Chromosomal separation of difficult species of Copris Geoffroy , 1762 and Onthophagus Latreille , 1802 ( Coleoptera , Scarabaeidae ) , with discussion of O . massai Baraud as a British Pleistocene fossil

Karyotype analysis shows that Copris hispanus cavolinii Petagna should be regarded as a separate species from C. hispanus Linnaeus, and that Onthophagus massai Baraud is a valid species, not a synonym of Onthophagus fracticornis Preyssler. Chromosomal variation between populations of O. fracticornis is discussed, and Spanish material is shown to be the most distinct of the populations studied, but it is not considered that it should be placed as a separate species. Pleistocene fossil O. massai and Bronze Age O. fracticornis from England are discussed and illustrated. Th e distinctive elytral puncturation of O. massai is shown by the Pleistocene material, while Bronze Age O. fracticornis resembles modern material of that


Introduction
A number of recent studies on Scarabaeoidea have demonstrated the usefulness of chromosomal analysis in establishing the limits of species which may be diffi cult to distinguish from one another morphologically.For example Wilson (2001) showed that the common and widespread European dung beetle Aphodius fi metarius (Linnaeus, 1758) (Aphodiidae) in fact comprised two distinct species, A. fi metarius itself and Aphodius pedellus (DeGeer, 1774) with completely diff erent karyotypes and with absolutely no evidence of hybridisation between them.Further work (Wilson and Angus, 2004) confi rmed the initial separation and extended the database.More recently, Angus (2008) was able to show that Onthophagus similis (Scriba, 1790) and Onthophagus opacicollis Reitter, 1893 (Scarabaeidae) are completely separate species with no evidence of hybridisation between them, contrary to the suggestion of Martín-Piera and Boto (1999), who used allozyme analysis.
Th e work on O. similis and O. opacicollis also included Onthophagus fracticornis (Preyssler, 1790), whose chromosomes were clearly very distinct from those of the other two species, and this suggests that a detailed comparison of the chromosomes of O. fracticornis and Onthophagus massai Baraud, 1975, a Sicilian endemic morphologically very similar to O. fracticornis would be useful, especially as O. massai has been recorded as a fossil from the Last Interglacial (about 120,000 years ago) in England (Coope, 2001).
Work by Angus et al. (2007) showed that the chromosomes of Copris hispanus hispanus (Linnaeus, 1764) do not match the published account of those of Copris hispanus cavolinii (Petagna, 1792) (Salamanna, 1972), raising the possibility that these are in fact separate species, a view supported by small diff erences in the form of their male genitalia (e.g.Baraud, 1992).Salamanna's work was done using squash preparations from testis, and his fi gures do not enable karyotypes to be assembled, so that fresh work is necessary.

Chromosome features used and rationale behind their use
Th e chromosome features used here, in addition to the total number of chromosomes present, are the size and shape of the chromosomes, expressed as Relative Chromosome Length (RCL: the length of each chromosome as a percentage of the total haploid autosome length in the nucleus) and Centromere Index (CI: the length of the short arm of a chromosome as a percentage of the total length of the chromosome), and the relative amounts and distribution of constitutive heterochromatin (repetitive DNA detected by C-banding).Th ese features allow a high level of distinction between different karyotypes, but care is needed when considering the implication of these differences.Diff erences in chromosome number may result from the presence of variable numbers of B-chromosomes, e.g.Pterostichus nigrita (Paykull, 1790) and Pterostichus rhaeticus Heer, 1837 (Angus et al., 2000), or may refl ect polyploidy often associated with parthenogenesis, as well as indicating diff erences between species.Fusion-fi ssion polymorphisms involving diff erent chromosomes are also known in Coleoptera, an example being Ilybius montanus Stephens, 1828 (Aradottir & Angus, 2004).Th e key to understanding these infraspecifi c diff erences is the occurrence of heterozygotes.Diff er-ences in CI may result from addition or deletion of heterochromatin, or pericentric inversions, and may occur as polymorphisms within species, e.g.autosome 5 of Aphodius pedellus (Wilson, 2001), and diff erences in RCL may refl ect diff erences in the amount of constitutive heterochromatin present, as in the long and short X chromosome of Helophorus grandis Illiger, 1798 (Helophoridae) (Angus, 1989).A key feature of these infraspecifi c chromosomal variations is that they are likely to occur as heterozygotes, though only if the two arrangements occur sympatrically.Intraspecifi c variation in chromosome sizes, revealed by diff erences in the sequence of RCLs along a karyotype, unless refl ecting diff erences in heterochromatic blocks, are likely to result from translocation of material between chromosomes.Such translocational diff erences may result in mispairings of chromosomes at fi rst division of meiosis, and hence reduced fertility (or even sterility), and are thus prima facie evidence that diff erent species are involved.Some caution is needed here: translocational diff erences will only be detectable if they result in noticeable alterations in a chromosome's size.Th ere may be reciprocal translocations which would not be detected but which would nevertheless result in reduced fertility or sterility of hybrids.In this sense the results of chromosomal studies are unidirectional -demonstrable diff erences may indicate that diff erent species are involved, but a lack of such diff erences does not prove conspecifi city.

