(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Natural history of the myrmecophilic spider, Masoncus pogonophilus Cushing, and its host ant, Pogonomyrmex badius (Latreille)"

NATURAL HISTORY OF THE MYRMECOPHILIC SPIDER, 
MASONCUS POGONOPHILUS CUSHING, AND ITS HOST ANT, 
POGONOMYRMEX BADIUS (LATREILLE) 



BY 
PAULA ELIZABETH CUSHING 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT 
OF THE REQUIREMENTS FOR THE DEGREE OF 
DOCTOR OF PHILOSOPHY 

UNIVERSITY OF FLORIDA 

1995 



UNIVERSITY OF FLORIDA LIBRA 



RIES 



DEDICATION 
I dedicate this to my parents, Paula M. Cushing and 
Joseph Cushing (deceased), who encouraged my curiosity and 
taught me to see the world from many different angles, and 
to my brother, Richard J. Cushing, and my sisters, Joan D. 
Cushing and Patricia C. Banta, for their humor and love. 



ACKNOWLEDGMENTS 

Thanks to my advisor, Dr. Jonathan Reiskind for his 
guidance and friendship. Thanks also to my committee 
members, Drs. John Anderson, Clifford Johnson, Michael 
Miyamoto, Sanford Porter, John Sivinski, and Robert Vander 
Meer, for providing advice, expertise, equipment and moral 
support. 

This work could not have been completed without the 
assistance of many people. I would like, in particular, 
to thank three undergraduates who spent countless hours in 
the field and the laboratory helping me: Michael Morgan, 
Joyce Thomas, and John Arnett. Their good humor and 
enthusiasm were greatly appreciated. Thanks also to the 
following for their assistance and their support: Joe 
Allen, Susan Moegenburg, Ray Moranz , Mark Stowe , Dr. Evan 
Chipouras, Dr. Dan Brazeau, Dr. Rich Buchholz, Dr. Laurie 
Eberhard, Ron Clouse, Dr. John Donald, Dr. Tes Toop, Dr. 
Pete Lahanas, Dr. David Evans, Dr. Doug Levey, Dr. Kent 
Vliet, Tom Workman, Erick Smith, Ryan Rodgers , Barbara 
Dixon, Thia Hunter, Rebecca Forsyth, Steven Craig, Sue 
Chien, Mark Hostetler, and Paul Wyness. 

A special thanks to Pete Ryschkewitsch and the rest 
of the staff in the stockroom as well as Jimmy Norton for 
being so helpful, good-humored, and patient with my many 

iii 



requests for supplies and equipment loans. Pete, 
especially, went out of his way to help me and I greatly 
appreciate it. Thanks also to the Department of Zoology 
for providing graduate students access to field vehicles 
and other research equipment. 

Thanks to Al Boning of the University of Florida 
Animal Nutrition labs for his help with the bomb 
calorimetry and lipid analyses and to Richard Fethiere of 
the Animal Nutrition Labs for running the protein analysis 
used in Chapter 5. Thanks also to Rodney Young of the 
USDA-APHIS Seed Examination Facility in Bethesda, 
Maryland, for his patience and enthusiasm in identifying 
many of the seed species from Chapter 5. 

This work was supported by a three-year Grinter 
Fellowship and an O'Neil Dissertation fellowship from the 
College of Liberal Arts and Sciences, a scholarship from 
the Florida Entomological Society, teaching assistantships 
from the Biological Sciences program, a Theodore Roosevelt 
Grant from the American Museum of Natural History, and a 
mini-grant from the Florida Entomological Society. 



IV 



TABLE OF CONTENTS 

page 

ACKNOWLEDGMENTS iii 

ABSTRACT vi i 

CHAPTERS 

1 INTRODUCTION 1 

2 DESCRIPTION OF THE SPIDER MASONCUS 

POGONOPHILUS . N. SP. (ARANEAE: LINYPHIIDAE) — 
A HARVESTER ANT MYRMECOPHILE 10 



GENETIC DIVERSITY WITHIN AND AMONG 

POPULATIONS OF MASONCUS POGONOPHILUS USING 
RANDOM AMPLIFIED POLYMORPHIC DNA (RAPD) 
FINGERPRINTING 

Introduction 

Materials and Methods 

Results 

Discussion 

THE ABILITY OF THE MYRMECOPHILIC SPIDER TO 
FOLLOW THE TRAILS OF ITS HOST ANT 

Introduction 

Materials and Methods 

Results 

Discussion 

FACTORS AFFECTING SEED SELECTION BY THE 

FLORIDA HARVESTER ANT POGONOMYRMEX BADIUS 
AT TWO NORTH FLORIDA SITES 

Introduction 

Materials and Methods 

Results 

Discussion 



26 

26 
30 
32 
35 



49 

49 
51 
58 
60 



68 

68 
71 
78 
82 



NEST DISPERSION AND INTER-NEST AGGRESSION OF 
POGONOMYRMEX BADIUS AND THE EFFECT OF HOST 
POPULATION STRUCTURE ON SPIDER INTEGRATION 
INTO COLONIES 



92 



Introduction 92 

Materials and Methods 95 

Results 100 

Discussion 102 

7 SUMMARY AND CONCLUSIONS Ill 

REFERENCES 120 

BIOGRAPHICAL SKETCH 130 



VI 



Abstract of Dissertation Presented to the Graduate 
School of the University of Florida in Partial Fulfillment 
of the Requirements for the Degree of Doctor of Philosophy 

NATURAL HISTORY OF THE MYRMECOPHILIC SPIDER, 
MASONCUS POGONOPHILUS CUSHING, AND ITS HOST ANT, 
POGONOMYRMEX BADIUS (LATREILLE) 

By 

Paula Elizabeth Cushing 

December 1995 

Chairman: Dr. Jonathan Reiskind 
Major Department: Zoology 

Masoncus poqonophilus Cushing, a small (2 mm long) 
spider in the family Linyphiidae, spends all life stages 
inside the nest chambers of the Florida harvester ant, 
Pogonomyrmex badius (Latreille) . The spiders appear to be 
commensals, taking advantage of the stable microclimate 
and abundant food available within the nests. However, 
the ecology of the host ant does directly affect the 
ecology of the spider. 

The fragility of M. poqonophilus as well as an 
apparent female-biased sex ratio suggested that dispersal 
of spiders between ant nests may be uncommon. As this 
would affect the genetic structure of spider populations, 
the genetic diversity within and among three populations 
was measured using the PCR-based Random Amplified 
Polymorphic DNA technique (RAPD-PCR) . Of the total genetic 
diversity from three populations of spiders, 77.4% was 

vii 



attributable to intra-population differences, 18.2% to 
differences between distant populations, and only 4.4% to 
differences between neighboring populations of spiders. 
Individuals living within one ant nest are not closely 
related and spiders from neighboring nests are more 
similar genetically than spiders from distant nests. 
Thus, dispersal events of spiders between neighboring 
colonies are interpreted as occurring at a significant 
rate. 

There is no evidence that the spiders use the trail 
pheromones of the ants to locate new host colonies. 
Although the host ants oriented significantly more to 
extracts of trail pheromones and to natural trails than to 
control trails, the spiders did not. 

The success of dispersal events among the spiders is 
affected by dispersion of the ant colonies. Dispersion of 
host colonies is, in turn, affected by habitat structure, 
resource availability and agonistic interactions between 
colonies. 

A detailed description is presented of the resources 
available to P. badius colonies in two different habitats 
as well as what resources (seeds) they store in their 
granaries. Differences in spacing patterns, densities, 
and aggressive interactions between P. badius nests in 
these two habitats are due to a complex set of factors 
including habitat structure and resource availability. 



vm 



CHAPTER 1 
INTRODUCTION 

Many arthropods have evolved symbiotic relationships 
with ants. Some are found at the periphery of the nest, 
either near the entrances or on refuse piles; others are 
found within the chambers of the nest, either in the 
peripheral chambers or deeper in the nest in the brood and 
storage chambers (Holldobler 1977). They range from tiny 
collembolans to beetles and caterpillars many times the 
size of their hosts (Holldobler and Wilson 1990). The 
formal study of myrmecophiles began with the work of 
Wasmann in 1894 who developed a classification system for 
myrmecophiles consisting of distinct categories, each 
suggesting increasing specialization and integration into 
the host colony. 

In this study, I explore different facets of the 
natural history and ecology of one myrmecophile and its 
host ant in an attempt to determine what factors are most 
important in the integration of this spider into the 
colony of its host. In general, ants live in complex 
societies in which only members are allowed. They 
communicate with their nestmates through chemical and 
tactile signals, and they tend to aggressively exclude 
intruders into their colony. The arthropod ant guests, or 
myrmecophiles, have evolved various adaptations enabling 

1 



them to exist in this hostile environment. Many of the 
myrmecophiles acquire cuticular hydrocarbons similar or 
identical to those of their hosts (Vander Meer and Wojcik 
1982, Vander Meer et al. 1989). This allows them to 
become integrated with hosts that are otherwise hostile to 
intruders with foreign, non-colony odors. Others, such as 
some staphylinid beetles and lycaenid caterpillars, have 
evolved specialized glands that produce appeasement 
substances (reviewed in Holldobler and Wilson 1990). 

In many myrmecophiles, the evolution of a symbiotic 
association can be intimated through an examination of 
extant species that show varying degrees of behavioral 
integration (Holldobler and Wilson 1990). For example, 
Akre and Rettenmeyer (1966) described species of 
staphylinid beetles that show varying degrees of 
association with army ants. Some species live only around 
the edges of the bivouacs or in the refuse piles but are 
not otherwise integrated into the colonies, others are 
found running along the edges and sometimes within the 
emigration columns of ants , and yet others are found 
directly in the midst of ants in the center of the 
emigration colonies. Some species even hitch rides on the 
booty or the brood carried by ants. Certain staphylinid 
species can only live within a narrow range of conditions 
found within colonies and die shortly after removal from 
the colonies. 



If each stage in this process of gradual integration 
into colonies is correlated with the evolutionary history 
of the lineages, then the various adaptations of the 
myrmecophiles leading to greater integration could be 
viewed as characters on the evolutionary tree (Brooks and 
McLennan 1991). Kistner (1979) takes this idea a step 
further by superimposing the phylogenies of termites in 
the family Rhinotermitidae with their associated 
termitophiles in the family Staphylinidae to illustrate 
the evolution of host specificity. Predation pressures may 
have triggered greater integration into the ant and 
termite societies in these staphylinid species as well as 
in other myrmecophiles and termitophiles since association 
with the aggressive hosts may afford some protection to 
the guests. Close association with the hosts itself may 
have led to integration within the colonies. Stable 
microclimatic conditions within the ant colonies as well 
as an abundant food supply (either in the form of host 
brood or other colony guests) would select for even 
greater integration into the colonies. 

Myrmecophilic spiders are unigue because their close 
relatives apparently have no preadaptations to a symbiotic 
lifestyle. Most spiders are solitary predators and 
symbiosis with other arthropod groups should be rare; yet 
myremcophilic spiders are found in the families 
Agelenidae, Aphantochilidae, Clubionidae, Gnaphosidae, 
Linyphiidae, Oonopidae, Salticidae, Theridiidae, 



Thomisidae, and Zodariidae (Donisthorpe 1927, Bristowe 
1939, Noonan 1982, Porter 1985, Holldobler and Wilson 
1990, Boeve 1992). Many of these spiders are specialized 
ant predators, but several, such as the clubionids in the 
genus Phrurolithus and the Linyphiid, Masoncus 
pogonophilus . are found in the company of the host ants 
and do not feed on the ants or their brood. Such species 
may be occasional visitors into ant colonies, using the 
entrance and upper chambers as temporary refuges, or they 
may be commensals that have become more dependent on the 
conditions present within the nest and spend their entire 
lives within this complex ecosystem. 

Within the colony chambers of the Florida harvester 
ant, Pogonomyrmex badius (Latreille) (Formicidae) , lives a 
small, approximately 2 mm long, species of spider, M. 
pogonophilus (Linyphiidae; Erigoninae) . This spider-ant 
association was first described by Porter (1985). All 
life stages of M. pogonophilus are found inside the ants' 
nest. They feed on collembolans (springtails) and perhaps 
other tiny symbiotic arthropods. All developmental stages 
of the spiders as well as spider eggsacs are found in all 
portions of the nests throughout the year (Fig. 1-1). 
During any given month, they are as likely to be found in 
shallow as in deep chambers or runways. Therefore, they 
are not occasional guests but true members of the nest 
community. 



The objectives of this project were to: 1) describe 
this species of myrmecophile, compare it to other members 
of the genus, and document its life cycle and life 
history; 2) determine how the population structure of the 
host ants affects the population structure of the 
myrmecophilic spider; 3) determine the dispersal mechanism 
of the myrmecophiles; and 4) investigate the factors, such 
as resource availability, habitat structure, and 
inter-nest competition that directly affect host ant 
population structure and indirectly affect dispersal of 
myrmecophiles and integration of myrmecophiles into new 
host colonies. 

Chapter 2 is a formal description of this previously 
undescribed spider. Only three other species of the genus 
Masoncus had been described prior to this study: M. 
arienus , M. conspectus , and M. dux . None of these species 
is known to be associated with ants. In Chapter 2 I 
compare M. pogonophilus with the described species and 
note their morphological differences. I also summarize 
what is known about the natural history and life cycle of 
M. pogonophilus based both on my own observations as well 
as those of Porter (1985). 

Both adult and juvenile spiders can be found in P. 
badius nests throughout the year. The spiders are common 
in nests within a given area so dispersal of spiders is 
evidently occurring. However, due to the spiders' 
susceptibility to desiccation outside the nests and the 



huge distances these tiny spiders would have to traverse 
to locate a new colony as well as to an apparent 
female-biased sex ratio among the spiders, I hypothesized 
that dispersal events are uncommon and that spider 
populations within ant colonies started with one or a few 
founding females. This type of population founding was 
suggested by Williams and Franks (1988) for a 
myrmecophilic isopod. They suggested that the 
myrmecophiles remained within a single colony for several 
generations until that colony died, at which time the 
myrmecophiles would disperse and a few would locate a new 
host nest. If this scenario were true for M. 
pogonophilus . then spiders within a nest should be 
genetically more similar to each other than to spiders 
from different nests. In Chapter 3, I test this 
hypothesis using the Random Amplified Polymorphic DNA 
(RAPD) fingerprinting technique. 

In Chapter 4, I attempt to determine the dispersal 
mechanism of the spiders. Pogonomyrmex badius emigrates 
to new nest sites, usually less than 5 m from the old 
site. Such emigrations are common in this species of 
harvester ant (Golley and Gentry 1964, Gentry and Stiritz 
1972, Gordon 1992). Between 60 and 97% of the colonies in 
an area migrate to a new nest site once a season, a few 
(under 30%) migrating two or three times (Carlson and 
Gentry 1973). When the ants move, so do the spiders and 
the symbiotic collembolans . The symbionts move in the 



emigration trail on their own accord and rarely veer out 
of the trail. I observed several such emigrations of the 
myrmecophiles. These field observations suggested that 
the spiders might detect and follow colony odors. I 
hypothesized, based on these observations, that spiders 
were locating new colonies by following trail pheromones 
of the ants, perhaps by following foraging trails away 
from the host nests until they located the foraging trail 
of a neighboring colony which they then followed to the 
new mound. I test this hypothesis using chemical 
bioassays of trail pheromones as well as using bioassays 
with a naturally laid trail. 

In Chapters 5 and 6, I shift the focus of the study 
toward the population structure and foraging ecology of 
the host ants since the population structure of the host 
ants indirectly affects the dispersal of the spider and 
integration of the myrmecophile into new nests. Closely 
spaced P. badius nests, for example, would be potentially 
easier for dispersing spiders to locate. However, if 
neighboring nests are aggressive towards one another and 
if spiders are absorbing host colony odors, then 
dispersing spiders arriving at a new nest may face a 
behavioral barrier from aggressive ants. The population 
structure of P. badius, in turn, is likely influenced by 
resource availability and habitat structure. Chapter 5 
deals exclusively with resource use by the harvester ants 
and the effect of habitat structure on resource 



availability. This information is used to explain 
patterns of nest dispersion and inter-nest aggression 
described in Chapter 6. 

In Chapter 6, I explore the dispersion of P. badius 
colonies and the possible influence of inter-nest 
aggression on nest dispersion. Inter-colony competitive 
interactions are implicated in many studies of nest 
dispersion among ants (de Vita 1979, Levings and Traniello 
1981, Harrison and Gentry 1981, Cushman et al. 1988). 
Inter-colony aggression between colonies of P. badius had 
not been guantitatively evaluated prior to this study. I 
considered this information crucial to understanding what 
behavioral barriers dispersing spiders might have to face 
from hostile ants. In Chapter 6, I also explore whether 
there is behavioral evidence to support the hypothesis 
that M. pogonophilus uses chemical mimicry to integrate 
itself into ant colonies. 



100' 



80- 



e 
u 



X 

Eh 

W 
Q 



60" 



T 
■ 
1 



40" 



20" 



T 

■ 

1 



I 

■ 

1 



T 

■ 

i 



T 

■ 

1 



T 
■ 

J. 



ill 



i — i — i — i — i — i — r 



i — r 



t — i — i — i — i — r 



pororonmnmLntNroroLntN 



Ol 


CT\ 


a\ 


CTl 


CT\ 


01 


o\ 


en 


o\ 


cn 


o\ 


CT> 


CT\ 


o\ 


1 


o 

1 


CO 
(M 

i 


O 

,-H 

1 


r-l 

1 


1 


ro 

1 


CO 
1 


1 


*J3 
t-H 

i 


in 

CM 

1 


cri 

1 


1 
CO 


in 

1 



O) (N <* <* 



EXCAVATION DATE 



o\ Q\ <J\ 



Figure 1-1. The mean depths + standard error of 
samples of spiders found during 17 different excavations. 
The excavation dates range from January 10th to October 
8th. The dates are listed in order by month rather than 
by actual order of excavation. The number of spiders 
making up each sample are shown above the error bars. 



CHAPTER 2 
DESCRIPTION OF THE SPIDER MASONCUS POGONOPHILUS . N. SP. 
(ARANEAE: LINYPHIIDAE) — A HARVESTER ANT MYRMECOPHILE 



Three species are included in the genus Masoncus 
Chamberlin 1948: M. arienus Chamberlin 1948, M. dux 
Chamberlin 1948, and M. conspectus (Gertsch and Davis 1936) 
(synonymized with M. nogales Chamberlin 1948 by Ivie 1967). 
The female holotype of M. dux has been lost and I was 
unable to locate any specimens of this species. The female 
holotype, male allotype and paratypes of M. nogales 
designated by Chamberlin (1948) have also been lost. 
However, the holotype of Tapinocyba conspecta is deposited 
at the American Museum of Natural History in New York, NY 
(AMNH) as are other representatives of this species. The 
holotype and paratypes of M. arienus designated by 
Chamberlin (1948) are also at AMNH. One male 
representative of M. arienus is deposited at the California 
Academy of Sciences in San Francisco, CA (CAS). 

No information was recorded either in the original 
species descriptions or on the collecting labels of the 
existing specimens regarding the natural history of the 
described species. Masoncus dux was described from a 
single female collected in northern Manitoba, Canada. All 
specimens of M. arienus were collected in Arizona. 
Masoncus conspectus was described from the male holotype 

10 



11 



and two male paratypes collected in Texas. Other records 
of this species include Arizona and Florida (the latter 
specimen collected by the shores of Newnan's Lake in 
Alachua County) . 

Masoncus pogonophilus n. sp. was originally collected 
by Sanford Porter from the nests of the Florida harvester 
ant, Pogonomyrmex badius (Latreille) (Hymenoptera: 
Formicidae) near Tallahassee, Florida in Leon County 
(Porter 1985). It is included in the genus Masoncus due to 
the presence of distinct cephalic pits and a straight, 
distally bifid embolic division in the males (see genus 
description below) . 

In the species description that follows, I use 
primarily carapace, genitalic, chaetotaxic, numeric, and 
palpal characters deemed most useful by Millidge (1980) for 
erigonine spiders. These characters include: 1) the 
overall conformation of the male palpal organ, 2) the shape 
of the embolic division, 3) the external appearance of the 
epigynum, 4) the number of dorsal trichobothria present on 
the palpal tibia of both sexes, 5) the number of dorsal 
tibial spines present (expressed by the formula a:b:c:d), 
6) the number of dorsal metatarsal trichobothria present 
(expressed by the formula I : II : III : IV) , 7) the relative 
position of the dorsal metatarsal trichobothrium on leg I 
(expressed by the formula Tml = distance from 
tibia-metatarsus joint to trichobothrium / distance from 
tibia-metatarsus joint to metatarsus-tarsus joint), and 8) 



12 



the relative stoutness of tibia I (expressed by the formula 
TibI = length of tibia / width of tibia viewed laterally) . 
Overall body size, body color, and number of setae on the 
carapace are also given. Certain of these characters as 
well as others used in Chamberlin's (1948) descriptions or 
obvious on the existing specimens are of particular value 
in separating M. arienus . M. conspectus , and M. 
pogonophilus (Table 2-1). All measurements were taken 
directly off the specimens using an ocular micrometer in a 
dissecting scope. Measurements were rounded to the nearest 
0.1 mm . 

Masoncus Chamberlin 1948 

The type species of the genus is M. arienus . The 
genus Masoncus is characterized by both cephalic pits in 
the males and a straight, distally bifid embolic division 
(Chamberlin 1948) (diagram of Linyphiid palpal structures 
in Millidge 1980) . 

Masoncus pogonophilus new species 
(Figs. 2-1 to 2-5) 

The male holotype was collected 23 cm below ground 
inside a nest chamber of the Florida harvester ant, P. 
badius in Archer Sandhills, 1.4 Km west of the Levy Co. 
line off of State Road 24. The female allotype was 
collected from the same P. badius nest. She was found in a 
nest chamber 46.5 cm below ground. Both were collected 25 
September 1994 and both will be deposited in the 
Arachnological collection at CAS. 



13 



The holotype, eleven male paratypes, the allotype, 
and 12 female paratypes were used in this species 
description. The collecting information as well as the 
future museum destination for these paratypes are presented 
in Table 2-2. 

Etymology . The specific epithet is derived from the 
generic name of the host ant. 

Holotype (male) . Total body length: 1.7 mm. 
Carapace length: 0.9 mm. Carapace width: 0.7 mm. Colors: 
Carapace orange; abdomen grey; legs orange; sternum orange. 
Number of setae along midline of carapace: three. Palp as 
in Fig. 2-3. Embolic division distally bifid with the 
proximal part of the bifurcation bent forward and extending 
over the most distal part (Fig. 2-4). Number of 
trichobothria on palpal tibia: two (Fig. 2-2). Number of 
dorsal tibial spines: 1:1:1:1. Number of dorsal metatarsal 
trichobothria: 1:1:1:0. Tml : 0.82. TibI : 7.0. 

Males (general) - (n = 12) . Total body length: 1.6 
- 2.1 mm (x = 1.8 ± 0.14). Carapace length: 0.8 - 0.9 mm 
(x = 0.9 ± 0.04). Carapace width: 0.6 - 0.8 mm (x = 0.7 ± 
0.05). Colors: Carapace yellow-orange to orange; abdomen 
grey; legs yellow-orange to orange; sternum yellow-orange 
to orange. The color seems to fade severely when specimens 
are kept in isopropanol rather than ethanol . Number of 
setae along midline of carapace (Fig. 2-1): variable, two to 
four (setae easily broken in preservation). Palp as in 
Fig. 2-3. Embolic division as in Fig. 2-4. Number of 



14 



trichobothria on palpal tibia: generally two (Fig. 2-2), 
however one male had two on the left palpal tibia and three 
on the right and another had three on the left and two on 
the right. Number of dorsal tibial spines: 1:1:1:1. 
Number of dorsal metatarsal trichobothria: 1:1:1:0. Tml : 
0.82 -0.88 (X = 0.84 ± 0.02). TibI : 6.5 - 7.7 (x = 7.0 ± 
0.35) . 

Females - (n = 13 ) . Total body length: 1.5 - 1.9 mm 
(x = 1.8 ± 0.13). Carapace length: 0.8 - 1.2 mm (x = 0.9 
± 0.11). Carapace width: 0.6 - 0.9 mm (x = 0.7 ± 0.09). 
Colors: same as males. Number of setae along midline of 
carapace: variable, two to five; females also had smaller 
setae scattered on either side of midline. Epigynum as in 
Fig. 2-5. Number of trichobothria on palpal tibia: 
generally three, however one female had two on both palps, 
three other females had three trichobothria on the left 
palpal tibia and two on the right. Number of dorsal tibial 
spines: 1:1:1:1. Number of dorsal metatarsal 
trichobothria: 1:1:1:0. Tml: 0.58 - 0.87 (x = 0.81 ± 
0.09). TibI: 6.5 - 7.9 (x = 7.1 ± 0.38). 

