<%BANNER%>

The Origin and Spread of the West Indian Drywood Termite Cryptotermes Brevis (Walker) in the Azores using Genetic Marker...

Permanent Link: http://ufdc.ufl.edu/UFE0022723/00001

Material Information

Title: The Origin and Spread of the West Indian Drywood Termite Cryptotermes Brevis (Walker) in the Azores using Genetic Markers, and Testing of Colony Foundation Preventative Measures to Control Its Further Spread
Physical Description: 1 online resource (144 p.)
Language: english
Creator: Bravo Ferreira, Maria Teresa
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: azores -- drywood -- microsatellites -- population -- termites
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Cryptotermes brevis (Walker) is a drywood termite of the Kalotermitidae family (Isoptera). The West Indian drywood termite, C. brevis, is a very destructive wood pest that causes high levels of damage to wood and wood structures throughout the world in tropical and subtropical areas. This species was first identified in the Azorean archipelago in 2000, and has since then become the number one urban pest in that archipelago. The Azores is a small group of nine islands located in the Atlantic Ocean, about half way between Europe and the United States. In order to identify the origin and spread of C. brevis in the Azores, genetic markers were used to elucidate if the introduction of the species to the islands was a consequence of a single introduction or multiple introductions. Mitochondrial DNA and microsatelites were used to statistically model the relationships between the populations in the Azores. Nine different locations were chosen to collect termites in four of the islands. The termite DNA was extracted, and 16SrRNA and Cytb mitochondrial genes were used to assess phylogenetic relationships between the populations. Furthermore, five loci for microsatellite DNA were used as well. The data showed that there were multiple introductions of the species and that they still appear to be occurring, seeing that the genetic diversity is high. This is important information to have in order to advise stricter regulations for the exchange of wood products into and between the islands. To determine the origin of these introductions, samples from 11 different countries were used. The same genetic markers were used to show that the two possible places of origin might be the endemic region of Chile and Peru, and Venezuela. Lastly, colony preventative treatments were studied in order to optimize control of the species in the Azores and possibly decrease the spread of the infestation. Different light wavelengths were studied to determine if there was a preference for a determined wavelength, and it was shown that lights in the white, blue, and green spectrum are preferred to red or no light. Also chemicals were tested against colony foundation and it was found that the two most common wood treatments used in the Azores, permethrin and cypermethrin, were effective at preventing infestations. This knowledge can now be applied to help reduce the spread of the infestation in the locations where it is already present.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Maria Teresa Bravo Ferreira.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Scheffrahn, Rudolf H.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0022723:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022723/00001

Material Information

Title: The Origin and Spread of the West Indian Drywood Termite Cryptotermes Brevis (Walker) in the Azores using Genetic Markers, and Testing of Colony Foundation Preventative Measures to Control Its Further Spread
Physical Description: 1 online resource (144 p.)
Language: english
Creator: Bravo Ferreira, Maria Teresa
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: azores -- drywood -- microsatellites -- population -- termites
Entomology and Nematology -- Dissertations, Academic -- UF
Genre: Entomology and Nematology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Cryptotermes brevis (Walker) is a drywood termite of the Kalotermitidae family (Isoptera). The West Indian drywood termite, C. brevis, is a very destructive wood pest that causes high levels of damage to wood and wood structures throughout the world in tropical and subtropical areas. This species was first identified in the Azorean archipelago in 2000, and has since then become the number one urban pest in that archipelago. The Azores is a small group of nine islands located in the Atlantic Ocean, about half way between Europe and the United States. In order to identify the origin and spread of C. brevis in the Azores, genetic markers were used to elucidate if the introduction of the species to the islands was a consequence of a single introduction or multiple introductions. Mitochondrial DNA and microsatelites were used to statistically model the relationships between the populations in the Azores. Nine different locations were chosen to collect termites in four of the islands. The termite DNA was extracted, and 16SrRNA and Cytb mitochondrial genes were used to assess phylogenetic relationships between the populations. Furthermore, five loci for microsatellite DNA were used as well. The data showed that there were multiple introductions of the species and that they still appear to be occurring, seeing that the genetic diversity is high. This is important information to have in order to advise stricter regulations for the exchange of wood products into and between the islands. To determine the origin of these introductions, samples from 11 different countries were used. The same genetic markers were used to show that the two possible places of origin might be the endemic region of Chile and Peru, and Venezuela. Lastly, colony preventative treatments were studied in order to optimize control of the species in the Azores and possibly decrease the spread of the infestation. Different light wavelengths were studied to determine if there was a preference for a determined wavelength, and it was shown that lights in the white, blue, and green spectrum are preferred to red or no light. Also chemicals were tested against colony foundation and it was found that the two most common wood treatments used in the Azores, permethrin and cypermethrin, were effective at preventing infestations. This knowledge can now be applied to help reduce the spread of the infestation in the locations where it is already present.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Maria Teresa Bravo Ferreira.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Scheffrahn, Rudolf H.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0022723:00001


This item has the following downloads:


Full Text

PAGE 1

1 THE ORIGIN AND SPREAD OF THE WEST INDIAN DRYW OOD TERMITE C ryptotermes brevis (WALKER) IN THE AZORES USING GENETIC MARKERS AND TESTING OF COLONY FOUNDATION PREVENTATIVE MEASURES TO CONTROL ITS FURTHER SPREAD By MARIA TERESA MONTEIRO DA RO CHA BRAVO FERREIRA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Maria Teresa Monteiro da Rocha Bravo Ferreira

PAGE 3

3 To my Dad, my Mom and my broth er who have always supported me

PAGE 4

4 ACKNOWLEDGMENTS I acknowledge and thank Dr. Rudolf Scheffrahn for guiding me on this scientific road, offering an environment that provided cultivation of knowledge. I also acknowledge the Portuguese Foundation for Science ( SFRH/BD/29840/2006), the Azorean DRCT (Project M221I 002 2009 TERMODISP), and the Structural School of Fumigation for providing me with the funds that allowed me to pursue this degree. I w ould like to thank all the teachers and scientists at the FLREC namely Dr. Robin Giblin Davis, Dr. Nan Yao Su, Dr. William Kern, Dr. Hartwig Hochmair and Dr. Nigel Harrison for helping me increase my knowledge and aiding with any questions that I had. I acknowledge Dr. Paulo Borges for initiating my path into the world of termites and for supporting me unconditionally throughout this whole process I would like to acknowledge my roommates, who provided a relaxing and fun environment at home, but also an i ntellectually stimulating environment. I would like to thank Dr Seemanti Chakrabarti and Ericka Helmick M Sc for all the help with the genetics laboratory as well as interpretation of results, and for their support, and friendship. I would like to acknow ledge Dr Thomas Chouvenc, and Dr Hou Feng Li for their friendship and inspiration. And f inally, I would like to thank my mother and my brother for being a huge support in my life even from far away, and to all my friends back home who supported me anywa y that they could from a distance. To all the people that one way or another have contributed to this work, I give a most sincere thank you.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES ........................................................................................................ 11 LIST OF ABBREVIATIONS ........................................................................................... 14 ABSTRACT ................................................................................................................... 15 CHAPTER 1 INTRODUCTION .................................................................................................... 17 Objectives ............................................................................................................... 33 2 POPULATION DIVERSITY OF C ryptotermes brevis IN THE AZORES: SINGLE VS MULTIPLE INTRODUCTIONS .......................................................................... 35 Material and Methods ............................................................................................. 37 Termites ........................................................................................................... 37 DNA extraction and amplification ..................................................................... 38 Data analysis .................................................................................................... 41 Results .................................................................................................................... 50 Mitochondrial DNA ........................................................................................... 50 Microsatellites ................................................................................................... 55 Discussi on .............................................................................................................. 60 3 ORIGIN OF THE AZOREAN POPULATION OF THE WEST INDIAN DRYWOOD TERMITE ............................................................................................ 67 Material and Methods ............................................................................................. 67 Termites ........................................................................................................... 67 DNA extraction and amplification ..................................................................... 68 Data analysis .................................................................................................... 69 Results .................................................................................................................... 81 Mitochondrial DNA ........................................................................................... 81 Microsatellites ................................................................................................... 83 Discussion .............................................................................................................. 94 4 CONTROL OF THE SPREAD OF ALATES USING ATRACTION TO DIFFERENT LIGHT WAVELENGTHS AND CHEMICAL TREATMENTS ............. 100 Materials and Methods .......................................................................................... 101

PAGE 6

6 Different light wavelength preference ............................................................. 101 In Florida .................................................................................................. 101 In the Azores ............................................................................................ 102 Data Analysis ........................................................................................... 104 Chemical prevention of colonization ............................................................... 104 Results .................................................................................................................. 107 Different light wavelength preference ............................................................. 107 In Florida .................................................................................................. 107 In the Azores ............................................................................................ 107 Chemical prevention of colonization ............................................................... 108 Discussion ............................................................................................................ 109 5 CONCLUSIONS ................................................................................................... 113 APPENDIX A EXTRA FIGURES ................................................................................................. 115 B EXTRA TABLES ................................................................................................... 121 LIST OF REFERENCES ............................................................................................. 133 BIOGRAPHICAL SKETCH .......................................................................................... 144

PAGE 7

7 LIST OF TABLES Table page 1 1 Food and nesting habits of the extant families and subfamilies of termites. ....... 18 2 1 Locations of collection sites in the Azores and number of individuals (N) used for analysis. ........................................................................................................ 38 2 2 PCR Cycles for the mitochondrial16S rRNA, and Cytb genes optimized for amplification in C. brevis .................................................................................... 39 2 3 PCR profiles used for all the C. brevis microsatellite loci. ................................... 40 2 4 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cyt b and 16S rRNA of C. brevis .................................. 51 2 5 Continuation of distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis .............. 52 2 6 Continuation of distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis .............. 52 2 7 Summary statistics for the AMOVA among and within subpopulations of the different Islands. ................................................................................................. 54 2 8 Size of alleles in number of base pairs per locus for the five microsatellite loci of C. br evis analyzed. ......................................................................................... 55 2 9 Allele frequencies in percentage for Csec6 locus for the subpopulations of C.brevis calculated by the Microsatellite tool kit for Excel. ................................. 5 5 2 10 Allele frequencies in percentage for Csec5 locus for the subpopulations of C.brevis calculated by the Microsatellite tool kit for Excel. ................................. 55 2 11 Allele f requencies in percentage for Csec4 locus for the subpopulations of C.brevis calculated by the Microsatellite tool kit for Excel. ................................. 56 2 12 Allele frequencies in percentage for Csec1 locus for the subpopulations of C.brevis calculated by the Microsatellite tool kit for Excel. ................................. 56 2 13 Allele frequencies in percentage for Csec3 locus for the subpopulations of C.brevis calculated by the Microsatellite tool kit for Excel. ................................. 56 2 14 Estimates of genetic diversity at 5 polymorphic microsatellite loci in C.brevis from the Azorean Islands. ................................................................................... 57 2 15 Matrix of FST p values between subpopulations of C. brevis for all microsatellite loci.. .............................................................................................. 58

PAGE 8

8 2 16 Estimated values for null allele frequencies per subpopulation, per locus for the C.brevis subpopulations sequenced in this study. ........................................ 58 2 17 Values of FST, FIS, and FIT calculated using three different methods for all loci for all subpopulations .......................................................................................... 58 3 1 Exact locations of collection sites from the world collection and number of individuals (N) used for analysis. ........................................................................ 68 3 2 Subpopulations included in the partitioned data used for the STRUCTURE analysis, ............................................................................................................. 71 3 3 Size of alleles in number of base pairs per locus for the five microsatellite loci of C. brevis analyzed. ......................................................................................... 83 3 4 Allele frequencies, in percentage, for the Csec6 locus for the subpopulations of C.brevis, generated by the Microsatellite tool kit for Excel. ............................ 83 3 5 Allele frequencies, in percentage, for the Csec5 locus for the subpopulations of C.brevis, generated by the Microsatellite tool kit for Excel. ............................ 83 3 6 Allele frequencies, in percentage, for the Csec4 locus for the subpopulations of C.brevis, generated by the Microsatellite tool kit for Excel. ............................ 84 3 7 Allele frequencies, in percentage, for the Csec1 locus for the subpopulations of C.brevis, generated by the Microsatellite tool kit for Excel. ............................ 84 3 8 Allele frequencies, in percentage, for the Csec3 locus for the subpopulations of C.brevis generated by the Micr osatellite tool kit for Excel. ............................ 84 3 9 Estimates of genetic diversity at 5 polymorphic microsatellite loci in C.brevis from the samples from various continents. ......................................................... 85 3 10 Estimated null allele frequencies for all loci for the subpopulations of C. brevis sampled from the University of Florida Collection .................................... 86 3 11 Matrix of FST p values between subpopulations. Significantly different values are indicated by after 190000 permutations. .................................................... 86 3 12 Continued matrix of FST p values between subpopulations. Significantly different values are indicated by after 190000 permutations. ........................... 87 3 13 Continued matrix of FST p values between subpopulations. Significantly different values are indicated by after 190000 permutations. ........................... 87 3 14 Continued matrix of FST p values between subpopulations. Significantly different values are indicated by after 190 000 permutations. .......................... 88

PAGE 9

9 3 15 Values of FST, FIS, and FIT calculated using three different methods for all locus ................................................................................................................... 88 4 1 Total and mean number of alates caught in light traps in Florida. .................... 107 4 2 Total and mean number of alates caught in light traps in the Azores. .............. 108 4 3 Mean SE per attic module of C. brevis liv e dealates, number of nuptial chambers, and number of eggs. ....................................................................... 109 B 1 Primer sequence for mitochondrial genes used for amplification. Adapted from Legendre et al. 2008. ............................................................................... 121 B 2 Primer sequence for microsatellite locus used for amplification. Adapted from Fuchs et al. 2003. ............................................................................................. 121 B 3 Variation sites for the 22 haplotypes of C. brevis sequences for the combined Cytb and 16SrRNA genes for the island samples. ............................................ 122 B 4 Variation sites for the 22 haplotypes of C. brevis sequences for the combined Cytb and 16SrRNA genes for the island samples. ............................................ 123 B 5 Variation sites for the 22 haplotypes of C. brevis sequences for the combined Cytb and 16SrRNA genes for the island samples. ............................................ 124 B 6 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 125 B 7 Distance matrix of esti mates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis .............................. 126 B 8 Distance matrix of estimates of evolutionary divergence between the combined sequenc es of Cytb and 16S rRNA of C. brevis ............................... 127 B 9 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ................................ 127 B 10 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 128 B 11 Distance matr ix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 129 B 12 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 130 B 13 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 131

PAGE 10

10 B 14 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 131 B 15 D istance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ............................... 132 B 16 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis ................................ 132

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Picture of Cryptotermes brevis showing soldiers, reproductives and pseudergates (adapted from Scheffrahn). .......................................................... 22 1 2 Map of the distribution of C. brevis in the world. (Adapted from Scheffrahn et al. 2008). ............................................................................................................ 26 1 3 Confirmed locations of C. brevis infestations in the Azorean Archipelago. Yellow dots mark the locations of the known infestation sites. ........................... 26 2 1 Location of the t owns where C. brevis was collected in the four islands. ........... 37 2 2 Scenario 1 of introduction to the Islands tested with the DIYABC software. ....... 44 2 3 Scenario 2 of introduction to the Islands tested with the DIYABC software.. ...... 45 2 4 Scenario 3 of introduction to the Islands tested with the DIYABC software.. ...... 46 2 5 Scenario 4 of introduction to the Islands tested with the DIYABC software. ....... 47 2 6 Scenario 5 of introduction to the Islands tested with t he DIYABC software. ....... 48 2 7 Scenario 6 of introduction to the Islands tested with the DIYABC software. ....... 49 2 8 Evolutionary relati onships of C. brevis subpopulations. ...................................... 53 2 9 Genealogical relationships among 22 haplotypes of C. brevis estimated by TCS. ................................................................................................................... 54 2 1 0 Distance diagram calculated using Genetic Distance Analysis. ......................... 59 2 11 Logistic regression of the simulated data probability for the different scenarios tested in DIYABC.. ............................................................................. 60 3 1 Scenario 1 of the origin of the Islands subpopulations tested with the DIYABC software. ............................................................................................... 74 3 2 Scenario 2 of the origin of the Islands subpopulations tested with the DIYABC software.. .............................................................................................. 75 3 3 Scenario 3 of the origin of the Islands subpopulations tested with the DIYABC software. ............................................................................................... 76 3 4 Scenario 4 of the origin of the Islands subpopulations tested with the DIYABC software. ............................................................................................... 77

PAGE 12

12 3 5 Scenario 5 of the origin of the Islands subpopulations tested with the DIYABC so ftware. ............................................................................................... 78 3 6 Scenario 6 of the origin of the Islands subpopulations tested with the DIYABC software.. .............................................................................................. 79 3 7 Scenario 2A of or igin of the populations tested with the DIYABC software.. ...... 80 3 8 The evolutionary history was inferred using the UPGMA method. ...................... 82 3 9 Distance diagram calculated using Genetic Distance Analysis. ......................... 89 3 10 Assignments of 104 C. brevis individuals sampled from 9 sites to genetic clusters inferred from STRUCTURE ................................................................... 90 3 11 Assignments of 274 C. brevis individuals sampled from 20 sites to genetic clusters inferred from STRUCTURE ................................................................... 91 3 12 Assignments of 184 C. brevis individuals sampled from 13 sites to genetic clusters inferred from STRUCTURE ................................................................... 92 3 13 Logistic regression of the simulated data probability for the different scenarios tested in D IYABC. .............................................................................. 93 3 14 Logistic regression of the simulated data probability for the different scenarios tested in DIYABC. .............................................................................. 94 4 1 Distribution of the different wavelengths in the boxes. ...................................... 103 4 2 Close up of the light chamber (green light 525nm). .......................................... 103 4 3 Attic modul e design. Mode of assembly and measures are shown Back boards are assembled in a clapboard pattern. .................................................. 106 4 4 Attic modules randomly distributed under U.V. light. ........................................ 106 A 1 Different castes of termites and life cycle (Scheffrahn). .................................... 115 A 2 Gel picture of PCR products of C. brevis amplified mtDNA for 16S rRNA and Cytb genes. ..................................................................................................... 115 A 3 Gel picture of amplified PCR products for four microsatellite loci (Csec1, 4, 5, and 6). .............................................................................................................. 116 A 4 Example of an alignme nt session using MEGA 5.0 where the sequences have been aligned using ClustalW method. ..................................................... 116 A 5 Example of scoring the base pair size of an allel e on the PeakScanner software. ........................................................................................................... 117

PAGE 13

13 A 6 Assignments of 104 C. brevis individuals sampled from 9 sites to genetic clusters inferred from STRUCTURE ................................................................. 117 A 7 Assignments of 104 C. b revis individuals sampled from 9 sites to genetic clusters inferred from STRUCTURE ................................................................. 118 A 8 Assignments of 274 C. brevis individuals sampled from 20 sites to genetic clusters inferred from STRU CTURE ................................................................. 118 A 9 Assignments of 274 C. brevis individuals sampled from 20 sites to genetic clusters inferred from STRUCTURE. ................................................................ 119 A 10 Assignments of 184 C. brevis individuals sampled from 13 sites to genetic clusters inferred from STRUCTURE ................................................................. 119 A 11 Assignments of 184 C. brevis individuals sampled from 13 sites to genetic clusters inferred from STRUCTURE ................................................................. 120

PAGE 14

14 LIST OF ABBREVIATION S 16S rRNA 16S ribosomal RNA mitochondrial gene. BLAST Basic Local Alignment Search Tool. Cytb Cytochrome b mitochondrial gene. DNA Deoxyribonucleic acid mtDNA Mitochondrial DNA. PCR Polymerase Chain reaction. RNA Ribonucleic acid. UPGMA Unweighted Pair Group Method with Arithmetic Mean

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Pa rtial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DETERMINING THE ORIGIN, AND SPREAD OF THE WEST INDIAN DRYWOOD TERMITE Cryptotermes brevis (WALKER) IN THE AZORES USING GENETIC MARKERS. AND TESTING COLONY FOUNDATION PREVENTATIVE MEASURES TO CONTROL F U RTHER SPREAD OF THE SPECIES. By Maria Teresa Monteiro da Rocha Bravo Ferreira December 2011 Chair: Rudolf H. Scheffrahn Major: Entomology and Nematology Cryptotermes brevis (Walker) is a drywood termite of the Kalotermitidae famil y (Isoptera). The West Indian drywood termite, C. brevis is a very destructive wood pest that causes high levels of damage to wood and wood structures throughout the world in tropical and subtropical areas. This species was first identified in the Azorean archipelago in 2000, and has since then become the number one urban pest in that archipelago. The Azores is a small group of nine islands located in the Atlantic Ocean, about half way between Europe and the United States. In order to identify the origin and spread of C. brevis in the Azores, genetic markers were used to elucidate if the introduction of the species to the islands was a consequence of a single introduction or multiple introductions. Mitochondrial DNA a nd micros atelites were used to statistic al ly model the relationships between the populations in the Azores. Nine different locations were chosen to collect termites in four of the islands. The termite DNA was extracted, and 16SrRNA and Cytb mitochondrial genes were used to assess phylogenetic relationships between the populations. Furthermore, five loci for microsatellite DNA were used as well. The data showed that there were

PAGE 16

16 multiple introductions of the species and that they still appear to be occurring, s eeing that the genetic diversity is hig h. This is important information to have in order to advise stricter regulations for the exchange of wood products into and between the islands. To determine the origin of these introductions, samples from 11 dif f erent countries were used. The same genetic markers were used to show that the two possible places of origin might be the endemic region of Chile and Peru, and Venezuela. Lastly colony preventative treatments were studied in order to optimize control of the species in the Azores and possibly decr ease the spread of the infestation. Different light wavelengths were studied to determine if there was a preference for a determined wavelength, and it was shown that lights i n the w hite, blue, and green spectrum are preferred to r ed or no light. Also chem icals were tested against colony foundation and it was found that the two most common wood treatments used in the Azores permethrin and cypermethrin, were effective at preventing infestations. This knowledge can now be applied to help reduce the spread of the infestation in the locations where it is already present

PAGE 17

17 CHAPTER 1 INTRODUCTION Termites are eusocial insects that comprise an order of their own, the Isoptera. These insects were first described to the Royal Society of London in 1781 by Henr y Smeathman (Howse 1970). Their name comes from the characteristic of the r eproductive adults having two pairs of wings of almost equal size (iso = equal, ptera = wings) There are over 2, 6 00 described species of termites ( Kambhampati and Eggleton 2000) wh ich have been traditionally distributed over seven families. More recently, termites have been classified into nine families with seven extant, two fossil families and one with uncertain status (Engel et al. 2009). These famil ies are the Mastotermitidae, T ermopsidae, Hodotermitidae, Archotermopsidae, Stolotermitidae, Kalotermitidae, Stylotermitidae, Rhinotermitidae, Serritermitidae, and Termitidae (Table 11 for details on ecology and common names of each famil y) Termites follow the species richness latitu dinal gradient pattern, with a higher number of species in tropical areas, and a lower number towards the temperate areas (Eggleton 2000) The iso pterans are closely related to the co ckroach order (Blattodea) (Inward et al. 2007) The termites are believed to have evolved from an extinct primitive type of wood feeding cockroach ancestor It has been acknowledged that termites share several characteristics with the wood roaches of the Cryptocercidae family within the genus Cryptocercus Some termites share several groups of symbiotic flagellates housed in their hindgut, with the wood roaches. Another characteristic that shows similarity is that wood roaches have a semi social behavior, displaying some brood care. Other cockroach hatchlings are independent at birth but woodroaches must acquire the intestinal symbionts by proctodeal feeding from their parents (Cleveland et al. 1934)

PAGE 18

18 The similarity between termites and cockroaches has been phylogenetically analyzed and proven with studies on the cocladogenes is between termites, cockroaches, and their symbionts (Lo et al. 2003) and it has even been proposed that the order Isoptera be sunk into a family within the order Blattodea ( Inward et al. 2007). Table 11. Food and nesting habits of the extant families and subfamilies of termites Family Subfamily Food Nesting Common names Mastotermitidae wood In tree trunks Giant Northe r n termite of Australia Hodotermitidae grass In soil Harvester termites Termopsidae wet wood In damp wood Dampwood termites Kalot ermitidae dry wood In dry wood Drywood termites Rhinotermitidae wood In soil Subterranean termites Serritermitidae nests In decayed wood and other termite nests Termitidae Apicotermitinae soil, dung, grass Few build nests Soldierless termites Macr otermitinae fungus Epigeal Fungus growing termites Nasutitermitinae wood, dung, soil Arboreal, epigeal, or subterranean Nasutes Termitinae wood, grass, dung Arboreal, epigeal, or subterranean Snappi n g or piercing termites Syntermitinae soil dung Ep igeal and other termite nests

PAGE 19

19 Termites are truly social insects. They possess the three main character istics to being eusocial : there is cooperative brood care in their colonies, they are divided into different castes which perform different tasks withi n the colony, and they have overlapping generations. In termites, the larvae not only must acquire symbionts from their nestmates but must also be fed for the first couple of instars because they cannot fe ed directly on wood themselves (Noirot and Noirot Timothe 1969) Termites are often separated into two groups, higher termites and lower termites. In the higher termites (Termitidae), w hich makes up 75% of all termite species, only bacteria are present in the gut. In the lower termites protozoan symbionts can be found in the gut in addition to bacteria (Krishna 1969 ). These symbionts help with the digestion of cellulose. The lower termites are generally more primitive, having simple galleries but not well formed nests (with the exception of some Rhinotermitidae which have mounds for nests) They have colonies without true workers and generally eat only wood. Lower termites usually oc cur in more temperate latitudes than higher termites H igher termites (Termitidae) are much more diverse ecological ly While many still consume wood, others have evolved different diets of herbage, grass, dung, humus, fungus, lichens, or organic material in soil. The higher termites rely either on internal digestion with gut bacteria or external digestion in fungus com bs (Edwards and Mill 1986) The higher termites often build large nests or mounds, and are common in tropical areas but are rare or absent in temperate areas. Termites can also be separated into groups according to their habitat types. They can be earthd welling termites and wooddwelling termites. There are several types of earthdwelling termites, such as the subterranean termites and mound building termites.

