Landcare Research - Manaaki Whenua

Landcare-Research -Manaaki Whenua

FNZ 59 - Erotylinae (Insecta: Coleoptera: Cucujoidea: Erotylidae) - Cladistic analysis

Skelley, PE; Leschen, RAB 2007. Erotylinae (Insecta: Coleoptera: Cucujoidea: Erotylidae): taxonomy and biogeography. Fauna of New Zealand 59, 59 pages.
( ISSN 0111-5383 (print), ; no. 59. ISBN 978-0-478-09391-9 (print), ). Published 07 Sep 2007
ZooBank: http://zoobank.org/References/351ADE1F-65D8-44E1-9F57-C94CACEA93DF

Cladistic analysis

Here we reconstruct the phylogenetic relationships of Cryptodacne to determine the placements of C. brounii and C. rangiauria. Cryptodacne is monophyletic based on the presence of dilated maxillary palpi, mentum not excavated, and absence of flight wings. All species of Cryptodacne were coded and entered into MacClade version 3 (Maddison & Maddison 1992) for character analysis. Tree searches were done in PAUP* version 4.0 (Swofford 2003). A thorough study of the dacnine genera has not been done, though the tribe was represented by Combocerus Bedel, Dacne Latrielle, and Cryptodacne in the morphological study by Wegrynowicz (2002), coded as a monophyletic group by Leschen (2003, based on representives of Cryptodacne, Dacne, Hoplepiscapha Lea, and an undescribed Australian genus), and by Dacne californica Horn in the molecular study by Robertson et al. (2004). In these studies Dacnini is placed in a basal position in Erotylinae, but the exact sister-relationships of Cryptodacne are unclear because a more complete phylogenetic study of the tribe is unavailable. The dilated maxillary palpus present in Cryptodacne is present also in Cnecosa, and this genus is a likely sister taxon. We rooted trees with other dacnines as outgroups: Cnecosa, Kuschelengis, and two species of Thallis (T. janthina Erichson (Australia) and T. nigroaenea Crotch (New Caledonia)). The settings used in PAUP* for heuristic tree searches include a random addition sequence (100 replicates) with steepest descent; character states were treated as unordered. A total of 19 characters (listed below) were coded and the data matrix is provided in Table 1. Confidence intervals for branches on a cladogram were determined by Bremer support (Bremer, 1988) as implemented in Autodecay 4.0.2’ppc (Eriksson 2000) and Bootstrap analysis (Felsenstein 1985, Sanderson 1995) with 1000 replications to determine support. Characters were optimised onto trees using standard ACCTRAN (accelerated transformation) and DELTRAN (delayed transformation) optimisations (Maddison et al. 1984).


Characters Used in Cladistic Analysis

  1. Dorsal setation of pronotum and elytra. 0, indistinct or absent, if present majority barely extend out of punctures (Fig. 4); 1, distinct, long enough for majority of setae to extend out of punctures (Fig. 9).
  2. Colour pattern of dorsal body. 0, absent and with uniform colour pattern (Fig. 1); 1, present, contrasting marks present (Fig. 2).
  3. Dorsal punctation. 0, coarse; 1, fine (Fig. 4).
  4. Body shape. 0, parallel sided; 1, elongate, sides arched, widest near basal third of elytra (Fig. 1).
  5. Terminal maxillary palpomere. 0, acuminate, cylindrical, with terminal sensory area very small and circular (Fig. 19); 1, dilated, with terminal sensory area elongate (Fig. 20). Character state 1 is present only in Cnecosa and Cryptodacne.
  6. Transverse gular groove. 0, incomplete, absent at middle (Fig. 21); 1, complete (Fig. 19).
  7. Pronotal shape. 0, sides evenly arcuate (Fig. 1); 1, sides parallel-sided (Fig. 2).
  8. Basal bead of pronotum: 0, incomplete, absent at middle (Fig. 5); 1, complete across base, fine (Fig. 4); 2, strong, complete, with punctures in basal groove. In Thallis janthina and Cnecosa insueta (Crotch) the marginal line is in the form of it a complete deep groove while in Kuschelengis it is distant and fine.
  9. Prosternal apex. 0, truncate (Fig. 19); 1, emarginate (Fig. 21); 2, lobed or rounded.
  10. Male genitalia: dorsal lobe on internal sac. 0, membranous (Fig. 24); 1, partly or entirely sclerotised (Fig. 31); 2, absent. The male genitalic characters require more detailed analysis, especially considering other taxa of Dacninae, and when considering the species presently included in Australian Thallis. For example, while the dorsal lobe of Kuschelengis is long and basally sclerotised (state 1) the internal sac of T. janthina has a sclerite but lacks the lobe and T. nigroaenea lacks both structures (Thallis is coded with state 2).
  11. Male genitalia: fleshy, ventral lobes on internal sac. 0, absent (Fig. 31); 1, present (Fig. 24).
  12. Male genitalia: arrangement of microsetae at middle of internal sac. 0, scattered, not discernable patch (Fig. 31); 1, ventral patch of setae (Fig. 30).
  13. Male genitalia: flagellar length. 0, short, much shorter than median lobe (Fig. 31); 1, long, as long or longer than median lobe (Fig. 27). The flagellum of Thallis nigroaenea is very short and peg-like.
  14. Male genitalia: sclerite at base of flagellum. 0, narrowly or not expanded (Fig. 31); 1, broadly rounded (Fig. 30).
  15. Male genitalia: sclerite at base of flagellum with an anterior projection, beyond where the internal sac joins with the sclerite. 0, absent (Fig. 31); 1, present (Fig. 24).
  16. Female abdominal segment IX surface structure: 0, distinct comb rows, full length of segment; 1, comb rows shortened, half length of segment, or distinct patch of asperites half length of segment; 2, absent or undefined patch of weak asperites.
  17. Wing development. 0, present; 1, reduced to membranous strap with terminal binding patch.
  18. Tubercles of male profemur. 0, indistinct or absent; 1, small but distinct.
  19. Sexual dimorphism of protibia. 0, sexes similar; 1, male protibia arched and tuberculate.

