J

gagix...

mm

fig

,1‘. ..L

.vi...;!..:

an
in

 

.h: .
afi: .q
:13 .
91b tin-.91
.45...»me

I

VIC L
.ulnm. :8. 3!
Lu ...\...:..rlir;€n u.

[291‘ 1. ,....«h . nfl,
:A r V
_ «Smut ..

I
v!

. “mm“ Pfiflhumwu
Annkhufio. 1rd .

3?? tr». 5.2“. . . .
, 93.5.2 cm, 3.1.! 1..

_

 

‘ummiltljiflillflilflflllllfltul *

 

LIBRARY
Mlehigan State
University .

 

 

 

This is to certify that the

thesis entitled

Effects of habitat succession on population
and reproduction of the Allegheny mound ant

presented by

Heather Cristina Rowe

has been accepted towards fulfillment
of the requirements for

 

 

M. S . degree in Entomology

Cage/uh; m swat

Major professor

 

Date 10 December, 1998

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN REFURN BOX to remove this checkout from your record.
TO AVOID FINI return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M mu

 

EFFECTS OF HABITAT SUCCESSION ON POPULATION AND
REPRODUCTION OF THE ALLEGHENY MOUND ANT

By

Heather Cristina Rowe

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Entomology

1998

ABSTRACT

EFFECTS OF HABITAT SUCCESSION ON POPULATION AND
REPRODUCTION OF THE ALLEGHENY MOUND ANT

By

Heather Cristina Rowe

Reliance on transient habitats makes Formica. exsectoides F orel, the Allegheny
mound ant, useful in the study of reproductive allocation as an adaptive strategy for
coping with environmental change. The sex ratio of F. exsectoides was determined to be
largely male-biased, yet variable among populations. Sex ratio estimates were influenced
by sexual dimorphism in weight, but not energy content, and stage of pupal development.
Population density and reproductive allocation were examined across a gradient of jack
pine stand ages, grouped into young (2-16 y), mature (37-45 y) and old (66-77 y) classes.
Forager abundance was greatest in mature stands, and declined in old stands.
Composition of ground arthropod communities shifted as stands matured; the strongest
change was an increase in predaceous beetles in old stands, possibly creating competition
between beetles and F. exsectoides for protein resources. While this may cause decline
of F. exsectoide.s' in old stands, light availability may also play a role. Light environment
was linked to mound reproductive status—mounds in low light produced only worker
brood, mounds in intermediate light produced sexual-caste brood, and mounds in low
light produced no brood. This suggests that low light makes older stands unsuitable for
brood production, and that production of sexual forms may be a dispersal response to
deteriorating conditions. Sex ratio was not associated with any measured variables

related to habitat succession, but was related to population age.

This is dedicated to next year’s brood, who will escape me.

iii

ACKNOWLEDGEMENTS

I would like to thank the people who have allowed me to run rampant through
their labs, use their equipment and supplies, and place my greedy fingerprints upon their
carefully accumulated wealth of knowledge. Phil Huber, of the US. Forest Service, gave
me data on stand age and made sure I didn’t get lost in the jack pine. Use of the bomb
calorimeter was generously granted by Terry Trier and Bill Mattson. Staining protocols,
allozyme advice, and freezer space were provided by Guy Bush, Jim Smith, and
Vanderlei Martins. I particularly thank Aram Stump for showing me the way of the
cello—gel. Fred Dyer allowed me to squash his bees. Dora Carmona and Jana Lee saw to
it that my beetles were identified, and Gary Coovert identified my ants. Rich Kobe
allowed me to run off into the woods with $1500 of new camera equipment, and m let
me clutter up his lab while processing the data.

Bryan Bishop and Carolyn Copenheaver provided field assistance in times of
need. Julie K. Purdom Lindblad sorted more insects than any other beer cart driver.

H. Governess, C. Vreisenslurp, and S. Frozen provided fabulously critical reviews
and love when needed. It will all come back to them some day.

The Michigan State Department of Entomology has provided support in the form
of travel grants, research funding, computer equipment, and faculty and staff resources. I
can think of no one who I asked for help who was not willing to give it.

I would like to thank my committee: Fred Dyer, Guy Bush, and Mark Scriber,
and my advisor, Cathy Bristow, who understands the value of positive encouragement.

And my parents, as always, were swell.

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii
LIST OF FIGURES ............................................................................... xi
CHAPTER 1
Sex ratio of Formica exsectoides F orel. the Allegheny mound ant ............................ 1
Introduction ................................................................................... 1
Material and Methods ....................................................................... 5
Site locations ........................................................................ 5
Brood collection and measurement ............................................... 5
Energy content ...................................................................... 7
Calculation of sex ratio ............................................................ 7
Results ....................................................................................... lO
Mound measurements and collection of brood ................................ 10
Weight and development ......................................................... 10
Energy content ..................................................................... 16
Sex ratios ........................................................................... 23
Discussion ................................................................................... 26
Mound measurements and collection of brood ................................ 26
Weight and development ......................................................... 27
Sex ratio ............................................................................ 28
CHAPTER 2
Age of jack pine stands influences abundance of the Allegheny mound ant
(Hymenoptera: F ormicidae) and other ground-dwelling arthropods ................ 32
Introduction ................................................................................. 32
Material and Methods ..................................................................... 35
Site selection and description .................................................... 35
Location and density of F. exsectoides mounds .............................. 35
Abundance of F. exsectoides and other ground-dwelling arthropods ...... 38
Data analysis ....................................................................... 38
Results ....................................................................................... 41
Location and density of F. exsectoides mounds .............................. 41
Abundance of F. exsectoides and other ground-dwelling arthropods. . ....47
Ants ......................................................................... 54
Herbivores ................................................................. 59
Predators .................................................................. 63
Scavengers ................................................................ 63
Correlation between feeding groups ............................................ 63
Discussion ................................................................................... 65
Location and density of F exsectoides mounds .............................. 65
Abundance of F. exsectoides and other ground-dwelling arthropods. .....66
Ants ......................................................................... 67

Herbivores ................................................................. 69

Predators .................................................................. 7O
Scavengers ................................................................ 71
Correlation between feeding groups ............................................ 71
Conclusions ........................................................................ 72
CHAPTER 3
The impact of jack pine habitat succession on reproductive strategy of the
Allegheny mound ant, Formica exsectoides F orel .................................... 74
Introduction ................................................................................. 74
Material and Methods ..................................................................... 79
Site selection and description .................................................... 79
Environmental variables ......................................................... 79
Light availability .................................................................. 79
Mound reproductive status ....................................................... 80
Sex ratios ........................................................................... 81
Statistical analyses ................................................................ 81
Results ....................................................................................... 83
Environmental and mound variables ........................................... 83
Mound reproductive status ....................................................... 83
Sex ratio ............................................................................ 90
Discussion ................................................................................. l 14
Environmental and mound variables .......................................... 114
Mound status ..................................................................... 1 15
Sex ratio ........................................................................... 1 17
APPENDIX A
Arthropods collected in jack pine (Pinus banksiana Lamb.) by pitfall trapping,
7 July—4 August, 1997 ............................................................... 123
APPENDIX B
Attempted search for diploid males (or Things That Didn’t Work) ......................... 126
APPENDIX C
Record of deposition of voucher specimens ................................................... 133
LIST OF REFERENCES ......................................................................... 136

vi

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

LIST OF TABLES

Pilot study site locations and descriptions ......................................................... 6

Sampling data and sex ratios, by site. Boomsma cost ratio is generated
by correcting dry weight ratios with Boomsma’s (1989) energetic cost
ratio: (weight males/weight total)07ll

Site means and standard deviations of surface area indices and distance
to nearest neighboring mound... ... .................................................................. 12

ANOVA test of H0: Surface area indices of mounds producing sexual
caste brood are equal to those of mounds producing worker brood
only ................................................................................................................... 15

ANOVA test of H0: Distance to nearest neighbor is equal for mounds
that produce sexual forms and mounds that produce workers only .................. 15

Analysis of covariance test of hypotheses H0: Weight of male and

female pupae are equal; H0: There is no significant relationship

between weight and developmental stage; and Ho: This relationship

is the same for males and females ..................................................................... l8

Nested ANOVA performed on energy content data, testing Ho: Hc is
equal for all sites and both sexes within sites. Hc = heat of combustion.

in calories per gram ........................................................................................... 22
ANOVA test of H0: Sex ratio is equal for all sites ........................................... 25
Stand locations and descriptions for Chapters 2 & 3 ........................................ 37

Table 10. Stand age class means and standard deviations for mound and tree

densities ............................................................................................................ 42

Table 11. Nearest neighbor analysis of mound dispersion (Clark and Evans 1954),

performed by calculating an Index of Nonrandomness (R) by the formula

R = 2r(p)0‘5. (Where r = distance to nearest neighbor (m), p = population
density (mounds/m2». Analyses were performed on individual mounds

and R was tested against a null expectation of random dispersion (R = 1)
using a t-test as described in Pielou (1969) ...................................................... 45

vii

Table 12.

Table 13.

Table 14.

Table 15.

Table 16.

Table 17.

Table 18.

Table 19.

Lloyd’s indices of mean crowding (X*) and patchiness. X* = mean

number of neighbors per individual in a sampling unit. Sampling units

were l250m2 (4 units/plot). The patchiness index indicates dispersion,

with a null expectation of 1 indicating random dispersion .............................. 45

Invertebrates captured in pitfall traps. Values presented are: %

of all traps containing at least one individual and mean number of

individuals per trap (standard deviation). P-values were generated

by one-way ANOVA on each taxonomic group. Entries in bold-face

were significantly different after sequential Bonferroni correction

for multiple comparisons .............................................................. 48

Univariate analyses testing Ho: Feeding guild proportions are equal
across stand age classes. Canonical coefficients were generated by
MANOVA test; only those discussed in the text are shown ............................ 53

Ant genera and species most commonly captured in jack pine. Values
presented are: % of all traps containing at least one individual, and

mean (standard deviation) of individuals captured per trap for each

stand age class. P-values were generated by one-way ANOVA for

each species; bold species were significantly different after Bonferroni
correction .................................................................................. 56

Univariate analyses testing H0: Proportions of ant species are equal
across stand age classes. Canonical coefficients were generated by
MANOVA test; only those discussed in the text are shown ............................ 57

Correlation between major invertebrate groups. “Correlation” =

Pearson correlation coefficient, generated by PROC CORR, SAS

Systems, Inc. Significance determined with a sequential Bonferroni
adjustment of p-values. ‘* ’ indicates values provided for only one

stand age class .................................................................................................. 64

Correlation between environmental and F. exsectoides population
characteristics. After correction for multiple comparisons, only

correlations marked ‘***’ (p < 0.001) were significant. Correlations

marked ‘**’ (p < 0.01) and ‘*’ (p < 0.05) were considered to be

‘marginal’. Direction of correlation is listed below significance.

“Mound size” refers to surface area indices, NOT an absolute

measure ..................................................................................... 85

Summary of mound status for each site and age class. Totals represent

all mounds present in a 2500m2 study plot. “Active” mounds

contained brood, either a mixture of reproductive caste and worker

brood or worker brood only. “Dead” mounds showed little or no ant

activity and contained no brood ....................................................................... 86

viii

Table 20.

Table 21.

Table 22.

Table 23.

Table 24.

Table 25.

Table 26.

Table 27.

Table 28.

ANOVA test of significance for the regression model: % dead

mounds = stand age, when “stand age” = years since major

disturbance; ANOVA test of significance for the regression

model: % dead mounds = stand age, when “stand age” = years

since last clear cut ........................................................................................ 87

ANOVA test of significance for the regression model: % of mounds
producing sexual forms = stand age, when “stand age” = years
since major disturbance ................................................................................... 89

Correlation between percentages of dead and sexual-producing

mounds and stand-specific variables. After correction for multiple
comparisons, only correlations marked ‘***’ (p < 0.001) and

‘**’ (p < 0.01) were significant. Correlations marked ‘*’ (p < 0.05)

were considered to be ‘marginal’. Direction of correlation is listed

below significance. For the three correlated variables, stand age,

stand density, and % open sky (see Table 18), shaded blocks indicate

the strongest correlation ................................................................................... 93

ANOVA test of Ho: Light environment (% open sky) is equal for

all mound status groups. Status groups were: absent = no mound

present at measurement location; dead == inactive, not producing

brood; workers only = producing only worker brood; reproductive

= producing both sexual caste and worker brood95

ANOVA test of hypotheses HO: Pupal dry weight is equal for males
and females and H0: Pupal dry weight does not differ among stand
age classes ........................................................................................................ 96

ANOVA test of H0: Sex ratio of young clear cut sites is equal to the
sex ratio ofall other Site3102

ANOVA test of significance for the regression model: sex ratio =
log-transformed abundance of F. exsectoides foragers captured in
pitfall traps ..................................................................................................... 109

ANOVA test of significance for the regression model: sex ratio =
F. fusca abundance in pitfall traps. a) all sites b) excluding clear
cut ......................................................................................... l 11

a) ANOVA test of significance for the regression model: sex ratio =

age + agez, when “age” = years since last major disturbance of stand.

b) ANOVA test of significance for the regression model: sex ratio =

age + agez, when “age” = years since clear cut of stand ............................... 1 l3

Table 29. Allozyme banding pattern for F. exsectoides; values represent cm
each band migrated from the origin. The column “workers” represents
a summation of all bands in all samples (each sample contained
multiple ants, see text). Columns Y1, Y4, and 04 contain concensus
banding patterns from male and female sexual forms sampled in
stands Y1, Y4, and O4—no within—stand variation in banding pattern
occurred at any locus ................................................................. 131

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

LIST OF FIGURES

Representative stages of development of male and female pupae ..............

Mounds that produced sexual forms were larger in comparative
surface dimensions than those that did not. Error bars represent

one standard deviation .................................................................................

Distance from nearest neighboring mound is not related to production

of sexual forms. Error bars represent one standard deviation .....................

Females were heavier than males. Error bars represent one standard

dev1ation

Pupal dry weight decreased over the course of development.
Regression slopes were significant at p = 0.0001;

ANOVA tests are provided in Table 6 ........................................................

Average stage of pupal development did not differ significantly

between sexes or among Sites

There is no significant difference in energy density (Hc) of
reproductive caste pupae between sexes or among sites. ANOVA

tests are provided in Table 7 .......................................................................

Sex ratio was predominantly male-biased ..................................................

a) Stands were located in Crawford and Oscoda counties, Michigan.
Letter indicates stand age class: 0 = old, m = mature,

Y = young. b) Plot layout and pitfall trap locations .................................

Stand density increased significantly with stand age. Density values
represent both average diameter of trees in plot and density of
individuals (stem density). No value is present for Y4, due to lack

of stem density data for portions of the plot ..............................................

a) Mound density was negatively correlated with stand density.
b) This correlation was strongest when only young stands were

considered ..................................................................................................

xi

...... 8

....13

..... 14

17

....19

.20

..... 21

..... 24

..... 36

..... 43

..... 44

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

Figure 18.

Figure 19.

Figure 20.

Figure 21.

a) Mound surface area indices were not larger in older stands.
b) Mean mound surface area indices increase with stand age,
possibly due to incomplete budding and a lack of small mounds ................... 46

Total number of individuals captured per stand age class, sorted

by taxon group and date. Taxon groups presented were present

in over 25% of all traps and accounted for more than 1% of all

invertebrates captured. These include: a) F. exsectoides,

b) other Formicids, c) Orthoptera, d) Lygaeidae, e) Thyreocoridae,

f) Aranae, g) predaceous Coleoptera, h) other Coleoptera,

i) Opiliones, j) Diplopoda, k) Diptera ........................................................... 49

Mound density was positively correlated to log-transformed
abundance of F. exsectoides foragers captured ............................................... 55

There was no significant relationship between abundance of
F exsectoides foragers per trap collection and abundance of
all other ants .................................................................................................... 58

a) There was a significant positive relationship between mean

number of F. filsca captured per trap and total number of morpho-

species captured in all traps at a site. b) This relationship remained
significant even when highly influential points (Y4, Y5) were

removed from the analysis .............................................................................. 60

Changes in composition of feeding guilds across stand age classes.
a) Herbivores b) Herbivores (part 2) c) Predators d) Scavengers ............... 62

Creation of relatedness asymmetry by haplo-diploidy, as described
by Trivers and Hare (1976)76

a) While correlated with stand age, light availability was highly
variable within stands. b) Status groups differed in their relationship
to the range of light available .......................................................................... 84

a) The percentage of mounds producing sexual caste brood increased

with stand age, when “stand age” = years since major disturbance.

b) This relationship was no longer significant when “years since clear

cut” was used as stand age. Arrow indicates burned stands .......................... 88

The percentage of mounds classified as “dead” (containing no brood.

inactive) increased significantly as open sky decreased. Values of
open sky are plot mean values ........................................................................ 91

xii

Figure 22. a) Sites with more mounds contained lower proportions of sexual-
producing mounds b) Plotting the interaction of this
relationship with stand age reveals that this relationship is strongest
in mature stands, while old and young stands are likely to produce
sexual brood in a constant proportion of mounds ........................................... 92

Figure 23. Light availability differed significantly among mounds of differing
“status”. “Absent” represents values of open sky measured where
no F. exsectoides mounds were present. “Dead” represents values
of light availability for mounds that were abandoned or producing
no brood. “Workers only” and “reproductive” indicate absence or
presence of reproductive-caste pupae in active, brood-producing
mounds. Values presented are least-squares means and standard
errors. All differences were significant at p < 0.001 ..................................... 94

Figure 24. Females were significantly heavier than males. Both sexes were
heavier in old stands than in young or mature stands. Values
presented are least-squares means and standard errors. Significance
tests are presented in Table 24 ........................................................................ 97

Figure 25. Pupal dry weight of sexual forms were not related to sex ratio
Regression slopes were not significantly different from zero for
either sex .................................................................................... 98

Figure 26. Sex ratio estimates for F. exscctoidcs populations in 15 jack pine
stands .................................................................................. 100

Figure 27. Confidence intervals for numerical sex ratio estimates. generated
with binomial probability functions .............................................. 101

Figure 28. Observed sex ratios were a poor fit to logistic regression models
based on light availability ............................................................................. 105

Figure 29. Observed sex ratios were a poor fit to logistic regression models
based on surface area indices ........................................................................ 107

Figure 30. a) Sex ratio was not significantly related to abundance of F.
emectoides foragers captured in pitfall traps. b) This relationship
became significant when sex ratio outliers (young clear cut stands)
were removed from the analysis ................................................................... 108

xiii

Figure 31.

Figure 32.

a) Sex ratio appeared to have a marginally significant negative

relationship with pitfall trap abundance of F. fusca. b) This

relationship disappeared when 2 young clear cut stands (outliers for

both sex ratio and F. fusca abundance) were removed from the analysis.
Values presented are mean F. fusca abundance for each stand and

total male allocation ratios for each stand. Significance tests for

regression analyses are presented in Table 27 a, b ...................................... 110

a) Sex ratio was not significantly related to stand age when “stand

age” = years since major disturbance. b) Sex ratio was significantly

related to stand age when “stand age” = years since clear cut. Curves

in both a) and b) were fit using second-degree polynomial regression ........ 112

xiv

CHAPTER 1

Sex ratio of Formica exsectoides Forel, the Allegheny mound ant

Introduction

Formica exsectoides Forel is a cosmopolitan and ecologically important ant
species in the eastern United States. Ranging from Nova Scotia to northern Georgia and
west to Ontario and Illinois, these ants are usually found in forest edges or openings
(Creighton 1950). Often recognized by their conspicuous mound architecture,
populations have been studied with respect to physical structure, placement, and
orientation of mounds (McCook 1877, Andrews 1926, Cory & Haviland 1938, Dimmick
1951)

F. exscctoides is an aggressive predator (Ayre 1963) and has been investigated for
its potential in biocontrol of forest pests (Allen et. al. 1970, Campbell 1991, Campbell et.
al. 1991). F. exsectoz'des is also considered a pest species, mainly due to its practice of
killing young trees in proximity to mounds (Andrews 1926, Wilson 1977, Patterson
1994), and its generally aggressive behavior (Levins and Traniello 1981, personal
observation). Investigations into its control include Price (1945), Haviland (1947), and
Weaver and Smith (1993). F. exsectoides has also been shown to affect the community
structure of phloem-feeding insects through mutualistic associations with Homopterans
(Andrews 1929, Bishop 1998). Despite this obvious interest in economic influences of F.

exscctoides, knowledge of its basic biology is lacking.