PAGE 20

20 The wooddwelling termites on the other hand, confine themselves to wood. Wooddwelling termites can be classified as, either drywood termites, attacking dry, sound wood, or dampwood termites attacking damp usual ly decaying wood (Kofoid 1934). Termite colonies are comprised of individuals that are separated into three main castes which are differenti ated morphologically and behaviorally. These castes are: 1) reproductives (king, queen, and unmated winged forms called alates); 2) soldiers; and 3) workers (higher termites) and false workers or pseudergates (lower termites) (Snyder 1926, Krishna 196 9 ) (F igure 11) A lates are usually dark er in color with a fully developed, pigmented chiti nous exoskeleton, and with well developed compo und eyes (Light 1934a). Alates have deciduous wings that are broken off along a basal suture after a dispersal flight. A fter losing their wings (becoming dealates) the dealates become the primary reproductives of a colony. A colony begins with a single pair of de alates, a male and a female that cohabitate with copulation occurring at different intervals throughout their lives (Krishna1969 ). Besides the primary reproductives, a colony of termites may produce secondary reproductives that develop from nymphs with external signs of wing buds (brachypterous) and tertiary reproductives with no signs of wing buds (apterous) of bo th sexes (Snyder 1926). S oldiers have modified heads and mandibles to protect the colony. There are several mechanisms of defense by soldiers. They can have phragmotic heads which serve to plug entry holes to the colony, they can use mandibular biting in which they will use their strong or sharp mandibles to bite enemies, and they can have mandibular snapping in which there is an elastic distortion of the mandibles that when released,

PAGE 21

21 snap, striking a hard blow to the approaching enemy Soldiers also have chemical defenses with glandular structures in their heads that exude chemical compounds. Glands can be l ocated in the front of the head ( frontal glands ) which are found in Rhinotermitidae and Termitidae, in the mouth area, or the abdomen, salivary glands and cibarial glands. The chemicals exuded can have different functions like keeping a wound from healing ( lipids ), causing irritation (irritants), causing toxicity (contact poisons), and acting like an entangling agent (glues) (Prestwich 1984). Some species of termites however do not have soldiers. T hese termitids (subfamily Apicotermitinae) have lost their soldiers and developed other forms of defense, like autothysis, which is the rupturing of the abdomen when in contact with a predator to exude their sti cky gut contents (Sands 1982). The remaining castes of workers, pseudergates and immature reproductives are usually light in color, with no specialized heads or mandibles. The immature reproductives (brachypterous nymphs) may have wing buds of different sizes. The pseudergates and the workers carry out all the work in the colony. The immature reproductives (nymphs) may also help with this (Edwards and Mill 1986) Work in the colony consists of taking care of the eggs and larvae and moving them when the nest is disrupted, foraging for food, feeding the larvae and the soldiers, building tunnels and nest structures and excavating wood and soil Termites are ecologically important ( Bignell and Eggleton 2000), being a key decomposer in numerous ecosystems, with their role in improving soil quality increasingly emerging ( Holt and Lepage 2000 ). Termites, however, can be important pests to crop and timber. Of all the species of termites described ( only 183 are

PAGE 22

22 known to cause damage to buildings with 83 species causing severe damage (Edwards and Mill 1986) The pest status of termites is recognized worldwide, and in the United States alone, the cost for termite control has been estimated to be between $1.02 and 1.5 billion (Edwards and Mill 1986 Su 1994 ). As f or Europe, in France alone the estimated cost for termite treatment is about 200 million euros per year (Bagnres et al. personal communication). Drywood termites are responsible for about 20% of the budget spent on termite control in the United States (Su and Scheffrahn 1990). The West Indian drywood termite (Figure 1 1) Cryptotermes brevis (Walker) is a drywood termite of the family Kalotermitidae. It was first described in Jamaica ( Walker 1853) and, with the exception of Asia, has a tropicopolit an distribution (Scheffrahn et al. 2008) (Figure 1 2 ) Altho ugh drywood termites are a widespread pest, they have certain temperature and moisture requirements (Edwards and Mill 1986). Cryptotermes brevis infests buildings and furniture, being mainly repor ted in the tropical and subtropical areas with some isolated occurrences in warmer temperate regions (Light 1934b; Edwards and Mill 1986). It is endemic to Chile and Peru where it occurs in nature, away from structural wood (Scheffrahn et al. 2008). Figure 11. Picture of Cryptotermes brevis showing soldiers, reproductives and pseudergates (adapted from Scheffrahn) Soldier Reproductive Pseudergates

PAGE 23

23 Cryptotermes brevis was found infesting structures in Portugal by Mateus and Goes (1953) in the Atlantic Island of Madeira. Later it w as reported in Mozambique (Carval ho 1972), which was then a Portuguese colony. For decades this species had n ot been reported anywhere else on Portuguese territory until it was identified as the cause of structural damage in the Azorean Island of Terceira in the year 2000 (Borges et al. 2004, Myles 2004). Since then, it has been further reported in the Islands of Faial, So Miguel Santa Maria (Borges 2007), So Jorge (personal observation), and Pico Island ( O. Guerreiro personal observation) all part of t he Azorean archipelago (Figure 1 3) The Azores is an archipelago of nine main islands located in the Atlantic Ocean about 1500 km from Li sbon, Portugal, and 3900 km from the East Coast of the United States of America. The archipelago was discovered in 1427 by the Portuguese and the first settlements on this uninhabited archipelago started in 1439. Due to its strategic position the Azores has been a stopping point for ships, yachts and airplanes which make it vulnerable to invasive pests. The Azores is at temperate latitude but due to sea currents enjoys a subtropical climate without extremes of either very hot or freezing temperatures. Because of the surrounding ocean the humidity is always high and therefore the climate is extremely favorable for termi tes, either of temperate or tropical origin. There are three more species of introduced termites know in the Azores that have been recently identified. One of these species is another Kalotermitidae that is common in the Mediterranean area of Europe, Kalot ermes flavicollis (Fabr.). A dampwood termite, it has been found in living trees in three of the islands in the Azores ( Terceira, So Miguel, and Faial), with few occurrences in wood structures (Myles et al. 2007a). The two other species that have been ide ntified in the Azores are both

PAGE 24

24 r hinotermitids of the genus Reticulitermes One species has been reported in the island of Faial where it was found infesting a coastal neighborhood in 2006 and it was identified as R grassei Clment a common termite in sou thern Europe, and was probably brought into the island from there. The sec ond is R. flavipes (Kollar) which was identified in an American Military a ir base in Terceira island about 20 years ago (Myles et al. 2007a), and more recently in the area surrounding the a ir base (personal observation). Reticulitermes flavipes is a common pest in the United States, and there was much traffic of goods from the U.S. into the Military base. Because of the location of the infestation it is believed that these termites may have been introduced through military shipment at the Lages Air base The main urban pest at this moment in the Azores is C. brevis The level of infestation in the city of Angra do Herosmo on the island of Terceira is very high, with about 43% of the buildings in the historical center infested, and of these, about 50% show high levels of infestation (Borges et al. 2004, Guerreiro et al. 2009). The origin of the Azorean population of this species is unknown, but was presumed to be either from the importation of infested wood, or dispersal flights from infested boats in Azorean harbors or both (Scheffrahn personal observation) Cryptotermes brevis is the most destructive drywood termite pest in the world. Because of its ability to withstand wood with low moisture content it is able to attack all kinds of wood in service including structural timbers, beams, studs, flooring, molding, doors, window frames and even wooden articles such as carvings, tools, picture frames, musical instruments, etc L ike other drywood termites, C. brevis is a cryptic species which nests in its food source, wood, spending almost all of its life cycle within the confines of wood. A colony of drywood termites can vary in number from hundreds to a

PAGE 25

25 few thousand termites (Nutting 1970) and several colonies can be found inside a single piece of wood. A C. brevis colony is estimated to have anywhere from 2 individuals (incipient colony) to 296 individuals with an average of 45 i ndividuals (Myles et al 2007a). The early development of a kalotermitid colony is slow and numbers are small in the first year (Nutting1969). For C. brevis, the number of eggs in an incipient colony is very low with an average of 4 eggs laid per pair of female and male dealates in the first month (Ferreira 2008). I n the first year there will be an average of 3.4 nymphs and no soldiers (McMahan, 1960). These low numbers indicate that many colonies of C. brevis can co habit a single piece of wood. Williams (1977) suggests that a colony of C. brevis will gradually move from highly infested timber into timber that is less than 30% damaged if it can do so, although it maintains activity in more damaged timbers. A majority of reports describe aggressive responses in nonkalotermitids when encounters between colonies of the same species occur. Violent agonistic responses, such as attack and dismemberment of intruders by workers have be en reported for several termite species (Thorne and Haverty 1991). However colony fusion has been observed for termite species like R eticulite rmes speratus (Kolbe) where members of a host colony accept a lower nymph ratio colony not showing agonistic behavior towards it (Matsuura and Nishida 2001). This is an important factor for invasive social species, because reduced genetic variability aft er an introduction can lower the intercolonial ag onism making possible for a larger unicolonial society to become dominant (Tsutsui and Suarez 2003).

PAGE 26

26 Figure 12. Map of the distribution of C. brevis in the world. (Adapted from Scheffrahn et al. 2008). Figure 13. Confirmed locations of C. brevis infestations in the Azorean Archipelago. Yellow dots mark the locations of the know n infestation sites.

PAGE 27

27 The life cycle of C. brevis begins with a dispersal flight where the alates leave their previous gallery system in order to form new colonies. The dispersal flights of termites occur at different times of the year depending on the species. For C. brevis the main flight season in South Florida occurs between April and July, with a secondary smaller flight sea son in November (Minnick 1973 ) In the Azores the flight season occurs between May and September (Borges et al. 2004) The dispersal flights are the only occasion that this species is found outside of wood (Kofoid 1934). Otherwise, it never leaves the nest to exploit new food sources (Korb and Katrantzis 2004) although they will bridge small gaps between wood members The new colony does not produce alates for about 5 years, at which point the colony is considered mature. The size of a colony and its maturi ty is important in determining the number of alates to be produced by a colony. It is known that a colony of C. brevis can survive with as few as four pseudergates and that this species can produce more neotenics than other Cryptotermes species (Williams a nd Morales 1980). The percentage of alates that leave the colony for some species has been determined, e.g : Pterotermes occidentis (Walker) 30%; Paraneotermes simplicicornis (Banks) 14 16%; Zootermopsis laticeps (Banks) 39%; Stolotermes vic toriensis Hill, 17.5%; and Nasutitermes exitiosus (Hill) 2.4% (Nutting 1969). Myles et al. (2007a) estimated that approximately 25% of C. brevis colony matures into alates and leav e the colony based on alates and alate nymphs present in active colonies inside wood. One of the main methods that have been used for controll ing this pest has been fumigation (Bess and Ota 1960). Sulfuryl fluoride eliminates existing colonies very effectively with no termites detected in a structure up to two years after treatment

PAGE 28

28 (Minnic k et al. 1972). However in places like the Azores, structural fumigation is not an option for remedial treatment of C. brevis due to logistical and European safety rules where there is a 10 m exclusion zone. In addition, fumigation does not prevent new in festations from occurring, and a residual termiticide should be applied for preventing new colony formation. Detection is a very important aspect of the local control of drywood termites. Due to their cryptic nature however, detecting exact locations for drywood termite colonies is difficult. The most used type of detection is by visual inspection (Scheffrahn et al. 1993). This involves either finding alates inside structures during the dispersal flight season, which provides evidence that there is an inf estation somewhere in the structure, finding frass near wood objects or structural wood, or tapping the wood surface and listening for hollow areas, or finding thin surfaces over galleries and kick out holes These methods are only reliable to identify an infestation when the colony is already well established. For the first years of the colony, there are few fecal pellets and this allows it to grow and consume the wood without being detected. Some of the other methods used for detection include the use of dogs to detect the s c ent of termites (Lewis et al. 1997). Dogs can be 97.3% accurate in detecting drywood termites in structures (Brooks et a l. 2003). Studies performed on acoustic emissions d etectors (AED) demonstrated that this is a reliable non destruct ive method for detecting drywood termites (Fuji et al. 1990, Scheffrahn et al. 1993, Lemaster et al. 1997, and Lewis et al. 2004). An AED can detect, depending on the position, as far as 8 to 80 cm inside the wood (Scheffrahn et al. 1993). This method is e ffective since wood structures are composed of wood that is not very thick. Some other detection methods involve

PAGE 29

29 infrared, laser x ray, and microwave technologies (Evans 2002, Grossman 2006, Lewis et al. 2005, Mankin 2004). These detection methods are imp ortant for remedial control in order to assess the extent of the existing infestation as well as the best methodology to apply. They are also important when using spot treatments, to help guide where the treatment should be applied. For remedial control of termites, there are several types of treatments which can be categorized into whole structure treatments compartmental or localized treatments. Ano ther method used for control for whole structure (besides fumigation) is heat treatment. Heating of struct ures has been shown to produce 96% mortality (Lewis and Haverty 1996). Field studies have shown that reaching a mean core temperature in the wood of a targeted 49C can take as little as 5h (Woodrow and Grace 1998a). Commercial heat operators use exposures of 54.4C for an hour as a standard for effective control (Quarles 2006). The localized treatments used against drywood termites can be of two kinds, either chemical or nonchemical. In the chemical treatments, there are several insecticides that are cur rently used. The most common method for applying localized chemicals is by drill and injection. In this method, the pest control operator drill s holes into the infested wood, and upon finding a gallery will inject the chosen chemical into the termites gal leries (Hickin 1971) or inject into kickout holes Some of the presently used chemicals in the United States are disodium octoborate t etrahydrate (DOT), imidacloprid, and fipronil (liquid, foam, and dry formulation), and dlimonene. All of these insecti cides can produce 100% mortality on contact ( Scheffrahn et al. 1997a, Myles et al. 2007b). In Europe, other insecticides used include cypermethrin, permethrin, and

PAGE 30

30 borates. A chemical being used currently in drywood termite control in California is dlimon ene, an essential oil extract from citrus However, t his active ingredient (AI) has shown a high repellency to termites (Myles et al. 2007b), which is a nondesirable characteristic if termites are to be exposed to it Another chemical being currently inve stigated against termites is chlorantraniliprole. This chemical is non repellent to subterranean termites (Yeoh and Lee 2007) as well as having a 90% mortality rate when R hesperus were exposed to it for an hour (Rust 2008, personal communication). This c hemical has shown high toxicity to R flavipes when applied topically (Spomer et al. 2009). Ano ther class of chemicals that has been tested against termites is juvenile hormone analogues like hydroprene. When tested against subterranean termites, hydroprene showed a low mortality rate (50%) after 14 days, but a high production of soldiers which is a non reproductive caste. This might affect the long term survival of a colony (Hrd et al. 2006). Chitin synthesis inhibitors are also used against termites. Hex aflumuron was tested against R. flavipes and Coptotermes formosanus Shiraki, demonstrating nondeterrence and a 90% mortality at concentration s of 15.6 62.5 ppm, and also nondeterrent making it a good choice for baiting subterranean termites (Su and Schef frahn 1998). N onchemical localized treatments against drywood termites include electrocution microwaves, heat, cold, carbon dioxide and entomopathogenic fungi. Electrical shock treatment has been shown to produce almost 100% mortality in laboratory test s (Creffield et al. 1997) 4 weeks after the electrical shock treatment (Lewis and Haverty 1996). However it is not considered to be an effective method for i nfested boards in houses, because it causes scorching and the number of holes drilled into wood can be

PAGE 31

31 up to 70% of a board (Lewis and Haverty 2001). Laboratory studies with microwave ovens revealed a mortality of over 84% (Lewis et al. 2000) although there is great variability in mortality across different studies (Lewis and Haverty 1996). For heat tre atments, Scheffrahn et al. (1997b ) showed that C. brevis has high mortality when exposed to temperatures above 50C for at least 15 min. Heat may be applied directly in infestation sites to increase the temperature of localized spots preventing an increase of temperature more than necessary and decreasing the danger of damage to structures by heat (Woodrow and G race 1998b). The use of liquid nitrogen as a cold treatment which is injected inside wall voids in structures can produce 100% mortality (Lewis and Haverty 1996). However, close monitoring must be done because i nsu lation materials such as fiberglass may deflect the n itrogen and prevent termites from being exposed to the freezing temperatures (Rust et al. 1997). The exposure to carbon dioxide in chambers has shown a high mortality for drywood termites (Borges et al. 2007). Finally, the use of entomopathogenic fungi as a nonchemical treatment against termites has shown high mortality, above 70%, when termites were exposed to two strains of fungi Metarh izium anisopliae and Verticillium indicum (Nars and Moein 1997) in the laboratory Although the contact mortality of termites is high, the indirect application of fungi to termites is not as effective. Untreated termites placed in arenas with treated termi tes did not produce an increase in mortality (Chouvenc et al. 2008). Preventative treatments may be used to reduce or prevent alates from flying into a structure and forming a new colony. Prevention of colonization can be of two types, either physical or chemical Physi cal prevention can be of various types, from traps for alates to resistant materials used in construction, mosquito screening, caulking, and

PAGE 32

32 other exclusion treatments These resistant materials may or may not be wood. Some types of wood are more resistant to termite attack than others and are not heavily consumed by drywood termites (Wolcott 1924, Williams 1934, Minnick et al. 1973, Gonalves and Oliveira 2006, Ferreira et al. 2007). Alates are attracted to light when flying (positive photo taxis) and the more intense light attracts a higher number of alates (Ferreira and Scheffrahn 2011). Light traps can be optimized by using high intensity lights to capture alates before they enter wood. Insects can be attracted to different wavelengths of light and using a preferred wavelength can bring a higher efficacy to a light trap. Blacklight lamps have been effectively used for attraction of insects for many years (Nabli et al. 1999). Minnick (1973) reported differences in wavelength of light preference by C. brevis He compared sunlight with U V light and incandescent light, and found that sunlight was most preferred followed by U V light and incandescent light. However no work on exact wavelength preference has been reported until now Insecti cides have long been used to prevent infestations by termites. Wood pressuretreated with preservatives increased the durability against termites (Randall and Doody 1934), and also preservatives can deter drywood termite infestations (Hunt 1959). Inert dus t materials that have been sprayed on wood as preventatives have also had some positive results (Wagner and Eb e ling 1959). As colony foundation preventatives, insecticides like disodium octoborate tetrahydrate and chromium copper arsenate when used as a su rface treatment are an unpreferred spot for colony establishment (Scheffrahn et al. 1998). Other chemicals that have been tested as colony prevention treatments were imidacloprid, silica gel and fipronil (Scheffrahn et al

PAGE 33

33 2006). These were shown to be e ffective and their use is recommended in wood structures, either in construction or applied to the wood surfaces that are exposed to infestation (Scheffrahn et al 2001). However these insecticides are not commercially available in Europe, and for insectic ides like cypermethrin and permethrin which are available, no tests against drywood termite colonization have been performed. Objectives Resolving the question of the origin of C. brevis in the archipelago, whether there was a single or multiple introduc tions can help make the case for new and stricter laws for handling import of wood, wood materials and maritime boat regulations Also by comparing the Azorean populations with populations from other countries and the putative countries of ori gin, this s tudy can help resolv e the quest ion about the geographic origin of this species in the Azores In order to apply an effective method of reducing dispersal during the flight season of C. brevis it is important to understand what happens during these flights Finding the percentage of the colony that actually leaves during dispersal flights can help understand how effective a method of control for the alates can be. If a very large percentage of the colony becomes alates and leaves it, then improving control methods can help reduce the numbers of termites infesting wood. Improvement of some control methods, like the use of different chemicals as preventative treatments and the use of light traps can be viable alternatives for use in Europe where chemical usag e is very restricted. The general aim of this dissertation is to study the origin, spread, and control of the West Indian drywood termite C brevis with application to the Azores Islands by testing the following hypotheses:

PAGE 34

34 Hypothesis 1 The introduction of C. brevis to the Azorean archipelago is widely spread, with six of the nine main islands being infested This suggests that there might have been multiple introductions of the species to the Archipelago. In order to test this hypothesis a genetic study will be performed comparing the different subpopulations of the islands Hypothesis 2 The origin of the introduction of C. brevis to the Azorean archipelago is unknown, and due to the position of the archipelago the introduction of this species might hav e come from various sources, including the American Continent (North, Central, an d South America), Africa, or even as far as Oceania. In order to test this hypothesis a genetic study will be performed comparing the island subpopulations to subpopulations f rom various countries. Hypothesis 3 Dispersal flights are the only time C. brevis is found outside of wood, with part of the colony leaving the wood as alates, and the remainder staying behind to continue the colony. After leaving the colony, the alates will fly in search of new places to colonize. During this time preventative treatments are important in reducing the risk of infestation. T he hypothesis being tested is that some light wavelengths are more attractive to alates tha n others for use in light traps And a lso insecticides can be used to prevent alates from establishing a new colony in wood.

PAGE 35

35 CHAPTER 2 POPULATION DIVERSITY OF C ryptotermes brevis IN THE AZORES: SINGL E VS MULTIPLE INTRODUCTIO NS The use of genetic markers in population analysis is becoming an established method for inferring relationships amongst different populations, ancestry, and relatedness. Mendels work showed an individual organism possesses both a heritable form of a gene (genotype), as well as individual variants (alleles) Population genetics is the study of the occurrence of alleles within and between populations. In population genetics, a population is considered a group of organisms of the same species living within a sufficiently restricted geographical area so that a ny member can potentially mate with any other member of the opposite sex (Hartl and Clark 1997) There are techniques used to detect genetic variability, each with some advantages and disadvantages. The study of mitochondrial DNA is widely used and it pro vides sufficient variation for studies at the individual, population, and species level, depending on the region of the mtDNA that is analyzed. The reasons for the adoption of mtDNA as a marker are that experimentally, mtDNA is relatively easy to amplify b ecause it appears in multiple copies in the cell. Mitochondrial gene s are maternally inherited, and are strongly conserved across animals Mitochondrial DNA is highly variable in natural populations because of its elevated mutation rate, which can generate conclusive data about population history over short time frames. Variable regions (e.g. the control region) are typically flanked by highly conserved ones (e.g. ribosomal DNA), in which PCR primers can be designed (Galtier et al. 2009) However, when look ing at the population level one genet i c marker is usually not enough. The variability may not be sufficient to show differences at a population level. And so, mitochondrial DNA studies of populations often involve the use of another genetic

PAGE 36

36 marker, micro satellites that have a high mutation rate and can show variability within a species Microsatellites are short segments of DNA that have repetitive sequences and are usually noncoding (Hamada et al. 1984) Over time, as animals in a subpopulation breed, t hey recombine their microsatellites during sexual reproduction and the subpopulation will maintain a variety of microsatellites that is characteristic for that subpopulation and distinct from other subpopulations which do not interbreed. For this reason the study of microsatellites is useful in subpopulation studies, and in the case of termites even at a colony level, to help understand the interactions between different colonies and possible interbreeding between closely located colonies (Husseneder et al. 2006) The Azorean Islands are a small ar chipelago in the Atlantic Ocean, where most towns have harbors with extensive marine traffic. It is therefore highly susceptible to invasive species introduction from organisms such as C br evis Presumed to be rec ently introduced to the islands, the question to whether there was a single introduction or multiple introductions of this species is still unresolved. For this study the genetic markers chosen to solve this question w ere two mtDNA markers, 16S rRNA and C ytb, and also microsatellites. Mitochondrial 16S rRNA was chosen because it is has clearly resolved many questions in termite phylogeny studies (Kambhampati et al. 1996, Legendre et al. 2008). Cytb gene has yield ed good results in Kalotermitid phylogenetic studies (Thompson et al. 2000), especially with the genus Cryptotermes (Legendre et al. 2008). The microsatellite loci used were amplified by primers developed for a cong eneric species C. secundus (Hill) (Fuchs et al. 2003). The objective of this study was to use these markers to determine if the infestation in the

PAGE 37

37 Azorean islands was caused by a single introduction, or if there were several introductions. Furthermore, in case there were multiple introductions, to determi ne if these are still occurring and if there is any inter island spread of the species. Material and Methods Termites Termites were collected during the month of August, 2009 in four of the nine Islands of the Azorean archipelago. The islands where the termites were collected were Terceira Island, in the city of Angra do Herosmo, So Miguel Island, in the city of Ponta Delgada, o n the Island of Faial, in the city of Horta, and o n the Island of Santa Maria, in the village of Maia (Figure 2 1) All termites were collected from infested struct ures. T o collect the termites access to private houses was received, and alates were gathered from spider webs or homeowner s light traps. In a few locations, infested wood was recovered from private houses and was later chopped open to collect termite s. The termites were immediately processed into 95% nondenatured ethanol vials at room 1 shows the collecting sites for the islands, where there were three collecting sites for Terceira Island (Angra), three sites for So Miguel (P.D.), two sites for Faial (Horta), and one site for Santa Maria Island (Maia). Only alates were used and t he number varied between a minimum of 4 and a maximum of 20 per site (Table 21). Figure 21. Location of the towns where C. brevis was collected in the four islands.

