Copyright is owned by the Author of the thesis.  Permission is given for 
a copy to be downloaded by an individual for the purpose of research and 
private study only.  The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 
SEQUENCES AND SIGNALS: EVOLUTIONARY HISTORIES 
OF NEW ZEALAND SKINKS. 
A thesis presented in partial fulfilment of 
the requirements for the degree of 
Doctor of Philosophy in Genetics 
at Massey University 
New Zealand 
Robert Eric Hickson 
May 1993 
Abstract 
The application of DNA sequencing to studies of the New Zealand biota is illustrated by investigations 
into the evolutionary relationships of skinks in the genus Leiolopisma. DNA sequences from a region of 
the mitochondrial 12S rRNA gene were determined for 20 taxa by use of the polymerase chain reaction. 
A vertebrate secondary structure model for this part of the gene was developed using comparative 
sequence analysis and calculations of RNA folding energies. Approximately one third of the molecule 
does not vary between the vertebrates examined, and there are similar patterns of sequence variability 
among the vertebrates. The secondary structure model was subsequently used to assist phylogenetic 
analyses of the skink sequence data set. 
Analyses of the mitochondrial DNA sequence information, using newly developed and more 
sophisticated algorithms, did not produce a fully resolved phylogenetic tree for all the skinks, though 
relationships between some taxa are less ambiguous. The lack of resolution does not appear to be due to 
limitations in the analytical methods, nor to the patterns of nucleotide substitutions in the skink 12S 
rRNA sequences. The skink sequence data set is unusual in that most of the taXa have similar numbers 
of nucleotide substitutions when compared to each other. These results are interpreted as reflecting a 
rapid divergence of the Leiolopisma group of skinks. Simulation studies support this interpretation. 
Three hypotheses are presented to account for the patterns of sequence differences between the skinks. 
One proposes a Gondwanan divergence of the Leiolopisma group, about 80 million years ago. Under 
this hypothesis the distribution of the skinks on islands in the Pacific and Indian oceans can be 
explained, in part, by continental drift. A second hypothesis suggests that New Zealand Leioiopisma are 
derived from a Miocene (15-25 million years ago) evolutionary radiation in New Zealand. This 
hypothesis however is inconsistent with observations of the sequence similarities between Mauritian, 
Australian, and New Zealand skinks. A third hypothesis, proposing several independent colonizations of 
New Zealand by Leioiopisma, is also not as well supported by the available sequence data. However, a 
close relationship between the l2S rRNA sequences of one New Zealand species L. inJrapunctatum and 
the Australian Lamprophoiis guichenoti suggests that at least two skink immigrations to, or emigrations 
from, New Zealand may have occurred. Predictions of the three hypotheses and strategies to test them 
are discussed. 
Some of the conclusions derived from analyses of the mitochondrial DNA sequences conflict with those 
obtained from allozyme information, though there are points of agreement. Comparison of the allozyme 
and sequence data do also revealed a case of hybridization between two sympatric species, L. n. 
polychroma and L. maccanni, at a site in Southland. Analyses of both data sets indicate that the 
morphological similarity of Leioiopisma species obscLI!es a large amount of genetic diversity, and the 
evolutionary histories of New Zealand Leiolopisma are older and more complex than previously 
considered. Further genetical and ecological studies of Leioiopisma are required, but this thesis 
emphasizes both the suitability and necessity of molecular genetic approaches for evolutionary 
investigations in New Zealand. 
Acknowledgments 
Barry Scott and Garry Latch are penultimately to blame for this thesis. None of us knew what we were 
really in for. 
:� Financial support was provided by the University Grants Committee, the Nga Manu Trust, Massey 
University, the Molecular Genetics Unit and Department of Microbiology and Genetics. Additional 
support came from the Dean of Science, the Royal Society of New Zealand, the Alfred P. Sloan 
Foundation (not once but twice!), and my parents. To all I am very grateful. 
ii 
lain Lamont and Liz Poole of Otago University provided hospitality and Liz showed me how to double­
strand sequence. Brad Shaffer, Centre for Population Biology, University of California Davis, permitted 
me time, space and consumables in his lab. Ross Sadlier of the Australian Museum, Sydney, provided 
tissue samples of New Caledonian skinks. 
. 
With Robert Fisher, Centre for Population Biology, University of California Davis, I enjoyed many 
hours of entertaining and interesting discussions, and he let me share a room with his lizards. Thanks 
Robert, but get on with your thesis! Mike Hendy, Mike Steel, Mike Charleston, Peter Waddell, and 
Chris Simon are also thanked for stimulating discussions and ideas. 
Charlie Daugherty, Victoria University, provided the initial inspiration for this study. He gave me access 
to the skink tissues and was ever willing to fill me in on the details of New Zealand skinks. 
Barry Scott, Tim Brown, and Geoff Malcolm provided a stimulating and friendly atmosphere in which 
to work, freedom to do as I pleased and to go where I wanted, and money to make it possible. Mister 
Tucker, Paul, Laura and Terri covered administration and clerical things. 
Thanks also to my travel agent Sue. But absolutely no thanks to Guido and the Linguine Gang, Milano 
Station, Italy, even though they too relieved me of a burden. 
Lots of people have been and gone, or are still around. Thanks to them for all sorts. Especially Carolyn 
and Miranda 
To the following, words are no recompense: 
David and Eric, both much more than superb and challenging supervisors and friends. 
Alan, Gina, and particularly Mary of "Capital City Campus". 
Stan, Pete, Trish and Liz of the Farside (and other places). 
But most of all I thank my parents, with much love. 
iii 
Table 0/ Contents 
Abstract .................................................................................................................................................. . 
Acknowledgements ................................................................................................................................... ii 
Table o/Contents .................................................................................................................................... iii 
List 0/ Figures ..................... ........ .............................................................. ............ ....... ............ ..... ......... vii 
List o/Tables ............................................................................................................................................. x 
Chapter One: Introduction •...•••...••.•.••••••••••••••••••...•....•••...•.•.•..•......••.••.••••••.••••••••..••••.•..••.•..•...•......•••.•..•... 1 
New Zealand in an Evolutionary Context ......................................................................................... 1 
The Oligocene ................................................................................................................................... 1 
The Pliocene ..................................................................................................................................... 2 
The Pleistocene ................................................................................................................................. 2 
The Holocene .................................................................................................................................... 2 
Evolutionary Studies in New Zealand ............................................................................................... 2 
Biogeographic Studies in New Zealand ............................................................................................. 3 
Assessing Taxonomic Relationships .................................................................................................. 4 
Secondary Metabolites ..................................................................................................................... 4 
Immunological Techniques .............................................................................................................. 4 
Chromosomal Studies ....................................................................................................................... 4 
Allozyme Studies ............................................................................................................................... 5 
DNA Fragments ................................................................................................................................ 5 
DNA Sequences ................................................................................................................................ 6 
Mitochondrial DNA and Evolution ................................................................................................... 6 
Rates of Evolution in Cytoplasmic Organelles ................................................................................. 7 
Limitations of Sequence Data ............................................................................................................ 8 
Rates of Sequence Evolution .............................................................................................................. 9 
Skinks as a Model Group for Evolutionary Studies ............ ............................................................ 9 
Views on the Origin of New Zealand Lizards ................................................................................... 9 
Ecological Aspects of Some Leiolopisma .......................................................................................... 10 
Genetical Investigations of Leiolopisma .......................................................................................... 11 
Chapter Two: Methods ......................•..••.........••...•••.•.....•......•••..•.....•....•.•••...•.•.•••..•...•.•.........•..•.•.••.......• 13 
The Polymerase Chain Reaction ...................................................................................................... 13 
Mitochondrial DNA Isolation .......................................................................................................... 14 
Total DNA Isolation .......................................................................................................................... 14 
Amplification of DNA by the Polymerase Chain Reaction ........................................................... 16 
Direct Sequencing of PCR Products ................................................................................................ 17 
RESULTS .................... ....................................................................................................................... 18  
iv 
Chapter Three: A Refined Secondary Structure Model 
for Domain III of Vertebrate 12S rRNA .................................................................................. 22 
Refining a Secondary Structure Model ........................................................................................... 22 
The Skink and Other Vertebrate 12S rRNA Sequences ............................................................... 23 
Approaches for Refining a Vertebrate Domain m Model . ........................................................... 23 
Energetics of the Folding of the rRNA Sequences ...................................................................... 23 
Identifying Individual Helices .......................................................................................................... 27 
Helix 35 .......................................................................................................................................... 27 
Helix 36 .......................................................................................................................................... 27 
Helix 37 .......................................................................................................................................... 27 
Helix38 .......................................................................................................................................... 29 
Helix39 .......................................................................................................................................... 29 
Helix40 .......................................................................................................................................... 29 
The Refined Secondary Structure Vertebrate Model .................................................................... 29 
Conserved Positions in the Secondary Structure Model ............................................................... 30 
Variability within Domain ITI .......................................................................................................... 30 
Nucleotide Substitutions in the Skink Structure ............................................................................ 30 
DISCUSSION .................................................................................................................................... 32 
Interactions Between Domain m and Ribosomal Proteins ...................................................... 32 
Tetra-loops .................................................................................................................................... 36 
Comparison of Secondary Structure Models ............................................................................. 36 
Limitations of the RNA Folding Algorithm ............................................................................... 37 
Variability bern'een Regions ........................................................................................................ 38 
Secondary Structure Models and Sequencing Precision .......................................................... 38 
Bias in Types of Substitutions ..................................................................................................... 39 
Chapter Four: Transitions and Transversions .......•.•...•............•....•....•...•.•...•..•....••..•.•.•.............•••••..•... 40 
Transitions and Transversions ........................................................................................................ 40 
Saturation of Transitions? ................................................................................................................ 40 
Simulations ......................................................................................................................................... 45 
Estimating Divergence from Transversions ................................................................................... 47 
Comparisons to Other Data Sets ............... ; ..................................................................................... 47 
Changes in the Context of Secondary Structure ............................................................................ 51 
DISCUSSION .................................................................................................................................... 52 
Rates of mtDNA Evolution .......................................................................................................... 53 
Estimating Degree of Divergence with Transversions .............................................................. 54 
Regions of Variability .................................................................................................................. 54 
v 
Chapter Five: Signals in the Data and Phylogenetic Analyses ............................................................. 56 
Spectral Analysis of DNA Sequences .............................................................................................. 56 
What is a Bipartition and How is it Represented? ......................................................................... 57 
Lack of Resolution in the Skink 12S rRNA Data Set ..................................................................... 60 
Spectral Analysis of other Vertebrate 12S rRNA Data Sets ......................................................... 60 
Parsimony and Neighbor-Joining Analyses of the Skink Data Set .............................................. 60 
Partially Resolved Relationships Using the Hadamard Conjugation ........................................... 61 
Weighting of Characters ................................................................................................................... 61 
Weighting o/Transversions .......................................................................................................... 62 
Weighting 0/ Paired Regions ........................................................................................................ 62 
Weighting 0/ More Variable Sites ................................................................................................. 64 
The Effect o/Constant Columns in Sequence Analysis .............................................................. 64 
The Effect on Skink Phylogeny o/Weighting o/Constant Columns .......................................... 65 
Summary of the Effect of Weighting Schemes on the Skink Phylogeny ...................................... 65 
Spectral Analysis of Simulated Sequences ...................................................................................... 66 
Relationships of Northern & Southern New Zealand Leiolopisma .............................................. 66 
Spectral Analysis of Skinks from Northern New Zealand ............................................................ 68 
The Complex Relationships of C. aenea .......................................................................................... 68 
Spectral Analysis of Southern New Zealand Skinks ...................................................................... 69 
Mitochondrial Lineages of the Skinks ............................................................................................. 69 
Addition of a New Caledonian Skink .............................................................................................. 69 
Cytochrome b Sequence Data .......................................................................................................... 71 
DISCUSSION .................................................................................................................................... 71 
"Rapid" Divergence of Leiolopisma ........................................................................................... 74 
Estimating Times of Divergence for the Skinks ........................................................................ 74 
Estimates from Immunological Data ......................................................................................... 75 
Estimates from Allozyme Data .................................................................................................. 75 
Estimates from Biogeographic Information .............................................................................. 76 
Hypothesis One: A Gondwanan Origin and Diversification .................................................... 76 
The Mauritian Leiolopisma are S till Enigmatic ....................................................................... 77 
Origins o/the New Zealand Skinks .............................................................................................. 77 
Diversification Within New Zealand .... : ..................................................................................... 78 
Hypothesis Two: Post-Oligocene Diversification ...................................................................... 78 
Phylogenetic Limitations of the Skink Allozyme Data ............................................................. 79 
Hypothesis Three: Multiple Leiolopisma Colonizations of NZ ................................................ 79 
Hypotheses for the testing ............................................................................................................ 80 
vi 
Chapter Six: Comparing Sequence and Allozyme Data ........................................................................ 81 
Conflict Between AUozyme and 12S rRNA Sequence Data ................................................ ... ....... 81 
Congruence Between Allozyme and 12S rRNA Sequence Data ...... ......... ........................ ....... ..... 81 
Hybridization .... .... ...... .................. ............. ................ ............................. ....... . ............... . . . ........ .. ....... 83 
DISCUSSION ....... .. ....................... .... ............................. ............................... .................................... 84 
Hybridization. Questions to Address .................. ....... ............. ................................. .... .......... '" 85 
Species Concepts ............... ... . .................... .... ............... .. ........................... .. . ................ ................. 87 
Origins of Some Specific New Zealand Leiolopisma .. ......... ............... ..................... ....... ........... 87 
Chapter Seven: Discussion ...................................................................................................................... 89 
Secondary Structure and Phylogenetic Analyses ......... ........ .... ......... . ........................................... . 89 
Evolution of the 12S rRNA Molecule ...... ................ .................................................................. ..... .  90 
Simulation Studies .................... ...... ................. .......... ......... ... . . ...................... . . ....... ........ ................... 91 
The Hadamard Conjugation and Spectral Analysis ........ ..... ............................. ........................ .... 92 
Incompletely Resolved Skink Phylogeny ................... . . .............. .... . . . ............................. ................. 92 
Hypotheses for the Origins of New Zealand Leiolopisma .............................. . ... ........................ .. .. 93 
Predictions of the Three Hypotheses ..... . . . . . . . . ......... . ...... . . ... ..... . . ................ . .. ... .. .... . ................. ........ 94 
Hypothesis One: Gondwanan Origins ........... ................. ................ ............................................... 94 
Hypothesis Two: Oligocene Drowning ... .......... .... .... ....... .... ................ ... ....... ............... ................. 94 
Hypothesis Three: Multiple Colonists . ............ . ........... .. ... .... ... ..... ................ . .... ..... .... . .. .. ....... ........ 94 
Testing the Hypotheses: What Other Skinks to Sample? ............................................. ..... .... ....... 95 
Testing the Hypotheses: What Other Genes to Sample? ............. .... . ........... . ................ .......... ...... 95 
Investigation of Nuclear Gene Sequences .......... ........ ... . .... . . ... ........ .... ... . ... . ...................... ............ .. 96 
Hybridization between L. n. polychroma and L. maccanni ...... .......... ...... ............. .. . ..................... 97 
The Potential of L. n. polychroma For Population Studies . . ... .................. . .... . . ........... ................. .. 98 
Rates of Molecular and Morphological Evolution in Skinks ...................... ........... ........ ....... . ....... 98 
Evolutionary Investigaiions of New Zealand Biota .. ............ ........................ .. ... .... ... .. ... .... .......... .. 99 
References ............................................................................................................................................. 101 
Appendix 1 ........................ . ................. ...................................................................................... on diskette 
a). Skink 12S rRNA sequences 
b). Skink cytochrome b sequences 
Appendix 2 .................................................................................. .............................................. on diskette 
a). Xantusiid 12S rRNA sequences 
b). Ratite 12S rRNA sequences 
c). Bovid 12S rRNA sequences 
d). Great Ape 12S rRNA sequences 
Figure 2.1 
Figure 2.2 
Figure 3.1 
Figure 3.2 
Figure 3.3 
Figure 3.4 
Figure 3.5 
Figure 3.6 
Figure 3.7 
Figure 4.1 
Figure 4.2 
Figure 4.3 
Figure 4.4 
Figure 4.5 
Figure 4.6 
Figure 4.7 
Figure 4.8 
Figure 4.9 
Figure 4.10 
Figure 4.11 
Figure 4.12 
Figure 4.13 
List of Figures 
Following page 
Map of New Zealand. 15 
Inferred phylogenetic relationships of vertebrates based on 12S rRNA sequences. 19 
Secondary structure of  prokaryote and mitochondrial small subunit rRNA. 
Potential pairings for helix 36 in vertebrate 12S rRNA. 
Secondary structure model for domain III of skink 12S rRNA. 
Potential alternative secondary structures for skink domain Ill. 
Conserved positions in vertebrate 12S rRNA domain III sequences. 
Distribution of variable sites in domain III of vertebrate 12S rRNAs. 
Regions of variability in the skink 12S rRNA sequence. 
Frequency distributions for total nucleotide differences, transitions 
and transversions in skink 12S rRNA sequences. 
Numbers of transversions versus numbers of transitions for skink 12S rRNA 
sequences. 
Total number of differences versus numbers of trans versions for skink 
12S rRNA sequences. 
Relationship between numbers of transitions and numbers of transversions 
for ratites and pecoran bovids. 
Proportion of transition substitutions in skink 12S rRNA sequences. 
Relationship between number of transversions and the transition/transversion 
ratio for the skink data set. 
Relationship between number of trans versions and the transition/transversion 
ratio for other data sets. 
Frequency distribution of transitions and trans versions in simulated sequence 
data. 
Relationship between numbers of transitions and numbers of transversions 
in simulated sequence data. 
Relationship between ilUmbers of transitions and numbers of transversions 
in simulated sequence data using 40 and 55 expected substitutions. 
Relationship between numbers of transitions and numbers of transversions 
for other simulations using Tsrrv ratios of 2/1 and 8/1. 
Frequency distributions of transitions and transversions in simuiated 
sequence data for a large number of iterations. 
Relationship between numbers of transitions and numbers of transversions 
in simulated sequence data when 500 iterations are performed. 
22 
28 
29 
29 
30 
31 
31 
42 
44 
44 
44 
44 
45 
45 
46 
46 
46 
46 
46 
46 
vii 
viii 
Figure 4.14 Means and standard deviations of estimates for the real number 
of substitutions. 47 
Figure 4.15 Means and standard deviations of estimates for the real numbers 
of substitutions for other simulations using 40 and 55 expected 
substitutions. 47 
Figure 4.16 Means and standard deviations of estimates for the real numbers of 
substitutions for other simulations using Ts/Tv ratios of 2/1 and 8/ l .  47 
Figure 4.17 Variable sites in the skink 12S rRNA domain III. 51 
Figure 4.18 Location of nucleotide substitutions in 12S rRNA for pairs of skink taxa 
with a small number of differences between them. 51 
Figure 4.19 Comparisons of regions of variability in the 12S rRN A sequence for 
pairs of closely and more distantly related skink taxa 51 
Figure 4.20 Comparisons of regions of variability in the 12S rRNA sequence for 
closely and more distantly related ratites, and great apes. 51 
Figure 5.1 Spectral analysis of skink 12S rRNA sequences. 60 
Figure 5.2 Spectral analyses of ratite and bovid 12S rRNA sequences. 60 
Figure 5.3 Maximum parsimony tree for skink 12S rRNA sequences. 61 
Figure 5.4 Neighbor-joining tree for skink 12S rRNA sequences. 
, 
61 
Figure 5.5 Inferred phylogenetic relationships of 20 skink taxa based on spectral ) 
analysis of 12S rRNA sequences. 61 
Figure 5.6 Effect of weighting transversions in skink data set 62 
Figure 5.7 Inferred phylogenetic relationships of 20 skink taxa based on spectral 
analysis of 12S rRNA sequences (transversions weighted four 
times higher than transitions). 62 
Figure 5.8 Effect of weighting upaired regions three-fold higher than unpaired regions 
in skink 12S rRNA. 63 
Figure 5.9 Effect of down-weighting the eight most variable sites in the skink 12S rRNA. 64 
Figure 5.10 Effect of removing the eight most variable sites. 64 
Figure 5.11 A resolved tree of 10 skink taxa. in which the two strongest 
signals are not in the tree. 65 
Figure 5.12 Effect on the 10 taxa spectrum of removing constant sites from the sequence. 65 
Figure 5.13 Spectral analysis of simulated sequences. 66 
Figure 5.14 Spectral analysis of northern and southern skink 12S rRNA sequences. 66 
Figure 5.15 Bipartitions in the optimal tree for 11 northern New Zealand skinks. 68 
Figure 5.16 Bipartitions in the optimal tree for 10 northern New Zealand skinks. 68 
ix 
Figure 5.17 Bipartitions in the optimal tree for 10 southern New Zealand skinks. 69 
Figure 5.18 Inferred phylogenetic relationships for northern and southern 
New Zealand skinks based on 12S rRNA sequences. 69 
Figure 5.19 Influence of the first 80 bases on spectral signals in the 
skink data seL 70 
Figure 5.20 Spectral analysis of 19 Leiolopisma skinks and the New Caledonian 
Tropidoscincus rohssii. 71 
Figure 5.21 Inferred phylogenetic relationships of four skink taxa from 
Southland using 12S rRNA and cytochrome b sequences. 72 
Figure 5.22 Three hypotheses for the origin of Leiolopisma skinks. 76 
Figure 6.1 Evidence for sexual hybridization in a skink population. 83 
Figure 6.2 Potential offspring from sexual hybridization. 84 
x 
List of Tables 
Page 
Table 2.1 Skink taxa from which DNA sequence infonnation was obtained. 15 
Table 2.2 Base composition of the skink 12S rRNA sequences. 21 
Table 3.1 Alignment of 20 skink 12S rRNA sequences. 24 
Table 3.2 Species and sources of 12S rRNA sequences used in Table 3.3 24 
Table 3.3 Alignment of vertebrate 12S rRNA sequences. 26 
Table 3.4 Local minimal free energies for folding of helices. 28 
Table 3.5 Distribution of conserved nuc1eotides in vertebrate 12S rRNA. 31 
Table 3.6 Changes occurring in paired and unpaired regions of skink 12S rRNA. 31 
Table 3.7 Substitutions among paired nucleotides in skink 12S rRNA. 33 
Table 3.8 Observed and expected numbers of transitions and transversions in 
the skink data set 34 
Table 3.9 Base composition of conserved residues in skink 12S rRNA. 35 
Table 4.1 Distance matrices for skink 12S rRNA sequences. 41 
Table 4.2 Mean numbers of total differences, transitions, transversions, 
and transitiOn/transversion ratios for each taxon. ' 43 
Table 4.3 Variability in total substitutions, transitions and Ts/Tv ratio in 
relation to numbers of transversions. 44 
Table 4.4 Parameters and results of simulations. 46 
Table 4.5 Distance matrices for 12S rRNA from other vertebrates. 48 
Table 4.6 Comparison of estimated times of divergence among the skinks. 50 
Table 5.1 Detennining a relative weighting factor for paired regions. 63 
Table 5.2 New Zealand Leiolopisma divided into northern and southern groups. 67 
Table 5.3 Mitochondrial lineages in skinks. 70 
Table 5.4 Distance matrices for cytochrome band 12S rRNA sequences from four 
skink taxa. 72 
Table 6.1 Comparisons of genetic distances inferred from 12S rRNA sequences 
and allozymes. 82 
pagel. 
Chapter One: Introduction 
New Zealand's biota is a largely untapped resource for evolutionary investigations. Despite the fact that 
our flora and fauna rank alongside life forms from the islands of Hawaii, Madagascar, New Guinea and 
the �pagos' in furnishing vivid examples of evolution in action (Hooker 1853, Hutton 1872, Wallace 
1880, Fleming 1958, Carlquist 1974), this potential has not yet been fully exploited. Some reasons for 
this are discussed below. This thesis illustrates the importance of New Zealand and its biota for 
addressing evolutionary questions. It does this by examining mitochondrial DNA sequences and from 
this data inferring aspects of the evolutionary history and relationships of skinks (Reptilia: Lacertilia: 
Scincidae) in the genus Leiolopisma. It also investigates what phylogenetic information can be obtained 
from DNA sequences by consideration of the sequences' secondary structure and by the use of new 
sequence analysis algorithms. 
The major conclusions of this thesis are that firstly, phylogenetic trees are hypotheses only, and close 
examination of DNA sequences is necessary to have confidence in the reliability of phylogenetic 
relationships. New DNA sequence algorithms are used which allow exploration of information in 
sequences. Secondly, the use of molecular genetic techniques are valuable for investigations of 
evolutionary problems in New Zealand, particularly when applied in conjunction with ecological and 
population genetic studies. Thirdly, that the evolutionary history of New Zealand skinks is more 
complex and fascinating than previously suspected. Use of model groups, such as New Zealand skinks, 
is essential for an understanding of the patterns and processes of evolution. 
New Zealand in an Evolutionary Context. 
New Zealand contains great geological variety and complexity, diverse topography and habitats, and has 
been, and is being, subjected to extensive geological, climatic and ecological disturbances (Gage 1961, 
Suggate et al. 1978, Fleming 1980, Burrows & Greenland 1979, Mildenhall1980, Stevens 1980, 
Pocknall1989, 1992). In addition to the origin and separation of New Zealand from Gondwanaland, 
four periods have been suggested to have had a significant effect on the biota. 
The Oligocene 
During the Oligocene ( 25-35 million years ago [MY AD, New Zealand was a series of low lying islands, 
and fluctuating sea levels changed their size and number (Suggate et al. 1978; Cooper, A., Chambers, 
G.K., Cooper, R.A. in prep.). Extinction of species may have been high at this time because of the due 
to rising sea levels (Cooper et al. in prep.). Subsequent tectonic uplift, and increases in land area and 
relief, temperature, and habitat diversity during and after the Miocene (15-25 MY A) may have promoted 
conditions suitable for diversillcation of species (see Suggate et al. 1978, Cooper et al. in prep.). 
Chapter I, page 2. 
The Pliocene 
The Pliocene (1-15 MY A), marked the start of the rise of the southern alps in New Zealand (Suggate et 
aI. 1978, Fleming 1980). With no previous alpine environment, the origins of the alpine flora has been 
considered problematical. Raven (1972) suggested the alpine flora came from Australia. but adaptation 
of indigenous New Zealand plants is more likely (Fisher 1965, Wardle 1978, Fleming 1980). During the 
Pliocene the north of the North Island was a series of archipelagoes, which changed in size and shape. 
The diversity of beetles and land snails now found there may be due to such geological changes (Powell 
1949, Holloway 1963, Climo 1978; see also Fleming 1980). 
The Pleistocene 
Pleistocene glaciations are considered to have had an important effect upon the distribution and 
speciation of many groups both here (powell 1949, Willett 1950, Dell 1955, Wardle 1963, 1988, 
Burrows 1965, Bigelow 1967, Irving 1%7, Petersen 1968, Bull & Whitaker 1975, Fleming 1980, Solem 
et al. 1981), and elsewhere (for example, Sylvester-Bradley 1963, Haffer 1969, Mayr & O'Hara 1986). 
In New Zealand, the existence of several "refugia" for cold intolerant species during glacial periods has 
been suggested (Burrows 1965, Wardle 1963, Fleming 1980). McGlone (1985) argues however that 
plant distributions in New Zealand may have been influenced more by pre-Pleistocene tectonic activity 
than by glacial advances. Evaluating these hypotheses requires accurate assessments of both taxonomic 
relationships and times of divergence. Genetic investigations of Australian and European fauna indicate 
that the effects of Pleistocene glaciation may have had less influence on the speciation of some groups 
than previously assumed (Roberts & Maxson 1985, Wallis & Arntzen 1989). Similar evaluations for the 
New Zealand biota are lacking, and the phylogenetic relationships of many New Zealand groups are still 
poorly known (Solem et al. 1981, Fife 1985, Wardle 1988). 
The Holocene 
Human settlement of New Zealand over the last millennium has also had a major influence on the biota, 
through direct action and indirectly via introduced plants and animals (Molloy et al. 1963, Anderson 
1983, McGlone 1983, Holdaway 1990). Evolutionary studies have relevance in this context. They can 
identify taxonomically important populations in urgent need of conservation (for example, Daugherty et 
aI. 1990a). Investigations of genetic diversity within and between populations is also essential to 
develop rational conservation strategies (Vrijenboek et al. 1985, Cohn 1990, May 1990, V ane-Wright et 
aI. 1991). 
Evolutionary Studies in New Zealand 
How and why species may change have been illustrated by studies in the Galapagos and Hawaiian 
islands (see for example Berry 1984, Simon 1987, Grant & Grant 1989), and also by detailed 
investigations of Drosophila species (see Lewontin et al. 1981, DeSalie & Hunt 1987). Despite 
recognition that New Zealand is an important place to investigate evolutionary patterns and processes 
(Hooker 1853, Hutton 1872, Wallace 1889, Cockayne 1911, Godley 1949, Hair 1966, Fleming 1958), 
the biota have not yet been examined in sufficient detail to complement Hawaiian and Galapagan 
studies. 
Chapter 1. page 3. 
New Zealand's antiquity, continental origins, and the large size of some of its islands make it different 
from the islands of Hawaii and the Galapagos' . Studies of the origins of New Zealand's biota and how 
they have evolved may therefore offer different perspectives on patterns and processes of evolution. 
Studies of evolutionary relevance have been conducted here but have been primarily descriptive and 
species-based, rather than being of a more theoretical and experimental nature. Few of the studies have 
directly examined the evolutionary concepts established by Lyell (1830), Darwin (1859), and Wallace 
(1889), and subsequently elaborated upon by Dobzbansky (1941), Stebbins (1950), Mayr (1963, 1982), 
Lewontin (1974), Kimura (1983) and Nei (1987). Exceptions are studies of alpine plants (Fisher 1965, 
Raven & Raven 1976, Ornduff 1964), morphological and genetic variation among introduced birds 
(Baker 1975, 1992, Ross 1983, Baker & Moeed 1987), floral biology (Lloyd & Yates 1982, Lloyd & 
Webb 1986), and behavioural ecology (for example, Craig 1984). Recent advances in molecular genetic 
techniques are however leading to renewed interest in evolutionary investigations here. 
Biogeographic Studies in New Zealand 
One of the major problems facing evolutionary studies in New Zealand is establishing from where, and 
when, many elements of our biota came. Biogeographic studies in New Zealand (see Fleming 1980, 
McGlone 1985, Craw 1989) laid the groundwork for examination of the evolution of species. As a crude 
distillation two, often competing, views of the history of New Zealand's biota have been presented. 
f4�+ 
"Traditional" biogeographers (Fleming 1980) have assum�much of the flora and fauna arrived here 
after the separation of the New Zealand land mass from Gondwanaland, approximately 80 MY A 
(Lawver et al. 1991), and various means and routes of dispersal have been proposed to account for their 
arrival (Fleming 1980, McGlone 1985, see also Kuschel 1975). Panbiogeographers (see Craw 1989) 
took the alternate view that present distributions may reflect more about past geographical connections 
than dispersal. Both perspectives are limited by uncertainty about the taxonomic relationships of some 
of the biota that they discuss. Inaccurate assessments of the degree of genetic separation between groups 
results in speculative accounts of dispersal routes (for example Hardy 1977), or biological scenarios in 
conflict with geological evidence (see Cooper 1989). 
The tendency to create "narratives" (Craw 1988) rather than testable hypotheses is also evident in other 
discussions about the evolution of New Zealand's biota . 1'B8 rale 6f hjlb:tidizatiOlJ for example" .A.C\ 
notable feature of many New Zealand plant groups is their readiness to hybridize (see Allan 1961, Fisher 
1965, Carlquist 1974). Rattenbury (1962) and Raven & Raven (1976) suggested that hybridization has 
played a major role in the evolution of New Zealand plants, but the true extent of hybridization has often 
been poorly investigated (Connor 1985). Failure to recognize that many of the hybrids occur in 
disturbed habitats may also have overemphasized the role of hybridization in New Zealand (Hair 1966). 
Wardle (1988, and Wardle et al. 1988) attempted to explain some disjunct plant distributions in terms of 
long distance pollen dispersal, interspecific hybridization and subsequent environmental selection to 
reconstitute a parental genotype. Genetic investigations to determine the taxonomic status of such 
Chapter 1. page 4. 
putative hybrids have not yet been conducted, so proposals involving ill defined selection processes are 
not warranted at this stage. Similar uncertainty over taxonomy is also evident in discussions of the 
evolution of some animal groups, such as land snails (Solem et al. 1981, Climo 1989). 
Assessment of hypotheses about distributions and speciation requires information on genealogical 
relationships. Taxonomic investigations in New Zealand have relied primarily upon morphological 
comparisons, and as already noted, have not always adequately resolved relationships. The first priority 
for evolutionary studies then, is to reliably determine phylogenies. 
Assessing Taxonomic Relationships 
Morphological variation is at the core of the theory of descent with modification by means of natural 
selection (Wallace 1858, Darwin 1859, 1875, Mayr 1942, 1963, S tebbins 1950). Morphological 
characters however, being only indirect manifestations of the genotype, can be uninformative or 
misleading about the closeness of evolutionary relationships (for example, King & Wilson 1975, Systma 
& Gottlieb 1986, Fitch & Atchley 1987, Wayne et al. 1989, Daugherty et al. 1990a, 1990b). 
Biochemical and cytological investigations of taxonomy are an essential complement to morphological 
studies (Hillis 1987, Patterson 1987, Hillis & Moritz 1990). These techniques are described below, and 
examples of their applications in New Zealand noted. 
Secondary Metabolites 
Analysis of secondary metabolites have been used for taxonomic purposes and provide an interesting 
adjunct to other phylogenetic methods (Cronquist 1980, Wright 1980). Chemotaxonomy has been used 
to study several New Zealand plant groups (for example, Taylor 1964, Markham & Godley 1972, 
Wilson 1984, see also Connor 1985), but genetic methods are required to resolve their relationships. 
Immunological Techniques 
. Immunological affinities (Champion et al. 1974, Maxson et al. 1990), and DNA-DNA hybridization 
(Sibley & Ahlquist 1987), are useful for establishing taxonomic affmities, particularly between more 
distantly related groups. They provide little insight however into the genetical processes at work within 
populations. The relationships of New Zealand Leiopelma frogs (Daugherty et al. 1981)  and the short­
tailed bat (Mystacina tuberculata; Pierson et al. 1986) have been investigated immunologically. DNA­
DNA hybridization studies have been conducted for ratites and a few New Zealand passerines (Sibley & 
Ahlquist 1981, 1987, Sibley et al. 1982). 
Chromosomal Studies 
Examination of chromosomal organisation can be useful both for systematics and for inferences about 
processes (Dobzhansky 1941, White 1978, Grant 1981,  Carson 1983, Moritz 1986, Systma 1990), but it 
is difficult to prove that such changes are causally related to the formation of new species (Bush et al. 
1977, White 1978) �ther than being subsequent effects of isolation (see Endler 1986). Chromosomal 
atlases for many New Zealand plant species have been compiled (see Frankel & Hair 1937, Hair 1966, 
Chapter I, page 5. 
1977, Rendle & Murray 1989), but the range and patterns of chromosomal changes have not been 
extensively examined. Hair (1966) noted that chromosomal instability was associated with advancing 
fronts ofpodocarp dispersal but this has not been investigated closely. Chromosomal studies of 
blackflies (McLea & Lambert 1985) and Leiopelma frogs (Green 1988, Green & SharbeI 1988), and 
preliminary descriptions of skink chromosomes (Hardy 1977) have been published, while karyological 
investigations of weta and other invertebrates are underway (M. Richards pers. comm.). 
Allozyme Studies 
Enzyme electrophoresis is also applicable to studies of evolutionary patterns and processes. The 
potential resolving power of a well conducted allozyme survey can exceed both restriction mapping and 
nucleotide sequencing studies if 30 or so loci are examined (Nei 1987). The relative simplicity, speed, 
cost effectiveness and the number and range of nuclear loci which can be surveyed by protein 
electrophoresis give it great practical value (Powell 1975, Buth 1984, Murphy et al. 1990). The extent of 
genetic variability revealed by this technique however depends upon assay conditions, such as 
temperature, pH and buffer solutions (see Nei 1987). Allozyme surveys can be used for both phylogeny 
reconstructions and for analyses of population structuring and dynamics (powell 1975, Spieth 1975, 
Gottlieb 1984, Murphy et al. 1990). Analysis of electrophoretic data for phylogeny reconstruction can 
present problems though because of the uncertainty over whether to use similarity or distance 
information, and whether to code the data in terms of presence or absence of alleles or in terms of allele 
frequencies (Buth 1984, Swofford & Olsen 1990). Assumptions used in calculations of genetic distance 
estimates from allozyme data can also be unrealistic and lead to inaccurate views of relationships (Hillis 
1984, Swofford & Olsen 1990). 
In New Zealand the taxonomic status of molluscs (Phillips & Lambert 1990), freshwater fish (Allibone 
1990), Leiopelma frogs (Daugherty et al. 1981, Green et al. 1989), skinks (Vos 1988, Daugherty et al. 
1990b, C.H. Daugherty & G.B. Patterson pers. comm.), tuatara populations (Daugherty et al. 1990a), 
and ducks (Hitchmough et al. 1990), have been investigated with allozymes. Relationships of New 
Zealand weta (M. Richards pers. comm.) and geckos are also being studied (R. Hitchmough pers. 
comm.). The techniques have also been used to assess the genetic diversity of molluscs (Freeth & Sin 
1986, Smith et al. 1989), marine fish stocks (e.g., Smith & Johnston 1985), introduced birds (Ross 1983, 
Baker & Mooed 1987, Baker 1992), the kakapo (Triggs et al. 1989), possum (Triggs & Green 1989), 
rimu (Dacrydium cupressinum; Hawkins & Sweet 1989) and beech (Nothofagus species; Wilcox & 
Ledgard 1983, Haase 1992). 
DNA Fragments 
Restriction mapping of nuclear or organelle DNA had a central role in the development of molecular 
systematics (Wilson et al. 1985, 1989, Nei 1987). Specillc pieces of nuclear DNA can be investigated 
(for example, Hillis & Davis 1987, Crowhurst et al. 1990), but it is often easier and more informative to 
isolate and use cytoplasmic genomes. Phylogenetic inferences from cytoplasmic DNA can be made by 
simply noting the occurrence and sizes of fragments cut by a range of enzymes, but more accurate 
Chapter 1, page 6. 
reconstructions are obtained by mapping the restriction endonuclease sites (Wilson et al. 1985, Crozier 
1990). Restriction analyses require relatively large amounts of DNA, and can have limited resolving 
power, particularly for more distant divergences (Wilson et al. 1989, Hillis & Moritz 1990). Recent 
developments for rapidly obtaining and sequencing specific DNA sequences (see below) are resulting in 
a decline in use of restriction analyses for phylogenetic studies. 
DNA Sequences 
Information from protein or nucleotide sequences has been extensively utilized to study taxonomic 
relationships over both vast tracts of evolutionary time (for example, Goodman et al. 1987, Gray 1989, 
Lake 1990, Martin et al. 1993) and much briefer periods (for example, Hahn et al. 1986, Koop et al. 
1989, Henderson et al. 1989, Vigilant et al. 1990, Miyamoto & Goodman 1990). The availability of 
sequence data permits direct examination of the the basis of genetic variation, and the widespread 
application of sequencing has led to a renaissance in systematics (patterson 1987 & 1990). 
The effort and expense involved in 'conventional' approaches to isolating and sequencing genes 
(Sambrook et al. 1989) prohibited their application to population studies, although some attempts have 
been made (e.g., Hahn et al. 1986). The development of the polymerase chain reaction ("PCR", Saiki et 
al. 1988, Ambeim & Erlich 1992) however now provides a system which can bring nucleotide 
sequencing to a level matching or exceeding the utility of allozymes in terms of ease, speed, and sample 
handling (Gyllensten & Erlich 1988, Kocher & White 1989, White et al. 1989, Arnheim et al. 1990). 
The procedure (see Chapter Two) is elegantly simple (in theory), allowing specific regions of DNA to 
be isolated and sequenced within hours, and from organisms for which no detailed genetic information 
is available. Consequently it is being rapidly employed for both taxonomic investigations and 
assessment of sequence diversity (for example, Kocher & White 1989, Kocher et al. 1989, White et al. 
1989, Ambeim et al. 1990, Simon et al. 1991), and it is the method used in this study. 
Mitochondrial DNA and Evolution 
The mitochondrial DNA (mtDNA) of animals in particular has several features which facilitate analysis 
of evolutionary divergence over periods of time ranging from one generation to about 100 million years 
(Moritz et al. 1987, Harrison 1989, Kocher et al. 1989). It is, firstly, the epitome o� economy - small and 
simple. Only 14 to 39 thousand nucleotide pairs ("kbp") in size, it contains information for 22 transfer 
RNAs, two ribosomal RNAs and about 13 open reading frames. In addition it lacks intervening 
sequences, and recombination appears to be infrequent ( Brown 1983, Moritz et al. 1987). Insertion and 
deletion events tend to occur in the control region (Moritz et al. 1987). Gene order can vary among 
invertebrates (Cantatore et al. 1987, Thomas et al. 1989), and among vertebrate groups (Desjardins & 
Morais 1990, Moritz 1991, Paabo et al. 1991). The only region of non -coding DNA is the control region 
or D-loop (D for displacement), around the site of initiation of mtDNA replication. Parts of this region 
are the most rapidly changing in the mitochondrial genome, though conserved domains occur within it, 
and consequently it is used for examination of intra- and inter-population diversity (Cann et al. 1984, 
Vigilant et al. 1991 ,  Wilkinson & Chapman 1991,  Ward et al. 1991). The protein-encoding and 
Chapter I, page 7. 
ribosomal RNA mitochondrial genes are being used to study older divergences (for example, Irwin et al. 
1991 ,  Simon et al. 1991,  Smith & Patton 1991, Ballard et al. 1992, Cooper et al. 1992, Hickson et al. 
1992, Pashley & Ke 1992). 
The predominance of maternal mitochondrial inheritance, absence of recombination and the apparent 
rzrity of different types ofmtDNA within an organism (but see Bermingham et al. 1986, Gyllensten et 
al. 1991, Hoeh et al. 1991) make analyses simpler than for nuclear loci (Brown 1983, Wilson et al. 
1985, Moritz et al. 1987). The relatively rapid rate of sequence evolution of animal mtDNA in 
comparison to nuclear rates is also advantageous when comparing closely related groups. In some 
vertebrates the rate of mitochondrial genome evolution can exceed that of single-copy nuclear DNA by 
up to ten-fold (Brown et al. 1979, 1982, Wilson et al. 1985), but this is not universal (Britten 1986, 
Vawter & Brown 1986, Thomas & Beckenbach 1989, Goddard et al. 1990, Palmer 1990). 
Rates of Evolution in Cytoplasmic Organelles 
Reasons for variation in the rate and types of changes in mtDNA are unknown, but such variability 
implies that the small mitochondrial genome size in itself is insufficient to explain rapid rates of change. 
Chloroplast genomes are also relatively small but change at a slower rate than animal mtDNA (palmer 
1990). Other possible causes for a fast rate of evolution, such as reduced replicative fidelity and 
translational constraints, and the absence of recombination (Cann et al. 1984, Wilson et al. 1985, 
Clayton 1982), have not been investigated in detail. 
Cytoplasmic genomes of plants and fungi are larger and more complex than animal mtDNA and do not 
exhibit such rapid sequence changes (although fungi are poorly studied; Gray 1989). Nor do they appear 
to have the large bias (10:1) in transitions (changes between nucleotides of the same class, for example 
adenine to guanine) over transversions (for example adenine to cytosine) found in some animal mtDNA 
(Brown 1983, Palmer 1990). Chloroplast DNA and fungal mtDNA can however be used for population 
comparisons (Taylor 1986, Soltis et al. 1989, Clegg & Durbin 1990, Palmer 1990). Plant mitochondrial 
genomes can be both very large (sizes up to 25 ()()() kbp have been reported) and can undergo rapid and 
extensive organisational rearrangements without accompanying sequence alteration (Palmer & Herbon 
1988). One reason for the latter may be due to the creation of a panmictic mtDNA population by fusion 
of mitochondria within cells with the consequent reduced probability of fixation of mutations (Lonsdale 
et al. 1988). Paternal transmittance of mitochondria and chloroplasts can occur (for example, Harrison & 
Doyle 1990, Kondo et al. 1990, Gyllensten et al. 1991), but maternal inheritance predominates (Wilson 
et al. 1985). 
Maternal inheritance and rapid rates of change in animal mtDNA make it possible to investigate gene 
evolution over relatively short periods of time, as well as permit examination of geographic structuring 
and gene flow in populations (reviewed by Wilson et al. 1985, Avise et al. 1987, Avise 1989a, 1991 ,  
Harrison 1989, Slatkin 1989). The haploid state and uniparental acquisition ofmtDNA also make it 
more sensitive than nuclear DNA to changes in population size, so low mtDNA variability can be 
Chapter I, page 8. 
indicative of past reductions in population size (Wilson el al. 1985). Examination of more distant 
relationships using mtDNA sequences include studies by Ballard et al. (1992), Meyer & Dolven (1992), 
Irwin el al. (1991), and Allard el al. (1992). Several molecular systematic studies are now underway in 
New Zealand. Published reports are available for ratites (Cooper el al. 1992) and skinks (Hickson el al. 
1992). DNA fingerprinting of sea bird populations have also been initiated (Millar el al. 1992) and 
satellite DNA has been examined in the native frog Leiopelma hochstetteri (Zeyl & Green 1992). 
Limitations of Sequence Data 
Analysis of mtDNA sequences is a convenient starting point to investigate relationships among animals, 
but it should be borne in mind that the uniparental inheritance of mtDNA means that only the 
evolutionary history of the matemal line (and only a small part of that) can be examined. Reliance on 
single genes can be misleading (penny et al. 1982, Wilson et al. 1987, Wyss el al. 1987). Examination 
of short regions of DNA also reflects the evolutionary history of that gene rather than the organism (Nei 
1987, Pamilo & Nei 1988, Martin el al. 1990). Confidence in a phylogeny requires congruence from 
other data sets, such as morphology and allozymes (Hillis et al. 1987, Patterson 1987, Hillis & Moritz 
1990). Sequence information from several thousand base pairs and from unlinked loci is also preferred 
(Saitou & Nei 1986, Nei 1987, Martin el al. 1990), but sequencing studies employing the PCR usually 
concentrate on sequences under one kilobase pairs in length (Kocher el al. 1989). 
Study of nuclear sequences with PCR may require separation of allelic variants by denaturing gel 
electrophoresis (Sheffield et al. 1990), use of allele specific primers (Gyllensten & Erlich 1988), or 
"single molecule PCR" (Jeffreys et al. 1990) but are technically more demanding. Use of nuclear 
ribosomal sequences (Hillis & Dixon 1991)  and intron sequences (palumbi & Baker submitted) are 
proving useful however for examination of older and more recent diveregences, respectively. 
Generating sequence data is now becoming routine. The primary difficulties lie in data analysis. Most 
phylogenetic trees derived from sequence data are probably incorrect (penny et al. 1990, Rohlf et al. 
1990), since no current algorithm meets all the necessary criteria of being fast, efficient, consistent, 
robust, and falsifiable (Henderson et al. 1989, Penny et al. 1990). Reviews of phylogeny reconstruction 
generally recognise the limitations, though users of the programs may not (Felsenstein 1988, Swofford 
& Olsen 1990, Cracraft & Helm-Bychowski 1991,  Nei 1991,  Penny et al. 1992, Stewart 1993). 
Phylogenetic trees should be regarded as hypotheses and subject to error. Statistical analyses are being 
developed to assess the reliability of phy logenies (Li & Gouy 1991). The emphasis of this thesis is on 
what information can be extracted from DNA sequences, and confidence in relationships is obtained by 
examination of conflicting associations of taxa using newly developed sequence analysis algortithms 
(penny et al. 1992, 1993). This is discussed in more detail in Chapter Five. 
Chapter 1. page 9. 
Rates of Sequence Evolution 
While patterns of evolution may be intrinsically interesting, understanding evolutionary processes also 
requires a temporal framework. The apparent regularity in the accumulation of amino acid and 
nucleotide substitutions led to the suggestion that evolutionary time can be measured from sequences 
(Zuckerkandl & Pauling 1965, Wilson et ai. 1977 & 1987, see also Ingram 1961). There is a growing 
awareness however that single genes or proteins may not be reliable as markers of time (see Wilson et 
ai. 1987, Easteal 1990), that rates are lineage dependent (fhotpe 1982, Goddard et ai. 1990, Palmer 
1990), and that models used to describe the process of substitution rates require more investigation 
(Gillespie 1986, Wilson et ai. 1987). Caution must be used if rate estimates are extrapolated from one 
group, such as mammalian mtDNA, to another (see Wilson et ai. 1985, 1987). 
Skinks as a Model Group for Evolutionary Studies 
Integrated approaches are essential to an understanding of evolutionary patterns and processes. One 
emphasis of this thesis is on developing a "model" system with which to investigate in detail the 
workings of evolution, as has been done for Drosophila (Lewontin et ai. 1981 ,  DeSalle & Hunt 1987), 
and Galapagos' finches (Grant & Grant 1989). 
Regrettably few studies have taken a comprehensive view of evolution by utilizing ecological, 
morphological, biochemical and genetic data for a specific group (but see Larson 1984, DeSalle & Hunt 
1987, Sytsma 1990). As well as enhancing our understanding of how organisms and genes evolve, this 
approach also has the advantage that congruence of diverse data sets encourages confidence in 
phylogenetic conclusions, while inconsistencies can identify false assumptions and limitations in models 
or data (Hillis 1987, Patterson 1987, Hillis & Moritz 1990, Sytsma 1990). As discussed above, no 
comprehensive evolutionary investigations of New Zealand taxa have been undertaken. New Zealand 
skinks are one of many groups which offer promise however (see also Fisher 1965, Raven 1988). 
Extensive morphological (McCann 1955, Hardy 1977) and allozyme studies of skink populations (Vos 
1988, Daugherty et ai. 1990b, C.H. Daugherty & G.B. Patterson pers. camm.) have provided a 
framework upon which to develop detailed evolutionary investigations. 
Information on the basic biology (Barwick 1959, Towns 1975a,b) and ecology (Barwick 1959, Whitaker 
1968, Towns 1975a, Gill 1976, Patterson 1985, Porter 1987), as well as some physiological data (Morris 
197 1 ,  Pollock & MacAvoy 1978, Werner & Whitaker 1978, Evetts & Grimmond 1982) is also available 
for various species of skinks. 
Views on the Origin of New Zealand Lizards 
New Zealand lizards (and the tuatara) trace their ancestry, in Maori mythology, back to Tangaroa, God 
of the Oceans (Andersen 1969). Seeking to escape the avenging wrath of Tawhiri-matea (Father of 
winds & storms) for the separation of Rangi (Earth) and Papa (Sky), Tu-te-wehiwehi and his 
descendants fled from the sea to the land adopting lizard-like forms. 
Chapter I, page 10. 
Three terrestrial groups of reptiles now survive here. The ancient reptilian order Sphenodontidae has two 
surviving Sphenodon species, both endemic to New Zealand (S. punctatus and S. guntheri, Daugherty et 
al. 1990a). Tuatara ecology and reproductive biology has been extensively studied (Crook 1975, Bell et 
al. 1985, Cree et al. 1992). Geckos and skinks are cosmopolitan families, each comprising about 600 
species, in approximately 82 and 40 genera respectively (Cogger & Heatwole 1981). On the basis of 
morphology, 18 species of gecko, in three endemic genera, are currently recognised in New Zealand 
(Towns 1985), but the validity of one genus is uncertain (Thomas 1982a). The New Zealand skinks pose 
a more difficult taxonomic problem because of their limited morphological diversification. Two species 
were originally described by J. E. Gray but more were recognised (see Dieffenbach 1843), and numbers 
have been continually revised upwards (McCann 1955, Hardy 1977). Two genera are currently 
described here, Leiolopisma Dumeril & Bibron and Cyclodina Girard, which currently contain 20 and 
six described species respectively (Towns 1985, Daugherty & Patterson 1990), but allozyme studies 
indicate that these are still underestimates (C.H. Daugherty pers. comm.). Leiolopisma skinks are 
generally diurnal insectivores found in grassland or coastal habitats. Cyclodina skinks tend to be slightly 
larger, nocturnal, forest dwellers, and are now found primarily on offshore islands in the Northern parts 
of New Zealand (Bull & Whitaker 1975, Pickard & Towns 1988). Rats, cats and collectors have 
considerably reduced lizard numbers in New Zealand (Whitaker 1978, Thomas 1982b). 
Our terrestrial reptilian fauna lacks a fossil record (Bull & Whitaker 1975) so there is uncertainty about 
the origins of New Zealand lizards. Tuatara seem to be long term residents, dating back about 80 million 
years when Gondwanaland was unfragmented (Crook 1975). Geckos and skinks have been suggested to 
be more recent colonists with affinities to New Caledonian and Australian groups (Bull & Whitaker 
1975, Hardy 1977, Towns et al. 1985). Morphological assessments suggest that New Zealand geckos 
have closest affmities to species from New Caledonia, and could have arrived here during the Miocene, 
up to 30 MY A (Kluge 1967). Members of the genus Cyclodina are known only from New Zealand but 
some morphological and karyotypic features suggest closer relationships to New Caledonian and 
Australian species than to New Zealand Leioiopisma (Hardy 1977, 1979), and Hardy (1977) proposed 
that Cyclodina may have arrived here independently, possibly during the early Pleistocene (up to 2.5 
MY A). Cyclodina has morphological similarities to the Lord Howe skink, L. lichenigerum (Bull & 
Whitaker 1975), and to L. alazon from Fiji (Zug 1985). Interpretation of allozyme data (V os 1988, C.H. 
Daugherty & G.B. Patterson pers. comm.) indicates that Cyclodina form a separate group, distinct from 
the New Zealand Lei% pisma. 
Ecological Aspects of Some LeioiopisT1Ul 
The widespread distribution and ecological flexibility of some New Zealand Leioiopisma is an 
interesting and important feature. Allozyme studies (Towns et al. 1985, Patterson & Daugherty 1990) 
corroborated ecological information (patterson 1985) which suggested that L. nigriplantare is a species 
complex which cannot be adequately resolved by anatomical characters. One subspecies, L. n. 
polychroma, is found in the lower half of the North Island and over many areas of the South Island, 
encompassing approximately six degrees of latitude and an altitudinal range of 1700 metres (pickard & 
Chapler 1. page 1 1 . 
Towns 1988). A morphologically distinct subspecies, L. n. nigripiantare, occurs on some of the 
Chatham Islands, 900 km east of the South Island, but it's allozyme profile is very similar to the 
mainland subspecies (Daugherty et al. 1990b). 
L. n. polychroma is apparently closely related to L. notosaurus from Stewart Island and to two other 
southern South Island species (L. maccanni, L. inconspicuum; Patterson & Daugherty 1990). It occurs 
sympatrically at several sites with these latter two but all are ecologically distinct (patterson 1985) and 
apparently reproductively isolated from each other (Daugherty et al. 1990b). The sympatric species can 
be distinguished on the basis of colour patterns ("striped", "spotted",  and "speckled"), though this is not 
always an infallible guide since the colour patterns are associated more with habitat type (for example 
"striped" is common in tussock habitats) than species (as identified on the basis of allozyme profiles; 
Patterson 1985, Daugherty et al. 1990b). 
Genetical Investigations of Leiolopisma 
The genus Leiolopisma is a particularly complex group however, and taxonomic relationships of 
members of the genus are uncertain. Preliminary immunological studies of albumin (Baverstock & 
Donnellan 1990, Hutchinson et al. 1990) support the view (Greer 1974 & 1982) that the genus is 
polyphyletic, some members being more closely related to other genera. Hutchinson et al. (1990) 
recommended restricting Leiolopisma to the type species L. telfairi from Mauritius, and they resurrected 
or created several generic epithets to accommodate the Australian species. Insufficient work has been 
done to clarify the affmities of New Zealand and New Caledonian species, but initial comparisons 
(Hutchinson et al. 1990) indicate that New Zealand and New Caledonian species are related to, but 
immunologically quite distinct from, the Australian taxa. A new genus for the New Zealand Leiolopisma 
is currently under review (C.H. Daugherty pers. comm.), but the term Leioiopisma is retained in this 
thesis for the New Zealand species as well as for other taxa which are or have been included in the 
genus. 
Members of the Leiolopisma group occur in New Zealand, Australia. Lord Howe Island, New 
Caledonia, Fiji, and also Mauritius in the Indian ocean. Morphologically, New Zealand Leiolopisma 
have affinities to some New Caledonian and Australian species, and the degree of similarity between 
them was used to suggest a 'recent' arrival here, within the last 5 million years (Towns 1975a, Hardy 
1977). This would have necessitated crossing hundreds of kilometers of open ocean. The Australian L. 
delicata, utilizing modem forms of transportation, became established here during the 1960's (Robb 
1980). 
The biochemical data (Baverstock & Donnellan 1990, Hutchinson et al. 1990, Daugherty et al. 1990b, 
C.H. Daugherty & G.B. Patterson pers. comm.) make it unlikely that the initial Leioiopisma 
colonisation(s) of New Zealand occurred as recently as the Pliocene (as suggested by Towns 1975a and 
Hardy 1 977), but the data is either insufficient or inadequate (see Avise & Aquadro 1982, Nei 1987 and 
Chapter Five) to determine precisely how long skinks have been here. A more detailed view of 
Otapter 1, page 1 2. 
Leiolopisma evolution based on DNA sequences is presented in this thesis. Similar analyses of 
Cyclodina are being conducted by others. 
