvifascia Weymer 1883 d 0 0 2 2 0 2,256,255 Methona grandior Forbes 1944 d 0 1 1 1 1 5,350,135 Napeogenes aethra Hewitson 1868 8,d 0 5 7 0 0 139,498 Napeogenes inachia Hewitson 1855 10,d 2 3 22 17 3 10,400,295 Napeogenes larina quadrilis Hewitson 1856 8,d 0 0 0 2 0 2,936,230 Nap eogenes pharo Felder & Felder 1862 9,d 0 2 0 2 0 3,717,056 Napeogenes rhezia Geyer 1834 9,d 0 3 0 0 0 6,083,537 Napeogenes sylphis Gurin Mneville 1844 10,d 1 1 16 9 3 5,196,811 Oleria agarista Felder & Felder 1862 d 0 0 0 0 1 994,144 Oleria assimilis Hae nsch 1903 d 1 0 6 1 0 1,907,958 Oleria gunilla Hewitson 1858 d 2 7 20 5 5 4,562,477 Oleria ilerdina Hewitson 1858 1,d 0 0 3 0 1 4,363,231 Oleria onega Hewitson 1852 1,d 0 13 8 8 0 3,565,964 Oleria sexmaculata Haensch 1903 d 0 7 0 0 2 929,668 Oleria tigilla Weymer 1899 d 1 0 0 0 0 1,418,977 Pseudoscada florula Hewitson 1855 6,d 3 15 10 5 3 8,830,916 Pseudoscada timna Hewitson 1855 6,d 0 1 2 3 0 3,825,920 Pteronymia primula Bates 1862 4,d 0 7 5 5 0 6,602,995 Pteronymia sao Hbner 1813 d 0 0 0 1 0 4,540,145 Pt eronymia vestilla Hewitson 1853 4,d 6 0 3 22 0 2,995,068 Scada zibia Hewitson 1856 d 0 7 15 18 1 3,237,947

PAGE 66

66 Thyridia psidii Linnaeus 1758 d 0 1 0 3 0 11,729,265 Tithorea harmonia Cramer 1777 11,d 1 2 0 1 0 13,407,300 Brevioleria seba Hewitson 1872 d n/a n/a n/a n/a n/a 3,489,241 Ceratinia neso Hbner 1806 d n/a n/a n/a n/a n/a 6,816,812 Dircenna dero Hbner 1823 d n/a n/a n/a n/a n/a 12,889,417 Dircenna loreta Haensch 1903 d n/a n/a n/a n/a n/a 1,019,544 Episcada mira Hewitson 1877 d n/a n/a n/a n/a n/a 156,9 84 Forbestra equicola Cramer 1780 d n/a n/a n/a n/a n/a 5,196,547 Godyris dircenna Felder & Felder 1865 d n/a n/a n/a n/a n/a 1,276,646 Hypoleria aureliana Bates 1862 d n/a n/a n/a n/a n/a 521,443 Hypothyris moebiusi Haensch 1903 d n/a n/a n/a n/a n/a 734, 176 Hypothyris ninonia Hbner 1806 n/a n/a n/a n/a n/a 10,773,151 Ithomia agnosia Hewitson 1855 d n/a n/a n/a n/a n/a 12,100,866 Ithomia arduinna d'Almeida 1952 d n/a n/a n/a n/a n/a 2,212,192 Mcclungia cymo Hbner 1806 d n/a n/a n/a n/a n/a 11,232,270 M elinea marsaeus Hewitson 1860 d n/a n/a n/a n/a n/a 5,132,662 Melinea mnasias Hewitson 1856 d n/a n/a n/a n/a n/a 6,491,833 Napeogenes duessa Hewitson 1859 d n/a n/a n/a n/a n/a 2,302,551 Pteronymia tucuna Bates 1862 d n/a n/a n/a n/a n/a 2,254,233 Scada r eckia Hbner 1808 n/a n/a n/a n/a n/a 8,398,236