PAGE 38

38 Table 21. L ocations of collection sites in the Azores and number of individuals (N) used for analysis. Island Town Sample name N Latitude Longitude Terceira Angra do Herosmo AA 7 38.655475 27.219519 Terceira Angra do Herosmo AB 20 38.656608 27.228125 Terceira Angra do Herosmo AC 4 38.654694 27.226614 So Miguel Ponta Delgada PDA 7 37.739322 25.669872 So Miguel Ponta Delgada PDB 5 37.742244 25.665503 So Miguel Ponta Delgada PDC 6 37.742300 25.662792 Faial Horta HA 20 38.540844 28.625658 Faial Horta HB 15 38.540731 28.625842 Santa Maria Maia M 20 36.939256 25.014958 DNA extraction and amplification Mitochondrial DNA. DNA was extracted by using a DNeasy Blood and Tissue kit (Qiag en), following the protocol for ext racting DNA from animal tissue from the DNeasy Blood and Ti ssue Handbook (2006). After extraction, DNA was amplified using PCR. The primers used for mitochondrial 16S rRNA and Cytochrome b were the same primers used by Legendre et al. (2008) (Table B 1), because they previously successful for C. brevis A maximum of five different specimens per collection site was used for the mtDNA amplification (wit h the exception of Maia where ten individuals were used because there was only one point of collection) and individually sequenced. A PCR for temperature gradient was run to determine the best temperature for annealing both 16S rRNA and Cytb primers and the optimum PCR profile for 16S rRNA and Cytb

PAGE 39

39 are described in Table 22. The template (3 of DNA) w as loaded in to a thermo cycler ( DNA Engine Thermal Cycler) with a PCR Master Mix formulation (Promega M7502) ( l), and the primers (1.5 of each forward and reverse primer at a concentration of 10 M ) according to the gene that was to be am plified and diluted with water (5 After the DNA was ampli fied 5 l of the PCR products were run on a polyacrylamide gel (8%) along with an exACTGene 50bp Mini DNA Ladder (Fisher) to ensure the products obtained were the right size (Figure A 2) The g els were then stained with Ethidi um Bromide (0. 5 ) and observed under U. V light The remainder of each of the samples was cleaned with a Montge PCR Centrifugal filter device (Millipore, Montge ) following the protocol for Millipore. After being cleaned the samples w ere sent to the University of Florida DNA Sequencing Laboratory, where they w ere sequenced using an ABI 3 130 automated sequencer. Table 22. PCR Cycles for the mitochondrial 16S rRNA, and Cytb genes optimized for amplification in C. bre vis Gene Heat Denaturation Annealing Extension Final E xtension Cycles 16S rRNA 94C (2min) 94C (1min) 50C (1min) 72C (1min 15s) 72C (7min) 40 Cytb 94C (2min) 94C (1min) 56C (1min) 72C (1min 15s) 72C (7min) 40 Microsatellites. The same extr acted DNA was used for the microsatellite data analysis. Up to 20 specimens were used per collection site (Table 2 1). The primers used for microsatellite amplification were adapted from Fuchs et al. (2003) These primers were developed for C secundus bu t proved to successfully amplify five loci in C. brevis (Table B 2). For all five loci a temperature gradient was run to assess the optimal annealing temperature. For two of the loci (Csec 5 and 6) a n optimum

PAGE 40

40 consistent annealing temperature was found, usi ng the PCR profiles provided in the literature (Table 23). For these loci the same PCR Mast er Mix and quantities were used as for the mitochondrial gene amplification. However for the remaining loci (Csec1, 3, and 4), due to inconsistencies in results the Terra PCR Direct Polymerase Mix (Clontech Industries Inc. CAT: 639270) was used with the specific PCR profiles recommended by the manufacturer. For the Terra PCR, 25 l of buffer, 9 l of water, 5 l of both forward and reverse primer, 1 l of Taq, and 5 l of DNA were loaded into the thermocycler. The PCR products for all loci were loaded into a 12% gel ( Hoeffer, GE Healthcare) following the same protocol as described above, to determine if the amplified products were in the right size range (Figure A 3) All the forward primers for the microsatellite were tagged with a fluorescent tag. Loci Csec 3, 4, 5, and 6 were tagged with FAM (Integrated DNA Technologies, Inc.), and Csec1 was tagged with HEX (Integrated DNA Technologies, Inc.) Loci Csec1 and Cs ec 4 were multiplexed. All PCR products were sent to the University of Florida Sequencing Laboratory for genotyping. Table 23. PCR profiles used for all the C. brevis microsatellite loci. Locus Heat Denaturation Annealing Extension Final extension Cycles Csec1 98C (2min) 98C (10s) 60C (15s) 68C (30s) 68C (7min) 40 Csec3 98C (2min) 98C (10s) 60C (15s) 68C (30s) 68C (7min) 40 Csec4 98C (2min) 98C (10s) 60C (15s) 68C (30s) 68C (7min) 40 Csec5 95C (1min3 0s) 95C (45s) 57C (45s) 72C (45s) 72C (7min) 35 Csec6 95C (1min 30s) 95C (45s) 57C (45s) 72C (45s) 72C (7min) 35

PAGE 41

41 Data analysis Mitochondrial DNA Phylogenetic and molecular evolutionary analyses for the 16S rRNA gene and Cytb gene were conducted using MEGA ver sion 5 (Tamura et al. 2011). Sequences were aligned in MEGA version 5 (Figure A 4) using the ClustalW method set at default, as recommended by the software. The sequences from 16S rRNA and Cytb were combined in order to create more robust sequence data. After the sequences were aligned they were run through the Gblocks software (Talavera and Castresana 2007) to eliminate poorly aligned positions divergent regions of the alignment and insertions and deletions D istance matrixes for number of nucleotides were created for all the islands The distance pvalue was then calculated to test if the data from the sequences were good. Similar sequences were eliminated from the analysis, and a representative sequence was chosen to represent those sequences. The re presentative sequences wer e then analyzed using the UPGMA method (Sneath and Sokal 1973). The bootstrap consensus tree was inferred from 5000 replicates (Felsenstein 1985). The evolutionary distances were computed using the Kimura 2parameter method (Kimura 1980) and are based on t he number of base substitutions per site. Sequences from K alotermes flavicollis for both genes were used as out groups. These representative sequences were obtained from GenBank by using BLAST for each gene. For mitochondrial 16S rRNA gene the K. flavicollis sequence used was reported in Szalanski et al. (2004) and identified by accession number AY486437. For Cytb gene, the K. flavicollis sequence used was the one reported in Legendre et al. (2008) and identified with the accession number EU253919. Haplotyp es for the individual termites were calculated using TCSv1.21 (Clement et al. 2000) and a tree was created showing the genealogical relationships between haplotypes. The prog ram

PAGE 42

42 ARLEQUIN 3.5.1.2 (Excoffier and Lischer 2010) was used to calculate the AMOVA (Analysis of Molecular Variance) between haplotypes of the different subpopulations. Microsatellites For the microsatellite analysis the Peak Scanner S oftware v1.0 ( Copyright 2006, Applied Biosystems) was used to score the microsatellites (Figure A 5 ) The EXCEL Microsatellite toolkit (Park 2001) was used to estimate the number and frequency of alleles as well as observed and expected heterozygosity An ANOVA was used to test if there were significant differences between expected and observed heterozy gosities and allele frequencies between populations using SAS (2003). FSTAT 2.9.3.2 (Goudet, 2001) and GENEPOPV4 ( F. Rousset) w ere used to calculate FST, FIS, and FIT to quan tify the respective loss in heterozygosity linkage disequilibrium, and the frequency of null alleles. They were also used to estimate the modified Wrights F statistics by Weir and Cockerhams (1984) ST, Neis (1987) GST, and Michalakis and Excoffiers (1996) ST The gene flow was quantified in terms of migrants exchanged among su bpopulations per generation (Nm), calculated using GENEPOPV4 ( F. Rousset). Genetic distance was calculated using Neis (1972) method and a genetic di stance matrix and tree was generated using Genetic Distance Analysis 1.1 (Lewis and Zaykin 2001) To test different scenarios of introduction the program DIYABC (Cornuet et al. 2010) was used. The program uses approximate Bayesian computations to infer origins considering bottleneck effects when introduction occurs The assumptions used for the models were that there was an ancestral population of size NA from where the introduced subpopulations originated. The size of the ancestral population was assumed to be higher than that of the introduced subpopulations N1 N4 (Figures 22 to 2 7) For the computations the subpopulations of each Island were considered as one

PAGE 43

43 single subpopulation per Island. T he data of subpopulations AA, AB, and AC were merged into Population 1 (default nomenclature in the program), subpopulations PDA, PDB, PDC were merged into Population 2, subpopulations HA and HB were merged into Population 3, and the M subpopulation was represented as Populati on 4. A generalized mutation model (GSM) was assumed for the microsatellite loci, with a gamma distribution for the mutation rate bounded betw een 5x104 and 5x103. The reference table was created for 50,000 simulations. The posterior probability of each scenario was assessed by a polychotomous weighted logistic regression that estimates the difference between simulated (from the reference table) and observed data sets. The number of observed data sets used was 500 and the number of simulated data used was 5,000. Six scenarios of introduction were tested with no admixture assumed. In scenario 1 Population 1 (Angra) and Population 2 (Ponta Delgada) were assumed to be part of an ancestral common population of size NA, and were then separately introduced to the Islands at an undetermined time (td). Population 3 (Hor ta) was assumed to split from Population 1 at a later time (ta), while Population 4 (Maia) split from Population 2 (Figure 22). In scenario 2, Populations 1 and 2 were assumed to be part of an ancestral common population of size NA, and were then separately introduced to the Islands at an undetermined time (td). Populations 3 and 4 were assumed to have both split from Population 1 at a later time (ta) making Population 1 the common ancestral population for Populations 3 and 4 (Figure 23).

PAGE 44

44 Figure 22. Scenario 1 of introduction to the Islands tested with the DIYABC software. The program assumed an effective random population size for each population N1, N2, N3, and N4. Computations were made based on the samples from each population (Sa 14). The splitting of populations was assumed to happen at a determined time for each event, t d and ta, with the condition that td Each Population correspond ed to a different Island with the different sites per Island merged into one single Population. Population 1 was the subpopulation of th e city of Angra, Population 2 was the subpopulation of the city of Ponta Delgada, Populat ion 3 is the subpopulation of the City of Horta, and Population 4 wa s the subpopulation of the city of Maia. The scenario assumed that populations 1 and 2 were part of one N4, and that subsequently population 3 split from population 1, while population 4 split from population 2. Horta Angra Maia Ponta Delgada

PAGE 45

45 Figure 23. Scenario 2 of introduction to the Islands tested with the DIYABC soft ware. The program assumed an effective random population size for each population N1, N2, N3, and N4. Computations were made based on the samples from each population (Sa 14). Each Population corresponded to a different Island with the different sites per Island merged into one s ingle Population. Population 1 was the subpopulation of t h e city of Angra, Population 2 was the subpopulation of the city of Ponta Delgada, Population 3 was the subpopulation of the Ci ty of Horta, and Population 4 was the subpopulation of the city of Maia. T he splitting of populations was assumed to happen at a determined time for each event, td and ta, with the condition t hat that populations 1 and 2 were part of one N4, and that subsequently Population 3 and Population 4 split from P opulation 1. Horta Ponta Delgada Angra Maia

PAGE 46

46 In scenario 3, Populations 1 and 2 we re assumed to be part of an ancestral common population of size NA, and were then separately introduced to the Islands at an undetermined time (td). Populations 3 and 4 were assumed to have both split from Population 2 at a later time (ta), making Populati on 2 the common ancestral population for populations 3 and 4 (Figure 24). Figure 24. Scenario 3 of introduction to the Islands tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N3, and N4. Computations were made based on the samples from each population (Sa 14). Each Population corresponded to a different Island with the different sites per Island merged into one s ingle Population. Population 1 was the subpopulation of the city of Angra, Population 2 wa s the subpopulation of the city of Ponta Delgada, Population 3 wa s the subpopulation of the City of Horta, and Population 4 was the subpopulation of the city of Maia. T he splitting of populations was assumed to happen at a determined time for each event, td and ta, with the condition t hat that populations 1 and 2 were part of one N4, and that subsequently Population 3 and Population 4 split from Population 2. Angra Ponta Delgada Maia Horta

PAGE 47

47 In scenario 4, all p opulations were assumed to have come from the same ancestral common p opulation of size NA, and were subsequently introduced from that one population into the different islands at an undetermined time (td). In this scenari o the different subpopulations of the islands are considered to have been introduced independently from each other (Figure 25). Figure 25. Scenario 4 of introduction to the Islands tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N3, and N4. Computations were made based on the samples from each population (Sa 14). Each Population corresponded to a different Island with the different sites per Island merged into one single Populati on. Popula tion 1 was the subpopulation of th e city of Angra, Population 2 was the subpopulation of the city of Ponta Delgada, Population 3 was the subpopulation of the Ci ty of Horta, and Population 4 was the subpopulation of the city of Maia. The splitting of populations was assumed to happen at a determine d time, td. The scenario assumed that populations 1, 2, 3, and 4 were all part of a single ancestral population of size NA with the N4, and were all introduced, from that ancestral population, separately to the islands. Maia Ponta Delgada Angra Horta

PAGE 48

48 In scenario 5, Population 1 and 2 were assumed to be part of an ancestral common population of size NA, and were then separately introduced to the Islands at an undetermined time (td). Population 3 was assumed to have s plit from Population 2 at time (ta) and subsequently Population 4 was assumed to have split from Population 3 with Population 1 being completely independent from the other populations (Figure 26). Figure 26. Scenario 5 of introduction to the Islands tested with the DIYABC softwar e. The program assumed an effective random population size for each population N1, N2, N3, and N4. Computations were made based on the samples from each population (Sa 14). Each Population corresponded to a different Island with the different sites per Island merged into one si ngle Population. Population 1 was the subpopulation of th e city of Angra, Population 2 was the subpopulation of the city of Ponta Delgada, Population 3 was the subpopulation of the Ci ty of Horta, and P opulation 4 was the subpopulation of the city of Maia. The splitting of populations was assumed to happen at a determined time for each event, td and ta, with the condition t hat that populations 1 and 2 were part of one N4, and that subsequently population 3 split from population 2 and population 4 split from population 3. Horta Ponta Delgada Angra Maia

PAGE 49

49 In scenario 6, Populations 1 and 2 were assumed to be part of an ancestral common population of size NA, and were then separately introduced to the Islands at an undetermined time (td). Population 3 was assumed to have split from Population 1 at time (ta) and Population 4 was assumed to have split from Population 3 with Population 2 being completely independent from the other populations (Figure 27) Figure 27. Scenario 6 of introduction to the Islands tested with the DIYA BC software. The program assumed an effect ive random population size for each population N1, N2, N3, and N4. Computations were made based on the samples from each population (Sa 14). Each Population corresponded to a different Island with the different sites per Island merged into one si ngle Population. Population 1 was the subpopulation of th e city of Angra, Population 2 was the subpopulation of the city of P onta Delgada, Population 3 was the subpopulation of the Ci ty of Horta, and Population 4 was the subpopulation of the city of Maia. The split ting of populations was assumed to happen at a determined time for each event, td and ta, with the condition that td that populations 1 and 2 were part of one ancestral population of size NA wiN4, and t hat subsequently population 3 split from population 1 and population 4 split from population 3. Horta Angra Ponta Delgada Maia

PAGE 50

50 Results Mitochondrial DNA A total of 49 samples of PCR product per gene were sent to the sequencing laboratory. Of these samples only 31 produced readable sequences for 16S rRNA, and 28 for the Cytb gene. Only complete sequences from both Cytb and 16S rRNA were used and combined with a total of 28 combined sequences for analysis. Of these sequences a total of 9 sequences for the combined 16S rRNA and Cytb were n ot used for the phylogenetic analysis because they were not different from other sequences as determined by the calculated distance matrix (Tables 2 4 to 2 6 ). For the remaining 19 combined sequences the pdistance was calculated at ere considered good for proceeding with the analysis. MEGA 5.0 produced a phylogenetic tree for the combined mtDNA gene s. For the combined sequences the UPGMA bootstrap consensus tree shows two distinct branches with 100% resolution (Figure 2 8 ). The subp opulations of Ponta Delgada (PD) and Maia (M) are placed together in one of the main branches. The subpopulation of Horta is placed in the other main branch. The subpopulation of Angra is placed in both of the br anches, with individuals in the branch that includes Horta and individuals in the branch that includes both Ponta Delgada and Maia. The analysis of the 28 combined sequences for haplotypes yielded a total of 22 different haplotypes (Table B 3 and 4) separated into two major groups that showed no gen ealogical steps between one another (Figure 29 ). O f the haplotypes 16 had one representative sequence while the remainder six, had two sequences with the same haplotype. The haplotypes of the Angra subpopulation show ed more distance from the haplotypes r epresented in the Ponta Delgada and Maia subpopulations than from the

PAGE 51

51 Horta subpopulations (Figure 29). The FST value for the AMOVA was 0.77576 with a pvalue of 0.0000.000 not being significantly different from 0. The percentage of variation among and b etween subpopulations is shown in Table 27 where th e variation among populations was above 77%. Table 24. Distance matrix of estimates of e volutionary d ivergence between the combined sequences of Cytb and 16S rRNA of C. brevis for the Azorean subpopulat ion s. The number of base differences between sequences is shown. The analysis involved 28 nucleotide sequences. AA 1 AA 3 AB 1 AB 2 AB 4 AB 5 AC 2 PDA 2 PDA 3 AA 1 AA 3 0 AB 1 40 39 AB 2 40 39 0 AB 4 6 6 39 39 AB 5 42 41 4 4 41 AC 2 41 40 3 3 40 1 PDA 2 37 36 8 8 38 13 12 PDA 3 44 43 2 2 43 7 6 9 PDA 5 42 40 5 5 42 10 9 5 6 PDB 1 37 36 8 8 38 13 12 2 11 PDB 2 39 38 4 4 40 9 8 4 7 PDC 1 40 39 4 4 41 9 8 5 7 HA 1 2 2 39 39 6 41 40 38 43 HA 2 2 2 4 1 41 8 43 42 38 45 HA 3 3 3 40 40 7 42 41 39 44 HA 4 1 1 38 38 5 40 39 37 42 HA 5 0 0 40 40 6 42 41 37 44 HB 1 1 1 38 38 5 40 39 37 42 HB 2 2 2 39 39 6 41 40 38 43 HB 3 2 2 39 39 6 41 40 38 43 HB 4 2 2 42 42 8 44 43 37 44 HB 5 0 0 39 39 6 41 40 36 43 M 1 41 40 6 6 42 11 10 6 9 M 2 39 38 4 4 40 9 8 4 7 M 5 39 38 5 5 40 10 9 4 8 M 7 39 38 4 4 40 9 8 4 7 M 9 42 41 7 7 43 12 11 7 10

PAGE 52

52 Table 25. Continuation of distance matrix of estimates of e volutionary d ivergence between the combined sequences o f Cytb and 16S rRNA of C. brevis for the Azorean subpopulation s. The number of base differences between sequences is shown. The analysis involved 28 nucleotide sequences. PDA 5 PDB 1 PDB 2 PDC 1 HA 1 HA 2 HA 3 HA 4 HA 5 PDB 1 5 PDB 2 1 4 PDC 1 2 5 1 HA 1 42 38 40 41 HA 2 42 38 40 41 4 HA 3 43 39 41 42 3 5 HA 4 41 37 39 40 1 3 2 HA 5 41 37 39 40 2 2 3 1 HB 1 41 37 39 40 1 3 2 0 1 HB 2 42 38 40 41 2 3 3 1 2 HB 3 42 38 40 41 2 3 3 1 2 HB 4 43 39 41 42 4 2 5 3 2 HB 5 40 36 38 39 2 2 3 1 0 M 1 3 6 2 3 42 42 43 41 41 M 2 1 4 0 1 40 40 41 39 39 M 5 1 4 0 1 40 40 41 39 39 M 7 1 4 0 1 40 40 41 39 39 M 9 4 7 3 4 43 41 44 42 42 Table 26. Continuation of distance matrix of estimates of e volutionary d ivergence b etween the combined sequences of Cytb and 16S rRNA of C. brevis for the Azorean subpopulation s. The number of base differences per sequence from between sequences are shown. The analysis involved 28 nucleotide sequences. There were a total of 707 positions in the final dataset. HB 1 HB 2 HB 3 HB 4 HB 5 M 1 M 2 M 5 M 7 M 9 HB 2 1 HB 3 1 0 HB 4 3 4 4 HB 5 1 2 2 2 M 1 41 42 42 43 40 M 2 39 40 40 41 38 2 M 5 39 40 40 41 38 2 0 M 7 39 40 40 41 38 2 0 0 M 9 42 43 43 42 41 5 3 3 3

PAGE 53

53 Figure 28 Evolutionary relationships of C. brevis subpopulations. The evolutionary history was inf erred using the UPGMA method. The bootstrap consensus tree inferred from 5000 replicates is taken to represent the evolutionary history of the s ubpopulations analyzed. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates are collapsed. The percentage of replicate trees in which the associated subpopulations clustered t ogether in the bootstrap test are shown next to the branches The evolutionary distances were computed using the Kimura 2parameter method and are in the units of the number of base substitutions per site. The analysis involved 20 nucleotide sequences. Cod on positions included were 1st+2nd+3rd+Noncoding. All ambiguous positions were removed for each sequence pair. There were a total of 701 positions in the final dataset. PDB 2 M 2 PDA 5 PDB 1 PDA 2 M 1 M 9 AB 5 AC 2 AB 2 PDA 3 AB 4 HA 3 HA 2 HB 4 AA 3 HA 1 HA 4 HB 2 Kalotermes flavicollis 76 100 95 86 58 55 100

PAGE 54

54 Table 27. Summary statistics for the AMOVA among and within subpopulations of the diff erent Islands. Haplotype data was used to determine the percentage of variation in the Arlequin 3.5 software. Source of variation d.f. Sum of squares Variance components Percentage of variation Among populations 8 246.921 9.35878 77.58 Within populations 19 51.4 2.70526 22.42 Total 27 298.321 12.06405 Figure 29 Genealogical relationships among 22 haplotypes of C. brevis estimated by TCS. Haplotypes in rectangular boxes were randomly chosen by software to root the tree Length of branches indicates number of changes in haplotype. Number in parenthesis represents the number of individuals sharing the haplotype.