A region of the mitochondrial l2S (small subunit) ribosomal RNA gene is examined using PCR and 
direct DNA sequencing (described in Chapter Two). No DNA sequence information was available for 
skinks before this s9 commenced so the 12S rRNA gene was selected for initial investigations since it 
is well conserved among vertebrates (Kocher et al. 1989, Palumbi et al. 1991). The 12S rRNA gene has 
been used for evolutionary investigations of other animal groups (Hixson & Brown 1986, Hedges et al. 
199 1 ,  Simon et aI. 1991, Allard et al. 1992, Ballard et al. 1992, Cooper et al. 1992, Meyer & Dolven 
1992). The skink sequences are analyzed with respect to a secondary structure model developed in 
Chapter Three. However, the relationships of the Leiolopisma skinks based on the sequence information 
are unclear, contrasting with the results from allozyme analyses (Chapters Five and Six). The patterns of 
DNA sequence differences are interpreted as suggesting a period of rapid Le iolopisma diversification 
and hypotheses to account for this are presented in Chapter Five. 
An important aspect of this study is that individuals of known allozyme profile were used, so DNA 
sequence and allozyme information from the same individuals was available. An advantage of this is 
illustrated in Chapter Six where comparison of the allozyme and sequence information leads to the 
suggestion that sexual hybridization has occlUTed at one site between two sympatric species, L. n. 
polychroma and L. maccanni. 
The evolutionary histories of the New Zealand skinks are not resolved and the DNA sequence 
information is suggestive of a complex set of relationships. Continued molecular investigations of this 
group are advocated, not only to aid in resolution of their phylogenetic relationships, but also to build 
upon and extend existing studies of Leiolopisma to explore broader ecological and evolutionary 
questions. Avenues for such research are discussed in the fmal chapter. 
page 13.  
Chapter Two: Methods 
This thesis uses the polymerase chain reaction (PC�) to obtain specific regions of the mitochondrial 
genome from frozen skink tail muscle. The sequences of these pieces of DNA are then determined and 
analyzed to make inferences about the evolutionary relationships between the skinks. While the 
p�ple ofPCR is elegantly simple (Saiki et al. 1988, Ambeim et al. 1990), there can be several 
difficulties in it's implementation. The sensitivity of PCR and consequent possibility of amplifying the 
wrong DNA make it necessary to check the accuracy and identity of the DNA sequences. The 
procedures used to extract amplify and sequence skink mitochondrial 12S rRNA sequences are 
described in this chapter, along with steps taken to insure the correct identity and accuracy of the 
sequences. A preliminary account of the methods was published in Hickson et al. (1992). 
The Polymerase Chain Reaction 
The polymerase chain reaction (PCR) is a simple and rapid means of amplifying specific regions of 
DNA from tissues, cells or other sources. Its versatility, the detaileq information which can be obtained, 
as well as the relatively low cost make it not only suitable, but essential, for the investigation of a variety 
of problems in population and evolutionary biology (Ambeim et al. 1990). The essence of the procedure 
is to have two short pieces (about 20 to 30 nucleotides long) of single-stranded DNA - the primers. 
These need sufficient sequence specificity to bind to the DNA bracketing the region of interest; one 
primer for each of the complementary DNA strands. The first step in the reaction is a high temperature 
(usually between 90 and 94°C) denaturation to separate the DNA strands, thus allowing the primers 
access to their complementary region(s). Rapid lowering of the temperature (to between 37 and 60°C) in 
the second stage permits the primers to bind with the complementary sequence. In the presence of the 
four nucleotide triphosphates the region of DNA between the pair of primers can be copied using a heat 
stable DNA polymerase, such as Taq polymerase (Saiki et al. 1988). These three steps are repeated 
sequentially 20-40 times. The amount of primers in the reaction is in sufficient excess so that the process 
is exponential and able to produce a million or more copies of the desired region in a few hours. 
Having appropriate DNA primers is critical to the success of the PCR This information is obtained 
either by conventionally cloning and sequencing the region bf interest, or by using information held in 
DNA sequence databases such as Genbank and EMBL. Relatively highly conserved regions of DNA -
such as the 1 2S ribosomal RNA encoding gene in animal mItochondria - can be isolated from a wide 
range of organisms using "universal primers" (e.g. Kocher et al. 1989), but faster evolving regions of 
DNA require more taxon-specific sequence data. This can pose difficulties for groups that have been 
less well characterized at the molecular level; for instance reptiles, invertebrates, and many plant groups. 
A difficulty of PCR is that its very power means that even slight contamination from other tissue or 
DNA can lead to the amplification of the wrong sample. Ex�eme care must therefore be taken to ensure 
that the correct sample of interest is amplified. A further difficulty was that no other reptilian sequences 
Chapter 2. page 14. 
were available to compare our gene sequences with. Considerable effort therefore was made in this 
study to ensure that we were obtaining skink gene sequences. 
Mitochondrial DNA Isolation 
To ensure that the primers amplified the correct mitochondrial region intact mitochondria were isolated 
and then the mtDNA purified. Laboratory rats were used as a test of the methods, then mitochondria 
were prepared from individual L. n. poiychroma. Mitochondria were prepared following the procedure 
of Fleischer et ai. (1979). Livers were removed from freshly killed animals, washed in ice cold 0.3 M 
sucrose, cut into small pieces, washed again, and ground in a glass tissue grinder in ice cold 0.3 M 
sucrose. Nuclear debris was removed by a low speed spin of 755 x g for 5 minutes at OOC. The 
supernatant was respun at 6800 x g for 15 min., OOC, and the brown mitochondrial pellet removed with a 
sterile glass rod, leaving behind the black nuclear debris. Mitochondria were resuspended in cold 0.3 M 
sucrose, and the purification repeated. A final spin at 7700 x g gave a pellet which subsequently proved 
to be largely free of nuclear DNA. This pellet was resuspended in cold TES buffer (10 mM Tris, pH 8, 1 
mM EDTA, 100 mM NaCI). Confirmation that mitochondria had been isolated was obtained by 
measuring respiratory activity using an oxygen electrode with sodium glutamate as the substrate (Lessler 
& Brierly 1969). The remaining suspensions were frozen (-20°C) as 250 J..Ll aliquots. 
Mitochondrial DNA was prepared from these aliquots by addition of sodium dodecyl sulphate (SDS) 
and proteinase K to final concentrations of 1 % and 50 Ilg/ml, respectively, and incubating at 37°C for 1 
hour. RNA was removed by a further 10 minute incubation with 4 Ilg/ml of RNase A. The solution was 
then sequentially extracted with, respectively, one volume each of phenol, phenol and chloroform, and 
fmally chloroform, to remove proteinaceous material. The supernatant was precipitated wiLh 95% 
ethanol and 0.3 M sodium acetate and resuspended in 25 J.1l sterile water. 
Total DNA Isolation 
If the PCR primers are specific enough for their target and do not cross react with other regions of the 
genome then highly purified DNA solutions however are not required for the PCR, it is simpler to obtain 
preparations of total DNA (containing both nuclear and mitochondrial DNA). For the main part of the 
study total DNA was extracted from frozen tail muscle tissue, obtained from the reptile collection of the 
National Museum in Wellington (Table 2. 1). These were some of the same specimens used by 
Daugherty et ai. (1990b; C.H. Daugherty & G.B. Patterson pers. comm.) for their allozyme studies. 
In addition, isolation of DNA from tissues preserved in other ways was also investigated. Extractions of 
DNA were performed on fresh tail muscle and tails preserved by desiccation with silica gel, storage in 
95% ethanol, or fixing in fonnalin. 
Oiapter 2, page 15. 
Table 2.1 . Skink taxa from which DNA sequence information was obtained. Species, locality, and 
National Museum of New Zealand reptile collection catalogue numbers are given. The three letter 
abbreviation for each taxon, which are used in later tables, are shown in parentheses. Figure 2.1 shows 
the New Zealand localities. 
Taxon 
"Stewart Island Green" l (SIG) 
Leiolopisma graru1e (Lgr) 
L. notosaurus (Lno) 
L. lineoocellatumlchloronotum2 (LIe) 
L. suteri (Lsu) 
L. nigripiantare nigriplantare(Lnn) 
L. n. polychroma (Lnp) 
L. microlepis (Lmi) 
Cyclodina aenea (Cae) 
L. acrinasum (Lac) 
L. inconspicuum (Lin) 
L. smithi (Lsm) 
L. zelandicum (Lze) 
L. fallai (Lfa) 
L. maccanni (Lma) 
L. injrapunctatum (Lfr) 
L. otagense (Lot) 
L. maco (Lmo) 
L. telfairi (Lte) 
Lampropholis guichenoti (Lag) 
Table Hill, St. Is. 
CentIal Otago 
Masons Bay, St Is. 
Tekapo, Otago 
Aorangi, Poor Knights Is. 
Chatham Islands 
Gorge Bum, Southland 
Twizel, Canterbury 
Taihape 
Somes Is., Wellington 
Fiordland 
Gorge Bum, Southland 
Ruamahua-iti, Alderman Is. 
Outer Chetwode Is., Marlb. 
Great Island, Three Kings 
Gorge Burn, Southland 
Stephens Is., Marlb. 
Central Otago 
Poor Knights Is. 
Round Is., MAURmUS 
AUSTRALIA 
a Locality abbreviations: St Is. = Stewart Island, Marlb. = Marlborough Sounds, 
Three Kings = Three Kings Island. 
1 Stewart Island Green has not yet been formally described. 
Cat. No. 
FT6 
CD1055 
CDI089 
CD1217, CD1218 
CD 1027 
CDI058, CD 1 060 
CD 1 1  10- 1 1 12 
CD2126 
CD2123 
CD1962 
CD826 
CD1 1OO, CDl lOl 
FT569 
CD1952 
FT598 
CD 1 1  06-CD 1 108 
CD535 
CDI053 
CD848, CD103 1 
CD2021 
CD536 
2 L. lineoocellatumlchloronotum; these individuals could not be identified defmitively on the basis 
of morphology and allozymes as either L. lineoocellatum or L. chloronotum (C.H. Daugherty pers. 
comm.). 
Fig. 2.1. Map of New Zealand showing locations of the skinks examined in this thesis. Latitude and 
longitude are indicated along the edges of the map. 
Fiordland 
1 72 '  1 73 1 74 '  1 75 '  1 76 '  1 77 '  1 78 
Wellington 
�� r � , I I 1 36  �' _ < 1 A;lderman Is. 
- 37 
1� �1 � 3-e 
Hf -' ;P-;T+ :: . 
Er l7 hr .  I :  ' :::1 " 
" ' I , J ' ' " i , . I I 
, : : " Taihape " ., / 
I I  I :: ' , ,�_t= :' I ;� � ' I U---�--L-�--�--��2 
" r----:-1-----;1-��-, �-: '!,-zh- -1'---'''';1-: - Marlborough 
1-- -1L' _- " " -(ff;*=-;lz'- 42 
r--r-...,-----,-J j' . ! it� +--/ I I ",( II '  I . ,  ' 3  I---I--- .r---t--.-- , -r , ' I ' i ,l"" J -
J-J7+J/-a,�,-�)i-_ '�:_�,�. , �. 1�"",-�, �\��c+a�nterbu rY f- ' , ' :? !! ChathaJll Is. , :  I I . , . , . 
-'" : : " . : . ) I I 
t'tt : . ' , : : ' : • � ?tago -1--
�5 
. I >; : ' I ' :: ;: ' :: I I I I :  1 " ..L. 2S thl d r- - ':o 
'{ �:�� -H0U an I _' 7 
Stewart Is. --� p. - -+---+--j 
I '  '1 I 1 I ! 
1 6 7 '  1 68 1 69 1 70 1 7 1  1 72 1 73 ' 7� 
Chapter 2. page 16. 
Small samples (about 0.05 g) of muscle were removed from near the base of the tail, cut into small 
pieces and incubated in 300 III of extraction buffer (10 mM Tris pH 7.5, 100 mM EDTA, 50 Ilg/ml 
proteinase K, and 1 % SDS) at 65°C for two to four hours (Sam�rook et al. 1989). The solution was then 
extracted with phenol and chloroform, precipitated, and resuspended in 25 III of sterile milliQ water. 
Amplification of DNA by the Polymerase Chain Reaction. 
The primers found to work best with skink DNA were: 
12SAR 5 '  -AAACTGGGAT TAGATACCCC ACTAT-3'  ( L1 0 9 1 )  and 
12SBR 5 ' -GAGGGTGACG GGCGGTGTGT-3 ' ( H147 8 ) , 
where L and H signify the light and heavy strands of the mitochondrial DNA (mtDNA) and the numbers 
refer to the 3' ends of the primers according to the complete human mtDNA sequence (Anderson et al. 
1981). These primers isolate an approximately 400 nucleotide long region in the second half of the gene 
and have been used to amplify 12S rRNA from a variety of vertebrate and invertebrate groups (palumbi 
et al. 1991)  
DNA amplifications were performed in  0.5 ml  reaction tubes using a DNA thermal cycler (perkin Elmer 
Cetus, Connecticut, USA). Initial trials were carried out in 20 jll volumes, while 80 jll was used for the 
production of DNA templates for sequencing. Reactions contained 200 )lM of each deoxynucleotide, 
reaction buffer (prom ega Corporation, Madison, USA; 50 mM KCl, 10 mM Tris-HCl, pH 8.8, 1 .5 mM 
MgC�, 0. 1% Triton X- 100), 0. 1 )lM of each primer, and 2 units of Taq polymerase (Promega). About 
100 ng of pure mitochondrial DNA or 100 to 200 ng of total DNA were used as the template for 
polymerase chain reactions. Samples were overlain with a drop of mineral oil to prevent evaporation. 
Several factors influence the success of the PCR (including the quality and quantity of the DNA 
template, primer sequence and concentration, magnesium concentration, the quality of the DNA 
polymerase, and the type of thermal cycler used). Various temperature and time settings need to be tried 
to obtain optimal specificity and amplification of the desired product (Innis & Gelfand 1990). We tried a 
variety of conditions and found that good amplifications for the 12S primers and skink DNA were 
obtained using the following parameters: 
94°C for 60 sec to separate DNA strands, 
54°C for 60 sec to allow primer annealing, and 
noc for 60 sec for DNA copying. 
Successful DNA sequencing requires good quality templates. Direct precipitation of the PCR reaction, 
either with or without phenol extraction, did not give clear sequencing gels. Initially, good sequences 
Olapter 2, page 17. 
were obtained by purifying PCR products through a 1 % Seaplaque agarose gel (FMC BioProducts, 
USA) prior to sequencing. Isolation of the product from the agarose used the hot phenol method of 
Thwing et al. (1975). After precipitation, DNA fragments were resuspended in 1 1  J.1l of sterile milliQ 
water. The availability of inexpensive centrifugal dialysis filters to remove excess primer DNA, 
unincorporated dNTPs, and other inhibitors from the PCR reaction proved more efficient and cost 
effective however. Most of the samples were processed using the Promega Magic PCR Preps TM DNA 
purification system (Promega Corporation, Madison, Wisconsin, USA). 
Mineral oil from the PCR reaction was removed by addition of approximately 30 J.1l chloroform and 
brief centrifugation. The PCR sample (approximately 70 J.1l) was mixed with 100 J.1l of direct 
purification buffer. This was mixed for one minute with 1 ml of DNA purification resin and then passed 
through the Magic PCR Preps TM mini-column using a syringe. The column was washed with 2 ml 80% 
isopropanol. Excess alcohol was removed by centrifugation for 20 seconds. Complete evaporation of the 
isopropanol was ensured by leaving the column at room temperature for a further two minutes. Thirty 
microlitres of sterile milliQ water was then added to the mini-column and the DNA allowed to 
resuspend for one minute. The sample was then collected by a 20 second centrifugation step. Adding the 
solution back to the mini-column and respinning increased the DNA recovery. Two microlitres of the 
purified PCR product was checked on a mini-gel prior to sequencing. Recovery of the product using the 
mini-columns was consistent and high. 
Direct Sequencing of peR Products 
Attempts to directly produce single-stranded DNA templates for sequencing by limiting one primer in 
the amplification reaction (Gyllensten & Erlich 1988) were neither consistent nor reliable. Smearing of 
DNA above the PCR product was usually seen when the single-stranded amplification reaction was 
checked on a gel. Increasing annealing temperature and decreasing the template DNA concentration in 
the PCR reaction had very little effect Consequently, a double-stranded sequencing procedure was tried 
based on the method of Casanova et al. (1990), using the modified T7 DNA polymerase (USB,  Ohio, 
USA) and 35S dA TP as the radiolabel source. Two microlitres of gel purified template (about 100 ng) 
was denatured at 96°C for 5 minutes in a 10 J.1l volume, containing 10 ng of one primer, and reaction 
buffer. This was then immediately snap frozen in liquid nitrogen to reduce reassociation of the DNA 
strands and then 5.5 Jll of the sequencing cocl1.ai1 added (as described in the Sequenase kit). 
Termination reactions witb the dideoxynucleotides were incubated at 37°C for two minutes. Sequencing 
reactions were run on 6% acrylamide gels, 8 M urea, then fixed, dried, and exposed to Kodak X-omat 
AR film (Eastman Kodak Co., Rochester, New York, USA). 
Four primers were used for sequencing. Two were the PCR amplification primers while two more were 
constructed after obtaining preliminary skink sequence data. This set of four primers allowed the 
complete sequence of botb strands to be determined, though use of primer 12SBR for sequencing 
consistently gave faint sequencing lanes. The skink-based primers, deSignated SK12SL and SK12SR, 
were situated approximately half-way along the PCR product and were complementary, designed to read 
OJapter 2, page 1 8. 
off opposite strands of the product (as shown in Appendix 1). Their sequences and corresponding 
location relative to the human mtDNA sequence are, respectively, 
SK12SL 5 '  -CTTCTI'TCAT AAGGTAGGC-3 ' (1.. 1408) and 
SK12SR 5 '  -GCCTACCTTA TGAAAGAAG-3 ' (H1390). 
Attempts to amplify and sequence another mitochondrial gene were less successful. Conserved primers 
to the cytochrome b gene (Kocher et al. 1989) did not consistently amplify the equivalent region in the 
skinks. Altering reaction temperatures and times (to both higher and lower levels), Mg2+ concentration, 
primer concentrations, Taq polymerase and buffer, did not result in improved yields. This suggests that 
either the cytochrome b primers of Kocher et al. (1989) may not match the skink cytochrome b gene 
target sites very well, or that there may be a secondary structure interaction inhibiting amplification. The 
same cytochrome b primers have been successfully used in other studies of vertebrates (e.g. Meyer et al. 
1990, Hedges et al. 1991, Irwin et al. 1991 ,  Moritz et al. 1993). Information was however obtained for 
three taxa and will be discussed in Chapter Five. The cytochrome b primers and their location relative to 
the human mtDNA sequence are: 
CBS 5 '  -GCTTCCATCC AACATCTCAG CATGATG-3 ' (1..148 13) and 
CB3 5 '  -GCAGCCCCTC AGAATGATAT TTGTCCTC-3 ' (HI5171) 
Both strands of the PCR product were sequenced and where possible sequences from several individuals 
from each population were obtained. Sequences were compared and aligned using the University of 
Wisconsin GCG sequence analysis software package, Version 6.2 (Devereux e t  al. 1984) and 
phylogenetic inferences made using the Hadamard (discrete Fourier) conjugation (Hendy & Penny 1989, 
1993, Penny et al. 1992). The Hadamard conjugation is described in more detail in Chapter Five, aB9-a 
regy�g versjgg f the 1:1 W� �f9gFaIB is i8elliaea i8 AjJ�eBa� 2 98 the Elisk@«e at the �aGk 9f tB8 
fftesis. 
RESULTS 
Oxygen electrode assays conflrmed the presence of mitochondria from the rat liver preparations, though 
respiratory activity could not be detected in any mitochondrial preparations from skinks. Digestion of 
the DNA with the restriction endonuclease BamHI gave three distinct fragments with a total size of 
approximately 16 000 base pairs, the expected size of the rat mitochondrial genome (Gadaleta et al. 
1989). The absence of background smearing in these preparations conflrmed that there was little 
contamination with nuclear DNA. Digestion of the skink DNA with both BamHI and EeoR! also 
confirmed that the preparations were free of nuclear DNA. Enzymatic digestion also indicated that the 
skink mitochondrial genome was at least 1 5  000 base pairs in size, but an accurate estimate of its size 
was not obtained. 
High molecular weight DNA was obtained from fresh, frozen, dried, and ethanol preserved tissues and 
all proved suitable substrates for PCR ampliflcation of the mitochondrial DNA fragment. DNA could 
not be obtained from tissues stored in formalin. Formalin rapidly cross-links DNA making it difflcult to 
Chapter 2, page 19. 
obtain DNA suitable for the PCR. even from samples stored for short periods in formalin (Greer et al. 
1991). Other sources of DNA. such as subfossil bones (eg. Cooper et al. 1992). preserved museum skins 
(e.g. Thomas et al. 1990). and herbarium samples (Rogers & Bendich 1985) can be suitable for such 
studies. 
The same single PCR product was obtained from preparations of both total DNA and purified mtDNA. 
Sequence analysis of the product showed it to be most similar to other vertebrate 12S rRNAs from the 
region of the gene to which the PCR primers anneal. The very sensitivity of the PCR however makes the 
procedure vulnerable to the amplification of contaminating DNA. such as that from the hands or skin of 
the investigator (Kitchin et al. 1990). When this study was begun no reptilian mtDNA sequences were 
available on the GenBank or EMBL sequence databases. so confrrrnation that skink sequences had been 
amplified was established indirectly in several ways. 
Amplification of a contaminating. non-skink. DNA template was discounted by the consistency of the 
skink DNA sequences. The same sequence was obtained from separate DNA preparations of the same 
individual. from separate amplifications of the same DNA sample. and from different individuals from 
the same population (see Table 2. 1). The latter observation is to be expected for a relatively highly 
conserved molecule such as 12S rRNA (MindeU & Honeycutt 1990). This consistency also indicates the 
fidelity of the Taq polymerase. Errors can be introduced however. In one instance (out of several 
hundred amplifications) a single nucleotide difference was found for separate amplifications of the same 
skink DNA. Such misincorporation errors by the Taq polymerase are known to occur at relatively low 
frequencies (Bloch 1991). As these results show. repeating amplification and sequencing reactions 
safeguards against these errors. Inadvertent amplification of rat 12S rRNA rather than skink sequence 
was also detected on one occasion. but in general contamination was not a problem. 
A final verification that the PCR products were not derived from other sources was performed by 
aligning a skink sequence against the same region of the 12S rRNA gene from other vertebrates and 
establishing their phylogenetic relationships. The alignment is presented in the next chapter in the 
context of secondary structure features (Table 3 .3). The phylogenetic relationships of the skink and ten 
other vertebrate sequences are shown in figure 2.2. The Hadamard conjugation using the two character 
state option (see Chapter Five) was used. Omitting one taxon to permit four character state analysis gave 
the same relationships for the remaining taxa. Parsimony analysis using PAUP 3.0s (Swofford 1990) 
also produced the same result (not shown). The skink sequence is most similar to the other reptile used. 
a xantusiid lizard from the Caribbean (Hedges et al. 1991). confirming that the correct target sequence 
had been obtained by PCR. 
In this analysis (Fig. 2.2). the birds are slighlty closer to the reptiles than to the mammals. The 
phylogenetic relationship of reptiles. birds and mammals to each other are still uncertain. Benton (1990) 
places birds closer to the reptiles than mammals on the basis of morphological characters. Sequence 
analyses of 18S rRNA. beta-haemoglobin and myoglobin group birds with mammals. but relationships 
Fig. 2.2. Inferred phylogenetic relationships of 1 1  vertebrate 12S rRNA sequences. Alignment was done with 
reference to sequence's secondary structure and analysis performed using the Hadamard conjugation. The skink 
sequence is mos,t similar to that of the xantusiid lizard. Xanlusia riversiana, confirming that reptilian DNA had 
been amplified by the polymerase chain reaction. Branch lengths are proportional to the probability of change along 
that branch. The tree is unreoted. 
�oad 
, 7 . .  , .�. ' . ,, ' . -. ... :. : Fish 
"?B't : , -
. , . . . .L . " M ouse r , .� 
OJapter 2, page 20. 
inferred from insulin, alpha-aystallin and alpha-haemoglobin sequences place birds with reptiles 
(Hedges et al. 1 990). Inferences about the order of deep branches in vertebrate evolutionary 
relationships are however, unlikely to be robust using the shOrt sequence of 1 2S rRNA analysed here. 
Two different 1 2S rRNA sequences were found for L n. polychroma in this study. The sequence of L. n. 
polychroma from Gorge Burn was identical to that of L maccanni at the same site, while the 12S rRNA 
sequence of the L n. polychroma individual from Twizel was quite different (see Table 3.1). Partial 
sequence information from other L. n. polychroma populations in the North and South Islands (data not 
presented) were the same as the sequence from Twizel, while partial sequence data from other 
populations of L. maccanni were the same as the sequence from Gorge Burn. DNA was extracted 
independently from several tissues (muscle and liver) and from several individuals for these taxa, and 
amplified and sequenced several times. The same sequences were obtained from each individual every 
time indicating that the identical sequences were not the result of contamination of tissue samples or 
PCR amplification reactions. This point will be discussed again in Chapter Six. 
The complete 12S rRNA sequence information for 20 skink taxa are given in Appendix 1, in a format 
suitable for analysis using the Hadamard conjugation. The sequences are also presented in the next 
chapter (Table 3.1).  The nucleotide frequencies for the 20 skink taxa which are considered in most detail 
in this thesis are very similar, with a slight predominance of adenine (Table 2.2). The results of the 
analyses presented in the subsequent chapters are therefore not affected by differing nucleotide 
compositions between taxa (see Lockhart et al. 1 992, Hasegawa & Hashimoto 1993). 
Oiapter 2, page 21. 
Table 2.2. Base composition of the skink 1 2S rRNA sequences. 
Frequency of 
Taxon A C G tl A+U C+G 
"St Is. Green" 0.36 0.27 0. 19 0.19 0.54 0.46 
L grande 0.35 0.26 0.19 0.20 0.55 0.45 
L notosaurus 0.35 0.27 0.19 0.19 0.54 0.46 
L linlchl 0.35 0.26 0.19 0.20 0.55 0.45 
L. suteri 0.35 0.26 0.19 0.20 0.55 0.45 
L n. nigriplantare 0.35 0.26 0.20 0.20 0.54 0.46 
L microlepis 0.34 0.27 0.20 0.19 0.53 0.47 
La. guichenoti 0.35 0.26 0.19 0.20 0.55 0.46 
C. aenea 0.34 0.28 0.20 0.1 8  0.52 0.48 
L. acrinasum 0.35 0.27 0.19 0.19 0.54 0.46 
L inconspicuum 0.34 0.28 0.20 0. 19 0.53 0.47 
L. smithi 0.34 0.28 0.20 0.19 0.52 0.48 
L. zelandicum 0.34 0.28 0. 19 0.19 0.53 0.47 
L· fallai 0.35 0.27 0.19 0.19 0.54 0.46 
L. maccanni 0.34 0.27 0.19 0.20 0.54 0.46 
L telfairi 0.34 0.28 0.19 0.19 0.53 0.47 
L. n. po/ychroma 0.34 0.26 0.20 0.20 0.54 0.46 
L. infrapunctatum 0.35 0.26 0.19 0.20 0.55 0.45 
L otagense 0.34 0.27 0.20 0.19 0.53 0.47 
L. maco 0.34 0.28 0.20 0.19 0.52 0.48 
Mean 0.35 0.27 0.19 0.19 0.54 0.46 
Standard Deviation 0.Q1 0.01 0.01 0.01 0.01 0.01 
elulpter Three: A Refined Secondary Structure Model 
for Domain 111 of Vertebrate 12S rRNA 
page 22. 
A DNA sequence is not a simple string of abstract letters. Appreciation of the higher order structure of 
DNA has several advantages for evolutionary analyses. Different positions and regions in the molecule 
have different functional and structural constraints, and so different rates and patterns of change (see for 
example Kimura 1983, Noller et al. 1990) which should be taken account of when-attempting-to 
reconstruct phylogenies. Knowledge of how and where the molecule varies, and where it is constrained, 
is also essential for understanding the patterns and processes of molecular evolution. The purpose of this 
chapter is to ftrstly refine a vertebrate secondary structure model for domain ill of mitochondrial 12S 
rRNA and then to determine patterns of sequence variability among the vertebrate groups. This 
information will then be used in analyses in subsequent chapters. 
Refining a Secondary Structure Model 
Comparative sequence analyses, free energy predictions, and experimental investigations (such as 
chemical modification, nuclease digestion and intra-RNA crosslinking) have been used to determine 
secondary structure models for ribosomal RNA (reviewed in Woese et al. 1983, Noller 1984). Three 
different secondary structure models exist for the prokaryotic small subunit (16S) rRNA, to which the 
mitochondrial 12S rRNA is related (Stiegler et al. 1981,  Zwieb et al. 1981,  Gutell et al. 1985). The three 
models are generally similar, though they differ in the size and or placement of some base-paired 
regions (Noller 1984, GuteU et al. 1985). 
Structurally, the small subunit rRNA is divided into four domains (labelled I, II, ill and IV), each 
separated by a highly conserved single-stranded region (Fig. 3.1a). Domains III and IV are the most 
highly conserved regions of the molecule (Noller 1984, Neefs et al. 1990). The mitochondrial form of 
the small subunit rRNA is a reduced structure, lacking some helices (base-paired regions) but still 
having all four domains (Fig. 3. 1b). Structural and sequence conservation among rRNAs is high (even 
from different kingdoms), with major changes in size involving insertion or deletion of blocks of 
sequence (Noller 1984). Comparative sequence analysis is therefore an effective way of estimating 
secondary structures of rRNA, the validity of which can be subsequently examined experimentally 
(Noller 1984). Several compilations of rRNA sequences in a structural context are available for a wide 
range of organisims (Dams et al. 1988, Gutell & Fox 1988, Gutell et al. 1990, 1992, Neefs et al. 1990, 
1991 ,  Specht et al. 1991, de Rijk et al. 1992). The small subunit rRNA compilation of Neefs et al. 
(1990) formed the basis for refining a secondary structure model for part of animal mitochondrial 12S 
rRNA. Neefs et al. ( 1991) now also provide the compilation in computer formal Models for the 12S 
rRNA structure are limited by the fact that only a few mitochondrial sequences have been used in 
previous compilations (Zwieb et al. 1981 ,  Clary & Wolstenholme 1985, Gutell et al. 1985, Neefs et al. 
1990). The PCR primers 12SAR and 12SBR (Chapter Two) amplify the last part of domain IT and all of 
domain III (Fig. 3 . 1b), though for simplicity in this thesis, the whole PCR fragment is usually referred to 
as domain III. In this chapter I present alignments for domain III of the 12S rRNA from all five classes 
Fig. 3.1. a. Secondary structure model for the prokaryote small subunit (16S) rRNA (from Neefs et al. 
1990). The mitochondrial small subunit ( 1 2S )  rRNA gene is derived from this. Helices are numbered, 
and regions which may be absent in some prokaryotes are drawn in thin lines and labelled VI-V9. b. 
Secondary structure model for human mitochondrial small subunit rRNA (from G utell et al. 1985). 
Helices are numbered as in 3.1a. The location of the PCR primers 12SAR and 12SBR are indicated. 
a. 
'8 '-__ ___ 
V2 
V7  
41 
47 \_ 
VB 
V9 
Homo sapiens mi1ochondtion 
OIapter 3, page 23. 
of vertebrates. Complete 12S rRNA sequences are available for only a few vertebrates, but the use of the 
polymerase chain reaction and universal l2S rRNA primers (Kocber et al. 1989, Palumbi et al. 1991) 
are leading to an increasing availability of domain III sequences from a wide range of taxa 
The Skink and Otber Vertebrate 12S rRNA Sequences 
The sequences from 20 skink taxa are presented in Table 3.1, overlain with Neefs et al. suggested 
belical (stem) regions. Other sources of vertebrate sequences used in this cbapter are listed in-table 3.2. 
The 12S rRNA sequences for several snakes (Knight & Mindell I993), and a complete seal 
mitocbondrial genome (Amason & Jobnsson 1992) were also recently published , but were obtained too 
late to be included in the analyses presented in this thesis. A more general structural model of domain ill 
for both vertebrate and invertebrate 12S rRNA is being developed (AJ. Cooper, R.E. Hickson, G.M. 
Lento, C. Simon & D. Penny in prep.). 
Approacbes for Refining a Vertebrate Domain ill Secondary Structure Model 
Three strategies were used to determine bow the region of the 12S rRNA between PCR primers 12SAR 
and 12SBR could fold. Comparative sequence analysis establisbed wbicb belices bad the potential to 
form (Table 3.3). Identification of fixed compensatory mutations (cbanges preserving base-pairing in 
helical regions) was especially useful for confuming belices and locating those wbose sequences varied 
between the different groups. The vertebrate sequences in table 3.3 were relatively easy to align wben 
done so in the context of secondary structure. Gaps to maintain alignments were introduced at the 
beginning or end of unpaired regions. 
Energetics a/the Folding a/the rRNA Sequences 
The third approacb was to compare minimal free energies of some of these suggested belix structures 
using the MFOLD algorithm (Zuker 1989, Zuker et al. 1991) in the University of Wisconsin Genetics 
Computer Group (GCG) Package, version 7.2. The program takes account of the energetic costs and 
benefits of potential base pairings, and the occurrence, locations and interactions among unpaired 
nucleotides (Zuker 1989, Zuker et al. 1991). The MFOLD program can sbow both optimal and 
suboptimal folded structures. Other folding programs are also available (e.g., Abrahams et aI. 1990) but 
were not used in this study. 
On their own, and in the absence of knowledge about tertiary or quaternary structure, sucb energy 
calculations may not indicate wbat can actually form, but when used in conjunction with the other two 
lines of inquiry provide a well-founded basis for suggesting secondary structure models. Structures 
determined on the basis of comparative sequence analysis are usually found to have a minimal free 
energy within 10% of the value of the optimal structure determined by MFOLD (Zuker et al. 1991). 
Table 3.1.  Alignment of 20 skink taxa for 384 bases of 12S rRNA. The PCR primers 12SAR and 12SBR are not included but start 
immediately before and after, respectively, the sequences shown. Conserved sites are in uppercase and variable positions are in lowercase, 
with trans versions in bold. The suggested helical (stem) regions of Neefs et al. (1990) are shown above the sequence alignment (a ' 
denotes the distal arm <;>f a helix). Taxa are arranged in order of increasing numbers of differences (see Table 4.1a), and the three letter 
abbreviation for each taxon is shown on the right. A dot indicates no difference from the consensus, while deletions are shown as dashes. 
The internal sequencing primers SK12SL and SK12SR anneal between bases 185 and 203. L. telfairi has two single base deletions 
(positions 25 & 60) and one single base insertion (between 202 & 203). L. maccanni also has a deletion at position 25. When these two 
taxa are included in phylogenetic analyses (Chapter Five), the sites of deletion and insertion are not included. 
St IsGreen 
L . grande 
L . notosaurus 
L . l inlchl 
L . suteri 
L . n . nigripl . 
L . microlepis 
La . guichenoti  
C. aenea 
L . acrinasum 
L .  inconspicuum 
L . smithi 
L . zelandicum 
L .  fa llai  
L .maccanni 
L .  telfairi 
L .  n . polychroma 
L . infrapunctatum 
L . otagense 
L .moco 
2 6  -li: _.:.2..;.0_' __ 3 0  
1 2 3 4 5 6 7 8 9 10  
123456789 0123456789 0 12345678901234567890 12345678901234567890 123 456789 0 123456789 0 1234567 89 0 123 456789 0 
GCucAGCCGUcAACAAAgAcAGuauaaaauACAauaCUgUUCGCCAGAGAAcUAcAAGcuAAaaCUcaAAACuCcAAGGACUUGGCGGUGCUCCAcAuCa 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . .  
· . . . . . .  , . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . .  
· . . . . . . . . . . . . . . . . . . . . .  cc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . .  c . .  
· . . u . . . . . .  a . . . . . . . . . . . . . . . . . .  c . . . . . . . .  a . . . . . . . . . . . . . . . . . . . .  c . .  gg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
· . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . .  
. . . . . . . . . . . . . . . . . . .  u . . . . . . . . .  c . . .  g . . . .  a . . . . . . . . . . . . . . . . . . . . . . .  g . . . .  g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
. . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . .  g . • . • . . . . . . • . . . . . . • . . . . .  a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
· . c  . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . .  c . . . . . . . . . . . . . . . . . . . . . . .  u . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
· . c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g . .  a . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
· . c . . . . . . . a . . . . . . . . . . . . .  a . . . . c . . . . c . . . . . . . . . . . . . . . . . . . . . . . . . . . gg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
. . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . g  
. . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 
· . .  u . . . . . .  u . . . . . .  a . . . . . . .  g . .  cc . . . .  c . . . . . . . . . . . . . . . . . . . . . . . .  a . . .  9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
I . . . . . . . . . . . .  ' . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . .  u . . . . . . . . . . .  g • .  a . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
. . . . . . . . . .  a . . . . . . a . . . . . .  - . . . . . . . .  g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . g  
. . . . . . . . . .  a . . . . . . a . . . . . .  - . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - . . . . . . .  g . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . .  . 
. . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . .  g . g  . .  a . . . . . . . . . . . . . . . . . . . .  a . .  g . . . . . . . . . . .  u . . . . . . . .  ,' . . . . . . . . . . . . . . .  . 
· . c  . . . . . . .  a . . . . . . . . . . . . . . . . . .  c . . . .  c . .  ; . . . . . . . . . . . . . . . . . . . .  u . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . 
· . c . . . . . . .  a . . . . . . . . . . .  c . c . . . . . . . . .  cg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • .  9 
· . . . . . . . . .  u . . . . . . . . . . .  c . . . . . .  c . . . .  cg . . . . . . . . . . . . . . . . . . . . . . . . . . .  9 . .  g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
Sig 
Lgr 
Lno 
L1c 
Lsu 
Lnn 
Lmi 
Lgu 
Cae 
Lac 
Lin 
Lsm 
Lze 
Lfa 
Lma 
Lte 
Lnp Q 
Lfr i 
Lot .l" 
Lmo :i (t 
N � 
StIsGreen 
L . 'grande 
L . notosaurus 
L . l in/chl 
L . suteri 
L . n . nigripl . 
L . microlepis 
L . guichenoti 
C . aenea 
L .  acrinasum 
L .  inconspicuum 
L . smi thi 
3 1  3 2  3 3  3 3 ' 3 4  3 6  3 7  3 8  3 8 ' 
1 1  12  13 1 4  1 5  1 6  17 18  1 9  2 0  
1234 56789 0123 456789 0 1234 56789012345678901234567890123456789 0 1234567890123456789 0 123456789 0 1234567890  
aCCUAGAGGAGCCUGUCCUAUAAUCGAUACCCCcCGAUCuACCuGAGGacUUUUUGAAacUCAGcCUAUAUACCGCCGUCGuCAGCcUACCUUaUGAaAG 
Sig 
Lgr 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . .  Lno 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  L1c 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lsu 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lnn 
c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lrni 
c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . • . • . . . . . • . • • . . . . • . • •  Lgu 
c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ' . . . . . .  c . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . .  " . . . . . . . . . . . . . . . .  g .  . . . . .  Cae 
u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . • . . . . . . . • • . .  Lac 
. . . . . . . • . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . • . .  u • . • . • • . . . • . . • . . . . • • • •  c • • . . • • • • • . • . . • . • • •  Lin 
c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lsm 
L .  zelandicum c .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lze 
L . fallai  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lfa 
L . maccanni . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . .  u . . . . . . . . .  u . . . . .. . . _ . . . . . . . . . . . . . . c . . . . . . . . . . .  g . . . . . .  Lrna 
L. telfairi g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g . . . . . .  Lte 
L . n .polychroma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . a . . . . . . . . . . . . . .  g . . . . . . . . .  gu . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . .  Lnp 
L .  infrapuncta tum c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . .  Lfr 
L . otagense u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . • . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g . .  Lot 
L . moco c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . .  gu . . . .  u .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lrno 
StIsGreen 
L . grande 
L . notosaurus 
L . l in/chl 
L . suteri 
L . n . nigripl . 
L . microlepis 
L . guichenoti 
C. aenea 
L . acrinasum 
L .  inconspicuum 
L . smi thi 
L . zelandicum 
L .  fallai  
L . maccanni 
L .  telfairi 
L .  n . polychroma 
L . infrapunctatum 
L . otagense 
L . moco 
4 0 ' 3 6 ' 34 ' 3 2 ' 4 3 4 3 ' 
2 1  22 23  24  2 5  2 6  27 28  29  
12 3 4 56789 0 123 45 6789 0 123456789 0 123 45678901234567890 1234567890 123456789 0 123456789 0 1234567890123456789  
aa-guauAGuAaGcaAAAuAguCaccaAcUAaAACGuCAGGUCAAGGUGUAGCAcAUaaaguGGaAGAGAUGGGCUACACUCUCUcCCcCAGAGaAcACg 
. a . c  • • • • • • • . • . . • • • • • • • . • . • . . . . . • •  c • . . • . • • • • • • • . • . • • . • • • . • • . . . . . • . . . • • • • • • • • . • • • • • •  a • •  u . . . . . . . . . . .  Sig 
• . •  c • • • • . . •  g • • • • . . • . • . . . . • . • • . • • •  c • • • . • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • . . • • • • • • • • • • •  u . . . . . . . • . . .  Lgr 
a . .  a • • • •  g • • . • • . . • • • • • • • • • . • • • . • • • . • . • • • • • • . • • • . . • • . • • . . • • • • . . . • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •  Lno 
. c . a  • • . • • • • • • • • • • . • • •  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . . .  a • • • • • • • • • • • • • • • • • • • • • • • • • • •  u . . . . . . . . . . .  Llc 
. . . . . . . . . .  u . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g • • . • •  Lsu 
ac . . . . • .  g • • • . • • . . • • • . • . • . • • • • . . • • • • • • • • • • • • • • • • • • • • • • • • • •  c • • • • . • . • • • • . . . • . . • • • • • • • • • •  u . . . . • . . • • . .  Lnn 
• • .  c • • • • . • • • • • • • • • • • . • • • • • • • • • • • • . . . • . • • • • • • • • . . • . . • • .  gu • • • • . • . • • • • • • • • • • • • • • • • • • • • • •  u .  • • • • • • • • • .  Lmi 
· a . • • . • • • • • • • • • • • • • • •  u . • • • • •  g • • • • • • • • • • • • • • • • • • • • • • • . • • • •  a • • . • • • • • • • • • • • • . • • • • • • • • a • •  u .  • • • • • • • • • •  Lgu 
. cg • • • • • • • • . • • • .  , • • • • • • . . . . • • • • . • . • • . • . • • • • . • • • • • • • • • • • • • • •  c • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •  Cae 
• . . • . . . . . • • • • • . • .  a • . . • • • .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  a • • . • • • • • • • • • • • • • . • • . • • • •  a • • • • • • . • • . • • • •  Lac 
a . .  a .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g • • • • c • • • • • • . . • . . • • • . • . . . . . • • • • • • • •  g .  . . . •  Lin 
· c . c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  gu • • . • • • . . . • . • . . . • . . . • • • . • . • • • •  u .  . . . . . . . . . .  Lsrn 
. g  a . . c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . .  Lze 
ac . c  • • . • . . . . • . • • • • • • . . . • . . . . • . • . • • • • . . . . • . . • • • • • • . .  u . . . . .  a • . . . . . • . • • • • • • • • • • • • • • • . . • • • • • • • • • • • • •  u Lfa 
g . .  c . �  . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  u . . . .  -. a  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 4 • • • • • • • • Lma 
. .  ac . .  a . . • . . • • • . • • • • •  c . c .  uu • • • • • . . . . • . . • • • • . • • • • • • • • • • • • • • • • •  c • • • • • • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • •  Lte 
· c . a . . . .  9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . .  u .  . . . . . . . . . .  Lnp 
aa • . • • • • • • • • • • • . . • . • .  u . . . . . .  g . • • • . . • • • • • • • • • • • • . • • • . • • • • •  a • • . • • • • • • • • . • . • . • . • • • • • • • • •  u . . . . . . . . . . .  Lfr 
. . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g • • • • • • • • • • • • • •  Lot 
. g  cc . . . •  c . . . • . • • •  c • . • • • • . . . . . • • • • • • • • • • . . . • • • • • • . . . . . . • . • • .  a • • . • . • • • . . • . • • . . • • • • • • • • • • . • . • • • • • • • • • •  Lmo 
St;IsGreen 
L . grande 
L . notosaurus 
L . l in/ehl 
L . suteri 
L . n . nigripl . 
L . mierolepis 
L. guiehenoti 
C. aenea 
L . aerinasum 
L .  ineonspieuum 
L . smi thi 
L . zelandieum 
L .  fal lai  
L . maeeanni 
L .  telfairi 
4 5 3 1 ' 4 6  4 6  ' 3 0 ' 
3 
o 3 1  3 2  3 3  3 4  3 5  3 6  3 7  3 8  
0 12 345678901234 56789 0 12345678901234567890123 456789 0 1234567890123456789 0 12 345678901234  
AAcaGCAucaAUGAAAcaCuGCucaAAGGuGgAUUUAGuAGUAAGAuaaaCaAGAgACuuaucUuAAAcCAGCcCUGGAGCGCGC 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . .  u . . . . . . . . . . .  Sig 
. . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  a . . . . . . . . . .  u . . . . . . . . . . .  Lgr 
. . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . 
Lno 
Llc 
. . . . . . . . . . . . . . . .  . c . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . .  a . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lsu 
. . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lnn 
• • • • • • •  c • • . . • • • • • • • • • • . • • • • . • • • • . • . • • •  c • . . . • . • • • •  g . . . . . . . .  c . • • • • • • • •  u .  • • . . . • • • • • • • • • •  Lrni 
. . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g . . . . . . . . . . . . . . . . . . . . . . .  u . . . . . . . . . . .  Lgu 
. . . . . . .  c . . . . . . . . . . .  '. . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . .  Cae 
· . . . . . . . . . . . . . . . .  c . c . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  g . . . . . . . . . . . . . . . . . . . . . . .  u .  . . . . . . . . . .  Lac 
. . . . . . . .. . . . . . . . . . . . . . . . u . . . . . c . . . . . . . . . . . . . . . . . .  g • . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . .  Lin 
· . . . . . . c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . .  9 . g . . . . . . . .  ccg . . . . . . .  u .  . . . . . . . . . . . . . . .  Lsm 
· . . . . . . c . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . .  g . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . .  Lze 
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . .  g . . .  a . . . . . . . . . . . . . . . . . . . .  Lfa 
· . . . . . . • • • . . • • . . . . . . . . .  ug . . . . . . . . . . . . . • . . . . . . . . •  g • • • . . . . . • • • . . • • .• • . . . . . . • . • • • • • • • • • • .  Lrna 
. . . . . . . .  c . . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lte 
L . n .polyehroma . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . .  c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lnp 
L . infrapunetatum . .  u . . . . .  ag . . . . . .  uu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . •  g . . . . . . . . . .  a . . . . . . . • . . . . . . . . . . . . . . . . .  Lfr 
L . otagense 
L . moco 
· • .  g . . . . . . . • • • . • . . . . . .  eu . . . . .  c . a .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Lot 
· . . . . . . . . . . . . . . . . . . . . . . .  9 . . • •  c • • • . • • • . • . . • . . • • • • • • •  g .  . . . . . . . • . • • • • • • . • • . . . . . . • • • • • • • .  Lrno 
O!apter 3, page 25. 
Table 3.2. Species and sources of 12S rRNA sequences used for Table 3.3. Whether the sequence was 
obtained by PCR and direct sequencing or by cloning is also indicated. An asterisk indicates a group for 
which several sequences are available (see Appendix 2). 
Genbank 
Taxon Ace. No. Ref. Note 
Coelocanth S I 1 1210 1 .  PCR product 
Lungftsh S 1 1 1206-S 1 1 1209 1 .  PCR 
Crossostoma lacustre . f f�hL M91245 2. cloned 
Rana castesbeiana (frog) X12841 3. cloned 
Xenopus laevis (toad) MI0217 4. cloned 
Xantusiid lizards 
* 
M651 1O-M651 16 5. PCR 
Chicken X52392 6. cloned 
Ratites 
* 
X67626-X67638 7. PCR 
Whale S79330 8. cloned 
Rat X14848 9. cloned 
Mouse J01420 10. cloned 
Cow 
* 
J01394 1 1 . cloned 
Human 
* 
J01415  12. cloned 
References: 
1. Meyer & Dolven (1992). 
2. Tzeng et al. (1992). 
3. Nagae et al. (1988). 
4. Roe et al. (1985). 
5. Hedges et al. (1991). See Appendix 2. 
6. Desjardins & Morais (1990). 
7. Cooper et al. (1992). See Appendix 2. 
8. Gadaleta et al. (1989). 
9. Bibb et al. (1981). 
10. Amason et al. (199 1). 
1 1. Anderson et al. (1982). See Appendix 2. 
12. Anderson et al. (1981). See Appendix 2. 
Table 3.3. Alignment of the region of vertebrate 12S rRNA between the PCR primers 12SAR and 12SBR. The primer sequences are not 
included, but 12SAR is immediately to the left, and 12SBR immediatedly to the right of the sequences shown. Where more than one taxon 
is available for a group (see Table 3.2), the positions that are variable among them are shown in lower case. Bold regions denote helical 
(stem) regions following Neefs et al. (1990), with the exceptions that stem number 39 is omitted. The regions into which the sequences 
have been divided for further analyses are indicated below the human sequence (S denotes a stem &/or hairpin region, and L a loop). 
Nucleotides conserved in all the vertebrates listed here are also shown below the alignment. Coelocanth and lungfish are included in the 
alignment but not used in the subsequent analyses because they are incomplete. Helix regions marked with a prime ( , ) denote the distal 
(3 ' )  arm of the helix. Gaps, introduced to maintain alignments, are indicated by " -" .  
Coelocanth 
Lungf i sh 
Fish 
Frog 
Toad . • 1 
Xantus 1.f 
Skink 
Chicken 
Moa 
Whale  
Rat 
Mouse  
MooCow 
Human 
Conserved 
2 5 ' 2 6  2 6 ' 2 0 ' 2 9 2 9 ' 2 ' 3 0  
GGAACAACAAGCCACA- -GCUUAAAACUCAAAGGACUUGGCGGUGCUUCAUA 
CCAGGAaCUACaAGCcCAa- -GCUuAAAACcCAAAGGACUUGGCGGUGCCUCAcA 
GCUCAGCUAUAAACCUAGACGUUU- -AAUCACAACAAACGUCCGCCAGGGUACUACGAGCGUCA- -GCUUAAAACCCAAAGGACUUGGCGGUGCCUUAGA 
GCCUAGCCGUAAACAA- - - - -UUA- -AUUUACACCAAUAAG-CGCCAGGGAAUUACGAGCAAU- - -GCUUAAAACCCAAAGGAUUUGACGGUGUCCCA- ­
GCCUAGCCAUAAAC- - - - - -UUUGACUACUUACGCAAA AUCCGCCAG- -AACUACGAGCCUAA--GCUUAAAACCCAAAGGACUUGGCGGUGCUCCAAA 
GCuuaACuguaAACcuAgacagC- -caaaaacauuugCuGucCGCCAGgGaAcUACAAGcAAAA- -gCUcaAAACuCaAAGGACUUGGCGGUGcuCuAuA 
GCucAGCCGUcAACAAAgAcAGu-auaaaauACAauaCUgUUCGCCAGAGAAcUAcAAGcuAAa- -aCUcaAAACuCcAAGGACUUGGCGGUGCUCCAcA 
GCCUAGCCCUAAAUCUAGAUACCg-CCCAUCACAC�UGUAUCCGCCUGAGAACUACGAGCACAAACGCUUAAAACUCUAAGGACUUGGCGGUGCCCCAAA 
GCUUaGCCcUAAAUCcaGaUaCU�--aCcccACac�AGUAuCCGCCcGAGAACUACGAGCACAAACGCUUAAAACUCUAAGGACUUGGCGGUGCCCcAAA 
GCUUAGUCGUAAACCCCAAUAGUC- -ACAAAACAAGACUAUUCGCCAGAGUACUACUAGCAACA--GCCUAAAACUCAAAGGACUUGGCGGUGCCUCAUA 
GCUUAGCCCUAAACCUUAAUAAUU-AAACCUACAAAAUUAUUUGCCAGAGAACUACUAGCUACA- -GCUUAAAACACAAAGGACUUGGCGGUACUUUAUA 
GCUUAGCCAUAAACCUAAAUAAUUAAAUUUAACAAAACUAUUUGCCAGAGAACUACUAGCCAUA- -GCUUAAAACUCAAAGGACUUGGCGGUACUUUAUA 
GCuuAGCCcUAAACAcAgaUaauu-acauaAACAAaAuUaUUCGCCAGAGuACUACuaGCaAca--GCuuaAAACUCAAaGGACUUGGCGGUGCUUuAuA 
GCuUAGCCCUAAACuUc�CAGUU-aAAUuAACAAaACUGCUCGCCAGAACACUACGAGCCACA- -GCUUAAAACUCAAAGGACCUGGCGGUGCUUCAUA 
GC AA 
825 ' i 826 
GCC A G 
i 820 '  i 829 
C AAAC C A GGA UG CGGU 
i 82 ' i 830 
A 
i 
Coelocanth 
Lungfish 
Fish 
Frog 
Toad . 
Xantusi'd 
Skink 
Chicken 
Moa 
Whale  
Rat 
Mouse 
MooCow 
Human 
Coe.locanth 
Lungfish 
Fish 
Frog 
Toad 
c 
Xantusid 
Skink 
Chicken 
Moa 
Whale 
Rat 
Mouse 
MooCow 
Human 
3 1  3 2  3 3  3 3 ' 34 " 3 5 '  3 6  3 7  
CCC-CCUAGAGGAGCCUGUUCUAGA!CCGAUAAACCCCGA�CAACCUCAACCACACUU-GC-UAu(rUCAGCCUAUAUACCGCCGUCGCCA�C 
CCCACCUAGAGGAGCCUGUuCUAgA!CCGAUAAUCCACG�uuACCcaA-CCuucccUgGC- -Aut'uCaGc�UACCgCCGUCGCCAGCCaAC 
CCCCCCUAGAGGAGCCUGUUCUAGA!CCGAUAACCCCCG�AAACCUCA-CCACUUCUaGU-CAUCCCCGCCUAUAUACCGCCGUCGUCAGCUUAC 
CCCCACUAGAGGAGCCUGUUCUAUAAUCGAUGAUCCCCGAUAUACCCGA-CCAUUUCUCGC-AUUAUCAGU�UACCUCCGUCG�GCUUAC 
CCCACCUAGAGGAGCCUGUUCUGUAAUCGAUACCCCUCGCUAAACCUCA-CCACUUCUUGC-CAAACCCGCCUAUAUACCACCGUCGCCAGCCCAC - -- -- - ----- -- ----- --- --
uCaAaCUAGAGGAGCCUGUCCcAUAAUCGAUAcCcCACGaUAaACCCgA-CCACucUUggaAuacUcCAGCCUAUAUACCGCCGUCacCAGccuAC 
uCaaCCUAGAGGAGCCUGUCCUAUAAUCGAUACCCCcCGAUCuACCuCA-CCgcUUUUUG- -AAacUCAGcCUAUAUACCGCCGUCGuCAGCcUAC 
CCCACCUAGAGGAGCCUGUUCUAUAAUCGAUAAUCCACGAUUCACCCAA-CCACCCCUUGC-CAGCACAGCCUAC�UACCGCCGUCG£CAGCCCAC 
CCCACCUAGAGGAGCCUGUUCUAUAAUCGAUAAcCCACGuUaCACCCga-CCAuCuCUuGC-CcaugCAGCCUAC�UACCGCCGUCc£CAGCcCGC 
CCCAUCUAGAGGAGCCUGUUCUGUA!CCGAUAAACCCCGAUCAACCUCA-CCAACCCUUGC-UACUUCAGUCUAUAUACCGCCAUC�CAQ£AAAC 
UCCGUCUAGAGGAGCCUGUUCUAUA!UCGAUAAACCCCG�CUACCUUA-££CCUUCUCGC-UAAUUCAGC�UACCGCC��CAGCAAAC 
UCCAUCUAGAGGAGCCUGUUCUAUA!UCGAUAAACCCCGCUCUACCUCA-CCAUCUCUUGC-UAAUUCAGCCUAUAUACCGCCAUC�CAGCAAAC 
ucCuuCUAGAGGAGCCUGUUCUaUaAUCGAUAAACCCCGAUAaACCUcA-CCAauuCUuGC-UAAUaCAGuCUAUAUACCGCCAUC�CAGCaAAC 
uCCCuCUAGAGGAGCCUGUUCUGUAAUCGAUAAACCCCGAUCAACCUCA-CcaCCuCUUGC- - - - -UCAGCCUAUAUACCGCCAUC�CAGCAAAC 
CUAGAGGAGCCUGU C A CGAU C CG U ACC CC U C G CU AUACC CC UC AG C 
L30 'i 831 i 832 i 833 i L33 ' i 834 i L34 ' i 836 i40i 83 7 
3 8  3 8 ' 37 ' 4 0  4 0  ' 3 6 '  3 4 ' 3 2 ' 
CCU--GUGAAGGAAAUACAAUGGGCAAA AUA!--�--AAAAUUAAAAACGUCAGGUCGAGGUGUAGCAAAUG�AUGGGA!£AAAUGGGCUACA 
cCC--cUGA-GGcccacuAGUugGCAaAAUaga- -�agc�---aCac�ca�CGAGGUGUAGCacAUGggagG�g�AaAUGGGCUACA 
CCU- -GUGAAGGCUCAAUAGUAAGCAAAGUGGG--CACAACCCA-- -AAACGUCAGGUCGAGGUGUAGCGUACGAAGUGGGAAGA�AUGGGCUACA 
£A�- -GUGA!cQGUUGC-AGUAGGCUUAAUGACCUAACACGUCA-- -AUACGUCAGGUCAAGGUGCAGCUUAAGAAAUGGGAAGUAAUGGGCUACA 
CUC--GUGAGAGAUUCUUAGUAGGCUUAAUGAU--UUUUCAUCA- - -ACACGUCAGGUCAAGGUGUAGCAUAUGAAGUGGGAAGAAAUGGGCUACA 
cuu--augAgaGcacAaaAGUaAGCAaAAcuGc--aaacaCaauU- - - -AcGcCAGGUcAAGGUGUAGCuuAcaggGUGG-Ag�A�AUGgGCUAca 
CUU--aUGAaAGaaguauAGuAaGcaAAAuAgu-- - -CaccaAcUA-aAACGuCAGGUCAAGGUGUAGCAcAUaaaguGG-aAGA�AUGGGCUACA 
CUCUAAUGAAAGAACAACAGUGAGCUCAA�---UCCCUCGCUA-AUAA�ACAGGUCAAGGUAUAGCCUAUGGGGUGG-GAGAAAUGGGCUACA 
CUa---UGAaAGAACAauAGCGAGCACAAcAGC- - - -cacccGCUA-aCAA�ACAGGUCAAGGUAUAGCauAuGagaUGG-aAGAAAUGGGCUACA 
CCU- - -AAAGGG-AGAAAAGUAAGCAUAACCAU- - - - -CCUACAUAAAAAC�UUAGGUCAAGGUGUAACCCAUGGGUUGGGAAGUAAUGGGCUACA 
CCU- -AAAAAGGCACUAAAGUAAGCACAAGAAC- - - - - -AAACAUAAAaAC�UUAGGUCAAGGUGUAGCCAAUGAAGCGGAAAGAAAUGGGCUACA 
CCU- -AAAAAGGUAUUAAAGUAAGCAAAAGAAU-- - - -CAAACAUAAAAAC�UUAGGUCAAGGUGUAGCCAAUGAAAUGGGAAGAAAUGGGCUACA 
CCU--aAAAAGGaaaAaaAGUAAGCauAA�-----gauacaUAAAaACQUUAGGUCAAGGUGUAaCCuAUGaaauGGgAAGAAAUGGGCUACA 
CCU-GAUGAAGGcuACaAAGUAAQ£gCAAguAC-- - - - -CCACGUAAAGACGUUAGGUCAAGGUGUAGCCcAUGagGuGGCAAGAAAUGGGCUACA 
A G A G A C AGGU A C A G A G GCUA 
83 7 i L38 '  i 83 7 i 8L4 0 i 836 iL34 ,i 834 i 832 i 
Coelocanth 
Lungf ish 
Fish 
Frog 
Toad . 