Table 1. Data matrix for cladistic analysis of Cryptodacne spp.

              0000000001111111111
              1234567890123456789

Kuschelengis  0001011101001001011
nui           0011101010101002111
lenis         0111100010100012100
brounii       001110001010001?100
synthetica    1111101010100012111
rangiauria    101110100001010?100
ferrugata     1011101010110102100
pubescens     1011101010010102100
Cnecosa       1100111220010001011
T. janthina   0000011222001000000
T. nigroaenea 0011011122000001000

Results and Discussion

The analysis resulted in three most-parsimonious trees (Tree Length 38, Consistency index = 0.60; Retention index = 0.66) shown in Fig. 34. Tree 3, which is also the same as a strict consensus tree is shown with support values. The conflict among the trees relates to the uncertain basal placements of C. synthetica and C. nui. The two clades C. brounii + C. lenis and C. ferrugata (C. pubescens + C. rangiauria) are consistent and supported by the characters mapped onto tree 2 (Fig. 35).

Sympatry and the Taxonomic Status of Cryptodacne brounii

Sympatric populations of Cryptodacne can vary extensively in any single character so that it may appear that certain individuals belong to separate species. Without considering the whole organism, series of specimens, male genitalia, and location of capture, it is often difficult to recognise a species without careful study. Unfortunately biological data, such as host fungus that may also provide important clues to species limits, is generally unavailable. Cryptodacne brounii differs from C. lenis in colour variation and by having a shorter base of the flagellum and these two species group consistently in the reconstructed trees. This suggests that C. brounii may be a variant of C. lenis, which it closely resembles, but we have not found aedeagi in populations of C. lenis that resemble C. brounii. However, because specific locality information for C. brounii is not available it is not possible to assess whether or not the characters in C. brounii are aberrant.

It is not clear if the relative similarity of the widespread Cryptodacne populations is an indication of speciation in a recent time frame. It is possible that species could have been isolated by geologic features, such as the deposition of ash and tephra mainly in the North Island or mountain building in the South Island, and diverged into separate species due to an interruption of gene flow, and later formed sympatric distributions through faunal mixing. It is also possible that current sympatric distributions are the result of allopatric speciation during a time when New Zealand was reduced to a series of smaller islands during the Oligocene (Fleming 1979, Cooper & Millener 1993, Cooper & Cooper 1995). Cryptodacne have the hind wing reduced to a narrow strap, and are presumed to be ancestrally flightless, and brachyptery may have been a factor that facilitated the speciation in the mainland forms. Being flightless and with potentially isolated populations over a broad range, any of the Cryptodacne species could contain cryptic species that could be recognised by further morphological or molecular study which will have to be considered in a more detailed analysis when more material and data are available.