Material and methods
Table 1 lists the material used in these analyses, with the localities of capture and the number of specimens analysed.Th is refers to the number of beetles from which distinctive chromosome spreads were used in the RCL and CI analyses.Additional material checked for distinctive chromosomes is given in parentheses.Th e localities are numbered, and their geographical locations are shown on the map in Fig. 1.Note that when two localities are fairly close together they have been given the same number.
Chromosome preparations are from mid-gut and testis of adult beetles, as described by Angus (1982) and Shaarawi and Angus (1991).Slides were stained in 1-2 % Giemsa, dried and photographed under oil immersion.Immersion oil was removed using xylene followed by absolute ethanol, and the 2-day old slides were C-banded using saturated barium hydroxide at room temperature (ca 22 o C).Treatment in barium hydroxide was for 3 minutes and was followed by washing in 3 changes of Sörensen at pH 6.8, and incubation in salt-sodium citrate (2 X SSC: 0.3 M sodium chloride and 0.03 M trisodium citrate) for 1 hour at 55 o C. Th e slides were then washed in a further 3 changes of Sörensen at room temperature, and stained in Giemsa as before.Th is enables the same nucleus to be studied both plain and C-banded, and has been done throughout this study.Photographs were printed at a magnifi cation of 3000 X, and the chromosomes were cut out and arranged as karyotypes.At this stage they were scanned into a computer and further arrangement and measurement done using Adobe Photoshop.Th e use of the total  1 for explanation of the numbers, and note that neighbouring sites may share the same number.Th e fossil material was photographed using a Zeiss photomicroscope with oblique surface illumination from a standard bench light.Th e elytra were very crumpled and only small portions were in focus at any one time.However, the resolution was good and the resulting photographs are suffi cient to show the diagnostic features.

Copris hispanus hispanus and C. h. cavolinii 2n = 16 + Xy (♂), 16 + XX (♀).
Plain (Giemsa stained) and C-banded karyotypes are shown in Fig. 2, while RCL and CI data are shown in Table 2, where values showing diff erences between the two forms signifi cant at the 95% level are indicated by yellow highlight.For practical reasons (diffi culty of accurate measurement) CI values below 15 are listed as acrocentric, without further analysis.Points to note are the RCL diff erences in autosomes 3, 6 and 7, and the y chromosome, and the CI diff erences in autosomes 4 and 7. Autosome 6 of C. h.hispanus appears to show a pericentric inversion in the specimen illustrated, but not in the other analysed material.In both forms the heterochromatic blocks are very small and confi ned to the centromere region, so diff erences in the RCLs of the chromosomes are almost certainly due to translocation.
Plain and C-banded karyotypes are shown in Fig. 3, while RCL and CI data are shown in Tables 3 and 4. In the tables instances where the values for chromosomes from different populations of O. fracticornis diff er at the 95% signifi cance level are indicated by yellow highlight, while cases in which the chromosomes of O. massai diff er from those of O. fracticornis are indicated by green highlight.Within O. fracticornis the most notable size diff erences are shown by autosomes 5 and 9 of Spanish material (Fig. 3a, b), and the y chromosome of the Italian material (Fig. 3k, l).Th e large Italian y chromosome is very striking in all preparations, resulting in no clear size diff erence between it and autosome 9. C-bands are heavy on all the autosomes and in some preparations small intercalated C-bands may be present, possibly sites of nucleolus organisers.Th e C-band of the y chromosome is clearly weaker than that of autosome 9. CI diff erences are shown by the y chromosome, with a fairly median (metacentric) centromere in Macedonian (Fig. 3 e-h) and Italian material (Fig. 3k, l) -and the single Swiss specimen (not shown here), and the subterminal (subacrocentric) centromere of Spanish and English material (Fig. 3a-d).In one Macedonian specimen (Fig. 3g, h) autosome 5 is heterozygous for a pericentric inversion, and this specimen also has one B-chromosome.
Th e chromosomes of O. massai (Fig. 3 m, n) show an extensive suite of RCL differences from those of O. fracticornis (Fig. 3a-l).Th us the RCL values for autosomes 1, 5, 8 and 9 are diff erent at the 95% level.Th e y chromosome is as large as that of Italian O. fracticornis, but diff ers from it in having a subterminal centromere.Th ere is also an extensive suite of CI diff erences, involving autosomes 1, 5 and 7-9, as well as the X chromosome.Th e subterminal centromere of the X chromosome is very distinctive.

