Genus Geochelone

Fitzinger, 1835
Typical tortoises

Recognition
This is the largest genus of tortoises, both in number of living species (24, but one probably extinct) and in size attained (to 130 cm, and 400 kg). It is pantropical with living species in Africa, Madagascar, India, Sri Lanka, Southeast Asia, Indonesia, South America, and on the oceanic Aldabra and Galápagos Islands. The species complex of the latter island group is discussed below. Formerly the genus was even more widespread, as fossil species are known from North America as far north as southern Canada, the British Isles, western Europe, northern Africa, the Middle East, Central Asia, and additional islands in the Indian Ocean (Auffenberg, 1974). Fossil Geochelone have even been found on Ellesmere Island (78°N) (Estes and Hutchison, 1980).
Adults are usually more than 25 cm in carapace length. The hingeless carapace varies in shape from oval to elongated, and from domed to somewhat dorsally flattened (especially in males of some species) to saddlebacked (narrower and elevated anteriorly, wider and lower posteriorly). Some form of cervical indentation or notch is usually present, and the anterior marginals may be flared or downturned. Both anterior and posterior marginals may be serrated. Usually 11 marginals lie on each side of the carapace, and the supracaudal is usually single and undivided. There are no submarginal scutes, but a cervical scute is present in most species. The anterior neurals are alternately octagonal or quadrilateral in shape, and usually two suprapygal bones are present: the larger anterior suprapygal bifurcates posteriorly and partially surrounds the smaller posterior. The plastron lacks a hinge and varies from narrow to broad. Its posterior notch may be deep, shallow, or nonexistent. The thickened gular scutes are paired and broader than long in all species but Geochelone yniphora, which has them fused into a long single element; the entoplastron is almost always anterior to the humero-pectoral seam. Axillary and inguinal buttresses are short, and barely touch the costals. The skull is short to moderate with a hooked, sometimes bi- or tricuspid upper jaw. Sides of the upper jaw are usually serrate. The triturating surface of the maxilla is strongly ridged, but no medial ridge occurs on the premaxilla. The maxillae do not contribute to the roof of the palate, and the anterior orbito-nasal foramina are small and concealed from ventral view. The temporal arch is weak to moderately wide. The prootic bone is well-exposed dorsally and anteriorly, and the quadrate bone usually encloses the stapes. The limbs are club shaped with no toe webbing; each forefoot has five claws. In a few species, the tail ends in a large horny scale.

Subgenera
Living species have been separated into several subgenera. Auffenberg (1974) recognized six:
1) Aldabrachelys Loveridge and Williams, 1957: G. gigantea;
2) Astrochelys Gray, 1873a: G. radiata, G. yniphora;
3) Chelonoidis Fitzinger, 1835: G. carbonaria, G. chilensis, G. denticulata, and the G. elephantopus complex;
4) Geochelone Fitzinger, 1835: G. elegans, G. pardalis, G. platynota, G. sulcata;
5) Indotestudo Lindholm, 1929: G. elongata, G. forstenii, G. travancorica;
6) Manouria Gray, 1852: G. emys, G. impressa.
However, we feel Indotestudo and Manouria are sufficiently different to warrant full generic rank, and Crumly (1982a) feels Indotestudo is monophyletic (see Diff. Geochelone/Indotestudo and discussion under Indotestudo).
Bour (1980a) elevated some of these subgenera to full generic status and partitioned the genus Geochelone. Under his 1980a system the living species were reassigned to the following genera: Aldabrachelys (gigantea), Astrochelys (radiata, yniphora), and Chelonoidis (carbonaria, denticulata, chilensis, and the various Galápagos taxa). The species formerly living on Mauritius, Rodriques and Réunion (the Mascarene group) are assigned to the genus Cylindraspis. Later, Bour (1982a) created a new genus, Dipsochelys, for the Aldabra, Seychelles and Madagascar Geochelone tortoises, because the type species of Aldabrachelys in fact is a Geochelone denticulata (Pritchard, 1986). Crumly (1982a) has shown that, based on 26 selected cranial features, Astrochelys, Chelonoidis, Geochelone and Manouria do not share derived features and are not monophyletic. Aldabrachelys /Dipsochelys (including several fossil forms, as well as G. gigantea) is probably monophyletic since its members share the derived vertically elongated external narial openings. Cladistic analyses by Crumly have indicated possible relationships between G. (Geochelone) pardalis, and G. (Astrochelys) yniphora, and among G. (Geochelone) elegans, G. (Aldabrachelys) gigantea, and G. (Astrochelys) radiata, but final decision must await further anatomical studies, inclusion and analysis of other tortoise genera, and consideration of functional hypotheses. The (sub)genera, as now recognized, are often no more than convenient geographical assemblages. Until Justin Gerlach has published his cladistic analysis of 120 skull characters of all tortoise (sub)genera, we concur with Crumly (1982a) and adopt the conservative view.

