Rodrigues Starling

The Rodrigues Starling (Necropsar rodericanus), alternatively spelled Rodriguez Starling, is an extinct and quite enigmatic songbird species. It is the only valid species in genus Necropsar, and provisionally assigned to the starling family (Sturnidae). This bird used to inhabit Rodrigues in the Mascarenes and at least one of its offshore islets. The record of its erstwhile existence is limited to an old travel report and a few handfuls of subfossil bones.^[2]

In 1726 or shortly thereafter,^{[Note 2]} the French naval officer Julien Tafforet of the La Ressource described his encounters with the bird on an offshore islet in the Relation d'île Rodrigue, which documented his 9-month stay in 1725. In 1874, Reverend Henry Horrocks Slater, a naturalist of the British Transit of Venus expedition, found subfossil bones of a starling-like songbird on Rodrigues proper, as had magistrate George Jenner shortly before. These are generally assumed to belong to the bird Tafforet wrote about. Some additional bones were found in 1974. Together, they represent most of the skeleton, except for the spine, pelvis and the small bones, and are mainly in the Cambridge Museum. The IUCN regards the Rodrigues Starling as a valid species, because Tafforet's report and the bones provide compelling evidence that it existed.^[1][2][3][4]

N. rodericanus was frequently confused with a supposedly related but smaller species "N. leguati", described with specimen D.1792, a skin in the World Museum Liverpool, as holotype. This was also called "White Mascarene Starling" and variously considered a distinct parapatric or allopatric species, an immature, a colour morph or a female. It led some to believe that the Rodrigues Starling became extinct only about 1830, when the Liverpool specimen was collected. D.1792, however, was eventually identified as an albinistic individual of the Martinique Trembler (Cinclocerthia gutturalis).^{[5][6][7][8][9]}

Contents 1 Description 2 Taxonomy and nomenclature 3 Systematics 4 Ecology 4.1 Feeding 4.2 Extinction 5 Notes 6 References 6.1 Bibliography

Description

According to Tafforet, the bird was slightly larger^{[Note 3]} than the Common Blackbird (Turdus merula) he knew from France. This would imply a total length of about 25–30 cm, and a weight of perhaps 90-130 g. Judging from the bones, the bird was about 10% smaller than the Bourbon Crested Starling (Fregilupus varius), its presumed closest relative. Consequently, its length would have been about 25–28 cm and its weight perhaps 100-120 g when adult.^[1][10][11]

Its skeleton largely agreed with F. varius in proportions, but differed in some details. The skull was shaped somewhat differently, being longer (about 29 mm long from the occipital condyle), narrower (21–22 mm), with the eyes set slightly lower, the upper rims of the eye sockets being some 8 mm apart. The interorbital septum was more delicate, with a larger hole in its center. The bill was about 36–39 mm long, less curved than in F. varius and proportionally a bit deeper, and seems to have had larger nostrils, with the nostril openings in the bone 12–13 mm in length. The mandible was about 52–60 mm long and 4–5 mm deep proximally. Its ulna was somewhat shorter by comparison – measuring 37–40 mm versus 32–35 mm for the humerus – and the keel on its sternum was a bit lower, but its power of flight was not reduced. It had strong quill knobs on the ulna, indicating the secondary remiges were well-developed. One coracoid measured 27.5 mm in length, and one carpometacarpus was 22.5 mm long. The leg and feet had the same proportions in both; in N. rodericanus the femur measured around 33 mm, the tibiotarsus 52–59 mm, and the tarsometatarsus 36–41 mm.^[1]

The head and body plumage was white; tail and wings were partially black.^{[Note 4]} Presumably, this included the flight feathers, which are generally rich in eumelanin polymers that make the feather more robust. The wing- and tail-tips are, for example, even black in the Bali Myna (Leucopsar rothschildi), a starling with otherwise entirely white plumage. It is not known whether the non-black (presumably white) part of the wings was a patch^{[Note 5]} or wing-stripe, or simply the wing coverts or fringes or one vane, as is commonly seen in birds in general. Reconstructions tend to show the tail with black tips and lighter feather bases. But while this is more likely than white-ended black rectrices,^{[Note 6]} these too are seen in some birds and thus the tail pattern is essentially conjectural; it might just as well have had white fringes or vanes for example. The beak, as well as the feet, were reported as yellow by Tafforet.^{[Note 7][Note 8]}

