Biogeographical Affinities of the New Caledonian Biota

SUPPORTING INFORMATION

Biogeographical affinities of the New Caledonian biota:

a puzzle with 24 pieces

Michael Heads

Journal of Biogeography

Appendix S1: Supplementary notes on New Caledonian biogeography

This appendix includes notes on methods and on each of the 24 tracks described in the main text, an analysis of the terrestrial isopods of New Caledonia, and documentation of correlated local and global patterns for New Caledonian taxa.

Methods in historical biogeography

Metapopulations and dispersal

Van Balgooy et al. (1996, p. 210) concluded their discussion of Pacific Cupanieae (Sapind.) by writing that: ‘One can only detect that dispersal is involved, because the geological history of the Pacific is not consistent with the idea that all Pacific island groups together with Australia have formed one united area which has broken up in the course of time’. This denies ordinary, local dispersal in a metapopulation, but accepts extraordinary long distance dispersal.

Modern centre of origin/dispersal theory

In their study of Pachycephalidae, Jønsson et al. (2010) calibrated the clock with endemics on Sangihe I., formed at 2-3 Ma, and Tanimbar I., formed at 1 Ma. To corroborate this first calibration, the authors used another rate, but this was also based on the assumption that an island endemic (Mimus graysoni) could be no older than its island (Socorro I. off western Mexico; Arbogast et al., 2006). As with the Indonesian islands, Socorro I. also occurs near an active plate margin and, again, the assumption that its endemics are no older than the island is very likely to give dates that are much too young. Any ‘corroboration’ between the two calibrations is meaningless. Although the spatial and chronological analyses of Jønsson et al. (2010) are not accepted here, the molecular phylogeny itself is of great interest.

The main clade in Pachycephala comprises three branches in an unresolved trichotomy: P. caledonica of New Caledonia (Grand Terre only, not Loyalty Is.), P. soror of New Guinea, and a large group ranging from Bangla Desh to Samoa, including New Guinea and New Caledonia (Grande Terre and Loyalty Is.). For P. caledonica, Jønsson et al. (2010) proposed a centre of origin in New Guinea and colonizing flights to New Caledonia, Bangla Desh, Samoa etc. But the centre of origin can be interpreted instead as a centre of differentiation in a widespread Bangla Desh – Samoa ancestor and long distance colonizing flights are unnecessary. Following initial vicariance of the three clades there has been secondary, local overlap in New Guinea and Grande Terre caused by terrane accretion, population expansion, or both.

In Pachycephala the second New Caledonian species is a form in the large Bangla Desh – Samoa group cited above, a subspecies of P. rufiventris found on Grande Terre. This was not sampled in the molecular study but probably has its closest relatives in Australia. Any overlap with P. caledonica within Grande Terre (distribution maps do not seem to be available) could be due to range expansion within the island following initial vicariance between P. caledonica and the large group that includes P. rufiventris.

The third New Caledonian Pachycephala also belongs to the large group cited above, and is a member of the diverse P. pectoralis species complex. This has a phylogeny: (Vanuatu (Bismarck Archipelago (Solomon Is. (Tanimbar I. (Australia))))). It also occurs in New Caledonia, but only on the Loyalty Is. and not on Grande Terre. Jønsson et al. (2010) interpreted the phylogeny as representing a colonization of the Melanesian arc from New Guinea (their Fig. 4 indicates colonization of Vanuatu from New Guinea via long distance dispersal north of the Bismarck Archipelago/Solomon islands, then the Bismarck Archipelago from New Guinea, then the Solomon Islands from New Guinea). Finally, there was a colonization of Australia from a centre of origin from somewhere in New Guinea/Pacific islands, where the basal grade occurs (cf. Rhipidura: track 6; monarch flycatchers, Filardi & Moyle, 2005). None of these dispersal events are necessary as the species are allopatric. In addition, dispersal does not explain the absence of the group from Grande Terre. The simplest explanation of the phylogeny is that it represents a series of differentiation events in an ancestor already widespread in Australia and the Pacific islands, including the Loyalty Is. but not Grande Terre. The biogeographical boundary between the Loyalty Is. and nearby Grande Terre is a standard one (cf. the plant Cyrtandra, Fig. 9) but is not explained in a dispersal model.

Documentation for the New Caledonia tracks

For each area, notes on the treatment in the main text are given followed by information on lower plants, seed plants, invertebrates and vertebrates. Family names for seed plant families are abbreviated. Data is included for liverworts (Miller et al., 1983), mosses (Miller et al., 1978), higher plants (Council of Heads of Australasian Herbaria, 2009), Diptera (Evenhuis, 2005a), spiders (Platnick, 2007), birds (Dickinson, 2003), and other groups as cited.

Molecular evidence was used wherever available, but only for phylogenetic and geographical information; the chronologies of molecular clock studies were not accepted. Patterns are usually cited only if the groups they belong to seem to be reasonably well-collected. For example, disjunctions are usually only recorded if closely related taxa have been collected in the gap.