PAGE 67

67 LIST OF REFERENCES Baguette, M. 2003. Long distance dispersal and landscape occupancy in a metapopulation of the cranberry fritillary butterfly. Ecography 26: 153 160. Beaumont, L.J., Hughes, L., & M. Pou lsen. 2005. Predicting species distributions: use of climatic parameters in BIOCLIM and its impact on predictions of species current and future distributions. Ecological Modelling 186: 250 269. Beccaloni, G.W. 1997. Vertical stratification of ithomiine bu tterfly (Nymphalidae: Ithomiinae) mimicry complexes: the relationship between adult flight height and larval host plant height. Biological Journal of the Linnean Society 62: 13 34. Beccaloni, G.W., Viloria, A.L., Hall, S.K., & G.S. Robinson. 2008. Catalog ue of the hostplants of the Neotropical butterflies/Catlogo de las plantas husped de las mariposas neotropicales. Vol. 8. Zaragoza: Monografas 3ercer Milenio. 536 p. Beck, J. & I.J. Kitching. 2007. Correlates of range size and dispersal ability: a comp arative analysis of sphingid moths from the Indo Australian tropics. Global Ecology and Biogeography 16: 341 349. Beck, J., Kitching, I.J., & K.E. Linsenmair. 2006. Measuring range size of South East Asian hawkmoths (Lepidoptera: Sphingidae): effects of scale, resolution and phylogeny. Global Ecology and Biogeography 15: 339 348. Berwaerts, K., Van Dyck, H., & P. Aerts. 2002. Does flight morphology relate to flight performance? An experimental test with the butterfly Pararge aegeria Functional Ecology 1 6: 484 491. Blackburn, T.M., Cassey, P., & K.J. Gaston. 2006. Variations on a theme: sources of heterogeneity in the form of the interspecific relationship between abundance and distribution. Journal of Animal Ecology 75: 1426 1439. Blackburn, T.M. & K. J. Gaston. 2002. Macroecology is distinct from biogeography. Nature 418: 723. Blackburn, T.M. & K.J. Gaston. 2002. Scale in macroecology. Global Ecology and Biogeography 11: 185 189. Blackburn, T.M., Gaston, K.J., Quinn, R.M., Arnold, H. & R.D. Gregory. 1997. Of mice and wrens: the relation between abundance and geographic range size in British mammals and birds. Philosophical Transactions of the Royal Society, London, B 352: 419 427. Blackburn, T.M., Harvey, P.H., & M.D. Pagel. 1990. Species number, pop ulation density and body size relationships in natural communities. Journal of Animal Ecology 59: 335 345.

PAGE 68

68 Blackburn, T.M., Jones, K.E., Cassey, P., & Losin, N. 2004. The influence of spatial resolution on macroecological patterns of range size variation : a case study using parrots (Aves: Psittaciformes) of the world. Journal of Biogeography 31: 285 293. Bock, C.E. & R.E. Ricklefs. 1983. Range size and local abundance of some North American songbirds: a positive correlation. American Naturalist 122: 295 299. Bowler, D.E. & T.G. Benton. 2005. Causes and consequences of animal dispersal strategies: relating individual behaviour to spatial dynamics. Biological Reviews 80: 205 225. Brower, A.V.Z., Freitas, A.V.L., Lee, M., Silva Brand o, K.L., Whinnett, A., & K.R. Willmott. 2006. Phylogenetic relationships among the Ithomiini (Lepidoptera: Nymphalidae) inferred from one mitochondrial and two nuclear gene regions. Systematic Entomology 31: 288 301. Brown, J. H. 1984. On the relationship between abundance and distribution of species. The American Naturalist. 124: 255 279. Brown, J.H., Gillooly, J.F., West, G.B., & V.M. Savave. 2002. The next step in macroecology: from general empirical patterns to universal ecological laws. In Proceedings of the XII National Congress of the Italian Ecological Society, Urbino, Italy. Brown, J.H. & B.A. Maurer. 1987. Evolution of species assemblages: effects of energetic constraints and species dynamics on the diversification of the North American avifauna. American Naturalist 130: 1 17. Brown, J.H. & B.A. Maurer. 1989. Macroecology: the division of food and space among species on continents. Science 243: 1145 1150. Brown, K.S. 1985. Chemical ecology of dehydropyrrolizidine alkaloids in adult Ithomiinae (Lepidoptera: Nymphal idae). Revista Brasileira de Biologia 44: 435 460. Chai, P. & R.B. Srygley. 1990. Predation and the flight, morphology, and temperature of Neotropical rain forest butterflies. The American Naturalist 135: 748 765. Chao, A.1984. Nonparametric estimation o f the number of classes in a population. Scandinavian Journal of Statistics 11: 265 270. Checa, M.F., Barragn, A., Rodrguez, J., & M. Christman. 2009. Temporal abundance patterns of butterfly communities (Lepidoptera: Nymphalidae) in the Ecuadorian Amaz onia and their relationship with climate. Annales de la Socit Entomologique de France 45: 470 486.