In Michigan, F. exsectoia’es is commonly found in jack pine (Pinus banksiana
Lamb.) ecosystems. F. exsectoides exhibits a patchy use of this habitat and can form
large, locally dense populations. F. exsectoides colonies possess multiple queens
(polygyny) (Bristow et. al. 1992 ) and may occupy multiple mounds (polydomy) (N ipson
1978, Oster and Wilson 1978). Colony expansion occurs by budding or production of
sexual forms. Sexual forms are produced once a year in Michigan, with pupation
occurring in late July/early August. No mating swarms have been observed—females
exhibit calling by pheromone (Holldobler and Wilson 1990), attracting males to mate on
low shrubs and saplings. Mated queens may found new colonies by temporary social
parasitism of F. firsca Linn., or may attempt to gain acceptance into an existing F.
exsectoides mound (Wheeler 1933; Creighton 1950, Starr 1979).

The sex ratio of F. exsectoides is unknown. Knowledge of sex ratio is important
for understanding the reproductive ecology of an organism. In ants, sex is determined by
one or more loci, where heterozygotes are female and hemi- or homozygous individuals
are male (Crozier 1971, Bull 1983). For simplicity, it can be generalized that females are
diploid (arising from fertilized eggs) and males are haploid. This system of sex
determination, coupled with caste differentiation displayed by highly eusocial
Hymenoptera, places sex ratio on a behavioral, rather than evolutionary, time-scale of
response. The lability of this trait makes it an important source of reproductive variation
among species.

The primary objective of this study was to estimate the sex ratio of F. exsectoidcs.
As well as providing long-overdue knowledge of the reproductive ecology of this species,

results may contribute to the growing power of sex ratio meta-analyses to inform on

important evolutionary questions. Haplo-diploidy creates elevated relatedness between
sisters, since sisters share an average of 75% of their genetic material, while parent and
offspring share 50%, and sisters and brothers share (on average) only 25%.

Due to these relatedness asymmetries between siblings, studies of Hymenopteran sex
ratios can provide valuable tests of kin selection theory, and insight into parent-offspring
conflict, the mechanisms of social evolution and the maintenance of sex ( Hamilton 1964,
Trivers and Hare 1976, Chamov 1982, for review see Nonacs 1986, Pamilo and Crozier
1996). Individual studies are of limited utility in approaching such general questions.
Syntheses of many sex ratio studies can be more informative, provided that the studies
incorporated are compatible and relevant. Sex ratio data from F. exsectoides may also
facilitate comparisons with Formica exsecta and the more extensively—studied European
wood ants of the Formica rufa group (Pamilo and Rosengren 1983, Rosengren and
Pamilo 1986).

A secondary objective of this study was to determine an appropriate method of
estimating F. exsectoides sex ratio, by addressing questions of sexual dimorphism and
intraspecific sex ratio variation. Measures of sex ratio generated by counting individuals
of each sex have been displaced by often sophisticated attempts to quantify allocation of
resources to each sex. The simplest of these is the use of weight ratios, correcting for the
possibility of sexual size dimorphism. Boomsma (1989) stated that female investment
had been “systematically overestimated”, as male ants generally have higher respiration
rates than females. His correction, supported by energetic data from Lasius niger
(Boomsma 1989) and additional species (Boomsma er. al. 1995), is generally accepted,

but should be used with the caveat that not all species will show the same extent of

physiological dimorphism, and that metabolic activity may vary over development. Few
studies (Boomsma and Isaaks 1985, MacKay 1985) incorporate the overall energy
content of sexual forms. Energy content may be expected to differ between sexes, to an
extent determined by dispersal patterns and founding biology. The investigation of these
questions is vital to the future utility of this study to larger analyses, since combining sex
ratio studies without reference to the method and organism of study can create misleading
conclusions. While it is not essential that all authors estimate sex ratio by the same
method, the products of differing methods must be biologically equivalent to facilitate

meaningful comparisons.

Material and Methods

Site locations. Study sites were located in Oscoda, Crawford, and Roscommon counties,
in the northern lower peninsula of Michigan (see Table 1 for details on site locations).
All sites were noncontiguous and located at least 1 km apart, so populations were
discrete. Vegetation consisted predominantly of jack pine (Pinus banksiana) with some
interspersed red pine (Pinus resinosa Aiton), Northern pin oak (Quercus ellipsoidalis
Hill), and cherry (Prunus spp.). Understory vegetation consisted primarily of blueberry
(Vaccinium angustifolium Aiton), sand cherry (Prunus pumila L.), bracken fern
(Pteridum spp.) and grasses (C arex spp.).

Brood collection and measurement. Partial excavations of 66 mounds from 6 sites
were performed from July 18 to July 26, 1996. For each mound, a comparative index of
mound surface area (longest slope * shortest slope) and the distance to the nearest
neighboring mound were recorded. Presence or absence of reproductivecaste brood was
determined by excavating the mound to approximately 0.5 m below ground level (or until
reproductive caste pupae were found). Reproductive caste pupae are easily distinguished
from worker pupae by size and coloration.

Samples of 30-90 pupae were collected from mounds found to contain reproductive caste
brood. Pupae were dried and weighed. Sex could usually be determined visually after
removing the pupal case. Some pupae had not developed enough for sex to be
distinguishable by morphology; these were scored as “undeveloped”. Stage of pupal

development was scored from 0 (undeveloped) to 3 (fully developed, pigmented, ready to

$5803 .08me 0w8_ 55 05a xofl. So 80» 8.3
0525, .3 “0080.8 9880—0
#0me 0H8..— 8 53, 28% 0:3 xofl. 20 80% on

8883-2 8» 8: 8080—0 03988.” 8008

0585 :28082 nwE .5080? $8823
agooéqsew E09033 55 05988.6 0.588

8908“ .3
80808 $5806 .8me .bofibwfi. 080 ”b08800

xoom $505558 28% 05a 08% 30 .80» 8

388% 0.888 20.80% E 85 x03 @5803
damn x08“ Em: @2388 5? .808 .888 £03

cowatomxo

2 .38 d: meF .60 .8080 0
Va .38 .m: meH .60 $580 w

own .3: .meF 60 E8285 #4

:8 .Bma .ZVNF .oU :oEEoomom m

o: .3: .239 .60 20.386 N

vmflma .3: .ZmNP :00 €0,385 #

c2802 0%

mcouatomoo 98 82802 0% .335 BE 4033.

eclose). Figure 1 presents representative photographs of pupal development stages.
Stages presented are 1, 2, and 3, but half-stages were used where development was
intermediate. Analysis of covariance was used to analyze pupal dry weights, with sex as
the primary analysis variable and stage of development as the covariate (PROC
ANCOVA, SAS Systems 1990).

Energy content. Caloric densities of pupae were measured using a semi-micro
calorimeter (Parr Instrument Company, Moline, IL). Heat of combustion in calories/ g

(He) was calculated by the following equation:

Hc = ((B * AT) — ((fl-f2)* 1400)) / sample weight

where B is a caloric constant, calculated through calibration with benzoic acid, AT is the
change in temperature occurring during combustion, and (fl-f2)*1400 is a correction
factor accounting for the amount of fuse wire used to ignite the sample.

Since measurement accuracy declines sharply in samples below 0.01g weight (Terry
Trier, personal communication), 3-5 individuals of each sex were measured together. At
least three samples from each site were analyzed to determine energy content of males; it
was often not possible to analyze three (or any) samples of females, as fewer females
were available. The data were analyzed using a nested analysis of variance, in the
following hierarchy: days (mounds (sex))). The variable “days” accounts for variation in
calorimeter performance during the course of the analysis.

Calculation of sex ratio. Sex ratio was estimated numerically (number of males/total
number of pupae) and by dry weight (dry weight of males/ dry weight of total sample).

“Undeveloped” pupae were not included. As pupae within samples were usually of the

female

 

Figure 1. Representative stages of development of male and female pupae

same developmental stage, weights were not corrected for developmental stage.
Numerical and weight-based sex ratio estimates were compared to each other using a
paired t-test. Weight-based sex ratios were corrected using Boomsma’s (1989) energetic
cost ratio, and these corrected ratios were compared between mounds and between sites
using a nested analysis of variance. Boomsma-corrected sex ratios were also correlated
to mound surface area and distance to nearest neighbor using PROC CORR (SAS

Systems, Inc., 1990).

Results

Mound measurements and collection of brood. Reproductive pupae were only
found in 37 of 66 partially excavated mounds. Of these 37, only 33 mounds contained
sufficient pupae (25-30) for sex ratio sampling. Table 2 provides a summary of
excavations, sampling, and sex ratios for each site. Sites 4 and 6 contained relatively few
mounds; all mounds within visual range were examined. A total of 1510 pupae were
collected across the six sites. Of these, 134 had not developed sufficiently for sex to be
morphologically distinguishable; these were classified as “undeveloped”, and are
subtracted from the total pupae sampled to provide the “# of pupae for ratios” shown in
Table 2.

Mound surface area ranged from 0.9 m2 to 12.5 m2. Distance to nearest
neighboring mound ranged from 1 m to 51 m. Only active/inhabited mounds were
counted as “neighbors”. Table 3 lists the mean and standard deviation of mound surface
area indices and distance to nearest neighbor recorded at each site. The mean surface
area index for mounds found to produce sexual brood was marginally greater than that for
mounds that produced only worker brood (p = 0.0560, Table 4), but this relationship was
variable among sites (Figure 3). Distance from nearest neighboring mound did not

appear to be associated with production of sexual brood (p = 0.8662, Table 5; Figure 4).

Weight & Development. Females were heavier than males. The mean female weight

(least squares mean, corrected for developmental stage using ANCOVA) across all sites

10

owed

 

mmod $06 co 8: m m 0
«end mood wmod mmm mmm o om m
Smd vwmd med 5 co m m w
3&6 55¢ omwd wen mwm N. E m
Sod mwwd mood in own m m_ m
Rad memo 35o 5m Sam N. o_ _
288 288 0:8 8:8 @0388 @0388 808.830 02m

88 88m800m £803 .30 80:08:: How 082$ 083 :39 £0868 $0888

3288 58038088 8303 ”288 88 088080 83: m.88m800m
:85 888 £903 .30 @8888 .3 “008888 a 28.8 800 88m8oom 0% .3 .888 x3 E8 8% w8888m .N 03:.

 

 

3.2 «3.3. m: 39m 0
3.2 $5.2 89m Noam m
2.3 802 mm; Sod v
mwé ROE SM 20m m
«2. 3a.: 3; 2:. m
ood— mo_.2 mo.m 2mm _
>080 .808 >0va 8808 02m
AEV Hops—m8: 8080: 8 888% $5 808 08.83 2:08

 

 

.0858 882330: 8080: 9 02886 28 808 008.88 00 828300 @8888 28 8808 2% .m 038%

12

 

.5
N

 

Ilsexuals [:1 no sexuals I

 

 

(D
n

 

 

average mound surface area (m2)
0)

 

 

 

 

 

Figure 2. Mounds that produced sexual forms were larger in comparative surface
dimensions than those that did not. Error bars represent one standard deviation.

13

50!

 

 

Ilsexuals Elno sexuals I .,

 

 

 

 

 

 

 

 

 

 

 

a

m I

‘3' 40- I

0

c A

3 g 30. "

8 a

s 2

.2 .9 20' ,_ " -

'U C)

0 C

O) 104

N

h

0 J.

>

a 0* - _l
1 2 3 4 5 6

site

Figure 3. Distance from nearest neighboring mound is not related to production of
sexual forms. Error bars represent one standard deviation.

14

Table 4. ANOVA test of Ho: Surface area indices of mounds producing sexual caste
brood are equal to those of mounds producing worker brood only.

 

 

Source of variation df mean square F p
mound status 1 28.55 3.84 0.056
error 46 7.43

total 47

Table 5. ANOVA test of H0: Distance to nearest neighbor is equal for mounds that
produce sexual forms and mounds that produce workers only.

 

 

Source of variation df mean square F p
mound status 1 4.44 0.03 0.8662
error 46 154.64

total 47

15

was 0.0101 (stderr = 0.00007; n=276). The mean male weight across sites was 0.0081
(stderr = 0.00004; n=1100).

While significant differences in pupal weights of both sexes existed among sites,
variation in weights was also found among mounds within sites. Male and female pupal
dry weights with standard errors at each site are shown in Figure 4.

Some of this weight variation was linked to variation in stage of pupal
development. Significant negative regression relationships existed between development
stage and pupal dry weight for both sexes (Table 6, Figure 5). The regression slopes
were similar between sexes, but development stage explained less variation in male
weight (r2 = 0.1787) than in females (r2 = 0.2934), due to the presence of outliers in the
male weight data. Mean stage of pupal development was not different among mounds or

among sites (Figure 6).

Energy Content. There was no significant difference (p = 0.5576) between energy
density (corrected Hc) of male and female pupae (Table 7 a,b; Figure 7). The least-
squares means of energy density for males was 5526.2 cal/g, and for females 5507.9
cal/g. The overall nested ANOVA model: Hc = day (site (sex))) was not significant at p
= 0.5001. As pupae could not be combusted individually, it was impossible to test for

differences in caloric density among stages of development.

16

0.015

 

 

 

 

 

 

 

§
.3 0.01 - § § (I)
3 i!
.
1:
7. 0005.
Q ,
a Ofemales
lmales
O I T I I I *I
1 2 3 4 5 6

site

Figure 4. Females were heavier than males. Error bars represent one standard
deviation.

l7

Table 6. Analysis of covariance test of hypotheses H0: Weight of male and female
pupae are equal; HO: There is no significant relationship between weight and
developmental stage; and H0: This relationship is the same for both sexes.

 

 

Source of variation df mean square F p
sex 1 0.0002445 160 0.0001
development 1 0.0003666 239.92 0.0001
sex*development 1 0.0000124 8.14 0.0044
error 1353 0.0000015

total 1356

18

 

male 0 female
- - - Linear (male) -—Linear (female)

0.016

 

0.014 -

0.012 -

0.010 -

0.008 1

0.006 -

 

pupal dry weight (g)

0.004 -

0.002 -

 

0.000

 

 

stage of development

Figure 5. Pupal weight decreased over development. Regression slopes were
significant at p = 0.0001; ANCOVA tests are provided in Table 6.

19

 

     

 

 

 

Imale
a: 4. Ofemale
a:
£3
to
C
d) 3-
E
a
2
24
1:
C
8
E 1-
O l r I I I I
1 2 3 4 5 6
site

Figure 6. Average stage of pupal development did not differ significantly between
sexes or among sites.

20

5800

 

 

 

 

 

5700-
2"?
E 5600.
3
o
I 5500-
m
c»
g 5400-
>
a

5300 - Ofemales

lmales
5200 fi I I r I I
1 .
0 2 3 sue 4 5 6

Figure 7. There was no significant difference in energy density (H) of

reproductive caste pupae between sexes or among sites. ANOVA tests are
provided in Table 7.

Table 7. Nested ANOVA performed on energy content data, testing Ho: Hc is equal for
all sites and both sexes within sites. Hc = heat of combustion, in calories per gram.

 

 

 

 

Source of variation df mean square F p
model 6 15298 0.9 0.5002
sex 1 5904 0.35 0.5576
site 5 74435 0.88 0.5021
error 74 17013
total 80

 

 

 

22

Sex Ratios. As males and females had no consistent differences in energy content,
caloric content data could not be used to estimate sex ratio. Numerical and dry-weight
estimates of sex ratio were statistically different (t = 5.64997, df = 32, p = 0.0001). Due
to the differences between male and female weights, using numbers of individuals to
estimate sex ratio overestimated investment in males by an average of 2.8% compared to
estimates based on weights. Whether this difference would be meaningful in a larger
context depends largely on the precision of theoretical predictions tested. Applying
Boomsma’s (1989) correction increases the male bias to a greater extent than using the
original numerical investment ratio. As sex ratio estimates based on dry weights seem
more accurate in determining investment, and the Boomsma correction is accepted by
convention, Boomsma-corrected weight ratios are discussed in the rest of this thesis.

Sex ratios were predominantly male-biased, but ranged from all-male (1.0) to
almost all-female (0.08). Although considerable variation in sex ratio was present within
sites, polydomy of F. exsectoia’es colonies makes mound to mound comparisons
questionable without further genetic information. Thus sex ratio estimates from all
mounds within a site are treated as samples of a single population, and only comparisons
between population averages were performed. Population mean numerical and weight-
based sex ratios are shown in Table 2.

Analysis of variance showed differences in sex ratio among sites (p = 0.0245,
Table 8; Figure 8). Sites fell into two categories: those that were male-biased (sites 1, 2,
3, & 6), and sites with sex ratios not differing from 50/50 or equal investment (sites 4 &
5). Sites 4 and 5 still contained male-biased mounds, but these were balanced by female-

biased mounds.

23

ratio (male weight I total

 

 

site

Figure 8. Sex ratio was predominantly male biased.

24

Table 8. ANOVA test of Ho: sex ratio is equal for all sites.

 

 

Source of variation df mean square F p
site 5 0.0806 3.10 0.0245
error 27 0.0230

total 32

25

Discussion

Mound measurements and collection of brood. The relatively low percentage of F.
exsectoides mounds containing reproductive brood may be simply explained in the
context of a polydomous colony—one or a few mounds within the colony may contain all
of the colony’s sexual offspring. While this may seem like “putting all of the pupae in
one basket”, several advantages are possible. The environment in some mounds might be
better suited for production (or pupation) of sexual offspring. Some gain in efficiency of
brood care might be attained by clustering sexual brood in a few mounds, rather than
scattering it among many. Alternatively, a lack of production of sexual forms may reflect
allocation to asexual colony expansion (budding). Budding should be a more successful
method of expansion within habitats, since it does not depend on availability of F. fusca
host nests. Alate sexual forms are usually vulnerable to predation during dispersal
(Holldobler and Wilson 1990), and may be an unnecessary investment in stable habitats.
While mound surface area is not a good surrogate for direct measures of mound
population (Cory and Haviland 1938), or even mound volume (Bristow et. al. 1992), the
measurement has some value in comparisons between established mounds and newly-
formed buds. As mound size increases over time (Andrews 1926, Haviland 1948), the
tendency for mounds producing sexual brood to be larger in surface dimensions than
mounds producing only workers may reflect differences in colony maturity (Figure 3).
Measurements of distance to nearest neighboring mound may be expected to reflect
crowding, local competition for resources, and frequency of budding, all factors that

might influence reproductive allocation within a mound. Knowledge of relatedness and

26

interaction behavior is necessary to understand which mounds are “neighbors” and
which function as parts of the same mound. Current data do not show an effect of

neighbor distance on reproduction.

Weight & Development. Differences in pupal development within and between mounds
may arise from environmental factors (mound temperature, placement of pupae within
mound), colony factors (time of egg-laying, maternal effects, nutrition), or genetic factors
influencing development time. The data obtained in this study give no indication that
males and females differ in development rates or emergence times (Figure 5, 6). While
sex of “undeveloped” pupae cannot be determined (inviting conjecture that differences in
development time may obscure the true sex ratio), relatively equal proportions of males
and females at each developmental stage lends support to the contention that
“undeveloped” pupae represent a random assortment of males and females (Figure 5).

Weight data indicated significant dimorphism between males and females. but
this dimorphism was small relative to that shown by many ant species (Boomsma 1989,
Crozier and Pamilo 1996). This is consistent with the founding biology of F.
exsectoide.s'—species with dependent (non-claustral) founding tactics, such as budding
and social parasitism, may invest less in individual females. Founding biology also
explains the lack of difference in male and female energy density, as females do not
require large fat reserves for founding new colonies.

