Karyotypes of some medium-sized Dytiscidae (Agabinae and Colymbetinae) (Coleoptera)

Abstract An account is given of the karyotypes of 29 species of medium sized Dytiscidae (Coleoptera). Of the 20 species of Agabus Leach, 1817, 18 have karyotypes comprising 21 pairs of autosomes and sex chromosomes which are either X0(♂) or XX (♀). These species are Agabus serricornis (Paykull, 1799), Agabus labiatus (Brahm, 1791), Agabus congener (Thunberg, 1794), Agabus lapponicus (Thomson, 1867), Agabus thomsoni (J. Sahlberg, 1871), Agabus confinis (Gyllenhal, 1808), Agabus sturmii (Gyllenhal, 1808), Agabus bipustulatus (Linnaeus, 1767), Agabus nevadensis Håkan Lindberg, 1939, Agabus wollastoni Sharp, 1882, Agabus melanarius Aubé, 1837, Agabus biguttatus (Olivier, 1795), Agabus binotatus Aubé, 1837, Agabus affinis (Paykull, 1798), Agabus unguicularis (Thomson, 1867), Agabus ramblae Millan & Ribera, 2001, Agabus conspersus (Marsham, 1802) and Agabus nebulosus (Forster, 1771). However two species, Agabus infuscatus Aubé, 1838 and Agabus adpressus Aubé, 1837, have developed a neo-XY system, with karyotypes comprising 21 pairs of autosomes and XY sex chromosomes (♂). No chromosomal differences have been detected between typical Agabus bipustulatus and Agabus bipustulatus var. solieri Aubé, 1837, nor have any been found between the three species of the Agabus bipustulatus complex (Agabus bipustulatus, Agabus nevadensis and Agabus wollastoni). The four species of Colymbetes Clairville, 1806, Colymbetes fuscus (Linnaeus, 1758), Colymbetes paykulli Erichson, 1837, Colymbetes piceus Klug, 1834 and Colymbetes striatus (Linnaeus, 1758) have karyotypes comprising 20 pairs of autosomes and sex chromosomes which are X0 (♂), XX (♀). Two of the species of Rhantus Dejean, 1833, Rhantus exsoletus (Forster, 1771) and Rhantus suturellus (Harris, 1828) have karyotypes comprising 20 pairs of autosomes and X0/XX sex chromosomes, but the other three species, Rhantus grapii (Gyllenhal, 1808), Rhantus frontalis (Marsham, 1802) and Rhantus suturalis (Macleay, 1825) have 22 pairs of autosomes and X0/XX sex chromosomes. Agabus congener and Rhantus suturellus may have one B-chromosome. Nine of the species have previously published karyotype data but for seven of these the data are wrong and are here corrected.

introduction When Smith and Virkki (1976) compiled their list of beetles whose chromosome numbers were known, they gave data for 2120 species, including 138 named species belonging to the suborder Adephaga. Of these 110 were Carabidae, 21 Dytiscidae and 7 Gyrinidae. By 1984 the number of carabid species whose chromosome numbers were known had increased to 426 (Serrano and Yadav 1984) and the number of Dytiscidae had reached 32, though five of these were unidentified . Interestingly, the total number of world species of Carabidae is given as "more than 40,000" (Wikipedia) while the number for Dytiscidae is about 4080 (Nilsson-Örtmann and Nilsson, 2010), so at this stage the proportion of species for which chromosome numbers are listed in the two families is about the same. Data have continued to accumulate, so that Galian and Moore (1994) give the number of carabid species whose chromosome numbers are known as "more than 800". Numbers for Dytiscidae have also continued to increase. Saleh Ahmed et al. (2000) gave data on 1 species of Hydrovatus Motschulsky, 1 Hydroporus Clairville and 3 Nebrioporus Régimbart (Hydroporinae), 1 Agabus Leach (Agabinae), 1 Colymbetes Clairville (Colymbetinae) and 1 Eretes Laporte and 1 Hydaticus Leach (Dytiscinae). Aradottir and Angus (2004) gave information on 7 species of Ilybius Erichson (Agabinae), Dutton and Angus (2007) described the karyotypes of 7 species of the "Stictotarsus griseostriatus (De Geer) group" (now in the genus Boreonectes Angus) (Hydroporinae), and Tatton and Angus (2011) reported on 30 species related to Deronectes Sharp (Hydroporinae), of which 27 had no previously published data, bringing to total number of dytiscid species with known chromosome numbers to about 82. This gives both the Carabidae and the Dytiscidae as having about 2% of their species with known chromosome numbers.
