THE REPRODUCTIVE BIOLOGY OF ZAMIA (CYCADALES: ZAMIACEAE) IN PUERTO RICO:
IMPLICATIONS FOR PATTERNS OF GENETIC STRUCTURE AND SPECIES CONSERVATION
By
JULIO C. LAZCANO LARA
A dissertation submitted to the
DEPARTMENT OF BIOLOGY
FACULTY OF NATURAL SCIENCES
UNIVERSITY OF PUERTO RICO
RÍO PIEDRAS CAMPUS
In partial fulfillment of the requirements for the degree of
DOCTOR IN PHILOSOPHY
May 2015
Río Piedras, Puerto Rico
i
This dissertation has been accepted by faculty of the:
DEPARTMENT OF BIOLOGY
FACULTY OF NATURAL SCIENCES
UNIVERSITY OF PUERTO RICO
RÍO PIEDRAS CAMPUS
In partial fulfillment of the requirements for the degree of
DOCTOR IN PHILOSOPHY
Dissertation Committee:
__________________________________________
James D. Ackerman, Ph.D.
Advisor
__________________________________________
Jess K. Zimmerman, Ph.D.
__________________________________________
Riccardo Papa, Ph.D.
__________________________________________
Ingi Agnarsson, Ph.D.
__________________________________________
Nico Franz, Ph.D.
ii
THE REPRODUCTIVE BIOLOGY OF ZAMIA (CYCADALES: ZAMIACEAE) IN PUERTO RICO:
IMPLICATIONS FOR PATTERNS OF GENETIC STRUCTURE AND SPECIES CONSERVATION
iii
A mi familia
iv
ACKNOWLEDGMENTS
I would like to thank my advisor James D. Ackerman for his mentoring and counseling
throughout my graduate studies. Jim’s mentoring was based on high academic standards,
freedom, accountability, self-regulation, availability, and full support combined with a generous
dose of constructive criticism; this was the best combination for me. I also thank the other
members of my committee: Jess K. Zimmerman, Riccardo Papa, Ingi Agnarsson, and Nico Franz
for their valuable comments, recommendations, and for their availability. Their prompt
response every time we had to coordinate a meeting allowed me to focus on research without
investing excessive time in other activities. I thank the professors with whom I took classes
during the graduate studies, especially I would like to acknowledge Jess K. Zimmerman, his
Communities and Ecosystems class opened my mind to research in tropical ecology; Tomas
Hrbek, his Speciation class opened my mind to research in evolutionary biology; and Carla
Restrepo with whom I took Tópicos en Biología Moderna and Seminario Doctoral, two
fundamental courses of the graduate program. While taking Tópicos I developed the idea of
studying spatial processes in Zamia, and during Seminario I produced my dissertation proposal,
both experiences were very fruitful. I would also like to thank my classmates, they provided an
example to follow and constructive criticism that helped me to improve my research. I thank the
Graduate Biology Program and the Department of Biology for their support and high quality
education.
Without the funds received from various sources I could not have completed my graduate
studies. My gratitude to the Dean for Graduate Studies and Research of the University of Puerto
Rico, Río Piedras Campus for its financial support through two fellowships that allowed me to
focus on my research. I send my gratitude to the Center for Applied Tropical Ecology and
v
Conservation for financing several field trips, equipment, and the processing of DNA samples. I
thank James D. Ackerman for funding the acquisition of materials and the processing of DNA
samples required to complete the second chapter of this dissertation. I also thank Riccardo Papa
who offered his Laboratory to carry out all the lab work required to complete the second
chapter of this dissertation.
I express my gratitude to María E. Ocasio Torres for her assistance in almost the entirety of
the field work, she endured prickles, barbs, thorns, spines, insect bites and stings, and regular
contact with three different toxic plants with grace and just a reasonable amount of
complaining. I also thank Paola B. Fernández Molinary and Luis de Jesús Pagán for their field
assistance during the early stages of this research. I send my gratitude to Jaime R. Pagán
Jiménez for taking me to two localities of Z. pumila, and inviting me to write as a collaborator a
communication on the distribution of this species on volcanic substrates. I thank Alfonso Carrero
for showing me the Z. pumila population at Ponce Landfill. I thank Marcos Caraballo for the
directions to find reproductive individuals of Z. portoricensis in Guánica. I also thank the local
residents and landowners that helped me find the zamias and allowed me to study them in their
private properties. I thank Yadira Ortiz, Silvia Planas, Ximena Velez Suazo, Fabiola Areces, and
Dania Rodríguez for their assistance in the laboratory during the population genetics study.
I thank Gladys M. Nazario for her guidance during my time as a teaching assistant at the
Botany laboratory. She helped me to approach botany teaching in a new, better way.
Finally I thank the administrative staff of the Center for Applied Tropical Ecology and
Conservation, the Graduate Biology Program, and the Department of Biology for their readiness
and assistance with the paperwork throughout my graduate studies.
vi
ABSTRACT
Variation in plant reproductive success is affected by how ecological conditions interact with
a plant’s sexual, mating, and pollination systems. Long-term viability of natural populations is
dependent on the maintenance of sufficient levels of genetic diversity. However, many
genetically rich species are continuously driven into threatened status by the cumulative effect
of human actions. This dissertation is focused on the hypothesis that the spatial distribution of
individuals within populations affects reproductive success and the patterns of intra-population
genetic structure. I used Zamia portoricensis, a short, dioecious gymnosperm, with an
underground stem, limited pollen and seed dispersal, as a model system. I answered the
following questions: (1) does pollen availability, expressed as the frequency of male and female
plants in the vicinity of a target female, affect seed set?, (2) how does distance to the nearest
mate affect seed set?, (3) is there a clear pattern of spatial genetic structure (SGS) that can be
observed at local, fine scale in Z. portoricensis, and (4) are genotypes spatially arranged in a
pattern that reflects limited pollen and seed dispersal? The relationship between spatial
distribution of individuals and reproductive success was addressed through observational field
studies in two populations of Z. portoricensis located on a serpentine outcrop of southern
Puerto Rico. The relationship between spatial distribution and the pattern of genetic structure
was addressed by the characterization of SGS in Z. portoricensis using microsatellite markers.
The data indicated that: (1) as a trend, female plants located in vicinities with more males than
females have higher reproductive success than those located in vicinities with more females
than males; (2) reproductive success is highly variable but, as a trend, it exhibits an inverse
relationship with the distance to the nearest male; (3) at the local, fine-scale used in this study,
Z. portoricensis exhibits a significant SGS; (4) there are identifiable discontinuities in the spatial
vii
distribution of genotypes. In addition, I evaluated the current conservation status of the genus
Zamia in the island using the International Union for the Conservation of Nature criteria. Newly
proposed Red List categories for the Puerto Rican zamias are: Zamia erosa Least Concern instead
of Vulnerable, Zamia portoricensis Vulnerable instead of Endangered, and Zamia pumila
Critically Endangered instead of Near Threatened.
viii
TABLE OF CONTENTS
Acknowledgements …………………………………................................................…………………………………. v
Abstract …………………………...…………………………………………………………………….………………………………. vii
List of Tables …………………………………………………………………………………………………………………...……..... xi
List of Figures ……………………..………..…………………………….……………………………………………………...….. xiii
Prologue ...………………………………………………………………………………………………………..……………..…..... xvi
Chapter 1: Best in the company of many males: female success in the threatened cycad Zamia
portoricensis ……………………………………………………………………………………………………………………..……... 1
Abstract ……………………………………………………………………………………………………………………….. 2
Introduction ...……………………………………………………………………..……………………………………….. 3
Materials and Methods ..…………………………………………………………………………………………..….. 5
Results …………………………………………………………………………………………………………………………. 9
Discussion …………………………………………………………………………………………………………………… 11
Literature Cited ...…………………………………………………………………………………….…………………. 18
Tables …………………………………………………………………………………………………………………….…… 26
Figures ……………………………………………………………………………………………………………………….. 30
Chapter 2: Small reproductive groups maintain high intra-population genetic diversity in Zamia
portoricensis ……………………………………………………………………………….………………………………………….. 36
Abstract ……………………………………………………………………………………………………………………… 37
Introduction ..………………………………………………………………………………..…………………………… 38
Materials and Methods ..…………………………………………………………………………………………..… 40
Results ……………………………………………………………………………………………………………………….. 45
ix
Discussion …………………………………………………………………………………………………………………… 46
Literature Cited ...…………………………………………………………………………………….…………………. 53
Tables …………………………………………………………………………………………………………………….….. 59
Figures ……………………………………………………………………………………………………………………….. 62
Chapter 3: Conservation status of Zamia in Puerto Rico ……………………………………………….………… 64
Abstract ……………………………………………………………………………………………………………………… 65
Introduction ..………………………………………………………………………………………………………..…… 66
Materials and Methods ..…………………………………………………………………………………………..… 68
Results ……………………………………………………………………………………………………………………….. 71
Discussion …………………………………………………………………………………………………………………… 72
Literature Cited ...…………………………………………………………………………………….…………………. 77
Tables …………………………………………………………………………………………………………………….…… 81
Figures ……………………………………………………………………………………………………………………….. 84
General Conclusions …………………………………………………………………………….…………………………………. 89
x
LIST OF TABLES
Table 1.1 Summary of reproductive and population structure data of the plants inventoried at
two populations of Zamia portoricensis. ET1, ET2 = plots at El Tamarindo population and SF1,
SF2 = plots at Susúa State Forest population; M = males; F = females; Sd = Number of seedlings
and early juveniles (i.e., less than 12 leaflets in the longest leaf); Juv = Number of juveniles (i.e.,
from 12 to 17 for ET2-SF1, or from 12 to 18 for ET1-SF leaflets); NonRA = Number of nonreproductive adults; %RA = Percentage of reproductively active adults (i.e., those that produced
cones) of the total of adult plants ………………………..……………………………………….……………………….. 27
Table 1.2. Size distribution of males and females of Zamia portoricensis at El Tamarindo and
Susúa State Forest sites. Size was measured as the number of leaflets in the longest leaf. ET1,
ET2 = plots at El Tamarindo population and SF1, SF2 = plots at Susúa State Forest population; M
= males; F = females; SE = standard error………………………………………………………………………….…..… 28
Table 1.3. Nested ANOVA analysis of the differences in size of male and female plants of Zamia
portoricensis at three levels: (1) between sites, (2) between plots within sites, and (3) between
sexes within plots. SS = Sum of squares; df = degree of freedom; MS = Mean squares ………..…. 29
Table 2.1 Descriptive statistics of population genetics and SGS for each plot. N = sample size; Na
= number of alleles; Ho, He = observed and expected heterozygosity; FIS = fixation index; F1 =
average kinship coefficient between individuals separated by 1.5 m or less; bLd = slope of the
regression of kinship coefficient values on the logarithm of the spatial distance between
individuals; Sp = extent of SGS; SE = standard error for each statistic. Values without SE are not
means ………………………….……………………………………………………………………………………………….………… 60
Table 2.2 Estimates of two gene dispersal parameters for each plot. De = effective density; Nb
= neighborhood size; σ = extent of gene dispersal (m); CI = 95% confidence interval ..……….…… 61
xi
Table 3.1 Population structure of three Zamia species of Puerto Rico. Data are from single 25 
20 m plots per locality. M = # of male plants; F = # of female plants; NRA = # of non-reproductive
adults; S-J = # of seedlings and juveniles; AOOp = area of occupancy per plot; AOOl = area of
occupancy per locality; NMsc = # of mature individuals based on size class per locality; NMcon =
# of mature individuals based on conning plants per locality .………………………..………………..….…. 82
Table 3.2 IUCN Red List categories, criteria and supporting information for Zamia in Puerto Rico.
Ze = Zamia erosa; Zpr = Zamia portoricensis, Zpm = Zamia pumila; NL = # of localities visited for
this study; EOO = extent of occurrence in km2; AOO = area of occupancy in km2; NMsc = # of
mature individuals based on size class; NMcon = # of mature individuals based on conning
plants; 2010 = present category (Stevenson, 2010); 2015 = category proposed in this study
……………………………………………………………………………………………………………………………….………….…… 83
xii
LIST OF FIGURES
Figure 1.1 Habitat physiognomy and examples of limited dispersal of Zamia portoricensis in the
south-central serpentine outcrop of Puerto Rico. (A) El Tamarindo site, showing the relatively
open, low canopy on steep slope; (B) Susúa State Forest site, showing the close, high canopy on
relatively flat terrain; (C-F) Four examples of limited dispersal in Z. portoricensis, (C) seeds
individually dispersed on a flat surface with no particular direction; (D) seeds individually
dispersed on steep surface with downhill direction; (E) seeds dispersed as a group, in a cone
fragment; (F) seeds remained as a group in a rotten but unbroken cone, at the mother
plant.………………………………………………………………………………………….………………………..…………………. 31
Figure 1.2 Maps of four 25 m x 20 m plots, showing the spatial arrangement of individuals of
Zamia portoricensis with 12 leaflets or more in the longest leaf. (A) ET1 and (B) ET2 plots at El
Tamarindo population; (C) SF1 and (D) SF2 plots at Susúa State Forest population. In each map,
the red circles represent female plants, blue triangles represent male plants, and black crosses
represent non reproductive plants. Scale bar = 5 m ………………………………………………….……………. 32
Figure 1.3 Analysis of spatial interaction between individuals in two populations of Zamia
portoricensis. Linhom, L transformation of inhomogeneous Ripley’s K-function (i.e. Kinhom). (A)
ET1 and (B) ET2 plots at El Tamarindo population; (C) SF1 and (D) SF2 plots at Susúa State Forest
population. In each chart, the solid black line shows test statistic, dashed straight red line shows
theoretical values of Kinhom(r) for CSR, and grey lines show significance envelopes for α=0.01
estimated from 499 replicates of Monte Carlo simulation ………………………………..……………………. 33
Figure 1.4 Relationship between reproductive success and sex ratio, males to females (M/F), in
Zamia portoricensis. Adjusted r2 = 0.08, F1,58 = 4.96, P < 0.03 …................................................... 34
Figure 1.5 Relationship between reproductive success and distance of the female plant to the
nearest male in Zamia portoricensis. Adjusted r2 = 0.18, F1,83 = 19.78, P < 0.001. The dashed line
xiii
represent a threshold distance after which the inverse relationship between reproductive
success and distance becomes significant ……………………………………………………………….……………… 35
Figure 2.1 Average kinship coefficient, F(d), plotted against logarithm of spatial distance, ln(d),
between individuals for each plot. ET1 = blue line; SF2 = red line ……………………………..……………. 63
Figure 2.2 Estimated cluster membership of individuals for each plot. (A) ET1; (B) SF2. Colored
ellipses indicate membership to the same group of genotypes. Scale bar = 5 m …………….….…… 63
Figure 3.1 Distribution map of Zamia in Puerto Rico. The dots represent extant localities
corroborated with field visits, and also represent the limits of the distribution of each species.
Blue dots are Zamia erosa localities, red dots are Zamia portoricensis localities, and yellow dots
are Zamia pumila localities. Three localities (green dots) were referenced from the literature,
one in Isabela, one in Río Abajo, and one in Peñuelas (Acevedo-Rodríguez & Strong, 2005;
Meerow et al., 2012) …………………………………………………………………………………………………………….… 85
Figure 3.2 General status of Zamia erosa in Puerto Rico. (A) Vega State Forest, complete
dominance of the species on the understory; (B) Road 675, several plants can be seen at the
edge of the limestone hill; (C) Julio E. Monagas National Park, the species is less dominant but
prominent on the understory; (D) Garrochales, seedlings of two different ages near the mother
plant ………………………………………..…………………………………………………………………………………...……….. 86
Figure 3.3 General status of Zamia portoricensis in Puerto Rico. (A) Susúa State Forest, a dense
group of plants at a relatively open portion of the forest; (B) El Tamarindo, a female cone with
ripe seeds ready to be dispersed; (C) general aspect of the southern serpentine outcrop of the
island; (D) Guánica 1, one of the many isolated adult plants present in the area; (E) Guánica 2, a
female cone with ripe seeds ready to be dispersed ……………………………………………….……………….. 87
Figure 3.4 General status of Zamia pumila in Puerto Rico. (A) Barrio Cuevas 1, a plant at a small
limestone outcrop, an island of high quality habitat; (B) Barrio Cuevas 4, a plant in an open area
xiv
affected by fire in the past; (C) Vertedero Ponce, one of the two healthy populations
documented; (D) Vertedero Peñuelas, a female and a male located side by side, the female cone
is almost ready to disperse the seed; (E) Yabucoa Alta, an isolated individual in volcanic
substrate …..……………………………………………………………………………………………………………………..….… 88
xv
PROLOGUE
Sexual systems (e.g. dioecy or monoecy), mating systems (e.g. selfing or outcrossing), and
pollination systems (e.g. generalist or specialist vector; reward or deception mediated
pollination) play a major role in the evolution of plants, and influence the effects that different
ecological conditions have on their reproductive success. Mating systems strongly influence the
patterns of genetic structure of populations. In fact, outcrossing and selfing leave distinctive
imprints on the genetic constitution of species, which is particularly notable in the levels of
genetic diversity. For example, obligate outcrossing species (e.g., self-incompatible and
dioecious species) are particularly affected by access to mates compared to those capable of
selfing. At the same time, access to mates is influenced by the pollination system and the
spatial distribution of individuals, among other factors.
Reproductive success in obligately outcrossing plant species is negatively correlated with
distance to the nearest mate. This negative correlation implies a reduction in the amount of
individuals that participate in sexual reproduction, which can be further reduced if pollination
success is limited by the characteristics of the pollen vector (e.g. generalist vs. specialist).
Populations affected by these limitations are particularly susceptible to genetic drift, and are
expected to exhibit low levels of genetic diversity.
Long-term viability of natural populations is dependent on the maintenance of sufficient
levels of genetic diversity. Genetic variation is an indicator of the capacity for populations to
cope with a changing environment, and for the last three decades has been widely used to
evaluate the conservation status of wild species.