Phylogenetic Placement of C. rangiauria and other Chatham Island Species

Examined in the context of the reconstructed phylogenies, C. rangiauria is sister taxon to C. pubescens and is nested in a clade of taxa that are widely distributed in the North and South Islands. Because C. rangiauria is not located at a basal position and is rather derived for the genus, this phylogenetic location supports other phylogenetic studies where it has been shown that the Chatham Islands fauna has recently dispersed from the mainland (e.g., Knox 1960, Trewick 2000, Arensburger et al. 2004, Stevens & Hogg 2004).

In the absence of molecular data, we can provide quantitative evidence for relative age (or placement in a cladogram) by examining phylogenetic position in a cladogram as calculated by measuring Relative Phylogenetic Position (RPP) which is the ratio of the node number of the taxon/longest path in the cladogram beginning at the root of the tree (Leschen 2005). An RPP < 0.50 is considered relatively basal, while an RPP > 0.50 is relatively derived. The RPP for C. rangiauria is 1.0 (5/5) indicating that this species (and C. pubescens) is one of the most derived members of the group. There are problems with this simple method (Leschen 2005) and biases include taxonomic level, numbers of terminals, multiple trees, and resolution of polytomies. Here polytomies were not reconstructed and trees derived from combined data were examined if multiple trees were provided in the original paper.

Calculating the RPP for Chatham Islands taxa in groups with rooted phylogenies from recent literature indicates that most species are relatively derived (Table 2). Interestingly, the population level studies had low RPP values (e.g., Austridotea). The high RPP value for most Chatham Islands endemics indicates that these had ancestors that were recent colonisers to the islands perhaps dating at the earliest from the Pliocene through to more recent times based on molecular clock data (Trewick 2000, Vink & Paterson 2003, Arensburger et al. 2004, Chinn & Gemmell 2004, Stevens & Hogg 2004, McGaughran et al. 2005), rather than an ancient connection dating to 70 my (Campbell et al. 1993), when the South Island was connected to the Chatham Islands as indicated by old Mesozoic continental crust making up the Chatham Rise.

Source Areas of the Chatham Islands Biota

Emberson (1995, 1998) surveyed the Chatham Islands beetle fauna and indicated 30% of the species are endemic to the Chatham Islands and there is a strong South Island connection. In previous papers, Craw (1988, 1989) hypothesised that the Chatham Islands fauna was a composite of northern and southern elements. Phylogenetic reconstructions showed that closest relatives were widespread taxa (Table 3), as supported by our data for Cryptodacne, and we can assume that widespread taxa are better dispersers, or have a higher chance of colonising offshore islands and splitting into daughter species.

We can determine the location of the source area by examining known phylogenies for Chatham Islands taxa. A null hypothesis of a widespread ancestor would be falsified if sister taxa are exclusively found in the South or North Island, or by having a restricted range on the mainland. If immediate sister-taxa are restricted in distribution then the higher level clade to which the sister taxa belong may consist of widely distributed species, providing evidence of an ancient ancestrally widespread species that gave rise to disjunct species on the Chatham Islands and elsewhere. We review recently published phylogenies based on traditional and/or molecular approaches (Table 3) to test the widespread ancestral species hypothesis. Note that if there are multiple endemic Chatham Island species in a single phylogeny then the reconstructed sister-relationships are treated as independent colonisation events.

Plants

Hebe (Scrophulariaceae) is a large group of plants in New Zealand, with over 100 species and varieties, with three endemic species on the Chatham Islands, and has been studied by Wagstaff et al. (2002, see also Wagstaff & Garnock-Jones 1998). In a strict consensus tree consisting of exemplar Hebe species Wagstaff et al. (2002) place the three Chatham Island species in a large polytomy consisting of New Zealand and non-New Zealand taxa. Two species are sister taxa (H. dieffenbachia (Benth.) Cockayne & Allan, and H. chathamica (Buchanan) Cockayne & Allan) while H. barkeri (Cockayne) Cockayne remains isolated with uncertain sister relationships. Based on this study it is not certain if there were one or two colonisation events to the Chatham Islands. An analysis of the 17 described species of Pseudopanax (Araliaceae) was presented by Mitchell & Wagstaff (1997) and in this work Pseudopanax chathamicus Kirk is placed in a trichotomy with the widespread taxa Pseudopanax crassifolius (Sol. ex A.Cunn.) K. Koch and Pseudopanax ferox Kirk, a relationship strongly supported by morphological characters in a combined analysis (Mitchell and Wagstaff 1997).