The present paper reports on 20 Agabus species, of which only four had previously published chromosome data (wrong for three of the species), 4 Colymbetes, all of which have previously published data, though for three of the species these data were wrong, and 5 Rhantus of which one species had published data, again wrong. This gives a net increase to over 100 in the number of dytiscid species for which information on chromosome numbers are available. The data have been gathered over more than 25 years, and include the results of research projects by three undergraduate students of Royal Holloway, University of London, supervised by R. B. Angus. D. E. Wenczek (1994) studied Rhantus Dejean, J. C. Carter (2001) Rhantus and Colymbetes, and M. J. Clery (2009) made a special study of the Agabus bipustulatus (Linnaeus) species group.

Material and Methods
The species studied, with their localities of origin, collectors and dates, as well as the number of specimens yielding successful preparations, is given in Table 1. Nomenclature and classification follow Nilsson and Hájek (2013, internet version). Where there is more than one locality for a given species the localities from which various prepara- Preparations were made from adult beetles, using mid-gut, testis and ovary, following the protocol given by Shaarawi and Angus (1991) and Dutton and Angus (2007). Treatment with colchicine and hypotonic KCl was for 12.5 min in each solution. C-banding was obtained using saturated Ba(OH) 2 at room temperature, followed by incubation in salt-sodium citrate (2 X SSC) at 60 o C. Treatment times varied, and the technique evolved over the more than 25 years of the study. If a treatment has been insufficient to produce C-banding, it may be repeated. Initially Angus used to clear the stain with a short immersion in 2X SSC at 60 o C, but later found this unnecessary. One set of early experiments with Agabus congener and A. lapponicus was particularly interesting: an initial treatment of 5 min in Ba(OH) 2 proved inadequate. A repeat treatment with 5 min in Ba(OH) 2 produced good centromeric C-bands, but if the second treatment was for 3 min the secondary constrictions were also stained (Fig 1 f, g, and k with the secondary constrictions, Fig. 1 j with just the centromeric C-bands).
Chromosome measurements were made on screen and were used for calculating Relative Chromosome Length (RCL), the length of each chromosome expressed as a percentage of the total haploid autosome length in the nucleus. This compensates for differing degrees of chromosome contraction shown in different nuclei. For the Agabus bipustulatus group the RCL data were subjected to statistical analysis using Student's t-test, but otherwise they are given as approximate values only, to indicate the size relationships of the different pairs of autosomes. Centromere Indices (CI) are not given in detail, but are assigned to their conventional categories. Based on Sumner (2003) the categories are: metacentric-CI 46-50; submetacentric-CI 26-45; subacrocentric-CI 16-25; acrocentric-CI 3-15.

Subgenus Acatodes C. G. Thomson, 1859
A. congener (Thunberg, 1794). Fig. 1 f, g (♂), Fig. 1 h (♀). Published information: none. 2 n = 42 + X0 (♂), 42 + XX (♀), 1 B-chromosome. The autosomes, all more or less metacentric, have RCLs ranging from about 7-4, with an even size decrease along the karyotype. The submetacentric X chromosome, RCL about 9, is clearly the longest in the nucleus. All the chromosomes have distinct centromeric C-bands, with some variation in strength between pairs, and autosomes 1 and 8 have secondary constrictions which may C-band, especially that on autosome 1. The C-banding reaction of the secondary constriction of autosome 8 is less pronounced, and the constriction may be apparent in only one of the replicates. The Swedish female ( Fig. 1 h) has one B-chromosome, about as long as autosome 1 and appearing uniformly partly heterochromatic.
