The Effects of Geomorphology and Primary Productivity on Neotropical Leaf Litter Herpetofauna: Implications for Amazonian Rainforest Conservation Jessica L

Home , Ahua (spider), Anomaloglossus stepheni

Louisiana State University LSU Digital Commons

LSU Doctoral Dissertations Graduate School

2009 The effects of geomorphology and primary productivity on neotropical leaf litter herpetofauna: implications for Amazonian rainforest conservation Jessica L. Deichmann Louisiana State University and Agricultural and Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_dissertations

Recommended Citation Deichmann, Jessica L., "The effects of geomorphology and primary productivity on neotropical leaf litter herpetofauna: implications for Amazonian rainforest conservation" (2009). LSU Doctoral Dissertations. 2582. https://digitalcommons.lsu.edu/gradschool_dissertations/2582

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU Digital Commons. For more information, please [email protected].

THE EFFECTS OF GEOMORPHOLOGY AND PRIMARY PRODUCTIVITY ON NEOTROPICAL LEAF LITTER HERPETOFAUNA: IMPLICATIONS FOR AMAZONIAN RAINFOREST CONSERVATION

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Doctor of Philosophy

The Department of Biological Sciences

by Jessica L. Deichmann B.S., Colorado State University, 2002 August 2009

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my advisor, Bruce Williamson. Despite innumerable challenges over the last few years, he has been more than generous with his time and biological expertise. Bruce, you have been a model academic father and this dissertation would have been impossible without you. Thank you for your timely responses to my many questions, for motivating ecological discussions, and for wonderful conversations about life in general. I have thoroughly enjoyed working with you and I thank you for helping me to become a better scientist.

I thank James Geaghan for his patience and guidance through several statistical consults.

I am also thankful to my other committee members who have always been helpful and responsive throughout this process - Phil Stouffer, Chris Austin, and Richard Stevens.

During my time at LSU, I have had the good fortune to meet many talented graduate students. I thank my past and present lab mates for their friendship and support, especially Jen

Cramer who taught me all the ins and outs of graduate school. She has been a wonderful friend and mentor who has been willing to lend much-appreciated advice even from afar. I also want to thank my friends for providing me with necessary distractions from work from time to time. I always looked forward to Wednesday nights and don‘t know what I would have done without you all. I want to especially thank Adriana Bravo, Cheryl Haines and Andrés Vidal Gadea who have been my partners in crime from our first weeks at LSU. For me, your friendship during our time in Baton Rouge has been priceless.

There is a long list of professors, coworkers and colleagues to whom I owe a debt of gratitude. Kelly Swing introduced me to the tropics. He is more of an inspiration to his students than he could ever imagine. Charlie Case showed me how to truly enjoy the job and life in general. Bill Duellman and Jeff Boundy shared data with me and also made invaluable

ii comments on manuscripts within this dissertation. Steven Whitfield, Tiffany Doan, Catherine

Toft, James Vonesh and Warren Allmon graciously allowed me pester them for data time and time again. Prissy Milligan and Jill Atwood have been indispensible with logistical help and I thank them for their patience and kindness. Albertina Lima was my Brazilian counterpart, without whom I could not have conducted research in Brazil.

This work would not have been possible without the help of all my field assistants and the staff at my field stations. From Tiputini, I would especially like to thank Mayer Rodriguez and Ima Nenquimo for help in the field and David and Consuelo Romo for logistical support; at

Yasuní, Amo Enomenga and Humberto Ahua in the field and Lucy Baldeón in the office; at

PDBFF, João de Deus Fragata in the field and M. Rosely C. Hipólito for logistical support; and at Nouragues, Peter Deichmann in the field and Philippe Gaucher for logistical support, field work and for teaching me to climb trees. I also must thank the various students and others who helped from time to time with the tedious work of quadrat sampling.

Generous funding was provided by the Conservation, Food and Health Foundation,

Centre National de la Recherche Scientifique, the Louisiana Governor‘s Office of Environmental

Education, Louisiana State University Department of Biological Sciences Graduate Student

Association (Biograds), and the Graduate School and Department of Biology at Louisiana State

University.

Finally, words cannot express how grateful I am to my family and friends back home. I want to thank my Mom and Dad, Sally and John Deichmann, for their unfaltering love, support and encouragement. Thank you for everything you have done and for all that you have sacrificed so that I could chase my dreams – and for understanding when those dreams seemed a little crazy at times! I also am grateful to my sister, Kerry, who has been my lifeline in Baton

Rouge, and my brother, Pete, who proved to be the best field assistant there ever was. Thank

iii you so much to my late grandparents, Bill and Maretta and Verna and Vernal, and all my aunts, uncles and cousins who have always supported me and played a major role in my life. I‘ve learned from all of you. The love and support of my family and friends has always made me feel secure and given me the courage to step outside the box. Lastly, I want to thank Ben Marks for his unwavering love and companionship, which has made an inevitably difficult process much easier to bear.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

ABSTRACT ...... x

CHAPTER 1 INTRODUCTION ...... 1 Cross-continental Comparisons ...... 1 Intra-continental Comparisons ...... 4 Focal Taxa ...... 11 Field Sites...... 12 Chapter Summary ...... 15 Literature Cited ...... 16

2 ANURAN ARTIFACTS OF PRESERVATION: 25 YEARS LATER ...... 23 Introduction ...... 23 Methods...... 24 Results ...... 25 Discussion ...... 30 Literature Cited ...... 32

3 PREDICTING BIOMASS FROM SNOUT-VENT LENGTH IN NEW WORLD FROGS ...... 35 Introduction ...... 35 Methods...... 36 Results ...... 39 Discussion ...... 45 Literature Cited ...... 50

4 EFFECTS OF GEOMORPHOLOGY ON AMAZONIAN LEAF LITTER HERPETOFAUNA ...... 53 Introduction ...... 53 Methods...... 55 Results ...... 59 Discussion ...... 62 Literature Cited ...... 70

5 NEOTROPICAL LEAF LITTER HERPETOFAUNA ACROSS A GRADIENT OF SOIL FERTILITY ...... 75 Introduction ...... 75 Methods...... 77 Results ...... 81

Discussion ...... 85 Literature Cited ...... 93

6 A NOTE ON AMPHIBIAN DECLINE IN THE NEOTROPICAL LOWLANDS ...... 97 Literature Cited ...... 102

7 CONCLUSION ...... 103 Literature Cited ...... 107

APPENDIX 1: SPECIES LIST ...... 108

APPENDIX 2: LETTER OF PERMISSION ...... 114

VITA ...... 115

LIST OF TABLES

Table 1.1 Leaf litter herpetofauna abundance data …………………………….……………2

Table 2.1 Size range of frogs measured and the difference in measurements between researchers ……………………………………………………………………….26

Table 2.2 Ranges and means of SVL1 (before preservation) and the change in SVL from the first measurement to the second measurement (SVL2-SVL1)……..……………26

Table 2.3 Pearson‘s Correlation Coefficient for the relationship between percent and numerical changes and SVL1 within species ……………………………………29

Table 2.4 SVL differences between 2 measurements taken after preservation…………….29

Table 3.2 Results of ANCOVA tests for regression differences between sexes and age classes within species.……………..……………………………………………..44

Table 4.1 Climate data for sampled sites on young (Tiputini and Yasuní) and old (BDFFP) soils…..…………………………………………………………………………..56

Table 4.2 Multiple linear regression models for the relationship between environmental variables and leaf litter herpetofauna abundance and biomass at sites on old and young soils …………………………………………...………………………….61

Table 5.1 List of sites included in this study. N represents the number of plots conducted on primary terra firme forest in the wet season at each site………..………………..78

Table 5.2 Contributions to the Χ2 values by size range (mm) and country for frogs found in litter plots………………………………………………………………………...84

Table 5.3 Contributions to the Χ2 values by size range (mm) and country for lizards found in litter plots……………………………………………………………………...84

vii

LIST OF FIGURES

Figure 1.1 Map of South America showing approximate locations of field sites (1=Tiputini Biodiversity Station and Yasuní Research Station in Ecuador, 2= Biological Dynamics of Forest Fragments Project in Brazil, 3=Nouragues Research Station in French Guiana) with insets of (A) a satellite image of the Amazon Basin and (B) the variety of landforms underlying Amazon forests…………………………5

Figure 2.1 The relationship between mean SVL of 14 species before preservation and (A) the mean percent of shrinkage and (B) the mean numerical degree of shrinkage after preservation ……………………………….……………………………………..28

Figure 3.1 Mass/SVL regression lines for (A) Hylidae, (B) Bufonidae and Ranidae, (C) Aromobatidae, Dendrobatidae, and Microhylidae, (D) Eleutherodactylidae and Leptodactylidae. ………………………………………………………..………..42

Figure 3.2 Mass versus SVL plots showing differences in males and females of (A) Scinax pedromedinae, (B) Hyla squirella, (C) Hypsiboas fasciatus, and (D) Pristimantis fenestratus…………………………………..……………………………………46

Figure 3.3 Mass versus SVL plot of individuals of the Rhinella ―margaritifera‖ complex (BuMa), demonstrating variation between geographic regions, sexes …..……...50

Figure 4.1 Abundance of (A) all secondary consumer reptiles and amphibians, (B) only amphibians and (C) only lizards and biomass of (D) all reptiles and amphibians, (E) only amphibians and (F) only lizards.……………………………………….60

Figure 4.2 Size ranges of frog species in Ecuador and Brazil: (A) males, Ecuador; (B) males, Brazil; (C) females, Ecuador; (D) females, Brazil …………………………...….63

Figure 4.3 Size range of lizard species in (A) Ecuador and (B) Brazil………………….…..64

Figure 4.4 Rarefaction curves with 95% confidence intervals for (A) all secondary consumers in leaf litter herpetofauna, (B) amphibians only and (C) lizards only encountered in quadrats in Ecuador (black lines) and Central Brazil (gray lines)……………………………………………………………………………...65

Figure 5.1 Average abundance with 95% confidence intervals of (A) frogs and (B) lizards and biomass of (C) frogs and (D) lizards in plots in South America....………….82

Figure 5.2 Size distributions of frogs captured in plots in (A) Ecuador, (B) Peru, (C) French Guiana, (D) Panama and (E) Brazil ……………………………………….…….83

Figure 5.3 Size distributions of lizards captured in plots in (A) Ecuador, (B) French Guiana, (C) Brazil……………………………..………………………………………….84

Figure 5.4 Species rarefaction curves for (A) leaf litter frogs and (B) lizards…..…………..86

viii

Figure 5.5 Dendrogram of site similarity based on cluster analysis of Bray-Curtis similarity indices.……………………………..…………………………………………….87

Figure 6.1 Average frog abundance with 95% confidence intervals within the whole of the BDFFP and at 3 sites within the BDFFP in 1985 and 2007……………………..99

Figure 6.2 Average biomass with 95% confidence intervals of leaf litter frogs within the whole of the BDFFP and at 3 sites in the BDFFP in 1985 and 2007….………...99

Figure 6.3 An overlay of the BDFFP frog density data on the data from primary forest at La Selva in Costa Rica.……………………….……………………………………100

ABSTRACT

The Amazon rainforest encompasses over one billion acres of South America and sustains remarkable biodiversity. Despite the large body of research stemming from this region, little is known about the effects of geomorphology and primary productivity on the fauna of

Amazonia, and on reptiles and amphibians in particular. In my dissertation, I examine differences in the abundance, biomass and species richness of secondary consumers in the leaf litter herpetofauna communities on young and ancient soils.

Herein, I develop methods to utilize existing data sets and museum collections in new studies involving community biomass. I found that, although the process of preservation does change the size of specimens, for most species the differences are less than 4%. Furthermore, these changes can be quantified and taken into account when applying measurements made on preserved specimens to studies of living individuals. I also derive equations which can be used to calculate mass from length in anurans. Snout-vent length (SVL) has been measured on thousands of herps, some released in the field and many housed in research collections.

Estimated using SVL measurements, mass of individuals can now be used to determine community biomass of populations that were initially sampled for other purposes.

In the second half of this dissertation, I assess the effects of geomorphology on leaf litter herpetofauna in Neotropical lowland rainforests. At sites with similar latitude, elevation and climate, litter herpetofauna abundance, biomass and species richness are twice as high on younger soils. Using methodology developed in the first chapters, this comparison is expanded to include forests differing in latitude and climate. In this study, the trend of increased density, biomass and diversity on younger soils holds, lending further support to the hypothesis that geomorphology and primary productivity drive leaf litter herpetofauna community dynamics.

These studies taken together provide the basis for a change in Amazonian management strategies. The prevailing notion that a few large reserves within the Amazon will be able to sustain Amazonian biodiversity is unfounded. Reserves and conservation policy must be designed around the local geologic history and forest dynamics of forest regions within the

Basin.

CHAPTER 1 INTRODUCTION

The Amazon Basin is a vast expanse of tropical lowland rainforest that is home to over

1000 species of amphibians and reptiles. In my dissertation, I compare frog and lizard communities at sites within the Basin that differ greatly in primary productivity as a result of soil age. My primary goal is to understand the role that geomorphology plays in structuring leaf litter herpetofauna communities in Neotropical forests so that more precise conservation management practices can be developed and implemented in tropical rainforests, particularly within

Amazonia.

CROSS-CONTINENTAL COMPARISONS

The tropical humid forest biome, found on five different continents across the globe, exhibits enormous differences in ecosystem structure and function (Williamson et al. 2005).

Given the complexity of tropical rainforests and their varied historical processes, it is no surprise that rainforest dynamics differ between and even within a single continent.

Over the last 40 years, ground-dwelling herpetofauna have been studied extensively in

Central America and Southeast Asia. Although litter frogs and lizards in the tropical lowland wet forests of both these regions encompass very similar ecological roles, using litter as a microhabitat and feeding on litter invertebrates (Inger 1980; Whitfield & Donnelly 2006), abundances of these animals at various sites throughout Central America are an order of magnitude greater than at sites surveyed in Southeast Asia (Table 1.1; Inger 1980; Scott 1976).

A number of hypotheses have been put forth to try to explain the discrepancies in litter herpetofauna densities between these regions. One hypothesis suggests there may be a relative increase in the number of predators or competitors of the secondary consumers in the litter herpetofauna community in Southeast Asia, which would lead to a decrease in the abundance of

Table 1.1 Leaf litter herpetofauna abundance data (from Allmon 1991, Inger 1980, Toft 1980 and Scott 1976) Rainfall Frogs Lizards Locality (mm/year) (individuals/100m2) (individuals/100m2)

Borneo (Nanga Tekalit) 6000 1.31 0.25

Borneo (Sarawak) 5500 1.09 n/a

Thailand 1500 0.12 1.03

Costa Rica (La Selva) 3600 14.70 2.80 Costa Rica (Osa) 4000 11.60 3.90 Peru (Rio Llullapichis) 2400 9.96 n/a

Central Brazil (BDFFP) 2500 4.76 0.10

frogs and lizards (Scott 1976). However, data from Borneo and Costa Rica suggest that snake abundances are closely correlated with the abundance of the rest of the litter herpetofauna (Scott

1976). As snakes are the most important predators of litter frogs and lizards (Duellman & Trueb

1986; Scott 1976), it is unlikely that predators are causing the incongruity in abundance between

Central America and Southeast Asia. Data on potential competitors such as birds, spiders and other predaceous arthropods are limited; however, anecdotal accounts suggest that differences in the densities of these organisms are also unlikely to account for the discrepancies in herpetofauna abundance (Scott 1976).

A second hypothesis suggests that leaf litter depth should affect the number of secondary consumers within the community. Leaf litter provides ground dwelling herpetofauna with shelter from predation as well as a site for reproduction. Because litter is not a homogeneous habitat, more litter could potentially provide additional spatial structure, allowing for higher densities of the animals that utilize this habitat (Scott 1976). Additionally, more leaf litter with rapid decomposition sustains a multitude of small arthropods, and therefore, should result in an

2 increase in the number of animals that depend on these decomposers as a source of energy

(Janzen 1974). If this hypothesis were true, we would expect to see increased litter fall and decomposition rates in Central America relative to Southeast Asia. Most accounts, however, suggest that no such differences in litter fall exist between these regions (Inger 1980; Kira et al.

1967; Ogawa 1978) although there are exceptions (Bray & Gorham 1964).

A third potential explanation is that the abundance differences may be caused by the different mechanisms of fruiting seen in Central America and Southeast Asia (Inger 1980). An increased amount of fruit in mast years in Southeast Asian forests should lead to a temporary increase in the abundance of seed predators and decomposers (Curran & Leighton 2000; Curran

& Webb 2000; Janzen 1974) as well as in consumers at other levels of the food chain. However, near starvation conditions during the intervening non-mast years causes dramatic reductions in animal densities at all trophic levels. Though mast fruiting occurs in some parts of the

Neotropics, the scale at which it occurs in the Dipterocarpaceae forests of Southeast Asia far exceeds that of Central and South America (Ashton et al. 1988; Brearley et al. 2007; Norden et al. 2007; Wright et al. 1999).

Supra-annual mast fruiting may explain abundance differences between Central America and Southeast Asia, but further studies encompassing other geographical areas show similar differences in abundances within continents as well as among them (Table 1.1). Allmon (1991) showed that the anuran abundance pattern exhibited in Central America varied from previously studied Neotropical sites. Whereas anuran abundances in the Peruvian Amazon are comparable to those in Central America (Allmon 1991; Toft 1980), Allmon (1991) found low litter frog abundances in the Brazilian Amazon, with densities comparable to Borneo. He speculated that these differences could be related to the geological age of the soils in different parts of the

Amazon. Unfortunately, the compared sites differ in numerous confounding factors that

3 preclude isolating the effect of soil age. An alternative approach would require examination of litter communities in tropical forests sharing similar climate and elevation, but differing in the geologic age of the soil.

INTRA-CONTINENTAL COMPARISONS

Plant Productivity Differences

It is well established that primary plant productivity varies widely between different habitats, but what drives these differences? The precise mechanism influencing productivity may vary depending on the ecosystems being compared. For example, across habitats there is strong evidence that precipitation influences productivity, with areas of increased precipitation generally being more productive (Kay et al. 1997). Productivity in the tropics has also been shown to change with elevation, first increasing to mid-elevations and then decreasing at higher elevations. The levels of photosynthetically active radiation acquired by mid-elevation plants is apparently comparable to levels accrued in low elevation forests, but lower night time temperatures at mid-elevations produce lower respiration demands, resulting in more conserved energy that can be allocated to growth (Janzen 1973). Primary productivity has long been recognized as having the potential to influence differences in faunal communities between tropical regions, although little has been done to provide evidence for this link. Among leaf litter herpetofauna, evidence exists to support both these drivers of productivity, with higher abundances of amphibians found at wetter sites and at mid-elevations (Scott 1976). However, rainfall and elevation cannot explain herpetofauna community differences across regions with similar rainfall and elevation. Large areas of continuous forest, such as the Amazon Basin, have generally been overlooked when exploring differences in litter herpetofauna communities.

How alike are the rainforests across the Amazon Basin? The perception of uniformity of the Amazon may stem from its relatively low elevation and a lack of barriers to dispersal;

6 however, this apparent homogeneity undermines striking differences in geology and forest types within the Basin (Tuomisto et al. 1995; Figure 1.1). Obviously, flooded forests, such as igapo and varzea, differ from upland (terra firme) forests. Even among upland forests across the

Basin, differences in soil age have profound effects on nutrient levels. Young soils, such as those found in the Ecuadorian Amazon, originated and were deposited from material produced during the rapid uplift of the Andes which began in the Cenozoic era and continue to this day

(Gregory-Wodzicki 2000). On the other hand, soils in Central Amazonia are derived from the

Brazilian Highlands and the Guiana Shield which originated in the Paleozoic and continued developing into the Mesozoic (Sombroek 2000). Age-associated weathering and mineralization have caused leaching of soil nutrients. The vast difference in bedrock age and weathered sediments between the west and central/eastern regions of Amazonia sets the stage for varied ecosystem dynamics.

Although the Amazon Basin contains the largest tract of intact rainforest on the planet, relatively little effort has been devoted to discriminating forest types (ter Steege et al. 2000;

Terborgh & Andresen 1998); however, recently there has been a surge in our knowledge of differences in forest dynamics within Amazonia. Variations in tree communities across

Amazonia include differences in abundance, composition and turnover and many of these discrepancies are closely linked to soils and geomorphology. For example, soil nutrient concentrations are closely correlated with the abundance of many tree species (Bohlman et al.

2008) and can account for a third of the variation in above ground biomass of trees in Central

Amazonia (Laurance et al. 1999). On younger, more fertile soils, course woody productivity is greater (Malhi et al. 2004) and tree turnover is twice that on older, more weathered soils (Phillips et al. 2004). Furthermore, tree species composition within the Amazon Basin is at least partially dependent on soil fertility (ter Steege et al. 2006; Terborgh & Andresen 1998). On a genetic

7 level, high diversity in tropical tree species has been suggested to be a result of high diversity in soil types (Fine et al. 2005; Korning et al. 1994; Sabatier et al. 1997).

There is clearly a strong link between edaphic conditions and forest productivity. Are soils as tightly linked to animal community dynamics? In my dissertation, I explore the effects of primary productivity on leaf litter herpetofauna under the hypothesis that the differences in soil age that influence primary production in the Amazon Basin will reverberate up the food chain, perhaps in a multiplicative fashion. The net result at higher trophic levels may be much greater animal biomass and perhaps species richness on younger, more productive soils.

Animal Differences

As previously mentioned, herpetofauna density is known to vary in rainforests across the globe with litter frog abundances in Central America and Peru being significantly greater than in

Southeast Asia and the Brazilian Amazon. These results are strongly suggestive that geologic age plays a crucial role in herpetofauna biomass, but for all these studies, the authors recorded numbers of individuals, not biomass, and each study employed its own sampling methodology.

Trends of increased densities on younger soils in the Amazon Basin have been shown in other taxa as well. Non-flying mammal abundance is positively correlated with soil fertility, with mammals attaining higher densities and species richness in Western Amazonia (Emmons

1984). Euglossine bees are captured at much lower rates in Central Amazonia than at sites in

Central America and Peru (Becker et al. 1991). Terrestrial insectivorous birds occur in lower densities in Central Brazil than in French Guiana or Peru (Stouffer 2007), and primates have higher biomass densities in Western than in Central Amazonia (Peres & Dolman 2000).

However, the sites compared in these studies differed in a number of confounding factors such as disturbance, fragmentation and/or hunting pressure. In my dissertation research, I used standardized measurements of litter herpetofauna biomass with a common methodology in

8 undisturbed primary forest tracts of at least 10,000 ha that share a comparable climate; the sites differing primarily in landform, as defined by Sombroek (2000).

Species Richness

If soils and productivity have an effect on biomass, do they also affect species richness?

Productivity has been shown to influence the number of species occurring in an area, though the mechanism propagating this pattern is not completely clear (Rosenzweig 1995). The species- energy theory contends that increased available energy can support more individuals and consequently more species in a given area (Brown 1981; Wright 1983). Although the species- energy theory may not explain variations in richness of all taxa (Latham & Ricklefs 1993), the relationship is generally strong for many species, across latitudes and climates (Hawkins et al.

2003). There is some evidence of the species-energy theory at play in the Neotropics: species richness of primates, particularly primary consumers, is highest in areas with the highest productivity (Kay et al. 1997).

One route by which energy directly influences species richness is through population sizes (Evans et al. 2005). Limited available energy in a system can limit resources available to consumers, thereby reducing both biomass and density of animals. Diminished population sizes lead to increased extinction risk (Pimm et al. 1988; Soulé et al. 1988) through reduction of populations below a critical minimum threshold (Shaffer 1981). It is possible that a low biomass of frogs and lizards in the Central Amazon could drive population densities so low that some species are unable to maintain reproductively viable populations, while higher densities in

Western Amazonia allow for the existence of more species.

Conservation Issues

Today, as deforestation throughout South America increases due to demand for lumber, oil, soybeans, and urbanization, developing efficient management techniques in the Amazon

Basin is especially important. There are many processes of environmental destruction in the

Amazon; most often cited are oil exploration, mining for gold, bauxite and other minerals, logging and agriculture. There are also more subtle, but equally damaging activities including illegal hunting and anthropogenic fires. This dissertation provides support for practical and responsible conservation management of litter herpetofauna in the Amazon Basin, the principles behind which may be applicable to all organisms within the ecosystem.