The sexual dimorphism in weight was partially masked by the presence of male
weight outliers. These were usually males that were as heavy or heavier than females.

These males were often also outliers with respect to appearance, possessing

27

disproportionately large heads. The presence of these “heavy” males might indicate a
dispersal polymorphism such as that described by Fortelius el. al. (1987) and Agosti and
Hauschteck-Jungen (1987) in Formica exsecta, in which small males dispersed while
larger males mated in the vicinity of the nest. The low frequency of these males among
F. exsectoides sampled argues against that explanation. Alternatively, these males could
be diploid. Nipson’s (1978) study, as well as the colony structure and mating behavior of
F. exsectoides, suggests that significant inbreeding may occur within populations.
Inbreeding may lead to the production of diploid individuals homozygous at sex-
determining loci—intended females who are morphologically male and reproductively
dysfunctional (Crozier 1971, Pamilo eta]. 1993). The presence of diploid males as a
significant fraction of the reproductive brood produced would lead an overestimation of

male investment, and should be investigated in the future.

Sex Ratio. The strong male bias in sex ratio displayed by F. exsectoides is surprising, as
relatedness asymmetries in eusocial Hymenoptera are expected to result in worker
preference for a female bias (Trivers and Hare 1976). The male bias observed is
consistent with observed sex ratios for other ant species that display colony expansion
through budding (Bourke and Franks 1995, Pamilo and Rosengren 1983). Unfortunately,
budding is associated with a suite of characteristics, such as polygyny and polydomy that
may equally influence sex ratio (Boomsma 1993). The extent of any or all of the above
traits may be related to genetic, social, or environmental factors.

Polygyny may lead to an increased worker preference for males, but the

functional extent of polygyny in this species is incompletely known. F. exsectoides

28

colonies are peculiar in that “foreign” queens are sometimes adopted into existing
mounds (Creighton 1950, Nipson 1978). This apparently selfless act of adopting a
stranger and raising her offspring might be explained if relatedness within populations is
high (decreasing the genetic level of polygyny) or if inbreeding has depressed the
population (by increasing diploid male production) to the extent that novel genetic input
is desirable.

In highly polygynous colonies, relatedness between sisters will decrease (unless
all queens are highly relatedl). Workers in these situations may attempt to produce their
own (male) offspring. However, workers should prevent reproduction by their less-
related sisters, leading reciprocal policing and a lack of worker-produced males.
Additionally, workers of many ant species lack ovarian function, or are inhibited by the
presence of a healthy queen. There are no firm reports of worker reproduction in queen-
right colonies of Formica species (Crozier and Pamilo 1996). F exsectoides has only
been observed to lay trophic eggs in queenless laboratory colonies (C. Bristow, personal
communication). All colonies found to contain sexual caste pupae also contained worker
pupae, so it is unlikely that these mounds were queenless. While the majority of males
in this study could be produced by workers, the production of such large numbers of
worker-laid males in queenright colonies could indicate a decline in colony stability and a
possible cause of colony failure (Starr 1977, Crozier and Pamilo 1996).

The difference in sex ratio between sites suggests an environmental influence.
Two sites located within (site 4) or near (site 5) recently clear-cut forests had sex ratios
not different from 50/50, while all sites located in mature forest had strongly male-biased

sex ratios (Figure 8). If this difference arises from a true environmental influence

29

(occurrence of aberrant sex ratios in these areas may be due to chance), two factors that
might encourage increased female investment are food and availability of nest sites.
Resource availability may contribute directly to sex ratio, either through developmental
effects (lack of food results in fewer sexuals, more workers (Deslippe and Savolianen
1995, Herbers and Banschbach 1998)) or by influencing reproductive allocation. The

theory of local resource competition (Clark 1978) states that, in a resource and dispersal

limited environment, male production will be favored since female offspring will
compete directly with the mother for resources. While this theory is not consistent with
observations in this study (clear-cut areas may be expected to have fewer resources than
mature forest), it is certainly possible that differences in sex ratio among sites represent
population-level strategies for coping with ecological conditions at each site. The effect
of environment on sex ratio may also be indirect—Herbers (1986, 1993) shows the
impact of ecological factors on queen number, which might be expected to affect sex
ratio by altering the relatedness structure of the colony. More detailed comparisons
between populations along a gradient of environmental conditions would be necessary to
determine the existence of these strategies and the factors regulating them.

Factors that cause Formica exsectoides to present particular difficulties in sex
ratio studies, such as polydomy, social parasitism, and specialization on transient
habitats, increase the need to perform such studies, since understanding sex ratio and
reproductive dynamics in only discrete, monogynous laboratory colonies gives an
incomplete picture of the interaction of genetic and environmental factors in shaping
reproductive strategy. While perhaps unfortunate from the standpoint of those hoping to

integrate sex ratio studies in large-scale evolutionary analyses, this study indicates that

30

intraspecific variation in sex ratio exists in F. exsectoides, whether due to environmental
or genetic influences. As significant variation in sex ratio occurred between sites within
a specific habitat (jack pine), whole species generalizations of sex ratio based on one or

two studies will surely misrepresent true sex ratio dynamics of many ant species.

31

CHAPTER 2

Age of jack pine stands influences abundance of the Allegheny mound ant

(Hymenoptera: F ormicidae) and other terrestrial arthropods

Introduction

Formica exsectoides (F orel), the Allegheny mound ant, has been presented by
Oster and Wilson (1978) as a primary example of a fully polygynous species. They
propose that F exsectoides fits an “adaptation syndrome” for habitats that are stable,
long-lived, and patchy. These habitats are primarily persistant grassy clearings.

The syndrome includes traits displayed by F. exsectoides that allow it to rapidly
expand to highly dense populations within a habitat. Such traits include polygyny
(possessing multiple queens per colony) and polydomy (occupying more than one nest
site per colony). Polygyny provides colonies with the potential for earlier production of
sexual forms than single-queen colonies (Keller 1991). Polydomy and polygyny allow
colonies to expand by mound budding or fission, rather than or in addition to the slower
and riskier process of establishing new colonies through dispersal of alate sexual forms.
F. exsectoides colonizes new habitats through temporary social parasitism of Formica
fusca colonies, possibly an adaptation that allows F. exsectoidcs to skip the vulnerable
early stages of claustral founding, creating lower rates of failure in young colonies and

allowing rapid growth of the population (Bourke and Franks 1995).

32

In Michigan, F. exscctoides occurs in high abundance in jack pine (Pinus
ban/cs‘iana Lamb.) (Bristow et. a1. 1992, Bishop 1998). An early successional tree
adapted to cold climates and poor, sandy soils, jack pine is valued for rapid
recolonization of burned areas and for covering land of poor soil quality. Jack pine
management is usually focused on pulpwood production, but in Michigan, jack pine is
managed extensively to create habitat for the Kirtland Warbler, an endangered species
(Rudolph and Laidly 1990). Jack pine stands are generally maintained by wildfire. In
the absence of fire, jack pine does not recruit well naturally—clear-cut stands are often
replanted (Beland and Bergeron 1993. Duchesne and Sirois 1995). Jack pine stands may
persist 80-100 years on poor soils. but it is a short-lived species and will eventually fall
prey to disease or insect damage.

No habitat is truly stable—in fact, as an early successional community, jack pine
stands may be considered quite transient. Little is known about causes of population
decline and colony mortality in F cxsectoides, but habitat changes may play a role.
Researchers studying F. cxscctoidcs populations (Andrews 1926, 1929; Haviland 1948,
Dimmick 1951) have reported observations suggesting that habitat suitability, particularly
the light environment, may decline over as short a time span as 10 years. While
distribution (Bristow et. al. 1992), mound placement (Andrews 1926) and population size
(Cory and Haviland 1938) have been investigated, these studies focus on single
populations of F. cxsectoidcs. rather than presenting comparisons between populations in
differing habitats.

I examine changes in F. cx.s'cct()idc.s' populations within various stages of this

transient habitat. The high density and aggressive nature of F. exscctoidcs (Creighton

33

1950) gives it many important roles in jack pine communities. While the effects of F.
cxscctoidcs populations on arboreal insect communities have been examined by Bishop
(1998), no data is available on the relationship between F. exscctoides and the ground-
dwelling communities of jack pine. The majority of arthropod study in jack pine is based
on the economic pests, such as Choristoncura pinus (jack pine bud worm) (Nealis and
Lomic 1994, Volney and McCullough 1994). Studies of insect communities have
generally focused on early stages of jack pine growth (Naylor and Bendell 1983, Spires
and Bendell 1983, Beaudry et. al. 1997). Accordingly, this study also examines the
ground arthropod community co-occurring with F. exsectoides and changes in abundance

and composition of feeding guilds as jack pine ages.

34

Materials and Methods

Site selection & description. All sites described in this study are located in the Huron
National Forest, Oscoda and Crawford Counties, Michigan (Figure 9a). Using maps
showing stand composition and date of origin (date since last cut), the Big Creek (Oscoda
Co. T25N, er) and Eldorado (Crawford Co. T25N, rlE) areas of the Huron National
Forest were surveyed for presence of F. exsectoides. Of stands containing significant
populations (>10 mounds), I selected five stands in each of three age classes for inclusion
in the study. The three age categories were: young (2-16 years), mid-regrowth (37-44
years), and old (64-77 years). Stand codes, names, United States Forest Service

designations, ages, and specific notes on sites are shown in Table 9.

Location and density of F. exsectoides mounds. Within each stand, I marked a 50m x
50m plot with wire stake flags. This plot was mapped using a grid of 5m2 increments.
compass, and surveyor’s rOpe. Locations of all trees and all anthills were recorded.
Longest slope and shortest SIOpe were measured with a tape measure, and multiplied to
yield a comparative measure of mound surface area in m2. Species, distance to mound
(m), and diameter at breast height (dbh) in cm were recorded for all trees greater than 1

cm in diameter within a 5m radius of each mound that showed activity (was not derelict).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.x
EL:
t...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b

Figure 9. a) Stands were located in Crawford and Oscoda counties. Michigan. Letter

indicates stand age class: Y = young, m = mature, o = old. b) Plot layout and pitfall trap
locations.

36

l\)

Table 9. Stand locations and descriptions for Chapters 2 & 3

 

Site USFS Age Stand Notes
Code compartment Class Age
& stand #

Y1 94/18 young 2 recent burn of 62 year-old redpine stand,
debris still on ground; open/grassy

Y2 128/21 young 9 area cleared by a 3-acre fire within 6l-yr old
jack pine; sparse/grassy

Y3 78/1 young 16 naturally regenerating clearcut, very sparse

Y4 77/25 young 15 dense jack pine plantation, all trees 2-2.5m
in height

Y5 75/7 young 8 dense jack pine plantation; all trees 1.5-2m
in height

M1 77/1 mature 43 . sparse stand, partially burned in 1966, many
mature trees; grassy

M2 94/ 5 mature 44 mixed understory of grasses/moss; high
recreation traffic

M3 56/ 1 9 mature 40 partially open, grassy

M4 49/35 mature 37 partially Open, grassy

M5 54/8 mature 42 extremely dense stand, much undergrowth,
oak understory

01 72/50 old 64 much storm damage during 1997

02 73/17 old 77 many openings due to tree fall, grassy areas

03 55/ 12 old 74 very small stand, bordering on young
jackpine plantation

O4 122/7 old 72 mature trees, primarily fern & moss
understory

05 75/6 old 72 mature trees, primarily fern & moss

understory

37

Abundance of F. exsectoides and other ground-dwelling arthropods. To estimate the
proportions and relative abundance of insect groups within these different jack pine
stands, I placed five pitfall traps in each plot in the arrangement shown in Figure 9b.
Pitfall traps consisted of two plastic cups (11 cm diameter, 7.5 cm in depth) nested inside
each other and buried with the opening flush with the soil surface. Traps were filled with
ethylene glycol to preserve samples until collection. Plastic lids held approximately 1cm
above the soil surface by roofing nails helped to minimize debris and small mammals in
traps. Trap lids were completely closed during non-collection periods. Traps were
installed June 28-30, 1997, but left closed for one week to reduce captures of curious ants
investigating the disturbance in the soil. Traps were opened for three one-week periods,
with collections taking place on July 8-9, July 15-16, and August 3-4, 1997. Traps were
emptied by removing the upper plastic cup and transferring the contents to a separate
container filled with 95% EtOH. allowing replacement of the cup with minimal
disturbance to the trap area.

Preserved pitfall samples were sorted and identified to either order or family level, with
the exception of ants and Carabid beetles, which were identified to genus or species level.
Ants were identified or confirmed by Gary Coovert, Dayton Museum of Discovery,

Dayton, OH. F. Purrington, Ohio State University identified carabid beetles.

Data analysis.
Location and density of F. exsectoides mounds. To test whether mounds were
clumped, randomly, or uniformly dispersed, I calculated Lloyd’s indices of mean

crowding (X*) and patchiness (Pielou 1969), using a quadrat size of 25 m2 (4

38

quadrats/plot). Since these indices are sensitive to quadrat size, I also performed a
nearest neighbor analysis as described by Clark and Evans (1954) and Pielou (1969).
Mound density was calculated as the sum of the total number of mounds divided by the
total plot area. Tree density for each plot was calculated as the mean dbh of trees within
the plot, multiplied by the number of trees per square meter, with the product variable in

units of cm (tree diameter)/m2 plot area.

Abundance of E exsectoides and other ground-dwelling arthropods. Maj or
taxonomic groups present in> 25% of samples and comprising > 0.1% of total
invertebrates sampled were analyzed for differences among stand age classes using
individual one-way analyses of variance (PROC GLM, SAS Systems, Inc.) A sequential
Bonferroni correction was then used to assess the significance of the results (Rice 1990).
Relative abundance of all major groups were also correlated to stand age, stand density,
and to each other using a correlation analysis (PROC CORR, SAS Systems, Inc.), also
corrected with a sequential Bonferroni procedure.

Composition of ground-dwelling insect communities. While families and genera
within some orders (e. g. Orthoptera) are ecologically equivalent (due to similarities in
feeding behavior), other higher taxonomic groups (such as Hemiptera and Coleoptera) are
more diverse. The groups are therefore subdivided according to feeding behavior, using
the categorizations of Herbivore, Predator, and Scavenger. Allocation of organisms to
feeding category was based primarily on family-level feeding ecology described in
Borror, Triplehom, and Johnson (1989) and Arnett (1993). Ants, because they feed at

multiple trophic levels and display complex behaviors such as homopteran tending, are

considered separately from these designations as per the ecological guild divisions
established by Moran and Southwood (1982). Parasitoids were noted but not included in
the analysis.

Relative proportions of each feeding group examined were arcsine transformed and
simultaneously compared within and among age classes using a multivariate analysis of
variance (MANOVA) (PROC GLM, SAS Systems, Inc. 1990), with canonical correlation
analyses performed to aid in interpretation of the analysis (Scheiner 1993). Similar
MANOVA tests were performed to compare relative proportions of ant genera/species
across stand ages. Relationships between feeding groups were examined by correlation
analysis. Overall analyses were performed, then correlation analyses for each stand age
class were performed separately to explore the interaction of age with feeding group

correlations.

Stand ages, mound and tree densities, mound surface area indices, and relative densities

of F. exscctoides (numbers of workers per trap) were correlated using PROC CORR

(SAS Systems, Inc. 1990).

40

Results

Location and density of F. exsectoides mounds. Age-class means and standard errors
for mound density and tree density are listed in Table 10. Average tree density increased
with stand age (Figure 10), but stand age explained less than 50% of the observed
variation in density (R2 = 0.4763, p = 0.0063).

Mound density (mounds/m2) was marginally correlated with stand density (R2 =
0.3603, p = 0.0232) (Figure 1 la). Further examination revealed an interaction of this
relationship with stand age class—mound density had a strong negative correlation with
stand density (R2 = 0.9948, p = 0.0026), but only in young stands. Density of F.
cxscctoia’es mounds was greatest in young burned (Y1 and Y2) and mature jackpine
stands (Figure 11a, b).

Nearest neighbor analyses showed that mounds were neither clumped nor
overdispersed at the plot level; the mound dispersion pattern was not significantly
different from the null expectation of randomness (p>0.05) (Table 11). Lloyd’s Index of
mean crowding showed an increase in number of mounds per 25 m2 patch from young to
mature stands, but a decrease from mature to old stands. Calculated values of Lloyd’s
patchiness index are similar across stand ages and approximately 1, indicating a random
distribution of mounds in all three stand age classes (Table 12).

Mound surface area estimates ranged from 0.43 m2 to 21.3 m2. Mounds did not
appear to be larger in older stands (Figure 12a), but site means for surface area estimates

increased significantly with stand age (R2 = 0.3542, p = 0.001) (Figure 12b).

41

Table 10. Stage age class means and standard deviations for mound and tree densities.

 

mound density tree density
stand age class (mounds/m2) (cm dbh/m2 plot)
young 0.0042 (0.0012) 0.53 (0.33)
mature 0.0047 (0.0012) 1.05 (0.54)
old 0.0030 (0.0010) 1.37 (0.22)
overall 0.0039 (0.0013) 1.02 (0.50)

42

 

 

 

 

 

A
w
2 1.51
a.
N
E
3 1
a I
.C
.a
1:
E 0.5a
A
=0.0063
A P
0 1 T u
0 20 40 60 80
stand age (y)

Figure 10. Stand density increased significantly with stand age. Density values
represent both average diameter of trees in plot and density of individuals (stem

density). No value is present for Y4, due to missing data on stem density for some
areas of the plot.

43

 

1.5-l

2..
05* R -0.3603

p = 0.0232

I I I

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007

 

 

cm dbh per m2 plot

 

 

O

mounds l m2

 

O young mature . old

 

 

‘6 2
a b
N 15 '

E

h

d) 1 .

0- Y5 Y3

fi 0 5 R2 = 0.9948

’3 ' ‘ p = 0.0026

E Y2

O I I I I I I

 

 

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007

mounds I m2

Figure 11. a. Mound density was negatively correlated with stand density. b.
This correlation was strongest when only young stands were considered.

44

Table 11. Nearest neighbor analysis of mound dispersion (clark & Evans 1954).
performed by calculating an index of Nonrandomness (R), by the formula R = 2r(p)"'5.

(Where r = distance to nearest neighbor, p = population density (mounds/m2»
Analyses were performed on individual mounds and R tested against a null expectation
of random dispersion (R = 1) using a t-test as prescribed in Pielou (1969).

 

young mature old
n 52 59 37
mean r (m) ' 14.7 12.6 16.3
mean r 0.0042 0.0047 0.0030
meanR 1.30 1.14 1.10
non-random? no no no

Table 12. Lloyd's indices of mean crowding (X*) and patchiness. X“ =
mean # of neighbors per individual within a sampling unit. Sampling units
were 1250m2 (4 units/plot). The patchiness index indicates dispersion, with
a null expectation of 1 indicating random dispersion.

 

young mature old overall
11 20 20 20 60
mean (mounds/25m2) 2.5 3.0 1.9 2.5
var (momds/25m2) 3.1 1 3.05 2.16 2.86
X" (mean crowding) 2.74 3.02 2.06 ' 2.64
patchiness 1.10 1.01 1.05 1.06
non-random? no no no no

45

25

 

 

 

 

 

 

 

 

 

 

 

 

 

NE a

m 20» — ~ ~ - — _ ,_ —. — .1 _ _ 2. , . L___.
2

g 15 .1- , . . , ~ , , ~ 02 7 i 7 2. - fl , L__n
8

t 10 _ __9:. _ . . .-____ -_~__ _ _ _

3 .1— . . 1 .m
‘0 e I. .
g 5-17 1 v L, L A. §~ k 7 z "— 2

o e g.

E o . . . .