A. thomsoni (J. Sahlberg, 1871). Fig. 1 l, m (♂). Published information: none. 2n = 42 + X0 (♂). The karyotype of this species is very similar to those of A. congener and A. lapponicus, but the longest autosome with a secondary constriction is placed as no. 2 as in this material it appears distinctly shorter than the longest autosome (pair 1). It is possible that additional material would show this not to be the case. As in the preceding two species, the secondary constriction on autosome 8 is more conspicuous in one of the replicates.
A. confinis (Gyllenhal, 1808). Fig. 1 n, o (♀). Published information: 2n = 40 + "XY" (sex chromosomes not identified) (Smith, 1953). 2n = 44 (♀), probably 42 + XX. The material available for study was three females, and although no intact chromosomal complement was obtained, the 43 chromosomes shown in Fig. 1 n exceed the number given by Smith. The suggestion that the X chromosome is the largest in the nucleus is based on comparison with the karyotypes of the three preceding species, all, like A. confinis, members of the A. congener group. In the interpretation given here, Fig. 1 n lacks one X chromosome while Fig. 1 o, from a different specimen, has both X chromosomes but lacks one replicate each of autosomes 1 and 2.
A. infuscatus Aubé, 1838. Figs 1q, 2 (♂). Published information: none. 2n = 42 + neo XY. The autosomes are nearly all either metacentric or submetacentric, but pairs 3 and 17 are subacrocentric. The RCLs of the autosomes range from about 7.9-2.9, and there is a fairly even size decrease along the karyotype, though with slightly sharper decreases between pairs 1 (RCL about 7.9) and 2 (RCL about 6.9), 11 (RCL about 4.3) and 12 (RCL about 3.6), and pairs 18 (RCL about 3.6) and 19 (RCL about 2.9). The subacrocentric X-chromosome (RCL about 7.2) has a distinct gap in its long arm and the Y chromosome, also subacrocentric, is smaller, RCL about 4.6, and matches the X chromosome minus the terminal section of its long arm. This is typical of a neo-XY system where the X chromosome fuses with an autosome to give neo-X, and the same autosome without the X fused to it becomes the neo-Y chromosome. First metaphase of meiosis ( Fig. 2) shows 22 bivalents with no suggestion of a B-chromosome behaving differently from the others. Although it is not possible to identify the neo-XY the behaviour of the chromosomes is entirely consistent with a neo-XY system.

Subgenus Gaurodytes C. G. Thomson, 1859
The A. bipustulatus group RCL data for this group are given in Table 2.
A. bipustulatus (Linnaeus, 1767). Fig. 3 a-f. Published information: 2n = 40 + Xy p (Suortti, 1971). (See comment on Suortti's work under A. sturmii.) 2n = 42 + X0 (♂), 42 + XX (♀). The X chromosome is the longest in the nucleus, though its RCL value can overlap that of autosome 1 (Table 2). Autosome 1 is characterised by a secondary constriction in its long arm, frequently picked out by C-banding (Fig. 3 b, d, f). The . Agabus (Gaurodytes) part 1, the A. bipustulatus group, mitotic chromosomes arranged as karyotypes. a-f A. bipustulatus: a-d ♂, Inari, testis a plain b the same nucleus C-banded c plain d the same nucleus C-banded e, f ♂, Woolmer, mid-gut e plain f the same nucleus C-banded g-k A. bipustulatus var. solieri: g, h ♂, Moiry, testis g plain h the same nucleus C-banded i-j, k Illsee, ♂, testis i plain j the same nucleus C-banded k a different nucleus C-banded l, m A. nevadensis, Sierra Nevada, mid-gut, plain, l ♂, m ♀ n, o A. wollastoni, ♂, Madeira, testis, plain; p-s A. melanarius: p, q ♂, testis p plain q the same nucleus C-banded r, s ♀, ovary, r plain, s the same nucleus C-banded. Bar = 5 μm. expansion or contraction of this constriction can drastically alter the apparent size of the chromosome (Fig. 3 a, b). The longer chromosomes (pairs 1-10) are submetacentric, while the smaller ones are more or less metacentric. The X chromosome is submetacentric to subacrocentric. The variation in the apparent size of this chromosome in different nuclei can be striking-it is about twice as long as autosome 1 in Fig. 3 a, b, but only slightly longer that autosome 1 in Fig. 3 c, d. Since these nuclei are from the same beetle the difference must be the result of different degrees of condensation of the chromosome.