Diagnosis . The carapace of male M. pogonophilus most 
resembles that of M. conspectus (Fig. 2-6 from Chamberlin 
1948 and Fig. 2-1). In both species, the cephalic pits 
extend beneath the posterior median eyes (p.m.e.) whereas 
in M. arienus the cephalic pits open behind the p.m.e. The 
embolic division of male M. pogonophilus n. sp. most 
resembles M. conspectus (Fig. 2-11 from Chamberlin 1948 and 



15 



Fig. 2-4) in that both are distally bifid with the proximal 
part of the bifurcation bent forward and extending over the 
most distal part of the bifurcation. However, in M. 
pogonophilus the most distal part of the bifurcation is, 
itself, bifurcated, whereas in M. conspectus it is 
flattened (although Fig. 2-11 from Chamberlin 1948 shows it 
to be pointed) . In M. arienus the embolic division is also 
bifid, but the bifurcation begins very close to the 
tailpiece and each segment of the bifurcation is coiled 
(see Fig. 2-14 from Chamberlin 1948). The male palpal 
tibia of the new species, as with M. conspectus and M. 
arienus . is fringed laterally with long setae (Fig. 2-2). 
Chamberlin (Fig. 2-15, 1948) does not show this fringe of 
setae on his drawing of M. arienus but it is evident on the 
preserved specimens. All three species have two 
black-tipped processes on the distal edge of the palpal 
tibia (Fig. 2-2). These processes are more widely spaced 
in M. arienus than in either M. conspectus or in M. 
pogonophilus . The black-tipped process in M. conspectus is 
found on a slight ridge that extends away from the surface 
of the tibia (Fig. 2-10 from Chamberlin 1948). 
Interestingly, M. conspectus is the only one of the three 
previously described congeners whose known distribution 
extends into northern Florida. The new species can be 
separated from the congeners based primarily upon 
characters described in Table 2-1 as well as upon overall 
size; the new species being somewhat smaller than M. dux, 



16 



M. arienus , and M. conspectus which are all between 2.10 
and 2.65 mm in length according to Chamberlin (1948) and 
Gertsch and Davis (1936). 

Natural History . Masoncus poqonophilus new species, 
lives within the nest chambers of the Florida harvester 
ant, P. badius . It is about one-guarter the size of its 7 
- 9 mm long host and feeds on collembolans found throughout 
the 1 - 3 m deep subterranean nests (Porter 1985). The ant 
nest provides a stable microclimate as well as an abundant 
food source for the spider. The spiders have never been 
collected away from the ant nests and cannot survive if 
placed on a hot substrate (such as the sand outside the 
nest in the middle of the day) or if placed in a vial 
without a constant supply of moisture. They appear, 
therefore, to be obligate ant symbionts, or myrmecophiles . 

Immigration to new nest sites is common in P. badius 
(Gentry and Stiritz 1972, Golley and Gentry 1964, Gordon 
1992). While observing six such colony migrations, each 
occurring either just after a summer shower, in the early 
morning when the surface temperature was cool and the 
humidity high, or during an overcast day, I saw both adult 
and immature spiders as well as collembolans moving from 
the old colony site to the new amidst their host ants 
within the emigration trails. None of these emigrations 
was over 5 m. Neither the spiders nor the collembolans 
veered out of the emigration trails suggesting that they 
were either able to visually follow their hosts to the new 



17 



nest sites (unlikely for either of these myrmecophiles) or 
they were following a trail pheromone laid by the ants (see 
Chapter 4). Analysis of the genetic structure of spider 
populations (Chapter 3) indicated that spiders disperse 
between neighboring ant nests. 

Both sexes of M. pogonophilus build prey capture webs 
in the lab and I have seen webs inside the ant nest 
chambers. Both males and females produce sticky silk. 
Therefore, males presumably retain the aggregate and 
flagelliform glands into adulthood; most adult male 
araneoid spiders lose these glands during the terminal molt 
and cannot subseguently produce sticky silk (Kovoor 1987). 
Maintaining the ability to produce sticky silk as adults 
may be common among male erigonine Linyphiids as I have 
observed such behaviors among other (unidentified) male 
erigonines. 

Female M. pogonophilus lay one to six eggs in a 
disk-shaped eggsac deposited in a depression in the wall of 
a nest chamber (n = 9 eggsacs , x = 2.9 ± 1.5 eggs/eggsac) . 
The eggsac is flush against the surface of the chamber 
walls. Juvenile spiders molt once inside the eggsac and 
pass through three additional molts before reaching 
maturity. Since males and females do not differ 
significantly in size, both sexes probably pass through an 
egual number of developmental stages and have similar life 
spans. The longest-lived adult kept in the lab was a 
mature male that lived for three months before escaping. 



18 



He was fed mites and collembolans placed in his web. I 
succeeded in raising one spider from just after hatching 
through maturity. It was also fed collembolans and mites 
placed in its web. This spider died after 2.5 months in 
captivity, soon after maturing. If these spiders are 
representative of free-living spiders and if males and 
females do have similar lifespans, then the minimum 
lifespan of individual M. poqonophilus is between five to 
six months. 

Juveniles are present inside the ant nests during all 
months of the year (Porter 1985, pers. obs.). Porter 
reported a 4:1 female-biased sex ratio among the spiders, 
while I have found an even more extreme 7.5:1 female-biased 
ratio (n = 53). Due to the difficulty of raising spiders 
in the laboratory, it has not been possible to determine 
whether this is a primary sex-ratio bias. 



19 



U3 
































X rH 


3 






















4h 






!T> (0 





> 



























(0 C rH 


c 


rH 
















0) 










rH \ O 0) 





0) 






0) 




co 


03 




rH P 




03 






XI C 3H+J- 


en 


rH 






£ 




r-l 


a) 




3 O 




<D 






O (B (N 


TO 









a 




3 


rH 




73 




P 






73 £ '>H I 


2 


co 


en 










a to 




O 03 




•H 


03 




0) 03 03 tO CM 




D 






rC 




w 


E 




C CD 




73 


0) 




o cu 3 -h c 


P 73 


J 






P 






Of 




r-t 




C 


rH 




(0 03 rH Xi • 


3 Q) 


M 




73 


03 




rH 


UH 




03 tO 




0) 


to 




Cm 03 4H -H O 1 


03 


X 




C 


0) 




10 






3 E 






E 




03 0) -P 0) -H 


4h (0 


cu 




(0 


C 




u 


r-l 









c 


0) 




U 0) CP4H 


Xl 


o 






01 




<d 







P C 




o 


4-1 




>i o o^4h c 


4H 


2 




CO X 











a) o 










rH r-l 73 -H 0) 


0] 


o 




c 






-H 


03 




C7> 




03 


r-l 




qj a a) r-i a) 


•H 


o 




0) 


ca 




rH 


0) 




•H P 


03 


u 


o 




p a) 4-i 03 


0) 


o 




Cm 73 




CD 


H 




-P 3 


a> 


3 






tO 73 rH o ^ 


(1) c 


ft 




o 


c 




x; 


03 




0) CtrH 


a 03 




rH 0) tO (0 rH 


P 






ai 




u 


E 




03 03 


(0 


03 


0) 




Q) a-P 4h ro 0) 


x -h 


• 


p 


p 












E 




H 




73 Cm 03 ^i P !T> 


p p 


SI 


-H 


X 




o 


C 




o u 


0) 


o 


03 




-H -H 3 0) 73 — 


a 




a, 


0) 




2 


O 




2 


4-1 


2 


E 




S -P 73 03 03 0) 00 


CJ1-H 


























■* 


C P 


























!T> 1 0^ 


•H U 




















73 






^ C CNH 


P 03 






<D 














C 






U C -H o 


<0 CD 






E 










03 




to 






to O 73 E • C 


P 73 






a 




• » 




03 


CD 










rH C O •- 0) tJVH 


(0 










03 




P 


rH 




03 


X 


c 


XI 03 0) Sh (0 CP-H rH 


a 03 






43 




a) 


>i 3 


to 




a) 


o-i 


o 


(1) -P 4h -H C 4-1 U 


Q) Q) 


CO 






p 




rH 


P 


a, 


•> E 




rH 


3 




73 03 X X) -H -— 0) 


CO -H 


D 




73 


03 




(0 


0) 


03 


03 a) 




(0 


o 


-p 


0) 03 0) >i-H !h X! 





H 




C 


0) 




E 


> 




0) 4H 




E S. 


u 


U (1) (0 P 4H Q) E 


C CD 


u 




(0 


c 








,* 


rH 






P 


c 


(0 O 0) > iT tO 


•h a 


w 






01 




C 





O 03 


(0 c 




c 


r-l 


■H 


O.0 !T'i0 4-irH73X 


co • 


Cm 




co xi 




O 


p 


to a) 


E o 







(0 


-p 


03 r-i 73 O (0 0) U 


rH 03 


W 




c 










rH r-^ 










03 


a a) >i +j 


3 a) a) 


2 




a) 


01 




-P 73 X! 


c tr> 




■p 


03 


-H 


>, rH Q) OJ rH g 


4-1 X >i 


O 




a.73 




C 


0) 


E 


o c 




c 


Q) 73 r-l 


rH73rHPO03t0O 


a) 4-> a> 


U 







c 




CD 





rH 0) 


•H 




Q) 


rH 


0) 


qj a) to x to s-i s-i 


co 






01 




03 


3 


rH 4-1 


P X 




03 


03 


03 P 


03 Cm-P !T4h !T Q) 4h 


3 73 C 


• 


p 


p 




CD 73 


(0 


3 




a> 


E 


03 P 


Cm 03 -H S-l C P 


C tO 


SI 


•H 


X 




P 


0) 


E C 


ft (0 




r-l 


OJ 


a) to 


r-\ -H -H rH 3 O tO O 


p <0 -H 




CU 


0) 




Cm 


P 


03 O 


Ifl rH 




Cm 


4-1 


rH <-\ 


U-P73 03 03rHrHrH 


03 73 




























O - (1) 




























g p e 


























\ c 


03 






p 


ai 


















X c s o 


CO o p 









E 


















CJ o 


M H O 




4h 


c 


a 








C 03 










(0 X •- 0) 


CD -H 




O 






03 






o a) 




03 






rH 03 03 (0 CP 


.p CD P 






• - 


P 


a) 






rH 




a> 






X 0) 3 -H C 


O^O) 




-* 


co 


a) 


r-t 






-P to 




rH 






03 rH X -H 


tO (13 P 


CO 




O 


0) 73 


(0 






C E 




(0 






73 03 4H -H p 


P co 


p 




(C 


>i c 


E 






0) Q) 




£ 






QJ QJ P 4H Q) 


rO X 





2 




X! 


0) 


3 




03 




03 4H 






03 




CJ U Q) IT 


x 3 


a 


U 










C 


0) 




0) 




C 


0) 




tO tT>4H rH 73 


73 




H 




03 


P 


tF 


O 


rH 




r-l 73 







r-l 




Cm P 73 O f0 0) 


II 


tl 




C 





C 




03 




Cm C 






ai 




03 Cm Q) P 


1—1 • 




a) 


•H 


•H 


-P 


£ 




(0 




p 


E 




Q) Q) rH 


(B 2IO 




a p 73 


c 


0) 




a) 




c 


0) 




>lT3 rH 03 (0 


E 


• 





Ol 


C 


a> 


4-i 




-H 03 




<D 


4-1 




rH QJ tO (0 P 


■h 4-i a 


SI 




P 


a) 


03 






3 0) 




03 






Q) Cm-P 4H CP 0) 


CP 




p 


Cfl 


p 


0) 73 




73 rH 




0) 


T3 




73 Cm 03 U C P 







-H 





X 


p 


G 




o to 




r-l 


C 




•H -H -H 3 tO 


Hffl- 




CU 


a a) 


04 


to 




2 E 




Cm 


03 




S P> 73 03 rH rH 


o c • 




























x a) a> 




























a e a 
















CP 






ini 




to 


P -H >, 




u 












M C 






O 




-H 


O -P 




•H 












(0 










X 


g CD 




rH 








X3 




4-1 




rH 


QJ 




-H 


an 




(0 








C 




0) 




<-{ 


T3 




P 


co o 




x 






03 


0) 




rH O 




(0 


•H 






x 




a 






r-l 






3 P 




E 


ca 




rH 


rH rH 




CD 






3 


rH 




73 




03 






(0 


1 ^ 0) 


en 


O 


CO 




a to 




O ^i 






rH 




Cm 


CN < rH 


K 




0) 




03 


P 




c o 




X 


(0 




rH 


^- (0 


U 


4-1 


rH 






03 




•H 




P 


p 




(0 


0) E 


H 


O 


(0 




rH 


-H 




03 U 




•H 


o 




a 


H 0) 


U 




E 




(0 


T3 




3 a) 




5 


0) 






x • 4-i 


< 


c 






r-l 






O -P 










4H 


(0 co 


3 


o 


c 




a) 


03 




^ c 




03 


c 




o 


Eh 0) 0) 


•H 


•H 




U 73 




a) to 


0) 


a) 


o 






•h x 


x 


-p 






•H 


U 




CP 


> 


p 






0) 


O P 


u 


(0 


ca 




rH 


(C 




-H ^ 





•H 


S-i 




Cm 


0) 







p 




a) 


5 




P 3 


o 


73 


3 


Cm 


tO 


a c 







•H 




x 


O 




0) Cm i-l 


C 


a 


■H 


X 




UI 









rJ 


a 




u 


P 




C/3 03 


CP 


ca 


w 


P 


CO 



20 









f^-p 






C Sh 






-H CO 






T3 a 






\ c 






5 4-1 +J CD rH 4H 






C -P rO H 






X3 CD X -P CU — 




CO 




■H -P X! CD CO c/i v? 




D 




4H Sh -h -P 1 




i-l 




•H CO C T3 TS -H CN 




H 




flftOC T3 




E 




■H CO -P - CD • 




ft 




>h-h .p co co -P cr> 




O 




rH CO CO TJ -H CO -H 




a 




rH £ SH £ &H 




o 




co -H Sh co X5 Sh 




o 




-PXdSrHOSO) 




o 




C0O4hShCD-H4hCD 




eu 




•h Sh -H > J3 -H CO 






Q SiXi 4h O 3 X! w 




• 


^-^. 




si 


fid w/ 
rt of 

bent 

extending 
istal part 
tself , 

(fig. 2- 
mberlin 
it pointed 




CO 




•H CO C T3 T3 -H 4h CO 




p 




J3 0.0 C 4H X3 CO 




Eh 




•H co -P - U 3 




U 




>irH +J CO CO 




w 




HtO'TJTJOH'OS^ 




a 




rH £ SH £ <DOC0 




CO 




co -H Sh co £ Sh Sh 




z 




-PXdSSHOCGMHCO 




o 




CO 4h Sh CD -H 3 ■* 




u 




H h-h O >X CHOi 






Q ttjQ 4H O 5 CO rH rH 




SI 


n begins 
ail-piece; 
nt of 
n coiled 

from 

1948) 




co 




-P CD ^r C 

•H £ -H rH -H 




g 




-P O tP-P I rH 




w 




co -P CD co CM Sh 




H 




CO CD 




§ 




Sh 0) Sh • Xj 

3 co a 3 en e 






4H O 4h -h cC 




• 


■r-\ r-i CO -H 4H X3 




SI 


ffluaia — u 


T3 






Q) 






3 






C 






-H 






-P 






c 




e 












1 
1 




•H 
CO 
•H 

> 




co 


•H 


1 




TJ 


CM 


Eh 


o 




CJ 


•H 


CD 


rij 


rH 


l-{ 
x» 


8 




XJ 


CCS 


a 


£ 


Eh 


u 




w 



21 



T3 
(fl 

01 ja 

0J 

a • 

>ifa|4-l 

+J o 

(C - 
M P 

05 c 
a to 



fa 



c 
o 

•H 

p 
u 

QJ 



o 
u 



CM 

I 

CM 



(0 



3 

CD 
2 



O 

a) 

2 



Q) 

P 
[fl 

a) 
> 
u 

(0 

<C 

TS 
■H 

O 



01 



CM 

0) 
X! 
+J 

O 

4H 

c 
o 

•H 
P 
(fl 

c 

■H 
P 
CO 

QJ 

£ 
3 
OJ 

0] 

e 
c 

C 

o 

■H O 
P M 

(13 4-1 

e 

Sh T3 
O CD 

4-1 +J 

c u 

-H 0) 



>1 

o 
p 

01 
•H 

n 



U 

3 
01 P 



(0 

03 U 



- C 
0) (fl 
CP u 
T3 -H 
■H h 



O 
U 
01 
■H 

U 

c 
(fl 

fa 

c 

03 

en 



a) 



01 
OJ 

o 

c 

<D 
•H 

-I U 
>iZ en 
CPS 

O < 4-1 

o 

>1 

e 

Q) 
T3 
[fl 
O 



o - 
O 03 

-H 



> 

•H 
P 
03 

03 

a, 

£ 
O 

u 



c 
rH 
o 



p 

I 01 

3 

N T3 

U C 

2 H 



- CO 
Sh u 



O 



s 

D 
W 
CO 

s 



fa 

o 



2 
D 
2 



« 2 

W M 

CQ U 



o 

Eh 
U 

w 
1-1 
fa 
o 
u 



Eh 

2 

D 
O 

u 

Q 

H 
PS 

o 
fa 



NNHHHlStflt/] 


H H H K N CO 


uufafacuz2<:< 


fa fa fa g o <C 


220002200 


Q Q Q 2 g U 


< < 


< 



CM rH 



CM 



[/] 


01 


Ifl 


rn 


£ 


£ 


O 


O 


43 x: 


H H 



u3 ^ 



rH rH ^ rH ^T 



01 
03 

£ 
O 
jC 
Eh 

i3 



3 3 


> 


O O 





X! £1 


X! 


c c 


C 


S-i S-i CP CP CP CP CP 


U CP CP CP CP 


J-lS-(33CCCCC 


3 C C C C 


<D OJ Eh Eh -H -H -H -H -H 


>i Eh -H -H -H -H 


P P X! XI £ £ X! 


aj x: x! x: x: 


i-i U » * 01 01 03 01 03 


rH - 01 01 01 01 


00>i>i33333 


rH >1 3 3 3 3 


fafaOJQJOOOOO 


OJ OJ CJ u u u 


rH i— 1 


X rH 




CQ i— 1 • • • • 


QQ OJ OJUWWUW 


aj w w w w 




■ M • • • • 


COCOCOCOfafafafafa 


fa CO fa fa fa fa 



03 




03 


01 




01 


c o 




O C E E 


o o 




O O 03 03 


C C P rH C 


>i >1 >1 >1 


rH 4-) C C C >1 


O O rH re O 


> > > > 


(fl rH O P P > 


OJ OJ 03 X OJ 


OJ OJ OJ OJ 


,* rC 0J 3 3 0J 


JhlSOhl 


hJ i4 i-l i-l 


O S fa fa fa fa 



W 




















Eh 




















<J 




OJ 


«sr o o 


CM CN 


CN 


r> 'j- 




o 


ONnn^r 


Q 




CO 


co as cti 


CP CP 


CP 


cp cp 




<P 


CP CP CP CP CP 






en 


(P CP CTi 


CP CP 


CP 


CP CP 




CP 


CP CP CP CP CP 


2 




H 


rH H H 


H H 


rH 


cH H 




rH 


H rH H H rH 


O 




1 


1 1 1 


1 1 


1 


I I 




1 


1 1 1 1 1 


H 




H 


M M H 


> > 


X 


H X 




H 


H > H H X 


Eh 




X 


X 






H H 


01 




H M H 


a 














a) 




> 


w 


01 


1 


1 1 1 


1 1 


1 


1 1 


rH 


1 


1 1 1 1 1 


fa 


a) 


o 


cp m ^ 


CP <D 


CO 


O IT) 


03 


CO 


n cp co to id 


fa 


rH 


rH 


rH H 


CN 




CM CM 


E 




H CM CM CM 


o 


rD 












OJ 






u 


2 












fa 









s 


cu 


•H 


> 


rH 


(0 H 


CO 0(0 


iM -H JM 


rH +J 


T3 O C 


£ ja cu 


cu g > 


-H CU 


O - > 


(0 5 - g 


aniH^ 3 


(0 -H (0 c 


^ > fc .» >, 


(0 Q) — IT 


-H P g -H 


to rc g a< 


d) CO rH <u 


rl ^ 0(N 


(0 O i-i • - 


£T3 ftoin 




H (I] 10 ~ 


3 3 0) E 


a jxi-h g 


■ H H (d 


CO (0 (0 O rH 


cu & a to • 


•H -^ O 


o <d <u 


rH rH — || 


a to to >i 


en s e h a) 


rH rH 


S P - (0 (0 


o) iM n p o 


G CU WW 


rH .„ .H >_ 


CO 


— TJ 


a 


<+H g S 


rH 


o g CU CU 


■H 


r-i "H 


jC 


to cm ja > 


a 


rH • -H 





rH O tO rH 


c 


CU -H (0 


o 


4J II > U 


CP 


(0 P 





a cu p c 


a 


rH to cu 


T3 (0 3 > 


CO 


c T-i 


=1 


(0 to to 


o 


^- c cu 


c 


(0 O g 





•H C0 -H 


Cfl 


n p to ■» 


n3 


•H -H -H tO 


2 


p a > 3 


•H & 


- in T3 rH 


• OJ (C (0 


lO H O ft 


1 — 3 -H 


(Mr,UH OJ 


£ H rH 


O g -H 42 (0 


p o g g 


<tf Q) 


rH • C -P — 


1 O --H Mh <+H g 


cm CU g 


II (0 rH 


CO -H C rH 


CU CU iM <H • 


iM rH £ -H O 


3 (0 P P 


CP O (0 C II 


■h en Xi o o 


fn — ' o u -h cu 


£ 3 tOH 


S CJIH-H (0 


CU -H -rH > 


•h u ja -h to 




> p ^T3 — 



23 




^f 




Figures 2-6 to 2-16 (from Chamberlin 1948). 6, 
Masoncus noaales new species. Cephalothorax, dorsal view. 
7, Masoncus nogales new species. Cephalothorax and 
chelicerae, lateral view. 8, Masoncus nogales new 
species. Epigynum. 9, Masoncus nogales new species. Male 
palpus, lateral view. 10, Masoncus nogales new species. 
Patella and tibia of male palpus. 11, Masoncus nogales 
new species. Mesoventrai view. 12, Masoncus arienus new 
species. Cephalothorax, dorsal view. 13, Masoncus 
arienus new species. Epigynum. 14, Masoncus arienus new 
species. Male palpus, ventral view. 15, Masoncus arienus 
new species. Patella and tibia of male palpus, dorsal 
view. 16, Masoncus dux new species. Epigynum. 



25 





(from Chamberlin 1948) 



CHAPTER 3 
GENETIC DIVERSITY WITHIN AND AMONG POPULATIONS OF 
MASONCUS POGONOPHILUS USING RANDOM AMPLIFIED 
POLYMORPHIC DNA (RAPD) FINGERPRINTING 

Introduction 

Many arthropods have evolved close symbiotic 

relationships with ants. These are referred to as 

myrmecophiles . Some are found at the periphery of the 

nest, either near the entrances or on the midden (refuse 

pile); others are found within the chambers of the nest, 

either in the peripheral chambers or deeper in the nest in 

the brood and storage chambers (Holldobler 1977). In 

general, ants live in complex closed societies in which 

only members are allowed. They communicate with nestmates 

through chemical and tactile signals, and tend to 

aggressively exclude intruders. Much work has been done 

studying the various adaptations myrmecophiles use to 

integrate into the hostile environment of an ant nest 

(reviewed in Kistner 1979 and Holldobler and Wilson 1990). 

However, little has been done to investigate the influence 

host population structure has on that of its guests, though 

Kistner (1982) proposed that the spacing and abundance of 

host nests in an environment plays an important role in the 

presence, abundance and population structure of 

myrmecophiles . 



26 



27 



My objective was to investigate the population 
structure of a myrmecophilic spider by examining genetic 
diversity within and among different populations and to 
determine the extent to which the distribution of the host 
ant nests influences the population structure of the 
myrmecophile. The spider Masoncus pogonophilus Cushing 
(Linyphiidae) (Chapter 2) lives within the nest chambers of 
the Florida Harvester ant, Pogonomyrmex badius (Latreille) 
(Hymenoptera: Formicidae). It is about one-guarter the size 
of its 7-9 mm long host and feeds on collembolans found 
throughout the 1-3 m deep subterranean nests (Porter 1985). 
The ant nest provides a stable microclimate as well as an 
abundant food source for the spider. The spiders have only 
been collected outside the ant nests when the hosts are 
migrating to a new nest site. The spiders are extremely 
susceptible to desiccation when removed from the nests. 
Therefore, they appear to be obligate ant symbionts. 

The following features of the natural history of the 
myrmecophile and its host suggest that inbreeding among 
spiders within colonies may be high (and, thus, genetic 
diversity low) and that dispersal of spiders between 
neighboring colonies may be low (Chapter 2): 1) The spider 
exhibits an extreme female-biased sex ratio ranging from 
4:1 (Porter 1985) to 7.5:1 (Chapter 2). Female-biased sex 
ratios among diploid organisms are often associated with a 
high level of inbreeding (Hamilton 1967, Rowell and Main 
1992). A female-biased sex ratio in such systems reduces 



28 



local mate competition between closely related males. 2) 
The host ants aggressively defend their territory against 
conspecifics from other colonies (Gentry 1974, Holldobler 
1976). Pogonomyrmex badius nests tend, therefore, to be 
overdispersed or evenly spaced in the landscape (Harrison 
and Gentry 1981, Chapter 6). Colonies are usually spaced 
between 8-16 m from one another (Harrison and Gentry 1981, 
pers. obs.). The high dispersion of P. badius nests coupled 
with the spiders' susceptibility to desiccation outside the 
nest would make dispersal of spiders a high risk activity 
during the daytime in the xeric environments in which their 
host is found. 3) Dispersal to or from nests at night is 
prevented by the host's habit of closing the nest entrance 
at that time. Williams and Franks (1988) suggested that a 
myrmecophilic isopod (also with a female-biased sex ratio) 
may remain within a host nest for several generations until 
the nest senesces and dies, at which time the isopods pulse 
out into the environment and become established in new 
nests. Queens of P. badius can live at least 15 years and 
some western congeners can live up to 30 years (Gentry 
1974, Porter and Jorgensen 1988). It is possible, 
therefore, that a population of M. pogonophilus . remains 
inside a single colony for several generations dispersing 
to a new colony only when the host nest begins to senesce. 