Many conservation organizations have taken steps over the last few years to address the threats facing intact ecosystems. For example, the International Union for Conservation of

Nature (IUCN), Secretariat of the Convention on Biological Diversity (SCBD) and others have recommended that at least 10% of each ecological region, or biome, be set aside for preservation

(SCBD 2004; IUCN 1993). Although generalizations are arguably necessary as first steps in conservation, a ‗biome‘ is an exceedingly broad classification. If we base conservation decisions on a ‗biome‘ scale, many microhabitats and their constituents will inevitably be excluded from protection. The research herein will lend support to the view that the tropical humid forest biome must be divided at a much finer scale when it comes to distributing that 10%. Furthermore, if in fact older soils support less biomass (and hence less energy) than younger soils, then proportionally larger tracts of land will need to be set aside on older soils in order to maintain comparable densities of animals.

Protected areas have become our ―cultural response‖ to the rapid loss of Earth‘s biodiversity (Chape et al. 2005) and are frequently established out of convenience which may make them inefficient and their utility as conservation tools questionable in these instances

(Bates & Rudel 2000). Often, large reserves founded in times of prosperity become too much of a burden for a country to manage and protect as time goes on (Bates & Rudel 2000). In this case, it may be better to initially create smaller reserves which would require less money to maintain

10 over time, but would still be sufficient to sustain viable populations. My research is a step toward understanding how primary productivity based on geomorphology affects animal biomass, thereby leading to efficient reserve design and more resourceful use of funds for conservation.

In addition, this research involves another important conservation issue, amphibian decline. Roughly 30% of all amphibian species on the globe are threatened with extinction. This number may be low given inadequate data for another 20%, mostly tropical species (IUCN et al.

2008; Stuart et al. 2004). The causes behind the decline are continually being debated (Collins &

Storfer 2003; Lips et al. 2005b; Stuart et al. 2004). Recent papers on global amphibian decline have highlighted the need for inventories of existing populations of frogs as well as repeated surveys of areas censused in the past (Eterovick et al. 2005; Lips et al. 2005a; Stuart et al. 2004).

Recently, Whitfield et al. (2007) have shown that amphibian decline in Central America is not restricted to montane habitats as previously believed, and that both lizards and frogs are declining. Here, I resurvey for frogs and lizards at a forest site in Brazil that was censused in

1983-84, documenting any changes potentially related to amphibian decline. Data within this dissertation also provide baseline numbers for sites not previously surveyed for leaf litter herpetofauna.

FOCAL TAXA

In my dissertation, I focus on the secondary consumers in the leaf litter herpetofauna community in the lowland forests of South America. These amphibians and reptiles serve as model organisms for the study of the effects of soil fertility and forest productivity because they constitute a cornerstone of the ecological community. Representing both predator and prey within an ecosystem, they are directly tied to the survival of taxa at multiple trophic levels.

Among litter frogs and lizards, we expect that some species may exhibit biomass differences between the Western and Central Amazon more than others, given the variety of life

11 histories among taxa. While lizards spend their entire life cycle in the leaf litter, often as secondary consumers, many tropical frogs, which have more than one life stage, do not. There are 21 different modes of reproduction used by Neotropical amphibians (Duellman & Trueb

1986). These range from placement of eggs in phytotelmata such as tree cavities, tank bromeliads or fruit capsules, to creating terrestrial foam nests with non-feeding tadpoles. Some species even use arboreal or terrestrial nests where direct development of eggs occurs in situ

(Caldwell 1993; Crump 1982; Duellman & Trueb 1986; Hödl 1990; Zimmerman & Simberloff

1996). During this larval period, most frogs are herbivorous, eating primarily algae, while some are detrivorous, carnivorous (even cannibalistic) or oophagous (Caldwell & Carmozina de

Araujo 1998; Diaz Paniagua 1985; Dutra & Callisto 2005; Jungfer & Weygoldt 1999). I expect that lizards, which are dependent on litter arthropods for a greater portion of their life cycle, will exhibit greater changes in biomass and biodiversity than most amphibians which practice prolonged herbivory during development.

FIELD SITES

To directly test my hypotheses, I considered upland forests in Central and Western

Amazonia and selected five sites, based on Sombroek's (2000, 2001) classification of landforms in the Amazon Basin. Tiputini Biodiversity Station and Yasuní Biological Station (Ecuador) are located in the young Amazon on the "Western Sedimentary Uplands". Cabo Frio, Dimona and

Km 41 are three continuous forest plots located within the Biological Dynamics of Forest

Fragments Project (BDFFP) (Brazil) and are in the old Amazon on the "Eastern Sedimentary

Uplands" (Sombroek 2000). These 5 sites were selected for this study because they lack a strong dry season, have adequate rainfall and similar elevation, latitude, and soil texture, but differ in their geological age. ―Western Sedimentary Uplands‖ (WSU) differ distinctly from the ―Eastern

Sedimentary Uplands‖ (ESU) in that the soils were less pre-weathered at the time of deposition,

12 the soils have higher ion-exchange capacities, and there remain weatherable minerals within the soil (Sombroek 2000). WSU are composed of ferric or haplic Acrisols and Alisols while the

ESU are characterized by xanthic or geric Ferralsols with some ferralic Arenosols (Sombroek

2000).

While I use Sombroek‘s landform classification which is based on the FAO/UNESCO soil taxonomic scheme for defining general soil types within the Amazon Basin, I will use the

USDA soil taxonomy classification scheme when describing the soils of my specific field sites throughout this dissertation. The USDA system includes a breakdown of soil groups and allows for soil descriptions at a finer resolution (Juo & Franzluebbers 2003).

Tiputini Biodiversity Station (TBS) and Yasuní Research Station (YRS) are located in lowland rainforest in Orellana province, Ecuador within the 1.5 million ha Yasuní Biosphere

Reserve. Tiputini Biodiversity Station (0°37‘S, 76°10‘W; 190–270m asl) was established in

1994 by the Universidad San Francisco de Quito. Yasuní Research Station (0°40'S, 76°24'W;

200-260m asl) was founded in 1994 by the Ecuadorian Government and is currently managed by the Pontificia Universidad Católica de Ecuador. Soils underlying TBS and YRS are largely udult

Ultisols and there is strong evidence of volcanic activity on soil formation of the region (Korning et al. 1994; Valencia et al. 2004). Average rainfall is approximately 2700-3000mm/year.

Although no month has less than 100mm precipitation, there is a 5-month drier period from

October to February (Karubian et al. 2005). Species richness among trees >10cm DBH is approximately 300/ha (Valencia et al. 2004).

The Biological Dynamics of Forest Fragments Project (BDFFP), initiated in 1979 as the

Minimum Critical Size of Ecosystems Project, is jointly run by Brazil‘s Institute for Amazonian

Research (INPA) and the Smithsonian Tropical Research Institute (STRI). The BDFFP (2°30'S,

60°W; 90-160m asl) is located in Amazonas state, Brazil, 80km north of Manaus, within an area

13 of approximately 500,000ha of relatively undisturbed, terra firme lowland rainforest (Lovejoy &

Bierregaard 1990). The BDFFP is composed of 3 main areas, or fazendas. Each fazenda,

Dimona, Esteio and Porto Alegre, has a number of reserves within. The soils of the BDFFP are predominately well-developed Oxisols (Laurance et al. 1999). Annual precipitation is between

2200-2700mm (Gascon & Bierregaard 2001; Radtke et al. 2007). The dry season generally lasts from June through October (Gascon & Bierregaard 2001), although precipitation at the BDFFP does not normally fall below 100mm in any month (Radtke et al. 2007). Tree species richness

(>10cm DBH) is approximately 285/ha (de Oliveira & Mori 1999).

These sites are ideal for this study because they share similar latitude, elevation and climate regimes, but have had different geologic histories. Their similarities will allow me to test specifically for differences in abundance, biomass and species richness of leaf litter herpetofauna due to soil age.

For the final chapter of this dissertation, I sampled 2 additional sites: Pararé (4°02'N,

52°41'W) and Inselberg (4°05'N, 52°41'W), both part of the Nouragues Biological Station in central French Guiana. These sites are not included in the primary comparison (Chapter 4) because they differ from the Ecuadorian and Brazilian sites in seasonality and latitude.

However, their addition lends an interesting comparison to the broader investigation of soil age

(Chapter 5) because Nouragues is located on soils which are intermediate to young Western and ancient Central Amazonian soils. Central French Guiana lies on soils classified by Sombroek

(2000) as ―Crystalline Shield Uplands‖ (CSU). In general, they are part of the ancient soils of the

Amazon Basin, although the crystalline soils differ somewhat from sedimentary soils in harboring more local variation in soil types [Cambisols, Acrisols and Ferralsols (USDA Soil

Taxonomy: Inceptisols, Ultisols and Oxisols)].

Nouragues Biological Station was officially established in 1986 with the help of the

French Environment Ministry and is currently managed by the Centre National de la Recherche

Scientifique (CNRS). The Nouragues area was declared a reserve in 1995 and includes over

100,000ha of continuous tropical rainforest. The soils of Nouragues are heterogeneous with rapid transitions between Inceptisols and Ferrasols (Grimaldi & Riéra 2001). The cation exchange capacity of the soils is intermediate to those at sites in Ecuador and Brazil (Fearnside

& Filho 2001; Grimaldi & Riéra 2001; Tuomisto et al. 2003). Annual precipitation at Nouragues is 2990mm with a distinct rainy season from December to July. May is generally the wettest month and September and October the driest, with less than 100mm rainfall (Grimaldi & Riéra

2001). Tree species richness at Nouragues reaches 260 species/ha (Chave et al. 2001).

CHAPTER SUMMARY

In the first part of my dissertation, I investigate the potential for using length measurements to estimate anuran biomass. In Chapter 2, I look at changes in frog morphology due to preservation. Chapter 3 focuses on the use of snout-vent length (SVL) measurements of frogs to estimate mass. Specifically, I develop precise regression equations for 36 anuran species. Museum collections potentially hold enormous stores of data on community biomass. If morphometrics are not altered by the preservation and storage process, and SVL can be used to accurately predict mass, then measurements made on preserved specimens, along with precise collection records, could be used to estimate biomass of past populations in all regions that have been previously studied.

The second part of my dissertation aims to answer the ecological question – how does soil age through productivity affect leaf litter herpetofauna? In Chapter 4, I use climactically similar sites to investigate the effects of geomorphology on leaf litter herpetofauna communities in the Amazon Basin. Sites are sampled using a standard methodology. Chapter 5 broadens the

15 comparison of leaf litter herpetofauna to lowland forest sites scattered throughout South

America. I use my own data along with existing data sets collected using quadrats to examine to what degree the pattern of increased biomass, abundance and species richness on young soils extends to other tropical lowland rainforests.

Existing quadrat data have proved useful for regional comparisons, but they can be used to explore other types of questions as well. As a byproduct of testing my productivity hypotheses, I was able to accumulate leaf litter herpetofauna data collected using quadrats at the

Biological Dynamics of Forest Fragments Project (BDFFP) in 1984-85. In my final chapter, I approach the question of amphibian decline through an investigation of changes in community structure of leaf litter herpetofauna over a 22 year period.

LITERATURE CITED

Allmon, W. D. 1991. A plot study of forest floor litter frogs, Central Amazon, Brazil. Journal of Tropical Ecology 7:503-522.

Ashton, P. S., T. J. Givnish, and S. Appanah. 1988. Staggered flowering in the Dipterocarpaceae: new insights into floral induction and the evolution of mast fruiting in the aseasonal tropics. The American Naturalist 132:44-66.

Bates, D., and T. K. Rudel. 2000. The political ecology of conserving tropical rain forests: A cross-national analysis. Society & Natural Resources 13:619-634.

Becker, P., J. S. Moure, and F. J. A. Peralta. 1991. More about Euglossine bees in Amazonian forest fragments. Biotropica 23:586-591.

Bohlman, S. A., W. F. Laurance, S. G. Laurance, H. E. M. Nascimento, P. M. Fearnside, and A. Andrade. 2008. Importance of soils, topography and geographic distance in structuring central Amazonian tree communities. Journal of Vegetation Science 19:863-874.

Bray, J. R., and E. Gorham. 1964. Litter production in the forests of the world. Pages 101-157 in J. B. Cragg, editor. Advances in Ecological Research.

Brearley, F. Q., J. Proctor, L. Nagy, G. Dalrymple, and B. C. Voysey. 2007. Reproductive phenology over a 10-year period in a lowland evergreen rain forest of central Borneo. Journal of Ecology 95:828-839.

Brown, J. H. 1981. Two decades of homage to Santa Rosalia: toward a general theory of diversity. American Zoologist 21:877-888.

Caldwell, J. P. 1993. Brazil nut fruit capsules as phytotelmata: interactions among anuran and insect larvae. Canadian Journal of Zoology 71:1193-1201.

Caldwell, J. P., and M. Carmozina de Araujo. 1998. Cannibalistic interactions resulting from indiscriminate predatory behavior in tadpoles of poison frogs (Anura: Dendrobatidae). Biotropica 30:92-103.

Chape, S., J. Harrison, M. Spalding, and I. Lysenko. 2005. Measuring the extent and effectiveness of protected areas as an indicator for meeting global biodiversity targets. Philosophical Transactions of the Royal Society B-Biological Sciences 360:443-455.

Chave, J., B. Riera, and M.-A. Dubois. 2001. Estimation of biomass in a Neotropical forest of French Guiana: spatial and temporal variability. Journal of Tropical Ecology 17:79-96.

Collins, J. P., and A. Storfer. 2003. Global amphibian declines: sorting the hypotheses. Diversity and Distributions 9:89-98.

Crump, M. L. 1982. Amphibian reproductive ecology on the community level. Pages 21-35 in N. J. Scott, editor. Herpetological Communities: A symposium of the Society for the Study of Amphibians and Reptiles and the Herpetologists' League, August 1977. United States Department of the Interior, Fish and Wildlife Service, Washington, D.C.

Curran, L. M., and M. Leighton. 2000. Vertebrate responses to spatiotemporal variation in seed production of mast-fruiting Dipterocarpaceae. Ecological Monographs 70:101-128.

Curran, L. M., and C. O. Webb. 2000. Experimental tests of the spatiotemporal scale of seed predation in mast-fruiting Dipterocarpaceae. Ecological Monographs 70:129-148. de Oliveira, A. A., and S. A. Mori. 1999. A central Amazonian terra firme forest. I. High tree species richness on poor soils. Biodiversity and Conservation 8:1219-1244.

Diaz Paniagua, C. 1985. Larval diets related to morphological characters of five anuran species in the Biological Reserve of Doñana (Huelva, Spain). Amphibia-Reptilia 6:307-322.

Secretariat of the Convention on Biological Diversity 2004. Decisions Adopted by the Conference of the Parties to the Convention on Biological Diversity at its Seventh Meeting. SCBD, Montreal.

Duellman, W. E., and L. Trueb 1986. Biology of Amphibians. McGraw-Hill Book Company, New York.

Dutra, S. L., and M. Callisto. 2005. Macroinvertebrates as tadpole food: importance and body size relationships. Revista Brasileira de Zoologia 22:923-927.

Emmons, L. H. 1984. Gographic variation in densities and diversities of non--flying mammals in Amazonia. Biotropica 16:210-222.

Eterovick, P. C., A. C. Oliveira de Queiroz Carnaval, D. M. Borges-Nojosa, D. L. Silvano, M. V. Segalla, and I. Sazima. 2005. Amphibian declines in Brazil: an overview. Biotropica 37:166-179.

Evans, K. L., P. H. Warren, and K. J. Gaston. 2005. Species-energy relationships at the macroecological scale: a review of the mechanisms. Biological Reviews 80:1-25.

Fearnside, P. M., and N. L. Filho. 2001. Soil and development in Amazonia: Lessons from the Biological Dynamics of Forest Fragments Project. Pages 291-312 in R. O. J. Bierregaard, C. Gascon, T. E. Lovejoy, and R. Mesquita, editors. Lessons from Amazonia: The Ecology and Conservation of a Fragmented Forest. Yale University Press, New Haven.

Fine, P. V. A., D. C. Daly, G. Villa Munoz, I. Mesones, K. M. Cameron, and S. Kalisz. 2005. The contribution of edaphic heterogeneity to the evolution and diversity of Burseraceae trees in the Western Amazon. Evolution 59:1464-1478.

Gascon, C., and R. O. J. Bierregaard. 2001. The Biological Dynamics of Forest Fragmentation Project: The study site, experimental design, and research activity in R. O. J. Bierregaard, C. Gascon, T. E. Lovejoy, and R. Mesquita, editors. Lessons from Amazonia. Yale University Press, New Haven.

Gregory-Wodzicki, K. M. 2000. Uplift history of the Central and Northern Andes: A review. Geological Society of America Bulletin 112:1091-1105.

Grimaldi, M., and B. Riéra. 2001. Geography and climate. Pages 9-18 in F. Bongers, P. Charles- Dominique, P.-M. Forget, and M. Théry, editors. Nouragues: Dynamics and Plant- Animal Interactions in a Neotropical Rainforest. Kluwer Academic Publishers, Boston.

Hawkins, B. A., R. Field, H. V. Cornell, D. J. Currie, J.-F. Guegan, D. M. Kaufman, J. T. Kerr, G. G. Mittelbach, T. Oberdorff, E. M. O'Brien, E. E. Porter, and J. R. G. Turner. 2003. Energy, water, and broad-scale geographic patterns of species richness. Ecology 84:3105- 3117.

Hödl, W. 1990. Reproductive diversity in Amazonian lowland frogs. Fortschritte der Zoologie 38:41-60.

Inger, R. F. 1980. Densities of floor-dwelling frogs and lizards in lowland forests of Southeast Asia and Central America. The American Naturalist 115:761-770.

IUCN, Conservation International, and NatureServe. 2008. An Analysis of Amphibians on the 2008 IUCN Red List in D. o. x. M. 2009., editor.

IUCN and The World Conservation Union 1993. Parks for Life: Report of the IVth World Congress on National Parks and Protected Areas. IUCN, Gland.

Janzen, D. H. 1973. Sweep samples of tropical foliage insects: effects of seasons, vegetation types, elevation, time of day, and insularity. Ecology 54:687-708.

Janzen, D. H. 1974. Tropical blackwater rivers, animals, and mast fruiting by Dipterocarpaceae. Biotropica 6:69-103.

Jungfer, K.-H., and P. Weygoldt. 1999. Biparental care in the tadpole-feeding Amazonian treefrog Osteocephalus oophagus. Amphibia-Reptilia 20:235-249.

Juo, A. S. R., and K. Franzluebbers 2003. Tropical Soils. Properties and Management for Sustainable Agriculture. Oxford University Press, New York.

Karubian, J., J. Fabara, D. Yunes, J. P. Jorgenson, D. Romo, and T. B. Smith. 2005. Temporal and spatial patterns of macaw abundance in the Ecuadorian Amazon. The Condor 107:617-626.

Kay, R. F., R. H. Madden, C. Van Schaik, and D. Higdon. 1997. Primate species richness is determined by plant productivity: Implications for conservation. Proceedings of the National Academy of Sciences 94:13023-13027.

Kira, T., H. Ogawa, K. Yoda, and K. Ogino. 1967. Comparitive ecological studies on three main types of forest vegetation in Thailand. IV. Dry matter production with special reference to the Khao Chong rain forest. Natural Life Southeast Asia 5:149-174.

Korning, J., K. Thomsen, K. Dalsgaard, and P. Nornberg. 1994. Characters of three Udults and their relevance to the composition and structure of virgin rain forest of Amazonian Ecuador. Geoderma 63:145-164.

Latham, R. E., and R. E. Ricklefs. 1993. Global patterns of tree species richness in moist forests: energy-diversity theory does not account for variation in species richness. Oikos 67:325- 333.

Laurance, W. F., P. M. Fearnside, S. G. Laurance, P. Delamonica, T. E. Lovejoy, J. M. Rankin- de Merona, J. Q. Chambers, and C. Gascon. 1999. Relationship between soils and Amazon forest biomass: a landscape-scale study. Forest Ecology and Management 118:127-138.

Lips, K. R., P. A. Burrowes, J. R. Mendelson, and G. Parra-Olea. 2005a. Amphibian declines in Latin America: Widespread population declines, extinctions, and impacts. Biotropica 37:163-165.

Lips, K. R., P. A. Burrowes, J. R. Mendelson, and G. Parra-Olea. 2005b. Amphibian population declines in Latin America: A synthesis. Biotropica 37:222-226.

Lovejoy, T. E., and R. O. J. Bierregaard. 1990. Central Amazonian forests and the Minimum Critical Size of Ecosystems project. Pages 60-71 in A. H. Gentry, editor. Four Neotropical Rainforests. Yale University Press, New Haven.

Malhi, Y., T. R. Baker, O. L. Phillips, S. Almeida, E. Alvarez, L. Arroyo, J. Chave, C. I. Czimczik, A. DiFiore, N. Higuchi, T. J. Killeen, S. G. Laurance, W. F. Laurance, S. L. Lewis, L. M. M. Montoya, A. Monteagudo, D. A. Neill, P. N. Vargas, S. Patino, N. C. A. Pitman, C. A. Quesada, R. Salomao, J. N. M. Silva, A. T. Lezama, R. V. Martinez, J. Terborgh, B. Vinceti, and J. Lloyd. 2004. The above-ground coarse wood productivity of 104 Neotropical forest plots. Global Change Biology 10:563-591.

Norden, N., J. Chave, P. Belbenoit, A. Caubère, P. Châtelet, P.-M. Forget, and C. Thébaud. 2007. Mast fruiting is a frequent strategy in woody species of eastern South America. PLoS ONE 2:e1079.

Ogawa, H. 1978. Litter production and carbon cycling in Pasoh forest. Malaysian Nature Journal 30:367-373.

Peres, C. A., and P. M. Dolman. 2000. Density compensation in neotropical primate communities: evidence from 56 hunted and nonhunted Amazonian forests of varying productivity. Oecologia 122:175-189.

Phillips, O. L., T. R. Baker, L. Arroyo, N. Higuchi, T. J. Killeen, W. F. Laurance, S. L. Lewis, J. Lloyd, Y. Malhi, A. Monteagudo, D. A. Neill, P. Núñez Vargas, J. N. M. Silva, J. Terborgh, R. Vásquez Martínez, M. Alexiades, S. Almeida, S. Brown, J. Chave, J. A. Comiskey, C. I. Czimczik, A. Di Fiore, T. Erwin, C. Kuebler, S. G. Laurance, H. E. M. Nascimento, J. Olivier, W. Palacios, S. Patiño, N. C. A. Pitman, C. A. Quesada, M. Saldias, A. Torres Lezama, and B. Vinceti. 2004. Pattern and process in Amazon tree turnover, 1976-2001. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 359:381-407.

Pimm, S. L., H. L. Jones, and J. Diamond. 1988. On the risk of extinction. The American Naturalist 132:757-785.

Radtke, M. G., C. R. V. da Fonseca, and G. B. Williamson. 2007. The old and young Amazon: dung beetle biomass, abundance and species diversity. Biotropica 39:725-730.

Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge.

Sabatier, D., M. Grimaldi, M.-F. Prévost, J. Guillaume, M. Godron, M. Dosso, and P. Curmi. 1997. The influence of soil cover organization on the floristic and structural heterogeneity of a Guianan rain forest. Plant Ecology 131:81-108.

Scott, N. J. 1976. The abundance and diversity of the herpetofaunas of tropical forest litter. Biotropica 8:41-58.

Shaffer, M. L. 1981. Minimum population sizes for species conservation. BioScience 31:131- 134.

Sombroek, W. 2000. Amazon landforms and soils in relation to biological diversity. Acta Amazonica 30:81-100.

Sombroek, W. 2001. Spatial and temporal patterns of Amazon rainfall - consequences for the planning of agricultural occupation and the protection of primary forests. Ambio 30:388- 396.

Soulé, M. E., D. T. Bolger, C. A. Allison, J. Wright, M. Sorice, and S. Hill. 1988. Reconstructed dynamics of rapid extinctions of chaparral-requiring birds in urban habitat islands. Conservation Biology 2:75-92.

Stouffer, P. C. 2007. Density, territory size and long-term spatial dynamics of a guild of terrestrial insectivorous birds near Manaus, Brazil. The Auk 124:291-306.