A. bipustulatus var. solieri Aubé, 1837. Fig. 3 g-k. Published information: none. 2n = 42 + X0 (♂), 42 + XX (♀). All the preparations illustrated are from the Swiss Alps, and are chosen because good plain and C-banded preparations were obtained from the same nuclei. The nuclei shown in Fig. 3 g-j are more condensed than the typical A. bipustulatus shown, but the one in Fig. 3 k shows a comparable degree of condensation. These karyotypes show no obvious difference from those of typical A. bipustulatus. The dark area at the end of the X chromosome in Fig. 3 k is where it overlapped one of the autosomes in the preparation. The extreme size difference between the two replicates of autosome 1 in Fig. 3 g, h is very striking, but C-banding (Fig. 3 h) shows that this size difference is entirely due to the degree of expansion of the secondary constriction.
A. nevadensis Håkan Lindberg, 1939. Fig. 3 l, m. Published information: none. 2n = 42 + X0 (♂), 42 + XX (♀) The preparations are from old material in R. B. Angus' archive, and no C-banding is available. The heavy short arm of one replicate of autosome 1 in Fig. 3  A. wollastoni Sharp, 1882. Fig. 1 n, o. Published information: none. 2n = 42 + X0 (♂). As with A. nevadensis, this is archive material and no C-banding is available. Only two karyotypes could be obtained, both from rather condensed nuclei, but the general arrangement of the chromosomes is very similar to, if not identical with, those of the species already discussed.
A. melanarius Aubé, 1837. Fig. 3 p-s. Published information: none. 2n = 42 + X0 (♂), 42 + XX (♀). The general layout of the karyotype is very similar to those of the A. bipustulatus complex described above, but there appear to be more secondary constrictions. Thus in the female (Fig. 3 s), where the C-banding is better displayed, secondary C-bands are clear in autosomes 1, 3, 6, 7 and 14, and even in the male (Fig. 3 q) the secondary C-bands are clear in autosomes 1, 6 and 14.

Other Gaurodytes species
A. biguttatus (Olivier, 1795). Fig. 4 a, b. Published information: 2n = 42 + X0 (♂), 22 + XX (♀) (Saleh Ahmed et al., 2000). The present material, from both Egypt and Sardinia, confirms the data of Saleh Ahmed et al. We have altered the position of the long chromosome with the secondary constriction from pair No. 3 to pair No. 1 as this matches the Sardinian specimen better, and there is sufficient variation in the RCL of this chromosome, due to opening of the secondary constriction to justify this move. The autosomes are all either metacentric or submetacentric with an even size decrease along the karyotype from RCL about 6 to about 3. The X chromosome has RCL about 6 and is more distinctly submetacentric than the larger autosomes, except of autosome 1 which has the secondary constriction. No C-banded preparation is available.
A. binotatus Aubé, 1837. Fig. 4 c. Published information: none. 2n = 42 + X0 (♂). The karyotype of this species appears very similar to that of A. biguttatus, with a similar spread of RCLs. However, autosomes 14-21 are clearly less metacentric than in A. biguttatus, in some cases approaching subacrocentric. The X chromosome, RCL about 8.5, is clearly the largest in the nucleus, thus distinctly larger than in A. biguttatus.
A. unguicularis (Thomson, 1867). Fig. 4 e. Published information: none. 2n = 42 + X0 (♂). The RCLs of the autosomes range from about 10-2.4. There is an abrupt size decrease between pairs 2 and 3 (RCLs about 9.4 and 7.6) and pairs 3 and 4 (RCL of pair 4 about 6.5), but apart from that the size decrease is fairly even. Most of the autosomes are metacentric or almost so, but a few are clearly submetacentric. The X chromosome, RCL about 6.5, is similar in size to autosome pair 4, but much more clearly submetacentric. No C-banded material is available.