Xantusi'd 
Skink 
Chicken 
Moa 
Whale 
Rat 
Mouse 
MooCow 
Human 
Coelocanth 
Lungfish 
Fish 
Frog 
Toad I 
Xantus id 
Skink 
Chicken 
Moa 
Whale 
Rat 
Mouse 
MooCow 
Human 
4 3 ' 4 5 4 5 ' 4 6  ---
U�ACAU--AGAAUAUU-- - - - - - - - - - -ACGAAAAAAACAG-CGAAACC�AC�-GAAGGAGGA�AGUA!AA�AAUAGAGAG- - ­
UUUUCUaC---- -GAAaAc-- - - - - - - - - - - -ACGgAcaaCcccA-uGAAAuugggGuu--ugAAGcUgga�uAGuA!G� 
UUUUCUACU-- -AGAAUAAG-- - - - - - - - - - -ACGAAUAGCAUCA-UGAAAACUUAAUGCUUGAAGGAGGA�UAGUAAAAAGGAAAUAGAGUG- - ­
AUUUCU-C- - - -AGAACAA-- - - - - - - - - - - -ACGAAAGACUAUA-UGAAAUUAUAAUCAU-GAAGGUGGAUUUAGUAGUAAAAAGAAA UAGAGUG- - ­
UUUUCUACCUU-AGAAUAA-- - - - - - - - - - - -ACGAAAGAUCUCUAUGAAACCAGAUCGAGAAAAGGCGGAUUUAGCAGUAAAGAGAAACAAGAGAG-- ­
uUUucUAAaac-agaauacgc- - - - - - - - - - -ACGgAa�CuuA-UGAAAaauaAcCu--aAaaggcgGAUUUAgCAGUAaaAuAaa-caAgaauu--­
CUCUCUcCC- - -AGAGaAc- - - - - - - - - - - - -ACgAAcaGCAucaAUGAAAcaCuGCuc- - -aAAGGuGgAUUUAGuAGUAAGAUaaa-CaAGAgA-- - ­
UUUUCUACU- - -AGAACAAA-- - - - - - - - - - -CGAAAAAGGAUG- -UGAAACCCGCCCU-UAGAAGGAGGAUUUAGCAGUAAAGUGAGAUCAUACCCCCU 
UUUUCUAacaU-AGAAcAccA- - - - - - - - - - -CGaAAGAgaAGa- -UGAAAcuC-UCc�-cagAAGGcGGA�AGUAAAaua�AcaAGAacG- - ­
UUUUCUACUA--AGAAC�UCCCCUAUACUCACACGAAAGUUUUUA-UGAAACUUAAAAACU-AAAGGAGGAUUUAGUAGUAAAUCAAGAGCAGAGUG- - ­
UUUUCUUCCC- -AGAGAACAUU-- - - - - - - - -ACGAAAC�--UGAAACUAAAGGAC--AAAGGAGGA�UAGUAAAUUAAGAAUAGAGAG- - ­
UUUUCUUCAA--AGAAC�UU-- - - - - - - - - - -ACUAUACCCUUUA-UGAAACUAAAGGACU--AAGGAGGAUUUAGUAGUAAAUUAAGAAUAGAGAG- - ­
UUcUCUacaccaAgAgaAucaagc- - - - - - - -ACGAAAGuuauuA-UGAAAccaauaACc- -AAAGGAGGAUUUAGcAGUAAaCUaagAAUAGAGuG-- ­
UUUUCUACC- - -AGAAAAcU- - - - - - - - - - - -ACGAUAaCCCUuA-UGAAAccUAAgGGUc-gAAGGUGGAUUUAGCAGUAAACUAAGAGUAGAGUG- - -
U U GAAA UUUA AG A A 
t 843 t L43 ' t 845 t L45 '  t 831 t , 846 
4 6 ' 3 0 ' 
- - -CCCCUCU -
- --UCCUUUUGAACC-CGGCUCU�A�-GCGCGU 
- --UUCUUUUUAACC-CGGCUCUGGG�ACGCGU 
- - -UUCC�AAAACGGCCCUGGA-GCGCGC 
- --�cuugaAGa-uugcuCUaGa-GcacGC 
- - -CuuaucUuAAAc-CAGCcCUGGA-GCGCGC 
AAGCUCACUUUAAGA-CGGCUCUGA�-GCACGU 
- - -cCcAuUUUAAgc-uGGCcCUgGG-GCACGU 
- - -CUUGAUUGAAUA-AGGCCA�GCACGC 
- - -CUUAAUUGAAUA-GAGCAAUGAA-GUACGC 
- - -CUUAAUUGAAUU-GAGCAAUGAA-GUACGC 
- - -CUuAGuUGAAuu-AGGCuAUGAA-GCACGC 
- - -CUUAGUUGAACA-gGGCCCUGAA-GCGCGU 
A U G 
t L46 '  t 830 t 
Oiapter 3, page 27. 
Identifying Individual Helices 
Helices 30-33 and 36 are well conserved among the vertebrates (Table 3.3), while six helices (26, 29, 34, 
43, 45, & 46) are quite variable in sequence but their pairings are easy to identify (Table 3.3). The helices 
26 and 29 are part of domain IT (Fig. 3.1). The area bounded by helix 34/34 ' shows the greatest potential 
for alternate structures (Table 3.3; Glotz & Brimacombe 1980, Dams et al. 1988, Neefs et al. 1990, Simon 
1991) and was examined in more detail. Helices 34, 36 and 37 tended to form spontaneously' in most of the 
vertebrates and so particular attention was paid to the effects of forcing other helices -to be included in the 
structure. 
Helix 35 
Examination of the base sequence between the proximal arms of helices 34 and 36 indicated that 
nucleotides in this region had the potential to pair in several of the vertebrates (see Table 3.3). This putative 
"helix 35" was not included in either Neefs et al. ' s (1990) compilation or in Hixson & Brown's ( 1986) 
primate model. Nucleotides which could participate in the helix are well conserve<l amongst members of a 
group (for instance skinks and ratites; Table 3.3). Except for the toad however, helix 35 did not form 
spontaneously using the MFOLD algorithm and its forced creation increased (made more positive) local 
free energies by more than 20% (Table 3.4). In addition, few fIxed compensatory mutations occur to 
support it's existence (fables 3.1 & 3.3). The 5 '  nucleotides for this helix could also be used in an 
extension of helix 34 (Table 3.3), so without experimental investigations (Noller 1984, Gutell et al. 1985) 
to test for its presence, helix 35 is not considered to be a general feature in vertebrate 12S rRNA. 
Helix 36 
This is one of the most conserved regions of domain ill (fable 3.3) but it can potentially pair in several 
ways (Figure 3.2). Structure a. in figure 3.2 is the one presented in table 3.3, but structures h. and c, tend to 
form in different taxa and/or when different helices are forced into the structure. Note that all forms of the 
helix have several unpaired "bulges". Such unpaired bases may act as protein recognition sites (Peattie et 
al. 1981 ,  Noller 1984, Gutell et al. 1985), and this point is discussed later. 
Helix 37 
Helix 37 can contain from four to six nucleotide pairs, though Neefs et al. (1991) only show it as a two 
base-pair structure. This helix may contain an internal unpaired bulge (Table 3.3). Conserved nucleotides 
and compensatory changes strongly support helix 37, and it forms without forcing using MFOLD (Table 
3 .4). 
O!apter 3, page 28. 
Table 3.4. Local minimal free energies for folding of helices 34, 36 and 37 in vertebrate 12S rRNA 
domain m. The additional energy cost required (less negative energy value) to include other helices - -
are shown, along with their percent energy differences from the minimal value. Energies were 
calculated using the MFOLD algorithm in the GCG package, version 7.2. The minimal energy for 
forming the alternate helix 40 in humans is also shown, and is lower than for placing helix 40 in the 
same position as other vertebrates. Energy differences more than 10% different (in kcal/mol) from 
the optimal value are considered less structurally feasible (Zuker et al. 1991 & pers. comm.). A value 
of 0% additional energy indicates that the helix forms without forcing. 
Forcing of helix: 
Taxon Energy 35 38 40 
(kcallmol) 
Fish -1 1.2 -7.3 -7.3 - 1 1 .2 
(35%) (35%) (0%) 
Toad - 2.5 -2.5 +0.2 -2.5 
(0%) (108%) (0%) 
Skink - 3.3 -2.5 - 1 .7 -3.3 
(25%) (49%) (0%) 
Chicken - 5.7 -3.3 -5.5 -5.7 
(42%) (3%) (0%) 
Moa - 2. 1 +0.3 +0.3 -2. 1 
(114%) (114%) (0%) 
Mouse - 6.2 +1 .4 -2.8 -6.2 
(123%) (55%) (0%) 
Cow - 6.8 -3.3 -6. 1 -2.9 
(52%) (10%) (57%) 
Human - 2.1 +2.9 -2. 1 -2. 1 
(238%) (0%) (0%) 
(alt S40) - 4.5 +1 .0 -4.5 -4.5 
(122%) (0%) (0%) 
Fig. 3.2. Potential pairings for helix 36 in vertebrate 11$ rRNA Structure a. is the one used in Neefs et 
al.(1990) compilation and shown in Table 3.3. The additio� guanine and pyrimidine nucleotides which 
can participate in the pairings are shown in bold in b. and c. Base pairings are indicated by "I" and "*" 
(for G*U), and gaps by .. _ n. The proximal (36) and diStal (36 ' ) arms of the heliX are labelled. Several 
sequences are available for the skink, moa, and cow groups, so variable sites in these groups are 
indicated by lower case letters. ' 
Fish 
36  
Toad 
Skink 
Moa 
Cow 
a. 
CUAUAUACCGCCGUC 
I I I  * I I I  I * * 1 
GAUGU_GGAGCUGG 
CUAUAUACCACCGUC 
I I I * I I I * *  I ·  
GAUGU_GGAACUGG 
CUAUAUACCGCCGUC 
I I I  * I I I  * *  I 
GAUGU_GGAACUGG 
CUAuAUACCGCCGUC 
I I I I I I I * *  I 
GAUAU_GGAACUGG 
CUAUAUACCGCCAUC 
I I I * I I I I * I 
GAUGU_GGAACUGG 
36 ' 
b. 
CUAUAUACCG_CCGGCGU 
I I I * I * I I  I I  I * *  
GAUGUGGAGCUGG_AcUG 
CUAUAUACC_ACCGuCGC 
I I I * I * I I I I * I 
GAUGUGGAACUGG_AcUG 
CUAUAUACCG_CCGUCGU 
I I I  * I * I I I  I * *  
GAUGUGGAACUGG_ACUG 
CUAuAUACCG_CCGUCGC 
I I I I 1 *  * I I I I 
GAUAUGGAACUGG_ACAG 
CUAUAUACCG_CCAUCuu 
I I I  * I * I I I I * 
GAUGUGGAACUGG_AuuG 
c. 
CUAUAUACCGCcGuCGU 
I I I  * 1 *  I I I I * *  
GAUGUGGAGCUGGACUG 
CUAUAUACCACCGUCGC 
I I I  * I * I I * I 
GAUGUGGAACUGGACUG 
CUAUAUACCGCCGUCGU 
I I I  * I * I I I * *  
GAUGUGGAACUGGAcUG 
CUAuAUACCGCCGUCGC 
I I I  I I * I I I I 
GAUAUGGAACUGGACAG 
CUAUAUACCGCCAUCUU 
I I I  * I * I I I * 
GAUGUGGAACUGGAuuG 
Otapter 3, page 29. 
Helix 38 
The energetic cost required to form helix 38 is high (Table 3.4). Although there is a large unpaired bulge 
between the distal aImS of helices 38 and 37 (Table 3.3), this bulge in itself may not be responsilbe for 
the increased energetic cost A similar size bulge occurs between helices 34 and 36 (Table 3.3). Helix 38 
is well conserved in sequence and size among the vertebrates however, and compensatory changes also 
occur (Table 3.3). Furthermore, in the skinks and many of the other vertebrates the adjacent unpaired 
region is one of the most variables part of the sequence (Tables 3.1 & 3.3), suggesting that the bases 
potentially able to form helix 38 are functionally constrained. Helix 38 is therefore included in the 
vertebrate secondary structure model. 
The hairpin loop at the end of helix 38 is often a "tetra-loop" (i.e. it consists of four bases; Table 3.3). 
tetra-loops can be unusually stable and frequently exist in the form of GNRA or UNCG (where N 
represents any base, and R is a G or A; Woese et al. 1990, Antao et al. 1991). They have been suggested 
to act as nucleation sites for organizing tertiary folding of the RNA (Tuerk et al. 1988). Fish, 
amphibians and lizards have an RTGA motif (Table 3.3). This also occurs in the great apes and birds, 
but the apes and chicken have an additional one or two bases in the loop respectively. Bovids and 
rodents have AAAA. Fur seaIs tend to have five adenines in this loop (G. Lento pers. comm.). The 
stability of GNRA tetra-loops are recognised in the MFOLD algorithm and assigned low (more 
negative) free energy values. The RTGA and AAAA motifs are not allocated similar bonus energies, but 
it may be possible that they too could confer increased stability in the molecule. 
Helix 39 
Helix 39, a two base pair structure in the Neefs et al. ( 1990) alignment, has little support in any of the 
vertebrates. Its suggested position falls within the most variable region of the skink molecule (Table 
3.1), and is also variable in lungfish, xantusiids, and bovids (Table 3.3). 
Helix 40 
Helix 40 is not shown for 12S rRNA in the Neefs et al compilation. In the vertebrates however a helix, 
usually four base pairs in size, can be formed starting four nucleotides after the end of helix 37 (and 
preceeded by two A's; Table 3.3). FIxed compensatory mutations (Table 3.3) and minimal energy 
calculations (Table 3.4) support this helix's presence and location in all but the rodents and primates. An 
energetically more favourable alternative helix 40 can be constructed closer to the 36 I helix for the 
primates (Tables 3.3 & 3.4). Additional experimental and sequence data are required to establish the 
existence of this alternate helix 40, but it cannot be formed in the other vertebrates (Table 3.3). 
The Refined Secondary Structure Vertebrate Model 
The model for domain III derived from these analyses is shown in Figure 3.3, using the skink sequence 
as a framework. This model differs from that of Glotz & Brimacombe (1980) and Dams et al. (1988) in 
the placement of some of the helices (Fig. 3.4). On the basis of number of bonds, free energy 
calculations (Fig. 3.4), and the occurrence of high sequence variability (Tables 3 . 1  & 3.3), neither the 
Fig. 3.3. Secondary structure model for L n. poJychroma 12S rRNA PCR product. Sites which vary among 
the 20 stink taxa are shown by lowercase letters. Every tenth base is marked by "." .  Dashes (-) and I signify 
nucleotide pairings (* for G*U), while a colon (":") deontes less certain pairings. Underscores U represent 
stylistic gaps used to simplify the diagram. For secondary structure analyses the molecule was divided into 
regions as indicated by the underlined numbers (S25 ' , S26, S20 ' , S2 ' & S29 are part of domain ll; see 
Table 3 .3). The positions of the PCR primers 12SA and 12SB are also shown. 
1 1 1 1  I I  
cAAcUGCCGAcuCG>12SA< 
A ill' 
A 
A aAA 
gAuAGuau a 
* I I I 1 * u .  S26 
UUaUCgug A 
-C A C 
-G 
-C 
-C 
• agc AaAGa a L38 ' 
-A S20 ' 
-G 
A 
G 
A .  
!!.!i' 
u U ill 
CC .A
CC 9 C 
C aCGAUCu u � A A � 
G : : : .  a 
U UUCC--GA a 
A--u 
� U--A .  
u* *g 
C- -G 
G--c 
A aA 
A 
c 
U 
A . C  I I I I  C .  A G .  • C A 
c A AGCUAAU ACCgcUUU
UUG cCUAUAUACCG_ccGuCGu 
AAGcaA U .A 1 1 1 1 1 1 1 : :  I I I  * I * I I I  I : :  
I I  S29  
cUCagA -­
a 
U- -AaGGugaaauAC_A_CGAUGUGGAACUGG_AcuG 
C--GA C 
A 
A .  
A 
C 
u 
C 
u g'  
A 
-A 
-G 
-G 
-A . 
C--G 
U--A 
S32 G* *U . 
U* *G 
C- -G 
C--G 
G- -C 
.A- -UA 
G 
G 
A--UA 
S 4 3  
CACUCUCUCC 
9 
. G  
I I I I I  
aGAGACu ( . 
A c A 
C 
C S30 
S31 G* *U 
L30 ' • A--U 
--uCa U--A A u C 9 L4 3 ' 
U • c a C--G 
C 
GG 
c 
A 
A .  
A 
Aa c c 
UGGCGGUGCUCCA 
I I  I * I I I I I I  
>12SB<CG_CGCGAGGU 
C 
G .  !!li' u--a 
c 
C 
U 
Au--A 
A U- -A 
G A c--G S4 6 
L4 6 '  ACc u--A 
.a--u 
. u--a 
u--a 
C a 
A C .  
9 a 
A GA 
c 
C--G 
. G- -C 
u--A 
C u 
c -· 
A 
A U 
AG 
A 
ill 
c 
a 
A • 
SL4 0 
A • C 
uAgu a 
I I I  I 
AUcA c 
a 
c 
Fig. 3.4. Potential alternative secondary structures for part �f skink domain ill bordered by helix 36. 
Structure a is based on the Dams et al. (1988) model, and slructure b on the model proposed by Glotz & 
Brimacombe (1980). Both these structures are energetically less favourable for the skink sequence than 
the model developed in this thesis (structure c), which has a local free energy of - 1 .7 Kca1Imol. Base 
pairs are shown by -, I, and *, while artificial gaps included to maintain alignments are indicated by 
underscores U. The underlined numbers correspond to helices shown in figure 3.3, and the periods (It.It) 
marie every tenth base. 
a. Dams et al. (1988) model 
Free Energy = 6.8 Kcallmol 
. 
GAa�_aagca 
II u l l l  * I a 
aUUCCA uGA 
· U--a . 
II 
3 6  
• • C 
u**g 
C--G 
G--c 
A aA 
A 
h. Glotz & BrimacoI?be (1980) 
free Energy = 13.1 Kcallmol 
gcaa 
a A 
a G 
. G* *u 
A A .  
a g 
A G 
G--c 
U--a 
• a A 
ACCUU--A 
U I I  � 
uCGA 
• • C 
CUAUAUACCG_CCGUCGU A • C 
uAgu a 
CUAUAUACCG_CCGUCGu 
I I I  * I * I I I I I I I * I * I I I I **  
GAUGUGGAACUGG_AcuG 1 1 1 1  
AUcA c 
GAUGUGGAACUGG_A c c 
C 
A a .  - a 
A 4 0  
c. Skink Model 
Free Energy = -1.7 Kcallmol 
• agc AaAGa a 
G i l l . a 
U UUCC- -GA a 
A- -u 
3 7  U--A. 
u* *g 
C--G 
G--c 
l§. A aA 
• • C A 
CUAUAUACCG_CCGUCGU A • C 
I I I * I * I I I  I I I 
GAUGUGGAACUGG_AcuG 
uAgu a 
I I  I I  
C AUcA c 
A A a 
a 
iQ 
c 
C a 
u- -A 
A 
g .  
a 
u 
C 
o.apter 3, page 30. 
Glotz & Brimacombe nor the Dams et al. models fits the skink or other vertebrate sequences as well as 
the new model presented (see also Cooper, Hickson, Lento, Simon & Penny in prep.). 
Conserved Positions in the Secondary Structure Model 
The model shown in figure 3.3 has 120 strictly conserved nucleotide positions among the vertebrates 
(human to fish; Fig. 3.5). Fifty (42%) of these bases are unpaired. While similar numbers of adenine, 
guanine, cytosine and uracil are conserved, conserved adenine residues are twice as Common at unpaired 
sites (fable 3.5), in agreement with observations of other small subunit rRNAs (Gutell et al. 1985). 
Variability within Domain m 
To investigate how the molecule varies I have divided domain ill into sections corresponding to 
stem+hairpin loops and other unpaired regions (see Table 3.3). The conserved vertebrate core of domain 
ill comprises helices 31 ,  32, 33 and 36 (Figs. 3 .3 & 3.5). Note also the generally conserved nucleotides 
in the unpaired region between helices 33 and 34 and between helices 45 and 3 1 ,  and the conserved 
UGAAA motif in the loop of helix 45 (fable 3.3, Fig. 3.3). The single-stranded regions (S20 ' & S2 , )  
linking domains II & ill are also well conserved. 
The six groups of vertebrates for which several sequences are available (skinks, xantusiid lizards, ratites, 
bovids, apes and mammals) show similar trends in variability (Fig. 3.6). Region S26 is one of the most 
variable, as determined by number of positions which vary. Other commonly varying regions are S34, 
SL37, SL40 (Figs. 3.6a-O. Greatest variability in ratites occurs in SUO (Fig. 3.6c), while S34 and S45 
have the most variable sites in the great apes (Fig. 3.6e). Except for the xantusiids, there are relatively 
few insertion/deletion (indel) events within members of these families. Small, one- or two-base, indels 
occur in unpaired regions in S26, S29, L34 ' , L38 ' , SUO, U3 ' & L46 '  in the skinks, ratites, and 
l , 
bovids (fable 3.1,  Appendix I). Xantusiids however have indels occurring in fourteen of the 23 regions 
of the model (Appendix 1). 
Nucleotide Substitutions in the Skink Structure 
The comparisons of regions shown in figure 3 .6 are a minimal estimate of variability, identifying only 
sites which have changed and not how frequently they may vary. When this is adjusted for in skinks (by 
multiplying each variable position by the number of taxa which vary at that site), regions L30 ' and 
L38 ' are the most variable (Fig. 3.7). These are both small 6-7 base internal loops (Fig. 3 .3). Eight 
positions in the skink sequences have more than two character states, and four of this very variable sites 
are in either L30 ' or L38 I (see Table 3. 1). Whether closely related taxa tend to differ at these very 
variable sites will be examined in the next chapter. 
In the secondary structure model for skinks (Fig. 3.3), significantly more changes <x?= 4.79, df = 1 ,  P < 0.05) 
�occur in unpaired regions than in paired regions (Table 3.6). There are 86 paired poSitions 
(excluding the bases in S25 I , S20 I , & S2 whose complementary pairs are not in the sequence 
examined). Twenty eight (33%) of these have changes in at least one taxon (fable 3 . 1). Nucleotide 
Fig. 3.5. Conserved <I> and non-conserved OD sites in vertebrate 12S rRNA domain III, based on 
comparisons of fish, amphibian, reptile, bird and mammal sequences (see Table 3.3). Every tenth base is 
marked by ".n.  Dashes (-) and I signify nucleotide pairings (* for G*U), while a colon (":") denotes less 
certain pairings. Underscores U represent stylistic gaps used to simplify the diagram. For secondary 
structure analyses the molecule was divided into regions as indicated by the underlined numbers (S denotes 
a stem region, and L a loop region). 
-I 
-I 
-I 
:::�: S 2 0 ' =. -
I S 2
' 
:::::: 
-I 
-I 
-I · 
I I I * I I I I I  I 
> 1 2  S B<:tLHt::I f fft::tl 
Hf' iJ- -t�: SL46 .:.:.: .:.:.: 
!d§.' 
L34 ' 
• 
• 
ill �H:--ri . 
• 
OIapter 3, page 31.  
Table 3.5. Distribution of conserved nucleotides at paired and unpaired sites for the vertebrate 12S 
rRNA model. Since 41 .7% of the conserved sites shown in figUre 3.5 are unpaired, the expected 
numbers of conserved nucleotides at unpaired sites were calculated by multiplying the total numbers of 
each conserved nucleotide by 0.417. Conserved adenine residues are more common at unpaired sites, 
while conserved guanines tend to occur more frequently at paired sites. 
No. Conserved 
A C G U Total 
Total 36 27 33 24 120 
Paired Obs. 12 17 24 17 70 
Unpaired Obs. 24 10 9 7 50 
Exp. 15.0 1 1 .2 13.8 10.0 
X2 = 8.10, P < 0.05 (3 degrees of freedom). 
Table 3.6. Changes occurring in paired and unpaired regions for the skink 12S rRNA data set (see Fig. 
3.3). Expected numbers of changes are calculated on the basis of numbers of paired and unpaired sites. 
No. Sites Which Vary 
Bases Observed Expected 
Paired 190 39 50 
Unpaired 194 62 51  
Total 384 101 
X
2
= 4.79, df = I ,  P < 0.05 
ZO 
1/1 
10 
/I 
0 
ZO 
1/1 
10 
'" � /I en 
.. 
;Q 0 co 0;: 
co ;> ... c .. u .. 20 .. I:lo 
1/1 
10 
/I 
0 
ZO 
1/1 
10 
/I 
0 
20 
1/1 
10 
/I 
0 
8. 
b. 
C. 
d. 
e. 
f. 
RegioDS of variability in domain m 
of �brate 12S rRNA. 
20 
15 
10 
Fig. 3.7. Regions of variability in the skink 12S rRNA sequence. This figure differs from figure 3.6 in 
that the degree of variability of each site was accounted for using the following fonnula: 
(number variable sites x number taxa which vary) x 1 0 0  
(total  number o f  sites x number of tax� ) 
In contrast to figure 3.6a. where S26 has the highest proportion of variable sites. the most variable 
positions in domain III occur in L30' and L38 '  . See also figure 4.17. 
926' &20' 
S28 S2t 
S2' CIO' 
SSG su 
-:�.: 
' ..... 
W' 
Olapter 3, page 32. 
pairing (including G-U) is retained in 56 (65%) cases, significantly more than would be expected to 
occur by chance (18%; ..,!= 84.80, degrees of freedom = I ,  P < 0.001; see also Table 3.7). Over half 
(32/56) of the substitutions which maintain complementary pairing are double changes. 
There are fewer than expected purine to purine (G H A) transitions (:x?= 4.11 ,  df = I ,  0.025 < P < 0.05; 
Table 3.8), and a lower number of G H C trans versions (Table 3.8). Although there is a tendency for 
more conserved areas (� 25% of sites vary) of the sequence to have a higher proportion of guanine 
residues (Table 3.9), the base composition of conserved nQcleotides is similar to the proportions 
observed for all sites in the skink sequence (Table 3.9). This suggests that differences in the types of 
transitions and transversions (Table 3.8) is not due to differences in base composition. 
DISCUSSION 
Two major points emerge from the refinement of a vertebrate secondary structure model for domain m 
of 12S rRNA. The first is that one third of the sites do not change. This conservation is associated with 
the single-stranded region linking domains II and m, and what can be termed the structural core of 
domain ill (helices 3 1 ,  32, 33, and 36; Fig. 3.5). The proportion of conserved sites in this model will 
decrease as more vertebrate sequences are added to the compilation, but the results presented here serve 
to indicate that many sites in the molecule may not be free to vary. Having part of the molecule 
effectively invariant has implications for models of sequence analysis and this will be considered in 
Chapter Five. 
The second point is that both paired and unpaired regions can have high levels of variability. Helix 26 
(and its hairpin loop), helix 34, and the unpaired region following helix 38 have many variable sites in 
skinks, xantusiids, ratites and bovids (fable 3.3, Fig. 3.6). The phylogenetic and functional significance 
of these changes are not yet clear, but complementary changes preserve the helices. 
Interactions Between Domain ill and Ribosomal Proteins 
Adenine residues tend to be more highly conserved in unpaired than paired regions (Table 3 .5). Gutell et 
al. (1985) suggested that, since adenine is the least basic nucleotide, high conservation of adenine in 
unpaired regions may reflect hydrophobic interactions between adenine and ribosomru proteins. This 
does not however account for the observation that nearly one quarter of the conserved A' s are located in 
the single-stranded regions linking domains II and ill (Figs. 3.3 & 3.5), where ribosomal proteins do not 
appear to specifically bind (Ehresmann et al. 1990, Noller et al. 1990). The ribosomal proteins S7, S9 
and S 19 bind to domain III (Stern et al. 1989, Ehresmann et al. 1990. Noller et aI. 1990), and the 
conserved residues in L33 ' and SL45 (Fig. 3 .5) may be involved in this binding. particularly with 
respect to S19. Noller & Woese (1981) suggested that some phylogenetically variable regions may be 
protein binding sites. Helix 33 is well conserved however, and appears to interact with S19 (Stern et al. 
1989, Ehresmann et al. 1990. Noller et al. 1990). Simon et aI. (1990) also noted the conservation of this 
helix in cicadas. Helix 34 in contrast, can be very variable (fables 3.1  & 3.3), but does not appear to be 
a primary site of protein binding (see Ehresmann et al. 1990). These observations suggest either that the 
ClIapter 3, page 33. 
Table 3.7. Substitutions among paired nucleotides in the skink 12S rRNA sequence data (helices 25 I , 
20 I & 2 � are excluded). Substitutions are categorized into changes maintaining base pairing (G*U pairs 
allowed), and those (in either direction) between complemeptary and non-complementary pairs. 
Expected numbers of changes were calculated separately for single and double substitutions using Dixon 
& Hillis' (1992) probabilities for maintenance of complementary pairings (0.125 and 0.256, 
respectively). For example, 52 single changes were observed, of which (52 x 0.125) = 6.5 are expected 
to maintain base pairing by chance. Of 17 pairs where both bases changed, 16 (representing 32 
substitution events) maintained complementary base pairing, whereas only (17 x 0.256) = 4.4 paired 
substitutions could be expected to occur. The differences between observed and expected numbers of 
fIXed compensatory changes are highly significant (P < 0.01) for both single and double substitutions. 
Type of Substitution 
Single 
base pair to base pair 
base pair to non-pair 
Double 
base pair to base pair 
base pair to non-pair 
Total 
Expected 
6.5 
45.5 
4.4 
12.6 
Observed X2
� 
24 
28 
53.8 
16 
1 
41.3 
86 
Chapter 3, page 34. 
Table 3.8a. Expected probabilities for each type of transversion and transition in the skink data set, 
based on observed nucleotide frequencies (numbers in parentheses). Assuming that the base 
composition of the sequence determines the frequency with which, for instance, an A changes to a C 
rather than to a U, then an A to C change should occur with a probability of [0.27/(0.27 + 0.19)] = 0.59. 
S imilarly, on the basis of the relative frequencies of A and G, a C is twice as likely to change to an A 
than it is to a G. On the basis of nucleotide frequencies, slightly more A to G transitions are expected 
than C to U. 
Transversions: 
(0.35) 
A 
0.67 
0.41 
U 
(0. 19) 
(0.19) 
G 
0.33 
0.59 
C 
(0.27) 
Transitions: 
(0.35) 
A 
U 
(0.19) 
0.54 
0.46 
(0.19) 
G 
-<:..:.. � c 
(0.27) 
Table 3.8b. Observed and expected (based on Table 3.8a) numbers of each type of transversion and transition 
among the 20 skink taxa. The direction of change is not assumed. 
Transversions : 
No. Obs. 
No. Exp. 
Transitions: 
No. Obs. 
No. Exp. 
A <-> C 
19 
18.2 
A <-> G 
40 
49.7 
A <-> U 
1 1  
7.6 
C <-> U 
52 
42.3 
G<-> C 
6 
1 1 .4 
G<->U 
2 
3.8 '1.2= 4.97, df = 3, P > 0.1 
'1.2= 4.1 1 ,  df = 1 , 0.025 < P < 0.05. 
Cllapter 3, page 35. 
Table 3.9. Base composition of conserved residues among regions of skink 12S rRNA domain ffi. 
Regions in bold highlight those where more than 75% of their sites are conserved. The expected 
proportions of each conserved nucleotide (based on nucleotide frequency; Table 2�1) is shoW? at the 
bottom of the table. 
No. Conserved Conserved 
Region Bases No. % % A  % C % G  % U  
S25' 17  14 82% 43 29 21 7 
S26 24 12 50% 50 17 8 25 
S20' 1 5  13 87% 46 23 23 8 
S29 10 6 60% 50 17  17  17  
S2' 18  14 78% 43 21 21 14 
S30 21 21  100% 10  33 44 14 
L30' 6 2 33% 50 50 0 0 
S31 1 5  14 93% 29 14 29 29 
S32 26 26 100% 27 23 3 1  19 
S33 18  17  94% 29 35 12 24 
L33' 8 6 75% 33 67 0 0 
S34 23 14 65% 14 21 21 43 
L34' 9 6 67% 50 17 17  17 
S36 34 32 94% 22 25 31  22 
S37&38 22 16 73% 25 19 3 1  25 
L38' 7 1 14% 100 0 0 0 
SL40 23 13 57% 69 23 0 8 
S43 16  13  81% 15 46 15 23 
L43' 8 5 62% 80 20 0 0 
S45 20 13  65% 39 30 23 8 
L45' 10  6 60% 50 0 50 0 
S46 25 12 48% 50 17 17 17  
L46' 9 7 78% 43 43 14 0 
384 283 Obsened: 34% 26% 23% 17% 
Expected: 35% 27% 19% 19% 
X2 = 1 . 1 1 ,  df = 3, not significant. 
Chapter 3, page 36. 
mitochondrial small subunit rRNA molecule may not have the same points of protein contact as the 
prokaryote molecule, or that not all protein binding regions are variable. The conserved nucleotides in 
the unpaired regions L33 ' and SLA5 (Fig. 3.5), and their probable interaction with S 19  indicate that the 
latter point may be the more likely interpretation. 
Helix 36 is the probable site of contact with the 50-S ribosome subunit in E. coli (Gutell et al. 1985). 
The unpaired bases in this helix, as well as the different potential alternate forms of it (Fig. 3 .4), may be 
associated with protein recognition sites (see Peattie et al. 1981 ,  Gutell et al. 1985, Noller 1990). 
Tetra-loops 
While the hairpin loop of helix 38 does not conform in either sequence (or size in some cases) to the 
common GNRA tetra-loop rwoese et al. 1990, Antao et al. 1991), it is interesting that the last 
nucleotide in the loop is generally an adenine (Table 3.3). Sea urchins (sequences not shown) do not 
conform to the GNRA or RTGA motif for loop 38 either, but they too have adenine as the final base. As 
already noted above, adenine residues have been found to be associated with protein binding (Gutell et 
aI. 1985), so the helix 38 hairpin loop and the adenine residue may have a critical role in the formation 
of the tertiary structure. The other tetra-loop in the 12S rRNA alignment presented here is in helix 29 
(Table 3 .3). Its sequence is not as conserved as that in helix 38, nor does it fit the more common tetra­
loop motifs rwoese et al. 1990), but the terminal base is also generally an adenine (Table 3 .3). 
- Insertion/deletions commonly occur in this hairpin loop however (Tables 3 . 1  & 3.3) so the significance 
of these putative tetra-loops is uncertain. The number of tetra-loops varies among vertebrate 12S rRNA 
sequences. There are four in human and seven in the rat (Gutell et al. 1985), suggesting that the tetra­
loops may not have a major functional role. 
Comparison of Secondary Structure Models 
The model presented here is very similar to Noller & Woese's (Gutell et al. 1985, Noller et al. 1990), 
but it provides a more detailed view of the vertebrate 12S rRNA domain III. Previous alignments and 
models used only human, cow, mouse, and frog 12S rRNA sequences (Gutell et al. 1985, Dams et al. 
1988, Neefs et al. 1990). Addition of reptile and bird sequences in the present study emphasizes 
differences in sites of variation between groups (Fig. 3.6). The structure in figure 3.5 is different from 
that proposed for the great apes (Hixson & Brown 1986). Helix 36 in the Hixson & Brown model is 
shifted slightly, and a five base pair helix (labelled 24 by Hixson & Brown) occupies the region 
encompassing helices 38 and 40 in figure 3 .5. Neither of these two features is supported by the 
comparative approach adopted here. 
Simon et al. (1990) found that 12S rRNA sequences from several species of cicada could fit both the 
Glotz & Brimacombe ( 1980) and Dams el al. (1988) models for the region between helices 37 and 40. 
Rather than being a conflict in structure, Simon el al. suggested that local switching between alternative 
structures may occur. Analysis of a wider range of 12S rRNA sequences indicate that this switching may 
not occur. On the basis of comparative sequence analysis, the presence of compensatory changes, and 
OJapter 3, page 37. 
examination of minimal free energy structures, one structure, common to all of the vertebrates 
examined, as well as some invertebrates, is most favoured (Fig. 3.5 and Cooper et al. in prep.). As 
Simon et al. noted though, and the present analyses indicate, experimental evidence for the structure of 
this region is desirable. Switching may occur in other parts of the rRNA, and play a role in ribosome 
assembly (Glotz & Bimacombe 1980, Zwieb et al. 1981). 
The degree of base conservation in a helix does not appear to be re,lated to whether the nuclootides of 
one side of the helix are close to (short-range) or further away from (long-range) the nucleotides 
forming the other side of the helix, as has been suggested by Simon (1991). Helix 33 is a short range 
stem and well-conserved, whereas helix 34 can be classed as a long-range stem but is more variable 
(Figs. 3.3 & 3.5). It is interesting to note that in the skinks, most of the variability in helix 34 occurs on 
the distal arm of the helix, though pairing is still maintained (fable 3.1). This could imply that the five 
uracil residues on the proximal arm are more constrained, but both sides of this helix vary in other 
vertebrates (fable 3.3). 
Limitations of the RNA Folding Algorithm 
The energetic calculations were of some use in discounting the existence of helix 35, but implied that 
helix 38 was unlikely to form (fable 3.4). Zuker et al. (1991) noted however that optimal energy 
solutions for domain III of 16S rRNA were not always in agreement with the structure determined by 
comparative sequence analysis, though the reasons for this are unclear. Experimental evidence (Glotz & 
Brimacombe 1980, Gutell et aI. 1985), or the occurrence of fixed compensatory mutations are required 
before helix 35 can be accepted. The potential for an alternative helix 40 to form in humans and other 
apes (fables 3.3 & 3.4) also requires further sequence and experimental analyses. The secondary 
structure folding of the 16S rRNA appears to be primarily determined by the sequence and experimental 
evidence does not indicate that ribosomal proteins have a significant role in stabilizing structures 
(Draper 1990, Noller et al. 1990). 
Another limitation in the folding algorithm (Zuker 1989, Zuker et al. 1991) is that non-standard base 
pairs (for instance A-C, and G-A) are not yet considered in the calculations, so minimal free energies 
could be underestimated. A-C pairs can be relatively common in rRNA and tRNA (Topal & Fresco 
1976, Anderson et al. 1981, de Bruijn & KIug 1983, Kraus et al. 1992), though they have larger free 
energies (Freier et al. 1986). A-G pairs may also occur, but their formation may be determined by 
adjacent nucleotides (Cheng et al. 1 992). Interactions between bases within loops can also occur (Heus 
& Pardi 1991), which may also affect the free energy level of the molecule. A-C pairings have the 
potential to occur in helices 36 and 37 (Table 3.3, Fig. 3.3), and could also increase the size of helix 40 
in mammals (Table 3.3). Taking account of A-C pairings may therefore support the more common helix 
40 over the alternative helix 40 in the great apes (see Tables 3.3 & 3.4, and Fig. 3.5). While non­
standard pairing may be a possibility in helix 36, the unpaired nucleotides may be important for protein 
binding, and experimental investigations suggest that unpaired bases occur in helix 36 (Gutell et al. 
1985, Noller et al. 1990). 
Cliapter 3, page 38. 
Variability between Regions 
Regions of relatively high or low variability may be more informative for examining more closely or 
distantly related taxa respectively (Kocher et al. 1989, Thomas & Beckenbach 1989, Simon 1991, 
Simon et al. 1991). Different regions in the 12S rRNA sequence do have different levels of variability 
(Figs. 3.6 & 3 .7). Comparisons between closely and more distantly related skink taxa in relation to the 
more variable regions and sites will be considered in Chapters Four and Five. 
Knowledge of which regions or sites are more conservative can be also be useful for determining the 
closeness of a relationship. A change shared by two taxa at a more conservative site in the molecule may 
be a more reliable indicator of true phylogenetic relationships than when taxa share a common 
substitution at more variable positions. L. fallai and C. aenea for example have the same change at 
position 73 (Table 3.1), a conservative region in vertebrates (Table 3.3). C. aenea, L fallai and L 
maccanni also differ from other taxa by a change in the otherwise conserved loop of helix 33 (Table 
3.1), suggesting that C. aenea and L fallai have a relatively close relationship. However, a change in the 
tetra-loop of helix 38 could also imply a strong relationship between C.aenea, L. maccanni and L. 
tel/airi (position 194; Table 3.1). These conflicts will be considered in Chapter Five. In contrast, L.fallai 
shares two substitutions with a New Caledonian skink, Tropidoscincus rohssii (see Chapter Five), but 
these are both at quite variable positions in helix 46, implying that these two taxa are not that closely 
related. 
Secondary Structure Models and Sequencing Precision 
Placing nucleotide sequences in the context of a secondary structure model has the additional advantage 
of identifying possible errors in sequence determination. There are several cases in Table 3.3 where one 
taxon differs from the rest at positions which are otherwise constant. These may be real differences or 
the results of cloning, PCR, andlor sequencing artifacts. Seven positions in the skink sequences are 
variable, while they are constant among the other vertebrates. At least three primers were used to 
sequence the skink PCR fragment and several sequencing reactions run. Rechecking the sequencing gels 
did not reveal ambiguities at these seven sites suggesting that errors during sequencing are probably not 
the cause. At five of these positions several skink taxa vary (Table 3.1) supporting the reliabilty of the 
amplification and sequencing reactions. At another site (position 233) however a gel reading error was 
found (and corrected) when the sequence was compared to the secondary structure model. 
During the compilation of table 3.3 it was noticeable that sequences obtained by cloning had more 
inconsistencies that sequences obtained by PCR. Changes in otherwise well conserved positions were 
particularly frequent in the frog and whale sequences. The frog sequence differs uniquely at seven 
locations, including a deletion in L30 I (Table 3.3) and additional sequencing of this region is required 
to confrrm the sequence. In the whale sequence there is a large, approximately 10 base, insertion in the 
unpaired region following helix 43 (Table 3.3). This unusual insertion is nol remarked upon by the 
Olapter 3, page 39. 
authors (Amason et al. 1991). In addition, the whale sequence has only three nucleotides in the loop of 
helix 38, whereas other mammals have at least four (fable 3.3). The whale also has an apparent 
insertion of a guanine residue into the 30 ' helix (fable 3.3). 
There are also problems with some of the other sequences. Three 1 2S rRNA sequences for Rattus 
norvegicus can be found in GenBank (accession numbers Xl4848, J01438 & V00680). The latter two 
- .-
differ at several sites though from XI4848, the one used in table 3.3. The quail 12S rRNA sequence 
(Desjardins & Morais 1989) was excluded from these analyses altogether because a block of about 50 
nucleotides was missing. Sequencing errors are a recognized problem in the databases (see for example 
Clark & Whittam 1992, States 1992). Corrections to the clolled toad 1 2S rRNA sequence have already 
been published (Dunon-Bluteau & Brun 1986). It is prudent to confIrm the nucleotide sequences 
discussed here by resequencing the relevant species. 
Bias in Types of Substitutions 
Bias in types of nucleotide substitutions has been observed in many sequencing studies (for example, 
Aquadro et al. 1984, Thomas & Beckenbach 1989, Marshall 1992, Knight & Mindell l993; see also 
Moritz et al. 1987). Thomas & Beckenbach (1989) attributed a C-+A transversion bias to preferential 
loss of guanine residues by depurination (Lindahl & Nyberg 1972). The direction of transversions 
cannot be determined for the skinks since an appropriate outgroup is not available, AHC transversions 
are the most common however (fable 3.8), and this bias does not appear to be related to base 
composition (Table 3.9). 
page 40. 
Chapter Four: TransitWns afUl Transversions 
Two major features of the skink sequence data set are that, fIrstly, many taxa are about equally divergent 
I 
from each other, but secondly, there is considerable variation in the numbers of transitions and 
transversions between them. These two features are examined in tqis chapter and compared with other 
vertebrate data sets and with simulated data. The patterns of differences are examined with respect to the 
secondary structure model developed in Chapter Three to unflerstaild both how the molecule changes 
and to identify the most variable regions. This information is subsequently used in Chapter Five. 
Transitions and Transversions 
There are 101 variable positions in the skink data set trable 3.1), of which 58 are phylogenetically 
informative (parsimony sites). Fifty of the nucleotide substitutions are singletons, that is one taxon 
differs from all the rest. Two single base deletions occur - at positions 25 (in both L. telfairi and L. 
maccannO and 6t(L. telfairi only). L. telfairi also has an extra base inserted between nucleotides 202 
and 203 (Table 3.1). 
Most pairs of taxa have between 20 and 25 observed differences (Table 4.1a, Fig. 4.1 a), and four to five 
transversions in pairwise comparisons (Table 4.1b, Fig. 4.tc). 1bere is Considerable variability however. 
Some taxa, for instance "Stewart Island Green", L. grande anq La. guichenoti have more variation 
associated with the number of transitions, while others, such as L. notosaurus, C. aenea and L. 
inconspicuum, show greater variation with respect to transversions (Table 4.2). L. infrapunctatum has 
the highest mean number of transversions and also the lowest trailsitionltransversion (Ts/Tv) ratio � 
(Tables 4.1 & 4.2). 
While there is a substitution bias in favour of transitions (Table 4.1 b), a very weak correlation exists 
between the number of transversions and the number of t:I1llsitions (,.2= 0.03; Table 4.3 ,  Fig. 4.2). 
Consequently, transversions are a poor predictor of the total number of differences between taxa 
(,.2= 0.26, Fig. 4.3), particularly when there are fewer than eight transversions (Table 4.3). As a 
comparison, stronger relationships exist between the n�bet of transitions and transversions in both 
ratite (,.2= 0.40, Fig. 4.4a) and bovid (,.2= 0.55, Fig. 4.41» data sets. 
Saturation of Transitions? 
There is no evidence to suggest that the skink sequences have become saturated with nucleotide 
substitutions. The skink sequences are less than 10% divergent and the proportion of transitions does not 
decrease as sequence divergence increases (Fig. 4.5), which is usually an indication of saturation 
(Wilson et aZ. 1985, DeSalie et aZ. 1987, Miyamoto & Boyle 1989, lrwin et aZ. 1991). 
OIapter 4, page 41. 
Table 4.1a. Distance matrices for skink 12S rRNA PCR fragment (384 base pairs). Total observed 
I 
nucleotide differences are given below the diagonal ahd perCentage transitions are above. Numbers in 
bold denote pairwise comparisons with 20 or fewer total differences. Underlined numbers identify 
where transitions comprised less than 75% of the differences. The mean numbef of nucleotide - . 
differences for each taxon is also given. Insertions and deletiorls when they bccurred in pairwise , , , 
I 
comparisons were not counted. Taxon abbreviations as in Table 3 . 1 .  
Number of Differences (below diagonql ) \  Percent Transitions ( above ) 
Sig Lgr Lno Llc Lsu Lnn Lmi Lag Ca. Lac Lin Lam Lz. Lfa Lma Lt. Lop Lfr Lot Lmo 
Sig �% 1l% 21% 75% 77% 72% 80% 82% 7 4 %  76%  75% 77% 74%  74%  83% 70%  7 %  88%  82% Sig 
Lgr 8 80% 80  79  85 75  71 87 1l 79 79  83 79 87 75  79 71 83 87 Lgr 
Lno 15 15 86%  7 5  75 76 72 8 1  � 92 79 78 74 90 78 81 6 8  7 8  80 Lno 
Llc 17 20 2 1  84% 79 74 67 7 9  7 8  85 75 73 6 8  95 83 84 73 87 77 Llc 
Lsu 16 14 16 19 90% 72 76 83  77  75 76  81  77  83  72 91 74 80  83 Lsu 
Lnn 17 20 16 14 20 79%  7 5  83 77 7 9  80 81 7 6  84 72 69 7 1  83 8 1  Lnn 
Lrni 18 16 2 1  2 3  18 24 74%  87 7 6  7 6  100 90 8 1  83  21 77 73  85  87 Lrni 
Lag 15 2 1  2 5  18 2 1  20 23 78%  7 1  72 7 6  77 21 75  67 75 67  8 6  77 Lag 
Cae 22  23 2 1  2 4  23 23 23 23 78%  78 88 87 8 8  84 68 79 7 5  87 96 Cae 
Lac 19 22 2 5  18 22 22 25 14 2 3  �% 7 9  76  1l 87 74  77 7 5  92 78  Lac 
Lin 2 1  19 12 27 20 24 25 29 23 29 79%  74 7 4  85 7 8  83 67 7 4  7 9  Lin 
Lsrn 20 19 24 24 2 1  25 5 25 24 28 28 91% 80 84 74 80 74 87 88 Lsrn 
Lze 22 24 23 22 2 6  26 20 27 23 25 23 23 7 9 %  83 70 84 2Q 87 84 Lze 
Lfa 2 3  24 23 22 2 6  25 26 24 17 26 27 25 24 76%  71 75 67 84 88 Lfa 
Lrna 24 22 20 2 1  23 25 30 28 25 30 20 31 30 25 89% 86 76 89 83 Lrna 
Lte 2 3  24  2 3  2 4  25  25 24  30  22 27 27  27 23  3 1  2 8  80% � 7 8  80  Lte 
Lnp 23 24  21 19 21 13 22 28 28 30 29 25 32 32  29 30 74%  7 8  77 Lnp 
Lfr 2 4  2 8  2 8  22  27  24 30  12 2 8  24 30 3 1  27  2 7  29  32 34  8 1% 72 Lfr 
Lot 26 23  23  30 25 29  26 28 23  24 23  31 31 32  28 27 32 32  85% Lot 
Lrno 27 30 2 5  2 6  30 26 3 0  2 6  22 27 2 9  33 25 2 4  29 30 3 1  2 9  27 Lrno 
Sig Lgr Lno L1c Lsu Lnn Lmi Lag Ca. Lac Lin Lam Lz. Lfa Lma Lt. Lnp Lfr Lot Lmo 
Means 20 . 0  20 . 9  21 . 7  22 . 6  23 . 2  24 . 5  25 . 1  26 .2  26 . 5  2 7 . 4  
20 . 8  21 . 6  22 . 0  23 . 0  24 . 2  24 . 7  25 . 4  26. 4 2 7 . 3  2 7 . 7 
Sig 
Lgr 
Lno 
LIe 
Lsu 
Lnn 
Lmi 
Lag 
Cae 
Lae 
Lin 
Lsm 
Lze 
Lfa 
Lma 
Lte 
Lnp 
Lfr 
Lot 
Lmo 
ClJapter 4, page 42. 
Table 4.tb. Pairwise comparisons of transitions (above diagotlal) and transversions (below). Bold , ' 
denotes comparisons with 20 or fewer total differences. The �eail riumbers of transitions and 
! ' 
transversions for each taxon are shown on the right and below the table, respectively. 
Transversions (below diagonal ) \ Transit ions ( above) 
Big Lgr Lno L1c Lsu LnD Lmi Lag Cae Lac Lin Lsm Lze Lfa Lma Lte Lnp Lfr Lot Lmo Mean Ts 
3 
4 
5 
4 
4 
5 
3 
4 
5 
5 
5 
5 
6 
4 
7 
6 
7 
3 
5 
5 11 
12 
3 
4 3 
3 4 
3 4 
4 5 
6 7 
3 4 
6 7 
4 1 
4 5 
4 5 
5 6 
3 2 
6 5 
5 4 
4 9 
4 5 
4 5 
12 
16 
1 8  
3 
3 
6 
6 
5 
4 
4 
6 
6 
7 
1 
4 
3 
6 
4 
6 
12 13 13 12 1 8  14 
11 17 12 15 2 0  16  
12 12  16  18  17  18  
16 11  17 12  1�  14 
18 13 16 1 9  17 
2 19 15 1 9  17 
5 5 17 2 0  19  
5 5 6 1 8  10 
4 4 3 5 18  
5 5 6 4 5 
5 5 6 8 5 8 
5 5 0 6 3  6 
5 5 2 6 3  6 
6 6 5 7 2 7 
4 4 5 7 4 4 
7 7 7 10 7 7 
2 4 5 7 6 7 
7 7 8 4 7  6 
5 5 4 4 3 2 
5 5 4  6 1 6 
16  
15  
11 
23  
15  
19  
19  
21  
18  
2 1  
6 
6 
7 
3 
6 
5 
10 
6 
6 
15 17 17  20  
is  20  19  19  
19  18  17 18 
18  16  15  20  
16 21  2 0  1 9  
20  2 1  19  2 1  
5 1 8  2 1  25  
19  2 1  17 2 1  
2 1  20  15  2 1  
2 2  19  + 9  2 6  
2 2  1 7  20  17 
2 1  20  26  
2 19 25 
5 5 19  
5 5 6  
7 7 9 3 
5 5 8 4 
8 8 9 7  
4 4 5  3 
4 4 3  5 
16  
1 8  
1 8  
20  
18  
18  
17 
20 
15 
20 
2 1  
2 0  
1 6  
2 2  
25  
6 
10  
6 
6 
17 
19 
17 
16 
19  
9 
17 
2 1  
22 
23 
24  
20  
27 
24 
25 
24  
9 
7 
7 
17 
20 
19 
16 
20 
17  
22  
8 
2 1  
1 8  
2 0  
2 3  
1 9  
1 8  
22  
22  
25  
6 
8 
2 3  22 Sig 
19  26 Lgr 
1 8  20 Lno 
2 6  20 LIe 
20 25 Lsu 
2 4  2 1  Lnn 
22  26  Lmi 
2 4  20 Lag 
2 0  2 1  Cae 
22  2 1  Lac 
17 23 Lin 
2 7  29 Lsm 
27 21 Lze 
27 21 Lfa 
2 5  24  Lma 
2 1  24  Lte 
25 24  Lnp 
2 6  2 1  Lfr 
23 Lot 
4 Lmo 
Big Lgr Lno LIe Lsu LnD Lmi Lag Cae Lac Lin Lsm Lze Lfa Lma Lte Lnp Lfr Lot Lmo 
14 . 3  
1 6 . 5  
1 6 . 3  
1 7 . 1  
1 7 . 2  
1 7 . 4  
1 7 . 8  
1 7 . 1  
1 9 . 1  
1 8 . 6  
1 8 . 9  
1 9 . 9  
2 0 . 2  
1 9 . 4  
22 . 0  
1 9 . 7 
21 . 0 
1 9 . 7 
23 . 0  
22 . 7  
MeanTv4 . 7  4 . 6  4 . 5  4 . 8  4 . 1  5 . 6  4 . 9  4 . 2  5 . 5  4 . 4  
4 . 3  4 . 5  4 . 6  5 . 9  5 . 6 4 . 8  6 . 0 6 . 7 7 . 6  5 . 0  
111 
a. 
-
II 
o 
111 b. 
" , 
0 
Fig. 4.1. Frequency distributions for pairwise comparisons of nucleotide substitutions in the 20 skink 
12S rRNA sequences. a. Tolal numbers of differences between taxa h. Numbers of transitions. c. 
Numbers of transversions. 
Total Differences 
. -
,I , 
.' 