Genetic diversity is generally expressed in terms of heterozygosity and allelic richness, which
are the base to calculate other descriptive statistics that have a pervasive presence in
xvi
population genetics studies. However, inferences made from classic population genetics models
are based on assumptions, (i.e., large population sizes, random mating, no population structure,
no migration, etc.) that rarely match the life history of the taxon under study. During the last
two decades methods have been developed to provide more realistic characterization of
patterns of genetic structure. They are based on the assessment of departure from random
spatial distribution of genotypes (i.e., spatial genetic structure, SGS) as the result of limited gene
dispersal (i.e., isolation by distance), and environmental heterogeneity (i.e., isolation by
resistance).
In this dissertation I focused on the hypothesis that the spatial distribution of individuals
within populations affects reproductive success and the patterns of intra-population genetic
structure. I also focused on the application of scientific results obtained at small scale studies to
evaluate the conservation status of threatened species at a larger scale. I used the genus Zamia
(Cycadales: Zamiaceae) in Puerto Rico to carry out evaluations at large scale, and the endemic
Zamia portoricensis, as a model system for studies at local, fine scale in two populations located
at the eastern portion of the southern serpentine outcrop of Puerto Rico.
Zamia is a genus of dioecious plants, which are pollinated by specialist insects, experience
pollen limitation, and have limited seed dispersal. The pattern of genetic diversity observed in
the Puerto Rican zamias, characterized by high intra-population genetic diversity and low interpopulation genetic differentiation, is generally associated with a reproductive scenario led by
panmixia at intra-population level and considerable gene flow between populations.
Paradoxically, the evidence accumulated thus far suggests that gene flow in these plants might
be limited by pollen and seed dispersal. How has Zamia overcome the expected constraints to
genetic diversity?
xvii
My objectives were: (1) to determine if pollen availability, expressed as the frequency of
male and female plants in the vicinity of a target female, affect reproductive success in Z.
portoricensis, (2) to evaluate if distance to the nearest mate affect reproductive success of
female plants, (3) to determine if there is a clear pattern of spatial genetic structure that can be
observed at local, fine scale in Z. portoricensis, (4) to detect if genotypes are spatially arranged
in a pattern that reflects limited pollen and seed dispersal, and (5) to assess the conservation
status of Zamia in Puerto Rico based on quantitative information on plant density, population
structure, and habitat range.
In chapter one I sought to determine whether variation in spatial interactions was an
indicator of success in sexual reproduction. The field studies revealed that Zamia portoricensis’
abundance at the study sites is among the highest reported for a cycad, and its population
structure suggests regular successful recruitment. Reproductive success of female plants is
highly variable in this species but, as a trend, it decreases with an increase in the distance to the
nearest male, a relationship that becomes statistically significant at distances beyond 190 cm.
Reproductive success, as a trend, is higher in female plants located at vicinities with more males
than females than in those located at vicinities with more females than males. As the individuals
are randomly located in relation to each other, being close to male plants and/or located within
a high pollen availability vicinity is a matter of chance. The effect of this stochastic process on
reproductive success is perhaps one of the factors that produce the high levels of intrapopulation genetic diversity found in this species.
In chapter two I assessed whether Z. portoricensis exhibits a significant pattern of SGS. I also
investigated the scale at which genetic material is effectively exchanged among individuals of
the species, and if limited gene dispersal produced identifiable clusters of genotypes. The study
xviii
using microsatellite markers showed that, at the small scale sampled in this study, this species
exhibits a significant deficiency in heterozygosity, while maintaining high levels of allelic
diversity. Gene dispersal estimates obtained (i.e., 2.4 and 3.1 m) are congruent with findings of
limited pollination and limited seed dispersal in this species and suggest that gene flow is very
restricted, and that at the fine-scale used in this study, Z. portoricensis exhibits a significant
spatial genetic structure, characterized by identifiable discontinuities in the distribution of
genotypes producing clusters with configurations that match some topographical features of the
habitat.
In chapter three I studied how the genetically diverse Puerto Rican zamias have coped with
the extensive habitat transformation that has affected the island in the past, and is affecting
part of these species range in the present. The field studies carried out at 36 localities on the
northern and southern limestone regions and also on the southern serpentine outcrop of Puerto
Rico, and the conservation assessment made using the International Union for the Conservation
of Nature (IUCN) Red List Categories and Criteria Version 3.1, demonstrated that of the three
species of Zamia present in the island, Z. erosa has the widest distribution both in terms of
range and number of localities, Z. portoricensis is an abundant species but it is not as widely
distributed as Z. erosa, and Z. pumila is the rarest native cycad of Puerto Rico. Although it has
been recorded at more places than Z. portoricensis, most Z. pumila populations consist of a
small number of adults that do not exhibit sexual reproduction. I propose new IUCN Red List
categories for the Puerto Rican zamias that are: Zamia erosa Least Concern instead of
Vulnerable, Zamia portoricensis Vulnerable instead of Endangered, and Zamia pumila Critically
Endangered instead of Near Threatened.
xix
CHAPTER ONE
BEST IN THE COMPANY OF MANY MALES: FEMALE SUCCESS IN THE THREATENED CYCAD
ZAMIA PORTORICENSIS
1
ABSTRACT
Variation in plant reproductive success is affected by how ecological conditions interact with
a plant’s sexual, mating, and pollination systems. For example, obligate outcrossing species are
particularly affected by access to mates compared to those capable of selfing. Here I address
the hypothesis that the spatial distribution of individuals within populations affects reproductive
success. Using the cycad Zamia portoricensis I sought to determine whether variation in spatial
interactions was an indicator of success in sexual reproduction. I focused on answering the
following questions: (1) does pollen availability, expressed as the frequency of male and female
plants in the vicinity of a target female, affect seed set?, and (2) how does distance to the
nearest mate affect seed set? I studied two populations of Z. portoricensis established on the
serpentine outcrop of southern Puerto Rico. I used four 25 x 20 m plots to characterize both
populations. In each plot, I recorded geographic location, vegetative, and reproductive data of
every individual with at least 12 leaflets in the longest leaf. I measured reproductive success as
the percentage of ovules that developed into seeds. To determine the relationship between
reproductive success and pollen availability I used circular plots of 1.5 m radius around 60
female plants. At each plot I counted the number of male and female plants and used a linear
regression analysis to assess the relationship between the seed set of the target female and the
sex ratio in its vicinity. Pollen availability was assumed to increase with an increase in the sex
ratio. I assessed the relationship between reproductive success of 85 female plants and the
distance to their nearest male using a linear regression analysis. Zamia portoricensis’ abundance
at the studied sites is among the highest reported for a cycad, and its population structure
suggests regular successful recruitment. Reproductive success is highly variable in this species
but, as a trend, it decreased with an increase in the distance to the nearest male, a relationship
2
that became statistically significant at distances beyond 190 cm. Reproductive success was
higher in plants located at vicinities with more males than females than in those located at
vicinities with more females than males. As the individuals are randomly located in relation to
each other, being close to male plants and/or located within a vicinity of high pollen availability
is a matter of chance. The effect of this stochastic process on reproductive success is perhaps
one of the factors that produce the high levels of intra-population genetic diversity found in this
species.
INTRODUCTION
Sexual systems (e.g. dioecy or monoecy), mating systems (e.g. selfing or outcrossing), and
pollination systems (e.g. generalist or specialist vector, reward or deception mediated
pollination) play a major role in the evolution of plants, and influence the effects that different
ecological factors have on their reproductive success (Barrett, 2003, 2013; Devoux et al., 2014).
Mating systems strongly influence the patterns of genetic structure of populations (Hamilton
2009). In fact, outcrossing and selfing leave a distinctive imprint on the genetic constitution of
species, which is particularly notable in their levels of genetic diversity (Hamilton 2009). For
example, obligate outcrossing species (e.g., self-incompatible and dioecious species) are
particularly affected by access to mates compared to those capable of selfing. At the same time,
access to mates is influenced by the pollination system and the spatial distribution of
individuals, among other factors (Barrett and Thomson, 1982; Heilbuth et al., 2001; Ghazoul,
2005).
Reproductive success in obligately outcrossing plant species is negatively correlated with
distance to the nearest mate (Barret and Thomson, 1982; Percy and Cronk, 1997; Metcalfe and
Kunin, 2006). This negative correlation implies a reduction in the amount of individuals that
3
participate in sexual reproduction, which can be further reduced if pollination success is limited
by the characteristics of the pollen vector (e.g. biotic vs. abiotic, generalist vs. specialist)
(Ghazoul, 2005). Populations affected by these limitations are susceptible to genetic drift, and
are expected to exhibit low levels of genetic diversity (Hamilton, 2009). At odds with this
expectation are Puerto Rican populations of Zamia (Cycadales: Zamiaceae), a genus of dioecious
plants, which are pollinated by specialist insects, experience pollen limitation (Newell, 1983,
Tang 1987a), and have limited seed dispersal (Eckenwalder, 1980; Negrón-Ortiz and Breckon,
1989a; Tang 1989). Paradoxically, these populations are characterized by high genetic diversity
(Meerow et al., 2012). How has Zamia overcome the expected constraints to genetic diversity?
Here I address the hypothesis that the spatial distribution of individuals within populations
affects reproductive success. Using Z. portoricensis Urb. as a model system, I sought to
determine whether variation in spatial interactions was an indicator of success in sexual
reproduction. I then focused on answering the following questions: (1) does pollen availability,
expressed as the frequency of male and female plants in the vicinity of a target female, affect
seed set?, and (2) how does distance to the nearest mate affect seed set?
Cycads have been considered living fossils (Chamberlain, 1919, 1935; Norstog and Nicholls,
1997) and are endangered around the world (Donaldson, 2003). Every year, to save many
species from extinction, active management involving introductions, reintroductions, and
translocations is becoming more needed beyond habitat protection. The results presented here
are valuable to design conservation programs addressed to promote self-sustainable, viable
populations of this extraordinary component of the world biodiversity.
4
MATERIALS AND METHODS
Study species— Zamia portoricensis are short plants with thick, bending upright, simple to
sparsely branched, underground stems. It is a perennial, dioecious gymnosperm endemic to the
south-central and southwestern serpentine outcrops and limestone zones of Puerto Rico
(Acevedo-Rodríguez and Strong, 2005). The reproductive cycle of the species generally starts
late in October with the appearance of the first male cones; female cones become receptive
from January to February and, if pollinated, they will support the growing seeds until cone
dehiscence around November. Ovules are pollinated by Zamia-specialist beetles, Pharaxonota
portophylla (Coleoptera: Erotylidae) (Franz and Skelley, 2008). This interaction involves a
sequential reward-deception pollination system, where the plants and the insects are mutually
dependent upon one another (Norstog et al., 1992). Beetles exclusively feed on pollen and male
cone tissue, probably mate inside the cones, and use them as a brood place (Norstog and
Nicholls, 1997). The life span of the insects is unknown but several generations are produced
during the host reproductive season (Norstog et al., 1992; Norstog and Nicholls, 1997).
Pollination occurs when receptive female plants produce volatile compounds that mimic male
scents and attract pollen-covered insects (Pellmyr et al., 1991; Terry et al., 2004; Proches and
Johnson, 2009). Female plants produce drops of a substance rich in sugars and amino acids at
micropyles; this substance could be considered a reward but it has been found that it is
produced in small amounts and at times when insects are not active inside female cones (Tang,
1987, 1993). Although the evidence is not conclusive, thus far female plants can be considered
rewardless (Donaldson, 1997). As other West Indies zamias, Z. portoricensis probably
experiences limited pollination (Newell, 1983; Tang, 1987) and limited seed dispersal
(Eckenwalder, 1980; Negrón-Ortiz and Breckon, 1989a; Tang, 1989). In the most recent
5
assessment of its conservation status, based on IUCN Red List Categories and Criteria, Z.
portoricensis was considered Endangered (EN) (Stevenson, 2010). Several field trips between
2011 and 2015 to different localities across the island revealed that many populations of the
species consist of only adult plants and exhibit no recruitment due to the extirpation of the
pollinator (Chapter 3).
Study sites—Two populations of Z. portoricensis, both established on the serpentine outcrop
of southern Puerto Rico, were included in this study. One population is located on both sides of
El Tamarindo route (Km 2.8 from intersection with route 368), Municipality of Sabana Grande
(N18o 04.368' W66 o 55.239') and the other population is located at Susúa State Forest, along
the dirt road between the main Camp and the Cabins, Municipality of Yauco (N18 o 03.972' W66o
54.076'). The zamias in El Tamarindo, hereafter named the ET population, are the most
abundant element on the understory of a low (<6 m), relatively open (canopy cover ranges from
31.4% to 90.9%, measured with a spherical densiometer), semi-deciduous forest growing on a
steep slope, where leaf litter is periodically flushed out by rains (Fig. 1A). The zamias in Susúa
State Forest, hereafter named the SF population, are also the most abundant element on the
understory of a high (>12 m), closed canopy (canopy cover ranges from 87.0% to 98.44%),
semideciduous forest growing on a nearly flat terrain, with a soil that periodically accumulates
organic matter (Fig. 1B). In both sites, Z. portoricensis has band-like, continuous distribution
along the northern slopes of the hills; bands are of approximately 80-100 m wide and extend for
about 0.4 km at ET and 2 km at SF. The species density is variable along those bands, sometimes
appearing as isolated individuals, and other times forming dense clusters of plants.
Population characterization—I established four 25 x 20 m plots, two per study site, to
characterize both populations. In each plot, I recorded plant size, measured as the number of
6
leaflets of the longest leaf and, when possible (i.e. when they produced cones or when debris of
previous cones were present), the sex of every reproductive individual. To describe the
population structure, I used the minimum size at which plants produced a cone to be the
division line between juvenile and adult size classes. Within the adult class, coning plants were
assigned to each sex, and the rest was counted as non-reproductive adults. Both populations
were regularly visited for three consecutive years from January, 2012 to December, 2014, thus
the data was frequently updated. Occurrence of biased sex ratios was tested using Chi-squared
test, and differences in size of males and females was tested using a nested ANOVA analysis.
Normality of the data was corroborated using diagnostic graphics and residual analysis. All the
analyses were performed using R software (R Development Core Team, 2014) but the nested
ANOVA was carried out in STATISTICA (StatSoft Inc., 2004).
Spatial patterns and interaction among individuals—To characterize the spatial
arrangement of the individuals of Z. portoricensis, I first recorded the geographic position of all
juveniles and adults and then mapped them. The spatial data was processed using Spatstat, an R
package for analyzing point patterns in two dimensions (Baddeley and Turner, 2005; Baddeley,
2010). Limited local dispersal of the West Indies zamias (Eckenwalder, 1980; Negrón-Ortiz and
Breckon, 1989a; Tang, 1989) (Fig. 1C-F), originates a non-stationary (i.e. inhomogeneous) point
process. Thus, to account for this inhomogeneity, an adjustment was applied to analyze the
interaction between points in the patterns obtained for each plot (Baddeley et al., 2000; Perry
et al., 2006; Law et al., 2009; Baddeley, 2010). To facilitate the visual assessment of the results, I
used an L transformation to plot the Poisson K function as a straight line. Rejection limits were
estimated as simulation envelopes based on a null hypothesis of Complete Spatial Randomness
7
(CSR) (Perry et al., 2006; Baddeley, 2010). I used 499 replicates of a Monte Carlo simulation to
obtain the significance envelopes for α = 0.01 (Perry et al., 2006).
Evaluation of reproductive success—Using mapped individuals, I initially selected 60 female
plants (i.e., 30 per site) spread across the plots and well separated from each other to assess
their reproductive success. However, only 33 of the 60 plants were fertile at the beginning of
this study (March 2013), so I added another 27 fertile plants from the plots and adjacent areas
to complete the 30 individuals per site. In October 2013, around seven months after pollination
took place and one month before seeds were completely ripe, the female cones were enclosed
in perforated plastic bags to prevent seed loss from shedding mature cones, and by the end of
November 2013, I recorded the total number of ovules per cone, and the number of seeds were
recorded. Only one plant produced more than a single cone and for this plant I averaged the
data. Reproductive success was measured as the percentage of ovules that developed into
seeds (i.e., seed set), which reflects the relationship between reproductive effort and pollination
success.
Reproductive success and pollen availability—To determine the relationship between seed
set and pollen availability I made a linear regression analysis of reproductive success on sex ratio
(i.e. males/females). I created circular plots of 1.5 m radius around each of the 60 previously
marked female plants, none of which overlapped. In each plot, male and female plants were
counted. Pollen availability was assumed to increase with an increase in the sex ratio. Normality
of the data was corroborated using diagnostic graphics and residuals analysis. The analyses were
performed using R software (R Development Core Team, 2014).
Reproductive success and distance to the nearest male—The relationship between seed set
and the distance to the nearest male was assessed through a linear regression analysis. I
8
included 85 plants in this study (i.e. 44 from El Tamarindo and 41 from Susúa State Forest).
Before combining the data from both sites I analyzed them separately, both were similar and
significant (i.e. ET: b = -0.19, r2 = 0.11, P < 0.05; SF: b = -0.33, r2 = 0.27, P < 0.001; Z = 1.26 n.s.).
To identify a threshold distance after which the effect of the separation between a female plant
and the nearest male results in a significant reduction of reproductive success I used an
approach similar to the one used by Ackerman et al. (2007). I first regressed reproductive
success on distance to the nearest male and then repeatedly re-ran the regression by
progressively eliminating the next longest distance until the relationship became unambiguously
non-significant (P > 0.1). I considered that last distance as the approximate threshold. Normality
of the data was corroborated using diagnostic graphics and residuals analysis. All the analyses
were performed using R software (R Development Core Team, 2014).
RESULTS
Summary of reproductive data and population structure—I inventoried 5689 plants (Table
1), approximately half of the plants were located in a single plot (ET1) and there was a five-fold
difference in plant density between this plot and the plot with the lowest number of individuals
(ET2).
The percentage of adult plants, inferred from vegetative data (e.g. size class), ranged
between 20% (plot SF1) and 50% (plot SF2). The number of reproductively active adults (i.e.,
plants that actually produced cones within the adult size class) ranged from 16% (plot SF2) to
52% (plot ET2) of the plants inventoried (Table 1).
Sex ratios were male biased in all plots (Table 1). However, the difference between number
of males and females was not statistically significant for one plot (ET2).