Tafforet makes no mention of marked sexual dimorphism, or of a pronouncedly different juvenile plumage. His 9-month visit is likely to have encompassed at least part of the breeding season (but see below), and if so almost certainly a time where postbreeding moult took place and when immature birds were around. Thus it may very well be that outward sexual and age differences were subdued or absent in N. rodericanus. The Bourbon Crested Starling had no pronounced sexual dichromatism either, but immature birds had a buffy tinge. The beak of its adult females was about one-tenth shorter than in males; it may be that this was due to niche partitioning to better utilize the limited resources of its island home (see also Huia). Too few skull remains of N. rodericanus have been found to date to assess whether it was similarly dimorphic. Among its limb bones – of which a larger number is known – the smallest do in fact measure 10% less than the largest.^[1][5][10]

The vocalizations are described by Tafforet as "a marvellous twitter [of] many and altogether different and all very joyful [calls]".^{[Note 9]}

Taxonomy and nomenclature

The bones found by Slater's team were the basis of the first scientific discussion by Albert Günther and Alfred Newton in 1879.^[1] According to Günther and Newton, the Rodrigues bird was closely related to the Bourbon Crested Starling (Fregilupus varius). Several authors have intimated that they would rather have placed the species in Fregilupus, but none seem to have actually done so. Even though James Greenway listed N. rodericanus as a junior synonym of F. varius, his description^{[Note 10]} indicates that this was an inadvertent error. The 1974 bones allowed for a more thorough assessment, verifying that the Rodrigues bird is appropriately placed in the monotypic genus Necropsar, established by Slater in 1879. In 1900, believing Necropsar to be a mis-spelling, G. E. Shelley "corrected" it to Necrospa and attributed it – erroneously – to Philip Sclater. He also gave the authors of the species name as Günther and Newton, but their original text assigns authorship to Slater.^{[1][5][12][13]}

Masauyi Hachisuka, believing the carnivorous habits described by Tafforet to be unlikely for a starling and the lack of a crest suggesting against a close relationship with Fregilupus. He was reminded of corvids by the black-and-white plumage, and assumed the bird seen in 1725 was some sort of "chough". These at his time were often still held to include the White-winged Chough (Corcorax melanorhamphos) – which is actually not a corvid – and the "ground choughs" (Podoces). Still, Hachisuka believed Günther's and Newton's assessment regarding the bones to be accurate, only that according to him Tafforet's record could not have pertained to that bird. Thus, in 1937 he described the hypothetical "chough" as Testudophaga bicolor ("Bi-coloured Chough"). Hachisuka's assumptions are generally disregarded today for a number of reasons (see below).^[3][9][14]

Systematics

Molecular phylogenetic analysis confirms Fregilupus varius to be a starling; the affiliations of N. rodericanus, presumably but not certainly closely related to it, have been open to more dispute. Mid-20th century studies found some similarities between the Bourbon Crested Starling and the Prionopidae (helmetshrikes and woodshrikes). These are in the fairly basal songbird superfamily Corvoidea, and thus among the Passeri not at all closely related to starlings, which are more advanced songbirds in superfamily Muscicapoidea of the infraorder Passerida. The Necropsar bones, while indeed reminiscent of Fregilupus (and the Prionopidae), remind much more of a starling's though.^{[3][15][16][17][18][19]}

In recent times, several songbirds of the Madagascar region, whose relationships were believed to have been resolved, were instead found to be part of the vangas (Vangidae) or the as of 2009 unnamed "Malagasy warblers". These have both undergone spectacular adaptive radiations and might conceivably have evolved an insular starling-like form. Phylogenetically, an affiliation between the "Mascarene starlings" and either of the two Malagasy radiations – or the Prionopidae, for that matter – cannot be discounted as yet, in particular as no detailed comparisons seem to have ever been made.^{[19][20][21][22]}