: 
.' � 
, � , 
Rnnn I� 
/ :1 , � 
n n � ' .: r'nn 
1 " I • II , ' 7 • • II' 11 U 11 U 111 I' U H H. u a au • •  n . a M U a � "  
Transitions 
0 1 ' S ' II ' 7 " " l1 U H U U H � H H . � a a u . u n  • •  
No. 'lnIIIltbII 
H,-----------------------------------------------------------------------�
c. Transversions 
Otapter 4, page 43. 
Table 4.2. Mean numbers (± standard deviation) of total observed differences, transitions, transversions, 
and transition/transversion ratios for each skink taxon. The five most variable taxa for each column 
(assessed as the largest standard deviations relative to their mearts) are shown ill bold. 
Taxon Total 
StIs.Green 20.0 ± 4.6 
L.grande 20.8 ± 5.1 
L.notosaurus 20.9 ± 4.3 
L.linlchl 21.6 ± 3.8 
L.suteri 21.7 ± 4.2 
L.n.nigriplantare 22.0 ± 4.4 
L.microlepis 22.6 ± S.8 
La.guichenoti 23.0 ± S.3 
C.aenea 23.2 ± 2.3 
L.acrinasum 24.2 ± 4.2 
L.inconspicuum 24.5 ± 4.7 
L.smithi 24.7 ± 6.1 
L.ze laruJicum 25. 1 ± 3.2 
L.fallai 25.4 ± 3.6 
L.maccanni 26.2 ± 3.7 
L.telfairi 26.4 ± 3.1  
L.n.polychroma 26.5 ± 5.5 
L.infrapunctatum 27.3 ± 4.8 
L.otagense 27.4 ± 3.3 
L.moco 27.7 ± 2.8 
-
Transitioils Transversions 
�S.3 ± 4.4 4.7 ± 1 .2 
16.5 ± 4.5 4.3 ± 1 .4 
16.3 ± 3.0 4.6 ± 1.8 
1 7. 1  ± 3.8 4.5 ± 1 .5 
17.2 ± 3.6 4.5 ± 1 .4 
17.4 ± 3.9 4.6 ± i.3 
17.8 ± 4.9 4.8 ± 1 .8 
17.� ± 4.3 5.9 ± 1 .6 
19.1 ± 2.0 4.i ± 1.6 
1 8.6 ± 3.7 5.6 ± 1 .4 
18.9 ± 3.3 S.6 ± 2.0 
19.9 ± 5.2 4.8 ± 1.8 
20.2 ± 3.2 4.9 ± 1 .5 
19.4 ± 2.9 6.0 ± 1 .8 
22.0 ± :tO 4.2 ± 1.5 
19.7 ± 2.9 6.7 ± 1.7 
21.0 ± 4.4 5.5 ± 1 .7 
19.7 ± 3.9 7.6 ± 1 .5 
23.0 ± 3 . 1  4.4 ± 1 .3 
22.7 ± 2.5 5.0 ± 1 .6 
TslTv 
3 .4 ± 1 .3 
4 . 1  ± 1 .4 
4.2 ± 2.2 
4.7 ± 4.0 � - --
4.2 ± 2.0 
4.0 ± 1 .5 
4.0 ± 1.8 
3 .0 ± 0.9 
S.7 ± 4.0 
3.7 ± 2.0 
3.9 ± 2.0 
4.5 ± 2.1 
4.7 ± 2.2 
3 .6 ± 1 .5 
6.3 ± 3.7 
3.2 ± 1 .5 
4 . 1  ± 1 .7 
2.6 ± 0.6 
5.7 ± 2.0 
S.6 ± 4.0 
Cltapter 4, page 44. 
Table 4.3. Variability (mean ± standard deviation) in tile tot.a1 number of observed nucleotide 
substitutions and numbers of transitions, and the transitioIJ./t.msversion (fstrv) ratio in relation to the 
I 
numbers of transversions in the skink data set There is tio significant incr� in the number of 
transitions as the number of transversions increases (F9 368= 1.81, P = 0.065). Th� correlatiol} 
coefficients between numbers of transversions and res�Jvely, the total number of differences and the 
numbers of transitions are also shown. 
No. 
No.Tv Obs. Total 
I 6 18.3 ± 4.9 
2 14 20.7 ± 2.2 
3 40 20.5 ± 5 . 1  
4 76 22.8 ± 5.7 
5 94 24.2 ± 3.7 
6 70 25. 1  ± 2.9 
7 48 26.1 ± 2.7 
8 16 29.4 ± 1 .5 
9 8 30.0 ± 2.9 
1 0  6 30.7 ± 1 .0 
All 24. 1  ± 4.9 
0.26 
Transitions 
17.3 ± 5.0 
18.7 ± 2.2 
17.5 ± 5 . 1  
1 8.8 ± 5.7 
19.2 ± 3.7 
19.i ± 2.9 
19.1 ± 2.7 
2 1 .4 ± 1 .5 
21.0 ± 2.9 
20.7 ± 1 .0 
1 8.9 ± 4.2 
I 
0,�3 
Tslfv 
1 7.3 ± 4.9 
9.4 ± 1 . 1  
5 .8 ± 1 .7 
4.7 ± 1 .4 
3.8 ± 0.7 
3.2 ± 0.5 
2.7 ± 0.4 
2.7 ± 0.2 
2.3 ± 0.3 
2.1 ± 0.1 
4.3 ± 2.5 
. .  --
Fig. 4.2. Numbers of transversions in pairwise comparisons of the skink 125 rRNA sequences. ploued 
against numbers of transitions. Note the variability iri numbers of transitions and trans versions. The 
correlation coefficient. ?-. is 0.03. 
11�-----------------------------------------------------------------------'
10-
• III III 
• 
'I III III • • III • • w. � w. � 
I :  '" . . '" . '" . . � . III '" • w. '" '" • • • • • III '" • III • 
w. � w. '" • III W. • III '" III • • '" � '" III '" • '" III 
HI '" '" '" '" 
I . '" ?fI .. 
1 
04---------�--------�--------�----------r_--------�--------_r--------_i 
• 10 1� 10 SO � 5 
S5 
so 
� 25  
� 
� 20 
� 
� 11\ 
10 
5" 
No. T.rauIitbJB 
Fig. 4.3. Total number of differences in pairwise comparisons of the skink 125 rRNA sequences plotted 
against numbers of transversions. There is only a weak correlation between the number of transversions 
and the total number of observed nucleotide substitutions (?- = 0.26). 
• 
'" 
• • • '" '" 
• III '" • '" 
• • • '" • • 
• • ill '" • 
• • • • '" • 
• '" • • • • 
• '" • • '" 
'" '" '" • '" • • .- • • 
'" • • • • • '" • • • • • 
• • • • • '" 
• '" • • 
• • '" 
101 • '" 
• '" '" 
'" 
• '" 
'" '" 
• 
'" '" 
'" 
0;
0
-----.-----,2r-----3
r-----.-----5.-----ar-----r7
-----r----�----�
10
----�
11 
No. 'I'nmsversions 
20 
-�. 
18 
UI 
a 
I" 10 
d 8 Z . 
8 
Fig. 4.4. Relationship between numbers of transitions and numbers of transversions for a. ralites (12 
taxa; ? = 0.40) and b. pecoran bovids (15 taxa; ? = 0.55). Compare to figure 4.2. 
• 
• • 
• • 
• 
• • • • 
• • 
• 
4 • 4 4 
• 
• 
• • 
• 
• • • • 
• • A 
• • •  
• • 
6o .  • • • •  
4 • 
• 
• A 4 
I A Ratit81 
100 
80-
80-
70-
! 
j 
80-
50-
1 
114 
41 
30-
20-
10-
0 
Fig. 4.5. Proportion of transition substitutions in pairwise comparisons of the skink 12S rRNA 
sequences. Note that total sequence divergence is less than 10%, and that in this data set there is no trend 
for the proportion of transitions to decrease as the degree of divergence increases (,.'2 = 0.(01). 
* .,. .. ,; 'l'raDsitionI 
iii II! . .  
--
II! II! iii .. !II w- .. iii iii iii !II • iii • • II! • iii II! II! !II II( • II( II( • iii 
II! iii • iii iii lI! .. iii • • • iii .. !II lI! .. .. II! II! ... .. • .. .. II! iii .. !II • II! • • iii JI! ;II II! ;Ii IIi it; lIC • • .. w !'< IIi • • M! iii ;Ii !4' II! 11( iii II! II! II! iii iii 
II! 
10 15 20 25 30 sG 
Total No. Differences 
Otapter 4. page 45. 
The expected Ts/Tv ratio if the sequences were saturated can be calculated from equation 1 1  in 
Holmquist (1983): 
(Aoo + Goo) (Coo + Uoo) 
(AooGoo) (CooUoo) 
where' �Goo etc, are the equilibrium base frequencies. Using the frequencies from table 2.2, the 
saturated Ts/Tv should be 0.47, well below the smallest observed ratio of 2.6 for L. injrapunctatum 
(Table 4.2). 
Although there is no strong trend for the number of t:raI1sversions to increase as more transitions occur 
(Fig. 4.2), the Ts/Tv ratio does decline as the number of transversions increases (Fig. 4.6). More 
importantly, the variation associated with Tsffv ratios also �eclines as the number of transversions 
increases (Fig. 4.6). This trend is also seen in other data sefS (Fig. 4.7) and suggests that this asymptotic 
decline in Ts/Tv ratio is due to greater variability in the proportion of transition substitutions when only 
a few transversions occur, rather than being indicative bf saturation of substitutions. 
Simulations 
I 
To examine the extent to which the variability in the skink data set is due to stochastic processes 
computer simulations were performed to randomly generate sequences. Parameters for the simulations 
are shown in table 4.4, the key variable being the Ts/Tv ratio. Values for sequence length (384) and 
number of expected changes (25) were chosen to correspond with the skink data. To generate a similar 
size transition and transversion matrix as the skinks (Table 4.ib), 200 iterations (runs) of the program 
were performed (Table 4.4). Longer iterations of 100,000 gave similar results (Table 4.4). 
The program creates a random sequence of the specifieq length. A new sequence is then derived from 
this with sites of substitutions chosen randomly. All sites in the sequence are free to vary. The 
probability of each change being a transition is determined �y the Ts/Tv ratio read in at the start of the 
program. 
The average number of substitutions between pairs of skink sequences is 25, and the Ts/Tv ratio about 
four (Table 4.2). In the simulation, when the expected number of changes is set to 25 and the Ts/Tv ratio 
is 4:1 then both the frequency distributions (Fig. 4.8) and numbers (Fig. 4.9) oftraDsitions and 
transversions are similar to the skink data (Figs. 4.1 & 4.2). The same pattern of variability in numbers 
of transitions and transversion is also seen when different numbers of substitutions are allowed (Fig. 
4.10), and when the Tsffv ratio is altered (Fig. 4.1 1).  The apparent skewness shown in the transition and 
transversion frequency distribution (Fig. 4.8, see also the skink distributions in Fig. 4.1) is a reflection of 
sampling error since this disappears when a larger number of simulation runs are performed (Fig. 4.12). 
The pattern of variability in numbers of transitions and transversions (Figs. 4.9, 4.10) is not due to 
sampling error however, since increasing the number of runs does not alter the general pattern (Fig. 
4.13). This point will be returned to shortly when other data sets are considered. 
Fig. 4.6. The relationship between number of transversions and the tran,;>itibn!transversion (Ts/Tv) ratio 
for the skink data set. Note the decreasing means and variability of the'Ts/Tv ratios 'as the number of 
trans versions increases. (See also Fig. 4.7). 
30,---------------------------------------------------------------------------------,
25-
• 
20- • -- .-
0 
� � 15-
� 
. " . .  � 
• 
• 
10- • 
• 
.. 
• • 
I 
it( 
5- I I • I I • I • • • 
0 
0 i 2 3 • 5 8 7 B g 10 11 
No. Transversions 
Fig. 4.7. The relationship belween number of transversions and the uansition/uansversion (fsITv) ratio 
for other dala selS. a. raliles (12 taxa), b. bovids (15 taxa), and c. simulaled sequences (20 taxa; see Figs. 
4.8 & 4.9, and the texl). All the dala sets show the same pallern of decline in mean and variabililY in the 
Tsffv ratio as the number of transversions increases. 
25 
! 
a. 
I .A. RaUt.e8 I II 20 
$ 
15 4 
4 
4 
10 4 
4 4 4 
5 • 
t • • 4 a • ,. 4. ,. 4 4. ,. ,. ,. 4 
0 
0 1 2 3 • 5 8 7 8 g 10 11 12  13 a 15 18  17 18 Ig 20 
No. 'l'rIDIneraioDI 
25 
! 
b. 
X BovidB 
II 20 
It 
:It 
$ % x 
15 
3E 
10 � � )( 
� ! I i 5 � I :lC: � � X * X 7. X X X A -.-
0 
0 1 2 3 • 5 6 7 8 9 10 . 1 1  12  13 a 15 16 17 18 19 20 
No. 'lnnsverBioDs 
40 
C. >< Simulations 
iii 
35 " 
" 
" 
" 
30 iii 
.. 
Ie 
Ie 
iii 
2S " 
iii 
0 . ., 
� 20  
� 
M 
" 
15 .. 
.. 
10 I " • I I I I 5 " • I I .. I I I i I • .. • .. 0 
18 19 20 
OJapler 4. page 46. 
Table 4.4. Parameters used in simulations of transitions and transversions. Values were assigned to be 
similar to the skink data; sequence length was 384 bases and the ntimber of changes expected to occur I 
was assigned a value of 25. The assigned transition/transversion (tsITransversion) value is given on the 
left and the mean ratio obtained from the simulations is shown in the last column:-
TslTv No. Transitions TransversionS Mean 
Ratio Runs (mean + Std) (mean + Std)' TslTv 
2 2 0 0  1 6 . 6  ± 3 . 6  9 . 0  ± 3 . 0 1 . 9  
1 0 0 , 0 0 0  1 6 . 7  ± 4 . 0  8 . 3  ± 2 . 9  2 . 0  
3 2 0 0  1 9 . 0  ± 3 . 8 6 . 6  ± 2 . 5 2 . 9  
10 0 , 0 0 0  1 8 . 7  ± 4 . 2  6 . 2  ± 2 . 5 3 . 0  
-
4 2 0 0  2 0 . 1  ± 4 . 6  5 . 0  ± 2 . 1  4 . 0  
1 0 0 , 0 0 0  2 0 . 0  ± 4 . 4  5 . 0  ± 2 . 2 4 . 0  
5 2 0 0  2 0 . 7  ± 4 . 2  4 . 3  ± 1 . 8  4 . 8  
1 0 0 , 0 0 0  2 0 . 8  ± 4 . 4  4 . 2  ± 2 . 0  5 . 0  
6 2 0 0  2 1 . 2  ± 4 . 5 3 . 6  ± 2 . 0 5 . 9  
1 0 0 , 0 0 0  2 1 . 4  ± 4 . 5  3 . 6  ± 1 . 9 5 . 9  
7 2 0 0  2 2 . 2  ± 4 . 5  3 . 3 ± 1 . 8  6 . 7 
1 0 0 , 0 0 0  2 1 . 9  ± 4 . 6  3 . 1  ± 1 . 8  7 . 1  
8 2 0 0  2 1 . 5  ± 4 . 4  2 . 8  ± 1 . 7 7 . 7 
1 0 0 , 0 0 0  2 2 . 2  ± 4 . 6  2 . 8  ± 1 . 7 7 . 9 
9 2 0 0  2 2 . 7  ± 4 . 8  2 . 6  ± 1 . 6  0 . 7 
1 0 0 , 0 0 0  2 2 . 5  ± 4 . 6  2 . 5  ± 1 . 6 9 . 0  
10 2 0 0  2 4 . 0  ± 4 . 9  2 . 5  ± 1 . 4  9 . 6  
1 0 0 , 0 0 0  2 2 . 7  ± 4 . 6  2 . 3  ± 1 . 5  9 . 9  
3. 
20 1 -
,-� 16 
-L -
-
o 
Fig. 4.8. Frequency distribution� of transitions (a) and transversions (b) in simulated sequence data. 
Sequences are 384 bases in length. Each sequence has on average 25 substitutions relative to other 
sequences, and a transition/transversion ratio of 4/1 . Two hundred simulation runs were performed. 
Compare these distributions with those of the skink data set in figure 4.1 b & c. 
- -
-
.. 
r:" 
1/ 
I:" 
-
I "  1 /  
I "  [/  :.- I >  , .., [/ 1 /  I." , , / 
n n n R  
> 
n n  · n n T I I I I I I I I .- .- � T :T I, T .- 1 T o 1 2 S • 6 I 7 8 • 10 11 12 13 1. 16 11 17 18 18 20 21 22 2S U 26 28 27 28 2t 30 31 32 3S 
No. 'lraDsltioDII 
26.---------------------------------------------------------------. 
b. 
20 
I � 16 
1 -i L 
I 
o 
Fig. 4.10. Comparison of the relationship between the numbers of transitions and numbers of 
transversions for other simulations using, a.40 expected substitutions and, b.SS expected substitutions 
between each sequence. Sequence lengths were 384 bases, and the transition/transversion ratio was 4/1. 
Correlation coefficients are, respectively, 0.00 and 0.03. Compare to figure 4.9, noting that the 
relationship between numbers of transitions and numbers of transversions does not change as more 
substitutions occur. This differs from the ratite and bovid data sets where there is a trend toward a linear 
relationship as more substitutions occur (Fig. 4.4). 
u�--------------------------�--------------�========� 
- I + '" c:baD&m 3. 
20 
• 
+ 
+ ++  
+ 
• 
++  + + ... 
... ... ... + 
+ ++  + + + .++  . 
.+  + + ++++++ + + +  
+ + ++++  + + 
++++ +++++++  ++++  + 
+++++++++++++++  + 
+ 
• 
+ 
• 
+ 
++++++ .+++++ ++  ++  + 
+ + .+ + +  .++  + ++  .+  • 
+ + + ++ + +  
+ • + .  
+ 
+ 
+ 
O;---------,-----�--._--------._--------._�------_.--------_r--------� 
10 20 so 40 70 
No. Tnnaiticm 
u,-----------------------------------------------------------------------�
b. 
20 
Iii 
� 10 
Ii 
!!IF;I(  
lIE 
;Ii It! lIE 
III 
!!! !!! lIE III .111*-* 
** *-••  III 
III lII. **  .'" 1It *  III 
!!!!!! "' !I!  !!! !I! 1II.IIt;lilt! III. 
III 
iii * 
!l!lt!1II III III!!! !!!...  III III lIE lIE 
III!!!!!! !!!!!! lIE !!! IIIIII !I! "'lIE  1lE1Ii!!! 
!!! !!! _ !I! !!! _ . _IIi!!! III lIE. 
* .1101( )If iii 
O;---------,----------r---------r---------r--------_.--------�--------� 
10 20 so 40 80 70 
No. Tnnwlticg 
� 
� 
Fig. 4.11. The relationship between the numbers of transitions and numbers of transversions for other 
simulations using, a. a Tsrrv ratio of 2/1 and, h. a Tsffv ratio of 8/1 .  Sequence lengths were 384 bases, 
and the expected number of substitutions in each sequence was 25. Correlation coefficients are, 
respectively, 0.024 and 0.002. Compare to figures 4.9 and 4.10, noting the similar patterns of 
relationships. 
20�-----------------------------------------------------------------------------'
3. x �/lT= 2 
1& x ::;:: x 
x x x x � 
x x x x x x x x 
x x x x x x x x x 
x x x x x x x x x x 
10 x x x x x x x x x x x x x 
x x x x x :It X X X X X X 
X X X X X X X X X X X :It 
--
X X X X X X X X X X X X X 
X X X X X X X X 
X X X X X X X X X X 
X X X X X X X 
X X 
04------------------.------------------.------------------r----------------� 
10 20 30 
No. Tnnsitiru 
2O�----------------------------------------------------------------------------_, 
b. !8l �/lT = 8 
1& 
10 
l1li 
III III l1li l1li III 
iii l1li III III III III III IIi1 III l1li IIi1 • III 
III III III III IQ III lIS • • • III iii 112 iii iii Ie 
• III III • • III III • • • • • " • • III a iii 
1lI l1li l1li • • III l1li " • • • l1li III • • III • III III III 
l1li • IIi1 IIIl III IZ lIS 
0 
10 20 31L 
No. 'rlmdtIaDI 
Fig. 4.12. Frequency distributions of transitions and transversions in simulated sequence data for a large 
number of iterations. In figure 4.8 200 runs were performed, while here the number of runs is 100,000. 
Increasing the number of simulation runs gives smoother frequency distributions, indicating that 
sampling error may account for the skewness in figures 4. 1 and 4.8. 
�.---------------------------------------------------------------�
3. 1 0 m  100, 000 l'UDII 
20-
� 15--
, '[> ., 
5-
��----------------------------------�========� b. I IZZ] TV, 100, 000 nms 
20 
20 
11 
cpo � 
Fig. 4.13. The relationship between numbers of transitions and numbers of transversions in the 
simulated sequence data set when 500 iterations are performed (larger runs exceeded the capacity of the 
plotter). Sequences are 384 bases in length, and each sequence has on average 25 substitutions relative 
to other sequences, and a transition/transversion ratio of 4/1 .  The correlation coefficient is 0.00, and this 
pattern of relationships is almost identical to that in figure 4.9 where 200 runs were performed. The lack 
of a strong correlation between numbers of transversions and numbers of transitions is, consequently, 
not due to insufficient numbers of comparisons (compare with the ratite and bovid plots in Fig. 4.4). 
x 
:.1)(",)( )(.-<  M 
)( M)(X )(XM X M 
x M 
)()()(M)()(M M )(XM x 
M�)(MXMXXX)(MXMXX M 
M YXXXX)(X�X)(XM"'MM MM 
)( MMMMMMX)(X)(MM"MXkM 
MX)(XM"'XMXXMMM MXM 
x MMMX)(X)()()(XXMMX MX 
X)()(M )(M)(XM)()(M�M)( M 
M M>t  )C)( xM)c x 
X 500 nmI 
Ol;----------r------�_r------�_r--------_r--------_.--------_.--------� 
10 20 so 40 70 
No. 'lDnait:i! OJ 
Olapter 4, page 47. 
Estimating Divergence from Transversions 
Due to the possibility of saturation of transitions obscuring total sequence divergence, numbers of 
transversions have been used to estimate the degree of divergence between taxa (Brown et al. 1982, 
Wilson et al. 1985, Miyamoto & Boyle 1989, Milinkovitch et al. 1993). As noted above however , the 
numbers of transversions in pair-wise comparisons can be very variable (Figs. 4.2, 4 .4 ,  4 .9 ,  4.10, 4.13). 
There will, therefore, be a higher variance associated with estimates of the degree of divergence based 
on transversions than when all substitutions are used . This is ill ustrated with the slDlUlation data, where 
the true values for total numbers of differences are known. Multiplying the number of  observed 
transversions by [1 + {TslTv ratio} ]  gives an estimate fo� total numbers of differences. For example, if 
the number of differences is set to be 25 and the TslTv ratio is 411, then on average there should be five 
transversions between each sequence , and 5x[1 + {4 } ]  will g ive 25 total s ubstitutions. The s tandard 
deviation for this estimate of mean numbers of s ubstitutions is however twice as large as that obtained 
from using the total numbers of observed differences , as shown in figure 4.14. This is true when both the 
numbers of differences (Fig 4.15) and the TslTv ratio (Fig. 4.16) are varied. Estimates of the mean 
divergence between taxa are less accurate however when the TslTv ratio is varied (Fig. 4.16) 
Comparisons to Other Data Sets 
The skink taxa are quite dist inct from each other, but does this imply recent and rapid divergence or 
slower change over a longer period of time? Lack of a foss il record makes it difficult to detemtine what 
period of time the observed sequence differences among the skinks reflect Variation in evolutionary 
rates between taxonomic groups makes estimates based on a general mtDNA evolutionary rate (Wilson 
et al. 1985, 1987) difficult This is illustrated in tables 4.5 and 4.6. Ratite and bovid data sets (Table 4.5a 
& 4.5b, respectively) show generally similar numbers of transitions and trans versions for the same 
region of 12S rRNA, despite having different times of separation (approximately 25 and 80 million 
years , respectively; Table 4.5). 
Xantus iid lizards from the Caribbean and Central America are much more divergent from each other 
(Table 4.5c), their sequences have many insertion/deletion events (Appendix 2), and substitutions occur 
in more conserved parts of the molecule (Fig. 3.6b). They therefore either diverged a long time ago or 
their rate of sequence evolution is high. Based on fossil records, the genera in table 4.5c separated 
between 35 and 70 million years (the original references for these were not seen, but are discussed 
Hedges et al. 1991). The date of 70 million years is derived from Paleocene fossils of Paleoxantusia, 
which appear to share derived features with Lepidophyma and Xantusia rather than Cricosaura (see 
Hedges et al. 1991). Separation between Lepidophyma and Xantusia may have occured in the Eocene 
(35-40 MY A), based on comparisons of Paleoxantusia foss ils . The phylogenetic relationships of the 
Paleoxantusia are not well known however , and there are few Xantusia and Lepidophyma fossils (see 
Hedges et al. 1991), so estimates of their divergence times , and hence rates of sequence evolution , are 
possibly less reliable than those for the birds and mammals . 
35 
30 
25 
� -
I" 
15 
10 
Fig. 4.14. The means and standard deviations of estimates for the real number of substitutions. One 
estimate ("Obs.") is based on the total number of observed differences, while the second ("Pred.") uses 
and the number of observed transversions to determine the total number of substitutions. Simulated 
sequence data were used, and the mean number of substitutions is expected to be 25. The 
transition/transversion ratio was set to 411, so the expected numbers of .§ubstitutiolls was calculated from 
transversions by multiplying the number of transversions by 5 .  The same mean number of differences 
are obtained when either the total numbers of observed differences or the numbers of transversions are 
used, but there is a larger standard deviation associated with the estimate derived from trans versions. 
There will, therefore, be greater uncertainty in the accuracy of the estimate of total sequence divergence 
between taxa when only the numbers of transversions are used (see the text). 
.r.: 
I � Obi. (25) � Pred. (25) I 
Mean ± Sid. Obs. 
Pred. 
25.3 ± 4.9 
25.9 ± 1 1 .3 
'10 '15 80 85 80 85 100 105 110 
i{ -
40 
_ 3. 
so -
-
-
10 -
,-
o 
40 
Fig. 4.15. The means and standard deviations for estimates of the real number of substitutions based on 
the total number of observed differences ("Obs. to) and the number of transversions ("Pred. to). Simulated 
sequence data was used, and the transition/transversion ratio was set to 411. In a, the expected number of 
substitutions is 40, and in b it is 55. Calculations are as described in figure 4.14. The estimate based on 
the number of transversions has a higher standard deviation than that derived from the observed total 
numbers of substitutions. 
I tE?l Obi. (40) IE33 Pral. (40) 
Mean ± Std. Obs. 40.3 ± 
Pred. 39.4 ± 
- -
--
- - --
W"1 � � jj W"1 In 113 In I I I I I I I I I I o 6 a � "  � � � 40 � � � � � � � " e  � H m a6 1a 
Total No. Differences 
6.3 
13.7 
I 
h. 
I D Obi. (55) [X) Pral. (55) I 
Mean ± Std. 
I-
� I-
i-
,-
� 
� - - � 10 
> v � � x 
> 1\ v V >< > 
- > � 1\ � X > >< D< - � � > > > >< > D< >< � � Jl rv  > 1\ X � r. >< x x � > > > [X >< >< IX r. I I I I I I o o 6 a � � � � � 40 � � � "  7 7 
Total No. Dit!erences 
Obs. 54.6 
Pred. 55.1 
n I)l ,.., n I I 1 1 11 
± 6.5 
± 16.9 
Fig. 4.16. The means and standard deviations for estimates of the real number of substitutions based on 
the total number of observed differences ("Obs.") and the number of transversions ("Pred."). Simulated 
sequence data was used, and the expected number of substitutions is 25. In a. , the transition/transversion 
. ratio was set to 2/1, and in b. it is 8/1 .  Calculations are as described in figure 4.14, except that 
transversions'are now multiplied by 3 and 9 in a and b, respectively. Once again (compare with Figs. 
4.14 & 4.15) the estimate obtained from numbers of transversions has a higher standard deviation than 
that derived from observed total numbers of substitutions. 
�,-----------------------------------------------------------------------�
i{ -
30 
25 
10 
110 
75 
70 
55 1-
I-
i-
1-
50 
515 
150 
, -
1 -
, -j: 
1 -30 
215 
20 
, -
, -
-
10 -
o 
a. 
o 3 8 
b. 
. 
I o II 
1 _ Ob& ('lJ/lV=2) E'2I Pred. (Ts/lV=2) I 
Mean ± Std. Obs. 
Pred. 
25.5 ± 4.4 
26.4 ± 8.4 
II 12 16 111 21 2" 27 30 33 38 311 � � 51  
� � )� 
';1' 
� I 
111 
. Total No. OOferences 
T 
� � 
� 
. � 
'� 
� 
� � . ·0 �  
.>{J 
� 
�?' �, .�� ',(S:' 
27 35 
Total No. DitfereDceB 
[ � Ob& (Ts/l'v=8) !HE Pred. (Ts/l'v=8) I 
Mean ± Std. Obs� 25.2 ± 5.2 
Pred. 22.9 ± 13 .9 
EHB -I I 
83 
OJapter 4. page 48. 
Table 4.5. Distance matrices for 12S rRNA sequences from other vertebrates. The same region of the 
gene (that bounded by the PCR prin!ers 12SAR and 12SBR) is compared for all groups. Transitions are 
shown above the diagonal and transversions below. Full species names and GenBank accession numbers 
are given in Appendix 2. 
Table 4.5a. Ratites. Familial divergence times up to 80 million years ago (Sibley & Ahlquist 1981, 
Cooper et aI. 1992). 
Transversions ( below diagonal )  \ Transit ions ( above ) 
Ram Rda Ost Cas Emu BrK RoK LSK Meg Din Ano Pac 
Rhea amer .  ( Ram) 
DarwinsRhea (Rda) 0 
Ostrich ( Ost ) 7 
Cassowary 
Emu 
Brown Kiwi 
Roa Kiwi 
LSpot Kiwi 
Megalapt . 
Dinornis 
Allomalopt . 
Pacbyornis 
( Cas )  6 
(Emu )  6 
( BrK) 1 0  
( RoK) 9 
( LSK) 9 
(Meg ) 8 
( Din)  7 
( Ano ) 7 
( Pac ) 7 
1 0  29  3 4  
3 1  3 6  
7 2 4  
6 3 
6 3 2 
1 0  7 6 
9 6 5  
9 6 5  
8 9 8  
7 8 7  
7 8 7  
7 8 7  
3 6  3 0  
37  3 0  
2 4  3 0  
17 22 
27  
6 
5 1 
5 1 
8 12  
7 1 1  
7 1 1  
7 1 1  
3 0  3 1  
3 0  3 1  
2 6  27  
1 8  1 9  
22  2 3  
1 5  1 3  
2 
o 
1 1  1 1  
1 0  1 0  
1 0  1 0  
10  1 0  
1 9  
18  
23  
28  
27  
29  
29  
3 0  
1 
1 
1 
2 5  
2 4  
2 8  
2 8  
27  
3 1  
2 9  
3 0  
1 0  
o 
o 
27  2 6  
2 6  2 5  
2 3  2 2  
2 6  �-27 
2 9  28  
2 9  2 8  
2 9  2 8  
3 0  3 0  
1 1  8 
1 1  1 0  
3 
o 
Table 4.5b. Pecoran bovids. Most of these bovids diverged from each other 23-28 million years ago 
(Kraus & Miyamoto 1991, Allard et al. 1992). 
B . tauru 
Ca .birc 
M.kirki 
K. ellip 
G. tboms 
D. dorca 
Ce .maxw 
Bo . trag 
Aa .mala 
O . virgi 
Mu .raav 
H. in arm 
C. unico 
A .amari 
T.napu 
Transvers ions ( below diagonal )  \ Transit ions ( above ) 
Bta Cah Mki Mki Kel Gth 000 Cam Bot Aem ovi Mur Hin Aam Tna 
3 
3 
3 
5 
5 
2 
2 
4 
5 
5 
6 
5 
7 
13  
27  
2 
2 
4 
4 
1 
3 
3 
4 
4 
5 
4 
6 
1 0  
23  25  3 1  3 1  22 1 9  2 9  2 3  25  2 5  
1 4  2 2  2 5  19  1 8  2 5  2 0  2 3  2 1  24  
2 0  1 6  2 1  17  2 1  1 6  14  1 5  1 9  
2 22  2 3  2 1  23  3 0  2 6  2 5  2 5  
4 4 2 6  19  2 1  2 3  2 1  24  2 7  
4 4 6 2 1  2 9  2 6  27  28  2 8  
1 1 3 3 2 1  2 0  22  23  2 1  
3 3 5 5 2 2 3  2 2  2 2  2 1  
3 3 5 5 2 4 2 1  29  2 9  
2 4 4 6 3 5 5 18  1 3  
4 4 4 6 3 5 5 2 17  
5 5 5 7 - 4 6 6 3 3 
4 4 4 6 3 5 5 2 2 3 
6 6 8 8 5 5 7 8 6 9  
10  12 1 4  10  1 1  13  1 1  12  1 4  1 5  
24  26  3 3  
2 3  2 8  3 9  
1 8  2 4  3 4  
28  3 0  3 5  
2 9  27  40  
3 0  3 0  4 1  
2 2  2 5  3 9  
22  28  3 7  
2 5  25  4 4  
1 2  2 7  3 3  
1 6  3 0  3 5  
17  27 3 3  
27 3 5  
8 3 2  
1 4  1 6  
Olapter 4. page 49. 
Table 4.Sc. Xantusiid lizards. Divergence of Cricosaura and Xantusiid lineages approximately 70 
million years ago, Lepidophyma and Xantusiids 40-35 million years ago (see Hedges et al. 1991). 
Transversions (below diagonal )  \ Transitions ( above ) 
Cty Lea Xbo Xhe Xvi Xri 
c .  typica 4 8  4 4  42  - 4 0 40  
Le . smi thii  55  4 8  4 8  48  4 5  
X. bolsonae 57 3 8  3 0  2 9  28  
X. henshawi 57 42  1 8  2 2  23  
X.  vigilis  55  3 9  1 1  1 1  1 1  
X. rivers . 5 1  37  9 1 1  4 
Table 4.5d. Great Apes (Hixson & Brown 1986). Orangutan-Gorilla divergence estimated at 13-15 
million years (Miyamoto et aI. 1988), and Gorilla-Chimpanzee-Human at 4-7 million years 
(Miyamoto & Goodman 1990). 
Tv (below diagonal ) \ Ts ( above ) 
Ora Gor PCh CCh Hum 
Orangutan 19  2 4  25  2 1  
Gorilla 1 1 1  1 0  1 0  
PygmyChimp 0 1 3 9 
CommonChimp 0 1 0 1 1  
Human 0 1 0 0 
Olapter 4. page SO. 
Table 4.6. Comparison of estimated times of divergence among the skinks based on rate estimates for 
the same 12S rRNA fragment from other vertebrate taxa (see Table 4.5). Numbers of differences 
between selected taxa from tables 4.5a-4.5d are sbown and the approximate nucleotide cbanges per 
million years are calculated from suggested divergence times. These rates are then applied to numbers of 
differences observed in the skink data set. The large differences in rates between the different data sets 
illustrates that rates can vary between taxonomic groups. ,Estimates �f divergence times for the skinks 
based on any of these comparisons are therefore likely to be unreliable (see text). 
No.2 Est.3 
Taxal Difr. Time 
Ost-Rama 36 80 
Cri-Xrib 9 1  70 
Les-Xric 82 35 
Bta-Aamd 3 1  25 
Ora-Hume 21 14 
CCb-Humf 10 5 
lComparisons between: 
a Ostrich & Rhea americana (fable 4.Sa) 
c L smithii & X riversiana (fable 4.Sc) 
e Orang-ulan & Human (fable 4.Sd) 
2 Number of observed differences bewteen taxa. 
Estimated timeS 
Changes per4 if Differences of 
Million Years 25 30 
0.5 50 60 
1 .3 19 23 
2.3 1 1  13  
1 .2 21 25 
1 .5 17  20 
2.0 12.5 15  
b C. typica & X riversiana (fable 4.Sc) 
d Bas taurus &A. americana (fable 4.5b) 
f Common chimpanzee & Human (fable 4.Sd) 
3 Estimated times of divergence in millions of years (see Table 4.5). 
4 Number of substitutions/million years calculated from Number of Differences and Estimated Divergence Times. 
35 
70 
27 
15 
28 
23 
17.5 
S Times of divergence (in million years) based on rates in previous column and applied to a range of observed numbers of nucleotide 
differences in skink pair-wise comparisons (see Table 4.2). 
OJapter 4, page 51 .  
Approximate rates of evolution for this region of 12S rRNA are presented in Table 4.6, using sequence 
differences and times of separation from tables 4.5a-d. These values are then used to calculate times of 
separation for skink taxa when between 25 and 35 nucleotide substitutions occur (fable 4.6). The rates of 
change for the 12S rRNA sequence vary between the different groups, so it is difficult to determine times of 
separations for skinks based soley on these comparisons, though they could have diverged 70 million years 
or more ago (Table 4.6). 
While the xantusiid lizards are also small lizards, using their rates of 12S rRNA evolution (Table 4.6) as 
estimates for a skink rate is liable to be inaccurate. As noted above, times of separation of the xantusiids may 
be imprecise. In addition, the xantusiids have many differences between them (fable 4.5c), whereas the 
skinks have comparatively few (fable 4.1 b), so the error associated with extrapolating rates of sequence 
evolution from xantusiids to skinks is likely to be large. Other ways of estimating times of separation of the 
skinks will be discussed in Chapter Five. 
Changes in the Context of Secondary Structure 
One quarter of the sites in the skink sequence are variable (Fig. 4.17). For the sites which do change it is 
important to establish if there are patterns in the rate at which they accumulate mutations. Rapidly changing 
sites or regions are expected to be most informative for relatively closely related taxa, while greater 
resolution for more divergent taxa can be obtained from consideration of more conservative sites (Simon 
1991,  Dixon & Hillis 1992). 
In the skinks most of the transversions, and the most variable positions (in terms of number of taxa which 
differ), occur in the unpaired portions of the most variable regions (S26, S34, L38 · & S46; Fig. 4.17). Four 
transversions occur in helices, but none of them have a corresponding compensatory change in the other 
member of the base pair (Fig. 4.17). Eight sites had both ti'ansitions and trans versions, and more than half the 
taxa are variable at four of these sites (positions 101, 203, 204 & 206; Table 3.1 ,  see also Fig. 4.17), 
suggesting that these locations may be sites of multiple changes. The effect that these highly variable sites 
have on phylogenetic analyses is investigated in Chapter Five. 
Five pairs of taxa have fewer than 15  differences between them (L. microlepis+L. smithi, L. grande+St. Is. 
Green, L. inconspicuum+L. notosaurus, L. n. polychroma+L. n. nigriplantare, and L injrapunctatum+La. 
guichenoti; Table 4.1 a). There is a trend for these pairs to differ at the most variable regions (S26, L38 ' and 
S46; Fig. 4.18). However, L. n. polychroma and L. n. nigriplantare differ by 13  substitutions but some of 
these occur in the more conserved regions of the molecule (S2 ' for example; Fig. 4.18). In addition, some of 
the most variable sites, position 101 for example, accumulate differences only between more distantly related 
taxa (Fig. 4.19). Other regions also tend to accumulate changes as the number of differences between taxa 
increases (e.g., L30 ' ,  S34, SL40, and U5 ' ;  Fig. 4.19). 
Similar analyses of closely related ratites suggest that regions S26, L34 ' and SL40 are the first to change in 
these birds (Fig. 4.20a). These regions are also the most variable in comparisons between more distantly 
Fig. 4.17. Variable sites in the skink 12S rRNA domain m. Constant sites are shown as it, transitions as .. 
and transversions by " . Sites at which both transitions and transversions occur are indicated by • .  Base 
pairings are shown by - and I (and * for G*U pairs), while a colon (": ") denotes less certain bonds. Every 
tenth base is marked by ".". Regions of the molecule are identified as either paired (S) or unpaired (L) as 
described in Chapter Three. 
I I I I I I  
t::
:t:rtt • tr:i:�:;:;::;::l.uSr?> 1 2 SA< 
l� 
t�� .l r:: · · · 
�: ::�mI::'A T • ·?
:Ii: T 
* I I I I * � .  S 2 6  
::tr:�: :::::t�.u:\ , }H -fW ;.;.;. �{: 
-@ S 2 0 ' 
I I I * I I I I I I 
> 1 2 SB<tJI tt:fJ: t t :::n:r:: 
. . 
T--::r S 4 6 
�--:t: 
.�--� 
. - -� 
�--� 
::g � 
A ...... . : .... : ... :.: .. :::.• ?�� � �:::: ....•. 
84 3 
L43 '  
)t 
:tA A Am:: 
L45 '  A--A 
:t:--f:: 84 5 
�--:::::: 
:::::: � 
T T 
� � 
10 
II 
8 
7 
I 
� 
I 
� 
� 
1 
Fig. 4.18. Location of nucleotide substitutions in 12S rRNA sequences for pairs of taxa which have only 
a small number of differences between them. The sequence was divided into regions as described in 
Chapter Three. Most of the substitutions for these pairs of taxa occur in 526, L38 '  , and around helices 
543-546. The L. n. polychroma+L. n. nigriplantare pair are unusual in that differences between them 
can occur in some of the more conserved regions of the molecule (see Figs. 3.5 & 4.17). The pairs of 
taxa and numbers of observed substitutions between them are: 
Lmi vs Lsm 
Lgr vs SIG 
Lin vs Lno 
Lfr vs Lag 
Lnp vs Lnn 
= 
= 
= 
= 
= 
L. micro Ie pis & L. smithi (5 substitutions). 
L. grande & Stewart Island Green (8). 
L. inconspicuum & L. notosaurus (12). 
L. infrapunctatum & Lampropholis guichenoti (12). 
L. n. polychroma & L. n. nigriplantare (13). . 
r" - .. __ ..... - . .. .... --... . . . .  _ . .. -- -'" . .. . . . .. - - . . .. .  .' 1 l_lml wl.sm�Icrwsr;  OOUnwlDo  i i9±lltrWt.,  �lDpl'lllDn I 
C-_.� __ .. __ ___ . . _. _ ___ . __ _ .. ___ _  ... _ ._ ... _ _ _  .-J 
_ .  
Fig. 4.19. Comparison of regions of variability in the 12S rRNA sequences for pairs of taxa which have 
few ("close") or many ("far") substitutions between them. The data for taxa with few differences 
between them is the same as presented in figure 4.18, though the y axis now represents the proportion of 
sites in the molecule which are variable. Comparisons between pairs of taxa which have larger numbers 
of substitutions between them ("far") were done in the following way. From the "Close" data -set the 
locations of nucleotide substitutions between L. n. polychroma and four of the other taxa (L. microlepis 
[22 substitutions], L. grande [24], L. inconspicuum [29], and L. injrapunctatum [34]) were noted. The 
total numbers of differences in each region for these comparisons were then used to calculate the 
proportion of variable sites. Note that region S26 has a similar proportion of variable sites when both 
"close" and "far" comparisons are made. L38 ' however, is relatively more variable among taxa which 
have few numbers of differences between them, while S34 and L46 ' have higher proportions of 
substitutions when more distantly related taxa are compared. Different regions of the skink 12S rRNA 
molecule therefore accumulate substitutions at different rates. 
��------------------------------------------------------�------------�
20-4 
-a 
� . 1 101 
p. , 
5� : 
1 
} ! 
o . 
925' 920' 
928 sag 
Fig. 4.20a. Comparisons of regions of variability in the 12S rRNA sequence for closcly and more 
distantly related ratites. Calculations are described in figure 4.19. For closely related ratites the numbers 
of differences between the following pairs of taxa were counted; 
Rhea americana & Pterocnemia pennata (Darwin's rhea) [10 subsititutions] 
Apteryx haastii (Roa kiwi) & A. owenii (Little spotted kiwi) [2 substitutions] 
Megalapteryx didinus & Dinomis novaezealandiae [ 1 1  substitutions] 
Dinornis novaezealandiae & Anomalopteryx didifonnis [ 1 1  substitutions] 
Anomalopteryx didiformis & Pac/zyornis elep/zantopus [3 substitutions] 
For comparisons between more distantly related ratites the A. haastii sequence was compared to each of 
the four moa sequences (see Table 4.5a). 
As in the skink sequences (Fig. 4. 19), S26 is one of the most variable regions in both closely and more 
distantly related ratite taxa. SlAO however is more variable between more closely related ratites than it 
is in skinks. Note that the region L38 I has very few substitutions in the ratites, in contrast with its 
variability .in the skinks (Fig. 4.19). uT7----------------------------��====��====� a. � aa. n.ttt. � Elwlll  Vaa 
20 
b. 
1 
� � � � � � � m � � � � 
_ a _ � m � _ � � � � 
Recian 
Fig. 4.20b. Comparisons of regions of variability in the 12S rRNA sequence for the most closely related 
great apes; the two chimpanzee species (pygmy chimpanzee and common chimpanzee; 3 sequence 
differences) and the common chimpanzee and human ( 1 1  substitutions; see Table 4.5d). There appears 
to be a tendency for substitutions to occur in regions S34, SlAO, S45 and lA5 I flfsL 
OIapter 4, page 52. 
related ratites as well (Fig. 4.20a), so it may be difficult to differentially weight regions for taxonomic 
comparisons. Among the great apes, regions S26, S34, SIAO, S45 and 1A5 ' vary (Fig. 4.20b). Similar 
comparisons were not made in the bovid data set since taxa differ from each other by more than 15  
substibJtions (Table 4.5c). 
DISCUSSION 
The 12S rRNA sequences from the skinks appear unusual because of the weak correlation between numbers 
of transitions and numbers of transversions (Fig. 4.2). A bias in favour of transitions is observed however, as 
has been observed for other groups and sequences (Fitch 1967, Vogel 1972, Wilson et al. 1985, Jukes 1987, 
Moritz et al. 1987). The cause of this bias is not clear though. Differences in the extent of the transition bias 
occur in mtDNA (see for example Vawter & Brown 1986, Thomas & Beckenbach 1989, DeSalle et al. 
1987), and transition biases are found in nuclear sequences as well (e.g., Fitch 1967, Jukes 1987), so the 
absence of a mis-match repair system in mitochondria (Wilson et al. 1985, Schaaper & Dunn 1987) may in 
itself be insufficient to account for a biased rate of fIxation of transitions. 
The variability in the proportions of transitions and transversions does not seem to be due to sawration of 
transition mutations. A common feature of mtDNA sequence evolution is the decline in the proportion of 
transitions as time since separation increases (Wilson et al. 1985, Moritz et al. 1987, Miyamoto & Boyle 
1989). The skink taxa have relatively few differences between them (less than 10%), and forty two percent 
of the substitutions in the skink data set are singletons. The proportions of transitions also do not show a 
tendency to decline as more differences accumulate (Fig. 4.5), which suggests that the degree of divergence 
between the skink taxa has not been obscured by many unobserved changes. 
The relative decline in numbers of transitions in skinks is apparent however when the Ts/Tv ratio is plotted 
in relation to the number of transversions (Fig. 4.6). This is a common trend though (Fig. 4.7), and the fact 
that the skink Ts/Tv ratios (fable 4.2) were much larger than the asymptotic value of 0.47 indicates that 
saturation of fIxed mutations has not occurred. 
The variation in the numbers of transitions and transversions among the skinks (Fig. 4.2) may be attributed 
in part to the stochastic nature of the substibJtion process. There is variability in both the ratites and bovids 
when fewer than 30 transitions occur (Fig. 4.4). As more transversions accumulate variability decreases 
(Table 4.3, Figs. 4.4, 4.6, 4.7), which also suggests that the variability is related to having a relatively small 
number of changes. The observation that there may be a relatively high number of transversions in 
comparisons between taxa with few differences, but not when more divergent taxa are compared (see for 
instance Stewart Island Green; Table 4. 1 b) could be related to sampling error. 
The simulation studies also indicated that numbers of transversions can vary in relation to the number of 
transitions (Fig. 4.9). However, the observation that this variability did not decrease as more substitutions 
occurred (Fig. 4.10), or as the Ts/Tv ratio was varied (Fig. 4. 1 1), suggests that there is a difference between 
Olapter 4, page 53. 
the simulation process and that which generated, for instance, the ratite pattern of differences. The sequences 
produced for the simulations were generated by a "star tree", or "tumbleweed", model: 
where all sequences are derived from the same point source (the initial starting sequence). In contrast, 
phylogenies such as the ratites are usually formed by a more consecutive branching order as-shown here: 
/ / 
, , 
The patterns of substitution in the sldnks seem to fit the star tree model quite well. Many of the sldnk taxa 
have similar numbers of nucleotide substitutions between them, and the pattern of variation in numbers of 
transitions and transversions (Fig. 4.2) is also similar to the simulation results (Fig. 4.9). These observations 
suggest that the sldnks (including the Australian La. guichenoti and the Mauritian L. telfain1 diverged from 
each other too rapidly for many phylogenetically informative changes to occur in this portion of the 12S 
rRNA. The Pecoran bovids diverged relatively rapidly from each other, over a period of five million years 
(Kraus & Miyamoto 1991 ,  Allard et al. 1992). The plot of transitions and transversions for the sldnks is 
more similar to the simulated data than to the bovid pattern (Fig. 4.4b) however, indicating that the sldnks 
may have diverged relatively more rapidly than the bovids. 
Rates of mtDNA Evolution 
Reliable estimates of times of divergence of the skinks cannot be obtained by comparison to other data sets 
(Table 4.6). Restriction enzyme digestion studies of mtDNA suggested a general rate of evolution of 
between 1-2% sequence difference per million years for primates and several other mammalian groups 
(Wilson et al. 1985, 1987), but the applicability of this rate to both specific genes, such as the slowly 
evolving 1 2S rRNA, and to other groups has not been validated (Vawter & Brown 1986, Moritz et al. 1987, 
Thomas & Beckenbach 1989, Goddard et al. 1990). Rates of change for specific sequences are also lineage 
dependent (Britten 1986, Gillespie 1986, Goddard et al. 1990, Palmer 1990, Satta & Takahata 1990). The 
slow rate of change in the ratite sequences, in comparison with the bovids (Table 4.6), supports the view that 
rates of bird sequence evolution may be slower than those of mammals (Kessler & A vise 1985), though 
Shields & Wilson (1987) suggested that geese have a mtDNA evolutionary rate similar to mammals. 
Rates of evolution could be affected by efficiency of DNA repair (Britten 1986), cell division rates (Wu & Li 
1985), generation time (WU & Li 1985, Li et al. 1987), and metabolic rate (Martin et al. 1992, Palumbi & 
Martin submitted). Recent studies have suggested that poikilothermic vertebrates may have slower rates of 
mtDNA evolution than birds or mammals (Thomas & Beckenbach 1989, Avise et al. 1992, Martin et al. 
1992, Martin & Palumbi submitted). These studies examined aquatic groups only (salmon, turtles and 
sharks), and the applicability to other groups such as terrestrial amphibians and reptiles is less certain. Some 
newts and salamanders however appear to have slow rates of sequence evolution (Larson & Wilson 1984, 
Wallis & Arntzen 1989), but the sequence data from xantusiid lizards (Table 4.6) could indicate that 
poikilotherms can have rates of sequence evolution as high as some homeotherms. 
OJapter 4, page 54. 
Slower rates of evolution in poikilothetmS may be related to their lower basal metabolic rate. A high 
metabolic rate (as in many birds and mammals) may result in a faster turnover ofmtDNA, and consequently 
an increased chance of DNA replication errors (Martin & Palumbi submitted). High body temperatures may 
also lead to greater levels of oxidative damage in DNA (Shigenaga et al. 1989). Some lizards, and other 
reptiles, however may not have a single basal metabolic rate, (e.g., Werner & Whitaker 1978; see also 
Schmidt-Nielsen 1986). If the skinks do have a slower rate of mtDNA evolution than mammals then this 
implies that many of the skink taxa diverged from each other much earlier than th� separatiol.l of the 
orangutan from the other great apes, more than 15 million years ago (see Tables 4.5d, 4.6). This period of 
time contrasts sharply with the suggestion, based on morphoiogical comparisons, of 5 million years 
separation between New Zealand and Australian Leiolopisma (Hardy 1977). 
Allozyme data (Daugherty et al. 1990b, C.H. Daugherty & G.B. Patterson pers. comm.) and immunological 
studies (Baverstock & Donnellan 1990, Hutchinson et aI. 1990) also indicate much greater times of 
separation of skinks than morphological features imply. Skinks have no fossil record in New Zealand (Bull 
& Whitaker 1975) and are poorly represented in fossil faunas elsewhere as well (Carrol 1969, Molnar 1991). 
Rates of skink molecular evolution cannot therefore be directly calibrated with geological time. The 
dynamics of reptilian molecular evolution have not been investigated so it may be inappropriate to 
extrapolate rates from other groups, such as birds or mammals. Reptilian groups with more substantial fossil 
records, such as crocodilians and iguanids (Carroll 1969, Estes & Price 1973, Pregill I989), require DNA 
sequencing studies so that rates of molecular evolution in reptiles can be investigated in more detail. The use 
of other information to infer times of divergence for the skinks will be considered in the next chapter after 
phylogenetic analyses of the sequences. 
&itimating Degree of Divergence with Transversions 
The variability in the number of transversions implies that estimates of total sequence divergence based on 
numbers of trans versions (Brown et al. 1982, Wilson et al. 1985, Miyamoto & Boyle 1989, Milinkovitch et 
aI. 1993) may have a high level of uncertainty (Figs. 4.14. 4.15. & 4.16). More comprehensive simulation 
experiments and analyses than those presented here however are required to examine this problem in detail. 
Regions of Variability 
Two points emerge from comparison of changes between closely related taxa (or those with the fewest 
number of differences). The flrst is that certain regions do have a tendency to vary flrst (Figs. 4. 19, 4.20). 
Helix 26, part of domain II, is very variable within and between vertebrate groups (Table 3.3), but still 
retains its helical structure. Helix 34 in skinks is also variable, more so than some of the unpaired regions 
(Fig. 4.17).  These two helices are not primary points of contact for ribosomal proteins, nor do they appear to 
be sites of tRNA binding (Gutell et al. 1985, Noller et al. 1990). Helix 33 in contrast is well conserved (Fig. 
4.17, and see also Table 3.3), a fact also noted by Simon (1991) for cicadas. This helix appears to be a point 
of contact for ribosomal protein S 19 (see Stem et al. 1989, Ebresmann et al. 1990, Noller et al. 1990) and so 
the conserved nucleotides may be associated with this. 
OIapter 4. page 55. 
The second point is that different groups of vertebrates may have different centres of variability. L38 ' is 
very variable in skinks, xantusiids and bovids but not in ratites or the great apes (Table 3.3, Fig. 3.6). These 
latter two groups show more variability in SUO and S45, respectively (Figs. 3.6 & 4.20). The fact that 
different groups tend to vary in different regions suggests either that there may be little functional 
significance to these patterns of variability, or that functional constraints differ between groups. Differences 
in sites and frequencies of Variability between groups will be affected by how many taxa are examined for 
each group, and also by how closely related they are, so comparisons of more species and groups are 
required to determine if there are taxonomic patterns of regional variability. 
Skink taxa with larger numbers of differences between Qle� accumulate substitutions in regions S29, L30 ' , 
S34 and SUO, as well as S26 (Fig. 4.19). A similar pattern occurs among the ratites (Fig. 4.20). Half of the 
eight most variable sites in skinks occur in S26 and L38 ' (Fig. 4.17), but the variable site (position 101) in 
L30 ' tends to differ only between more distantly related taxa (for example between L. n. polychroma and L. 
infrapunctatum; Table 3.1). 
L. n. polychroma and L. n. nigriplantare differ from each other in some of the more conserved regions of the 
molecule (Fig. 4.18), despite the fact that they have relatively few differences between them (Table 4.1a). 
While changes in more conserved regions may imply a more distant relationship between these two taxa, this 
is not supported by morphological or genetic information (Daugberty et aI. 1990b), nor by analyses of the 
sequence data (see Chapter Five). Errors in PCR or sequencing could explain why these two taxa differ in 
more conserved regions, but this seems unlikely since at least two individuals were sequenced for both L. n. 
polychroma and L. n. nigriplanrare (Table 2.1) and several amplification and sequencing reactions were 
performed for eacb sequence. The locations of the differences between them may just reflect the fact that not 
all sites in "more conserved" regions are equally conserved, just as some sites in "more variable" regions 
(e.g., position 207 in L38 ' , Fig. 4.17) can be more conserved than their neighbours. 
The effects on phylogenetic analyses of both a transition substitution bias and different levels of variation 
bewteen sites in the sequence are investigated in the next chapter. 
page 56. 
Chapter Five: Signals in the DaM and Phylogenetic A nalyses 
Aligned macromolecular sequences contain sites of variability, the patterns of which (the "signals") can 
be evaluated by clustering or phylogenetic algorithms. Different phylogenetic analysis algorithms use 
differing amounts of this information and have different optimality criteria for evaluating it and 
constructing evolutionary trees (see Felsenstein 1988, Swofford & Olsen 1990, Nei 199 1 ,  and Penny et 
aI. 1992 for reviews). With biological data conflicting signals, which may arise from, for insrance, 
convergence, can prevent taxa being placed into mutually exclusive groups. Such signals can complicate 
analyses, but they may be reconciled. As an example, a part of the skink data set yields the following 
signals (some of which are underlined): 
L. notosaurus 
L. Unlchl. 
L. suteri 
L. n. nigriplantare 
aaau�uguaugua£ 
aag�uauauguau 
aag.!!'luuauauguau 
aa�cauuguaugua£ 
The signals in this region provide support for L. notosaurus grouping with all of the three other taxa L. 
notosaurus has the most signals in common with L. n. nigriplantare however, and in this example these 
two would have the strongest relationship. Quantifying the signals supporting different sets of 
relationships in this way indicates the support, and hence reliability, for phylogenetic relationships. 