Stuart, S. N., J. S. Chanson, N. A. Cox, B. E. Young, A. S. L. Rodrigues, D. L. Fischman, and R. W. Waller. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306:1783-1786. ter Steege, H., N. C. A. Pitman, O. L. Phillips, J. Chave, D. Sabatier, A. Duque, J.-F. Molino, M.-F. Prévost, R. Spichiger, H. Castellanos, P. v. Hildebrand, and R. Vásquez. 2006. Continental-scale patterns of canopy tree composition and function across Amazonia. Nature 443:444-447. ter Steege, H., D. Sabatier, H. Castellanos, T. Van Andel, J. Duivenvoorden, A. A. De Oliveira, R. Ek, R. Lilwah, P. Maas, and S. Mori. 2000. An analysis of the floristic composition and diversity of Amazonian forests including those of the Guiana Shield. Journal of Tropical Ecology 16:801-828.

Terborgh, J., and E. Andresen. 1998. The composition of Amazonian forests: patterns at local and regional scales. Journal of Tropical Ecology 14:645-664.

Toft, C. A. 1980. Feeding ecology of thirteen sympatric species of anurans in a seasonal tropical environment. Oecologia 45:131-141.

Tuomisto, H., A. D. Poulsen, K. Ruokolainen, R. C. Moran, C. Quintana, J. Celi, and G. Canas. 2003. Linking floristic patterns with soil heterogeneity and satellite imagery in Ecuadorian Amazonia. Ecological Applications 13:352-371.

Tuomisto, H., K. Ruokolainen, R. Kalliola, A. Linna, W. Danjoy, and Z. Rodriguez. 1995. Dissecting Amazonian biodiversity. Science 269:63-66.

Valencia, R., R. G. Condit, R. B. Foster, K. Romoleroux, G. Villa Muñoz, J. C. Svenning, E. Magard, M. Bass, E. C. Losos, and H. Balslev. 2004. Yasuní Forest Dynamics Plot, Ecuador. Pages 609-628 in E. C. Losos, and E. G. J. Leigh, editors. Tropical forest diversity and dynamism: Findings from a large-scale plot network. University of Chicago Press, Chicago.

Whitfield, S. M., K. E. Bell, T. Philippi, M. Sasa, F. Bolaños, G. Chaves, J. M. Savage, and M. A. Donnelly. 2007. Amphibian and reptile declines over 35 years at La Selva, Costa Rica. Proceedings of the National Academy of Sciences of the United States of America 104:8352-8356.

Whitfield, S. M., and M. A. Donnelly. 2006. Ontogenetic and seasonal variation in the diets of a Costa Rican leaf-litter herpetofauna. Journal of Tropical Ecology 22:409-417.

Williamson, G. B., M. G. Radtke, and J. L. Deichmann. 2005. How many tropical rainforests? Tropinet 16:1-3.

Wright, D. H. 1983. Species-energy theory: an extension of species-area theory. Oikos 41:496- 506.

Wright, S. J., C. Carrasco, O. Calderon, and S. Paton. 1999. The El Nino Southern Oscillation, variable fruit production, and famine in a tropical forest. Ecology 80:1632-1647.

Zimmerman, B. L., and D. Simberloff. 1996. An historical interpretation of habitat use by frogs in a central Amazonian forest. Journal of Biogeography 23:27-46.

CHAPTER 2 ANURAN ARTIFACTS OF PRESERVATION: 25 YEARS LATER*

INTRODUCTION

Approximately 25 years ago, Julian Lee (1982) published the first-ever article detailing changes in linear measurements and allometries of an anuran, Chaunus marinus. Despite wide recognition and accolades of the study, there have been no subsequent studies of frog changes in preservation (Hayek et al. 2001). Perhaps more problematic is that researchers have continued the long history of assuming that character transformations between live and preserved specimens are negligible or that they are commensurate across individuals of a species (e.g.,

Emerson 1978, Withers and Hillman 2001).

Such ongoing assumptions emerge from a practical reality. Preserved specimens are important tools for scientific studies of evolution, ecology and systematics, and museum wet collections house well over one million anuran specimens worldwide (HerpNET 2007).

Although these specimens have been collected and preserved using a variety of techniques over the years, many collectors have moved towards more uniform methods of preservation over the last few decades, and today most anurans are fixed in formalin and then transferred to alcohol for storage (Simmons 2002, National Park Service 2006).

While measurements of preserved specimens continue to be applied to live animals, little is known about the effects of preservation on morphology of frogs. Obviously, biomass of preserved frogs can easily misrepresent live biomass, but how much do linear measurements such as snout-urostyle length change during preservation? Lee's (1982) seminal study of a single species, Chaunus marinus, revealed that all 14 linear measurements changed after preservation.

Some measurements increased while others decreased.

*Currently in review in the journal Phyllomedusa 23

Preservation artifacts are better known in fishes, although relatively few species have been studied, mostly those of commercial value (e.g., Burgner 1962, Parker 1963, Stobo 1972,

Engel 1974, Yeh and Hodson 1975, Billy 1982, Leslie and Moore 1986, Jennings 1991, Sagnes

1997, Fey 1999). The types of osmotic influences and autolysis that change lengths in fishes in preservation may produce similar effects in frogs, given that lengths of both taxa encompass skin and cartilage over an internal skeleton. Changes in fish length with preservation are not uniform and tend to be species-specific and vary among size classes and the preservative used (Billy

1982, Jennings 1991, Fey 1999). The general trend is a decrease in length with most of the shrinkage occurring within the first 40 days after preservation (Jennings 1991, Sagnes 1997).

Here, we evaluate the single most widely measured character in anurans, snout-vent length (SVL), before and after preservation, for 14 species of frogs to test for interspecific differences in response to preservation. Whereas Lee (1982) measured numerous characters on one species, we have measured one character on numerous species. We address four questions:

 To what degree does SVL change with preservation?

 Does the amount or direction of change vary by species?

 Is the change proportionally different in large or small frogs?

 Do specimens continue to change over time in preservative?

METHODS

We examined and re-measured SVL of frogs that had been stored at the Louisiana State

University Museum of Natural History (LSUMZ). Initially the live frogs had been anesthetized with chloretone and snout-urostyle length (SVL1) was measured to the nearest 0.1 mm using dial calipers. Then, the frogs were fixed using 10% formalin and subsequently transferred to either

70% ethanol or 55% isopropanol for long term storage. The preserved frogs were measured a second time (SVL2) between August and October, 2006. This represented at least one month and 24 not more than 5 years after the initial measurement (SVL1). Then, SVL3 was measured in

November 2007 (13 months after SVL2) for 3 species to monitor secondary storage changes, aside from initial preservation.

Initial SVL1 was measured by JB, whereas SVL2 and SVL3 were measured by JLD. In order to monitor potential researcher bias in measurements, both JB and JLD were asked to measure the same set of 10 frogs for 3 species of different sizes at SVL2--Acris gryllus, Hyla cinerea and Incilius nebulifer.

Because the time between SVL1 and SVL2 was not uniform, we regressed numerical change in preservation (SVL2-SVL1) on time in preservation in order to determine whether time since initial preservation had an effect on overall SVL change per individual frog. We then used paired t-tests to test for differences between pairs of measurements of SVL1, SVL2 and SVL3 and between researchers for SVL2. Pearson‘s correlation coefficient was used to investigate changes in SVL as a function of frog size among species and within species. Unless otherwise indicated, all statistical tests were performed using SAS 9.1 (Cary, NC).

RESULTS

For the three species measured by both researchers, there was a detectable, but statistically insignificant difference in measurement bias. The mean percent difference in measured lengths between the researchers (JLD-JB) was -0.68% for Incilius nebulifer, -0.46% for Hyla cinerea and -1.09% for Acris crepitans (Table 2.1). These differences were not significant in any of the three (Table 2.1, P=0.36, 0.25, and 0.14 respectively), nor were they significant taken together (P>0.10 Fisher's Combined Probability Test, Sokal and Rohlf, 1969).

Despite statistical insignificance, the differences represent a mean difference in measurements of

0.74%, which may be biologically important (Hayek and Heyer 2005).

Table 2.1 Size range of frogs measured and the difference in measurements between researchers (JLD-JB).

Family Species N SVL range (mm) Difference range (mm) Mean difference (SD) P-value Mean % difference Bufonidae Incilius nebulifer 10 33.7 - 70.3 1.9 - -2.2 -0.41 (1.33) 0.356 -0.68 Hylidae Hyla cinerea 10 24.4 - 55.0 0.9 - -0.9 -0.18 (0.46) 0.246 -0.46 Hylidae Acris gryllus 10 15.6 - 26.5 1.0 - -0.3 -0.20 (0.39) 0.138 -1.09

Table 2.2 Ranges and means of SVL1 (before preservation) and the change in SVL from the first measurement to the second measurement (SVL2-SVL1). Taxonomy from Frost et al. 2006. SVL Mean Difference Species range SVL range Mean Mean % Family Species ID N (mm) (mm) (mm) difference P-value difference Bufonidae Incilius nebulifer INNE 26 36.0 - 79.9 64.73 3.9 - -3.7 -0.71 0.046 -1.32 Ranidae Lithobates sphenocephala LISP 15 34.6 - 82.6 62.09 1.0 - -5.5 -2.17 0.002 -3.55 Bufonidae Anaxyrus fowleri ANFO 12 43.4 - 73.8 61.99 2.4 - -3.7 -1.38 0.010 -2.29 Scaphiopodidae Scaphiopus holbrookii SCHO 6 46.6 - 63.6 56.57 -1.3 - -5.2 -2.63 0.012 -4.56 Ranidae Lithobates clamitans LICL 56 23.6 - 82.1 45.71 2.5 - -7.3 -1.83 0.001 -3.99 Hylidae Hyla cinerea HYCI 23 25.0 - 56.3 44.45 0.3 - -4.4 -1.60 0.001 -3.58 Hylidae Hyla chrysoscelis HYCH 27 19.6 - 52.5 41.39 3.0 - -5.5 -1.38 0.001 -2.96 Hylidae Hyla avivoca HYAV 13 25.2 - 44.3 36.31 0.9 - -4.6 -1.78 0.001 -4.89 Hylidae Hyla femoralis HYFE 8 28.6 - 37.0 32.46 0.5 - -2.1 -0.60 0.080 -1.90 Hylidae Hyla squirella HYSQ 35 22.0 - 35.1 29.83 2.1 - -2.0 -0.37 0.044 -1.05 Microhylidae Gastrophryne carolinensis GACA 17 24.2 - 32.3 27.35 0.4 - -4.1 -1.79 0.001 -6.36 Hylidae Pseudacris crucifer PSCR 14 13.4 - 33.1 26.36 1.1 - -2.5 -0.57 0.046 -2.14 Hylidae Acris gryllus ACGR 24 16.5 - 27.5 22.33 0.8 - -1.5 -0.50 0.001 -2.19 Hylidae Acris crepitans ACCR 52 11.1 - 25.6 20.28 2.5 - -2.5 -0.65 0.001 -2.85

Tests for differences due to preservation (SVL2-SVL1) in each species resulted in negative differences (i.e., shrinkage) for all 14 species (Table 2.2). The mean shrinkage was statistically significant in 13 of the 14 species measured and nearly significant in the 14th species, Hyla femoralis (P=0.08), whose sample size was small with only 8 individuals (Table

2.2). Over the 14 species, the magnitude of shrinkage in SVL varied from 1.05% to 6.36%

(Table 2.2). Shrinkage was independent of time in preservation across all individuals (Pearson‘s

R = 0.08, P = 0.14).

Across species, there was no indication that smaller species shrank proportionately more or less than larger species (Pearson‘s R = 0.0002, P = 0.96; Figure 2.1A). However, numerical shrinkage did increase in larger anuran species, with a small coefficient of determination

(Pearson‘s R = 0.28, P = 0.05; Figure 2.1B). Sample size was not associated with numerical shrinkage (Pearson‘s R = 0.02, P = 0.64) nor with percent shrinkage (Pearson‘s R = 0.06, P =

0.38).

Within species, results were somewhat mixed, but generally showed no consistent evidence for size-related shrinkage (Table 2.3). Nine species showed neither a significant correlation between numerical shrinkage and SVL nor between percent shrinkage and SVL.

Three species showed both significantly more shrinkage and more percent shrinkage with greater size, respectively: Acris crepitans (Pearson‘s R =0.17 for numerical shrinkage and Pearson‘s R

=0.14 for percent shrinkage), Gastrophryne carolinensis (Pearson‘s R =0.56, Pearson‘s R

=0.45), and Hyla chrysoscelis (Pearson‘s R =0.21, Pearson‘s R =0.26). One species, Lithobates clamitans, showed significantly more numerical shrinkage with size (Pearson‘s R =0.16), but not more percent shrinkage with size. Another species, Incilius nebulifer, showed significantly less percent shrinkage with size (Pearson‘s R =0.21), but no difference in absolute shrinkage with size.

A SVL1 (mm) 0.000 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 -1.000

-2.000

-3.000 Percent change Percent

-4.000

-5.000

-6.000

-7.000

B SVL1 (mm) 0.000 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00

-0.500

-1.000

-1.500 Numerical change (mm) change Numerical -2.000

-2.500

-3.000

Figure 2.1 The relationship between mean SVL of 14 species before preservation and (A) the mean percent of shrinkage and (B) the mean numerical degree of shrinkage after preservation.

Table 2.3 Pearson‘s Correlation Coefficient for the relationship between percent and numerical changes and SVL1 within species. See Table 2.2 for species ID; * = P<0.05.

Correlation coefficient Percent Numerical Family Species ID change change Bufonidae INNE 0.46* 0.23 Ranidae LISP 0.08 -0.27 Bufonidae ANFO 0.16 -0.08 Scaphiopodidae SCHO -0.43 -0.53 Ranidae LICL -0.01 -0.41* Hylidae HYCI -0.05 -0.20 Hylidae HYCH -0.50* -0.46* Hylidae HYAV -0.01 -0.07 Hylidae HYFE 0.31 0.26 Hylidae HYSQ -0.28 -0.29 Microhylidae GACA -0.67* -0.75* Hylidae PSCR -0.04 -0.25 Hylidae ACGR -0.23 -0.37 Hylidae ACCR -0.37* -0.41*

Changes over the additional 13 months of storage, measured as differences between

SVL2 and SVL3, were greatly reduced relative to the original shrinkage in preservative. For the three species measured, there was no significant change in Acris gryllus, a significant 0.30% increase in Incilius nebulifer and a significant 0.45% decrease in Hyla cinerea (Table 2.4).

Table 2.4 SVL differences between 2 measurements taken after preservation (SVL3-SVL2).

Difference Mean P- Mean % Family Species N range (mm) difference value difference Bufonidae Incilius nebulifer 26 1.10 - -0.50 0.223 0.008 0.30 Hylidae Hyla cinerea 23 0.05 - -0.60 -0.196 0.001 -0.45 Hylidae Acris gryllus 24 0.40 - -0.40 -0.042 0.304 -0.22

DISCUSSION

Because the original live SVL1 measurements were performed by JB, and the recent

SVL2 measurements by JLD, we tested for differences in inter-observer mensuration. These differences were not statistically significant although perhaps larger sample sizes would have proven them to be. However, our goal here was to determine the potential magnitude of researcher bias in order to compare it to the measured shrinkage, not to prove that researcher bias can exist (Hayek et al. 2001). Mean shrinkage from the original SVL1to SVL2 in the three co- measured species could be adjusted accordingly for each of those species as follows: Incilius nebulifer (1.32%-0.68=0.64% shrinkage), Hyla cinerea (3.58%-0.46%=3.12% shrinkage), and

Acris gryllus (2.19%-1.09%=1.10% shrinkage). For the remaining species, we do not have co- measurements from both observers, but we suggest that the percent shrinkage values in Table 2.2 can be reduced by the mean inter-observer bias of 0.74%. Doing so would adjust the range of percent shrinkage values to 0.31%-5.62% across the 14 species and adjust the mean percent shrinkage to 2.38% (3.12%-0.74%).

One other study of inter-observer measurements on a set of 88 individuals of Vanzolinius discodactylus, showed a statistically significant difference of 1.4% between two observers for

SVL (Hayek et al. 2001). That study found significant differences in 13 of 14 characters studied, although the variable measured most consistently and with the greatest precision was SVL

(Hayek et al. 2001). Here, we have achieved less inter-observer variability because we, in fact, attempted to standardize our SVL measurements by having the observers converse and compare preliminary measurements. In short, our goal was to minimize inter-observer differences, not monitor independent measurements as was the case in Hayek et al. (2001). It should be no surprise then, that our inter-observer difference (0.74%) is about half that (1.4%) found by

Hayek et al. (2001).

Across the 14 species studied here adjusted shrinkage averaged 2.38% of SVL, and ranged from 0.31% to 5.62%, a range which is quite comparable to the range of shrinkage in fishes of 1% to 6.8% (Lee 1982). In contrast, Lee's measurements on Chaunus marinus showed an increase in SVL with preservation, not a decrease (+1.9% in males and +1.2% in females). Of his 14 characters, Lee (1982) reported 6 increases and 8 decreases in preservation. Given our records for 14 frog species and numerous reports for fishes, the increase in length in preservation of Chaunus marinus appears to be unusual among both frogs and fishes, and as such it may be particular to that species, as Chaunus marinus is among the largest and "hardiest" of anurans.

However, it should be no surprise to find species-specific differences in frogs at least as great as those in fishes, given the magnitude of morphological variation among species in both taxa. For example, our greatest degree of adjusted shrinkage, 5.62%, occurred in Gastrophryne carolinensis, which coincidentally was the only species where size explained about half the shrinkage; clearly, this species is more susceptible to shrinkage than the others we studied, as it shrank more and showed more size-related shrinkage. Further investigations examining species- specific rates of change in preservation are necessary to prevent potential biases in conclusions drawn from studies involving museum specimens.

For our 14 species, variation in shrinkage was not correlated with species size, sample size, time in preservation, nor was it associated with family or genus. Within species there was some tendency for shrinkage to increase with frog size, but the majority, 9 of 14, showed no evidence for shrinkage as a function of frog size (Table 2.2). These results contradict the general trend in fishes where shrinkage is proportionately greater in smaller fish (Burgner 1962, Stobo

1972, Yeh and Hodson 1975, Billy 1982). Consequently, the only working hypothesis for the degree of shrinkage associated with preservation in anurans is that there are species-specific differences.

This hypothesis offers both positive and negative factors for anuran biologists. On the negative side, we are unable to simply assume no changes occur in SVL with preservation, nor can we assume some constant proportional shrinkage across all species. On the positive side, we can establish useful guidelines for SVL measurements on preserved specimens:

1. SVL measurements on preserved specimens are likely to be different from SVL measurements on live individuals, on the order of 0-6%.

2. SVL measurements on preserved specimens can be adjusted by a species-specific proportional factor that can be determined by measuring SVL on live, anaesthetized individuals, preserving them through standard protocol, and re-measuring SVL after a moderate period of time—the time is likely to be 2-3 months, based on data here and from Lee (1982).

3. As the proportional factor is likely to be in the range of 0-6%, applications using preserved

SVL can either accept such error without adjustment, or provide proportional adjustment as needed. For example, in systematic studies, where frog measurements are used to distinguish cryptic species, this error may not be tolerable. However, in ecological studies, such as estimation of anuran community biomass (Deichmann et al. 2008), a 6% error in SVL measurements would translate into approximately a 2% error in mass estimates for most species, or 2 grams out of 100. This difference in SVL measurements may be acceptable, depending on the ultimate goals of the study.

These guidelines are obviously tentative as we have measured preservation effects on only 14 of the globe‘s 5000 species. Still, expanding the universe of monitored preservation effects from one (Lee 1982) to 14 merits a modicum of tentative generalization.

LITERATURE CITED

Billy, A. J. 1982. The effects of formalin and isopropyl alcohol on length and weight measurements of Sarotherodon mossambicus Trewavas. Journal of Fish Biology 21:107- 112.

Burgner, R. L. 1962. Studies of red salmon, Oncorhynchus nerka, smolts from the Wood River Lakes, Alaska. Pages 251-316 in T. S. Koo, editor. Studies of Alaska red salmon. University of Washington Fisheries Publications, Seattle.

Deichmann, J. L., W. E. Duellman, and G. B. Williamson. 2008. Predicting biomass from snout- vent length in New World frogs. Journal of Herpetology 42:238-245.

Emerson, S. B. 1978. Allometry and jumping in frogs: helping the twain to meet. Evolution 32:551-564.

Engel, S. 1974. Effects of formalin and freezing on length, weight and condition factor of cisco, Coregonus artedi, and yellow perch, Perca flavescens. Transactions of the American Fisheries Society 1:136-138.

Fey, D. P. 1999. Effects of preservation technique on the length of larval fish: methods of correcting estimates and their implication for studying growth rates. Archive of Fishery and Marine Research 47:17-29.

Frost, D. R., T. Grant, J. Faivovich, R. H. Bain, A. Hass, C. F. B. Haddad, R. O. De Sa, A. Channing, M. Wilkinson, S. C. Donnellan, C. J. Raxworthy, J. A. Campbell, B. L. Blotto, P. Moler, R. C. Drewes, R. A. Nussbaum, J. D. Lynch, D. M. Green, and W. C. Wheeler. 2006. The Amphibian Tree of Life. Bulletin of the American Museum of Natural History 297:1-370.

Hayek, L.-A. C. and W. R. Heyer. 2005. Determining sexual dimorphism in frog measurement data: integration of statistical significance, measurement error, effect size and biological significance. Anais da Academia Brasileira de Ciências 77:45-76.

Hayek, L.-A. C., W. R. Heyer, and C. Gascon. 2001. Frog morphometrics: a cautionary tale. Alytes 18:153-177.

HerpNET. 2007. HerpNET data portal. http://www.herpnet.org.

Jennings, S. 1991. The effects of capture, net retention and preservation upon lengths of larval and juvenile bass, Dicentrarchus labrax (L.). Journal of Fish Biology 38:349-357.

Lee, J. C. 1982. Accuracy and precision in anuran morphometrics: Artifacts of preservation. Systematic Zoology 31:266-281.

Leslie, J. K. and J. E. Moore. 1986. Changes in lengths of fixed and preserved young freshwater fish. Canadian Journal of Fisheries and Aquatic Sciences 43:1079-1081.

Parker, R. P. 1963. Effects of formalin on length and weight of fishes. Journal of the Fisheries Research Board of Canada 20.

Sagnes, P. 1997. Potential artefacts in morphometric analyses of fish: effects of formalin preservation on 0+ grayling. Journal of Fish Biology 50:910-914.

National Park Service 2006. National Park Service Museum Handbook - Appendix T: Curatorial care of biological collections. http://www.nps.gov/history/museum/publications/MHI/mushbkI.html.

Simmons, J. E. 2002. Herpetological collecting and collections management. Herpetological Circular 31:1-153.

Stobo, W. T. 1972. Effects of formalin on the length and weight of yellow perch. Transactions of the American Fisheries Society 101:362-364.

Withers, P. C. and S. S. Hillman. 2001. Allometric and ecological relationships of ventricle and liver mass in anuran amphibians. Functional Ecology 15:60-69.

Yeh, C. F. and R. G. Hodson. 1975. Effects of formalin on length and weight of bluegill and white crappie from Lake Nasworthy, Texas. The Southwestern Naturalist 20:315-322.

CHAPTER 3 PREDICTING BIOMASS FROM SNOUT-VENT LENGTH IN NEW WORLD FROGS*

INTRODUCTION

In moist, warm environments in temperate and tropical regions, anurans are a major component of ecosystems. Their roles in ecological pyramids are substantial (Whiles et al. 2006), although data on biomass generally are unavailable. Instead, anuran communities continue to be characterized by species richness and relative abundances. These variables are incorporated into classical ecological diversity indices, but alone, they do not encompass the dynamics portrayed by biomass and secondary productivity, which reveal the role that taxa or guilds play in ecosystem energy flow (Smith and Smith, 2001). Furthermore, local extinction of some species seems to be a global phenomenon, but how such declines affect anuran community biomass is essentially unexplored (Pounds, 2001; Collins and Storfer, 2003; Stuart et al., 2004; Lips et al.,

2005).