PAGE 55

55 Microsatellites A total of 33 alleles were detected in the populations over the five loci. Alleles per locus f or all subpopulation s are given in T able 2 8 and allele frequencies per locus are given in Tables 29 to 2 1 3 The allele 234 for the locus Csec1 was unique for the AA subpopulation, and for the Csec3 locus the allele 134 was unique for the subpopulation HA, while the al lele 138 was unique for the subpopulation M. The loci Csec 1, 3 and 4 had the higher number of alleles for the Azorean subpopulations, with nine, ten and seven alleles, respectively. The Csec6 and Csec5 loci had three and four alleles respectively. The m ean number of alleles varied between 2.6 and 4. The frequencies of alleles varied between 0100% for the loc i (Table s 2 9 to 2 1 3 ). The allele frequencies were not significantly different between sub populations (F=0.11, pvalue=0.999). Table 28 Size of a lleles in number of base pairs per locus for the five microsatellite loci of C. brevis analyzed. Locus Alleles (bp) Csec6 143 146 149 Csec5 122 125 128 131 Csec4 129 144 148 154 157 166 169 Csec1 208 214 216 220 224 22 6 230 232 234 Csec3 116 118 120 122 126 128 130 132 134 138 Table 29. Allele frequencies in percentage for Csec6 locus for the subpopulations of C. brevis calculated by the Microsatellite tool kit for Excel. Allele AA AB AC PDA PDB PDC HA HB M 143 78.6 77.5 75.0 100.0 100.0 75.0 95.0 93.3 95.0 146 21.4 22.5 25.0 0.0 0.0 16.7 5.0 3.3 5.0 149 0.0 0.0 0.0 0.0 0.0 8.3 0.0 3.3 0.0 Table 210. Allele frequencies in percentage for Csec5 locus for the subpopulations of C.brevis calculated by the Microsatellite tool kit for Excel Allele AA AB AC PDA PDB PDC HA HB M 122 7.1 30.0 12.5 71.4 50.0 33.3 40.0 20.0 12.5 125 92.9 60.0 62.5 28.6 50.0 50.0 60.0 80.0 85.0 128 0.0 7.5 25.0 0.0 0.0 8.3 0.0 0.0 2.5 1 31 0.0 2.5 0.0 0.0 0.0 8.3 0.0 0.0 0.0

PAGE 56

56 Table 211. Allele frequencies in percentage for Csec4 locus for the subpopulations of C. brevis calculated by the Microsatellite tool kit for Excel Allele AA AB AC PDA PDB PDC HA HB M 129 1 4.3 17.5 12.5 78.6 40.0 0.0 0.0 0.0 5.0 144 0.0 0.0 0.0 14.3 20.0 0.0 0.0 0.0 0.0 148 0.0 0.0 0.0 0.0 20.0 0.0 2.5 0.0 0.0 154 21.4 20.0 37.5 7.1 0.0 8.3 0.0 0.0 0.0 157 64.3 55.0 50.0 0.0 10.0 25.0 2.5 13.3 0.0 166 0.0 2.5 0.0 0.0 10.0 16.7 57.5 50.0 72.5 169 0.0 5.0 0.0 0.0 0.0 50.0 37.5 36.7 22.5 Table 212. Allele frequencies in percentage for Csec1 locus for the subpopulations of C. brevis calculated by the Microsatellite tool kit for Excel Allele AA AB AC PDA PDB PDC HA H B M 208 50.0 87.5 87.5 50.0 40.0 75.0 62.5 66.7 62.5 214 0.0 2.5 0.0 0.0 0.0 0.0 2.5 0.0 0.0 216 0.0 7.5 0.0 0.0 10.0 0.0 0.0 0.0 0.0 220 0.0 0.0 0.0 7.1 20.0 0.0 2.5 0.0 2.5 224 0.0 0.0 0.0 42.9 30.0 0.0 0.0 0.0 2.5 226 14.3 0.0 12.5 0.0 0.0 8.3 22.5 10.0 7.5 230 7.1 0.0 0.0 0.0 0.0 8.3 10.0 16.7 15.0 232 14.3 2.5 0.0 0.0 0.0 8.3 0.0 6.7 10.0 234 14.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Table 213. Allele frequencies in percentage for Csec3 locus for the subpopulations of C. brevis calculated by the Microsatellite tool kit for Excel Allele AA AB AC PDA PDB PDC HA HB M 116 0.0 0.0 0.0 0.0 0.0 0.0 2.5 0.0 0.0 118 0.0 0.0 25.0 57.1 60.0 41.7 62.5 66.7 57.5 120 0.0 5.0 0.0 0.0 0.0 0.0 2.5 3.3 0.0 122 64.3 75.0 50.0 21 .4 20.0 41.7 12.5 16.7 17.5 126 0.0 0.0 0.0 7.1 20.0 0.0 12.5 10.0 15.0 128 14.3 7.5 0.0 0.0 0.0 0.0 0.0 3.3 0.0 130 0.0 5.0 25.0 14.3 0.0 16.7 0.0 0.0 5.0 132 21.4 7.5 0.0 0.0 0.0 0.0 5.0 0.0 2.5 134 0.0 0.0 0.0 0.0 0.0 0.0 2.5 0.0 0.0 138 0.0 0.0 0 .0 0.0 0.0 0.0 0.0 0.0 2.5

PAGE 57

57 Table 21 4 Estimates of genetic diversity at 5 polymorphic microsatellite loci in C. brevis from the Azorean Islands. Sub p op Mean Expected Hz Mean Observed Hz A NOVA 's F statistics p value for ANOVA Mean No of Alleles per locus F IS p value* AA 0.47 0.10 0.54 0.08 0.16 0.702 3 1.22 0.16 0.9613 AB 0.44 0.07 0.39 0.05 0.24 0.635 4 1.22 0.13 0.0681 AC 0.54 0.09 0.45 0.11 0.26 0.627 2.6 0.55 0.182 0.2132 PDA 0.42 0.11 0.37 0.08 0.07 0.802 2.6 1.1 4 0.114 0.2954 PDB 0.56 0.15 0.4 0.10 0.72 0.421 3 1.58 0.304 0.0394 PDC 0.59 0.06 0.57 0.09 0.04 0.849 3.6 0.55 0.05 0.4262 HA 0.46 0.09 0.43 0.05 0.02 0.883 4 2.12 0.059 0.2835 HB 0.43 0.09 0.48 0.06 0.14 0.715 3.4 1.14 0.124 0.9446 M 0.40 0.10 0.46 0.05 0.14 0.722 4 1.87 0.148 0.9915 *p value for F larger than the observed, in randomization tests. The mean observed heterozygos ity was not significantly different from the expected heterozygos ity (Table 214), for all t he subpopulations. The values of FIS were low, with some values being negative, and were not significantly different from 0 (Table 2 14). The Hardy Weinberg test for disequilibrium was significant (pvalue=0.0076 for Score U test) with linkage disequilibri um found at loci Csec4 and Csec6. The null hypothesis that there is no defici ency of heterozygosity was rejected. The p values of FST between subpopulations are presented in Table 215, with the significant values indicated. The Angra subpopulations were n ot significantly different between one another. O ne of the Angra subpopulations (AC) was not different from the Ponta Delgada subpopulation s. The subpopulations of Horta and Maia were not significantly different from each other and from th e PDC subpopulat ion (Table 215). The estimated frequency of null alleles f or the Csec4 locus varied between 0.07 and 0.4259, with null alleles present in seven of the nine subpopulations (Table 216 ). The Csec 1 loci had two subpopulations with null alleles, while the C sec 3, 5 and 6 loci had three subpopulations with null alleles. The calculated gene flow (Nm) was 3.77626

PAGE 58

58 migrants per generation. The calculated values of FST, FIS, and FIT for the different methods are presented in Table 21 7 The distance diagram calcul ated for the microsatellite data is shown in Figure 2 10. The subpopulations of Maia and Horta were shown to be closer to each other than to the other sub populations. Table 21 5 Matrix of FST p values between subpopulations of C. brevis for all microsate llite loci. p values obtained after 36000 permutations, Indicative adjusted nominal level (5%) for multiple comparisons is 0.001389. Significant values indicated by *. Sub pop AB AC PDA PDB PDC HA HB M AA 0.033 0.08 4 0.001 0.00 2 0.000* 0.000* 0.000* 0.0 00* AB 0.236 0.000 0.000* 0.00 1 0.000* 0.000* 0.000* AC 0.003 0.02 5 0.554 0.000* 0.000* 0.000* PDA 0.67 0.001* 0.000* 0.000* 0.000* PDB 0.002 0.000* 0.000* 0.000* PDC 0.015 0.138 0.00 7 HA 0.519 0.045 HB 0.52 4 T able 216. Estimated values for null allele frequencies per sub population, per locus for the C. brevis sub populations sequenced in this study Locus AA AB AC PDA PDB PDC HA HB M Csec1 0 .000 0 .000 0 .000 0 .000 0.188 0 .000 0 .000 0.05 1 0 .000 Csec3 0 .000 0 .00 0 0.107 0.072 0 .000 0 .000 0.039 0 .000 0.0 30 Csec4 0.07 6 0.07 0 0 .000 0.189 0.189 0.42 6 0 .000 0.2 30 0.149 Csec5 0 .000 0.01 7 0 .000 0 .000 0 .000 0.11 4 0.259 0 .000 0 .000 Csec6 0 .000 0.167 0.290 No inf No inf 0.251 0 .000 0 .000 0 .000 Table 21 7 Values of FST, FIS, and FIT calculated using three different methods for all loc i for all sub populations Weir & Cockerham (1984) Michalakis and Excoffier (1996) Nei's (1972) F ST 0.176 0.2453 0.147 F IS 0.015 0.1081 0.049 F IT 0.188 0.1637 0.162

PAGE 59

59 Figure 2 10. Distance diagram calculated using Genetic Distance Analysis. Length of branches represents genetic distance according to Neis identity method. The longer the branch compared to the scale, the larger the distance between the subpopulations All the subpopulations sampled from the Islands are represented. The scenarios tested with DIYABC showed a low probability of being the scenario of introduction of the subpopulations to the islands for all except for scenarios 3 and 5, which showed higher probability a fter 5,000 simulated data were added in to the logistic regression. S cenario 5 had a 50% probability while the Scenario 3 had a 40% probability of being the introduction scenario (Figure 211). Angra A 0.1 Angra B Angra C Ponta Delgada A Ponta Delgada B Ponta Delgada C Horta A Horta B Maia

PAGE 60

60 Figure 211. Logis tic regression of the simulated data probability for the different scenarios tested in DIYABC. Probability varies between 0 and 1. All scenarios are within a 95% confidence interval for the simulations. Discussion The phylogenetic tree obtained from the mtDNA showed a clear pattern where the combi ned genes for the sub populations were separated into two major branches. T he sub populations from Ponta Delgada and Maia were grouped in a closely related group. This is not surprising because the Islands of So Miguel and Santa Maria, where these towns are located respectively are geographically close (Figure 2 1) and there is a lot of traffic of goods and visitors between the m The Horta subpopulation showed a distinct separation from all the other subpopulations being only closer to one of the Angra samples (AB 4). The Angra subpopulations were distributed between the two main branches showing close relationships to both the Ponta Delgada and the Horta subpopulations. Based upon the mtDNA data alone that there is such a cleft between the two major branch es would suggest that there were multiple introductions of this species to the islands with at least two major introductions. Furthermore, it seems that Number of data simulated Probability

PAGE 61

61 the species was first introduced in the two major cities (Angra and Ponta Delgada) and from there s pre ad out to the smaller towns in the other islands The genealogical relationships among the 22 hapl o types found for C. brevis in the Azorean subpopulations show a similar pattern to the phylogenetic tree, where two very distinct haplotype clusters were obs erved, the first connect ing the Ponta Delgada subpopulations amongst each other and close to Maia, while the Angra sub populations appeared as a more distant group with more changes in the haplotype, compared with the Ponta Delgada populations. Contrastingl y Horta show ed a separate cluster with the haplotypes very close amongst each other but more distant from the one haplotype present in the AB population. These results confirm what was found with the phylogenetic analysis. The high number of haplotypes indicate d that there is high genetic diversity in the se subpopulations. This is further confirmed by the AMOVA results that show ed a high value for haplotype FST, which was significantly different from 0, showing a high percentage of variance among populatio ns. The high genetic diversity among subpopulations suggests that there was a multiple introduction scenario, because if this had been a single introduction, the diversity would be very l ow due to a founder effect where only a portion of the diversity of t he original subpopulation is represented in the founder subpopulation, the overall diversity of the separated subpopulation is lower than that of the original subpopulation. Another common reason for loss in diversity is th e bottleneck effect which happens when a population goes through a period where the effective size (Ne) is reduced and this can result in a loss of heterozygosity which is not the case here A common founder effect event in nature occurs when a small group of emigrants from an establish ed subpopulation goes on to found a new subpopulation. The mtDNA data indicates that there is high diversity

PAGE 62

62 among subpopulations and that therefore the subpopulations are not under the founder effect, and this can be explained by the fact that ther e were multiple introductions. The Azorean subpopulations however, may still be exchanging genes. This is corroborated by the fact that the gene flow of Nm= 3.7 per generation is high. For isolated populations this value is usually bellow 1.6 (Hartl and Clark 2007). Over time, a high rate of gene flow can lead to maintenance of genetic similarity between subpopulations, but because this is a recent introduction the high rate of gene flow between the different islands subpopulations is not reflected in the genetic diversity, which is high. Also the mean expected heterozygosity was not significantly different from the observed heterozygosity, which suggests that there was no decrease in heterozygos ity which would be expected in isolated subpopulations and shows that there is gene flow between the subpopulations. FST values show the proportion of the total genetic variance contained in a subpopulation relative to the total genetic variance and high values mean considerable degree of difference among subpopulations (H artl and Clark 2007). T hat there was great genetic diversity in the sub populations studied herein, as shown by the high overall values of FST 0.15 for all methods calculated) and that there was a very low level of inbreeding for each population ( FIS 0) suggests that there is gene flow. The FIS values show the proportion of the variance in the subpopulation contained in an individual as compared to its subpopulation, and the higher the value the higher the level of inbreeding (Hartl and Clark 2007). A hi gh level of inbreeding would be expected if a subpopulation were completely isolated, so if the subpopulations of the islands were truly isolated from each other the inbreeding values would be different from 0. The FIT values take into account both the FST and FIS values to calculate the probability of

PAGE 63

63 autozygosity of an inbred individual relative to the subpopulation as a whole (Hartl and Clark 20 07). The low values obtained were again indicative that there is high diversity and low inbreeding as would be expected if there was still exchange of genes between the subpopulations, or if there were gene flow from outside subpopulations This supports the hypothesis that there were several introductions that may still be occurring. The distance diagram calculate d using the GDA software (Figure 2 10) showed the Angra subpopulation separated f rom the other subpopulations. This is different from wha t was found with the mtDNA data, where the Angra subpopulations were close to both the Ponta Delgada and the Horta subpopulations. In the distance diagram the Horta subpopulations were close to the Maia subpopulation and the Ponta Delgada subpopulation. When comparing the FST values between the subpopulations (Table 215) the p values encountered showed a similar pattern w here the Horta and Maia subpopulations did not have significant differences between each other, while the Angra and Ponta Delgada subpopulations had values that were not different between some of the subpopulations. Even though the microsatellite results s eem to contradict the findings of the mtDNA, the pattern of multiple introductions is still supported. In both cases the data showed that there is high diversity. Mitochondrial DNA is maternally inherited while the microsatellite loci are codominant markers with each allele being contributed by each parent. Over evolutionary time, microsatellites can reflect potential homoplasy. H owever because the C. brevis introduction is recent, the microsatellites are a good marker to study the relationships between t he subpopulations. This can account for the differences encountered when calculating the genetic distance between subpopulations. However, in both mtDNA and microsatellites, the data showed that

PAGE 64

64 there were two distinct introductions in the two major islands and from there they may have spread to the smaller islands. The scenarios tested with the DIYABC softw are were created based on what was probable given the previous results. Of all the scenarios tested, the two that showed the highest probability were s cenarios 3 and 5 where the subpopulations of Horta and Maia were considered to have separated from the subpopulation o f Ponta Delgada, while the Angra subpopulations were independent. T he fact that scenario 4, where all of the subpopulations were assumed t o have come from a common ancestral population, had near 0 probability confirms that the hypothesis of multiple introductions cannot be rejected. The data showed that there have been at least two separate introductions of C. brevis into the Azorean archipelago, and that the infestations are spreading through the islands. The reason why the probability of the scenarios is not higher might be due to the continuous flow o f genes into the subpopulations from multiple introductions that might confound the analys is The present study has confirmed that multiple introductions have occurred to the Azorean islands yet f urther studies need to be made to resolve this issue completely The reason why two different genetic markers were used in this study, and are normally used, is because the whole aspect of the population has to be taken into account. Using only mtDNA can induce an error if a set of populations of a species happens to have high variability for the chosen gene and because it is only maternally inherited The opposite can also happen where there is very little variability for a particular gene that masks genetic divergence between two populations. M ore importantly the statistical tests used to calculate these relationships always present the caveat of sam ple size. When looking at studies of introducti ons of species, namely termites that have such a

PAGE 65

65 complex life cycle, sample size is crucial. Not only in terms of actual numbers of termites, but also in terms of number of loci /genes and number of alleles. M any population genetics studies with termites have emphasized the caveat that a large sample is necessary ( Husseneder et al. 2002, Austin et al. 2 006, Vargo et al 2006). However, the number of alleles found in the subpopulations of C. brevis from the Azore s is similar to the number of alleles found by Fuchs et al. (2003) for C. secundus showing that the choice to use C. secundus microsatellite primers for C. brevis is a good choice for these studies. Cryptotermes brevis has a cryptic nature, spending most of its life cycle inside wood. Collecting such a species in urban environment where it is mostly found, proves difficult. The permission of homeowners is usually implied and dispersal flight season happens only once a year. For this reason it is hard to gather large amounts of specimens for this species. Population genetic studies of C. brevis have not been performed previously and so there were no primers specific ally designed for the species, that would allow for better resolution of results This stud y presents itself as the first step in studies of C. brevis population genetics. The results obtained from both the mtDNA and the microsatellites helped resolve a long standing question, despite the caveats. The Azorean population was introduced multiple t imes and it is most likely still receiving gene flow from periodic new introductions. I t has also been shown by this study that the species is spreading from island to island. For these aforementioned reasons it has already been recommended and taken into action to have a more rigorous regulation over what materials enter the islands, especially wood item s, like crates, or even furniture which can potentially harbor termites ( Decreto Legislativo Regional n. 22/2010/A de 30 de Junho de 2010)

PAGE 66

66 Future direct ions call for a more detailed study with primers specifically designed for this species which would allow analysis over a large number of loci and the inclusion of a higher number of termites per sample. Also one could possibly include the recently identif ied introductions to other islands in the Azores over fixed periods of time to have a better grasp and better comprehension of the population dynamics within and between the islands. Recent introductions are a very interesting phenomenon to study because t he evolution of the population dynamics can be followed in real time, and can help further the knowledge of a complex breeding system like the termites. Understanding the evolutionary processes that introduce and maintain genetic variability among subpopul ations can help predict the dispersal pattern and therefore assist with the identification of potential spread.

PAGE 67

67 CHAPTER 3 ORIGIN OF THE AZOREA N POPULATION OF THE WEST INDIAN DRYWOOD TERMITE Cryptotermes brevis is a wood pest and is well distributed in t he tropical and subtropical areas (Figure 1 2) This species is believed to be endemic to Chile and Peru, where it can be found infesting living trees, away from any structure (Scheffrahn et al. 2008) It is possible that this is the point of origin for ma ny of the current infestations of C. brevis around the world, through the elevated traffic of wooden ships from the Spanish sailors that ruled the sea in the 1500s, transporting goods and possibly termites with them (Scheffrahn et al. 2008). This mode of dispersion, through maritime routes may have been the cause of the introduction of the species in the Azorean archipelago. The association between termite invasiveness through marina areas has already been shown (Hochmair and Scheffrahn 2010) for subterranean termites. Also, the import of wood for construction may be another avenue of entry for invasive drywood termite species into islands (Scheffrahn and Crowe 2011) This study aims to resolve the possible origin of the populations in the Azores using genetic markers to compare the Azorean populations with populations from various parts of the world. For this to be achieved, microsatellite analysis along with mtDNA analysis needs to be combined. Material and Methods Termites Termites from the Azores were collected as described in the previous chapter. The additional specimens used were selected from the UF termite collection at the Fort Lauderdale Research and Education Center, Davie Florida. The se specimens were collected by various coll aborators in recent decades from various parts of the globe, and

PAGE 68

68 are preserved in vials in 8 5% non denatured ethanol at room temperature. S pecimens were chosen from the database to cover different parts of the world (Table 31). Table 31. Exact locations of collection sit es from the world collection and number of individuals (N) used for analysis. Continental Area Country Sample name Database Code N Latitude Longitude Collector Africa Republic of Congo AFR AFR302 5 4.3022 15.3052 Nixon Oceania Australia AUS AUS77 5 25.5405 152.7155 Sch South America Venezuela VZ SA092 5 10.4980 66.9212 Helm South America Chile CLA CL3 20 27.8013 70.1333 Ripa South America Chile CLB CL47 20 33.0233 71.6166 Paola South America Peru PEA PE101 20 11.9790 77.0900 Krecek So uth America Peru PEB PE152 20 13.0870 76.3970 Garcia North America United States FLD FL98 20 24.5551 81.7795 Sch Central America Honduras HN HN666 20 14.0087 87.0144 Sch* Central America Costa Rica CR CTA127 15 9.9277 84.0771 Suiter Central Am erica El Salvador ELS CTA112 20 13.6633 89.2580 Nixon Sch* Scheffrahn DNA extraction and amplification The procedures for DNA extraction and amplification were the same as described in the previous chapter. Five termites per population were used individ ually for the mtDNA work, while the microsatellite analysis used between five to twenty termites per population based on the availability of samples in the collection (Table 31). The PCR products for the mtDNA work were sequenced at Laragen Inc. (Culver C ity, C A)

PAGE 69

69 Data analysis Mitochondrial DNA Phylogenetic and molecular evolutionary analyses for mitochondrial 16S rRNA gene and Cytb gene were conducted using MEGA ver sion 5 (Tamura et al. 2011). Sequences were aligned in MEGA version 5 using the ClustalW method set at default, as recommended by the software. The sequences from 16S rRNA and Cytb were combined in order to create more robust sequence data. After the sequences were aligned they were run through the Gblocks software (Talavera and Castresana 2007) to eliminate poorly aligned positions and divergent regions of the alignment. Distance matrixes for number of nucleotides were created for all the islands and equal sequences were eliminated from the analysis, and a representative sequence was chosen t o represent those sequences. The distance pvalue was then calculated to test if the data from the sequences were good to analyze. The remaining sequences were then analyzed using the UPGMA method (Sneath and Sokal 1973). The bootstrap consensus tree was i nferred from 5000 replicates (Felsenstein 1985). The evolutionary distances were computed using the Kimura 2parameter method (Kimura 1980) and are based on the number of base substitutions per site. Sequences from K flavicollis for both genes were used as an out group. These representative sequences were obtained from GenBank by using BLAST for each gene. For mitochondrial 16S rRNA gene the K. flavicollis sequence used was reported in Szalanski et al. (2004) and identified by accession number AY486437. For Cytb gene, the K. flavicollis sequence used was the one reported in Legendre et al. (2008) and identified with the accession number EU253919. Microsatellites. For the microsatellite analysis the Peak Scanner Software v1.0 ( Copyright 2006, Applied B io systems) was used to score the microsatellites The

PAGE 70

70 EXCEL Microsatellite toolkit (Park 2001) was used to estimate the number and frequency of alleles, as well as observed and ex pected heterozygosit ies An ANOVA was used to test if there were significant differences between expected and observed heterozygosities and allele frequencies between populations using SAS (2003). FSTAT 2.9.3.2 (Goudet, 2001) and GENEPOPV4 ( F. Rousset) were used to calculate FST, FIS, and FIT, to respectively quantify the loss in heterozygosity, calculate linkage disequilibrium, and establish the frequency of null alleles. They were also used to estimate the modified Wrights F statistics by Weir and Coc kerhams (1984) ST, Neis (1987) GST, and Michalakis and Excoffiers (1996) ST The gene flow was quantified in terms of migrants exchanged among subpopulations per generation (Nm), calculated using GENEPOPV4 ( F. Rousset). Genetic distance was calculat ed using Neis (1972) method and a genetic distance matrix and tree was generated using Genetic Distance Analysis 1.1 (Lewis and Zaykin 2001). The program STRUCTURE (Pritchard et al. 2000) was use d to evaluate the subpopulation genetic structure. This meth od uses the individual multilocus genotypic data to evaluate models assuming different numbers of genetic clusters (K), and each individuals genome is partitioned into fractions that represent the ancestry in each inferred cluster. The models used assumed some degree of admixture between subpopulations but considered the allele frequencies independent. The other parameters of the models were set at the default of the program. The models were run for a length of 100,000 data collection generations and 50,000 Markov chain Monte Carlo (MCMC) generations. Several subsets of data were analyzed. The number of clusters used varied between K=2 to K= 10. The partitioned data evaluated was : Islands O nly data; All S ubpopulations data ; and I slands and E ndemic O nly data (Table 32)

PAGE 71

71 Table 32. Subpopulations included in the partitioned data used for the STRUCTURE analysis, codes of the populations are given according to the codes previously described. Partitioned data Subpopulations included Islands only AA, AB, AC, PDA, PDB, PDC, HA, HB, M Islands and Endemic AA, AB, AC, PDA, PDB, PDC, HA, HB, M, CLA, CLB, PEA, PEB All Subpopulations AA, AB, AC, PDA, PDB, PDC, HA, HB, M, CLA, CLB, PEA, PEB, HN, VZ, CR, FLD, ELS, AFR, AUS To test different scenarios of origin th e program DIYABC (Cornuet et al. 2010) was used. The assumptions used for the models were that there was an ancestral population of size N 2 from where the subpopulations originated. The size of the ancestral population was assumed to be higher than that of the other subpopulations N2>N1, N3, N4, N5, and N6 (Figures 3 1 to 3 6 ). For the computations the subpopulations of geographical region were considered as one single subpopulation region. The data of the Azores subpopulations AA, AB, AC PDA, PDB, PDC,HA, HB, and M were merged into Population 1 (default nomenclature in the program), the data for the South American subpopulations CLA, CLB, PEA, PEB, and VZ were merged into Population 2, Central American subpopulations HN, CR, and ELS into Population 3, the North American subpopulation FLD was represented as Population 4, the Australian subpopulation AUS was represented as Population 5, and the African subpopulation was represented as Population 6. A generalized mutation model (GSM) was assumed for the micr osatellite loci, with a gamma distribution for the mutation rate bounded between 5x104 and 5x103. The reference table was created for 6 0,000 simulations (10,000 per scenario) The posterior probability of each scenario was assessed by a polychotomous wei ghted logistic regression, and t he number of simulated data used for

PAGE 72

72 the logistic regression was 6,000 with 600 real data used for comparison. Six scenarios of introduction were tested (scenarios 1 to 6) with no admixture assumed. In all scenarios, Populat ion 2 was assumed to have an effective size N2 for some time (td) and eventually at a time (ta) the populations were assumed to split from this ancestral population. Population 2 is assumed to be the ancestral population because the subpopulations merged i n this population include the endemic subpopulations of Chile and Peru. In scenario 1, f or simplicity of the model all of the p opulations were assumed to split at the same unknown time (ta), with Population 2 continuing to exist after the spilt (Figure 3 1 ).In scenario 2, a t time tb Population 1 (Azores) split s from Population 2, with it continui ng to exist after the two split The remaining populations split form Population 2 at an earlier time (ta). Populations 1, 3, 4, 5, and 6 were assumed to have a ra ndom effective size smaller than Population 2 (N2 N 6 ) (Figure 3 2). In scenario 3, a t time tb Population 1 split s from Population 3, with population 3 continuing to exist after the split. The remaining populations split form Population 2 at an earlier time (ta.) P opulations 1, 3, 4, 5, and 6 were assumed to have a random effective size smaller than Population 2 (N N6), and Population 1 was assumed to have a random effective size smaller or equal to N3 (Figure 33). In scenario 4, at time tb Population 1 split s from Population 4, with P opula tion 4 continuing to exist after the split. The remainder populations split form Population 2 at an earlier time (ta). P opulations 1, 3, 4, 5, and 6 were assumed to have a random effective size smaller than Population 2 (N N6), and Population 1 was assumed to have a random effective size smaller or equal to N4 (Figure 34). In scenario 5, at time tb Population 1 split from Population 5, with P opulation 5 continuing to exist after the split. The remaining populations split fro m Population 2 at an ear lier time (ta). Populations 1, 3, 4, 5, and 6

PAGE 73

73 we N6), and Population 1 wa s assumed to have a random effective size smaller or equal to N5 (Figure 35). In scenario 6, at time tb Population 1 split from Population 6, with P opulation 6 continuing to exist after the split. The remaining populations split form Population 2 at an earlier time (ta). Populations 1, 3, 4, 5, and 6 were assumed to have N6), and Population 1 w a s assumed to have a random effective size smaller or equal to N6 (Figure 36). An extra scenario was added to test the validity of the Population 2 as the ancestral population (scenario 2A). In scenario 2A, all the Populations we re assumed to have emerged from a common ancestral population of size NA, in which NA was assumed to be higher than the effective size of the remainder populations 6) All the populations including Population 2 were assumed to have diverged from the ancestral population at a determined time (td), with the ancestral population not continuing to the present time (Figure 37). The posterior probability of each s cenario was assessed by a polychotomous weighted logistic regression. The reference table created for the analysis had 20,000 simulations (10,000 per scenario) and the number of simulated data used for the logistic regression was 2,000 with 200 real data used for comparison.