A. ramblae Millan et Ribera, 2001. Fig. 4 f. Published information: none. 2n = 42 + X0 (♂), 42 + XX (♀).The RCLs of the autosomes range from about 7-2.9, with a fairly even decrease in length along the karyotype. The autosomes are a mixture of metacentrics and submetacentrics (some at the extreme end of the range), with autosomes 10-12, 15, 16 and 20 subacrocentric. The X chromosome is about the same size as autosome 1, but more clearly submetacentric. No C-banded material is available.
A. conspersus (Marsham, 1802). Fig. 4 g. Published information: 2n = 38 + XY . 2n = 42 + X0 (♂). The RCLs of the autosomes range from about 6.1-3.6, with an even decrease in chromosome size along the karyotype. The autosomes are all either metacentric or submetacentric, and autosome 3 has a prominent secondary constriction in its long arm and autosome 15 has what appears to be a terminal NOR at the end of its short arm. The X chromosome, RCL about 5.6, is submetacentric and similar in size to autosomes 4-6. No C-banded material is available. This karyotype is clearly very different from that reported by . They report a number of nuclei supporting their conclusions, so the most likely explanation is that they were working with a different species. It may be noted that Marsham (1802)   conspersus from England so the material here may be regarded as true A. conspersus.

Yadav et al. worked with Indian material.
A. nebulosus (Forster, 1771). Fig. 4 h, i. Published information: none. 2n = 42 + X0 (♂), 42 + XX (♀). The general layout of the karyotype in terms of RCLs of the autosomes is very similar to that of A. conspersus. Autosome 3 has a similar secondary constriction in its long arm, but the small chromosome with the terminal apparent NOR is relatively larger than in A. conspersus, and is placed as pair 12 as against 15. The X chromosome, RCL about 7.3, appears relatively larger than that of A. conspersus, and is metacentric. The Tenerife specimen whose chromosomes are shown in Fig. 4 i is of a form whose dark pronotal spots are absent or scarcely apparent, but the chromosomes clearly associate it with the British well-spotted A. nebulosus rather than A. conspersus which lacks the pronotal spots.
A. adpressus Aubé, 1837. Fig. 4 j-l. Published information: none. 2n = 42 + XY (♂). The autosomes are all either metacentric or submetacentric, with RCLs ranging from about 7.2-3.1 and with an even decrease in size along the karyotype. Autosome 2 has a secondary constriction in its long arm and autosome 8 has one in its short arm. The X chromosome is submetacentric (almost metacentric), about as long as autosome 1. The Y chromosome, RCL about 5, looks like the X chromosome with most of one arm missing. C-banding (Fig. 4 l) shows considerable variation in the centromeric Cbands of the autosomes. Autosome 1 lacks any C-band, 2 and 3 have strong C-bands and 4 has a weak one. Autosome 5 lacks a C-band and that on autosome 6 is very weak. Autosomes 7-9 have strong centromeric C-bands and 10-13 have weaker ones. Pair 14 has very weak bands. Pairs 15-21 have strong C-bands. The secondary constriction of autosome 2 shows as a C-band, but that of autosome 8 appears to be merged with the strong centromeric C-band. The sex chromosomes both have very large strong centromeric C-bands, which is a powerful piece of evidence that this is a neo-XY system rather than an X0 system and a B-chromosome. Unfortunately no meiotic preparation is available.
C. piceus Klug, 1834. Fig. 5 e. Published information: 2n = 40 + X0 (♂), 40 + XX (♀) (Saleh Ahmed et al., 2000). The karyotype shown in Fig. 5 e is the one published by Saleh Ahmed et al. and is included for comparison with the other species. The RCLs of the autosomes range from about 8.2-1.9, with a fairly even decrease in chromosome size along the karyotype. Autosomes 2, 8, 12, 14 -17 and 19 are submetacentric, pair 9 is subacrocentric, and the remainder are metacentric. Autosomes 6 and 7 have secondary constrictions in their short arms. The X-chromosome, RCL about 8.2, is similar in size to autosome 1, but is less evenly metacentric. No C-banded material is available.
