CHAPTER SEVEN

CYTOTAXONOMY

Many biosystematic studies over the past 40 years have included

cytological observations, but all too frequently the authors have not
derived the maximum amount of information from their preparations.
Jackson (1984)

7.1 Introduction

Chromosomes have been recognized as taxonomic characters for a

very long time but, as indicated by Jackson (1984), a large percentage
of cytological investigations have unfortunately failed to provide
taxonomists with as much useful information as they could have, even
though the workers may have put considerable effort into their studies.
Thus if they were principally interested in cytogenetics they would
frequently fail to record information pertinent to taxonomy. Con-
versely, if the work was conducted by taxonomists, their frequent lack
of cytogenetic training means that they often missed many potentially
informative pieces of data.

7.2 Karyotypes

In eukaryotic organisms, the nuclear genome is organized into

chromosomes, the number of which varies between species from a
haploid number (n) of 1, for example, in the ant, Myrmecia pilosula,
and the nematode worm, Proascaris, to more than 200 in some
butterflies and as many as 250 in some ferns (Imai and Taylor, 1989;
Sivarajan, 1991). Whereas genetically controlled variation in other
characters is ultimately the result of differences in nucleotide
sequences within expressed genes, major physical rearrangements of
the genome within chromosomes will usually have no top of this,effect
apparent the
on the organism's
hypotheses that systematics aboutphysiology
development,
normal generates or behaviour
the evolutionary relation-as
as the relative
long between
ships organisms thefunctional
form of
numbers genes
basis of the remain undisturbed
comparative method. a

D. L. J. Quicke, Principles and Techniques of Contemporary Taxonomy

© Springer Science+Business Media Dordrecht 1993
 CYTOTAXONOMY 137

and they are not separated from their control regions. However,
chromosomal rearrangements can have a profound influence on the
fertility of offspring resulting from crosses between individuals of
different karyotypes, and therefore they have considerable evolutio-
nary significance (Jackson. 1976, 1982; Maynard Smith, 1989).
Furthermore, karyotypic information effectively forms an independent
data set for phylogenetic analysis and has probably been most useful in
the investigation of groups of closely related and morphologically
similar organisms.
The interpretation, characterization and identification of a cell's
complete chromosome set is referred to as karyotyping and is the first
stage in the process of using chromosomal characters for systematics.
For many organisms karyotyping is relatively easy if an appropriate
protocol is followed, however, in some groups the presence of large
numbers of small chromosomes or of generally very small cells can
make this a more complicated process. For example, most birds have a
small number of normal sized chromosomes but also have a large
number of microchromosomes which may be extremely difficult to
count, let alone identify.
Probably because of the widespread occurrence of reduced fecundity
in crosses between individuals of different karyotype (hybrid disadvan-
tage), karyotypes within interbreeding populations of a species are
usually remarkably constant although a few marked exceptions occur.
However, it is apparent that at least in small isolated populations,
modified karyotypes may become fixed with apparent ease (see
below), and therefore karyotypes can provide much useful information
about the evolution of a group, the more so as geographic isolation is
probably the most important mode of speciation.
Karyotypic variation may involve several types of change. Whole
chromosome sets may be gained through polyploidy events (which
themselves can take several forms). Rearrangements of chromosome
sections can occur (e.g. translocations and inversions; see section 7 .5)
or individual chromosomes may be lost or gained either by errors in
their migration to daughter cells during meiosis or, probably more
important, the fusion of one chromosome with another (combined with
subsequent loss of the extra centromere) or, less commonly, chromosome
fragmentation.
The degree of variation in chromosome number between closely
related species is subject to great variation. In some groups even
closely related species can be widely dissimilar as epitomized by the
138 PRINCIPLES AND TECHNIQUES OF CONTEMPORARY TAXONOMY