9
Male plants were reproductively active at sizes below 20 leaflets in the longest leaf (Table 2)
and this was similar in all plots. Female plants matured at different sizes (e.g., 24-37 leaflets)
and started producing cones at sizes between 21 to 30 leaflets in the longest leaf at ET plots,
and at sizes above 30 leaflets at SF plots (Table 2). The difference between the average size of
male and female plants was statistically significant for each plot (Tables 2 and 3), no difference
was detected between plots from the same site (Table 3), but significant difference between
sites was significant (Table 3).
Spatial patterns and interactions between individuals—Although several plants appeared to
be grouped (Fig. 2), the analysis of each spatial point pattern revealed that, at the scale
considered in this study, there is no significant spatial interaction among the individuals of Z.
portoricensis. The observed values of Linhom are not statistically different from the theoretical
Poisson values that indicate complete spatial randomness (Fig. 3). Consequently, individuals are
independent of each other (i.e., randomly distributed) and do not exhibit association (i.e.,
clustering) or repulsion (i.e., regularity) at the scale measured and regardless of their size or sex.
Reproductive success and pollen availability—The regression analysis showed that there is a
significant positive relationship between the reproductive success of a target female plant and
the ratio of males to females the target female (Adjusted r2 = 0.08, F1,58 = 4.96, P < 0.03) (Fig. 4).
Although statistically significant, this relationship is not very strong as reproductive success was
highly variable (Fig. 4).
Reproductive success and distance to the nearest male—The regression analysis showed
that there is a significant inverse relationship between seed set and the distance to the nearest
male (Adjusted r2 = 0.18, F1,83 = 19.78, P < 0.001) (Fig. 5). However reproductive success was
10
highly variable at distances up to 190 cm when seed set dramatically dropped off and remained
less variable (Fig. 5).
DISCUSSION
Population characterization—Zamia portoricensis is abundant at the sites studied; its
average density, excluding the seedlings and early juveniles, of 198 plants/100 m2 at SF, and 370
plants/100 m2 at ET show that the species is a prominent understory element at those sites (Fig.
1A, B). Studies on other cycads show densities as low as 0.25 plants/100 m2 in Microcycas
calocoma (Lazcano-Lara, 2004), and also densities of 3–8 plants/100 m2 in Z. melanorrhachis
(López-Gallego, 2008); 14–24 plants/100 m2 in Cycas armstrongii (Watkinson and Powell, 1997);
11–31 plants/100 m2 in Ceratozamia matudai (Pérez-Farrera et al., 2000), and densities as high
as 46 plants/100 m2 in Dioon edule (Vovides, 1990). In Puerto Rico, large densities have been
also reported for Z. erosa, 350 and 335 plants/100 m2 (Newell, 1983; Negrón-Ortiz and Breckon,
1989b) suggesting that in the island, at some localities, habitat conditions have been favorable
enough to allow these species to thrive. However, density estimates just provide an idea of the
relative abundance of the species at a local level; they are heavily influenced by the life form of
the species studied and the sampling method used to estimate them; thus no further
comparisons should be made using this data.
The population structure of cycads, as in many long-lived perennials, is generally
characterized by a “J inverted” shape, with large numbers of individuals at seedling and early
juvenile size classes and very few individuals in the late adulthood size class (Negrón-Ortiz and
Breckon, 1989b; Vovides, 1990; Watkinson and Powell, 1997; Pérez-Farrera et al., 2000, 2006;
Lazcano-Lara, 2004; Yáñez-Espinosa and Sosa-Sosa, 2007; López-Gallego, 2008; Gerlach, 2012).
11
At El Tamarindo and Susúa State Forest, Z. portoricensis exhibits the same population structure
pattern reported for other cycads. Thus, sexual reproduction is occurring at both sites (Table1).
In spite of theoretical expectations of an equilibrium 1:1 sex ratio, biased operational sex
ratios are common in populations of dioecious species (Field et al., 2012), and the populations
of Z. portoricensis studied here are no exception. Three of four plots reported here were
significantly male-biased. However, in cycads, male-biased sex ratios seems to be as frequent as
equal sex ratios. Studies on nine cycad species have reported populations with more males than
females (Ornduff, 1985, 1987; Clark and Clark, 1987; Tang, 1990; Negrón-Ortiz and Gorchov,
2000; Pérez-Farrera et al., 2000; Hall et al., 2004; Kono and Tobe, 2007; Proches and Johnson,
2009), while studies on another seven species have reported statistically equal proportions of
males and females (Ornduff 1987; Watkinson and Powell, 1997; Raimondo and Donaldson,
2003; Lazcano-Lara, 2004; Pérez-Farrera et al., 2006; Yáñez-Espinosa and Sosa-Sosa, 2007). Thus
far, female biased sex ratio may be the rarest one because it has only been reported in two
species (Newell, 1983; Mora et al., 2013).
Cost of reproduction for females is generally higher than for males (Obeso, 2002).
Differential reproductive investment, particularly in long-lived species, has been proposed as the
main cause behind operational male-biased sex ratios (Field et al., 2012). Indeed, Clark and Clark
(1987), and Tang (1990) demonstrated that female cycads invest more in reproduction than do
males and might not have the resources to cone every year. However, an increase in habitat
quality, mainly in light availability, may compensate for the differences between sexes in
reproductive investment (Clark and Clark, 1987), so that more females produce cones (i.e.,
fewer non-reproductive adult females) resulting in a 1:1 operational sex ratio as Ornduff (1987)
found in Z. pumila. In the present study, habitat quality also may be affecting sex ratios. The
difference in sex ratios was less pronounced at El Tamarindo plots (Table 1) where plants
12
received more light than those at Susúa (mean canopy cover = 63% and 72%, versus 94%, and
93% for Susúa). Finally, precocious male coning also contributed to the bias in sex proportions as
male plants reached maturity at smaller sizes than females (Tables 2 and 3).
Spatial patterns and interactions between individuals—Limited dispersal and environmental
heterogeneity are expected to result in an aggregated spatial arrangement of individuals but, in
these cases, clustering does not necessarily imply dependence or attraction between them (Law
et al., 2009).
Aggregated distribution of individuals in cycad populations occurs in: Dioon edule (Vovides,
1990), Ceratozamia matudai (Pérez-Farrera et al., 2000), Ceratozamia mirandae (Pérez-Farrera
et al., 2006), and Macrozamia ridlei (Gerlach, 2012). Limited dispersal and habitat heterogeneity
were hypothesized as major factors influencing the spatial pattern, although, in all the cases,
animal dispersal of a portion of the seeds was suspected (Vovides 1990, Pérez-Farrera et al.,
2000; 2006) or corroborated (Gerlach, 2012). Vovides (1990) hypothesized that competition
among the cycad D. edule and angiosperms was the main cause of the aggregated pattern, as
rapid growing angiosperms were using spaces with deep soil layer, excluding the cycads to
patches with a thin soil layer and rock surfaces. In the present study, such interspecific
competition is not a probable cause for clustering because Z. portoricensis is clearly the
dominant element on the understory (Fig. 2). Interestingly, Gerlach (2012) reported that a
Thomas model provided the best description for the spatial pattern of M. ridlei. Its pattern may
be a result of a Poisson cluster process in which plants stay aggregated after successive Poisson
(i.e. random) processes; thus, aggregation does not necessarily imply dependence or attraction
between individuals (Baddeley and Turner, 2005; Baddeley, 2010). As Vovides (1990) and PérezFarrera et al. (2000; 2006) only used a first order statistic based on differences in plant densities
13
(Perry et al., 2006; Law et al., 2009), it is impossible to determine whether aggregation is the
result of true dependence among individuals or just a lack of homogeneity in the spatial point
process.
The spatial pattern exhibited by Z. portoricensis, relatively grouped individuals without any
interaction between them (Figs. 2, 3), has two notable implications for the reproductive biology
of the species in these populations. First, proximity among individuals might provide stability for
a deception-based pollination system with rewardless female plants, since pollinators will have
an increased probability of finding a reward (i.e., a male plant with an available cone) after being
deceived which reduces the cost for the insects and may make it unnecessary to increase their
ability to identify deceptive females (Renner, 2006); and second, as the lack of dependence or
repulsion among individuals means that the presence of both sexes at the same place is
completely random, access to mates and, consequently, reproductive success will be subjected
to a high levels of stochasticity.
Reproductive success and pollen availability—Pollination systems with a deception
component, as the one exhibited by cycads, are based on a balance between the reward
obtained by the pollinators from using male sexual structures, and the cost of visiting deceitful,
non-rewarding female sexual structures (Norstog and Nicholls, 1997; Dufaÿ and Anstett, 2003).
This mutual exploitation appears to be sufficiently stable for long-term population persistence,
even when in various cycad species, seed set values rarely go over 60% of the ovules developed
into seeds (Newell, 1983; Tang, 1987; Terry, 2001; Wilson, 2002; Kono and Tobe, 2007; Proches
and Johnson, 2009; Suinyuy et al., 2009) with many plants showing evidence of receiving few or
no pollen at all (Newell, 1983; Tang, 1987a; Kono and Tobe, 2007; Proches and Johnson, 2009).
Half of the plants assessed in this study had seed set below 50%, and 16% of the plants did not
14
produce any seeds. While definitive experiments have not yet been done, the indirect evidence
indicates pollen limitation is common and often severe.
In cycads, pollen transfer requires the simultaneous presence of a male plant with a
dehiscent cone hosting a group of pollinators, and a receptive female within flying distance from
the male. Assuming that zamias experience some degree of pollen limitation, receptive females
are competing for pollination service. Pollen availability will be higher for females that have
access to more males and less competition from other females. In this study, I found a positive
relationship between reproductive success of a target female plant and the immediateneighborhood sex ratio (Fig. 4). Assuming equal survival rates, plants located in vicinities with
more males than females should have higher fitness than plants located in vicinities with more
females than males.
Reproductive success and distance to the nearest male—In deceptive pollination systems,
pollination is a collateral benefit for female plants because insects do no profit from visiting
female sexual structures. Intersexual mimicry plays an important role in maintaining this type of
interaction between plants and their insect pollinators (Dufaÿ and Anstett, 2003). In cycads,
females mimic the fragrance profile and thermogenesis pattern of males but this mimicry is not
perfect, and females emit odors in lower amounts, with small variations in the volatile
compounds (Pellmyr et al., 1991; Terry et al., 2004; Donaldson, 2007; Proches and Johnson,
2009; Suinyuy et al., 2010). As in other model-mimicry systems, the closer the mimic is to the
model, the better the reproductive success (Peter and Johnson, 2008). In Z. portoricensis, the
probability of female plants being detected by beetles should be higher the closer they are to
male plants. At the same time, proximity to a male should reduce the cost of visiting a
rewardless female for the deceived insects. Proches and Johnson (2009), in their study of
15
Stangeria eriopus pollination, found no effect of distance to the nearest male on fecundity (r2 =
0.043, df = 61, P = 0.114), whereas I showed that seed set in Z. portoricensis is significantly
associated with the proximity to males. Stangeria eriopus is pollinated by generalist insects that
feed on pollen but do not use male cones as a brood place, these insects cover longer distances
while foraging, and perhaps that is why Proches and Johnson did not find a relationship
between the distance to the nearest male and fecundity in this cycad.
Pellmyr et al. (1991) showed that odor emissions in Zamia are weak compared to other cycad
genera. This might explain why, in the absence of strong signaling mechanisms, a mate
proximity is more critical for the pollination of Z. portoricensis than for other cycads. However,
the relative low strength (Adjusted r2 = 0.18) of this relationship suggests that either a high level
of stochasticity exists or other factors in addition to mate proximity affect seed set in Z.
portoricensis. For example, pollination will not occur if nearby male plants fail to attract and
host pollinators, or if receptivity of female cones is asynchronous with pollen release by male
cones (Norstog and Nicholls, 1997).
Conclusions— Zamia portoricensis is well established at the eastern portion of the southern
serpentine outcrop of Puerto Rico. Its abundance in those places is among the highest reported
for a cycad, and its population structure suggests regular successful recruitment.
The reproductive success of Z. portoricensis, at the individual level, is affected by the spatial
distribution of the plants. Although reproductive success is highly variable, as a trend, it
decreases with an increase in the distance to the nearest male, a relationship that becomes
statistically significant at distances beyond 190 cm. Reproductive success also tend to be higher
for female plants at vicinities with more males than females than for those located at vicinities
with more females than males. As the individuals are randomly located in relation to each other,
16
being close to male plants and/or located within a high pollen availability vicinity is a matter of
chance. The effect of this stochastic process on reproductive success is perhaps one of the
factors that produce the high levels of intra-population genetic diversity found by Meerow et al.
(2012) in this species.
The remarkable long-term persistence of the cycads indicates the effectiveness of their
sequential nursery-deception pollination system. The variability of its effectiveness at the
individual level in Z. portoricensis could be interpreted as the result of the stochasticity
associated with random spatial distribution of individuals of both sexes and the lack of
synchronization between pollen release and female receptivity. But cycads are long-lived plants
so net reproductive success is perhaps best assessed via long-term studies to capture
fluctuations in fitness.
Puerto Rico has been heavily impacted for more than 200 years leaving the forest cover to
about 5% of its original size by the 1940s. While a significant recovery of the forest cover has
occurred during the last 70 years (Grau et al., 2003; Helmer, 2004; Kennaway and Helmer, 2007)
many biological interactions may not have experienced the same recovery. In accordance with
the results presented here, any management plan addressed to promote viable populations of
this species should be based on establishing populations with densities that will facilitate
effective access to mates, and with an adequate number of plants of both sexes to maximize
reproductive success.
17
LITERATURE CITED
ACEVEDO-RODRÍGUEZ, P. AND M.T. STRONG. 2005. Monocotyledons and Gymnosperms of Puerto
Rico and the Virgin Islands. Contributions from the United States National Herbarium
Volume 52: 1–415.
ACKERMAN, J.D., J.C. TREJO-TORRES, AND Y. CRESPO -CHUY. 2007. Orchids of the West Indies:
predictability of diversity and endemism. Journal of Biogeography 34: 779–786.
AXELROD, F.S. 2011. A Systematic Vademecum to the Vascular Plants of Puerto Rico. Botanical
Research Institute of Texas, Texas, USA.
BADDELEY , A. 2010. Analysing spatial point patterns in R. Workshop Notes Version 4.1.
Commonwealth Scientific and Industrial Research Organization, CSIRO. 232pp.
BADDELEY , A., J. MOLLER AND R. WAAGEPETERSEN. 2000. Non- and semiparametric estimation of
interaction in inhomogeneous point patterns. Statistica Neerlandica 54: 329–350.
BADDELEY , A. AND R. TURNER. 2005. Spatstat: an R package for analyzing spatial point patterns.
Journal of Statistical Software 12: 1–42.
BARRETT , S.C.H. 2003. Mating strategies in flowering plants: the outcrossing-selfing paradigm
and beyond. Philosophical Transactions of the Royal Society of London, B, Biological
Sciences 358: 991–1004
BARRETT , S.C.H. 2013. The evolution of plants reproductive systems: how often are transitions
irreversible? Proceedings of the Royal Society of London, B, Biological Sciences 280:
20130913.
BARRETT , S.C.H. AND J.D. THOMSON. 1982. Spatial patterns, floral sex ratios, and fecundity in
dioecious Aralia nudicaulis (Araliaceae). Canadian Journal of Botany 60: 1662–1670.
CHAMBERLAIN, C.J. 1919. The Living Cycads. The University of Chicago Press. Chicago. IL.
18
CHAMBERLAIN, C.J. 1935. Gymnosperms: Structure and Evolution. University of Chicago Press,
Chicago. IL.
CLARK, D.A. AND D.B. CLARK. 1987. Temporal and environmental pattern of reproduction in
Zamia skinneri, a tropical rainforest cycad. Journal of Ecology 75: 135–149.
DEVAUX C., C. LEPERS AND E. PORCHER. 2014. Constraints imposed by pollinator behaviour on the
ecology and evolution of plant mating systems. Journal of Evolutionary Biology 27: 1413–
1430.
DONALDSON, J.S. 1997. Is There a Floral Parasite Mutualism in Cycad Pollination? The Pollination
Biology of Encephalartos villosus (Zamiaceae). American Journal of Botany 84: 1398–1406.
DONALDSON, J.S. 2003. Status Survey and Conservation Action Plan: Cycads. IUCN/SSC Cycads
Specialist Group. IUCN, Gland, Switzerland and Cambridge, UK.
DONALDSON, J.S. 2007. Hot and Smelly Sex! Theoretical Considerations of the Adaptive
Significance of Cone Thermogenesis and Volatile Odors in Cycad Reproduction. Memoirs of
the New York Botanical Garden 97: 308–-325.
DUFAŸ, M. AND M-C. ANSTETT . 2003. Conflicts between plants and pollinators that reproduce
within inflorescences: evolutionary variations on a theme. Oikos 100: 3–14.
ECKENWALDER , J.D. 1980. Dispersal of the West Indian Cycad, Zamia pumila L. Biotropica 12: 79–
80.
FIELD, D.L., M. PICKUP AND S.C.H. BARRETT . 2012. Comparative analyses of sex ratio variation in
dioecious flowering plants. Evolution 67: 661–672.
FRANZ, N.M. AND P.E. SKELLEY . 2008. Pharaxonota portophylla (Coleptera: Erotylidae), new
species and pollinator of Zamia (Zamiaceae) in Puerto Rico. Caribbean Journal of Science
44: 321–333.
19
GERLACH, M.A. 2012. The population structure and dynamics of Macrozamia ridlei within the
Perth region. B.Sc. dissertation, School of Environmental Science, Murdoch University, Australia.
GHAZOUL, J. 2005. Pollen and seed dispersal among dispersed plants. Biological Reviews 80: 413–
443.
GRAU, H.R., AIDE, T.M., ZIMMERMAN, J.K., THOMLINSON, J.R., HELMER, E. AND ZOU, X. 2003. The
Ecological Consequences of Socioeconomics and Land-Use Changes in Postagriculture
Puerto Rico. BioScience 53: 1159–1168.
