Spring19 Palm Tree Proposal

INVESTIGATING SEED SIZE AS A DRIVER FOR RODENT CAUSED MORTALITY
ON BARRO COLORADO ISLAND
Project Lead
Justin Tirrell—Biology Department
Advisors
Dr. Noelle Beckman—Utah State University
Dr. Carol Garzon-Lopez—Los Andes University (Colombia)
1

PEAK 2019: Investigating seed size as a driver for rodent caused mortality on barro colorado island 2
Project Description
Introduction
Analysis in spatially explicit ecology has a long history (Terborgh 2012), and understanding mechanisms
driving the spatial distributions of ecosystems remains a longstanding goal in ecology (Chesson 2000).
Previous investigations in spatial ecology have focused on understanding whether patterns of biodiversity
can be better explained by natural resource gradients or by the effects of species-species interactions.
These investigations have concluded that environmental gradients alone don’t provide a sufficient
explanation for existing biodiversity (Garzon-Lopez 2015). Further research in biodiversity revealed
the importance of the predator. Predators shape the structure of populations at the community level by
selecting individuals for removal and consumption (Terbough 2012; Janzen 1970; Connel 1971). This
mechanism has been primarily modeled using specialist predators, which feed on a single species of prey
(Janzen 1970; Connel 1971). However, models have been recently developed and tested for generalist
predators, which consume multiple species of prey (Garzon-Lopez 2015). The Garzon-Lopez study
found that models which based their predictions on mechanisms of generalist predators explained the
distribution of seeds on Barro Colorado Island significantly better than models based on the predictions
of specialist predators (Garzon-Lopez 2015). In this project, I will be using linear regression and
model averaging techniques to analyze data that I collect in one of the most studied tropical
forests in the world, Barro Colorado Island, located in Central Panama. The purpose of this
study is to uncover how variation in seed density and seed size influence patterns of seed survival in a
tropical tree system with shared predators. The results of this project will increase our understanding of
mechanisms shaping patterns of plant distributions and diversity, specifically the importance of within
versus among species trait variation in mediating seed predation by generalists versus specialists.
Spatial Mechanisms
Specialized predators can maintain coexistence in theory (Mordecai 2011) and are hypothesized to
maintain the high diversity of plant diversity observed in tropical forests (Janzen 1970; Connell 1971).
A population of specialist predators is tied to their food source, and they flourish where the seeds are
the most abundant (Terborgh 2012). As a result, seeds in isolation tend to survive to adulthood more
often than seeds in high density populations. This phenomenon, known as the Janzen-Connell Effect,
limits the aggregate density of a single species of tree in a community (Janzen 1970; Connell 1971).
This Janzen-Connell Effect is hypothesized to result in specific spatial patterns, where seedlings densities
are low near the parent tree and higher away from the tree. If this pattern persists until adulthood,
reproductive trees are expected to be regularly distributed. The spacial pattern of reproductive adult
trees is not as predictive for seeds which are consumed by generalist species because generalists have
more mobility than specialists (Terborgh 2012). The resulting spatial model has proposed that generalist
species attack seeds more frequently as the heterospecific density (all prey species) of trees increases
(Garzon-Lopez 2015). Additionally, generalist species are thought to have preferences about which
species of seed they consume. These species-based preferences cause trees to have complex effects on
each other’s seeds. For instance, a population of trees which are highly disliked by a generalist predator
may disguise the presence of preferred seeds sharing the same habitat, increasing their survival.

Seed Traits
The key to developing models for generalist predators may be understanding what mechanisms drive
their preferences among seeds within the same species. As a result, research has focused on how the
traits of seeds increase or decrease a predator’s preference to consume the seeds. Studies have been
performed to determine how tannin content, beetle infestation, and seed size affect the preferences
of generalist rodent predators in Central America (Sallabanks & Courtney 1992; Garzon-Lopez 2015;
Kuprewicz 2019). There is an opportunity to expand work being done on seed traits by comparing
the explanatory strength (with specialized linear regression techniques) of each variable in a single
ecosystem. This project aims to compare the explanatory strength of seed size variation within and
among species to the explanatory strength of the model used in Garzon-Lopez (2015) which included
species identity only for modeling generalist predation within the context of Barro Colorado Island.
Previous research in Costa Rica has found that tropical rodents prefer to store and consume seeds based
on size, the largest seeds being the most preferred (Kuprewicz 2019). This project will advance the field
of spatial ecology by determining how effective seed size is at predicting the attacks of rodent generalist
predators (native to Central America) on seeds of the following trees: Astrocaryum standleyanum,
Dipteryx oleifera, and Attalea butyracea. This project requires that I travel to the STRI Barro Colorado
Island research facility in Panama to collect data on seed size and seed mortality.
Feasibility
Round-trip travel expenses and equipment costs will be obtained from the USU URCO grant ($1000
award) and the USU College of Science minigrant ($1000 award). Living expenses while on Barro
Colorado Island can be covered through stipends granted by the USU Peak Summer Felowship award
payment ($4000 award). Housing options for this project include living at the Smithsonian Tropical
Research Institute (STRI) facility on Barro Colorado Island; commuting from Gamboa, Panama; or
commuting from Panama City, Panama depending on availability of funds and housing at STRI.