Historically, anuran biomass may have been neglected in field studies given the lack of accurate scales and balances at remote sites. However, snout-vent lengths (SVL) have been recorded from living individuals in the field or preserved ones in the laboratory, with little difference between measurements of living and preserved individuals (Lee 1982). Consequently, there are extensive datasets of anuran SVLs, often without accompanying mass data. An exception is the anuran community at Cusco Amazónico, Peru (Duellman, 2005).

Herein, we explore the relationship between SVL and mass and its utility for estimating anuran biomass (excluding larval stages). Length-weight relationships have been useful in estimating biomass for a variety of organisms, including insects (Rogers et al. 1977, Schoener

*Reprinted by permission of Journal of Herpetology. The manuscript has been lightly edited for recent changes in taxonomy.

1980), spiders (Sage 1982, Brady and Noske 2006), sea turtles (Georges and Fossette 2006), marine mammals (Trites and Pauly 1998), and fish (Kohler et al. 1995, Martin-Smith 1996,

Froese and Palomares 2000). In this study, we employ linear regression to determine how well

SVL estimates mass for 36 species of frogs, and we discuss the implications of these regressions for estimates of community biomass. Because most anurans exhibit sexual dimorphism (Shine

1979) and complex development with larval, juvenile and adult stages (Duellman and Trueb

1986), we also investigate differences in the SVL/mass relationships between sexes, between gravid and non-gravid females, and between juveniles and adults.

METHODS

This study is based on data on anurans that we collected in the USA, Ecuador, and Peru.

We included only species where SVL and mass were measured for at least 10 individuals.

In the USA, specimens were collected in Louisiana between August 2002 and September

2006, using a combination of opportunistic sampling and directed effort for particular species.

For individuals collected and released on site, SVL (to nearest 0.1 mm) was measured with dial calipers and mass (to nearest 0.01 g) with an electronic balance (Ohaus Scout Pro SP202). For these individuals, all frogs were captured and placed in zip-lock bags until all were measured.

Frogs were only released once sampling of the site was complete. Collected specimens were taken live to Louisiana State University, anaesthetized in chloretone, and then measured with dial calipers (SVL to 0.1 mm) and weighed on an electronic balance (0.01 g) (Sargent-Welch

SWE-500). These specimens were fixed on a surface saturated with formalin and then draped with same. All specimens were stored in 55% isopropanol or 70% ethanol at the Louisiana State

University Museum of Natural Science (LSUMZ).

In Peru, specimens were collected at Cusco Amazónico, Departamento de Madre de

Dios, discontinuously from January 1986 through December 1991. Mass (0.1 g) was measured

36 in the field with Pesola scales. Snout-vent length was measured (0.1 mm) with dial calipers between 1995 and 1997 on the preserved specimens. Similarly, specimens from Departamento de Loreto, Peru were collected in June and July 1993, weighed in the field, and SVL measured on preserved specimens in December 1993. All specimens were preserved in 10% formalin and stored in 70% ethanol at the University of Kansas Natural History Museum (KU).

In Ecuador, specimens were captured at Tiputini Biodiversity Station and Yasuní

National Park, Provincia de Orellana, from April through May, 2005, and from February through

April, 2006, using both randomly located litter plots and opportunistic sampling. Before being released on site, individuals were measured (SVL) in the field with dial calipers (0.1 mm) and weighed on an electronic balance (0.01 g) (Ohaus Scout Pro SP202). As in Louisiana, individuals in a plot or given area were placed in zip-lock bags and not released until all frogs had been measured. Therefore, individuals were not sampled more than once.

Mass and SVL were log10 transformed for least squares linear regressions for each of the

36 species, following the equation: log Mass = log a + b log SVL. From 1,492 specimens from the three countries, 15 were removed as outliers, based on studentized residuals (R Student).

Generally, when the absolute value of an R Student observation is greater than 2, the observation is considered suspicious and its validity questionable (SAS Help and Documentation, SAS

Institute Inc., Cary, North Carolina). In order to be conservative, we removed only observations with an R Student larger than 4. We believe these extreme outliers to be a result of measurement or recording error.

Analyses of covariance were performed to test for differences in the SVL/mass (log/log) relationship between three sets of pairwise classes: juveniles vs. adults, gravid females vs. non- gravid females, females vs. males. Class (sex or stage) was used as a factor in the ANCOVA.

We first tested for a difference in the SVL*class interaction which corresponds to the slope. If

37 the slopes were not different, we then tested for a difference in class which corresponds to the intercept. When gravid females differed significantly from non-gravid females, only the latter were compared to males; otherwise, all females were included in the female-male comparison.

Sex and life stage were not recorded for all individuals in the data set. Therefore, only species for which more than 5 individuals were available in the data set for each sex or life stage were used in ANCOVA analyses.

For each species, we employed the significance level 0.05 to determine biological importance, reporting both the mass/SVL regression equation and its R2 and P values. However, for the ANCOVAs that compared the three subclasses within a species, we pursued one additional criterion to determine biological importance. When significant differences existed between juveniles and adults, the equation developed from all individuals (the ‗species equation‘) was used to predict mass for each individual. Then, the predicted mass for individuals of each stage, juvenile and adult, was compared separately to the actual mass of individuals of that stage in a paired t-test. This same method was applied to ANCOVA differences between gravid and non-gravid females.

When females and males showed statistically different mass/SVL regressions through

ANCOVA, we visually inspected the data with groups that may be sexually dimorphic, because males and females may not completely overlap in size (Hayek and Heyer 2005). To determine biological importance of the separate sex regressions, we examined the overlap of the 95% confidence intervals of the regression equation for one sex with the data points for the opposite sex—i.e., did the data for the opposite sex fall inside or outside the 95% confidence limits? All statistical tests were performed with SAS software (SAS software, Version 9.1.3, 2004).

The classification of anurans is in a state of flux. Herein, we follow the taxonomy of

Hylidae proposed by Faivovich et al. (2005) and for Aromobatidae and Dendrobatidae proposed

38 by Grant et al. (2006). For Bufonidae, Leiuperidae and Ranidae, we follow the classification proposed by Frost et al. (2006), despite arguments against this scheme given by Wiens (2007).

We do not recognize the family ―Brachycephalidae‖ proposed by Frost et al. (2006) and instead follow the classification of the Terrarana proposed by Hedges et al. (2008), recognizing the family Strabomantidae.

RESULTS

We examined a total of 36 species of anurans belonging to seven families (Table 3.1).

For each species, the linear regression of mass on SVL was highly significant (P < 0.001)

(Figure 3.1A–D). Log SVL explained 75-100% of the variation in log mass for 32 of the 36 species (Table 3.1). In the other four species, SVL explained at least 50% of the variation in mass: Allobates trilineatus (R2 = 0.65), Scinax ictericus (R2 = 0.50), Leptodactylus rhodonotus

(R2 = 0.62), and Elachistocleis ovalis (R2=0.59).

Sufficient data were available from six species for analysis of covariance between juveniles and adults in mass/SVL (Table 3.2). Statistically significant differences were detected in 3 species, with differing slopes between juveniles and adults in Leptodactylus didymus (F1,41 =

13.22, P = 0.001) and differing intercepts in two species, Ameerega bilinguis (F1,53 = 10.01, P =

0.003) and Phyllomedusa vaillanti (F1,49 = 11.86, P = 0.001). However, in none of these three cases (six tests, P > 0.05) was there a significant difference between the observed mass of one stage class and the mass predicted from the species regression equations.

Adequate sample sizes for nine species permitted ANCOVA tests of differences in the mass/SVL relationships between gravid and non-gravid females (Table 3.2). Only two of the nine showed significant differences, one in the intercept, Leptodactylus didymus (intercept F1,16

= 6.67, P = 0.020), and one in the slope, Phyllomedusa tomopterna (slope F1,6 = 11.48, P =

0.015). However, paired t-tests showed no differences (four tests, P > 0.05) between the

Table 3.1 Size ranges and linear regression coefficients for 36 New World frog species, and for sexes separately in those species exhibiting important regression differences between males and females. All species regressions were significant (P < 0.001). All male and female regressions were significant (P<0.005) with the exception of Pristimantis fenestratus males (P = 0.21). SVL (MM) Mass (g) Species Species ID N Min. Max. Min. Max. R2 Slope SE Intercept Bufonidae: Anaxyrus fowleri BuFo 13 43.40 73.80 7.42 39.24 0.99 3.01 0.10 -4.07 Anaxyrus terrestris BuTe 13 43.50 72.90 6.31 33.25 0.93 3.05 0.25 -4.14 Incilius nebulifer BuNe 67 12.30 79.90 0.15 46.64 1.00 2.94 0.02 -4.00 Rhinella “margaritifera” (Ecuador) BuMaE 67 7.40 67.90 0.04 26.95 0.99 3.00 0.03 -4.04 Rhinella “margaritifera” (Peru) BuMaP 66 40.50 67.20 5.00 29.50 0.93 3.04 0.10 -4.13 Aromobatidae: Allobates trilineatus CoTr 49 11.40 18.60 0.20 0.54 0.65 1.95 0.22 -2.76 Allobates femoralis AlFe 28 8.50 26.60 0.09 1.85 0.94 2.64 0.12 -3.56 Dendrobatidae: Ramitomeya duellmani DeDu 11 9.20 16.70 0.11 0.48 0.83 2.04 0.31 -2.90 Ameerega bilinguis EpBi 56 10.50 22.60 0.12 1.15 0.89 2.82 0.13 -3.80 Hylidae: Acris crepitans AcCr 75 11.10 27.70 0.15 1.69 0.95 2.87 0.08 -3.89 Dendropsophus leucophyllata DeLe 42 28.30 40.20 1.10 7.50 0.76 4.17 0.37 -5.96 Non-Gravid Females 11 32.70 40.20 1.90 7.50 0.72 4.68 0.89 -6.73 Males 31 28.30 35.30 1.10 2.60 0.28 2.31 0.70 -3.19 Hyla cinerea HyCi 24 25.00 56.30 0.86 10.82 0.94 3.13 0.17 -4.47 Hyla squirella HySq 36 22.00 35.10 0.57 2.88 0.89 2.99 0.18 -4.16 Hypsiboas fasciata HyFa 45 33.90 51.20 2.00 6.80 0.88 2.75 0.16 -3.92 All Females 10 42.40 51.20 4.00 6.80 0.65 2.33 0.61 -3.21 Males 35 33.90 40.30 2.00 3.20 0.24 1.40 0.43 -1.81 Osteocephalus taurinus OsTa 36 48.60 93.90 5.10 50.00 0.86 2.97 0.20 -4.27 Phyllomedusa tomopterna PhTo 32 41.50 62.30 3.30 9.80 0.76 2.26 0.23 -3.11 Phyllomedusa vaillanti PhVa 52 20.30 76.20 0.30 35.50 0.95 3.06 0.10 -4.40 40

(Table 3.1continued)

Scinax ictericus ScIc 108 26.30 36.70 0.90 2.70 0.50 3.05 0.30 -4.32 Scinax pedromedinae ScPe 71 20.30 31.50 0.50 2.20 0.81 2.70 0.16 -3.81 Strabomantidae: Hypodactylus nigrovittatus ElNi 15 14.10 23.70 0.32 1.27 0.96 2.82 0.17 -3.74 Pristimantis fenestratus ElFe 46 14.10 51.00 0.20 10.00 0.89 2.32 0.12 -2.99 Non-Gravid Females 15 38.70 51.00 4.80 10.00 0.79 2.16 0.43 -2.69 Males 19 23.40 33.30 1.60 3.30 0.09 -0.60 0.39 1.28 Pristimantis ockendeni ElOc 30 7.30 29.50 0.03 1.87 0.96 2.85 0.11 -3.92 Pristimantis peruvianus ElPe 50 15.70 41.40 0.50 5.80 0.90 3.05 0.15 -4.19 Pristimantis toftae ElTo 62 17.00 27.30 0.20 1.40 0.91 3.30 0.14 -4.56 All Females 26 18.90 27.30 0.60 1.40 0.80 2.62 0.24 -3.62 Males 35 17.00 19.80 0.20 0.60 0.73 5.03 0.58 -6.76 Oreobates quixensis IsQu 22 16.70 54.10 0.20 18.50 0.96 3.23 0.15 -4.42 Leiuperidae: Edalorhina perezi EdPe 28 24.60 36.70 1.30 4.63 0.90 3.17 0.21 -4.27 Engystomops petersi PhPe 36 15.50 37.40 0.30 5.44 0.87 2.90 0.20 -3.96 Leptodactylidae: Adenomera hylaedactyla AdHy 22 7.20 27.80 0.10 2.23 0.81 1.97 0.21 -2.72 Leptodactylus didymus LeDi 46 17.30 58.60 0.40 16.50 0.98 2.88 0.07 -3.82 Leptodactylus rhodonotus LeRh 26 62.90 83.10 21.00 55.00 0.62 2.40 0.39 -2.89 Microhylidae: Chiasmocleis bassleri ChBa 15 16.80 29.90 0.50 3.87 0.85 2.83 0.33 -3.74 Elachistocleis ovalis ElOv 25 30.00 43.80 2.70 5.60 0.59 1.48 0.26 -1.74 Gastrophryne carolinensis GaCa 25 11.90 32.30 0.14 2.59 0.98 2.86 0.08 -3.89 Hamptophryne boliviana HaBo 68 24.00 36.70 1.30 6.10 0.81 2.87 0.17 -3.77 Ranidae: Lithobates clamitans RaCl 78 23.60 82.10 0.96 47.90 0.99 3.04 0.04 -4.14 Lithobates sphenocephalus RaSp 21 22.00 82.60 1.74 65.70 0.95 2.57 0.14 -3.31

A B 2 2 AcCr BuFo DeLe BuTe 1.5 HyCi 1.5 BuNe HyFa BuMaP BuMaE 1 HySq 1 OsTa RaCl PhTo RaSp 0.5 PhVa 0.5 ScIc

ScPe LogMass (g) LogMass (g) 0 0 1 1.2 1.4 1.6 1.8 2 1 1.2 1.4 1.6 1.8 2 -0.5 -0.5

-1 Log SVL (mm) -1 Log SVL (mm)

C D 2 2 CoTr ElFe ElNi AlFe 1.5 1.5 ElOc DeDu ElPe EpBi ElTo IsQu 1 ChBa 1 ElOv AdHy LeDi GaCa 0.5 LeRh HaBo 0.5 PhPe

EdPe Log Mass(g)Log 0 Mass(g)Log 0 1 1.2 1.4 1.6 1.8 2 1 1.2 1.4 1.6 1.8 2 -0.5 -0.5

-1 Log SVL (mm) -1 Log SVL (mm)

Table 3.2 Results of ANCOVA tests for regression differences between sexes and age classes within species. Significant P-values are in bold. J=Juveniles; A=Adults; F=Females; G=Gravid females; M=Males; F(G)=Females and gravid females pooled. See Table 3.1 for Species ID. Slope Intercept Species Comparison N F P F P BuMaE J VS A 37 / 28 0.01 0.909 0.12 0.735 ElFe J VS A 8 / 38 1.04 0.315 0.16 0.689 EpBi J VS A 13 / 43 0.77 0.383 10.01 0.003 IsQu J VS A 9 / 13 0.71 0.411 0.03 0.866 LeDi J VS A 15 / 31 13.22 <0.001 PhVa J VS A 9 / 43 0.07 0.795 11.86 0.001

CoTr F VS G 13 / 8 1.72 0.207 0.71 0.411 ElOv F VS G 12 / 6 2.15 0.165 0.00 0.984 ElPe F VS G 6 / 7 3.62 0.090 0.75 0.406 ElTo F VS G 16 / 10 1.81 0.192 0.19 0.663 HaBo F VS G 24 / 10 0.01 0.927 1.75 0.196 HyFa F VS G 5 / 5 2.09 0.198 1.46 0.266 LeDi F VS G 10 / 9 0.00 0.947 6.67 0.020 PhTo F VS G 5 / 5 11.48 0.015 ScIc F VS G 22 / 5 1.57 0.222 0.35 0.557

BuMaP F VS M 17 / 49 2.54 0.116 0.00 0.980 BuNe F VS M 12 / 23 0.17 0.686 0.33 0.571 CoTr F(G) VS M 21 / 22 2.80 0.104 1.78 0.191 DeLe F VS M 11 / 31 4.36 0.044 ElFe F VS M 15 / 19 22.24 <0.001 ElPe F(G) VS M 13 / 30 3.47 0.070 1.39 0.245 ElTo F(G) VS M 26 / 36 14.71 <0.001 HaBo F(G) VS M 34 / 34 3.11 0.083 8.11 0.006 HyFa F(G) VS M 10 / 32 1.56 0.219 10.58 0.002 HySq F VS M 14 / 19 5.56 0.025 LeDi F VS M 10 / 12 0.01 0.936 0.02 0.882 LeRh F VS M 11 / 13 3.40 0.080 0.00 0.959 OsTa F VS M 12 / 19 0.36 0.552 2.65 0.115 PhPe F VS M 15 / 13 0.05 0.834 1.32 0.262 ScIc F(G) VS M 27 / 81 3.09 0.082 47.53 <0.001 ScPe F VS M 28 / 39 0.31 0.579 4.51 0.038

44 observed mass of individuals of each group of females and the mass of individuals predicted from the species equation.

Sufficient sample sizes allowed comparison of mass/SVL regressions between females and males in 16 species (Table 3.2). Although eight of the 16 species exhibited statistically significant differences between males and females, four of these showed strong overlap of the

95% confidence intervals of the separate mass/SVL regressions for one sex and the points of the opposite sex (Figure 3.2 A-B). The remaining 4 species demonstrated notable sexual dimorphism with little to no overlap of regression confidence intervals (Figure 3.2 C-D). These latter four species may, in fact, require separate equations to estimate male and female mass/SVL relationships, so regressions for each sex are presented separately as well as combined (Table

3.1). All separate sex regressions are significant with the exception of male Pristimantis fenestratus (Table 3.1).

DISCUSSION

Estimating biomass from linear dimensions has been a useful technique in fisheries science (Ricker 1973), dietary studies (Beaver and Baldwin 1975), and conservation management (Trites and Pauly 1998, Braccini et al. 2006) across a variety of organisms.

Overwhelmingly for the anurans studied here, SVL predicted mass for individual species, suggesting that SVL data from past or future herpetological studies, combined with SVL/mass regressions and population densities, can functionally predict community biomass across geographic and temporal scales. Such data could document changes in community biomass as well as serve as a baseline for changes in individual taxa. For example, in cases where an individual species has disappeared or become rare, it is critical to know to what degree the entire community has changed in composition and biomass. With the ability to estimate community biomass, changes in each taxon can be compared to overall community changes, and changes at 45

Figure 3.2 Mass versus SVL plots showing differences in males and females of (A) Scinax pedromedinae, (B) Hyla squirella, (C) Hypsiboas fasciatus, and (D) Pristimantis fenestratus. Open circles represent females and solid triangles represent males.

A B 0.4 0.5 0.45 0.3 0.4 0.2 0.35

0.1 0.3 0.25 0 0.2

1.25 1.3 1.35 1.4 1.45 1.5 1.55 MassLog(g) 0.15 Log MassLog(g) -0.1 0.1 -0.2 0.05 -0.3 0 1.35 1.4 1.45 1.5 1.55 1.6 -0.4 Log SVL (mm) Log SVL (mm) C D 0.9 1.2 0.8 1 0.7 0.6 0.8 0.5 0.6

0.4 Log MassLog(g) Log MassLog(g) 0.3 0.4 0.2 0.2 0.1 0 0 1.5 1.55 1.6 1.65 1.7 1.75 1.3 1.4 1.5 1.6 1.7 1.8 Log SVL (mm) Log SVL (mm)

47 any one site compared to baseline data at other sites—for example, tropical montane sites to tropical lowland sites. Such information might prove useful for determining the relative sensitivity of different taxa to disease and environmental changes. In fact, monitoring through time may be one component of an overall strategy to track amphibian decline.

Estimates of community biomass, however, will only be as accurate as the underlying mass/SVL relationships and relative species contributions to the community. If the most abundant and/or largest species exhibit the worst mass/SVL regressions, the community biomass will be subject to error. In the four poorest regressions in this study, only 50–75% of mass was explained by SVL. Each of these four species is in a different family (Aromobatidae, Hylidae,

Strabomantidae, and Microhylidae), so there seems to be no taxonomic association with the poorer regression fits. Furthermore, these four species were represented by 26–108 individuals

(Table 3.1), demonstrating no particular tendency toward rarity or abundance.

Likewise, different mass/SVL relationships for sex or size classes within a species could complicate community biomass estimates, particularly if sex/size dimorphisms occur in the more abundant or larger species in a community. Here, we concluded that only four cases showed biologically important differences within a species—Dendropsophus leucophyllatus, Hypsiboas fasciatus, Pristimantis fenestratus and P. toftae (Table 3.1). All exhibit strong sexual size dimorphism (Duellman 1978, Bartlett and Bartlett 2003, Duellman 2005). In cases in which such sexually dimorphic species dominate a community, separate mass/SVL regressions may be important in estimating community biomass. In the present study, we tested relationships using data already in hand. In the future, it would be prudent to pursue individuals of both sexes across the range of sizes during data collection.

Although we limited our study to species-specific relationships of frogs, further analyses with larger datasets and more species could compare relationships for higher taxa, such as genera

48 and families. For example, a posteriori review of our dataset (Table 3.1), suggests that the slopes of the log-log mass/SVL regressions appear to be more uniform within the Bufonidae

(2.9-3.0) than within the Hylidae (2.3-4.2), a result that probably reflects the uniform body shape of bufonids relative to hylids (Figure 3.1A and 3.1B) (Duellman and Trueb 1986).

Here, we were restrained in predicting mass/SVL relationships for higher taxa, given potential disparities in our dataset, due to different researchers with different mensuration tools.

Still, the strength in the mass/SVL relationship over-rode such variation in sampling protocol.

For example, our Rhinella ―margaritifera‖ were collected either by JLD in Ecuador in 2005-

2006 and weighed on an electronic balance or by WED in Peru in 1986-1991 and weighed with pesola scales. In addition, R. ―margaritifera‖ apparently is a complex of several species across neighboring countries and perhaps within sites (Cisneros-Heredia 2006). Despite these opportunities for variation, our individuals of R. ―margaritifera‖ exhibited a remarkably strong mass/SVL relationship (Table 3.1, Figure 3.3), adding credence to the prospect of biomass analyses across different regions where data are contributed by different researchers.

In order to assure the accuracy of community biomass estimates, two steps are recommended. First, it is necessary to establish the mass/SVL relationships for many additional anuran species through field measurements on large numbers of individuals with care to include all sexes and sizes, moving toward more accurate mass measurements with electronic field balances rather than the traditional use of spring scales. Second, community studies that collect frogs according to standardized sampling protocols such as litter plots or transects should always include measurements of SVL, and of mass for undocumented species. Where such data are standardized, a user friendly database needs to be compiled, monitored and made accessible to the scientific community.

1.6

1.4

1.2

1 BuMaE_A 0.8 BuMaP_F

0.6 BuMaP_M Log Mass (g) Mass Log 0.4

0.2

0 1.35 1.45 1.55 1.65 1.75 1.85 1.95 Log SVL (mm)

Figure 3.3 Mass versus SVL plot of individuals of the Rhinella “margaritifera” complex (BuMa), demonstrating variation between geographic regions (E = Ecuador, P = Peru), sexes (A = adults of unknown sex, M = males, F = females). Regression lines for each group visually overlap at this scale, so they are not shown here. Such a database could also incorporate prior studies where the amphibian community was sampled adequately with standardized methodologies. For example, litter plots have been employed across many tropical regions and often repeated through time at some sites (Scott

1976, Inger 1980, Allmon 1991). Using modern mass/SVL relationships, community biomass could be determined for these older studies if SVL data were recorded for the individuals captured. For some sites, these historical data over 50 years would span the entire recent history of amphibian decline.

LITERATURE CITED

Allmon, W. D. 1991. A plot study of forest floor litter frogs, Central Amazon, Brazil. Journal of Tropical Ecology 7:503-522.

Bartlett, R. D. and P. P. Bartlett. 2003. Reptiles and Amphibians of the Amazon. The University Press of Florida, Gainesville. 50

Beaver, D. L. and P. H. Baldwin. 1975. Ecological overlap and the problem of competition and sympatry in the Western and Hammond's Flycatchers. Condor 77:1-13.

Braccini, J. M., B. M. Gillanders, and T. I. Walker. 2006. Total and partial length-length, mass- mass and mass-length relationships for the piked spurdog (Squalus megalops) in south- eastern Australia. Fisheries Research 78:385-389.