HALL, J.A., G.H. W ALTER, D.M. BERGSTROM AND P. MACHIN. 2004. Pollination ecology of the
Australian cycad Lepidozamia peroffskyana (Zamiaceae). Australian Journal of Botany 52:
333–343.
HAMILTON, M.B. 2009. Population Genetics. Wiley-Blackwell, Chichester, West Sussex, UK.
HEILBUTH, J.C., K.L. ILVES AND S.P. OTTO. 2001. The consequences of dioecy for seed dispersal:
modelling the seed-shadow handicap. Evolution 55: 880–888.
HELMER, E.H. 2004. Forest conservation and land development in Puerto Rico. Landscape
Ecology 19: 29–40.
KENNAWAY , T. AND HELMER, E.H. 2007. The Forests Types and Ages Cleared for Land Development
in Puerto Rico. GIScience & Remote Sensing 44: 356–382.
KONO, M., AND H. TOBE.2007. Is Cycas revoluta (Cycadaceae) wind or insect pollinated? American
Journal of Botany 94: 847–855.
LAW, R., J. ILLIAN, D. F. R. P. BURSLEM, G. GRATZER , C. V. S. GUNATILLEKE AND I. A. U. N.
GUNATILLEKE . 2009. Ecological information from spatial patterns of plants: insights from
point process theory. Journal of Ecology 97: 616–628.
20
LAZCANO-LARA, J.C. 2004. Biología poblacional y aspectos ecológicos de Microcycas calocoma
(Miq.) A.DC. en la Sierra de La Caoba, Viñales. M.Sc. dissertation, Universidad de La Habana,
Cuba.
LÓPEZ-GALLEGO , C. 2008. Demographic variation in cycad populations inhabiting contrasting
forest fragments. Biodiversity Conservation 17: 1213–1225.
MEEROW, A., J. FRANCISCO -ORTEGA, M. CALONJE , M.P. GRIFFITH, T. AYALA-SILVA, D.W. STEVENSON
AND K. N AKAMURA . 2012. Zamia (Cycadales: Zamiaceae) on Puerto Rico. Asymmetric genetic
differentiation and the hypothesis of multiple introductions. American Journal of Botany
99: 1828–1839.
METCALFE , D.B. AND W.E. KUNIN. 2006. The effects of plant density upon pollination success,
reproductive effort and fruit parasitism in Cistus ladanifer L. (Cistaceae). Plant Ecology 185:
41–47.
MORA, R., L. YÁÑEZ-ESPINOSA, J. FLORES AND N. NAVA-ZÁRATE. 2013. Strobilus and seed production
of Dioon edule (Zamiaceae) in a population with low seedling density in San Luis Potosí,
Mexico. Tropical Conservation Science 6: 268–282.
NEGRÓN-ORTIZ, V. AND G.J. BRECKON. 1989a. A Note on the dispersal of Zamia (Zamiaceae) in
Puerto Rico. Caribbean Journal of Science 25: 86–87.
NEGRÓN-ORTIZ, V. AND G.J. BRECKON. 1989b. Population structure in Zamia debilis (Zamiaceae) I.
Size classes, leaf phenology, and leaf turnover. American Journal of Botany 76: 891–900.
NEGRÓN-ORTIZ, V., D.L. GORCHOV AND G.J. BRECKON. 1996. Population structure in Zamia
(Zamiaceae) in northern Puerto Rico. II. Seed germination and stage-structured population
projection. International Journal of Plant Sciences 157: 605–614.
21
NEGRÓN-ORTIZ, V. AND D.L. GORCHOV . 2000. Effects of fires season and postfire herbivory on the
cycad Zamia pumila (Zamiaceae) in slash pine savannah, Everglades National Park, Florida.
International Journal of Plant Sciences 161: 659–669.
NEWELL , S.J. 1983. Reproduction on a natural population of cycads (Zamia pumila L.) in Puerto
Rico. Bulletin of the Torrey Botanical Club 110: 464–473.
NORSTOG, K.J., P.K.S. FAWCETT AND A.P. VOVIDES. 1992. Beetle pollination of two species of
Zamia: Evolutionary and ecological considerations. The Palaeobotanist 41: 149–158.
NORSTOG, K.J. AND T.J. NICHOLLS. 1997. The Biology of the Cycads. Cornell University Press,
Ithaca, NY.
OBESO, J.R. 2002. The costs of reproduction in plants. New Phytologist 155: 321–348.
ORNDUFF, R. 1985. Male biased sex ratios in the cycad Macrozamia ridlei (Zamiaceae). Bulletin of
the Torrey Botanical Club 112: 393–397.
ORNDUFF, R. 1987. Sex Ratios and Coning Frequency of the Cycad Zamia pumila L. (Zamiaceae) in
the Dominican Republic. Biotropica 19: 361–364.
PELLMYR, O., W. TANG, I. GROTH, G. BERGSTROM AND L.B. THIENS. 1991. Cycad Cone and
Angiosperm Floral Volatiles: Inferences for the Evolution of Insect Pollination. Biochemical
Systematics and Ecology 19: 623–627.
PERCY, D.M. AND Q.B.C. CRONK. 1997. Conservation in relation to mating system in Nesohedyotis
arborea (Rubiaceae), a rare endemic tree from St Helena. Biological Conservation 80: 135–
145.
PÉREZ-FARRERA, M.A., P.F. QUINTANA-ASCENCIO , B.S. IZABA AND A.P. VOVIDES. 2000. Population
dynamics of Ceratozamia matudai Lundell (Zamiaceae) in El Triunfo Biosphere Reserve,
Chiapas, Mexico. Journal of the Torrey Botanical Society 127: 291–299.
22
PÉREZ-FARRERA, M.A., A.P. VOVIDES, P. OCTAVIO-AGUILAR, J. GONZÁLEZ-ASTORGA, J. DE LA CRUZRODRÍGUEZ, R. HERNÁNDEZ -JONAPÁ AND S.M. VILLALOBOS-MÉNDEZ. 2006. Demography of the
cycad Ceratozamia mirandae (Zamiaceae) under disturbed and undisturbed conditions in a
biosphere reserve of Mexico. Plant Ecology 187: 97–108.
PERRY, G.L.W., B.P. MILLER AND N.J. E NRIGHT . 2006. A comparison of methods for the statistical
analysis of spatial point patterns in plant ecology. Plant Ecology 187: 59–82.
PETER, C. I. AND S. D. JOHNSON. 2008. Mimics and magnets: the importance of color and
ecological facilitation in floral deception. Ecology 89: 1583–1595.
PROCHES S. AND S.D. JOHNSON. 2009. Beetle pollination of the fruit-scented cones of the South
African cycad Stangeria eriopus. American Journal of Botany 96: 1722–1730.
R DEVELOPMENT CORE TEAM. 2014. R: A language and environment for statistical computing.
Version 3.1.2. The R Foundation for Statistical Computing, Vienna, Austria.
RAIMONDO , D.C. AND J.S. DONALDSON . 2003. Responses of cycads with different life histories to
the impact of plant collecting: simulation models to determine important life history stages
and population recovery times. Biological Conservation 111: 345–358.
RENNER, S. S. 2006. Rewardless flowers in the angiosperms and the role of insect cognition in
their evolution. In N. M. Waser and J. Ollerton [eds.], Plant-Pollinator Interactions: From
Specialization to Generalization, 123–144. University of Chicago Press, Chicago.
STATISTICA. 2004. StatSoft, Inc., Tulsa, Oklahoma.
STEVENSON, D.W. 2010. Zamia erosa, Z. portoricensis, Z. pumila. In IUCN 2013. IUCN Red List of
Threatened Species. Version 2013.2. Website http://www.iucnredlist.org [accessed on 10
February 2014].
SUINYUY , T.N., J.S. DONALDSON AND S.D. JOHNSON. 2009. Insect pollination in the African cycad
Encephalartos friderici-guilielmi Lehm. South African Journal of Botany 75: 682–688.
23
SUINYUY , T.N., J.S. DONALDSON AND S.D. JOHNSON. 2010. Scent chemistry and patterns of
thermogenesis in male and female cones of the African cycad Encephalartos natalensis
(Zamiaceae). South African Journal of Botany 76: 717–725.
TANG, W. 1987. Insect Pollination in the Cycad Zamia pumila (Zamiaceae). American Journal of
Botany 74: 90–99.
TANG, W. 1989. Seed dispersal in the cycad Zamia pumila in Florida. Can. J. Bot. 67: 2066-2070.
TANG, W. 1990. Reproduction in the Cycad Zamia pumila in a Fire-Climax Habitat: An Eight-Year
Study. Bulletin of the Torrey Botanical Club 117: 368–374.
TANG, W. 1993. Nectar-like secretions in female cones of cycads. The Cycad Newsletter 16: 10–
13.
TERRY, I. 2001. Thrips and Weevils as Dual, Specialist Pollinators of the Australian Cycad
Macrozamia communis (Zamiaceae). International Journal of Plant Sciences 162: 1293–
1305.
TERRY, I., C.J. MOORE , G.H. WALTER, P.I. FORSTER , R.B. ROEMER, J.S. DONALDSON AND P.J. MACHIN.
2004. Association of cone thermogenesis and volatiles with pollinator specificity in
Macrozamia cycads. Plant Systematics and Evolution 243: 233–247.
TERRY, L.I., G.H. WALTER, J.S. DONALDSON, E. SNOW, P.I. FORSTER AND P.J. MACHIN. 2005.
Pollination of Australian Macrozamia Cycads (Zamiaceae): Effectiveness and Behavior of
Specialist Vectors in a Dependent Mutualism. American Journal of Botany 92: 931–940.
VOVIDES, A. P. 1990. Spatial distribution, survival, and fecundity of Dioon edule (Zamiaceae) in a
tropical deciduos forest in Veracruz, Mexico, with notes on its habitat. American Journal of
Botany 77: 1532–1543.
WATKINSON, A.R. AND J.C. POWELL . 1997. The life history and population structure of Cycas
armstrongii in monsoonal northern Australia. Oecologia 111: 341–349.
24
WILSON, G.W. 2002. Insect Pollination in the Cycad Genus Bowenia Hook. ex Hook. f.
(Stangericaceae). Biotropica 34: 438–441.
YÁÑEZ-ESPINOSA, L. AND F. S OSA-SOSA. 2007. Population structure of Dioon purpusii Rose in
Oaxaca, Mexico. Neotropical Biology and Conservation 2: 46–54.
25
CHAPTER ONE
TABLES
26
Table 1. Summary of reproductive and population structure data of the plants inventoried at
two populations of Zamia portoricensis. ET1, ET2 = plots at El Tamarindo population and SF1,
SF2 = plots at Susúa State Forest population; M = males; F = females; Sd = Number of seedlings
and early juveniles (i.e., less than 12 leaflets in the longest leaf); Juv = Number of juveniles (i.e.,
from 12 to 17 for ET2-SF1, or from 12 to 18 for ET1-SF leaflets); NonRA = Number of nonreproductive adults; %RA = Percentage of reproductively active adults (i.e., those that produced
cones) of the total of adult plants.
Site
Total
M
F
M/F1
Sd
ET1
3099
250
187
1.34a
2100
ET2
612
59
46
SF1
1065
73
34
SF2
913
83
24
NonRA
%RA
220
342
44
410
22
75
52
2.14b
804
43
111
41
3.46c
258
194
354
16
1.28
Juv
Notes:
1
Sex ratio and significant Chi-square test results for null hypothesis males to females proportion
will be 1:1. a χ2 = 9.08, df = 1, P = 0.003; b χ2 = 14.21, df = 1, P < 0.001; c χ2 = 32.53, df = 1,
P << 0.001.
27
Table2. Size distribution of males and females of Zamia portoricensis at El Tamarindo and Susúa
State Forest sites. Size was measured as the number of leaflets in the longest leaf. ET1, ET2 =
plots at El Tamarindo population and SF1, SF2 = plots at Susúa State Forest population; M =
males; F = females; SE = standard error.
ET1
Size class
ET2
SF1
SF2
M
F
M
F
M
F
M
F
10-20
9
0
13
1
3
0
5
0
21-30
72
12
13
7
8
0
27
0
31-40
102
58
21
9
32
10
34
2
41-50
57
81
10
16
22
14
11
14
51-60
9
27
2
12
7
10
5
6
61-70
1
7
0
1
1
0
1
0
> 70
0
2
0
0
0
0
0
2
250
187
59
46
73
34
83
24
35.33
43.33
31.47
42.41
38.89
45.29
34.10
49.5
8.76
9.13
10.23
11.13
9.81
7.57
9.81
10.77
% 10 - 20
3.6
0
22.03
2.17
4.11
0
6.02
0
% 21 - 30
28.88
22.03
15.22
10.96
0
32.53
0
Total
Mean size
SE
6.42
28
Table 3. Nested ANOVA analysis of the differences in size of male and female plants of Zamia
portoricensis at three levels: (1) between sites, (2) between plots within sites, and (3) between
sexes within plots. SS = Sum of squares; df = degree of freedom; MS = Mean squares.
SS
df
MS
F
P
Site
1602.0
1
1602.0
18.488
< 0.001
Plot (Site)
477.6
2
238.8
2.756
0.064193
Sex (Site*Plot)
15313.9
4
3828.5
44.182
< < 0.001
Error
64815.8
748
86.7
29
CHAPTER ONE
FIGURES
30
Figure 1. Habitat physiognomy and examples of limited dispersal of Zamia portoricensis in the
southcentral serpentine outcrop of Puerto Rico. (A) El Tamarindo site, showing the relatively
open, low canopy on steep slope; (B) Susúa State Forest site, showing the close, high canopy on
relatively flat terrain; (C-F) Four examples of limited dispersal in Z. portoricensis, (C) seeds
individually dispersed on a flat surface with no particular direction; (D) seeds individually
dispersed on steep surface with downhill direction; (E) seeds dispersed as a group, in a cone
fragment; (F) seeds remained as a group in a rotten but unbroken cone, at the mother plant.
31
Figure 2. Maps of four 25 m x 20 m plots, showing the spatial arrangement of individuals of
Zamia portoricensis with 12 leaflets or more in the longest leaf. (A) ET1 and (B) ET2 plots at El
Tamarindo population; (C) SF1 and (D) SF2 plots at Susúa State Forest population. In each map,
the red circles represent female plants, blue triangles represent male plants, and black crosses
represent non reproductive plants (i.e., juveniles and non-reproductive adults). Scale bar = 5 m.
32
Figure 3. Analysis of spatial interaction between individuals in two populations of Zamia
portoricensis. Linhom, L transformation of inhomogeneous Ripley’s K-function (i.e. Kinhom). (A)
ET1 and (B) ET2, plots at El Tamarindo population; (C) SF1 and (D) SF2, plots at Susúa State
Forest population. In each chart, the solid black line shows test statistic, dashed straight red line
shows theoretical values of Kinhom(r) for CSR, and grey lines show significance envelopes for
α=0.01 estimated from 499 replicates of Monte Carlo simulation.
33
Figure 4. Relationship between reproductive success and sex ratio, males to females (M/F), in
Zamia portoricensis. Adjusted r2 = 0.08, F1,58 = 4.96, P < 0.03.
34
Figure 5. Relationship between reproductive success and distance of the female plant to the
nearest male in Zamia portoricensis. Adjusted r2 = 0.18, F1,83 = 19.78, P < 0.001. The dashed line
represent a threshold distance (i.e. 190 cm) after which the inverse relationship between
reproductive success and distance becomes significant.
35
CHAPTER TWO
SMALL REPRODUCTIVE GROUPS MAINTAIN HIGH INTRA-POPULATION GENETIC DIVERSITY IN
ZAMIA PORTORICENSIS
36
ABSTRACT
Genetic diversity is an indicator of the capacity for populations to cope with a changing
environment. Methods have been developed to provide more realistic characterization of
patterns of genetic structure, based on the assessment of departure from random spatial
distribution of genotypes as the result of limited gene dispersal and environmental
heterogeneity. Population genetic studies on the Puerto Rican zamias have revealed a pattern of
genetic variation characterized by high intra-population diversity, and low inter-population
differentiation. Several questions associated with this pattern are unanswered. For example, at
what scale is gene dispersal effective in these species; how many individuals are required to
maintain high levels of genetic variation within a population; and is this pattern evident across
different generations of individuals? To provide explanations grounded on the natural history of
the group I assessed whether Zamia portoricensis exhibits a significant pattern of spatial genetic
structure (SGS). I also investigated the scale at which genetic material is effectively exchanged
among individuals of the species, and also attempted to detect discontinuities in the spatial
distribution of genotypes. I studied two populations of Z. portoricensis established at the
eastern portion of the southern serpentine outcrop of Puerto Rico. I selected six polymorphic
microsatellite loci with no heterozygote deficiency, no linkage disequilibrium, no null alleles, and
a high Shannon’s Information Index for genotyping. I assessed the SGS, and estimated the extent
of gene dispersal for each data set using procedures implemented in the software SPAGeDI. I
identified discontinuities in the spatial distribution of genotypes using a method that performs
inferences from spatially referenced individuals, genotyped at a multilocus level, implemented
in the R package GENELAND. Gene dispersal estimates, 2.4 and 3.1 m, are congruent with
findings of limited pollination and limited seed dispersal in this species and suggest that gene
37
flow is very restricted. A significant departure from Hardy-Weinberg equilibrium due to a
deficiency in heterozygosity, while maintaining high levels of allelic diversity was detected. At
the local, fine-scale used in this study, Z. portoricensis exhibits a significant SGS, characterized by
identifiable discontinuities in the distribution of genotypes producing clusters with
configurations that match some topographical features of the habitat.
INTRODUCTION
High levels of genetic diversity has been found to be crucial for the long-term viability of
natural populations (Frankham, 2005). Genetic diversity is an indicator of the adaptive potential
of populations (Kramer and Havens, 2009), and for the last three decades has been widely used
to evaluate the conservation status of wild species (Groom et al., 2006).