However, M. pogonophilus are consistently found 
within mature P. badius nests in a given habitat so 
dispersal events, although perhaps infreguent, must be 



29 



occurring. Each P. badius nest is established by a single 
inseminated queen; the colonies do not split into new 
colonies (Cole 1968). Therefore, unless the spiders are 
phoretic, hitching rides on the bodies of their hosts (an 
unlikely dispersal mechanism for reasons explained in 
Chapter 4), they must be finding their way to new colonies 
by other means — perhaps by eavesdropping on the chemical 
signals of their host ants and following trail pheromones 
(see Chapter 4). Nevertheless, dispersal events among the 
myrmecophiles could be infrequent (i.e., occurring only 
when the host nest senesces) yet still maintain a 
relatively high occurrence of spiders among nests as 
spiders dispersing from the dying host nest find their way 
to new colonies. Given the potential lifespan of a single 
colony and the short generation time of spiders (Chapter 
2), inbreeding among the spiders should be high. 

To test the hypothesis that dispersal events of 
myrmecophilic spiders between neighboring nests are 
infrequent and that genetic diversity within nests is low 
(perhaps due to inbreeding), I used the PCR-based Random 
Amplified Polymorphic DNA, or RAPD's, technique (Williams 
et al. 1990, Welsh and McClelland 1990, Hadrys et al . 1992) 
to measure the genetic diversity among spiders within P. 
badius nests as well as among spiders from different nests. 

If dispersal of spiders between nests is infrequent 
and inbreeding among spiders within each nest is high, then 
I would expect genetic diversity between individuals within 



30 



each nest (i.e., within each population) to be low. 
Furthermore, I would expect genetic diversity between 
spider populations from neighboring nests as well as 
between populations from distant nests to be approximately 
equal and to account for a greater percentage of total 
genetic diversity than within-population differences. 

Materials and Methods 
Population Sampling 

In June 1993, three P. badius nests were excavated at 
two sites in north Florida. Nests 1 and 2, approximately 
12 m apart, were located at Archer Sandhills (ASH1 and 
ASH2 , respectively) 25 km west of Gainesville, FL in Levy 
County. Nest 3 was located 55 km east of Archer Sandhills 
at the Katherine Ordway Preserve/ Swisher Memorial 
Sanctuary (0RD3) in Putnam County FL. Approximately 1 m 
deep excavation pits were dug adjacent to the nest 
entrance. The soil was scraped away from the pit wall to 
expose the nest chambers and the spiders within. Fifteen 
adult spiders were collected from ASH1 , 16 from ASH2 , and 9 
from 0RD3 . 
DNA Isolation 

Spiders were placed in vials and chilled at -80 °C 
for a few seconds to kill them. Each was washed with 20 pi 
of sterile STE buffer (pH 8.0) ( Sambrook et al. 1989) to 
remove any sand grains or debris, then transferred to a new 
sterile microcentrifuge tube and homogenized with the 
rounded end of a sterilized glass pipette in 50 ul of cold 



31 



(1 - 2 °C) STE buffer. Tubes were incubated overnight (15.5 
hrs.) at 55 °C with 2.5 ul 20% SDS and 2.5 ul of Proteinase 
K (50 ug/ml) (Sambrook et al . 1989). Samples were then 
extracted with phenol: chloroform (50:50) and the DNA 
precipitated with 95% ethanol. DNA was resuspended in 20 
pi of 0.1 X TE (pH 8.0) (Sambrook et al. 1989) and 
concentration determined spectrophotometrically . Each 
sample was diluted with buffer to give a final DNA 
concentration of 5 ng/ul. 
DNA Amplification 

A Perkin Elmer Cetus DNA Thermocycler was used for 
DNA amplification. The cycling protocol was 1 min. at 94 
°C; 1 min. at 50°C; and 2 min. at 72 °C for 45 cycles. Each 
reaction was carried out in a total volume of 50 ul 
containing 0.5 X Stoffel buffer (Perkin Elmer Cetus), 100 
uM of each dNTP, 1.75 mM MgCl 2 , 0.05 uM of primer, 0.05 
U/ul of ampliTag DNA polymerase Stoffel fragment (Perkin 
Elmer Cetus), and 5 ng/ul of template DNA. DNA 
amplification bands were separated in 1.2% agarose gels in 
1.0 X TBE buffer (Sambrook et al. 1989). Bands were 
visualized under UV light after staining with ethidium 
bromide at a final concentration of 0.5%. 

Eight random 10-mer primers (DNA Synthesis Lab, 
Gainesville, FL) were screened (Table 3-1). Each 
amplification with one primer was replicated three times. 
Representative products from ASH1 , ASH2 , and 0RD3 were run 
side by side. Negative controls containing all the reagents 



32 



except the template DNA for each primer were also conducted 
to ensure the fidelity of the results. For each spider, 
bands were scored as present ( 1 ) or absent ( ) . RAPD bands 
generated by the 8 primers that were consistently 
reproduced in at least two replicate PCR reactions were 
counted in the final analysis. 
Statistical Analysis 

The Analysis of Molecular Variance technique (AMOVA) 
was used to analyze genetic diversity within and among the 
three populations (Excoffier et al. 1992, Huff et al. 
1993). This analysis was designed to handle different 
types of molecular data and uses no a priori assumptions 
regarding gene flow or population structure (Excoffier et 
al. 1992). Although first used to analyze mitochondrial 
DNA haplotype data, it has since been applied to RAPD data 
(Huff et al. 1993). The analysis is based upon pairwise 
comparisons of banding patterns between all 35 spiders. 
These genetic distances were expressed as Relative Band 
Distances = 100 * [(# different bands)/ (total # bands)] 
for each pair. 

Results 

Three of the 15 ASH1 , two of the 16 ASH2, and one of 
the nine 0RD3 spiders were males. All 15 of the ASH1 
spiders, 11 of the 16 ASH2 spiders (including one male), 
and all nine of the ORD3 spiders were used for the final 
analysis. The depths of the chambers or runways from which 
the spiders were drawn are shown in Table 3-2. 



33 



Each spider yielded between 35 - 300 ng/ul of total 
DNA (x + s.d.: 116.3 ± 53.7). The individual primers 
yielded between 5 to 19 polymorphic bands (Table 3-1). 
Primer #3 resulted in three different sets of individual 
banding patterns for the three replications. Since no 
individual showed consistent banding patterns at least 
twice, I excluded these data from the analysis. Primer #5 
resulted in a monomorphic pattern across all three 
populations. Since this provided no useful information 
about population subdivision (i.e., a highly conserved 
region of the genome was amplified) , I also excluded these 
data from the final analysis. Six of the eight primers 
screened yielded a total of 67 bands. Of these, 14 were 
monomorphic for all 3 5 spiders and were excluded from the 
final analysis. Of the 53 polymorphic bands, 10 (19%) were 
unigue to ORD3 . Six bands (11%) were found in both ASH 
populations but were absent in the ORD3 population. Thus, 
16 bands (30%) distinguish the populations at the ASH site 
from the population at the ORD site. In contrast, only two 
bands (4%) were unigue to ASH1 and only three bands (6%) 
were unigue to ASH2. One band was present in one 
individual from ASH2 and three individuals from 0RD3 . 

The relative band differences between spiders within 
each population as well as between individuals from 
different populations are presented in Tables 3-5 to 3-10. 
A summary of the pairwise intra-nest and inter-nest 
relative band distances are presented in Table 3-3. An 



34 



unbiased estimate of the standard error of the mean, 
corrected for the nonindependence of pairwise comparisons, 
was calculated based upon the formula in Miyamoto et al. 
(1994) as modified from Lynch (1990): 

SE = 100 ( (2D[1-D] [1+5] ) / (n[3+D])) °' 5 . 
In this equation, 5 equals mean relative band distances for 
all possible pairs in the analysis and n refers to the 
average number of scored bands per individual. 

The AMOVA results (Table 3-4) indicate significant 
genetic differences between the ASH and the ORD sites (p < 
0.005) as well as between populations at the ASH site (p < 
0.04). There is also significant genetic diversity within 
each of the three populations (p < 0.005). Of the total 
genetic diversity, 77.4% was due to individual differences 
within the three populations; 18.2% was due to differences 
between ASH and ORD; and 4.4% was due to differences 
between ASH1 and ASH2 (i.e., within their region). 

To further segregate the patterns of genetic 
differences among the three populations, three separate 
pairwise comparisons of the populations were conducted 
(Table 3-4). For all of these comparisons, the largest 
component of genetic diversity is attributable to within 
population differences (75 -94%, p < 0.03). Although 
between 19 - 24% of the total genetic diversity is 
attributable to differences between distant nests, only 
about 6% of the total genetic diversity is attributable to 
differences between the neighboring nests. 



35 



Discussion 

All three nests show high intra-nest genetic 
diversity. Only 14 of the 67 total bands (or 20.9%) were 
monomorphic for all three populations. The variation among 
the remaining 53 polymorphic bands results more from 
intra-nest genetic diversity rather than inter-nest 
diversity. However, genetic diversity between either of the 
ASH populations and the ORD population is higher than the 
genetic diversity between the ASH1 and ASH2 populations. 
This suggests that gene flow may be great enough to offset 
the diversifying effects of genetic drift between 
neighboring nests in contrast to the geographically 
separated populations (Slatkin 1994). 

If the neighboring populations had been separated for 
many generations and if one or both resulted from a single 
foundress (perhaps arriving from a neighboring colony) , 
then the foundress effect and genetic drift should have 
decreased intra-population genetic diversity and increased 
inter-population genetic differences. Instead, these data 
indicate that dispersal events of spiders between 
neighboring nests in the same habitat are occurring at a 
significant rate within the lifetime of an ant colony. 
Spiders and ants may take advantage of cool mornings in the 
winter or periods after summer showers to disperse. 
Spiders migrating with host ants to a new nest site may 
wander off the emigration trail and find their way to a new 
host nest instead. It may be that dispersing spiders 



36 



follow foraging or orientation trail pheromones to the edge 
of the hosts' territory and then search for the trail of a 
neighboring colony which they then follow until they get to 
the new nest. The existence of chemical trails for 
foraging, recruitment, and homing has been well documented 
in various species of Pogonomyrmex ants including P. badius 
(Holldobler 1971, Holldobler and Wilson 1970, Regnier et 
al. 1973). Observations of spiders emigrating with their 
hosts or moving from the periphery of a foreign P. badius 
mound directly to the mound entrance suggest that spiders 
may have evolved the capacity to follow chemical trails of 
the ants (see Chapter 4). Dispersal to new ant nests may be 
a mechanism for avoiding inbreeding depression. Or, it may 
be triggered by conditions, such as increased resource 
competition, within the nest. Since the ecosystem in which 
spiders live is destined to go extinct upon the death of 
the queen ant, it may be adaptive for some individuals in 
the population to risk the hazards associated with 
dispersal in order to locate potentially younger ecosystems 
(i.e., ant nests with a relatively longer life expectancy 
due to the presence of a younger gueen) . 

If it is primarily male M. pogonophilus that disperse 
to other colonies, then high mortality of the dispersers 
when they venture into the xeric environment in search of a 
new colony 10-15 m away would explain the apparent 
female-biased sex ratio. However, a model proposed by 
Bulmer and Taylor (1980) suggests that the sex ratio should 



37 



be biased in favor of the dispersing sex. Because only 
five of the spiders used for the present study were males, 
I could not determine whether there were significant 
differences in banding patterns between the males and 
females within a single nest. It is common among male 
web-building spiders for mature males to wander about in 
search of females. Such wandering often leads to apparent 
sex ratio bias due to high mortality among the males 
(Vollrath and Parker 1992). However, in such species, the 
sexes are usually dimorphic, the male maturing at a much 
smaller size and after fewer molts than the female 
(Vollrath and Parker 1992), and the males are unable to 
build prey capture webs (Kovoor 1987). Neither trait holds 
true for M. pogonophilus males (see Chapter 2). 

If, however, the female-biased ratio is real and not 
an artifact of differential mortality, then the traditional 
explanation of such ratios — high inbreeding leading to 
reduced local mate competition through production of fewer 
sons than daughters — is not supported since the RAPD's data 
do not indicate high levels of genetic relatedness within 
spider populations. Instead, this system may fit a model 
proposed by Colwell (1981, 1982) and Wilson and Colwell 
(1981) in which female-biased sex ratios are established 
regardless of the level of inbreeding or local mate 
competition. Colwell shows that in a sub-population made 
up both of females that skew the sex ratio of their 
offspring towards daughters, "Hamiltonian females", as well 



38 



as females that produce an equal number of sons and 
daughters, "Fisherian females", the Fisherian females will 
have greater fitness within a sub-population (in this case, 
within a single ant nest) but Hamiltonian females will have 
greater fitness at the larger scale of the population as a 
whole because they produce a greater number of dispersing 
daughters or foundresses. The greater the number of 
Hamiltonian daughters dispersing to establish new 
sub-populations, the greater the frequency of Hamiltonian 
females even within any given sub-population (within an ant 
nest). Therefore, the female-biased sex ratio observed 
among M. pogonophilus may be the result, not of inbreeding 
and local mate competition, but of the greater fitness of 
"Hamiltonian females" due to the dispersal patterns of 
spiders between ant nests . 



39 



Table 3-1. Eight 10-mer primers screened. Primers 
1, 2, 4, 6, 7, and 8 generated the 53 polymorphic bands 
used in the final analysis. 



PRIMER 


NUCLEOTIDE 


NUMBER OF 




SEQUENCE 


POLYMORPHIC 




5' to 3' 


BANDS 


1 


CTGAAGCGGA 


19 


2 


ATCAAGCTGC 


5 


3 


AGCTGAAGAG 


* 


4 


GCCCTGATAT 


5 


5 


CAGGACATCG 


* 


6 


ACAGGGAACG 


10 


7 


GACCCAGAAG 


5 


8 


CGACCAGAGC 


9 



* The bands resulting from amplification 
with primers 3 and 5 were excluded from 
the data set (see text). 



40 



Table 3-2. Depths of the subterranean chambers or 
runways from which the spiders used in the analysis were 
drawn. Numbers in parenthesis indicate how many of the 
spiders collected from that location were males. 



Nest 


Depth 


of Chamber 


#Spiders 




or Runway (cm) 


(#Males) 


ASH1 




30.5 


1 


ASH1 




39.0 


2 


ASH1 




46.5 


3(1) 


ASH1 




55.0 


1 


ASH1 




55.0 


1 


ASH1 




70.0 


3 


ASH1 




71.0 


1 


ASH1 




94.0 


KD 


ASH1 


bottom of exc. pit 


2(1) 


ASH2 




23.0 


KD 


ASH2 


35 


.0 - 41.0 


8 


ASH2 




50.0 


1 


ASH2 




54.0 


1 


ORD3 




3.5 


1 


ORD3 




13.0 


KD 


ORD3 




30.5 


1 


ORD3 




37.0 


1 


ORD3 




37.5 


3 


ORD3 




47.0 


1 


ORD3 


80 


.0 - 84.0 


1 



41 



Table 3-3. Summary of means + standard errors and 
ranges for the intra- and inter-population Relative Band 
Distances . 

POPULATION COMPARISON #PAIRS X ± SE RANGE 



I Within Populations 














Intra-ASHl 


105 


29, 


.56+8, 


.24 


0.00-70, 


.00 


Intra-ASH2 


55 


35, 


. 40±9. 


.03 


5.26-78, 


.26 


Intra-ORD3 


36 


35, 


,92±8, 


.54 


3.57-60, 


.00 


II Between Populations 














ASH1 vs ASH2 


165 


34, 


. 38±8, 


.75 


0.00-75, 


.00 


ASH1 vs 0RD3 


120 


43, 


. 83±9, 


.10 


19.05-75, 


.00 


ASH2 vs 0RD3 


99 


44, 


, 41±9, 


.30 


15.79-76, 


,00 



42 



73 4H 






Q) 




* 


01 U tfl 




* 


H (8 cu g 




w 


ro a w o 3 




D 


3 £ (8 C CO 




J 


73 3 <D 




< 


■HO M II 




> 


> >i C CD 




1 


-H — »P 0) 73 4-1 Q 




(X 


73 < -H 0) 0) 4-1 CO 






C •— CQ 3 g -H CO 






-H U 4-1 Sh 73 






<C CD 0) 




* 


\D> >UIH O • 




rJ 


(nj o -h h p w 




< 


2 73 73 CD C C 




o\° H 


CD < C QrH 




o 


J3 O (0 -H 




E-i 


P 0) -H CD >iP 






C P - S-l 4-> (0 






H O CD W CD -H rH 






O CCS013 






4-i (DO Sh a 




Eh 


• tP-H ^- Q) o 




w a 


0) <D CTQ > G* 




u w 


P > CD CD -H 




s s 


rH Sh X! S-i 1 73 C 




< o 


3 CD E-» -H 




H Ph 


WW Cffl U£ 




§ o 


Q) Q) -H v_ -H P 




Sh Sh • XI 4-1 -H 




> u 


Ph >iP to O S 






— » ro -h •» C 






< >i 5 3 < Q) W 






> (0 73 > &irH 






O 3 Sh W O ro 






2 73 O C 2 CU 3 




Q 


< H O < X! 73 


w 


co 


— ' O 73 -H P -H 


cu 


2 


C -P rH > 


o 




(1) g ro rC ro rji-H 


c 




O O rH C C 73 


cu 




C U W 3 O -H C 


S-i 




ro 4-1 rH a-H C -H 


cu 




•H rH O P O 


4-1 




s-i w -h a-H -h c 


4-1 


Q 


ro rH jC 73 P CU 


•H 


CO 


> rfl 73 C 73 -h CU 


Q 


CO 


3 C CU ro P 3 






U 73 ro <D S-i P 73 




ro -H en 3 CD ro CD 


cu 




rH > P CD axi 


Sh 




3 -H Sh <D Sh 


ro 




O 73 CU a Xi - W 


3 




a) c x! E-" w cd 


CF 


4-1 


rH -H U - CO 


CO 


73 


O S-i w o c 






2 en <c c ■ -H d) 


c 




O W -P S-i 


ro 




4-i CD C -H C ro 0) 


0) 




ox: cu en o rH q-i 


2 




P a; cu -h 3 4-i 






w 3 sh p a-H 


II 




-H 73 P ro O 73 






w c cu c rH a 


Q 




>, ro Xi CU 3 73 


co 




rH cu a 4-i c 2 




ro W >i 3 O O ro 






c rH p p a 


• » 




< H -H 0) S-I W 


w 


is 


•H W Xi C -H C 


cu 


o 


jC S-i -H ro O 





H 


• 73 cu o x! a-H 


c 


Eh 


^r c > P P P 


cu 


< 


1 ro -H c -H X! ro 


S-l 


H 


ro CO 73 -H 3 O -H 


cu 


3 


r0 3 4-i 


CU S-i O 73 W CU a 4-i 


> 


H <D -H CU rH O 


•H 




X! x! p c ro tr aa 


Pn 


rC O CU O 3 G 




O 


Eh Sh C -H 73 -H C 73 




< CU P -H S-i O 


cu 


w 


cp-h > r0 cu 


S-i 


u 


g p -h a 3 


ro 


« 


O CU Sh 73 g P 


3 


5 


s-i a ro c o cu 


tr 


o 


4h p a-H o xi 


CO 


CO 



IT) 




un 






IT) 




ID 






O 


"* 


o 


ro 




O 




O 






o 


o 


o 


O 




O 




O 






• 


• 


• 


• 




• 




• 






o 


o 


o 


o 




o 




O 






V 


V 


V 


V 




V 




V 






rH 


«tf 


>x> 


o 


o 


kO 


"* 


•3* 


IX) 




CN 


^r 


ro 


co 


CN 


CN 


r» 


r> 


CM 




00 


^ 


r^ 


\f) 


^J" 


-* 


m 


CP 


o 




H 




r^ 




CTi 


CM 


r^ 


H 


CO 




CO 


in 


CTl 


co 


O 


O 


>* 


CO 


CM 




CO 


en 


<* 


en 


o 


rH 


cr> 


CO 


CO 




ro 


o 


>X> 


o 


IS 


in 


LO 


<tf 


r^ 


P 






H 




■H 




H 




rH 


C 
CU 

c 
o 
a 
g 
o 


<tf 


CTi 


CP 


CP 


O 


IS 


■* 


CO 


CM 





CTi 


■* 


^ 


-* 


O 


ro 


CP 


H 


CO 


X! 


r> 


CO 


>& 


co 


VD 


ro 


ID 


iH 


t> 


o 


r^ 


CN 


r-l 


CM 


rH 


[^ 


H 


^ 


H 


ro 

CU 

>1 

XI 


*# 


CTi 


O 


en 


ro 


IX) 


O 


co 


p> 




en 


■3* 


>£) 


-* 


CP 


ro 


VO 


H 


lO 


73 


r^ 


CO 


r~- 


co 


CO 


ro 


o 


H 


o 


P 


r> 


CM 


CM 


CN 


CO 


t*v 


in 


iX> 


CM 


3 






Lfl 




CO 




CO 




CO 


X! 

-H 

Sh 

p 

c 


iH 


rH 


CN 


H 


<tf 


^ 


CN 


rH 


CO 


O 






ro 




CN] 




CN 




r-\ 




CU 


c 

rO 






W 




w 




W 




W 


•H 






C 




c 




C 




C 


S-l 






o 









o 







ro 






•H 




•H 




•H 




■H 


> 




W 


P 




-P 




P 




P 






c 


ro 




ro 




ro 




ro 


rH 




o 


rH 




r-i 




rH 




H 


ro 




-H 


3 




3 




3 




3 


P 




CP 


a 




a 




a 




a 


O 




cu 


o 




o 












P 




S-i 


a 




a 




a 




a 






\ 


\ 


CN] 


\ 


ro 


\ 


ro 


\ 


4H 




w 


w 


a 


w 


Q 


w 


Q 


w 


o 


Q 


c 


rH 


CO 


rH 


« 


rH 


« 


rH 




P4 


o 


(0 


< 


(0 


O 


ro 


o 


(0 


cu 


O 


-H 


3 




3 




3 




3 


CP 




p 


T3 


w 


73 


w 


73 


w 


73 


ro 


w 


ro 


•H 


> 


•H 


> 


•H 


> 


-H 


P 


> 


rH 


> 




> 




> 




> 


c 




3 


•H 


H 


•H 


H 


•H 


CM 


■H 


cu 


X 


a 


73 


X 


73 


K 


73 


X 


73 





CO 


o 


C 


CO 


C 


CO 


C 


CO 


C 


S-i 


< 


Cm 


H 


< 


M 


< 


H 


rffl 


H 


cu 

Cm 


. 






# 




• 




• 






< 






rx 




u 




Q 




* 



cu 
o 
c 

ro 

X! 
O 

>m 
Xi 

p 
c 
cu 
c 
o 
a 
g 
o 
o 

cu 
o 

c 

(0 

-H 
S-i 
ro 

> 

CO 

g 

<D 

Sh 
P 
X 
<D 

<D 

Sh 
O 

g 

ro 
CP 

c 

•H 

c 

•H 

ro 
P 
XI 

o 

o 

>1 
p 



XI 
ro 
XI 
O 

Sh 
CU 



* 

-K 



43 



Table 3-5. Relative band distances between all pairs 
of M. pogonophilus individuals from ASH1 . The individual 
identification numbers are listed along the horizontal and 
vertical axes. 



1 

2 

3 

4 

.5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 



3 



8 9 10 11 12 13 14 15 





53 

52 19 

50 16 10 

52 19 10 

57 20 23 27 23 

52 15 5 23 

50 19 13 16 13 10 14 

23 42 42 33 42 45 40 37 

52 29 13 16 13 14 10 13 37 

58 44 39 41 39 33 29 33 42 39 

70 40 41 45 41 21 41 30 58 33 52 

55 26 15 12 15 10 14 9 38 9 39 29 

52 24 13 16 13 30 10 19 37 24 39 48 21 

44 36 31 28 31 29 32 26 33 26 47 40 22 31 



44 



Table 3-6. Relative band distances between all pairs 
of M. pogonophilus individuals from ASH2. The individual 
identification numbers are listed along the horizontal and 
vertical axes. 



16 
17 
18 
19 
20 
21 
22 
23 
24 
25 
26 



16 17 18 19 20 21 22 23 24 25 26 


64 
32 56 
43 62 32 
38 60 35 30 
38 69 35 27 27 
32 47 29 15 10 10 
35 64 32 27 38 13 5 
22 70 20 37 38 24 11 24 

30 70 27 24 34 26 19 26 16 

31 78 45 28 45 41 28 45 41 42 



45 



Table 3-7. Relative band distances between all pairs 
of M. pogonophilus individuals from 0RD3 . The individual 
identification numbers are listed along the horizontal and 
vertical axes. 



27 

28 
29 
30 
31 
32 
33 
34 
35 



27 28 29 30 31 32 33 34 35 


50 
31 57 
56 39 35 
39 58 25 42 
27 49 24 46 14 
29 50 27 47 17 4 
33 50 25 47 28 21 23 
50 60 32 44 22 33 30 29 



46 



Table 3-8. Relative band distances between all pairs 
of M. pogonophilus individuals from ASH1 and ASH2 . The 
individual identification numbers are listed along the 
horizontal and vertical axes. 