PAGE 74

74 Figure 31. Scenario 1 of the origin of the Islands subpopulations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N3, N4, N5 and N6. Computations were made based on the samples from each population (Sa 16). The Population 2 was assumed to be the ancestral population existent for a determined time (td). The splitting of populations was assumed to happen at a determined time (ta), a. Each Population corresponds to a different Geographical Region with the different sites per region merged into one single Population. Population 1 wa s the subpopulation of the Azores, Population 2 wa s the subpopulations of the South American samples, Po pulation 3 wa s the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The scenario assumed that populations 1, 3, 4, 5, and 6 were part of one ancestral population of size N2 with the condition that N6, and that subsequently they split into the different populations with Population 2 continuing its existence. Africa Australia North America Central Ameri ca South America Azores

PAGE 75

75 Figure 32. Scenario 2 of the origin of the Islands subpopulations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N3, N4, N5 and N6. Computations were made based on the samples from each populat ion (Sa 16). The Population 2 was assumed to be the ancestral population existent for a determined time (td). The splitting of populations was assumed to happen at a determined time (ta and tb), with the condition that td different Geographical Region with the different sites per region merged into one s ingle Population. Population 1 was the subpopulati on of the Azores, Population 2 was the subpopulations of the South American sample s, Population 3 was the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The scenario assumed that populations 3, 4, 5, and 6 were part of one ancestral population of size N2 with the condition that N2 N6, and that subsequently they split into the different populations at time ta with Population 2 continuing its existence, while Population 1 split at a later time from Population 2 at time tb North America South America Azores Central America Australia Africa

PAGE 76

76 Figure 33. Scenari o 3 of the origin of the Islands subpopulations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N 3, N4, N5 and N6. Computations were made based on the samples from each population (Sa 16 ). The Population 2 was assumed to be the ancestral population existent for a determined time (td). The splitting of populations was assumed to happen at a determined time (ta and tb), with the condition that td different Geographical Region with the different sites per region merged into one s ingle Population. Population 1 was the subpopulati on of the Azores, Population 2 was the subpopulations of the South American sample s, Population 3 was the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The scenario assumes that populations 3, 4, 5, and 6 were part of one ancestral population of size N2 with the condition that N2 N6, and that subsequently they split into the different populations at time ta with Population 2 continuing its existence while Population 1 split at a North America South A merica Central America Azores Australia Africa

PAGE 77

77 Figure 34. Scenario 4 of the origin of the Islands subpopulations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N 3, N4, N5 and N6. Computations were made based on the samples from each population (Sa 16) The P opulation 2 was assumed to be the ancestral population existent for a determined time (td). The splitting of populations was assumed to happen at a determined time (ta and tb), with the condition that td different Geographical Region with the different sites per region merged into one s ingle Population. Population 1 was the subpopulati on of the Azores, Population 2 was the subpopulations of the South American sample s, Population 3 was the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The scenario assumes that populations 3, 4, 5, and 6 were part of one ancestral population of size N2 with the condition that N2 N6, and that subsequently they split into the different populations at time ta with Population 2 continuing its existence while Population 1 split at a North America Central America South America Azores Australia Africa

PAGE 78

78 Figure 35. Scenario 5 of the origin of the Islands subpopulations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N 3, N4, N5 and N6. Computations were made based on the samples from each population (Sa 16). The Population 2 was assumed to be the ancestral population existent for a determined time (td). The splitting of populations was assumed to happen at a determined time (ta and tb), with the condition that td different Geographical Region with the different sites per region merged into one s ingle Population. Population 1 was the subpopulati on of the Azores, Population 2 was the subpopulations of the South American sample s, Population 3 was the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The scenario assumes that populations 3, 4, 5, and 6 were part of one ancestral population of size N2 with the condition that N2 N6, and that subsequently they split into the different populations at time ta with Population 2 continuing its existence while Population 1 split at a North America Central America South America Australia Azores Africa

PAGE 79

79 Figure 36. Scenario 6 of the origin of the Islands subpopulations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N 3, N4, N5 and N6. Computations were made based on the samples from each population (Sa 16). The Population 2 was assumed to be the ancestral population existent for a determined time (td). The splitting of populations was assumed to happen at a determined time (ta and tb), with the condition that td different Geographical Region with the different sites per region merged into one s ingle Population. Population 1 was the subpopulati on of the Azores, Population 2 was the subpopulations of the South American sample s, Population 3 was the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The scenario assumes that populations 3, 4, 5, and 6 were part of one ancestral population of size N2 with the condition that N2 N6, and that subsequently they split into the different populations at time ta with Population 2 continuing its existence while Population 1 split at a North America Australia Azores Africa South America Central America

PAGE 80

80 Figure 37. Scenario 2A of origin of the populations tested with the DIYA BC software. The program assumed an effective random population size for each population N1, N2, N3, N4, N5, and N6. Computations were made based on the samples from each population (Sa 16). Each Population corresponds to a different Geographical Region with the different sites per region merged into one s ingle Population. Population 1 was the subpopulati on of the Azores, Population 2 was the subpopulations of the South American samples, Population 3 was the subpopulation of the Central American samples, Population 4 was the subpopulation of the North American samples, Population 5 was the subpopulation of Australia, and Population 6 was the subpopulation of Africa. The splitting of populations was assume d to happen at a determined time td. The scenario assumes that populations 1, 2, 3, 4, 5 and 6 were all part of a single ancestral population of size NA with the condition that N6, and were all introduced, from that ancestral population, separately to the different geographical regions A North America Australia Africa Central America Azores South America

PAGE 81

81 Results Mitochondrial DNA A total of 55 samples were successfully am plified for each gene and PCR products were sent for sequencing. Of th e 55 samples, 39 sequences were readable and analyzed with MEGA 5 for the combined 16S rRNA and Cytb. Twenty eight of the Azorean sequences (from the previous chapter) were added to the alig n ment making it a total of 67 sequences. Five of the sequences were eliminated from analysis after being determined to be equal in the distance matrix (Tables B 6 to B 16) and one sequence was eliminated because it was too di fferent and assumed to be a contaminated sequence (FLD 1 in Table B 7). The phylogenetic bootstr ap consensus tree is shown in Figure 38 There were two main axes with a r esolution below 70. The endemic subpopulations were distributed through the two main axis (highlighted in yellow) as were the Azorean subpopulations (highlighted in gray) Half of the Angra subpopulations we re clustered with the Ponta Delgada and Maia subpopulations in a separate branch from the other subpopulations with poor support The other half of the Angra subpopulations were clustered in a branch with a resolution of 90, along with subpopulations from Africa, Australia, Chile, Peru, Costa Rica, and El Salvador, as well as with the Horta subpopulations. The subpopulations from the endemic region show ed a separation as far as subpopul ations from the same country were concerned. The C h ilean subpopulation A was clustered in the upper branch, while Chiles subpopulation B appear clustered in the lower branch of Figure 38. As for Peru the B subpopulation wa s clustered toget her, while the A subpopulation was divided between the upper and lower main branch of the phylogenetic tree.

PAGE 82

82 Figure 38 The evolutionary history was inferred using the UPGMA method. The bootstrap consensus tree was inferred from 5000 replicates and taken to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates we re collapsed. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test are shown next to the branches. The evolutionary distances were computed using the Kimura 2parameter method and are in the units of the number of base substitutions per site. The analysis involved 62 nucleotide sequences. All ambiguous positions were removed for each sequ ence pair. There were a total of 686 positions in the final dataset. Highlighted subpopulations are Azorean (grey ) and endemic (yellow). AFR 2 FLD 3 AB 1 AB 2 PDA 3 AC 2 PDB 2 PDB 5 PDC 1 PDB 2 M 2 M 7 M 1 PDB 1 M 9 AB 5 VZ 1 VZ 5 CLA 1 CLA 2 CLA 3 CLA 5 PEA 4 CR 2 CLA 4 AFR 5 AFR 3 AUS 2 AUS 5 AUS 1 CR 1 CLB 2 AA 3 HB 5 HA 5 AA 1 PEB 1 PEB 4 PEB 2 HA 2 HB 4 HA 1 HA 4 HB 1 HB 2 HB 3 HA 3 CR 3 HN 1 PEA 3 HN 5 ELS 1 ELS 3 ELS 4 ELS 5 AFR 4 CLB 1 CLB 3 AB 4 CR 4 HN 4 Kaloterm es flavicollis 96 71 70 65 90 67 61 62 66 61 66 64 53 61

PAGE 83

83 Microsatellites A total of 3 7 alleles were detected in the populations over the five loci. Alleles per locus are giv en in table 3 3 The loci Csec 1, 3, and 4 had the most alleles with 10, 12, and seven alleles respectively. The Csec6 and Csec5 loci had less alleles with three and five alleles respectively. The estimated allele frequencies are given in Tables 34 t o 3 8 No unique alleles were found for the world samples subpopulations. The frequencies varied between 0 and 100%, with no significant differences between the subpopulations (F=0.61 with a pvalue=0.902 for an 0.05 value of significance). Table 33 Size of alleles in number of base pairs per locus for the five microsatellite loci of C. brevis analyzed. Locus Alleles (bp) Csec6 143 146 149 Csec5 122 125 128 131 134 Csec4 129 144 148 154 157 166 169 Csec1 208 214 21 6 218 220 224 226 230 232 234 Csec3 116 118 120 122 124 126 128 130 132 134 136 138 Table 34 Allele frequencies, in percentage, for the Csec6 locus for the subpopulations of C. brevis, generated by the Microsatellite tool kit for Excel. Allele CLA C LB PEA PEB CR ELS VZ HN FLD AUS AFR 143 85.0 90.0 47.5 62.5 73.3 87.5 100.0 85.0 60.0 100.0 100.0 146 12.5 10.0 47.5 22.5 26.7 12.5 0.0 10.0 30.0 0.0 0.0 149 2.5 0.0 5.0 15.0 0.0 0.0 0.0 5.0 10.0 0.0 0.0 Table 35 A llele frequencies in percentage, f or the Csec5 locus for the subpopulations of C. brevis generated by the Microsatellite tool kit for Excel Allele CLA CLB PEA PEB CR ELS VZ HN FLD AUS AFR 122 7.5 2.5 12.5 20.0 26.7 0.0 0.0 2.5 32.5 50.0 0.0 125 87.5 35.0 67.5 80.0 56.7 92.5 100.0 97.5 67.5 50.0 100.0 128 5.0 40.0 17.5 0.0 6.7 7.5 0.0 0.0 0.0 0.0 0.0 131 0.0 5.0 2.5 0.0 10.0 0.0 0.0 0.0 0.0 0.0 0.0 134 0.0 17.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

PAGE 84

84 Table 36 Allele frequencies in percentage, for the Csec4 locus for the subpopulatio ns of C. brevis generated by the Microsatellite tool kit for Excel. Allele CLA CLB PEA PEB CR ELS VZ HN FLD AUS AFR 129 55.0 85.0 10.0 25.0 93.3 75.0 100.0 50.0 57.5 60.0 70.0 144 5.0 0.0 0.0 2.5 0.0 2.5 0.0 2.5 15.0 0.0 0.0 148 5.0 0.0 5.0 17.5 0.0 0. 0 0.0 0.0 10.0 0.0 0.0 154 35.0 2.5 12.5 45.0 3.3 2.5 0.0 47.5 15.0 10.0 0.0 157 0.0 10.0 0.0 0.0 3.3 17.5 0.0 0.0 0.0 0.0 30.0 166 0.0 0.0 52.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 169 0.0 2.5 20.0 10.0 0.0 2.5 0.0 0.0 2.5 30.0 0.0 Table 37 Allele freq uencies in percentage, for the Csec1 locus for the subpopulations of C. brevis generated by the Microsatellite tool kit for Excel. Allele CLA CLB PEA PEB CR ELS VZ HN FLD AUS AFR 208 42.5 67.5 55.0 67.5 70.0 85.0 50.0 57.5 67.5 80.0 90.0 214 27.5 22.5 5.0 2.5 20.0 7.5 10.0 0.0 2.5 0.0 0.0 216 0.0 0.0 0.0 0.0 0.0 2.5 0.0 22.5 0.0 0.0 0.0 218 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.0 0.0 0.0 220 15.0 0.0 5.0 20.0 0.0 0.0 20.0 2.5 0.0 0.0 0.0 224 7.5 2.5 5.0 2.5 0.0 0.0 20.0 17.5 0.0 0.0 0.0 226 2.5 7.5 10. 0 2.5 10.0 5.0 0.0 0.0 5.0 20.0 10.0 230 5.0 0.0 2.5 5.0 0.0 0.0 0.0 0.0 20.0 0.0 0.0 232 0.0 0.0 17.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 234 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Table 38 Allele frequencies, in percentage, for the Csec3 locus for the subpopulations of C. brevis generated by the Microsatellite tool kit for Excel Allele CLA CLB PEA PEB CR ELS VZ HN FLD AUS AFR 116 0.0 17.5 0.0 0.0 0.0 10.0 0.0 7.5 12.5 0.0 0.0 118 32.5 2.5 35.0 42.5 3.3 2.5 80.0 32.5 17.5 0.0 0.0 120 7.5 12.5 0.0 0.0 3.3 12.5 0.0 10.0 17.5 0.0 0.0 122 12.5 15.0 0.0 17.5 20.0 40.0 0.0 27.5 12.5 40.0 30.0 124 0.0 2.5 2.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 126 27.5 5.0 5.0 12.5 0.0 2.5 20.0 5.0 7.5 0.0 0.0 128 0.0 17.5 2.5 2.5 23.3 12.5 0.0 5.0 0.0 10.0 0.0 130 0.0 2.5 32.5 17.5 10.0 10.0 0.0 10.0 12.5 0.0 10.0 132 15.0 17.5 2.5 2.5 13.3 7.5 0.0 2.5 7.5 40.0 30.0 134 0.0 2.5 10.0 2.5 0.0 0.0 0.0 0.0 0.0 10.0 0.0 136 5.0 0.0 0.0 0.0 0.0 2.5 0.0 0.0 2.5 0.0 0.0 138 0.0 5.0 10.0 2.5 26.7 0.0 0.0 0.0 10.0 0.0 30 .0

PAGE 85

85 The mean expected and observed heterozygosities were not significantly different for a ll the subpopulations (Table 39 ). The mean number of alleles per subpopulation varied between 1.8 and 5. 4 The FIS values ranged from 0.079 to 0.302 (Table 39 ). The values were not significantly different from 0, with only Venezuela, Australia and Africa having FIS values significantly different from 0. Table 39 Estimates of genetic diversity at 5 polymorphic microsatellite loci in C. brevis from the samples from various continents. Pop Mean Expected Hz Mean Observed Hz ANOVA F stat s P value for ANOVA Mean No of Alleles per locus Fis p* value CLA 0.52 0.12 0.42 0.05 0.31 0.593 4.4 1.52 0.199 0.0032 CLB 0.51 0.13 0.45 0.05 0.11 0.746 5.2 3.42 0.118 0.047 2 PEA 0.64 0.05 0.48 0.05 2.77 0.135 5.4 2.07 0.249 0.0001 PEB 0.57 0.08 0.44 0.05 1.80 0.217 4.8 2.39 0.236 0.0005 CR 0.49 0.12 0.35 0.06 0.70 0.428 3.8 1.92 0.302 0.0006 ELS 0.37 0.12 0.32 0.05 0.09 0.776 4.4 2.88 0.142 0.0321 VZ 0.22 0.15 0.16 0.07 0.11 0.751 1.8 1.30 0.289 0.1676 HN 0.45 0.13 0.38 0.05 0.13 0.726 4 2.35 0.166 0.0224 FLD 0.61 0.08 0.53 0.05 0.72 0.421 4.8 2.68 0.13 0.0303 AUS 0.45 0.13 0.48 0.10 0.02 0.884 2.4 1.14 0.079 0.7564 AF R 0.29 0.15 0.24 0.09 0.08 0.789 2 1.22 0.2 0.2784 *p value for larger F than the observed for randomization tests The estimated null allele frequency for all the subpopulations sampled was low for Csec 5, and 1 with values never above 0.13 (Table 3 10). The locus Csec3 showed the highest frequency of null alleles wi th values of 0.44 and 0.45 for C R and AFR respectively. The matrix of FST p values between subpopulations had many significantly different values (Tables 311 to 314). The subpopulat ions of V Z and A US were not significantly different from the Angra and Ponta Del gada subpopulations (Tables 312 and 313). The subpopulation AC was not significantly different from any of the other subpopulations. Subpopulation PEB was not different from the Ponta Delgada subpopulations (Table 312).

PAGE 86

86 Table 310. Estimated null allele frequencies for all loci for the subpopulations of C. brevis sampled from the University of Florida Collection calculated by GENEPOP software Locus CLA CLB PEA PEB CR ELS VZ HN FLD AUS AFR Csec5 0.10 0.30 0.09 0.02 0.00 0.13 No inf 0.00 0.00 0.00 No inf Csec6 0.16 0.11 0.25 0.31 0.30 0.18 No inf 0.11 0.01 No inf No inf Csec4 0.20 0.00 0.26 0.24 0.00 0.00 No inf 0.24 0.14 0.38 0.00 Csec1 0.06 0.00 0.02 0.04 0.00 0.00 0.16 0.00 0.00 0.00 0.00 Csec3 0.00 0.12 0.15 0.13 0.44 0.00 0.00 0.09 0.29 0.00 0.45 The Hardy Weinberg test for disequilibrium was significant (pvalue=0.000 for Score U test) with linkage disequilibrium found at all the loci. The null hypothesis that there is no def iciency of heterozygos ity was rejected. The calculated gene flow was Nm=0.605876 migrants per generation. The calculated values of FST, FIS, and FIT for the three different met hods are presented in Table 315 The distance diagram calculated for the microsatellite data is shown in Figure 39. The subpopulations of Angra are clustered in a separated branch from all the other subpopulations. Maia and Horta subpopulations are clustered together closer to the Ponta Delgada subpopulations. The Ponta Delgada subpopulations are calculated to be close to the Venezuela and Peruvian subpopulations. Table 311. Matrix of FST p values between subpopulations. Significantly different values are indicated by after 190000 perm utations. Indicative adjusted nominal level (5%) for multiple comparisons is 0.000263. Sub pop AB AC PDA PDB PDC HA HB M AA 0.03 0.085 0.0006 0.00118 0.00059 0.00001* 0.00001* 0.00001* AB 0.2361 0.00001* 0.00005* 0.00055 0.00001* 0.00001* 0.00001* AC 0.003 0.02428 0.55392 0.00014* 0.00052 0.00012* PDA 0.66685 0.001 0.00001* 0.00002* 0.00001* PDB 0.00209 0.00037 0.00028 0.00009* PDC 0.01416 0.13522 0.00679 HA 0.52179 0.04845 HB 0.52604

PAGE 87

87 Table 312. Continued matrix of FST p values between subpopulations. Significantly different values are indicated by after 190000 permutations. Indicative adjusted nominal level (5%) for multiple comparisons is 0.000263 Sub pop CLA CLB PEA PEB CR ELS VZ AA 0.00002* 0.00001* 0.00001* 0 .00002* 0.00004* 0.0002* 0.00117 AB 0.00001* 0.00001* 0.00001* 0.00001* 0.00001* 0.00001* 0.00004* AC 0.00354 0.00129 0.00247 0.03205 0.00776 0.01845 0.00785 PDA 0.00002* 0.00001* 0.00002* 0.00066 0.00003* 0.00001* 0.00253 PDB 0.00372 0.00003* 0.00013* 0.01324 0.00005* 0.00001* 0.07152 PDC 0.00001* 0.00002* 0.00705 0.00074 0.00002* 0.00001* 0.00219 HA 0.00001* 0.00001* 0.00001* 0.00001* 0.00001* 0.00001* 0.00003* HB 0.00001* 0.00001* 0.00001* 0.00001* 0.00001* 0.00001* 0.0001* M 0.00001* 0.00001* 0. 00001* 0.00001* 0.00001* 0.00001* 0.00007* CLA 0.00001* 0.00001* 0.00032 0.00001* 0.00001* 0.17167 CLB 0.00001* 0.00001* 0.00337 0.00123 0.00007* PEA 0.00001* 0.00001* 0.00001* 0.00003* PEB 0.00001* 0.00001* 0.00387 CR 0.00074 0.00009* ELS 0.00001* Table 313. Continued matrix of FST p values between subpopulations. Significantly different values are indicated by after 190000 permutations. Indicative adjusted nominal level (5%) for multiple comparisons is 0.000263 Sub po p HN FLD AUS AFR AA 0.00001* 0.00001* 0.00127 0.00623 AB 0.00001* 0.00001* 0.00033 0.00022* AC 0.00426 0.00149 0.00789 0.04801 PDA 0.00006* 0.00005* 0.00259 0.0013 PDB 0.00006* 0.00071 0.00797 0.00792 PDC 0.00001* 0.00004* 0.00225 0.00233 HA 0.00001 0.00001* 0.00002* 0.00002* HB 0.00001* 0.00001* 0.0001* 0.00008* M 0.00001* 0.00001* 0.00002* 0.00001* CLA 0.00002* 0.00002* 0.00008* 0.00005* CLB 0.00001* 0.00001* 0.00262 0.01386 PEA 0.00001* 0.00001* 0.00002* 0.00002* PEB 0.00002* 0.00009* 0.000 42 0.00003* CR 0.00001* 0.00001* 0.02756 0.08136 ELS 0.00001* 0.00001* 0.00087 0.16771 VZ 0.00408 0.00004* 0.00787 0.00789

PAGE 88

88 Table 314. Continued matrix of FST p values between subpopulations. Significantly different values are indicated by after 190 000 permutations. Indicative adjusted nominal level (5%) for multiple comparisons is 0.000263 Sub pop HN FLD AUS AFR HN 0.00001* 0.00003* 0.00002* FLD 0.00049 0.00061 AUS 0.03228 Table 315. Values of FST, FIS, and FIT calculated using three different methods for all locus Weir & Cockerham (1984) Michalakis and Excoffier (1996) Nei's (1972) F ST 0.17 0.2528 0.165 F IS 0.128 0.0062 0 .123 F IT 0. 277 0. 3015 0.17 2 The STRUCTURE results from the clusters tested from K =2 to K=5 showed a clear distribution of the individuals per cluster (Figures 310 to 31 2 ) for all the subsets of data tested. The K=6 to K=10 clusters did not show a clear pattern and were not considered further for the analysis (Figures A 6 to A 1 0 ). For the Islands o nly a sub set of the number of clusters K=3 divided the subpopulations into three clear clusters with the Angra Population in one cluster (green), PDA and PDB in a distinct cluster ( red ) and the remaining subpopulations in the third cluster (blue) (Figure 3 10) The All S ubpopulations subset STRUCTURE result (Figure 311) showed that for the number of clusters K=4, Angra subpopulations sharing a cluster with El Salvador and Africa (blue), while Horta and Maia shared the Peru A cluster (yellow). Chile subpopulation B and Costa Rica shared a cluster (Red), and the rest of the subpopulations (PDA C, CL A, PEB, VZ, HN, and FLD) shared the fourth cluster (green). For Islands and Endemic only subset of data, the K=4 clusters had the subpopulation of Angra isolated from all other subpopulations (yellow), and the CLB subpopulation also isolated from the others (red).