Genetic diversity is generally expressed in terms of heterozygosity and allelic richness, which
are the basis for calculating other descriptive statistics that have a pervasive presence in
population genetics studies (Hamilton, 2009). However, inferences made from classic population
genetics models are based on assumptions, (i.e., large population sizes, random mating, no
population structure, no migration, etc.) that rarely match the life history of the taxa under
study (Storfer et al., 2007). During the last two decades, methods have been developed to
provide more realistic characterization of patterns of genetic structure based on the assessment
of departure from random spatial distribution of genotypes (i.e., spatial genetic structure, SGS)
as the result of limited gene dispersal (i.e., isolation by distance), and environmental
heterogeneity (i.e., isolation by resistance) (Vekemans and Hardy, 2004; Storfer et al., 2007,
2010; Guillot et al., 2012; Shirk and Cushman, 2014).
38
SGS can be quantified by statistics based on the relationship between pairwise relatedness
coefficients and the spatial distance between individuals (Vekemans and Hardy, 2004). Careful
selection of the molecular markers and the spatial dimension used to observe and quantify
processes is fundamental when carrying out spatially explicit genetic studies. Large-scale studies
can be useful to assess the results of major evolutionary events (Storfer et al., 2007, Segelbacher
et al., 2010; Guillot et al., 2012) but may fail to capture the effects of limited gene dispersal (i.e.
pollen and/or seed dispersal), life history traits (e.g., mating system, pollination system), local
demography (e.g., population structure and its dynamics), and habitat heterogeneity, that can
be more accurately assessed with fine-scale studies (Heuertz et al., 2003; Vekemans and Hardy,
2004; Born et al., 2008; Nazareno et al. 2013).
Population genetic studies on the Puerto Rican zamias have revealed a pattern of genetic
variation characterized by high intra-population diversity and low inter-population
differentiation (Meerow et al. 2012). This pattern is generally associated with a reproductive
scenario characterized by panmixia at intra-population level and considerable gene flow
between populations (Hamilton 2009). This is unexpected considering that these plants show a
pattern of differential reproduction from low to highly variable seed sets (Newell, 1983; Chapter
1), a decline in seed set when potential pollen donors are at distances longer than 1.9 m
(Chapter 1), and a pattern of limited seed dispersal (Negrón and Breckon, 1989b; Chapter 1)
suggesting that gene flow might be limited either by pollen or seed dispersal. Under these
circumstances, one expects to find small effective population sizes leading to genetic
homogenization through the effects of genetic drift (Hamilton 2009).
Several questions associated with the pattern of genetic structure of the Puerto Rican zamias
remain unanswered: (1) at what scale is gene dispersal effective in these species? (2) what is the
39
minimum number of individuals required to maintain high levels of genetic variation within a
population?, and (3) is the pattern of genetic variation evident across different generations of
individuals? To provide explanations grounded on the natural history of the group for the
patterns of genetic variation found by Meroow et al. (2012), I assessed whether Zamia
portoricensis, one of the species studied by Meerow et al., exhibits a significant pattern of SGS. I
also investigated the scale at which genetic material is effectively exchanged among individuals
of the species. Finally, I attempted to determine if genotypes are spatially arranged in a pattern
that reflects limited pollen and seed dispersal, by detecting and mapping discontinuities in the
spatial distribution of genotypes of Z. portoricensis.
The results presented here will be invaluable for designing effective conservation strategies
to promote self-sustainable, long-term viable populations of this endemic and threatened
species.
MATERIALS AND METHODS
Study species— Zamia portoricensis are short plants with thick, bending upright, simple to
sparsely branched, underground stems. It is endemic to the south-central and southwestern
serpentine outcrops and limestone zones of Puerto Rico (Acevedo-Rodríguez and Strong, 2005).
The plant is pollinated by Pharaxonotha portophylla (Coleptera: Erotylidae) (Franz and Skelley,
2008). The relationship between Pharaxonotha and Zamia involves a nursery, deception-based
pollination system, where the specialist beetles are involved in a strict mutualism with the host
plants (Norstog et al., 1992). The beetle’s life cycle is closely associated with male plants; they
feed on pollen and tissue of male cones, probably mate inside the cones, and use them as a
brood place. Attraction is olfactory so that pollination occurs by deception when pollen-laden
beetles visit female cones, which emit similar odors as males (Norstog and Nicholls, 1997). As
40
other West Indian zamias, Z. portoricensis is not only pollination limited (Newell, 1983, Tang
1987a, Chapter 1), but also has limited seed dispersal (Eckenwalder, 1980; Negrón-Ortiz and
Breckon, 1989a; Tang 1989, Chapter 1). In the most recent assessment of its conservation
status, Z. portoricensis was considered Endangered (EN) (Stevenson, 2010). Several field trips
between 2011 and 2015 to different localities across the island revealed that many populations
of the species only consist of adult plants, and exhibit neither sexual reproduction nor
recruitment likely due to extirpation of its pollinator (Chapter 3).
Study sites and sampling—My research sites were two previously studied populations of Z.
portoricensis (Chapter 1), both established at the eastern portion of the southern serpentine
outcrop of Puerto Rico. One population is located on both sides of Calle El Tamarindo (km 2.8 E
of intersection with route 368), Municipality of Sabana Grande (N18° 04.368' W66° 55.239'), and
the other population is located in the Susúa State Forest, along the dirt road between the main
Camp and the Cabins, Yauco County (N18° 03.972' W66° 54.076') (see Chapter 1 for detailed
descriptions of the populations). A total of 250 individuals, 125 per site, were sampled for the
study. All individuals were collected from two previously characterized 25  20 m plots,
hereafter called ET1 for the plot at El Tamarindo and SF2 for the plot at Susúa State Forest, for
which data on geographic position of each individual, population structure, and density was
known (Chapter 1). Plots were selected to represent the range of habitat occupied by the
species. Samples were collected using a clustered random sampling (Storfer et al., 2007) to
include all size classes (i.e. seedling and early juveniles, juveniles, and adults), different
generations (i.e. clear difference in the number of leaflets on the longest leaf), and, when
possible, both sexes. One leaflet per individual was collected and placed in silica gel for later
DNA extraction.
41
Genotyping— I extracted the DNA using a DNeasy® Plant Mini Kit (QIAGEN®, Inc.). I selected
six polymorphic microsatellite loci that have previously exhibited no heterozygote deficiency, no
linkage disequilibrium, no null alleles, and a high Shannon’s Information Index for genotyping:
Zam10, Zam16, Zam53, Zam59, Zfg02, and Zfg16 (Meerow et al., 2012). PCR reactions were
performed using the Taq PCR Master Mix Kit (QIAGEN®, Inc.) in a final reaction volume of 10 µL
(5 µL of 2x Taq PCR Master Mix, 0.4 µL of primer mix, 1 µL of BSA, 2.6 µL of H2O, and 1 µL of DNA
template). Primers for Zam10, Zam16 and Zam53 were marked with 5’-/56FAM/ and primers for
Zam59, Zfg10 and Zfg16 were marked with 5’-/5HEX/ (IDT®). PCR conditions were adapted from
those used by Meerow et al. (2012) and were set as follows: 2 min denaturation at 94 ⁰C
followed by 30 cycles of 30 s of initial denaturation at 94 ⁰C, 60 s of annealing at 50 ⁰C for
Zam10-Zam16, at 55 ⁰C for Zfg02-Zfg16, and at 60 ⁰C for Zam53-Zam59, 90 s of extension at 72
⁰C and 10 min of final elongation at 72 ⁰C. Amplifications were carried out in a PTC-100® Thermal
Cycler (Bio-Rad Laboratories, Inc.). PCR products from Zam10-Zam59, Zam16-Zfg02, and Zam53Zfg16 were combined for each individual sampled. One microliter of each PCR product, 0.25 µL
of GeneScanTM 500 ROXTM size standard, and 7.75 µL deionized formamide were pooled to
complete 10 µL and were analyzed on an Applied Biosystems 3130xl Genetic Analyzer. Scoring
was automated using GeneMapper version 4.0 software (Applied Biosystems) and allele sizing
was subsequently checked by hand. The total number of individuals genotyped was 247 since
DNA template of three individuals, one from ET1 and two from SF2, did not amplify with any
marker. Twenty-five percent of individuals had missing data for Zam59, 6% had missing data for
Zfg16, and 1.6% had missing data for Zam10; the other loci had a complete data set. Only 3.6%
of the individuals had missing data in two loci.
42
Genetic data analyses—I used GENEPOP 4.3 (Rousset, 2008) to obtain the basic information
on population genetics then I calculated the observed (Ho), and expected heterozygosity (He). I
tested for heterozygote deficiency using the Markov chain procedure (parameter values:
dememorization number = 10 000, number of batches = 300, number of iterations per batch =
5000). I calculated FIS values for each plot and tested their statistical significance using a
procedure based on the permutation of the distribution of alleles within plots.
Characterization of spatial genetic structure—I assessed the spatial genetic structure for
each data set using the procedure described in Vekemans and Hardy (2004), based on the
pairwise kinship coefficient between individuals, Fij, and implemented in the software SPAGeDI
1.5 (Hardy and Vekemans, 2002). Nason’s estimator of kinship coefficients (Loiselle et al., 1995)
was selected because of its robust performance even when rare alleles occur (Vekemans and
Hardy, 2004). Fij values were averaged over a set of distance classes (d) (i.e., 1.5, 3, 4.5, 6, 9, 12,
20 and 28 m) giving F(d), then Fij was regressed on the natural logarithm of the distance, ln(dij),
to obtain the regression slope bLd. Standard errors were obtained by jackknifing data over loci.
To visualize SGS, F(d) values were plotted against the set of spatial distance (d). To test for
statistical significance of SGS, the spatial positions of the individuals were permuted 9999 times
to obtain the frequency distribution of b under the null hypothesis that Fij and dij are not
correlated (cf. Mantel test). The amount of SGS was estimated by Sp = bLd / (F1−1) (Hardy et al.,
2006), where F1 is the average kinship coefficient between individuals separated by the first
distance class (i.e., d ≤ 1.5 m).
Estimation of gene dispersal—The extent of gene dispersal, σ, can be derived from estimates
of Wright’s neighborhood size, Nb = 4πDeσ2, where De is the effective density and σ2 is half the
mean squared axial parent-offspring distance. Wright’s neighborhood size can be estimated as
43
Nb = −(1− F1)/b, where b is the restricted regression slope of Fij on ln(dij) in the range σ > dij >
20σ, and σ can be estimated from Nb knowing De, σ = √Nb/(4πDe). An iterative procedure to
estimate Nb and σ is implemented in SPAGeDI 1.5 (Hardy and Vekemans, 2002) based on the
repeated estimation of Nb until the successive Nb estimates converge. For a fixed De lower and
upper bounds for 95% confidence interval (CI) of Nb were calculated as (F1− 1) / (b − 2SEb) and
(F1− 1) / (b + 2SEb) respectively, where SEb is the standard error of the b estimates obtained by
jackknifing over loci. The 95% CI of σ were obtained as √Nb/(4πDe) using the lower and upper
Nb bounds (Hardy et al., 2006).
Identification of discontinuities in the spatial arrangement of genotypes—To identify
discontinuities in the spatial distribution of genotypes I used a method that performs inferences
from spatially referenced individuals genotyped at a multilocus level. It uses the individual as the
operational unit of study without any a priori knowledge on the populational units and borders
(Guillot et al., 2012). The method uses a Bayesian approach for a spatial statistical model and
the inferences are performed via simulation of the posterior distribution of parameters by
Markov chain Monte Carlo (MCMC) techniques. The method is implemented in the R package
GENELAND version 4.0.0 (Guillot et al., 2005, 2012; R Development Core Team, 2015).
I first analyzed each set of data by processing 15 independent MCMC runs of the correlated
frequency model (CFM). Each run included 100 000 iterations of which every hundredth was
saved, with prior number of clusters, K, values between 1 and 15. This procedure allowed me to
detect the most frequent value of K for each data set among all simulated values. Then, with the
estimated values of K, I reanalyzed the data and selected a map of population membership as
the output option from a post-process number of pixels in the spatial domain of 250  200 and a
burn-in of 100.
44
RESULTS
Descriptive statistics of genetic structure—Both plots exhibited similar average number of
alleles (13 and 13.7), similar levels of heterozygosity (0.785 and 0.794), and significant
heterozygote deficiencies, as measured by FIS, although the value for ET1 was slightly higher
than the value for SF2 (Table 1).
Characterization of spatial genetic structure—In both plots the regression slope of the
kinship coefficient on the natural logarithm of the distance (bLd) was significant (Table 1) as
expected from an SGS pattern resulting from isolation by distance, and the relationship between
kinship coefficient and distance was very similar (Fig. 1). In ET1, as well as in SF2, there was a
dramatic drop in kinship coefficient values from the first distance class (1.5 m) to the second (3
m), also in both cases kinship values fluctuated between the second distance class and the sixth
(12 m) to experience a second, but less pronounced, descent after it (Fig. 1). The extent of SGS
was likewise very similar for both plots as reflected by the Sp estimates (Table 1).
Estimation of gene dispersal—Estimates of neighborhood size (Nb) and gene dispersal (σ,
the square root of half the mean squared axial parent-offspring distance) were very similar for
both plots but slightly smaller for ET1 (i.e. 104 individuals and 2.4 m) than for SF2 (i.e. 120
individuals and 3.1 m) (Table 2).
Identification of discontinuities in the spatial arrangement of genotypes—Fifteen
independent runs of the correlated frequency model showed that the most frequent number of
clusters (K) for both plots was ten. The number of individuals assigned to each cluster was
variable for both plots with SF2 showing smaller and more fragmented clusters than ET1 (Fig. 2).
45
DISCUSSION
Genetic diversity—Although I did not characterize the genetic structure of Z. portoricensis
throughout its range, which would have required a different sampling scheme, I compared the
results presented here with those previously reported for the same loci in this species but
assessed at a larger scale (e.g., three populations) (Meerow et al., 2012 Appendix S1). The mean
number of alleles registered here is not different from the value reported by Meerow et al., 13
vs 12; the observed heterozygosity was just slightly smaller, 0.785 and 0.794 vs 0.849, but the
difference in the fixation index (FIS) values was notably higher in this study, 0.060 and 0.044 vs
−0.032. I attribute this dissimilarity to differences in the sampling regimes. While I sampled 123
and 124 individuals, Meerow et al. sampled 97; however my sampling was conducted at a finer
geographical scale (i.e., 30 m was the longest distance between any pair of individuals). Meerow
et al. sampled fewer individuals per site (e.g., 39, 28, and 30) but they were well separated
(several meters apart) and the values they report are pooled from three different populations.
Certainly the inclusion of several relatives in my study influenced heterozygosity values. Despite
significant heterozygote deficiency in the plots I studied, allelic diversity was the same as
reported by Meerow et al. This is relevant because the allelic diversity is more sensitive to
stochastic processes associated with small effective population sizes than heterozygosity (Nei et
al., 1975; Eckert et al., 2008).
SGS and gene dispersal—Mating system, life form, pollen and seed dispersal vectors, and
spatial distributions of individuals are attributes that, combined with habitat features, influence
gene flow in plants (Wright, 1943; Chung and Epperson, 2000; Dutech et al., 2002; Fenster et al.,
2003; Heuertz et al., 2003; Vekemans and Hardy, 2004; Hardy et al., 2006; Born et al., 2008;
López-Gallego and O’Neil, 2009; Debout et al., 2011).
46
Vekemans and Hardy’s (2004) meta-analysis of SGS studies on 48 species revealed that
mating system (i.e., selfing, mixed mating, outcrossing, self-incompatible) and life form (i.e.,
herbaceous, small trees, and trees) affect SGS with 1) significantly higher Sp values for selfing
and herbaceous species, 2) significantly lower Sp values for outcrossing, self-incompatible and
trees species, and 3) intermediate values that nonetheless were significantly different from the
above mentioned systems for mixed-mating and small trees. Pollen dispersal vectored by
animals or wind and seed dispersal via gravity, wind or animals did not have significant effects
on SGS although they found that SGS had a tendency to be stronger for species with animaldispersed pollen and gravity-dispersed seeds. Zamia portoricensis is a dioecious plant with
limited pollen dispersal by specialist insects (Chapter 1) and currently limited seed dispersal by
gravity (Chapter 1) a combination of traits that was not well represented in the study by
Vekemans and Hardy (2004). However, they studied 8 species with a very similar combination of
traits (i.e., self-incompatible, short, trunkless plants, animal-dispersed pollen, and gravitydispersed seeds) allowing some comparison of our results. Mean Sp for those species was 0.013,
which is similar to my Sp estimates of Z. portoricensis (Table 1). In contrast, Z. fairchildiana
exhibited higher Sp values for subpopulations in an undisturbed forest (Sp = 0.015) and a lower
value (Sp = 0.008) for subpopulations in a disturbed forest (López-Gallego and O’Neil, 2010). The
main difference between the two forest types was higher values of canopy openness in the
disturbed habitat; in the present study, the differences in canopy openness between plots (i.e.,
ET1 > SF2) did not result in a difference of Sp values as large as that reported by López-Gallego
and O’Neil (2010).
The values obtained here for gene dispersal (σ), 2.4 m and 3.1 m, are congruent with
previous findings that pollination success in Z. portoricensis is dramatically reduced at distances
47
larger than 1.6 m between potential mates (Chapter 1). As gene dispersal is inversely related to
effective density (De), which was slightly higher in ET1 (Table 2), I expected this plot to exhibit
shorter gene dispersal distances because the distance between potential mates is shorter than
in SF2. Vekemans and Hardy (2004) and Theim et al. (2014) found this relationship for species
with different life forms, breeding systems, and pollination syndromes, and both invoked
pollinator behavior as a possible explanation (Vekemans and Hardy, 2004; Theim et al., 2014).