16 
17 
18 
19 
20 
21 
22 
23 
24 
25 
26 



7 8 9 10 11 12 13 14 15 



44 28 38 35 38 26 38 26 38 30 50 39 33 38 18 

40 70 64 65 64 59 50 65 63 64 62 75 68 69 70 

55 25 35 32 35 24 35 24 50 27 45 43 30 35 41 

50 22 3 13 3 27 5 16 45 16 42 45 18 16 34 

52 42 27 34 27 23 38 52 32 43 41 38 32 45 

52 34 24 26 24 23 34 52 29 44 41 31 29 50 

47 6 10 5 10 24 10 15 33 19 30 43 23 19 33 

50 29 24 21 24 27 5 34 47 29 48 45 31 29 46 

47 34 34 26 34 15 19 24 42 29 44 35 26 34 41 

52 31 21 18 21 14 10 16 44 16 41 33 12 26 33 

40 35 30 32 30 48 35 35 50 41 57 60 41 35 38 



47 



Table 3-9. Relative band distances between all pairs 
of M. pogonophilus individuals from ASH1 and 0RD3. The 
individual identification numbers are listed along the 
horizontal and vertical axes. 



27 

28 
29 
30 
31 
32 
33 
34 
35 



7 8 9 10 11 12 13 14 15 



70 46 36 38 36 42 30 46 59 36 56 58 42 41 57 

71 66 55 60 55 61 52 59 65 55 62 75 59 59 66 
62 29 23 26 23 32 19 34 43 29 45 50 35 29 37 
65 53 44 46 44 44 40 44 58 40 55 60 45 49 42 
53 31 26 23 26 36 24 31 39 26 47 55 32 31 39 
64 35 25 22 25 42 23 35 43 25 50 58 31 30 42 
64 37 27 24 27 42 23 37 45 27 52 58 33 32 44 
59 41 31 33 31 44 26 41 50 36 52 60 42 31 44 
55 42 38 39 38 39 27 42 52 42 53 57 47 42 45 



48 



Table 3-10. Relative band distances between all 
pairs of M. poqonophilus individuals from ASH2 and 0RD3 . 
The individual identification numbers are listed along the 
horizontal and vertical axes. 



27 
28 
29 
30 
31 
32 
33 
34 
35 



16 17 18 19 20 21 22 23 24 25 26 
57 75 39 34 40 36 39 41 46 38 53 
74 76 64 54 60 59 59 62 66 60 67 
40 64 29 27 32 39 20 34 44 41 47 
45 68 42 47 53 53 42 49 53 50 56 
37 61 41 29 40 36 16 31 36 38 38 
50 67 46 28 39 35 24 30 40 37 47 
50 68 46 30 41 37 24 32 42 39 48 
52 57 35 29 34 41 27 36 46 43 48 
40 57 43 41 36 47 20 42 47 49 50 



CHAPTER 4 
THE ABILITY OF THE MYRMECOPHILIC SPIDER TO FOLLOW THE 

TRAILS OF ITS HOST ANT 

Introduction 

Masoncus pogonophilus Cushing (Linyphiidae; 

Erigoninae) is a common symbiont within the nest chambers 

of the Florida seed harvesting ant, Pogonomyrmex badius 

(Latreille) (Porter 1985). Adult spiders are 1.8 mm in 

length (Chapter 2), about one-third the size of the 

formicid host. All developmental stages of the spider, as 

well as the eggsacs of M. pogonophilus can be found 

throughout the chambers of the subterranean nest (Chapter 

2). 

A study by Cushing (Chapter 3) using the Random 

Amplified Polymorphic DNA (RAPD) fingerprinting technigue 

indicated that spiders from neighboring P. badius colonies 

were genetically similar. This finding suggested that 

dispersal of these small spiders between nests was more 

frequent than expected. This study was an attempt to 

determine the mechanism by which spiders disperse. 

Dispersal among many species in the family 

Linyphiidae, especially among the tiny spiders in the 

subfamily Erigoninae, is by ballooning. Juvenile and adult 

spiders climb up to a high point — a blade of grass or a 

fence post — and release a strand of silk (Gertsch 1979). 

49 



50 



The silk acts as a kite, lifting the spiderling into the 
air. Species in the family Araneidae that disperse by 
ballooning normally produce large numbers of 
offspring — more than 1000 per female (Tolbert 1977, 
Reichert and Gillespie 1986). This high fecundity 
presumably offsets the risks of this dispersal strategy. 
However, M. pogonophilus produces, at most, six eggs per 
eggsac with a mean of 2.9 eggs (Chapter 2). Most araneoid 
species that produce multiple clutches, produce no more 
than five (Vollrath 1987). If M. pogonophilus females 
produce multiple clutches, their lifetime reproductive 
output would still only average 15 offspring. Such low 
fecundity is common among the tiny erigonine spiders of the 
family Linyphiidae (Bristowe 1958, Roberts 1995). In 
general, fecundity in spiders is positively correlated with 
the body size of the female (Peterson 1950, Kessler 1973, 
Wise 1975) . 

The risks of ballooning are high in areas where 
suitable habitat is patchy. In such cases, ballooning is 
rare and has been selected against as a viable dispersal 
mechanism (Janetos 1986). The ant nests where M. 
pogonophilus makes its home are spaced an average of 12.1 m 
from one another according to a study by Harrison and 
Gentry (1981). At my study sites, Archer Sandhill in Levy 
County, Florida and the Ordway-Swisher Preserve in Putnam 
County, Florida, the mean nearest neighbor distances were 
11.45 m and 20.00 m respectively (Chapter 5). Due to the 



51 



low fecundity of M. pogonophilus , the high spacing between 
P. badius nests, and the susceptibility of M. pogonophilus 
to desiccation outside the nests, ballooning is an unlikely 
dispersal mechanism for the spiders. 

I hypothesized that the spiders were, instead, 
locating new nests by using the chemical signals laid down 
by the host ants. Although neighboring P. badius 
colonies, as well as colonies of western species of 
Pogonomyrmex harvesters, often partition their foraging 
territories and locate their foraging trails in such a way 
as to reduce contact between foragers from different 
colonies (Holldobler 1976, Harrison and Gentry 1981), 
foraging trails of neighboring colonies sometimes do 
intersect (personal observation). I hypothesized that 
dispersing spiders follow foraging trails away from their 
host nests until they located, through random searching, 
the foraging trail of a neighboring nest. Although P. 
badius does not produce the distinct trunk trails seen in 
western species of the genus (Holldobler 1974, Holldobler 
1976), it does produce three or four primary trails that 
persist for several months (Harrison and Gentry 1981). The 
ability to follow host trails has been demonstrated for a 
variety of myrmecophilic arthropods (Moser 1964, Akre and 
Rettenmeyer 1968, Schroth and Maschwitz 1984). 

Materials and Methods 

Pheromones from the poison gland and Dufour's gland 
are involved with trail marking, recruitment to food 



52 



sources, and homing in P. badius (Holldobler and Wilson 
1970, Holldobler 1971). The chemical composition of the 
poison gland secretion was determined by Schmidt and Blum 
(1978). It is an enzyme rich substance containing high 
concentrations of phospholipase A 2 and B, hyaluronidase, 
acid phosphatase, lipase, and esterases (Schmidt and Blum 
1978). The composition of the Dufour's gland secretion 
(for other members of the genus Poqonomyrmex ) was 
determined by Regnier et al. (1973) who found it to consist 
of various hydrocarbons. There is no evidence of colony 
specificity for either of these glandular secretions 
(Holldobler and Wilson 1970, Holldobler 1971). Therefore, 
no effort was made in the following experiments to control 
for nest identity of the ants from which glands were 
excised or the ants used in the trials. 

Ants were collected from two field colonies and one 
laboratory colony for use in the following experiments. Ten 
spiders were collected from an excavated colony at Archer 
Sandhills in Levy County. These included two females, 
seven males, and one juvenile. (This was the only 
excavation in which more males than females were 
collected.) Eleven spiders were collected from the 
emigration trail of another P. badius colony at the 
Ordway-Swisher Preserve in Putnam County which included 
five females, one male and five juveniles. 

To test whether M. pogonophilus could follow the 
trail pheromones of its host, I conducted three types of 



53 



experiments: olfactometer experiments using poison gland 
and Dufour's gland extracts; choice experiments using 
artificial trails laid with gland extracts; and choice 
experiments using naturally laid trails. Due to 
difficulties in keeping the spiders alive in the lab, only 
16 were available for the artificial trail experiments and 
only 12 were available for the natural trail experiments. 
The spiders used for each experiment were guite active when 
removed from their vials. Any spiders that were listless 
or appeared to be sick or dying were not used for the 
experiments. Before experimenting with the spiders, I 
conducted bioassays with the host ants to ensure that the 
host ants, themselves, would respond to these chemical 
cues. 
Olfactometer Experiments 

I dissected poison glands and Dufour's glands from P. 
badius workers. These glands are located near the tip of 
the abdomen and release their contents through the sting. 
A drawing of these two glands is reproduced from Schmidt 
and Blum (1978) in Fig. 4-1. The olfactometer consisted of 
Y-shaped glass tubing with two enlarged bulbs on the arms 
of the Y (Fig. 4-2) (Vander Meer et al . 1988). 

For each trial, I made one-gland eguivalent and 
0.1-gland eguivalent solutions of the poison gland and of 
the Dufour's gland using hexane as a solvent. The 
one-gland eguivalent solution consisted of one crushed 
gland per 10 ul of hexane; the 0.1-gland eguivalent 



54 



solution consisted of one crushed gland per 100 ul of 
hexane. These solutions were in the same range as those 
used by Holldobler and Wilson (1970) and Holldobler (1971). 
For each trial, 10 ul of one solution or the other were 
pipetted onto a 3 X 7 mm rectangle of absorbent paper. Ten 
ul of hexane were used as a control. The paper saturated 
with the sample was placed in the enlarged bulb of one arm 
of the Y and the paper saturated with the control was 
placed in the other arm (see Fig. 4-1). An equal current 
of air was blown through the arms of the Y. 

Sixty ants were used for each trial . Ten ants at a 
time were placed in a small holding vial with rubber tubing 
at one end which could be fitted onto the long arm of the 
Y. I waited a few minutes for the ants to settle down 
before joining the holding vial to the glass tubing of the 
olfactometer. The ants then left the holding vial and, 
when they reached the junction of the Y, they either chose 
one arm or the other immediately or paused and antennated 
in both directions before choosing a side. The arm chosen 
was recorded. In a few instances, one or two ants refused 
to leave the holding vial; data for these individuals were 
not included. After every group of 10 ants, the 
olfactometer was rinsed with acetone, fresh extracts of 
sample and control were pipetted onto new pieces of paper, 
and the side of the Y-arm in which the sample and control 
were placed was reversed to eliminate directional bias. 
The same procedure was followed for the spiders except the 



55 



spiders were introduced into the apparatus individually. 
After each introduction, I attempted to sweep out at least 
the long stem of the Y-tube with a camel-hair brush to 
remove any stray draglines. After 4-5 spiders, the 
apparatus was cleaned and the locations of the sample and 
control were reversed. 
Artificial Trail Experiments 

Spiders not only can detect airborne chemicals via 
tarsal organs on their legs but can also detect chemicals 
from the substrate via contact chemoreceptors, or taste 
hairs, on the distal segments of the legs and palps (Foelix 
1982). I tested whether the spiders could detect trail 
pheromones from the substrate by laying an 11 - 12 cm trail 
of 0.1 gland-equivalent poison gland extract on a piece of 
chromatographic paper. An 11 - 12 cm hexane trail was laid 
at approximately 45° from the poison gland trail as a 
control . 

As with the olfactometer experiments, I first ran 
bioassays with P. badius to ensure that such 
substrate-bound trails were biologically meaningful. Four 
groups of 10 ants were placed in plastic vials and allowed 
to calm down since agitated ants tend to move randomly and 
do not readily orient towards trails (Holldobler and Wilson 
1970, personal observation). Each vial was then tipped on 
its side and opened, allowing the ants to move out onto one 
or the other trail. 



56 



After each group of ants, the chromatographic paper 
was changed, the metal tray in which the experiment was 
conducted was wiped with acetone, and the location of the 
trails was reversed. I repeated the experiment using the 
16 spiders still alive (from the original 21). The same 
procedure as above was followed, except the spiders were 
allowed to choose trails individually. The paper was 
changed, the tray cleaned, and the trails reversed after 
every four spiders . For both the ants and the spiders , a 
choice was considered a distinct directional movement along 
one trail or the other. 
Natural Trail Experiment 

Although the poison gland has been shown to be a 
major source of the trail pheromones of P. badius, it is 
possible that the trails are actually a mixture of 
glandular secretions from both the poison and Dufour's 
glands. To test whether the spiders could follow naturally 
laid trails of P. badius . I collected between 90 - 100 P. 
badius workers from a field colony and established them in 
a 45 X 65 cm tray. The ants moved readily into a test tube 
half -filled with water, plugged with cotton and covered 
with red acetate. On the opposite side of the tray from the 
test tube, I placed two flat metal dishes, each 6.5 cm in 
diameter, on 1.4 cm high pedestals. On one of these 
elevated dishes, I placed seeds and on the other I placed a 
mixture of the Bhatkar ant diet (Bhatkar and Whitcomb 
1970). Since P. badius are not adept at crawling up 



57 



vertical surfaces, the only way they could reach the food 
was to travel up a 1.5 X 9 cm strip of chromatographic 
paper that served as a bridge. The paper bridge was kept in 
place for four days, and the ants readily used it to 
forage. As the number of ants moving from the opposite 
side of the tray directly to and up the bridge increased 
over that time period, I assumed they had marked a trail 
both on the tray surface as well as on the paper bridge. 

To test this, I positioned two glass tubes in a metal 
tray at approximately 45° from one another. In one of the 
tubes, I placed a strip of chromatographic paper cut the 
same size as the strip used for the bridge. In the other, 
I placed the bridge. Both strips extended 1 cm beyond the 
end of the glass tubes and both ends were folded down so 
the paper was flush with the tray surface. Six groups of 
10 ants were placed in vials and allowed to calm down. As 
with the artificial trail experiments, each vial was then 
tipped on its side and opened. I recorded which glass tube 
the ants entered. After every group of 10 ants, I cut a 
new strip of control paper, rinsed the glass tubes and the 
metal tray with acetone, and reversed the placement of the 
paper strips. 

The experiment was repeated with the 12 remaining 
spiders. As before, the spiders were allowed to choose 
individually. To prevent the tiny spiders from evading 
both tubes altogether, I built a small arena (approximately 
3X3 cm) out of pieces of stiff acetone placed in front of 



58 



and between the two tubes. The spiders were allowed to 
move about in the arena until they moved up one piece of 
paper or the other. After each spider, I brushed the arena 
with a camel hair paint brush to remove stray dragline 
silk. After six spiders, I cut a new strip of control 
paper, rinsed the tray and the glass tubes with acetate and 
reversed the order of the paper strips. For all 
experiments, the data were analyzed using X 2 tests. 

Results 

Ants were highly attracted to extracts of the poison 
gland (Table 4-1). These results are consistent with those 
of Holldobler and Wilson (1970), Holldobler (1971) and 
Regnier et al. (1973). However, my results indicate that 
P. badius is not attracted to extracts of the Dufour's 
gland (Table 4-1). These results contradict the published 
accounts of the above authors. However, Holldobler 
(personal communication) assured me that orientation 
towards the poison gland extract and not towards the 
Dufour's gland extract is in accord with his expectations 
for Pogonomyrmex harvester ants. Since my olfactometer 
experiments indicated that the Dufour's gland extract was 
not a useful bioassay for trail-following, it was not used 
for any subseguent experiments with the ants or spiders. 

Since the 1-gland eguivalent poison gland extract was 
found to be a biologically meaningful concentration for the 
host ants, I used this higher concentration extract first 
in my olfactometer experiment with the spiders. I reasoned 



59 



that if either the 1-gland equivalent or the 0.1-gland 
equivalent mixtures were not biologically (physiologically) 
meaningful concentrations, I would have seen some 
indication of this in the behavior of the ants in the form 
of reduced orientation towards the sample (Attygalle and 
Morgan 1985). Since both the 1-gland equivalent and the 
0.1-gland equivalent poison sac extracts elicited 
significant orientation responses from the ants, I 
considered both concentrations biologically meaningful. The 
spiders, unlike their hosts, did not orient more towards 
the poison gland extract than towards the control (Table 
4-2) . 

Since the myrmecophilic spiders may respond more 
readily towards substrate-bound chemicals than towards 
airborne odors, I used the artificial trail experiment to 
determine if they might be better able to follow such a 
trail. I used the 0.1-poison gland equivalent solution for 
this experiment. Although the host ants showed a clear 
preference for the artificial trail over the control trail 
(Table 4-1), the spiders did not (Table 4-2). 

Because the spiders were not attracted to the pure 
poison gland extracts however the extracts were presented 
to them, I thought they might, instead, be responding to 
some other chemical element of the natural trails. 
However, the experiment with the natural trail indicated 
that this was not the case (Table 4-2). Although the host 



60 



ants oriented towards the natural trail (Table 4-1), the 
spiders did not. 

In all experiments, the numbers of males, females, 
and juvenile spiders orienting towards the sample versus 
the control was approximately equal. There was also no 
correlation between spiders collected from the emigration 
trail versus spiders collected from inside the nests and 
their orientation towards either the extracts or the 
controls. 

Discussion 

The results do not support the hypothesis that M. 
pogonophilus spiders use trail pheromones of host ants to 
disperse from and locate new host nests. However, the 
context of laboratory experiments may not be adequate for 
testing this hypothesis. Perhaps the spiders do not 
respond in captivity as they would otherwise respond in the 
field. Or perhaps they are only responsive to host 
pheromones during certain times in their life. However, if 
this were the case, the spiders collected from the 
emigration trail should have been more sensitive to either 
the artificial or natural trails than the spiders collected 
from inside the excavated nest. This was not the case. 
Rather than cueing in on trail pheromones, spiders may 
locate new colonies by sensing airborne colony odors 
(osmochemotaxis) . Holldobler (1969) demonstrated that a 
myrmecophilic staphylinid beetle, Atemeles pubicollis . was 
able to locate new host colonies via such airborne cues. 



61 



The spiders are able to follow the host ants from the 
old nest sites to the new nest sites when the hosts 
emigrate (Chapter 2). I believe this is the time when 
spiders disperse from the host colony and make their way to 
new colonies. The present study suggests that the spiders 
may not be using trail pheromones to accomplish either 
feat. It is highly unlikely that they are following the 
host ants visually since the majority of araneoid spiders 
have notoriously poor eyesight (Foelix 1982). If M. 
pogonophilus is an exception to this rule, I should have 
seen some evidence of visual acuity in the prey capture 
behavior of the spiders. However, when I feed collembolans 
to the spiders in the lab, they only attempt to capture the 
springtails that have either directly contacted the spider 
or wandered next to the spider (or been caught in the webs 
many of the spiders build in the lab) . These observations 
suggest that the spiders respond primarily to vibratory, 
rather than visual stimuli. 

If the spiders are not responding to trail pheromones 
or visual signals when locating a new nest site, then they 
might simply be following the vibrations or movements of 
the host ants themselves — literally swept along with the 
ants in the emigration trails or in the foraging trails. 
However, I have seen spiders in emigration trails moving 
towards the new nest sites even when no host ants are in 
their immediate vicinity. I have never seen spiders in 
foraging trails. 



62 



Another possible mechanism of dispersal is phoresy, 
or hitching a ride on the body of a newly inseminated queen 
or a forager. Phoresy has been seen in other ant symbionts 
such as many species of mites (Holldobler and Wilson 1990). 
Some species of pseudoscorpions are phoretic on a variety 
of other arthropods as well as on vertebrates (Weygoldt 
1969). However, it is unlikely that M. pogonophilus is 
using phoresy to disperse from nest to nest. Since each P. 
badius nest is established by a single inseminated queen, 
the spider would have to ride on the body of a queen as she 
flies away from the natal nest to find a nest site of her 
own. Immediately prior to leaving the natal nest, the 
female alate is mated by both siblings as well as males 
flying onto the nest from surrounding colonies (however, 
personal observations of mating in P. badius indicates that 
most matings are between siblings). Mating among male and 
female P. badius alates occurs on the surface of the 
mounds. Several males surround and clamber onto the bodies 
of females who often try to dislodge the males (Turner 
1909, Van Pelt 1953, Harmon 1993, personal observation). In 
the mating activity I witnessed, I saw no symbionts 
— collembolans, mites, or spiders — clinging to the bodies 
of the females. The spiders are so delicate and so 
unprotected by any type of thickened cuticle that I find it 
unlikely that, if they were phoretic, they would survive 
the mating frenzy of the alates. The spiders may be 
phoretic on the bodies of foragers. However, after hundreds 



63 



of hours of observing ants in the field, I have never seen 
a single such phoretic spider. 

The ants are probably not actively transporting their 
guests themselves. After observing many emigrations of P. 
badius , all the M. pogonophilus spiders I saw in the 
emigration trails were moving of their own accord; none 
were being carried by the host ants. 

In sum, the mechanism by which M. pogonophilus is 
dispersing to new host nests remains a mystery. Although 
this study argues against the hypothesis that the spiders 
are using the chemical signals of the host ants, it may be 
that the laboratory bioassays, although adequate for the 
host ants, were not meaningful for the symbionts. For 
example, the texture of the substrate may be of critical 
importance in cuing the spider in to the presence of a 
foraging or emigration trail. 



64 



-H 
03 
Sh 

P 



03 
M 

-P 
03 
C 

T3 

C 
03 



03 
P 



03 
•H 
U 



w 
D 

> 

I 



o 

Eh 
Z 
O 

u 
o 

Eh 



r4 


fj 


oo 


m 


(N 


CN 


o 


o 


CN 


en 


o 


O 


o 


o 


<D 


r^ 


o 


O 


o 


o 


o 


o 


o 


O 


V 













o 

rH 



in 



CM 



en 

CM 



vO 



If) 



u 


w 














•H 


hJ 














-P 


04 














03 


§ 


en 


IT) 


«5 


en 


■* 


IT) 




CO 


■* 


<tf 


Ol 


CN 


co 


x* 


^H 


o 














(U 


Eh 














P 
















CD 


=*fe 














B 












P 

















C 




p 












CU 

















r-i 




03 










p 


03 




4-1 • 










C 


> 




rH Cfl 






p 


P 


CU 


•H 




M 






c 


C 


r-i 


3 




CU 




P 


CU 


CU 


03 


tr 




Q) M 




c 


H 


r-{ 


> 


CU 




£ Sh 




CU 


03 


03 


•H 






P 




i-H 


> 


> 


3 


TD 




> 




03 


•H 


•H 


CP 


c 




4-1 




> 


3 


3 


CU 


03 




O Cfl 




-H 


D 1 


1 




r-i 




3 




3 


CU 


CU 


T3 


D> 




CO -H 




CT 






C 


1 




P T3 




CU 


T3 


T3 


03 


rH 




H 03 






C 


C 


r-t 


• 




3 X! 




73 


03 


03 


D> 


o 




CO 


W 


C 


rH 


rH 


1 






CU • 


hJ 


03 


CT> 


IT 


rH 






« Ch 


Oj 


p— 1 


1 


1 


• 


• • 






s 


CP 


r-t 


r^ 


o 


r-t 




CX 


frf 


1 


• 






■H 




• c 


1/1 


rH 


o 






03 




rH -H 








• • 


• • 


Sh 




1 Cfl 


Pn 






T> 


T3 


Eh 




^ 3 


O 


• • 


• • 


C 


C 




r-i 






T3 


T3 


03 


0! 


T3 


■H 


CU cfl 


■z 


C 


c 


r-\ 


^ 


C 


03 


r-t 4-> 


o 


03 


0! 


o 


O 


03 


Sh 


J3 C 


H 


rH 


rH 






r-i 


Eh 


03 CU 


Eh 


u 


o 


Cfl 


Cfl 


a 




Eh g 


H 






V 


N 




rH 


•rH 


W 


c 


c 


M 


Sh 


c 


03 


u 


o 


o 


o 


3 


3 


o 


Sh 


CU 


& 


Cfl 


Cfl 





O 


CO 


3 


a 


g 


•H 


■rH 


4H 


4H 


-H 


P 


X 


o 


o 


o 


3 


3 





03 


CU 


u 


0< 


&4 


Q 


Q 


04 


Z 



65 



01 



en 



•H 




w 


o 




o 


03 




O 


V 


o 


V 


M 




r-1 


a 


01 


a 


P 




<: 


V 


• 


V 






> 


If) 


o 


IT) 


rH 




1 


• 


A 


• 


03 




ft 


o 




O 


!h 












3 












P 












03 












C 




j 








T3 




o 








C 




PS 








(0 




25 








»- 




O 


o 


CO 


IT) 


rH 




U 


rH 






-rH 












03 




o 








!h 




Eh 








p 




=tfc 








rH 












03 












•H 












O 












-H 












4H 




W 








-H 




i-l 








■P 




ft 








Sh 




g 








03 




en 


rH 
rH 


CO 


r> 


5h 




o 








CD 




Eh 








P 












0) 


• 


=*fc 








e 


CD 






P 







U 






C 




p 


0) 






<u 




O TJ 






>-i 




03 


•H 






03 




4H 


a 






> 




rH 


03 






•H 













3 






co 




P 


O 1 




0) 


3 




c 


0) 




£ 


rH 




a) 






P 


•rH 




rH 


T3 






-C 




03 


C 




MH 


a 




> 


03 












•rH 


rH 






c 




3 


0i 




CO 


o 




1 


1 




P 


Di 




0) 


r-^ 




rH 


o 






• 




3 


a 




T3 


o 




CO 




W 


C 






0) 


• 


i-l 


03 






a s 


ft 


r-t 


• • 






Di 


^ 


Oi 
1 


rH 

■H 




• 


c 


CO 


rH 


03 




(M 


H 






Sh 




1 


tn 


ft 




Eh 




-tf 


a 


o 


• • 




rH 








T5 


T3 


-rH 


0) 


tn 


25 


C 


C 


03 


rH P 


o 


03 


03 


U 


X! 


c 


H 


rH 


rH 


Eh 


03 


0) 


Eh 


u 


o 




&H 


e 


H 






rH 




H 


CO 


c 


c 


03 




r4 


O 


o 


o 


Sh 




a) 


ft 


CO 


CO 


3 




a 


s 


-H 


•rH 


P 




X 


o 


o 


o 


03 




<D 


u 


ft 


ft 


25 



66 




Figure 4-1 (From Schmidt and Blum 1978). Venom 
apparatus of the harvester ant P. badius . Abbreviations: 
DG, Dufour's gland; FF , free filaments; SS, sting shaft; 
VR, venom reservoir. 