PAGE 69

69 Clobert, J., Danchin, E., Dhondt, A.A., & J.D. Nichols (eds.). 2001. Dispersal. Oxford University Press, Oxford, U.K. Collins, S.L. & S.M. Glenn. 1990. A in grassland vegetation. American Naturalist 135: 633 648. Colwell, R.K. 2009. EstimateS: statistical estimation of species richness and shared cation published at: http://purl.oclc.org/estimates Colwell, R.K. & J.A. Coddington. 1994. Estimating terrestrial biodiversity through extrapolation. Philosophical Transactions of the Royal Society London B 34 5: 101 118. Colwell, R.K. & D.C. Lees. 2000. The mid domain effect: geometric constraints on the geography of species richness. Trends in Ecology and Evolution 15: 70 76. Conradt, L., Bodsworth, E.J., Roper, T.J., & C.D. 2000. Thomas. Non random dispersa l in the butterfly Maniola jurtina : implications for metapopulation models. Proceedings of the Royal Society of London B 267: 1505 1570. Cook, L.M., Dennis, R.L.H., & P.B. Hardy. 2001. Butterfly hostplant fidelity, vagrancy and measuring mobility from dis tribution maps. Ecography 24: 497 504. Cowley, M.J.R., Thomas, C.D., Wilson, R.J., Le n Cort s, J.L., Guti rrez, D., & C.R. Bulman. 2001. Density distribution relationships in British butterflies. II. An assessment of mechanisms. Journal of Animal Ecology 70: 426 441. Currie, D.J. 1993. What shape is the relationship between body size and population density? Oikos 66: 353 358. Damuth, J. 1981. Population density and body size in mammals. Nature 290: 699 700. Dennis, R.L.H., Donato, B., Sparks, T.H., & E. Pollard. 2000. Ecological correlates of island incidence and geographical range among British butterflies. Biodiversity and Conservation 9: 343 359. de Silva, D.L., Day, J.J., Elias, M., Willmott, K., Whinnett, A., & J. Mallet. 2010. Molecular phylogen etics of the Neotropical butterfly subtribe Oleriina (Nymphalidae: Danainae: Ithomiini). Molecular Phylogenetics and Evolution 55: 1032 1041. DeVries, P.J., Murray, D., & R. Lande. 1997. Species diversity in vertical, horizontal, and temporal dimensions o f a fruit feeding butterfly community in an Ecuadorian rainforest. Biological Journal of the Linnean Society 62: 343 364. Dudley, R. & Srygley, R.B. 1994. Flight physiology of Neotropical butterflies: allometry of airspeeds during natural free flight. Jou rnal of Experimental Biology 191: 125 139.