0 20 40 60 80
stand age (y)
0 young ® mature . old I

“7‘ 8

E

I? b o 9 9

2 6- A
(I

§ '0 '
m

0

g 21 o R2 = 0.3542

0 =

E p 0.001

5 0 . . -

E 0 20 40 60 80

stand age (y)

Figure 12. a. Mound surface area indices were not larger in older stands.
b. Average mound surface area indices increased with stand age, possibly due to
incomplete budding and a lack of small mounds.

46

Abundance of F. exsectoides and other ground-dwelling arthropods. A list of
arthropods captured in jack pine is presented in Appendix A. Table 13 provides the mean
number of individuals per trap and standard deviation of major invertebrate orders and
families found in jack pine, as well as the percentage of all traps containing at least one
individual from the group. Probability values reflect a test of the null hypothesis of no
difference between stand age groups; 8 of 14 taxonomic groups tested were significantly
different across stand age classes. Changes in total abundance of these same groups over
collection dates and stand age are shown in Figure 13.

Arthropods were partitioned into the following groups: Ants, Herbivores,
Predators, Scavengers, and Parasitoids. Organisms not included in these analyses due to
low abundance or low frequency of occurrence included: gastropods (snails and slugs),
oligochaetes (earthworms), and isopod crustaceans. Collembola were abundant in several
samples but could not be counted accurately. Flying insects, particularly scavenging
Diptera such as the Sepsidae and Lauxaniidae, were considered incidental. While these
insects were probably attracted to the traps by odor cues, their representation in pitfall
traps may not represent their normal range of ground activity.

Multivariate analysis of variance showed stand age class differences in
proportions of feeding groups captured (Pillai’s Trace: F = 5.5409, df = 10, p = 0.0001;
Roy’s Greatest Root: F = 7.4226, df = 5, p = 0.0001). Examination of standardized
canonical coefficients revealed that changes in proportions of predators accounted for
most of the variation between age classes, followed by scavengers, then ants (Table 14).
Groups showing significant changes in univariate analyses were Predators (F = 14.33, p =

0.0001), and ants (F = 6.34, p = 0.0021). Changes in scavenger proportions across stand

47

.88 voEuommcmbc: 2m “38:39:. 22530: 235% can 282 .98 c~:_._o.:m:m:-wo_ .8 £3.35. 38:2 2:57: e

 

«as... Ame: S... 5.: a... 3.: :.~ :. €2.23:
:8... and m5 5.: 4m... 2...: 2.... a 82:25
~83 a»: 8.6 an: 36 and New ea 85%

Sigma

3.8 as and Ed A8. : one an 2%: 828203

$23 90: Ed 82: ”no so: 8d mm 2238:
82:. A8: 8... see a... an: mu... m 32.88::
2.26 Geo: we: at: 2.8 5.”: $2 a 82653

883862

88... 5.: NS 36.: a: an: and em sees

:36 c n: 3% 3e. : as $6.: 8: a. case: 22828

88... an: 5% ea: :4 an: mm... mm 223323 2638.6

38... an: «.2 en: mi. 63: e2 2. 22555
88... an: we... 5.: m2. 5.: a... NM 29%
.88... as: ”2: $2: 36 E2: mad mm 8225:..— see
.88... 3...: mm 3: new A8: SN 2: essence 8.3.5

anocoeehm

a no": Eu: own: 33.x.
Bo 8:ng wade»

.meomtanEoo 2:22: 8: weapons?“ Eobomeom Essen—com baa “magnum: baseman new 68:43.3 E $85 96%
03888» come no <>OZ< .3365 3 wosfioeow again Accuses: :32va gab .6: mfiaEzwfi .8 $382. :88 was
.132sz one ammo— .m wEEmES was: an :0 .x. 63 Become“: 82m> dag ENE: E 682:8 $8565.6— .m_ «3.2.

48

 

25000
a. F. exsectoides
20000 -
15000 -

10000 -

total collected

5000 -

 

 

 

O

 

 

7/15 8/3

sample date

350
300 _ b. other Formicids
250 -
200 T
150 -
100 -
50 -
0

 

total collected

 

 

 

 

 

 

 

 

   

7/15

sample date
IDyoung Imature loldj

 

 

Figure 13. Total number of individuals captured per stand age class, sorted by
taxon group and date. Taxon groups presented were present in over 25% of all
traps and accounted for more than 1% of all invertebrates captured. These include:
a) F. exsectoides, b) other Formicids, c) Orthoptera, d) Lygaeidae, e)
Thyreocoridae, f) Aranae, g) predaceous Coleoptera, h) other Coleoptera, i)
Opiliones, j) Diplopoda, k) Diptera.

49

350
300 . c. Orthoptera
250 -
200 -
150 -
100 -

50 -

 

total collected

 

 

 

 

 

 

7/7 7/15 8/3
sample date

1000

 

d. Lygaeidae
800 -

600 ~

400 -

total collected

200 .

 

 

 

 

 

7/7 7/15 8/3
sample date

160
140 . e. Thyreocoridae
120 -
100 -
80 -
60 -
40 -
20 .
o .

 

total collected

 

 

 

7/7 7/15 8/3
sample date

 

IEIyoung Imature Iold

 

Figure 13. (CONTINUED)

50

 

 

 

 

 

 

 

 

 

f. Aranae
-o
0
‘6'
2
3
U
:3
.9
7/7 7/15 8/3
sample date
120
100 g. predaceous Coleoptera
U I
2
U
2
‘o
0
B
.9
7/7 7/15 8/3
sample date
40
1: h. other Coleoptera
0
‘6
2
'6
0
£3
.9

 

 

 

7/7 7/15 8/3

sample date

 

IDyoung Elmature Iold

 

Figure 13: (CONTINUED)

51

 

i. Opiliones

total collected

 

 

 

7/7 7/1 5 8/3

sample date
80
7o .. j. Diplopoda
60 -
50 -
40 -
30 -
20 -
10 -

 

total collected

 

 

 

 

 

7/15 8/3
sample date
140
120 q k. Diptera
100 .
80 -
60 -
40 -
20 -

 

 

total collected

 

 

 

 

 

 

 

 

7/15 8/3
sample date

 

fiyoung Imature Iold I

 

Figure 13. (CONTINUED)

52

 

 

mom; 5%.: 83.0- omwod £8: cod mmm Snood mMmOZm><Um
$3.: $3.. T 833. 2.3:- 58.: m3; NNN 98:: mmOH<dem
o2 mg Seed- moood- memo: comm: 02 «mm 038.: mmmoamxmm
Novad mime- wmad mwovd Soc: 3.: man 38:: mHZ<
Am: 85:82.9,“ CV 352556 Am: 5253 2V 5233 a m d: mm:
$56508 .8288 <>OZ< 035,83

856508 .880ch .wommflu own :53 $88 :33 us 33:88: 23m weave»: ”om mafia mug—m5 8396: .2 «ES.

 

 

.5505 8a :8. 05 E @896va 085 3.5 smog <>OZ<E 3 Sunbeam v.53

age classes were marginally significant (F = 3.90, p = 0.0216). Organisms assigned to

each group and changes within feeding groups are presented below.

Ants. Ants were the most abundant group of organisms, comprising 87-93% of all
invertebrates captured. Relative abundance of F. exsectoides foragers was not linearly
correlated with stand age, mound surface area estimates, or tree density, but was
significantly and positively correlated with mound density (R2 = 0.51, p = 0.0027)
(Figure 14). F. exsectoides foragers were most abundant in mature stands.

Only 2.2% of ants captured were species other than F. exsectoides. Ant
genera/species captured in more than 3 traps are listed in Table 15, as well as the mean
number of individuals per trap, standard deviation, and the percentage of all traps
containing at least one individual of that genus/species. Probability values reflect a test
of the null hypothesis of no difference between stand age groups; only Formica fusca (p
= 0.0005) and Aphaenogaster spp. (p = 0.0001) were significantly different in average
abundance per trap after correction for multiple comparisons.

MANOVA tests of stand age class differences in proportions of ant species were
highly significant (Pillai’s Trace: F = 5.2734, df = 8, p = 0.001; Roy’s Greatest Root: F
= 8.5993, df = 4, p = 0.0001). Standardized canonical coefficients indicated that these
differences were largely due to shifts in abundance of Aphaenogaster spp. and
Dolichoderus plagiatus, and that the correlation between groups shifted across age
classes (Table 16).

Abundance of F. exsectoides was not correlated with abundance of other ants as

a group (R2 = 0.0126, p = 0.1387) (Figure 15). In individual species comparisons,

54

01

 

 

‘3‘

'a

34-

2

a

a.

83-

2

o

O!

gzl

l.-

E

:11- R2=O.51
3 p=0.0027

 

 

O

 

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007
mounds I m2

 

0 young 69 mature . olcfl

 

Figure 14. Mound density was positively correlated to log-transformed abundance
of F. exsectoides foragers captured.

55

.828 BEofifibfi. Bob Bigot 2w 823330 @8283 was 282 .83 vogofigbm2 .«o mix—Ea 88:2 o2m>d ...

 

$86 698 23 899 8d 28.: one a. Essa. 55.2.58:
9:3 E8 3o 83: cod 5.8 N; v ..._s§e§ $8855
”33 5.8 So Gas 86 5.: 23 m an; sci
”2 3 808 8.0 868 83 so. 3 as 2 2:283 ...:.%§%Q
398 so: So 29: 36 so. 3 So on $.23 $853
28... A26 2: and NS 88.8 8... mm 838.2:5‘
SS... 818 2.” :93 83 :39 gm mu .9: saunaagiw
.88... SE ”3.8 :3 53mm 3.5 3.3” 2: 33°33» 8..ch

a 20 BBQ: wadox mash .x. 95950

dozooboo Eobomcom “can “Gustav 3:805:me 2a 360% Eon
823% some 8m < >0 Z< 33.28 .3 “888.8% 963 mos—«Si .mmso own :08 com @633 58E ES ._m:E>6:_ use 5.8—

a wafigcoo mag =w .«o a\o ”03 3680a 333/ .ocfi xomm E “82393 35868 ~88 86on Ba Socuw :_< .2 «Ex...

56

 

 

8S2- N33- MES- 83¢ ”22 So 5 288.0 22.328
5.x 3:. ~32 38.0- $35 8.0 «2 8286 8.3.x m
«~32- 830. 982 #8.? have m; ”.2 $83 85%....sz
23- 82 2.53- 23o 58d 3: V2 @883 8323.3 m
S 85855;... E @3655; S c8258 8 5058 a a we mm:
356E000 E06058 <>OZ< Saginaw

 

 

.535 93 “x3 06 E 33:86 085 Eco um“: <>OZ<2 .3 33.2% 2b.:
macaw—oboe fiumcoswu .mommflo own 23% 328 3:3 as 360% Ed .8 mcomtoaoa ”om wcumou 892mg magic: .3 «Bah

57

 

70
g- 60. R2=0.O126 o
f p>0.05 o
m 50"
0- o
3 40- o
5 o
3 30-1 0 0 °
:5 0 0° 00 00 O
o 204 % 0<8
5
2

 

 

 

 

total F. exsectoides per trap

Figure 15. There was no significant relationship between abundance of F.
exsectoides foragers per trap collection and abundance of other ants.

58

F. exsectoides forager abundance showed no significant relationship to any species but
C. herculeanis. Abundance of C. herculeanis was significantly but weakly correlated to
abundance ofF. exsectoides (R2 = 0.0456, p = 0.0001). Weak but significant
relationships between relative abundance and stand age were found in Aphaenogaster
spp. (R2 = 0.1211, p = 0.0001), and F. fusca (R2 = 0.0445, p = 0.0011). Additionally,
abundance of F. fusca was strongly correlated to number of species captured (R2 =
0.7467, p = 0.0001) (Figure 16a). While the large trap catches of F. fusca obtained in Y4
and Y5 (both dense jack pine plantations) lend those two data points strong influence, the
relationship is still highly significant and reasonably strong if they are removed from the

analysis (R2 = 0.4569, p = 0.0001) (Figure 16b).

Herbivores. The Herbivore grouping contains taxa known to feed only on plant
material, by chewing or sucking. These include Orthoptera, Hemiptera (Thyreocoridae,
Lygaeidae), Homoptera, Lepidoptera, Coleoptera (Curculionidae). While overall
proportions of herbivores did not vary significantly among stand age classes (F = 1.39,
dfc = 222, p = 0.2560, Table 14), the composition of the herbivore group changed
dramatically (Figure 17a,b). Abundance of Orthoptera declined from young to mature
stands, then dropped sharply in old stands. Thyreocoridae were present in low numbers
in young and mature stands, but increased from 1% of all herbivores in young and mature
stands to 19% of the herbivores in old stands. Lygaeidae, the most abundant herbivores

in all stand age classes, did not differ with stand age.

59

 

 
 

O

   

# ant morpho-spp.
OJ

 

 

 

 

 

 

 

2 .
R2 = 0.7467
1
l p = 0.0003
0 I I I I I
O 2 4 6 8 10 12
mean fusca per trap
0 young mature . old I
4
b C
d.
a
‘9
o
.C
E-
o
E
'57 1
* l R2 = 0.4569
0 I l l

 

 

 

O 1 2 3 4

mean fusca per trap

Figure 16. a) There was a significant positive relationship between mean number
of F. fusca captured per trap and total number of ant morpho-species captured in
all traps at a site. b) This relationship remained significant even when highly
influential points (Y4 and Y5) were removed from the analysis.

60

Figure 17. Changes in composition of feeding guilds across stand age
classes. a) herbivores b) herbivores (part 2) c) predators d) scavengers

Herbivores are presented in two graphs due to considerations of scale.
Abbreviations of taxa are as follows: Or = Orthoptera, He(L) = Hemiptera
(Lygaeidae), He(T) = Hemiptera (Thyreocoridae), Ho = Homoptera, La =
soft-bodied Lepidopteran and Hymenopteran larvae, Le = Lepidoptera, Co =
ColeOptera, Ar = Araneae, Chi = Chilopoda, Op = Opiliones, Di = Diptera,
Dp = DipIOpoda. Values shown are total pitfall trap captures, summed across
traps, dates, and sites within stand age classes.

6]

 

 

1800 300

E; 1600. ‘3

a 250-
:; 1400.

$3

a 1200- 200-
:5 1000-

.- 150-
.2 600.

i?

i: 500' 100.
0

3k 400-

" 50-
%; 200-

H

 

 

 

0.

 

 

     

He(T) Ho La Le

 

 

 

 

total # of individuals captured

 

400
3504
300(
250‘
200-
1504
100-
50-
0.

 

 

 

total individuals captured

 

 

 

 

 

 

09

Figure 17. Changes in composition of feeding guilds across stand age classes. a)
herbivores b) herbivores (part 2) c) predators d) scavengers

62

Predators. The primary group of non-ant predators captured was predaceous
Coleoptera. These included individuals from the families Carabidae, Staphylinidae,
Histeridae, Cicindellidae. Other predators were spiders (Aranae), centipedes
(Chilopoda), and Hemiptera (Pentatomidae).

Abundance of predaceous Coleoptera increased dramatically in old stands (Figure
17c). The most commonly captured carabid beetles were Myas cyanescens Dejean and
Pterostichus pennsylvanicus LeConte, both reported by Beaudry et. al. (1997) as having
reduced abundance in recently clear cut areas.

Spiders as a group did not differ in relative abundance across stand ages;
compositional shifts among spider families may have occurred but were not investigated.

Centipedes comprised only 0.02% of arthropods sampled.

Scavengers. While many predators scavenge opportunistically, I have restricted the
“scavenger” category to invertebrates feeding primarily on dead or decaying plant or
vegetable matter. These include arachnids of the order Opiliones (daddy long-legs),
Diptera (Mycetophilidae, Bibionidae, Lauxmonidae, Chironomidae, Muscidae, and
Sepsidae), Coleoptera (primarily Scarabaeidae), and Diplopoda. Other scavengers
present but not analyzed included collembolans and isopod crustaceans. Primary changes
in scavenger abundance were due to an increase in Opiliones in old stands and decreases

in Diptera and Diplopoda as stands aged (Figure 17d).

Relationships between feeding groups. Table 17 provides a summary of significant

correlations between major arthropod orders and families of different feeding groups.

63

BO 8 18888 .883 3:0» 5 88:8me ”mom

 

8 $8: $86 a _ .o 666%.:
EV Soaoflou

85% mag» 8 8688 .28 ”mom... emu» .. 586 85.6.6 afimd 288%:

on 3838 83 $86 83. 552:0

88:88 8 83388 $88.» .20 8 8:86 ”mom no» 586 38d #36 fit 8880200
88:80

on a» 88: 83° 2.3. Q 63828

8888 8 83388 .880» .20 E 8:86 we» we» Sood Some End. 88:80

69% mesa 5 8e; 9 amass am an; a» good 886 62.0 226658
8888888 .m

283888 owe $5058me 8 =828 «a 82:28.80 8:08

.888 own 28% 28 Eco 8.: 8288 329 8:885...
8:378 .8 8:888:98 Eotomcom 388:ch a £3» 38888: 888$:me .05 .wESmxm m<m .mmou DOME .8
88828w 8085008 8828.80 888$ u ecouflohoue .883 3828898 8.88 88389 cote—880 .2 035.

()4

Discussion

Location and density of F. exsectoides mounds. Stand age and tree density were
correlated, but stand history and other stochastic affects may have more important
impacts on vegetation than the passage of time alone (Figure 10). Much variation existed
among young sites. due to differences in stand treatment/histories (Table 9). Y1 and Y2
were quite dense, due to recent reforestation by the United States Forest Service. Y3. a
naturally regenerating clear-cut, was extremely sparse because jack pine is adapted for
establishment after burning and may not regenerate well after harvest (Beland and
Bergeron 1993, Zoladeski and Maycock 1991). Y4 and Y5 were of medium density,
with high vegetational heterogeneity due to the patchiness of natural burn regeneration.
Even within mature stands, much variation may exist due to small fires, tree mortality
caused by disease or herbivory, storm damage, and thinning treatments. Old stands are
increasingly vulnerable to damage and have greater probability of impact from random
mortality factors simply by virtue of having been around longer.

F. exsectoides has been described as an opportunist species—although dispersal
between patches may be risky and limited, once a suitable patch is colonized colonies
quickly expand to control the entire territory (Oster and Wilson 1978). This dispersal
limitation may be the key factor explaining high densities and smaller average mound
surface area in young (especially clear-cut) stands (Figure 12b). In clear-cut stands,
removal of trees and some ground vegetation by mechanical harvesting may be followed
by several years of delay before replanting, potential cause for colony mortality (personal

observation). Mechanical disturbance by machinery used to re-plant the stand may

65

further diminish surviving populations, leaving open territories for new colonies.
Recolonization of the site by F. exsectoides may be delayed by several factors. Colonists
arriving at early stages of vegetation regeneration (before prey populations have
established) may lack adequate resources for colony growth. Additionally, F. exsectoides
queens found new colonies through temporary social parasitism of Formica of thefusca
species group, and thus depend on the prior presence of a suitable host species to succeed
in entering a ‘new’ habitat (Wheeler 1928, Wilson 1971).

The larger average mound size in old stands may be related to age of the
population (older mounds may be larger) (Andrews 1926), but it also reflects the
presence of fewer small mounds, possibly new mounds formed by budding or mound
fission (Figure 12a,b). A large number of abandoned mounds and a lack of new mound
establishment show the decline of F. exsectoides populations in old jack pine stands.
Mound distribution was not different from a null expectation based on random
dispersion. Isozyme studies performed by Nipson ( 1978) suggested that most
Fexsectoides populations were the result of a single initial founding event, followed by
budding of the primary colony and occasional acceptance of immigrant queens into these
offshoot colonies. Given budding or colony fission as the usual method of within-site
colony expansion, a null expectation of mound clumping seems reasonable. A lack of
apparent clumping may therefore actually indicate some repellence between mounds, or
specific nest sight requirements (such as microclimate effects) that may overrule any
territory requirements for individual mounds.