muntjack deer with the Chinese subspecies having a haploid number

(n) of 23, while the Indian subspecies has only 3 plus one sex
chromosome (Benirschke and Kumamoto, 1991). Even more extreme
is the variation encountered in the ant genus Myrmecia in which n
ranges from 1 to 42 (Imai and Taylor, 1989). In marked contrast to
these examples, considerable consistency in chromosome number can
sometimes be observed throughout major groups of organisms. Thus,
for example, in the Lepidoptera (butterflies and moths) and the
Trichoptera (caddisfties), which diverged from one another at least 65
million years ago, during the late Cretaceous, the commonest
chromosome number is approximately the same (n = 30 or 31). Even
when chromosome number differs significantly from these values, the
total genomic DNA content is actually stable, indicating that
polyploidy has not been a significant factor in these insects. A similar
consistency in number is also shown by the bony fish in which n = 24 is
by far the most common value although the overall range goes from
n = 11 to n = 60 (Hinegardner and Rosen, 1972).
A particular problem in this respect can crop up when autopolyploidy
of various degrees occurs, as in quite a number of plants. In those
cases in which morphological characters can also be found to
distinguish between plants of different ploidy level it is standard to
regard the different karyotypes as different species. However, in some
cases no morphological difference can be found between individuals
with widely different chromosome numbers (e.g. the bedstraw, Galium
aparine includes individuals with 2n = 22, 44, 66 and 88). In these
cases they are sometimes retained in the same species even though the
consequences for interbreeding are often not fully understood.
Chromosome number alone is typically a poor indicator of
phylogeny because changes can originate through multiple roots, e.g. it
will be reduced by one no matter which of a set of chromosomes is lost
or which of probably many possible pairwise combinations become
fused. Thus. in order to use chromosome number more meaningfully,
it is necessary to be able to identify individual chromosomes.
Identification of homologous chromosomes in specimens from the
same or closely related species can be difficult. Fortunately, in many
species. some or all of the chromosomes may have distinctive and
consistent morphological and/or chemical features that can enable
recognition between individuals. Morphological differences in form
include chromosome size and the locations of their centromeres. In the
salivary glands of some flies, particularly large polytene chromosomes
 CYTOTAXONOMY 139

are found which have a natural banding pattern that can enable
recognition not only of individual chromosomes but also of small
segments of chromosomes. In many other species, differences in the
densities of chromatin and heterochromatin along chromosomes and
differences in the relative amounts of A-T (adenine-thymine) rich or
C-G (cytosine-guanine) rich DNA along chromosome arms can be
revealed by appropriate staining procedures, again permitting the
recognition of particular chromosomes or parts of chromosomes.
A cursory consideration of the karyotypes of a few different
organisms will soon show that while some are dominated by
metacentrics others have high proportions of acrocentrics and/or
telocentrics (Figure 7.1). Such differences are often apparent even
between quite closely related taxa such as species within a genus or
genera within a tribe. Unfortunately, changes in chromosome number
or, in the case of plants, ploidy level may make it effectively
impossible to score the character state of homologous chromosomes in
the different taxa and therefore if this variation is to be used for
taxonomic purposes it is frequently necessary to obtain an index of
chromosome arm asymmetry. An appropriate measure is the intra-
chromosomal asymmetry index, A 1 (Zarco, 1986), where

chromatids

Metacentrics Acrocentrics Telocentric

Figure 7.1 Diagrams of three different chromosome morphologies.
140 PRINCIPLES AND TECHNIQUES OF CONTEMPORARY TAXONOMY

where bi and Bi are the lengths of the shortest and longest arms of the
n chromosomes in the karyotype. This index is framed so as to be close
to zero if all the chromosomes are metacentric and near one if all are
telocentric.

7.3 Chromosome banding

DNA is inevitably non-homogeneous along the arms of a chromosome.

During mitosis and meiosis, local variations in the degree of
condensation of the chromatin to form heterochromatin may be
apparent in metaphase preparations. Other differences such as the
ratio of adenine + thymine to cytosine + guanine, and the
concentration. composition and lengths of repetitive sequences may be
visualized by more specific staining procedures.
The banding patterns of the giant polytene chromosomes from the
salivary glands and certain other tissues of larval Diptera (true flies)
have proved particularly useful in both distinguishing cryptic species
and even for reconstructing their phylogenies (Dunbar, 1966). Such
work is of considerable importance as one of the groups most
extensively studied to date is the blackfly family, Simulidae, which
includes many species that are vectors of serious diseases, especially in
the tropics. Cytotaxonomy of these flies has been greatly facilitated by
their polytene chromosome banding patterns. Within the Simulidae,
chromosome number is constant with a haploid number of 3. The
Eusimulion aureum group are atypical in having n = 2, but examina-
tion of their banding patterns shows that in this species group the two
smaller simuliid chromosomes have become fused.
Despite their immense usefulness in the cytotaxonomy of true flies,
polytene chromosomes occur only rarely in other organisms (for
example. in the springtails (Collembola) and in dinoflagellate Protista)
and therefore recognition of chromosomes or parts thereof in other
groups depends largely on the use of more sophisticated staining
procedures.
Banding patterns can be revealed in many non-polytene chromosomes
by appropriate staining protocols such as C-, G- and Q-banding, brief
accounts of which are given in the glossary. These have had a patchy
usc in cytotaxonomic studies to date. Nevertheless. where they have
been employed they have often helped considerably in identifying
 CYTOTAXONOMY 141