PAGE 70

70 Elias, M., Gompert, Z., Jiggins, C., & K. Willmott. 2008. Mutualistic interactions drive ecological niche convergence in a diverse butterfly community. PLOS Biology 6: 2642 2649. Elias, M., Joron, M., Willmott, K., Silva Brando, K.L., Kaiser, V., Arias, C.F., Gomez Pierez, L.M., Uribe, S., Brower, A.V.Z., Freitas, A.V.L., & C.D. Jiggins. 2009. Out of the Andes: patterns of diversification in clearwing butterflies. Molecular Ecology 18: 1716 1729. Engen, S., L ande, R., Walla, T., & P.J. DeVries. 2002. Analyzing spatial structure of communities using the two dimensional poisson lognormal species abundance model. The American Naturalist 160: 60 73. Erwin, T.L. 1982. Tropical Forests: their richness in Coleopter a and other arthropod species. The Coleopterists Bulletin 36: 74 75. Felsenstein, J. 1985. Phylogenies and the comparative method. The American Naturalist 125: 1 15. Fisher, R.A, Corbet A.S., & C.B. Williams. 1943. The relation between the number of spec ies and the number of individuals in a random sample of an animal population. Journal of Animal Ecology 12: 42 58. Foggo, A., Frost, M.T. & M.J. Attrill. 2003. Abundance occupancy patterns in British estuarine macroinvertebrates. Marine Ecology Progress S eries 265: 297 302. Gaston, K.J. 1994. Rarity. Chapman and Hall, London, U.K. Gaston, K.J. 1996. Species range size distributions: patterns, mechanisms and implications. Trends in Ecology & Evolution 11: 197 201. Gaston, K.J. 1998. Species range size distributions: products of speciation, extinction and transformation. Philosophical Transactions of the Royal Society of London B 353: 219 230. Gaston, K.J. 1996. The multiple forms of the interspecific abundance distribution relationship. Oikos 76: 211 220. Gaston, K.J., Blackburn, T.M., & R.D. Gregory. 1997. Interspecific abundance range size relationships: range position and phylogeny. Ecography 20: 390 399. Gaston, K.J., Blackburn, T.M., & J.H. Lawton. 1997. Interspecific abundance range size relati onships: an appraisal of mechanisms. Journal of Animal Ecology 66: 579 601.

PAGE 71

71 Gaston, K.J. & F. He. 2002. The distribution of species range size: a stochastic process. Proceedings of the Royal Society of London B 269: 1079 1086. Gaston, K. J. & J.H. Lawton 1988. Patterns in the distribution and abundance of insect populations. Nature 331: 709 712. Gaston, K.J. & T.M. Blackburn. 1999. A critique for macroecology. Oikos 84: 353 368. 380. Guo, Q., Brown, J.H. & T.J. Valone. 2000. Abundance and distribution of desert annuals: are spatial and temporal patterns related? Journal of Ecology 88: 551 560. Gutirrez, D. & R. Menndez. 1997. Patterns in the distribution, abundance and body size of cara bid beetles (Coleoptera: Caraboidea) in relation to dispersal ability. Journal of Biogeography 24: 903 914. Hagemeijer, W.J.M. & M.J. Blair (Eds.). 1997. The EBCC atlas of European breeding birds: their distribution and abundance. T. & A.D. Poyser, Britai n. Harcourt, H., Coppeto, S.A., & S.A. Parks. 2005. The distribution abundance (density) relationship: its form and causes in a tropical mammal order, Primates. Journal of Biogeography 32: 565 579. Hawkins, B.A. & E.E. Porter. 2003. Relative influences o f current and historical factors on mammal and bird diversity patterns in deglaciated North America. Global Ecology and Biogeography 12: 475 481. Hawkins, B.A. & P.J. DeVries. 2009. Tropical niche conservatism and the species richness gradient of North Am erican butterflies. Journal of Biogeography 36: 1698 1711. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., and A. Jarvis. 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25: 1965 1978. Hill, J.K., Thomas, C.D., & D.S. Blakely. 1999. Evolution of flight morphology in a butterfly that has recently expanded its geographic range. Oecologia 121: 165 170. Hill, J.K., Thomas, C.D., & O.T. Lewis. 1996. Effects of habitat patch size an d isolation on dispersal by Hesperia comma butterflies: implications for metapopulation structure. Journal of Animal Ecology 65: 725 735. Hubbell, S.P. 2001. The unified neutral theory of biodiversity and biogeography. Princeton University Press, Princet on, N.J. IUCN. 2001. IUCN Red List Categories and Criteria: Version 3.1. IUCN