The neighborhood size estimates reported in this study, 104 and 120 (Table 2), must be
carefully interpreted; Nb, Wright’s neighborhood size (i.e. genetic neighborhood), is defined as
the number of breeding individuals in the local neighborhood in a continuous population under
isolation by distance. That is, within this neighborhood the effect of gene flow is high relative to
drift (Wright, 1946). At a theoretical level it is a concept easy to comprehend but its estimates
from field studies have been difficult to interpret (Fenster et al., 2003). In species with
contrasting biological traits, e.g. Adansonia digitata, a mixed mating tree with bat-dispersed
pollen and animal-dispersed seeds, and Primula elatior, an outcrossing herb with insectdispersed pollen and gravity-dispersed seeds, relatively similar values (i.e., 45 and 55
respectively) have been reported (Van Rossum and Triest, 2006; Kyndt et al. 2009). Meanwhile
species with similar biological traits, can have dissimilar Nb values. Fraxinus excelsior and Pinus
strobus are outcrossing trees with wind-dispersed pollen and seeds, yet the former has Nb = 55
and the latter has Nb = 100 (Heuertz et al., 2003; Walter and Epperson, 2004). It seems that Nb
estimates are strongly context dependent (Shirk and Cushman, 2014). Considering that Z.
portoricensis is a dioecious plant with very limited pollen and seed dispersal, and that
reproductive success is inversely related to the distance to the nearest mate (Chapter 1), within
48
the neighborhood sizes estimated here, at least two clusters of relatively inbred individuals
could be accommodated resulting in a departure from the original expectation of panmixis.
Spatial arrangement of genotypes—The choice of a model to assess discontinuities in the
spatial distribution of genotypes has been a subject of discussion (Guillot et al., 2012). In some
cases there is a considerable difference between the number of genetic clusters identified by
the uncorrelated frequency model (UFM) and the correlated frequency model, with CFM
identifying larger values of K than UFM. Guillot et al. (2012) suggested that the selection of
either model should be case specific. As pollen and seed dispersal is limited in Z. portoricensis,
evidenced here by gene dispersal estimates of 2.4 and 3.1 m, it is highly probable that present
time clusters originated from the split of an ancestral cluster several generations ago so that the
use of the CFM is justified (Falush et al., 2003; Guillot, 2008; Guillot et al., 2012). Geneland has
been used at scales where genetic clusters have been associated with operational units such as
species or subspecies and demographic populations or subpopulations (Born et al., 2008; Joseph
et al., 2008), which are larger than the scales assessed here but this does not alter the results
(i.e., a set of identifiable clusters based on genetic similarity).
Groups of genotypes at SF2 are more dispersed than at ET1, which is congruent with the
gene dispersal estimates for each plot (Table 2; Fig. 2). The position of the clusters within each
plot (Fig. 2) match some physical characteristics of their microhabitat. For example, ET1 is
located at a steep slope and when seeds are released from female cones some of them may be
dispersed downward from south to north at a distance that can reach a few meters under heavy
rain, so a vertical structure may emerge from this dispersal. At the same time, rocks and the
base of trees and shrubs are barriers that limit downward dispersal resulting in seed
accumulation and horizontal structure. Plot SF2, is located on flat terrain and has a diagonal
49
drain of an ephemeral stream running from the southwest side to the northeast side of the plot
but under heavy rain water moves like a front from south to north that might explain why
similar clusters are located at diagonal extremes of it (Fig. 2).
Patterns of genetic differentiation in Z. portoricensis—Meerow et al. (2012) proposed that
high intra-population differentiation in Puerto Rican Zamia could be attributed to a pattern
typical of microsatellite marker data sets, retention of genetic variation in long-lived species
with long generation times, high rates of mutation and drift after new populations are founded,
and the relatively short span of time that Puerto Rico has been above sea level. The first two are
plausible explanations but the third and the fourth are not. Microsatellites are neutral,
codominant markers, and their polymorphism mainly results from variability in length rather
than in the primary sequence; thus, it is common to find many loci that are characterized by
high heterozygosity and multiple alleles (Ellegren, 2004). However, the populations of Z.
portoricensis sampled by Meerow and collaborators (Susúa State Forest and adjacent areas, and
Guánica State Forest) contain thousands of individuals and each cover about two kilometers in
range (Chapter 3) suggesting that they are very old populations, thus any pattern associated to
founder effect should be currently erased. Furthermore, the relatively short span of time that
Puerto Rico has been above sea level (5 Ma BP) cited by Meerow et al. has been challenged by
other authors (MacPhee and Iturralde-Vinent, 2005; Hedges, 2006; Iturralde-Vinent, 2006) who
suggest that the island has been above sea level for approximately 40 Ma.
In addition to the previously discussed explanations, I suggest that the patterns of genetic
structure found in Z. portoricensis can be explained by constraints to gene flow as the result of
limited pollen and seed dispersal, and by the effects of stochastic processes, like the random
spatial distribution of individuals, on reproductive success (Chapter 1). Both may prevent
50
genetic homogenization from occurring, the first by producing relatively isolated clusters of
individuals with similar genotypes, and the second by reducing the probability for particular
alleles to become fixed at the population level as a result of high variability in reproductive
success. In this way inter-population differentiation is constrained but allelic diversity is
preserved within populations. This scenario is plausible for species that exhibit very limited
pollen and seed dispersal, and that are not undergoing a rapid evolutionary change, as is the
case for cycads (Salas-Leiva et al., 2013).
Conclusions—At the small scale sampled in this study Z. portoricensis exhibits a significant
deficiency in heterozygosity, while maintaining the same high levels of allelic diversity as has
been reported across multiple populations.
Gene dispersal estimates, 2.4 and 3.1 m, are congruent with findings of limited pollination
and limited seed dispersal in this species and suggest that gene flow is very restricted. Estimates
of neighborhood size of 104 and 120 individuals should be carefully interpreted. As
neighborhood size is directly proportional to the effective density, a relative high abundance of
reproductive adults result in high values, but in Z. portoricensis, reproductive success is
influenced by the spatial distribution of individuals (Chapter 1), and certainly not all the
reproductively active adults are contributing to recruitment, which is far from the expectation of
panmixia in Wright’s neighborhood model.
At the local, fine-scale used in this study, Z. portoricensis exhibits significant spatial genetic
structure, characterized by identifiable discontinuities in the distribution of genotypes producing
clusters with configurations that match some topographical features of the habitat.
51
At El Tamarindo and Susúa State Forest, Z. portoricensis has been able to maintain an
adequate population structure (Chapter 1) and high levels of genetic diversity; both populations
can be used as models to set management goals for future conservation programs.
52
LITERATURE CITED
ACEVEDO-RODRÍGUEZ, P., AND M.T. STRONG. 2005. Monocotyledons and Gymnosperms of Puerto
Rico and the Virgin Islands. Contributions from the United States National Herbarium
Volume 52: 1–415.
BORN, C., O.J. HARDY , M-H. CHEVALLIER , S. OSSARI, C. ATTÉKÉ, E.J. WICKINGS, AND M. HOSSAERT MCKEY. 2008. Small-scale spatial genetic structure in the Central African rainforest tree
species Aucoumea klaineana: a stepwise approach to infer the impact of limited gene
dispersal, population history and habitat fragmentation. Molecular Ecology 17: 2041–2050.
CHUNG, M.G., AND B.K. EPPERSON. 2000. Clonal and spatial genetic structure in Eurya emarginata
(Theaceae). Heredity 84: 170–177.
DEBOUT , G.D.G., J-L. DOUCET, AND O.J. HARDY. 2011. Population history and gene dispersal
inferred from spatial genetic structure of a Central African timber tree, Distemonanthus
benthamianus (Caesalpinioideae). Heredity 106: 88–99.
DUTECH, C., J. SEITER , H.I. PETRONELLI, AND P. JARNE . 2002. Evidence of low gene flow in a
neotropical clustered tree species in two rainforest stands of French Guiana. Molecular
Ecology 11: 725–738.
ECKENWALDER , J.D. 1980. Dispersal of the West Indian Cycad, Zamia pumila L. Biotropica 12: 79–
80.
ECKERT , C.G., K.E. SAMIS, AND S.C. LOUGHEED. 2008. Genetic variation across species’ geographical
ranges: the central-marginal hypothesis and beyond. Molecular Ecology 17: 1170–1188.
ELLEGREN, H. 2004. Microsatellites: simple sequences with complex evolution. Nature Reviews
Genetics 5: 435–445.
53
FALUSH, D., M. STEPHENS, AND J. PRITCHARD. 2003. Inference of population structure using
multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164: 1567–
1587.
FENSTER, C.B., X. VEKEMANS, AND O.J. HARDY. 2003. Quantifying gene flow from spatial genetic
structure data in a metapopulation of Chamaecrista fasciculata (Leguminosae). Evolution
57: 995–1007.
FRANKHAM, R. 2005. Genetics and extinction. Biological Conservation 126: 131-140.
FRANZ, N.M., AND P.E. SKELLEY . 2008. Pharaxonotha portophylla (Coleptera: Erotylidae), new
species and pollinator of Zamia (Zamiaceae) in Puerto Rico. Caribbean Journal of Science
44: 321–333.
GROOM, M.J., G.K. MEFFE, AND C.R. C ARROLL . 2006. Principles of conservation biology, 3rd ed.
Sinauer Associates, Inc. Sunderland, MA.
GUILLOT , G. 2008. Inference of structure in subdivided populations at low levels of genetic
differentiation. The correlated allele frequencies model revisited. Bioinformatics 24: 2222–
2228.
GUILLOT , G., F. MORTIER, AND A. ESTOUP. 2005. GENELAND: a computer package for landscape
genetics. Molecular Ecology Notes 5: 712–715.
GUILLOT , G., S. RENAUD, R. LEDEVIN, J. MICHAUX, AND J. CLAUDE. 2012. A unifying model for the
analysis of phenotypic, genetic, and geographic data. Systematic Biology 61: 897–911.
HAMILTON, M.B. 2009. Population Genetics. Wiley-Blackwell, Chichester, West Sussex, UK.
HARDY, O.J., AND X. VEKEMANS. 2002. SPAGeDi: a versatile computer program to analyze spatial
genetic structure at the individual or population levels. Molecular Ecology Notes 2: 618–
620.
54
HARDY, O.J., L. MAGGIA, E. BANDOU, P. BREYNE, H. CARON, M-H. CHEVALLIER , A, DOLIGEZ, ET AL .
2006. Fine-scale genetic structure and gene dispersal inferences in 10 Neotropical tree
species. Molecular Ecology 15: 559–571.
HEDGES, S.B. 2006. Paleogeography of the Antilles and origin of the West Indian terrestrial
vertebrates. Annals of the Missouri Botanical Garden 93: 231–244.
HEUERTZ, M., X. VEKEMANS, J-F. HAUSMAN, M. PALADA, AND O.J. HARDY. 2003. Estimating seed
versus pollen dispersal from spatial genetic structure in the common ash. Molecular
Ecology 12: 2483–2495.
ITURRALDE -VINENT, M.A. 2006. Meso-Cenozoic Caribbean Paleogeography: implications for the
historical biogeography of the region. International Geology Review 48: 791–827.
JOSEPH, L., G. DOLMAN, S. DONNELLAN, K. M. SAINT , M. L. BERG, AND A.T. BENNETT . 2008. Where
and when does a ring start and end? Testing the ring-species hypothesis in a species
complex of Australian parrots. Proceedings of the Royal Society of London, B, Biological
Sciences 275: 2431–2440.
KRAMER, A.T. AND K. HAVENS. 2009. Plant conservation genetics in a changing world. Trends in
Plant Science 14: 599–607.
KYNDT, T., A.E. ASSOGBADJO , O.J. HARDY, R.G. KAKAÏ, B. SINSIN, P. VAN DAMME , AND G. GHEYSEN.
2009. Spatial genetic structuring of Baobab (Adansonia digitata, Malvaceae) in the
traditional agroforestry system of West Africa. American Journal of Botany 96(5): 950–957.
LOISELLE , B.A., V.L. SORK, J. NASON, AND C. GRAHAM. 1995. Spatial genetic structure of a tropical
understory shrub, Psychotria officinalis (Rubiaceae). American Journal of Botany 82: 1420–
1425.
55
LÓPEZ-GALLEGO, C., AND P. O’NEIL. 2009. Life-history variation following habitat degradation
associated with differing fine-scale spatial genetic structure in a rainforest cycad.
Population Ecology 52: 191–201.
MEEROW, A., J. FRANCISCO -ORTEGA, M. CALONJE , M.P. GRIFFITH, T. AYALA-SILVA, D.W. STEVENSON,
AND K. N AKAMURA . 2012. Zamia (Cycadales: Zamiaceae) on Puerto Rico. Asymmetric genetic
differentiation and the hypothesis of multiple introductions. American Journal of Botany
99: 1828–1839.
MACPHEE, R.D.E. AND M. ITURRALDE-VINENT. 2005. The Interpretation of Caribbean
Paleogeography: Reply to Hedges. In ALCOVER, J.A. and BOVER, P. (eds.): Proceedings of
the International Symposium “Insular Vertebrate Evolution: the Palaeontological
Approach”. Monografies de la Societat d’Història Natural de les Balears, 12: 175-184.
NAZARENO , A.G., A.L. ALZATE -MARIN, AND R.A.S. PEREIRA. 2013. Dioecy, more than monoecy,
affects plant spatial genetic structure: the case study of Ficus. Ecology and Evolution 3:
3495–3508.
NEI, M., T. MARUYAMA, AND R. CHAKRABORTY . 1975. The bottleneck effect and genetic variability
in populations. Evolution 29: 1–10.
NEGRÓN-ORTIZ, V., AND G.J. BRECKON. 1989a. A Note on the dispersal of Zamia (Zamiaceae) in
Puerto Rico. Caribbean Journal of Science 25: 86–87.
NEWELL , S.J. 1983. Reproduction on a natural population of cycads (Zamia pumila L.) in Puerto
Rico. Bulletin of the Torrey Botanical Club 110: 464–473.
NORSTOG, K.J., P.K.S. FAWCETT , AND A.P. VOVIDES. 1992. Beetle pollination of two species of
Zamia: Evolutionary and ecological considerations. The Palaeobotanist 41: 149–158.
NORSTOG, K.J., AND T.J. NICHOLLS. 1997. The Biology of the Cycads. Cornell University Press,
Ithaca, NY.
56
R DEVELOPMENT CORE TEAM. 2015. R: A language and environment for statistical computing.
Version 3.1.3. The R Foundation for Statistical Computing, Vienna, Austria.
ROUSSET , F. 2008. GENEPOP’007: a complete re-implementation of the GENEPOP software for
Windows and Linux. Molecular Ecology Resources 8: 103–106.
ROUSSET , F., AND M. RAYMOND. 1995. Testing heterozygote excess and deficiency. Genetics 140:
1413–1419.
SALAS-LEIVA, D.E., A.W. MEEROW, M. CALONJE , M.P. GRIFFITH, J. FRANCISCO -ORTEGA, K. NAKAMURA,
D.W. STEVENSON, ET AL . 2013. Phylogeny of the cycads based on multiple single-copy
nuclear genes: congruence of concatenated parsimony, likelihood and species tree
inference methods. Annals of Botany 112: 1263–1278.
SEGELBACHER , G., S.A. CUSHMAN, B.K. EPPERSON, M.J. F ORTIN, O. FRANCOIS, O.J. HARDY, R.
HOLDEREGGER , ET AL . 2010. Applications of landscape genetics in conservation biology:
concepts and challenges. Conservation Genetics 11: 375–385.
SHIRK, A.J., AND S.A. CUSHMAN. 2014. Spatially-explicit estimation of Wright’s neighborhood size
in continuous populations. Frontiers in Ecology and Evolution 2: 62.
STEVENSON, D.W. 2010. Zamia erosa, Z. portoricensis, Z. pumila. In IUCN 2013. IUCN Red List of
Threatened Species. Version 2013.2. Website http://www.iucnredlist.org [accessed on 10
February 2014].
STORFER, A., M.A. MURPHY , J.S. EVANS, C.S. GOLDBERG, S. ROBINSON, S.F. SPEAR, R. DEZZANI, ET AL .
2010. Putting the ‘landscape’ in landscape genetics. Heredity 98: 128–142.
STORFER, A., M.A. MURPHY , S.F. SPEAR, R. HOLDEREGGER , AND L.P. WAITS. 2010. Landscape
genetics: where are we now? Molecular Ecology 19: 3496–3514.
TANG, W. 1987a. Insect Pollination in the Cycad Zamia pumila (Zamiaceae). American Journal of
Botany 74: 90–99.
57
TANG, W. 1989. Seed dispersal in the cycad Zamia pumila in Florida. Canadian Journal of Botany
67: 2066–2070.
THEIM, T.J., R.Y. SHIRK, AND T.J. GIVNISH. 2014. Spatial genetic structure in four understory
Psychotria species (Rubiaceae) and implications for tropical forest diversity. American
Journal of Botany 101: 1189–1199.
VAN ROSSUM, F., AND L. TRIEST . 2006. Fine-scale genetic structure of the common Primula elatior
(Primulaceae) at an early stage of population fragmentation. American Journal of Botany
93: 1281–1288.
VEKEMANS, X., AND O.J. HARDY. 2004. New insights from fine-scale spatial genetic structure
analyses in plant populations. Molecular Ecology 13: 921–935.
WALTER, R., AND B.K. EPPERSON. 2004. Microsatellite analysis of spatial structure among seedlings
in populations of Pinus strobus (Pinaceae). American Journal of Botany 91: 549–557.
WESTENGEN, O.T., M.A. OKONGO, L. ONEK, T. BERG, H. UPADHYAYA, S. BIRLELAND, S.D.K. KHALSA, ET
AL .
2014. Ethnolinguistic structuring of sorghum genetic diversity in Africa and the role of
local seed systems. Proceedings of the National Academy of Sciences 111: 14100–14105.
WRIGHT , S. 1943. Isolation by distance. Genetics 28: 114–138.
WRIGHT , S. 1946. Isolation by distance under diverse systems of mating. Genetics 31: 39–59.
58
CHAPTER TWO
TABLES
59
Table 1. Descriptive statistics of population genetics and SGS for each plot. N = sample size; Na =
number of alleles; Ho, He = observed and expected heterozygosity; FIS = fixation index; F1 =
average kinship coefficient between individuals separated by 1.5 m or less; bLd = slope of the
regression of kinship coefficient values on the logarithm of the spatial distance between
individuals; Sp = amount of SGS; SE = standard error for each statistic. Values without SE are not
means.