The signals present in the skink 12S rRNA sequence are examined in this chapter using the Hadamard 
conjugation, a new method for spectral analysis (Hendy & Penny 1993). This method shows that with 
New Zealand skinks the data set contains many conflicting signals and a completely resolved phylogeny 
cannot be obtained with the available information. Furthermore, differential weighting of both 
trans versions and more variable sites in the molecule has little effect in clarifying the skink 
relationships. Rapid diversification of the skinks appears to be the cause of many of the conflicts in the 
data. Three hypotheses to account for the origins of New Zealand skinks are proposed. 
Spectral Analysis of DNA Sequences 
The Hadamard conjugation uses a discrete Fourier transform to adjust for unobserved nucleotide 
changes in DNA sequences (Hendy & Penny 1993). Application of this method gives a "spectrum" 
which provides a quantitative measure of conflicting signals in the data which can support different 
phylogenetic hypotheses. A unique property of the Hadamard conjugation is it's "invertibility" which 
allows investigation of the properties in the data with respect to their fit to explicit assumptions of the 
model. The model has three essential elements (penny et al. 1992, 1993), namely 
a tree, 
a mechanism for sequence change, and 
probabilities of changes (lengths) along branches of the tree. 
Otapter 5, page 57. 
Given the model it is possible with the Hadamard conjugation to predict the data that would petfectly fit 
the tree. The number of patterns in the observed data is compared to the number of patterns in the 
predicted data and statistically evaluated A statistically poor fit between the observed and predicted data 
can lead to an incompletely resolved set of relationships. That is, the model is a poor explanation of the 
observed data. Details of the Hadamard conjugation are sboWIi in Box 5 .1 .  
There are two principal programs for the Hadamard conjugation: 
PREPARE - which reads sequences in a variety of fOfJIlats (interleaved, block, Nexus), and can 
calculate base composition. and frequencies of patterns in the sequences. Output files can be 
prepared suitable for analysis by Hadtree, PAUP, and Phylip. 
HADTREE - reads the partition frequency data from Prepare. corrects for multiple changes (using 
Cavender's 1978 model), and searches for an optiIp.al tree using the closest tree criterion (Box 5.1;  
Hendy & Penny 1989). It also has the ability to use other optimaiity criteria, such as parsimony. 
Both four- and two-character state analyses can be done using Hadtree. The four-character state model 
uses Kimura's (1981) three parameter model of evolution (one dass of transition mutations, and two 
classes of transversions) to analyse signals on the basis of all four types of nucleotides. 
Quadripartitions of the data are created in this way, groupmg taxa whicb share patterns of A, C, G, or 
U, respectively at a site. Storage limitations currently restrict fOUf-character state analysis to 10 or fewer 
taxa Two parameter models divide the signals into bipartitions, segregating for example taxa having a 
purine (A or G) from those baving pyrimidines (C or 1) at a specific site, as shown in Box 5.2. Currently I 
up to 20 taxa can currently be analyzed using the two-chkcter state procedure. The order of taxa in the 
data file does not influence the results. 
In this chapter the skink data set of 20 taxa (Table 3.1) is analyzed using the two character state 
condition. The "sum of seven" option (Box 5.2) was selected to find all possible signals. Use of the 
"AT/Ge" and "AC/GT" options give similar spectra anq resul\S. Too few transversions occur to use the 
"R/Y" option. 
What is a Bipartition and How is it Represent�d? 
Taxa can be grouped into bipartitions with O. 1 . 2, . ..•  n taxa The number of occurrences for each 
bipartition are summed for the sequence and the frequency of the bipartition is calculated by dividing by 
the total sum of all bipartitions in the data set (inclpding the bipartition which contains all the taxa; i.e. 
constant sites). The frequencies of bipartitions are cailed signals and the collection of signals for a 
sequence is referred to as a spectrum (Hendy & Penny 1993). Indiv�dual signals in the spectrum may be 
referred to as a spectral signal in this thesis. Bipartitions with only one taxon in them correspond to 
pendant brancbes on a tree - the brancb leading only to that taxon. The other bipartitions (excluding the 
one with all taxa in �t) are potential internal branches (or edges) in the pbylogenetic trees. 
Olapter 5, page 58. 
Box 5.1. Interpretation of components of the Hadamard conjugation. The Hadamard 
conjugation operates on the signals in observed data and a model (tree + mechanism + 
probabilities of change). It's invertible properties allow the prediction ofpattem�in data tha.t 
would perfectly fit the evolutionary model. It also permits correction for unobserved changes 
in sequence data before such data is either viewed directly (as the 'Y spectruin) or best-fitted to 
a tree, so corrections are independent of the tree. Probabilities of observing changes along 
edges of a tree (P) are specified by the model. Corrections to p for unobserved changes at sites 
are then made to generate inferred partition values, 'Y, or expected edge lengths in the optimal 
tree. The Hadamard transform, [H}, converts these values to p, a vector which represents 
corrected path lengths (distances between taxa in the tree) for subsets of taxa in the tree. The 
correction for unobserved changes is then reversed to generate observed distances, T. The 
Hadamard transform is then reversed, [HJI t to produce St the vector containing expected 
frequencies of observed patterns in all possible trees. With real data the process is reversed by 
going from s to 'Y. The closest tree optimality criterion uses the 'Y values to select the best fit 
between the real data to a tree that has edge lengths p. 
Model I 
p (probabilities of change on edges) 
I (corrections for unobserved changes) 
'Y (corrected estimate of edge lengths) 
I [HJ 
P (corrected path lengths between all taxa) 
I (uncorrect for unobserved changes) 
T (uncorrected distances between all taxa) 
i [HJI 
J, 
S (uncorrected estimate of edge 
i lengths in all possible trees). 
i 
Sequence Data 
Olapter 5, page 59. 
Box 5.2. Options for two-character state analysis of sequence data using the Hadamard 
conjugation. Comparisons between taxa are performed for each site in the sequence. The tilde 
(-) corresponds to the mathematical expression "not". 
1 .  AI-A Group taxa with adenine at a site and 
group those which have a differeht 
nucleotide. . . . 
2. G/- G Group taxa with guan�ne at a site. 
3. C/- C Group taxa with cytosine at a site. 
4. T/- T Group taxa with thymine (or uracil). 
5. AC/GT Group taxa with adenine or cytosine at a 
site and group taxa with guanine and 
thymine (or uracil). 
6. AT/GC Group taxa with adenine or thymine. 
7. RIY Compare on the basis of transversions 
only (AG/Cn. 
8. All 7 Performs each of the above analyses 
indi vidually. 
9. Sum of7 Performs the fIrst 7 comparisons and 
sums the frequencies pf the signals. 
ClJapter 5, page 60. 
These internal edge bipartition signals are ranked in order of descending frequency. The "strongest 
signal" has the highest frequency, and means that a relativeiy large number of sites in the sequence 
support the grouping of the taxa in this bipartition. If thi� sighal is in the tree, its branch would have the 
largest internal length. Conflicting signals however may mean that the strongest signal is not included in 
a resolved tree. This is illustrated later in the Chapter. 
Each bipartition can be uniquely identified by giving each taxbn a binary number and assigning the 
bipartition the sum of the taxa it contains. For example, with five taxa the binary numbers 1 ,  2, 4, 8, and 
15 describe each of the five taxa respectively (the last taxon's number is one less than its real binary 
number so that all bipartition numbers are unique). The bipartition corresponding to a grouping of the 
first two taxa can be described as either (1+2) = 3, or as (4+�+15) = 27. By convention the smaller of the 
two numbers is used to label the bipartition (PeIUlY et al. 1993). A second example illustrates why the 
last taxon is identified by a number one less than its true binary number. The bipartition of the first four 
taxa corresponds to (1+2+4+8) = 15, which is the same as sayihg that the fifth taxon is in a bipartition of 
its own. 
Lack of Resolution in the Skink 12S rRNA Data Set 
Analysis of the 20 skink taxa does not yield a completely resolved tree. A graphical representation of the 
spectrum (Fig. 5 .1)  shows that there is much conflict in the data, most of the bipartition frequencies are 
the same. There are only four relatively distinct and strong signals (Fig. 5.1).  However, for 20 taxa there 
are {219 - 21 } or 524,267 possible bipartitions, but only 66 qccur ainong the skinks. For a functional 
gene not all bipartitions will be present because of ftihctional constraints, but such a small proportion of 
bipartitions in the skink data set suggests that the sequences do contain phylogenetic history. A data set 
with no conflicting signals would, for 20 taxa, produce 17 signals, each representing an internal edge of 
the phylogenetic tree (in addition there will also be 20 o�er signals corresponding to each pendant 
branch). 
Spectral Analysis of other Vertebrate 12S rRNA Data Sets 
To investigate whether the skink 12S rRNA molecule was unsuitabie for the study, spectra were 
determined for two other vertebrate 12S rRNA data sets. The ratite data set (12 taxa) gives a resolved 
phylogeny using other methods (Cooper et al. 1992), while the Pecoran bovid data set produces a less 
resolved tree (Kraus & Miyamoto 1991, Allard et al. 1992). The ratite 1 2S rRNA sequences have a 
greater number of strong spectral signals and a resolved phylogeny is produced (Fig. 5 .2a). The bovid 
data set however produces a spectrum similar to the skinks; and relationships cannot be completely 
resolved (Fig. 5.2b). 
Parsimony and Neighbor-Joining Analyses of the skink Data Set 
Spectral analysis of the ratite and bovid sequences confirmed the results from previous phylogenetic 
analyses (see Cooper et al. 1992, Allard et al. 1992). However, to conlnm that aspects of the Hadamard 
conjugation were not a cause of lack of resolution among the skink sequences, maximum parsimony 
Ii I 
i 
� 31 
� : II) i 
I .J 
i 
1 
! 
I I 
i 
1� 
I 
Fig. 5.1. Spectral analysis of 382 bases of 12S rRNA from 20 Leiolopisma and related taxa. In this 
example the bipartitions are ranked. rather than each bipartition being identified by it's unique number. 
The taxa in the four most well supported bipartitions are identified; 
-. 
Lmi = L. micro/epis Lsm = L. smithi 
Lfr = L. injrapunctatum Lag = La.. guichenoti 
Cae = C. aenea Lfa = L· fallai 
Lze = L. zelandicum Lmo = L. maco 
Note that most of the other bipartitions bave similar amounts of support in this data set No fully 
resolved tree is produced. 
-- -- ------
+ Lag 
I I 5 7 8 11 13 15 1 18 21 23 25 27 8 31 33 5 7 38 '1 " 45 '  " 1 3 55 7 8 81 83 85 
8 10 12 a 18 18 20 22 U 28 28 30 32 3' 38 38 .0 � « .a .a 50 52 5' 58 58 80. 82 8' 88 
Rank of Bipartition 
8 
3. 
7-
I-
1 "-
,-
S-
2-
1-
0 
o ' 
Fig. 5.2a. Spectral analysis of 380 bases of 12S rRNA from 12  ratite birds (data from Cooper et aL 
1992). A resolved tree is produced using HADTREE and is shown in the upper right comer of the 
figure. Compare this spectrum with that of the skinks in figure 5 . 1 .  In particular, note the different 
relative frequencies of bipartitions in the ratite data set Full species names and sequences are given in 
Appendix 2. 
Ratites 
Roa Kiwi nomloptery:c 
>-�--�,---�--�,­L.spot. KiWI 
Brown Kiwi 
Ostrich 
Fig. 5.2b. Spectral analysis of 387 bases of 12S rRNA from 15 bovids (data from Kraus & Miyamoto 
1991  and Allard et al. 1992). A fully resolved tree is not produced by HADTREE. As with the skinks, 
many of the bipartitions have similar frequencies. The bovid data set is listed in Appendix 2. 
Bovids 
O!apter 5, page 61. 
(Swofford 1991) and neighbor-joining (Saitou & Nei 1987) analyses were also performed. Both of these 
latter two methods give similar results (Figs. 5.3 & 5.4). Internal branches are short and most of these 
are collapsed when bootstrapping is performed. The unbootstrapped trees are presented in figures 5.3 
and 5 .4 however since they at least provide clues for possible relationships among the taxa. Both 
parsimony and neighbor-joining produce similar estimates of relationships of the skinks (Figs. 5.3 & 
5.4). 
Partially Resolved Relationships Using the Hadamard Conjugation 
Although a completely resolved set of relationships for the Leiolopisma is not obtained, the spectrum 
does indicate partially resolved relationships (Fig. 5.5). In tb:is case every taxOD is linked to at least ODe 
other but an unambiguous branching order for all taxa cannot be determined. The four strongest signals 
in the spectrum (Fig. 5.1) correspond to the pairs L. microlepis+L. smithi, L. infrapunctatum+La. 
guichenoti, L./aUai+C. aenea and L zelandicum+L. "!Oco respectively, though support for the latter two 
is comparatively weak. The same clustering pattern of the first three of these pairs is also evident in the 
parsimony (Fig. 5.3) and neighbor-joining trees (Fig. 5.4). 
It is not the purpose of this thesis to compare in detaii the Hadamard conjugation with other tree I 
reconstruction algorithms, and more detailed comparisons between the trees are not considered here. 
Discussion of the Hadamard conjugation in relatiop to other methods can be found in Penny et al. (199 1 ,  
1992). 
Weighting of Characters 
The preceeding analyses treated all sites and types of changes in the sequences equally. Differentially 
weighting characters or regions of the sequeDce can be used to take account of different rates of change 
or violations in assumptions used in the model. However deciding what to weight and appropriate 
weighting values is often subjective and problematical (Swofford & Olsen 1990, Cracraft & Helm­
Bychowski 199 1 ,  Mindell 1991). Weighting of sites on the basis of type or frequency of change have 
been suggested. 
Transversions often accumulate less rapidly than transitions in DNA sequences (see Jukes 1980, Wilson 
et al. 1985, Avise et al. 1987, Moritz et al. 1987). This sjower rate may mean that they are less likely to 
be obscured by multiple substitutions than transitions, �d so m�y be more phylogenetically informative 
when comparing distantly related taxa Giving transversions tnore phylogenetic importance by 
weighting them is therefore sometimes used (see MindeIl 1991). A seCond weighting scheme takes 
account of the fact that, for rRNA sequences, fixation of tnutations in base-paired regions may not be 
independent If one member of a base pair changes, selection for the maintenance of the pair bond could 
increase the chance of fixation of a complemenatry charige iti the other member of the bond. Paired 
regions may therefore be weighted differently from unpaired regions in phylogenetic analyses (Mindell 
& Honeycutt 1990) .. A third method, as discussed by Simon (1991), is that regions of different levels of 
sequence variability in a molecule may be more informative for different depths of phylogenetic 
Fig. 5.3. Consensus of nine equally parsimonious (shortest) trees for the 382 bases of 12S rRNA 
sequence from 20 Lei% pisma and related taxa The tree shown is a 50% majority rule tree, that is, 
only associations of taxa which occur in at least half the trees are indicated. Lengths of edges are 
proportional to the number of changes along each edge. Note that many of the internal edges are 
short relative to pendant branches. Thick lines in the tree denote those branches which do not 
collapse after 10  bootstrap replications (more could not be performed because of the long 
computational time required). PAUP version 3 .0s was used for the analyses. 
L. maccanni 
L. notosaurus 
L. smitlzi 
L. telfairi L. microlepis 
L. n. nigriplantare 
L. n. polychroma 
L. infrapunctatum 
St. Is. Green 
L. grande _______ L. fallai 
1---"::;;;;- C. aenea 
...----L. acrinasum 
L. otagense 
Fig. 5.4. Neighbor-joining tree for the 382 bases of 12S rRNA sequence from '20 Lei% pisma and related 
taxa. (phylip version 3.4 was used). Percentage values on internal edges correspond to the number of times 
that edge occurred in 100 bootstrap replications. Only four of the internal edges occur in more than 80% of 
the bootstrapped trees. Lengths of branches are proportional to the probability of change along that branch. 
Note that internal edges are relatively short in comparison to the pendant branches. The tree is unrooted. 
Compare this tree with the parsimony tree in figure 5.3 and the Hadamard tree in figure 5.5. 
L . notosaurus 
55% 
37% L. inconspicuum 
L. macca nni 
I""'"" 45% 
C. aenea 
L.fallai 
� - L. moco 
L. ota gense 
L. zelandicum 
L. tel/airi 
- L. lin/chl 
36% 
87% 
L. n. nigriplantare 
"--- L. n. polychro ma 
84% 
La. guichenoti 
59% L. l 'nfrapunctatum 
\ L. acrinasum 
J suteri 
l' L. microlepis 
100% \ L. smithi 
St.Is.Gree� 
87% 
L. rande g 
Fig. 5.5. Inferred phylogenetic relationships for the 20 skink taxa based on spectral analysis of 382 
nucleotides from the 12S rRNA gene. The frequency of each grouping of taxa (bipartition) is shown next to 
the branch leading to that group. The clusters cannot be reconciled into a single phylogenetic tree and this is 
indicated by the broken lines. Low frequency values indicate that few shared nucleotide substitutions occur to 
support that bipartition and the true phylogenetic relationships for groupings of such taxa are therefore 
uncertain. Transitions and trans versions are given equal weighting. Branch lengths � proportional to the 
degree of divergence between taxa. 
L. grande 
L mOkO \ 
U 
L. zelandicum � ..... ..... 
L. smithi ) ... , -;' ... 
L. microlepis 
L. suteri 
1 .2 
I 
I 
0.6 
I 
L. in!rapunctatum __ -' 
..... 
L. notosaurus 
.i"---L. inconspicuum 
L. maccanni 
1 . 1  ..... 
..... ..... 
(L. tetrairi 
..... 
L. otagense 
..... ..... ( c. aenea 1 .8 
L.fallai 
L. acrinasum 
-r---L. n. polychroma 
L. n. nigriplantare 
'The Gods of the earth and sea 
Sought thro' nature to find the Tree; 
B uJ their search was all in vain: 
There grows one in the Human Brain . "  
(Blake) 
OJapter 5, page 62. 
branching and could be weighted differently to reflect Ibis. These three weighting options were 
investigated for the skink sequences to determine if they e�an�ed or reduced signals and lead to better 
phylogenetic resolution. 
Weighting of Transl'ersions 
Transitions outnumber transversions in the skink data set by, on average, four to one. A four-fold 
weighting factor applied to transversions did not greatly alter the strongest signals in the spectrum but it 
did change some of the other signals (Fig. 5.6). The relationships betWeen L. grande and Stls. Green, 
I 
and between L. n. polychroma and L n. nigriplantare are lost for example, because many of the 
differences between these pairs of taxa are transversions (Table 3.1). When the subtrees for this 
differential weighting are drawn it can be seen that L. grande, L acrinasum and L otagense cannot be 
linked to any other taxa (Fig. 5.7), in contrast to the more resolved relationships in the equally weighted 
analysis (Fig. 5. 5). Given this result, and the fact that high numbers of transversions do not signify more 
distant relationships (Table 4.3, Fig. 4.2), transitions and trarisversions are given equal weighting in all 
further comparisons. 
Weighting of Paired Regions 
Wheeler & Honeycutt (1988) reported that for both animal and plant groups, analysis of paired regions 
in 5S rRNA were phylogenetically misleading whereas phylogenies derived from unpaired regions were 
I 
in more agreement with relationships inferred from morphological data. This they attributed to selection 
maintaining base pairings in rRNA. They suggested that a c�ange in one member of a base pair 
increases the chance of flxation of a complementary change in the other base pair, and Wheeler & 
Honeycutt consequently recommended assigning weights of 0.5 to nucleotides involved in pairings; that 
is, a change at an unpaired site can be considered twice as informative as a change in a nucleotide 
which is involved in a base pair. This level of weighting is only appropriate however if complete 
compensation occurs in paired regions, but this is not the case for skinks (Table 3.6) or many other 
groups (Smith 1989, Hedges et al. 1990, Mindell & Honeycutt 1990, Hillis & Dixon 1991, Simon 1991, 
Dixon & Hillis 1992). Nevertheless, paired regions can be phylogenetically informative. Dixon & Hillis 
(1992) examined relative weighting of paired regions for 28S rRNA and found that a weighting factor 
closer to 1 .0 than 0.5 was appropriate for their study since more changes in paired regions were 
uncompensated than compensated. 
I 
At least 54% of the nucleotides in the skink sequences are involved in base pairing (Fig. 3 .3), and 39% 
of the variable sites OCCur in paired regions (Table 4. 1 7). A high proportion of changes have both pairing 
partners changed and pairing is maintained (Table 3.7). 
Dixon and Hillis (1992) adopted a linear scaling approach to weighting paired regions. With this method 
the weighting given to paired nucleotides is calculated by comparison to the number of changes 
expected by chance and with no compensation. Following tbeir scheme, a relative weighting of 0.7 1  is 
obtained for the paired regions in the skink molecule (Table 5.1). This equates to a weighting of 
10 
II I-
I I-
.-7 
8 1-
J · 1-
,-
3 1 -
2 :-
1 -
o 
1 
Fig. 5.6. The effect of weighting transversion substitutions four-fold higher than transitions in the skink 
data set. The major difference from the equally weighted spectrum (Fig. 5 .1)  is that with differential 
weighting there is less support for the C. aenea+L. faIlai and the L zelandicum+L. moco bipartitions. 
Minor differences occur in the strength of support for other bipartitions. but as in figure 5.1 most of the 
bipartitions have similar levels of support. Resolution of the skink relationships is not improved by 
giving trans versions greater pbylogenetic importance than transitions (see also Fig. 5.7). 
(Transversions)*4 
Lmi  + Lsm ' . , 
Lfr + Lag 
Cae + Lfa 
Lze  + Lmo 
:11: 
II 
, 
3 5 7 II 1 3 il i� II 1 � � � 1 � II 41 j3 &: 4 D 1 3 �j II 1 � U 
2 4 8 I 10 12 a 18 11 20 22 24 28 28 30 32 34 31 31 jf j2 .. ... ... 50 52 54 58 511 80 82 84 118 
Rank of Bipartition 
Fig. 5.7. Inferred phylogenetic relationships for 20 skink taxa based on spectral analysis of 382 nucleotides 
from the 12S rRNA gene. The frequency for each bipartition is shown next to the branch leading to it. 
Transversions are weighted four times higher than transitions. Compare these relationships and signal strengths to 
Fig. 5.5. Note that using this weighting scheme, L.grande, L. acrinasum and L. otagense cannot be placed with 
other taxa , and that L. n. nigriplantare now groups with L. telfairi rather than with L. n. polychroma. 
L. n. polychroma 
L. smithi L. infrapunctatum 
L. suteri 
i--L. microlepis -s---La. guichenoti 
1 .8 2.3 
L. notosaurus 
L. inconspicuum _-� 
1 .8 
L. lelfalry. nigriplllnlllre 
L. /allai 
C. aenea 
L. linlchl ---.... 
1 .8  
L. grande L. acrinasum L. otagense 
I I I 
OIapter S, page 63. 
Table 5.1. Determining a relative weighting factor for pHylogenetic analysis of paired regions in the 
skink 1 2S rRNA fragment B� 888 99g�P,,!88 BtHB6ers ee eB8Bges ill I'EHre8 FegieBS are frem Table 
�. If changes are random all base substitutions will be �ndependeiit and therefore no differential 
weighting should be applied. Alternatively, if all chang� ill paired regions were compensatory, then 
each substitution should be given a weighting of 0.5. Ol;mges in the skink helices fall nearly halfway 
between the two extremes. 
Random Complete 
Expectations Observed Com,jensation 
No. Changes 1 5.3 56 85 
.J, .J, .J, 
% Dependence 0 %  58.4% 100 % 
.J, .J, ' .J, 
Relative Weighting 1 .0 - �.71 0.5 
10 
8 
7 
e 
J 5 
:I 
2 
1 
0 
Fig. 5.8. The effect of a three-fold higher weighting factor for substitutions in unpaired regions of the 
skink 12S rRNA sequence three-fold higher than changes in paired regions (see text and Table 5.1). The 
bipartition of L. inJrapunctatum+La. guichenoti is now the most strongly supported, indicating that 
substitutions supporting this bipartition occur mostly in unpaired regions. This is confIrmed by 
examining the sequence alignment (Table 3. 1). In contrast, substitutions supporting the L. microiepis+L. 
smithi bipartition are more common in paired regions (Table 3.1), and the spectral signal for this 
bipartition has decreased (compare with figure 5. 1). No fully resolved tree is produced with this 
weighting scheme. 
Lfr + Lag 
I 1ft (Unpaired SitaI)t3 I 
. e 8 W � U " " � H U H e � ft U H � � � " � � � H M � e " n " " " � n  
Rank of Bipartition 
Chapter 5, page 64. 
approximately 3: 1 for unpaired versus paired regions. The spectrum obtained using this value differs 
little in general from that obtained by equal weighting - no fully resolved tree is produced (Fig. 5.8) .  The 
four major signals remain, though the signal for the bipartition of L. inJrapunctatum+La. guichenoti is 
now the strongest, while the signal for L. microlepis+L. smithi is reduced. The three sites supporting the 
L. inJrapunctatum+La. guichenoti bipartition are all in unpaired regions, while two of the three positions 
which support the L. microlepis+L. smithi bipartition are in helix 34 (see Table 3 . 1). The signals ranked 
from 5 to 27 in figure 5.7 are proportionally higher than in the unweighted spectrum, implying that they 
are derived from unpaired sites (compare Fig. 5.8 to Fig. 5. 1). The Hadamard spectrum and a relative 
weighting scheme are therefore an easy way to see the extent to which paired regions contribute to 
bipartition signals. 
Weighting of More Variable Sites 
Eight positions in the skink data set have both transitions and transversions (Fig. 4. 1 7). These sites are 
potentially rapidly changing positions which could be useful for resolving relationships amongst closely 
related taxa but could mislead more distant phylogenetic relationships because of unobserved 
substitutions (Mindell & Honeycutt 1990, Simon 199 1). This does not however appear to be the case for 
the skinks. The taxa which have the fewest numbers of differences between them also tend to group 
together in figure 5.5, so small numbers of nucleotide substitutions between taxa seems to be a good 
indicator of closeness of relationships (see also the neighbor-joining tree, Fig. 5.4). Taxa classified on 
this basis do not always differ at the most variable sites. L. microlepis and L. smithi for instance, differ 
from each other at five positions, but only one of these (position 204) is classed as a highly variable site 
(Table 3 . 1 ). 
Down-weighting the eight most variable sites by a factor of two or four has little effect on the spectrum 
(Fig. 5.9, compare to Fig. 5.1). However, excluding these eight sites entirely from the analysis results in 
the loss of 14 (21 %) bipartitions (Fig. 5. 10). These lost bipartitions represent groups of seven, eight or 
more taxa, rather than pairs of taxa, indicating that removal of these sites has only a minor effect on 
resolution of pairs of taxa. The reduction in the number of bipartitions when the most variable sites are 
removed does suggest though that these sites are a major source of conflicting signals. A resolved tree is 
still not produced however. 
The Effect of Consront Columns in Sequence Analysis 
The correction for unobserved changes in the Hadamard conjugation assumes that all sites are free to 
vary. This is not the case for functional genes (Fitch & Margoliash 1967, Shoemaker & Fitch 1989). As 
shown in Chapter Three, one third of the region of the 12S rRNA molecule under scrutiny here is 
conserved across the vertebrates. Assuming that all sites are free to vary will mean that corrections for 
unobserved changes will underestimate the true numbers of substitutions. This is because the frequency 
of change per site is lower if averaged across the whole sequence. If only a proportion of the sites can 
change then the actual frequency of substitutions at variable sites is higher, and so the probability of 
multiple changes sho.uld be higher. Lengths of some potential internal edges in the tree will then be 
longer and may result in greater resolution. 
15 
,-
13 -
12 ,-
11 -
10 -
-
-
-J : 
8 -
-
-
s -
2 -
1 -
o 
1 
Fig. 5.9. The effect of giving a lower weighting to the eight most variable sites in the skink 12S rRNA 
sequence. (Both transition and transversion substitutions occur at these sites). Weighting these positions 
either two-fold (5.9a) or four-fold (5.9b) lower than the other sites in the sequence has little effect on the 
relative differences between bipartitions. The weighting of sites was done by simply duplicating. prior to 
analysis. the 374 less variable sites once (5.9a) or thrice (5.9b) with respect to the other 8 sites. 
Consequently. the frequencies of the bipartitions are proportionally higher than in the original spectrum 
(Fig. 5.1).  
3. 
b. 
I � 8 lites 1ft. = 0.5 
-, 
, 
"III, 
rrrrrrrrrm 
1 3 5  7 It 1 3 l� It 21 � � )0: � 1 � I! l� III 41 4S . 41 1 � I! AoJ AJ 81 as � 
2 4 8 I a � u a H � � U U U � � U B B � � " � � H U " M H H a " H  
Rank fA Bipartition 
[1118 sits 1ft. = 0.25 
4 S 8 10 12 U IS 18 20 22 U 2S 28 30 32 34 3S 38 � � U 4S 48 50 52 54 5S 58 SO S2 S4 SS 
RJmk of BiDartition 
I 
5 
�-
:-
2 :-
1 -
o 
Fig. S.10. The effect of removing the eight sites in the skink 12S rRNA data set which have both 
transition and transversion substitutions. Both the original spectrum (from Fig. 5 .1) and the edited 
sequence spectrum are shown. Removal of the 8 sites has a greater effect than just giving them a lower 
weight in analyses (Fig. 5.9). Support for 14 bipartitions is lost when the eight most variable sites are 
removed, but the frequencies of other bipartitions are largely unchanged. The lost bipartitions generally 
contain eight or more taxa and so resolution of the relationships of more closely related taxa (those with 
few nucleotide differences between them) does not appear to be affected. Removal of the most variable 
sites, and consequent reduction in the number of bipartitions, does however indicate that such sites may 
be contributing noise to the analyses of more distantly related taxa. 
In 
� all sites r2'l minus 8 sites 
Ill_It 
1 3 5 7 I 1 13 15 17 11 1 � 7 1 31 It· 35 7 31 U .:s � .7 ., 1 3 55 57 1 81 83 85 
2 • 8 8 10 12 U 18 18 20 22 2. 28 28 30 32 3. 38 38 .a � u .a .a 50 52 U 58 58 80 82 U sa 
RB:nk of Bipartition 
I 
ClJapter 5, page 65. 
The presence of an invariant core in the molecule and its affect on the Hadamard transform can be 
approximated by reducing the number of constant sites in the data. As the number of columns is reduced 
an expectation is that the strength of the signals should increase since there will be a greater correction 
for unobserved changes. Signals whose strength is indepe�dent of the number of constant columns are 
likely to be spurious and have little phylogenetic infonnatioQ. 
The Effect on Skink Phylogeny of Weighting of Consllmt Columhs 
To examine this problem a subset of 10  skink taxa were chosen so that the effects are easier to detect. 
The taxa were chosen because they illustrate an extreme c� where the two largest spectral signals are 
not included in the optimal tree (Fig. 5 . 1 1 ;  see also Penny ef al. 1993). Are some of the conflicting 
signals in this data set a result of inaccurate correctiohs for upobserved changes? 
The frequencies of each bipartition in a data set are calculated by the P� ARE program. This 
information is written into a separate file which is then used by HADTREE to generate the spectrum. It 
was simpler to adjust for constant columns by changing paraineters in this frequency file than to remove 
constant columns from the original data set Only two numbers need to be changed to achieve this; one 
represents the total number of columns of data, and the other the number of these columns which 
represent constant sites in the data. The effect of progressively removing constant columns (1/3, 1/2, and 
213) prior to the Hadamard conjugation was investigated for the 10 taxa data set Figure 5.12a shows that 
as more constant sites are removed most of the biparti�oil frequencies increase slightly, though some do 
not. The increase in size of the bipartition frequencies ls relatively small (Fig. 5.12a), and all of the 
seven original internal edges in the optimal tree are retained even when two-thirds of the constant sites 
have been removed (Fig. 5.12b). This indicates that the majqrity of the inferred phylogenetic 
relationships are robust and not due to invalid an assumption in the model (namely, that all sites are free 
to vary). 
The last few bipartitions in the spectrum in figure 5.12 have little change in ¢eir frequencies as more 
constant sites are removed, suggesting that they are probably spurious signals caused by the correction 
for unobserved changes. None of these signals in fact exist as observed changes, they are all produced 
by the correction for unobserved changes. 
Summary of the Effect of the Weighting Schemes on the Skink Phylogeny 
Adjusting the analyses to take account of a transition bias, differences in frequency of changes between 
paired and unpaired regions, and different frequencies of nucleotide substitution do not enhance 
resolution of the skink phylogenetic relationships. There are no large changes in the spectral signals for 
individual bipartitions which suggests that the bipartitions with the most support do represent relatively 
close phylogenetic relationships rather than being artifacts due to noise or limitations of analyses. 
II ._. __ ._  .. _ .  __ .__  .__ ....  __ .._----
G-
4-
J s- o 
2-
1 
A IIII .. 2G8 '1111 78 147 82 11111 �88 Sill III� 811 SII� 77 m WI 102 510 130 1" 40G 117 104 271 247 l1li 224 182 1110 2114 40S SSO 418 42S sal 
BipartJ.tJon 
1 
3>+ 
1 
J�nnrmnrllll-rnn� 10 
-?,0 8 
2 -( 
1 
3 
7 
rn � m n m m � " � m � � m � m 42S W  
1. St. Is. Green .(1) 
2. L. /ill/chi (2) -
3. L. graTUle (4) 
4. L. micro/epis (8) 
S. C. aenea (16) 
BIpart.it.ton 
6. L. inconspicuum (32) 
7. L. ze/andicum (64) 8. L. maccann; (128) 
9. L. n. po/ychroma (256) 
10. L. otagense (5/2) 
9 
8�------------------------------------------------------------------------' 
a. Constant Sites Remaining 
5-
2 
1-
258 5 88 78 �4.7 82 185 388 351 "3 88 38" 77 317 328 -402 10 
130 1" "85 87 10" 271 2"7 ee 22" 182 180 28" -'03 SSO "18 -'23 381 
3,-------------------------------------------------------------------------�
b. 
2 
1 
1-· 
258 5 88 78 
130 1" "85 
1 "3 118 
88 22" 1112 
Bipartition 
OJapter 5, page 66. 
Spectral Analysis of Simulated Sequences 
The preceeding analyses have demonstrated that lack of pbylogenetic resolution among the sldnk taxa is 
not due to the HADTREE program, nor to insufficient variable sites, nor biases in sites or types of 
substitutions. In Cbapter Four it was sbown that randomly generated sequences gave a similar pattern of 
transitions and transversions as the skink: data (Figs. 4.2 & 4.9). In this section spectral analyses of 
randomly generated sequences are performed and compared to the skink: spectrum (Fig. 5.1). 
Using the program described in Cbapter Four, sets of twenty random sequences were made. These 
sequences were 384 nucleotides long, and bad an average of 25 total cbanges and five transversions 
between eacb sequence, corresponding to the mean differences between the skink: taxa (Table 4.2). 
Spectral analysis of these sequences gave spectra like the one shown in figure 5.l3a. This is similar to 
the skink: spectrum (Fig. 5.1) in that there is no step-wise decrease in signal frequency (as in Fig. 5.2), 
and there are large numbers of signals with the same frequencies. There are differences bowever 
between the skink: and simulated sequence spectra. The skink: spectrum bad only 66 signals but the 
simulated sequence spectrum (Fig 5.13a) has 157. Tbere are also more variable sites (75% as compared 
with 26% in the skinks) and more singleton changes (62% versus 42%) in the simulated sequences than 
in the skink: sequences. 
The simulated data bave an important difference from the skink: data however - every site is free to 
cbange. In the vertebrate sequence only two-thirds of the molecule appears to be free to vary (Fig. 3.5). 
To more closely emulate the skink: data set the simulation was modified so that the 25 cbanges occwred 
in 264 bases (two-thirds the length of the original sequence). Tbe new spectrum (Fig. 5.l3b) is generally 
the same as that wben all 384 sites are free to vary (Fig. 5. 13a), but the proportion of singleton changes 
(46%) is now more similar to the frequency of singletons in skink: data set When invariant sites are 
taken account of therefore, the simulated data describe the pattern of variation in the skink data quite 
well. The implications of this are discussed later in this chapter. 
Relationships of Northern and Southern New Zealand Lewlopisma 
The relationsbips for all of the taxa are poorly resolved (Fig. 5.5), but can more information be obtained 
from subsets of the taxa? The New Zealand species can be sub-divided into two groups, based on their 
occurrence north or south of the Nelson-Marlborough region (Table 5.2; see Pickard & Towns 1988). 
Biogeographic connections exist between the southern North Island and Nelson-Marlborougb, reflecting 
their connection during the Pleistocene (Fleming 1980, McGlone 1985). Clines in allozyme frequency 
for skinks also occur across Cook Strait supporting this past continuity (C.H. Daugberty pers. COIIllil.). 
L. n. polychroma is widely distributed and so is included in both groups. L. n. nigriplantare from the 
Chatham Islands shares close genetic similarity to L. n. polychroma (Table 4.1 ,  Fig. 5 .1 ,  Daugherty et 
aI. 1990b), and so is also included in both northern and southern groupings. Members of the L. 
lineoocellatumlchloronotum complex occur in both the north and the south, and along with L. telfairi 
and La. guichenoti are included in both subsets as well. 
J 
J 
Fig. 5.13. Spectral analysis of 20 randomly generated sequences (see Chapter Three for details of the 
simulation program). a. Shows a sample spectrum obtained when sequences are 384 bases long, the 
mean transition/transversion (TsfTv) ratio is 4/1, and an average of 25 substitutions occur (as in the 
skink data set). As with the skink spectrum (Fig. 5.1), there are many bipartitions of similar frequency, 
but the simulated sequence spectrum has more bipartitions. The star-tree model which generated the 
simulated sequence data is shown in the top right comer of the figure. In h. the spectrum of another 
simulated spectrum is shown. In this example the effect of having sites not free to vary is approximated 
by reducing the sequence length by one-third. to 264 bases, while still having on average 25 
substitutions and a TsfTv ratio of 4/1 .  The overall spectrum is similar to that in 5 .13a. though more 
bipartitions now occur. 
s�------------------------------------------------------------------------------
a. 
384 Bases 
2 
1 
158 
264 Bases 
1 
O+----.JfHffIo 
10 
a 
'-7 
e 
,-
:I 
2 
1 
o 
a. 
Fig. 5.14. Spectral analyses of 12S rRNA sequences from a. 1 1  slunk taxa which occur in northem New 
Zealand, and b. 10 slUnk taxa which occur in southem New Zealand. The AusLralian Lampropholis 
guichenoti and the Mauritian LeioiopisllUl teljairi are also included in both data sets. All the spectral 
signals are shown in these examples, but UJe only labelled bipartitions are those corresponding to the 
branches leading to individual taxa (these signals are independent of the tree). The other bipartition, 
representing potential intemal edges in trees, are identified in figures 5.15 and 5.17. There are two major 
points to note in figure 5.14. Firstly, the number of changes on branches leading to individual taxa varies 
greatly. Secondly, the differences between intemal edge bipartitions are smaller in LlJe spectrum for LlJe 
southem taxa LlJan in LlJe northem taxa, indicating poorer resolution of LlJe relationships of the southern 
taxa (see also Figs. 5.16 & 5. 18). 
1= 
3 5 7 9 11 13 
1 4 6 S 10 12 
, 
Northern Taxa 
TTIJ .. 
Bipartition 
10l -·-·- ·-·- --- -·- ----.. ----·-------.. ··----------·-
b Southern Taxa . 
• ,-
• I-
,-7 
• I- � 
>-
1-
. a 
-
o , 
1 I I 7 • 11 
a , • • 10 11 
� 
� t< " 
T 
' . 
I 
'�nr rJ 
::lnl IT. . -I 
i 
1 
I 
Chapter S. page 67. 
Table 5.2. The New Zealand Leiolopisma taxa for which 12S rRNA sequence data is available 
subdivided on the basis of present geographic distribution (see Pickard & Towns 1988). Stewart Island 
taxa (L. notosaurus and St.Is. Green) are included with the South Island species. Taxa occuring north of 
Nelson are listed as being in the North Island. L. n. polychroma. L. n. nigriplantare. and L. 
lineoocellatumlchloronotum are included in both islands (see text). 
South Island North Island 
L.n. polychroma L.n.polychroma 
L. n. nigriplantare L. n. nigriplantare 
L. linlchl. L. linlchl. 
L. grande L. zelandicum 
StIs. Green L. infrapunctatum 
L. maccanni L. microlepis 
L. inconspicuum L. smithi 
L. notosaurus L. suteri 
L. otagense L. fallai 
L. acrinasum L. moco 
C. aenea 
<llapter 5, page 68. 
Spectral Analysis of Skinks from Northern New Zealand I 
The northern New Zealand skinks have stronger and more sffUctured spectral signals than the southern 
group (compare Figs. 5. 14a & 5. 14b), indicating that there is less ambiguity in the phylogenetic 
relationships of northern taxa This is also shown by the relatively longer internal branch lengths in the 
optimal tree for the northern taxa (Fig. 5. 15). 
The Complex Relationships of C. aenea 
C.aenea clusters with L telfairi in this example (Fig. 5.15), though in the spectra for all 20 taxa (Fig. 
5 .1), C.aenea groups with L fallai. C.aenea and L. fallai differ by only 17 differences, and the 
bipartition of these two taxa is supported by Jhree shared substitutions (at poSitions 67, 73 & 140; Table 
3 .1). These substitutions are in the relatively conserved si , and 1.,33 ' regions (Figs. 3.3 & 4.17), which 
could imply C.aenea and L. fallai share a common ances�or. lbe bipartition of C. aenea and L. telfairi is 
also supported by three substitutions (positions 194, 263, & 331 ;  Table 3.1), one of which is in the tetra­
loop of stem 38. In addition to these bipartitions, C. aenea, L. fallai, and L. telfairi share other 
substitutions among themselves and with other taxa. One substitution (at position 291)  is shared by C. 
I 
aenea, Lfallai, andL. telfairi (along with L suteri and L. nwco; Table 3 . 1). Three observed 
substitutions group C. aenea non-exclusively with L fallai (positions 30, 35, & 64), while C. aenea and 
L. telfairi have a substitution in common with three other taxa at position 309 (Table 3 . 1). This 
illustrates the complexity of the relationships among some of the skinks. 
The case of C. aenea serves to demonstrate another feature of HADlREE. The bipartition favouring C. 
aenea+L.fallai (#288) has about the same strength of suppo� as that favouring C. aenea+L. telfairi 
(#544; Fig. 5.15). An option in HADlREE allows specific bipartitions to be forced into the optimal tree. 
When the C. aenea+L fallai bipartition is forced however, no resolved tree is produced. This indicates 
that the pattern of sequence relationships in this data set are ��tet explained by Cyclodina having a 
more distant relationship to the New Zealand Leiolopisma. 
C. aenea can however be made to group with L. fallai if L. lin/chi is omitted (Fig. 5 . 16a). The L. 
lin/chi individual used in this study was collected from Otago, the main range of L. chloronotum (see 
Pickard & Towns 1988), so there is a rationale for excluding it for this example. Forcing the C. 
aenea+L. telfairi bipartition (#272) into the tree changes six of the nine intemal edges, and some of the 
internal edges in the u'ee shoJ1en (Fig. 5 . 16b). Given that some of the shared nucleotide substitutions 
grouping C. aenea both with L. fallai and with L. telfairi are at relatively well conserved sites, the 
phylogenetic relationships of C. aenea are uncertain. 
Why L. lin/chI affects the grouping of C. aenea is not clear. Exclusion of L. lin/chi also increases the 
strength of the L. n. nigriplantare+L. n. polychroma bipartition (compare figures 5 . 15  & 5.16). The 
conflicting signals associated with inclusion of L lin/chi were also noted in the analyses shown in figure 
5.1 1  
Fig. 5.15. Bipartitions included in the optimal tree for 1 1  northern New Zealand and two overseas skink 
taxa (Only internal bipartitions are sbown in the grapb, and are given in their binary number notation). 
The optimal tree is illustrated on the right (see also Fig. 5.18), and bipartitions excluded from the tree 
are plotted below the x axis. The relatively strong signal supporting the bipartition of C. aenea+L. fallai 
(#288) is not in the tree, and C. aenea groups with L. telfairi (bipartition #544). No fully resolved tree is 
formed if bipartition #288 is forced to be included in the tree . 
. ,-----------------------------------------------------------------------------�
10 
6 
4 ... 'V
I3 5 
7 I 
'12 1 
9 2 3
1 
11 
1. L. lin/chi (1) 8. L zelandicum (128) 
2. L. suteri (2) 9. L.fallai (256) 
3. L. n. nigriplantare (4) 10. L. telfairi (512) 
4. L. microlepis (8) 11 .  L n. polychroma (1024) 
5. La. guichenoti (16) 12. L. inJrapunctatum (2048) 
6. C. aenea (32) 13. L. moco (4096) 
7. L. smith; (64) 
: 3. 1-
Unforced 
; 
! I .� ! 2 I I 7� 
i 
I 2 
I'" 
10 / I , 
i :� 
I :� 
12 
10 
1 
1 
0 ---
-I 
�� 
... -!- I I 
10 
Forced 
• b. 
9 
• 
7 
2 
• IOr- I O  
I 
J :  
I 
1 
0 
-1 
� 
... 
1. L. suteri (1) 7. L zelandicum (64) 
2. L n. nigriplantare (2) 8. LJaiiai (128) 
3. L microiepis (4) 9. L. telfairi (256) 
4. La. guichenot; (8) 10. L n. po/ychroma (512) 
5. C. aenea (16) 1 1. L. inJrapunctatum (1024) 
6. L smithi (32) 12. L moco (2048) 
Olapter 5, page 69. 
Spectral Analysis of Southern New Zealand Skinks 
In comparison with the northern taxa (Figs. 5 .15 & 5.16) the relationships of the southern taxa (Fig. 
5 . 17) are less certain. All of the internal edges in the latter tree are sbort. The grouping of L. grande with 
St Is. Green bas the most support. L. n. polychroma does not appear to be closely related to other 
members of the cryptic species complex (L. maccanni, L. inconspicuum, and L. notosaurus) as described 
by Daugberty et aI. 1990b. This point will be considered again in Chapter Six. 
L. n. polychroma and L. n. nigriplantare form a distinct cluster in the tree ()f northern species (Fig. 
5.18a). This is not the case with respect to southern taxa wbere these two taxa bave affinities to L. 
lineoocellatumlchloronotum and L. acrinasum (Fig. 5 . 18b). A division between northern and southern 
taxa (with L. n. nigriplantare and L. n. polychroma clustering with southern species) is also apparent in 
the allozyme data (C.H. Daugberty & G.B.  Patterson pers. comm.; and see Chapter Six). L. n. 
polychroma may therefore bave spread no�wards. 
Mitochondrial Lineages of the Skinks 
On the basis of both the numbers of differences between taxa (Table 4.1) and the relationships inferred 
by spectral analysis (Figs. 5.5, 5.15, & 5.18), twelve Leiolopisma lineages can be identified (Table 5.3). 
Taxa whicb have the strongest spectral signals differ by fewer than 20 substitutions (less than 5% 
sequence divergence), so I bave cbosen this value as the criterion for identifying a lineage. As an 
example, L. n. nigriplanJare and L. lin/chi differ from L. n. polychroma by 13 and 19  substitutions 
respectively (Table 4. 1), while 20 or more differences separate L. n. polychroma from the other taxa 
These three taxa are therefore classed as one lineage, and 1 1  lineages occur in New Zealand (Table 5.3). 
Ten of these lineages are no more different from eacb other �an any of them are from L. tel/airi (Table 
5.3), suggesting that New Zealand bas several independent skink mitocbondrial lineages. 
Addition of a New Caledonian Skink 
A New Caledonian representative of the Leiolopisma group was not obtainable for this study. Sequence 
information was obtained bowever from another New Caledonian skink, Tropidoscincus rohssii, a 
member of the Eugongylus group to wbicb the Leiolopisnia skinks are related (Greer 1970). Only 300 
bases of 12S rRNA could be reliably determined for T. rohssii so this taxon was excluded from the 
I 
earlier analyses (Appendix 1). It is useful to include it bere 110wever. Since HADTREE can currently 
only bandle up to 20 taxa, addition of T. rohssii required the removal of another taxon. L. microlepis 
was dropped since it is very similar to L. smithi. The frrst 8d bases, wbicb include the variable belices 26 
and 29, as well as the single-stranded region linking domains iI and III (Fig. 3.3), are omitted from the 
data set wben T. rohssii is included. Tbis reduction appears tei remove all pbylogenetically informative 
signals (Fig. 5.19a), but separate analysis of the first 80 bases sbows that this region bas low resolution 
on it's own (Fig. 5. 19b). This latter point was also sbo� by the secondary structure comparisons of 
both closely and more distantly related taxa in Chapter Four. Most of the variable sites occur towards 
the other end of the 1II0iecuie (Figs. 3.7 & 4.17). Omission of L. micro Ie pis also removes one of the 
strongest signals (see Fig. 5.1) .  
Fig. 5.17. Spectrum and optimal tree for 10 southern New Zealand Leioiopisma, and the Australian and 
Mauritian species. In contrast to the spectra for the northern taxa (Figs. 5.15, 5.16), the internal branches 
of the optimal tree for southern taxa have much less support Note also that some of the strongest 
bipartitions included in the tree are for groups of five taxa rather than for pairs of taxa. 
ID�l------- ------- ------ --------- ----------- ------ ---------------- ----1 
• 
.J 
o� 
7J 
I 
'1 
i 1S....l 
I 
� ! 
I 'I 
1 s� :J 
o I 
6 
4 
10  :-
�i 9 _  [0 
""-i - y� 'il 
711 S as loes 1"' II 
1. St. Is. Green (1) 
2. L. grande (2) 
3. L. notosaurus (4) 
4. L. lin/chi (8) 
5. L. n. nigriplantare (16) 
6. La. guichenoti (32) 
10 
-.c......1--- 12 
8 
11 
2 
7. L. acrinasum (64) 
8. L. inconspicuum (128) 
9. L. maccanni (256) 
10. L. te/fairi (512) 
11 .  L. n. polychroma (1024) 
12. L. otagense (2048) 
Fig. 5.18a. Inferred phylogenetic relationships for Leiolopisma skinks found in northern New 
Zealand. The Australian Lampropho/is guichenoti and the Mauritian L. telfairi are also included. 
Branch lengths are proportional to the probability of a nucleotide substitution along that branch. 
L. smithi 
L. telfairi 
C. aenea 
L. zelandicum . moco 
L. lin/chl 
L.fallai L. suten L. n. nigriplantare 
. guichenod 
L. infrapunctatum 
L. n. polychroma 
Fig. 5.18b. Inferred phylogenetic relationships for Leiolopisma skinks found in southern New 
Zealand. The Australian Lampropholis guichenoti and the Mauritian L. telfairi are also included. 
Branch lengths are proportional to the probability of a nucleotide substitution along that branch. The 
greater uncertainty about the relationships of the southern taxa compared with the northern group is 
reflected in the relatively shorter internal edges in the tree for the southern group. 
L. telfairi 
;------.L. otagense 
L. inconspicuum 
L. n. polychroma 
OIapter S, page 70. 
Table 5.3. Mitochondrial lineages in skinks. A lineage is defined as taxa whose 1 2S rRNA sequences 
differ by less than 5%. The mean (± standard deviation) bumbets of�ucleotide differences between the 
first member of each lineage and the other lineages are presented in the right-hand column. The mean 
numbers of differences between all lineages are significantly different (Fl l,l20 = 3.25, P < 0.01). 
However if lineage 1 is excluded, numbers of nucleotide differe�ces bewteen the other lineages are not 
significantly different from each other (F 10,1 10 = 1 .9, P > 0.05), indicating that they may have diverged 
I , 
from each other at about the same time. The same resWts are obtained when distances are corrected for 
unobserved changes using the Jukes & Cantor (1969) correction. Using other taxa within a lineage for 
the comparisons also bad little effect. 
Lineage & Taxa 
1 .  L. graru1e 
St Is. Green 
L suteri 
2. L zelandicum 
3.  L inconspicuum 
L. notosaurus 
4. L acrinasum 
5. L. smithi 
L. microlepis 
6. L.fallai 
C. aenea 
7. L. maccanni 
8. L. otagense 
9. L. telfairi 
10. L. moco 
1 1. L injrapunctatum 
La. guichenoti 
12. L n. polychroma 
L. n. nigriplantare 
L. lin.lchl. 
Inter-Lineage 
DifterenceS 
(mean T sid) 
22.6 ± 3.3 
25.Q ± 3.5 
25.2 ± 4.0 
25.5 ± 3.3 
26.0 ± 4.0 
26.4 ± 3.3 
27. 1 ± 3.3 
27.3 ± 3.4 
27.8 ± 2.8 
27.9 ± 2.5 
28.7 ± 2.9 
28.7 ± 3.3 
a. 
1 
b. 
s 
I 
2 
1 
Fig. 5.19a. The effect of removing the flrst 80 sites of the 12S rRNA sequence from the analysis, before 
adding the New Caledonian Tropidoscincus rohssii sequence. The New Zealand taxon L micro/epis has 
also been excluded from the data set (see text). Note the loss of all strong spectral signals (see Figs. 5 .1 ,  
5.19b). 
l _ lOnUB 1st 80 bIIIII I 
, 8 8 10 12 l' 18 18 20 22 2' 28 28 30 32 3' 38 38 '" .a " " 
Rank of Bipartition 
Fig. S.19b. Spectral analysis of the flrst 80 sites of the 12S rRNA sequence for 19 taxa This spectrum 
indicates that some of the support for the bipartitions c.aenea+L. lallai and L injrapunctatum+La. 
guichenoti is derived from this region of the sequence (see Table 3.1). Use of this region alone however 
provides less phylogenetic resolution than when the full sequence is analyzed (see Fig. 5. 1). 
Cae + Lfa 
.� Lfr + Lag 
2 8 8 10 12 a 
Rank of Bipartitlon 
18 20 
OIapter S, page 71. 
The strongest signal in this comparison links T. Tohssii with L. fallai (Fig. 5.20), but the number of pair­
wise differences for the T. Tohssii sequence, and the tocation of the substitutions linking it with L. fallai 
suggest that this is not a close phylogenetic relationship. T. rohssii has more substitutions (an average of 
32) in pair-wise comparisons than do the 4iolopisma skinfs. For example, L. fallai and L. 
infrapunctatum have for the same region of 300 bases on average 18.3 and 20.6 differences, 
respectively, from other taxa. Furthermore, the bipartition containing T. rohssii and L. fallai is suppoited 
by two observed substitutions, both of which occur in helix 46 (positions 346 & 364), which is one of 
the most variable regions in the molecule (Fig. 4.17). 
Cytochrome b Sequence Data 
Sequence information (254 nucleotides) from the mitochondrial cytochrome b gene was obtained for 
four skink taxa. This small data set has 13  informative (parsimony) sites and 40 singleton changes 
(Appendix 1). Eighty nine percent of the substitutions occur at third codon positions. As noted in , I 
Chapter Two, the 12S rRNA sequences of L. n. polychroma �dividuals from G?rge Bum are identical 
I 
to those of L. maccanni from the same location but are quite different from sequences obtained from L. 
n. polychroma at other sites. The Gorge Burn populations Of L. n. polychroma and L. maccanni differ by 
three transitions in their cytochrome b sequences, all at third codon positions, and are quite distinct from 
the Twizel L. n. polychroma sequence (Table 5.4). 
The cytochrome b sequence has changed more rapidly than the 12S rRNA sequence (Table 5.4). In 
comparisons between L. n. polychToma (Twizel) and L. I'IU1{:canni there are about 50% more 
substitutions in the cytochrome b than in the 12S rRNA sequence, even though a shorter length of the 
former was examined. The relative lengths of the branch!!S leading to the four taxa are similar for both 
the 12S rRNA and cytochrome b sequence data (Fig. 5.21), which suggests that there is nothing atypical 
in the relative rate of change of the 12S rRNA sequence wi� �espect to other skink mitochondrial 
I 
sequences. Consequently, the more well supported relationships determined from the 12S data (for 
example L. microlepis+L. smithi, La. guichenoti+L. infrapunctatum, St Is. Green+L. grande, and L. n. 
polychroma+L. n. nigTiplantare; figures 5.15, 5.16 & 5.17) are expected to be confirmed by other 
mitochondrial sequence data. The very close sequence similarity between the Gorge Bum populations of 
L. n. polychToma and L. maccanni will be discussed iri the next chapter. 
DISCUSSION 
The skink 12S rRNA sequence data set has proven to be v�ry useful for demonstrating aspects of the 
HADlREE algorithm, particularly with respect to the iuql�g of conflicting signals (Figs. 5.5 & 5. 16). 
Spectral analysis is a direct and easy way for assessing conflicts and noise i� sequences and for 
determining alternative sets of relationships. Unl�e boots�pping (Felsenstein 1985) or analysis of the 
distribution of tree lengths (Fitch 1984, Hillis 1991 ,  IGUlersjo et al. 1992), spectral analysis can directly 
show where the conflicting signals in the data set come from. 
I 
Fig. 5.20. Spectral analysis of 304 base pairs of 12S rRNA for 19 Leioiopisma and the New Caledonian 
species T. rohssii. The only strongly supported bipartition is that grouping the New Zealand L. fallai 
with T. rohssii, but see the text. 
5�-------------------------------------------------------------------.
4-
3-
2-
1-
L. fallai + T. rohssii 
1 3 5 7 8 11 13 15 17 18 21 23 25 27 28 31 33 35 37 38 41 4S 45 .7 
2 4 8 8 10 12 a lIS 18 20 22 24 28 28 30 32 3. 38 38 .a � U 48 .a 
Rank of Bipartition 
. 