Spiders

Vink & Paterson (2003) reconstructed the relationships of all lycosid spider species contained in the genus Anoteropsis L. Koch, two species of which are endemic to the Chatham Islands. In the combined analysis of two genes and morphology (based on the earlier work by Vink 2002), there were two separate colonisations of the Chatham Islands from widespread taxa (one sister comparison was assessed by examining the relationship of the Chatham Island species to two different clades, all three were part of basal trichotomy, see Vink & Paterson (2003, Fig. 6). In the morphological tree (Vink 2002) the two relationships of the Chatham Islands species are as follows: A. okatainea Vink, North Island (A. senica (L. Koch), widespread (A. insularis Vink, Chatham Islands (A. ralphi (Simon), Chatham Islands (A. hilaris (L. Koch), widespread)))). If true, then this relationship suggests that the Chatham Islands was a sister area to the derived widespread distribution present in A. hilaris. We prefer the combined tree because it explains all of the data and supports a more parsimonious hypothesis based on a mainland origin of the species with two colonisation events (the ancestor of A. ralphi and A. hilaris is more derived). The relationships among the ND1 partition show that A. insularis is sister taxon to 14 taxa which also contains the derived sister pair A. ralphi + A. hilaris. In summary, the Anoteropsis data show widespread sister taxa to the Chatham Islands species, but in one partition, one species is sister taxon to a Southland species.

Isopods

The relationships of the endemic freshwater isopod genus Austridotea (containing 3 spp., Idoteidae) were reconstructed by McGaughran et al. (2005). There were two colonisation events from regions in the South Island to the Chatham Islands: One colonisation event was by A. annectens Nicholls, with basal populations located on Pitt Island, Chatham Islands, (the species is sister taxon to A. benhami Nicholls found in Otago). Within the species A. lacustris (Thomson), the basal-most population is found in Fiordland and this is sister to populations present on Pitt Island with populations present also in Otago, and on Stewart and Campbell Islands). Though not a phylogenetic study, Stevens & Hogg (2004) demonstrate that the Chatham Islands populations of the amphipod Paracorophium excavatum Thomson (Corophiidae) share alleles with southern North Island and widespread South Island populations.

Insects

Arensburger et al. (2004) reconstructed the phylogeny of Kikihia Dugdale cicadas (10 of 11 described species, Cicadidae) and showed that the Chatham Island species K. longula (Hudson) is sister taxon to an undescribed species from Nelson, and these two are sister taxa to a species from Kaikoura (K. paxillulae Fleming). This relationship was supported in the two trees they presented.

Saprosites Redtenbacher (Scarabaeidae, Aphodiinae) is a relatively diverse scarab beetle genus distributed in Australia, Central and South America, and New Zealand (Stebnicka 2005). In her cladistic study, Stebnicka (2005) included all eight of the mainland New Zealand species, the Chatham Islands S. sulcatissimus (Broun), three South American species, and one Australian species introduced to New Zealand. Determining the relationships of the Chatham Island species to other taxa is ambiguous because there is a basal polytomy of seven taxa (New Zealand and South America) with a monophyletic group composed of S. sulcatissimus, the Australian species, and the remaining New Zealand species.

Craw (1999) reconstructed the phylogeny of Molytini weevils. The genus Hadramphus Broun composed of 4 species has one species found on Chatham Islands (H. spinipennis Broun), and it is sister taxon to a species found in Fiordland and the Snares (H. stilbocarpae Kuschel).

Trewick (2000) provided partial phylogenies for four insect groups in his study. Trewick (2000) sampled eight of the 13 species of cockroaches in the genus Celatoblatta Johns. The Chatham Islands species C. brunni (Alfken) is shown as sister taxon to a South Island species C. quinquemaculata Johns. A more detailed study of the South Island taxa by Chinn & Gemmell (2004) showed that the C. brunni was sister taxon to C. penninsularis Johns, a species endemic to Banks Peninsula. The Chathams + South Island pattern holds (Trewick 2000), though the ancestral area can be reconstructed more precisely in the more complete study by Chinn & Gemmell (2004). Three of the five species the cave weta genus Talitropsis Bolivar (Rhaphidophoridae) were included in the Trewick (2000) study. In the unrooted network, the two Chatham Islands species T. megatibia Trewick and T. crassicrurus Hudson are monophyletic and are derived from a polytomy consisting of the North Island populations of T. sedilloti Bolivar.