Plot
N
ET1
124
SF2
123
Na
Ho
He
FIS1
F1
bLd 2
Sp
Mean
13
0.785
0.834
0.060***
0.029
-0.011***
0.012
SE
3.29
0.023
0.011
0.006
0.002
0.005
Mean
13.17
0.794
0.830
0.042
-0.014***
0.018
SE
3.43
0.032
0.021
0.011
0.003
0.014
0.044*
Notes:
1
Significant departure from HWE based on U test for heterozygote deficiency (Rousset and
Raymond, 1995). P values: NS for P > 0.05; * for 0.05 ≥ P > 0.01; ** for 0.01 ≥ P > 0.001; *** for
P ≤ 0.001.
2
P values: NS for P > 0.05; * for 0.05 ≥ P > 0.01; ** for 0.01 ≥ P > 0.001; *** for P ≤ 0.001.
60
Table 2. Estimates of two gene dispersal parameters for each plot. De = effective density;
Nb = neighborhood size; σ = extent of gene dispersal (m); CI = 95% confidence interval.
Plot
De
Nb (CI)
σ (CI)
ET1
1.4
104 (66-135)
2.4 (1.9-2.8)
SF2
1.0
120 (48-128)
3.1 (2.0-3.2)
61
CHAPTER TWO
FIGURES
62
Figure 1. Average kinship coefficient, F(d), plotted against logarithm of spatial distance, ln(d),
between individuals for each plot. ET1 = blue line; SF2 = red line.
0.05
0.04
0.03
F(d)
0.02
0.01
0
-0.01
-0.02
-0.03
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
Distance (m)
Figure 2. Estimated cluster membership of individuals for each plot. (A) ET1; (B) SF2. Colored
ellipses indicate membership to the same group of genotypes. Scale bar = 5 m.
63
CHAPTER THREE
CONSERVATION STATUS OF ZAMIA IN PUERTO RICO
64
ABSTRACT
Cycads are an ancient lineage with only ten extant genera distributed across the tropics and
subtropics. Zamia is the only cycad genus widely distributed in the West Indies. In Puerto Rico
there are three species: Z. erosa, the endemic Z. portoricensis, and Z. pumila. In the most recent
conservation assessment of Puerto Rican zamias using the International Union for the
Conservation of Nature (IUCN) criteria, Z. portoricensis was considered Endangered, Z. erosa
Vulnerable, and Z. pumila Near Threatened. Thus far, there is no official recognition of the
threatened status of Zamia in Puerto Rico, and none of its species are included in the United
States Fish and Wildlife Service regional list of endangered and threatened species. In this study
I produce the first conservation assessment of the genus in Puerto Rico based on plant density,
population structure, and habitat range. I located populations using herbarium records accessed
online, available literature, data from interviews with persons with field experience on the
island, and Google Earth to identify suitable but unexplored habitats. For each species I
estimated the extent of occurrence, the area of occupancy, and the number of mature
individuals by using polygons in Google Earth, and randomly set 25  20 m plots in which I
recorded vegetative and reproductive data of every individual. I carried out the Red List
assessment using IUCN Red List Categories and Criteria Version 3.1, second edition. New
categories are proposed for the three species assessed. Zamia erosa should be considered Least
Concern, Z. portoricensis should be regarded as Vulnerable, and Z. pumila should be considered
Critically Endangered in Puerto Rico. Habitat destruction and a past eradication campaign are
likely major factors that drove Z. portoricensis and Z. pumila to a threatened status since extant
populations, at undisturbed sites, are healthy, well-structured, and self-sustainable. I provide
baseline information that can be used in the future to monitor the conservation status of the
65
group, and I recommend six critical actions that should be taken to provide this extraordinary
group of plants with an opportunity for long-term persistence.
INTRODUCTION
Extant cycads belong to a lineage that has existed for about 300 million years (Chamberlain,
1919, 1935). As the most ancient seed plants living today (Norstog and Nicholls, 1997) they
provide unique opportunities to understand fundamental processes that have led to the
evolution of the seed and the early origins of animal-plant interactions (Klavins et al., 2005).
During the last two decades, and because of their prehistoric beauty, these plants have gained
popularity among garden and landscape designers, horticulturists, and private collectors (Jones,
2002). In addition, they have also become relevant to medical and veterinary research because
of their link to neurodegenerative diseases that affect humans and cattle (Reams et al., 1993;
Cox et al., 2007; Shaw et al., 2007).
Cycads are dioecious, insect-pollinated gymnosperms that exhibit two general types of plantpollinator interactions, an obligate mutualism that involves specialist insects that feed on pollen,
or male cone tissue, and use male cones as a brood place (Tang, 1987; Norstog et al., 1992;
Donaldson et al., 1995; Mound and Terry, 2001; Wilson, 2002; Hall et al., 2004), and a
facultative mutualism that involves generalist insects that feed on pollen but do not use cycad
cones as brooding sites (Kono and Tobe, 2007; Proches and Johnson, 2009). It is clear that the
long-term survival of the plants is linked to the presence of their pollinators, and that the longterm survival of the pollinators is linked to the persistence of their hosts, at least for the first
type of mutualism which is the most common and the one found in Zamia.
Cycads are very resilient due to a suite of morphological, anatomical and life history traits,
which include: (1) aerial or subterranean stems able to survive periodic fires; (2) stems with
66
starch reserves that allow individuals to enter dormancy for more than two years; (3)
mycorrhizae, and specialized roots that host cyanobacteria able to actively fix atmospheric
nitrogen; (4) salt tolerance; (5) production and storage of secondary metabolites that provide
chemical defense against herbivores, (6) tough and leathery leaves that provide mechanical
defense; (7) longevity, living at least several decades sometimes reaching hundreds of years, and
(8) asexual reproduction (Chamberlain 1919, 1935; Newell, 1983; Vovides, 1990; Norstog and
Nicholls, 1997).
There are ten living cycad genera distributed across the tropics and subtropics (Haynes,
2011). Zamia, with about 70 species, is one of the four genera present in the New World, and is
also the only cycad genus widely distributed in the West Indies. Despite the problematic
interspecific taxonomy of Zamia, West Indian species form a distinct, monophyletic group
known as the Zamia pumila complex (Norstog and Nicholls, 1997; Caputo et al., 2004).
Unfortunately, there is no consensus about the number of species assigned to this group, and it
ranges from a single polymorphic species (Eckenwalder, 1980) to as many as nine species
(Stevenson, 1987; González Géigel, 2003). Nevertheless, the more recent studies involving
Puerto Rican zamias are congruent, recognizing three species reported for Puerto Rico: Z. erosa,
distributed across the northern limestone region; Z. portoricensis, the only endemic, inhabiting
the south-central and south-western limestone regions and serpentine outcrops; and Z. pumila,
distributed in the south-central limestone region with few individuals in volcanic substrates
(Acevedo-Rodríguez & Strong, 2005; Axelrod, 2011; Meerow et al., 2012; Pagán-Jiménez and
Lazcano-Lara, 2013).
The latest conservation assessment of the Puerto Rican zamias proposed that the endemic Z.
portoricensis is Endangered (EN), while Z. erosa is Vulnerable (VU), and Z. pumila is Near
67
Threatened (NT) (Stevenson, 2010). In all cases, the assessment process was mainly based on
rough estimates of population size and population decline. Thus far, there is no official
recognition of the threatened status of Zamia in Puerto Rico; none of its species are included in
the United States Fish and Wildlife Service (2015) regional list of endangered and threatened
species, and consequently none are protected under the Endangered Species Act (USFWS,
2015).
Like many other species, zamias have been severely affected by habitat destruction. The
landscape of southern Puerto Rico has been altered for many years by the establishment of
pastures, quarries, crop fields, roads, urbanizations, and landfills. In addition to habitat
destruction all species of Zamia were subjected to an eradication campaign in the 1960’s
promoted by the Department of Agriculture (Hernández de Rivera, 1966) in response to an
increase in cases of cattle poisoning after the ingestion of cycad leaves (Reams et al., 1993).
In this study I produce the first assessment of the conservation status of Zamia in Puerto Rico
based on quantitative information on plant density, population structure, and habitat range. I
also provide baseline information that may be used to monitor the conservation status of the
group, and recommend six critical actions that should be taken to provide this extraordinary
group of plants with an opportunity for long-term persistence.
MATERIALS AND METHODS
Species distribution and population structure—To locate populations of Zamia in Puerto
Rico I used two different sources for known occurrences, at least in the past: herbarium records
accessed online (i.e., UPRRP and MAPR, 2010-13), and available literature. I used two other
sources to identify existing locations and others with suitable habitat but with previously
68
unrecorded occurrences: interviews with persons with field experience on the island (i.e.,
Alberto Areces, interviewed in 2009; Carlos Pacheco in 2009; James D. Ackerman in 2010;
Franklin S. Axelrod in 2010; Jaime R. Pagán-Jiménez in 2011; Omar Monsegur in 2012; Marcos
Caraballo in 2014) and Google Earth.
With a list of actual and potential localities I carried out several field trips and was able to
verify the presence of Zamia in 36 localities (Table 2). At 22 of these populations (i.e., 8 Z. erosa,
6 Z. portoricensis, and 8 Z. pumila) I randomly set a 25  20 m plot, and in each plot, I recorded
plant size, measured as the number of leaflets of the longest leaf and, when possible, the sex of
every individual. To describe population structure, I considered the division line between
juvenile and adult size classes as the minimum size at which a plant produced a cone. Within the
adult class, I noted the sex of coning plants, and counted the remainder as non-reproductive.
Additionally I georeferenced the approximate limits of each population to estimate the area
occupied by the species at each of the 36 verified localities.
Estimation of Extent of Occurrence (EOO)—For each species I created polygons in Google
Earth to calculate the area encompassing all the actual or potential sites where each species can
be present, while excluding non-suitable habitats, such as urban and agricultural areas (IUCN,
2012).
Estimation of Area of Occupancy (AOO)—Based on the stem diameter and the fact that
some adults have branched stems, I assigned an approximate area of occupancy of 0.16 m2 for
an individual adult and of 0.04 m2 for an individual seedling or juvenile. I calculated the area of
occupancy of the species in each 500 m2 plot, AOOp = Na*0.16 + Nsj*0.04, where, Na is the
number of adults and Nsj is the number of seedlings and juveniles. I then extrapolated the AOOp
to the range of the species at each locality using the following formula AOOl = (AOOp*At)/500,
69
where AOOl is the area of occupancy of the species at each locality, and At is the range of the
species at each locality in square meters. For the 14 localities of Z. erosa where I did not set a
plot I calculated AOOl by applying a conservative approach using the data of the plot with the
second lowest values of Na and Nsj. The AOO for each species was the sum of all the values per
locality.
Estimation of number of mature individuals (NM)—To estimate the number of mature
individuals I used two different NM values, NMsc which is the number of adult plants obtained
from vegetative data (i.e., size class) including those plants that were not producing cones at the
time of the study, and NMcon which is the number of adults that were actually showing
reproductive structures during the survey. Following the Red List guidelines, at those localities
where there is no recruitment, NMsc and NMcon were considered null (IUCN, 2012). At each
plot, NMsc was the sum of males + females + non-reproductive adults, and NMcon was the sum
of males + females. The values per plot (i.e., 500 m2) were extrapolated to the range of the
species at each locality. For the 14 localities of Z. erosa, where I did not set a plot I calculated
NMsc, and NMcon by applying a conservative approach using the data of the plot with the
second lowest values of NMsc and NMcon (i.e., Candelaria). The number of mature individuals
for each species was the sum of all the values per locality.
Categorization—I carried out the Red List assessment process following the instructions
established in the International Union for the Conservation of Nature (IUCN) Red List Categories
and Criteria Version 3.1, second edition (IUCN, 2012). Considering that Puerto Rico is a small
archipelago with an area of about 9 106 km2, the criterion B (i.e., geographic range) was
carefully used to support the assessment but no category was assigned using it as the main
criterion for categorization.
70
RESULTS
General status of Zamia in Puerto Rico—Of the three species in the region, Zamia erosa has
the widest distribution both in terms of range and number of localities (Fig. 1). It is also the most
abundant species (Table 1). At some locations it is so numerous that it dominates the
understory perhaps competitively excluding other vascular plants from becoming stablished
(Fig. 2A, B). This species also recruits on a regular basis as plants at 20 of the 22 localities visited
during this study produced seeds and had seedlings (Fig. 2C, D).
Zamia portoricensis is also an abundant species (Table 1; Fig. 3A) but it is not as widely
distributed as Z. erosa (Fig. 1). Most Z. portoricensis are reproductively active individuals located
on the eastern portion of the southern serpentine outcrop of Puerto Rico (Fig. 3B, C). The
southern limestone region between the municipalities of Guánica and Yauco are also host to a
large number of plants but most of them are adults and recruitment is very limited in this area
(Table 1; Fig. 3D, E).
Zamia pumila is the rarest native cycad in Puerto Rico (Table 1). Although it has been
recorded at more places than Z. portoricensis (Fig. 1), most Z. pumila populations consist of a
small number of adults that do not exhibit sexual reproduction (Fig. 4).
All the individuals registered in the field were healthy, and did not show any major signs of
damage due to herbivory or pest attack. For all the species, the absence of sexual reproduction
recorded at some localities is likely caused by the extirpation of the pollinator from those places.
New IUCN Red List categories for Puerto Rican Zamia—Based on the information gathered
from the field I propose to change the IUCN Red List categories previously assigned to the
Puerto Rican zamias (Stevenson, 2010). Zamia erosa category should be changed from
Vulnerable (VU) to Least Concern (LC) (Table 2), and Zamia portoricensis category should be
71
changed from Endangered (EN) to Vulnerable (VU) (Table 2). In the case of Zamia pumila,
Stevenson (2010) considered the species as possibly extinct in Puerto Rico, and based its
category assignation as Near Threatened (NT) on estimates from Cuba and Dominican Republic.
Fortunately, the species has not been eradicated from Puerto Rico, but at the country level it
should be considered as Critically Endangered (CR) (Table 2).
DISCUSSION
Conservation status of Zamia in Puerto Rico—Zamia portoricensis is the only cycad endemic
to Puerto Rico. If geographic range is used as the main criterion to establish conservation
priorities for the cycads on the island, then this species should receive top priority. The largest
group of individuals occurs at Susúa State Forest and surrounding areas in the municipalities of
Sabana Grande and Yauco. Susúa State Forest has an area of 13 km2 and the land has been
protected since 1935 when it was acquired by the Puerto Rico Reconstruction Administration
(DRNA, 2007). This cycad is thriving in the area. Along one forest trail it is possible to walk for
about 2 km continuously observing Z. portoricensis of all sizes, in variable densities (Fig. 3A).
Outside the protected area, on privately owned land, plants are also abundant. Most individuals
in these areas grow on steep slopes not suitable for development where they are expected to
continue thriving undisturbed (Fig. 3B).
The second largest group of Z. portoricensis occurs in the Guánica State Forest but the
situation is quite different from that of the Súsua. In Guánica, adult plants are scattered and in
many areas no recruitment is taking place (Fig. 3D). Although the land has been protected since
1919 (DRNA, 2008), the pollinator seems to have been extirpated from areas that were heavily
disturbed in the past and has not been able to recolonize them. Fortunately, sexual
72
reproduction is still occurring in some places (Fig. 3E), and insects from those zones could be
used in future for management actions.
There is no report of other places, besides the two sites already mentioned, where large
groups of Z. portoricensis are present, and that is why the species should be considered
Vulnerable even when the number of mature individuals suggests that it should not be
considered as Threatened.
Zamia pumila is the rarest native cycad in Puerto Rico but historically it may have been very
abundant. Its current distribution is as broad as that of Z. portoricensis, and the existence of a
neighborhood named after the plant suggests that at least locally, it may have been a significant
component of the vegetation (Pagán-Jiménez and Lazcano-Lara, 2013) (Fig. 1). Habitat
destruction to establish pastures, quarries, crop fields, roads, urbanizations, and landfills is
extensive within the range of the species (Fig. 4B). In addition, Z. pumila seems to have been
particularly affected by an eradication campaign promoted by the Department of Agriculture
(Hernández de Rivera, 1966) due to an increase in cases of cattle poisoning after the ingestion of
cycad leaves (Reams et al., 1993) (Fig. 4E). I found only two populations where plants are
regularly recruiting; these populations are located within the Peñuelas and Ponce landfills and
quarries (Fig. 4C, D). Although they are not currently threatened, that status is not assured in
the long-term.
Zamia pumila also occurs in central Cuba and the Dominican Republic, so the category I
propose for it (i.e., Critically Endangered, CR) is only valid for Puerto Rico. However, the
taxonomic history of the species has been complicated (Stevenson, 2010), and the taxonomic
limits among several members of the Zamia pumila complex are still not clear. Further studies
73
may find that Puerto Rican populations are distinct; thus, until this is resolved, the conservation
of this species should also receive top priority (Donaldson, 2004).
Zamia erosa is perhaps one of the most abundant cycads in the world since the majority of
the extant cycad species exist as small, scattered populations (Donaldson, 2003). Regardless of
its status in other parts of its range (Cuba, Jamaica), the data presented here are sufficient to
assign it the Least Concern (LC) category. The three largest surveyed populations of this cycad
are located in protected areas (i.e., Julio E. Monagas National Park, Vega State Forest, and
Cambalache State Forest) (Fig. 4A-C). The species is expected to be located at many more sites
that will be explored and documented in the near future.
Cycads are able to survive many disturbances including fire, drought, and hurricanes (Norstog
and Nicholls, 1997) but at least some of their pollinators appear to be more sensitive to
disturbance than the plants. Pharaxonotha portophylla (Coleoptera: Erotylidae), the pollinator
of Z. portoricensis (Franz and Skelley, 2008), and Z. pumila (Tang, 2013, pers. com.) has been
extirpated from many locations in southern Puerto Rico (Table 1), a zone severely affected by
intentional fires. As a result, several populations solely consist of adult plants, which regularly
produce cones but, without pollinators, they are effectively sterile. The long-term survival of
these populations is clearly linked to the persistence of their pollinators; thus, management
actions should also be directed to protect this critical interaction.