OIapter S, page 72. 
Table 5.4. Numbers of transitions (above diagonal) and transversions (below diagonal) for 254 base 
pairs of the cytochrome b gene. The 12S rRNA distance matrix for the same four taxa (383 bases) is also 
shown. L. n. polychroma (Lnp-GB), L. maccanni (Lmac), and L. inconspicuum (Linc) are sympatric at 
Gorge Burn, Southland, while the other L. n. polychroma (Lnp-Tw) population is from Twizel, 
Canterbury. 
Cytochrome b Transversions\Transitions 
Lnp-Tw Lnp-GB Lmae Line 
Lnp-Tw 22 25 29 
Lnp-GB 14 3 25 
Lmae 14 0 24 
Line 12 6 6 
12S rRNA Transversions\Transitions 
Lnp-Tw Lnp-GB Lmae Line 
Lnp-Tw 17 17 24 
Lnp-GB 3 0 25 
Lmae 3 0 25 
Line 5 4 4 
Fig. 5.21. Inferred phylogenetic relationships of four skink taxa from Southland. Trees derived from 
383 nucleotides of 12S rRNA and 254 nucleotides of cytochrome b sequence. The L. n. polychroma 
sequences come from two populations: Twizel and Gorge Bum. L. maccanni and L. inconspicuum are 
sympatric with the Gorge Bum population of L. n. polychroma (patterson & Daugherty 1990). The 
resulting tree when both the 12S rRNA and cytochrome b sequences are combined for analysis is also 
shown. Trees are unrooted and branch lengths are proportional to the number of changes expected to 
occur down that branch. The 12S rRNA sequences of L. n. polychroma (Gorge Bum) and L. 
maccanni are identical. 
L. n. polychroma 
(Gorge Burn) 
L. maccanni 
12S rRNA 
L. n. polychroma 
(Twizel) 
L. inconspicuum 
Cytochrome b 
L. n. polychroma 
>----c.. (Gorge Burn) 
L. maccanni 
Combined seqyences 
L. n. po/ychroma 
(Twizel) 
L. inconspicuum 
L. n. polychroma 
>-----c.. (Gorge Burn) 
. maccanni 
Olapter 5, page 73. 
Forcing specific bipartitions to be included in the tree (Fig. 5.16) is one way of determining the 
likelihood of alternative groupings of taxa Continued developIilent and refinement of the programs are 
expanding their utility, enabling more detailed questioris to De asked of the data. Of particular use will 
be the ability to obtain and compare suboptimal trees so that potential phylogenetic relationships can be 
analyzed in more detail. 
Domain ill of the skink 12S rRNA is not able to resolve relationships among many of the taxa examined 
here. This portion of the 12S rRNA gene has been useful for investigating and resolving phylogenetic 
relationships for a wide range of animal groups (for example, Simon et al. 1991 , Hedges et al. 1991, 
Hillis & Dixon 1991, Cooper et al. 1992) so it  was a reasonable premise that it would be suitable for 
skinks as well. Analysis of the secondary structure (Chapter Three) demonstrated that the sidDle 
sequence is evolving in a similar way to other vertebrqtes. Lack of phyiogenetic resolution is therefore 
not attributable to an unusual pattern of molecular evolqtion. Nor is it a result of having too few variable 
sites - the numbers of differences among the skinks are s�i1at to those occurring in other 12S rRNA 
data sets (Table 4.5a-c). 
Removal of constant columns from the HADTREE fr�uency input file is a preliminary, and in some 
ways crude, approximation to addressing the problem of invariant (constant) sites. It serves to illustrate 
however that most of the spectral signals, particularly those �cluded in the optimal tree, are robust and 
not artifacts of analysis. While it is important to develop more realistic models to take account of 
unchanging sites, assuming all sites are free to vary does not appear to be cause for poor resolution of 
the skink relationships. 
Four sets of relationships in the skink data set seem robust, in that they have strong spectral signals 
and/or the taxa consistently grouped together (Figs. 5.5, 5.15, 5.16, & 5.17). These taxa are L 
microlepis+L smithi (which differ by 5 nUcleotide subsqtutions), L grande+StIs. Green (S differences), 
L inJrapunctatum+La. guichenoti (12 differences), and L. n. polychroma+L. n. nigriplantare (13 
differences; Figs. 5.5, 5 . 16, & 5.17). These are all supported by allozyme analyses (C.H. Daugherty & 
G.B. Patterson pers. COmm.; and see Chapter Six). L i1lConspicuum, L. notosaurus, and L. maccanni 
tend to cluster together (Figs. 5.5 & 5.17; though see Chaptet Six). As already noted, the relationship 
between L.fallai and C. aenea (17 nucleotide differehces) is less certain (Figs. 5.15 & 5.16). The 
number and location of shared nucleotide substitutions for these two taxa suggest however that C. aenea 
may have closer phylogenetic relationships to L. fallai than to L telfairi. Additional sequence data is 
I '  
required to investigate this. 
The spectral signals and the short internal edge lengths for comparisons among the southern taxa (Fig. 
5.1Sb) suggest that they have had a more complex evolutionary history than the northern species (Figs 
5. 15, 5.16 & 5.1Sa). Furthermore, the distinctiveness of L. n. polychroma (widely distributed in New 
Zealand) and L. n. nigriplantare (from the Chatham Islands) from the northern groups (Fig. 5. 14) 
suggest a southern origin for them. 
Cllapter 5, page 74. 
"Rapid" Divergence of LeiolopislTUl 
The most significant features of the skink data set are that most of the taxa have similar numbers of I I 
substitutions (Tables 4.1a, & 5.5), and a resolved phylogeny for all of them is not produced. The model 
of evolution used in the Hadamard conjuagation has thtee parts; a tree, a mechanism of change, and 
edge lengths (probabilities of nucleotide substitutions) for the branches of the tree. The mechanism of 
change assumed all sites in the sequence were free to vary, but this is incorrect (Fig. 3.3). However 
removal of constant columns, to simulate the presence of invariant sites, did not change the sequence 
spectrum (Fig. 5 .13) and suggests that the assumption was not misleading analyses of the sequence data. 
Furthermore, the Hadamard conjugation does produce a resolved tree for the ratite 12S rRNA data set 
(Fig. 5.2a) under the same assumption of all sites are free to vary. The mechanism of change assumed by 
the model does not therefore account for in.complete resolution of tbe skink sequence data. 
I 
Analyses of simulated sequence data (Chapter Four and Fig. 5.13) gave very similar results to those I ! 
obtained for the skinks. The simulated sequences were derived using a star-tree model, whereas the tree 
assumed by the model in the Hadamard conjugation is a binary, or bifurcating, tree (see Chapter Four). 
The inability to resolve the skink phylogeny therefore appe� related to how the taxa diversified. Rather 
than a slow diversification over time, the Leiolopisma sldnks examined here appear to have diversified 
from each other more rapidly. The bovid data set also has a sequence spectrum (Fig. 5.2b) similar to the 
sldnks (Fig. 5.1), and it also failed to produce a completely resolved phylogeny (see also Kraus & 
Miyamoto 1991, Allard et al. 1992). With support from fossil evidence, Allard et al. (1992) suggested 
rapid diversification of the pecoran bovids over a 5 miIiion year period. 
There are no fossils for the sldnks to directly place a �e on their origin or diversification (Carrol 1969, 
Molnar 1991). As discussed in Chapter Four, the rate of seql!e�ce evolution in the sldnks may be slower 
than that in the bovids, so "rapid" diversification of the sldnks may still have occured over millions of 
years. The fact that L. telfairi has similar numbers of differences from New Zealand taxa as these taxa 
have among themselves (Table 5.3) implies that this Mauritian species diverged from them at about the 
same time. 
L. telfairi has two unique insertion/deletion events, bo� involving single bases (Table 3.1). Single-base 
insertion/deletion events do not necessarily indicate greater genetic divergence however since such 
indels occur between closely related ratites and among the bovids (see Appendix 2). Immunological 
studies also support approximately equal genetic divergence between the Leioiopisma (Hutchinson et ai. 
1990). The New Caledonian T. rohssii has a greater number of substitutions relative to all the 
Leioiopisma however, so it probably represents an earlier split in the Eugongyius group of sldnks. 
Estimating Times of Divergence for the Skinks 
Analyses of immunological (Baverstock & Donnellan 1990, Hutchinson et ai. 1990), allozyme 
(Daugherty et al. 1990, C.H. Daugherty & G.B. Patterson pers. comm.; and see Chapter Six), and now 
OIapter 5, page 75. 
sequence data all support an older origin for New Zealand skinks than suggested by morphological 
studies (five million years or so, Hardy 1977, Towns et al. 1985). There is no evidence indicating that 
the skink sequences are changing more rapidly than other vertebrate gtoups (Fig. 3.6) so the degree of 
sequence divergence between the skinks implies periods qf separation in excess of 15  million years 
(Chapter Four, and also Hickson et al. 1992). This sectioQ examines how more precise estimates for 
times of separation can be inferred. 
Estimates from Immunological Data 
As already noted, there are no Leiolopisma fossils to indicate when they originated. The immunological 
and allozyme data could however provide a means for establ�hing a time frame for Leiolopisma 
evolution. In the immunological studies of Australian Leiolopisma one New Zealand species, L grande, 
was included (Baverstock & Donnellan 1990, Hutchinson et aI. 1990). Assuming an albumin molecular 
clock, Baverstock & Donnellan tentatively proposed that L. grande may have diverged from Australian 
species 20 million years ago. L grande and the Australian La. guichenoti differ by 21 nucleotide 
substitutions (Table 4.1a) so, using the immunological clock estimate, this would imply approximately 
one nucleotide substitution per million years. Applying tllis rate, most of the New Zealand taxa would 
have diverged between 12 and 35 million years ago. 
The reliability of this immunological data for estimating time is uncertain however. Complete reciprocal 
matrices were not obtained for the albumin immunological differences (AID; Baverstock & Donnellan 
1990) so phylogenetic inferences made from this data may l>e unreuable (Maxson & Maxson 1990). 
AID rates must also be calibrated for each taxon (Thorpe 1982), which has not yet been done for skinks. 
Hutchinson et al. (1990), in a more extensive study, did not use the immunological distances to estimate 
times of separation for Leiolopisma. 
Estimates from Allozyme Data 
Preliminary analyses of the allozyme data for New Zealand skinks do not indicate a rapid divergence of 
the skinks from each other, and La. guichenoti and L. telfairi represent much earlier divergences 
(Daugherty et al. 1990b, C.H. Daugherty & G.B. Patterson pers. comm.). Estimates of divergence time 
based on allozyme data have been made by calibrating genetic distance values (for example Nei's D 
metric value) against an AID clock. Maxson & Maxson (1979) for instance, used the AID of 
salamanders to calculate that a Nei's D value of 1 .0 represented 14-15 million years separation. 
Applying this calibration to the skink allozyme data would indicate a diversification of New Zealand 
Leiolopisma commencing 1 1-12 million years ago (Daugherty et al. 1990b, and C.H. Daugherty pers. 
comm.). If this time frame is used, then if there are about 25 nucleotide differences between New 
Zealand skink taxa (Tables 4.la, & 5.5), there have been approximately 2.1-2.3 nucleotide substitutions 
since divergence 
per million year, L. mlCTolepis and L. smithi, with five <lifferences between them, could therefore have 
diverged just over two million years ago, while the split between L. n. polychroma and L. n. 
nigriplantare (13 differences) may have occurred about six million years ago. These dates disagree 
however with those estimated from the allozyme data (C.H. Daugherty & G.B. Patterson pers. comm., 
and see Chapter Six). 
Olapter 5, page 76. 
As with the immunological data, the accuracy and validity of using Maxson & Maxson's (1979) 
calibration for groups other than their salamanders has not been examined (Thorpe 1982, Nei et al. 
1983) and there is no evidence for a common multilocus electrophoretic clock (Avise & Aquadro 1982). 
Furthermore, in the skink allozyme study only 17 presumptive loci provided information for all the taxa 
(C.H. Daugherty & G.B. Patterson pers. comm.). The accuracy of both the topology and branch lengths 
of phylogenetic trees determined from allozyme data is low when less than 30 loci are used (Nei et aL 
1983, Nei 1987), so phylogenetic inferences based on the skink allozyme data set are likely to be 
unreliable. 
EstimaJesfrom Biogeographic Information 
A third way of estimating divergence times between the � is hy taking account of their geographic 
I ,  
distribution. Australia. New Caledonia. and New Zealand all contain fragments of continental land and 
were once part of Gondwanaland (Mayes et aL 1990, Lawver ef al. 1991). Fiji is also continental in 
origin and another Leiolopisma, L. alazon, occurs there and lias morphological similarities to New 
Zealand taxa, particularly Cyclodina (Zug 1985). Pacific members of the Leiolopisma group of skinks 
therefore all occur on continental fragments. 
L. telfairi is a biogeographic anomaly, in that it is the only member of the Leiolopisma group outside of 
the Pacific. Sub-fossils of another Mauritian skink "Didosaurus" mauritianus have been found, which 
Arnold (1980) considered to be a sister species of L telfairi. Leiolopisma have not been found in Asia or 
Africa Africa rifted from Gondwanaland prior to 140 million years ago, while India separated at least 
75 million years before present (McKenzie & Sclater 1973, Tarling 1988, Lawver et al. 1991). India 
passed near to the current location of Mauritius and the Mascarene �lateau (McKenzie & Sclater 1973), 
so ancestors of L telfairi and L mauritianus may have h�� the potential to be transported there from the 
Australian region. Fragmentation of Gondwanaland abo�� &0 million years ago (Mayes et al. 1990, 
Lawver et al. 1991) and subsequent continental drift could �erefore explain the origin and dispersal of 
Leiolopisma. 
Hypothesis One: A Gondwanan Origin and Diversification of Leiolopisma I 
Both immunological (Hutchinson et al. 1990) and sequence data caD be interpreted in the framework of 
a Gondwanan origin and diversification. A Gondwanan divergence for Leiolopisma implies a slow rate 
of sequence evolution (30 nucleotide substitutions over 80 million years yields a rate of 0.38 per million 
years, or one substitution per 2.7 million years, on average). This figure is not too disimilar from that for 
the ratites (Table 4.6). Some of the New Zealand Leiolopisma are as distinct from each other as they are 
from L telfairi (Table 5.3, Figs. 5.5 & 5.18). A Gondw'!llan origin therefore implies that several distinct 
and old skink lineages inhabit New Zealand, as shown in figure 5.22a 
Under this hypothesis 12S rRNA sequence data from New Caledonian Leiolopisma and the Fijian L. 
alazon should reveal similar numbers of sequence differences and patterns of change as the New 
Zealand, Australian and Mauritian taxa Further investigations of New Caledonian skinks (there are 10 
Fig. 5.22. Hypotheses for the origin of Lei% pisma skinks. A) A Gondwanan origin of the taxa implies that 
their present distribution is due to the breakup of Gondwanaland during the Creataceous (approximately 80 
million years ago). B) The "Oligocene drowning" hypothesis suggests that most of New Zealand's extant skink 
fauna are the result of diversification after the Oligocene (36-25 million years ago), in response to increasing 
land area and habitat diversity (see the text). C) The "Multiple colonists" scenario indicates that diversification 
of the skinks occurred sometime after the fragmentation of Gondwanaland, and that several distinct groups of 
skinks reached New Zealand independently and still survive here. These hypotheses also have relevance to other 
Pacific and New Zealand biota as discussed in the text 
A. Gondwanan diversification 
Australia 
Lord Howe Is. 
Mauritius ----... New Zealand 
Fiji 
New Caledonia 
B. QIi�ocene drownin� 
Australia 
Mauritius --------4�-----.. ;.-.- New Zealand 
New Caledonia 
C. Multiple colonists 
Australia 
Mauritius -----.-------- New Zealand 
New Caledonia 
Olapter 5, page 77. 
genera; Sadlier 1986) and other members of the Eugongylus group of skinks may shed light on the 
hypothesis of a general radiation of this group about the time that Gondwanaland fragmented. 
Immunological studies suggest that Australian members of the Leiolopisma group are not monophyletic 
but are derived from a variety of other genera (Baverstock & Donnellan 1990, Hutchinson et al. 1990). 
The morphological taxonomy of the Leiolopisma group has peen misleading (Towns et al. 1985, 
Daugherty et al. 1990b) so L. telfairi could be derived from a different source than the Pacific group, 
possibly an Asian Mabuya species, the ancestral group for Leiolopisma (Greer 1970). Sequence and 
I 
immunological data indicate however that there are still close phylogenetic relationships between L. 
telfairi and the Pacific Leiolopisma. A close relationship among the Leiolopisma group of skinks is also 
supported by comparison of the relationships of the New Caledonian skink. T. rohssii has more 
sequence differences from any of the Leiolopisma than the �tter have among themselves, and T. 
rohssii's sequence similarities to L.fallai (Fig. 5.19) ate in � variable region so may be coincidental. L 
telfairi on the other hand shares changes with C. aenea and other Leiolopisma at more conservative sites 
in the 12S rRNA molecule (Table 3. 1), which suggests that L telfairi does have close phylogenetic 
relationships to the Pacific Leiolopisma. 
Other animal groups with representatives in New Zealand have been considered to have had an origin in 
Gondwanaland. For example, the Leiopelmid frogs (BUll & Whitaker 1975), the tuatara (Crook 1975), 
and the moa (Cooper et al. 1992). The distribution of some of New Zealand's insect fauna may also be 
accounted for by a Gondwanan origin (Dugdale 1989). A Gondwanan origin for geckos has also been 
suggested on the basis of karyological evidence (King 1979, 1987, King & Mengden 1990). Bauer 
(1990), on morphological biogeographical criteria, also advocates a Gondwanan origin for some of the 
geckos. The Leiolopisma skinks may therefore be just one of several groups of old New Zealand 
inhabitants. 
The Mauritian Leiolopisma are Still Enigmatic 
The Mascarene islands, to which Mauritius belollgs, are volcanic and were formed 7-8 million years ago 
(McDougall & Cbamalaun 1969). No Leiolopisma-like skinks have been recorded from India or 
Madagascar, 700 kilometers west of Mauritius. India may also have separated from Gondwanaland over 
100 million years ago (Lawver et al. 1991) and Leiolopisma may not have been present then. Even 
under a Gondwanan origin hypothesis, the occurrence of L. telfairi on Mauritius is still enigmatic. It is 
possible that the ancestral group of L. telfairi and L. mauritianus had a post-Gondwanan origin in the 
Pacific region and subsequently reached Mauritius by its own dispersal abilities. This is in effect a null 
hypothesis, to be retained if additional sequence and biochemical data cannot establish an independent 
and old origin for L. te lfairi. 
Origins of the New Zealand Skinks 
Even if there was a Gondwanan origin for the skinks, New Zealand may have been colonized by them 
after separation fro� Gondwanaland. If this was the case there are two alternatives for the current 
diversity of them in New Zealand. New Ze
a
Iand Leiolopisma may be a monophyletic group which 
Cliapter S, page 78. 
diversified in New Zealand from a common ancestor (Fig. 5.22b). Or. several groups of skinks may 
have reached and colonized New Zealand independently (Fig. 5.22c). Both of these hypotheses have 
some support from analyses of the sequence or allozyme data. 
Diversification Within New Zealand 
As the water barrier between New Zealand and Australla and New Catedonia increased following 
separation from Gondwanaland, the chances of colonization of New Zealand by flightless terrestrial 
animals would diminish. Many independent colon�tions of New Zealand by skinks would therefore 
seem unlikely. As discussed above. the occurrence of L. telfairi on MaUritius may not be due to 
continental drift. The New Zealand skinks may therefore hav� a more recent diversification. Analyses of 
allozyme data for New Zealand Leiolopisma by C.H. Daugberty & G.B. Patterson (pers. comm.) suggest 
that they have diversified in New Zealand, and are riot as closely related to L. telfairi and La. guichenoti 
as the 12S rRNA sequence analyses indicate. Limitations of both the allozyme and sequence data will be 
discussed below. but a monophyletic origin for most New Zealand Leiolopisma cannot yet be excluded. 
Comparison of the 12S rRNA sequence differences among the skinks to those found among other 
I 
vertebrates indicates that the New Zealand Leiolopisma diverged from each other at least 15 million 
years ago. and probably much earlier (fable 4.6). The timing of this diversification may be related to 
major changes in New Zealand's geography during and after the Oligocene (36-25 million years ago). 
Hypothesis Two: Post-Oligocene Diversification of New Zeaiaild Leiolopisma 
New Zealand was a low-lying landmass in -the Oligocene ll!ld rising sea levels created a series of islands. 
Total land area was 80% less than at present and fluctuating �a Jevels altered the number and size of 
these islands (Suggate et al. 1978, AJ. Cooper, G.K. Chambers, & R.A. Cooper in prep.). Cooper et aI. 
proposed that this may have resulted in population "bottlenecks" and extinctions (Fig. 5.22b). Post­
Oligocene increases in land area and elevation would have led to the creation of many new habitats and 
consequent diversification among terrestrial biota. 
If the New Zealand lineages (fable 5.3) are the result of Miocene (i5-25 million years ago) 
diversification from one or a few taxa which survived the Opgoceile (Fig. 5.22b). then it is expected that I 
they would have fewer differences between themselves than they would to L. infrapunctatum and L. 
telfairi. This is not the case (fable 5.3). 
A post-Oligocene radiation has been suggested for both �e moa and the kiwi (Cooper et al. in prep.). 
The number of observed nucleotide differences between the moa genera and between the kiwi species 
are about 1 1  and 15, respectively (Table 4.5a), substantially fewer than the 25 differences between the 
skink taxa. A Miocene diversification of the skinks would therefore ihdicate a more rapid rate of 
sequence evolution in the skinks than in the ratites. Post-Oligocene speciation may account for the 
diversity of other N�w Zealand fauna, such as Leiopelma frogs (Daugherty et al. 1982, and CH. 
Daugherty pers. comm.). 
OIapter 5, page 79. 
Population bottlenecks will decrease the number of mitocbondrial lineages, but under the neutral 
framework of molecular evolution, it should not affect the rate of sequence divergence (Kimura 1983, 
Wilson et al. 1985, Neigel & Avise 1986). Under a "nearly neutral" hypothesis however, population size 
could affect the rate of molecular evolution (Ohta 1976, 1992). Analyses of Drosophila lineages 
suggested that groups which probably passed through repeated popu1ation bottlenecks had a higher rate 
of molecular evolution than lineages which had more stable popuiation sizes (DeSalle & Templeton 
1988). However, the dates of separation of these Drosophila species have not yet been reliably 
established, nor bas the fact that some did pass through severe bottlenecks. With respect to the 
Leiolopisma, even if bottlenecks can affect rates of molecuJat evolution, this does not alter the 
conclusion from the sequence data that the skinks appear to have diversified relatively rapidly, and that 
the Australian and Mauritian taxa are not distinct from the New Zealand taxa 
Phylogenetic Limitations of the Skink Allozyme Data 
The skink allozyme data (Daugherty et al. 1990b, C.H. Daug!lerty & G.B. Patterson pers. comm.) could 
indicate a monophyletic origin of New Zealand Leiolopisma and Cyc1odina. The number of 
electrophoretic loci (seventeen) is too few however to reliably infer phylogenetic relationships. Ideally 
at least 30 loci should be used (Nei et al. 1983). More rigourous analyses (such as parsimony analysis, 
Swofford & Berlocher 1987) of the data have yet to be performed. The measure of genetic distance used 
by Daugherty et aI. (1990b), Nei's unbiased D value (see Nei 1987), relies on the assumption of equal 
rates of substitution at all iocl, a condition unlikely to be met (Swofford & Olsen 1990), and this 
measure may therefore not be an accurate estimate of genetic divergence between taxa A modified 
version of Nei's genetic distance (Hillis 1984), or Cavalli-Sforza & Edwards' (1967) distance measure 
may be more appropriate (Swofford & Olsen 1990), though the limited number of loci must still be 
borne in mind. Limitations of the sequence data will be discussed in Chapter Six. A post-Oligocene 
diversification of the New Zealand Lei% pisma (Fig. 5.22b) does not account for the 12S rRNA 
sequence similarity within and between the New Zealand and the Australian and Mauritian taxa 
Hypothesis Three: Multiple Leiolopisma Colonizations of New Zealand 
Similar genetic distances between the New Zealand skink �a do not by themselves indicate a Miocene 
diversification in New Zealand. While the skinks may have 4iversified from each other relatively 
rapidly, different groups could have reached New Zealand independently. L. infrapunctatum and La. 
guichenoti seem most closely related to each other on the basis of sequence similarity (Table 5.3, Figs. 
5. 1 & 5.5, 5 .15, 5.16). If a Gondwanan origin for the oilier skink lineages is assumed, then 25-30 
nucleotide differences implies a rate of 0.3-0.4 substitutions per million years. (The error associated with 
these estimates is likely to be high, as discussed in Chapter Four). The 12 differences observed between 
L. infrapunctatum and La. guichenoti 12S rRNA sequence could indicate a separation of 20-30 million 
years. If this is the case then the L. infrapunctatum lineage either entered New Zealand more recently 
than other taxa, or n:tembers of it dispersed west to Australia The most likely route into, or out of, New 
Zealand would be via New Caledonia and the Norfolk Island ridge, which was once part of the ancestral 
Cliapter S, page 80. 
New Zealand landmass (Suggate et al. 1978, Stevens lQ80). Land connections along this route 
disappeared by the Miocene (approximately 30 million ye� ago; Stevens 1980), which could place an 
upper bound on time of entry for many flightless anim�s. Estimates for the divergence of the ancestors 
of L. infrapunctatum and La. guichenoti are compatible )Vitb a pre-Miocene entry. 
An independent arrival of L. infrapunctatum in New Zealand is still compatible with both a Gondwanan 
diversification of Leiolopisma (Fig. 5.22a), as well as diverstflcation of other Leiolopisma in New 
Zealand (Fig. 5.22b). Under the multiple colonists hypothesis the rtlnnber of different mitochondrial 
lineages of Leiolopisma (Table 5.3) requires many multiple independent colonizations of New Zealand 
I 
by skinks. Considering the distances and oceans to cross, this is perhaps less likely than the other two 
hypotheses. 
Hypotheses for the testing 
The phylogenetic relationships of the New Zealand Leiolopisma are largely unresolved by the current 
sequence data. but they support suggestions (Towns et al. i985, Daugherty et al. 1990) of a much longer 
history in New Zealand than indicated by morphological similari�es. Analysis of the sequence data also 
suggests a rapid divergence of Leiolopisma skinks. Whether this divergence reflects a Gondwanan or 
more recent evolution requires additional sequence informaqon. Strategies to test the three hypotheses 
are discussed in Chapter Seven. Regardless of which bypo�esis receives the most support, the 
evolutionary history of the skinks, and probably other New Zeaiand hiota. are much more complex, and 
hence interesting, than previously imagined and offer many opportUnities for more detailed 
investigations of evolutionary processes. 
page 81. 
Chapter Six: Comparing Sequence and Allozyme Data 
One of the primary reasons for selecting the skinks to address evolutionary questions in New Zealand 
was that allozyme information was available for virtually all of the described taxa, as well for taxa yet to 
be formally named (Towns et ai. 1985, Patterson & Daugherty 1990, C.H. Daugherty & G.B. Patterson 
pers. comm.). DNA sequence information was obtained from individuals which contributed to the 
allozyme study so direct comparison bewteen the two �ta Sets is possible. In this chapter the 
relationships of the New Zealand Leioiopisma skinks as inferred from the sequence data are compared 
with the allozyme divergence data. The data sets do not contain all �e same taxa so only the 16 New 
Zealand Leioiopisma common to both are compared. 
Preliminary results of the allozyme study have been published (Daugherty et ai. 1990b), but manuscripts . 
containing the full analyses are in preparation (C. H. Oaugherty & G. B.  Patterson pers. comm.). 
Allozyme information could only be obtained from 17 presumptive loci for the sldOO, though most taxa 
were distinguished by ftxed allelic differences. The accuracy of both the topology and branch lengths of 
phylogenetic trees determined from allozyme data is low when less than 30 loci are used (Nei et aI. 
1983), so comparisons between the sequence and allozyme data only address the question of how 
similar are their estimates of genetic distance. 
Conflict Between Allozyme and 12S rRNA Sequence Data 
Using Nei' s unbiased D (Nei 1972) as a measure of genetic qistance Daugherty & Patterson (pers. 
comm.), identifted nine sets of relationships with D values less than 0.4. Some of these groupings are 
found in the sequence analyses, but D values do not correlate well with numbers of nucleotide 
differences (Table 6.1). Stewart Island Green and L. gratUie, and L maccanni and L. inconspicuum are 
the most closely related pairs on the basis of allozymes (Table 6.1), but their 12S rRNA sequences differ 
by 8 and 20 substitutions, respectively. In contrast, only 5 nucleotide differences were found between L. 
microlepis and L. smithi but they have a D value of approximately 0.55 (Table 6.1). L. fallai and L. 
suteri, and L. infrapunctatum and L. otagense have no spectral signals linking them together (see 
Chapter Five) despite havi�g relatively low D values (Table 6.1). Both L. telfairi and La. guichenoti 
have D values in excess of 1 .0 when compared to each other and to the New Zealand Leiolopisma (C.H. 
Daugherty & G.B. Patterson pers. comm.), in contrast to their sequence similarities (Table 4.1a). 
Congruence Between Allozyme and 12S rRNA Sequence Data 
The allozyme and sequence data are not in complete disagreement however. Members of the taxon pairs 
L. n. polychroma/L. n. nigriplantare, L. grandeISt. ls. Green, and L. microlepislL.smithi are each others 
closest relatives on the basis of both DNA (Table 4.1a, Fig. 5.5) and allozyme analyses (Table 6. 1 ;  C.H. 
Daugherty & G.B. Patterson pers. comm.). 
Olapter 6, page 82. 
!fable 6.1. Comparisons of genetic distances inferred from 12S rRNA sequence analysis and allozymes 
(C.H. Daugherty & G.B. Patterson pers. comm.). The relationsbips of taxon pairs 1 , 2, and 4 are 
(:ongruent in that both allozyme and sequence information identify the members of the pair as most I ,  - -
liimilar to each other. The relative degree of genetic divergence between taxa differ however when the 
Jlumber of nucleotide differences and Nei's unbiased D vaiue (c.lI. Daugherty & G.B. Patterson pers. 
t;omm.) are compared (Spearman's rank correlation, rs = i>.245, P < 0.05). 
Seque{Jces AUozymes 
No. Unbiased 
Dirt. 
I 
Taxon Pair Rank D Value Rank '.' 
- --
I. L. microlepislL. smithi 5 1 0.55 9 
2. St.ls.GreeniL. grande 8 2 0.175 1.5 
3. L. inconspicuumlL.notosaurus 12 3 0.25 4.5 
4. L.n.nigri.lL.n.polychroma 13 4 0.2 3 
5. L. inconspicuumiL. maccanni 20 5.5 0.175 1 .5 
6. L maccannilL.notosaurus 20 5.5 0.25 4.5 
7. L. fallailL maco 24 7 0.35 7 
8. L. fallailL. suteri 26 8 0.275 6 
9. L. inJrapunctatumlL.otagense 32 9 0.4 8 
OJapter 6, page 83. 
These differences between the sequence and allozyme data may be due to limitations in both the DNA 
and allozyme data sets. The mitochondrial sequence da� reflect maternal relationships only while the 
allozymes provide information on both maternal and paternal inherited genomes (Wilson et al. 1985, 
A vise et al. 1987). Sampling errors due to the short length of DNA sequences and the small number of 
allozyme loci scored may also contribute to the differing c�:mclusions in each data set Collection of 
more mitochondrial and nuclear DNA sequence informa,qon is required to determine the accuracy of the 
I - -
phylogenetic relationships. Differing phylogenies based on mitochondrial and nuclear sequence data 
will indicate whether the mtDNA of Leiolopisma is suitable for reconstructing organismal relationships. 
The allozyme data has only been analyzed using clustetipg al�orithpls (Weighted Pair Group Method 
using Arithmetic means, Sneath and Sokal 1973) so fat (C.H. Daugherty & G.B. Patterson pers. comm.) 
and further analyses (such as parsimony) are required to defcimine the robustness of the relationships 
between the skink taxa (see Swofford and Olsen 1990). 
Hybridization 
Comparison of the allozyme and the mitochondrial sequence information also illustrate the benefits of a 
broad approach to evolutionary studies. A difference betweep the DNA sequences and allozyme profiles 
of L n. polychroma populations suggests that interspecifiC hybridization has occured. 
The identical 12S rRNA and very similar cytochrome b sequences for L. n. polychroma and L maccanni 
at Gorge Burn were presented in the previous chapter (fable 5.4). Several observations show tha this 
sequence similarity was not due to taxonomic mis-identificapon nor to mixing of tissue or DNA 
samples. The allozyme profiles for L. n. polychroma individuals at Gorge Burn are consistent with those 
of L. n. polychroma at other locations, and there appears to be nothing incongruous about the L. 
maccanni allozyme data at Gorge Burn either (Daughetty et al. 1990b). DNA was extracted from 
several individuals and from several tissues at different times for both species. Both PCR amplification 
and DNA sequencing were also performed several times for each sample. Identical DNA sequences 
were obtained in all cases (Hickson et al. 1992). That �e �ytochrome b sequences for L. n. polychroma 
and L. maccanni differ at three positions (Table 5.4) supports the view that the identical l2S rRNA 
sequences in the Gorge Burn populations is not a result of contammation. 
The conflicting relationships for the Gorge Burn popuiation pf L. n. polychroma are shown in figure 6.1 .  
Note that the relationship of L notosaurus presented iIi figUre 6.1 differs from tllose in figures 5.5 and 
5. 18, where more taxa were included. The different placemeht of L. notosaurus is a reflection of the 
small number of shared sequence signals between th� taxa. For example, L. notosaurus and L. 
inconspicuum have fewer nucleotide substitutions when compared to each other than either do when 
compared to L. maccanni (Table 4.1a), which may suggest that the latter species is more distantly 
related to them. In the comparison of six taxa shown in figure 6.1 ,  the bipartition (sites where they both 
differ from the other taxa) of L. notosaurus+L. inconspicuum is supported by two signals, while the L. 
inconspicuum+L. maccanni bipartition has three signals. When more taxa are added L. maccanni, 
Fig. 6.1. Evidence for sexual hybridization in a New Zealand skink population. Twizel and Gorge Bum 
populations of L. n. polychroma have the same allozyme profile (Daugherty et al. 1990b), but quite 
different mtDNA sequences. The discrepancy between the allozyme and sequence data can be reconciled 
by proposing that L. maccanni females have mated with L. n. polychroma males (see text and figure 6.2). 
---
� 
<::> 0'1 0'1 
� 
-: 
� l:::I - -� � 
0 » -� s-
5 � ».c N �  o � - � - 0  < --
� -� 
0 
� c.J 
C � � 
C" � 
en 
< 
Z 
� s-
en M 
� 
� 'i -
� i � j � t::. 
k 
.... Q .;:. ... ... .a:: -.� � ;:: Q 
s:: Q � 
...j s:: 
...j 
� 
{ 
� s ;:: .s 
.� .� ;:: 
s:: 
...j 
I'"'"' 
= 
.... e .., ;:: e ::s ;:: ::s ... 8 ::s ::s ::s � � � � .... .., � � ;:: Q - ;:: .s Q 8 <I> ;:: ...j N .5 ...j ...j ...j 
J - .... 
� J § 
t k § eE - � 
oJ ...j e -,.... ::s e ::s � ::s .... .!: 
S 
� � ;:: ;:: ... 8 .s ::s .5 � � .J .J -
Q ;:: 
...j 
.... -
I 
OIapter 6, page 84. 
because of it's larger number of differences, has more conflicpng signals to other taxa This means that 
the number of signals supporting L inconspicuum+L. maccanni to the exclusion of all other taxa 
decreases, and the relationship between L. notosaurus and L. inconspicuum consequently strengthens 
(Figs. 5.5, 5.18). The phylogenetic position of L. notosaurus does nOt however affect the main 
conclusion shown in figure 6. 1 ;  that 12S rRNA sequences from the Gorge Bum population of L. n. 
rngriplantare are identical to those from the sympatric L. maccanni population. 
Disparity between mitochondrial and nuclear markers is often indicative of sexual hybridization (Wilson 
et al. 1985, Moritz et al. 1987, Barton & Hewitt 1989, Arnolli 1992) and this seems the simplest 
explanation for the Gorge Bum population of L. n. polychroma. The alternative hypothesis that the 
Gorge Bum L. n. polychroma mitochondrial sequences only reflect a different mitochondrial haplotype 
(see Avise et al. 1987, Moritz et al. 1987, Avise 1989a) within the L. n. polychroma species appears 
unlikely. There are 29 differences between the 12S rRNA sciJuences of the Gorge Bum and the Twizel 
populations of L. n. polychroma and 36 differences iIt their cytochrome b sequences (Table 5.4). This 
level of inter-population sequence divergence has not peen reported even for more rapidly evolving 
regions ofmtDNA (Thomas et aI. 1990, Smith & Patton 1991). Numbers of differences in the 12S 
rRNA sequences between species can be as few as three for »oth kiwi (Table 4.5a) and great apes (Table 
4.5d). 
The L n. polychroma individuals from Gorge gum have the L. maccanni type riltDNA and either 
L.maccanni or L. n. polychroma allozyme profiles. No individuals had electromorphs ("alleles") from 
both parents. Since mtDNA is primarily maternally inherited (see Wilson et al. 1985) this suggests that 
one or more female L. maccanni mated with L. n. polych.roma males, introducing the L. maccanni 
mtDNA into the L. n. polychroma gene pool, as shoWn in figure 6.2a. Loss of L. maccanni alleles in the 
hybrids can be accounted for by the female hybrids subsequently mat.iilg with L. n. polychroma males 
for several generations (Fig. 6.2b). When this hybridizappn occurred cannot be accurately determined 
from the sequence data at present However if we assume d Gondwanan origin for the L. n. polychroma 
and L. maccanni lineages (see Chapter Five), then the 39 differences in the cytochrome b sequences of 
L. n. polychroma and L. maccanni (Table 6.1) could equate to 0.5 substitutions per million years 
(39/80). The three differences between the Gorge Bum L. n. polychroma and L. T1Ul;Ccanni populations 
could therefore indicate 6 million years since hybridization. A more recent separation of L. n. 
polychroma and L. maccanni, say 20 million years, wou1d �ply hybridization occured about 1 .5 million 
years ago. 
DISCUSSION 
Lack of congruence between allozyme and sequence data is often attributable to hybridization (see 
Moritz et al. 1987), though it may also suggest variation in rates of change between the two types of 
data (Dowling & Brown 1989). Allozyme data can be more suitable for examining relationships 
amongst closely rel�ted taxa while sequences may be more informative at greater levels of separation 
(Buth 1984, Hillis 1987, Hillis & Moritz 1990). The 12S rRNA sequence data indicates that few of the 
Fig. 6.2a. Potential offspring from sexual hybridization between two species 
with different nuclear (allozyme) and mitochondrial markers.The nuclear 
phenotype is shown by the shading in the large oval while the mitochondrion is 
represented as the smaller oval. The offspring should have the mitochondrial 
type of its mother but nuclear markers reflecting both parents (as indicated by 
cross-hatching). 
L. n. polychroma 
Hvbrid Offspring 
(L fllQccanni mtDNA) 
~ 
My �male 
Hvbrid Offspring 
(L. n. polychrofIIQ mIDNA) 
L. maccanni 
Fig. 6.2b. Effects of subsequent breeding of hybrid with members of the parental 
populations. Mitochondrial type remains unchanged but allozyme profile should 
tend towards the non-hybrid parents markers over several generations. 
L. maccanni 
Hybrid 
�( 
(L. fllQccanni mtDNA) 
/ L. n. polychroma 
L. n. polyclzroma 
Otapter 6, page 85. 
taxa are closely related (Tables 4.1 & 4.2, Fig. 5.5). If, in addition, many of the skinks diverged from 
each other at about the same time (see Chapter Five), Uten these two factors may explain why the 
sequence and allozyme data sets conflict. Nei et al. (1983) in simulation experiments showed that 
relationships derived from fewer than 20 allozyme loci were less likely to be accurate than when more 
than 30 loci were used. Allozyme information from otller presumptive loci could not be obtained for the 
skinks (C.H. Daugherty pers. comm.), so further investigation of the nuclear gen<:me will r�uire other 
strategies. 
mtDNA provides information about maternal ancestors, aliq mfDNA sequence divergence may predate 
speciation so analysis of mtDNA, particularly the conservative 12� rRNA, may fail to identify more 
I 
recent population subdivisions or hybridizations (Avise et al. 1987, Moritz et al. 1987, Arnold 1992). 
This is not the case with the skinks however, the sequence arid aIiozyme data identify differences in all 
the taxa The conflict is in estimates of the degree of genetic divergence. Except for L. microlepis and L. 
smithi, the sequence data tend to show greater divergence betWeen the taxa than the allozymes (Table 
6. 1). Nei's unbiased D value is calculated under the assumption of equal rates of changd>etween 
different loci. This is probably incorrect and a modified version of Nei' s genetic distance (Hillis 1984), 
or Cavalli-Sforza & Edwards' (1967) arc distance measure may be more appropriate (Swofford & Olsen 
1990). 
As discussed in Chapter Five, interpretations of immu�blogicaI data from Leiolopisma suggest that L. 
tel/airi is similarly distant from New Zealand and Australian Leiolopisma (Hutchinson et al. 1990). This 
is not apparent in the allozyme analyses (C.H. Daugherty & G.B. Patterson pers. comm.), and may 
reflect the relatively small number of loci examined. A wider sampling of the mitochondrial genome, 
either by restriction mapping or the sequencing of other genes, would indicate the accuracy of the 12S 
rRNA sequence data Sequencing of nuclear genes, such as the 18S rRNA or intragenic spacer regions 
in the rRNA cluster (Hillis & Dixon 1991), can be used to test the validity of the mtDNA and allozyme­
based phylogenies. 
Both the allozyme and mtDNA sequence data suggest however that the genetic diversity of New 
Zealand Leiolopisma is much greater than their morphological diversity, and the skinks are consequently 
much older than previously (Bull & Whitaker 1975, Hardy 1977) suggested. Similar observations have 
I 
been made for amphibians (Cherry et al. 1978, Roberts & Maxson 1985, Wallis & Arntzen 1991) and 
East African jackals (Wayne et al. 1989). 
Hybridization - Questions to Address 
The role and significance of hybridization in evotution is not well understood, though it has been 
suggested to be a major source of genetic diversity and innovation (Huxley 1938, Dobzhansky 1940, I 
Rattenbury 1962, Barton & Hewitt 1989, Arnold 1992, Grant & Grant 1992). Hybrids are often 
associated with habi_tat disturbance (see Hair 1966, Arnold 1992). The putative hybridization at Gorge 
Bum is therefore an important result and requires more detailed studies. There are several questions to 
address; 
OIapter 6, page 86. 
1). What is the extent of hybridization? Only a few indlviduals front Gorge Bum were examined in this 
study, many more need to be sampled. Do both hybrid and non-hybrid L. n. polychroma co-exist at " 
Gorge Bum? L. inconspicuum is sympatric with L n. polychroma and L. maccanni at several sites 
(Patterson 1985, Daugherty et al. 1990b) and these species and populations should be examined closely 
as well. 
2). The available data implies that there is directional exchange of mtDNA; L. maccanni females have 
mated with L. n. polychroma males. This would implica� behavioural and reproductive factors affecting 
the hybridization (Ferris et aI. 1983, Lamb & Avise 1986). Examination of more individuals from Gorge 
Bum is necessary to determine in hybrids with L. n. polychroma IhtDNA also occur. 
3). The large number of sequence differences between L n. polychroma and L. maccanni (Table 6.1) 
suggests that they have been separated for a considerable period of time, perhaps up to 80 million years 
(Chapter Five), though the allozyme data could reflect a closer relationship between them (C.H. 
Daugherty & G.B. Patterson pers. comm.). Hybridizaqon between animal species is generaliy more 
common among closely related taxa, that is those separated by a few million years (see Barton & Hewitt 
1989, Arnold 1992). Frogs however, have a greater ability to l1ybridize across larger phylogenetic I ' 
distances (tens of millions of years) than mammals (Wilson et aI. 1974). The compositional organization 
of the reptilian nuclear genome ma:y be more similar to amphibians (Bernardi & Bernardi 1991,  Olmo 
1991) than mammals (Bernardi et aI. 1988), and this maY be re�ected in a greater ability to hybridize. 
4). The potential for skinks species to interbreed, and the viability and reproductive success of hybrid 
offspring, can be investigated by captive breeding, as � been done for species of Drosophila (Aubert 
& Solignac 1990). Close study of hybrids may also indicate if there is heterozygote advantage or 
disadvantage (see for example Paige et al. 1991,  Grant & Grant 1992). lf it is shown that L. n. 
polychroma, L. maccanni and L. inconspicI:tum can interbreed and produce viable offspring, despite their 
large genetic differences, then strong ecological and/or pe�vioural structuring between the species may 
have prevented frequent interbreeding. A geological, cliqlatic or other disturbance may therefore have 
been the cause of hybridization at Gorge Bum. The preliIpi� calculations in this chapter suggest 
hybridization may have occurred in the Pliocene, up to six million years ago. The hybridization may 
have represented a flISt meeting of L. maccanni and L. n. po!ychroma at that site. 
Other New Zealand skink species are also taxonomically proplematical, for example, L. waimatense, 
and the L. lineoocellatumiL chloronotum complex (Pickard & 'towns 1988; C.H. Daugherty pers. 
comm.). These may also provide examples of hybridization. 
Hybridization has been suggested for many groups in New Zealand, both plant (Cockayne 191 1 ,  Allan 
1961, Rattenbury 1962, Fisher 1965, Webby et aI. 1987, Wardle 1988) and animal (Powell 1949, 
Bigelow 1965, Batc�elar & McLennan 1977, Climo 1978, Solem et al. 1981, Hitchmough et al. 1990). 
With the exception of Hitchmough et aI. (1990), these have not been investigated genetically. Hybrids in 
OJapter 6, page 87. 
kiwi and in shearwaters have also recently been detected by DNA analyses (AJ. Cooper pers. comm.), 
so hybridization of both plants and animals in New Zealand may be relatively common. 
The ability to extract sufficient DNA for PCR analyseS frpm small tissue samples makes it feasible to 
screen large numbers of individuals without the rieed to kill �em. Removal of a portion of the skinks' 
tail should not adversely affect it's survival. The development of PCR primers suitable for examination - -
of both conserved and variable regions ofnucleat rRNA (Hillis & Dixson 1991) will facilitate 
comparison of mitochondrial and nuclear derived phylogenies, as well as permit more detailed studies of 
the nuclear genome and the processes of evolution. 
Species Concepts 
Does the hybridization between L. n. polychroma and L. maccanni imply that they are not distinct 
species? The difficulty in distinguishing these taxa. and L. inconspicuum (patterson & Daugherty 1990), 
may imply that there is just one variable species. On the basis of allozymes however, they appear to be 
reproductively isolated (Daugherty et al. 1990b). Genetic distance measures cannot be lise(! as the sole 
criterion for identifying species (see Nei 1987, Coyne 1992), but the fact that L. n. polychroma, L. 
maccanni, and L. inconspicuum occur sympatrically and have distinct allozyme proflles provides strong 
evidence that they do not frequently interbreed. They are also ecologically differentiated (Patterson 
1985). The number of differences between their 12S rRNA sequences (Tables 4. 1a, and 5.3) are also 
very high (compare to the kiwi (Table 4.5a) and the apes <Table 4.5d), supporting their taxonomic 
distinctiveness. 
Origins of Some Specific New Zealand LeiolopislTUl 
L. n. nigriplanJare is found only on the Chatham islands, 800 km east of Christchurch. The Chathams 
have long been separated from mainland New Zealand, though there is evidence for more extensive land 
between them in the Eocene (45 MY A; Suggate et al. 1978, Hay et al. 1970, Craw 1988). The land area 
of the Chatham Islands has been greatly reduced ill the past by rises in sea-level (Hay et al. 1970) and 
consequently some of its biota may have arrived there in more recent times (see Fleming 1980, Lloyd 
1982, Craw 1988). There are 13  nucleotide differences between L. n. nigriplantare and L. n. 
polychroma. If it is assumed that skinks with 25-30 differences diverged during the. fragmentation of 
Gondwanaland (about 80 MY A; see Chapter Five), then the nucleotide substitution rate is 0.31-0.38 
substitutions/site/million years. This would plaCe the time of qivergence between the lineages leading to 
L. n. nigriplanJare and L. n. polychroma at between 34 and 42 million years, which may coincide with 
the Eocene land route. A Miocene origin for New Zealand skinks (Fig. 5 .22b), say 20 MY A, could 
indicate a separation of 9-10 million years between the Chatham Island and mainland skinks. 
L. suteri is the only egg-laying skink in New Zealand (Whif4ke� 1968, Towns 1975b) and the lack of a 
close relationship between it and La. guichenoti, which also lays eggs, indicates that this is a 
physiological trait which can be gained or lost independently. The predominance of ovovivipary 
(bearing live young) in New Zealand skinks was considered to Pc an adaptation to cold periods, such as 
Chapter 6, page 88. 
occurred during the Pleistocene (Whitaker 1968, Towns 1975a). L. suteri is now found in the north of 
the North Island, and it has been suggested (Whitaker 1968, Towns 1975a) that it retained the egg­
laying trait because it did not experience the same cold Pleistocene glacial climates as other New 
Zealand skinks. Many other skinks are found in the far north of New Zealand however (pickard & 
Towns 1988) but do not lay eggs. Analyses of both allozyme (C.H. Daugherty & G.B. Patterson pers. 
comm.) and the 12S rRNA sequence data (Figs. 5.5, 5 .18) indicate thatL. suteri may have relatively 
close relationships to some of the ovoviviparous New Zealand taxa. Egg-laying by L. suteri may (in 
some as yet unknown way) be associated with it's adaptation to living near the intertidal zone (Towns 
1975a). La. guichenoti is a widespread and common s� in garden� of south-east Australia (Cogger 
1986), and does not inhabit a similar niche as L. suteri. 
Some skink groups are reported to have good dispersal ability because apparently closely related taxa 
can be widely separated (Greer 1974), but there is circularity in this argument since the morphological 
similarities are assumed to imply recent origins. The present study, and those based on allozymes 
(Daugherty et al. 1990b, C.H. Daugherty & G.B. Patterson pers. comm.) and immunologjc� affinities 
(Baverstock & Donnellan 1990, Hutchinson et al. 1990) have demonstrated morphological similarity to 
be a poor guide to genetic relationships in the Leiolopisma group of skinks. Lack of appreciation of the 
taxonomic status and antiquity of species can lead to inappropriate species management and 
consequently, extinction (Avise 1989b, Daugherty et al. 1990a). Most of the New Zealand skinks have 
restricted distributions, and are rare or endangered (Pickard & Towns 1988). Many are restricted to 
offshore island habitats. Rats, cats and habitat destruction have contributed to their decline (Whitaker 
1978, Thomas 1982b). Information about skink evolutionary relationships and genetic diversity should 
be considered in conservation strategies (Vrijenhaek et al. 1985, Avise 1989b). 
page 89. 
Chapter Seven: Discussion 
The New Zealand skinks present a much more interesting and challenging evolutionary problem than 
previously suspected and there is much potential for further study. In this fmal chapter the principal 
results are restated and suggestions for future work discussed. 
The seven major features of this thesis are: 
1). The demonstration of the utility of DNA sequ�nces for e�amining evolutionary 
questions. 
2). A secondary structure model for domain ill of vertebrate 125 rRNA was developed 
(Chapter Three), and subsequently used to examine nucleotide substitutions in the 
context of secondary structure (Chapters Four and Five). One third of this region of the 
125 rRNA molecule does not vary among the vertebrate sequences examined. _ .= 
3). Comparison of vertebrate 125 rRNA sequences also showed that there are similarities 
in the patterns of variability within the molecule. There may however also be taxon 
specific regions of variation. 
4). The sequence data were analyzed using a new sequence analysis approach termed . 
Spectral Analysis. One of the primary advantages of the spectral approach is that it gives 
a visual and quantitative representation of conflicting patterns of relationships in the data. 
5}. Most of the skink taxa have similar numbers of nucleotide substitutions when 
compared to each other an
J
phylogenetic analyses of the 125 rRNA sequences did not 
produce a completely resolved evolutionary tree. 
6). This lack of phylogenetic resolution was interpreted as reflecting old and rapid 
divergence among the Leiolopisma skinks. Three hypotheses were presented to account 
for the origin of New Zealand skinks. 
7). Hybridization between sympatric populations of two New Zealand skink species was 
inferred from comparisons of allozyme and mitochondrial DNA sequence information. 
Secondary Structure and Phylogenetic Analyses -
The refinement of the secondary structure model of domain ill of vertebrate 125 rRNA is a useful step 
for phylogenetic analyses of this commonly analyzed region of mtDNA. Comparison of the vertebrate 
sequences provides information on both the number of sites potentially free to vary and the pattern of 
variability in the molecule within and between groups. 
Oiapter 7, page 90. 
That not all sites in a gene change at the same rate has long been known (FilCh & Margoliash 1967, 
Fitch & Markowitz 1970), and models to take account of different rates of change among the codon 
positions are available (e.g. ,  Holmquist & Pearl 1980, Nei & Gojobori 1987). However, as FilCh & 
Margoliash (1967) and Shoemaker & Fitch (1989) demonstrated, rates of change between sites can be 
more complex. Dividing codons into just fast (third codon position) and slower (frrst+second) classes 
may not be sufficient to explain the patterns of variability. Methods to take account of different levels of 
variability within non-protein encoding sequences, such as rRNA, have not attracted the same degree of 
attention. 
While failure to take account of invariant sites can result in underestimates of the numbers of 
unobserved nucleotide substitutions (Palumbi 1989, Shoemaker & Fitch 1989), this does not appear to 
mislead phylogenetic inferences for the skinks (Fig. 5.12). Removal of constant columns from the data 
set (Fig. 5.12) is a simplistic way of investigating this problem however, and mathematical methods for 
adjusting for variable rates of change among sites in a sequence are required. Manske & Cllapman 
(1987) developed a model for 5S rRNA which assigns probabilities of substitution for each site based on 
relative nucleotide variance. Palumbi (1989) applied a simple model for corrections to mitochondrial 
restriction enzyme fragment mapping data. For mammalian D-loop sequences Saccone et al. (1991) 
calculated times of divergence using only those variable sites which fitted a Poisson distribution. 
Golding (1983) however, showed that gamma and lognormal distributions can be better than a Poisson 
distribution for describing variable mutation rates among sites in sequences. A gamma distribution 
model is being investigated for use in the Hadamard conjugation algorithms (PJ. Waddell pers. comm.). 
Weighted parsimony, which weights sites on the basis of relative numbers of substitutions, is more 
likely to give the correct tree than uniform weighting of characters in simulation studies (Williams & 
Fitch 1989, FilCh & Ye 1991). The method is tree-dependent however and results are sensitive to initial 
parameters of the tree (Williams & FilCh 1989). Tree-independent methods for taking account of sites of 
differing variability are currently being considered for implementation in HADTREE. 
Evolution of tbe 12S rRNA Molecule 
Knowledge of secondary structure is important for investigating the structural and functional basis of 
sequence variability between groups. The skink data set presented here is not a particularly useful one 
for examining variability of regions in relation to increasing taxonomic distance, since most of the taxa 
are equally divergent from each other and the majority of variable positions have only one or two taxa 
varying (Table 3. 1). Skinks, xantusiids, ratites and mammals each showed similar patterns of variability 
however(Fig. 3 .6), indicating that the skink molecule is not evolving in a different way than other 12S 
rRNA genes. In general, region S26 has a high proportion of variable sites in the vertebrates. The paired 
and unpaired regions around helices 43, 45 and 46 also accumulate many substitutions. In addition, it 
was suggested in C�apter Three that there may be taxon specific regions of variation. Skinks for 
example, had many variable in region L38 ' (Fig. 3.6, Table 3. 1), whereas ratites, bovids, and the great 
Otapter 7. page 91 .  
apes had relatively more variability in  SL40 (Figs. 3.6. 4.20). Further comparative sequence analyses for 
a larger vertebrate data set is required to determine how general these patterns of variability are. If taxon 
specific regions of variability do exist it may suggest differences in secondary structure and/or tertiary 
interactions among the vertebrates. 
Comparative sequence studies form the basis for experimental investigations of the secondary structure 
(Woese el al. 1983, Noller 1984). Most of the experimental examinations of the small subunit rRNA 
have been directed at E. coli (reviews in Gutell el al. 1985, Hill el al. 1990), which has the advantages of 
ease of culturing and amenability to extensive genetic manipulations. The mitochondrial small subunit 
rRNA lack many of the prokaryote helices (see figure 3.1),  and the functional and strutural significance 
of this is not well understood. The secondary structure of the region bounded by helix 36 requires 
further study (see Figs. 3.2, 3 .3, 3.4), and experimental examinations of the 12S rRNA molecule are 
important in this respect (see Gutell et aI. 1985, Hill et al. 1990). 
In a more general context. the small subunit rRNA offers a good system with which to investigate 
molecular evolution. Functional rRNA genes occur in both the mitochondrial and nuclear genomes. 
Sequences of the nuclear small subunit (18S) gene are available for many groups (see de Rijk et al. 
1992), and so detailed studies of the molecular evolution of the small subunit rRNA are possible. The 
mitochondrial and nuclear rRNA genes differ in size and evolve at different rates (Gutell et aI. 1985, 
Mindell & Honeycutt 1990, Hillis & D�1991). Refmement of secondary structure models for these 
genes permits patterns of sequence evolution to be related to the structure and function of the molecule. 