chromosomal homologies, inversions, translocations, etc., and have,

like other characters, been employed to produce cladograms.
The ability of chromosome banding to distinguish between otherwise
similar chromosomes has also revealed some interesting facets of the
organization of mitosis and meiosis in polyploid hybrids. Cell division
turns out to be even more sophisticated and organized than might at
first be imagined, as in hybrid cells the chromosomal sets from each
parent are not simply randomly assorted during division but at
metaphase the chromosomes from one parent tend to form a ring
around those from the other parent (Linde-Laursen and van Bothmer,
1988).

7.4 Chiasma frequency

Chiasmata or crossovers are evident in mewtlc cell divisions in

diplotene, when they provide visual evidence of genetic exchange
between sister chromosomes. Chiasma frequency is itself under genetic
control and the number of crossovers observable in meiotic cell
preparations tend to be relatively constant within subspecies or species
and often differ between species or subspecies. Their frequency can
therefore be used as a taxonomic character.
Chiasma frequency can also be used as an estimator of the homology
between the two sets of chromosomes occurring in hybrid individuals
by comparing them with the numbers found in each of the parental
species (Jackson. 1984). If the two sets of chromosomes have a high
level of homology, then no depression in the number of chiasmata
would be expected. Generally, it has been found that interspecific
hybrids, and even hybrids between different strains of a single species,
show a reduced frequency of crossing over.

7.5 Inversions, translocations and their significance

Inversions and translocations are two forms of chromosomal rearrange-

ment with considerable implications for the genetic compatibility of
sexually reproducing organisms. In an inversion a segment of a
chromosome becomes reinserted in the same chromosome but the
opposite way around (Figure 7.2) while in a translocation, a
chromosomal segment is removed from one place and reinserted
142 PRINCIPLES AND TECHNIQUES OF CONTEMPORARY TAXONOMY

somewhere else in the genome, either in the same or in some other

chromosome. Further, inversions can either involve only one chromo-
somal arm or they can incorporate the centromere, so-called
paracentric and pericentric inversions, respectively. Inversions and
other chromosomal rearrangements have perhaps been most thoroughly
investigated in the true flies (Diptera) because of their naturally
banded giant polytene chromosomes which make them particularly
easy to recognize and allow accurate definition of the chromosomal
segments involved.
The significance of these chromosomal rearrangements to taxonomy
is two-fold. Firstly they are characters that can be recognized, for
example. in polytene chromosomes or in other material with suitable
histological staining procedures, and then used in phylogenetic
analyses. Secondly, pairing during meiosis between a normal chromo-
some and one with a sufficiently large translocation or inversion can
potentially have a profound effect on the fertility of hybrid organisms.
During meiosis pairing takes place between the homologous chromo-
somal arms of maternally and paternally derived haploid chromosome
sets and chromosome segments are exchanged between these pairs.
Under normal circumstances each of the resulting recombined
chromosome arms contains the same genetic complement but with the
genes from the two parental sets having been reshuffled between
homologous chromatids. However, when a chromosome in one of the
haploid sets possesses an inversion with respect to its homologue,

paracentric pericentric intrachromosomal

inversion inversion translocation

Figure 7.2 Diagrams showing the difference between two forms of chromosomal
inversion and a translocation.
 CYTOTAXONOMY 143

pairing and resulting exchange of chromosomal segments may result in

one of the recombined chromatids missing some genes and the other
having two sets. Any zygote formed involving gametes that either
lacked or had duplicated sets of genes would be unlikely to survive
because of their unbalanced genomes. In the case of pericentric
inversions even greater problems occur for hybrids that were
heterozygous for the inversion as one of the chromatids could end up
with two centromeres while the other would have none, with the result
that chromosome segregation would be disrupted.
Whether chromosomal rearrangements per se account for many
instances of the reduced fecundity often observed in interspecific
hybrids is still debated after many years. Certainly there are plenty of
examples in which the chromosomes of the members of species pairs
differ quite markedly in terms of inversions, translocations or
chromosome fusions, yet hybrids between them show little or no
reduction in fertility, for example, various Drosophila or grasshoppers.
Nevertheless, White (1982) argues that these may be exceptions and
that there are also plenty of counter examples in which small
chromosomal rearrangements have dramatic effects. In any case, any
chromosomal rearrangement after its first occurrence must exist as a
polymorphism within a population, and therefore if this rearrangement
were to be disadvantageous in meiotic pairing with non-rearranged
chromosomes then it would be selected against from the beginning and
would never spread through the population. If this were so, then only
chromosomal rearrangements that were not disadvantageous would
ever persist or become fixed in a population. However, several
arguments and mathematical models suggest that the above argument
is too simplistic. In short. rearrangements that could be disadvan-
tageous in the heterozygous form (i.e. as in interspecific hybrids) could
become fixed if either populations were effectively subdivided into
small units (demes) and/or there is considerable inbreeding and/or
individuals homozygous for the new arrangement are at an advantage.
The importance of chromosomal rearrangements for phylogenetic
analysis stems from their supposed uniqueness. In this respect it has
been argued that since a chromosome could split and rearrange at any
of many possible positions, it is extremely unlikely that a given
chromosome inversion or translocation could undergo a reversal back
to exactly its same original state. It might also be argued that it is
unlikely that identical chromosomal rearrangements could arise more
than once in the evolution of a species. However, while this may be
144 PRINCIPLES AND TECHNIQUES OF CONTEMPORARY TAXONOMY