Conservation actions recommended to protect the Puerto Rican zamias—To provide the
Puerto Rican Zamia with an opportunity for long-term persistence the following actions are
recommended:
74
1. Submit the results of this study to the pertinent authorities to encourage adequate
protection for this group of plants under state and federal laws, and gain official
recognition of its relevance as a wildlife resource.
2. Assure the long-term protection of at least one of the two extant self-sustainable
populations of Z. pumila by establishing a protected area or a conservation agreement
with landowners.
3. Translocate the isolated individuals of Z. portoricensis and Z. pumila that are located in
very disturbed, non-protected habitats to protected areas to found new self-sustainable
populations.
4. Restore sexual reproduction where it is not occurring due to the extirpation of the
pollinator, insufficient plant density, and/or inadequate spatial arrangement of plants of
both sexes.
5. Establish a monitoring program to periodically collect data on parameters relevant for
the conservation of the group (i.e. health, population structure, habitat quality, and seed
production as an indicator of pollinator presence).
6. Carry out a local-scale awareness campaign to educate landowners about the relevance
of this group of plants, and promote its protection at privately owned localities.
Conclusions—Based on quantitative data on plant density, population structure, and habitat
range, I have reassessed the conservation status of Zamia in Puerto Rico, and have proposed
new IUCN Red List categories for its three species. Zamia erosa category should change from
Vulnerable (VU) to Least Concern (LC), Z. portoricensis category should change from Endangered
(EN) to Vulnerable (VU), and Z. pumila should be considered Critically Endangered (CR) in Puerto
Rico.
75
Habitat destruction and a past eradication campaign are the major factors that drove Z.
portoricensis and Z. pumila to a threatened status since extant populations, at undisturbed sites,
are healthy, well-structured, and self-sustainable.
Because of the current threatened status of the southern Puerto Rican zamias, there is a
need for conservation actions to provide these populations with an opportunity for long-term
persistence.
76
LITERATURE CITED
ACEVEDO-RODRÍGUEZ, P., AND M.T. STRONG. 2005. Monocotyledons and Gymnosperms of Puerto
Rico and the Virgin Islands. Contributions from the United States National Herbarium
Volume 52: 1–415.
AXELROD, F.S. 2011. A Systematic Vademecum to the Vascular Plants of Puerto Rico. Botanical
Research Institute of Texas, Fort Worth, TX.
CAPUTO, P., S. COZZOLINO , P. DE LUCA, A. MORETTI, AND D.W. STEVENSON. 2004. Molecular
Phylogeny of Zamia (Zamiaceae). In T. Walters and R. Osborne [Eds]: Cycad Classification:
Concepts and Recommendations. CAB Publishing, Cambridge, MA.
CHAMBERLAIN, C.J. 1919. The Living Cycads. The University of Chicago Press. Chicago. IL.
CHAMBERLAIN, C.J. 1935. Gymnosperms: Structure and Evolution. University of Chicago Press,
Chicago. IL.
COX, P.A., S.A. BANACK, AND S.J. MURCH. 2007. Cyanobacteria, Cycads, and Neurodegenerative
Disease among the Chamorro People of Guam. Memoirs of the New York Botanical Garden
97: 253–285.
DONALDSON, J.S. 2003. Status Survey and Conservation Action Plan: Cycads. IUCN/SSC Cycads
Specialist Group. IUCN, Gland, Switzerland and Cambridge, UK.
DONALDSON, J.S. 2004. Saving Ghosts? The Implications of Taxonomic Uncertainty and Shifting
Infrageneric Concepts in the Cycadales for Red Listing and Conservation Planning. In T.
Walters and R. Osborne [Eds.], Cycad Classification, Concepts and Recommendations. CABI
Publishing, Cambridge, MA.
DONALDSON, J.S., I. NÄNNI, AND J.D. BÖSENBERG. 1995. The role of insects in the pollination of
Encephalartos cycadifolius. In P. Voster [ed.], Proceedings of the third international
77
conference on cycad biology. Cycad Society of South Africa, Stellenbosch, South Africa.
423–434.
DRNA. 2007. Hojas de Nuestro Ambiente. Departamento de Recursos Naturales y Ambientales.
Estado Libre Asociado de Puerto Rico. P–022.
DRNA. 2008. Hojas de Nuestro Ambiente. Departamento de Recursos Naturales y Ambientales.
Estado Libre Asociado de Puerto Rico. P–034.
ECKENWALDER , J.D. 1980. Taxonomy of the West Indian Cycads. Journal of the Arnold Arboretum
61: 701–722.
FRANZ, N.M., AND P.E. SKELLEY . 2008. Pharaxonotha portophylla (Coleptera: Erotylidae), new
species and pollinator of Zamia (Zamiaceae) in Puerto Rico. Caribbean Journal of Science
44: 321–333.
GONZÁLEZ GÉIGEL , L.G. 2003. Zamiaceae. Flora de Cuba 8: 3–22.
HALL, J.A., G.H. W ALTER, D.M. BERGSTROM, AND P. MACHIN. 2004. Pollination ecology of the
Australian cycad Lepidozamia peroffskyana (Zamiaceae). Australian Journal of Botany 52:
333–343.
HAYNES, J.L. 2011. World List of Cycads: A Historical Review. IUCN/SSC Cycad Specialist Group.
Unpublished. 36 pp. Document retrieved from www.cycadsg.org/pages/worldlist.
HERNÁNDEZ DE RIVERA, B.S. 1966. Advierten sobre planta venenosa que causa parálisis en
animales. Agricultura al Día 12: 17–18.
IUCN. 2012. IUCN Red List Categories and Criteria: Version 3.1, 2nd ed. Gland, Switzerland and
Cambridge, UK.
JONES, D.L. 2002. Cycads of the World: Ancient Plants in Today’s Landscape, 2nd ed. Smithsonian
Institution Press, Washington, D.C.
78
KLAVINS, S.D., D.W. KELLOGG, M. KRINGS, E.L. TAYLOR , AND T.N. T AYLOR. 2005. Coprolites in a
Middle Triassic cycad pollen cone: evidence for insect pollination in early cycads?
Evolutionary Ecology Research 7: 479–488.
KONO, M., AND H. TOBE.2007. Is Cycas revoluta (Cycadaceae) wind or insect pollinated? American
Journal of Botany 94: 847–855.
MEEROW, A., J. FRANCISCO -ORTEGA, M. CALONJE , M.P. GRIFFITH, T. AYALA-SILVA, D.W. STEVENSON,
AND K. N AKAMURA . 2012. Zamia (Cycadales: Zamiaceae) on Puerto Rico. Asymmetric genetic
differentiation and the hypothesis of multiple introductions. American Journal of Botany
99: 1828–1839.
MOUND, L.A., AND I. TERRY. 2001. Thrips Pollination of the Central Australian Cycad, Macrozamia
macdonellii (Cycadales). International Journal of Plant Science 162: 147–154.
NEWELL , S.J. 1983. Reproduction on a natural population of cycads (Zamia pumila L.) in Puerto
Rico. Bulletin of the Torrey Botanical Club 110: 464–473.
NORSTOG, K.J., P.K.S. FAWCETT , AND A.P. VOVIDES. 1992. Beetle pollination of two species of
Zamia: Evolutionary and ecological considerations. The Palaeobotanist 41: 149–158.
NORSTOG, K.J., AND T.J. NICHOLLS. 1997. The Biology of the Cycads. Cornell University Press,
Ithaca, NY.
PAGÁN-JIMÉNEZ, J.R., AND J.C. LAZCANO-LARA. 2013. Toponymic Data Helps to Reveal the
Occurrence of Previously Unknown Populations of Wild Zamia pumila L. on Volcanic
Substrates in South Central Puerto Rico. Ethnobiology Letters 4: 52–58.
PROCHES S., AND S.D. JOHNSON. 2009. Beetle pollination of the fruit-scented cones of the South
African cycad Stangeria eriopus. American Journal of Botany 96: 1722–1730.
79
REAMS, R.Y., E.B. J ANOVITZ, F.R. ROBINSON, J.M. SULLIVAN, C.R. CASANOVA, AND E. MÁS. 1993.
Cycad (Zamia puertoriquensis) toxicosis in a group of dairy heifers in Puerto Rico. Journal of
Veterinary Diagnostic Investigation 5: 488–494.
SHAW, C.E., J.M.B. WILSON, R. CRUZ-AGUADO, S. SINGH, E.L. HAWKES, V. LEE, AND T. MARLER. 2007.
Cycad-Induced Neurodegeneration in a Mouse Model of ALS-PDC: Is the Culprit Really
BMAA, or is a Novel Toxin to Blame? Memoirs of the New York Botanical Garden 97: 286–
307.
STEVENSON, D.W. 1987. Again the West Indian Zamias. Fairchild Tropical Garden Bulletin 42: 23–
27.
STEVENSON, D.W. 2010. Zamia erosa, Z. portoricensis, Z. pumila. In IUCN 2013. IUCN Red List of
Threatened Species. Version 2013.2. Website http://www.iucnredlist.org [accessed on 17
April 2015].
TANG, W. 1987. Insect Pollination in the Cycad Zamia pumila (Zamiaceae). American Journal of
Botany 74: 90–99.
USFWS. 2015. Endangered and Threatened Plant Species. U.S. Fish and Wildlife Service.
Ecological Services in the Caribbean, Southeast Region. Website:
http://www.fws.gov/caribbean/es/Endangered-Plants.html. [Accessed on 17 April 2014].
VOVIDES, A. P. 1990. Spatial distribution, survival, and fecundity of Dioon edule (Zamiaceae) in a
tropical deciduos forest in Veracruz, Mexico, with notes on its habitat. American Journal of
Botany 77: 1532–1543.
WILSON, G.W. 2002. Insect Pollination in the Cycad Genus Bowenia Hook. ex Hook. f.
(Stangericaceae). Biotropica 34: 438–441.
80
CHAPTER THREE
TABLES
81
Table 1. Population structure of three Zamia species of Puerto Rico. Data are from single 25 
20 m plots per locality. M = # of male plants; F = # of female plants; NRA = # of non-reproductive
adults; S-J = # of seedlings and juveniles; AOOp = area of occupancy per plot (m2); AOOl = area of
occupancy per locality (m2); NMsc = # of mature individuals based on size class per locality;
NMcon = # of mature individuals based on coning plants per locality.
Species / Locality
M
F
NRA
S-J
AOOp
AOOl
NMsc NMcon
Manatí
19
15
57
64
17.12
171.2
910
340
Cambalache
38
37
411
170
84.56
10570
60750
9375
Cueva Ventana
55
153
174
188
68.64
686.4
3820
2080
Garrochales
68
35
420
136
89.12
2495.36
14644
2884
Candelaria
15
24
58
127
20.60
206
970
390
Design Building
41
57
222
214
59.76
597.6
3200
980
J.E. Monagas
57
110
388
316
101.44
1521.6
8325
2505
Bosque Estatal de Vega
97
138
296
510
105.36
13591.44
68499
30315
Guánica 1
3
4
18
0
4.00
1024
0
0
Guánica 2
13
8
29
27
9.08
1743.36
9600
4032
Carr. El Tamarindo
59
46
75
432
46.08
5160.96
20160
11760
Bosque Estatal de Susúa
73
34
111
847
68.76
27504
87200
42800
Sierra Bermeja
4
2
0
0
0.96
9.6
0
0
Cabo Rojo
3
4
0
0
1.12
11.2
0
0
Barrio Cuevas 1
1
0
8
0
1.44
0.0864
0
0
Barrio Cuevas 2
4
2
5
0
1.76
2.816
0
0
Barrio Cuevas 3
7
7
0
0
2.24
5.6
0
0
Barrio Cuevas 4
8
8
0
0
2.56
8.2432
0
0
La Rita
2
2
12
0
2.56
32.768
0
0
Vertedero Ponce
16
13
44
78
14.80
59.2
292
116
Vertedero Peñuelas
13
10
39
63
12.44
298.56
1488
552
Tallaboa Alta
4
5
1
0
1.60
2.88
0
0
Carr. 382-Carr. 383
1
2
24
3
4.44
22.2
135
15
Zamia erosa
Zamia portoricensis
Zamia pumila
82
Table 2. IUCN Red List categories, criteria and supporting information for Zamia in Puerto Rico.
Ze = Zamia erosa; Zpr = Zamia portoricensis, Zpm = Zamia pumila; NL = # of localities visited for
this study; EOO = extent of occurrence in km2; AOO = area of occupancy in km2; NMsc = # of
mature individuals based on size class; NMcon = # of mature individuals based on conning
plants; 2010 = present category (Stevenson, 2010); 2015 = category proposed in this study.
Species
NL
EOO
AOO
NMsc
NMcon
2010
Criteria
201
Criteria
2010
5
2015
Ze
22
395
0.046
240173
80654
VU
C1
LC
n/a
Zpr
6
50
0.034
116960
58592
EN
B2ab(iii,v);
VU
B1a+2a;
C1
Zpm
8
53
0.0004
1915
683
NT
n/a
D2
CR
A4c;
B1a+2a;
C1
83
CHAPTER THREE
FIGURES
84
Figure 1. Distribution map of Zamia in Puerto Rico. The dots represent extant localities
corroborated with field visits, and also represent the limits of the distribution of each species.
Blue dots are Zamia erosa localities, red dots are Zamia portoricensis localities, and yellow dots
are Zamia pumila localities. Four localities (green dots) were referenced from the literature, one
in Isabela, one in Río Abajo, one in Punta Guaniquilla, and one in Peñuelas (Acevedo-Rodríguez
& Strong, 2005; Meerow et al., 2012). Scale bar = 10 km.
85
Figure 2. General status of Zamia erosa in Puerto Rico. (A) Vega State Forest, complete
dominance of the species on the understory; (B) Road 675, several plants can be seen at the
edge of the limestone hill; (C) Julio E. Monagas National Park, the species is less dominant but
prominent on the understory; (D) Garrochales, seedlings of two different ages near the mother
plant.
86
Figure 3. General status of Zamia portoricensis in Puerto Rico. (A) Susúa State Forest, a dense
group of plants at a relatively open portion of the forest; (B) El Tamarindo, a female cone with
ripe seeds ready to be dispersed; (C) general aspect of the southern serpentine outcrop of the
island; (D) Guánica 1, one of the many isolated adult plants present in the area; (E) Guánica 2, a
female cone with ripe seeds ready to be dispersed.
87
Figure 4. General status of Zamia pumila in Puerto Rico. (A) Barrio Cuevas 1, a plant at a small
limestone outcrop, an island of high quality habitat; (B) Barrio Cuevas 4, a plant in an open area
affected by fire in the past; (C) Vertedero Ponce, one of the two healthy populations
documented; (D) Vertedero Peñuelas, a female and a male located side by side, the female cone
is almost ready to disperse the seed; (E) Yabucoa Alta, an isolated individual in volcanic
substrate.
88
GENERAL CONCLUSIONS
The combined results of these studies indicate that female reproductive success in Z.
portoricensis is highly variable and affected by the proximity to male plants. Reproductive
success is highly variable but when the distance to the nearest male exceeds 1.9 m it
dramatically drops to very low levels (Chapter 1). Limited pollen dispersal in association with
limited seed dispersal explains the small values of gene dispersal, 2.4 and 3.1 m, as well as the
occurrence of significant spatial genetic structure at the local, fine-scale used here (Chapter 2).
Female reproductive success is also related to neighborhood sex ratios. In circular vicinities
of 1.5 m radius, the higher the male bias, the higher the seed set for females located at the
circle center. The relationship is significant, but highly variable (Chapter 1) suggesting that even
within effective pollen dispersal distances successful pollination is not guaranteed. As the
individuals are randomly located in relation to each other (Chapter 1), being close to male plants
and/or located within a high pollen availability vicinity is a matter of chance. The effect of this
stochastic process on reproductive success is perhaps one of the factors that produces the high
levels of intra-population genetic diversity found in Z. portoricensis.
Limited gene dispersal, as evidenced by a dramatic drop in the kinship coefficient values from
1.5 m to 3 m (Chapter 2), is perhaps the main cause of the occurrence of identifiable
discontinuities in the distribution of genotypes resulting in clusters of related individuals
(Chapter 2). The size and configuration of these clusters appear to be influenced by
topographical features of the habitat (Chapter 1 and 2).
The remarkable long-term persistence of the cycads indicates the effectiveness of their
pollination system. Its variability in effectiveness in Z. portoricensis (Chapter 1) could be
interpreted as the result of the stochasticity associated with random spatial distribution of
89
individuals of both sexes and the lack of synchronization between pollen release and female
receptivity but cycads are long-lived plants; thus, the net reproductive success of a particular
individual can only be assessed via long-term studies.
Of the three species of Zamia present in the island, Z. erosa is not only abundant, but also
has the widest distribution both in terms of range and number of localities. Zamia portoricensis
is an abundant species but is not as widely distributed as Z. erosa. And Z. pumila is the rarest
native cycad in Puerto Rico. Although it has been recorded at more places than Z. portoricensis,
most Z. pumila populations consist of a small number of adults that do not exhibit sexual
reproduction (Chapter 3).
Based on quantitative data on plant density, population structure, and habitat range, I have
reassessed the conservation status of Zamia in Puerto Rico, and have proposed new
International Union for the Conservation of Nature Red List categories for its three species
(Chapter 3). Zamia erosa category should change from Vulnerable (VU) to Least Concern (LC), Z.
portoricensis category should change from Endangered (EN) to Vulnerable (VU), and Z. pumila
should be considered Critically Endangered (CR) in Puerto Rico. Habitat destruction and a past
eradication campaign are the major factors that likely drove Z. portoricensis and Z. pumila to a
threatened status since extant populations, at undisturbed sites, are healthy, well-structured,
and self-sustainable (Chapter 1 and 3).
In this study I used neutral molecular markers, therefore no particular prediction can be
made about the species long-term capabilities to adapt to a dynamic environment in the future,
but the results obtained here show that at El Tamarindo and Susúa State Forest, Z. portoricensis
has been able to maintain an adequate population structure (Chapter 1) and high levels of
90
genetic diversity (Chapter 2); thus, both populations can be used as models to set management
goals for future conservation programs.
Because of the current threatened status of the southern Puerto Rican zamias, there is a
need for conservation actions to provide these populations with an opportunity for long-term
persistence.
91