Copris hispanus hispanus and C. h. cavolinii
Th e RCL diff erences between the karyotypes of these two beetles, taken in combination with their very small heterochromatic regions, similar in size in the two taxa, suggest very strongly that the karyotypes diff er as a result of translocation of material    between non-homologous chromosomes, and thus provide good evidence that they should be considered as separate species, Copris hispanus and C. cavolinii.
Copris cavolinii was described, as a species, by Petagna (1792) on the basis of material from the Naples area, but most subsequent works, including the Catalogue of Palaearctic Coleoptera (Löbl et al., 2006a) place it as a subspecies of C. hispanus.
Th e morphological and geographical characteristics of C. hispanus and C. cavolinii (which she regarded as a subspecies of C. hispanus) were discussed in depth by Rommel (1965).She fi gured details of the cephalic horns and pronotal carinae, but not the genitalia.She gave the distribution of C. hispanus as extending from southern France, Corsica and Sardinia, via the Iberian Peninsula to North Africa, where its range is shown as extending as far as Egypt.No local variation was noted.C. cavolinii was noted from Italy, the Balkans, Turkey, Israel, and extending eastwards through northern Iran to the former Middle Asian republics of the USSR.She distinguished three forms.Th e western form occurs in Italy and the former Yugoslavia, and Albania, the eastern form in Greece (including Crete), Turkey, Cyprus and Israel, and the northeastern form occurs in middle Asia.Dellacasa (1968) reviewed the morphological characteristics of C. hispanus and C. cavolinii and, although he left them as subspecies, pointed out that C. cavolinii was in fact very distinct.
As far as C. cavolinii is concerned, it is important to note that our data refer only to Italian (Sicilian) material, and therefore to Rommel's western form.In the case of C. hispanus, the situation at fi rst appears more straightforward as no geographical variation was noted.However Ebied et al. (2000) record and fi gure a completely diff erent karyotype, with 20 mainly metacentric chromosomes, from Egyptian C. hispanus.Th is is so diff erent from those reported here that it cannot refer to C. hispanus, and means that, unless their material is misidentifi ed, the species passing as C. hispanus in Egypt is something entirely diff erent.
It would be useful to study material from a wider area.Th e C. hispanus localities lie at the apices of an equilateral triangle with 40 km sides, but at least Spanish material must be considered typical of C. hispanus.Th e two Sicilian localities are only about 16 km apart, but nevertheless the results from all the material are consistent so that there is no reason to doubt their validity.Th e main unanswered question is whether study of material from a wider area, especially of the diff erent forms of C. cavolinii, would reveal the presence of other species.Th is question is given added weight by the diff erences between some of the populations of O. fracticornis, to be discussed next.