C. striatus (Linnaeus, 1758). Fig. 5 f, g. Published information: 19 -21 pairs + Xy p ? (Suortti, 1971). 2n = 40 + X0 (♂). The RCLs of the autosomes range from about 8-2.7, with a more noticeable decrease in length between autosomes 1 and 2 (RCL about 6.1) than in the other species, but otherwise with a fairly even decrease in chromosome length along the karyotype. Autosomes 2, 3, 5, 7 and 9 are submetacentric, but the others are more or less metacentric. Autosomes 4, 8 and 9 have secondary constriction in their short arms. The X chromosome, RCL about 9, is similar in length to autosomes 2-4, more nearly metacentric than pairs 2 and 3, but less so than pair 4. No C-banded material is available. For Suortti's data, see comment under C. paykulli.
R. suturalis (Macleay, 1825). Fig. 5 n-p. Published information: none. 2n = 44 + X0 (♂). The RCLs of the autosomes range from about 6.1-3.5. The rate of decrease along the karyotype is very even with many of the adjacent pairs appearing more or less the same size. Most of the autosomes are more or less metacentric but pairs 13, 14 and 22 are clearly submetacentric and pair 20 is subacrocentric. C-banding (Fig. 5 o) shows all the chromosomes with centromeric C-bands, of varying strengths. Pairs 1, 8 and 14 have secondary constrictions on their short arms. The X chromosome, RCL about 5.8, is metacentric with a rather weak centromeric C-band. The Kuwaiti material (Fig. 5 p) shows no differences from the British.
R. suturellus (Harris, 1828). Fig. 5 q-t. Published information: none. 2n = 40 + X0 (♂), 40 + XX (♀), sometimes with 1 B-chromosome. The RCLs of the autosomes range from about 7-2.7, with a fairly even decrease in chromosome length along the karyotype. Autosomes 5,7,9,11,12,[15][16][17]19 and 20 are clearly submetacentric, with the remainder more or less metacentric. Pairs 4, 8 and 9 have secondary constrictions in their short arms. The X chromosome, RCL about 5.3, is metacentric and similar to chromosomes 4-8. No C-banded material is available. This karyotype is unusual in having a B-chromosome, a small metacentric, RCL about 3, which has so far been found in Studland Heath material. The first Studland Heath material, in 1993, comprised a male with a B-chromosome and a female without one, giving the impression that this species had an XY sex chromosome system. However, the 2000 material, a mail from Pinail lacking the B-chromosome and a female from Studland Heath with the B-chromosome, revealed the true nature of the situation.

Discussion
In considering the data presented here, two aspects are of particular note: the extent to which the different genera have characteristic karyotypes and details of any deviations from generic karyotypes; and the extent to which the karyotypes of related species show clear differences.
In Agabus 18 of the 20 species reported have a karyotype involving 21 pairs of autosomes and sex chromosomes which are X0 (♂) and XX (♀), but the remaining 2, A. infuscatus and A. adpressus, have 21 pairs of autosomes and sex chromosomes which are XY (♂) and XX (♀). These two species are not closely related (they are placed in different subgenera), but appear to have evolved similar neo-XY sex chromosomes. What makes this particularly surprising is that, since the development of a neo-XY system involves fusion of the original X chromosome with an autosome, there should be an initial reduction by one in the number of pairs of autosomes. However, both the species involved here show no such reduction, so have presumably undergone fission of one autosome to give two and hence restore the original number. It may be noted that  describe their "Agabus conspersus" as having 38 autosomes (19 pairs) and XY sex chromosomes. Assuming their chromosome data are correct and they are working with an Agabus species, this one has a reduced number of autosomes as well as an XY system.
Among the Agabus species reported here, there are two groups of particularly close relatives, A. congener, lapponicus and thomsoni, and the A. bipustulatus group. A. congener and lapponicus show no interspecific chromosomal differences despite a good number of high-quality preparations. A. thomsoni may show a slight difference in the RCL of the longest secondary constriction-bearing autosome, but more material would be needed to confirm this.