67 




<Z> 



Figure 4-2. Y-shaped Olfactometer. A, air flow 
over the sample or over the control. P, chromatographic 
paper saturated with sample or with control. Ants or 
spiders are introduced into the long stem of the Y. Dra.wn 
approximately to scale. 



CHAPTER 5 
FACTORS AFFECTING SEED SELECTION BY THE FLORIDA HARVESTER 
ANT, POGONOMYRMEX BADIUS , AT TWO NORTH FLORIDA SITES 

Introduction 

Although it is widely acknowledged that granivorous 
ants influence habitat structure (Coffin and Lauenroth 
1990, Crist and MacMahon 1992, Gentry and Stiritz 1972, 
Harmon and Stamp 1992, Hobbs 1985, Kelrick et al . 1986, 
Rissing 1981), there is little agreement as to what 
influence habitat structure has on resource use by 
granivorous ants. My objective was to investigate the 
influence of habitat structure on seed use by the Florida 
harvester ant, Pogonomyrmex badius (Latreille) 
(Formicidae) , and to determine if seed selection is 
correlated with seed nutritional guality. Pogonomyrmex 
badius inhabits the Gulf Coastal states of the U.S. east 
of the Mississippi River from Louisiana to North Carolina. 
It is found in more xeric habitats with well-drained soils 
and is an important arthropod granivore in these 
environments (Cole 1968). 

Fewell (1988) showed that when resources are 
clumped, P. badius moves its foraging trails to maximize 
resource use. Such clumping of resources may be reflected 
in the proportional representation of that plant species 
harvested by the ants (Davidson 1977, Whitford 1978). 

68 



69 



Where seed sources are more variable or more evenly 
dispersed, the proportions of different seeds brought into 
each colony may reflect this difference in resource 
availability (Fewell 1988). Although some researchers have 
found such a correlation (de Vita 1979, Whitford 1978), 
others have not (Crist and MacMahon 1992, Gordon 1993, 
Hobbs 1985) . 

Seed selection can also be influenced by seed 
characteristics such as shape (de Vita 1979, Pulliam and 
Brand 1975, Whitford 1978), size (Bailey and Polis 1987, 
Rissing 1981), and nutritional quality (Fewell 1990, 
Gordon 1980, Kelrick and MacMahon 1985, Kelrick et al. 
1986). However, researchers do not agree on which, if 
any, of these factors are most important in seed selection 
by ants. Some have indicated either no correlation or 
negative correlation between size or quality and seed 
selection by ants (Crist and MacMahon 1992, Pulliam and 
Brand 1975) . 

Here, I was specifically interested in addressing 
the following questions: 1) Are there attributes of 
habitat structure and seed availability in different 
habitats that influence seed selection by ants? 2) Are 
there differences in seed selection by colonies found in 
the same habitat that may reflect either seasonal 
differences in resource availability or differences in 
seed availability within the different foraging ranges of 



70 



the colonies? and 3) Does the nutritional quality of seeds 
influence seed selection by the ants? 

The study was conducted at two different sites in 
north Florida: Archer Sandhills (ASH) in Levy County and 
the Ordway-Swisher Preserve (ORD) in Putnam County. The 
habitat at ASH is dominated by Florida rosemary, Ceratiola 
ericoides (Empetraceae) , whereas the ORD site consists of 
an old field habitat with no one plant species 
predominating. This field was characterized by a wide 
variety of grasses and forbs (Franz and Hall 1991). If the 
apparent difference in habitat structure between these two 
sites is reflected in the resources available to the ants 
from the seed bank (i.e., dropped seeds lying in the 
soil), and if the ants are selecting seeds based on 
availability, a preponderance of rosemary seeds should be 
found in the seed chambers, or granaries, of the P. badius 
colonies at ASH and a more variable representation of 
seeds in the granaries of the colonies at ORD. In 
addition, a measurement of the proportional availability 
of seeds in the seed bank should correspond to the 
proportions of these seeds collected by the ants. 
(Reichman (1979) has shown that Pogonomyrmex harvester 
ants primarily collect seeds from the surface.) 

Furthermore, if individual foragers or individual 
nests specialize on certain types of seeds, then 
differences in seed selection should be observed between 
colonies within a site. Such differences in seed selection 



71 



may also result from seasonal differences in seed set by 
plants within the each colony's foraging range. Finally, 
if the ants are selecting seeds, not based on availability 
but on guality, then correlations should exist between 
seeds harvested and some measure of quality (e.g., seed 
size, caloric, lipid, protein, or carbohydrate content). 

Materials and Methods 
Seed Collection and Comparison of Seeds Harvested 

Seven P. badius nests were excavated at Archer 
Sandhills (ASH) in Levy County, Florida 26 Km SW of 
Gainesville on 8 Oct 1992 (ASH15), 20 Feb 1993 (ASH18), 10 
Apr 1993 (ASH20), 21 Apr 1993 (ASH21), 4 June 1993 
(ASH20B), 13 June 1993 (ASH22), and 25 Sept 1993 (ASH24). 
ASH20 and ASH20B represent two excavations of the same 
nest. Four P. badius nests were excavated at the 
Katherine Ordway Preserve-Swisher Memorial Sanctuary (ORD) 
in Putnam county, Florida 35 Km SE of Gainesville on 10 
Jan. 1993 (ORD16), 23 Jan. 1993 (ORD17), 28 Feb. 1993 
(ORD19), and 25 July 1993 (ORD23B) . ORD17 and ORD23B 
represent two excavations of the same nest. The contents 
of two to five granaries were collected from each nest, 
dried, sorted, and counted. If several granaries were 
encountered during an excavation, we collected several at 
different depths. Florida rosemary seeds, C. ericoides, 
were so abundant in the ASH nests that we estimated the 
number present in each granary using the weight of 1000 
seeds. Rosemary seeds were stored both as individual 



72 



seeds as well as fruits (each of which contains two 
seeds). (Harvester ants were observed to collect both 
individual seeds as well as fruits.) Therefore, fruits 
were counted as two seeds. A little over half of the 1000 
seeds weighed were in the form of fruits since this is the 
approximate representation of fruits versus individual 
seeds found in the granaries. The seeds were identified, 
as far as possible, using the seed collection at the 
University of Florida Herbarium as well as floral lists of 
ASH and ORD (Franz and Hall 1991 and W. Judd, pers . 
comm. ) . Representatives of all seed types collected were 
then sent to the USDA APHIS Seed Examination Facility in 
Beltsville, Maryland for further identification or 
verification. Voucher specimens of all seed types will be 
kept for future reference. The average seed weight was 
determined for each species using at least 10 seeds (when 
possible). All seeds were kept under the same storage 
conditions throughout the study. 

Chi-sguare analyses were used to determine if any 
differences existed in the proportions of different seed 
species harvested between the ASH nests or between the ORD 
nests. For each X 2 analysis, seed types that had expected 
numbers of five or fewer were combined. Since the nests 
were excavated at different times of the year, I suspected 
that seasonal differences in seed availability might be 
reflected as differences in granary contents between the 
nests. 



73 



Measuring Seed Availability 

To determine whether seeds harvested were correlated 
with seed availability, six samples from the seed bank at 
ASH and six samples at ORD were collected in May 1995. 
Each sample was collected by placing a ring 17 cm in 
diameter onto the soil and collecting the soil within the 
ring down to a depth of 0.75 cm. Three samples at each 
site were collected 2 m from a P. badius nest while the 
other three samples were collected 4 m from the same three 
nests. Therefore, all six samples were collected well 
within the foraging range of the ants. The three nests 
chosen were greater than 30 m from one another in order to 
sample a broad range of resources available to the ants in 
that habitat. The samples were dried and sifted through 
successively smaller mesh sizes, the smallest being 0.50 
mm. Coffin and Lauenroth (1989) found seasonal 
differences in the seed composition of the seed bank. To 
reduce the potential temporal bias inherent in collecting 
seed bank samples only one time of the year, seed 
fragments were included in the samples as long as the 
fragments were identifiable. However, to be conservative 
in the estimates of seed numbers present, seed fragments 
were counted as 1/2 a seed. The data from the six samples 
at each of the two sites were pooled and Spearman's rho 
was used to determine if overall proportions of different 
seeds from the ASH and from the ORD nests were correlated 
with seed availability. 



74 



Measuring Nutritional Value of Seeds 

Nutritional value was determined for only the ten 
seed species for which sufficient material was available. 
Of those ten, six were among the most abundant species 
listed in Table 5-2. The remaining four species 
represented very minor portions of the contents of the ASH 
and ORD nests. However, these seeds were abundant in the 
granaries of a nest excavated at a separate site in North 
Florida (San Felasco Preserve) and were included in the 
nutritional analyses because they were part (if only a 
minor part) of the contents of the ASH and ORD nests. 
Insufficient material (< one gram) was available to run 
nutritional analyses for the following seeds, listed in 
Table 5-2 as being some of the most abundant species 
collected by the ASH and ORD nests: golden aster, 
Pityopsis ( Chrysopsis ) cf graminif ola (Asteraceae) ; 
centipede grass, Eremochloa ophiuroides (Poaceae); an 
unidentified species of grass (Poaceae); and jointweed, 
Polygonella sp. (Polygonaceae) . The nutritional guality of 
the euphorb Crotonopsis linearis , or rushfoil, was also 
measured. This species was abundant in the San Felasco 
nest and was used for a seed choice experiment (see 
below) . 

Caloric content was measured using a Parr Model 1261 
Isoperibol Calorimeter (Paine 1971, Parr Instrument Co. 
1984). Between 0.25 - 0.65 g of macerated material was 
used for each run. For each run, benzoic acid was used as 



75 



a standard. Two bombs were used alternately. Bomb #1 had 
a relative standard deviation of 0.130827. Bomb #2 had a 
relative standard deviation of 0.120286. I did not 
correct for ash content since all samples were less than 
one gram. Sufficient material was available to do two or 
three replicate runs for five of the 11 major seed types. 
For these replicates, the largest measurement of caloric 
content did not differ from the smallest measurement by 
more than 1.27%. Therefore, I was confident that even for 
the six seed types for which no replicates were made, the 
measure of caloric content was accurate. The caloric 
content per gram of material was converted to calories per 
seed using the seed weights. For the five seed types for 
which replicates were made, the mean value for all 
analyses was used. 

An ether extraction technigue was used to measure 
the lipid content of the 11 major seed species according 
to the procedure described in AOAC (1990). Lipid content 
was recorded as % lipid / dry weight. For this technigue, 
sufficient material was available to run replicates of 
five seed species. The higher replicate for each species 
did not differ by more than 1.10% from the lower value. 
Therefore, the mean % lipid was used for all analyses. 

A protein digestion technigue using a Technicon Auto 
Analyzer was used to determine % Crude Protein in the 11 
seed species following the protocol in Gallaher et al 
(1975) and Hambleton (1977). Approximately 0.25 g of 



76 



macerated material was used for each sample. This 
analysis was conducted by the Forage Evaluation Support 
Laboratory at the University of Florida. 

Once the total caloric density of the seeds was 
determined, as well as the % lipid and % crude protein 
composition, a crude estimate of % carbohydrate content 
was calculated using values of the catabolic yield for 
seed carbohydrates, lipids and proteins: 4100 cal/g, 9300 
cal/g, and 4200 cal/g respectively (Cristian & Lederle 
1984, Hill 1976) . 

Spearman's nonparametric rho (Sokal and Rohlf 1981, 
Siegel 1956) was used to test for correlations between 
seed size (i.e., average seed weight) and caloric content, 
% lipid, % crude protein, and % carbohydrates. Rho was 
also used to determine if there were any correlations 
between seeds harvested (i.e., proportions of each seed 
species collected by the ants) and any of the following 
measures of seed value: size, caloric content, % lipid 
content, % crude protein, and % carbohydrate. 
Seed Choice Experiment 

Because of the preponderance of rosemary seeds, both 
in the granaries as well as in the seed bank samples at 
ASH, I decided to do a seed choice experiment to determine 
if individual foragers will choose seeds other than 
rosemary if given a choice. Thirteen foragers from eleven 
nests were presented with a choice of one rosemary seed, 
one rushfoil (C. linearis ) seed and one unidentified sedge 



77 



seed ( Cyperus sp.#2, Cyperaceae). The rushfoil and sedge 
seeds were considered novel since neither had been found 
in any of the excavated ASH nests. Rushfoil was a larger 
seed than either the sedge or rosemary seeds and had 
higher nutritional values than rosemary seeds for all 
measures of nutritional quality except estimated % 
carbohydrates (Table 5-4). The sedge was smaller than the 
rosemary seed and had lower values than rosemary caloric 
content, % dry matter, and % lipids. It had slightly 
higher % crude protein and estimated % carbohydrates 
(Table 5-4) . 

A different cluster of the three seeds were used for 
each trial so that no seed was used twice. The three 
seeds were placed atop a small amount of sand on a 2.5 cm 2 
cardboard held in front of a forager with forceps. In all 
trials, the three seeds were tightly clustered to increase 
the likelihood that the foragers would contact all of 
them. The foragers were from 0.7 - 5.25 m from their 
respective nests when intercepted. The ants did not appear 
to be disturbed by the presence of the cardboard and 
readily walked onto its surface. Only instances when a 
forager chose a seed were recorded. Foragers often walked 
over the cardboard without inspecting the seeds. In 
addition to recording what seed was chosen by a forager, I 
also recorded the order of inspection (i.e., antennation) . 

If individual foragers prefer novel seeds when 
encountered as Fewell and Harrison (1991) suggest, then 



78 



the rushfoil and sedge seeds should be chosen more often 
than the rosemary seed. If, however, individual foragers 
form a search image and specialize on seeds commonly 
encountered (Briese and Macauley 1981, Crist and MacMahon 
1991, Hobbs 1985, Whitford 1976) then the rosemary seed 
should be preferred over the rushfoil or sedge seeds. 
Finally, if individual foragers select seeds of high 
nutritional guality (when presented with a choice), as 
suggested by several authors (Cristian and Lederle 1984, 
Fewell 1990, Gordon 1980, Kelrick and MacMahon 1985, 
Kelrick et al. 1986, Whitford 1978) then the rushfoil seed 
should be chosen over either the sedge or the rosemary 
seeds since it was both of higher nutritional guality as 
well as a larger seed than either the rosemary or sedge 
seeds. 

Results 
Seed Collection and Comparison of Seeds Harvested 

The contents of 29 granaries, ranging in depth from 
24 - 94 cm, were collected from the seven ASH nests. The 
granaries each contained from 155 - 18000 seeds. The 
contents of 13 granaries, ranging in depth from 28 - 124 
cm, were collected from the four ORD nests. They 
contained from 527 - 4767 seeds. Fifty species of seeds 
in at least 19 different families were identified from the 
granaries (Table 5-1). At ASH, Florida rosemary made up 
94.2 - 99.9% of the seeds collected by the seven nests 



79 



(Table 5-2). At ORD, the colonies showed a more varied 
selection (Table 5-2). 

Differences existed in the proportions of the 
various seed species collected by the seven ASH nests 
(X 2 =11163 .12, P<<0.001). Differences also existed in the 
proportions of the various species collected by the four 
ORD nests (X 2 =13436 . 53 , P<<0.001). Separate analyses were 
made to determine if ASH20 and ASH20B and 0RD17 and ORD23B 
(the same nests excavated at two different times of the 
year — after the colonies had repaired damage from the 
first excavations) contained similar proportions of the 
various seed species. They did not (X 2 =58.52 and 916.98 
respectively, P<<0.001). The probability that the same 
granaries were collected during the second excavation of 
either of these pairs of nests is minimal. Two pairs of 
nests at ASH (ASH20 and ASH21 as well as ASH20B and ASH22) 
were analyzed that were excavated during the same month. 
Both pairs differed in the proportions of seed species 
collected (X 2 =248.47 and 514.04 respectively, P<<0.001). 
0RD16 and ORD17 were also compared. These nests, too, 
were excavated during the same month. They also differed 
in their granary contents (X 2 =789.48, P<<0.001). 
Measuring Seed Availability 

Four of the six species of seeds found in the seed 
bank samples at ASH were also found in the granaries 
(Table 5-3): Florida rosemary; twining milk-pea, Galactia 
volubilis (Fabaceae); threeawn grass; and arrowfeather 



80 



threeawn grass, A. purpurascens (Poaceae). These four 
species made up 99.8% of the seeds in the seed bank. The 
remaining seeds in the seed bank were two small 
unidentified species. Rosemary seeds accounted for 98.9% 
of the seeds in the seed bank samples and 98.0% of the 
seeds found overall in the granaries. No significant 
correlation existed between the proportions of the four 
species found in the seed bank samples and the overall 
proportions of these species represented in the ASH nests 
(R s =0.800, P>0.05). However, this lack of correlation was 
due entirely to differences in the abundance of the minor 
seed species. 

Sixty-five percent of the seeds found in the seed 
bank samples from ORD were also species collected by the 
ants (Table 5-3). The remaining 35% consisted of six 
unidentified species, two of which accounted for 26% of 
all the seeds found. These two seed species had 
distinctive feathery awns which de Vita (1979) suggested 
made seeds difficult for seed harvesting ants to 
transport. In fact, none of the species collected by the 
ASH or the ORD nests had any such awns. Fifteen species 
found in the granaries were also represented in the soil 
bank and included the following most collected species: 
sedge, Cyperus sp. #2; threeawn grass, Aristida sp. ; 
centipede grass, Eremochloa ophiuroides ; Pensacola bahia 
grass, P. notatum; jointweed, Polygonella sp. ; and Poor 
Joe, D. teres (Table 5-2). A significant positive 



81 



correlation existed between the proportions of these 15 
species found in the seed bank samples and the overall 
proportions of these species represented in the ORD nests 
(R s =0.509, P<0.05). 
Measuring Nutritional Value of Seeds 

The nutritional characteristics of six abundant seed 
species (Table 5-2) as well as four additional seed 
species that make up only a small proportion of the 
granaries of the ants are summarized in Table 5-4. Seed 
size (i.e., seed weight) is positively correlated with 
calories / seed (R s =0.997, P<0.01). Seed size is not 
correlated with any other measure of nutritional quality. 

The measures of the nutritional quality of the seeds 
were compared with the relative proportions of those 
species found in each of the eleven nests. Since seed 
weights had been recorded for all the species found in 
each of the nests, these data were used to determine if 
any correlations existed between seeds harvested and seed 
size. The proportions of seeds collected by the ants were 
negatively correlated with seed size for 0RD17 (R s = 
-0.348, P<0.05); ORD23B (R s = -0.418, P<0.05); ASH20 (R s = 
-0.743, P<0.05); and ASH22 (R s = -0.628, P<0.05). In 
ASH18, proportions of seeds collected were negatively 
correlated with % crude protein (R s = -1.00, P<0.05) and 
positively correlated with estimated % Carbohydrates 
(R s =1.00, P<0.05). In ASH21, proportions of seeds were 
also negatively correlated with % crude protein (R s = 



82 



-1.00, P<0.05). No other significant correlations were 
found between the proportions of seed species harvested 
and any measure of seed quality. By chance alone, a 
significant difference is expected in 5% of the 
correlation tests. Chance would account for some but not 
all instances in which significant correlations were 
found. The various measures of nutritional quality were 
converted to absolute amounts (i.e., total dry matter by 
weight, total lipids by weight, etc.) and these values 
compared with the proportions of the seeds harvested by 
the ants. No significant correlations were found. 
Seed Choice Experiment 

For the seed choice experiment conducted at ASH, 12 
of the 13 foragers selected the rushfoil seed and one 
forager chose a rosemary seed (X 2 =20.62, P<0.001). Of the 
12 foragers that chose the rushfoil seed, nine clearly 
antennated at least one other seed before making a choice. 
Three appeared to select the rushfoil immediately upon 
encountering it. The ant that chose the rosemary seed 
antennated only the sedge seed before making its choice. 

Discussion 

The two habitats explored in this study were 
characterized by different overall structure and floral 
composition. The resources available to P. badius at the 
Archer Sandhills site were clumped with seeds concentrated 
under rosemary bushes rather than scattered evenly 
throughout the habitat, and were homogeneous, consisting 



83 



predominantly of the seeds of Florida rosemary. In 
contrast, the Ordway site was much more heterogeneous with 
a greater variety of seed plants available to the ants. 

The lack of correlation between the seed bank sample 
and the granary contents at ASH was due to differences in 
the abundance of those seed species making up less than 3% 
of either the seed bank samples or the granaries. No 
appreciable difference was found between the proportion of 
rosemary seeds found in the granaries and the proportion 
found in the seed bank. At the ORD site, a positive 
correlation existed between the seeds found in the seed 
bank samples and the seeds found in the granaries (when 
the data from all the nests were pooled) . These data 
support the hypothesis that this species of harvester ant 
is selecting seeds based primarily upon availability. 

However, 26% of the seeds found in the seed bank at 
ORD were never found in any of the granaries. The bulk of 
these seeds had structures which are apparently difficult 
for ants to handle (de Vita 1979). Therefore, the ants 
probably do reject certain seeds due to morphological 
characteristics . 

The fact that significant differences were found 
among the nests at each site supports the hypothesis that 
the timing of seed set and seed deposition into the seed 
bank by different species of plants has a significant 
impact on resource availability (Coffin and Lauenroth 
1989). This is further supported by the observation that 



84 



the nests (one at ASH and one at ORD) which had each been 
excavated twice differed significantly between the first 
and second excavations. The most likely explanation for 
this is a seasonal difference in resource availability as 
suggested by Coffin and Lauenroth (1989). These 
observations further support the hypothesis that ants 
select seeds based primarily upon availability. 

Significant differences in the granary contents 
among nests excavated during the same month (but at 
different locations in the habitat) suggest a difference 
in seeds available to the colonies within each of their 
foraging ranges. In fact, when the seed bank samples 
collected 2 and 4 m from the same P. badius nests are 
pooled and the resulting three seed bank data sets (at 
each site) are compared, a significant difference is found 
in seeds available to different P. badius nests (X 2 = 
1119.3, p << 0.001 for ORD samples; X 2 = 32.4, p << 0.001 
for ASH samples). For both these analyses, all seed types 
with expected values < 5 were collapsed into one category. 
Therefore, plant phenology as well as the structure of the 
habitat (in terms of where different seed plants are 
growing in relation to the colonies) influences resource 
use and seed selection. 

Finally, the colonies, in general, did not seem to 
be cuing in on any measure of seed guality. In fact, when 
significant correlations were found, these were primarily 
negative correlations. This contradicts the observation 



85 



by many workers that harvester ants do use measures of 
nutritional quality when selecting seeds (Fewell 1990, 
Gordon 1980, Kelrick and MacMahon 1985, Kelrick et al. 
1986). However, Fewell (1990) and Kelrick et al. (1986) 
determined seed preference by offering ants clumps of 
seeds from which to choose (analagous to our seed choice 
experiment except we offered only three seeds to 
individual foragers rather than dishes of many seeds of 
different species to foragers as a whole). As pointed out 
by Crist and MacMahon (1992), individual foragers probably 
rarely have the opportunity to choose among a variety of 
seed species of varying quality. Therefore, at the level 
of individual foragers, it is unrealistic to present ants 
with a choice of seeds and assume that their choice of 
higher quality seeds means that, in their daily foraging 
bouts, they are actively using some criteria of seed 
quality in their selection of resources. The seed choice 
experiment leaves little doubt that, when presented with a 
choice, harvester ants are able to discriminate among 
seeds based on size or some measure of quality. This may 
be important on occasions when foragers encounter patches 
of seeds. However, this ability to discriminate is 
probably not relevant to the decisions the majority of 
foragers make when they encounter seeds since our data 
indicate they are basing choice more on availability than 
on quality. 