Studies of the evolution of the rRNA genes can also be related to tertiary interactions by investigations 
of the evolution of the ribosomal binding proteins. While the ribosomal proteins which interact with the 
nuclear and mitochondrial molecules are all encoded by nuclear genes, the mitochondrial ribosomal 
proteins are the products of a different set of genes than the nuclear ribosomal proteins. The 
mitochondrial ribosomal proteins also appear to be evolving more rapidly than their nuclear counterparts 
(O'Brien et al. 1985). Mitochondrial ribosomal proteins. and their genes, however have not been studied 
in as much detail as nuclear ribosomal proteins, which in turn are less well studied than the equivalent 
prokaryote proteins (Wool et aI. 1990. Wittmann-Liebold et al. 1990). 
Simulation Studies 
There were two important results from the simulation study (Chapters Four and Five). The first was that 
a star tree (as used in the simulations) is a useful null model to investigate the skink diversification. This 
star model describes the pattern of transitions and transversions among the skinks very well (compare 
Figs. 4.2, & 4.9). Validity of the model was supported when corrections for invariant sites were made. In 
the skink data set 41 % of the nucleotide substitutions were singletons (occurring on pendant edges). 
Assuming that all sites were free to vary in the initial simulation model resulted in approximately 60% 
of substitutions being singletons. When the adjustments to accoWlt for invariant sites were made 
(Chapter Five), 46� of the changes were singletons, agreeing well with the skink data. Further study of 
the applicability of this model is required however. For instance, the lengths of branches leading to 
ClJapter 7, page 92. 
skink taxa vary in size (see figures 5.3, 5.4, 5.18), which is not expected for a star tree model, unless 
there is variation in the rate of m�lecular evolution between the lineages. 
The second important aspect of the simulation study relates to calculating sequence divergence from 
transversion substitutions. Saturation of transition nucleotide substitutions occurs with increasing 
genetic divergence (Brown et al. 1982, Wilson et al. 1985, Miyamoto & Boyle 1989), so trans versions - -
have been used to obtain estimates for the degree of sequence divergence (Miyamoto & Boyle 1989, 
Kraus & Miyamoto 1991 ,  Milinkovitch et al. 1993). The preliminary analyses presented here indicate 
that such estimates will have a larger variance than estimates derived from total observed substitutions 
(Figs. 4.14, 4.15, 4. 16). Inferences of divergence based on transversion substitutions must therefore be 
made with caution. Proper statistical analyses of the simulation data are required to investigate this 
problem in more detail. A binary tree model also needs to be simulated to assess the relevance of the 
results obtained from the star tree model. 
The Hadamard Conjugation and Spectral Analysis 
As the skink data set used in this thesis illustrates, molecular sequences contain much information. 
Development of more detailed mathematical models to examine the information still remains a major 
challenge. The Hadamard conjugation and spectral analysis provide a powerful tool for examining 
information within DNA sequences. Data sets such as the skink one have been important and influential 
in rerming the algorithms. As demonstrated bere, support for taxonomic associations and the strength of 
conflicting signals can be directly assessed using spectral analysis. The technique of forcing specific 
groups of taxa to be included in the tree (Fig. 5. 16) is an important means for testing the likelihood of 
alternative bypotheses. Further refinements to the programs are being directed toward comparison of 
sub-optimal trees, so that the effects of conflicting signals in the data can be more accurately evaluated. 
Incompletely Resolved Skink Phylogeny 
The lack of phylogenetic resolution among the skinks was not simply due to there being too few variable 
or informative sites. Similar numbers of variable sites occur in the ratite data set of Cooper et al. 1992 
(Table 4.5a), and a resolved tree is produced with this data set (Fig. 5.2a). Nor was the complexity of the 
skink spectrum (Fig. 5.1) an artifact related to the presence of either invariant or highly variable sites. 
Weighting or removal of some or all of these sites did not improve resolution (Figs. 5.9, 5.10). 
Simulation studies (Figs. 4.9, 4. 13), as well as the bovid data set (Fig. 4.4b; Kraus & Miyamoto 1991 ,  
Allard et  al. 1992), indicate that a star tree model of divergence may be a good first approximation for 
the evolution of the skinks. A star-like divergence of the Leiolopisma may also be supported by 
immunological data (Hutchinson et al. 1990). With such a model, the many conflicting signals in the 
skink data set are due to rapid diversification of the taxa, as occured in Pecoran bovids (Kraus & 
Miyamoto 1991 ,  Allard et al. 1992). As noted above bowever, further testing of this and other models 
are required. 
Olapter 7, page 93. 
On the basis of the 12S rRNA sequence infoxmation the 17 New Zealand Leiolopisma examined in this 
thesis do not appear to have a recent and monophyletic origin in New Zealand Strongest support in the 
sequence data is found for the grouping of five pairs of taxa (Fig. 5.5), implying five more recent 
speciation events. These taxa are: L. microlepis and L. smithi; L. grande and "Stewart Island Green"; L. 
n. polychroma and L. n. nigriplantare; L. inconspicuum and L. notosaurus; and L. infrapunctatum with 
Lampropholis guichenoti (Figs. 5 .1 ,  5.5, & 5.18). A relatively close relationship between L. fallai and C. 
aenea was also suggested (Figs. 5 .1 ,  5 .16), but sequence similarities between C. aenea and the 
Mauritian L. teifairi (Fig. 5. 1 5) indicate that the phylogenetic position of Cyclodina is still uncertain. 
Hypotheses for the Origins of New Zealand Leiolopisma 
The occurrence of Leiolopisma on islands in both the Indian and Pacific oceans, and their apparent 
divergence from each other at about the same time (Fig. 5.5) may reflect a Gondwanan origin for this 
group of skinks (Fig. 5.22a). Continental drift, rather than �e skinks own powers of dispersal could 
account for their present distribution. The suggested close morphological similarity of the Fijian L. 
alazon to New Zealand Cyclodina (Zug 1985) is important in this respect since Fiji also 'appears to be a 
continental fragment (see Stevens 1980). Australian, New Zealand (L. grande, and Cyclodina oliverI), 
and New Caledonian (L. nigrofasciolatum and Caledonescincus autrocaledonicus) species all have 
similar immunological distances from L. teifairi (Hutchinson et al. 1990). 
Two other hypotheses for the origin of Leiolopisma were presented in Chapter Five. One proposed that 
the New Zealand Leiolopisma are the result of diversification within New Zealand during the Miocene 
(Fig. 5.22b). Analysis of allozyme data provides support for this hypothesis (C.H. Daugherty & G.B. 
Patterson pers. comm.), but as discussed in Chapters Five and Six, the small number of electrophoretic 
loci which could be sampled may make the phylogenetic relationships derived from the skink allozyme 
data less reliable than the pattern of relationships based on the sequence data. The third hypothesis 
suggests that many of the New Zealand Leiolopisma are derived from several independent colonizations 
of New Zealand (Fig. 5 .22c). Studies of the genetic variation of geckos on Pacific islands indicate that 
some species have only recently colonized some islands (Moritz et al. 1993), indicating that they have 
good oceanic dispersal abilities. Similar studies are required for the skinks. If New Zealand skinks did 
colonize New Zealand after the breakup of Gondwanaland, then the degree of diffe�nces between the 
12S rRNA sequences of the New Zealand Leiolopisma indicates that several independent colonizations 
would have had to have occurred (Fig 5.22c), pOssibly across large tracts of open ocean. The fact that 
the sequences from the Australian and Mauritian species are no more distinct from the New Zealand 
data set than the sequences of New Zealand taxa are from each other (Table 5.3), makes this hypothesis 
less well supported than the other two hypotheses. The relative closeness of the relationship between the 
New Zealand L. infrapunctatum and the Australian La. guichenoti could however imply that there have 
been at least two colonizations of (or departures from) New Zealand by Leiolopisma. 
Predictions of the Three Hypotheses 
Hypothesis One: Gondwanan Origins 
OIapter 7. page 94. 
If there was rapid divergence of Leiolopisma-like skinks in Gondwanaland, then further immunological 
and sequencing studies of other Leiolopisma-like species are expected to show similar levels of 
immunological and sequence divergence among the taxa as Hutchinson et al. (1990) and this study have 
found. Analyses of sequence (and immunological) data from other groups of plants and animals with 
limited powers of dispersal, such as geckos and land molluscs, could also be expected to show that taxa 
on several different islands of continental origin diverg�from each other at about the same time. 
Hypothesis Two: Oligocene Drowning 
If many of the New Zealand Leiolopisma speciated within New Zealand (Fig. 5 .22b), then they should 
be more similar genetically to each other than to New Caledonian or Australian species. This hypothesis 
is not supported by the available 12S rRNA sequence data. Most New Zealand Leiolopisma skinks are 
as equally distinct from each other as they are from the Australian and Mauritian species (Tables 4.1a, 
5.3). More Australian and New Caledonian species need to be examined though. 
Initial analyses of allozyme data have been interpreted as indicating a diversification of Leiolopisma 
within New Zealand, with La. guichenoti and L. telfairi much more distantly related (C.H. Daugherty & 
G.B. Patterson peTS. comm.). The different views of the evolutionary history of New Zealand skinks 
presented by allozyme data on the one hand and mitochondrial sequence data on the other may be a 
reflection of the small number of loci and sequence length examined in each case, as discussed in 
Chapter Six. 
If the evolutionary relationships of New Zealand skinks are similar to those indicated by analyses of 
allozymes and the Oligocene drowning hypothesis (Fig. 5.22b), then both immunological studies and 
sequence analysis of nuclear genes should group New Zealand Leiolopisma together, distinct from other 
Leiolopisma. 
Hypothesis Three: Multiple Colonists 
The third hypothesis, proposing multiple colonization events in New Zealand, predicts that taxa in New 
Caledonia and/or Australia will have closer phylogenetic relationships to at least some New Zealand 
species than the New Zealand species have among themselves. The relatively close relationship between 
L. injrapunctatum and La. guichenoti gives support to this hypothesis. The proximity of New Caledonia 
to New Zealand suggests that New Zealand skinks should show genetic closer similarity to New 
Caledonian than Australian species, if they arrived by this route. If Leiolopisma did diversify in 
Gondwanaland, then a date of separation between the ancestors of L. injrapunctatum and La. guichenoti 
of approximately 40 million years is implied (Chapter Five). Land connections between New Caledonia 
and New Zealand were more extensive during this period (Suggate et al. 1978), so DNA sequence 
studies of New Caledonian skinks may reveal taxa there with close relationships to both L. 
inJrapunctatum and La. guichenoti. 
Olapter 7, page 95. 
Testing the Hypotheses: What Other Skinks to Sample? 
More Pacific skinks need to be examined to test all three hypotheses. Several New Zealand skinks also 
have still to be studied at the DNA level. These include L. waimatense (South Island), L. Iwmalonotum 
(Poor Knights Islands), and more members of the South Island L. lineoocellatum-L. chloronotum 
complex, as well as several as yet undescribed species. Sequence studies of Cyclodina are already 
underway. The most critical overseas species to examine are L. alazon (Fiji), L. lichenigerum (Lord 
Howe), Leiolopisma species from New Caledonia (four are described; Sadlier 1986), and Australian 
species of the Leiolopisma group (which are now being reassigned to several new genera; Hutchinson et 
al. 1990). Other groups of skinks also need to be sampled to determine whether there has been more 
widespread skink diversification. Australia has about 30 genera of skinks (Cogger 1986), New 
Caledonia has ten (Sadlier 1986), New Guinea about sixteen (Allison 1982), and many more genera and 
species are distributed throughout the Pacific and Asia (Greer 1974, 1979). Of particular interest would 
be examination of other members of the Eugongylus group, such as Emoia and Morethia, and the New 
Caledonian genera Tropidoscincus, Geoscincus, Plwboscincus, and others (Sadlier 1986). Other genera 
within Lygosominae, for example Notoscincus (Australia), and Mabuya, a predominantly Asian group 
which Greer (1979) regards as the stock for the Lygosominae, could also be investigated. Emoia (which 
is distributed across the Pacific), is not too distant, immunologically, from the Leiolopisma (Hutchinson 
et al. 1990). 
Testing the Hypotheses: What Other Genes to Sample? 
Maternal inheritance, lack of recombination, and the effectively haploid state of mitochondria (Brown 
1985, Wilson et al. 1985, Moritz et al. 1987) make mtDNA a simpler system than the nuclear genome to 
investigate at the DNA sequence level. It may not be suitable however for investigating the deep 
branches in the skinks. The 12S rRNA sequence data for Leiolopisma is notable for the many conflicting 
phylogenetic signals it contains (Fig. 5.1). Resolution of the evolutionary relationships of the 
Leiolopisma may therefore require examination of more conservative sequences than domain ill of 12S 
rRNA, as is also required for the Pecoran bovid phylogeny (Kraus & Miyamoto 1991). Analysis of the 
complete 12S rRNA gene may provide greater taxonomic resolution and, as discussed above, is 
important for more general studies of molecular evolution. Cloning of the skink mitochondrial genome 
is underway and sequencing of the 12S rRNA gene is one of the first objectives. The mitochondrial 
large subunit (16S) rRNA may be another suitable candidate gene for examining the relationships of 
more distantly related taxa, though alignment of mammalian sequences in the GenBank database (data 
not shown) suggest that it changes at a similar rate as the mitochondrial small subunit gene, so it may 
not have greater resolving power for deeper branches in the evolutionary tree. 
Other mitochondrial genes, such as cytochrome oxidase I and cytochrome b, are widely used, and 
useful, for phylogenetic comparisons (for example, Kocher et al. 1989, Irwin et al. 1991 ,  Palumbi et al. 
1991 ,  Simon et al. 1.991,  Smith & Patton 1991, Pashley & Ke 1992). They tend to have more rapid rates 
of change than the rRNA genes thougb (Moritz et al. 1987), and so may be inadequate to resolve the 
Cllapter 7. page 96. 
deep branching patterns of skink relationships. However, the limited cytochrome b data for the skinks 
(Chapter Five, and Appendix 1), indicates that most of the differences between the skink sequences 
occur in third codon positions. If there are too many conflicting signals from third codon position sites, 
analysis of the more conservative ftrst and second codon positions may provide greater phylogenetic 
resolution. Third codon positions may be useful for studying the relationships between more closely 
related taxa, such as L. micro1epis and L. smithi. As shown in Chapter Five, the cytochrome b sequence 
-
did differentiate between the Gorge Burn L. maccanni and the putative hybrid (Table 5.4). 
Consistent ampliftcation of the cytochrome b gene sequence could not be achieved using the 
cytochrome b universal PCR primers of Kocher et al. (1989; see Chapter Two). A more efftcient, and 
ultimately more useful, strategy would be to clone and sequence the skink mitochondrial genome, and 
this is currently being attempted. Suitable PCR primers appropriate for various taxonomic or population 
comparisons can then be designed. Few molecular sequence studies of reptiles have been undertaken 
and the organization of the reptilian mitochondrial genome is not known, so such a project has general 
utility. Mitochondrial gene organization differs between birds, marsupials and placental m.iUnmals 
(Desjardins & Morais 1990, Paabo et al. 1991), and knowledge of gene order in reptile mtDNA may 
help resolve the relationships among reptiles, birds and mammals (see Chapter Two, and Marshall 
1992). 
Investigation of Nuclear Gene Sequences 
More conserved sequences for studying the branching order of the skinks can be obtained from nuclear 
genes, and analysis of nuclear rRNA and their intragenic spacer regions may be the most useful. PCR 
primers for some of these genes (e.g., 18S rRNA), and which work for a range of vertebrates, have 
already been designed (Hillis & Dixon 1991). As noted above, analysis of the nuclear 18S rRNA gene 
(the equivalent of the mitochondrial 12S rRNA molecule) would also complement molecular 
evolutionary studies of the 12S rRNA gene. 
Examination of the phylogenetic relationships and evolutionary histories of skinks using sequences from 
nuclear genes is also important because of the apparent conflicts between the mit�hondrial and 
allozyme studies. Investigation of nuclear protein-encoding genes in this regard would be informative. It 
would be particularly interesting to obtain sequence data from some of the loci used in the skink 
allozyme survey. Phosphoglucomutase, for example, had the largest number of alleles among the skinks 
(ten; C.H. Daugherty & G.B. Patterson pers. comm.), and may contain sufficient variable nucleotide 
sites to be phylogenetically informative. Speciftc PCR ampliftcation of introns by using PCR primers 
which anneal to conserved exon sequences may be the most suitable method to achieve this (Palumbi & 
Baker 1993, submitted). Glucose-6-phosphate dehydrogenase has received close molecular genetic 
scrutiny in other studies (Vulliamy et al. 1992) and is another potential nuclear gene to be considered for 
study in the skinks (six alleles were detected for this locus by C.H. Daugherty & G.B. Patterson). If the 
intron sequences are evolving rapidly then they may only be useful for resolving more recent 
Otapter 7. page 97. 
divergences (for example, the separation of L n. po/ychroma and L n. nigrip/antare, L microlepis and 
L. smithi, and possibly the hybrid and parental populations at Gorge Burn). 
Hybridization between L n. polychroma and L. maccanni 
Hybridization has been considered to have an important role in the evolution of plants (Cockayne 191 1 ,  
Rattenbury 1962, Grant 1981,  Arnold 1992), but it's significance in  animals may_be underra!ed (Mayr 
1963, Arnold 1992, Grant & Grant 1992, Dowling & DeMarais 1993). Dowling & DeMarais (1993) 
investigated hybridization among freshwater fish in the genus Gila, and suggested that interspecific 
hybridization has been a major factor in generating morphological and genetic diversity. Interspecific 
hybridization may also be common among some groups of birds, aDd hybrids may, temporarily at least, 
have greater reproductive success (Grant & Grant 1992). Hybridization between Lei% pisma therefore 
requires much closer examination. 
The three nucleotide differences between L. maccanni and the L. n. polychroma cytochrome b sequences 
from Gorge Burn suggest hybridization may have occurred in the Pliocene, over a million years ago 
(Chapter Six). That hybridization was detected by examination of only a few individuals indicates that 
hybridization may be relatively common. Whether there is a hybrid zone (Barton & Hewitt 1989) or just 
a localized introgression requires further examination of individuals from Gorge Burn and adjacent 
populations (see Patterson 1985, Daugherty et al. 1990b). Patterson's (1985) study of the ecology of the 
Central Otago skinks is the most comprehensive one of the gtoup so far, but the discovery of hybrid 
skinks emphasises the need for additional detailed ecological studies of both hybrid and non-hybrid 
populations to determine the behavioural and ecological features which usually keep them from 
interbreeding. 
Of more general significance is the fact that hybridization between the skinks can occur despite tens of 
millions of years separation from a common ancestor. Reports of naturally occuring hybridization 
between taxa which diverged more than five million years ago are uncommon in the literature (see 
Chapter Six), though frogs which diverged up to 35 million years ago have the potential to hybridize 
(Wilson et al. 1974). As discussed above, it is important to obtain more refined estimates for the 
divergence times of the skinks so that their ability to hybridise can be more accurately related to their 
degree of divergence. 
Ecological studies to investigate the viability and reproductive success of the skink hybrids (see Barton 
& Hewitt 1989, Grant & Grant 1982) are an essential complement to the genetic analyses. Captive 
b�eeding of skinks is a means to examine the potentia.l for interspecific hybridization and the viability 
and competitive ability of hybrids. Such experimental studies are essential for an understanding of the 
processes of speciation (Endler 1986, Coyne 1992). 
Oaapter 7, page 98. 
The Potential of L n. polychroma For Population Studies 
A strong emphasis of this thesis has been to being together ecological and evolutionary studies in New 
Zealand. L. n. polychroma, and other New Zealand skinks, offer many opportunities for such research. 
L. n. polychroma is a wide ranging species, encompassing six degrees of latitude (from the central North 
Island to Southland), and living in a variety of habitats from sea level to 1700 m (pickard & Towns 
1988, Patterson & Daugherty 1990). Wide-ranging and common species are ofte� cODSide� to be more 
variable (Wallace 1889, Mayr 1963) but genetic evidence in support of this is meagre. Reduced 
phenotypic variability has also been noted for populations at the geographic or ecological margins of a 
species' range (Mayr 1963, Lewontin 1974, Brossard 1984). StUdies of genetic variation within and 
between populations of ecologically and geographically distinct L. n. polychroma may be informative in 
this respect 
Analyses of the sequence data (Fig. 5.18) led to the suggestion in Chapter Five thatL. n. polychroma has 
an origin in Southern New Zealand. It now inhabits many areas considered to have been glaciated or 
ecologically depauperate during the Pleistocene (see Fleming 1980, Stevens 1980). C.H: Daugherty 
(pers. comm.) noted clines in allele frequencies for L. n. polychroma, and investigation of sequence 
differences between populations may permit examination of the origins and the directions and timings of 
dispersal of L. n. polychroma. If current distribution is due in part to post Pleistocene dispersal then 
sequence analysis of the most variable region of mtDNA, the D-loop (Moritz et al. 1987), could be 
suitable for studying this (see for example, Smith & Patton 1990, Thomas et al. 1990, Ward et al. 1991). 
Chromosomal studies (see Hillis & Moritz 1990) and DNA fingerprinting (Jeffreys et al. 1991) may also 
be informative for interpopulation differences. 
Rates of Molecular and Morphological Evolution in Skinks 
A Gondwanan origin for Leiolopisma skinks implies a low nite of sequence evolution (25 to 30 
substitutions (Table 4.la) per 80 MY, or 0.3 to 0.375 substitutions per MY). The ratites have 
approximately 0.5 substitutionslMY, and bovids a rate of 1 .2IMY (Table 4.6). A slower rate of 
molecular evolution in lizards than in birds or mammals may be related to a lower metabolic rate in 
reptiles (Martin & Palumbi submitted), though the xantusiid lizards may have a high rate of mutation 
(Table 4.6; Hedges et al. 1991). A low metabolic rate may result in both less oxidative damage to DNA 
(Richter et al. 1988, Shigenaga et al. 1989, Martin & Palumbi submitted) and slower turnover of DNA. 
and consequently slower rates ofmtDNA change (Martin & Palumbi submitted). Non-invasive analysis 
of oxidative damage is now possible using HPLC (Shigenaga et aI. 1992), so more detailed 
examinations of the relationship between metabolic rate and degree of DNA damage may be feasible. 
A Gondwanan, or even Miocene (15-25 MY A), origin of the skinks also suggests a low rate of 
morphological change in the skinks. Marked differences in rates of morpbological and genetic change 
are well documented (Wilson et al. 1974, King & Wilson 1975, Cherry et al. 1978, Wayne et al. 1989). 
The lack of morphological diversification in New Zealand skinks is unusual however both with respect 
to their ecological and latitudinal ranges (Bull & Whitaker 1975, Pickard & Towns 1988, Daugherty et 
OIapter 7. page 99. 
aI. 1990b), and their suggested period of separation from each other. T:heif m9fJ'helogieftlllRifeRBity 
may s9we't'er ge "Ia�d to ecologiGal RaWR&. Australian skinks are morphologically very diverse 
(Ricklefs et ai. 1981), and this may be due to a combination of factors, such as habitation in infertile and 
arid environments, which have few avian or mammalian predators, and the abundance of termites as a 
food source (Morton & James 1988). New Zealand geckos are morphologically more varied than skinks, 
though the degree of genetic divergence among geckos appears to be smaller � for skinks _(R. 
Hitchmough pers. comm.). A combination of exploitation of forest habitats and adoption of nocturnal 
lifestyles by some New Zealand geckos may have been a contributory factor in their greater 
morphological diversity (R. Hitchmough pers. comm.). -
Perhaps, as in the case of African jackals (Wayne et aI. 1989), an open and fertile grassland habitat. and 
potential abundance of prey may reduce both interspecific competition and development of 
morphological divergence among the New Zealand skinks. The existence of sympatric species of skinks 
(patterson 1985, Daugherty et al. 1990b) however suggest that subtle microhabitat differentiation and 
behavioural features must be important for maintaining separate gene pools. Some of the-sequence data 
presented here indicates that such differentiation may break down occasionally. 
Evolutionary Investigations of New Zealand Biota 
Many of the problems which perplexed early biologists in New Zealand are still unresolved. The chief 
difficulty has been accurate determination of evolutionary relationships among groups. Biochemical and 
molecular approaches, as demonstrated here and elsewhere (Daugherty et ai. 1982, Green et al. 1989, 
Daugherty et al. 1990a,b, Cooper et al. 1992) are essential for evolutionary investigations of New 
Zealand's biota. 
As aspects of the present study, and those of Patterson (1985) and Daugherty et al. (1990b) have shown, 
evolutionary and ecological questions are linked. The coordination of ecological and evolutionary 
studies of New Zealand skinks provide general insights into both the biological and geological history of 
New Zealand, as well as processes of molecular and organismal evolution. 
Application of molecular techniques to evolutionary questions has contributed much to an understanding 
of phylogeny (Wilson et ai. 1985, Hillis 1987, Patterson 1987, Wilson et al. 1989, Moritz & Hillis 
1990). They have been helpful not only in resolving relationships which have been difficult to address 
by other means, but can also provide a more reliable indicator of the degree of phylogenetic divergence 
(for example, King & Wilson 1975, Roberts & MaxsoIi 1985, Wayne et al. 1989, Meyer et al. 1990, 
Martin et al. 1993, see also Hillis 1987, Moritz & Hillis 1990). Molecular systematics has lead to a 
renaissance in morphological taxonomy (Patterson 1987), and satisfyingly, molecular and 
morphological phylogenies are more often in agreement than conflict (Hillis 1987, Moritz & Hillis 
1990). Where morphological and molecular data are in apparent conflict (for example with the 
I MASSEY U .  ' ITY
-
I US  f. -:. ( '-- -
OJaptel' 7. page 100. 
mammals; Wyss et al. 1987, Gauthier et al. 1988, Hedges et aI. 1991, Marshall 1992), emphasis should 
be on examination of the advantages and limitations of each data set, rather than which data set is 'right' 
(Moritz & Hillis 1990). 
Phylogenetic trees should be regarded as hypotheses, and steps along the way to understanding 
evolutionary processes. Mayr's observation (1963, page 9) that ... 
..... even today there still are some zoologists to whom the term 'evolution' signifies little 
more than the determination of homologies, common ancestors, aru:I phylogenetic trees. " 
should not be forgotten. 
References 
Abrahams, J. P., van den Berg, M., van Batenburg, E., Pleij, C. (1990). Prediction of RNA secondary sttucture, including 
pseudolcnotting, by computer simulation. Nucleic Acids Res. � 3035-3044. 
Allan, H. H. (1961). Flora of New Zealand. VoL 1. Government Printer, Wellington. 
page 101. 
Allard, M W., Miyamoto, M. M., Jarecki, L., Kraus, F., Tennant, M. R. (1992). DNA systematics and evolution of the artiodactyl 
family Bovidae. Proc. NatL Acad. Sci. USA 89, 3972-3976. 
Allibone, R. M. (1990). Studies of genetic variation and systematics of New Zealand galaxiids using allozyme electrophlXesis. 
Unpublished MSc Thesis, University of Otago, Dunedin. 
Allison, A. (1982). Distribution and ecology of New Guinea lizards. In: Gressitt, J.L. (ed.) Biogeography and ecology of New Guinea. 
Vol. 2. Dr W. Junk b.v. Publishet's, The Hague. pp 803-813. 
Andersen, J. C. (1969). Myths &: legends of the Poly�ians. Charles E. Tuttle Company. 
Anderson, A. (1983). Faunal depletion and subsistence change in the early prehistory of southern New Zealand. ArchaeoL Oceania � 
1- 10. 
Anderson, S., Bankier, A.T., Barrell, B.G., de Bruijn, M.H., Coulson, A.R., Drouin, J., Eperon, I.e., Nierlich, D.P., Roe, B.A., Sanger, 
F., Schreier, P.H., Smith, A.J., Staden, R., Young, I.G. (1981). Sequence and organization of the human mitochondrial genome 
Nature 290, 457-465. 
Anderson, S., de Bruijn, M.H., Coulson, A.R., Eperon, I.C., Sanger, F., Young, I.G. (1982). Complete sequence of bovine mitochondrial 
DNA: conserved features of the mammalian mitochondrial genome. J. Mol. BioL 156, 683-717. 
Antao, V. P., Lai, S. Y., TInoco, I. (1991). A thermodynamic study of unUSUally stable RNA and DNA hairpins. Nucleic Acids Re.r. 19, 
5901-5905. 
Aquadro, C. F., Kaplan, N. Risko, K. 1. (1984). An analysis of the dynamics of mammalian mitochondrial DNA sequence evolution. 
Mol. BioL Evol . .!.' 423-434. 
Amason, U., Gullberg, A., Widegren, B. (1991). The complete nucleotide sequence of the mitochondrial DNA of the fin whale, 
Baloenoptera physalus. J. Mol. EvoL 33, 556-568. 
Amason, U., Johnsson, E. (1992). The complete mitochondrial DNA sequence of the harbIX seal, Phoca vitulina. J. MoL EvoL � 493-
505. 
Arnheirn, N., Erlich, H. (1992). Polymerase chain reaction strategy Ann. Rev. Biochem. 61, 131-156. 
Arnheim, N., White, T., Rainey, W. E. (1990). Application ofPCR: organismal and population biology. BioScience 40, 174-182. 
Arnold, E. N. (1980). Recently extinct reptile populations from Mauritius and Reunion, Indian Ocean. J: ZooL, Lond. 191, 33-47. 
Amold, M. L. (1992). Natural hybridization as an evolutionary process. Ann. Rev. &oL Syst. 23, 237-261. 
Aubert, J., Solignac, M. (1990). Experimental evidence for mitochondrial DNA introgression between Drosophiw species. Evolution 
44, 1272-1282. 
Avise, J. e. (1986). Mitochondrial DNA and the evolutionary gen�cs of higher animals. Phil. Trans. R. Soc. Lond. B 312, 325-342. 
Avise, J. e. (1989a). Gene trees and IXganismal histories: a phylogenetic approach to population biology. Evolution 43, 1 192-1208. 
Avise, 1. e. (1989b). A role for molecular genetics in the recognition and conservation of endangered species. TREE 1 279-281. 
Avise, J. e. (1991). Ten I!northodox perspectives on evolution prompted by comparative population genetic findings on mitochondrial 
DNA. Ann. Rev. Genet. 25, 45-69. 
page 102. 
Avise, J. c., Aquadro, C. F. (1982). A comparative summary of genetic distances in the vertebrates. Panerns and conelations. EvoL 
Bioi. 15, 151-185. 
Avise, J. C., Arnold, J., Ball, R. M, Bermingham, E., Lambr, T., Neigel, J. E., Reeb, C. A., Saunders, N. C. (1987). Intraspecific 
phylogeograpby: the mitochondrial DNA bridge between population genetics and systematics. Ann. Rev. Ecol. Syst. ll, 489-522. 
Avise, J. C., Bowen, B. W., Lamb, T., Meylan, A. B., Benningham, E. (1992). Mitochondrial DNA evolution at a turtle's pace: 
evidence for low genetic variability and reduced microevolutiolW}' rate in Testudines. Mol. BioL Evol. �, 457-473. 
Baker, A. J. (1975). Morphological variation, hybridization and systematics of New Zealand oyster catchecs (CharaQriiformes: 
Haematopodidae). l. ZooL Lond. 175, 357-390. 
Baker, A. J. (1992). Genetic and morphometric divergence in ancestral european and descendent New Zealand populations of 
chaffinches (Fringilla coekbs). Evolution 46, 1784-1800. 
Baker, A. 1., Moeed, A. ( 1987). Rapid genetic differentiation and founder effect in colonizing populations of common mynas 
(Acridotheres tristis). Evolution 41,  525-538. 
Ballard, J. W. O., Olsen, G. J., Faith, D. P., Odgecs, W. A., Rowell, D. M., Atkinson, P. W. (1992). Evidence from 125 ribosomal RNA 
sequences that onychophorans are modified arthropods. Science 258, 1345- 1348. 
Barton, N. H., Hewitt, G. M. (1989). Adaptation, speciation and hybrid zones. Nature 341,  497-503. - '-
Barwick, R.E. (1959). The life history of the common New Zealand skink uiolopisma l.eJandjca (Gray, 1 843). Trans. Roy. Soc. N.Z. 
86, 331-380. 
Batcheler, C. L., McLennan, M. J. (1977). Craniometric study of allometry, adaptation and hybridism of red deer (CeNUs elapluu 
scoticus L) and wapiti (C. e. nelsoni, Bailey) in Fiordland, New Zealand. Proc. N.Z £CoL Soc. 24, 57-75. 
Bauer, A. M. (1990). Phylogenetic systematics and biogeography of the Carpbodactylini (Reptilia: Gekkonidae). Bonner Zoologische 
Monographien 30, 1-128. 
Baverstock, P. R., Donnellan, S. C. (1990). Molecular evolution in Australian dragons and skinks: a progress report. Memoirs 
Queensland Mus. 29, 323-33 1 .  
Bell, B. D., Newman, D .  G., Daugherty, C .  H .  (1985). The ecological and biogeography of the archaic New Zealand he:petofauna 
(Leiopelmatidae, Sphenodontidae). In: Grigg, G., Shine, R., Ehmann, H. (eds.), Biology of Australasianfrogs and reptiles. Royal 
Zoological Society of New South Wales. pp. 99-106. 
Benton, M. J. (1990). Phylogeny of the major tetrapod groups: morphological data and divergence dates. J. Mol. EvoL 30, 409-424. 
Benningham, E., Lamb, T., Avise, J. C. (1986). Size polymorphisms and heteroplasmy in the mitochondrial DNA of lowec vectetntes. 
1. Heredity 77, 249-252. 
Bernardi, G., Becnardi, G. (1991). Compositional properties of nuclear genes from cold-blooded vertebrates. J. MoL Evol. 33, 57-67. 
Bernardi, G., Mouchiroud, D., Gautier, c., Bernardi, G. (1988). Compositional patterns in vertebrate genomes: conservation and change 
in evolution. J. Mol. Evol. 28, 7-18. 
Berry, R. 1.  (ed.). (1984). Evolution in the Galapagos Islands. Academic Press, LondonlLinneaD Society of London. Symposium Series 
no. 1 2. 
Bibb, MJ., van Etten, R.A., Wright, C.T., Walberg, M.W., Clayton, D.A. (1981). Sequence and gene organization of mouse 
mitochondrial DNA. Cell 26, 167- 1 80. 
Bigelow, R. S. (1965). Hybrid zones an 'd reproductive isolation. Evolution 19, 449-458. 
Bigelow, R. S. (1967). The grasshoppecs (Acrididae) of New Zealand. Their taxonomy and distribution. University of CanJerbury Publ. 
2, 1 - 1 1 1 .  
Bloch, W .  (1991). A biochemical perspective of the polymerase chain reaction. Biochemistry 30, 2735-2747. 
page 103. 
Britten. R. J. (1986). Rates of DNA sequence evolution differ between taxonomic groups. Science 231. 1 393-1398. 
Brown. W. M .• 1983. Evolution of animal mitochondrial DNA. In: Nei. M .• Koehn, R.K. (eds.). Evolution of genes and proteins. 
Sinauer Associates. Inc. Sunderland, Massachusetts. Pp. 62-88. 
Brown. W. M .•  George. M .• Wilson. A. C. (1979). Rapid evolution of animal mitochondrial DNA. Proc. NatL Acad. Sci. USA 76. 1967-
1971 .  
Brown. W.  M . •  Prager. E. M . •  Wang. A.. Wilson. A. C .  (1982). Mitochondrial DNA sequences of primates: tempo and mode of 
evolution. J. MoL Evol. J!. 225-239. 
Brossard. P. I. (1984). Geographic patterns and environmental gradients: the central-marginal model in Drosophila revisited. Ann. Rev. 
&oL Syst. 15. 25-64. 
Bun. P. C .• Whitaku. A. H. (1975). lbe amphibians. reptiles. birds and mammals. In: Kuschel. G. (ed.). Biogeography and ecology in 
New Zealand. Dr W. Junk b.v. Publishers. lbe Hague. Pp. 231-276 
Burrows. C. J. (1965). Some discontinuous distributions of plants within New Zealand and their ecological significance. Part n: 
disjunctions between Otago-Southland and Nelson-Marlborough and related distribution patterns. Tuatara Q. 9-29. 
Burrows. C. J •• Greenland, D. E. (1979). An analysis of the evidence for climatic change in New Zealand in th.: �  thousand years: 
evidence from divuse natural phenomena and from instrumental records. J. Roy. Soc. N.Z. 2.. 321-373. 
Bush. G. L. Case. S. M .• Wilson. A. C .• Patton. J.L (1977). Rapid speciation and chromosomal evolution in mammals. Proc. Natl. 
Acad. Sci. USA 74. 3942-3946. 
Buth. D. G. (1984). The application of electrophoresis data in systematic studies. Ann. Rev. &oL Syst. 12. 501-522. 
Cann, R. L. Brown. W. M .• Wilson. A. C. (1984). Polymorphic sites and the mechanism of evolution in human mitochondrial DNA. 
Genetics 106. 479-499. 
Cantatore. P .• Gadaleta. M. N .• Roberti. M .• Saccone. c.. Wilson. A. C. (1987). Duplication and remoulding oftRNA genes during the 
evolutionary rearrangement of mitochondrial genomes. Nature 329. 853-855. 
Casanova. J.-L. Pannetier. C .• Jaulin, c.. Kourilsky. P. (1990). Optimal conditions for directly sequencing double-stranded PCR 
products with Sequenase. Nucleic acids research J!. 4028. 
Carlquist, S. (1974). Island Biology. Columbia University Press. New York. 
Carrol. R. L (1969). Origin of reptiles. In: Gans, C. (ed.), Biology of reptilia. Vol. I. Morphology A. Academic press. Pp. 1-44. 
Carson, H. L (1983). Chromosomal sequences and interisland colonizations in Hawaiian Drosophila. Genetics 103, 465-482. 
Cavalli-Sforza, L L. Edwards. A. W. F. (1967). Phylogenetic analysis: models and estimation procedures. Evolution 21. 550-570. 
Cavender. J. A. (1978). Taxonomy with confidence. Math, Biosci. 40, 271-280. 
Champion. A. B .• Pragu. E. M .• Wachter. D .• Wilson, A. C. (1974). Microcomplement fixation. In: Wright, C.A. (ed.) Biochemical and 
imnumological taxonomy 0/ animals. Academic Press. London. pp. 397-416. 
Cheng. J.-W., Chou. S.-H., Reid, B. R. (1992). Base pairing geometry in GA mismatches depends entirely on the neighboring sequence. 
J. MoL BioL 228. \037- 1041. 
Cherry. L. M . •  Case. S. M .• Wilson. A. C. (1978). Frog puspective on the morphological differences between humans and 
chimpanzees. Science 200, 209-21 1 .  
Qark, A. G . •  Whittam, T. S .  (1992). Sequencing errors and molecular evolutionary analysis. Mol. BioI. EvoL 2, 744-752. 
Qary. D. O .• Wolstenholme. D. R. (1985). The ribosomal genes of Drosophila mitochondrial DNA. Nucleic Acids Res. Q. 4029-4045. 
page 104. 
Oayton, D. A. (1982). Replication of animal mitochondrial DNA. CeU 28, 693-705. 
Oegg, M. T., Durbin, M. L. (1990). Molecular approaches to the study of plant biosystematics. Aust. Syst. Bot. lo l-8. 
Oimo, F. M. (1978). The Powellip/Janta gilliesi-traversi-hochsterteri-rossiana-lignaria-superba ring species (Mollusca: Pulmonata). 
N.Z J. 200L 1, 289-294. 
Oimo, F. M. (1989). The panbiogeography of New Zealand as illuminated by the genus Fectola lredale, 19 15  and subfamily 
Rotadiscinae Pilsbry, 1927 (Mollusca: Pulmonata: Punctoidea: Cbaropidae). N.Z J. 2001. 16, 587-649. 
Cockayne, L (191 1 ). Observations concerning evoluuon, derived from ecological studies in New Zealand. Trans. Proc. N.Z 1nst. 44, I­
SO. 
Cogger, H. G. (1986). RepriJes and amphibians of Australia. Reed Books Pty Ltd. 
Cogger, H. G., Heatwole, H. (1981). The Australian reptiles: origins, biogeography, distribution patterns and island evolution. In: Keast, 
A. (ed.), Ecological biogeography of Australia. Dr. W. Junk b.v. Publishers, The Hague. pp. 1333-1373. 
Cohn, J. P. (1990). Genetics for wildlife conservation. BioScience.1Q. 167-171 .  
Connor, H.  E.  (1985). Biosystematics of higher plants in  New Zealand 1965-1984. N.Z J. Bot. 23, 613-644. 
Cooper, A. J., Chambers, G. K., Cooper, R. A. (in prep.). The Oligocene bottleneck and New Zealand ratite phyiogeny. 
Cooper, A., Mourer-Chauvir6, C., Chambers, G. K., von Haeseler, A., Wilson, A. C., Piiiibo, S. (1992). Independent origins of New 
Zealand moas and kiwis. Proc. Natl. Acad. Sci. USA 89, 8741-8744. 
Cooper, R. A. (1989). New Zealand tectonostratigraphic terranes and panbiogeogra�y. N.Z J. Zool. 16, 699-712. 
Coyne, J. A. (1992). Genetics and speciation. Nature 355, 5 1 1-5l5. 
Cracraft, J.,  Helm-Bychowski, K. (1991). Parsimony and phylogenetic inference using DNA sequences: some methodological 
strategies. In: Miyamoto, M.M., Cracraft, J. (eds.), Phylogenetic analysis of DNA sequences. Oxford University Press. pp. 1 84-220. 
Craig, J. (1984). Are communal pukeko caught in the prisoner's dilemma? Behav. &oL Socio. Biol. ll, 147-150. 
Craw, R. C. (1988). Continuing the synthesis between panbiogeography, phyiogenetic systematics and geology as illustrated by 
empirical studies on the biogeography of New Zealand and the Chatham Islands. Syst. ZooL 37, 291-3 10. 
Craw, R. (1989). New Zealand biogeography: a panbiogeographic approach. N.Z J. Zool. 16, 527-547. 
Cree, A., Cockrem, J. F., Guillette, L J. (1992). Reproductive cycles of male and female tualara (Sphenodon punctatus on Stephens 
Island, New Zealand. J. Zool., Lond. 226, 199-217. 
Cronquist, A., 1980. Chemistry in plant taxonomy: an assesment of where we stand. In: Bisby, F.A., Vaughan, J.G., Wright, c.A. (eds.), 
Chemosystematics: principles and practice. pp. 1-27. 
Crook, I. G. (1975). The tuatara. In: Kuschel, G. (ed.), Biogeography and ecology in New Zealand. Dr W. Junk b.v. Publishers The 
Hague. pp. 331-352. 
Crowhurst, R. N., Lints, R., Atkinson, R. G., Gardner, R. C. (1990). Restriction fragment length polymorphisms in the genus Actinida 
(Actinidiaceae). Plant Syst. Evol. 172, 193-203. 
Crozier, R. H. (1990). From population genetics to phylogeny: us� and limits of mitochondrial DNA. Aust. Syst. Bot. 1, °1 1 1-124. 
Darns, E., Hendriks, L, van der Peer, Y., Neefs, J.-M., Smits, G., Vandenbempt, I., De Wachter, R. (1988). Compilation of small 
ribosomal subunit RNA sequences. Nucleic Acids Res. 16, Suppl., r87-rI73. 
Darwin, C. (1859). On thf origin of species by means of natural selectio,," or the preservation of favoured races in the struggle for life. 
John Murray, London. 
page 105. 
Darwin, C. (1875). The variation of animals and plants under domestication. Second edition. Iohn Murray, London. 
Daugherty, C. H., Bell, B. D., Adams, M., Maxson, L. R. (1981). An electrophoretic study of genetic variation in the New Zealand frog 
genus Leiopelma. N.Z. J. ZooL !, 543-550. 
Daugherty, C. H., Cree, A., Hay, 1. M., Thompson, M. B. (1990a). Neglected taxonomy and continuing extinctions of tuatara 
(Sphenodon). Nature 347, 177-179. 
Daugherty, C. H., Patterson, G. B., Thorn, C. I., French, D. C. (1990b). Diffeieiltiation of the members of the New Zealand Lewlcpisma 
nigriplanrare species complex (Lacertilia: Scincidae). Herpet. Monogr. 1, 61-76. 
de Bruijn, M. H. L., KIug, A. (1983). A model for the teniary structure of mammalian transfer RNAs lacking the entire ' dihydrouridine' 
loop and stem. £MBO J. �, 1309-1321. 
de Rijk, P., Neefs, I.-M., van de Peer, Y., de Wachter, R. (1992). Compilation of small ribosomal subunit RNA sequences. Nucleic 
Acids Res. 20, Suppl., 2075-2089. 
Dell, R. K. (1955). The land mollusca of Fiordland, south-west Otago. Trans. Roy. Soc. N.Z. 82, 1 1 35-1 148. 
DeSalle, R., Freedman, T., Prager, E. M., Wilson, AC. (1987). Tempo and mode of sequence evolution in mitochondrial DNA of 
Hawaiian Drosophila. J. MoL Evol. 26, 157-164. 
DeSalle, R., Hunt, 1. A (1987). Molecular evolution in Hawaiian drosophi1ids. Trends &01. Evol. �, 212-216. 
DeSalle, R., Templeton, A R. (1988). Founder effects and the rate of mitochondrial DNA evolution in Hawaiian Drosophila. Evolution 
42, 1076-1084. 
Desjardins, P., Morais, R. (1990). Sequence and gene organization of the chicken mitochondrial genome: a novel gene order in higher 
vertebrates. J. Mol. BioI. 212, 599-634. 
Desjardins, P., Morais, R. (1991). Nucleotide sequence and evolution of coding and noncoding regions of a quail mitochondrial 
genome. J. MoL Evol. 32, 153-161. 
Devereux, I., Haeberli, P.,  Smithies, O. (1984). A comprehensive set of sequence analysis programs for the V AX. Nucleic acids 
research 12, 387-395. 
Dieffenbach, E. (1843). Travels in New Zealand. Vol. II. Iohn Murray, London. 
Dixon, M. T., Hillis, D. M. ( 1992). Ribosomal RNA secondary structure: compensatory mutations and implications for phylogenetic 
analysis Mol. Bioi. Evol. 10, 256-267. 
Dobzhansky, T. (1941). Genetics and the origin of species. Second edition. Columbia Univecsity Press, New York. 
Dowling, T. E., Brown, W. M. (1989). Allozymes, mitochondrial DNA, and levels of phylogenetic resolution among four minnow 
species (Notropis: Cyprinidae). Syst. Zool. 38, 126-143. 
Dowling, T. E., DeMarais, B. D. (1993). Evolutionary significance of evolutionary hybridization in cyprinid fishes. Nature 362, 444-
446. 
Draper, D. E. (1990). Structure and function of ribosomal protein-RNA coinplexes: thermodynamic studies. In: Hill, W.E., Dahlberg, 
A., Garrett, R.A., Moore, P.B., Schlessinger, D., Warner, I.R. (eds.), The ribosome: structure, functwn, and evolutwn. American 
Society for Microbiology, Washington, D.C. pp. 160- 167. 
Dugdale, 1. S. (1989). New Zealand lepidoptera: basic biogeography. N.Z. J. Zool. .!2, 679-687. 
Dunon-Bluteau, D., Brun, G. (1986). The secondary structures of the Xenopus Iaevis and human mitochondrial small ribosomal subunit 
RNA are similar. FEBS Letters 198, 333-337. 
Easteal, S. (1990). The pattern of mammalian evolution and the relative rate of molecular evolution. Genetics 124, 165-173. 
page 106. 
Ehresmann, B., Ehresmann, C., Romby, P., Mougel, M. Baudin, F., Westhof. E., Ebel, J.-P. (1990). Detailed structures ofrRNA.s: new 
approaches. In: Hill, W.E., Dahlberg, A., Garrett, R.A., Moore, P.B., Schlessinger, D., Warner, 1.R. (eds.), The ribosome: strvcture, 
function. and evolution. American Socie.ty for Microbiology, Washington, D.C. Pp. 148-159. 
Endler, J. A. (1986). Natural selection i" the wild. Princeton Univasity Press, Princeton, New Jasey. 
Estes, R., Price, L. I. (1973). Iguanid lizard from the upper cretaceous of Brazil. Science 1 80, 748-751 .  
Evetts, P.M., Grimmond, N.M. (1982). 'llie effects oC temperarure on oxygen consumption in Leiolopisma nigripiantare maccanni and 
Hoplodactylus maculatus. In: Newman, D.G. (ed.), New Zealand herpetology. New Zealand Wildlife Service Occassional PubUcatWn �, 
409-415. 
Felsenstein, J. (1988). Phylogenies from molecular sequences: infaence and reliability. Ann. Rev. GeMt. 22, 521-565. 
Felsenstein, J. (1991). PHYUP (phylogmy inference packLlge), version 3.4. Department of Genetics, Univ. Washington, Seattle. 
Fetris, S. D., Sage, R. D., Huang, e.-M. , Nielsen, J. F., Rille, U., Wilson, A. e. (1983). Flow of mitochondrial DNA across species 
boundary. Proc. NatL Acad. Sci. USA ��, 2290-2294. 
Fife, A. 1. (1985). Biosystematics of the cryptogamic flora of New Zealand: bryophytes. N.Z. J. Bot. 23, 645-662. 
Fisher, F. J. F., 1 965. The alpine RanuncuU of New Zealand. New Zealand Department of Scientific and Indus��Research Bulletin 
165. 
Fitch, W. M. (1967). Evidence suggesting a non-random character to nucleotide replacements in naturally occurring mutations. J. Mol. 
BioI. 26, 499-507. 
Fitch, W. M. (1984). Oadistic and other methods: problems, pitfalls, and potentials. In: Duncan, T., Stuessy, T.F. (eds.), Cladistics: 
perspectives on the reconstrvction of evolutionary history. Columbia University Press, New York. Pp.221-252 
Fitch, W. M., Atchley, W. R. (1987). Divergence in inbred strains of mice: a comparison of three diffaent types of data. In: Patterson, 
C. (ed.), Moucules and morphology in t'Volution: conflict or compromise? Cambridge University Press, Cambridge. Pp. 201-216. 
Fitch, W. M., Margoliash, E. ( 1967). Construction of phylogenetic trees. Science 155, 279-284. 
Fitch, W. M., Markowitz, E. (1970). An imprOVed method for determining codon variability in a gene and its application to the rale of 
fixation of mutations in evolution. Biochem. Genet. i, 579-593. 
Fitch, W. M., Ye, J. (1991). Weighted parsimony: does it work? In: Miyamoto, M.M., Cracraft, J. (eds.), PhylogeMtic Q1IQlysis of DNA 
sequences. Oxford University Press. pp. 147-154. 
Fleischer, S., McIntyre, J.O., Vidal, J.e. ( 1979). Large-scale preparation of ral liva mitochondria in high yield. Meth. Enzym. 55, 32-39. 
Fleming, C. A. (1958). Darwinism in Nc'w Zealand: some examples, influences and developments. Trans. Proc. Roy. Soc. N.Z M. 65-
86. 
Fleming, C. A. (1980). The geological history of New Zealand and its Ufe. Auckland University Press. 
Frankel, O. H., Hair, J. B. (1937). Studic,..; on the cytology, genetics, and taxonomy of New Zealand Hebe and Veronica. (part 1). N.Z J. 
Sci. Tech. ll, 669-687. 
Freeth, A. L, Sin, F. Y. T. (1986). Enzyme variation in marine and estuarine populations of the chiton Sypharochiton pelliserpentis. 
N.Z J. Mar. Freshwater Res. 20, 561-568. 
Freier, S. M., Kierzek, R., Jaeger, J. A., Sugimoto, N., Caruthers, M. H., Neilson, T., Turna, D. H. (1986). Improved free-energy 
parameters for predictions of RNA duplc:x stability. Proc. NatL Acad. Sci. USA 83, 9373-9377. 
Gadaleta, G., Pepe, G., de Candia, G., Quagliariello, e., Sbisa, E., Saccone, C. (1989). The complete nucleotide sequence of the Rattus 
norvegicus mitochondrial. genome: cryplic signals revealed by comparative analysis between vertebrates. J. Mol. Evol. 28, 497-516. 
Gage, M. (1961). New Zealand glaciations and the duration of the Pleistocene. J. Glaciology �, 940-943. 
page 107. 
Gauthier, J. A., Kluge, A. G., Rowe, T. (1988). Amniote phylogeny and the importance of fossils. Cladistics �, 105-209. 
Gibson, A. R., Baker, A. J., Moeed, A. (1984). Morphometric variation in introduced populations of the common myna (Acridotheres 
tristis): an application of the jaclclcnife to Irincipal component analysis. Syst. Zool. 33, 408-421. 
Gill, B. J. (1976). Aspects of the ecolog)" morphology, and taxonomy of two skinks (Reptilia: Lacertilia) in the coastal Manawatu area 
of New Zealand. N.Z. J. ZooL �, 141-157. 
Gillespie, 1. H. (1986). Rates of molecul:u evolution. Ann. Rev. EcoL Syst. E, 637-665. 
Glotz, c., Brimacombe, R. (1980). An eRperimentally-derived model for the secondary structure of the 16S ribosomal RNA from 
Escherichia coli. Nucleic Acids Res. !, 2377-2395. 
Goddard, K., Caccone, A., Powell, J. R. (1990). Evolutionary implicatioils of DNA divergence in the Drosophila obscura group. 
Evolution 44, 1656-1670. 
Godley, E. J. (1949). Cytology and geneltics and their application to New Zea1ai1d plants. Tuatara �, 109-1 15.  
Golding, G. B. (1983). Estimates of DNA and protein sequence divergence: an examination of some assumptions. MoL Bioi. EvoL !, 
125-142. 
Goodman, M., Miyamoto, M. M, CzelwmiaJc. 1. (1987). Pattern and process in vertelrate phylogeny revealed Dy �volution of 
molecules and morphology. In: Patterson, C. (ed.), Molecules and morphology in evolution: conflict or compromise? Cambridge 
University Press, Cambridge. pp. 141-176. 
Gottlieb, L D. (1984). Isozyme evidenCtl and Iroblem solving in plant biosystematics. In: Grant, W.P. (ed.), Plant biosystematics. 
Academic Press. pp. 343-357 
Grant, B. R., Grant, P. R. (1989). Evolutionary dynamics of a natural population. The large cactus finch of the Galapagos. Univeristy of 
Chicago Press. 
Grant, P. R., Grant, B. R. (1992). Hylri(�zation and bird species. Science 256, 193-197. 
Grant, V. (1981). Plant speciation. Secol'ld edition. Columbia University Press. 
Gray, M. W. (1989). Origin and evolution of mitochondrial DNA. Ann. Rev. Cell BioI. �, 25-50. 
Green, D. M (1988). Cytogenetics ofthl: endemic New Zealand frog, LeiopJema hochsteneri: extraordinary supernumerary 
chromosome variation and a unique sex chromosome system. Chromosoma 97, 55-70. 
Green, D. M., Sharbel, T. F. (1988). Comparative cytogenetics of the primitive frog Leiopelma archeyi (Anura; Leiopelmatidae). 
Cytogenet. CelL Genet. 47, 212-216. 
Green, D. M., Sharbel, T. F., Hitchmou�,h, R. A., Daugherty, C. H. (1989). Genetic variation in the genus Leiopelma and relationships 
to other primitive frogs. Z. 1.001. Syst. Eliolut.-forsch. 27, 65-79. 
Greer, A. E. (1970). A subfamilial classification of scincid lizards. Bull. Mus. Comp. 2001. 139, 151-184. 
Greer, A. E. (1974). The generic relatioliships of the Scincid lizard genus Leiolopisma and its relatives. Aust. J. Zool., Suppl. Ser.ll, 1 -
67. 
Greer, A. E. (1979). A phylogenetic subdivision of Australian skinks. Rec. Aust. Mus. 32, 339-37 1 .  
Greer, A. E. (1982). A new species of L..'iolopisma (Lacertilia: Scincidae) from Western Australia, with notes on  the biology and 
relationships of other Australian specieS. Rec. Aust. Mus. 34, 549-573. 
Greer, C.E., Peterson, S.L, Kiviat, N.B., Manos, M.M. (1991). PCR amplification from paraffin-embedded tissues. Effects of fixative 
and IlXation time. Americ� journal of clinical pathology 95, 1 17-124. 
Gutell, R. R., Fox, G. E. (1988). A compilation of large subunit RNA sequences Iresented in a structural format. Nucleic Acids Res. 16, 
Supp\., rI75-r269. 
page 108. 
Gutell, R. R., Schnare, M. N., Gray, M. W. (1990). A compilation of large subunit (23S-like) ribosomal RNA sequences presented in a 
secondary structure format. Nucleic AcUilr Res. !!, Suppl., 23 19-2330. 
Gutell, R. R., Schnare, M. N., Gray, M W. (1992). A compilation of large subunit (23S- & 23S-like) ribosomal RNA structures. Nucleic 
Acids Res. 20, Suppl., 2095-2109. 
Gutell, R. R., WeiSel", B., Woese, C. R., Noller, H. F. (1985). Comparative anatomy of 16-S-like ribosomal RNA Prog. Nucleic Acids 
Res. 32, 155-216. 
Gyllensten, U. B., Erlich, H. A (1988). Generation of single-stranded DNA by the polymerase chain reaction and its application to 
direct sequencing of the HlA-DQ). locu;s. Proc. Natl. Acad. Sci. USA 85, 7652-7656. 
Gyllensten, U., Wharton, D., Josefsson, A., Wilson, A C. (1991). Paternal inheritance of mitochondrial DNA in mice. Nature 352, 255-
257. 
Haase, P. (1992). Isozyme variability and biogeography of Nothofagus truncata (Fagaceae). N.Z J. Bot. 30, 3 15-328. 