true at the molecular leveL there is still the practical problem of being
able to recognize the difference between two different chromosomal
rearrangements that involve similar but non-identical transformations.
Nevertheless. the proposed uniqueness of any particular inversion
appears to be largely supported by other evidence although there may
be occasional exceptions. As explained by Farris (197R), unique
acquisition of an inversion does not rule out reversal because while the
back inversion may be extremely unlikely, a population could be
polymorphic for the inversion and from this situation the new inversion
could either become fixed or could be selected against and ultimately
eliminated. Phylogenetic analysis of chromosomal rearrangements
might therefore be better conducted under the polymorphism par-
simony criterion rather than with the assumption that chromosomal
rearrangements arc irreversible (see chapter 2 and Figure 2.10).

7.6 In situ hybridization

The in situ hybridization procedure relies on the fact that under

appropriate circumstances a length of single-stranded DNA will bind
to an homologous stretch of DNA in a cell nucleus or chromosome and
so make it possible to locate whereabouts in an organism's chromosome
set a particular gene sequence occurs. In practice, an artificially
synthesized probe sequence of DNA is labelled either radioactively or
nowadays often with a highly fluorescent chemical group. It is then
allowed to hybridize with chromosomal DNA in a cytological
preparation on a microscope slide and after washing off the excess
unbound probe the cytological preparation can be stained and either
autoradiographed or examined under UV to reveal the location of the
bound probe among the chromosomes.
In generaL this technique is usually used to locate genes that are
represented within the genome by many copies, for example ribosomal
RNA genes, tRNA genes, certain repeated sequences, etc. because the
level of radioactivity or fluorescence available from a single probe
molecule is low and so will not give a clear signal above background.
Nevertheless, some studies with radio-labelled probes have shown that
by counting silver grains over the chromosomes in many spreads, the
location of a gene represented by a single copy (or perhaps a few
copies) can be deduced.
A useful taxonomic role of in situ hybridization has been in the
 CYTOTAXONOMY 145

verification of hypothesized chromosomal homologies based on general

morphology and banding procedures as, for example, is illustrated by
the study of Steinemann et al. (1984) with chromosomes of the
Drosophila obscura group. This sort of confirmatory role can greatly
increase confidence in phylogenies based on putative chromosomal
rearrangements. Another growing use of in situ hybridization is in
genome painting, a technique which is gaining in use for the detection
of taxa that have resulted from interspecific hybridization events such
as appears to be particularly common among the higher plants
(Schwarzacher eta!., 1989; Leitch eta!., 1991; Bennett eta!., 1992). In
this case use is made of naturally extracted and subsequently labelled
DNA from one species which is suspected of being one of the parents
of the supposed hybrid. The DNA, after purification, is cleaved into
fairly short fragments that are then labelled at one end with a
fluorescent marker. Subsequent hybridization of these labelled frag-
ments with cytological chromosome preparations may show more or
less even binding indicating generally spread sequence homology, or it
may show that the preparation's chromosomes fall into two distinct
groups which could indicate that the species under investigation is
indeed a hybrid with one part of its genome derived from the species
providing the DNA probe. Further confirmation can be obtained by
investigating which chromosomes appear to have been derived from
the other parent, and it is even possible to use double labelling
procedures with DNA from one parent labelled with a fluorochrome of
one colour while a diffcrentlv coloured marker is used for DNA from
the other parent. Further. such studies have helped reveal that in
hybrids, chromosomes do not intermingle at random during cell
division but rather that those of one parent tend to surround those of
the other during metaphase.