The carabid beetle Mecodema alternans Laporte de Castelnau is present on the Chatham Islands, the southern portion of the South Island, and the Snares and this species is shown as a sister taxon to the widespread South Island species M. rugiceps Townsend in Trewick (2000). There are over 50 species of Mecodema Blanchard (Larochelle & Lariviere 2001), and seven species were included in Trewick (2000) with one from the North Island. Lastly, three species (one of which is undescribed) of the lucanid Geodorcus Holloway from The Sisters, Chatham Islands, and the South Island were sampled by Trewick (2000); but, note there are 10 described and undescribed species from North and South Islands (Holloway 1961, 1996), and this group is not considered for this study.

Discussion

Of the 13 taxa surveyed with unambiguous area-reconstructions, Chatham Island has six closely-related taxa that are widespread (Hebe, Pseudopanax, Anoteropsis spp., Paracorophium, Cryptodacne), six closely-related taxa that are from the South Island (Austridotea spp., Celatoblatta, Kikihia, Mecodema, Hadramphus), and one closely-related taxon from the North Island (Talitropsis). Four distributions of the South Island sister-areas are relatively restricted and one was uninformative (Saprosites). While half of the sister-comparisons show a South Island source for the Chathams Islands fauna as suggested by Emberson (1998), about half show widespread distributions supporting older hypotheses listed by Craw (1988), resulting in no real consensus for accepting the widespread ancestral area hypothesis.

Part of the problem with the test we provide is that some of the studies do not have rigorous sampling of species or populations. Most molecular studies suffer from incomplete taxon sampling, either by having limited samples of populations of the ingroup, or by having no outgroups to identify the roots of the trees. Incomplete sampling is a further problem because exact sister-species or population cannot be determined (compare the two studies of Celatoblatta). This is also exemplified in Stebnicka’s (2005) phylogeny where the single Australian species is grouped with the Chatham Island species of Saprosites, and one is left to wonder if there were multiple origins of the New Zealand fauna, highlighting the importance of sampling outside of the group of interest and including more outgroups to better root the tree.

The morphological studies of Hadramphus and Cryptodacne included all of the available taxa that allows for complete assessment of relationships. However, ancestral population-areas cannot be located in widespread sister-species, which can only be determined in molecular studies that have adequate population sampling. A molecular analysis of Cryptodacne would be useful to determine if populations of C. rangiauria are more closely related to South Island populations of C. pubescens than to North Island populations. This is similar to the situation in corophiid amphipods where North and South Island populations of Paracorophium excavatum shared alleles with those in Chatham Islands (Stevens & Hogg 2004), but in this case, characters useful for cladistic analysis are needed to reconstruct the phylogeny of the group.

Biogeographic Summary

Analytical and data-set issues aside, the biogeographic information indicate that there may be several factors that facilitated the arrival of colonising species to the Chatham Islands. Different source areas, separate arrivals in the spider data (Vink & Paterson 2003), and variance among molecular dates (compare Trewick 2000, Vink & Paterson 2003, Arensburger et al. 2004, Chinn & Gemmell 2004, Stevens & Hogg 2004, and McGaughran et al. 2005) indicate independent times of colonisation events. The range of molecular dates for nodes containing Chatham Islands endemic species or populations are from Pliocene and post-Pliocene indicating that mainland dispersers arrived during or after the formation of the Manawatu Strait (or Pliocene Sea Strait) present during the lower Pliocene during a time of submergence (Fleming 1979, Cooper & Millener 1993, Lewis & Carter 1994). When the Manawatu Strait was present, ocean currents driven by westerly forcing may have facilitated movement of the first colonisers to the Chatham Islands, like the separate ancestors that gave rise to Anoteropsis insularis, Paracoriphium excavatum, and Austridotea lacustris. More recent colonisers may have used intervening islands as stepping-stones, island hopping to the Chatham Islands (Fleming 1979) during periods of more recent glaciation. Because dispersal is an on-going process that occurs over great distances (e.g., Hoare 2001), dates provided by molecular studies need to be examined prudently.

More complete phylogenetic studies for all Chatham Islands species and their relatives are needed for a biogeographic synthesis, but here we offer a scenario for Cryptodacne. The trans-Cook Strait coastal distribution of C. pubescens presently occupies areas that were submerged during the Pliocene, including, significantly, what was submerged during the presence of the Manawatu Strait. It is tempting to suggest that there may have been a widespread ancestor that gave rise to the widely distributed C. ferrugata and the ancestor of C. pubescens + C. rangiauria prior to the development of the Manawatu Strait. Ancestral populations of the species C. pubescens + C. rangiauria colonised newly emerged lands and dispersed to the Chatham Islands forming C. rangiauria after the Pliocene. Such a “near coastal” or “Manawatu Strait” ancestor of Chatham Islands fauna could be present in other lineages.

Purchase this publication