Background
The project will require travel to Barro Colorado Island in Panama. The details of the location and
focal species are outlined below.
Focal Species
The trees being considered are Astrocaryum standleyanum, Dipteryx oleifera, and Attalea butyracea. These
trees can be effectively used to model how predators effect the spatial distribution of plant populations
because they have relatively few predators that have the ability to break through the stony endocarps
that protect the seeds (Bradford & Smith 1977; Janzen 1971). These trees were chosen because they
share three rodent predators, the Central American Agouti (Dosyprocta puntata), the Central American
Spiny Rat (Proechymus semispinous), and thee Red Tailed Squirrel (Scurrious granatensis) (Janzen
et al). The seeds of these trees don’t fruit at the same time of the year, but the seeds of Attalea and
Astrocaryum are dormant during the year and remain as a source of food for seed predators. The seeds
of Dipteryx are present during the Spring during Panama’s dry season. The project will orient itself
as an expansion to previous work done on the island with an emphasis placed on the importance of
the role of seed size variation on the level of predation experienced by these three species of plants.
Barro Colorado Island
Barro Colorado Island is a lowland tropical moist forest in the middle of Gatun Lake, a section of the
Panama Canal (Leigh 1999). The region experiences a dry period between December and early May
(Leigh 1999). Dipteryx fruits in the dry season, while Attalea and Astrocaryum fruit at separate times
during the wet season (Garzon Lopez 2012). May through June is an optimal time period to perform
this experiment because the seeds from Attalea and Astrcaryum have multi-year dormancy, making
them available to predators throughout the year (Garzon-Lopez 2012).
Significance
This project will advance the field of spatial ecology by promoting our understanding of how seed
size variation within species and heterospecific tree density are related to each other as mechanisms
that increase and maintain biodiversity. Generalist rodent predators have been deemed one of the
most important vertebrate seed predators of Attalea butyracea seeds in Panama (Galvez 2007), and
the focal tree species of this project are the most abundant species of trees on Barro Colorado Island
(Garzon-Lopez 2015). This project will be performed with the assistance of international collaborators
who have dedicated years of their life to understanding mechanisms underlying tropical biodiversity.
As a consequence, the data collected will be readily available to individuals in the scientific community
who can use it to advance programmatic and analytic models. Advancing models for projecting plant
population growth provides allows us to effectively manage plant populations.

Research Question & Hypotheses
Research Question: How does variation in seed size within species mediate interactions of three tree
species with shared generalist predators?
Hypothesis: Seed size will be better than species identity at explaining the frequency of rodent predation
for the largest seeds, while species identity will be better than seed size for medium and small sized
seeds. Preferences for larger seeds should be more evident in plots with high tree densities where
food availability is abundant, and become less evident in densities of low tree abundance, where food
availability is more scarce. I hypothesize that rodents prefer Astrocaryum and Attalea over Dipteryx
due to their high oil content, and that rodents prefer Astrocaryum over Attalea due to their higher
sugar content in the pulp. Dipteryx fruits before Astrocarym and Attalea, and is eaten when other fruit
are not available. Attalea tend to be larger than Astrocaryum seeds. Because rodents are choosing the
largest seeds for the energy content the seed itself provides, I hypothesize that seed size will be better
in explaining rodent predation for the largest seeds.
Overview
Seed data will be collected following procedures modified from the Garzon-Lopez (2012) examination
of the indirect interactions of palm populations through shared predators. The data will be collected
daily from 25 pre-determined plots over 2 months. The diameter of each seed collected will be recorded
in addition to the suspected predator. Rodent predation can be uniquely identified by a distinctive claw
mark scarring that they leave on the endocarp (shell) of the seed. Subsequent to the data collection,
data will be analyzed using regression techniques. Additional statistical analysis can be provided by
collaborators Dr. Beckman and Dr. Garzon-Lopez to ensure that the findings will meet standards for
publication. I predict that the rodent predators will exhibit a preference for large seeds that overshadows
their preference for between seeds of differing species. This leads me to the following hypothesis.
Objectives
• Collect seed caching data following the procedures of Wright & Duber (2001) and Visser (2011)
in locations predetermined by Garzon-Lopez (2012).
• Determine whether the frequency of seed predation is explained by the Gazon-Lopez model for
shared predators.
• Analyze the data, comparing the correlation between the generalist predator model and the size
of the seeds.
• Make the data set available to international researchers for further analysis.