Abundance of E exsectoides and other ground-dwelling arthropods. Even at the

coarse resolution of order-level comparisons, it is obvious that major shifts in ground

66

arthropod communities occur as jack pine stands age. Significant changes in proportions
of ants and predators occurred across stand age groups. This is inconsistent with the
findings of Heatwole and Levins (1972), that trophic stucture remains constant on
mangrove islands even while composition of trophic guilds varies. As this study only
incorporates ground dwelling fauna, it may be that trophic guild stability is maintained
when arboreal arthropods are considered. Proportions of herbivores and scavengers did

not change, but alterations in the composition of these groups were evident.

Ants. The numerical dominance of F. exsectoides is in keeping with published accounts
of their large population sizes and pugnacious natures (McCook 1877, Creighton 1950).
Variation in numbers of F exsectoides captured was substantial, and particularly large
samples may have been caused by recruitment of foragers to trap areas. This variation
may have concealed a correlation between stand age and F. exsectoides forager
abundance, as differences were significant when stands were grouped as age classes
(Table 15). Overall, F. exsectoides foragers were most abundant in mature stands, and
least abundant in old stands.

The measurement of mound surface area provides only a comparative index of surface
dimensions. As mound surface dimensions are only weakly correlated to overall nest
volume (Bristow er. al. 1992), it is not surprising that this measurement was not
correlated with forager abundance. Mound density, however, was correlated with forager
abundance, lending support to the intuitive assumption that more mounds means more

ants (Figure 14).

67

 

Levins and Traniello (1981) state that F. exsectoides is not a “truly territorial” ant
(defending “territory” or potential foraging area), but defends only the nest and persistent
food resources such as aphid colonies. Given the density of mounds and the extent of
polydomy in F. exsectoides populations (Holldobler and Wilson 1990), their defense
behavior seems functionally equivalent to that of a territorial speCies. As would be
expected in that case, abundance of other ants was quite low. The absence of significant
correlation between abundance of F. exsectoides and abundance of other ants (either as a
group or as individual species) may arise from the general scarcity of non-exsectoides.
Other species may be suppressed so effectively that responses to fluctuations in F.
exsectoides abundance are not statistically detectable; a true test would require sampling
in areas where F. exsectoides is absent.

The ants most commonly co-occuring with F. exsectoides were Aphaenogaster
spp. and F. fusca (Table 15). Abundance of both showed weak but significant correlation
to stand age, with F. fusca declining as stands matured and Aphaenogaster increasing.
As the host species for F. exsectoides colony founding, establishment of F. fusca in a
habitat must precede colonization by F exsectoides (Wheeler 1928, Creighton 1950). F.
fusca can then be expected to decline as F. exsectoides colonizes young habitats by
parasitizingfusca nests. The relationship might be stronger if stands were sampled
earlier in the colonization process. Aphaenogaster increases primarily in old stands; this
may be due to a habitat preference or the decline of F. exsectoides.

F. fusca is described as “the submissive ant” (Savolianen 1990), or “singularly
lacking in pugnacity” (Creighton 1950). This “docility”, including a relative lack of even

nest defense, may contribute to the large number of parasitic and slave-making species

68

utilizing F. fusca as a host (Holldobler and Wilson 1990). Since Creighton (1950) also
considers F. fusca is also considered to be “exceptionally adaptable in their
environmental relationships”, it seems more likely that their presence is limited by the
presence of other species than by changes in habitat. The strong correlation between
abundance of F. fusca and number of ant species found at each site may indicate that F.
fizsca is particularly amenable to coexistence, or that habitats where F. fusca has been
suppressed are also unsuitable for more aggressive ant species.

It is interesting to note that, of eight ant species found to occur in jack pine, four
are considered “pest” species (Patterson 1994). F. exsectoides alone is considered a pest
within its natural setting, because of the damage it causes to young pine trees (Wilson
1977). C. herculeanis, T. sessile, and M. minimum are pests of human dwellings; T.
sessile and M. minimum are particularly known for their cosmopolitan distribution and
variety of habitats (Smith 1947). Factors contributing to this success may allow them to

coexist with F. exsectoides in a habitat that is comparatively species-poor.

Herbivores. The overall decline in herbivores captured in pitfall traps is correlated to
stand age. This may represent an overall shift in herbivore activity from ground
vegetation to trees as trees mature and ground vegetation becomes less abundant due to
shading or increased soil acidity caused by the decomposition of pine needles (Hendrix
et. al. 1988). The sharp decline in numbers of orthopterans as stands mature is not quite
matched by the sudden increase in Hemiptera in old stands. This change is almost

entirely due to the increase of the small, well-armored Thyreocoridae (Galgupha spp.),

69

generalist feeders on weedy plants. The presence of phloem-feeding insects, particularly

homopterans, in pitfall traps is likely due the individuals being carried by ants.

Predators. Aside from F. exsectoides, the most numerically abundant predator in stands
of all age classes were Aranae (spiders). The abundance of spiders was relatively
constant across stand age classes (Table 13). Halaj et‘. al. (1997) found a negative impact
of ants on hunting spiders in Douglas fir stands, but no relationship between ants and
web-spinning spiders. A similar pattern may exist in this system, but spiders were not
divided for analysis.

The primary source of variation in proportions of feeding groups between stand
age classes was changing predator abundance; this change was largely caused by the
sharp increase in predaceous Coleoptera in old stands. The three families of predaceous
beetles found frequently in jack pine were Carabidae, Staphylinidae, and Histeridae.
Carabidae and Staphylinidae are both very diverse families, occupying diverse habitats.
It is surprising that these families are so poorly represented in young and mature stands.
Beaudry et. al. (1997), in comparing carabid communities in burned and clear cut jack
pine stands, found similar abundance and diversity of carabids in both stand types, but
different species composition. The presence of F. exsectoides is therefore the more likely
factor limiting carabid presence in young and mature stands, rather than habitat
suitability. Hister beetles prey primarily on organisms attracted to decomposing plant or
animal material (Borror, Triplehom, and Johnson 1989), and thus may have been

attracted to pitfall traps as good foraging locations.

70

Scavengers. Scavengers were especially likely to be attracted to pitfall traps by the odor
of decomposition. Significant changes in scavenger abundance across stand age classes
included a rise in numbers of Opiliones captured as stands aged and declines in Diptera
and Diplopoda from young to old stands (Table 13, Figure 17d). Some Opiliones are
predaceous. Moving Opiliones from the Scavenger to the Predator group creates a
stronger downward trend in scavenger abundance as stands age and a stronger increase in
predators in old stands. Scavengers, particularly those consuming decaying plant
material, may follow the same trend as Herbivores—a decline in ground vegetation
suggests a decline in rotting vegetation on the ground, and acidic conditions presented by
decaying jack pine needles may discourage detritovores. This may explain a decline in
Diplopoda, but explaining decline in the more diverse scavenging Diptera may require

detailed analyses focused on that group.

Correlations between feeding groups. Of correlations significant after correction for
multiple comparisons, several are explained by shared correlation to habitat: Opiliones
and predaceous Coleoptera are both found primarily in old habitats, and are both
negatively correlated to F. exsectoides abundance, which declined in old stands.
Orthoptera were found primarily in young habitats, and were negatively correlated to
Opiliones (Table 17). More interesting are the positive relationship between F.
exsectoides and Orthoptera, and the negative relationships between F. exsectoides and
predaceous Coleoptera and Opiliones. Although the correlations were weak and could be
explained by habitat as above, all three of these correlations were stronger than the

correlation between F. exsectoides and stand age.

71

F. exsectoides is an extremely generalist predator, feeding on anything that it can
catch (Dimmick 1951, Ayre 1970). F. exsectoides have been observed feeding on
Orthoptera, and I encountered Orthopteran fragments during nest excavations performed
for sex ratio studies (see Ch. 1 . 3). Where Orthoptera are abundant, they could provide an
abundant protein source for F exsectoides brood development. Both predatory beetles
and Opiliones may compete with F. exsectoides for living or dead protein resources.
Alternatively, this too may be a habitat-associated correlation, where F exsectoides and
predaceous Coleoptera/Opiliones have opposing responses to an unmeasured

environmental variable relating to stand age.

Conclusions. F exsectoides was numerically dominant in all age classes of jack pine,
but had a greater abundance of foragers and higher mound density in mature stands.
Smaller F exsectoides presence in young stands is probably due to the recent
establishment of these populations, while the lower abundance and density in old stands
is more likely a product of population decline. This suggests that F. exsectoides is
limited by the “life span” of its habitat. Populations show the “dense, thorough
occupation” of habitats described by Oster and Wilson (1978), but lower forager
abundance and mound density in young stands indicates that F exsectoides still requires
substantial time investment to capitalize on a newly-colonized habitat. While the
population decline seen in old stands may be exacerbated by competition from increasing
numbers of predaceous arthropods, it is difficult to separate this influence from the

impact of abiotic factors such as temperature and shading on brood development.

72

The ground arthropod community of jack pine changes significantly both in
proportions and composition of feeding groups. Many of these changes seem likely to be
driven by changes in ground vegetation as the trees mature and the canopy closes, while
others may have a less direct relationship, being influenced by age-correlated changes in
abundance of predators or competitors. The ant community exemplifies this pattern, as
F. exsectoides, F fizsca, and Aphaenogaster spp. responded significantly to stand aging.
C. herculeanis responded significantly to abundance of F exsectoides, and the diversity
of ants increased with the abundance of F fusca.

Despite the well-described aggressiveness of F exsectoides, this study provided
no strong evidence that F exsectoides excludes other ant species. Roberts and Gilliam
(1995) review the proposed role of disturbance in maintaining species diversity by
preventing competitive exclusion. If such an effect was present in this system, it might
be revealed by more thorough examination of young jack pine stands (both burned and
clear cut), before and after colonization by F exsectoides. Additional data on
colonization success, dispersal patterns, and habitat availability is necessary to
comprehend the constraints and adaptations arising from specialization on transient

environments.

73

CHAPTER 3

The impact of jack pine habitat succession on reproductive strategy of the Allegheny
mound ant, Formica exsectoides F orel

Introduction

The study of allocation of reproductive efforts is essential to understanding the
ecology and evolution of all organisms. Intrinsic to this study is the concept of the
“trade-off”: energy devoted to reproduction must be diverted from other tasks, such as
growth, defense, or storage (Cody 1966, Lande 1982, Charnov 1993). Successful
strategies are expected to possess features that will optimize allocation to reproduction
versus other life functions.

Evaluating this trade-off for sexually reproducing organisms is more complicated.
The cost of sexual reproduction (or “the cost of males” (Maynard Smith 1978, Williams
1975)) is assumed to be offset by an advantage gained through recombination, though the
nature of that advantage (and its existence) remains in debate (Hurst and Peck 1996,
Kondrashov 1993, Crow 1994). Organisms possessing the ability to reproduce by both
sexual and asexual means are expected to reproduce sexually only when there is a benefit
to doing so.

In addition to alternating between sexual and asexual strategies, reproductive
strategies can include adjustment of the sex ratio, or allocation to males versus females.
While it is intuitively obvious that the “cost of sex” may be minimized by producing

fewer males, Fisher’s (Fisher 1930, Trivers 1985) classic theory on the evolution of sex

74

ratios proposes an explanation for the existence of equal investment sex ratios in so many
diploid species. The basic argument is economic in nature—the supply of either sex will
vary with the cost of production and the “demand” (a measure of mating success,
dependant on availability of the opposite sex). When males and females are equal in
cost, sex ratio should approach an equilibrium value of equal investment, since over-
production of a rare sex will always be advantageous and overproduction of a common
sex will be detrimental. This theory has been elaborated to include ecological and
genetic factors leading to skewed sex ratios, such as Local Mate Competition (Hamilton
1967), Local Resource Competition (Clark 1978), differential mortality (usually of
males), and inbreeding (Thomhill 1993). The theory of kin selection (Hamilton 1963,
1964, Trivers and Hare 1976) has led to an explosion of sex allocation studies of the
eusocial Hymenoptera, particularly ants (for review see Nonacs 1986, Crozier and Pamilo
1996). The bulk of this study has been aimed at testing genetic relatedness hypotheses,
with far fewer studies focused on the role of sex ratio as an environmentally variable
reproductive strategy.

The genetic-relatedness hypothesis of sex ratio is based on the relatedness
asymmetry between workers, males, and queens that is created by the haplo-diploid
genetic structure of Hymenoptera (Figure 18). Elevated relatedness between female
siblings provides not only an explanation for the evolution of eusociality (workers gain
more by rearing their sisters than by producing their own offspring), but is expected to
lead to female-biased sex ratios in colonies where workers control sex allocation (Trivers

and Hare 1976). Elaboration of the genetic relatedness hypothesis to incorporate

75

 

 

0.25

 

 

 

 

Figure 18. Creation of relatedness asymmetry by haplo-diploidy, as described by Trivers
and Hare (1976).

polygyny, polyandry, and worker reproduction are reviewed in Bourke and Franks (1995)
and Crozier and Pamilo (1996).

Theories regarding environmental influences on sex ratio are primarily based on
resource-availability, although proposed mechanisms differ. The theory of local resource
competition (Clark 1978) has been invoked as a possible explanation for male-biased sex
ratios (Crozier and Pamilo 1996). In areas where resources are limited. female
production would be expected to decrease if female offspring will compete directly with
their mothers for resources. This theory is presented in a more general form by Cannings
and Cruz Drive (1975), who predict a bias toward the more dispersive sex. Deslippe and
Savolainen (1995) focused on food as the limited resource. but suggested a mechanism

based on nutritional physiology (Wheeler 1994) rather than competition. They observed

76

an increased allocation to female production in colonies of Formica podzolica that had
been offered increased food and protein resources, and hypothesized that female
allocation is often limited by colony nutrition (since underfed diploid offspring will
develop as workers, while haploid offspring will always develop as males). Elmes and
Wardlaw (1982a,b) hypothesize a relationship between female production and nest
insolation, but present no specific data. Environmental factors may also have indirect
effects on sex ratio, as Herbers (1993) describes evidence that ecological factors such as
habitat and food availability may encourage polygyny, thereby altering the relatedness
structure of colonies and potentially sex ratio.

Formica exsectoides F orel presents a particularly interesting challenge in the
study of reproductive strategies in ants. While many features of its biology make the
prospect of testing genetically-based models of sex ratio quite daunting (see Chapter 1:
Discussion), it presents an excellent opportunity for investigating environmental impacts
on reproductive strategy.

Reproduction in F. exsectoides can occur by colony fission/budding, or by
production of alate sexual forms (Creighton 1950). Mated females attempt to re-enter
established colonies, or found new colonies through temporary social parasitism of
Formica fusca colonies (Wheeler 1928). F exsectoides colonies are long-lived (Andrews
1929) and are not obligated to produce sexual forms for colony expansion within habitats.

F exsectoides inhabits transient, early successional forest habitats. Colonies
appear to decline as the forest ages (Andrews 1929, Chapter 2: Location and density of F
exsectoides mounds, Abundance of F exsectoides). Several authors have hypothesized

that insolation is an important factor (Cory and Haviland 1938; Wheeler 1928; Andrews

77

1926, 1929), but this has not been quantified. Numerous features of the habitat will be
expected to vary along successional gradients, including abiotic factors (such as light
availability) and biotic factors (such as composition of plant and animal communities).
This variation follows predictable patterns, allowing adaptive changes in life history
patterns to take place. Organisms inhabiting transient habitats should possess strategies
to quickly exploit optimal habitats, and to alter or escape suboptimal habitats. Since
colony growth may occur by sexual or asexual processes, and dispersal to new habitats is
dependent on winged sexual forms, both allocation to sexual forms and variations in sex
ratio represent potential strategic responses to environmental change.

The objective of this study is to determine whether differences in reproductive
(brood production) status of F exsectoides mounds and variation in sex ratio are
consistently related to habitat age. In addition to stand age, I examined light availability
which, while correlated to stand age, may be expected to vary within stands. Population-
level variables, such as mound density and forager abundance are examined as potential

links between changes in the forest environment and changes in reproductive allocation.

78

Methods

Site selection and description. Populations of F exsectoides were studied in stands of
jack pine (Pinus banksiana Lamb.) within the Huron National Forest, Michigan. Five
sites in each of 3 jack pine age categories were selected to represent a gradient of jack
pine succession. Further details on sites are presented in Chapter 2, with locations and

descriptions of sites listed in Table 9.

Environmental variables. One 2500 m2 plot was placed at random within each of the 15
study sites. These plots were mapped as a grid of 5m x 5m squares, using a compass,
marker flags, and surveyor’s rope. The location of all anthills was noted, even those that
appeared to be newly-formed buds. A comparative index of mound surface area was
calculated as longest slope x shortest slope in meters. Mounds were numbered and
marked with wire-stake flags. Location of all trees > 1 cm diameter and overall stand
density (cm dbh/m2 plot) was recorded, as described in Chapter 2. Pitfall traps were used
to measure relative abundance of F exsectoides foragers and other ground-dwelling

arthropods; trap placement and collection regimes are presented in Chapter 2.

Light availability. Light availability was measured using hemispherical photography. A
Nikon camera mounted with an 8mm Sigma fish-eye lens was placed above each anthill
and oriented north. Slides were taken above each anthill sampled for sex ratio, five
additional anthills that were not sampled for sex ratio (if available), and above each
pitfall trap. Pitfall traps were placed in fixed positions with respect to plot boundaries,

therefore they were placed randomly with respect to anthills and provide a null

79

expectation or random measurement of stand light availability. Slides were digitized and
analyzed using the program GLI/C (Canham 1995), with defined latitude of 44.55°N and

a growing season from 90 to 304 Julian days.

Mound reproductive status. Reproductive allocation was examined at the pupal stage,
when the sex of individuals is identifiable by morphology but confusion caused by
dispersal or loss of individuals to predation is not yet a factor. Whether mounds are
producing sexual forms is also best determined at this stage, as pupation occurs in the
upper portions of the mound. Sexual forms in the egg or larvae stage may never reach
maturity, so observing these stages might provide a misleading picture of mound
reproductive output. Mounds outside plot boundaries were periodically examined to
check for pupation of reproductive forms, avoiding disturbance of study mounds before
sampling. After pupation had occurred, all mounds within a plot were partially excavated
to record the presence or absence of reproductive caste brood. Reproductive brood was
considered absent if excavation of the mound to approximately 0.5m below the soil
surface and sifting of the mound material through a large strainer revealed no
reproductive caste pupae, prepupae, larvae, or alate adults. Mounds containing neither
sexual per worker brood were classified as “dead” even if adult workers were present on
the mound surface. Mounds were excavated in an order based on computer-generated
random number series to avoid biasing data by preferentially selecting large mounds for

excavation.

8O

Sex ratios. Samples of 30 reproductive-caste pupae were collected from the first 5
sexual-producing mounds excavated within each plot. Pupae were individually weighed
and stored in a 10°C incubator for 10 days. This allowed development to continue,
avoiding weight differences due to stage of development and problems with inability to
determine sex of “undeveloped” pupae (see Chapter 1). The heaviest 5% of each sample
was removed and frozen at -80°C for genetic analyses (see Appendix B: Attempted
search for diploid males). These pupae were later returned to their original samples.
Pupal cases were removed and the sex of each pupa recorded. Pupae were dried for 3
days at 70°C and re-weighed. A nested analysis of variance was used to analyze
differences in pupal dry weights between sexes and among sites. Sex ratios were
calculated using Boomsma’s (1989) Energetic Cost Ratio (Bourke and Franks 1995), and

sex ratios are expressed as proportion of male allocation.

Statistical analyses. An initial correlation analysis was performed to investigate the
relationship between environmental variables. Mound density (mounds/m2), stand
density (cm dbh/m2 plot), stand age (y), and light availability (% open sky) were analyzed
using PROC CORR (SAS Systems 1990).

The percent of all living mounds within each stand that produced sexual caste
pupae was correlated with the above variables. When significant correlation occurred
between the proportion of sexual-producing mounds and two or more correlated
variables, only comparisons with highest explanatory value are considered for discussion.