Haffer, J. (1969). Speciation in Amazonian forest birds. Science 165, 13 1-137. 
Hahn, B. H.,  Shaw, G. M.,  Taylor, M. E., Redfield, R. R., M¥kham, P. D.,  Salahuddin, S. Z., Wong-Staal, F., Gallo, R. C., Parks, E. S., 
Parks, W. P. (1986). Genetic variation in H1LV-IIIILAV over time in patiehts with AIDS or at risk for AlDS.!��ce 232, 1548-1553. 
Hair, J. B. ( 1966). Biosystematics of the New Zealand flora, 1945- 1964. N.Z J. Bot. �, 559-595. 
Hair, J. B. (1977). Contribution to a chromosome atlas of the New Zealand flora- 1 8  Epilobium (Onagraceae). N.Z. J. Bot.�, 5-15. 
Hardy, G. S. (1977). The New Zealand Scincidae (Reptilia: Lacertilia); a taxonomic and zoogeographic study. N.Z J. Zool. � 221-325. 
Hardy, G. S. (1979). The karyotypes ofawo scincid lizards, and their bearing on relationships in genus Leiolopisma and its relatives 
(Scincidae: Lygosominae). N.Z J. ZooL §, 609-612. 
Harrison, R. G. (1989). Animal mitochondria) DNA as a genetic marker in population and evolutionary biology. Trends EcoL EvoL i, 6-
1 1. 
Harrison, R. G., Doyle, J. J. (1990). Redwoods break the rules. Nature 344, 295-296. 
Hasegawa, M, Hashimoto, T. (1993). Ribosomal RNA trees misleading? Nature 361 , 23. 
Hawkins, B. J., Sweet, G. B. (1989). Genetic variation in rimu-an investigation using isozyme analysis. N.Z J. Bot. 27, 83-90. 
Hay, R. F., Mutch, A R., Watters, W. A. (1970). Geology of the Chatham Islands. N. Z GeoL Survey Bulletin 83, 1-86. 
Hedges, S. B., Bezy, R. L., Maxson, 1.. R. (1991). Phylogenetic relationships and biogeography of xantusiid lizards, inferred from 
mitochondrial DNA sequences. Mol. BioI. Evol. �, 767-780. 
Hedges, S. B., Moberg, K. D., Maxson, L. R. (1990). Tetrapod phylogeny inferred from 1 8S and 28S ribosomal RNA sequences and a 
review of the evidence for amniote relationships. Mol. BioI. Evol. Z, 607-633.  
Helm-Bychowski, K. M, Wilson, A.  C.  (1986). Rates of nuclear DNA evolution in  pheasant-like birds: evidence from restriction maps. 
Proc. Natl. Acad. Sci. USA 83, 688-692. 
Henderson, I. M., Hendy, M. D., Penny,. D. (1989). Influenza viruses, comets and the science of evolutionary trees. J. Theor. Bioi . .!iQ. 
289-303. 
Hendy, M. D., Penny, D. (1989). A framework for the quantitative study of evolutionary trees. Syst. Zoal. 38, 297-309. 
Hendy, M. D., Penny, D. (1993 in press). Spectral analysis of phylogenetic data. Cladistics 10. 
Heus, H. A., Pardi, A (1991). Structural features that give rise to the unusuai stability of RNA hairpins containing GNRA loops. 
Science 253, 191-194. 
page 109. 
Hickson, R. E., Penny, D., Scott, D. B. (1.992). Molecular systematics and evolution in New Zealand: applications to cryptic skink 
species. N.Z. J. ZooL 19, 33-44. 
Hill, W. E., Dahlbug, A., Garrett, R.A., Moore, P.B., Schlessinger, D., Warner, 1.R. (eds.) ( 1990). The ribosome: structure,function, 
and evolution. American Society fOt Microbiology, Washington, D.C. 
Hillis, D. M (1984). Misuse and modification of Nei's genetic distance. Syst. ZooL 33, 238-240. 
Hillis, D. M. (1987). Molecular versus m.orphological approaches to systematics. Ann. Rev. &01. Syst . .!!' 23-42. 
Hillis, D. M. (1991). Discriminating between phylogenetic signal and rail(!om noise in DNA sequences. In: Miyamoto, M.M., Cracraft, 
1. (eds.), Phylogenetic analysis of DNA sequences. Oxford University Press. Pp. 278-294. 
Hillis, D. M, Davis, S. Ie. (1987). Evolution of the 28S ribosomal RNA-ge�e in anurans: regions of variability and their phylogenetic 
implications. MoL BioL EvoL i, 1 17-125. 
Hillis, D. M., Dixon, M. T. (1991). Ribosomal DNA: molecular evolution and phylogenetic inference. Quart. Rev. Bioi. 66, 41 1-453. 
Hillis, D. M, Moritz, C. ( 1990). An overview of applications of molecular systematics. In: Hillis, D.M., Moritz, C. (eds.), Molecu14r 
systematics. Sinauer Associates, Inc. SUDlderland, Massachusetts. pp. 502-5 15. 
Hitchmough, R. A., Williams, M., Daugherty, C. H. (1990). A genetic analysis of maIlards, grey ducks, and their hybrids in New 
Zealand. N.Z. J. ZooL g, 467-472. 
Hixson, 1. E., Brown, W. M. (1986). A comparison of the small ribosomal RNA genes from the mitochondrial DNA of the great apes 
and humans: sequence, structure, evolution, and phylogenetic implications. Mol. Bioi. Evol. 2, 1-18 .  
Hoeh, W. R., Blakely, Ie. H., Brown, W. M.  (1991). Heteroplasmy suggests limited biparental inheritance of Mytilus mitochondrial 
DNA. Science 251 ,  1488-1490. 
Holdaway, R. N. ( 1990). Qlanges in the diversity of New Zealand forest birds. N.Z. J. Zool. 17, 309-321. 
Holloway, B. A. (1963). Wing development and evolution of New Zealand Lucanidae (Insecta: Coleoptera). Trans. Roy. Soc. N.Z. �, 
99- 1 16. 
Holmquist, R. (1983). Transitions and trllosversions in evolutionary descent: an approach to understanding. J. MoL Evol. 19, 1 34-144. 
Holmquist, R., Pearl, D. (1980). Theoretical foundations fOt quantitative pa1eogenetics. Part IIi. The molecular divergence of nucleic 
acids and proteins for the case of genetic events of unequal probability. J. Mol. Evol. 16, 21 1-267. 
Hooker, I. D. (1 853). Introductory essay .. Flora Novae-Zelandiae. 1 Cramer, Weinheim (1963 reprint). 
Hotz, H., BeerH, P., Spolsky, C. (1992). Mitochondrial DNA reveals formation of nonhybrid frogs by natural matings between 
hemic10nal hybrids. Mol. Bioi. Evol. 2, 610-620. 
Hudson, G. V. (1920). On some examplc:s of New Zealand insects illustrating the Darwinian principle of sexual selection. Trans. Proc. 
N.Z. Inst. 52, 43 1-438. 
Hutchinson, M. N., Donne\1an, S. C., Baverstock, P. R., Krieg, M., Simms, S., Burgin, S. (1990). Immunological relationships and 
generic revision of the Australian lizards assigned to the genus Leiolopisma (Scincidae: Lygosominae). Aust. J. Zool. 38, 535-554. 
Hutton, F. W. (1872). On the geographical relations of the New Zealand fauna. Trans. Proc. N.Z. lnst. �, 227-256. 
Huxley, I. (1938). Evolution, the mode/7il synthesis. Allen & Unwiii, London. 
Ingram, V. M (1961). Gene evolution in haemoglobins. Nature 1 89, 704-708. 
Innis, M.A., Gelfand, D.H. (1990). Optimization of PCRs. In: Innis, M.A., Gelfand, D.H., Sninsky. 1.1., White, T.1. (eds.), peR 
protocols: a guide to methods and applications. Academic Press, Inc. Pp. 3-12. 
-
page 1 10. 
Irving, G. 1. (1967). Distribution and po!sible factors in evolution of some New Zealand grasshoppers. Zool. Hons. Pan III Thesis, 
University of Canterbury. 
Irwin, D. M., Kocher, T. D., Wilson, A. C. (1991). Evolution of the cytochrome b gene ofmamrna1s. J. MoL Evol. 32, 128-144. 
Jeffreys, A. 1., Neuman, R., Wilson, V. (1990). Repeat unit sequence variation in minisatellites: A novel source of DNA polyJDOqirism 
for studying variation and mutation by si.ngle molecule analysis. CeU 60, 473-485. 
Jukes, T. H. (1980). Silent nucleotide sulbstitutions and the molecular evolutionary clock. Science 210, 973-978. 
Jukes, T. H. (1987). Transitions, transveuions, and the molecular evol�tionary clock. J. Mol. Evol. 26, 87-98. 
Jukes, T. H., Cantor, C. R. (1969). Evolution of protein molecules. In: Mumo, H.N. (ed.), Mammalian protein metabolism. Academic 
Press, New York. Pp. 21-132. 
Killersjo, M., Farris, J. S., Kluge, A. G., Bull, c. (1992). Skewness and permutation. CkuJistics �, 275-287. 
Kessler, L G., Avise, J. C. (1985). A comparative description of mitochondrial DNA differentiation in selected avian and other 
vertebrate genera. MoL BioI. Evol. �, 109-125. 
Kimura, M (1981). Estimation of evolullionary distances between homologous nucleotide sequences. Proc. NatL Acad. Sci. USA 78, 
5773-5777. 
Kimura, M. (1983). The neutral theory afmolecular evolution. Cainbridge University Press, Cambridge. 
King, M. (1979). Karyotypic evolution in Gehyra (Gekkonidae: Reptilia). I. The Gehyra variegata-punctata compleL Aust. J. Zaol. 27, 
373-393. 
King, M. (1987). Chromosomal evolution in diplodactylinae (Gekkonidae: Reptilia). I. Evolutionary relationships and patterns of 
change. Aust. J. Zaol. 35, 507-531 .  
King, M., Mengden, G.  (1990). Chromosomal evolution in  diplodactylinae (Gekkonidae: Reptilia). II. Chromosomal variability between 
New Caledonian species. Aust. J. ZaoL :�, 219-226. 
King, M.-C., Wilson, A. C. (1975). Evolution at two levels in humans and chimpanzees. Science 1 88, 107- 1 16. 
Kitchin, P.A., Szotyori, Z., Frornholc, c., Almond, N. (1990). Avoidance offalse PCR positives. Nature 344, 201 .  
Kluge, A. G.  (1967). Systematics, phyJc'geny, and zoogeography of the lizard genus Diplodactylus Gray (Gekkonidae). Aust. J. ZooL 
�, 1007- 1 108. 
Knight, A., Mindell, D. P. (1993). Substitution bias, weighting of DNA sequence evolution, and the phylogenetic position of Fea's 
viper. Syst. BioL 42, 18-3 1 .  
Kocher, T. D., Thomas, W. K., Meyer, A., Edwards, S .  V., Pubo, S., Villablanca, F. X., Wilson, A. C.  (1989). Dynamics of 
mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proc. NatL Acad. Sci. USA M. 6196-
6200. 
Kocher, T. D., White, T. J. (1989). Evolutionary analysis via PCR. In: Erlich, H.A. (ed.), PCR technology. Principles and applications 
for DNA amplification. Stockton Press, New York. Pp. 137-147. 
Kondo, R., Satta, Y., Matsuura, E. T., hhiwa, H., Takahata, N., Chigusa, S. 1. (1990). Incomplete maternal transmission of 
mitochondrial DNA in Drosophila. Genetics 126, 657-663. 
Koop, B. F., Tagle, D. A., Goodman, M., Slightom, J. L (1989). A molecular view of primate phylogeny and important systematic and 
evolutionary questions. Mol. BioL EvoL Q, 580-612. 
Kraus, F., Jarecki, L, Miyamoto, M. M., Tanhauser, S .  M., Laipis, P. J. (1992). Mispairing and compensational changes during the 
evolution of mitochondrial ribosomal RN A. Mol. BioI. EvoL 2, 770-774. 
page I l l .  
Kraus, F., Miyamoto, M. M (1991). Rapid c1adogenesis among the pecoran ruminants: evidence from mitochondrial DNA sequences. 
Syst. Zool. 40, 1 17- 130. 
Kuschel, G. (ed.) (1975). Biogeography ,and ecology in New Zealand. Dr W. Junk b. v. Publishers, The Hague. 
Kwiatowski, J., Skarecky, D., Hernande:� S., Pham, D., Quijas, F., Ayala, F. J. (1991). High fidelity of the polymecase chain reaction. 
Mol. BioL Evol. !, 884-887. 
Lake, J. A. (1990). Origin of the MetazoL Proc. NaIL Acad. Sa. USA 87, 763-766. 
Lamb, T., Avise, J. C. (1986). Directionnl introgression of mitochondrial DNA in a hybrid population of tree frogs: the influence of 
mating behavioc. Proc. NatL Acad. Sa. USA. 83, 2526-2530. 
Larson, A. (1984). Neontological inferences of evolutionary pattern and process in the salamander family P1ethodontidae. EvoL BioL 17, 
1 19-217. 
Larson, A., Wilson, A. C. (1989). PattenlS of ribosomal RNA evolution in salamanders. MoL BioI. EvoL �, 1 3 1-154. 
Lawver, L. A., Royer, J.-Y., Sandwell, D. T., Scotese, C. R. (1991). Evoluuon of the Antarctic continental margins. In: Thomson, 
M.R.A., Crame, J.A., Thomson, J.W. (eds.), Geological evolution of Antarctica. Camlxidge University Press, Cambridge. pp. 533-539. 
Lessler, M.A., Brierly, G.P. (1969). Oxygen electrode measurements in biochemical analysis. Meth. Biochem. Anal.!1, 1-29. 
Lewontin, R. C. (1974). The genetic bas/oS 0/ evolutionary change. Columbi� University Press, New Yorlc. 
Lewontin, R. C., Moore, J. A., Provine, W. B., Wallace, B .  (cds.), (1981). Dobzhansky's Genetics of natural populations I-XllII. 
Columbia University Press. 
Li, W.-H., Gouy, M. (1991). Statistical methods for testing molecular phylogenies. In: Miyamoto, MM., Cracraft, J. (cds.), 
Phylogenetic analysis 0/ DNA sequenceJ'. Oxford University Press. pp. 249-277. 
Li, W.-H., Tanimura, M., Sharp, P. M (1987). The molecular clock fUns more slowly in man than in apes and monkeys. Nature 326, 93-
96. 
Lindahl, T., Nyberg, B. (1972). Rate of depurioation of native deoxyribonucleic acid. Biochem. !!, 3610-3618. 
Uoyd, D. G. (1982). Variation and evolution of plant species on the outlying islands of New Zealand, with particular reference to 
Cotulafeatherstonii. Taxon n, 478-487. 
Uoyd, D. G., Webb, C. J. (1986). The avoidance of interference between the presentation of pollen and stigmas in angiosperms. I. 
Dichogamy. N.Z 1. Bot. 24, 1 35- 162. 
Uoyd, D. G., Yates, J. M. A. (1982). IntJ:asexual selection and the segregation of pollen and stigmas in hermaphrodite plants, 
exemplified by Wahlenbergia albomargl'nata (Campanulaceae). Evolution 36, 903-913. 
Lockhart, P. J., Peony, D., Hendy, M. H., Howe, C. J., Beanland, T. J., Larkurn, A. W. D. (1992). Controversy on cblocoplast origins. 
FEBS Lett. 301,  127- 13 1 .  
Lonsdale, D .  M, Brears, T., Hodge, T. P., Melville, S. E., Rottman, W.  H .  (1988). The plant mitochondrial genome: homologous 
recombination as a mechanism for generating heterogeneity. PhiL Trans. Roy. Soc., LoruL B 3 19, 149-163. 
Lyell, C. (1830). Principles of geology or the tntXkm change of the earth and its inhabitants considered as illustrative of geology. John 
Murray, London. 
Manske, C. L., Chapman, D. J. (1987). Nonuni!ormity of nucleotide substitution rates in molecular evolution: computer simulation and 
analysis of 5S ribosomal RNA sequences. 1. Mol. EvoL 26, 226-251.  
Markham, K.  R.,  Godley, E.J. (1972). Olemotaxonomic studies in Sophora. 1 .  An evaluation of Sophora microphyUa Ait. N.Z 1. Bot. 
10, 627-640. 
page 1 12. 
Marshall, C. R. (1992). Substitution bias, weighted parsimony, and amniote phylogeny as inferred from 18S rRNA sequences. MoL 
Bioi. EvoL 2, 37()'373. 
Martin, A. P., Kessing, B. D., Palumbi, S. R. (1990). Accuracy of estimating genetic distance between species from short sequences of 
mitochondrial DNA. MoL BioL EvoL 2, 485-488. 
Martin, A. P., Naylor, G. 1. P., Palumbi, S. R. (1992). Rates of mitochondrial DNA evolution in sharks are slow compared with 
mammals. Nature 357, 153- 155. 
Martin, A. P., Palumbi, S.  R. (1993 submitted). Body size, metabolic rate, genecation time and the m�ecular clock. !roc. NatL Acad. 
Sci. USA. 
Martin, W., Lydiate, D., BrinIcmann, H., Forkmann, G., Saedler, H., Cerff, R. (1993). Molecular phylogenies in angiosperm evolution. 
Mol. BioL Evol. 10, 14()'162. 
Maxson, L R., Maxson, L. R. (1979). Comparative albumin and biochemical evolution in Plethodontid salamanders. Evolution 33, 
1057-1062. 
Maxson, L R., Maxson, R. D. (1990). Proteins II: immunological techniques. In: Hillis, D.M., Moritz, C. (cds.), Molecular systematics. 
Sinauer Associates, Inc. Sunderland, Ma5sachusetts. Pp. 127-155. 
May, R. M. (1990). Taxonomy as destiny. Nature 347, 129-130. 
Mayes, C. L., Lawver, L. A., Sandwell, D. T. (1990). Tectonic history and new isochron chart of the South Pacific. J. Geophys. Res. 95, 
8543-8567. 
Mayr, E. (1942). Systematics and the origin 0/ species. From the viewpoint 0/ a zoologist. Columbia University Press, New York. 
Mayr, E. (1963). Animal species and evolution. Belknap Press of Harvard University Press, Cambridge, Mass. 
Mayr, E. (1982). The growth 0/ biologicGrl thought. The Belknap Press of Harvard Univecsity Press, Cambridge, Massachusetts. 
Mayr, E., O'Hara, R. I. (1986). The biog,wgraphic evidence supporting the t'Ieistocene forest refuge hypothesis. Evolution 40, 55-67. 
McCann, C. (1955). The lizards of New Zealand. Gekkonidae and Scincidae. Dom. Mus. BulL E, 1-127. 
McDougall, I., Chamalaun, F. H. (1969). Isotopic dating and geomagnetic polarity studies of volcanic rocks from Mauritius, Indian 
Ocean. Bull. Geol. Soc. Am. 80, 1419-14t2. 
McGlone. M.S. (1983). Polynesian defor,l:Station of New Zealand: a pceliminary synthesis. Archaeol. Oceania ll, 1 1-25. 
McGlone, M. S. (1985). Plant biogeography and the late Cenozoic history of New Zealand. N.Z. J. Bot. 23, 723-749. 
McKenzie, D. P., Sclater, 1. G. (1973). 11le evolution of the Indian Ocean. Sci. Am. 228, 62-72. 
McLea, M. C., Lambert, D. M. (1985). Cytogenetics of New Zealand blackflies of the genus Austrosimulium (Diptera: Simuliidae). II. 
Heterozygote deficiency and non-random association of inversion hetecozygotes. Genetica 66, 203-21 1 .  
Meyer, A., Dolven, S .  I .  (1992). Molecul,�, fossils, and the origin of tetrapods. J. MoL Evol. 35, 102- 1 1 3. 
Meyer, A., Kocher, T. D., Basasibwaki, P., Wilson, A. C. (1990). Monophyletic origin of Lake Victoria cichlid fishes suggested by 
mitochondrial DNA sequences. Nature �!7., 55()'553. 
Mildenhall, D. C. (1980). New Zealand h.te Cretaceous and Cenozoic plant biogeography: a contribution. Palaeogeogr. PalaeoclimatoL 
PalaeoecoL 11, 197-233.  
Milinkovitch, M. C., Orti, G., Meyer, A.  (1993). Revised phylogeny of whales suggested by mitochondrial ribosomal DNA sequences. 
Nature 361, 346-348. 
Millar, C. D., Lambert, O. M., Bellamy, A. R., Stapelton, P. M., Young, E. C. (1992). Sex-specific restriction fragments and sex ratios 
revealed by DNA fingerprinting in the brown skua. J. Heredity 83, 35()'355. 
MindeIl, D. P. (1991). Aligning DNA sequences: homology and phylogenetic weighting. In: Miyamoto, M.M., Cracraft, 1. (cds.), 
page 1 13. 
Phylogenetic analysis of DNA sequenceJ'. Oxford University Press. pp. 73-89. 
Mindel!. D. P .• Honeycutt, R. I.. (1990). Ribosomal RNA in vertebrates: evolution and phylogenetic applications. Ann. Rev. EcoL Syst. 
21. 541-566. 
Miyamoto. M. M .• Boyle. S. M. (1989). 'The potential importance of mitochondrial DNA sequence data to eutherian mammal 
phylogeny. In: Fernholm, B .• Bremer. K . •  Jomvall. H: (eds.). The hierarchy of life. Molecules and mcrphology in phylogenetic analysis. 
Elsevier Science Publishers b.v. Amsterdam. Pp. 437-450. 
Miyamoto. M. M .• Goodman. M. (1990)" DNA systematics and evolution of priinates. Ann. Rev. EcoL_Syst. 21.  197:2220. 
Molloy. B. P. J .• Burrows. C. J .• Cox. 1. E .• Johnston, 1. A .• Wardle. P. (1963). Distribution of subfossil forest remains. eastern South 
Island, New Zealand. N.Z. J. Bot. !. 68-77. 
Molnar. R. E. (1991). Fossil reptiles in Australia. In: Vickers-Rich. P .• Monaghan. J.M .• Baird, R.F .• Rich. T.H. (eds.). VertebraJe 
paleontology of Australasia. Pioneer De.\:ign Studio Pty Ltd. Melbourne. Pp. 605-702. 
Moritz, C. (1986). The population biology of Gehyra (Gekkonidae): chromosome change and speciation. Syst. ZooL 35. 46-67. 
Moritz, C. (1991). Evolutionary dynamics of mitochondrial DNA duplications in parthenogenetic geckos. Heteronotia binoei. Genetics 
129. 221-230. 
Moritz, C .• Case. T. J., Bolger. D. T .• Dol nellan, S. (1993). Genetic diversity and the history of pacific island h'ou1'e geckos 
(Hemidactylus and Lepidodactylus). BioI. J. Linn. Soc. 48. 1 13-133. 
Moritz, C., Dowling. T. E .• Brown. W. M. (1987). Evolution of animal mitochondrial DNA: relevance for popUlation biology and 
systematics. Ann. Rev. EcoL Syst. !!. 269-292. 
Moritz, C .• Hillis. D. M. (1990). Molecular systematics: context and controversies. In: Hillis. D.M .• Moritz, C. (eds.). Molecular 
systematics. Sinauer Associates. Inc. Sunderland. Massachusetts. pp. 1- 10. 
Moritz, C .• Schneider. C. 1., Wake. D. B. (1992). Evolutionary relationships within the Ensarina eschscholtzji complex confirm the ring 
species interpretation. Syst. BioL �. 273·291.  
Moritz, C. C . •  Wright, J .  W . •  Brown. W.M. (1989). Mitochondrial-DNA analyses and the origin and relative age of parthenogenetic 
lizards (Genus Cnemidophorus). III. C. vdox and C. exsanguis. Evolution 43. 958-968. 
Morris. R. W. (198 1). Effect of of a broad temperature range on the metabolic responses of the eurythermic lizard Leiolopisma 
zelandica. Compo Biochem. PhysioL 70A, 623-626. 
Morton. S. R.. James, C. D. (1988). The ctiversity and abundance of lizards in arid Australia: a new hypothesis. Am. Nat. 132. 237-256. 
Murphy. R. W .• Matson, R. H. (1986). Evolution of isozyme characters in the tuatara, Sphenodon punctatus. N.Z. J. ZooL Q. 573-58 1 .  
Murphy. R. W . •  Sites. 1 .  W .• Buth. D. G . •  Haufler. C. H .  (1990). Proteins I :  isozyme electrophoresis. In: Hillis. D.M. Moritz, C. (eds.). 
Molecular systematics. Sinauer Associates, Inc. Sunderland. Massachusetts. Pp. 45-126. 
Nagae. Y .• Fujii. H .• Yoneyama. Y .• Goto. Y .• Okazaki. T. (1988). Nucleotide sequence of the Rana catesbeiana mitochondrial small 
(12S) and large (16S) ribosomal RNA gelles. Nucleic Acids Res. 16. 10363. 
Neefs. J.-M .•  Van de Peer, Y .• Hendriks. L. De Wachter, R. (1990). Compilation of small ribosomal subunit RNA sequences. Nucleic 
Acids Res . .!!. SuppJ.. 2237-23 17. 
Neefs. J.-M. Van de Peer. Y .• De Rijk, P., Goris. A .• De Wachter. R. (1991). Compilation of sma)) ribosomal subunit RNA sequences. 
Nucleic Acids Res. 19. SuppJ.. 1987-2015. 
Nei. M. (1972). Genetic distance between populations. Am. Nat. 106. 283-292. 
Nei. M. (1987). Molecular evolutionary g.�netics. Columbia University Press. 
page 1 14. 
Nei, M. (1991). Relative efficiencies of different tree-making methods for molecular data. In: Miyamoto, M.M., Craaaft, J. (cds.), 
Phylogenetic analysis of DNA seqUUlces. Oxford University Press. pp. 90-128. 
Nei, M, Gojobori, T. (1987). Simple methods for estimating the numbecs of synonymous and nonsynonymous nucleotide substitutions. 
Mol. Bioi. Evol. �, 418-426. 
Nei, M., Tajima, F., Tateno, Y. (1983). Accuracy of estimated phylogenetic trees from molecular data. II. Gene frequency data. J. MoL 
EvoL 19, 153-170. 
Neigel, J. E., Avise, 1. C. (1986). Phylogenetic relationships of mitochondrial DNA under various demographic models of speciation. 
In: Karlin, S., Nevo, E. (cds.), Evolutionary processes and theory. Academic Press, Orlando. Pp. 515-534. 
Noller, H. F. (1984). Structure of ribosomal RNA. Ann. Rev. Bwchem. 53, 1 19- 162. 
Noller, H. F. (1991). Ribosomal RNA and translation. Ann. Rev. Bwchem. 60, 191-227. 
Noller, H. F., Moazed, D., Stem, S., Powers, T., Allen, P. N., Robertson, J. M., Weiser, B., Triman, K. (1990). Structure of rRNA and 
its functional interactions in translation. In: Hill, W.E., Dahlberg, A., Garrett, R.A., Moore, P.B., Schlessinger, D., Warner, J.R. (cds.), 
The ribosome: structure, junction, and evolution. American Society for Microbiology, Washington, D.C. pp. 73-92. 
Noller, H. F., Woese, C. R. (1981). Secondary structure of 16S riboso� RNA. Science 212, 403. 
O'Brien, T. W., Pietromonaco, S. F., Piatyszek, M. A., Gillevet, P. M. (1985). Genes for mammalian mitoribosomal proteins. In: 
Quagliariello, E., Slater, E.C., Palmieri, F., Saccone, c., Kroon, A.M. (cds.), Achievements and perspectives of mitoclwndri4l research. 
Volume II. Biogenesis. Elsevier Science Publications b.v. (Biomedical Division). pp. 379-387. 
Ochman, H., Wilson, A. C. (1987). Evolution in bacteria: evidence for a universal substitution rate in cellular genomes. J. MoL Evol. 26, 
74-86. 
Ohta, T. (1976). Role of very slightly deleterious mutations in molecular evolution and polymorphism. Theor. Pop. Bioi. 10, 254-275. 
Ohta, T. (1992). The nearly neutral theory of molecular evolution. Ann. Rev. &01. Syst. 23, 263-286. 
Olmo, E. (1991). Genome variations in the transition from amphibiallS to reptiles. J. MoL Evol. 33, 68-75. 
Omduff, R. (1964). Evolutionary pathways of the Senecw /autus alliance in New Zealand and Australia. N.z. 1. Bot . .!!. 349-360. 
Paabo, S., Thomas, W. K., Whitfield, K. M., Kumazawa, Y., Wilson, A. C. (1991). Rearrangements of mitochondrial transfer RNA 
genes in marsupials. J. MoL Evol. 33, 426-430. 
Paige, K. N., Capman, W. C., Jennetten, P. (1991). Mitochondrial inheritance patterns across a cottonwood hybrid zone: cytonuclear 
disequilibria and hyirid zone dynamics. Evolution 45, 13@1369. 
Palmer, J. D. (1990). Contrasting modes and tempos of genome evolution in land plant organelles. Trends Genet. 2, 1 15-120. 
Palmer, J. D., Herbon, L. A. (1988). Plant mitochondrial DNA evolves rapidly in structure but slowly in ·sequence. J. MoL EvoL 28, 87-
97. 
Palumbi, S. R. (1989). Rates of molecular evolution and the fraction of nucleotide positions free to vary. 1. MoL Evol. 29, 1 8�187. 
Palumbi, S. R., Baker, C. S. (1993 submitted). Opposing views of humpback whale poplation structure using mitochondrial and nuclear 
DNA sequences. Proc. Natl. Acad. Sci. USA. 
Palumbi, S., Martin, A., Romano, S., McMillan, W. O., Stice, L, GraboWSki, G. (1991). The simple fool's guide to peR. Version 2.0. 
Dept. Zoology and Kewalo Marine Laboratory, University of Hawaii. 
Pamilo, P., Nei, M. (1988). Relationships between gene trees and species trees. Mol. Bioi. Evol. �, 568-583. 
Pashley, D. P., Ke, L. D. P992). Sequence evolution in mitochondrial ribosomal and ND-I genes in Lepidoptera: Implications for 
phylogenetic analyses. Mol. BwL EvoL 2, 1061-1075. 
page 1 15. 
Patterson, C. (1987). Introduction. In: Patterson, C. (ed.), Molecules and morphology in evolution: conflict or compromise? Cambridge 
Univel"sity Press, Cambridge. 
Patterson, C. (1990). Reassessing relationships. Nature 344, 199-200. 
Patterson, G. B. (1985). The ecology &lid taxonomy of the common skink Leiolopisma nigriplantare maccanni in tussock grasslands in 
Otago. Unpublished PhD thesis, University of Otago. 
Patterson, G. B., Daugherty, C. H. (1990). Four new species and one new subspecies of skinks, genus.Leiolopisma (Reptilia: Lacertilia: 
Scincidae) from New Zealand. J. Roy. Soc. N.7.. 20, 65-84. 
Peattie, D. A., Douthwaite, S., Garrett, R. A., Noller, H. F. (1981 ). A "bulged" double helix in a RNA-protein contact site. Proc. NatL 
Acad. Sci. USA 78, 733 1-7335. 
Penny, D., Foulds, L R., Hendy, M. D. (1982). Testing the theory of evolution by comparing phylogenetic trees constructed from five 
different protein sequences. Nature 297, 197-200. 
Penny, D., Hendy, M. D., Steel, M. A. (1991). Testing the theory of descenL In: Miyamoto, M.M., Cracraft, J. (eds.), Phylogenetic 
analysis of DNA sequences. Oxford University Press. pp. 155-183. 
Penny, D., Hendy, M. D., Steel, M. A. (1992). Progress with methods for constructing evolutionary trees. TRE!
_
�73�79. 
Penny, D., Hendy, M. D., Zimmer, E. A., Hamby, R. K. (1990). Trees from sequences: panacea or Pandora's box? Aust. Syst. Bot. 2., 21-
38. 
Penny, D., Hickson, R. E., Lockhart, P. J., Watson, E. E. (1993 submitted). Recent progress with methods for evolutionary trees. N.7.. J. 
Bot. 
Petel"son, A. J. (1968). Geographical variation in Paprides nitidus (Hutton) (Orthoptera: Acrididae). N.7.. J. Sci . .!!, 693-705. 
Phillips, N. R., Lambert, D. M. (1990). A cladistic analysis of species of the molluscan genus Potamopyrgus based on aIlozyme data. 
N.7.. J. ZooL 17, 257-263. 
Pickard, C. R., Towns, D. R. (1988). Atlas of the amphibians and reptiles of New Zealand. Conservation Sciences Publication No. 1 .  
Science and Research Directorate, Department of Conservation, Wellington . .  
Pierson, E. D., Sarich, V. M., Lowenstein, J. M., Daniel, M. J., Rainey, W. �. (1986). A molecular link between the bats of New 
Zealand and South America. Nature 323, 60-63. 
Pocknall, D. T. (1989). Late Eocene to early Miocene vegetation and climate history of New Zealand. J. Roy. Soc. N.7.. 19, 1 -18. 
Pocknall, D. T. (1990). Palynological evidence for the early to middle Eocene vegetation and climate history of New Zealand. Rev. 
Palaeobot. Palynol. 65, 57-69. 
Pollock, M., McAvoy, E.S. (1978). Morphological and metabolic manges in muscles ofhibemating lizards. Copeia 1978, 412·416. 
Portel", P. (1987). An ecological comparison of two Cyclodina skinks (Reptilia: Lacertilia) in Auckland, New Zealand. N.Z. J. ZooL 14, 
493·507. 
Powell, A. W. B. (1949). The species population in New Zealand land snails. Trans. Roy. Soc. N.7.. 77, 202·208. 
Powell, J. R. (1975). Protein variation in natural populations of animals. Evo!. Bioi. �, 79·1 19. 
Pregill, G. K. (1989). Prehistoric extinction of giant iguanas in Tonga. Copeia 1989, 505·508. 
Rattenbury, J. A. (1962). Cyclic hybridization as a survival mechanism in the New Zealand forest flora. Evolution 16, 348· 363. 
Raven, P. H. (1972). Evo!ution of subalpine and alpine plant groups in New Zealand. N.7.. 1. Bot . .!!' 177·200. 
Raven, P. H. (1988). Onagraceae as a model of plant evolution. In: Gottlieb, LD., Jain, S.K. (cds.), Plant evolutionary biology. 
Olapman & Hall Ltd, London. pp. 85-107. 
page 1 16. 
Raven, P. H., Raven, T. E. (1976). The genus EpiJobium in Australasia: a systematic and evolutionary study. N.Z. Dept. Sci. Res. Bull. 
216, 1-321.  
Rendle, H., Murray, B. G.  (1989). Olromosome relationships and breeding barriers in New Zealand species of Ranunculus. N.Z. J. Bot. 
27, 437-448. 
Richter, C., Park, J.-W., Ames, B. N. (1988) Nonnal oxidative damage to mitochondrial and nuclear DNA is extensive. Proc. Natl. 
Acad. Sci. USA 85, 6465-6467. 
Ricklefs, R. E., Cockran, D., Pianka, E. R. (1981). A morphological analysis of the structure of communities of lizards in desert 
habitats. Ecology 62, 1474-1483. 
- -
Robb, J. (1980). New Zealand amphibians and reptiles in colour. Collins. 
Roberts, J. D., Maxson, L. R. (1985). Tertiary speciation models in Australian anurans: molecular data challenge Pleistocene scenario. 
Evolution 39, 325-334. 
Roe, B. A., Ma, D.-P., Wilson, R. K., Wong, J. F.-H. (1985). The complete nucleotide sequence of the Xenopus laevis mitochondrial 
genome. J. BioL Chem. 260, 9759-9774. _ ._ 
Rogers, S. 0., Bendich, A. J. (1985). Extraction of DNA from milligram amounts of fresh, herbarium and mummified plant tissues. 
Plant molecular biology 1, 69-76. 
Rohlf, F. 1., Olang, W. S., Sokal, R. R., Kim, J. (1990). Accuracy of estimated phylogenies: effects of tree topology and evolutionary 
model. Evolution 44, 1671- 1684. 
Ross, H. A. (1983). Genetic differentiation of starling (Stumus vulg'lris: Aves) popUlations in New Zealand and Great Britain. J. ZooL, 
Land. 201, 351-362. 
Saccone, C., Pesole, G., Sbisa, E. (1991). The main regulatory region of mammalian DNA: structure-function model and evolutionary 
pattern. J. MoL EvoL 33, 83-9 1 .  
Sadlier, R.  A. (1986). A review of the scincid lizards of New Caledonia. Rec. Aust. Mus. 39, 1-66. 
Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Hom, G.T., Mullis, K.B., Erlich, H.A. (1988). Primer-directed 
enzymatic amplification of DNA with a thennostable DNA polymerase. Science 239, 487-491 .  
Saitou, N., Nei, M.  (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. MoL BioL EvoL i. 406-
425. 
Sarnbrook, J., Fritsch, F. F., Maniatis, T. (1989). Molecular cloning. A laboratory manual. Second edition. Cold Spring Harbor 
Laboratory Press. 
Satta, Y., Talcahata, N. (1990). Evolution of Drosophila mitochondrial DNA and the history of the melanogaster subgroup. Proc. Natl. 
Acad. Sci. USA 87, 9558-9562. 
Schaaper, R. M., Dunn, R. L (1987). Spectra of spontaneous mutations in Escherichia coli strains defective in mismatch correction: the 
nature of in vivo DNA replication errors. Proc. Natl. Acad. Sci. USA 84, 6220-6224. 
Schmidt-Nielsen, K. (1986). Scaling. Why is animal size so imp0ryJnt? Cambridge University Press. 
Sheffield, V. C., Cox, D. R., Myers, R. M. (1990). Identifying DNA polymorphisrns by denaturing gradient gel electropha-esis. In: 
Innis, M.A., Gelfand, D.H., Sninski, JJ., White, T.J. (cds.), PCR protocols: A guide to methods and applications. Academic press, Inc. 
San Diego. Pp. 206-218. 
Shields, G. F., Wilson, A. C. (1987). Caliocation of mitochondrial DNA evolution in geese. J. Mol. Evol. 24, 212-217. 
page 1 17. 
Shigenaga, M. K., Gimeno, C. J., Ames, B. N. (1989). Urinary 8-hydroxy-2'-deoxyguanosine as a biological marka of in vivo oxidative 
DNA damage. Proc. NalL Acad. Sci. USA 86, 9697-9701 .  
Shigenaga, M. K., Park, J.-W., Cundy, K .  C., Gimeno, C. J., Ames, B.  N .  (1992). In vivo oxidative DNA damage: measurement of 8-
hydroxy-2'-deoxyguanosine in DNA and urine by high-performance liquid chromatography with electrochemical detection. Meth. 
Enzymol. 186, 521-530. 
Shoemaker, J. S., Fitch, W. M. ( 1989). Evidence from nuclear sequences that invariable sites should be considered when sequence 
divagence is calculated. Mol. BioL Evol. �, 270-289. 
Sibley, C. G., Ahlquist, J. E. (1981). The phylogeny and relationships of the ratite birds as indicated by DNA-DNA hybridization. In: 
Swdder, G.G.E., Reveal, J.L (eds.), Evolution today. Carnegie-Mellon University, Pittsburg. pp. 301-335. 
Sibley, C. G., Ahlquist, J. E. (1987). The relationship of four species of .New Zealand passerine birds. ETTUI 87, 63-66. 
Sibley, C. G., Williams, G. R., Ahlquist, J. E. (1982). The relationships of the New Zealand wrens (Acanthisittidae) as indicated by 
DNA-DNA hybridization. Notomis 29, 1 13- 130. 
Simon, C. (1987). Hawaiian evolutionary biology: an introduction. TREE�, 175-178. 
Simon, C. (1991). Molecular systematics at the species boundary: exploiting conserved and variable regions of the mitochondrial 
genome of animals via direct sequencing from amplified DNA. In: Hewitt, G.M., Johnston, A. W.B., Young, J.P.W. (eds.), Molecular 
techniques in taxonomy. NATO ASI Series, Vol. H57. Pp. 33-71. - .-
Simon, C., Franke, A., Martin, A. (1991). The polymerase chain reaction: DNA extraction and amplification. In: Hewitt, G.M, 
Johnston, A. W.B., Young, J.P.W. (eds.), Molecular techniques in taxonomy. Nato AS1 Series, Vol. H57. pp. 329-355. 
Simon, C., Piiiibo, S., Kocher, T. D., Wilson, A. C. (1990). Evolution of mitochondrial ribosomal RNA in insects as shown by the 
polymerase chain reaction. In: Oegg, M.T. O'Brien, SJ. (eds.), UClA Symposw on Molecular and Cellular Biology, New Series. 122, 
235-244. 
Simpson, G. G. (1944). Tempo and mode in evolution. Columbia University Press, New York. 
Smith, A. B. (1989). RNA sequence data in phylogenetic reconstruction: testing the limits of its resolution. Cladistics�, 321-344. 
Smith, M. F., Patton, J. L (1991). Variation in mitochondrial cytoduome b sequence in natural populations of South America 
akodontine rodents (Muridae: Sigmodontinae). Mol. BioI. EvoL !, 85-103. 
Smith, P. J., Johnston, A. D. (1985). GIucosephosphate isomerase and a-glycerophosphate dehydrogenase electromorph frequencies in 
groper Polyprion oxygeneiosis from central New Zealand. N.Z. J. Mar. Freshwater Res. 19, 173-177. 
Smith, P. J., MacArthur, G. J., Michael, K. P. (1989). Regional variation in electromorph frequencies in the tuatua, Paphies 
subtriangulala, around New Zealand. N.Z. J. Mar. Freshwater Res. 23, 27-33. 
Sneath, P. H. A., Sokal, R. R. (1973). Numerical ta:J:onomy. The principles and practice o/numerical classificalion. W.H. Freeman and 
Comapny, San Francisco. 
Solem, A., Climo, F. M., Roscoe, D. J. (1981). Sympatric species diversity of New Zealand land snails. N. Z. J. Zool. !, 453-485. 
Soltis, D. E., Soltis, P. S., Ness, B. D. (1989). Chloroplast DNA variation and multiple origins of autopolyploidy in Heuchera micrantha 
(Saxifragaceae). Evol. 43, 650-656. 
Specht, T., Wolters, J., Erdmann, V. A. (1991). Compilation of 5S rRNA �d 5S rRNA gene sequences. Nucleic Acids Res. 19, 2189-
2191 .  
Spieth, P .  T. (1975). Population genetics of allozyme variation in  Neurospora intennedw. Genetics 80, 785-805. 
States, D. J. (1992). Molecular sequence aCaJracy: analysing imperfect data. Trends Genet. !, 52-55. 
Stebbins, G. L. (1950). Variation and evolution in plants. Columbia University Press, New York. 
page 1 18. 
Steel, M. A., Hendy, M. D., Penny. D. (1993 submitted). Parsimony can be consistent! Syst. Bioi. 
Stern, S., Powers, T., o.angchien, L-M., Noller, H. F. (1989). RNA-protein interactions in 30S ribosomal subunits: folding and 
function of 16S rRNA. Science 244, 783-790. 
Stevens, G. R. (1980). New Zealand adrift. The tMory of continwaJ drift in a New Zealand sening. A.H. &. A.W. Reed, Wellington. 
Stewart, C.-B. (1993). The powers and pitfalls of parsimony. Nature 361, 603-607. 
Stiegler, P., Carbon, P., Ebel, J. P., Ehresmann, C. (1981). A general secondary-structure model for procaryotic and eucaryotic RNAs of 
the small ribosomal subunits. Eur. J. BiocMm. 120, 487-495. 
Suggate, R. P., Stevens, G. R., Te Punga, M. T. (cds.) (1978). The ge0I:!gicaJ histoTY. of New Zealand. Volumes I and II. Government 
Printer, Wellington. 
Swofford, D. L. (1991). PAUP manua� version 3.0. lliinois Natural History Survey, Champaign. 
Swofford, D. L., Berlocher, S. H. (1987). Inferring evolutionary trees from gene frequency data under the principle of maximum 
parsimony. Syst. ZooL 36, 293-325. 
Swofford, D. L., Olsen, G. J. ( 1990). Phylogeny reconstruction. In: Hillis, D.M., Moritz. C. (cds.), Mokcular systematics. Sinauer 
Associates Inc., Sunderland, Massachusetts. Pp. 41 1-501. 
Sylvester-Bradley, P. C. (1963). Post-Tertiary speciation in Europe. Nature 199, 126-130. 
Sytsma, K. J. (1990). DNA and morphology: inference of plant phylogeny. Trends EcoL Evol. �, 104-1 10. 
Sytsma, K. J., Gottlieb, L. D. (1986). o.loroplast DNA evolution and phylogenetic relationships in Clarkia sect. Peripetasma 
(Onagraceae). Evolution 40, 1248-1261.  
Tarling, D.  H., 1988. Gondwanaland and the evolution of the Indian Ocean. In:  Audley-OJarles, M.G., Hallam, A.  (cds.), Gondwana 
and Tethys. Geological Society Special Publication No. 37. Pp. 61-77. 
Taylor, A. O. (1964). A biochemical approach to some taxonomic problems in the genus Coprosma. New PhytoL 63, 135-139. 
Taylor, J. W. (1986). Fungal evolutionary biology and mitochondrial DNA. Exp. MycoL 10, 259-269. 
Thomas, B. W. (1982a). New Zealand gekkonidae: evolutionary trends and taxa. Wildlife Service Occ. Pub/. �, 333-337. 
Thomas, B. W. (1982b). A review of the herpetofauna of southern New Zealand with taxonomic considerations. Herpetofauna 14, 22-
34. 
Thomas, W. K., Beckenbach, A. T. (1989). Variation in salmonid mitochondrial DNA: evolutionary constraints and mechanisms of 
substitution. J. Mol. EvoL 29, 233-245. 
Thomas, W. K., Paabo, S., Villablanca, F. X., Wilson, A. C. (1990). Spatial and temporal continuity of kangaroo rat populations shown 
by sequencing mitochondrial DNA from museum specimens. J. Mol. EvoL ll, 101-1 12. 
Thorpe, J. P. (1982). The molecular clock hypothesis: biochemical evolution, genetic differentiation and systematics. Ann. Rev. EcoL 
Syst. g, 139-168. 
Thuring, R. W. J, Sanders, 1. P. M., Borst, P. (1975). A freeze-squeeze method for recovering long DNA from agarose gels. Analytical 
biochemistry 66, 213-220. 
Topal, M. D., Fresco, J. R. (1976). Complementary base pairing and the origin of substitution mutations. Nature 263, 285-289. 
. \ 
Towns, D. R. (1975a). Ecology of the black shore skink, Leiolopisma suteri (Lacertilia: Scincidae), in boulder beach habitats. N.Z. J. 
ZooL �, 389-407. 
Towns, D. R. (1975b). Reproduction and growth of the black shore skink, Leiolopisma suteri (Lacertilia: Scincidae), in north-eastern 
New Zealand. N.Z. J. Zool. �, 409-423. 
page 1 19. 
Towns, D. R. (1985). Afield guide to the lizftrds of New Zealand. New Zealand Wildlife Service Occ. Pub!. 7. Wellington. 
Towns, D. R., Daugherty, C. H., Newman, D. G. (1985). AIl overview of the ecological biogeography of the New Zealand lizards 
(Gekkonidae, Scincidae). In: Grigg, G., Shine, R., Ehmann, H. (eds.), Biology of Australasian frogs and reptiles. Royal Zoological 
Society of New South Wales. pp. 107-1 15. 
Triggs, S. J. ,  Green, W.Q. (1989). Geographic patterns of genetic variation in brushtail possums TrichosuTUS vulpecula and implications 
for pest control. N.Z 1. Ecol. g, 1 - 10. 
Tuerk, c., Gauss, P., Therrnes, C., Groebe, D. R., Gayle, M., Guild, N., Stormo, G., d'Aubenton-Carafa, Y., Uhlenbeck, O. C., TInoco, 
I., Brody, E. N., Gold, L (1988). CUUCGG hairpins: extraordinarily stable RNA secondary structures associated with various 
biochemical processes. Proc. Nad. Acad. Sci. USA 85, 1364-1368. 
Tzeng, C.-S., Hui, C. F., Shen, S. C., Huang, P. C. (1992). The complete nucleotide sequence of the Crossostoma /acustre mitochondrial 
genome: conservation and variations among vertelx-ates. Nucleic Acids Res. 20, 4853-4858. 
Vane-Wright, R. I., Humphries, C. J., Williams, P. H. (1991). What to protect? Systematics and the agony of choice. BioL Cons. 55, 
235-254. 
Vawter, L, Brown, W. M. (1986). Nuclear and mitochondrial DNA comparisons reveal extreme rate variation in the molecular clock. 
Science 234, 194-1%. 
Vigilant, L, Stoneking, M., Hupending, H., Hawkes, K., Wilson, A. C. (1991). African populations and the evolution of human 
mitochondrial DNA. Science ill, 1503- 1507. 
Vogel, F. (1972). Non-randomness of base replacement in point mutation. 1. MoL EvoL 1, 334-367. 
Vos, M. E. (1988). A biochemical, morphological and phylogenetic review of the genus Cyclodina. Unpublished BSc(Hons) thesis, 
Victoria University of Wellington. 
Vrijenhoek, R. c., Douglas, M E., Meffe, G. K. (1985). Conservation genetics of endangered fish populations in Arizona. Science 229, 
400-402. 
Vulliamy, T., Mason, P., Luzzalto, L. (1992). The molecular basis of glucose-6-phosphate dehydrogenase deficiency. Trends Genet. !, 
1 38-143. 
Wake, D. B. (1981). The application of aIJozyme evidence to problems in the evolution of morphology. In: Scudder, G.G.E., Reveal, 
J.L. (eds.), Evolution Today. pp. 257-270. 
Wallace, A. R. (1880). Island life. MacMillan & Co., London. 
Wallace, A. R. (1889). Darwinism. An exposition of the theory of natural selection with some of its applications. MacMillan & Co. 
London. 
Wallis, G. P., Arntzen, J. W. (1989). Mitochondrial-DNA variation in the crested newt superspecies: limited cytoplasmic gene flow 
among species. Evolution 43, 88-104. 
Ward, R. H., Frazier, B. S., Dew, K., Piiibo, S. (1991). Extensive mitochondrial diversity within a single Amerindian tribe. Proc. Nall. 
Acad. Sci. USA 88, 8720-8724. 
Wardle, P. (1%3). Evolution and distribution of the New Zealand flora, as affected by quaternary climates. N.Z. J. Bot. 1, 3-17. 
Wardle, P. (1978). Origin of the New Zealand mountain flora, wi� special reference to trans-Tasman relationships. N.Z. 1. Bot. 16, 535-
550. 
Wardle, P. (1988). Effects of glacial climates on floristic distribution in New Zealand. 1. A review of the evidence. N.Z 1. Bot. 26, 541-
555. 
Wardle, P., Harris, W., B!lxton, R. P. (1988). Effects of glacial climates on floristic distribution in New Zealand. 2. The role of long­
distance hybridisation in disjunct distributions. N.Z 1. Bot. 26, 557-564. 
page 120. 
Wayne, R. K., Van, Valkenburg B., Kat, P. W., Fuller, T. K., Johnson, W. E., O'Brien, S. J. (1989). Genetic and morphological 
divC7gence among sympatric canids. J. Heredity 80, 447-454. 
Webby, R. F., Markham, K. R., Molloy, B. P. J. (1987). The characterisation of New Zealand Podocarpus hybrids using flavonoid 
markC7S. N.Z J. Bot. 25, 355-366. 
Werner, Y. L., Whitaker, A. H. (1978). Observations and comments on the body temperatures of some New Zealand reptiles. N.Z J. 
Zool. �, 375-393. 
Wheeler, W. C., Honeycutt, R. L. (1988). Paired sequence difference in ribosomal RNAs: evolutionary and phylogenetic implications. 
Mol. BioL Evol. �, 90-96. 
Whitaker, A. H. (1968). Leiolopisma suteri (Boulenger), an oviparous skink in New Zealand. N.Z J. Sci. !!, 425-432. 
Whitaker, A. H. (1978). The effects of rodents on amphibians and reptiles. In: Atkinson, LA.E., Dingwall, P.R., Hoy, C. (cds.), The 
ecology and control of rodents in New Zealand nature reserves. N.Z Dept. Lands and Survey, Infonnation Smes�, 75-88. 
White, M. J. D. (1978). Modes of speciation. H. Freeman & Co. San Francisco. 
White, T. J., Arnheim, N., Erlich, H. A. (1989). The polymerase chain reaction. Trends Genet. �, 185- 189. 
Wilcox, M D., Ledgard, N. J. (1983). Provenance variation in the New Zealand species of Nothofagus. N.Z J. Ecol. 2, 19-31 .  
Wilkinson, G .  S., Chapman, A. M .  (1991). Length and  sequence variation in  evening bat D-loop mtDNA. Genetics 128, 607-617. 
Willen, R. W. (1950). The New Zealand Pleistocene snow line, climatic conditions, and suggested biological effects. N.Z J. Sci. 
TechnoL 32B, 1 8-48. 
Williams, P. L., Fitch, W. M. (1989). Finding the weighted minimal cHange in a given tree. In: Fernholm, B., Bremer, K., Jornvall, H. 
(cds), The hierarchy of life. Molecules and morphology in phylogenetic analysis. Elsevier Science PublishC7s b.v. Amsterdam. pp. 453-
470. 
Wilson, A. c., Cann, R. L., Carr, S. M., George, M., Gyllensten, U. B., Helm-Bychowski, K. M., Higuchi, R. G., Palumbi, S. R., Prager, 
E. M., Sage, R. D., Stoneking, M. (1985). Mitochondrial DNA and two perspectives on evolutionary genetics. BioI. J.l..inn. Soc. 26, 
375-400. 
Wilson, A. c., Carlson, S. S., White, T. J. (1977). Biochemical evolution. Ann. Rev. Biochem. 46, 573-639. 
Wilson, A. C., Maxson, L R., Sarich, V. M. (1974). Two types of molecular evolution. Evidence from studies of interspecific 
hybridization. Proc. Natl. Acad. Sci. USA n, 2843-2867. 
Wilson, A. c., Ochman, H., Prager, E. M. (1987). Molecular time scale for evolution. Trends Genet . .l, 241-247. 
Wilson, A. C., Zimmer, E. A., Prager, E. M, KochC7, T. D. (1989). Restriction mapping in the molecular systematics: a retrospective 
salute. In: Fernholm, B., Bremer, K., Jornvall, H. (cds) The hierarchy of life. Molecules and morphology in phylogenetic analysis. 
Elsevier Science PublishC7s b.v. Amsterdam. Pp. 407-419. 
Wilson, R. D. (1984). ChemotaxoDomic studies of the Rubiaceae. 2. Leaf flavonoids of New Zealand Coprosmas. N.Z J. Bot. 22, 195-
200. 
Wittrnann-Liebold, B., Kopke, A. K. E., Arndt, E., Kromer, W., Hatakeyama, T., Wittmann, H.-G. (1990). Sequence comparison and 
evolution of ribosomal proteins and their genes. In: Hill, W.E., Dahlberg, A., Garrett, R.A., Moore, P.B., Schlessinger, D., Warner, J.R. 
(eds.), The ribosome: Structure,functian, and evolution. American Society for Microbiology, Washington, D.C. pp. 598-616. 
Woese, C. R., Gutell, R., Gupta, R., Noller, H. F. (1983). Detailed analysis of the higher-order structure of 16S-like ribosomal 
ribonucleic acids. Microb. Rev. 47, 621-669. 
Woese, C. R., Winker, S., Gutell, R. (1990). Architecture of ribosomal RNA: constraints on the sequence of tetra-loops. Proc. NatL 
Acad. Sci. USA 87, 8467-8471.  
page 121. 
Wool, I. G., Endo, Y., a.an, Y.-L., Gluck, A. (1990). Structure, function, and evolution of mammalian ribosomes. In:  Hill, W.E., 
Dahlberg, A., Garrett, R.A., Moore, P.B., Schlessinger, D., Warner, J.R. (eds.), The ribosome: structure, function, and evolution. 
American Society for Microbiology, Washington, D.C. Pp. 203-214. 
Wright, C. A., 1980. Chemosystematics: paspective, problems and prospects for the zoologist In: Bisby, F.A., Vaughan, J.G., Wright, 
C.A. (cds.), Chemosystematics: principles and practice. pp. 29-38. 
Wu, C.-I., Li, W.-H. (1985). Evidence for higher rates of nucleotide substitution in rodents than in man. Proc. NatL Acad. Sci. USA 82, 
1741- 1745. 
Wyss, A. R., Novacek, M 1., McKenna, M. C. ( 1987). Amino acid sequence versus morphological data and the interordinal 
relationships of manunaIs. MoL BioL EvoL 10 99-1 16. 
leyl, C. W., Green, D. M. (1992). Heteromorphism for a highly repeatid sequence iii the New Zealand frog Leioplema hochstetteri. 
Evolution 46, 1891-1899. 
Zuckerkandl, E., Pauling, L (1965). Molecules as documents of evolutionary history. J. Theor. BioL !, 357-366. 
Zug, G. R. (1985). A new skink (Reptilia: Sauria: Leiolopisma) from Fiji. Proc. BioL Soc. Wash. 98, 221-231 .  
Zuker, M.  (1989). On finding all suboptimal foldings of an RNA molecule. Science 244, 48-52. 
Zuker, M., Jaeger, J. A., Turner, D. H. (199l). A comparison of optimal and suboptimal RNA secondary structures predicted by free 
energy minimization with structures determined by phylogenetic comparison. Nucleic Acids Res. 19, 2707-27 14. 
Zwieb, C., Glotz, C., Brimacombe, R. (1981). Secondary structure comparisons between small subunit ribosomal RNA molecules from 
six different species. Nucleic Acids Res. 2, 3621-3640.