86 



Q) 
43 
-P 

E 
O 
U 



p 

0] 

QJ 
G 

w 

■H 

73 

03 

43 



73 
0) 
4J 
O 
0) 

rH 

o 
o 

0,1 

73 ~ 

QJ Q 

Q) « 

03 O 



14-1 

O 

c 
o 

-H 

-p 

<0 
P 

c 

0) 
03 

<D 

Sh 



QJ 
> 

U 
QJ 
03 
QJ 

& 

>i 
(0 
5 

73 



Q) O 
U 

73 

rH C 



33 
CO 



p 

o 
a 

O 03 

U H 

an 

•H 

73 43 
C 73 

id c 

10 

>ita 
■p 

■H sh 

P QJ 

C £ 

Q) O 

73 M 

h <; 

QJ 
• 43 

rH P 

I 

IT) 14-1 

O 

QJ 

<H 03 
42 QJ 

10 -H 

(0 
C 
(0 
M 



o\° 



w 

E-t 

H 
CO 



W 



z 
o 



o 
u 



CO 

w 

H 

u 

H 

o* 

CO 



H 

s 
< 



73 
QJ 

a) 
s 

c 
o 
p 
p 
o 
u 



QJ 

n3 
QJ 
(0 

a) 

u 

10 

p 

c 

(0 
M 
(0 

e 



o 



o 
o 



<H 

o 
o 

V 



33 
CO 

< 



33 

CO 

< 



S3 
CO 

< 



u 

(0 

E 
3 
co 



u o 

(0 (0 

E E 

3 3 

CO CO 





=»fc 


f0 




43 


• 


CJ 


a 


■H 


03 


H 




0) 


03 





3 


M 


43 


Cn 


K 



CN 

=tfc 



03 






03 



(XI (XI 



QJ 
(0 
QJ 
U 
(0 
■H 
73 
Sh 
(0 
O 
(0 

c 
< 



33 
CO 

< 



U 
QJ 
-P 
03 
< 

C 
0) 
73 
rH 

o 

o 



03 
■H 

03 

Ch 1 

o 



o 

V 



Q 
(X 

o 



73 

!h 
(0 
-P 
03 

3 
£ 



co 

o 



Q 
X 
O 



03 
3 
-P 
U 
«3 

u 
u 

(0 
0) 

a 



4* 

o 

•H 

!h 
Oh 



o 



33 
co 
< 



o 
o 



IT) 



a 
a 
o 



a 
x 
o 



0) 






(0 








>1 


^ 




03 








U 


rH 




3 




rH 




43 


o 




4-4 




=**= 




U 


14-4 


• 


•H 


TJ 






^-H 


a 


E 


QJ 


• 




c 


03 


3 


•H 


SX 




03 


-H 




43 


<4H 


03 


r\] 


■H 


E 


id 




-H 




=«= 


CO 


(0 


o 


(0 


-P 


03 




Q4 


u 


-H 


•H 


c 


3 


• 


O 


01 


0) 


-P 


QJ 


U 


a 


> 




03 


C 


73 


QJ 


03 


P 


• 


(0 


3 


•H 


^ 




■H 


<u 


U 


Q4 


c 


>■ 




ft 


o 


CQ 


O 


3 


U 


Ul 








QJ 












(0 








QJ 




QJ 








(0 




U 






QJ 


QJ 




10 


QJ 




to 


O 


QJ 


H 


(0 




Q) 


03 


(0 


3 


QJ 




O 


O 


QJ 


> 


O 




(0 


■H 





rH 


10 


Z 


Sh 


03 


(0 


o 


u 




0) 


03 


-P 


> 


QJ 




■p 


(0 


CJ 


c 


a 




03 


h 


(0 


o 


>i 




< 




CO 


CJ 


o 


u 





ro 
O 



co 



33 
CO 

< 



03 

QJ 
T3 
■H 

O 
U 



o 



Q 

X 

o 





0J 




rd 


QJ 


QJ 


(0 


u 


QJ 


(0 


O 


•H 


(0 


43 


Sh 


Sh 


■P 


O 


QJ 


43 


Q< 


a 


s 


3 


w 


W 



o 
\ 



o 

V 



a 
x 
o 
\ 

33 

co 

< 



rH 

o 



o 

V 



o 

V 



33 
CO 



33 
CO 

< 







U 










(0 










E 










QJ 










03 










O 










X 




73 






(0 




QJ 






73 


QJ 


QJ 


QJ 


QJ 


■H 


r- 1 


3 


CP 


0"> 


Sh 


-P 


QJ 


73 


73 


O 


P 


> 


QJ 


QJ 


H 


QJ 


O 


CO 


CO 


(in 


2 


a 



QJ 


03 






3 


• 


rc 


i— i 


a 


rH 


O 


03 





U 




■H 


03 


C 


P 


O 





rrj 


73 


p 


(H 


•H 





QJ 


c 


)H 


u 


u 


u 



73 
QJ 

-H 

<4H 

■H 
-P 

c 

QJ 

73 

•H 

c 

3 



a 

03 
QJ 

c 

QJ 

£ 
O 

c 

43 
O 
03 
0) 

< 



QJ 
(0 
QJ 
O 
(0 
43 
(0 



87 



























CN 






CO 






























CO 






m 










in 


rn 


rH 


rH 


rH 


•3* 


rH 


co 


rH 


(X) 


o 


CN 


rH 


en 


o 


n 






0> 


rH 


O 


o 


O 


rH 


O 


CM 


o 


rH 


\ 


o 


o 


rH 


C?i 


in 




o\° 






















rH 


• 


• 


\ 




. 






o 


O 


O 


o 


o 


o 


O 
V 


o 


o 


o 


O 

O 
V 

a 


o 


o 

V 


CO 

o 
o 

a 

Oh 


o 


rH 




w 


Q 


Q 


Q 


a 


a 


Q 


X 


Q 


a 


a 


o 


Q 


X 


o 


3C 


Q 




E-t 


« 


« 


« 


« 


a 


04 


CO 


a 


« 


o. 


\ 


a 


CO 


\ 


CO 


« 




M 


O 


O 


O 


o 


o 


O 


< 


o 


o 


o 


X 


o 


*fl 


X 


< 


O 




co 




P 

c 

CO 

rH 


co 


rH 


ITS 

a 
a 
i 






CO 


>1 






C/l 

< 


CD 

C 




CO 

< 


c 

CO 
CD 
CD 
U 

x: 
p 

S-i 


CO 
CO 
CO 
Sh 




u 




a, 


rH 


-H 


rH 






u 








•H 






CD 


t? 




s 


x: 




rH 





•H 






(0 








CU 






x: 






S3 


u 


a> 


CD 


4-4 


2 






o 






73 




CO 




p 


0) 




S 


p 


> 


XI 


CD 








X 






<u 


4-1 


E 


c 


03 


73 






<D 


•H 


1 


Sh 


<7> 












CD 


CO 


CD 


s 


CD 


CD 




2 


> 


P 


P 


P 


C 


CD 




c 






£ 


CD 


P 


(0 


4h 


a, 




O 


P 


•r-( 


•H 


1 


-H 


C 


u 


a) 








rH 


CO 


CD 


s 


•H 




s 


C 


co 


-Q 


^ 


c 


-H 


0) 


73 






0) 


CP 


CD 


CD 





P 




§ 


•H 


c 


XX 





•H 


a 


•H 


rH 






.* 


c 


3 


U 


u 


c 




o 


o 


0) 


<0 


•H 


5 


3 


Sh 


o 






o 


o 


rH 


x: 


u 


CD 




u 


1-0 


1/1 

en 
c 

(0 

-p 

•H 
P 

u 


ft 

(0 
■H 
rH 

o 

4-1 
■H 
73 

C 


s 

•H 

rH 

o 

4-1 


Eh 

co 

•H 

rH 


hJ 


OQ 


o 

to 

rH 
rH 

>t 

xt 
a 






CU 

(0 

c 

(0 


■H 


r-1 


CQ 


in 


< 


U 

CO 
QJ 
73 

•H 



Sh 

3 








-H 


3 


•H 


•H 




a 


o 


rH 


IN 


Sh 


co 






CO 


•H 








c 


P 


3 


.Q 




CO 


01 


=tfc 


=**= 


cu 


•H 


. 




c 


.C 













C 


3 






>i 






e 


Sh 


a 




CD 


tt 






(0 


03 


!-i 


0) 


rH 




'— » 


u 


73 


73 


CO 


P 


CO 


• 


o 









rH 


P 




P 


o 


* 


o- 





0) 


a) 




CO 




a 


CO 






CO 


3 


CO 


(0 




> 


CL 


— - 




•H 


•H 


CO 


3 


c 


CO 


to 


(0 




w 


73 


•H 


•H 


e 




co 




CO 


4H 


4-1 





rH 







Sh 







M 


•H 


Sh 


!h 


3 


05 




(0 


•H 


•H 


•H 





(0 


C31 


(0 


3 


rH 




u 








(0 


-H 


•H 


en 


•H 


CQ 


P 


P 


(0 


CU; 





73 


a 


x: 




w 


CO 


a) 


rH 


73 


P 


3 


^ 


O 


c 


c 


rH 




QJ 


•H 


Sh 





• 


cu 


•H 


(0 











C 


n 


rH 


a) 


0) 


o 


CO 


O 


p 


3 


o 


73 


1/1 


> 


E 


p 


e 


(0 


■H 


(0 


x: 


73 


73 


p 


3 


Sh 


CO 


Q. 


£ 


CD 






03 


o 


CO 


rH 


a 


u 


a 


-H 


•H 


>■ 


C 


73 


•H 




CD 


3 




• 


£ 


Sh 


CD 


(0 


3 


XI 


cu 


c 


c 


J= 


•H 


C 


U 


. 


Sh 


C 

-H 
P 




< 


u 


U 


a 


u 


J 


C/l 


Eh 


3 


3 


cu 


Cm 


<: 


< 


< 


W 




































c 




































o 




































a 

i 
i 
























CD 
CO 
CD 
O 
CO 












H 
1 




CD 




















u 




CD 










in 




03 
CD 




















(0 

rH 


(0 
<D 


ai 

to 








CD 


H 























o 


U 


CD 








rH 


s 


(0 




















p 


(0 





2 


z 


^ 


.Q 


r 3 ? 


ja 




















>l 


c 


(0 








(0 


Pn 


CO 




















x: 


•H 











Eh 




fa 




















CU 


CU 


CU 









88 



























r-\ 








rH 
















to 




m 










o 








o 












CM 




«* 




o 










• 








• 












r*. 




• 




• 










o 








o 










en 


• 


>£> 


O 


r-l 


CN 


^ 


CTi 


"tf 




V 




CM 


rH 


V 


r^ 


rH 






VO 


o 


H 


\ 


O 


V. 


00 


^f 


IT) 




\ 




O 


o 


\ 


o 


O 




o\° 


• 


\ 


• 


H 


• 


H 


• 


• 


• 




•-i 




• 


• 


H 


■ 


• 






o 


H 
O 

O 

Q 


CM 
*3* 


O 

O 
V 

Q 


o 


O 

O 

V 

Q 


H 


O 


rH 




o 
o 

V 

a 

PC 




o 


o 


o 
o 

V 

a 

ft 


o 


o 

V 




W 


Q 


O 


Q 


O 


Q 


o 


Q 


ffi 


p 




o 




a 


a 


o 


a 


ffi 




H 


« 


\ 


tf 


X. 


c< 


\ 


ft 


CO 


« 




\ 




« 


c< 


\ 


c< 


CO 




H 


O 


K 


O 


K 


o 


X 


o 


rtl 


o 




w 




o 


o 


K 


o 


< 




CO 




en 

< 


01 

to 

O 

(0 
-H 

to 


CO 

e 

CJ 


01 
01 
fO 

u 
o 

c 
to 

•H 

73 


co 

< 








r< 
O 


CO 

rf! 

(0 

rH 
r-\ 
-H 

M 

fO 




(0 

-p 

cu 

CT 1 

•H 
^1 

•H 
Cm 




CO 

< 








W 






CQ 


(0 

ft 


c 

H 






>1 




O 

o 


ft 

fO 




e 

CU 












5! 






(0 


01 






73 


r) 




& 


01 




P 












£ 






iH 


fO 


73 




(U 


M 


cu 


i 


u 




CO 














s 


e 


o 


Cm 


0) 




CU 


QJ 


o 


CO 


to 
















s 


3 


3 







73 




> 


.Q 


b 


Xi 


CO 




Xi 












o 








(0 


4h 


•H 




-P 


X 




1 






p 


cu 










s 


■H 


-H 


03 


4H 


01 




c 


O 


Ih 


73 73 




o 


ft 










j§[] 


C 


c 


c 


•H 


ft 




•H 


(0 


o 


rH 


<-\ 




o 


(0 










o 


(0 


to 


CU 


P 


o 




o 


iH 


o 


•iH 


-H 




e 


r< 










u 


Cm 


Cm 


Cm 


CO 


J 




r} 


CQ 


Cm 


s s 




CO 


a 


































to 
























g 










rJ 




c 
























;3 










o 




fO 
























73 














•H 
























G 










to 




c 




















6 




3 






01 




-p 




•H 




















3 




O 




• 







(0 




t-i 
















H 




P 




CD 




ft 


■H 




rH 




o 
















=tfe 




fO 

-p 


3 


01 


73 


01 


rH 




01 


o 




(0 
















• 


CM 





0) 


6 


0) 


(0 


4H 


a) 


■H 







(0 














a 


=*fc 


c 





3 


•H 


r-t 


a) 


u 


>H 






U 












co 


01 






fO 


Jh 


4-1 


r-{ 


c 


a) 


3 


CO 


CO X 


• 










w 




• 


e 


-p 


-P 


■H 


0) 


c» 


p 


(0 


o 


p 


CO 


ft 










H 


6 


ft 


3 


0) 


01 


P 


c 


o 






d 


d) rH 


01 










u 


3 


01 


rH 


01 


fO 


C 


o 




(0 


X 


CO 


d 


CP 










• 


w 







(0 




si 


0) 


tp 


01 


•rH 


(0 rH 


O 1 




01 








73 


Cm 


•H 


• 


ft 


• 


en 


73 


>1 


2 


73 


rH 


en 


-H 


• 


-H 








CU 


1/1 


c 


G 


01 


01 


U 


•H 


rH 


■Q 


O 


•rH 




r4 


rJ 


P 








3 




to 


to 


fO 


fO 


o 


c 


o 


3 


•H 


£ 


• 


-H 


fO 


-H 




in 


r*. 


C 

-H 
P 




Cm 


Cm 


Cm 


Cm 


CO 


3 


Cm 


Pi 


Q 


CO 


CO 


Cm 


> 


> 




=*= 


=**= 


































I 


i 


c 






































o 
































H 


CM 


ix) 


u 

1 
















CU 
fO 
Q) 
















0) 

-H 


TS 
CU 

•H 


73 
QJ 

•H 


H 
















CJ 
fO 


0) 


0) 
(0 


cu 

CO 




CU 
CO 




CU 


4H 
-rH 


4H 
•H 


4H 
•rH 


in 


X 


CU 












c 


(0 


a) 


0) 




CU 




(0 


P 


P 


-P 




h-l 


CO 












o 


0) 


u 







O 




CU 


c 


c 


c 


(U 


H 


01 












CP 





to 


fO 




CO 




o 


cu 


a) 


cu 


r-\ 


s 















>1 


(0 


•H 


<-i 




C 




(0 


ts 


73 


73 


£t 


<j 


to 












rH 


01 


A 


-H 




u 




-p 


-H 


-H 


•H 


(t3 


Ph 


o 












o 


o 


CJ 


g 




CJ 




•H 


c 


C 


C 


En 




Cm 












Cm 


a 


« 


CO 




B 




> 


ID 


D 


D 



e 
o 
u 



89 



CO u 
P o 

C 4-1 
<0 

TJ 

>i 0) 

£ N 

>i 

TD H 

(D <o 

P c 

u (0 
CD 

rH 10 
rH CD 

O -H 

O U 

(1) 

co a 
0) co 

•h 

o - 

CD* 

a 
ca 

TJ CO 
CD C 
<1) O 

co -h 
p 

4-> (0 
C > 

tO <0 

n o 
c x 

3 CD 

(0 Q 

Dh 

P O 

CO 

o 

e 



U 

3 

o 

CD 4-1 
P CD 

4-1 P 

o 

X3 

CO c 

CD <0 
D> 

(0 CO 

P C 

C O 

CD -H 

CJ P 

r4 (0 • 

CD > P 

CU (0 C 

U CD 

X P 

CD C 
O 

X u 



CM 

I 

in en 

<: 



CD 

rH 

Eh 



to 

c 
o 



c 

CD 

> -H 

CD P 

CO -H 

CD P 



cm 

X 
en 



CN 
CM 

X 

en 

< 



en 
o 

CM 
X 

Cfi 

< 



E-l rH 

CO CM 

W DC 

s en 



o 

CM 

X 
en 



oo 

r-i 

X 

en 



IT) 
rH 

K 

en 



£3 

P 



en 

w 

H 

u 

w 
cu 
en 



CM 


CM 


•* 


• 


• 


• 


ro 


<n 


r-i 




<y\ 


CX) 




• 


• 


O 


in 

en 


n 



ai 



en 

en 



in 



en 



^> 



en 
en 



en 
en 



CM 



oo 
en 



o 

v 



CQ 
n 

(M 

a 
a 
o 



en 


rH 


r-l 


r-t 




m 


en 


■* 


rH 


• 


• 


• 




• 


• 


• 


a 


r-i 


CM 


o 


o 


tM 


>£> 


in 


a 


eO 


CM 


V 










o 

















r^ 



CO 




* 


rO 




co 


r— 1 


•X 


c 


o 


CO 


CD 


4-1 


CD 





•rH 


T3 


CO 


c 


-H 


(0 


■H 





u 


£ 


u 


3 


m 


-rH 


a 


M 


r4 


u 


en 


CD 


3 



en 


rc 




■H 


<-\ 


tO 


CO 


o 


T3 


U 


-H 


•H 


o 


P 


P 


>, 


tO 


CO 


p 


u 


•H 


-rH 


CD 


u 


04 


u 


< 



O ex) 

V 



<M 
^0 



in 



in 



Q 


o 


CM 


n 


CM 


CC 




CM 




<tf 


o 










^o 




CO 


o 


CO 


rH 




• 


• 


• 


Q 


o 


r-i 


rH 


fN 


« 








[^ 


o 











CM 



a 

CO 

Cfl 

3 

M 
CD 

a 
>i 
u 



CO 

to 

T3 

-H 

P 
CO 
-H 

!h 



eoi 

CD 

TJ 
■H 

O 

Sh 

■H 

CM 

o 

rc 
O 

r— I 

a 
o 

o 

e 

CD 
Sh 

W 



g 

P 
(0 
P 

o 

c 

g 

3 

r— I 

tO 
Qi 
CO 
(0 

a, 



r^ 



CM 






ro 
O 



CM 
CM 



en 



CD 
(0 
CD 
O 

tO 

o 
a, 



p 

c 

CD 
X5 

•H 

c 

D 



ft 

CO 



CD 
C 

o 



* 
CO 
CD 

rH 

CD 
P 

f0 
-H 
X3 

O 

■i-\ 

a 



90 



Table 5-3. Seeds collected from the seed bank 
samples at ASH and ORD. 



SPECIES 


SITE 


#COLLECTED 


Froelicha floridana 


ORD 


14 


Opuntia humifusa 


ORD 


1 


Cyperus sp. 


ORD 


109 


Ceratiola ericoides 


ASH 


2165 


Chamaecrista nictitans 


ORD 


3 


Galactia volubilis 


ASH 


1 


Lupinus sp. 


ORD 


1 


Pinus palustris 


ORD 


3 


Aristida sp. 


ASH/ORD 


11/153 


Aristida purpurascens 


ASH 


7 


Eremochloa ophiuroides 


ORD 


274 


Panicum sp. #1 and #2 


ORD 


457 


Paspalum notatum 


ORD 


32 


Paspalum setaceum 


ORD 


12 


Sorahastrum secundum 


ORD 


32 


Polyaonella sp. 


ORD 


23 


Diodia teres 


ORD 


19 


Unidentified 


ASH 


5 


Unidentified 


ORD 


610 



91 



-P 

C 
Q) 

CO • 
QJ CO 
Sh P 

a c 
^ e 



qj 
u 

3 

co 



OJ 
Q) 

CO W 

a) 

P 3 
C rH rc 
fT3 (T3 0» 

-o > e 

c 

3 QJ 

ja • p 

ns co rc 
-P u 

P CO -H 

CO QJ rH 

O C ft 

e 

Q 

aj a 
xi o m -P 
p ox 

T5 <w QJ 
<H C P 
O « 0) 

i-t a) 
X I J2 QJ 
■H CO (0 03 

LO <C rH »^ 



qj 



M-l C 
O -H 

CO 73 

O C 

•H 3 
P O 
CO <4H 
•H 

M co 
OJ Q) 

P -H 

u o 
rc Q) 

s-i a 

03 CO 

O H 
fO 

rH C 

03 O 
C -H 
O P 

•H -H 
■P 

•H 
>H 
P 

3 
C 

0) 

,c 

EH 



I 

IT) 



03 P 

> C 

(0 QJ 

e 

CO -H 

(0 ^ 

5 Q) 

a 

r-< X 

rc QJ 

•H 

k OJ 

QJ U 

P -H 

rC O 

g £1 

U 

P 

C T3 

Q) Q) 

-rH d) 

u co 

H 

■W QJ 

4H £ 

3 P 
CO 

C U 

■H O 

1H 

a) 

U TS 

Q) 0) 

CO -C CO 

rc S 3 

H P CO 
H ft 10 

QJ QJ 5 



QJ CO 

(C 



U 
X 
QJ 



X! 

rC CO QJ 



CO 

rH 

U 

CO 03 



QJ 



QJ !T> C 

-H 03 -rH 

O rH rH 

QJ QJ 

a > • 

CO rc U| 



m 

Eh 



CO 

u 

< 

a 

> 

o 



co m 
u 



w 

Q 
D 



o\° Cm 



a 

H 



o\° 



a e-< 

Q Eh 
o\° kC 



CO Q 

W W 

H W 

a co 
o 

3 a 

< w 

u ft 



Q S 
w — 
w 

CO 



co 

w 

H 

u 

w 

CO 



ID 


CO 


CO 


'd* 


ID 


IT) 


r- 


^ 


>X) 


en 


CD 


IX) 


o 


X) 


^r 


CO 


rH 


r-i 


co 


m 


o 


ro 


^r 


CM 


-* 


CM 


CD 


■* 


■* 


^r 


r-< 


m 


rji 


CO 


CO 


r> 


ID 


<X> 


r> 


CO 


(D 


CO 


m 


t 



«5 



CM 



00 



CM 



CM 

cd 



ryi 

oo 



o 



CM 



CO 

CO 
3 
in 
QJ 

a 
>i 
u 



o 

CD 



in 
n 



ID 



CD 



O 



CO 

o 



CO 
QJ 
T3 
■H 
O 
U 
■H 
Sh 
QJ 



CD 
CD 



iX3 
CM 



■H" 

CO 



X3 



CM 
CD 



CO 
CD 






1X3 



rH 



01 



o 
cm 



in 



co 



co 



in 



CM 
CM 



X3 



>X3 



CD 



CO 

o 



CD 



co 



X3 
LD 



LT) 



in 

>X3 



CM 



X3 
1X3 



o 



o 

ID 



CM 



ID 



1X3 



X3 



■H< 

CD 



n 

r^ 



CM 
CD 



CM 

r> 



CD 



00 
IX) 



n 

CD 



00 



CM 
CD 



ID 



CD 



co 

CD 






CM 



vo 



X3 



r^ 



o 



*XJ 

o 



ID 



03 



>i 

ft 
O 
CO 

u 
o 



ft 

CO 



(C 




rC 




l-{ 


rc 


■H 


03 





■H 


CO 


T3 


•rH 


P 





•H 


P 





u 


P 


rc 


(0 


£ 


CO 


Sh 


rH 


a 


•H 


QJ 


(C 


0) 


rH 


U 


O 


EH 


<c 



CO 

c 

QJ 
O 

CO 

rC 

rH 

3 
ft 

rH 

3 
ft 



<\ 



rc 

CM 

CO 

03 



o 

00 



ID 



CO 
CM 



X) 



CD 
CO 



CM 
CD 



<3* 

X3 



CM 

n 





co 






3 






■H 






rH 






o 


CO 


e 


4H 


QJ 


3 


QJ 


U 


QJ 


C 


OJ 


O 


3 


p 


rC 







P 




rC 


QJ 


CO 


-rH 


CO 


3 


T3 




-Q 





• 


3 


•rH 


Ph 


K 


Q 



CD 



ID 
CM 



CM 



CM 



>X) 



CD 



ID 
iX> 



CM 



CM 



CO 
■H 
CO 
ftl 

o 

c 
o 
p 
o 
Sh 
u 



CHAPTER 6 
NEST DISPERSION AND INTER-NEST AGGRESSION OF 
POGONOMYRMEX BADIUS AND THE EFFECT OF HOST POPULATION 
STRUCTURE ON SPIDER INTEGRATION INTO COLONIES 

Introduction 

The spatial arrangement of ant colonies in a habitat 
may provide insight into intraspecif ic competition for 
shared resources . Many workers have found that 
intraspecif ic ant colonies as well as interspecific 
colonies of ecologically similar species are overdispersed 
(de Vita 1979, Levings and Traniello 1981, Harrison and 
Gentry 1981, Cushman et al. 1988). This overdispersion 
probably reflects interference and exploitative 
competition (Holldobler and Wilson 1990). 

Resource requirements of intraspecif ic colonies will 
overlap to a greater extent than resource requirements of 
interspecific colonies found in the same habitat. The 
more limited the resource, the greater the advantage of 
niche partitioning (i.e., increasing colony spacing). 
Therefore, it would be adaptive for colony members to be 
able to differentiate between competitors ( intraspecif ic 
neighbors or interspecific neighbors who share a limited 
resource) and non-competitors (members of more distant 
colonies or interspecific foragers of a non-competitor). 
In fact, Gordon (1989) found that western harvester ants 
of the species Pogonomyrmex barbatus can distinguish 

92 



93 



between a neighboring colony's foragers (competitors) and 
a more distant colony's foragers (stragglers or lost 
foragers, i.e., non-competitors). When neighboring ants 
were encountered in the foraging trail of a nest, foraging 
intensity dropped to a greater extent than when ants from 
a more distant colony (strangers) were encountered . in the 
foraging trail. Gordon cited foraging intensity as a 
measure of the colony's reaction to the presence of alien 
ants. Thus, she concluded that encounters with neighbors 
deterred foraging to a greater extent than encounters with 
strangers. 