As ‘young’ stands were of heterogeneous origin (3 clear-cut, 2 burned (see Table

9)), analyses involving stand age were also performed using date since last cut (rather

81

than date since major disturbance, i.e. cut or burn) for values of stand age. As F
exsectoides populations may be reduced or destroyed by clear-cutting but are in high
abundance after fires (see Chapter 2 Discussion: Location and density of F exsectoides
mounds), date since last cut provides a more reasonable estimate of the age of the ant
population, while date since major disturbance provides a better estimate of patch age.

Analysis of variance was used to analyze differences between light environments
within stands. Specific comparisons were made between mounds producing sexual forms
and mounds producing only workers, mounds producing brood and dead mounds, and
mounds producing brood and random forest.

A nested analysis of variance was used to separately analyze differences in
weights of each sex between sites and between mounds within sites. Binomial
probability curves were generated to provide confidence curves for sex ratio estimates.

As comparisons were of greater interest than estimating actual values of sex ratio
and mound-level sex ratio data were distributionally-challenged, a logistic regression
(PROC CATMOD, SAS Systems) of raw count data was used to analyze the effect of
mound-specific variables on sex ratio (Juliano 1993). The variables used in these
analyses were light availability (% open sky) and mound surface area indices.

For analyses involving site-level variables, site averages for sex ratio were
analyzed using simple regression. These analyses used stand age, % open sky, mound
density, and F exsectoides forager abundance as predictor variables. Additionally, sex
ratio was regressed on abundance of Formica fusca L. As F fusca is the host species for
nest founding by F exsectoides queens, female investment would be expected to increase

in areas where F. fusca (and therefore nest sites) are abundant.

82

Results

Environmental and mound variables. Age—class means and standard deviations for
mound density and tree density are presented in Chapter 2, Table 10. Age class means
and standard deviations for abundance of ants are presented in Chapter 2, Table 15.

Light availability ranged from 11.8 to 82.4 percent open sky. While correlated
with stand age, light environment was highly variable within stands (R2 = 0.4834, p =
0.0001; Figure 19a).

The results of correlation analyses including mound density, stand density, stand
age, light availability, and forager abundance values are presented in Table 18. Percent
open sky is significantly and negatively correlated to both stand age and stand density.
Relative abundance of F exsectoides foragers was significantly correlated to mound
density alone. Marginal correlations were present between mound surface area indices

and stand age/open sky; all other correlations were not significant.

Mound reproductive status. Collections of reproductive caste pupae took place from
July 19 to July 24, 1997. Of 241 mounds examined, 93 (38.5%) were dead and
reproductive caste pupae were found in only 91 (38.1%). A few mounds in older stands
(O4 and 01) contained only reproductive caste larvae at the time of initial excavation.
Table 19 provides a summary of mound status for each site and each stand age class.
The percent of mounds examined that were classified as ‘dead’ increased
significantly with stand age (p = 0.0006, Table 20), as did the proportion of mounds

producing sexual forms (p = 0.0107. Table 21; Figure20). The percentage of dead

83

 

    
     

 

 

 

90
g R2 = 0.4834
g 60 - °
0
2 o e
0 8
3' ° 3 : 6
°\o 30 - 3 2
3 0 ’ s
o " G ‘:'
Go
a
0 I I I I I I I
0 1o 20 30 4o 50 60 7o 80
stand age (y)
90

 

 

absent

dead

workers only
reproductive

— - - -Linear (absent)
Linear (dead)

Linear (workers only)
----- Linear (reproductive)

OIX>

% open sky

 

 

 

 

 

 

 

 

 

60 80

stand age (y)

Figure 19. a) While correlated with stand age, light environment was highly variable
within stands. b) Status groups differed in their relationship to the range of light
available

84

 

 

 

 

 

 

 

 

 

A - V A + V A + V A - V A - V
8 I. 8 8 8 88858.. 88.8
A - V A - V A + V A + V
8 ... 8 .. .86 8.88
A + V A - V A - V
8 .. 8 .38vo .6888
A - V A - V
.12.. ..:.:.. 5? 825$.
A + V
.1. .88vo «.88

 

 

 

88m 8888 .888 .888 .3? 88 .x. .8808 88% «we 88%

88808
820QO 8. HO Z .8868 88 8885 0388888 o. 888 :86 8882.. 888.88%
323 8%: w. 8830.88 .8 .8885 28888.. on 2 88288 «83 AmAVAV

v QV 1.. 8 AAAVAV v QV 2...... @888 8888800 .8onch 88388 on :8 Good v 8V
.1... @888 80:22.80 .88 888888 28:38 88. 83888 8%.. 82.82888“.
88288 828888 .m 88 8888838 8038.. 8:38.80 .3 «EPA.

85

Table 19. Summary of mound status for each site and age class. Totals represent

all mounds present in a 2500 m2 plot. "Active" mounds contained brood, either a
mixture of reproductive caste and worker brood or worker brood only. "Dead"
mounds showed little or no ant activity and contained no brood.

 

 

 

 

 

 

 

mounds mapped active mounds dead mounds
reproductive workers only

young 59 26 26 7
Y] 14 6 7 1
Y2 l7 7 7 3
Y3 11 4 6 1
Y4 9 5 2 2
Y5 8 4 4 O

mature 101 34 25 42
M1 20 5 ll 4
M2 18 6 5 7
M3 13 9 2 2
M4 20 7 6 7
M5 30 7 1 22

old 81 31 6 44
Ol 12 5 1 6
02 26 6 4 16
O3 15 6 1 8
O4 15 10 O 5
05 13 4 O 9

86

Table 20. ANOVA test of significance for the regression model: % dead mounds =
stand age, when “stand age” = years since last major disturbance.

 

 

Source of variation df mean square F p
stand age 1 4621 19.88 0.0006
error 13 232

total 14

ANOVA test of significance for the regression model: °/o dead mounds = stand age,
when “stand age” = years since last clear cut.

 

 

Source of variation df mean square F p
stand age (cut) 1 2539 6.47 0.0245
error 13 393

total 14

87

100
901
801
70-
60-
50-
404 ”a

30. O
20-
10-

 

 

R2 = 0.4163
p = 0.0107

% mounds producing sexuals

 

 

 

0 I I I I I I I

0 10 20 30 40 50 60 70 80
stand age (years since disturbance)
'0 young mature . old I

 

 

100
90. b
80-
70-
60-
50-
40. O /" O
30- O

20-
10-

 

 

R2 = 0.1904
p>005

 

 

% mounds producing sexuals

 

O r I I I I I I

0 1O 20 30 40 5O 60 7O 80

stand age (years since cut)

Figure 20. a) The percentage of mounds producing sexual caste brood increased with
stand age, when "stand age" = date since major disturbance. b) This relationship was
no longer significant when "years since clear-cut" is used as stand age. Arrows
indicate burned stands.

88

Table 21. AN OVA test ofsignificance for the regression model: "/0 of mounds
producing sexual forms = stand age, when “stand age” = years since last major

disturbance.

 

 

Source of variation df mean square F p
stand age 1 2772 8.86 0.0107
error 13 313

total 14

89

mounds was, however, more strongly related to light availability (p = 0.0004, Figure21).
The percent of mounds producing sexual forms was dependent on mound density in
mature stands, but relatively independent of mound density in both young and old stands
(Figure22 a, b).

Percent of mounds producing sexual forms was correlated to stand age, stand
density, and light availability (Table 22), but stand age explained the highest proportion
of variance.

Light environments differed significantly with mound status (Table 23, F igure23).
There was no significant difference between light environment in locations where
mounds were absent and locations where mounds were dead (p = 0.0672), but light
environments of inhabited mounds had significantly greater percentages of open sky than
either absent or dead (0.0001). Additionally, mounds containing worker brood only were
found in locations with significantly greater percentages of open sky than mounds

producing sexual forms (p = 0.0037).

Sex ratio. Pupal dry weights differed between sexes and between stand age classes
(Table 24, p = 0.0001) (Figure 24). Dry weights of males and females were highly
correlated to each other within stands. Dry weight was not related to sex ratio (Figure
25).

Sex ratios were male-biased at all sites, but sites differed in the composition of sex ratio
(F igure26). Sites Y2, M3, M4, M5, and 02 display patterns indicative of “split” sex ratio
(Boomsma and Grafen 1991; Godfray and Werren 1996), while other sites display a less-

severe but more consistent male bias. While sites Y2 and M3 had sex ratios not

90

 

80
70 -
60 -
50 .
40 -
30 ' R2 = 0.6278
20 -
10 -

% dead mounds

p = 0.0004

 

 

 

 

80

mean % open sky

 

'0 young mature . old I

 

Figure 21. The percentage of mounds classified as "dead" (containing no brood,
inactive) increased significantly as open sky decreased. Values of open sky are plot
mean values.

91

 

 

 

 

 

 

 

 

 

 

 

T. 100 O t
3
5‘:
a) 80-
U!
.E
5 m 60'
9 E
0.3 40..
(0
'5
§ 20
o I
E a
as o . . T
0 5 10 15 20
total mounds in plot
a:
.2
g 80-
X
G)
U)
E 60-
U
1’:
o 404
L-
O.
3 20
§ . b
o
E 0 t I I
0
°\ 0 5 10 15 20

total mounds in plot

0 young 9 mature 0 old
------ Linear (young) -- -Linear (mature) —Linear (old)

 

 

Figure 22. a) Sites with more mounds contained lower proportions of sexual-
producing mounds. b) Plotting the interaction of this relationship with stand age
reveals that this relationship is strongest in mature stands, while old and young stands
are likely to produce sexual brood in a constant proportion of mounds.

92

 

 

 

 

 

 

 

 

 

A - V A - V A - V A + V A + V
m: * * * * *
A-V A-V A-V A+V A+V
m: * as; .13., .13..
8:35.? bwfiow 56:26 2:68 .9? some @8226 €me own 98%

com: a :otflotoQ V0 .8885 ..._m:_meE= on 9 382.88 263 God v 3

$558 0263on8 X.

€558 Row o\o

.fiwcobm 6onon 5V? coca—880 05 8865 @303
wonmzm .AM: 033. oomV 5? some Axe tam .bmmcov 29m .313 23% £231?» 338.50 mph: 05 .8 ”A 088$:me 323

Cw.

5 «v8.88 mucus—250 4:85:me

35238 on :8 290 v QV :1: Ba AAAVAVAV v 3 5w * *. @812: mcosfloboo Eco .2892”an 2&st 8V :ouootoo
Sada .mofimcg £63??me USN €556 963699838 can 3% V0 mowficoeom c8389 5:220:00 .NN «Sun.

93

Sod v a a ado—mama Boa moocoéfic =< .Eobo ESESw 93 SSE moan—03mm“: Bu.

REESE 323/ .wccsoE wEosuoaéoob .oéow 5 633m 038 6383689: Vo 8:305 Ho 853m 8865
..o>:o=uoa2._ 28 12:0 Eco—53.. 6003 o: mEosvoa Ho 3:033.“ 263 $5 €558 5V bzfiazgm

Em: V0 85:; mucosa»: E89. .2583 803 3:208 832838 .k o: 823 @2582: 3m como V0 8315
mucomoaoe .2527}. ...m=§m= manage “0 2:59: w=o§ VAssmowVEwAm @2850 55238.8 Ems .3 “95m:—

wfluwuw UCSOE

twat 9303359. 2:0 99:95 “coma—w

 

NV":
now

.ON

.om

 

.ov

 

.om

 

 

 

U
l--t

.oo

 

 

 

on

Ms uado % ueaw

94

Table 23. ANOVA test of H0: Light environment (% open sky) is equal for all mound
“status” groups. Status groups were: absent = no mound present at measurement
location; dead = inactive, not producing brood; workers only = producing only worker
brood; reproductive = producing both sexual caste and worker brood.

 

 

Source of variation df mean square F p
status 3 4345.5 17.3 0.0001
error 228 25 1 .2000

total 231

95

Table 24. ANOVA test of hypotheses Ho: Pupal dry weight is equal for males and-
females and H0: Pupal dry weight does not differ among stand age classes.

 

 

Source of variation df mean square F p
model 3 0.00013049 1 15.01 0.0001
age 2 0.0000895] 78.89 0.0001
sex I 0.00022857 201.45 0.0001
error 1735 0.000001 13

total 1738

96

.VN Bank 5 @8535 2m 33“ oofiouEEm

.muobo Emonflm 9a 2on 8:33-68. 08 3:23on 835/ .monfim 282: Mo masoVA E :2:
€55 20 E 333: 083 85% Atom .338 A85 538: 39:85:me 203 why—Scum .vm “Earn

E

 

. Fooo
. wood
1 wood
.. voo.o
. mooo
.. oooo
. nooo
wood

 

 

 

 

 

97

(5) tufiiaM Mp

0.009

 

 

 

 

 

 

 

 

 

 

0.008 .
3 0.007 -
E .- -. """ O ' ‘ ’
o 0.006 - G
.5 E." [I
g 0.005 - I’Efi
3 0.004 -
m
3
Q 0.003 -

0.002 -

0.001 . IO young mature . old

0 I I I I
0.5 0.6 0.7 0.8 0.9 1
sex ratio (males/total)
I male 0 female
Linear (male) ------ Linear (female)

 

 

Figure 25. Pupal dry weights of sexual forms were not related to sex ratio.
Regression slopes were not significantly different from zero for either sex. Each pair

of male (Cl) and female (0) weights represents one stand.

98

AR 2%: 38.0 n

a a £3me 20 can .239: .Am> 45 35:3 55 common 0E8 208 bagucmcwfi 803 Am> can
6% .36 mocfim So .520 .mofifi 8 conga—Va 35358 >on “838$ 9 5382?. $58.52
3E3 .252: 28 $5332 Snow a 22:3 met some 65% 28 355298 Snow 53

.26qu 0:3 xofl 2 E 803238 332930 E .8 0:8 5m .3 «Law:—

99

clearcut

   

@111

 

 

 

 

  

 

100

Figure 26. Sex ratio estimates for F. exsectoides populations in 15 jack pine stands.

0.25

 

 

0.23 - young
021 d mature — - —

old —

 

 

0.19-

 

0.17-
0.15-

pr (ratio = x)

0.13 -
0.11 -
0.09 -
0.07 -

 

 

 

0.05 . , I ~~ .
0 0.2 0.4 0.6 0.8 1

 

male allocation ratio

Figure 27. Confidence intervals for numerical sex ratio estimates, generated with
binomial probability functions.

10]

significantly different from 50-50 investment, and all other sites were male-biased, there
were no significant differences in sex ratio among stand age classes (Figure 27). Sites
Y3, Y4, and Y5 (three young clear-cut stands) were significantly more male-biased than
any other stands (Table 25; Figure 26).

Logistic regression of raw counts of males and females on percent open sky for
each mound yielded a cubic equation. Maximum likelihood tests indicated a probability
of 0.0001 that the model produced an adequate description of the observed sex ratio data
(F igure28). A similar analysis using mound surface area indices as the predictor variable
yielded similar results. Analysis at the quadratic level yielded significant non-zero
parameter estimates, but values generated by the equation did not fit the data (p = 0.0001,
F i gure29).

Sex ratio was not significantly related to mound density. Sex ratio was
marginally related to abundance of F. exsectoides foragers (Figure 30a,b; Table 26), and
also to relative abundance of F. fusca (Table 27, F igure3 1a). As the correlation with F.
fusca appears to be based on two outlying data points, both clear-cut stands, and
disappears when those stands are removed from the analysis (F igure3 1b). Sex ratio was
not significantly related to stand age when stand age was determined by date since major
disturbance (Table 32a, Figure 30a), but was related to stand ages determined by date

since last cut (Table 28b, Figure 32b).

102

Table 25. ANOVA test of 110: Sex ratio of young clear cut sites is equal to the sex ratio
of all other sites.

 

 

Source of variation df mean square F p
stand 1 0.2914 6.08 0.0162
error 70 0.0480

total 71

103

.5 Stop a 8 Bow Eco Emaotmo 3:35:me 2225.8:

023:: 9 :25 328:8 VamEoFAAo: oonoofifisoo Ba .oBmooégoom 53:3 .Aoooo

u a an 38%: 33 San coinage 05 338.0.vo i833?“ E58 $5 8:: w:::% 32:an

:s: 05. .AAoooo n :V Sou Bot “5.6%: >_H::owm:w_m 203 £80883 z: .802 ..u:_ £8893
m<m .QOSEuQU DOMEV 228:3 @0053: :58:me mi? 33:58 203 330883

AAmfimesoood + N333: 5.3 + Axmtoammd + ~52: one -_ .1. 292035 2a.. 5m
”mm? 329 @2068: Bacofim 8 3m: scams? 2:.

5:528;
Em: :o :83 225:: :23“:on 0532 9 E So: a 825 moon: x% 83080 .3 2:»:—

104

om

bzfimzmé Em: no woman £088 :23“:on ”Eamo— 9 5 Son 5 8m moufi 5% 82030 .wN PEME

 

_ vmuofimal 82390 0 _

 

a.» :30 a}
om on om om ov om om or o

I I I I I I I I

 

 

. md

 

 

 

 

 

(lama/99mm) one: xas

105

.5 Stop a .8 88. Got Embobww
338$:me 380888 3:68 8 36¢ 328:3 380828 oohwovéhfi 28 50:5 .5006
n a E 38o? 83 8% 33030 05 Uoptomuc 32.3on ESE £5 at wcsflm £85093

:3: 05. .2890 n 3 83 89a Eastmw £2856me 803 $80883 :m .803 :2: ,mEBmxm
m<m .QOEH<U 00%: 8083 c0258.: E:E_me mam: 338:8 053 $80833

.8065 88
Sufism no 393 £288 .568“on 2632 8 E Sou a 803 8:8 xom @3530 5N flaw:—

106

mm

3065 $8 8&2; co woman £258 :o_mm2w8 ucm_wo_ 9 5 Son a 8a .858 5% BiomnO .aN mgwwm

_ umfifimfil 82630 0 _

 

 

xqu no; moat—5 950:.

ON m_, or m o

 

 

(mouseuaw) one: xas

 

 

 

107

 

 

 

 

 

 

 

 

 

 

 

 

A a
:3
2
7a
2
N
E. 0.5 -
.2
E
3 R2 = 0.1136
"’ p = 0.0929

0 I I I I I I

O 1 2 3 4 5 6 7
log (total foragers/plot)
I 0 young mature - old I

1
A b
a
33
7o
2
E; 0.5 -
.2
E
a: R2 = 0.3257
0)

0 I I *I I I I

0 1 2 3 4 5 6 7

log (total foragers/plot)

Figure 30. a) Sex ratio was not significantly related to abundance of F. exsectoides
foragers captured in pitfall traps. b) This relarionship became significant when sex
ratio outliers (young clear cut stands) are removed from the analysis.

108

Table 26. ANOVA test of significance for the regression model: sex ratio = log-
transformed abundance of F. exsectoides foragers captured in pitfall traps.

 

 

Source of variation df mean square F p
forager abundance 1 0.0263 3.29 0.0929
error 13 0.0080

total 14

l09

 

 

 

 

 

 

 

 

 

 

 

 

B
E
.2
‘5
§ 0.7 J 0
g Q R2 = 0.2842
E 0.6 - p = 0.0464

0.5 . , . . .

0 2 4 6 8 10 12
mean fusca captured per trap
I 0 young mature . old
1.0 -
b U
,2 0.9 .
m
E A
E o 3 . __ O O
2 _

5 O R - 0.0351
‘5’ 0 6 - p = 0.8510

0.5 t 1 .

0 1 2 3 4

mean fusca captured per trap

Figure 31. a) Sex ratio appeared to have a marginally significant negative
relationship with relative abundance of F. fusca. b) This relationship disappeared
when 2 young clear cut stands (outliers for both fusca abundance and sex ratio) are
removed from the analysis. Values presented are mean F. fusca abundances for each
stand, and total male allocation ratios for each stand. Significance tests for regression
analyses are presented in Table 27a,b.