Biogeographically, the Mascarenes avifauna has either affiliations with birds from Madagascar and Africa, or with South Asian lineages. The latter, however, is more common in older endemics, such as the dodos (Raphinae). While an African origin of the "Mascarene starlings" is thus more likely than a South Asian one, even this does not help much to narrow their relationships, as all possible relatives are native to Africa, and in case of the Sturnidae and Prionopidae also to South Asia.^[19][23]

As regards plumage, the White Helmetshrike (Prionops plumatus) bears an uncanny resemblance to an imagined "hybrid" between the two "Mascarene starlings", with an overall coloration recalling Tafforet's description and a crest like F. varius. Among the vangas, species of Artamella, Falculea, Leptopterus and Vanga have a similar coloration. But while most of the Afro-Asiatic lineage of starlings are either brownish-grey or have dark metallic hues, there are some light-bodied black-winged species found among them too, e.g. in the genera Creatophora, Speculipastor and Sturnia. Consequently, the striking color pattern of Tafforet's birds is of little use in determining their relationships (except perhaps by suggesting against a close relationship to the "Malagasy warblers", which are predominantly greenish).^{[21][24][25][26][27]}

Anatomy is quite characteristically autapomorphic among Passeriformes in general, but a few phylogenetic patterns can be discerned. Naturally, only Fregilupus varius can be studied in detail today. What can be said is that this species' overall myology and osteology and pterylosis is altogether passeriform. Pterylosis is no reliable trait to discern more detailed relationships, and in fact the Bourbon Crested Starling in this regard resembled a – clearly not very closely related – sickletail bird of paradise (Cicinnurus) more than either the White-headed Vanga, the basal sturnid Aplonis tabuensis (Polynesian Starling), or the more advanced European Starling (Sturnus vulgaris). The wing myology of passerine birds is rather uniform; compared to the three other taxa F. varius is fairly autapomorphic and does not agree consistently either with the starlings or the vanga. The pelvis and leg muscles of F. varius are even more strongly autapomorphic versus the comparison taxa, otherwise agreeing more with starlings than with Artamella. Its foot and toe muscles are notably more similar to starlings, at least where no different perching styles have obscured any phylogenetic pattern.^[15][28]

Fregilupus varius seems to have some synapomorphies with the more advanced Passerida – Muscicapoidea and Passeroidea. Its tenth (outermost) primary remex is quite reduced for example. This feature is characteristically found in Passerida, but varies in extent, with the basal lineages have little reduction in length, and many nine-primaried oscines of the Passeroidea having lost that feather altogether. The pattern of scales on its feet also indicates a position among the Passerida at least. But these traits cannot be known for the Rodrigues bird of course. However, as F. varius, Muscicapoidea and Passeroidea have an apparently apomorphic divided fossa tricipitalis of the proximal humerus head. This serves as attachment point for the triceps brachii muscle. In basal Passeri as well as Sylvioidea (which are basal Passerida), there is only one fossa or at most a small second one. In Muscicapoidea and Passeroidea, there always seems to be a second fossa of some size present dorsal of the first. The humerus of Necropsar was said to "differ nowise"^[29] from that of Fregilupus or advanced starlings, and in the drawings of the fossils, the double fossa is visible. Judging from the anatomical studies, F. varius – and presumably also N. rodericanus – were probably not basal Passeri and quite likely belonged to the Muscicapoidea, and may indeed have been true starlings.^[9][15][24]

That Tafforet compared his birds to a thrush and not to some other similarly-sized bird – a quail or a grosbeak for example – at least suggests that they were rather thrush-like in habitus, and thus probably belonged to the basal Passeri or the Muscicapoidea, and not to the Sylvioidea, let alone the often colourful and sexually dimorphic and/or "finch"-like Passeroidea. Even the latter have, however, evolved an island taxon with highly unusual feeding habits (the Vampire Finch Geospiza difficilis septentrionalis). On account of the vocalisations, a placement with the Sturnidae seems indeed more likely than with more basal Passeri: the latter generally have harsh voices, while starlings' songs are indeed composed of a lively, diverse and usually quite musical chatter. It is interesting to note, though, that Tafforet did not compare the vocalizations more explicitly to those of e.g. the European Starling, a very widespread and common bird he almost certainly knew.^[27][30]