Methods
Data Collection
6 plots have been selected for each combination of high and low adult density of the 3 focal tree species
(Garzon-Lopez 2012). This has resulted in 54 plots which are within 100 meters of trails (Garzon-Lopez
2012). This project will aim to collect data from 27 of the plots pre-determined by the Garzon-Lopez
study. Correlation analysis has previously revealed that the increased tree density is correlated to an
increase in seed density (Garzon-Lopez 2012). Each plot is 1 hectare; 10 random points will be generated
in arcGIS. These points will then be loaded into a handheld receiver so the locations can be found in the
field. 1−m2
quadrats will be set up at each random point. Seeds at the soil surface and buried up to 5 cm
deep in the soil will be collected according to the methods used in Wright and Duber (2001) and Visser
et al. (2011). After seeds have been collected, the status of the seed will be recorded. Seeds consumed
by rodents have distinctive claw marks, and beetle predators leave large emergence holes which are
easily observed (Wright & Duber 2001). Seeds that have escaped predation will be recorded as ’intact.’
Analysis
All statistical analysis will be done in R v.3.5 (R Core Team 2013). The data for the largest seeds will be
analyzed separately from the complete data set to see if the explanatory strength of the model persists
for the largest seeds. The analysis will follow the approach of Burnham and Anderson (2002), which
has also used in the parent study of this experiment (Garzon 2012). This approach involves using
logistic regression and model averaging to determine how well the data set meets the predictions of
each model. The single species specific predator model (Janzen; Connell) and the generalist predator
model (Garzon-Lopez 2012) will be used for the analysis. The goodness of the fit will be evaluated
by comparing calculated Akaike weights, which are representative of the goodness of fit of each model
(Garzon-Lopez 2012). Dr. Garzon-Lopez has agreed to assist with the technical aspects of implementing
these tests. Her assistance will ensure that this study produces quality results.

Education Plan
I am confident beyond doubt that this project will be a source of radical growth in my life. Designing this
experiment has challenged me to develop my skills as a scientific author, to engage Ecology literature
with purpose, and to think critically about the effects of predators on an ecosystem. Based on the
growth that I have already experienced, I anticipate that executing this experiment will be effective
in fostering my development as an Ecologist and a researcher. Specifically, my objectives are to 1)
collaborate with international scientists, 2) collect and analyzing my own data, and 3) gain an authentic
understanding of how research is performed in a Moist Tropical Forest.
During the data collection phase of this project, I will be living on Barro Colorado Island (BCI), a STRI
research station in the Panama Canal. By living with and collaborating with international scientists,
I will gain a deeper understanding of what it means to be doing research in Ecology. This is an
opportunity for me to gain exposure to the ways that I can engage in Ecology research. This is a way
for me to understand what research questions I am most interested in so that I can pursue a graduate
degree in an area that I am passionate about.
Collecting and analyzing my own data is meaningful to me. It is evidence that I have the drive and
capacity to produce results. The finished products of this project will give prospective mentors at
graduate schools confidence in choosing to take me on as a PhD or Masters student. In my mind, this
project is the beginning of my life’s work.
I’ve heard it said that a person can never fully understand the experience of being in a place by reading
the accounts of others. Travelling to Panama to do research will expose me to the realities behind
the scientific literature that I’ve read. By physically seeing the interactions between the rodents, birds,
beetles, and seeds; I can develop new research questions to expand Ecology in new and unexpected ways.
Dissemination
The idea of this project will be presented in concept at the USU Spring Student Research Symposium.
Pending acceptance and travel funding, the results of this project will be presented at the Ecological
Society of America conference in August. After final analysis, the results will be presenting at the USU Fall
Student Research Symposium and in the form of a scientific paper, which I plan to submit for publication.
Acknowledgements
I’d like to thank the following individuals for revising this proposal: Dr. Noelle Beckman, Dr. Garzon-
Lopez, Dr. Will Pearse, Dr. Robert Schaeffer, Ryan Tarver, Binod Borah, Rylee Jensen, Sarah Bogen,
and Eric Sodja.