PAGE 89

89 The Horta and Maia subpopulations were clustered with the PEA subpopulation (Blue), while the PDA C, CLA and PEB were clustered together (green) (Figure 312). Figure 39 Distance diagram calculated using Genetic Distance Analysis. Length of branches represents genetic distance according to Neis identity method. The longer the branch compared to the scale, the more distant the su bpopulations. All sampled subpopulations are represented. 0.01 Angra A Angra B Angra C Chile A Honduras Peru B Florida Australia Chile B Costa Rica El Salvador Africa Ve nezuela Ponta Delgada A Ponta Delgada B Ponta Delgada C Peru A Maia Horta A Horta B

PAGE 90

90 Figure 310. Assignments of 104 C. brevis individuals sampled from 9 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K=2 to 5 are shown. Labels identi fy the subpopulations from each Island from where samples were collected.

PAGE 91

91 Figure 311. Assignments of 274 C. brevis individuals sampled from 20 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, der ived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K=2 to 5 are shown. Labels identify the subpopulations from each subpopulation from where samples were collected.

PAGE 92

92 Figure 31 2 Assignments of 184 C. brevis individuals sampled from 13 sites to genetic clusters inferred from STRUCTURE simulations (based on average member ship coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. C lusters K=2 to 5 are shown. Labels identify the subpopulations from the Azores and the endemic region from where samples were collected.

PAGE 93

93 The DIYABC software results for the probability of the scenarios showed that for the comparison of the six scenarios o f possible origin of the Azorean population, scenario 1, where all the populations were assumed to have diverted from the ancestral Population 2, was the most probable scenario with a 50% probability (Figure 313) while all the other scenarios had less than 20% probability of being the right scenario for the origin of the Azorean subpopulation with a 95% confidence interval. When compared with a scenario where the populations were all part of an ancestral unknown Population of size NA (scenario 2A) scenar io 1 was again the most probable scenario with an 80% probability (Figure 314). Figure 313. Logistic regression of the simulated data probability for the different scenarios tested in DIYABC. Probability varies between 0 and 1. All scenarios are within a 95% confidence interval for the simulations. Probability Number of simulated data

PAGE 94

94 Figure 314. Logistic regression of the simulated data probability for the different scenarios tested in DIYABC. Probability varies between 0 and 1. All scenarios are within a 95% confidence interval for the simulations. Discussion The mtDNA results showed a pattern where both the Azorean subpopulations and the endemic subpopulations were divided between two main branches. Most of the Azorean subpopulations were clustered togeth er in one branch, while some were in a completely different branch, confirming the previous findings of different introductions of the various subpopulations. The results show ed that the endemic subpopulations are highly diverse, with the subpopulations spread through the two main branches. However when focusing on some of the other subpopulations for these genetic markers the subpopulations are spread out throughout the phylogenetic tree (Figure 38) e.g the African and Costa Rican subpopulations This can be explained by the high variability of the 16S rRNA gene. Kambhambati et al. (1996) stated that 16S rRNA can have low resolution when the variability of the organisms being studied is very high and it will not Probability Number of simulated data 2A

PAGE 95

95 produce a reliable phylogenetic analysis. On a smaller scale wher e the populations being studied are small and closely related (as previously found) the results are reliable. However, when analyzing the relationship between very distinct populations from a great geographical range, the 16S rRNA can produce a phylogeneti c tree that might have higher variability than expected for these subpopulations even though a robust method with a high number of replications was used for analysis The mitochondrial data does indicate that there is high diversity for both the Azorean a nd endemic subpopulations which would be expected given that the Azorean subpopulation is recently introduced from multiple points and the endemic subpopulation is itself a source of genetic diversity. T he micro satel lite data showed that the subpopulations do not have significant differences in the frequency of alleles between them, and that there are no unique alleles in the subpopulations of the world (Tables 33 to 37). Although the gene flow is low between these subpopulations, the alleles that are present are shared by several subpopulations, indicating their common origin. The low gene flow is indicative of subpopulation isolation (Hartl and Clark 2007) which would be expected considering that the locations where these subpopulations were sampled ar e geographically isolated, and these subpo pulations have been well established. The FIS values were not significantly different from 0 for most subpopulations except for three of them. The African, Australian and Venezuelan subpopulations had FIS values significantly different from 0, showing high levels of inbreeding for these subpopulations. That there were only five individuals sampled for the African, Australian, and Venezuelan subpopulations may have biased the results producing the appearance of hig her levels of inbreeding. Even though these are long established

PAGE 96

96 subpopulations, termites have been shown to have low inbreeding rates even within colonies ( Husseneder et al. 1999) and have strategies to avoid inbreeding depression (Husseneder et al.2006) It would be thus expected to see these low levels of inbreeding for all the subpopulations. However for some of the subpopulations the frequency of null alleles was high and this may have caused the linkage disequilibrium to be found for all loci. The ov erall FST, FIT and FIS values calculated with the three methods showed high diversity and low inbre e ding, confirming the previous findings. When comparing the FST p values between all the subpopulations the results show ed that one of the endemic subpopulat ions (PEB) has values that do not differ from the Ponta Delgada subpopulations which would be expected, due to the fact that a similar value of high diversity would be expected to be found in endemic populations as well as populations recently introduced t hat are still getting gene flow from other sources (as previously found for the Azorean subpopulations). Also the genetic diversity found in the Venezuelan and Australian populations were not significantly different from the one found in Angra and Ponta D elgada. This could be indicative of high diversity in the subpopulations of Venezuela and Australia, even though the sample size for these subpopulations was low. The GDA results show a very clear pattern in terms of genetic distance between all the subpop ulations (Figure 39). The Ponta Delgada, Horta, and Maia subpopulations were closer to Peru and Venezuela subpopulations, whereas the Angra subpopulations appear to be closer to the other world sample d subpopulations. This suggests that there might be a p oint of origin for at least the Ponta Delgada, Maia, and Horta subpopulations in Venezuela and Peru. The Angra subpopulations may just

PAGE 97

97 be too diverse to determine a point of origin, or the world sample was not big enough to find a match The STRUCTURE ana lysis showed a clear pattern of ancestry for the different subset s of subpopulations analyzed. The Azorean subpopulations Islands Only subset showed that there were three distinctive ancestry clusters. The Angra subpopulations shared common ancestry, while the Ponta Delgada subpopulations were divided between two ancestries, one unique to the Ponta Delgada subpopulations and one shared with the Horta and Maia subpopulations (Figure 310). When compared with all the world subpopulations, the Azorean subpopul ations continued to be distributed between three clusters of ancestry The Angra subpopulations shared their ancestry with El Salvador and Africa, while the Ponta Delgada subpopulations shared their ancestry with Chile, Peru, Venezuela, Honduras and Florida. The Maia and Horta subpopulations share d their ancestry with the Peru A subpopulation. These results strongly support that there are several points of origin for the Azorean subpopulations. Most of the subpopulations appear to originate from South Ameri ca. When compared with the endemic regions only, the Azorean subpopulations are divided into two distinct points of origin. The Angra subpopulations were completely separated, while the Ponta Delgada, Horta and Maia share their ancestry with the Peru and C hile subpopulations. These results substantiate what was previously observed from the GDA distance calculation, where the Angra subpopulations were isolated from the other Azorean subpopulations. Also, the origin of these other Azorean subpopulations was l ocated in South America. This wa s confirmed by the scenarios tested with DYI ABC software Of all the scenarios tested the one that had the highest probability was the scenario where all the subpopulations originated from the South America continental area. Scheffrahn

PAGE 98

98 et al. (2008) had found that the endemic populations of C. brevis were located in South America, namely in Chile and Peru. This study has confirmed this to be true. When comparing a scenario where all the subpopulations originate from an unknown ancestral population with the scenario where all the subpopulations originate from the South American group, it is shown that the endemic populations of South America we re the original population for the world subpopulations. Combining the results from the STRUCTURE analysis with the DYI ABC software it wa s apparent that the points of origin for the subpopulations in the Azores in the towns of Ponta Delgada, Maia and Horta were from South America ( Chile, Peru and Venezuela) The Angra subpopulation, how ever, was too diverse to account for an accurate point of origin. T hat the Azorean subpopulations have a high level of genetic diversity and share ancestry with the endemic subpopul ations can have two explanations. Firstly, the Azorean subpopulations may have originated from the endemic populations of Chile and Peru. Secondly, they may be showing a high level of divers ity because this was a recent introduction from multiple locations that still shows a high level of diversity throughout the subpopulations However, with the analysis performed it was shown that even though there were multiple introductions to the Islands, most of the subpopulations originated from South America. The Angra subpopulations had a completely different point of origin from the ot her subpopulations, which was not found in this study Like all of these studies further investigations and research needs be done. W orking with a small number of samples and loci can only give a certain level of resolution to this story Future directions with this study would be surveys of infested homes in the Azores, where the origin of the lumber and furniture can be assessed. Different populations should be genotyped to further the amount of data and possibly corroborate the

PAGE 99

99 hypothesis that the orig in of the Azorean populations of C. brevis is amongst others, in countries like Venezuela, Chile, and Peru. More samples from other regions of the world would be necessary, as well as a larger number of termites collected per site. One way to circumvent th e difficulty in obtaining a large number of specimens per collecting site would be to identify and create species specific primers for a large number of microsatellite loci. Analysis made with the red imported fire ant, Solenopsis invicta (Buren), have used as m any as 67 different markers, which allowed for only one specimen to be used per colony, per location (Ascunce et al. 2011). Due to the cryptic nature of C. brevis investing in augmenting the number of species specific markers could bring an even clearer picture of the point of origin of the Azorean subpopulations, as well as the pattern of spread of this recent introduction.

PAGE 100

100 CHAPTER 4 CONTROL OF THE SPREA D OF ALATE S USING ATRACTION TO DIFFERE NT LIGHT WAVELENGTHS AND CHEMICAL TREATME NTS Cryptotermes brevis has become a serious pest in the Azorean islands. The levels of infestation in the major cities are elevated and some houses are at risk of structural failure. This species has also been spreading to other smaller islands of the archipelago. Measur es have been taken to begin to bring this inter island spread to a halt with stricter laws for import and export of goods However, the spread within cities continues with new infestation sites being discovered every year (personal observation). The seaso nal flights of C. brevis are the only time this species is seen outside wood, making it a good time to apply measures of control. Light traps to capture insects have been long used. C. brevis is known to fly towards light during its dispersal fli ghts, and high intensity lights attract more alates of this spec ies, than lesser intense lights, and the use of light traps has been recommended as a measure of prevention of infestation against this termite species (Ferreira and Scheffrahn 2011) However, there ar e no data reporting if there is any preference by C. brevis for a specific light wavelength. One of the objectives of this study is to determine if there is a preferred light wavelength that will attract more alates of C. brevis into traps, and if this pre ference is true in different populations. A common preventative measure of control is the use of insecticides Many studies have been done throughout the years with different insecticides against drywood termites in the US. However drywood termites are not a common pest in Europe so very little information is available about the efficacy of some of the commercially available products in Europe. Some of the products available in Europe are used as general wood preservatives (Wocosen, and Xylophene), others are used as

PAGE 101

101 insecticides (Borowood, Gentrol, and XT 2000). With this in mind, the second objective in this study was to test these commercially available insecticides against drywood termites, and to determine how effective they are in preventing colonizat ion by this species. Materials and Methods Different light wavelength preference Experiments were conducted at the Fort Lauderdale Research and Education Center (FLREC) Davie, Florida, and in a privately owned attic in Angra do Herosmo, Terceira, Azores. The experiments were set up in dark rooms kept at ambient temperature (T) and relative humidity (RH) (Florida: T 25.5C, RH 73.6%; Terceira: T 20.5C, RH 79.9%). In both Florida and the Azores the rooms were filled with C. brevis infested wood. The Florida experiment was conducted between April and June of 2009, during C. brevis dispersal flight season for South Florida, and the experiment in the Azores was conducted during July and September of 2009, the dispersal flight season for C. brevis in the Azores. In Florida Twenty one transparent plastic boxes (36x23x28 cm) served as light wavelength preference chambers. Six different wavelengths were used (395 nm (ultraviolet), 460 555 nm (white), 470 nm (blue), 525 nm (green), 590 nm (yellow), 625 n m (red)) with three replicates per wavelength. Each box was wrapped in aluminum foil in order to isolate the light from one box to the other. Light Emitting Diodes (LED) light bulbs were used for each wavelength (Ultraviolet model# YA UV5N30N; White model# SS5W4UAEC; Blue model# SS5B4SEEC; Green model# SS5G4UAEC; Yellow model# SS5Y4UAEC; Re d model# SS5R4SDEC ). Nine light bulbs were mounted in a series

PAGE 102

102 circuit in clusters of three so that there would be three replicates of one wavelength with three light bulbs per replicate. Every cluster of light bulbs was five feet apart from the next one. The series circuit was mounted for each wavelength with 22Gauge stranded hook up wire ( RadioShack technology plus ), Full Wave bridge rectifiers (4Amp. 50Volts Rad ioShack 276 1146) plug and connector (Leviton 15A PR 926781002 0B. 125Volts), 10K Ohm resistor ( Watt 5% tolerance RadioShack 2711335), 4.7K Ohm resistor ( Watt 5% toleranc e RadioShack 271 1124), and viny l electrical tape medium grade (Scotch ). T he boxes were placed in a room infested with C. brevis with an open end facing the room. The different wavelengths were randomly distributed (Figure 41). Three of the boxes had unlit lights to serve as control and the remaining boxes had the different L EDs hung in the middle of the box. Hefty EZ Foil cake pans (21.5 cm diam. and 3.8 cm deep) were placed inside the boxes underneath the light bulbs and filled up to three quarters with water (Figure 42). These pans served as water traps to capture the alates flying to the lights. The alates were collected and counted every day for all the boxes, throughout the flight season. In the Azores For the Azorean set up the same light bulbs were used and connected with an adaptor for 240V input (standard use in Portugal). They were also randomly distributed inside 21 boxes to replicate the set up in Florida. This set up was assembled in an attic area filled with wood infested by C. brevis The lights ran from 7 pm to 9:30 am everyday and alates in the traps were collected and counted daily.

PAGE 103

103 Figure 41. Distribution of the different wavelengths in the boxes. The lights were distributed randomly with 3 replicates per light and 3 replicates with no light which were the controls. Figure 42. Close up of the light chamber (green light 5 25 nm). The bundle of three light bulbs is placed above the pan which was filled up with water. Termite alates that flew into the chamber were trapped in the water.

PAGE 104

104 Data Analysis The data collected from the experiments in the Azores and in Florida were separately analyzed using ANOVA (SAS 2003) to test if there was any difference in the variances of the number of alates per color light trap p er day Color was used as the factor for the ANOVA anal ysi s. The data had to be transformed by ln ( data+k) where k=2 in order t o meet the assumptions for ANOVA (the data did not have a normal distribution) The null hypothesis being tested was that there were no differences between the variances for the number of alates caught in the light traps per day. Further analysis using Tukey s method (SAS 2003) was used to determine which wavelengths had significantly different variance of the number of alates caught in the trap for each location, separately Chemical prevention of colonization Experimental set up. Experiments were conducted in a private attic in Angra do Herosmo, Terceira, Azores, and were conducted between July and September 2009. To test different chemicals as preventatives for colonization of this species, bioassays ( attic ) models that mimic the type of construction used in attics in the Azores were assembled These attic modules were treated with the commercially available products in the Azores, Xylophene (cypermethr in 0.07%), Wocosen (permeth r in 0.24%), Gentrol (hydroprene 9.0%), Bor owood ( dis odium tetra borate decahydrate 10%), and XT 2000 (d limonene 92%). The commonly used c onstruction grade Criptomria ( Cryptomeria japonica) was used to construct the modules. Boards 39x 13x 13 cm and 39x13x2.5 cm were purchased from a local lumber ya rd and assembled in the desig n and dimensions given in Figure 43 The sections of the boards were joined with a total of eight 3.5 cm long screws. The boards were assembled in a clapboard pattern. The attic modules

PAGE 105

105 were treated on the surface with label recommended rates of the insecticides and three replicates per insecticide w ere used, with t he three replicates for controls receiving no treatment for a total of 18 modules. After treatment the mo dules were left to dry for 48h and then placed individuall y in a shallow plastic container, to prevent cross contamination. The modules were then placed in a C. brevis infest ed room under a source of light for alate attraction (Figure 4 4) The source of light consisted of 70 watt power ultra v iolet lights common ly used in solariums The modules were randomly distributed and left to be colonized. After one month, the modules were removed from the light source and kept for another month. At the end of the second month the modules were disassembled and survival and colony development was recorded. Although mortality was also registered, it was not used for the analysis because dead dealates found in one attic unit might have landed on other units before dealating, and so only survival was considered. Data anal ysi s. A one way ANOVA (SAS 2003) was used to test the difference of variances between the number of live dealates found in the units with the treatment as the ANOVA factor. The same was used to test the difference of variance in the number of chambers, and number of eggs with the treatment as the factor. All the data was subjected to a ln (data+k ) transformation where k=2, to comply with the ANOVA assumptions (the data were not normally distributed) The null hypotheses being tested were: there were no differenc es in the variances of the number of live dealates in the attic units, per treatment; there were no differences in the variances of number of nuptial chambers in the attic units, per treatment; and there were no differences in variances in number of eggs i n the attic units, per treatment. In order to test which treatments were

PAGE 106

106 significantly different from one another a post ANOVA Tukey s analysis was performed, Figure 43. Attic module design. Mode of assembly and measures are shown Back boards are assembled in a clapboard pattern. Figure 44. Attic modules random ly distributed under U V light. 39cm 3.5 cm screw 39cm 13 cm 13 cm 13 cm 2.5 cm Overlapping boards

PAGE 107

107 Results Different light wavelength preference In Florida A total of 991 alates were caught in the light traps in Florida T he blue light had the highest total number of alates in the traps (343) but it was not significantl y different from white and green and r ed and control had the lowest total number of alates with a total of 6 and 2, respectively (Table 41) There were significant differences between the number of alates Both y ellow and U.V. lights had no significant differences to the lights with the lowest mean number of alates per trap per day, and the lights with the highest mean number of alates per trap per day (Table 41). Table 41. T otal and mean number of alates caught in light traps in Florida Values with same letter were not significantly different after Tukey s Wavelengths Total number of alates per waveleng th Mean number of alates per tr ap per day (SE ) Control 2 0.10 0.02 a 625 nm ( Red ) 6 0.29 0.06 a 590 nm ( Yellow) 26 1.10 0.24 a,b 395 nm ( U.V ) 110 4.80 1.05 a,b 460550 nm ( White ) 262 11.86 2.58 b 525 nm ( Green) 298 13.62 2.97 b 470 nm ( Blue ) 343 15.42 3.37 b In th e Azores A total of 1835 alates were caught in the light traps in the Azores. T he blue light had the highest total number of alates in the traps (388) whereas r ed and control had the lowest total number of alates with a total of 79 and 97, respectively (T able 42 ). There were significant differences between the number of alates i n the traps per

PAGE 108

108 wavelength (F=5.4 lights had no significant differences to the lights with the lowest mean number of alates per trap per day (red and control), and the lights with the highest mean number of alates per trap per day (green, white, and blue) (Table 42 ). Table 42. Total and mean number of alates caught in light traps in the Azores Values with same letter were not significantly differe nt after Tukey s analysis for Wavelengths Total number of alates per wavelength Mean number of alates per trap per day (SE ) Control 79 3.29 0.63a 625 nm ( Red ) 97 4.04 0.95a 590 nm ( Yellow) 228 9.50 2.37a,b 395 nm ( U.V .) 304 12.67 3.05a,b 525 nm ( Gree n ) 362 15.08 3.78b 460550 nm ( White ) 377 15.71 3.17b 470 nm ( Blue ) 388 16.17 4.017b Chemical prevention of colonization A total of 3,896 dead dealates were collected on all trays and modules. There were significant differences between treatments (Table 43) for the number of live dealates, number of chambers, and number of eggs. For both permethrin and cypermethrin no live dealates were found, and they were significantly different from all the other treatments. Dead dealates seemed to be evenly distributed throughout the treatments with no observable differences. The d limonene treatment was not significantly different from the control (Table 43). No larvae were found in the chambers. Most of nuptial chambers were occupied by a single pair of dealates, as observed in situ Not all of the live dealates were found in nuptial chambers. Some dealates were found crawling on the substrate with no nuptial chamber close by.

PAGE 109

109 Table 43. Mean SE per attic module of C. brevis live dealates, number of nuptial chambers, and number of eggs. Values with the same letter were not significantly different after Tukey s Treatment Number of dealates Number of chambers Number of eggs Control 23.33 2.91a 11.67 1.45 a 5.00 2.00 a D limonene 26.00 4.73 a 12.67 2.33 a 4.00 1.53 a Dis odium Tetrab orate Decahydrate 12.67 2.03 a,b 6.00 1.15 a,b 1.00 0.58 a,b Hydroprene 7.67 2.96 b 3.67 1.45 b 1.67 0.33 a,b Permethrin 0 0 c 0 0 c 0 0 b Cypermethrin 0 0 c 0 0 c 0 0 b F 40.01 32.51 8.62 df1 5 5 5 df2 12 12 12 p 0.000 0.000 0.001 Discussion The alates flying in both the Azores and Florida showed a similar behavior towards the wavelength of preference. When faced with choice chambers the lights in the wavelength of the blue, green and white had a higher incidence of alates tra p ped compared to the remaining wavelengths. However for both Florida and the Azores these three wavelengths did not have significantly mor e alates caught than y ellow or U V lights Minnick (1973) had observed an attraction of the alates for U.V. light, preferring it to the incandescent light. Ultra v iolet light is commonly used in insect traps, and the common notion is that this wavelength is the most preferred by insects in general. However, more studies are looking into other wavelengths and their attraction to insects. Chang et al. (2001) looked at alate attraction of C. formosanus to different light

PAGE 110

110 wavelengths and similarly found that blue (367583 nm) and green (525648 nm ) lights attracted significantly more alates of this species than the red (600733 nm) lights and the control. Previously, Yamano (1987) had also found that winged adult response to colored lights reached the maximum with the blue (400420nm) lights. With the ease of access to simple and inexpensive LEDs, these are being tested for traps for insects. Nakamoto and Kuba (2004) tested the effectiveness of green LEDs in light traps against the West Indian sweet potato weevil Euscepes postfasciatus (Fairmaire ) They found that when given the choice the weevil would choose the green light as opposed to blue, yellow, and red, and made a case for using LEDs. The experiments for light wavelength were performed with two different populations, one in Florida and one in the Azores. T hese two differ ent populations were kept in different environments under different characteristics with the population in Florida kept in a small closed, completely dark room while the Azorean popul ation was kept in a big open room, wit h some competing sunlight. However, and despite the different conditions the results were comparable, showing that this may be a preferen ce for the species itself and not just for a particular population. The main reason that it was important to look at t he light wavelength preferences for C. brevis was to optimize light traps for C brevis, to be used during the dispersal flight season. Using simple LED lights that are inexpensive can be a good way to get the public involved in preventative treatments for C. brevis in the Azores. Also there is lower fire hazard, because LEDs are cold lights and do not overheat. A simple light trap composed of regular Christmas tree lights (most of these products are made up of LEDs nowadays) and a sticky trap, or a container with water underneath c ould be

PAGE 111

111 recommended for use in a household. The fact that the controls got significantly less termites flying into it shows that having a light trap is helpful in reducing infestation pressure. In the Azores the main infestation sites are in the dark often unused or cluttered attics Promoting the use of these simple traps among homeowners can help slow the further spread of this species during the dispersal flight season. T he insecticide treatment results show that the two mos t commonly used insecticides for wood treatment permethrin and cypermethrin were most effective preventing colony foundation. B oth the di sodium tetra borate decahydrate and the hydroprene showed no significant differences in the number of eggs with permethr in and cypermerthrin, with a low number of eggs. This can be indicative that although sodium borate and hydroprene do not work in preventing colony foundation by killing the dealates, they may work on a physiological level preventing the new founded coloni es from producing enough progeny to survive as a colony. Studies have shown that the average number of eggs laid for C. brevis in a nuptial chamber is four (McMahan 1960, Ferreira 2008), with the colonies surviving to at least the following year (McMahan 1 960). The results showed an average of one to two eggs for disodium tetraborate decahydrate and hydroprene, while dlimonene showed an average of four eggs. However there were no significant differences between the average number of eggs laid in the disodi um tetraborate decahydrate and hydroprene treatments, and the dlimonene and control treatments. These results may be due to the fact that a C brevis incipient colony has a very low production of eggs in the first year (Ferreira 2008), and significant diff erences are hard to identify with these low numbers. Cypermethrin and permethrin were the most effective insecticides by causing 100% mortality with no survivors to start

PAGE 112

112 a nuptial chamber or lay eggs. This is important information because it confirms that these two popular wood preservatives can be used to prevent C. brevis colony foundation. Future directions can include testing the product degradation of cypermethrin and permethrin to determine what the frequency of treatments should be. The use of a combination of both light traps and an insecticide wood treatment with cypermethrin or permethrin might have a synergetic effect and could be a more effective way to reduce the spread of C. brevis not only within the households where there are infestations already, but as well as in the cities where this pest is already present. Also, the light traps can be used to monitor the efficacy of the pesticide treatment. Reducing the spread of C. brevis could help to keep the population levels low er even if there is no complete eradication. However, more work is needed.