The corvid hypothesis advanced by Masauyi Hachisuka can probably be dismissed outright as a somewhat bizarre case of confirmation bias. First, his reasoning is generally tenuous,^{[Note 11]} and even if it is accepted applies equally or more so to the better candidates – e.g. Prionopidae or (if one takes into account adaptation^{[Note 12]} to the harsh island habitat) Sturnidae. Second, the White-winged Chough (Corcorax melanorhamphos) is nowadays known neither to be a real chough (Pyrrhocorax) nor a corvid at all, and any perceived similarities of this bird to Tafforet's description cannot be used to argue for a corvid relationship. Third, corvids appear to be of Southeast Asian origin, and though they are widely found throughout the Holarctic and the Americas, they are not very diverse in the regions from where the Mascarenes avifauna originated – and all their species found in southern India and southeastern Africa today^{[Note 13]} belong to the crow and raven genus Corvus. Though these are known to be susceptible to island dwarfing^{[Note 14]} and to developing black-and-white plumage^{[Note 15]} they still make an unlikely ancestor for a thrush-like bird with a musical song (a fact that Hachisuka disregarded). Finally, the general improbability of two similarly-sized and somewhat carnivorous passerines (as can be deduced from Tafforet's account and the beak shape of the fossils) evolving on such a small island with its limited food resources suggests very strongly that Tafforet's bird was the same as the subfossil species. This also agrees with the fact that the native passerine birds of Rodrigues had largely non-overlapping ecological niches, and that introductions of passerine birds to the island have met with below-average success even to this day.^{[9][18][19][26][31][32]}

Ecology

Though there is no direct evidence that the bones are from the same species as Tafforet's birds, the absence of another suitable extinct or extant candidate and the severe competitive exclusion on such a resource-poor island like Rodrigues make it rather likely. No similar bird is known from elsewhere in the Mascarenes, excluding the possibility that the Rodrigues records refer to an ephemeral population of vagrants.^[33][34]

The birds found by Tafforet's party lived on the Île au Mât (today Île Gombrani – sometimes transcribed "Combrani", "Gombranis" or "Mombrani"), an offshore islet of Rodrigues. The bones were found in caves on the Plaine Corail, the island's southwestern calcarenite limestone plateau east of today's Sir Gaëtan Duval Airport. Île Gombrani is due southeast offshore the Plaine Corail; consequently, the bird must have been fairly numerous at least on the island's southwestern end in former times. The close geographical and ecological association of Tafforet's record and the bones further supports the assumption that all records refer to a single species.^{[35][36][37][3][38]}

In his report, Tafforet asserted however that he did not encounter his Île Gombrani bird on mainland Rodrigues.^{[Note 16]} He mentions some that were kept by his party, perhaps indicating that the breeding season so coincided with his 9-month visit for easily-caught fledglings to be present. No mention is made of nesting sites on Gombrani, but even 30 years earlier few small birds were found to nest on mainland Rodrigues on account of the rats (see below). While N. rodericanus might have been a ground-nester, this is not so likely; its presumed relatives are generally tree-nesting birds and starlings in particular are often cavity-nesters. There is, however, some woodland on Île Gombrani even today, and the abundant shrubs would have provided ample nesting sites. It is sometimes asserted as certain that the birds kept by Tafforet's party were not adult,^{[Note 17]} but that is conjecture; the original text is ambiguous in that respect.^{[Note 18]} With no report of nesting whatsoever, all that can be said is that Tafforet's encounter probably coincided with the end of the breeding season, after the young had fledged – if N. rodericanus was not easily caught even as adults, which is well possible and was noted in many Mascarene birds (including Fregilupus varius).^[27][39]