Many of the studies investigating ant colony spacing 
and its relation to competition have focused on seed 
harvesting ants of the genus Poqonomyrmex whose colonies 
are often overdispersed (de Vita 1979, Levings and 
Traniello 1981, Harrison and Gentry 1981). Inter-colony 
worker-worker aggression has been reported in harvester 
ants (Holldobler 1976, de Vita 1979, Gordon 1991) as has 
aggression between workers and non-colony queens 
(Holldobler 1976). Inter-colony worker-worker avoidance 
seems to be another common method of ensuring maximum 
colony spacing (Harrison and Gentry 1981, Gordon 1991). 

The purpose of this study was to investigate the 
factors that affect colony spacing of the Florida 
harvester ant, P. badius and to explore what effect this 
underlying population structure has on the integration of 
the myrmecophilic spider, Masoncus poqonophilus , into the 



94 



ant colonies. If, as Gordon suggested (1989), harvester 
ants can distinguish neighbors, or potential competitors, 
from strangers, or non-competitors, due, perhaps, to the 
greater frequency of encounters with foragers from 
neighboring colonies, then this difference may be 
reflected by differences in the level of aggression shown 
by ants towards neighbors versus strangers introduced onto 
their mounds. In addition, patterns of aggression towards 
neighboring colonies should influence colony dispersion if 
dispersion is a mechanism for reducing competition between 
nests. If the frequency of encounters between foragers 
from neighboring nests is high, then aggression towards 
neighbors introduced onto the nest mound should also be 
high and the population should tend towards even 
dispersion as a mechanism for reducing competitive 
interactions. Conversely, if encounters between foragers 
from neighboring nests are low, perhaps due to 
super-abundant resources or resources that are clumped 
very close to each nest, then aggression towards neighbors 
introduced onto the mound should be low and dispersion of 
nests should be random. 

Ants distinguish nestmates from non-nestmates 
through chemically-based phenotype matching (Holldobler 
and Wilson 1990). These chemical recognition cues can be 
genetically derived, acquired from the queen, acquired 
from other nestmates, and/or derived from the environment 
(Carlin and Holldobler 1986, Holldobler and Wilson 1990). 



95 



Workers antennate each other and match the cuticular cues 
of the encountered individual with a sensory template 
based upon a learned set of cues likely to be possessed by 
nestmates (see Holldobler and Wilson 1990 for an extensive 
coverage of this topic and associated references). 

Many myrmecophilic arthropods take advantage of this 
system of nestmate recognition among their hosts by 
mimicking host colony odor (probably through passive 
absorption into the cuticle). Such chemical mimicry has 
been documented in a myrmecophilic beetle (Vander Meer 
1982), a parasitoid wasp (Vander Meer et al. 1989), 
syrphid fly predator of a formicine ant (Howard et al. 
1990), and another syrphid fly predator of a myrmicine ant 
(Howard et al. 1990). If the myrmecophilic spider, M. 
pogonophilus also absorbs the colony odor of its host 
ants, then spiders introduced onto the mounds of 
neighboring (non-host) nests should be recognized and 
treated as intruders whereas spiders re-introduced onto 
the mounds of their host colonies should not be treated 
with aggression. If true, then host population structure 
and aggression between neighboring colonies would prove a 
formidable barrier to dispersing spiders. 

Materials and Methods 
Site Descriptions 

This study was conducted at two different sites in 
North Florida: Archer Sandhills in Levy County and the 
Ordway-Swisher Preserve in Putnam County. These sites 



96 



differed both in floral composition as well as in overall 
habitat structure (see Chapter 5). The Archer Sandhills 
site was dominated by dense stands of Florida Rosemary 
bushes ( Ceratiola ericoides) and scattered turkey oaks 
(Querelas laevis.) . Rosemary seeds made up between 94 - 99% 
of the seeds found in the nest granaries (see Chapter 5). 
In contrast, the Ordway Preserve site was an old field 
habitat characterized by many species of grasses and herbs 
with no dominant species (Franz and Hall 1991). This more 
varied resource base is reflected by the seeds stored in 
the nest granaries (see Chapter 5). 
Nest Dispersion 

Nearest neighbor distances of P. badius nests were 
measured in a 60 X 60 m plot at Archer Sandhills and a 110 
X 110 m plot at the Ordway Preserve. Bordering areas 
around the demarcated plot were searched for other P. 
badius colonies that might be nearest neighbors to those 
within the plots to avoid bias from edge effects (Krebs 
1989). Distances were accurate to within 0.5 m. The 
Clark-Evans Nearest Neighbor Method was used to determine 
whether the dispersion pattern was random, clumped, or 
overdispersed (Krebs 1989). 
Inter-Nest Aggression 

To test the hypothesis that P. badius workers can 
distinguish potential competitors (near neighbors) from 
non-competitors (workers from more distant nests), 
cross-introduction experiments using four focal nests at 



97 



Archer Sandhills and four focal nests at Ordway Preserve 
were performed. For each of these focal nests, the 
behavioral responses of ants to individual P. badius 
workers from either the nearest neighbor nest or a distant 
nest placed on the mound of the focal nest were recorded. 
Each of the focal nest/stranger nest pairs had at least 
one other P. badius nest located between them. For each 
pair of nests, 20 total cross-introductions of individual 
ants were performed; i.e., 20 neighboring ants and 20 
strangers were introduced onto the mound of the focal 
nest. The order of introduction onto the mound was: 1) 
neighboring ant, 2) ant from the distant mound, and 3) a 
control ant from the focal nest that was collected and 
reintroduced onto the mound. The reintroduction of a 
nestmate back onto the mound was used to control for the 
possible effects of general alarm, or aggression on the 
part of the focal ants in response to disturbance during 
the experiment. Behavioral responses were recorded as 
non-aggressive if the ants only antennated each other but 
showed no other interest or if the focal ants oriented 
suddenly towards the introduced ant with their mandibles 
agape but did not proceed to bite or attempt to remove the 
introduced ant from the mound. This latter behavior 
(sudden orientation with mandibles agape) was recorded as 
non-aggressive because it was often shown towards a 
control ant in response to the control ant's ( nestmate 's) 
agitation (and presumed release of alarm pheromone) . 



98 



Behavioral responses were recorded as aggressive if, after 
antennation, the ants bit or grappled with one another or 
if the host ants picked up the introduced ant with their 
mandibles and proceeded to remove the foreign ant from the 
mound. Introduced ants were removed only after 
antennating at least 3 host ants (unless an earlier 
encounter resulted in biting and grappling) . Vials used 
to capture and introduce ants onto mounds were wiped out 
with an ethanol-soaked cloth to reduce the chances of 
passive absorption by ants of non-colony odors. Each of 
the eight sets of neighbor versus stranger pairs was 
analyzed using a contingency table analysis (Sokal and 
Rohlf 1981). Two linear regression analyses were also 
performed to determine if aggression towards neighbors and 
towards strangers decreased with increased nest spacing. 
Distances between neighbors or between strangers were used 
as the independent variables with the arcsin transformed 
proportion of aggressive responses as the dependent 
variables. Transformation of the dependent variable was 
necessary since it was not a continuous variable 
(Kleinbaum and Kupper 1978, Sokal and Rohlf 1981). 
Response of Ants Towards Introduced Myrmecophilic Spiders 

Masoncus pogonophilus spiders were removed from 
their host colonies and subsequently reintroduced to those 
same colonies and/or to neighboring colonies known to be 
aggressive towards the host colony. If the spider had 
been separated from their host colonies for more than one 



99 



week, they were housed in a vial closed only with metal 
screening and cheese cloth (to allow free air — and 
odor — flow) and the vial placed in a closed plastic 
shoebox housing 60 - 70 ants from the host nest as well as 
material from the mound of the host nest which Gordon 
(1984) has shown is saturated with colony odor. The host 
ants and the spiders were kept together in this manner for 
at least two days to allow possible reabsorption of the 
colony odor into the spider's cuticle. After this time, 
the spiders were reintroduced to the mounds of the host 
colony and/or introduced to mounds of neighboring 
colonies. All spiders housed with ants were placed in 
fresh vials before introducing them to an ant mound to 
ensure that any colony odor lingering on the vial, itself, 
would not trigger a behavioral response from the ants. 

A total of seven spiders were reintroduced to host 
nests and 11 spiders were introduced to a neighboring 
nest. Of the 11 spiders introduced onto foreign nests, 
three had previously been re-introduced to their host 
nests and subseguently recaptured. The behavior of the 
ants to the spider was recorded as non-aggressive if the 
ants antennated or contacted the spider but showed no 
reaction to the presence of the spider. The behavior of 
the ants was recorded as aggressive if the ant attacked 
and attempted to bite the spider with its mandibles or if 
it made a sudden movement toward the spider with its 
mandibles agape. Care was taken to ensure that the ants 



100 



showed no visible signs of agitation or alarm prior to 
introduction of the spiders. The latter behavior on the 
part of the ants (sudden orientation with mandibles agape) 
was regarded in this experiment as aggressive because it 
was a distinct change in the behavior of an ant that, 
prior to encountering the spider, was antennating the 
midden or going about some other nest maintenance task. 

Results 
Nest Dispersion 

Eleven P. badius nests were found in the 60 X 60 m 
plot at Archer Sandhills resulting in a nest density of 
0.003 P. badius nests/m 2 . The mean nearest neighbor 
distance was 11.45 m (+ 2.35). The index of aggregation, 
R (Krebs 1989), was 1.27 which was not significantly 
different from 1 (0.05 < p < 0.1 for a one-tailed test). 
Therefore, the dispersion of P. badius colonies at this 
site was random. 

Sixteen P. badius nests were found in the 110 X 110 
m plot at the Ordway Preserve for a nest density of 0.001 
nests/m 2 . The mean nearest neighbor distance was 20.00 m 
(+ 9.29). The index of aggregation was 1.45 which was 
significantly different than 1 (p < 0.005 for a one-tailed 
test). Therefore, P. badius nests at the Ordway site are 
over-dispersed . 
Inter-Nest Aggression 

At Archer Sandhills, three of the focal nests were 
significantly more aggressive toward workers from 



101 



neighboring colonies than workers from distant colonies 
(X 2 = 4.51, 8.18, and 10.23 respectively, p < 0.05, Table 
6-1A) . However, one focal nest (#26ST in Table 6-1A) was 
relatively unaggressive towards workers from both the 
neighboring nest as well as toward workers from the more 
distant nest (X 2 = 0.11, p > 0.5). At Ordway Preserve, 
none of the focal nests showed higher aggression towards 
neighbors than towards strangers (all X 2 < 2.0, p > 0.2, 
Table 6-1B). The closer the eight focal colonies were to 
their nearest neighbors, the more aggressive focal ants 
were towards foreigners introduced onto their mounds (r 2 = 
.567, t = -2.804, d.f. = 6, p = 0.031, Fig. 6-1). However, 
distance was not a good predictor of aggression towards 
strangers introduced onto the mounds of the focal colonies 
(r 2 = 0.175, t = -1.131, d.f. = 6, p = 0.301, Fig. 6-2). 
Response of Ants Towards Introduced Myrmecophilic Spiders 

Of the seven spiders reintroduced to their host 
nests, two (29%) elicited an aggressive response from the 
hosts. Of the eleven spiders introduced to aggressive 
neighboring nests, seven (64%) elicited aggressive 
responses from the non-host ants (Table 6-2). One of the 
spiders was attacked and killed by a non-host worker. The 
difference in behavioral responses elicited by host ants 
versus non-host ants, although suggestive, was not 
significant (X 2 = 0.935 p = 0.334, Table 6-2). 



102 



Discussion 

Dispersion of P. badius nests at these two sites is 
not a good predictor of inter-nest aggression. At Ordway, 
the nests are over-dispersed. If this over-dispersion 
was triggered by more frequent encounters between 
neighboring nests, i.e., higher competitive interactions 
between neighboring nests, as suggested by Harrison and 
Gentry (1981), then neighboring nests should be more 
aggressive towards one another than nests spaced further 
apart. This is not the case: the four focal nests at 
Ordway showed less aggression overall towards foreign ants 
(neighbors or strangers) introduced onto their mounds than 
three of the four focal nests at Archer Sandhills. 

The greater aggression of focal ants at Archer 
Sandhills towards neighbors than towards strangers (at 
least for three of the four focal nests) would suggest a 
higher frequency of encounters of neighbors (competitors) 
than of strangers (non-competitors) as suggested by Gordon 
(1989). If these competitive interactions trigger 
colonies to migrate away from one another to reduce the 
overlap of foraging ranges (as suggested by Harrison and 
Gentry 1981) then nests at Archer should be overdispersed. 
This is not the case. 

However, the linear regression analyses suggest 
that, at least for neighboring nests (potential 
competitors), distance is a good predictor of aggression 
(Fig. 6-1). The aggression of the four focal nests at 



103 



Ordway towards their neighbors may be low simply because 
these nests are spaced further than the neighboring nests 
at Archer. This is further supported by the fact that the 
focal nest at Archer spaced furthest from its nearest 
neighbor (#26ST, Table 6-1) is the least aggressive of any 
of the four focal nests towards their neighbors. 

The lack of correlation between aggression of focal 
ants towards foreigners from distant colonies (Fig. 6-2) 
and the relatively low proportions of aggressive responses 
of focal nests towards strangers (Table 6-1) suggests 
that, once a colony is outside the foraging range of the 
focal nest, distance is no longer a good predictor of 
aggression. It suggests that if the foraging ranges of 
two nests do not overlap, it is irrelevant to what extent 
they do not overlap. However, it is important to note 
that three of the eight focal nests (#26ST, #6, and #R1 , 
Table 6-1) reacted somewhat more aggressively towards 
strangers than towards neighbors. Nevertheless, of these 
three, two (#26ST and #R1) showed little aggression 
towards any ant, neighbor or stranger, placed on the mound 
and none of the three were significantly more aggressive 
towards strangers than towards neighbors. 

The difference in dispersion patterns at these two 
sites can be explained both by the differences in nest 
densities as well as by differences in habitat structure 
and resource availability. The density of nests at Ordway 
(0.001 nests/m 2 ) is three times less than the density of 



104 



nests at Archer Sandhills. At Archer Sandhills, there is a 
superabundance of rosemary seeds available to the ants; 
98-99% of the seeds in the seed bank are rosemary seeds 
(Chapter 5). The higher density of nests at Archer 
Sandhills may be explained by the availability of this 
predictable and extremely abundant resource. Rosemary 
seeds are also particularly high in lipid content (Chapter 
5) and are, therefore, not only an abundant resource but a 
nutritionally high quality resource. At Ordway, there is 
considerably more variation in the seeds available to the 
harvester ants and more variance in the nutritional 
quality of the seeds available at any given season 
(Chapter 5). This seasonal variation in resources 
available to the ants at Ordway may increase the mortality 
of incipient P. badius nests, thus keeping the density of 
nests low (as compared to the Archer Sandhills 
population). At Archer Sandhills, the apparent 
availability of clear, open areas of sand for nest 
construction is less than the apparent availability of 
areas for nest construction at the open field habitat of 
Ordway due to the density of rosemary bushes at Archer and 
the lack of shrubs or bushes at Ordway (P. badius does not 
establish nests beneath trees or bushes — it requires open 
areas with well-drained soils). If the habitat at Archer 
Sandhills was more open, I contend that the nest 
dispersion would be more even. 



105 



The "passivity" of nests at Ordway towards ants from 
neighboring mounds may simply be a reflection of the low 
density and increased spacing of colonies in this habitat. 
In other words encounters, even between the closest nests, 
are probably infreguent. 

The results of the spider introduction experiments 
are ambiguous. Although some non-host ants reacted guite 
aggressively towards the spiders (one ant even killing a 
spider), many others showed no reaction to the 
myrmecophile. And not all the host ants reacted to the 
spiders as if they "belonged" on the mound; a few of the 
hosts even reacted aggressively towards them. The data do 
not preclude the possibility that the spiders absorb host 
colony odors. However, if they do, there may be a wide 
variation in the extent to which the hydrocarbons are 
absorbed into the cuticle of different individuals. I 
believe the spiders become integrated into colonies, not 
through chemical mimicry, but by being guick and sneaky. 
They are adept at evading the ants, moving rapidly away 
when contacted by a leg or an antennae. When placed on a 
P. badius mound, they wander around apparently aimlessly 
until they reach an area where the sand slopes downward 
towards the colony entrance, at which point they move 
directly downward and enter the colony. They move into 
the entrance when no ants are either entering or leaving 
the nest and usually enter upside down, walking on the 
ceiling of the entrance tunnel rather than on the floor 



106 



(where they would be more likely encountered by a worker). 
Once inside the nests, where they are surrounded by air 
saturated with the colony odor, absorption of colony odor 
may then play a role in maintaining the integration of the 
spiders with the hosts. If this is the case, a dispersing 
spider arriving on the mound of. a neighboring colony may 
be recognized as an intruder and attacked before it can 
make its way to the nest entrance and sneak inside. 



107 



Table 6-1. Distances between neighboring nests and 
distant nests for each focal pair. The proportion of 
aggressive encounters out of the 20 total encounters 
between neighbors or strangers is presented as are the X 2 
values. A. Archer Sandhills nests. B. Ordway Preserve 
nests. 



COLONY 
PAIRS 



DISTANCE 
(m) 



PROPORTION 

AGGRESSIVE 

RESPONSES 



A. 25R/23R 


7.05 


.90 


4.51* 


25R/24R 


11.50 


.55 




2 6ST/OAK1 


14.80 


.30 


0.11 ns 


26ST/13 


23.00 


.35 




23/20 


5.85 


.70 


8.18** 


23/12 


20.75 


.20 




12/15 


11.80 


.85 


10.23** 


12/17 


34.10 


.30 




B. 27/26 


8.35 


.55 


0.00 ns 


27/37 


22.20 


.55 




19/20 


15.00 


.35 


0.12 ns 


19/10 


30.90 


.25 




6/14 


10.20 


.45 


1.64 ns 


6/UNK 


27.45 


.70 




R1/R2 


16.25 


.15 


0.16 ns 


Rl/18 


50.65 


.25 





significant at p<0.05 
significant at p<0.005 
ns not significant 



108 



Table 6-2. Chi-Square table showing the responses 
of host and non-host ants towards M. pogonophilus spiders 
introduced onto their mounds. 



HOST ANTS 
NON-HOST ANTS 



#AGGRESSIVE 
RESPONSES 



2 

7 



#NON-AGGRESSIVE 
RESPONSES 



5 
4 



109 



80-i 




DISTANCE (m) 



Figure 6-1. Linear regression of the distances 
between focal nests and their nearest neighbors and the 
arcsin transformed proportion of aggressive responses of 
focal ants towards their neighbors (r 2 = 0.567, t= -2.804, 
d.f . = 6, p = 0.031) . 



110 



o 

o w 

M Eh 

Eh Z 

PS W 

O > 

o 

OS W 

a. > 

M 

Q W 

w w 

2 W 

o a 



60 n 



50 - 



40- 



£ < 30 - 



20 



10 



- r - 
20 



- 1 - 
30 



- r - 
40 



i 
50 



- 1 
60 



DISTANCE (m) 



Figure 6-2. Linear regression of the distances 
between focal nests and strangers, or more distant nests, 
and the arcsin transformed proportion of aggressive 
responses of focal ants towards strangers (r 2 = 0.175, t 
-1.131, d.f . = 6, p = .301) . 



CHAPTER 7 
SUMMARY AND CONCLUSIONS 

As Holldobler and Wilson (1990) propose, an ant 
colony can be considered an isolated ecosystem. Arthropods 
that have evolved mechanisms for integrating themselves 
into this specialized community are greeted with a stable 
microclimate, abundant food, and protection from predators 
and parasites. This study has investigated two members of 
one such ecosystem: the myrmecophilic spider, M. 
pogonophilus . and its host ant, P. badius . I have 
determined some of the adaptations involved in this 
association and have raised many additional guestions. 

I demonstrated that M. pogonophilus is 
morphologically distinct from previously described 
congeners and is dependent upon the ant nest ecosystem 
(Chapter 2). These myrmecophiles deposit their eggsacs in 
depressions in the ceilings of the nest chambers and spend 
all stages of their lives inside the nests. Furthermore, 
when the host colony emigrates to a new nest site, the 
myrmecophilic spider moves with them. 

I rejected the hypothesis that spider populations 
within ant nests represent semi-isolated demes , or 
metapopulations (Chapter 3). Gilpin (1991) described 
metapopulations as isolated local populations whose 
heterozygosity is low because of decreased gene flow 

111 



112 



between deities. Masoncus pogonophilus , in contrast, 
shows high genetic variation within local populations 
(i.e., among spiders within nests) and low genetic 
variation between local populations indicating high rates 
of dispersal between neighboring ant nests within the 
lifetime of the host nests. 

I used the Random Amplified Polymorphic DNA, or RAPD, 
fingerprinting technigue to measure the degree to which 
populations of spiders were isolated. The RAPD technigue 
proved effective in testing guestions concerning the 
isolation of separate populations. However, it proved less 
useful in determining the extent of gene flow per 
generation, or NM (Wright 1951). Because RAPD markers are 
inherited as dominant alleles, it is difficult to estimate 
heterozygosity since banding patterns represent either 
homozygote dominants or heterozygotes (Welsh and McClelland 
1990, Williams et al. 1990). Without an adeguate measure 
of heterozygosity, it is difficult to estimate population 
subdivision, or F ST (Wright 1951). Lynch and Milligan 
(1994) described algorithms for measuring F ST with RAPD 
markers. However, their technigue reguired an assumption 
of Hardy Weinberg eguilibrium which was not a valid a 
priori assumption for my study. Their technigue also 
estimated heterozygosity using an assumption that "null" 
alleles (or absence of bands at a locus) represent 
recessive alleles. However, even Milligan and Lynch (1991) 
admit that "null" alleles may have multiple causes such as 



113 



loss of primer sites or insertions. Despite these 
limitations, RAPD's is a useful molecular technique for 
measuring the degree to which populations are isolated. 

The mechanism of gene flow or dispersal of spiders 
between local populations remains unknown. Laboratory 
experiments did not support the hypothesis that spiders 
were able to follow trail pheromones (Chapter 4). However, 
it may be that, although the host ants readily follow 
artificial and natural trails in the laboratory, the 
spiders require additional stimuli that are missing in the 
laboratory before they can cue in on these chemical 
signals. Alternatively, it may be that M. pogonophilus 
locates new colonies, not via trail pheromones, but by 
sensing airborne colony odors (osmochemotaxis) . The 
ability to locate new host colonies via airborne cues was 
documented by Holldobler (1969) for the myrmecophilic 
staphylinid beetle, Atemeles pubicollis . 

The dispersion of host nests in an environment may 
affect the ability of dispersing spiders to become 
integrated into new host colonies in that more distant 
colonies may be more difficult for spiders to locate and 
colonies located very close to the former host colony may 
be more aggressive towards the spiders. Aggressiveness of 
P. badius colonies towards one another is correlated with 
proximity of and, presumably, with increased competitive 
interactions between neighboring colonies (Chapter 6). 
Aggressiveness of ants towards the spiders depends on the 



114 



ability of the ants to recognize the spiders as intruders. 
I presented circumstantial evidence that dispersing spiders 
may be recognized as intruders by ants of neighboring, 
non-host colonies and may, therefore, absorb colony odors 
into their cuticles. However, the data are too ambiguous 
to strongly support this hypothesis (Chapter 6). 

The dispersion of P. badius nests seems to be 
influenced primarily by habitat structure and resource 
availability. However, interference competition between 
neighboring nests may play a secondary role in nest 
dispersion. Resource availability, itself, varies 
depending upon the floral composition of the habitat in 
which populations of P. badius are found. I presented 
detailed information about resource use by colonies of P. 
badius at two different sites in north Florida and showed 
that foragers are collecting seeds (their primary food 
source) based primarily upon availability rather than on 
some measure of seed guality (Chapter 5). These data 
suggest that this species of seed-harvesting ant is not 
following predictions of optimal foraging theory which 
state that organisms should choose resources based upon 
some assessment of the costs and benefits of collecting 
that resource (Krebs and Kacelnik 1991). 
Coevolutionary scenarios 

The host ant, P. badius is the only member of this 
seed-harvesting genus found east of the Mississippi (Cole 
1968). During the late Pliocine and early Pleistocene, 



115 



about 10,000 years ago, a continuous band of arid habitats 
linked Florida with western North America (Webb 1990). 
During this time, present-day western relicts such as 
western pocket gophers, Thomomys spp. , and scrub jays, 
Aphelocoma coerulescens coerulescens , reached Florida (Webb 
1990). It was probably also during this period that the 
ancestor of P. badius became established in Florida. The 
closest relative of P. badius is P. comanche whose range is 
contiguous but not sympatric with that of P. badius (Taber 
1990). P. comanche has been found in western Louisiana, 
Texas, western Kansas, western Oklahoma, and western 
Arkansas (Cole 1968). P. badius has been found only east 
of the Mississippi in Florida, Alabama, Mississippi, 
Louisiana, Georgia, North Carolina, and South Carolina 
(Fig. 7-1, Cole 1968). The development of extensive 
wetlands around the Mississippi basin by the 
mid-Pleistocene, due to the rise in sea level during 
interglacial periods, likely served to divide an ancestral 
population of Pogonomyrmex . If this is true, then 
allopatric speciation led to the evolution of P. comanche 
and P. badius . Therefore, P. badius could be considered a 
geologically young species. 