Brady, C. J. and R. A. Noske. 2006. Generalised regressions provide good estimates of insect and spider biomass in the monsoonal tropics of Australia. Australian Journal of Entomology 45:187-191.

Cisneros-Heredia, D. F. 2006. La Herpetofauna de la Estacion de Biodiversidad Tiputini, Ecuador: Diversidad y Ecologia de los Anfibios y Reptiles de una Comunidad Taxonomicamente Diversa. Universidad San Francisco de Quito, Ecuador.

Duellman, W. E. 1978. The biology of an equatorial herpetofauna in Amazonian Ecuador. Miscellaneous Publications, Museum of Natural History, University of Kansas 65:1-352.

Duellman, W. E. 2005. Cusco Amazónico: The lives of amphibians and reptiles in an Amazonian rainforest. Comstock Publishing, Ithaca.

Duellman, W. E. and L. Trueb. 1986. Biology of Amphibians. McGraw-Hill Book Company, New York.

Froese, R. and M. L. D. Palomares. 2000. Growth, natural mortality, length-weight relationship, maximum length and length-at-first-maturity of the coelacanth Latimeria chalumnae. Environmental Biology of Fishes 58:45-52.

Georges, J. Y. and S. Fossette. 2006. Estimating body mass in leatherback turtles Dermochelys coriacea. Marine Ecology-Progress Series 318:255-262.

Hedges, S. B., W. E. Duellman, and M. P. Heinicke. 2008. New World direct-developing frogs (Anura: Terrarana): Molecular phylogeny, classification, biogeography, and conservation. Zootaxa 1737:1-182.

Inger, R. F. 1980. Relative abundances of frogs and lizards in forests of Southeast Asia. Biotropica 12:14-22.

Kohler, N. E., J. G. Casey, and P. A. Turner. 1995. Length-weight relationships for 13 species of sharks from the western North-Atlantic. Fisheries Bulletin 93:412-418.

Lee, J. C. 1982. Accuracy and precision in anuran morphometrics: Artifacts of preservation. Systematic Zoology 31:266-281.

Martin-Smith, K. M. 1996. Length/weight relationships of fishes in a diverse tropical freshwater community, Sabah, Malaysia. Journal of Fish Biology 49:731-734.

Ricker, W. E. 1973. Linear regressions in fishery research. Journal of the Fisheries Research Board of Canada 30:409-434.

Rogers, L. E., R. L. Buschbom, and C. R. Watson. 1977. Length-weight relationships of shrub- steppe invertebrates. Annals of the Entomological Society of America 70:51-53.

Sage, R. D. 1982. Wet and dry-weight estimates of insects and spiders based on length. American Midland Naturalist 108:407-411.

SAS Help and Documentation, Version 9.1.3, 2004. SAS Institute Inc., Cary, NC.

SAS software, Version 9.1.3, 2004. SAS Institute Inc., Cary, NC.

Schoener, T. W. 1980. Length-weight regressions in tropical and temperate forest-understory insects. Annals of the Entomological Society of America 73:106-109.

Scott, N. J. 1976. The abundance and diversity of the herpetofaunas of tropical forest litter. Biotropica 8:41-58.

Shine, R. 1979. Sexual selection and sexual dimorphism in the Amphibia. Copeia 1979:297-306.

Trites, A. W. and D. Pauly. 1998. Estimating mean body masses of marine mammals from maximum body lengths. Canadian Journal of Zoology 76:886-896.

Whiles, M. R., K. R. Lips, C. M. Pringle, S. S. Kilham, R. J. Bixby, R. Brenes, S. Connelly, J. C. Colon-Gaud, M. Hunte-Brown, A. D. Huryn, C. Montgomery, and S. Peterson. 2006. The effects of amphibian population declines on the structure and function of Neotropical stream ecosystems. Frontiers in Ecology and the Environment 4:27-34.

CHAPTER 4 EFFECTS OF GEOMORPHOLOGY ON AMAZONIAN LEAF LITTER HERPETOFAUNA

INTRODUCTION

Relative to the rest of South America, the Amazon Basin appears to be a relatively homogeneous expanse of forest, but closer observations reveal a very different picture. The geomorphology of Amazonia has shaped historical differences in productivity and composition of the lowland forests.

Across the Amazon Basin, weathering and mineralization cause leaching of soil nutrients, the extent of which is associated with the age of the soils (Sombroek 2001). Much of Central and

Eastern Amazonia lies on ancient, low nutrient Oxisols (Van Wambeke 1992). These soils originated in the Guiana Shield and the Brazilian Highlands (>300 mya) and have no remaining weatherable mineral reserves (Sombroek 2000). On the other hand, a large extent of Western

Amazonia lies on much younger soils, mainly Ultisols (Valencia et al. 2004a), having eroded from the Andean uplift which began approximately 30 mya and continues to this day (Gregory-

Wodzicki 2000).

The paucity of available mineral nutrients in Central and Eastern Amazonia are expected to affect primary productivity. In fact, forests in Western Amazonia have been shown to support

50% higher coarse woody productivity than forests on older soils of Central and Eastern

Amazonia (Malhi et al. 2004). Furthermore, tree turnover in Western Amazonia is twice as high as tree turnover in the Central part of the Basin (Phillips et al. 2004), indicating faster regeneration and higher mortality. Differences in soil age also have been shown to have effects on floristic composition. For example, tree species composition has been shown to depend, at least in part, on soil fertility (Terborgh and Andresen 1998) and the major gradient in tree composition across Amazonia and the Guiana Shield emulates the gradient in soil age (ter Steege

53 et al. 2006). These differences in primary productivity due to soil age should have repercussions at higher trophic levels as well.

Although our understanding of tropical forest productivity has increased dramatically over the last few years, little is known about the implications of productivity on fauna (but see

Kay et al. 1997, Radtke et al. 2007, Peres 2008). Substantial data show that leaf litter herpetofauna abundance varies in rainforests across the globe (Scott 1976, Inger 1980, May

1980). Abundances of litter frogs and lizards in Central America are an order of magnitude greater than at Southeast Asian sites (Scott 1976, Inger 1980). Within the Neotropics, litter frog abundances in the Brazilian Amazon (Allmon 1991) may be low relative to the Peruvian

Amazon and Central America (Scott 1976, Toft 1980, Lieberman 1986). These results are strongly suggestive that plant productivity may play a crucial role in herpetofaunal biomass; however, for all these studies, the authors recorded numbers of individuals, not biomass, and sampling methodology varied from study to study so site to site comparisons may be biased.

Other studies have also shown general trends of increased animal abundances in areas with younger soils (Emmons 1984, Becker et al. 1991, Peres and Dolman 2000), but once again, these comparisons sampled sites differing in a number of confounding factors such as disturbance, fragmentation and/or hunting pressure. Here we investigate the affect of soil age, a known determinant of forest productivity (Malhi et al. 2004, Phillips et al. 2004, ter Steege et al. 2006), on the biomass, abundance and species richness of secondary consumers in the leaf litter community of large, undisturbed tracts of wet tropical forest, using a standardized methodology for sampling herpetofauna.

We predict that differences in forest productivity due to soil age will reverberate up through higher trophic levels, resulting in lower biomass and abundance of secondary consumers in the herpetofauna community on ancient soils. Biomass is a better measure of energy flow in a

54 system than abundance which simply reflects the raw number of individuals. For this reason, we expect biomass to display a stronger response to differences in soil age than abundance. We also expect that reptiles may exhibit more pronounced responses to differences in soil age than amphibians. Amphibians and reptiles share similar roles in the litter ecosystem as secondary consumers. However, because most amphibians undergo an aquatic life stage, often as herbivores or omnivores, reptiles spend a larger portion of their lives as predators on the forest floor. Furthermore, differences in density and biomass may push populations below minimum viable thresholds, resulting in fewer species in regions with ancient soils (Wright 1983, Evans et al. 2005).

METHODS

Study Sites

We selected 3 sites on ancient soils and 2 sites on young soils (Sombroek 2000). The ancient soil sites are located in the reserves of the Biological Dynamics of Forest Fragments

Project (BDFFP), which is a joint effort of the Instituto Nacional de Pesquisas da Amazônia

(INPA) and the Smithsonian Tropical Research Institute (STRI). Surrounded by well over

10,000 ha of continuous tropical lowland rainforest, the BDFFP is comprised of 3 reserves covering 1,000 km2 of both continuous and fragmented forest plots and is located approximately

100 km north of Manaus, Brazil (2°24‘S, 59°44‘W). Within the BDFFP, we sampled continuous forest in each of the 3 reserves: Dimona, Cabo Frio and KM41. The soils underlying this area are classified by Sombroek (2000) as Eastern Sedimentary Uplands, derived from pre-weathered crystalline parent material of the Guyanan Shield. The Eastern Sedimentary Uplands (ESU) are well drained and contain no remaining mineral reserves (Sombroek 2000).

The young soil sites sampled are Tiputini Biodiversity Station and Yasuní Research

Station, located within or adjacent to the Yasuní Biosphere Reserve in Eastern Ecuador. Tiputini

Biodiversity Station (0°37‘S, 76°10‘W) is a 650 ha reserve established in 1995 by the

Universidad San Francisco de Quito and Boston University. Yasuní Research Station (0°40'S,

76°24'W) is managed by the Pontificia Universidad Católica de Ecuador. The two sites are located approximately 30km from one another and are surrounded by extensive (>900,000 ha) continuous tropical lowland rainforest. This area is classified as Western Sedimentary Uplands

(WSU) by (Sombroek 2000) with soils derived from the Andean Uplift. They characteristically hold much more weatherable mineral reserves and have higher cation-exchange capacities than the ESU.

Table 4.1 Climate data for sampled sites on young (Tiputini and Yasuní) and old (BDFFP) soils. Months w/ Elevation Temperature Rainfall Length of rainy <100mm Site Latitude (m) (˚C)a (mm) season (# months) precipitationa BDFFP 2°24‘S 100 26 2651a 7c 0 Tiputini 0°37‘S 200 27 2740b 7b 0 Yasuní 0°40'S 200 25 2826a 7b 0 a - Radtke (2007); b - Karubian (2005); c - Gascon and Bierregaard (2001)

Floristic composition varies among sites, but forest structure is similar with emergent trees reaching between 45-55m (Laurance et al. 1998, Valencia et al. 2004b) and species richness exceeding 250 tree species ≥10cm DBH per hectare (Rankin-de Merona et al. 1992, Valencia et al. 1994, de Oliveira and Mori 1999). All sites are similar in latitude, elevation, rainfall and seasonality but differ in the geologic age of the soils (Table 4.1). Each site was sampled during its rainy season in order to avoid potential biases resulting from natural temporal fluctuations in herpetofauna population densities (Duellman 1995). We sampled quadrats in Ecuador from

April-May 2005 and again from February-March 2006. Plots in Brazil were sampled from

February-May 2007. Sampling years exhibited normal rainfall for each of the sites.

Data Collection

We established and raked a 1m border around 5x5m quadrats on primary terra firme forest sites. Plot selection was somewhat haphazard, based on the following criteria: quadrats were located at least 200m from any permanent body of water, at least 100m from the edge of a plateau, at least 200m from a forest edge, did not contain any excessively large trees (occupying over 1/4 of the space in the plot), and had no standing water. Within each plot, we measured eight environmental variables: the number of trees >10cm DBH, number of logs >10cm in diameter, percent canopy cover, leaf area index (LAI), litter depth, ambient temperature and humidity, and elevation. Plots were sampled during the day between 0700 and 1600 and were searched by teams of 2-4 individuals. All amphibians and reptiles encountered in the quadrats, with the exception of turtles and venomous snakes, were captured by hand and placed in zip-lock bags until the plot was completely sampled. Although sampling time depended greatly on the amount of litter and logs in the plot and the number of herpetofauna encountered, sampling of a single quadrat generally took 25-40 minutes. Captured individuals were identified to species and photographed from dorsal, ventral and lateral views. The snout-vent length (SVL) was measured with calipers to the nearest 0.01mm and all animals were weighed on a top-loading field balance

(Ohaus Scout Pro) to the nearest 0.01g. Individuals were released after processing. For observed individuals that escaped capture, the size of the animal was approximated and the mass was estimated using SVL/mass regression equations (Deichmann et al. 2008). Biomass per plot was calculated as the sum of the actual and estimated mass of all individuals encountered in the plot. We excluded only non-secondary consumers and non-leaf litter species from the analyses.

Data Analyses

Unless otherwise noted, we used SAS 9.1.2 (Cary, North Carolina) to conduct all statistical analyses.

We used multiple regression to test for significant effects of environmental variables on the abundance and biomass of herpetofauna found in the plots both for young soil sites in

Ecuador and ancient soil sites in Brazil. We used PROC REG with stepwise selection and

Akaike Information Criterion (AIC) in order to choose the best regression model for predicting leaf litter herpetofauna biomass and abundance from the environmental variables at young soil and at ancient soil sites. We chose P<0.15 as the criterion for inclusion of independent variables in the model.

PROC MIXED was used to test for differences in abundance and biomass of amphibians and reptiles, separately and combined. In order to assure no differences between sites within areas of similar soil age and to support our grouping of sites, we first tested for differences between all sites sampled. We then tested for differences between soils of different ages.

We estimated species richness using Estimate S (Colwell 2006). Individual based rarefaction curves with 95% confidence intervals were calculated for amphibians, lizards and the two groups combined.

Our hypotheses about differences in leaf litter herpetofauna between ancient and new soils are based in ecology, but if such differences exist they could simply reflect differences in phylogenetic histories of herpetofaunas of Ecuador and Brazil. We used two approaches to tease out any potential effects of phylogeny. First, we selected all frog and lizard species that are known to inhabit the leaf litter on ancient and new soils at the BDFFP, Tiputini and Yasuní, including species we did not find in our own sampling. Using the literature, we determined the average adult size for males and females of each species of frog, and because lizard size is less well documented, we found the average size of each lizard species irrespective of sex. We then used a two sample Kolmogorov-Smirnov test to test for differences in the sizes of frog and lizard species found in Ecuador versus Brazil. The size of the individuals we encountered in our

58 samples did not differ appreciably from the sizes reported in the literature. Second, we considered the average biomass contributed to the samples by genera which occur in both

Ecuador and Brazil (in-common) and those that are not common to all sites (not-in-common).

Genera were used as the criteria for commonality instead of species because of the 56 total species of amphibians and reptiles encountered in litter plots in Ecuador and Brazil, only 6 were common to both, whereas 10 of 26 genera of frogs and lizards were found in both countries

(Appendix 1). We used a factorial ANOVA to test for an effect of the interaction between country and commonality in the biomass of frogs and lizards in Ecuador and Brazil.

RESULTS

Environmental Variables

Of the 8 independent variables used in the multiple regression analysis, the best model for predicting abundance at ancient soil sites included canopy cover, leaf litter depth and number of trees in the quadrat (F3,255=4.50, P=0.004; Table 4.2). Despite significance, the coefficient of multiple determination was small (R2=0.050), suggesting that these variables are overall poor predictors of leaf litter herpetofauna abundance. For young soils, elevation was the only variable included in the model for predicting abundance (F1,195=3.33, P=0.069). Once again, very little variance in herpetofauna biomass is explained by this independent variable (R2=0.012). No significant models were selected for predicting herpetofauna biomass from environmental variables at either old or young soil sites. The multiple regression models given here were selected as the best-fit by both the stepwise and AIC selection methods.

Abundance and Biomass

Statistical analyses revealed no differences in abundance (F3,460=0.76, P=0.520) or biomass (F3,460=0.85, P=0.468) between sites within ancient and young soil regions. This result confirmed our grouping of sites within each soil region for comparison between regions.

A D 8 16 7 14

2 6 12

5 2 10 4 8

3 m g/100 6

Individuals/100 m Individuals/100 2 4 1 2 0 0 Central Brazil Ecuador Central Brazil Ecuador

B E 7 16

6 14 2 5 12

2 10 4 8 3 g/100m 6 2

Individuals/100m 4 1 2 0 0 Central Brazil Ecuador Central Brazil Ecuador

C F 1.6 2.5 1.4 2 2 1.2

1 2 1.5 0.8

0.6 g/100m 1

Individuals/100m 0.4 0.5 0.2 0 0 Central Brazil Ecuador Central Brazil Ecuador

Table 4.2 Multiple linear regression models for the relationship between environmental variables and leaf litter herpetofauna abundance and biomass at sites on old and young soils. Abbreviations: cc= canopy cover; elev = elevation; dep = litter depth; tree = number of trees

Model Adj R2 R2 P F (df) Total Abundance Old cc + dep + tree 0. 039 0.050 0.004 4.5 0 (3,255) Young elev 0.012 0.017 0.069 3.33 (1,195) Total Biomass Old none significant Young none significant

Combined amphibian and reptile abundance (F1,460=7.39, P=0.007) was significantly different between ancient and young soils (Figure 4.1). We found an average of 4.94 ± 0.75 herps/100m2 (Mean ± 95% CI) at ancient soil sites and 6.38 ± 0.96 at young sites. Separately, amphibians were less abundant on ancient soils (3.88 ± 0.62) than on young soils (5.20 ± 0.78;

F1,460=8.70, P=0.003). Lizard abundance was not different between the two soils (ancient soils

1.03±0.27, young soils 1.06±0.33; F1,460=0.00, P=0.996).

Overall secondary consumer herpetofaunal biomass was greater on younger soils

2 2 (11.37±3.84 g/100m ) than on ancient soils (4.25±1.84 g/100m ; F1,460=37.00, P<0.001).

Amphibians had more than twice as much biomass at sites on young soils (young soils 9.83±3.77

2 2 g/100m , ancient soils 3.99±1.85 g/100m ; F1,460=26.75, P<0.001). Lizard biomass on young soils greatly exceeded that on ancient soils (young soils 1.41±0.85 g/100m2, ancient soils

2 0.27±0.17 g/100m ; F1,460=16.72, P<0.001).

Phylogenetic Consideration

We found no significant differences in size distribution of male or female frogs of species common to our sites in Ecuador and Brazil (males, P=0.694; females, P=0.909; Figure 4.2), nor did we find significant differences in lizard size distribution (P=0.0.793; Figure 4.3). We also

61 found no effect of the interaction between country and commonality of genera on frog

(F1,925=1.85, P=0.175) or lizard biomass (F1,925=2.06, P=0.152).

Species Richness

Species accumulation curves showed significantly greater richness at young soil sites, with no overlap of 95% confidence intervals, for overall herpetofauna as well as amphibians and lizards separately (Figure 4.4).

DISCUSSION

Sites on young soils support higher biomass, abundance and species richness within the leaf litter herpetofauna community. The measured microhabitat variables appear to play little role in determining the differences displayed in these regions. Appropriate models show independent variables explained only 5% of the variation in abundance on ancient soils and 1% of abundance on young soils. Furthermore, no combination of the measured environmental variables explains the differences in biomass at ancient or young soil sites. It appears that soil age is a much better predictor of the leaf litter herpetofauna community.

A number of hypotheses have been put forth to explain the well-documented differences in leaf litter herpetofauna abundances between Central America and Southeast Asia (Scott 1976,

Inger 1980). Some of these could potentially be alternatives to our geological age hypothesis.

For example, perhaps an increased number of predators or competitors in Brazil relative to

Ecuador could explain lowered herpetofauna abundances on ancient soils. Another alternative is increased accumulation of leaf litter at the Ecuadorian sites that could lead to higher herpetofauna densities through augmented structure for shelter and reproductive sites. Although both these hypotheses may provide reasonable explanations for the differences between

Southeast Asian and Central American litter herpetofauna abundances, they cannot explain the differences seen in this study. While the number of snakes we encountered in our quadrats was

A B 18 18 16 16 14 14 12 12 10 10

8 8

Frequency Frequency 6 6 4 4 2 2 0 0 20 40 60 80 100 120 140 More 20 40 60 80 100 120 140 More C D 16 16 14 14 P = 0.47 12 12 10 10

8 8 Frequency Frequency 6 6 4 4 2 2 0 0 20 40 60 80 100 120 140 More 20 40 60 80 100 120 140 More SVL (mm) SVL (mm)

Figure 4.2 Size ranges of frog species in Ecuador and Brazil: (A) males, Ecuador; (B) males, Brazil; (C) females, Ecuador; (D) females, Brazil.

A 10 9 8 7 6 5

4 Frequency 3 2 1 0 25 50 75 100 125 150 More

B 9 8 7 6 5 4

Frequency 3 2 1 0 25 50 75 100 125 150 More SVL (mm)

Figure 4.3 Size range of lizard species in (A) Ecuador and (B) Brazil. low at all sites, more were found in plots on new soils. Similarly, spider abundances were greater in plots on young soils. These data indicate that predator and competitor densities are likely increased in Ecuador, and therefore cannot account for the herpetofauna differences on young and old soils. Furthermore, leaf litter depth was not different between quadrats sampled in Ecuador and Brazil, providing no indication of increased structure on young soils.

Primary productivity differences between the ancient and young soils have come to light via recent studies on forest dynamics (Malhi et al. 2004, Phillips et al. 2004, ter Steege et al.

2006), thus confirming Sombroek's (2000) view that the ancient soils of the Amazon are 64

A 40

16 Number species Number of 8

0 0 50 100 150 200 250 300

B 30

10 Number Number species of 5

0 0 50 100 150 200 250

C 14

4 Number species Number of

0 0 10 20 30 40 50 60 70 Number of individuals

65 extremely poor in weatherable minerals. Such a difference at the base of the food chains portends differences at higher trophic levels, but how such differences are mediated is unknown.

As tree mortality and recruitment in the western Amazon are about twice that in the central and eastern Amazon (Phillips et al. 2004), how does productivity and biomass scale up to herbivores and predators?

Two principles argue against simple linear transitions of primary productivity up the food chain. First, the allocation of energy by plants to growth and reproduction is non-linear.

Plants often provide energy for growth before devoting energy to reproduction. Where productivity is limited, reproduction may be minimal or delayed until sufficient resources have been accumulated. Alternatively, where productivity is enhanced, there may be a disproportionate shift toward reproduction as the needs of growth have all been met and excess production is channeled into flowers and fruits. It is the flowers, fruits and seeds that provide the most nutritious resources for herbivores. The actual differences in "productivity" recorded in recent studies of forest dynamics across the Amazon were measured as differences in vegetative growth without regard for reproductive effort. Quite probably, differences in production of flowers and fruits could have been even more exaggerated. Although many studies have shown litter to increase in tropical forests under artificial fertilization regimes (Herbert and Fownes

1995, Mirmanto et al. 1999, Yang et al. 2007), few studies have divided the litter into reproductive and vegetative components. However, one very large, long term soil fertilization experiment in Panama on relatively mineral rich soils produced no increase in litter fall of leaves and twigs, but a 43% increase in reproduction (Kaspari et al. 2008).

Second, declines in primary productivity may result in local extinction of species at higher trophic levels. If so, then a given percent decline in primary production may trigger a greater decline in secondary production. The relationship between productivity and species

66 richness is largely unexplored in tropical taxa; however, it has been hypothesized that large vertebrates with small population sizes may be subject to extirpation where productivity reduces population density below a viable minimum threshold (Wright 1983). Data comparisons for primates (Kay et al. 1997, Peres and Dolman 2000), mammals (Emmons 1984), and fishes

(Henderson and Crampton 1997), all strongly suggest that reduced primary productivity in the

Central and Eastern Amazon results in declines in species richness. However, the methodologies used in these studies were not consistent and the sampled sites differed in a number of confounding factors. Furthermore, in the most standardized comparison, dung beetles showed no difference in species richness despite higher abundance and biomass in Ecuador than in

Central Brazil (Radtke et al. 2007). Prior arguments have suggested that arthropod populations, in contrast to mammals, may not dip below critical minima when primary production is reduced.

Arthropods, occupying lower trophic levels and having smaller body sizes, are expected to maintain larger populations.

Within the conservative parameters of 95% confidence intervals (Payton et al. 2003), our data illustrate a striking difference in species richness of secondary consumers (frogs and lizards) between sites of differing productivity. Therefore, even at this lower trophic level, the difference in available energy generates species richness differences. Here we used standardized measurements of litter herpetofauna biomass with a common methodology in undisturbed primary forest tracts of at least 10,000 ha that share a comparable climate, the sites differing primarily in landform, as defined by Sombroek (2000).