REFERENCES 8
References
[1] Bradford, D. F. & Smith, C. C. 1977. Seed predation and seed number in Scheelea palm fruits. Ecology
58:667–673.
[2] Burnham, K.P. & Anderson, D.R. (2002).Model Selection and Multi-Model Inference: A Practical
Information-Theoretic Approach. SpringerVerlag, New York, pp. 514.
[3] Chesson, P. (2000). Mechanisms of Maintenance of Species Diversity. Annual Review of Ecology and
Systematics, 31(1), 343366. https://doi.org/10.1146/annurev.ecolsys.31.1.343
[4] Connell JH (1971) On the role of natural enemies in preventing competitive exclusion in some marine
animals and in rain forest trees, in Boer PJD, Gradwell GR, Editors. Dynamics of populations. Wageningen,
Netherlands: Centre for Agricultural Publication and Documentation. pp. 298312.
[5] Gálvez, D., & Jansen, P. (2007). Bruchid beetle infestation and the value of Attalea butyracea endocarps
for Neotropical rodents (Vol. 23). https://doi.org/10.1017/S0266467407003975
[6] Gálvez, D., Kranstauber, B., Kays, R. W., & Jansen, P. A. (2009). Scatter hoarding by the Central
American agouti: a test of optimal cache spacing theory. Animal Behaviour, 78(6), 13271333.
https://doi.org/10.1016/j.anbehav.2009.08.015
[7] Garzon-Lopez, C. X., Ballesteros-Mejia, L., Ordoñez, A., Bohlman, S. A., Olff, H., & Jansen, P. A. (2015).
Indirect interactions among tropical tree species through shared rodent seed predators: a novel mechanism
of tree species coexistence. Ecology Letters, 18(8), 752760. https://doi.org/10.1111/ele.12452
[8] Janzen, D. H. 1970. Herbivores and number of tree species in tropical forests. American Naturalist (Vol.
104). https://doi.org/10.1086/282687
[9] Janzen, D. H. 1977. Why fruits rot, seeds mold, and meat spoils. American Naturalist 111:691–713.
[10] Kuprewicz, E. K., & García-Robledo, C. (2019). Deciphering seed dispersal decisions: Size, not
tannin content, drives seed fate and survival in a tropical forest. Ecosphere, 10(1), e02551.
https://doi.org/10.1002/ecs2.2551
[11] Mordecai, E. A. (2011). Pathogen impacts on plant communities: unifying theory, concepts, and empirical
work. Ecological Monographs, 81(3), 429–441. https://doi.org/10.1890/10-2241.1
[12] Mueller, R., D Wade, B., Gehring, C., & G Whitham, T. (2005). Chronic herbivory negatively impacts
cone and seed production, seed quality and seedling growth of susceptible pinyon pines (Vol. 143).
https://doi.org/10.1007/s00442-005-0029-0
[13] R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical
Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/.
[14] Sallabanks, R. & Courtney,S.P.1992.Frugivory,seedpredation, and insect-vertebrate interactions. Annual
Review of Entomology 37:377–400.
[15] Terborgh, J. (2012). Enemies Maintain Hyperdiverse Tropical Forests. The American Naturalist, 179(3),
303314. https://doi.org/10.1086/664183
[16] Visser, M.D., Muller-Landau, H.C., Wright, S.J., Rutten, G. & Jansen,P.A. (2011). Tri-trophic interactions
affect density dependence of seedfate in a tropical forest palm.Ecol. Lett., 14, 1093–1100
[17] Wright, S. J. and Duber, H. C. (2001), Poachers and Forest Fragmentation Alter Seed Dispersal, Seed
Survival, and Seedling Recruitment in the Palm Attalea butyraceae, with Implications for Tropical Tree
Diversity1. Biotropica, 33: 583-595. doi:10.1111/j.1744-7429.2001.tb00217.x
[18] Zimmerman, J., & Giles Leigh, E. (2000). Tropical Forest Ecology: A View from Barro Colorado Island
(Vol. 81). https://doi.org/10.2307/177121

Spring19 Palm Tree Proposal

More Related Content

What's hot (20)

Similar to Spring19 Palm Tree Proposal (20)

More from JustinTirrell1 (7)

Recently uploaded (20)

Spring19 Palm Tree Proposal