As regards habitat, today the Plaine Corail is largely pasture and other grassland, with sparse settlement. In the late 19th century it was even less accessible, and had an abundant growth of shrubs and small trees. Much of its original vegetation is shared by the islets offshore western and southern Rodrigues, and this allows to assess what plants grew in the birds' habitat. Among trees and shrubs that might have been used for nesting, native species almost certainly present were the dicots Gastonia rodriguesiana (Apiales), Grand Devil's-claws (Pisonia grandis, Caryophyllales), Bay Cedar (Suriana maritima, Fabales), Conkerberry (Carissa spinarum) and Antirhea bifida (Gentianales), Tree Heliotrope (Heliotropium foertherianum, Lamiales), Ludia mauritiana (Malpighiales), Portia Tree (Thespesia populnea, Malvales), Pemphis acidula (Myrtales), Zanthoxylum paniculatum (Sapindales), and Lycium mascarenense (Solanales). Monocots were not certainly present in significant numbers; they might have included the Hurricane Palm (Dictyosperma album var. aureum, Arecales) and Pandanus heterocarpus of the Pandanales.^[36]

With the forest on the Plaine Corail less dense than on the rest of the island, ground cover must have been abundant even when more trees were present. Native herbaceous dicots that provided cover, possibly nesting grounds and maybe supplementary food for the birds include Psiadia rodriguesiana and Rhamphogyne rhynchocarpa (Asterales), Aerva congesta and Achyranthes aspera var. argentea (Caryophyllales), Hypoestes inconspicua and Nesogenes decumbens (Lamiales), Oldenlandia sieberi var. congesta (Gentianales), and White-edged Morning Glory (Ipomoea nil) and I. rubens (Solanales). Euphorbia thymifolia (Malpighiales) and Ipomoea leucantha might be native, or introduced by early visitors. Poales of note on Îsle Gombrani and the Plaine Corail are the Tropical Fimbry (Fimbristylis cymosa ssp. cymosa, a sedge), and Bermuda Grass (Cynodon dactylon) and Stenotaphrum sp. which are true grasses. The only significant basal tracheophyte in this habitat is the spikemoss Selaginella balfourii.^[36]

Tafforet was not sure why the birds were absent from Rodrigues proper, though he was inclined to attribute it to birds of prey scaring off N. rodericanus.^{[Note 19]} This might refer to migrant falcons (Falco), vagrant Réunion Harriers (Circus maillardi) which were more widespread in the Mascarenes in former times, and of course to the resident Rodrigues Owl (Mascarenotus murivorus). The former two, with their presence only temporary, are unlikely to have exerted a marked evolutionarly pressure on such a sizable endemic resident landbird. The owl on the other hand was not much larger than N. rodericanus and though probably able to subdue them would – in particular its smaller (Morepork-sized) males – have preferred less hefty victims. These were abundant among the native birds, namely the Rodrigues Warbler (Acrocephalus rodericanus) and Rodrigues Fody which both still survive, and perhaps the bulbul^{[Note 20]} and the "Old World babbler"^{[Note 21]} which went extinct at an unknown (but probably early) date. Even the small pigeons that used to inhabit Rodrigues^{[Note 22]} would likely have made as good prey to any raptor as did Tafforet's large passerine with its meat-ripping bill. Thus, predation pressure by carnivorous landbirds was probably less significant for N. rodericanus than for the other native passerines. On the other hand, frigatebirds were earlier noted to frequent the mainland to hunt hatchling sea turtles^{[Note 23]} and to bully boobies for food^{[Note 24]} at dusk, so it may be those Tafforet had in mind. It is unlikely that the large passerine observed by Tafforet was originally confined to the islet, though seasonally it may have well been so. Overall, the absence of the bird from Rodrigues proper in 1725 is somewhat puzzling, but it may indicate an already-precarious state of the mainland fauna of Rodrigues (see also below).^[16][40]