To our knowledge, the only other study to use standardized methodology in a comparison of fauna on ancient vs. young soils involved dung beetles, which were used as an indicator of mammal biomass (Radtke et al. 2007). The authors of this study found a nearly three-fold increase in dung beetle biomass on young soils. Because most dung beetles depend primarily on

67 the excrement of large mammals including primates, it is reasonable to assume that mammal biomass follows the pattern exhibited by dung beetles. Although the degree to which mammal biomass increases on younger soils cannot be inferred directly from the dung beetle study, a recent meta-analysis of primate biomass over a range of Amazonian soil fertilities suggested a five-fold increase between the least and most fertile soil sites examined (Peres 2008). If dung beetles (detritivores), primates (herbivores and omnivores) and leaf litter herpetofauna

(insectivores) are all affected by differences in soil fertility and primary productivity, this effect is potentially felt at all levels of the ecological pyramid.

While the biomass of amphibians on young soils was twice that on older soils, lizard biomass was nearly five times greater on young soils. This difference in response strength may reflect ontogenetic differences in these taxa with lizards spending their entire lives dependent on the leaf litter ecosystem, while many amphibians do not begin their lives feeding in the litter.

Tropical frogs use a variety of habitats for reproduction that temperate frogs do not (Duellman and Trueb 1986). Just a few of these include the use of arboreal phytotelmata, creation of terrestrial foam nests with non-feeding tadpoles, and terrestrial or arboreal nests with direct developing eggs (Crump 1971, Duellman and Trueb 1986, Hödl 1990). Among tadpoles that do feed, some are carnivorous or oophagous, but, although macroinvertebrates can occasionally be found in the guts of many species, most tropical larval anurans are herbivores, feeding primarily on algae and detritus in the water column (Dutra and Callisto 2005). Consequently, tadpole survival is generally independent of resources on the forest floor.

Overall herpetofauna biomass was nearly 3 times greater on young soils, far exceeding the difference exhibited by abundance. Biomass reflects energy movement through a system and is more responsive to plant productivity than raw numbers of individuals. For this reason, it should be no surprise that biomass exhibits a stronger response to differences in soil age than

68 abundance. This phenomenon is even more pronounced in lizards which showed no difference in abundance, but greatly increased biomass on young soils. The majority of the aforementioned studies showing trends of increased densities on younger soils focused on differences in abundance and disregarded biomass. Our study suggests that important community differences may be overlooked when focusing exclusively on abundance and that the noted differences may be even more drastic when biomass is taken into account. Biomass is a source of energy which is transferred through the ecosystem by means of various trophic interactions (Saint-Germaine et al. 2007). Truly efficient conservation regimes can only be developed using this type of data which links the constituents of the ecological community.

In addition to looking at varying effects of forest productivity, variations in biomass, rather than simply abundance, may be of particular interest in the study of amphibian decline.

The loss of anuran consumers from an area has obvious ecological implications (Ranvestel et al.

2004, Whiles et al. 2006). From the perspective of the ecological community, it would be interesting to know if impervious species compensate in some way, either in number or biomass, for the loss of vulnerable species from a region. That is to say, over time can resilient species fill the ecological roles voided by species more susceptible to the causes of amphibian decline?

We cannot ignore the possibility that the biomass variation seen within the leaf litter herpetofauna community may be due to phylogeography. Our analyses, however, suggest that although different anuran and lizard species occur at sites on ancient and young soils, there is no difference in the size distributions of these species and hence, no phylogenetic constraints on size. Frog and lizard species on old soils in Brazil attain the same SVL as the frog species on young soils in Ecuador. Additionally, the average biomass contributed by in-common and not- in-common frog and lizard genera to the total biomass sample is consistent between countries, suggesting that phylogeny is not a confounding factor in this study.

These results have major implications for future conservation efforts across the Amazon

Basin. Proportionally larger tracts of land in the old Amazon may be required to maintain comparable biomass and abundance (population densities) of herpetofauna than on smaller tracts of land in the young Amazon. This area effect may be especially important as landscapes become increasingly fragmented: for a given sized fragment there will be lower population sizes on the ancient Amazonian soils than on the young ones. This information must be incorporated in the process of developing practical and responsible conservation and management of litter herpetofauna in the Amazon Basin, the principles behind which may be applicable to all organisms within the ecosystem.

LITERATURE CITED

Allmon, W. D. 1991. A plot study of forest floor litter frogs, Central Amazon, Brazil. Journal of Tropical Ecology 7:503-522.

Becker, P., J. S. Moure, and F. J. A. Peralta. 1991. More about Euglossine bees in Amazonian forest fragments. Biotropica 23:586-591.

Colwell, R. K. 2006. EstimateS: Statistical estimation of species richness and shared species from samples. http://viceroy.eeb.uconn.edu/EstimateS. Version 7.5.

Crump, M. L. 1971. Quantitative analysis of the ecological distribution of a tropical herpetofauna. Occasional Papers of the Museum of Natural History, The University of Kansas 3:1-62. de Oliveira, A. A. and S. A. Mori. 1999. A central Amazonian terra firme forest. I. High tree species richness on poor soils. Biodiversity and Conservation 8:1219-1244.

Deichmann, J. L., W. E. Duellman, and G. B. Williamson. 2008. Predicting biomass from snout- vent length in New World frogs. Journal of Herpetology 42:238-245.

Duellman, W. E. 1995. Temporal fluctuations in abundances of anuran amphibians in a seasonal Amazonian rain-forest. Journal of Herpetology 29:13-21.

Duellman, W. E. and L. Trueb. 1986. Biology of Amphibians. McGraw-Hill Book Company, New York.

Dutra, S. L. and M. Callisto. 2005. Macroinvertebrates as tadpole food: importance and body size relationships. Revista Brasileira de Zoologia 22:923-927.

Emmons, L. H. 1984. Gographic variation in densities and diversities of non--flying mammals in Amazonia. Biotropica 16:210-222.

Evans, K. L., P. H. Warren, and K. J. Gaston. 2005. Species-energy relationships at the macroecological scale: a review of the mechanisms. Biological Reviews 80:1-25.

Gascon, C. and R. O. J. Bierregaard. 2001. The Biological Dynamics of Forest Fragmentation Project: The study site, experimental design, and research activity.in R. O. J. Bierregaard, C. Gascon, T. E. Lovejoy, and R. Mesquita, editors. Lessons from Amazonia. Yale University Press, New Haven.

Gregory-Wodzicki, K. M. 2000. Uplift history of the Central and Northern Andes: A review. Geological Society of America Bulletin 112:1091-1105.

Henderson, P. A. and W. G. R. Crampton. 1997. A comparison of fish diversity and abundance between nutrient-rich and nutrient-poor lakes in the Upper Amazon. Journal of Tropical Ecology 13:175-198.

Herbert, D. A. and J. H. Fownes. 1995. Phosphorus limitation of forest leaf-area and net primary production on a highly weathered soil. Biogeochemistry 29:223-235.

Hödl, W. 1990. Reproductive diversity in Amazonian lowland frogs. Fortschritte der Zoologie 38:41-60.

Inger, R. F. 1980. Densities of floor-dwelling frogs and lizards in lowland forests of Southeast Asia and Central America. The American Naturalist 115:761-770.

Karubian, J., J. Fabara, D. Yunes, J. P. Jorgenson, D. Romo, and T. B. Smith. 2005. Temporal and spatial patterns of macaw abundance in the Ecuadorian Amazon. The Condor 107:617-626.

Kaspari, M., M. N. Garcia, K. E. Harms, M. Santana, S. J. Wright, and J. B. Yavitt. 2008. Multiple nutrients limit litterfall and decomposition in a tropical forest. Ecology Letters 11:35-43.

Laurance, W. F., L. V. Ferreira, J. M. R.-d. Merona, and S. G. Laurance. 1998. Rain Forest Fragmentation and the Dynamics of Amazonian Tree Communities. Ecology 79:2032- 2040.

Lieberman, S. S. 1986. Ecology of the leaf litter herpetofauna of a neotropical rain forest: La Selva, Costa Rica. Acta Zoologica Mexicana 15:1-72.

May, R. 1980. Why are there fewer frogs and lizards in Southeast Asia than in Central America? Nature 287:105.

Mirmanto, E., J. Proctor, J. Green, L. Nagy, and Suriantata. 1999. Effects of nitrogen and phosphorus fertilization in a lowland evergreen rainforest. Philosophical Transactions of the Royal Society B: Biological Sciences 354:1825-1829.

Payton, M. E., M. H. Greenstone, and N. Schenker. 2003. Overlapping confidence intervals or standard error intervals: what do they mean in terms of statistical significance? Journal of Insect Science 3:34-39.

Peres, C. A. 2008. Soil fertility and arboreal mammal biomass in tropical forests. Pages 349-364 in W. P. Carson and S. A. Schnitzer, editors. Tropical Forest Community Ecology. Blackwell Publishing, Oxford.

Peres, C. A. and P. M. Dolman. 2000. Density compensation in neotropical primate communities: evidence from 56 hunted and nonhunted Amazonian forests of varying productivity. Oecologia 122:175-189.

Radtke, M. G., C. R. V. da Fonseca, and G. B. Williamson. 2007. The old and young Amazon: dung beetle biomass, abundance and species diversity. Biotropica 39:725-730.

Rankin-de Merona, J. M., J. M. Prance, R. W. Hutchings, M. F. Silva, W. A. Rodrigues, and M. E. Uehling. 1992. Preliminary results of a large-scale inventory of upland rain forest in the central Amazon. Acta Amazonica 22:493-534.

Ranvestel, A. W., K. R. Lips, C. M. Pringle, M. R. Whiles, and R. J. Bixby. 2004. Neotropical tadpoles influence stream benthos: evidence for the ecological consequences of decline in amphibian populations. Freshwater Biology 49:274-285.

Saint-Germaine, M., C. M. Buddle, M. Larrivée, A. Mercado, T. Motchula, E. Reichert, T. E. Sacket, Z. Sylvain, and A. Webb. 2007. Should biomass be consiered more frequently as a currency in terrestrial arthropod community analyses? Journal of Applied Ecology 44:330-339.

Scott, N. J. 1976. The abundance and diversity of the herpetofaunas of tropical forest litter. Biotropica 8:41-58.

Sombroek, W. 2000. Amazon landforms and soils in relation to biological diversity. Acta Amazonica 30:81-100.

Sombroek, W. 2001. Spatial and temporal patterns of Amazon rainfall - consequences for the planning of agricultural occupation and the protection of primary forests. Ambio 30:388- 396. ter Steege, H., N. C. A. Pitman, O. L. Phillips, J. Chave, D. Sabatier, A. Duque, J.-F. Molino, M.-F. Prévost, R. Spichiger, H. Castellanos, P. v. Hildebrand, and R. Vásquez. 2006. Continental-scale patterns of canopy tree composition and function across Amazonia. Nature 443:444-447.

Terborgh, J. and E. Andresen. 1998. The composition of Amazonian forests: patterns at local and regional scales. Journal of Tropical Ecology 14:645-664.

Toft, C. A. 1980. Feeding ecology of thirteen sympatric species of anurans in a seasonal tropical environment. Oecologia 45:131-141.

Valencia, R., H. Balslev, and C. G. Paz y Mino. 1994. High tree alpha-diversity in Amazonian Ecuador. Biodiversity and Conservation 3:21-28.

Valencia, R., R. G. Condit, R. B. Foster, K. Romoleroux, G. Villa Muñoz, J. C. Svenning, E. Magard, M. Bass, E. C. Losos, and H. Balslev. 2004a. Yasuní Forest Dynamics Plot, Ecuador. Pages 609-628 in E. C. Losos and E. G. J. Leigh, editors. Tropical forest diversity and dynamism: Findings from a large-scale plot network. University of Chicago Press, Chicago.

Valencia, R., R. B. Foster, G. Villa, R. Condit, J. C. Svenning, C. Hernandez, K. Romoleroux, E. Losos, E. Magard, and H. Balslev. 2004b. Tree species distributions and local habitat variation in the Amazon: large forest plot in eastern Ecuador. Journal of Ecology 92:214- 229.

Van Wambeke, A. 1992. Soils of the Tropics. Their Properties and Appraisal. McGraw-Hill, New York.

Wright, D. H. 1983. Species-energy theory: an extension of species-area theory. Oikos 41:496- 506.

Yang, X. D., M. Warren, and X. M. Zou. 2007. Fertilization responses of soil litter fauna and litter quantity, quality, and turnover in low and high elevation forests of Puerto Rico. Applied Soil Ecology 37:63-71.

CHAPTER 5 NEOTROPICAL LEAF LITTER HERPETOFAUNA ACROSS A GRADIENT OF SOIL FERTILITY

INTRODUCTION

Quadrat samples, or litter plots, have been used extensively by herpetologists in tropical regions (Jaeger & Inger 1994). As a result, large collections of specimens have amassed in museums, culminating in the creation of extensive data sets for some areas, particularly Central

America and Southeast Asia. Using data sets from Costa Rica, Panama, the Philippines and

Borneo, Scott (1976) found abundances of litter frogs and lizards in the tropical lowland wet forests of Central America to be an order of magnitude greater than in Southeast Asia, although species richness was not different between the two regions. Later, Allmon (1991) suggested that the abundance pattern shown for Central America was not characteristic of all the Neotropics.

He found low litter frog abundances in the Brazilian Amazon relative to studies in the Peruvian

Amazon and Central America (Allmon 1991). Many hypotheses have been proposed to explain these discrepancies in abundance without satisfactory support for any particular one (Allmon

1991; Inger 1980; May 1980; Scott 1976). For all these comparisons, abundance and species richness were the only response variables of interest. Furthermore, each study employed its own sampling methodology, therein precluding a standardized and comprehensive test of any hypothesis accounting for community differences in leaf litter herpetofauna.

Only recently has a convincing argument to explain these discrepancies been tested.

Deichmann et al. (in prep; Chapter 4) demonstrated that the geological age of soils influences the leaf litter herpetofauna community in Amazonian lowland rainforests, with higher biomass, abundance and species richness on younger, more productive soils. How applicable are these findings to herpetofauna communities in lowland wet forests throughout the Neotropics?

South America has experienced a complex geologic history. After separating from

Gondwana, the continent of South America was isolated for well over 50 million years during which time it underwent a major orogenic event, the beginning of the uplift of the Andes. The rise of the Andes provided new parent material from which the bulk of the soils of the Western

Amazon Basin have been derived (Gregory-Wodzicki 2000; Sombroek 2000). On the other hand, soils in Central and Eastern Amazonia originated in the Guiana Shield and the Brazilian

Highlands and are ancient, low nutrient soils (Sombroek 2000). Within these ‗old‘ soil areas, there exists further division of soils based on geomorphology. Soils from Central to Eastern

Brazil, largely along the Amazon River, are composed of old Oxisols and have no remaining mineral reserves. Soils in the northern Central and Eastern parts of the Basin into the Guiana shield region, although still ancient, are more variable with transitions between Inceptisols,

Ultisols and Oxisols over short distances.

Because of the differences in soil age, we propose that leaf litter herpetofauna are more abundant and have higher biomass in Western South America with its younger soils than Central

Amazonia and the Guianan region with older soils. We also expect that there may be some differentiation within Eastern South America, with the intermediate soils of the Guiana region being able to support higher biomass and abundance than the older soils of Central Brazil. These density and biomass differences may indirectly affect species richness through population size, where reduced densities could push populations below a minimum viable threshold. Here, we use data from 11 different Neotropical lowland rainforest sites to investigate patterns in leaf litter herpetofauna communities. Unlike previous comparisons, we standardize these data sets to include only data collected during a single season (wet) from a single forest type (terra firme) using quadrats of similar size (5x5 or 6x6m).

METHODS

Sites

Abundance, biomass and species richness data were compiled from various South

American lowland forest sites including Cabo Frio, Dimona and Km 41 at the Biological

Dynamics of Forest Fragments Project (BDFFP) near Manaus, Brazil, Tiputini Biodiversity

Station and Yasuní Research Station in Orellana Province, Ecuador, Pararé and Inselberg sites at

Nouragues Biological Station in Central French Guiana, Panguana Biological Station in Huánuco

Department, Peru and Pipeline road, Carti Road and Rio Guanche, all in Colon Province,

Panama (Table 5.1). Although Panama is technically part of Central America, we included it in these comparisons because it does share some leaf litter herpetofauna with South American sites and it lies on young volcanic soils. Site descriptions and basic data collection methods can be found in the following: Brazil, Ecuador - Deichmann et al. in prep; French Guiana - Chapter 1;

Panama - Toft 1980b; Peru - Toft 1980a. Data on anurans were available for all sites; however, only 7 of the 11 sites included collection of data on lizards (Table 5.1).

A problem inherent in comparative studies, especially over large geographical regions, is that they are often confounded by a number of extraneous factors. Here, we have edited all data sets to include only diurnally sampled quadrats of similar size on terra firme forest during the wet season. In this way we hope to exclude potential complications due to variation in time of day, quadrat size, habitat type and seasonality.

Biomass Estimates

Data sets from Ecuador, Brazil and French Guiana have accurate mass measurements for individual frogs and lizards (measured with an Ohaus Scout Pro electronic balance to 0.01g). In these cases, biomass per plot was calculated as the sum of the actual mass of each individual found in a plot.

Table 5.1 List of sites included in this study. N represents the number of plots conducted on primary terra firme forest in the wet season at each site. Site Country Latitude Elevation (m) Rainfall (mm) Plot Size (N) Data Available Cabo Frio Brazil 2° 24‘ S 100 2600 5x5 (60) Frogs and lizards Dimona Brazil 2° 2‘ S 100 2600 5x5 (100) Frogs and lizards Km 41 Brazil 2° 26‘ S 100 2600 5x5 (105) Frogs and lizards Tiputini Ecuador 0° 37‘ S 200 2800 5x5 (100) Frogs and lizards Yasuní Ecuador 0° 40' S 200 2800 5x5 (100) Frogs and lizards Pararé French Guiana 4° 2' N 50 3000 5x5 (50) Frogs and lizards Inselberg French Guiana 4° 5' N 100 3000 5x5 (50) Frogs and lizards Panguana Peru 9° 35' S 200 2200 6x6 (12) Frogs Carti Road Panama 9° 20' N 300 3500 6x6 (5) Frogs Pipeline Road Panama 9° 5' N 30 2200 6x6 (7) Frogs Rio Guanche Panama 9° 30' N 10 3000 6x6 (3) Frogs

For all other sites, the mass of frogs and lizards was estimated from snout-vent length

(SVL). In most cases, snout-vent length was measured for each captured individual. For individuals with a measured SVL, but no mass, we used species-specific SVL/mass equations to estimate individual mass (Deichmann et al. 2008). For some species, where a species-specific equation was not available, we used equations derived from the genus or family. Plot biomass was then calculated by adding the estimated mass of each individual encountered in a given plot.

During litter plot sampling, it is inevitable that some individuals will escape before capture. For this reason, there were no measurements for some recorded individuals in the data sets; however several of these escapees were recorded as ‗juveniles‘ or ‗adults‘. In order to estimate mass for this entire subset of individuals (those with life stage status specified and those without), we found the average size of all individuals of the species at the site, the average of only juveniles and the average of only adults. We then applied the average juvenile mass to the

‗juveniles‘, the average adult mass to the ‗adults‘, and the overall average mass to the individuals for which neither SVL, mass or life stage was recorded. This method follows other biomass studies which have used the average mass of a species and multiplied it by the species‘ abundance, given the difficulty of capturing all individuals for more accurate measurements (ex.

Peres & Dolman 2000).

Unless otherwise noted, we used SAS 9.1.2 (Cary, North Carolina) to conduct statistical analyses.

Biomass and Abundance

Because our interest lies in differences in herpetofauna abundance and biomass between regions of lowland rainforest, we wanted to group sites by country. In order to do this, we first tested for differences in biomass or abundance using site nested in country as a factor (PROC

MIXED). If no significant effect of site was found within countries, abundance and biomass

79 were compared over the main effect of country alone. Where biomass or abundance was significantly different within the analysis, differences between sites or countries were examined using Tukey multiple comparisons.

Size Distributions

In order to investigate differences in the size structure of the leaf litter communities between countries, we compared the size distribution of individuals found in plots in each region with a Chi square test for independence (PROC FREQ). We ran this analysis separately for frogs and for lizards where data were available.

Species Richness

Estimate S (Colwell 2006) was used to calculate individual based rarefaction curves with

84% confidence intervals for frogs and lizards. For the most comprehensive test of the species richness in each area, we included all samples from terra firme forest and did not exclude dry season samples. This raises sample sizes for Panguana, Carti Road and Pipeline Road provided in Table 1.1 to 24, 8 and 11 respectively.

Community Similarity

We also used Estimate S to determine the species similarity between sites. Because we do not have lizard data for all sites, only anuran species were used in this analysis. We determined the Bray-Curtis similarity indices between all sites and converted them to a distance matrix. PROC CLUSTER was used to run an average linkage cluster analysis in SAS to determine the most closely related sites in terms of species composition. Here we included leaf litter frog data from La Selva Biological Station in Costa Rica as a Central American site to help root the dendrogram.

RESULTS

Biomass and Abundance

For amphibian abundance, there was no significant effect of site within country

(F7,580=1.47, P=0.174), but there was a significant difference in amphibian abundance between countries (F4,580=16.04, P<0.001), with the Panamanian and Peruvian sites having higher abundance than all other sites (Figure 5.1A). Ecuador frogs were more abundant than Brazil frogs (t=2.56, df = 580, Tukey adj. P=0.033) with French Guiana intermediate. For lizard abundance, there was no significant effect of site within country (F5,557=1.75, P=0.122) or country (F2,557=0.01, P=0.992; Figure 5.1B).

For anuran biomass, there was no significant effect of site within country (F7,580=1.38,

P=0.212), but there was a significant country effect (F4,580=4.61, P=0.001) with Ecuador having significantly higher biomass than Brazil. Although not significant, the data show a trend parallel to abundance with increased biomass in Western South America relative to Central Brazil with

French Guiana intermediate (Figure 5.1C). For lizard biomass, there was no significant effect of site within country (F5,557=1.00, P=0.418), but there was a significant country effect (F2,557=5.32,

P=0.005) with Ecuadorian lizard biomass significantly greater than that of Brazil, and French

Guiana intermediate (Figure 5.1D).

Size Distributions

The size distributions of frogs and lizards found within each country showed significant different (frogs, Χ2=130.74, df=24, P<0.001, Table 5.2; lizards, Χ2=58.79, df=10, P<0.001,

Table 5.3). The frog size distribution in Brazil is highly skewed toward small frogs (Figure 5.2).

Panama, Peru and French Guiana have similar size class distributions with the most frogs in the samples between 10-20mm SVL and some large (>40mm) frogs. Ecuador has a relatively even distribution of frogs across size classes with a peak in frogs between 15-20mm SVL.

A 25.00 B 1.60 A B 1.40 20.00 1.20

15.00 1.00 0.80 C CD D 10.00 0.60

Number lizardsNumber of 0.40 Number amphibians Number of 5.00 0.20

0.00 0.00 Panama Peru Ecuador French Guiana Brazil Ecuador French Guiana Brazil

C 16.00 AB AB D 2.50 A 14.00 2.00 AB 12.00 A AB 10.00 1.50 8.00 B 6.00 1.00 B 4.00

Amphibian Amphibian biomass (g) 0.50 2.00 Lizard biomass (g/100m2)

0.00 0.00 Panama Peru Ecuador French Brazil Ecuador French Guiana Brazil Guiana

Figure 5.1 Average abundance with 95% confidence intervals of (A) frogs and (B) lizards and biomass of (C) frogs and (D) lizards in plots in South America. Groups with distinct letters indicate significant differences in group means detected by post-hoc Tukey comparisons.