1. New Caledonia – south-west Indian Ocean

Notes on the main text

Although Davis et al. (2002) invoked long-distance dispersal for Acridocarpus from Madagascar to New Caledonia, Davis et al. (2004) insisted that wind-dispersed Malpighiaceae fruit such as those of Acridocarpus ‘do not easily travel for long distances over water’ which argues against long-distance dispersal from Madagascar to New Caledonia, 10 000 km away. Another wind-dispersed member of Malpighiaceae, Hiptage myrtifolia, is a Fiji endemic most closely related to H. luzonica of the Philippines and Sulawesi, 6000 km away (Jacobs, 1955). Davis et al. (2002) considered the possibility that Acridocarpus has gone extinct in Australia due to aridity, but this is unlikely in a group that has flourished in Africa. Instead, populations may have been extirpated during subsidence of basins around the Timor Sea – Torres Strait region. Despite the problems with both long-distance dispersal and extinction, Davis et al. did not mention the main evidence for a vicariance explanation, the fact that the south-west Pacific Ocean – south-west Indian Ocean connection is a standard pattern.

Good’s (1950) account of ‘Madagascar and New Caledonia’ is a useful discussion of the topic but dispersalists have seldom mentioned the pattern; Thorne (1965) wrote simply that ‘Little significance can thus be accorded those taxa restricted to New Caledonia and the Malagasian region beyond their relict status...’. This does not account for the frequent vicariance observed between New Caledonia – south-west Indian Ocean groups and their northern relatives.

Lower plants

In algae, Millar & Bolton (2004) cited several species restricted to eastern Australia (New South Wales) and South Africa. The liverwort Cololejeunea ceatocarpa is in Réunion, New Caledonia and Hawaii, and C. cuneata is in Mauritius and New Caledonia (Miller et al., 1983). In liverworts, Cololejeunea cardiocarpa is in New Caledonia, Madagascar, southern central and East Africa, and southern United States/Greater Antilles.

Seed plants

In Phyllanthus (Phyllanth.), matK sequences link some New Caledonian members of subgen. Gomphidium with another New Caledonian clade of Phyllanthus (groups 6 and 7 of Aubréville et al., 1967-) (Kathriarachchi et al., 2006). The ITS tree instead places the New Caledonian Gomphidium with the Madagascan P. betsileanus (bootstrap support 92%). Kathriarachchi et al. (2006) regarded the New Caledonia – New Caledonia tie as ‘biogeographically more plausible’ than the New Caledonia – Madagascar disjunction and so they removed the ITS data of the Madagascan species from the combined (ITS and matK) matrix, nevertheless, the ITS affinity again fits the standard pattern.

Groups centred around the Indian Ocean include the trees in tribe Caletieae (Picrodendr.) of South Africa, Australia, Malesia, New Caledonia and Fiji (Radcliffe-Smith, 2001). The group is well-represented in New Caledonia, with four genera: Austrobuxus, Canaca, Longetia, Scagea. The last three are endemic to New Caledonia, while Austrobuxus has most of its species (14 out of 20) there.

In other angiosperms, Arthropodium (Antheric.) is in Madagascar, Australia, New Zealand, New Caledonia, and New Guinea (van Balgooy, 1971). Pigea etc. (= Hybanthus p.p.) (Viol.; Fig. 13) has a related pattern.

Disjunct clades between the south-west Pacific and south-west Indian Oceans include a group in Crossosomatales, Geissolomataceae (south-west South Africa) and its sister clade Strasburgeria (New Caledonia)/Ixerba (New Zealand). Korthalsella salicornioides (Visc.) is in Madagascar, New Zealand and New Caledonia (I. des Pins and Kouaoua, southern Grande Terre) (Aubréville et al., 1967; Molvray, 1997). The mainly Malesian/Pacific species Geniostoma rupestre (Logan.) has an outlier in the south-west Indian Ocean, G. r. var. borbonicum of the Mascarenes. It is closest to G. r. var. glaberrimum from New Caledonia and distinguishing the two ‘is often extremely difficult’ (Conn, 1980).

In other plants (data from Aubréville et al., 1967-), Agrostophyllum (Orchid.) is in Malaysia, Indonesia, Melanesia and Polynesia, with a single species in the Seychelles (A. occidentale). This last species is ‘very similar’ to the single species in New Caledonia. Cohnia (Lomandr.) is often accepted as a separate genus for the free-tepalled species otherwise placed in Cordyline. It occurs in New Caledonia and the Mascarenes (Conran, 1998). Cynorkis fastigiata (Orchid.) is known from the south-west Pacific (Fiji and Futuna) and the south-west Indian Ocean (Madagascar, Mascarenes, Seychelles) (Renz & Vodonaivalu, 1989). Cossignia (Sapind.) is in north-eastern Queensland, New Caledonia and Fiji, and disjunct in the Mascarenes (Smith, 1979-1996). Albizia guillainii of New Caledonia is not related to Asian or Australian congeners, but to African ones (sect. Zygia).