PAGE 72

72 Species Survival Commission. IUCN, Gland, Switzerland and Cambridge, UK. ii + 30 pp. Johnson, M.L. & M.S. Gaines. 1990. Evolution of dispersal: theoretical models and empirica l tests using birds and mammals. Annual Review of Ecology and Systematics 21: 449 480. Kerr, J.T., Kharouba, H.M., & D.J. Currie. 2007. The macroecological contribution to global change solutions. Science 316: 1581 1584. Komonen, A., Grapputo, A., Kaital a, V., Kotiaho, J.S., & J. Pivinen. 2004. The role of niche breadth, resource availability and range position on the life history of butterflies. Oikos 105: 41 54. Kunin, W.E. & Gaston, K.J. 1993. The biology of rarity: patterns, causes and consequences Trends in Ecology and Evolution 8: 209 301. Lamas, G. 2004. Ithomiinae. In: Heppner, J.B. (Ed.), Checklist: Part 4A. Hesperioidea Papilionoidea. Atlas of Neotropical Lepidoptera, Vol. 5A. Association for Tropical Lepidoptera / Scientific Publishers, G ainesville, FL, USA: 172 191. Leite, S.J. F.S. Lopes. 2001. Local abundance and regional distribution of tree species of forest fragments in Brazil: a test of models. Revista de Biologa Tropical 49: 489 500. Loder, N., Gaston, K.J., Warren, P.H., R. Hen ry. 1998. Body size and feeding specificity: macrolepidoptera in Britain. Biological Journal of the Linnean Society 63: 121 139. Mackey, B.G. & D.B. Lindenmayer. 2001. Towards a hierarchical framework for modelling the spatial distribution of animals. Jou rnal of Biogeography 28, 1147 1166. Magurran, A.E. 2005. Measuring Biological Diversity. Blackwell Science Ltd., Malden, MA, USA. Mallarino, R., Bermingham, E., Willmott, K.R., Whinnett, A., & C.D. Jiggins. 2005. Molecular systematics of the butterfly genus Ithomia (Lepidoptera: Ithomiinae): a composite phylogenetic hypothesis based on seven genes. Molecular Phylogenetics and Evolution 34: 625 644. Mallet, J., Willmott, K.R., & Huertas, B. In: Tropical Andean Butterfly Diversity Project (TABDP). 2007. Darwin Database of Andean Butterflies. http://www.andeanbutterflies.org/database.html. Malmqvist, B. 1999. How does wing length relate to distribution patterns of stoneflies (Plecoptera) and mayflies (Ephemeroptera)? Biological Conservation 93: 271 276.

PAGE 73

73 May, R.M. 1975. Patterns of species abundance and distribution. In Ecology and Evolution of Communities (M.C. Cody & J.M. Diamond eds.): 81 120. Belknap Press, Cambridge, MA. Motomura, I. 1932. A statistical treatment of associations, J apanese Journal of Zoology 44: 379 383. M ller, F. 1879. Ituna and Thyridia : a remarkable case of mimicry in butterflies. Proceedings of the Entomological Society of London: xx xxix. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., & J. Kent. 2000. Biod iversity hotspots for conservation priorities. Nature 403: 853 858. Nix, H. 1986. A biogeographic analysis of Australian Elapid snakes. In: Longmore, R. (Ed.). Snakes: atlas of Elapid snakes of Australia. Bureau of Flora and Fauna, Canberra: 4 10. Pearso n, R.G. & T.P. Dawson. 2003. Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful? Global Ecology and Biogeography 12: 361 371. Peters, R.H. & K. Wassenberg. 1983. The effect of body size on anima l abundance. Oecologia 60: 89 96. niche modeling. The Condor 103: 599 605. Phillips, S.J., Anderson, R.P., & R.E. Schapire. 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling 190: 231 259. Phillips, S.J., Dud k, M., & R.E. Schapire. 2004. A maximum entropy approach to species distribution modeling. In: Proceedings of the Twenty First International Conference on Machine Learning: 655 662. Pollard, E. 1977. A method for assessing changes in the abundance of butterflies. Biological Conservation 12: 115 124. Pollard, E. 1979. A national scheme for monitoring the abundance of butterflies: the first three years. Proceedings of the Ent omological and Natural History Society 12: 77 90. Pollard, E. & T.J. Yates. 1993. Monitoring butterflies for ecology and conservation. Chapman & Hall, London, U.K. Preston, F.W. 1948. The Commonness, and Rarity of Species. Ecology 29: 254 283.