I had hoped to be able to construct a phylogeny of 
the spider genus Masoncus . However, too few specimens of 
the previously described congeners, M. arienus , M. 
conspectus , and M. dux, were available to make this goal 
feasible. However, from the collection locales of the few 



116 



existing specimens, I can propose two intriguing scenarios. 
Of the three previously described congeners, I was able to 
compare the morphological traits of two, M. arienus and M. 
conspectus . with the newly described species, M. 
pogonophilus . Of these two, M. pogonophilus most closely 
resembles M. conspectus in several features: the location 
of the cephalic pits, the shape of the embolic division, 
and the shape of a black-tipped process on the distal edge 
of the palpal tibia (Chapter 2). If this morphological 
resemblance reflects evolutionary relatedness, then M. 
conspectus may be the sister species of M. pogonophilus . It 
is also the only one of the three previously described 
congeners whose known range overlaps that of M. 
pogonophilus (see Fig. 7-2). One of the two scenarios I 
propose here is that a sub-population of the direct 
ancestor of M. conspectus established a symbiotic 
relationship with P. badius after P. badius itself had 
become established as a distinct species (i.e., after the 
Pleistocene allopatric speciation event). In other words, 
the morphological characters that distinguish M. 
pogonophilus from M. conspectus were due to genetic drift 
and evolved after a sub-group of the ancestor of these two 
species became established as a symbiont inside the nests 
of P. badius . 

If, however, the symbiotic association between ant 
and spider became established prior to the allopatric 
speciation event that resulted in the evolution of P. 



117 



comanche and P. badius, then excavations of the nests of 
extant P. comanche might very well reveal a fourth member 
of the spider genus Masoncus . Unfortunately, very little 
information is available in the literature concerning the 
natural history, nest structure, or symbiotic associates of 
P. comanche . 

If the first scenario is true, that the symbiotic 
association between P. badius and M. poaonophilus was 
established after P. badius had speciated, then the lack of 
certain integrative mechanisms in M. pogonophilus may be 
related to its relatively brief (in evolutionary terms) 
association with P. badius . In other words, the apparent 
inability of spiders to follow trail pheromones and the 
lack of other integrative mechanisms, may be related to its 
brief association with the host ant. 



118 




119 




CO 

3 

T3 
CO 



0>l 

x: 
o 

c 

CO 

E 
o 
o 



CO 




CD 




D) 




C 




CO 




cr 




^. 




CO 




CO 




CD 




v* 




CD 




O 




o 




T3 




C 




CO 


CO 


00 


3 






■ 


o 


c 


c 


1^ 


o 


<T> 


o 


E 


o 
a 


CO 

sz 


si 


o 


T3 


E 


C 
CO 



J= CO 

■o 2 

CD O 

I g 



CM 




1^ 


2 


CD 




O) 


to 


LL 


a> 



CO 



REFERENCES 

Akre RD, Rettenmeyer CW (1966) Behavior of Staphylinidae 
associated with army ants (Formicidae: Ecitonini). 
Journal of the Kansas Entomological Society, 
39:745-782. 

Akre RD, Rettenmeyer CW (1968) Trail-following by guests 
of army ants (Hymenoptera: Formicidae: Ecitonini). 
Journal of the Kansas Entomological Society, 
41:165-174. 

Association of Official Analytical Chemists (AOAC) (1990) 

Method #920.39: Fat (crude) or ether extract in 

animal feed. In: Official Methods of Analysis, 15th 
ed. 

Attygalle AB, Morgan ED (1985) Ant trail pheromones. 
Advances in Insect Physiology , 18:1-31. 

Bailey KH, Polis GA (1987) Optimal and central place 

foraging theory applied to a desert harvester ant, 
Pogonomvrmex r.al ifornicus . Oecologia, 72:440-448. 

Bhatkar AP, Whitcomb WH (1970) Artificial diet for rearing 
various species of ants. Florida Entomologist. 
53:229-232. 

Boeve J-L (1992) Association of some spiders with ants. 
Revue Suisse Zoologia . 1:81-85. 

Briese DT, Macauley BJ (1981) Food collection within an 
ant community in semi-arid Australia, with special 
reference to seed harvesters. Australian Journal of 
Ecology . 6:1-19. 

Bristowe WS (1939) The Comity of Spiders . Ray Society No. 
126, London. pp. 162-168. 

Bristowe WS (1958) The World of Spiders . Collins New 
Naturalist Series, London. 

Brooks DR, McLennan DA (1991) Phylogenv. Ecology, and 

Behavior . The University of Chicago Press, Chicago 
and London, 4 34 pp. 

Bulmer MG, Taylor PD (1980) Dispersal and the sex ratio. 
Nature, 284:448-449. 

120 



121 



Carlin NF, Holldobler B (1986) The kin recognition system 
of carpenter ants ( Camponotus spp.) I. Hierarchical 
cues in small colonies. Behavioral Ecology and 
Sociobiology , 19:123-134. 

Carlson DM, Gentry JB (1973) Effects of shading on the 
migratory behavior of the Florida harvester ant, 
Poqonomyrmex badius . Ecology . 54:452-453. 

Chamberlin RV (1948) On some American spiders of the 
family Erigonidae. Annals of the Entomological 
Society of America . 41:483-562. 

Christian DP, Lederle PE (1984) Seed properties and water 
balance in desert granivores. The Southwestern 
Naturalist . 29( 2) : 181-188 . 

Coffin DP, Lauenroth WK (1989) Spatial and temporal 

variation in the seed bank of a semi-arid grassland. 
American Journal of Botany . 76(l):53-58. 

Coffin DP, Lauenroth WK (1990) Vegetation associated with 
nest sites of Western harvester ants ( Pogonomyrmex 
occidentalis Cresson) in a semi-arid grassland. 
American Midland Naturalist . 123:226-235. 

Cole AC, Jr. (1968) Pogonomyrmex Harvester Ants : a Study 
of the Genus in North America . The University of 
Tennessee Press, Knoxville. 

Colwell RK (1981) Group selection is implicated in the 
evolution of female-biased sex ratios. Nature , 
290:401-404. 

Colwell RK (1982) Female-biased sex ratios: Colwell 
replies. Nature . 298:494-496. 

Crist TO, MacMahon JA (1991) Individual foraging 

components of harvester ants: movement patterns and 
seed patch fidelity. Insectes Sociaux, 38:379-396. 

Crist TO, MacMahon JA (1992) Harvester ant foraging and 

shrub-steppe seeds: interactions of seed resources 
and seed use. Ecology . 73( 5) : 1768-1779 . 

Cushman JH, Martinsen GD, Mazeroll AI (1988) Density and 
size-dependent spacing of ant nests: evidence for 
intra-specif ic competition. Oecologia , 77:522-525. 

Davidson DW (1977) Foraging ecology and community 

organization in desert seed-eating ants. Ecology . 
58:725-737. 



122 



de Vita J (1979) Mechanisms of interference and foraging 
among colonies of the harvester ant Pogonomyrmex 
ral -ifornicus in the Mojave desert. Ecology, 
60(4) :729-737. 

Donisthorpe HStJK (1927) The Guests of British Ants, Their 
Habits and Life Histories . George Routledge and 
Sons, London. 

Excoffier L, Smouse PE, Quattro JM (1992) Analysis of 

molecular variance inferred from metric distances 
among DNA haplotypes: application to human 
mitochondrial DNA restriction sites. Genetics, 
131:479-491. 

Fewell JH (1988) Variation in foraging patterns of the 

Western harvester ant, Pogonomyrmex occidentalis in 
relation to variation in habitat structure. In: 
Inter individual behavioral variability in social 
insects (ed. Jeanne RL) , pp. 257-287. Westview 
Press, Boulder, CO and London. 

Fewell JH (1990) Directional fidelity as a foraging 
constraint in the western harvester ant, 
Pogonomyrmex occidentalis . Oecoloaia , 82:45-51. 

Fewell JH, Harrison JF (1991) Flexible seed selection by 
individual harvester ants, Poaonomvrmex 
occidentalis . Behavioral Ecology and Sociobiology , 
28:377-384. 

Foelix RF (1982) The Biology of Spiders . Harvard 
University Press, Cambridge. 

Franz R, Hall DW (eds) (1991) Vegetative Communities and 
Annotated Plant Lists for the Katharine Ordway 
Preserve-Swisher Memorial Sanctuary , Putnam County , 
Florida ( 1983-1991) . Ordway Preserve Research 
Series, Report #3, Florida Museum of Natural 
History, University of Florida, Gainesville, FL. 

Gallaher RN, Welden CO, Futral JG (1975) An aluminum 

digester for plant and soil analysis. Soil Science 
Society of America , Proceedings , 39:803-806. 

Gentry JB (1974) Response to predation by colonies of the 
Florida harvester ant, Pogonomyrmex badius. 
Ecology . 55:1328-1338. 

Gentry JB, Stiritz KL (1972) The role of the Florida 

harvester ant, Pogonomyrmex badius , in old field 
mineral nutrient relationships. Environmental 
Entomology , 1:39-41. 



123 



Gertsch WJ (1979) American Spiders . 2nd ed. Van Nostrand 
Reinhold Company, New York. 

Gertsch WJ, Davis LI (1936) New spiders from Texas. 
Amererican Museum Novitatae , 881:1-21. 

Gilpin M (1991) The genetic effective size of a 

metapopulation. Biological Journal of the Linnean 
Society . 42:165-175. 

Golley FB, Gentry JB (1964) Bioenergetics of the southern 
harvester ant, Pogonomyrmex badius . Ecology , 
45:217-225. 

Gordon DM (1984) The harvester ant ( Pogonomyrmex badius ) 

midden: refuse or boundary? Ecological Entomology . 
9:403-412. 

Gordon DM (1989) Ants distinguish neighbors from 
strangers. Oecologia . 81:198-200. 

Gordon DM (1991) Behavioral flexibility and the foraging 
ecology of seed-eating ants. American Naturalist . 
138:379-411. 

Gordon DM (1992) Nest relocation in harvester ants. 

Annals of the Entomological Society of America f 
85:44-47. 

Gordon DM (1993) The spatial scale of seed collection by 
harvester ants. Oecologia , 95:479-487. 

Gordon SA (1980) Analysis of 12 Sonoran desert seed 
species preferred by the desert harvester ant. 
Madrono . 27(2):68-78. 

Hadrys H, Balick M, Schierwater B (1992) Applications of 

random amplified polymorphic DNA (RAPD) in molecular 
ecology. Molecular Ecology . 1:55-63. 

Hambleton LG (1977) Semi-automated method for simultaneous 
determination of phosphorus, calcium and crude 
protein in animal feeds. Journal of the Association 
of Official Analytical Chemists, 60:845-852. 

Hamilton WD (1967) Extraordinary sex ratios. Science . 
156:477-488. 

Harmon G (1993) Mating in Pogonomyrmex badius 

(Hymenoptera: Formicidae). Florida Entomologist . 
76:524-527. 

Harmon GD, Stamp NE (1992) Effects of postdispersal seed 

predation on spatial ineguality and size variability 



124 



in an annual plant, Erodium cicutarium 
(Geraniaceae) . American Journal of Botany, 
79(3) :300-305. 

Harrison JS, Gentry JB (1981) Foraging pattern, colony 
distribution, and foraging range of the Florida 
harvester ant, Pogonomvrmex badius . Ecology, 
62:1467-1473. 

Hill RW (1976) Comparative physiology of animals: an 
environmental approach . Harper and Row, NY pp. 
22-35. 

Hobbs RJ (1985) Harvester ant foraging and plant species 
distribution in annual grassland. Oecologia, 
67:519-523. 

Holldobler B (1969) Host finding by odor in the 

myrmecophilic beetle Atemeles pubicollis Bris. 
(Staphylinidae) . Science , 166:757-758. 

Holldobler B (1971) Homing in the harvester ant 

Pogonomvrmex badius . Science , 171:1149-1151. 

Holldobler B (1974) Home range orientation and 

territoriality in harvesting ants. Proceedings of 
the National Academy of Sciences . USA, 71:3274-3277. 

Holldobler B (1976) Recruitment behavior, home range 

orientation and territoriality in harvester ants, 
Pogonomyrmex . Behavioral Ecology and Sociobiology , 
1:3-44. 

Holldobler B (1977) Communication in social Hymenoptera. 
In: How Animals Communicate (ed. Sebeok TA) , pp. 
418-471. Indiana University Press, Bloomington, 
Indiana. 

Holldobler B, Wilson EO (1970) Recruitment trails in the 
harvester ant Pogonomvrmex badius . Psyche , 
77:385-399. 

Holldobler B, Wilson EO (1990) The Ants . Belknap Press, 
Cambridge, Massachusetts. 

Howard RW, Akre RD, Garnett WB (1990) Chemical mimicry in 
an obligate predator of carpenter ants (Hymenoptera: 
Formicidae) . Annals of the Entomological Society of 
America , 83:607-616. 

Howard, RW, Stanley-Samuelson DW, Akre RD (1990) 

Biosynthesis and chemical mimicry of cuticular 
hydrocarbons from the obligate predator, Microdon 
albicomatus Novak (Diptera: Syrphidae) and its ant 



125 



prey, Myrmica incompleta Provancher (Hymenoptera : 
Formicidae). Journal of the Kansas Entomological 
Society , 63:437-443. 

Huff DR, Peakall R, Smouse PE (1993) RAPD variation within 
and among natural populations of outcrossing 
buffalograss [ Buchloe dactyloides (Nutt.) Engelm. ] . 
Theoretical Applied Genetics , 86:927-934. 

Ivie W (1967) Some synonyms in American spiders. Journal 
of the New York Entomological Society . 75:126-131. 

Janetos AC (1986) Web-site selection: are we asking the 

right guestions? In: Spiders : Webs , Behavior . and 
Evolution (ed. Shear WA) , pp. 9-22. Stanford 
University Press. 

Kelrick MI, MacMahon JA (1985) Nutritional and physical 

attributes of seeds of some common sagebrush-steppe 
plants: some implications for ecological theory and 
management. Journal of Range Management . 

Kelrick MI, MacMahon JA, Parmenter RR, Sisson DV (1986) 
Native seed preferences of shrub-steppe rodents, 
birds and ants: the relationships of seed 
attributes and seed use. Oecologia , 68:327-337. 

Kessler A (1973) A comparative study of the production of 
eggs in eight Pardosa species in the field (Araneae, 
Lycosidae). Ti jdschrift voor Entomologie , 
116:23-41. 

Kistner DH (1979) Social and evolutionary significance of 
social insect symbionts. In: Social Insects, vol. 
1 (ed. Hermann HR) , pp. 339-413. Academic Press, 
New York. 

Kistner DH (1982) The social insects' bestiary. In: 

Social Insects, vol. 3 (ed. Hermann HR) , pp. 1-244. 
Academic Press, New York. 

Kleinbaum DG, Kupper LL (1978) Applied Regression Analysis 
and Other Multivariate Methods . Duxbury Press, 
Boston, Massachusetts. 

Kovoor, J. 1987. Comparative structure and 

histochemistry of silk-producing organs in 
arachnids. In: Ecophysiology of Spiders (ed. W. 
Nentwig) , pp. 160-186. Springer-Verlag, Berlin, 
Germany. 

Krebs CJ (1989) Ecological Methodology . Harper & Row, New 
York, pp. 125-167. 



126 



Krebs JR, Kacelnik A (1991) Decision-making. In: 

Behavioural Ecology : An Evolutionary Approach (ed. 
Krebs JR, Davies NB) , pp. 105-136. Blackwell 
Scientific Publication, London. 

Levings SC, Traniello JFA (1981) Territoriality, nest 
dispersion, and community structure in ants. 
Psyche , 88:265-319. 

Lynch M (1990) The similarity index and DNA 

fingerprinting. Molecular Biology and Evolution , 
7:478-484. 

Lynch M, Milligan BG (1994) Analysis of population genetic 
structure with RAPD markers. Molecular Ecology, 
3:91-99. 

Millidge AF (1980) The erigonine spiders of North America. 
Part 1 . Introduction and taxonomic background 
(Araneae: Linyphiidae) . Journal of Arachnology , 
8:97-107. 

Miyamoto MM, Allard MW, Moreno JA. (1994) Conservation 
genetics of the plain pigeon (Columba inornata) in 
Puerto Rico. Auk, 111:910-916. 

Moser JC (1964) Inquiline roach responds to trail-marking 
substance of leaf -cutting ants. Science, 
143:1048-1049. 

Noonan GR (1982) Notes on interactions between the spider 
Eilica puno (Gnaphosidae) and the ant Camponotus 
inca in the Peruvian Andes. Biotropica , 14:145-148. 

Paine RT (1971) The measurement and application of the 
calorie to ecological problems. Annual Review of 
Ecology and Systematics . 2:145-164. 

Parr Instrument Company (1984) Analytical Methods for 
Oxygen Bombs . Parr Publications #207m. 

Peterson B (1950) The relation between size of mother and 
number of eggs and young in some spiders and its 
significance for the evolution of size. 
Experientia , 6:96-98. 

Porter SD (1985) Masoncus spider: a miniature predator of 
Collembola in harvester ant colonies. Psyche, 
92:145-150. 

Porter SD, Jorgensen CD (1988) Longevity of harvester ant 
colonies in southern Idaho. Journal of Range 
Management , 41:104-107. 



127 



Pulliam HR, Brand MR (1975) The production and utilization 
of seeds in plains grassland of Southeast Arizona. 
Ecology . 56:1158-1166. 

Regnier FE, Nieh M, Holldobler B (1973) The volatile 
Dufour's gland components of the harvester ants 
Poqonomyrmex ruqosus and P. barbatus . Journal of 
Insect Physiology , 19:981-992. 

Reichman OJ (1979) Desert granivore foraging and its 
impact on seed densities and distributions. 
Ecology . 60:1085-1092. 

Riechert SE, Gillespie RG (1986) Habitat choice and 

utilization in web-building spiders. In: Spiders : 
Webs, Behavior , and Evolution (ed. Shear WA) , pp. 
23-48. Stanford University Press. 

Rissing SW (1981) Foraging specializations of individual 
seed-harvesting ants. Behavioral Ecology and 
Sociobiology , 9:149-152. 

Roberts MJ (1995) Spiders of Great Britain and Northern 
Europe . Harper Collins Publishers, London. 

Rowell DM, Main BY (1992) Sex ratio in the social spider 
Diaea socialis (Araneae: Thomisidae). Journal of 
Arachnology . 20:200-206. 

Sambrook J, Fritsch EF, Maniatis T (1989) Molecular 

Cloning: a Laboratory Manual . 2nd ed. Cold Spring 
Harbor Laboratory Press, New York. 

Schmidt JO, Blum MS (1978) A harvester ant venom: 

chemistry and pharmacology. Science . 200:1064-1066. 

Schroth M, Maschwitz U (1984) Zur Larvalbiologie und 

Wirtsfindung von Maculinea teleius (Lepidoptera : 
Lycaenidae), eines Parasiten von Myrmica laevinodis 
(Hymenoptera : Formicidae). Entomologia Generalis . 
9:225-230. 

Siegel S (1956) Nonpar ametric Statistics for the 

Behavioral Sciences . McGraw-Hill Book Co., Inc., NY. 

Slatkin M (1994) Gene flow and population structure. In: 
Ecological Genetics (ed. Real LA), pp. 3-17. 
Princeton University Press, Princeton, New Jersey 

Snider RJ (1987) Class and order Collembola. In: Immature 
Insects (ed. Stehr FW) , pp. 55-64. Kendall/Hunt 
Publishing Co., Dubuque, Iowa. 



128 



Sokal RR, Rohlf FJ (1981) Biometry . 2nd ed. WH Freeman 
and Co. , NY. 

Taber SW (1990) Cladistic phylogeny of the North American 
species complexes of Poaonomvrmex (Hymenoptera: 
Formicidae) . Annals of the Entomological Society of 
America , 83:307-316. 

Tolbert WW (1977) Aerial dispersal of two orb-weaving 
spiders. Psyche , 84:13-27. 

Turner CH (1909) The mound of Poaonomvrmex badius Latrl . 
and its relation to the breeding habits of the 
species. The Biological Bulletin , 17:161-170. 

Van Pelt A (1953) Notes on the above-ground activity and a 
mating flight of Pogonomyrmex badius (Latr.), 
Journal of the Tennessee Academy of Science . 
28:164-168. 

Vander Meer RK, Wojcik DP (1982) Chemical mimicry in the 

myrmecophilous beetle Myrmecaphodius excavaticollis . 
Science , 218:806-808. 

Vander Meer RK, Alvarez F, Lofgren CS (1988) Isolation of 
the trail recruitment pheromone of Solenopsis 
invicta . Journal of Chemical Ecology , 14:825-838. 

Vander Meer RK, Jouvenaz DP, Wojcik DP (1989) Chemical 

mimicry in a parasitoid (Hymenoptera: Eucharitidae) 
of fire ants (Hymenoptera: Formicidae). Journal of 
Chemical Ecology , 15:2247-2261. 

Vollrath F (1987) Growth, foraging and reproductive 
success. In: Ecophysiologv of Spiders , (ed. 
Nentwig W) , pp. 357-370. Springer-Verlag, Berlin. 

Vollrath F, Parker GA (1992) Sexual dimorphism and 
distorted sex ratios in spiders. Nature , 
360:156-159. 

Wasmann E (1894) Kritisches Verzeichniss des myrmecophilen 
und termitophilen Arthropoden. Felix Dames , Berlin. 
xi + 231 pp. 

Webb DS (1990) Historical biogeography . In: Ecosystems 
of Florida , (ed. Myers RL, Ewel JJ) , pp. 70-100. 
Central Florida Press, Orlando. 

Welsh J, McClelland M (1990) Fingerprinting genomes using 
PCR with arbitrary primers. Nucleic Acids Research , 
18:7213-7218. 



129 



Weygoldt P (1969) The Biology of Pseudoscorpions . Harvard 
University Press, Cambridge, Massachusetts. 

Whitford WG (1976) Foraging behavior of Chihuahuan desert 
harvester ants. The American Midland Naturalist . 
95(2) :455-458. 

Whitford WG (1978) Foraging in seed-harvester ants 
Pogonomyrmex spp. Ecology , 59 ( 1 ): 185-189 . 

Williams JGK, Kubelik AR, Livak KJ, Rafalski JA, Tingey SV 
(1990) DNA polymorphisms amplified by arbitrary 
primers are useful as genetic markers. Nucleic 
Acids Research , 18:6531-6535. 

Williams T, Franks NR (1988) Population size and growth 
rate, sex ratio and behaviour in the ant isopod, 
Platyarthrus hof fmannseggi . Journal of Zoology . 
London, 215:703-717. 

Wilson DS, Colwell RK (1981) Evolution of sex ratio in 
structured demes . Evolution . 35:882-897. 

Wise DH (1975) Food limitation of the spider Linyphia 

marginata : experimental field studies. Ecology , 
56:637-646. 

Wright S (1951) The genetical structure of populations. 
Annals of Eugenics , 15:323-354. 



BIOGRAPHICAL SKETCH 
Paula E. Cushing was born in Alexandria, Virginia, 
on February 17, 1964, to Paula M. and Col. Joseph Cushing. 
When she was in high school, she decided to become a 
biologist and to go as far as she could in the pursuit of 
her interests. At 17, she told herself that she would 
receive her Ph. D. by the time she was 30, and she almost 
made it. 

She graduated from high school in 1982 and received 
her Bachelor of Science degree in biology from Virginia 
Polytechnic Institute and State University in Blacksburg, 
Virginia in 1985. She remained at V.P.I, and S.U. for her 
Master of Science degree in zoology under the tutelage of 
Dr. Brent D. Opell. For her master's degree, she 
investigated disturbance behaviors in spiders and their 
possible role as predator avoidance strategies. It was 
then that she first came to be known by family and friends 
as the "Spider Lady." She received her master's degree in 
1988. 

For two years she assuaged her wanderlust by 
traveling to Europe, living and working in Panama, and 
visiting the western United States. In between trips, she 
worked as a research assistant for Dr. Opell and published 
her master's thesis. In August 1990, Paula entered the 

130 



131 



Doctor of Philosophy program at the University of Florida 
in Gainesville under the tutelage of Dr. Jonathan 
Reiskind. She looks forward with enthusiasm to her career 
as a professional biologist. She intends to go even 
further in pursuit of her interests. 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor^of—^Philosophy . 





an Reiskind, Chairman 
'ate Professor of Zoology 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adeguate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




John 7 Andersdn 

Associate Professor of Zoology 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree^-ef) Doctor of Philosophy. 







'c^lJ 



Clifford^ <Johnson' / 
Professor of Zoology 

I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adeguate, in scope and guality, 
as a dissertation for the degree of Doctor of Philosophy. 





Michael Miyamoto 
Professor of Zoology 

I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adeguate, in scope and guality, 
as a dissertation for the degree of Doctor of Philosophy. 




Sanford Porter 
Assistant Professor of 

Entomology and Nematology 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, 
as a dissertation for the deqree of Doctor of Philosophy. 




John Sivin"! 
rsistant Professor of 

Entomoloqy and Nematoloqy 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, 
as a dissertation for the—deqree of Doctor-"0~f) Philosophy. 




Robert Vander Meer 
Assistant Professor of 

Entomoloqy and Nematoloqy 

This dissertation was submitted to the Graduate 
Faculty of the Department of Zooloqy in the Colleqe of 
Liberal Arts and Sciences and to the Graduate School and 
was accepted as partial fulfillment of the requirements 
for the deqree of Doctor of Philosophy. 



December 1995 



Dean, Graduate School 



1995 



Imin^l'.Tr 0F FLORIDA 

illlilll 

3 1262 08557 0934