Invertebrates

The large landsnail Placostylus ranges widely around the Tasman and Fiji basins (Lord Howe I., northern New Zealand, New Caledonia, Vanuatu, Fiji, and the Solomon Is.). Its sister, Bothriembryon, is in Tasmania and south-western Australia. The sister group of the pair is Prestonella of southern Africa (Herbert & Mitchell, 2008). The authors supported a vicariance interpretation for the pattern.

In the leafhopper family Myerslopiidae (Homoptera), one tribe (Sagmatiini) ranges in Madagascar, eastern Queensland, and New Caledonia, one is in vicariant in New Zealand/Chile, and one is endemic to Juan Fernandez Is. (related families are mainly diversified in continental Africa) (Hamilton, 1999). Hamilton concluded that the New Zealand – Chile tribe, at least, is ‘virtually certain’ to date from the break-up of Gondwanaland and so it would seem likely that the Sagamatiini are also a Mesozoic group.

The spider genus Pachyballus (Salticidae) is restricted to New Caledonia and Africa (in the south, east and west). New Caledonian members of Philoliche (Diptera: Tabanidae) are ‘unequivocally’ related to Afrotropical species and have their closest relative in P. pennata of Mauritius (Burger, 1995). Aselgeoides (Homoptera: Fulgoroidea) occurs in New Caledonia, and also in the Seychelles, Madagascar and Africa (Fennah, 1969). In Psocoptera, species of Phlotodes ‘Group II’ occur on New Caledonia, Fiji and New Guinea, and also in Madagascar (Thornton, 1981). A trochiform landsnail of Norfolk Island, south of New Caledonia, is closest to a Madagascan species (F.M. Climo, pers. comm., 2002).

The bivalve genus Funafutia (Lucinidae) occurs in the south-west Pacific (Queensland, New Caledonia, Kiribati, and Tuvalu) and also around Madagascar (Glover & Taylor, 2001) but is absent from western Australia (where other lucinid genera such as Pillucina occur).

The centipede Dichelobius shows a pattern: (Queensland (Western Australia + New Caledonia)) (Edgecombe & Giribet, 2004) which may be related to this pattern, and the spiders Epimecinus (Desidae) and Syrorisa (Desidae), and the ostracod Vestalenula marmonieri (Martens & Rossetti, 2002) are also restricted to New Caledonia and Western Australia.

A New Caledonia/India disjunction may also be related and is shown by several spiders: Penionomus (Salticidae) of New Caledonia and India, Stergusa (Salticidae) of New Caledonia and Sri Lanka, and Thomiscus leucaspis (Thomisidae) of New Caledonia and India, with the gap filled by T. spectabilis: India to Australia. Thomiscus otherwise ranges east to Java/Sulawesi, and is also in the Americas.

Vertebrates

Fricke (2002) cited tripterygiid fish species with ‘bipolar west-east’ distributions, disjunct between the west Indian Ocean and the west Pacific Ocean, and ‘tripolar’ species (such as Enneapterygius elegans) which range around the Indian Ocean, Taiwan and the South Pacific, but have a broad gap in south-east Asia, western New Guinea and northern Australia.

This and similar examples indicate that the New Caledonia – Madagascar direct track is probably related to another: New Caledonia – New Guinea – Madagascar/Mauritius, which skirts Australia to the north, e.g. the group Weinmannia/Pancheria/Cunonia (Cunon.). Likewise, the marine mollusc Microvoluta joloensis ranges in New Caledonia, Fiji (Lau)/Tonga, and Wallis & Futuna (with related fossil forms in New Zealand), and is disjunct in the Philippines and Madagascar (Bouchet & Kantor, 2003).

2. Tethyan tracks: New Caledonia – Mediterranean

Invertebrates

The lithistid sponge Siphonidium occurs in Florida, the Caribbean, the Azores, Western Africa, the Mediterranean, Indonesia, and New Caledonia (Manconi & Serusi, 2008).

3. Southern South Pacific track: Australia – New Zealand – New Caledonia – southern South

America

Notes on the main text

Atherospermataceae (Laurales) are in eastern Australia/Tasmania (three genera), New Guinea (one genus), New Caledonia (one genus), and New Zealand-southern Chile (two genera). Fossil material is also known from the Kerguelen Plateau, the Antarctic Peninsula, Africa and Europe, but not America. The sister group is the family Gomortegaceae, with one genus in Chile (Heads, 2009c).

In the beetle group Adeliini (see main text) none of the four major areas occupied is confined to a single part of the phylogenetic tree and ‘This suggests that diversification had already occurred before the first area separated’ (Matthews, 1998). Fossil-based phylogeography and other dispersal analyses would instead infer that this indicated dispersal, but Matthews’ suggestion seems more likely.

In the Mesozoic, the fauna of the Murihiku terrane in New Zealand and faunas of equivalent terranes in New Caledonia and Argentina define a centre of endemism (Damborenea & Manceñido, 1992) which has persisted through the Cenozoic and is seen in many modern groups.