PAGE 113

113 CHAPTER 5 CONCLUSIONS Cryptotermes brevis in the Azores is a n important urban pest problem It is at this point well established and will only continue to spread to more islands. Since the beginni ng of this work in 2009, two more spots of infestation o n a different island (Pico) have been identified in 2011 (Figure 13) (O. Guerreiro personal observation). This species has been introduced several times and it is still being introduced. T his study has shown that the multiple introductions scenario was the most probable. T he points where this species might have been introduced from are Venezuela Chile and Peru. Even though more studies are necessary, this is an early assessment of this species population movement and dynamics. It was established in this study that the Azorean populations are still subject to high gene flow, not only from the outside but also between islands. It is therefore important to work on ways to prevent this species from spreading further, either within the islands where it is already present, or to other islands. Studies on preventative control and ways to improve this preventative control are important for the future i n order to have lower levels of infestation. Combining t he use of i nsecticides with light traps could help decrease the spread of this species within the Islands. The knowledge of the point of origin of this species and the fact that it is still being introduced can be used to convey a higher level of regulatio n in the import o f wood materials to the islands in order to decrease the introduction of this pest. Furthermore, this invasive species is only one of the termite species present in the Azores that is causing structural damage. This study helps to elucidat e what is

PAGE 114

114 happening with this particular species but more studies into the other species present o n the islands are also needed for the future.

PAGE 115

115 APPENDIX A EXTRA FIGURES Figure A 1. Different castes of termite s and life cycle (S cheffrahn) Figure A 2. Gel picture of PCR products of C. brevis amplified mt DNA for 16S rRNA and Cytb genes. A p proximate sizes of the products are shown in number of base pairs using the exACTGene Mini Ladder as a measurement aid.

PAGE 116

116 Figure A 3. Gel pic ture of amplified PCR products for four microsatellite loci (Csec1, 4, 5, and 6). White arrows indicate the approximate base pair size of the alleles, and yellow arrows indicate examples of a homozygote (single band) and a heterozygote (two bands) for the same loci. Figure A 4. Example of an alig n ment session using M EGA 5.0 where the sequences have been aligned using ClustalW method. The asterisks indicate matches in the sequences.

PAGE 117

117 Figure A 5. Example of scoring the base pair size of an allele on the Pe akScanner software. The peak is highlighted and it shows an homozygote with the information about the number of base pairs provided on the table at the bottom of the screen. Figure A 6. Assignments of 104 C. brevis individuals sampled from 9 sites to gen etic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color segment representing the individual s genomic membership to that cluster. Black lines separate sample sites. Clusters K=6 to 8 are shown. Labels identify the subpopulations from each Island from where samples were collected.

PAGE 118

118 Figure A 7. Assignments of 104 C. brevis individuals sampled from 9 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K=9 and K=10 are shown. Labels identify the subpopulations from each Island from where samples were collected. Figure A 8. Assignments of 274 C. brevis indiv iduals sampled from 20 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color s egment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K=6 to 8 are shown. Labels identify the subpopulations from each subpopulations from where samples were collected.

PAGE 119

119 Figure A 9 Assignment s of 274 C. brevis individuals sampled from 20 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a single vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K=9 and K=10 are shown. Labels identify the subpopulations from each subpopulation from where samples were collected. Figure A 10. Assignments of 184 C. brevis individuals sampled from 13 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci). Each individual is represented by a s ingle vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K= 6 to 8 are shown. Labels identify the subpopulations from the Azores and the endemic region from where samples were collected.

PAGE 120

120 Figure A 1 1 Assignments of 184 C. brevis individuals sampled from 13 sites to genetic clusters inferred from STRUCTURE simulations (based on average membership coefficient, Q, derived from 5 microsatellite loci ). Each individual is represented by a single vertical line, with the length of each color segment representing the individuals genomic membership to that cluster. Black lines separate sample sites. Clusters K= 9 and 10 are shown. Labels identify the subpopulations from the Azores and the endemic region from where samples were collected.

PAGE 121

121 APPENDIX B EXTRA TABLES Table B 1. Primer sequence for mitochondrial genes used for amplification. Adapted from Legendre et al. 2008. Gene primers Sequence (5' 3') Source 16S rRNA 16SAr CGC CTG TTT ATC AAA AAC AT Xiong and Kocher, 1991 16SF TTA CGC TGT TAT CCC TAA Kambhampati, 1995 Cytb cytb612 CCA TCC AAC ATC TCC GCA TGA TGA AA Kocher et al., 1989 cytb920 CCC TCA GAA TGA TAT TTG GCC TCA Kocher et al., 1989 Table B 2. Primer sequence for microsatellite locus used for amplification. Adapted from Fuchs et al. 2003. Locus primers Sequence (5' 3') Csec1 Forward AACGCTTGGATTATGGGTTC Reverse TATCTGTCTGTCTGTCTGTCG Csec3 Forward TGTACTTCTGACATCTCAGG Rev erse GTCGGTACGGTCCACTTTGC Csec4 Forward TGTTAGAAGGCTACCAGCGC Reverse TCTTTCCTCTGCGAACTGTC Csec5 Forward TGAAAGCCAGTGGGGCAGCTGC Reverse TGCCTACAGTCAGAGCTCAAGC Csec6 Forward ACAGTTTGATCAGTGCCTTGG Reverse ACTGGCATTAGGGTTAGGTAC

PAGE 122

122 Table B 3 V ariation sites for the 22 haplotype s o f C. brevis sequences for the combined Cytb and 16SrRNA genes for the island samples Dots are nucleotides equal to the first row. Numbers indicate the position of the variation sites on the combined sequences. Haplotypes 7 1 0 12 13 14 15 16 17 18 19 21 23 31 35 44 47 51 56 74 AA 1 C C T A G G A A T C C A A T A T C C T AA 3 . . . . AB 1 T G G T T T G T G C G C T T C AB 4 A . T . G . AB 5 T G G A T T A A T G C G C T T C AC 2 T G G A T T A G T G C G C T T C PDA 2 T G T C T T G C G C T T C PDA 3 T G G A T T G T T G C C G C T T C PDA 5 T G T C T T G C C G C T T C PDB 1 T G T C T T G C G C T T C PDB 2 T G T C T T G C G C T T C PDC 1 T A G T C T T G C G C T T C HA 1 . T . . . HA 2 . . . . HA 3 . T . . . HA 4 . T . . . HA 5 . . . . HB 2 . T . . . HB 4 . . . . M 1 T G T C T T G C G C T T C M 2 T G T C T T G C G C T T C M 9 T G T C T T G C G C T T C

PAGE 123

123 Table B 4. Variation sites for the 22 haplotypes of C. brevis sequences for the combined Cytb and 16SrRNA genes for the island samples Dots are nucleotides equal to the first row. Numbers indicate the position of the variation sites on the combined seq uences. 92 107 110 122 128 143 146 155 170 191 203 206 257 258 272 275 278 280 AA 1 T T C T G T C T C T C C T T A A A T AA 3 . . . . AB 1 C C T C A C T C T C T A C . AB 4 . . . C G AB 5 C C T C A C T C T C T A C . AC 2 C C T C A C T C T C T A C . PDA 2 C C T C A C T C T C T A C . PDA 3 C C T C A C T C T C T A C . PDA 5 C C T C A C T C T C T A C . PDB 1 C C T C A C T C T C T A C . PDB 2 C C T C A C T C T C T A C . PDC 1 C C T C A C T C T C T A C . HA 1 . . . . HA 2 . . . T G HA 3 . . . . HA 4 . . . . HA 5 . . . . HB 2 . . . . A HB 4 . . . T M 1 C C T C A C T C T C T A C . M 2 C C T C A C T C T C T A C . M 9 C C T C A C T C T C T A C T C

PAGE 124

124 Table B 5. Va riation sites for the 22 haplotypes of C. brevis sequences for the combined Cytb and 16SrRNA genes for the island samples Dots are nucleotides equal to the first row. Numbers indicate the position of the variation sites on the combined sequences. 327 3 28 345 349 350 372 396 397 401 405 434 475 585 592 646 647 652 654 AA 1 A T G A G T T T T C G C T A T C T C AA 3 . . . . AB 1 C A G A C C C C T A T . AB 4 . . T . AB 5 C A G A C C C C T A T . AC 2 C A G A C C C C T A T . PDA 2 C C C C C T A T T . PDA 3 C A G A C C C C T A T T . PDA 5 C A G A C C C C T A T . PDB 1 C C C C C T A T . A PDB 2 C A G A C C C C T A T . PDC 1 C A G A C C C C T A T . HA 1 . . . . A HA 2 . . . . HA 3 . . . C A HA 4 . . . . HA 5 . . . . HB 2 . . . . HB 4 . . . T . M 1 G C A G A C C C C T A T G . M 2 C A G A C C C C T A T . M 9 C A G A C C C C T A T .

PAGE 125

125 Table B 6. Distance matrix of e stimates of e volutionary d ivergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base d ifferences between sequences is shown. The analysis involved 67 nucleotide sequences AFR2 AFR3 AFR 4 AFR5 AUS1 AUS2 AUS5 CLA1 CLA2 CLA3 CLA4 CLA5 CLB1 CLB2 AFR2 AFR3 29 AFR4 33 4 AFR5 17 39 40 AUS1 26 4 6 38 AUS2 27 3 5 37 3 AUS5 29 2 4 39 2 1 CLA1 11 30 34 24 27 28 30 CLA2 10 27 31 22 24 25 27 2 CLA3 8 25 29 23 22 23 25 8 7 CLA4 13 34 36 23 31 32 34 14 13 12 CLA5 15 15 19 29 13 14 15 16 13 11 20 CLB1 29 7 7 39 7 8 7 30 27 25 32 14 CLB2 29 4 4 39 4 5 4 30 27 25 32 15 3 CLB3 30 6 6 40 6 7 6 31 28 26 33 15 1 4 CLB4 29 7 7 39 7 8 7 30 27 25 32 14 0 3 CLB5 29 7 7 39 7 8 7 30 27 25 32 14 0 3 CR1 26 7 7 36 3 6 5 27 24 22 29 16 8 5 CR2 11 22 24 23 20 21 22 12 9 7 14 9 21 20 CR3 29 4 4 39 4 5 4 30 27 25 32 15 3 2 CR4 18 15 17 30 13 16 15 2 2 20 16 21 20 16 15 ELS1 31 4 4 41 4 5 4 32 29 27 34 17 5 2 ELS3 30 5 5 40 3 6 5 31 28 26 33 16 6 3 ELS4 32 7 7 40 5 8 7 33 30 28 35 18 8 5 ELS5 27 8 8 37 5 6 8 28 25 23 30 15 9 6

PAGE 126

126 Table B 7 Distance matrix of estimates of evolutionary divergence bet ween the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base differences between sequences is shown. The analysis involved 67 nucleotide sequences AFR2 AFR3 AFR4 AFR5 AUS1 AUS2 AUS5 CLA1 CLA2 CLA3 CLA4 CLA5 CLB1 CLB2 FLD1 123 134 135 129 132 133 134 129 127 124 124 131 135 134 FLD3 13 37 39 24 34 35 37 19 20 17 19 24 35 35 HN1 29 4 4 39 4 5 4 30 27 25 32 15 3 2 HN3 29 4 4 39 4 5 4 30 27 25 32 15 3 2 HN4 45 20 20 54 20 21 20 46 43 41 48 31 19 18 HN5 30 5 5 40 5 6 5 31 28 26 33 16 4 3 PEA3 29 4 4 39 4 5 4 30 27 25 32 15 3 2 PEA4 14 21 21 28 19 20 21 18 16 14 17 10 18 17 PEB1 31 4 4 41 4 5 4 32 29 27 32 17 5 2 PEB2 31 4 4 41 4 5 4 32 29 27 34 17 5 2 PEB4 31 4 4 41 4 5 4 32 29 27 32 17 5 2 PEB5 31 4 4 41 4 5 4 32 29 27 32 17 5 2 VZ1 14 33 35 28 30 31 33 19 18 16 21 19 31 31 VZ5 12 27 29 25 24 25 27 16 15 10 18 17 25 25 AA 1 31 4 4 41 4 5 4 32 29 27 34 17 5 2 AA3 30 3 3 40 3 4 3 31 28 26 33 16 4 1 AB1 8 34 36 19 31 32 34 14 15 13 16 20 32 32 AB 2 8 34 36 19 31 32 34 14 15 13 16 20 32 32 AB4 31 9 9 41 9 10 9 32 29 27 34 16 4 5 AB5 10 35 37 23 32 33 35 17 18 15 19 22 33 33 AC2 8 34 36 19 31 32 34 14 15 13 16 20 32 32 PDA2 9 34 36 20 31 32 34 15 16 13 17 20 32 32 PDA3 9 35 37 20 32 33 35 15 16 14 17 21 33 33 PDA5 8 33 35 19 30 31 33 14 15 12 16 19 31 31 PDB1 9 34 36 20 31 32 34 15 16 13 17 20 32 32 PDB2 8 33 35 19 30 31 33 14 15 12 16 19 31 31 PDC1 8 33 35 19 30 31 33 14 15 12 16 19 31 31

PAGE 127

127 Table B 8 Distance matrix of estimates of evolution ary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base differences between sequences is shown. The analysis involved 67 nucleotide sequences AFR2 AFR3 AFR4 AFR5 AUS1 AUS2 AUS5 CL A1 CLA2 CLA3 CLA4 CLA5 CLB1 CLB2 HA1 32 5 5 42 5 6 5 33 30 28 35 18 6 3 HA2 32 5 5 42 5 6 5 33 30 28 35 18 6 3 HA3 33 6 6 43 6 7 6 34 31 27 36 19 7 4 HA4 31 4 4 41 4 5 4 32 29 27 34 17 5 2 HA5 30 3 3 40 3 4 3 31 28 26 33 16 4 1 HB1 31 4 4 41 4 5 4 32 29 27 34 17 5 2 HB2 32 5 5 42 5 6 5 33 30 28 35 18 6 3 HB3 32 5 5 42 5 6 5 33 30 28 35 18 6 3 HB4 32 5 5 42 5 6 5 33 30 28 35 18 6 3 HB5 30 3 3 40 3 4 3 31 28 26 33 16 4 1 M1 9 34 36 20 31 32 34 15 16 13 17 20 32 32 M2 8 33 35 19 30 31 33 14 15 12 1 6 19 31 31 M5 8 33 35 19 30 31 33 14 15 12 16 19 31 31 M7 8 33 35 19 30 31 33 14 15 12 16 19 31 31 M9 10 35 37 21 32 33 35 14 15 14 18 21 33 33 Table B 9 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base differences between sequences is shown. The analysis involved 67 nucleotide sequences CLB3 CLB4 CLB5 CR1 CR2 CR3 CLB4 1 CLB5 1 0 CR1 7 8 8 CR2 22 21 21 17 CR3 2 3 3 5 20 CR4 15 16 16 10 15 13

PAGE 128

128 Table B 10 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base differences between sequences is sh own. The analysis involved 67 nucleotide sequences CLB3 CLB4 CLB5 CR1 CR2 CR3 CR4 ELS1 ELS3 ELS4 ELS5 FLD1 FLD3 HN1 HN3 HN4 ELS1 4 5 5 5 22 2 15 ELS3 5 6 6 4 21 3 14 1 ELS4 7 8 8 6 23 5 16 3 2 ELS5 8 9 9 7 20 6 15 4 3 5 FLD1 135 135 135 130 127 133 124 135 134 133 133 FLD3 36 35 35 32 18 35 25 37 36 38 34 130 HN1 2 3 3 5 20 0 13 2 3 5 6 133 35 HN3 2 3 3 5 20 0 13 2 3 5 6 133 35 0 HN4 18 19 19 21 36 16 29 18 19 21 22 147 51 16 16 HN5 3 4 4 6 21 1 14 3 4 6 7 132 36 1 1 17 PEA3 2 3 3 5 20 0 13 2 3 5 6 133 35 0 0 16 PEA4 19 18 18 20 11 17 20 19 18 20 17 128 22 17 17 33 PEB1 4 5 5 5 22 2 15 2 3 5 6 135 37 2 2 18 PEB2 4 5 5 5 22 2 15 2 3 5 6 133 37 2 2 18 PEB4 4 5 5 5 22 2 15 2 3 5 6 135 37 2 2 18 PEB5 4 5 5 5 22 2 15 2 3 5 6 135 37 2 2 18 VZ1 32 31 31 30 17 31 26 33 32 34 29 132 20 31 31 47 VZ5 26 25 25 24 13 25 18 27 26 28 23 126 18 25 25 41 AA1 4 5 5 5 22 2 15 2 3 5 6 135 37 2 2 18 AA3 3 4 4 4 21 1 14 1 2 4 5 134 36 1 1 17 AB1 33 32 32 29 14 32 21 34 33 35 30 127 7 32 32 47 AB2 33 32 32 29 14 32 21 34 33 35 30 127 7 32 32 47 AB4 5 4 4 10 23 5 18 7 8 10 11 137 35 5 5 21 AB5 34 33 33 30 16 33 23 35 34 36 32 130 7 33 33 48 AC2 33 32 32 29 14 32 21 34 33 35 30 127 9 32 32 47 PDA2 33 32 32 29 14 32 21 34 33 35 30 126 9 32 32 47 PDA3 34 33 33 30 15 33 22 35 34 36 31 127 8 33 33 48

PAGE 129

129 Table B 1 1 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base differences between sequences is shown. The analysis involved 67 nucleotide sequences CLB3 CLB4 CLB5 CR1 CR2 CR3 CR4 ELS1 ELS3 ELS4 ELS5 FLD1 FLD3 HN1 HN3 HN4 PDA 5 32 31 31 28 13 31 20 33 32 34 29 127 8 31 31 46 PDB 1 33 32 32 29 14 32 2 1 34 33 35 30 127 9 32 32 47 PDB 2 32 31 31 28 13 31 20 33 32 34 29 127 8 31 31 46 PDC 1 32 31 31 28 13 31 20 33 32 34 29 127 8 31 31 46 HA 1 5 6 6 6 23 3 16 3 4 6 7 136 36 3 3 19 HA2 5 6 6 6 23 3 16 3 4 6 7 136 38 3 3 19 HA3 6 7 7 7 24 4 17 4 5 7 8 137 37 4 4 20 HA4 4 5 5 5 22 2 15 2 3 5 6 135 35 2 2 18 HA5 3 4 4 4 21 1 14 1 2 4 5 134 36 1 1 17 HB1 4 5 5 5 22 2 15 2 3 5 6 135 35 2 2 18 HB2 5 6 6 6 23 3 16 3 4 6 7 136 36 3 3 19 HB3 5 6 6 6 23 3 16 3 4 6 7 136 36 3 3 19 HB4 5 6 6 6 23 3 16 3 4 6 7 13 5 38 3 3 19 HB5 3 4 4 4 21 1 14 1 2 4 5 134 36 1 1 17 M1 33 32 32 29 14 32 21 34 33 35 30 126 9 32 32 47 M2 32 31 31 28 13 31 20 33 32 34 29 127 8 31 31 46 M5 32 31 31 28 13 31 20 33 32 34 29 127 8 31 31 46 M7 32 31 31 28 13 31 20 33 32 34 29 127 8 31 31 46 M9 34 33 33 30 15 33 22 35 34 36 31 129 10 33 33 48

PAGE 130

130 Table B 12 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The number of base differen ces between sequences is shown. The analysis involved 67 nucleotide sequences HN5 PEA3 PEA4 PEB1 PEB2 PEB4 PEB5 VZ1 VZ5 AA1 AA3 AB1 AB2 AB4 AB5 AC2 PEA3 1 PEA4 18 17 PEB1 3 2 19 PEB2 3 2 19 2 PEB 4 3 2 19 0 2 PEB5 3 2 19 0 2 0 VZ1 32 31 18 33 33 33 33 VZ5 26 25 18 27 27 27 27 12 AA1 3 2 19 2 2 2 2 33 27 AA3 2 1 18 1 1 1 1 32 26 1 AB1 33 32 17 34 34 34 34 16 15 34 33 AB2 33 32 17 34 34 3 4 34 16 15 34 33 0 AB4 6 5 20 7 7 7 7 32 27 7 6 32 32 AB5 34 33 20 35 35 35 35 18 16 35 34 4 4 33 AC2 33 32 17 34 34 34 34 16 15 34 33 2 2 34 6 PDA2 33 32 18 34 34 34 34 17 15 34 33 3 3 34 6 3 PDA3 34 33 18 35 35 35 35 17 16 35 34 1 1 33 5 3 PDA5 32 31 17 33 33 33 33 16 14 33 32 2 2 33 5 2 PDB1 33 32 18 34 34 34 34 17 15 34 33 3 3 34 6 3 PDB2 32 31 17 33 33 33 33 16 14 33 32 2 2 33 5 2 PDC1 32 31 17 33 33 33 33 16 14 33 32 2 2 33 5 2 HA1 4 3 20 3 3 3 3 34 28 3 2 33 33 6 34 35 HA2 4 3 20 3 3 3 3 34 28 3 2 35 35 8 36 35 HA3 5 4 21 4 4 4 4 35 29 4 3 34 34 7 35 36 HA4 3 2 19 2 2 2 2 33 27 2 1 32 32 5 33 34 HA5 2 1 18 1 1 1 1 32 26 1 0 33 33 6 34 33 HB1 3 2 19 2 2 2 2 33 27 2 1 32 32 5 33 34

PAGE 131

131 Table B 13 Distance matrix of estimates of ev olutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The num ber of base differences between sequences is shown. The analysis involved 67 nucleotide sequences HN5 PEA3 PEA4 PEB1 PEB2 PEB4 P EB5 VZ1 VZ5 AA1 AA3 AB1 AB2 AB4 AB5 AC2 HB2 4 3 20 3 3 3 3 34 28 3 2 33 33 6 34 35 HB3 4 3 20 3 3 3 3 34 28 3 2 33 33 6 34 35 HB4 4 3 20 3 3 3 3 34 28 3 2 35 35 8 36 35 HB5 2 1 18 1 1 1 1 32 26 1 0 33 33 6 34 33 M1 33 32 18 34 34 34 34 17 15 34 33 3 3 34 6 3 M2 32 31 17 33 33 33 33 16 14 33 32 2 2 33 5 2 M5 32 31 17 33 33 33 33 16 14 33 32 2 2 33 5 2 M7 32 31 17 33 33 33 33 16 14 33 32 2 2 33 5 2 M9 34 33 19 35 35 35 35 18 16 35 34 4 4 35 7 4 Table B 14 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The num ber of base differences between sequences is shown. The analysis involved 67 nucleotide sequences PDA2 PDA3 PDA5 PDB1 PDB2 PDC1 HA1 HA2 HA3 HA4 HA5 PDA3 2 PDA5 1 3 PDB1 2 4 1 PDB2 1 3 0 1 PDC1 1 3 0 1 0 HA1 35 34 34 35 34 34 HA2 35 36 34 35 34 34 4 HA3 36 35 35 36 35 35 3 5 HA4 34 33 33 34 33 33 1 3 2 HA5 33 34 32 33 32 32 2 2 3 1 HB1 34 33 33 34 33 33 1 3 2 0 1

PAGE 132

132 Table B 15 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The num ber of base differences between sequences i s shown. The analysis involved 67 nucleotide sequences PDA2 PDA3 PDA5 PDB1 PDB2 PDC1 HA1 HA2 HA3 HA4 HA5 HB1 HB2 HB3 HB4 HB5 HB2 35 34 34 35 34 34 2 3 3 1 2 1 HB3 35 34 34 35 34 34 2 3 3 1 2 1 0 HB4 33 34 34 35 34 34 4 2 5 3 2 3 4 4 HB5 33 3 4 32 33 32 32 2 2 3 1 0 1 2 2 2 M1 2 4 1 2 1 1 35 35 36 34 33 34 35 35 35 33 M2 1 3 0 1 0 0 34 34 35 33 32 33 34 34 34 32 M5 1 3 0 1 0 0 34 34 35 33 32 33 34 34 34 32 M7 1 3 0 1 0 0 34 34 35 33 32 33 34 34 34 32 M9 3 5 2 3 2 2 36 34 37 35 34 35 36 36 34 34 Table B 1 6 Distance matrix of estimates of evolutionary divergence between the combined sequences of Cytb and 16S rRNA of C. brevis for all subpopulations sampled. The num ber of base differences between sequences is shown. The analysis involved 67 nucleotide sequences M1 M2 M5 M7 M2 1 M5 1 0 M7 1 0 0 M9 3 2 2 2

PAGE 133

133 LIST OF REFERENCES Ascunce, M. S., C.C. Yang, J. Oakey, L. Calcaterra, W.J. Wu, C. J. Shih, J. Goudet, K. G. Ross, and DW. Shoemaker. 2011. Global Invasion History of the f ire ant Solenopsis invicta. Science 331: 1066 1068 Austin, W. J., A. L. Szalanski, R. H. Scheffrahn, M. T. Messenger, J. A. McKern, and R. E. Gold. 2006. Genetic evidence for two introductions of the Formosan subterranean termites, Coptotermes formosanus (Isoptera: R hinot e r mitidae), to the United States. Florida Entomol. 89: 183193. Bess, H. A., and K. Ota. 1960. Fumigation of buildings to control the dry wood termite, Cryptotermes brevis J. Econ. Entomol. 53: 503510. Bignell D E ., and P. Eggleton. 2 000 Termites in ecosystems pp367 387 In Abe T D E Bignell, M. Higashi (eds) Termites: evolution, sociality, symbioses, ecology. Kluwer Academic Publishers, Dordrecht Borges, A., O. Guerreiro, M. Ferreira, T. Myles, and P. A. V. Borges. 2007. Tratam ento das mobilias atacadas por Cryptotermes brevis com calor, fum igantes slidos e gases inertes, pp 8595. In P. Borges, T. Myl es (eds.) Trmitas dos Aores. Principia, Estoril. Borges, P. A. V. 2007. Introduo. p p 11 14. In P. Borges, T. Myles (eds.) Trmitas dos Aores. Principia, Estoril. Borges, P.A.V., D. Lopes, A. Simes, A. Rodrigues, S. Bettencourt, and T. Myles. 2004. Determinao da Distribuio e Abundnca de Trmitas (Isoptera) nas Habitaes do Concelho de Angra do Herosmo. Departamento de Cincias Agrrias, Universidade dos Aores, Angra do Herosmo. Unpublished report Brooks, S. E., F. M. Oi, and P. G. Koehler. 2003. Ability of Canine Termite Detectors to Locate Live Termites and Discriminate them from NonTermite Material. J. Econ. En tomol. 96: 12591266 Carvalho, J. H. B. 1972. Cryptotermes brevis (Walker) in Mozambique. First notice (Isoptera, Kalotermitidae). Rev. Cinc. Agronmicas 5: 4950. Chang, L., E. L. Hsu, and W. J. Wu. 2001. Study on the phototaxis of alates of Coptoterm es formosanus Shiraki (Isoptera: Rhinotermitidae). Formosan Entomol. 21: 353363. Chouvenc T., N. Y. Su, and M. L. Elliott. 2008. Interaction Between the Subterranean Termite Reticulitermes flavipes (Isoptera: Rhinotermitidae) and the Entomopathogenic Fungus Metarhizium anisopliae in Foraging Arenas. J. Econ. Entomol. 101: 885893.