110

Table 27. ANOVA test ofsignificance for the regression model: sex ratio = abundance
of F. jz'zsca. a) all sites b) excluding clear cut sites

 

 

 

 

a) Source of variation df mean square F p
fusca abundance 1 0.0354 4.84 0.0464
error 13 0.0073
total 14

b) Source of variation df mean square F p
fitsca abundance 1 0.00023 0.04 0.851

error 10 0.00606

total I 1

III

 

 

 

 

 

 

1 -fl
0
\ O
S 9 .
g \29 0
m
2
a
E 0.5 d
.2
g R2 = 0.0937
5‘»
m p>005
a
O l I I
0 20 40 60 80
stand age (y)
l 0 young mature - old I

 

 

05-

R2 = 04535
p = 0.0266

sex ratio (females/total)

 

 

 

0 20 40 60 80

stand age (years since cut)

Figure 32. a) Sex ratio was not significantly related to stand age when "stand age" =
years since major disturbance. b) Sex ratio was significantly related to stand age when
"stand age" = years since clear cut. Curves in both a) and b) were fit using second-
degree polynomial regression equations.

112

Table 28. a) ANOVA test of significance for the regression model: sex ratio = age +
agez, when “age” = years since last major disturbance of stand. b) ANOVA test of
significance for the regression model: sex ratio = age + agez, when “age” = years since
last clear cut of stand.

 

 

 

 

Source of variation df mean square F p

model 2 0.0079 0.83 0.4593
age 1 0.00960 1.01 0.3355
agez 1 0.00579 0.61 0.4509

error 12 0.00954

total 14

Source of variation df mean square F p

model 2 0.0295 4.98 0.0266
age (cut) 1 0.0435 7.34 0.0190
age (cut)2 1 0.0312 5.26 0.0407

error 12 0.0059

total 14

113

Discussion

Environmental and mound variables. The biological concept of succession focuses on
a progression of general changes that will take place as a community matures (Horn
1974). As forests grow and the canopy closes, light availability on the forest floor will
decrease, causing composition changes in understory plant communities and related shifts
in herbivore and predator communities and abundance. The responses of single species
are more difficult to predict, especially given the amount of variability associated with
any biological process, and the increasing likelihood of stochastic mischief with the
passage of time. This is exemplified by the high variation in light environment within
stands—older stands are especially vulnerable to mortality from insect attack, storm
damage, or 01d age, so older stands possess greater variation in light environment (Figure
19).

While relative forager abundance was highest in mature stands, suggesting that
mature stands represent a peak of F. exsectoides populations, this pattern was not
significant. In fact, forager abundance was correlated significantly only to mound density
(see Chapter 2). The stand variables in this study appear to have only an indirect effect
on forager abundance, through their relationship with mound density (Table 18). Cory
and Haviland (193 8) failed to find a relationship between sunlight and ant activity,

although this negative result may reflect a lack of quantitative data.

114

Mound status. The increase in the proportion of dead mounds in older stands is
consistent with the observations of earlier authors that F. exsectoides populations occur in
openings or forest edges, where sunlight is available (Haviland 1948). Andrews (1926)
notes that mounds of F. exsectoides “arise and pass away in rhythms harmonious with
phases of forestation”, and offers two explanations for the apparent decline of F.
exsectoides populations in mature forest. The first is that reduced light availability
creates conditions unsuitable for brood production. The second is based on the argument
that older trees will provide less food. Andrews speculates that young trees will sustain
more aphids that will yield greater amounts of honeydew than old trees, and that ants will
have further to travel to obtain the honeydew in old stands. While homopteran honeydew
provides valuable carbohydrates to ant populations, it is not essential for ant survival
(Bradley and Hinks 1968).

In a later study (Andrews 1929), Andrews examines the relationship of sunlight to
mound temperatures. He concludes that mounds are heated by sunlight (rather than
decay or ant activity), and that mound structure helps to retain heat. Andrews observed
ants moving brood from areas of lower temperature to areas of higher temperature within
the mound. These observations, combined with data from this study, lend support to
Andrews’ first explanation of F exsectoides decline in old stands—prepupal stages of
reproductive forms were found in shaded mounds in old stands (01 and 04) as late as 22
July, when brood in mounds at stand edges or in young stands had already pupated.
Additionally, old stands had higher pupal dry weights (Figure 24), which were explained
by the inferior stage of pupal development in these samples, rather than greater

investment per sexual (see Chapter 1: Weight and Developmental Stage). If lower light

115

availability leads to delayed brood development, mounds in old stands may have
insufficient time to complete their yearly cycle of brood production, and colonies would
be expected to move to a more favorable location or become extinct.

The pattern displayed by mounds with respect to light availability (Figure 23)
suggests that F. exsectoides prefers to construct mounds in areas with high insolation, and
will expand the colony through budding while light conditions remain favorable. In
studies of the related species Formica ulkei Emery, Scherba ( 1958) observed that new
buds were formed primarily in non-shaded areas. In habitats with few suitable locations
for budding, production of reproductive brood may be an adaptive response that will
allow dispersal to younger habitats.

Elmes and Wardlaw (1982a,b), in population studies of Myrmica sabuleti and
Myrmica scabri, stated that colonies producing sexual forms contain more workers than
colonies that do not. They further cite the assertion of Brian and Brian (1951) that
colonies of Myrmica ruginodis receiving high levels of insolation had higher worker
numbers and larger worker size. Pamilo (1991) and Herbers and Stuart (1998) present
the possibility that queen-worker conflict may occur over allocation to growth versus
reproduction as well as sex ratio. If workers, having a shorter life expectancy, prefer
greater allocation to reproduction, colonies with a larger workforce may be more likely to
produce sexual forms.

Production of reproductive brood may also be a function of colony maturity. Old
stands may contain higher proportions of mounds producing sexual forms because there
are fewer young mounds in these stands. If budding represents an intermediate life

history stage between ergonomic growth (worker production) and reproduction by sexual

116

forms, the switch from mound fission to dispersal of sexual forms may represent age-
specific reproductive response, rather than an adaptive response to the colony
environment. This explanation seems less likely, since re-analysis of the data with
burned stands aged by years from last cut (rather than years since fire event) yielded an
inferior fit to the data (Figure 20a,b).

If colonies are polydomous and a fixed number of mounds in a population will
produce sexual brood, the proportion of mounds producing reproductive forms might also
be expected to increase in old stands, because there are fewer mounds. It is intriguing to
observe that the relationship between the percentage of mounds producing sexual forms
and the total number of mounds per plot is not consistent among stand ages (Figure 22).
In both old and young stands, the proportion of mounds producing sexual brood declines
only slightly with increasing mound density, while the relationship is more dramatic (and
statistically significant) in mature stands. This implies that populations in young and old
stands will increase output of reproductive forms as mound density increases, while a
fixed number of mounds in mature stands will produce sexual brood without regard to

increases in mound density.

Sex ratio. Sex ratio appeared more affected by factors intrinsic to the colony, such as
age, size, and forager abundance, than by environmental influences. The extreme male
bias of populations in clear cut stands may relate to resource shortages, although there is
no indication that clear cut stands contain a lower abundance of invertebrate prey

(Chapter 2). As previous populations are reduced or destroyed by cutting and replanting

117

of stands, populations in clear-cut stands are more likely to represent recent colonization
by F. exsectoides. Young colonies may specialize in the production of males.

The lack of a detectable influence of light environment on sex ratio may be due to
a lack of response range—if there is a range of light environments within a given stand in
which F. exsectoides mounds produce sexual forms (versus producing workers or moving
out), this range might not be wide enough to allow a gradient of response in sex ratio. Or
it may be that light environment is unimportant. Experimental manipulation of light
environments would be useful, but logistical difficulties intervene.

Elmes and Wardlaw (1982a,b) draw the conclusion that, while production of
males is dependent on the number of workers in the colony, production of females is
dependent on the level of nest insolation. While the premise that separate factors may
regulate allocation to each sex is intriguing, the fact that the authors have also linked
colony size to level of insolation casts doubt on the independence of these factors. Elmes
and Wardlaw were able to collect entire colonies to account for total production of each
sex. Similar data for F. exsectoides colonies might suggest the presence (or absence) of
such patterns, but collection of even a single colony of F. exsectoides in its entirety
would require more resources than were available for this study.

It does appear, when age of population is taken into account (Figure 32b), that F.
exsectoides increase allocation to females in mature stands. Mature stands present an
environment where sexual production is less linked to mound density, indicating that
production of sexual forms is not as strongly favored as in old stands. Since mature
stands also contain the greatest abundance of foragers, more resources may be available

per sexual form produced, encouraging increased investment in females. This

118

explanation would be supported by higher pupal weights in mature stands, but this was
not observed (Figure 24). Mature stands could represent an earlier phase of dispersal
from the changing habitat (when the lifeboats are being provisioned and filled in an
orderly fashion), while the increase in male allocation in old stands might represent a
mass exodus (when the passengers are jumping onto anything that might float). Tests of
this possibility would consider changes in overall output of sexual forms as a function of
sex ratio.

Changes in sex ratio related to population (colony) age may reflect an adaptive
life history progression. Oster and Wilson (1978) define the main stages of ant life
history as foundation, ergonomic growth (production of only worker-caste brood).
reproduction, and orphanage. Young colonies leaving the ergonomic stage may begin the
reproductive stage by producing male-biased sex ratios, since males are usually the
“cheaper” sex and tend to be overproduced by small colonies (Crozier and Pamilo 1996).
Mature colonies may increase production of females at the “peak” of the reproductive
stage, but may shift back toward male production as the colony senesces or habitat
conditions become less favorable.

It is impossible to separate direct and indirect, genetic and environmental sources
of the male bias found in F. exsectoides populations without detailed knowledge of
population genetic structure. Given current knowledge of F. exsectoides biology, it is,
however, possible to discuss some of the more likely explanations, and suggest direction
(and caveat!) to future research.

Under the genetic relatedness hypothesis, increased male bias is expected under

conditions of both polygyny and polyandry (both conditions which will decrease

119

relatedness between female siblings) (N onacs 1986). In inbred populations, however,
multiple-mating and polygyny will be unlikely to decrease relatedness between sisters to
the same extent as multiple mating to unrelated males or polygyny with unrelated queens.
Further analysis of F. exsectoides population structure using hypervariable genetic
markers is necessary, as allozyme studies indicate a lack of variation at many loci
(Nipson 1971, Appendix B). This lack of variation may be indicative of high levels of
inbreeding and high intercolony relatedness.

Local Mate Competition is popularly invoked as an explanation for female-biased
sex ratios (N onacs 1986, Crozier and Pamilo 1996), but it may as easily be invoked to
explain male bias in F. exsectoides. If multiple mating is the norm for F. exsectoides
queens, and males are the more dispersive sex (as is indicated by observed mating
behavior, Bishop 1998), it is reasonable to expect a male bias in sex ratio. Each female
produced will require several males to mate, and males will be more vulnerable to
predation and random mortality factors during their dispersal. Given the lack of apparent
advantage, for either workers or queens, to producing millions of males of which only a
small fraction will mate successfully, female competition for sperm seems quite likely to
be an important factor in this system.

While polygyny, polyandry, polydomy, poor survival in lab colonies, and unusual
modes of colony founding and expansion may sound like a recipe for disaster in studies
of reproductive allocation, these very traits and the apparent success of F. exsectoides as
a eusocial species makes it important to understand how that success is shaped by
reproductive strategies. Life history strategies of ants are, to date, poorly understood.

Attempts to examine life history variation within species have rarely included age-

120

specific variation within stages, although F. exsectoides variation in sex ratio within the
reproductive stage indicates that this variation exists and must be taken into account in
sex ratio studies. While sex ratio was not shown to respond to influences external to the
colony, differential investment in sexual reproduction in response to environmental
factors supports the possibility that such interactions may occur, and merit investigation

in other species.

121

APPENDICES

122

APPENDIX A

123

APPENDIX A

Arthropods collected in jack pine (Pinus banksiana Lamb.) by pitfall trapping,

July 7—August 4, 1997.

CLASS INSECTA
Collembola
Orthoptera: Acrididae (Cyrtacanthacridinae, Oedipodinae)

Blattaria

Hemiptera:

Homoptera:

Neuroptera:

Coleoptera:

Tettigoniidae
Gryllacrididae (Rhaphidophorinae)
Gryllidae

Lygaeidae

Pentatomidae

Thyreocoridae (Galgupha spp.)
Miridae

Tingidae

Aphididae
Membracidae

Cercopidae
Cicadellidae

Myrmeleontidae

Carabidae
Notiophilus aenus (Hbst.)
Calosomafrigidum Kirby
Myas cyanescens Dej.
Pterostichus pennsylvanicus LeC.
Calathus gregarius (Say)
Synuchus impunctatus (Say)
Syntomus americanus (Dej.)

Staphylinidae

Histeridae

Elateridae

Nitidulidae

Tenebrionidae

Mordellidae

124

Diptera:

Lepidoptera

Hymenoptera:

ARACHNI DA
Araneae

Acari
Opiliones
DIPLOPODA
CHILOPODA

ISOPODA

Bruchidae
Curculionidae
Scarabaeidae

Mycetophilidae
Cecidomyiidae
Culicidae
Simuliidae
Therevidae
Dolichopodidae
Bombyliidae
Syrphidae
Pipunculidae
Sepsidae
Lauxaniidae
Muscoidea

Braconidae

Ichneumonidae

Chrysididae

Apidae

Mutillidae

Pompilidae

F ormicidae
Formica exsectoides F orel
Formica firsca Linn.
Aphaenogaster spp.
Dolichoderus plagiatus
Camponotus herculeanis
Tapinoma sessile
Monomorium minimum
Lasius spp.

125

APPENDIX B

126

APPENDIX B

Attempted search for diploid males (or Things That Didn’t Work)

Introduction

Inbreeding in ant populations is known to lead to the production of diploid males—
although diploid eggs are generally female, homozygosity at sex determining loci can
cause the egg to develop into an infertile male (Crozier 1971, Pamilo et. al. 1994).
Founder events and habitat isolation increase the likelihood that queens and males will
possess sufficient genetic similarity to cause a significant number of diploid males to be
produced by the colony (Kukuk and May 1990, Ross and Fletcher 1986). Analysis of
male weights in F. exsectoides showed a tendency toward dimorphism (possibly
indicative of male diploidy) (Chapter 1). F ortelius et al. (1987), in a study of Formica
exsecta, described a strong bimodal dimorphism of males, which they believed
represented differing reproductive strategies: small males tended to disperse farther
while large males mated in proximity of the nest. Since my data did not show such a
strong bimodality, and the presence of a significant production of diploid males may
distort the observed sex ratio, I felt that it was important to investigate this possibility.
Diploid males are produced accidentally from eggs intended to be females; thus colonies
producing a large percentage of diploid males could appear to exhibit a male bias when,

functionally and intentionally, the bias may actually be toward females.

127

Methods

Diploid males can be detected with several techniques, including staining and counting of
chromosomes, allozyme electrophoresis, and measuring total weight of DNA. I have

chosen to use the former two options.

A. Chromosome staining

Brain tissue was dissected from reproductive caste pupae of F. exsectoides. Staining
protocol followed Imai et. al. (1977). Tissues were placed in a hypotonic solution (1%
sodium citrate) with 0.005% colchicine (measured in weight/volume) for 10-20 minutes,
fixed with absolute ethyl alcohol and glacial acetic acid in the mixtures described by Imai
et. al., and stained using a 10% solution of commercially prepared Giemsa stain (Sigma-

Aldrich, St. Louis, MO).

B. Allozyme electrophoresis

Detection of diploid males through allozyme electrophoresis is dependent upon the
presence of substantial levels of heterozygosity within populations. As homozygous
diploid males will be indistinguishable from haploids, the extent of male diploidy among
homozygotes must be estimated by multiplying the proportion of diploid individuals
discovered among heterozygotes by the proportion of individuals homozygous at the loci
under consideration (Owen and Packer 1994).

All allozyme assays were performed using cellulose-acetate electrophoresis, with agar
overlay stains as outlined in Richardson et. al. (1986). Initial screenings for enzyme

polymorphism were performed on fresh F. exsectoides workers from a queenless colony

128

maintained in the Michigan State University Bug House. Abdomens were removed from
workers to prevent disruption of buffer pH by formic acid (Nipson 1978). Approximately
10 workers were homogenized together in each lane of the cellulose acetate plate,
allowing the screening of large numbers of workers for enzyme variation with lower
investment in labor and materials. Approximately 30 workers were screened for
variation at each of the following loci: PGM, PGD, PGK, MDH, ME, MPI, IDH, G6PD,
LDH, ACO, ADH, GPI, GLDH, and HK. As cellulose acetate electrophoresis requires a
very small amount of sample per plate, a single worker homogenate provided enough
material for the assay of 5-8 enzymes. Plates were not multiply stained, but did have
multiple origins in some cases.

Although homogenizing 10 individuals into a single lane removes the possibility of
distinguishing between heterozygosity and within-species polymorphism, the only
interest in this case was to locate loci with potentially informative variation. Enzyme loci
showing uniform banding patterns across all workers screened were not considered for
further investigation.

Using stains for the remaining enzymes, 1 then examined sexual caste pupae that had
been collected from mounds outside of my study plots at sites Y3, Y2 and 03 (see
Chapter 3: Methods). These pupae had been held in a 10°C incubator for 10 days, then
frozen at -80°C. Two females and 3 males from each site were electrophoresed and

stained for the following enzymes: ACO, GPI, HK, IDH, MDH, ME, PGM, and PGD.

129

Results

A. Chromosome staining
Although chromosomes could be clearly observed in some cells, no observed cells were

in appropriate mitotic phases to determine chromosome number.

B. Allozyme electrophoresis

PGK, G6PD, and GLDH showed uniform single bands in all workers screened. MP1,
ADH, and LDH presented technical difficulties in staining that would have been too
time-consuming to resolve for this study.

Banding patterns produced for ACO, GPI, HK, IDH, MDH, ME, PGM, and PGD are
presented in Table 29. Although electrophoresis of workers indicated some species-level
variation at each of the loci tested, analysis of sexual forms from the three populations
sampled revealed no within-population variation in banding patterns (except occasional

distance variation caused by inadequately controlled running time).

130

Table 29. Allozyme banding patterns for F. exsectoides, values represent cm each band
migrated from the origin. The column “workers” represents a summation of all bands in
all samples (each sample contained multiple ants, see text). Columns Y1, Y4, and 04
contain consensus banding patterns from male and female sexual forms sampled in stands
Y1, Y4, and O4—no within-stand variation in banding pattern occurred at any locus.

 

 

 

Enzyme banding pattern (cm from origin)
workers Y1 Y4 O4
ACO 0. 1.8 0, 1.8 0.3, 1.8 O, 1.8
GPI 0. 4 0, 3.5 0, 3.5 0, 4
HK 1.5, 2.1, 2.5 1.3, 2 1.3, 2 1.8
IDH 1.1, 1.3, 1.7 1 0.5-0.9 0.4-0.8
MDH (-0.2), 2.5 0-0.5, 2.5-3.3 0-0.5, 2.5-3.4 0-O.5, 2.5-3.5
ME 2.5, 3.0, 3.5 2.2 2.2 2.2
PGM 0.9. 1.5 1.5 1.5 2.5
POD 2, 3.1 2 2 2

131

Discussion

Although diploid males may be present in F. exsectoides populations, this study did not
provide sufficient data to justify any conclusion on the matter. The lack of appropriate
mitotic phases for karyotype determination is doubtlessly due only to the
inappropriateness of the late pupal stage of development for this analysis (a lack of cell
division), and could be remedied by substitution of a different life stage. As sex cannot
be determined by morphology at the egg or larval stages, identification of diploid males
by chromosome staining will require adults.

The uniformity of banding pattern within populations is more interesting, and may be
consistent with inbreeding as suggested by Nipson (1978).