A B 50.00 50.00 40.00 40.00 30.00 30.00 20.00 20.00

10.00 10.00

Relative Frequency Relative Relative Frequency Relative 0.00 0.00 0-10 10-15 15-20 20-25 25-30 30-40 >40 0-10 10-15 15-20 20-25 25-30 30-40 >40

C D 50.00 50.00 40.00 40.00 30.00 30.00 20.00 20.00

10.00 10.00

Relative Frequency Relative Relative Frequency Relative 0.00 0.00 0-10 10-15 15-20 20-25 25-30 30-40 >40 0-10 10-15 15-20 20-25 25-30 30-40 >40 SVL (mm) E 50.00 40.00 30.00 20.00

10.00 Relative Frequency Relative 0.00 0-10 10-15 15-20 20-25 25-30 30-40 >40 SVL (mm)

Figure 5.2 Size distributions of frogs captured in plots in (A) Ecuador, (B) Peru, (C) French Guiana, (D) Panama and (E) Brazil. 83

Table 5.2 Contributions to the Χ2 values by size range (mm) and country for frogs found in litter plots. 0-10 10-15 15-20 20-25 25-30 30-40 >40 Brazil 17.06 0.09 0.86 0.40 3.82 4.80 0.30 Ecuador 0.02 13.96 0.00 2.15 6.56 1.13 3.74 French Guiana 15.07 0.55 6.53 1.42 0.23 4.68 0.05 Panama 7.75 24.54 0.39 2.65 0.09 0.17 1.56 Peru 0.17 6.58 0.39 2.65 0.43 0.01 1.56

Table 5.3 Contributions to the Χ2 values by size range (mm) and country for lizards found in litter plots. 0-20 20-25 25-30 30-35 35-40 >40 Brazil 11.84 0.88 2.79 3.66 1.56 7.23 Ecuador 10.32 1.05 2.09 2.72 2.56 6.67 French Guiana 2.30 0.04 0.80 1.07 0.00 1.18

A 60.00 B 60.00

40.00 40.00

20.00 20.00

Relative Frequency Relative Relative Frequency Relative 0.00 0.00 0-20 20-25 25-30 30-35 35-40 >40 0-20 20-25 25-30 30-35 35-40 >40

C 60.00

40.00

20.00 Relative Frequency Relative 0.00 0-20 20-25 25-30 30-35 35-40 >40 SVL (mm) Figure 5.3 Size distributions of lizards captured in plots in (A) Ecuador, (B) French Guiana, (C) Brazil.

For lizards, the Ecuador size distribution is skewed toward larger (>40mm SVL) individuals with fewer small lizards, while Brazil again has a disproportionate number of very

84 small (<20mm) lizards (Figure 5.3). French Guiana has a fairly even distribution of lizards across the range of sizes.

Species Richness

Brazil has the lowest species richness both for frogs and for lizards (Figure 5.4). Ecuador and Panama had the highest anuran species richness, whereas Brazil was the lowest and Peru and

French Guiana were intermediate. For lizards, French Guiana and Ecuador had higher species richness than Brazil (Figure 5.4).

Community Similarity

There is a clear grouping of South American sites in terms of species composition (Figure

5.5). Within South America, the Peruvian and Ecuadorian sites are the most similar. This

Western Amazonian group is then most closely allied with French Guiana sites, followed by sites in Brazil. The Panamanian sites group with the Central American site, but sites within Panama are more similar to one another than any is to La Selva.

DISCUSSION

Anuran Abundance and Biomass

Frog abundance in Brazil is significantly lower than all other sites with the exception of

French Guiana (Figure 5.1A). These sites are considered to be ‗old soil‘ sites. Likely due to large confidence intervals, we see surprisingly few significant differences in frog biomass across regions. Only Brazilian and Ecuadorian samples differ significantly from one another; however, the data do suggest a trend of higher biomass in Western South America relative to the Guianan region and Central Amazonia (Figure 5.1B).

The low biomass at the Brazilian and French Guianan sites is expected. Relative to the other sites, Central Amazonia and the Guianan region lie primarily on very old Oxisols which have no remaining mineral reserves (Sombroek 2000). These old soils support forests with

A 25 Panama Peru French Guiana 20 Ecuador Brazil

10 Number species Number of

0 0 50 100 150 200 250 300

12 Brazil Ecuador 10 French Guiana

Number species Number of 4

0 0 10 20 30 40 50 60 70 80

Number of Individuals Figure 5.4 Species rarefaction curves for (A) leaf litter frogs and (B) lizards.

Panguana

Tiputini

Yasuní

Pararé

Inselberg

Dimona

Km 41

Cabo Frio

Rio Guanche

Carti

Pipeline

La Selva

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Distance

Figure 5.5 Dendrogram of site similarity based on cluster analysis of Bray-Curtis similarity indices. Colors represent sites within countries: gray = Peru, blue = Ecuador, green = French Guiana, peach = Brazil, purple = Panama, yellow = Costa Rica.

87 lower productivity than forests on younger soils in Western Amazonia (Malhi et al. 2004). We expect low primary productivity in Central Amazonia and the Guianas to be reflected through reduced biomass at higher trophic levels (Deichmann et al. in prep). It is interesting to note that

French Guiana has a biomass of both frogs and lizards which is intermediate to Western South

America and Central Brazil. Although both are largely derived from ancient parent material, the soils between our sites in Brazil and French Guiana are distinct with French Guianan soils having more local variation in soil conditions and available minerals (Sombroek 2000). Soils at

Nouragues have a much higher cation-exchange capacity that those at the BDFFP (Fearnside &

Filho 2001; Grimaldi & Riéra 2001). For this reason, through primary productivity, we would expect our sites in French Guiana to be able to support a higher biomass of secondary consumers in the leaf litter community than Brazil, but less than Western South America. Although the biomass differences between sites in Panama, Peru, French Guiana and Brazil exhibited in this study are not statistically significant, we strongly believe that the trend shown warrants further investigation as it may reflect a true pattern which can only be unveiled through additional sampling.

Lizard Abundance and Biomass

No significant differences were found in lizard abundance; however lizard biomass was significantly higher in Ecuador than in Brazil. We suggest that younger soils in Central America and Western Amazonia relative to Central Amazonia can explain this difference. The sites sampled in French Guiana lie on soils that are intermediate in terms of their age and mineral content to the soils described above. The intermediate lizard biomass fits in nicely with the soil age hypothesis.

Size Distributions

In general, the data suggest that older soils sustain smaller frogs and lizards while younger soils tend to support more intermediate to large frogs. This pattern does not appear to be an artifact of phylogeny, because the average size attained by species known to occur at these sites does not differ (Deichmann et al. in prep, Chapter 4). Primary productivity is higher on younger soils. Consequently, the resource base available to secondary consumers should be greater and more individuals should be able to grow and survive to adulthood on younger soils.

In fact, in Ecuador and Brazil, juveniles make up 48 and 57% respectively of all frogs found in quadrats. Although the abundance of frogs is similar, adults constitute a smaller proportion of the sample on old soils. This suggests that juveniles on older soils may encounter an insufficient supply of resources necessary to reach the adult stage.

Species Richness

The pattern shown by these leaf litter frog data sets largely reflects what is already known about general patterns of herpetofauna species richness in Central and South America. For example, Western Amazonia is known to have the highest anuran species richness among the lowland forests of Central and South America (Duellman 1999). The Ecuadorian data in this study support this for leaf litter herpetofauna; however, the Peruvian data set displays rather low species richness which is not representative of the generally high anuran diversity known from

Upper Amazonia (Duellman 1988). Although we used individual based accumulation curves to try to reduce the effect of sampling effort on the comparison, the low species richness in

Panguana (Peru) may be due to the low sample size (N=24).

According to these data, Panamanian frog species richness rivals that of Western

Amazonia (Figure 5.4A). Whether due to temporal or ecological restrictions, Panama is the northern limit of some species of South American herpetofauna and the southern limit of many

Middle American species (Duellman 1988; Savage 1982). Some of these include the leaf litter inhabiting species belonging to the Dendrobatid and Leptodactylid families (Campbell 1999).

Therefore, it is no surprise that when looking only at the guild of leaf litter frogs, Panama has high species richness. However, the exceedingly high richness shown here, nearly equaling that of the Ecuadorian samples, could be due to the fact that three different sites were included in the

Panama data set. The inclusion of more sites, though all within the same province, may result in increased species richness by broadening the sample area (Rosenzweig 1995).

Community Similarity

Amazonian sites are more similar in species composition to one another than they are to

Panamanian sites. This supports a previous analysis on community similarity between BCI,

Panama, Santa Cecilia, Ecuador, Manu, Peru and Manaus, Brazil (Duellman 1990). Slightly surprising is the closer association of the leaf litter frog community at sites in French Guiana to that of Western Amazonia rather than its geographic nearest neighbors, the Brazilian sites. This distinction differs from a previous analysis by Ron (2000) which shows a closer affiliation of species composition between sites in French Guiana and Central Brazil with a distant relation of that cluster to Western Amazonia. In his study, Ron (2000) considered the entire anuran species assemblage at each site. Some research has suggested that analyses of entire assemblages may obscure patterns that are guild-specific because different functional groups of anurans display different regional correlation patterns in terms of community composition (Ernst et al. 2006;

Ernst & Rödel 2008). Perhaps our data, which include only species found in the leaf litter, suggest some underlying similarity of the leaf litter guild between Western Amazonia and

French Guiana.

Alternatively, perhaps our community similarity results are obscured by cryptic diversity.

Although there are some pervasive Amazonian species with little genetic differences among

90 geographic regions, high levels of divergence exist within many species previously considered to be widespread (Fouquet et al. 2007). In accordance with areas of species endemism in

Amazonia (Cracraft 1985; Haffer 1985; Lynch 1979), Guiana and Central Brazil north of the

Amazon River should be more similar to one another than either site should be to the Napo region in Western Amazonia (Ecuador sites) or the Inambari region (Peruvian site). If we were able to base our analyses on genetic species distinctions rather than morphological ones, perhaps we would see a different pattern. For now, it appears that further studies are required to elucidate true patterns of community similarity.

Although leaf litter herpetofauna data sets exist for La Selva Biological Station in Costa

Rica and Barro Colorado Island (BCI) in Panama, we did not include them in abundance and biomass analyses here. We excluded the La Selva data primarily because recent evidence points to a decline in populations of leaf litter herpetofauna at that site (Whitfield et al. 2007).

Furthermore, Central America has a very complex geological history, with part of the continent from Southern Nicaragua to Colombia having been underwater for much of the Tertiary (Savage

1982). In addition to the old marine sediment underlying much of Costa Rica and Panama, this volcanic region has been covered by a number of lava flows, lahars and ash since its rise from the ocean (Sollins et al. 1994). Consequently, there is extreme heterogeneity in soil types throughout southern Central America with rapid changes in parent material over short distances

(Sollins et al. 1994; Yavitt 2000). Other sites within Central America would lend interesting additions to testing the hypothesis of soil age, but because the current state of leaf litter herpetofauna at La Selva appears to be affected by enigmatic decline, inclusion of the data here would have lead to inaccurate conclusions regarding the effects of soil age on abundance, biomass and species richness of frogs and lizards.

On BCI, a 1500-ha island, there are at least three different types of parent material, mostly Oligocene in origin. Although soils derived from these parent materials all sustain old growth tropical rain forest, they support different forest dynamics within a relatively small area

(John et al. 2007; Yavitt 2000). Because of the soil profile, BCI could potentially be an interesting addition to testing our hypotheses; however, BCI is known to have lower densities of leaf litter herpetofauna compared to mainland Panama (Heatwole & Sexton 1966). Although it has only been isolated for approximately 100 years, it is possible that BCI is transitioning to an ecosystem regulated by island dynamics. Populations on islands can have altered size structure

(Van Valen 1965) or unusual densities (Macarthur et al. 1972; Macarthur et al. 1973).

Alternatively, reduced densities of leaf litter herpetofauna could be a result of predation by unusually high densities of medium size mammals (Terborgh 1992). Regardless, inclusion of data from BCI would likely confound our tests for the effects of geomorphology on leaf litter herpetofauna community dynamics.

In the present study, we are limited by available data. Young soil sites, primarily in

Western South America, have been sampled much more extensively than old soil sites. In order to strengthen the evidence for geomorphological and primary productivity effects on leaf litter herpetofauna, it is necessary to sample systematically more sites throughout the lowland tropics and in particular on the geomorphologically ancient soils. Additionally, because of the extreme heterogeneity in parent materials in some areas such as Central America and the Guiana region, it would be prudent to systematically sample areas of different soils within countries for more direct investigations into the effects of geomorphology on leaf litter herpetofauna.

Although we took great care to compile data collected using relatively similar methods, it is difficult to distinguish true differences and trends because of inherent variation in sampling techniques. In developing our comparisons, some data sets had to be excluded because of

92 complications due to researcher bias in terms of skill and personal preferences for sampling leaf litter quadrats. This disorder highlights the need for communication between investigators in the data collection and data sharing processes as well as the need for caution when interpreting results of meta-analyses.

LITERATURE CITED

Allmon, W. D. 1991. A plot study of forest floor litter frogs, Central Amazon, Brazil. Journal of Tropical Ecology 7:503-522.

Campbell, J. A. 1999. Distribution patterns of amphibians in Middle America. Pages 111-210 in W. E. Duellman, editor. Patterns of Distribution of Amphibians: A Global Perspective. The Johns Hopkins University Press, Baltimore.

Colwell, R. K. 2006. EstimateS: Statistical estimation of species richness and shared species from samples. http://viceroy.eeb.uconn.edu/EstimateS. Version 7.5.

Cracraft, J. 1985. Historical biogeography and patterns of differentiation within the South American avifauna: areas of endemism. Ornithological Monographs 36:49-84.

Deichmann, J. L., W. E. Duellman, and G. B. Williamson. 2008. Predicting biomass from snout- vent length in New World frogs. Journal of Herpetology 42:238-245.

Duellman, W. E. 1988. Patterns of species diversity in anuran amphibians in the American tropics. Annals of the Missouri Botanical Garden 75:79-104.

Duellman, W. E. 1990. Herpetofaunas in Neotropical rainforests: comparative composition, history, and resource use. Pages 455-505 in A. H. Gentry, editor. Four Neotropical Rain Forests. Yale University Press, New Haven.

Duellman, W. E. 1999. Distribution patterns of amphibians in South America. Pages 255-327 in W. E. Duellman, editor. Patterns of Distribution of Amphibians: A Global Perspective. The Johns Hopkins University Press, Baltimore.

Ernst, R., K. E. Linsenmair, and M.-O. Rödel. 2006. Diversity erosion beyond the species level: Dramatic loss of functional diversity after selective logging in two tropical amphibian communities. Biological Conservation 133:143-155.

Ernst, R., and M.-O. Rödel. 2008. Patterns of community composition in two tropical tree frog assemblages: separating spatial structure and environmental effects in disturbed and undisturbed forests. Journal of Tropical Ecology 24:111-120.

Fearnside, P. M., and N. L. Filho. 2001. Soil and development in Amazonia: Lessons from the Biological Dynamics of Forest Fragments Project. Pages 291-312 in R. O. J. Bierregaard,

C. Gascon, T. E. Lovejoy, and R. Mesquita, editors. Lessons from Amazonia: The Ecology and Conservation of a Fragmented Forest. Yale University Press, New Haven.

Fouquet, A., A. Gilles, M. Vences, C. Marty, M. Blanc, and N. J. Gemmell. 2007. Underestimation of species richness in Neotropical frogs revealed by mtDNA analyses. PLoS ONE 2:e1109.

Gregory-Wodzicki, K. M. 2000. Uplift history of the Central and Northern Andes: A review. Geological Society of America Bulletin 112:1091-1105.

Haffer, J. 1985. Avian zoogeography of the Neotropical lowlands. Ornithological Monographs 36:113-146.

Heatwole, H., and O. J. Sexton. 1966. Herpetofaunal comparisons between two climatic zones in Panama. American Midland Naturalist 75:45-60.

Inger, R. F. 1980. Densities of floor-dwelling frogs and lizards in lowland forests of Southeast Asia and Central America. The American Naturalist 115:761-770.

Jaeger, R. G., and R. F. Inger. 1994. Quadrat Sampling. Pages 97-102 in W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L.-A. C. Hayek, and M. S. Foster, editors. Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Smithsonian Institution Press, Washington.

John, R., J. W. Dalling, K. E. Harms, J. B. Yavitt, R. F. Stallard, M. Mirabello, S. P. Hubbell, R. Valencia, H. Navarrete, M. Vallejo, and R. B. Foster. 2007. Soil nutrients influence spatial distributions of tropical tree species. Proceedings of the National Academy of Sciences 104:864-869.

Lynch, J. D. 1979. The amphibians of the lowland tropical forests. Pages 189-215 in W. E. Duellman, editor. The South American herpetofauna : its origin, evolution, and dispersal. Monographs of the Museum of Natural History. University of Kansas, Lawrence.

Macarthur, R. H., J. M. Diamond, and J. R. Karr. 1972. Density compensation in island faunas. Ecology 53:330-&.

Macarthur, R. H., J. Macarthur, D. Macarthur, and A. Macarthur. 1973. The effect of island area on population densities. Ecology 54:657-658.

Terborgh, B. Vinceti, and J. Lloyd. 2004. The above-ground coarse wood productivity of 104 Neotropical forest plots. Global Change Biology 10:563-591.

May, R. 1980. Why are there fewer frogs and lizards in Southeast Asia than in Central America? Nature 287:105.

Peres, C. A., and P. M. Dolman. 2000. Density compensation in neotropical primate communities: evidence from 56 hunted and nonhunted Amazonian forests of varying productivity. Oecologia 122:175-189.

Ron, S. R. 2000. Biogeographic area relationships of lowland Neotropical rainforest based on raw distributions of vertebrate groups. Biological Journal of the Linnean Society 71:379- 402.

Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge.

Savage, J. M. 1982. The enigma of the Central American herpetofauna: dispersals or vicariance? Annals of the Missouri Botanical Garden 69:464-547.

Scott, N. J. 1976. The abundance and diversity of the herpetofaunas of tropical forest litter. Biotropica 8:41-58.

Sollins, P., F. Sancho M., R. Mata Ch., and R. L. Sanford. 1994. Soils and soil process research. Pages 34-53 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva, a Nature Reserve and Field Station in Costa Rica. University of Chicago Press, Chicago.

Sombroek, W. 2000. Amazon landforms and soils in relation to biological diversity. Acta Amazonica 30:81-100.

Terborgh, J. 1992. Maintenance of diversity in tropical forests. Biotropica 24:283-292.

Toft, C. A. 1980a. Feeding ecology of thirteen sympatric species of anurans in a seasonal tropical environment. Oecologia 45:131-141.

Toft, C. A. 1980b. Seasonal variation in populations of Panamanian litter frogs and their prey: a comparison of wetter and drier sites. Oecologia 47:34-38.

Van Valen, L. 1965. Morphological variation and width of ecological niche. The American Naturalist 99:377-390.

Yavitt, J. B. 2000. Nutrient dynamics of soil derived from different parent material on Barro Colorado Island, Panama. Biotropica 32:198-207.

CHAPTER 6 A NOTE ON AMPHIBIAN DECLINE IN THE NEOTROPICAL LOWLANDS

Amphibian decline is a global phenomenon and is occurring at an alarming rate (Stuart et al. 2004). Typically, this enigmatic decline results in reductions or extinctions primarily in populations of stream-breeding species with ranges at mid to high elevations (Lips et al. 2005).

Decline has been attributed to a number of proximal causes (Pounds et al. 2006, etc., Alford et al.

2007, Di Rosa et al. 2007, Lips et al. 2008), although it is likely that more than one mechanism is driving amphibian decline (Collins and Storfer 2003).

Until recently, there was no evidence for declines in lowland rainforest populations of frogs within protected habitat in the Neotropics. Whitfield et al. (2007) studied quadrat data sets of leaf litter herpetofauna density spanning a 35 year period at La Selva Biological Station in

Costa Rica. They suggested that there has been a steady decline in populations of leaf litter frogs in primary forest at La Selva. They also found a parallel decline in lizards and suggested that these lowland declines in herpetofauna may be widespread.

To address the issue of lowland rainforest herpetofauna decline here, we present a comparison of data from sites in Central Brazil sampled in 1984-85 and again in 2007. As a bi- product of testing my productivity hypotheses, I was able to accumulate leaf litter herpetofauna data collected using quadrats at the Biological Dynamics of Forest Fragments Project (BDFFP) in 1984-85. Changes in leaf litter anuran abundance and biomass in Central Brazil were compared using data collected from continuous forest sites at the BDFFP. The BDFFP is divided into 3 main areas, or fazendas: Dimona, Porto Alegre and Esteio. Each of these is composed of a number of reserves. Initial collecting was done in 1984-85 using 5x5m quadrats

(Allmon 1991). Among reserves in continuous forest at the time of the study, Allmon sampled 2 at Dimona, 1 at Porto Alegre and 4 within Esteio. I sampled the BDFFP again in 2007, also

97 using 5x5m quadrats. Plots were located within continuous forest at the following sites: Dimona

(not within an established reserve), Cabo Frio within Porto Alegre and KM 41 within Esteio

(Deichmann et al. in prep). We compared the 1984-85 data set to data collected in the wet season of 2007. Although Allmon's 1984-85 data set encompasses samples taken throughout the year, only data from the wet season (November-May) were used in order to avoid seasonal bias.

For both sampling protocols, all frogs and lizards were captured and snout-vent length was measured (0.1mm). For the 2007 samples, individuals were also weighed (0.01g).

Biomass was calculated by summing the mass of all individuals found in each plot.

Because SVL but not mass was available for the 1984-85 samples, biomass was estimated using species-specific SVL/mass regressions. For individuals that were identified, but escaped capture, their mass was recorded as the average mass of all individuals of the species found at that site.

I conducted an overall comparison of abundance and biomass with the main effect of year and site nested in year in PROC MIXED. Although it is not an effect of main interest, I tested the nested effect of site in order to investigate differences within sites between years.

Because not all of the same reserves were sampled in both 1984-85 and 2007, I examined abundance and biomass at the sites that were sampled in both data sets: Km 41 in Esteio and

Cabo Frio in Porto Alegre. Although at Dimona, Allmon sampled 2 reserves which were later isolated as fragments, I also included Dimona in this comparison because of the proximity of those sites to the areas of continuous forest I sampled at Dimona. I tested the mean abundance and biomass at sites between years using post-hoc Tukey multiple comparisons.

There was no difference in either abundance (F1,444=0.95, P=0.331; Figure 6.1) or biomass (F1,444=0.05, P=0.817; Figure 6.2) between leaf litter frog samples in 1984-85 and 2007 at the BDFFP. Post hoc tests also indicated no difference in either frog abundance (Cabo Frio,

2 7

5 1985 4 2007

Individuals/100m 3

0 All BDFFP Dimona Km 41 Cabo Frio

Figure 6.1 Average frog abundance with 95% confidence intervals within the whole of the BDFFP and at 3 sites within the BDFFP in 1985 and 2007.

12.00

10.00

8.00

6.00 1985

g/100m2 2007 4.00

2.00

0.00 All BDFFP Dimona Km 41 Cabo Frio

Figure 6.2 Average biomass with 95% confidence intervals of leaf litter frogs within the whole of the BDFFP and at 3 sites in the BDFFP in 1985 and 2007.

99 t=2.83, df=444, Tukey Adj. P= 0.109; Dimona, t=0.14, df=444, Tukey Adj. P= 0.999; Km 41, t=0.68, df=444, Tukey Adj. P=0.999) or biomass (Cabo Frio, t=1.37, df=444, Tukey Adj. P=

0.910; Dimona, t=0.47, df=397, Tukey Adj. P= 0.999; Km 41, t=0.40, df=444, Tukey Adj.

P=0.999) for the reserves within the BDFFP that were sampled in both years.

The comparison of leaf litter frog abundance and biomass at the BDFFP in 1984-85 and

2007 show no significant change over the last 22 years. This differs strikingly from La Selva

Biological Station in Costa Rica where Whitfield et al. (2007) found that populations of all frog and lizard species declined by an average of 75% over a 35 year period (Figure 6.3). My results suggest that amphibian decline is not universal to Neotropical lowland rainforests.

Figure 6.3 An overlay of the BDFFP frog density data on the data from primary forest at La Selva in Costa Rica. This figure is modified from Whitfield et al. 2007. Gray circles are BDFFP data; Black circles are La Selva data.

Unfortunately, with the available data, I am unable to evaluate changes in the lizard community. In 498 plots in the 1984-85 data set, Allmon (1991) found 13 lizards (0.10/100m2).

In 265 plots in the 2007 data set, I found 72 (1.09/100m2). Because lizards can be difficult to sample with litter plots, we believe this difference reflects a discrepancy in sampling techniques 100 rather than a true 10 fold increase in lizard abundance over the last 22 years. In any event, even these lizard numbers represent an increase, not a decrease, from densities in 1984-85 to 2007.