PAGE 134

134 Clement M., D. Posada, and K. A. Crandall. 2000. TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9: 16571660. Cleveland L R S. R. Hall E. P. Saunders a nd J. Collier 1934. The woodfeeding roach Cryptocercus its protozoa, and the symbiosis between protozoa and roach. Mem Am Acad Sci 17: 185 342 Cornuet J.M., V. Ravign, and A. Estoup. 2010. Inference on population history and model checking using DNA sequence and microsatellite data with the sof t ware DIYABC (v1.0). S ubmitted. Creffield, J. W., J. R. J. French, N. Chew. 1997. A laboratory assessment on the effectiveness of the ElectroGun in eradicating adults of the indigenous Australian drywood termite, Cryptotermes primus (Isoptera: Kalotermitidae). CSIRO Forestry and Forest Products. Client Report n1636pp. Edwards R., and A. E. Mill. 1986. Termites in buildings, their biology and control. Rentokil Limited, East Grinstead, 261p. Eggleton, P. 2000. Global diversity patterns, pp 2551. In : T. Abe, D. E. Bignell, M. Higashi (eds.). Termites: Evolution, Sociality, Symbiosis, Ecology. Kluwer Academic Publishers. 488 Pp. Engel M S D A Grimaldi, K. Krishna. 2009. Termit es (Isoptera): their phylogeny, classification, and rise to ecological dominance. Am Mus Novit 3650:1 27. Evans, T. A. 2002. Assessing efficacy of Termatrac TM; a new microwave based technology for nondestructive detection of termites (Isoptera). Sociobiology 40: 575583. Exc offier, L. and H. E. L., Lischer. 2010 Arlequin suite ver. 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Res 10: 564 567. Felsenstein J. 1985 Confidence limits on phylogenies: An approach usi ng the bootstrap. Evolution 39:783791. Ferreira, M.T. 2008. Dispersal flight, post flight behavior, and early colony development of the West Indian Drywood termite Cryptotermes brevis (Walker) (Isoptera: Kalotermitidae). Masters thesis. University of Fl orida. 54p. Ferreira, M. T., and R. H. Scheffrahn. 2011. Light attraction and subsequent colonization behaviors of alates and dealates of the West Indian drywood termite (Isoptera: Kalotermitidae). Florida Entomol. 94: 131136.

PAGE 135

135 Ferreira, M., T. Myles, A Borges, O. Guerreiro, and P. A. V. Borges. 2007. Consumo de madeiras e produo de particulas fecais pelas espcies de trmitas aorianas da familia Kalotermitidae: Kalotermes flavicollis e Cryptotermes brevis p p 4761. In P. Borges, T. Myles (eds.) T rmitas dos Aores. Principia, Estoril. Fuchs, A., J. Heinze, C. Reber Funk, and J. Korb. 2003. Isolation and characterization of six microsatellite loci in the drywood termite Cryptotermes secundus (Kalotermitidae). Mol. Ecol. Notes 3: 355357. Fuji, Y. M. Noguchi, Y. Imamura, and M. Tokoro. 1990. Using acoustic emission monitoring to detect termite activity in wood. Forest Prod. J. 40: 3436. Galtier, N., B. Nabholz, S. Glmin, and G. D. D. Hurst. 2009. Mitochondrial DNA as a marker of molecular diver sity: a reappraisal. Molecular Ecol. 18: 45414550. Gonalves, F. G., and J. T. S. Oliveira. 2006. Resistncia ao ataque de cupim demadeira seca ( Cryptotermes brevis ) em seis espcies florestais. Cerne Lavras, 12: 8083. Goudet, J. 2001. FSTAT, A Pro gram to Estimate and Test Gene Diversities and Fixation Indices (version 2.9.3). http://www.unil.ch/izea/softwares/fstat.html Grossman, J. L. 2006. Advanced techniques in IR thermography as a tool for the pest management professional. In : Proce e dings of SPIE. 6205: 112. Guerreiro, O., A. Borges, and, P.A.V Borges. 2009. A trmita de madeira seca Cryptotermes brevis (Walker) na cidade de Angra do Herosmo: monitorizao e controle dos voos de disperso e preveno da colonizao. Departamento de Cincia s Agrrias, Universidade dos Aores, Angra do Herosmo. 30 pp Hamada, M M. G. Petrino T. Kahunaga M. Seidman and B. D. Stollar 1984. Characterization o f genomic poly (dT dG) poly (dC dA) sequences: Structure, organization, and conformation. Mol Cell Biol 4: 2610 1621. Hartl, D. L., and A. G. Clark. 200 7. Principles of population genetics. Fourth Edition. Sinauer Associates, Sunderland, MA Pp 652 Hickin, N. E. 1971. Termites A World Problem. Hutchinson & CO Ed. Pp232. Hochmair, H. H., and R. H. Scheffrahn. 2010. Spatial Association of Marine Dockage With LandBorne Infestations of Invasive Termites (Isoptera: Rhinotermitidae: Coptotermes) in Urban South Florida J. Econ. Entomol. 103: 13381346.

PAGE 136

136 Holt J A and M. Lepage. 2000 Termites and soil properties, pp 389 407. In Abe T .,D. E. Bignell, M. Higashi. (eds) Termites: evolution, sociality, symbioses, ecology. Kluwer Academic Publishers, Dordrecht. Howse, P. E. 1970. Termites: a study in social behavior. Hutchinson, London, 150p. Hrd, I., J. Kuldov, R. Hanus, Z. Wimmer. 2006. Juvenile hormone III, hydroprene and a juvenogen as soldier caste differentiation regulators in three Reticulitermes species: potential of juvenile hormone analogues in termite control. Pest Manag. Sci. 62: 84885 4. Hunt, R. W. 1959. Wood Preservatives as deterrents to drywood termites in the Southwest. J. Econ. Entomol. 52: 12111212. Husseneder, C., E. L. Vargo, and J. K. Grace. 2002. Multilocus DNA fingerprinting and microsatellite genotyping: complementary molecular approaches to investigating colony and population genetic structure in subterranean termites. Sociobiol ogy 39: 1 10. Husseneder, C., D. M. Simms, and D. R. Ring. 2006. Genetic diversity and genotypic differentiation between the sexes in swarm ag gregations decrease inbreeding in the Formosan subterranean termite. Insect. Soc. 53: 212219. Husseneder, C., R. Brandl, C. Epplen, J. T. Epplen, and M. Kaib. 1999. Within colony relatedness in a termite species: genetic roads to eusociality? Behaviour 1 36: 10451063. Inward D. G. Beccaloni P. Eggleton. 2007 Death of an order: a comprehensive molecular phylogenetic study confirms that termites are eusocial cockroaches Biol. Lett 3 331 335 Kambhampati, S., 1995. A phylogeny of cockroaches and related insects based on DNA sequence of mitochondrial ribosomal RNA g enes. Proc. Natl. Acad. Sci. 92: 2017 2020. Kambhampati, S ., K. M. Kjer, and B. L. Thorne. 1996. Phylogenetic relationship among termite families based on DNA sequence of mitochondrial 16S ribosomal RNA gene. Insect Mol. Biol. 5: 229238 Kambhampati S., P Eggleton. 2000. Taxonomy and phylog eny of termites, p p 1 23 In T Abe, D.E. Bignell, M. Higashi ( eds .) Termites: evolution, sociality, symbioses, ecology. Kluwer Academic Publishers, Do rdrecht. 488p. Kimura M. 1980 A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16:111120.

PAGE 137

137 Kocher, T.D., W. K. Thomas, A. Meyer, S. V. Edwards, S. Pbo, F. X. Villablanca, and A. C. Wilson 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved pri mers. Proc. Natl. Acad. Sci. 86: 6196 6200. Kofoid, C. A. 1934. Biological backg rounds of the termite probl em, p p112. In C.A. Kofoid, S. F. Light, A. C. Horner, M. Randa ll, W. B. Herms and E. E. Bowe (eds.) Termites and Termite Control. Univ. California Press, Berkeley, CA. 795p. Krishna, K. 196 9 Introduction, pp 117. In K. Krishna and F. M. Weesner (eds.) Biology of termites.Volume I. Academic Press, London and New York. 598p. Korb, J., and S. Katrantzis. 2004. Influence of environmental conditions on the expression of the sexual dispersal phenotype in a lower termite: implications for the evolution of w orkers in termites. Evol. Develop. 6: 342352. Legendre, F., M. F. Whiting, C. Bordereau, E. M. Cancello, T. A. Evans, P. Grandcolas. 2008. The Phylogeny of termites (Dictyoptera: Isoptera) based on mitochondrial and nuclear markers: Implications for the evolution of the worker and pseudergate castes, and foraging behaviors. Mol. Phylogenet. Evol. 48: 615627. Lemaster, R. L., F. C. B eal, and V. R. Lewis. 1997. Detection of termites with acoustic emission. Forest Prod. Jl. 47: 7579. Lewis, P. O., and D. Zaykin 2001. Genetic Data Analysis: Computer program for the analysis of all elic data. Version 1.0 (d16c). Free program distributed by the authors over the internet from http://lewis.eeb.uconn.edu/lewishome/software.html Lewis, V. R. 1997. Alternative control strategies for termites. J. of Agricult. Entomol. 14: 291307. Lewis, V. R., and M. I. Haverty. 1996. Evaluation of six techniques for control of the Western Drywood Termite (Isoptera: Kalotermitidae) in structures. J. Econ. Entomol. 89: 922934. L ewis, V. R., and M. I. Haverty. 2001. Lethal Effects of Electrical Shock Treatments to the Western Drywood Termite (Isoptera: Kalotermitidae) and Resulting Damage to Wooden Test Boards. Sociobiology. 37(1): 163183. Lewis, V. R., C. F. Fouch e, and R. L., Lemaster. 1997. Evaluation of dog assisted searches and electronic odor devices for detecting the western subterranean termite. Forest Prod. J. 47: 7984.

PAGE 138

138 Lewis, V. R., A. B. Power, M. I. Haverty. 2000. Laboratory Evaluation of Microwaves fo r Control of the Western Drywood Termite. Forest Prod. J 50(5): 7986. Lewis, V. R., A. B. Power, and G. M. Getty. 2005. Field evaluation of thiamethoxam 2sc for management of the Western drywood termite (Isoptera: Kalotermitidae), pp.331336. In : C. Y. Lee and W. H. Robinson (eds), Proceedings of the Fifth International Conference on Urban Pests, Suntec, Singapore, July 1013, 2005, Printed by Perniagaan Phng@ P & Y Desing, Network, Malaysia. Light, S. F. 1934a. T he external anatomy of termites, p p 50 57. In C.A. Kofoid, S. F. Light, A. C. Horner, M. Randall, W. B. Herms and E. E. Bowe (eds.) Termites and Termite Control. Univ. California Press, Berkeley, CA. 795p Light, S. F. 1934b. Dry wood termites, their cl assification and distribution, p p 20620 9. In C.A. Kofoid, S. F. Light, A. C. Horner, M. Randall, W. B. Herms and E. E. Bowe (eds.) Termites and Termite Control. Univ. California Press, Berkeley, CA. 795p Lo, N., C. Bandi, H. Watanabe, C. Nalepa, and T. Beninati. 2003. E vidence for cocladogenesis between diverse dictyopteran lineages and their intracellular endosymbionts. Mol. Biol. Evol. 20: 907913. Matsuura, K., and T. Nishida. 2001. Colony fusion in a termite: what makes the society open?. Insects Soc. 48:378383. Mankin, R. W. 2004. Mi crowave radar detection of storedproduct insects. J. Econ. Entomol. 97:11681173. Mateus, T., and E. R. Goes.1953. Sobre uma termita das madeiras secas, Cryptotermes brevis Walker Lab. Nac. de Eng. Civil, Lisboa. 44: 1 56 McMahan, E. A. 1960. Laborator y studies of Cryptotermes brevis (Walker) (Isoptera, Kalotermitidae) with special reference to colony development and behavior. Doctoral Dissertation. University of Hawai i 224 p. Michalakis, Y., and L. Excoffier. 1996. A genetic estimation of population subdivision using distances between alleles with special reference for microsatellite loci. Genetics. 142: 10611064. Minnick, D. R. 1973. The flight and courtship behavior of the drywood termite, Cryptotermes brevis J.Environ. Entomol. 2: 587591. Min nick, D. R., S. H. Kerr, and R. C. Wilkinson. 1972. Control of Cryptotermes brevis J. Econ. Entomol. 65: 15771579.

PAGE 139

139 Minnick, D. R., R. C. Wilkinson, S. H. Kerr. 1973. Feeding Preferences of the Drywood Termite, Cryptotermes brevis Environ. Entomol. 2: 481 484. Myles, T. G. 2004. Report on termites in the Azores with emphasis on Cryptotermes brevis and its control. Unpublished report. Myles, T. G., P. A. V. Borges, M. Ferreira, O. Guerreiro, A. Borges, and C. Rodrigues. 2007a. Filogenia, biogeografia e eco logia das trmitas dos Aores, p p 15 28. In P. Borges, T. Myles (eds.) Trmitas dos Aores. Principia, Estoril. Myles, T. G., A. Borges, M. Ferreira, O. Guerreiro, and P. A. V. Borges. 2007b. Eficcia de diferentes insecticidas no combate a Cryptoterm es brevis p p 6275. In P. Borges, T. Myles (eds.) Trmitas dos Aores. Principia, Estoril. Nabli, H., W. C. Bailey, and S. Necibi. 1999. Beneficial insect attraction to light traps with different wavelengths. Biol. Control. 16: 185188. Nakamoto, Y., a nd H. Kuba. 2004. The effectiveness of a green light emitting diode (LED) trap at capturing the West Indian sweet potato weevil, Euscepes postfasciatus (Fairmaire) (Coleoptera: Curculionidae) in a sweet potato field. Appl. Entomol. Zool. 39: 491495. Nar s, F. N., and S. I. M. Moein. 1997. New Trend of the use of Metarhizium anisopliae (Metschnikoff) Sokorin and Verticillium indicum (Petch) Gams as entomopathogens to the termite Cryptotermes brevis (Walker) (Isoptera, Kalotermitidae). Anz. Schdlingskde, P flanzenschutz, Umweltschutz 70: 1316. Nei, M. 1972. Genetic distance between populations. Am. Nat. 106: 283292. Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York, N Y 512p. Noirot, C., and C. Noirot Timothe. 1969. Th e digestive system, pp 4988. In K. Krishna and F. M. Weesner (eds.) Biology of termites.Volume I. Academic Press, London and New York. 598p. Nutting, W. L. 1969. Flight and colony foundation, p p 233 282. In K. Krishna and F. M. Weesner (eds.) Biolog y of termites.Volume I. Academic Press, London and New York. 598p. Nutting, W. L. 1970. Composition and size of some termite colonies in Arizona and Mexico. Ann. Entomol. Soc. Am. 63: 11051110. Park, S. D. E. 2001. Trypanotolerance in West African Cattl e and the Population Genetic Effects of Selection [ Ph.D. thesis (in prep.)], University of Dublin.

PAGE 140

140 Prestwich, G. D. 1984. Defense Mechanisms of Termites. Ann. Rev. Entomol. 29: 201232. Pritchard J. K., M. Stephens and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945 959 Quarles, W. 2006. Thermal pest eradication in structures. The IPM Practitioner. 28: 18. Randall, M. and T. C. Doody. 1934. Wood Preservatives and Protective Treatments p p 38541 6. In C.A. Kofoid, S. F. Light, A. C. Horner, M. Randall, W. B. Herms and E. E. Bowe (eds.) Termites and Termite Control. Univ. California Press, Berkeley, CA. Rust, M., E. O. Paine, and D. A. Reierson. 1997. Evaluation of Freezing to Control Wood Dest roying Insects (Isoptera, Coleoptera). J. Econ. Entomol. 90: 12151221. Sands, W. A. 1 982. Agonistic behavior of African soldierless Apicotermitinae (Isoptera, Termitidae). Sociobiology. 7: 6172. SAS. 2003. Version 9.1. SAS Institute, Cary, NC Schef Y. Su. 2001. Chemical prevention of colony foundation by Cryptotermes brevis (Isoptera: Kalotermitidae) in attic modules. J. Econ. Entom. 94: 915919. Scheffrahn, R. H., P. Busey, J. Kre cek, and B. Maharajh 2006. Prevention with Fipronil of Colony Foundation by the West Indian Drywood Termite, Cryptotermes brevis (Isoptera: Kalotermitidae), in Attic Modules. Unpublished report. Scheffrahn, R. H., and W. Crowe. 2011. Ship borne termite ( Isoptera) border interceptions in Australia and onboard infestations in Florida, 19862009. Florida Entomol. 94: 57 63. Scheffrahn, R. H Endemic origin and vast anthropogenic dispersal of the West Indian drywood termite. Biol. Invasions 11: 787799 Scheffrahn, R. H., W. P. Robbins, P. Busey, N.Y. Su, and R. K. Mueller. 1993. Evaluation of a novel handheld acoustic emissions detector to monitor termites (Isoptera: Kalotermitidae, Rhinotermitidae) in Wood. J Econ. Entomol. 86: 17201729.

PAGE 141

141 Scheffrahn, R. H., N. Y. Su, P. Busey. 1997 a Laboratory and Field Evaluation of Selected Chemical Treatm ents for Control of Drywood Termites (Isoptera: Kalotermitidae). J.Econ. Entomol. 90: 492502. Scheffrahn, R. H., N Y. Wheeler. 1998. Prevention of colony foundation by Cryptotermes brevis and remedial control of drywood termites (Isoptera: Kalotermitidae) with selected chemical treatments. J. Econ. Entom. 91: 13871396. Scheffrahn, R. H., G. S. Wheeler, and N. Y. Su 1997b Heat tolerance of structureinfesting drywood t ermites (Isoptera: Kalotermitidae) of Florida. Sociobiology. 29: 237245. Sneath P.H.A., and R. R. Sokal. 1973 Numerical Taxonomy. Freeman, San Francis co. Snyder, T. E. 1926. The biology of the termite castes. Quart. Rev. Biol. 1: 522552. Spomer, N. A., S. T. Kamble, B. D. Siegfried. 2009. Bioavailability of chlorantraniliprole and indoxacarb to Eastern subterranean termites (Isoptera: Rhinotermitida e) in various soils. J. Econ. Entomol. 102: 19221927. Su, N. Y. 1994. Field evaluation of hexaflumuron bait for population suppression of subterranean termites (Isoptera: Rhinotermitidae). J. Econ. Entomol. 87: 389397. Su, N. Y., and R. H. Scheffrahn. 1990. Economically important termites in the United States and their control Sociobiology 17: 7794. Su, N. Y., and R. H. Scheffrahn. 1998. A review of subterranean termite control practices and prospects for integrated pest management programmes. Integr ated Pest Manag. Reviews. 3: 113. Szalanski, A. L., R. H. Scheffrahn, J. A. Austin, J. Krecek, and N.Y. Su. 2004. Molecular phylogeny and biogeography of Heterotermes (Isoptera: Rhinotermitidae) in the West Indies. Ann. Entomol. Soc. Am. 97: 556566 Talavera, G., and Castresana, J. 2007. Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Systematic Biol 56: 564 577. Tamura K D. Peterson N. Peterson G. Stecher M. Nei and S. Kumar 2011. MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Molecular Biology and Evolution ( in press )

PAGE 142

142 Thorne, B. L., and M. I. Haverty. 1991. A review of intracolony, intraspeci fic, and interspecific agonism in termites. Sociobiology 19: 115145. Thompson, G. J., L. R. Miller, M. Lenz, and R. H. Crozier. 2000. Phylogenetic analysis and trait evolution in Australian lineages of drywood termites (Isoptera, Kalotermitidae. Mol. Phy lo. Evol. 17: 419429. Tsutsui, N. D., and A. V. Suarez. 2003. The colony structure and population biology of invasive ants. Conserv. Biol. 17: 4858. Vargo, E. L., C. Husseneder, D. Woodson, M. G. Waldvogel, and J.K. Grace. 2006. Genetic analysis of co lony and population structure of three introduced populations of the Formosan subterranean termite (Isoptera: Rhinotermitidae) in the Continental United States. Environ. Entomol. 35: 151166. Walker F 1853 List of specimens of Neuropterous insects in the collection of the British Museum. Part III. British Museum. London, UK. Pp 501 529 Wagner, R. E. and W. Ebe ling. 1959. L ethality of inert d ust m aterials to Kalotermes minor Hagen and t heir r ole as preventives in structural pest c ontrol. J. Econ. Ento mol. 52: 208212. Weir, B. S., and C. C. Cockerhams. 1984. Estimating F statistics for the analysis of population structure. Evolution. 38: 13581370. Williams, R. M. C. 1977. The ecology and ph y siology of structural wood destroying isoptera. Mater. Org 12: 111 140. Williams R. M. C., and J. V. P. Morales. 1980. The effect of group size on the survival and feeding economy of pseudoworkers of building damaging Cryptotermes spp. (Isoptera, Kalotermitidae), p p 219233. In Jaisson, P. (ed) Social insects in the tropics. Volume 2. Paris, Universit de Paris Nord. Williams, O. L. 1934. Wood preference tests p p 572573, In C.A. Kofoid, S. F. Light, A. C. Horner, M. Randa ll, W. B. Herms and E. E. Bowe (eds.) Termites and Termite Control. Univ. California Press, Berkeley, CA. Wolcott, G. N. 1924. The comparative resistance of woods to the attack of the termite, Cryptotermes brevis Walker. Bull. 33. Univ. Porto Rico Agri. Exp. Sta. 15p. Woodrow, R. J., J. K. Grace. 1998 a. Field studies of high temperature to control Cryptotermes brevis (Isoptera: Kalotermitidae). Sociobiology. 32: 2749. Woodrow, R. J., J. K. Grace. 1998 b. Laboratory evaluation of high temperatures to control Cryptotermes brevis (Isoptera: Kalotermitidae). J. Econ. Entomol. 91: 905909.

PAGE 143

143 Xiong, B., and T. D. Kocher 1991. Comparison of mitochondrial DNA sequences of seven morphospecies of black flies ( Diptera: Simuliidae). Genome 34: 306 311. Yamano, K. 1987. Physical control of the Formosan Subterranean termite, Coptotermes formosanus Sh iraki. Proceedings of the international symposium on the Formosan Subterranean termite 67th meeting of the Pacific Branch Entomological Society of America, Honolulu, Hawai i 83: 4347. Yeoh B. H., C. Y. Lee. 2007. Tunneling responses of the Asian subterranean termite, Coptotermes gestroi in termiticide treated sand (Isoptera: Rhinotermitidae). Sociobiology. 50: 457468.

PAGE 144

144 BIOGRAPHICAL SKETCH Maria Teresa Monteiro da Rocha Bravo Ferreira was born in 1982, in Lisbon, Portugal, to a middle class family of a schoo l teacher and a computer technic ian. Being a school teacher daughter she quickly learned the discipline and love for learning. Being a straight A student, she never gave grievance to her parents and completed her early studies with honors. Between t he years of 1994 and 2000, she attended Liceu Cames, a very respected and traditional high school in Lisbon. There, she developed her love for science by taking several different laboratory courses in chemistry and biology. In 2000, she graduated from Hig h School and went on to be a freshman in college pursuing a B iology major. At the Animal Biology Department at the Faculdade de Cincias da Universidade de Lisboa (Science College of the University of Lisbon) she joined the small e ntomology group in 2003 doing volunteer work in insect capturing and sorting. She did a year study in the Azorean Isle of So Miguel working with Diptera diversity in her final year in college. She graduated from college in July 2005, but continu ed her volunteer work with the ent omology group until early 2006. In 2006, she moved to the Azorean Isle of Terceira where she began work as a technician on a project to determine management tools for the Cryptotermes brevis infestation in the Azores. She worked in this project until the end of that year. In 2007, she entered the graduate program at the University of Florida in the Department of Entomology and Nematology where she graduated with a m aster s degree in August 2008. She is currently seeking her PhD in e ntomology at the University of Florida.