Onthophagus fracticornis and O. massai
In the case of O. fracticornis, in contrast to those of the two Copris species, material from populations over a wide area of Europe has been studied.While, in terms of RCL at least, this has revealed a considerable level of stability, the Spanish material, in particular, shows some signifi cant diff erences: autosome 5 (recognisable in all populations because of its low CI) is signifi cantly smaller than in other populations, while autosome 9 is signifi cantly larger.One eff ect of this is that in some preparations autosome 5 actually appears shorter than autosome 6. Th e centromeric C-bands of Spanish material appear similar in size to those of other populations, so it is diffi cult not to believe that some interchromosomal translocation of material has taken place.It would therefore seem logical to suggest that there is a prima facie case for regarding the Spanish material as representing a separate species.However, we have detected no morphological diff erence between Spanish and other material, so that for the moment it seems prudent to leave it as O. fracticornis, but note the problem.At this stage it is interesting to note that Angus (2008) found that autosome 5 of Spanish O. opacicollis was signifi cantly larger than that of Sardinian and Cyprus material.It may be appear a curious coincidence that this same chromosome is involved in both cases, but there is a simple explanation: autosome 5 in all the species concerned has a distinctly lower CI than those of autosomes 4 and 6, so autosome 5 is clearly recognisable.Th is raises the question as to whether this autosome is homologous in all the species, and whether the observed diff erences in its length are the only ones involved.As mentioned in the introduction, only translocational diff erences resulting in obvious changes to the length of a chromosome can be detected -and one requirement for this is that the chromosome concerned is itself clearly identifi able.Altering the RCLs of metacentric autosomes occupying adjacent positions in a karyotype might simply reverse the order in which they were placed, without this being apparent.
Apart from the Spanish situation, the Italian material has a signifi cantly larger y chromosome than those of other populations (only females were available from the Czech material, so we have no data on its y chromosome), meaning that, without C-banding, which shows the small heterochromatic block on the y, it could be diffi cult to distinguish from autosome 9.It is diffi cult to assess the signifi cance of this larger y chromosome.Th e sex chromosomes of most Polyphaga pair via a cytoplasmic vesicle (the parachute or Xy p association, cf.Smith and Virkki, 1978), and the small y chromosomes of these Onthophagus species suggest that they are likely to follow this pattern.Th is would mean that no impaired meiosis need be involved in hybrids, and we have no idea what, if any, extra genes the Italian y chromosome may be carrying.
When the CI data are considered, the only variation is found in the y chromosome, more or less metacentric in Macedonian and Italian material, as well as the single Swiss example, but subacrocentric in English and Spanish material.Th is is of no taxonomic signifi cance as it would not aff ect the Xy p pairing at meiosis.It is worth noting that the sequence of CI values along the karyotype of Spanish material does not diff er from those of the other populations, suggesting that the amount of chromosomal diff erence between Spanish and other material is small.
Comparison of the karyotypes of O. fracticornis and O. massai reveals a very diff erent situation, with four of the nine pairs of autosomes of O. massai having signifi cantly diff erent RCLs from their apparent counterparts in O. fracticornis.When the CIs are compared, fi ve pairs of autosomes appear diff erent, as does the X chromosome.Th is degree of diff erence is clearly far more than that shown by populations of O. fracticornis and vindicates the placing O. massai as a separate species.
Onthophagus massai was described, as a distinct species, by Baraud (1975) on the basis of material from the Piano Battaglia in the mountains of the Parco delle Madonie in northern Sicily.We were unable to fi nd it on the Piano Battaglia in early November 2008, but it was present on the Piano Zucchi slightly lower down the same mountains.Th e species status of O. massai was denied by Palestrini (1981), who placed it as a synonym of O. fracticornis.Baraud (1992) reasserted its species status and reviewed the morphological distinctions, especially as regards the sculpture of the elytral striae, between it and O. fracticornis.Subsequent authors (e.g.Carpaneto and Piatella (1995), Sparacio (1995), Pesarini (2004) and Lapiana and Sparacio (2006)) have followed Baraud's assessment, and this view is maintained in the Palaearctic Catalogue of Löbl et al. (2006b).Th us our chromosomal data are in agreement with the current taxonomic consensus.