Four enzymes, ACO, GPI, HK, and PGM, showed population—level variation in banding
pattern. In two of these four cases (GPI and PGM), O4 is the differing stand. 04 is more
geographically distant from Y1 and Y4 than Y1 and Y4 are from each other, but a sample
of three populations is not adequate for speculation about dispersal distances.

While a highly inbred population would seem to be a prime target for the investigation of
the frequency of male diploidy and its effects on sex ratio, this same level of inbreeding
leads to a reduction in genetic variation, making the identification of diploid males a
more trying task. While the late pupal stage is more convenient for collection and
determination of sex by morphology, it is inappropriate for karyotype analysis. Further
attempts to investigate male diploidy in F. exsectoides must rely on hypervariable genetic
markers, which can provide a wealth of information on the genetic structure of inbred

populations and greater resolution than allozyme techniques.

132

APPENDIX C

133

APPENDIX C

Record of Deposition of Voucher Specimens*
The specimens listed on the following sheet(s) have been deposited in the named
museum(s) as samples of those species or other taxa which were used in this research.

Voucher recognition labels bearing the Voucher No. have been attached or included in
fluid-preserved specimens.

Voucher No.:i1998-09

Title of thesis or dissertation (or other research projects):

Effects of habitat succession on pOpulation and reproduction of
the Allegheny mound ant

Museum(s) where deposited and abbreviations for table on following sheets:

Entomology Museum, Michigan State University (MSU)

Other. Museums:

Investigator’s Name(s) (typed)
Heather C. Rowe

 

 

Date 12/09/1998

* Reference: Yoshimoto, C. M. 1978. Voucher Specimens for Entomology in North
America. Bull. Entomol. Soc. Amer. 24:141-42.

Deposit as follows:
Original: Include as Appendix C in ribbon copy of thesis or dissertation.
Copies: Include as Appendix C in copies of thesis or dissertation.
Museum(s) files.
Research project files.

This form is available from and the Voucher No. is assigned by the Curator, Michigan
State University Entomology Museum.

134

 

ouwn neuwuso mm\mo\mfi m a mo

mug m§é\m.s\\wM\.

9.3.5335 3.3m adwanuda 0.5 a.“ ufimoaov
you 30.5003 wean: gone any vufiuuom 98 g
g .02 nanoso> Apoguv Amvamz mLOumemwéH

Axummmwooc «a muoocm Hmcofiufippm mwuv

 

 

 

 

 

Pages

 

of1

w

Voucher Spec inert Data

Page 1

 

 

 

 

 

 

 

 

 

 

9.8 .I
mama 32, m
mmd. day JEN...
Sm m9 ggng nggxmfigd
J 0.. panacea—av 28 can: no 33033 aoxuu uefio uo mofiumam
m e r r m u e M. u unosfloonn you dump Honda
r 0.0 m 1 1 a D. 8
m e ”I“ t c“ u p w w 00
M .mdi 0 A M an” N h h...
"we hung—z

 

 

 

 

REFERENCES CITED

136

REFERENCES CITED

Agosti, D.; Hauschteck-Jungen, E. 1987. Polymorphism of males in Formica exsecta
Nyl. (Hymenoptera: Formicidae). Insectes Sociaux. 34(4): 280-290.

Allen, Douglas C.; Knight, Fred B.; F oltz, John L. 1970. Invertebrate predators of the
Jack-Pine Budworm, Choristoneura pinus, in Michigan. Annals of the
Entomological Society of America. 63(1): 59-64.

Andrews, E. A. 1926. Sequential distribution of Formica exsectoides F orel. Psyche. 33:
127-150.

Andrews, E. A. 1927. Ant mounds as to temperature and sunshine. Journal of
morphology and physiology. 44(1): 1-20.

Andrews, E. A. 1929. The mound-building ant, Formica exsectoides F., associated with
tree-hoppers. Annals Entomological Society of America. 22: 369-391.

Arnett, Ross H. 1993. American insects: a handbook of the insects of America north
of Mexico. The Sandhill Crane Press, Inc.: Gainesville, FL.

Ayre, G. L. 1963. Laboratory studies of the feeding habits of seven species of ants
(Hymenoptera: Formicidae) in Ontario. The Canadian Entomologist. 95: 712-
715. ‘

Beaudry, S.; Duchesne, L. C.; Cote, B. 1997. Short-term effects of three forestry
practices on carabid assemblages in a jack pine forest. Canadian Journal of
Forest Research. 27(12): 2065-2071.

Beland, Martin; Bergeron, Yves. 1993. Ecological factors affecting abundance of
advanced growth in jack pine (Pinus banksiana Lamb.) stands of the boreal forest
of northwestern Quebec. The Forestry Chronicle. 69 (5): 561-568.

137

Bishop, D. B. 1998. Differential effects of the Allegheny mound ant, Formica
exsectoides, F orel, on aphid, scale, and predator populations and their
interactions in jack pine forests. Ph.D. dissertation, Michigan State University.

Boomsma, J .J . 1989. Sex investment ratios in ants: has female bias been systematically
overestimated? American Naturalist. 133: 517-532.

Boomsma, J .J . 1993. Sex ratio variation in polygynous ants. Queen number and
sociality in insects. (ed. L. Keller) Oxford University Press: New York.

Boomsma, J .J .; Grafen, A. 1991. Colony-level sex ratio selection in the eusocial
Hymenoptera. Journal of Evolutionary Biology. 3: 383-407.

Boomsma, J .J .; Isaaks, J. A. 1985. Energy investment and respiration in queens and
males of Lasius niger (Hymenoptera: Formicidae). Behavioral Ecology and
Sociobiology. 18: 19-27.

Boomsma, J .J .; Keller, L.; Nielsen, M. G. 1995. A comparative analysis of sex ratio
investment parameters in ants. Functional Ecology. 9: 743-753.

Borror, Donald J .; Triplehom, Charles; Johnson, Norman F. 1989. An introduction to
the study of insects (6th edition). Saunders College Publishing: Forth Worth, TX.

Bourke, A. F. G.; Franks, N. 1995. Social Evolution in Ants. Princeton University
Press: Princeton, NJ.

Bradley, G.A.; Hinks, J .D. 1968. Ants, aphids. and jack pine in Manitoba. The
Canadian Entomologist. 100: 40-50.

Brian, M. V.; Brian, A. D. 1951. Insolation and ant population in the west of Scotland.
Transactions of the Royal Entomological Society of London. 102: 303-330.

Bristow, C. M.; Cappaert, D.; Campbell, N. J .; Heise, A. 1992. Nest structure and
colony cycle of the Allegheny mound ant, Formica exsectoides F orel
(Hymenoptera: F ormicidae). Insectes Socieaux. 39: 385-402.

138

Bull, J .J . 1983. Evolution of sex determining mechanisms. Benjamin Cummings,
Menlo Park, CA.

Campbell, N. J. 1991. An evaluation of Formica exsectoides Forel as a potential
biological control agent of insect pests in pines. Ph.D. dissertation, Michigan
State University.

Campbell, N. J .; Bristow, C. M.; Ayers, G.; Simmons, G. 1991. Design and field test of
portable colonies of the predaceous ant, Formica exsectoides (Hymenoptera:
Formicidae). Journal of the Kansas Entomological Society. 64: 116-120.

C anham, Charles D. 1995. GLI/C: Software for calculation of light transmission
using color fisheye photography. Institute for Ecosystem Studies: Millbrook,
NY.

Cannings, C.; Cruz Orive, Luis, M. 1975. On the adjustment of the sex ratio and the
gregarious behavior of animal populations. Journal of Theoretical Biology. 55:
1 15-136.

C harnov, Eric L. 1993. Life history invariants: some explorations of symmetry in
evolutionary ecology. Oxford University Press, New York.

Charnov, Eric L. 1982. The theory of sex allocation. Princeton University Press,
Princeton, NJ.

C lark, A. B. 1978. Sex ratio and local resource competition in a prosimian primate.
Science. 201: 163-165.

Clark, P. J .; Evans, F. C. 1954. Distance to nearest neighbor as a measure of spatial
relationships in populations. Ecology. 60: 316-317.

Cody, Martin L. 1966. A general theory of clutch size. Evolution. 20:174-184.

Cory, Ernest N.; Haviland, Elizabeth E. 1938. Population studies of Formica
exsectoides Forel. Annals Entomological Society of America. 31: 50-56.

139

Creighton, W. S. 1950. The ants of North America. Bulletin of the Museum of
Comparative Zoology, Harvard University. 104: 1-585.

Crow, J. F. 1994. Advantages of sexual reproduction. Developmental Genetics. 15(3):
205-213.

Crozier, R. H. 1971. Heterozygosity and sex determination in haplo-diploidy. American
Naturalist. 105(945): 399-412.

Crozier, R. H.; Pamilo, P. 1996. Evolution of Social Insect Colonies: Sex Allocation
and Kin Selection. Oxford University Press: New York.

Deslippe, Richard J .; Savolainen, Riita. 1995. Sex investment in a social insect: the
proximate role of food. Ecology. 76(2): 375-382.

Dimmick, John F. 1951. An ecological study of Formica Formica exsectoides. Illinois
Academy of Science Transactions. 44: 197-204.

Duchesne, S.; Sirois, L. 1995. The first stage of post-fire regeneration in subarctic
coniferous populations. Canadian Journal of Forest Research. 25(2): 307-318.

Elmes, G.W.; Wardlaw, J .C. 1982. A population study of the ants Myrmica sabuleti and
Myrmica scabrinodis, living at two sites in the south of England. I. A comparison
of colony populations. Journal of Animal Ecology. 51: 651-664.

Fisher, R. A. 1930. The genetical theory of natural selection. Oxford University
Press: Oxford, England.

F ortelius, W.; Pamilo, P.; Rosengren, R.; Sundstrom, L. 1987. Male size dimorphism
and alternative reproductive tactics in Formica exsecta ants (Hymenoptera:
Formicidae). Ann. 2001. Fennici. 24: 45-54.

Godfray, H.C.J.; Werren, J .H. 1996. Recent developments in sex ratio studies. Trends
in Ecology and Evolution. 11(2): 59-63.

140

Halaj, J .; Ross, D. W.; Moldenke, A. R. 1997. Negative effects of ant foraging on
spiders in Douglas-fir canopies. Oecologia. 109(2): 313-322.

Hamilton, W. D. 1963. The evolution of altruistic behavior. American Naturalist. 97:
354-356.

Hamilton, W.D. 1964. The genetical evolution of social behaviour. Journal of
Theoretical Biology. 7: 1-32.

Hamilton, W.D. 1967. Extraordinary sex ratios. Science. 156: 477-488.

Haviland, E. E. 1947. The biology and control of the Allegheny mound ant. Journal of
Economic Entomology. 40: 413-419.

Haviland, EB. 1948. Mound changes after ten years in colonies of Formica exsectoides
Forel. Annals of the Entomological Society of America. 41: 438.

Heatwole, Harold; Levins, Richard. 1972. Trophic structure stability and faunal change
during recolonization. Ecology. 53(3): 531-534.

Hendrix, S.D.; Brown, V.K.; Dingle, H. 1988. Arthropod guild structure during early
old field succession in a new and old world site. Journal of Animal Ecology.
57: 1053-1065.

Herbers, Joan M. 1986. Effects of ecological parameters on queen number in
Leptothorax longispinosis (Hymenoptera: Formicidae). Journal of the Kansas
Entomological Society. 59(4): 675-686.

Herbers, Joan M. 1993. Ecological determinants of queen number in ants. in Queen
Number and Sociality in Insects (ed. L. Keller) Oxford University press: New
York, p.262-293.

Herbers, Joan M.; Banschbach, Valerie. 1998. Food supply and reproductive allocation
in forest ants: repeated experiments give different results. Oikos. 83: 145-151.

141

Herbers, Joan M.; Stuart, Robin J. 1998. Patterns of reproduction in slave-making ants.
Proceedings of the Royal Society of London, Series B. 265: 875-887.

Hollbobler, Burt; Wilson, E. O. 1990. The Ants. Harvard University Press: Cambridge,
MA.

Horn, HS. 1974. The ecology of secondary succession. Annual Review of Ecology
and Systematics. 5: 25-37.

Hurst, L. D.; Peck, J. R. 1996. Recent advances in understanding the evolution and
maintenance of sex. Trends in Ecology and Evolution. 11(2): 46-52.

Imai, H. T.; Crozier, R. H.; Taylor, R. W. 1977. Karyotype evolution in Australian ants.
Chromosoma. 59: 341-393.

Juliano, Steven A. 1993. Nonlinear curve fitting: predation and functional response
curves. pp. 159-182 in Design and analysis of ecological experiments. (ed.
Samuel M. Scheiner and Jessica Gurevitch) Chapman & Hall, New York.

Keller, L. 1991. Queen number, mode of colony founding, and queen reproductive
success in ants (Hymenoptera: F orrnicidae). Ethology, Ecology, and Evolution.
3: 307-316.

Kondrashov, A. S. 1993. Classification of hypotheses on the advantage of amphimixis.
Journal of Heredity. 84(5): 372-387.

Kukuk. R; May, B. 1990. Diploid males in a primitively eusocial bee, Lasioglossum
(Dialictus) chhyrum (Hymenoptera: Halictidae). Evolution. 44: 1522-1528.

Lande, R. 1982. A quantitative genetic theory of life history evolution. Ecology. 63:
607-615.

Levins, S.C.; Traniello, J FA. 1981. Territoriality, nest dispersion, and community
structure in ants. Psyche. 88: 265-319.

142

MacKay, W. P. 1985. A comparison of the energy budgets of three species of
Pogonomyrmex harvester ants (Hymenoptera: Formicidae). Oecologia. 66: 484-
494.

Maynard Smith, J. 1978. The Evolution of Sex. Cambridge University Press:
Cambridge, England.

McCook, H. C. 1877. Mound-making ants of the Alleghenies, their architecture and
habits. Transactions of the Entomological Society of America. 6: 253-296.

Moran, V.C.; Southwood, T.R.E. 1982. The guild composition of arthropod
communities in trees. Journal of Animal Ecology. 51: 289-306.

 

Naylor, B. J .; Bendell, J. F. 1983. Influence of habitat diversity on the abundance and
diversity of small mammals in jack pine forests in Ontario. pp.295-307 in
Resources and Dynamics of the Boreal Zone. (ed. Wein, Riewe, Methven)
Association of Canadian Universities for Northern Studies.

Nealis, V.G.; Lomic, P.V. 1994. Host-plant influence on the population ecology of the
jack pine budworrn, Choristoneura pinus (Lepidoptera: Tortricidae). Ecological
Entomology. 19: 367-373.

Nipson, H. E. 1978. Inbreeding in the ant species Formica exsectoides. Ph.D.
dissertation. Harvard University.

Nonacs, P. 1986. Ant reproductive strategies and sex allocation theory. Quarterly
Review of Biology. 61: 1-21.

Oster, George F .; Wilson, Edward O. 1978. Caste and ecology in the social insects.
Princeton University Press, Princeton, NJ.

Owen, R. E.; Packer, L. 1994. Estimation of the proportion of diploid males in
populations of Hymenoptera. Heredity. 72: 219-227.

Pamilo, P.; Rosengren, R. 1983. Sex ratio strategies in Formica ants. Oikos. 40: 24-35.

143

Pamilo, P.; Sundstrom, L.; F ortelius, W.; Rosengren, R. 1994. Diploid males and
colony-level selection in Formica ants. Ethology, Ecology, and Evolution. 6:
221-235.

Patterson, RS. 1994. Biological control of introduced ant species. in Exotic ants:
biology, impact, and control of introduced species. (ed. D. F. Williams)
Westview Press: Boulder, CO, p.293-307.

Pielou, E. C. 1969. Mathematical Ecology. John Wiley: New York.

Price, W. A. 1945. The Allegheny mound ant and its control. Journal of Economic
Entomology. 38(6): 706.

Rice, W. R. 1990. A consensus combined p-value test and the family-wide significance
of component tests. Biometrics. 46: 303-308.

Richardson, B. J .; Baverstock, P. R.; Adams, M. 1986. Allozyme Electrophoresis: a
handbook for animal systematics and population studies. Academic Press.

Roberts, Mark R.; Gilliam, Frank S. Patterns and mechanisms of plant diversity in
forested ecosystems: implications for forest management. Ecological
Applications. 5(4): 969-977.

Rosengren, R.; Pamilo, P. 1986. Sex ratio strategy as related to queen number, dispersal
behavior and habitat quality in Formica ants (Hymenoptera: Formicidae).
Entomol. General. 11: 139-151.

Ross. K. G.; Fletcher, D. J. C. 1986. Diploid male production—a significant colony
mortality factor in the fire ant, Solenopsis invicta (Hymenoptera: F ormicidae).
Behavioral Ecology and Sociobiology. 19: 283-291.

Rudolph, T. D.; Laidly, P. R. 1990. Pinus banksiana Lamb (Jack Pine). pp.280-293 in
Silvics of North America: 1. Conifers. (ed. Burns, Russell M.; Honkala,
Barbara H.) Agriculture Handbook 654. US. Department of Agriculture, Forest
Service, Washington, DC.

144

Savolainen, Riita. 1990. Colony success of the submissive ant Formica fusca within
territories of the dominant Formica polyctena. Ecological Entomology. 15: 79-
85.

Scheiner, S. M. 1993. MANOVA: Multiple response variables and multispecies
interactions. pp. 94-112 in Design and Analysis of Ecological Experiments.
(ed. Samuel M. Scheiner, Jessica Gurevitch) Chapman & Hall: New York.

Scherba, Gerald. 1958. Reproduction, nest orientation and population structure of an
aggregation of mound nests of Formica ulkei Emery (Formicidae). Insectes
Sociaux. 5(2): 201-213.

Spires, S.; Bendell, J .F. 1983. Early postfire effects on some invertebrates, small
mammals and birds in north-central Ontario. pp.308-318 in Resources and
Dynamics of the Boreal Zone: Proceedings of a conference held at Thunder
Bay, Ontario, August 1982. Association of Canadian Universities for Northern
Studies.

Smith, M. R. 1947. A generic and subgeneric synopsis of the United States ants, based
on workers. American Midland Naturalist. 37: 521-647.

Starr, C. K. 1979. Origin and evolution of insect sociality: a review of modern theory.
in Social Insects, vol.1 (ed. Hermann, H. R.). Academic Press: New York, p.35-
80.

Thomhill, Nancy Wilmsen.(ed.) 1993. The natural history of inbreeding and
outbreeding. University of Chicago Press, Chicago, IL.

Trivers, R. L. 1985. Social Evolution. Benjamin Cummings: Menlo Park, CA.

Trivers, R. L.; Hare, H. 1976. Haplodiploidy and the evolution of the social insects.
Science. 191: 249-263.

Volney, W.J.A.; McCullough, D.G. 1994. Jack pine budworrn population behavior in
northwestern Wisconsin. Canadian Journal of Forest Research. 24: 502-510.

145

Weaver, J.E.; Smith, BB. 1993. Allegheny mound ant control, Grant Co, WV, 1992.
Insecticide and Acaracide Tests. 18: 337.

Wheeler, Diana E. 1994. Nourishment in ants: patterns in individuals and societies.
pp.245-277 in Nourishment and Evolution in Insect Societies. (ed. James H.
Hunt and Christine A. Nalepa) Westview Press: Boulder, CO.

Wheeler, W.M. 1928. The social insects: their origin and evolution. Harcourt, Brace
and Company, New York.

Wheeler, W.M. 1933. Colony-founding among ants. Harvard University Press:
Cambridge, MA.

Williams, G. C. 1975. Sex and Evolution. Princeton University Press: Princeton. NJ.

Wilson, E. O. 1971. The Insect Societies. Harvard University Press: Cambridge, MA.

Wilson, Louis F. 1977. A guide to insect injury of conifers in the Lake States.
USDA Forest Service, Agriculture Handbook No. 501: Washington, DC.

Zoladeski, C. A.; Maycock, P. F. 1990. Dynamics of the boreal forest in northwest
Ontario. American Midland Naturalist. 124(2): 289-300.

146

 

 

 

"111111111111111“