PAGE 74

74 Reaka Kud la, M.L., Wilson, D.E., & E.O. Wilson (eds). 1997. Biodiversity II: understanding and protecting our biological resources. Joseph Henry Press, Washington, D.C., USA. Ronce, O. 2007. How does it feel to be like a rolling stone? Ten questions about dispersa l evolution. Annual Review of Ecology, Evolution, and Systematics 38: 231 253. Rundle, S.D., Foggo, A., Choiseul, V., & D.T. Bilton. 2002. Are distribution patterns linked to dispersal mechanism? An investigation using pond invertebrate assemblages. Fresh water Biology 47: 1571 1581. Sauer, J.R., Hines, J.E., Fallon, J.E., Pardieck, K.L., Ziolkowski Jr., D.J., & W.A. Link. 2011. The North American breeding bird survey, results and analysis 1966 2009. Version 3.23.2011. USGS Patuxent Wildlife Research Cente r, Laurel, MD. Schultz, S., Beccaloni, G., Brown, K.S., Boppr M., Freitas, A.V.L., Ockenfels, P., & R. Trigo. 2004. Semiochemicals derived from pyrrolizidine alkaloids in male ithomiine butterflies (Lepidoptera: Nymphalidae: Ithomiinae). Biochemical Sys tematics and Ecology 32: 699 713. Stenseth, N.C., Mysterud, A., Ottersen, G., Hurrell, J.W., Chan, K., & M. Lima. 2002. Ecological effects of climate fluctuations. Science 297: 1292 1296. Stevens, G.C. 1989. The latitudinal gradients in geographical ra nge: how so many species co exist in the tropics. American Naturalist 133: 240 256. Stevens, V.M., Turlure, C., & M. Baguette. 2010. A meta analysis of dispersal in butterflies. Biological Reviews. doi: 10.1111/j.1469 185X.2009.00119.x Stockwell, D. & D. Peters. 1999. The GARP modelling system: problems and solutions to automated spatial prediction. International Journal of Geographical Inf ormation Science 13: 143 158. Trigo, J.R. & K.S. Brown. 1990. Variation of pyrrolizidine alkaloids in Ithomiinae: a comparative study between species feeding on Apocynaceae and Solanaceae. Chemoecology 1: 22 29. Verbruggen, H., Tyberghein, L., Pauly, K. Vlaeminck, C., Van Nieuwenhuyze, K., Kooistra, W.H.C.F., Leliaert, F., & O. De Clerck. 2009. Macroecology meets macroevolution: evolutionary niche dynamics in the seaweed Halimeda Global Ecology and Biogeography 18: 393 405. Wikelski, M., Moxley, J., E aton Mordas, A., L pez Uribe, M.M., Holland, R., Moskowitz, D., Roubik, D.W., & R. Kays. 2010. Large range movements of Neotropical orchid bees observed via radio telemetry. Plos One 5: e10738. doi:10.1371/journal.pone.0010738

PAGE 75

75 Willmott, K.R. & A.V.L. Fre itas. 2006. Higher level phylogeny of the Ithomiinae (Lepidoptera: Nymphalidae): classification, patterns of larval hostplant colonization and diversification. Cladistics 22: 297 368.

PAGE 76

76 BIOGRAPHICAL SKETCH Geoffrey Gallice grew up in Laurel, Maryland an d received his b in biology from the University of Maryland in College Park. He currently lives in Gainesville, Florida When not investigating butterfly ecology, he enjoys travelling and photography.