Onthophagus massai as a Pleistocene fossil in England
As mentioned in the introduction, O. massai has been recorded as a fossil from the Last (Eemian or Ipswichian) Interglacial in England.Coope (2001) reviewed its occurrences and showed that it occurs, sometimes abundantly, in deposits of that interglacial, but not in the immediately preceding one.
Validation of O. massai as a species separate from O. fracticornis, rather than as a local variant of it, means that there is no theoretical diffi culty with its fossil distribution, since many species of beetle have altered their geographical ranges on a dramatic scale in response to the glacial/interglacial oscillations (Coope, 2001).Nevertheless the occurrence in Britain of what is now an endemic confi ned to a small region of northern Sicily is so unexpected that it requires special verifi cation.It is therefore appropriate to consider the characters on which the identifi cation was and is based.Russell Coope fi rst encountered this species in material from Trafalgar Square (see Franks et al., 1958), where at least 49 individual specimens were represented.Th e material clearly belonged to the O. fracticornis group on details of the head and pronotum, and the strength of the basal portion of the cephalic horn of the males was too great for either O. opacicollis or O. similis.Th e specimens appeared consistently small and dark when compared with O. fracticornis, but did match a small series of O. massai.Th e fossil pronota had the punctures large, especially towards the basolateral edge, matching the O. massai, but not the O. fracticornis available for study, and the basal portion of the cephalic horns of the males seemed more like the O. massai than the O. fracticornis.Coope also felt that the puncturation of the elytral interstices was somehow coarser in the fossils and in O. massai than in O. fracticornis, but at the time he did not know that this elytral puncturation character was the one Baraud now uses (e.g.Baraud, 1992) to key out O. massai.Much of the Trafalgar Square fossil material is now in a parlous state having been dry-mounted on cards for many years, but in some cases the elytral puncturation is adequately preserved.It should at this stage be noted that in Pleistocene fossil material the lipid components of the cuticle are lost and the structure tends to collapse on drying.However, as well as the Trafalgar Square fossil material, we have been able to study fossil O. fracticornis from Bronze Age deposits (aged about 4000 years) from Wilsford in Wiltshire, England (see Osborne, 1969).Th is material, though much younger than that from Trafalgar Square, shows a similar fragility due to loss of the lipid components.
Details of modern and fossil elytra of O. fracticornis and O. massai are shown in Fig. 4. Th e presetal granules of O. fracticornis (indicated by white-bordered black arrows) are very clear, and partial collapse of the fossil material serves merely to enhance them.In the case of the O. massai these granules are less conspicuous, but the perisetal punctures (indicated by white arrows) are clear and distinct in both the modern and fossil material.Th e fact that this character, unknown to Russell Coope when he originally studied the material, leads to the same identifi cation, gives ample support for the recognition of O. massai as a Pleistocene fossil, and graphically illustrates how modern restricted distributions may not refl ect the former ranges of the species concerned.

Figure 1 .
Figure 1.Map showing the collection sites of the material used in this paper.See Table1for explanation of the numbers, and note that neighbouring sites may share the same number.
in RCL calculations follows the procedure used with human chromosomes (ParisConference, 1971) and means that, although the X and y chromosomes can have calculated RCL values, it is the RCL values of the autosomes that should add up to 100.Th e CI calculations, again following the Paris Conference, are the basis of morphological classifi cation of the chromosomes.Based on  Sumner (2003), these are: metacentric, CI 50-46; submetacentric, CI 46-26; subacrocentric, CI 25-15; and acrocentric (including telocentric), CI less than 15.Th e beetles from which the preparations were obtained were card-mounted and are in R. B. Angus' collection.

Table 1 .
Material used, localities, map numbers, numbers of specimens used for chromosome measurements, with additional checked material given in parentheses.

Table 2 .
Copris hispanus and C. h.cavolinii, chromosome parameters.Mean, 95% confi dence intervals by t-test, number of chromosomes analysed.Signifi cant diff erences are indicated by gray background.

Table 3 .
Onthophagus fracticornis and O. massai, Relative Chromosome Length.Mean, 95% confi dence intervals by t-test, number of chromosomes analysed.Signifi cant diff erences between populations of O. fracticornis are indicated by light gray background, and those between O. fracticornis and O. massai by dark gray background.