Feeding

Tafforet also recorded some details on the birds' feeding habits. The food of the adults is described as "nothing else but [seabird] eggs or some turtles dead of hunger",^{[Note 25]} and they apparently made effective use of the stout beak in foraging, tearing turtle flesh from the shells^{[Note 26]} and presumably breaking open eggs. The (possibly young) birds kept by his party were given seeds of some tree^{[Note 27]} and chopped-up cooked meat^{[Note 28]} – possibly salt pork, beef or mutton which were the standard fare for ocean voyages at that time, but perhaps more likely turtle and tortoise meat procured on the island, as his group was effectively stranded on Rodrigues. Tafforet notes that the birds preferred the meat.^{[Note 29]} Since it is unclear what vegetable food precisely Tafforet's party offered the birds, nothing can be said for certain on this. However, it is clear that the birds were at least not exclusively vegetarian. The skull shows an attachment scar above the temporal fossa which indicates it was used for "gaping" and forceful probing as in starlings. The supraoccipital ridge on the skull is quite strongly developed and a biventer muscle attachment in the parietal region below it is conspicuous. This indicates strong neck and jaw muscles, agreeing well with how Tafforet describes the birds' procuring meat from dead turtles. The presumably related Bourbon Crested Starling (Fregilupus varius) had also notably robust neck and jaw muscles and a good "gaping" ability.^[1][15][41]

These "turtles dead of hunger" may have been the giant tortoises then found on Rodrigues,^{[Note 30]} but perhaps more likely Green Turtles (Chelonia mydas) and Hawksbill Turtles (Eretmochelys imbricata). The abundant vegetation on Île Gombrani is unlikely to have allowed land tortoises to starve even if they were able to reach the islet (which is not certain). Seabirds whose eggs would have been technically available as food^{[Note 31]} to N. rodericanus were the Brown Noddy (Anous stolidus), Lesser Noddy (A. tenuirostris), White Tern (Gygis alba) and Roseate Tern (Sterna dougallii) which still are found on and around Rodrigues, as well as the Indopacific/Indian Ocean Sooty Tern (Onychoprion fuscatus nubilosus), Abbot's Booby (Papasula abbotti) and the Red-footed Booby (Sula sula) which are now extirpated or uncertain breeders on Rodrigues but were recorded by Tafforet and others. The Great Frigatebird (Fregata minor) and perhaps also the Lesser Frigatebird (F. ariel) were formerly recorded on Île Gombrani, but only to prey on other birds' eggs; though they bred on nearby Île Frégate, these aggressive birds are unlikely to have their eggs attacked successfully by a passerine. Bourne's Petrel (? Pterodroma sp.) which is entirely extinct, as well as the extirpated Barau's Petrel (P. baraui) and Mascarene Petrel (Pseudobulweria aterrima), seem to have nested near the top of Rodrigues' mountains in burrows, and their eggs were thus inaccessible to avian predators. Whether the eggs of the Wedge-tailed Shearwater (Puffinus pacificus), White-tailed Tropicbird (Phaeton lepturus) and Red-tailed Tropicbird (P. rubricauda) – cavity-nesters still breeding on offshore islets of Rodrigues – would have been readily available prey for N. rodericanus is doubtful. With Île Gombrani formerly noted for its abundance of Sooty Terns, it seems that this species yielded most of the "eggs of the fishing birds" Tafforet mentions as staple food of his songbird. Particularly the noddies, noted for their extremely docile^{[Note 32]} behaviour, make notable secondary candidates.^[42][43]

But although the seabirds managed to raise three broods in some years at least (as attested by Leguat^{[Note 33]}), given that eggs cannot always have been available and considering that Tafforet does not note his birds to refuse the seeds entirely, N. rodericanus obviously ate some other food. Notable native plants of its habitat, which might have furnished fruits or seeds to complement its diet, are listed above. Apart from the occasional hatchling gecko or turtle, invertebrates probably made up much of the remaining animal food of this bird. The birds are likely to have preyed on the larger-sized of the native arthropod fauna – spiders,^{[Note 34]} millipedes,^{[Note 35]} centipedes,^{[Note 36]} beetles,^{[Note 37]} butterflies^{[Note 38]} and moths^{[Note 39]} – or more likely their caterpillars – Orthoptera,^{[Note 40]} Hymenoptera,^{[Note 41]} Hemiptera,^{[Note 42]} damselflies,^{[Note 43]} the Madagascan Marbled Mantis (Polyspilota aeruginosa), the stick insect Xenomaches incommodus – which is apparently also extinct today –, and presumably the antlion Myrmeleon obscurus.^{[44][45][46][47][48]}