Allmon (1991) sampled three different ‗subhabitat‘ types within sites at the BDFFP: slopes, stream valleys and terra firme. We were not able to distinguish among these sites in the data set thus data from all 3 microhabitats are included in the present analysis, even though the

2007 data come primarily from flat terra firme forest. Previous studies in the area suggest that topography on a local scale does not affect frog abundance and that most terrestrial species are habitat generalists (Allmon 1991, Menin et al. 2007). Therefore, despite microhabitat differences, these data sets should be comparable. Likewise, because they were collected using otherwise similar methodology, the data sets used here likely include less inter-researcher bias than those included in the La Selva analysis which were collected using a number of different sampling regimes.

Here, we show no evidence of amphibian decline at a lowland rainforest site in Central

Brazil which contrasts with evidence for decline from a site in Costa Rica. Although the authors suggest global warming has contributed to the declines at La Selva, there is another, possibly more parsimonious explanation. Many seminal studies in tropical ecology have been developed through work at La Selva. As such, La Selva is a well-studied site which attracts a large number of visitors each year. Habitat degradation caused by high visitor traffic in primary forest could be a factor in the decline in leaf litter herpetofauna populations.

Obviously, La Selva and the BDFFP represent a mere 2 sites in Neotropical lowland forests and further research is needed to determine why these two herpetofauna community patterns are contradictory. The lack of data sets is problematic, but priority should be given to re-sampling sites that have been sampled for leaf litter herpetofauna in the past. Additionally, researchers conducting re-sampling should communicate deliberately with those who first

101 collected the data in order to duplicate methods used in the initial collection. In this way, we can standardize data collection and make more accurate comparisons of data sets.

LITERATURE CITED

Alford, R. A., K. S. Bradfield, and S. J. Richards. 2007. Ecology: Global warming and amphibian losses. Nature 447:E3-E4.

Allmon, W. D. 1991. A plot study of forest floor litter frogs, Central Amazon, Brazil. Journal of Tropical Ecology 7:503-522.

Collins, J. P., and A. Storfer. 2003. Global amphibian declines: sorting the hypotheses. Diversity and Distributions 9:89-98.

Di Rosa, I., F. Simoncelli, A. Fagotti, and R. Pascolini. 2007. Ecology: The proximate cause of frog declines? Nature 447:E4-E5.

Lips, K. R., P. A. Burrowes, J. R. Mendelson, and G. Parra-Olea. 2005. Amphibian declines in Latin America: Widespread population declines, extinctions, and impacts. Biotropica 37:163-165.

Lips, K. R., J. Diffendorfer, J. R. Mendelson, and M. W. Sears. 2008. Riding the wave: reconciling the roles of disease and climate change in amphibian declines. PLoS Biology 6:e72.

Menin, M., A. P. Lima, W. E. Magnusson, and F. Waldez. 2007. Topographic and edaphic effects on the distribution of terrestrially reproducing anurans in Central Amazonia: mesoscale spatial patterns. Journal of Tropical Ecology 23:539-547.

Pounds, J. A., M. R. Bustamante, L. A. Coloma, J. A. Consuegra, M. P. L. Fogden, P. N. Foster, E. La Marca, K. L. Masters, A. Merino-Viteri, R. Puschendorf, S. R. Ron, G. A. Sanchez- Azofeifa, C. J. Still, and B. E. Young. 2006. Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439:161-167.

102

CHAPTER 7 CONCLUSION

The Amazon Basin represents the largest expanse of intact tropical rainforest on the planet, harboring an incredible diversity of amphibians and reptiles. As large expanses of the forest are lost through habitat destruction, it is becoming increasingly crucial that we understand the dynamics of intact ecosystems in order to be able to develop responsible management practices for remaining habitat. Although there has been much investigation into the forests as well as into various fauna within Amazonia, virtually nothing is known about the effects of plant productivity on animals in tropical forests. Differences in forest dynamics between regions should have a direct effect on taxa at higher trophic levels. Because primary productivity is strongly influenced by geomorphology, geomorphology should also play a role in determining community dynamics of herpetofauna. My dissertation reveals the effects and consequences of geomorphology and primary productivity on abundance, biomass and species richness of secondary consumers in the leaf litter herpetofauna community.

The first half of this dissertation explores the potential of using existing data sets to answer questions concerning changes and differences in leaf litter herpetofauna communities. In

Chapter 2, I examined the effect of preservation on the most common measurement of frogs, snout-vent length (SVL). Because measurements made on preserved anuran specimens are often used in studies of systematics, ecology and evolution, it is essential to determine whether the act of preservation and storage cause fundamental changes in the dimensions of specimens.

Preservation had significant effects on the SVL of 13 of the 14 species of North American frogs included in this study, with all species decreasing in SVL by a factor of 0.31-5.62%. Smaller frog species did not shrink proportionally more or less than larger species; however, the numerical shrinkage was correlated with SVL and was greater in larger species. Within species, percent

103 shrinkage was not correlated with SVL in 10 species, was greater for larger individuals in 3 species, and decreased with size in 1 species. Numerical shrinkage was greater for larger individuals in 4 species. My results agree with studies of morphological permutations in fish which show that most preservation-related changes take place within the first few months after initial preservation. I suggest that the potential consequences of using preserved specimens in research must be considered and that future studies continue to examine preservation effects, not only on frogs, but on all preserved specimens used in scientific investigations.

In Chapter 3, I determined how well the snout-vent length (SVL) of anurans estimated their mass for 36 species in the New World. Linear regressions of log mass on log SVL were highly significant for all species, explaining more than 75% of the mass variation in most species, and over 50% of the mass variation in all species. I also investigated differences in the mass/SVL relationship within species, comparing juveniles to adults, females to gravid females, and males to females, to determine the importance of developing separate regressions for sex or life stage classes. Three of six tests between juveniles and adults, and two of nine tests between females and gravid females indicated statistically significant differences, though these differences had only minor effects on mass estimates. More statistical differences in regression equations occurred between males and females; again, these differences were unimportant for estimates of mass in some cases, but they were important where there was strong sexual size dimorphism within a species. Continued collection of both SVL and mass data in new field studies of anurans will provide broader analyses of mass/SVL regressions. These species regressions along with data on density can be used to determine anuran community biomass.

The second half of this dissertation is a direct examination of the effects of geomorphology and primary productivity on the leaf litter herpetofauna communities of lowland

Neotropical rainforests. In Chapter 4, I hypothesized that secondary consumers in the leaf litter

104 herpetofauna community on ancient soils of Central Amazonia would exhibit reduced biomass compared to those found on younger soils of Western Amazonia, and that population densities on ancient soils could be driven below viable thresholds, reducing species richness. I compared herpetofauna abundance, biomass and richness on young soils in Ecuador to those on ancient soils in the Brazilian Amazon. Abundance, biomass and overall richness of the herpetofauna were significantly greater on younger soils. Separately, amphibians were only slightly more abundant, but their biomass on younger soils was twice that on ancient soils. Even more impressive was the variation exhibited by lizards: abundance was not significantly different, but biomass was five times greater on younger soils. Diversity of both taxa was greater on the young soils. The most important driver of differences in herpetofauna biomass, abundance and possibly diversity across the Amazon Basin appear to be underlying geomorphologic differences between regions. These differences are evident in other taxa as well, such as mammals and dung beetles.

Reduced primary productivity on ancient soils reverberates up the food chain, leaving fewer resources for higher trophic levels. This study highlights the importance of using biomass in addition to abundance as a standard measure in herpetofauna sampling.

In Chapter 5, I compiled data from 11 Neotropical lowland sites to further investigate effects of geomorphology on leaf litter frogs and lizards. I used methodology developed in

Chapters 2 and 3 to derive biomass data from existing records. Results demonstrate that leaf litter herpetofauna attain higher abundances and biomass in Panama and Western Amazonia relative to Central Amazonia and the Guiana region. French Guianan sites appear to be intermediate in frog and lizard biomass to Western Amazonia and Brazil. This is likely a result of intermediate soil fertility at the French Guianan sites. Western Amazonia and Panama have higher species richness and larger frogs and lizards than Central Brazilian sites, lending further support to the hypothesis that geologically younger soils can support more biomass, higher

105 densities and more species of secondary consumers among the leaf litter herpetofauna community.

Can the tropical rainforests in the Amazon be considered a single ecosystem whose conservation areas should employ common management practices? Although it has been stated that the ―preservation of relatively few, large areas in the lowlands [of Amazonia] will preserve a large percentage of the amphibian fauna‖ (Duellman 1999), it is clear that Amazonian conservation biology will be more complex. This dissertation reveals that the unique geological history of rainforest regions within Amazonia plays a large role in determining the structure of the secondary consumer leaf litter herpetofauna communities within. These findings likely apply to other taxa as well. Consequently, management strategies need to be tailored to geomorphological landforms. Proportionally larger tracts of land in the old Amazon will be required to maintain population densities of herpetofauna comparable to the young Amazon, especially as continuous forests become fragmented by anthropogenic activities.

Finally, in Chapter 6, I compared abundance and biomass of frogs at Brazilian sites sampled in 1985 to the same sites sampled 22 years later. I found no changes in leaf litter frog density or biomass over that time period. This result differs from evidence of amphibian decline at La Selva Biological Station in Costa Rica. The issue of global amphibian decline is pressing.

Over 30% of all amphibian species on the planet are threatened with extinction and most of these are tropical montane, stream-breeding species (Stuart et al. 2004). Whitfield et al. (2007) shed new light on these enigmatic declines by suggesting that they are not restricted to montane habitats and that lowland frogs as well as lizards are also experiencing decline. The results herein suggest that lowland amphibian decline is not a ubiquitous phenomenon throughout lowland rainforests. It is imperative that more research be done, particularly re-sampling of

106 areas with existing density records, in order to determine if lowland decline is occurring in other regions, or if the phenomenon at La Selva may have some other explanation unique to that site.

LITERATURE CITED

107

APPENDIX 1 SPECIES LIST

Appendix 1A Abundance of each species found in litter plots at Cabo Frio (CF), Dimona (DI), KM41 (KM) at the BDFFP in Brazil, Tiputini Biodiversity Station (TBS) and Yasuní Research Station (YRS) in Ecuador and Inselberg (IN) and Pararé (PA) at Nouragues Research Station in French Guiana. Individuals not identified to species are not included in this table. ORDER SUBORDER FAMILY GENUS SPECIES CF DI KM TBS YRS IN PA Anura Aromobatidae Allobates femoralis 2 1 1 Anura Aromobatidae Allobates granti 3 Anura Aromobatidae Allobates marchesianus 2 2 Anura Aromobatidae Allobates trilineatus 5 Anura Aromobatidae Anomaloglossus baeobatrachus 2 5 Anura Aromobatidae Anomaloglossus stepheni 22 27 20 Anura Bufonidae Atelopus franciscus 9 6 Anura Bufonidae Dendrophryniscus minutus 8 Anura Bufonidae Rhaebo guttatus 1 Anura Bufonidae Rhinella castenoetica 11 3 Anura Bufonidae Rhinella margaritifera 23 15 6 5 Anura Bufonidae Rhinella proboscidea 1 4 2 Anura Dendrobatidae Ameerega bilinguis 25 36 Anura Dendrobatidae Ameerega hahneli 1 Anura Dendrobatidae Dendrobates tinctorius 1 1 Anura Dendrobatidae Ranitomeya duellmani 10 Anura Dendrobatidae Ranitomeya ventrimaculata 1 Anura Hylidae Osteocephalus oophagus 2 Anura Hylidae Osteocephalus planiceps 2 Anura Hylidae Osteocephalus taurinus 1 1 Anura Hylidae Osteocephalus yasuni 2 1 Anura Leiuperidae Edalorhina perezi 1 Anura Leiuperidae Engystomops petersi 1 4 5 Anura Leptodactylidae Leptodactylus andreae 34 50 51 15 4 Anura Leptodactylidae Leptodactylus heyeri 1 Anura Leptodactylidae Leptodactylus hylaedactyla 1

108

Anura Leptodactylidae Leptodactylus mystaceus 7 Anura Leptodactylidae Leptodactylus rhodomystax 3 Anura Leptodactylidae Leptodactylus stenodema 1 Anura Microhylidae Chiasmocleis bassleri 2 3 Anura Microhylidae Chiasmocleis hudsoni 1 1 Anura Microhylidae Synapturanus mirandariberoi 1 1 Anura Microhylidae Synapturanus salseri 1 Anura Strabomantidae Hypodactylus nigrovittatus 4 8 Anura Strabomantidae Oreobates quixensis 1 Anura Strabomantidae Pristimantis altamazonicus 3 Anura Strabomantidae Pristimantis chiastonontus 3 Anura Strabomantidae Pristimantis conspicillatus 4 3 Anura Strabomantidae Pristimantis fenestratus 2 1 Anura Strabomantidae Pristimantis gutturalis 1 Anura Strabomantidae Pristimantis lanthanites 2 2 Anura Strabomantidae Pristimantis marmoratus 1 Anura Strabomantidae Pristimantis martiae 1 1 Anura Strabomantidae Pristimantis ockendeni 1 16 9 Anura Strabomantidae Pristimantis orphnolaimus 1 Anura Strabomantidae Pristimantis peruvianus 4 Anura Strabomantidae Pristimantis sp. 2 6 Anura Strabomantidae Pristimantis zeuctoctylus 2 Anura Strabomantidae Pristimantis zimmermanae 1 Anura Strabomantidae Strabomantis sulcatus 1 1 Caudata Plethodontidae Bolitoglossa equatoriana 3 Caudata Plethodontidae Bolitoglossa peruviana 1 Squamata Sauria Gekkonidae Arthrosaura kocki 1 Squamata Sauria Gekkonidae Coleodactylus amazonicus 10 19 32 5 3 Squamata Sauria Gekkonidae Gonatodes coccinatus 2 Squamata Sauria Gekkonidae Gonatodes humeralis 1 2 1 1 Squamata Sauria Gekkonidae Lepidoblepharis heyerorum 1 Squamata Sauria Gekkonidae Pseudogonatodes guianensis 5 4 Squamata Sauria Gymnophthalmidae Alopoglossus atriventris 6 1

109

Squamata Sauria Gymnophthalmidae Arthrosaura reticulata 2 2 1 Squamata Sauria Gymnophthalmidae Iphisa elegans 1 Squamata Sauria Gymnophthalmidae Leposoma cf guianense 1 1 Squamata Sauria Gymnophthalmidae Leposoma guianense 2 3 Squamata Sauria Gymnophthalmidae Leposoma parietale 8 10 Squamata Sauria Gymnophthalmidae Prionodactylus argulus 1 1 Squamata Sauria Gymnophthalmidae Tretioscincus agilis 1 Squamata Sauria Hoploceridae Enyalioides cofanorum 3 Squamata Sauria Polychrotidae Anolis fuscoauratus 1 1 Squamata Sauria Polychrotidae Anolis nitens 1 3 1 3 3 Squamata Sauria Polychrotidae Anolis ortonii 1 Squamata Sauria Polychrotidae Anolis trachyderma 1 Squamata Serpentes Aniliidae Anilius scytale 1 Squamata Serpentes Colubridae Atractus cf collaris 1 Squamata Serpentes Colubridae Atractus flammigerus 1 Squamata Serpentes Colubridae Taeniophallus nicagus 1 TOTAL ABUNDANCE: 68 128 118 125 126 74 54 TOTAL NUMBER OF SPECIES: 5 18 14 29 26 21 17

110

Appendix 1B Biomass of each species found in litter plots at Cabo Frio (CF), Dimona (DI), KM41 (KM) in Brazil, Tiputini (TI), Yasuní (YA) in Ecuador and Inselberg (IN) and Pararé (PA) in French Guiana. Individuals not identified to species are not included in this table. *Biomass analyses remain significant with the removal of largest individual from each site, as well as with removal of snakes and salamanders from the analyses of total biomass of secondary consumers among the leaf litter herpetofauna. ORDER SUBORDER FAMILY GENUS SPECIES CF DI KM TBS YRS IN PA Anura Aromobatidae Allobates femoralis 0.18 1.72 1.19 Anura Aromobatidae Allobates granti 1.31 Anura Aromobatidae Allobates marchesianus 0.78 0.89 Anura Aromobatidae Allobates trilineatus 2.54 Anura Aromobatidae Anomaloglossus baeobatrachus 0.94 1.49 Anura Aromobatidae Anomaloglossus stepheni 6.63 10.36 7.21 Anura Bufonidae Atelopus franciscus 6.89 6.54 Anura Bufonidae Dendrophryniscus minutus 2.72 Anura Bufonidae Rhaebo guttatus 76.60 Anura Bufonidae Rhinella castenoetica 27.85 1.28 Anura Bufonidae Rhinella margaritifera 117.81 87.01 84.38 2.49 Anura Bufonidae Rhinella proboscidea 18.29 35.17 15.97 Anura Dendrobatidae Ameerega bilinguis 12.95 18.93 Anura Dendrobatidae Ameerega hahneli 0.77 Anura Dendrobatidae Dendrobates tinctorius 0.30 5.65 Anura Dendrobatidae Ranitomeya duellmani 3.17 Anura Dendrobatidae Ranitomeya ventrimaculata 0.06 Anura Hylidae Osteocephalus oophagus 18.32 Anura Hylidae Osteocephalus planiceps 21.79 Anura Hylidae Osteocephalus taurinus 8.30 8.50 Anura Hylidae Osteocephalus yasuni 26.74 6.97 Anura Leiuperidae Edalorhina perezi 3.91 Anura Leiuperidae Engystomops petersi 3.89 12.74 4.75 Anura Leptodactylidae Leptodactylus andreae 12.36 14.87 15.55 6.09 1.84 Anura Leptodactylidae Leptodactylus heyeri 0.59 Anura Leptodactylidae Leptodactylus hylaedactyla 2.23 Anura Leptodactylidae Leptodactylus mystaceus 2.31 Anura Leptodactylidae Leptodactylus rhodomystax 79.11

111

Anura Leptodactylidae Leptodactylus stenodema 1.02 Anura Microhylidae Chiasmocleis bassleri 3.17 7.33 Anura Microhylidae Chiasmocleis hudsoni 0.11 0.15 Anura Microhylidae Synapturanus mirandariberoi 3.43 2.74 Anura Microhylidae Synapturanus salseri 2.42 Anura Strabomantidae Hypodactylus nigrovittatus 3.71 5.90 Anura Strabomantidae Oreobates quixensis 0.00 9.03 Anura Strabomantidae Pristimantis altamazonicus 0.00 5.47 Anura Strabomantidae Pristimantis chiastonontus 6.89 Anura Strabomantidae Pristimantis conspicillatus 5.46 9.87 Anura Strabomantidae Pristimantis fenestratus 3.24 1.57 Anura Strabomantidae Pristimantis gutturalis 0.63 Anura Strabomantidae Pristimantis lanthanites 0.87 0.06 Anura Strabomantidae Pristimantis marmoratus 0.71 Anura Strabomantidae Pristimantis martiae 0.66 0.29 Anura Strabomantidae Pristimantis ockendeni 0.51 12.50 4.99 Anura Strabomantidae Pristimantis orphnolaimus 1.36 Anura Strabomantidae Pristimantis peruvianus 7.32 Anura Strabomantidae Pristimantis sp. 0.67 2.06 Anura Strabomantidae Pristimantis zeuctoctylus 0.35 Anura Strabomantidae Pristimantis zimmermanae 0.74 Anura Strabomantidae Strabomantis sulcatus 0.12 0.31 Caudata Plethodontidae Bolitoglossa equatoriana 2.94 Caudata Plethodontidae Bolitoglossa peruviana 1.37 Squamata Sauria Gekkonidae Arthrosaura kocki 0.22 Squamata Sauria Gekkonidae Coleodactylus amazonicus 1.10 2.25 4.31 1.07 0.56 Squamata Sauria Gekkonidae Gonatodes coccinatus 0.00 5.96 Squamata Sauria Gekkonidae Gonatodes humeralis 1.22 1.78 0.62 0.32 Squamata Sauria Gekkonidae Lepidoblepharis heyerorum 0.42 Squamata Sauria Gekkonidae Pseudogonatodes guianensis 1.66 1.33 Squamata Sauria Gymnophthalmidae Alopoglossus atriventris 15.02 0.69 Squamata Sauria Gymnophthalmidae Arthrosaura reticulata 0.91 4.96 0.75 Squamata Sauria Gymnophthalmidae Iphisa elegans 1.56

112

Squamata Sauria Gymnophthalmidae Leposoma cf guianense 0.43 0.72 Squamata Sauria Gymnophthalmidae Leposoma guianense 1.38 2.68 Squamata Sauria Gymnophthalmidae Leposoma parietale 6.31 7.48 Squamata Sauria Gymnophthalmidae Prionodactylus argulus 0.24 0.64 Squamata Sauria Gymnophthalmidae Tretioscincus agilis 0.22 Squamata Sauria Hoploceridae Enyalioides cofanorum 15.05 Squamata Sauria Polychrotidae Anolis fuscoauratus 1.31 0.34 Squamata Sauria Polychrotidae Anolis nitens 0.20 9.81 1.45 8.06 7.57 Squamata Sauria Polychrotidae Anolis ortonii 1.86 Squamata Sauria Polychrotidae Anolis trachyderma 0.42 Squamata Serpentes Aniliidae Anilius scytale 5.88 Squamata Serpentes Colubridae Atractus cf collaris 0.00 4.15 Squamata Serpentes Colubridae Atractus flammigerus 1.80 Squamata Serpentes Colubridae Taeniophallus nicagus 1.46 TOTAL BIOMASS: 46.68 177.28 65.06 281.64 276.03 160.62 39.76

113

APPENDIX 2 LETTER OF PERMISSION

From: Jessica Deichmann [mailto:[email protected]] Sent: Wednesday, March 18, 2009 9:23 PM To: Matthew James Parris (mparris) Subject: Permission MS 07-073

Dear Dr. Parris, I am a PhD candidate at Louisiana State University. Today I am writing to request permission to include the following manuscript in my doctoral dissertation: Deichmann, J. L., W. E. Duellman and G. B. Williamson. 2008. Predicting biomass from snout-vent length in New World frogs. Journal of Herpetology 42:238-245. The article will appear as the third chapter of my dissertation. I will be the sole author of the dissertation, which will be submitted to the university no later than 1 June 2009.

Thank you for your assistance. Please do not hesitate to contact me if you have any questions or concerns regarding this request.

Sincerely,

Jessica

-- Jessica L. Deichmann PhD candidate Department of Biological Sciences 107 Life Sciences Building Louisiana State University Baton Rouge, LA 70803 (225)578-8494

From: Matthew James Parris (mparris) Sent: Thursday, March 19, 2009 4:32 PM To: Jessica Deichmann [mailto:[email protected]] Subject: RE: Permission MS 07-073

Jessica:

This is fine. It would be good to include some type of acknowledgement in your dissertation about the work having been published in the Journal of Herpetology.

Best wishes.

Matthew Parris, Ph.D. Editor, Journal of Herpetology Associate Professor Department of Biology University of Memphis Memphis, TN 38152 (901) 678-4408 [email protected] https://umdrive.memphis.edu/mvenesky/public/parrislab/parrishome.html

114

VITA

Jessica Lee Deichmann was born in April, 1980, to John and Sally Deichmann in St.

Louis, Missouri. Her love of nature was initially inspired through camping trips with her parents and identifying animals in picture books with her grandfather, William D. Deichmann, Jr. She grew up in St. Louis and graduated from Nerinx Hall High School in 1998. It was her passion for the outdoors that took her from the Mississippi River Valley to the Rocky Mountains at

Colorado State University where she earned her Bachelor of Science degree in zoology and a minor in Spanish in May 2002. While at Colorado State University, Jessica spent a semester abroad in Ecuador studying tropical ecology through a course sponsored by Boston University.

She also spent summers as an intern with the research department at the St. Louis Zoo (1999) and U.S. Fish and Wildlife‘s Southeast Louisiana National Wildlife Refuges (2001, 2002).

These research and field experiences convinced Jessica to pursue a doctorate in biology. In

August 2003, Jessica entered the Graduate School of Louisiana State University under the supervision of Dr. G. Bruce Williamson. After defending her dissertation, Jessica will travel to

Costa Rica to coordinate the Organization for Tropical Studies‘ graduate course in tropical ecology.

115