The A. bipustulatus group comprises A. melanarius and the A. bipustulatus complex within which the overriding impression from the present investigation is the extreme similarity between the karyotypes of the species. In the case of A. bipustulatus and A. bipustulatus var solieri this is not surprising as these are regarded as conspecific. The case of A. nevadensis is perhaps more interesting as this is currently regarded as a distinct species in spite of the lack of clear morphological characters to distinguish it from A. bipustulatus. The karyotype of A. wollastoni also shows no obvious difference from those of the other species, but in this case the species does have a very clear morphological character to distinguish it from A. bipustulatus-the inner anterior tarsal claw of the male is simple, not expanded to give the "scooped-out" appearance characteristic of A. bipustulatus, solieri and nevadensis. Only A. melanarius, not really a member of the A. bipustulatus complex, shows some karyotype differences, most clearly in the more extensive development of heterochromatic (C-banding) regions on the chromosomes. These findings may be considered in the light of the phylogenetic trees obtained by Drotz et al. (2010) from their studies of mitochondrial DNA of these beetles. Drotz et al. place the A. bipustulatus group as a complex within a slightly larger A. tristis Aubé group. Their Fig. 5 shows a strict consensus phylogenetic tree of the group. This figure is particularly interesting: A. melanarius is shown to be among the most isolated of the A. tristis group species, with it plus A. tristis placed as a sister taxon to all the rest combined. The remaining species, including A. wollastoni, comprise the A. bipustulatus complex, within which A. wollastoni is the first to come out, being placed as sister to all the others. It is at once apparent that the karyotypes of all these A. bipustulatus complex species are the ones showing no difference from one another. A. melanarius does show chromosomal differences, and it would be very interesting to know whether this is also true of A. tristis. However, this is a Nearctic and east Palaearctic species, not available for study here.
Examination of the material of A. bipustulatus, solieri and nevadensis included in their study shows how they came to their conclusions as to their taxonomic status. They are concerned with forms in which the primary reticulation (the fine meshes inside the larger elongate secondary meshes) is progressively reduced. These forms are referred to the varieties dolomitanus Scholz, 1935, falcozi Guignot, 1932, kiesenwetteri Seidlitz 1887and pyrenaeus Fresneda and Hernando, 1989. The most striking thing is that these various solieri forms come out in a number of different places, often with ordinary bipustulatus from neighbouring areas. A. nevadensis, with its very restricted distribution, almost inevitably comes out in only one place, but very closely associated with a population of solieri (kiesenwetteri) from France. The claim of A. nevadensis to species status appears weak. The mitochondrial DNA separation is very slight, the karyotype appears identical with those of other A. bipustulatus forms, and the morphological characteristics are less clear than those of solieri.
The case of A. wollastoni is interesting. This species is isolated on Madeira and has had time to diverge from other A. bipustulatus, both in its mitochondrial DNA and also in its morphology-simple inner anterior tarsal claws of males, and generally larger size. Only the chromosomes show no difference.
The four species of Colymbetes share the same basic karyotype with 2n = 40 + X0 (♂), with the X chromosome a large more or less metacentric. There are minor differences in the RCL sequences between the species, which may or may not stand up to more detailed analysis if more material becomes available. Autosome 1 of C. striatus appears larger than in the other species.
The karyotypes of the Rhantus species are interesting in showing two different numbers, with 2n = 40 + X0 (♂) in R. exsoletus and R. suturellus, but 2n = 44 + X0 (♂) in the other species studied. Interestingly, this number difference does not reflect the subgeneric classification. The B-chromosome of R. suturellus is interesting in that it could be confused with a neo-XY sex chromosome system comparable with that of Agabus infuscatus and A. adpressus.
The Kuwaiti material of R. suturalis is interesting as it shows no differences from British material. Balke et al. (2009) demonstrated that this "supertramp" species almost certainly originated in the highlands of New Guinea from where it extended its range in two separate lineages, one southern going into Australia and New Zealand, and the other northern, going into Asia and Europe. Clearly Kuwaiti and European material belong to this northern lineage, but it is good to see the absence of chromosomal differences between specimens from these areas supporting the integrity of this species.