It is not certain whether the viviparous flesh-flies (Sarcophagidae) Leguat noted^{[Note 44]} were native or early introductions; the mysterious and supposedly native Sarcophaga mutata might simply have been the globally widespread Red-tailed Flesh-fly (S. haemorrhoidalis). Otherwise, perhaps apart from the mid-sized Aegophagamyia remota, Odontomyia nigrinervis and Villa sexfasciata, no Diptera large and lumbering enough to make easy prey for Tafforet's bird seem to have been indigenous to Rodrigues. Terrestrial crustaceans were largely absent too; perhaps only the apparently native woodlouse of genus Oniscus or Porcellio would have been available. There were similarly few annelids; at least in the woodlands the large earthworm Amynthas rodericensis was abundant, but it is unclear whether the birds foraged in such habitat often.^{[47][49][50][51]}

The larger Muscicapoidea – mimids, trushes and starlings – are noted for including an unusual amount of land snails in their diet; some are expert in cracking open snail shells using their bill and tools such as small stones. Such potential prey was rather abundant on Rodrigues in former times. Namely the genera Gonospira^{[Note 45]} of the Streptaxidae and Tropidophora^{[Note 46]} of the Pomatiidae have undergone adaptive radiation on the island. Also large enough to furnish food for Tafforet's birds and native to Rodrigues were the possibly extinct Achatinellidae Elasmias jaurffreti, the Euconulidae Dancea rodriguezensis (extant) and Plegma bewsheriana (extinct), an extinct large Pupilla sp. (Pupillidae), and the and the slug Vaginula rodericensis (Veronicellidae). The now-extinct snail species are plentiful in the same caves on the Plaine Corail where the bones of N. rodericanus were found. It has never been studied whether some of the snail shell pieces recovered were smashed by birds; for the time being, it can thus only be noted that none of the other native birds of Rodrigues are known or suspected to have utilized the abundant terrestrial Gastropoda of the island as food resource.^{[27][37][52][53][54]}

Extinction

The cause for this intriguing bird's extinction and its date are not known, but can be deduced with fair certainty. With the dating of Tafforet's observation – and the identification of him as the author, which had long eluded researchers^[13] –, a terminus post quem for the extinction is established. It is more difficult to assess final extinction dates, as some remnant population may linger on for a long time (see also Lazarus taxon). However, Rodrigues is a small island, and was often visited throughout the 18th century to capture tortoises and sea turtles (until Cylindraspis peltastes and C. vosmaeri went extinct about 1800, too). By 1740, a major trade in these had developed, as their relatives on the other Mascarene islands were already gone or nearly so. The island must thus have been surveyed to a considerable degree by the mid-18th century. When A. G. Pingré during the French 1761 Transit of Venus expedition observed the seabird colonies of Île Gombrani in 1761, he found them still healthy; no trace of Tafforet's bird was noted, however, while the burgeoning Testudines-based economy of the island offered little other food^{[Note 47]} than tortoise dishes in the nearly 15 weeks^{[Note 48]} he stayed. Neither did Philibert Marragon see the birds some decades later; he already remarked upon the tortoises' impending extinction though. Quite likely, N. rodericanus became extinct within one decade of 1740.^{[55][56][57][58]}

That Tafforet did not encounter it on the main island, with a population essentially confined to one or a few islets and vulnerable to any stochastic catastrophe such as a tsunami or Mauritius cyclone, is suggestive of an extinction shortly thereafter. It is uncertain, however, whether his survey of the inhospitable^{[Note 49]} Plaine Corail region was thorough enough to find any birds remaining there. François Leguat, who was marooned with his companions in 1691–1693 on the north coast where now is Port Mathurin, does not explicitly mention the birds either in his extensive writings. Though his party visited most accessible parts of the island, he does not mention the striking karst-like landscape^{[Note 50]} of the Plaine Corail – a major tourist attraction today –, and probably only saw it from a distance. He discusses a low-lying plain^{[Note 51]} of several square kilometers, but as he extols the richness of its deep soil^{[Note 52]} and ample tree-cover^{[Note 53]} the La Ferme region – extensively logged nowadays, but then still pristine –, where several small rivers have deposited abundant sediments, makes a far better match for this locality.^{[13][59][60][61]}

This may be taken to suggest that the birds were in fact confined to the island's southwest, and perhaps only common on offshore islets, as early as 1690. Of the latter, Leguat visited several, almost certainly including Île aux Diamants and the sandy islet next to Port Mathurin's harbour, perhaps Île aux Fous north of these at the reef edge, as well as some well-wooded^{[Note 54]} islets which were most likely Île aux Cocos and Île aux Sables northwest of the mainland. He described the multitude of birds nesting on the latter.^{[Note 55]} Gombrani is hard to reach on foot without crossing the Plaine Corail with its numerous sinkholes and caves (neither of which is mentioned in Leguat's report), and though having some trees is at least today mainly covered in lower vegetation. Also, Leguat does not discuss the Rodrigues Warbler (Acrocephalus rodericanus), which must in his time have inhabited the mainland where it is still present today (albeit precariously rare), and thus there are some places on the mainland he did not visit.^[62]

He mentions seeing the Rodrigues Fody (Foudia flavicans) but no other songbirds^{[Note 56]} except "very few swallows"^{[Note 57]} – most likely vagrant Mascarene Martins (Phedina borbonica). These resemble the grey-brown Crag Martin (Ptyonoprogne rupestris) that is not rare in France. However, with the colour pattern described by Tafforet, N. rodericanus may well have resembled the black-and-white House Martin (Delichon urbicum) – the most common swallow in European cities and towns – if seen in flight from below. Thus, it is not entirely certain that Leguat never encountered the "starling". What can be said is that the absence of Tafforet's bird in Leguat's detailed testimony is best explained by its general absence from the Port Mathurin region – and perhaps the rest of the island, or most of it – in 1691–1693.^[11][63]

In general, the cause of the birds' disappearance was almost certainly some introduced species, rather than^{[Note 58]} habitat destruction or overhunting, as the reports of the first long-term inhabitants contain no mention of the bird any more. Feral goats and other likely candidates were only present after Leguat's visit, but he noted the large numbers of rats, for which he had no satisfying explanation^{[Note 59]} (but see below). The initial rat population has been identified from old bones as Black Rats (Rattus rattus), though starting in the late 18th century these were probably replaced by the Brown Rat (R. norvegicus) which predominates today. As a tree-climbing predator of nestlings and eggs, the Black Rat would certainly represent a major threat to any insular endemic songbird. Species likely to have carried avian diseases to Rodrigues were not introduced until after the disappearance of N. rodericanus, and thus it seems as good as certain that Tafforet's bird was killed off by the Black Rat, succumbing when these finally reached the offshore islets. At Leguat's time, he noted that the islands he visited seemed rat-free,^{[Note 60]} and he reasoned that the Rodrigues Grey Pigeon ("Alectroenas" rodericana) nested only there because on the mainland rats were already too numerous.^{[Note 61][9][64][65]}

It is not clear when Black Rats arrived on the island. Leguat's report is the first comprehensive one, and he mentions the main island overrun with them – and he does not mention the bulbul and presumed Old World babbler that once lived there. While Arabic, Swahili and maybe Pandyan sailors knew Rodrigues as (probably) Diva Harab ("Desert Island")^{[Note 62]} before the 16th century, the Late Medieval voyages of the Indian Ocean trade seem quite long ago even for the extreme overpopulation of rats observed by Leguat to develop. The island was relocated by Portugal in 1507 and visited (mainly by Dutch ships) occasionally from 1601 on; Leguat mentions inscriptions cut by Dutch sailors in the bark of several trees.^{[Note 63]}

With N. rodericanus as rare as to be overlooked for two years in 1693, still extant in 1725,^{[Note 64]} but extinct shortly thereafter, and some mainland birds apparently already extinct by 1691, it was probably roughly between 1550 and 1650 that the Black Rat settled Rodrigues.^{[36][66][67][68]}

Notes

References

Bibliography