The Galápagos tortoise complex
The Galápagos Archipelago (Galápagos), 965 km west of Ecuador on the Equator in the eastern Pacific Ocean, consists of 121 islands, with 13 of these larger than 10 km2 (Snell et al., 1995). The climate of the islands is dominated by the ocean currents. During the hot season, from December to May, the warm El Niño current flowing from Panama is dominant. Both the sea and the air temperatures are high. During the cool season, from June to November, the cold, southern Humbolt current is dominant and both sea and air temperatures are much cooler. The annual rainfall is extremely variable with both drought and flood years. The majority of the precipitation falls during the hot season, while a heavy mist, known locally as "garúa", is common during the cool season, especially at higher elevations. The lower parts of the islands are very dry; however, some of the higher (over 300 m) of the numerous craters are often cloud covered and have a damp climate and more lush vegetation, especially on the windward sides of the islands.
Nowadays the giant tortoises are confined to Pinta (Abingdon) Island, Pinzón (Duncan) Island, Española (Hood) Island, San Cristóbal (Chatham) Island, Santa Cruz (Indefatigable) Island, Santiago (James) Island, and Isabela (Albemarle) Island; the Fernandina (Narborough) Island population probably is extinct. They probably first reached these islands by rafting or simply floating from mainland South America. Marlow and Patton (1981) have shown in their electrophoretic studies that the Galápagos species share no greater similarity with mainland Geochelone carbonaria, G. chilensis, or G. denticulata than they do among themselves, so the most direct common ancestor of the Galápagos species is undoubtedly extinct. The closest living relative of the Galápagos tortoise is Geochelone chilensis. The split between G. chilensis and the Galápagos tortoises probably occurred 6-12 million years ago (Caccone et al., 1999).
Captive Galápagos tortoises are often confused with the Aldabra tortoise, Geochelone gigantea, but there are several easily recognized pertinent differences. The Aldabra tortoise has a thicker, domed carapace usually with a cervical scute; that of the Galápagos tortoise is thinner, lacks a cervical scute, and often is saddlebacked. Also, some of the Galápagos species reach a greater size. The heads are much different. That of the Aldabra tortoise is more pointed and wedge shaped with the two halves of the large, divided prefrontal scale lying parallel and not diverging posteriorly; the frontal scale is relatively small. The head of the Galápagos species is not wedgeshaped, but instead more blocklike or squared off. Its prefrontal scale is divided into two short, broad scales which diverge posteriorly, and the following frontal scale is about the same size.

One of the most confusing situations in turtle systematics and biology involves the status of these giant tortoises inhabiting the Galápagos Archipelago. All are members of the genus Geochelone Fitzinger, 1835, but are they separate species or merely subspecies of a single species? Most recent systematic accounts have treated them as subspecies of either Geochelone elephantopus (Harlan, 1827) or G. nigra (Quoy and Gaimard, 1824b) (Wermuth and Mertens, 1961, 1977; Hendrickson, 1966; McFarland et al., 1974a; Pritchard, 1979, 1996a; Iverson, 1992; Pritchard, 1996a), but Boulenger (1889), Bour (1980a), Fritts (1983), and Ernst and Barbour (1989), and Siaca and Fritts (1995) considered them valid species. Most are allopatric with no evidence of interbreeding, and, when cross breeding occurs in captivity, a large proportion of abnormal individuals results (Robert P. Reynolds, pers. comm.). We decided to follow Fritts (1983), who considers each geographically isolated population a separate species, until a genetic analysis is completed.
Several forms of evidence, other than the obvious differences in shell shape (domed versus saddlebacked) and carapace lengths, strengthen the separate species approach to the various island populations. When captives are kept in mixed species groups subadult courtship is normally directed only toward members of the same island population (Márquez et al., 1988; Thomas H. Fritts, pers. comm.). Captive crosses between tortoises from different island populations exhibit lower fertility rates and higher mortality rates than crosses between members of the same island population (MacFarland et al., 1974b; Márquez et al., 1988; Thomas H. Fritts, pers. comm.). Hatchlings from interpopulation crosses have an abnormally high incidence of abnormalities, including albinism (Robert P. Reynolds, pers. comm., Thomas H. Fritts, pers. comm.), and Linda J. Cayot has informed us that, although there are no written records to confirm this, local anecdotal knowledge is that about 33% of the offspring from such early crosses were albino. Apparently offspring that survive to headstart size are from pure genetic crosses between tortoises of the same population. Some of the successful "hybrid crosses" may be merely crosses between subspecies of the same species, such as between Geochelone vicina and G. guentheri (whose taxonomy is totally confused; Linda J. Cayot, pers. comm.) and the closely related G. microphyes and G. vandenburghi. Differences also exist between populations in the number of eggs per clutch, the number of clutches laid each season, and the seasonality of nesting (MacFarland et al., 1974a, 1974b; Swingland, 1989a). Published studies on the biochemical relationships among the populations indicate a common ancestor for all Galápagos tortoises, and overall genetic similarity; but some low variation (8%) does exist (Marlow and Patton, 1981; Patton, 1984). Surprisingly, the low level of genetic variation does not support the obvious morphological variation between populations. Of the island populations studied by Patton (1984), G. chathamensis is the most divergent, followed next by G. hoodensis and G. abingdoni, while G. nigrita, G. vicina, G. darwini, and G. ephippium seem to form one complex. Finally, wild territoriality occurs in saddleback species (Fritts, 1983; Fowler de Neira and Roe, 1984), but not in those with domed shells. This has always be explained as a result of food limitations in the dry saddleback island habitats, but territoriality is a major genetically controlled behavior adaptation.

Further confusing the issue, the name G. elephantopus has been misapplied to tortoises occurring on the islands of Isabela (Wermuth and Mertens, 1961, 1977), Pinta and Pinzón (Garman, 1917), when it really belongs to the extinct saddlebacked form of Santa Maria (Charles) Island. Here we treat all 12 possibly extant island forms (MacFarland et al., 1974a) as full species, following a slightly modified version of Van Denburgh's (1914) species list for the names: abingdoni (Pinta), becki (Volcan Wolf on Isabela), chathamensis (San Cristóbal), darwini (Santiago), ephippium (Pinzón), guentheri (Volcan Sierra Negra on Isabela), hoodensis (Española), microphyes (Volcan Darwin on Isabela), nigrita (Santa Cruz), phantastica (Fernandina), vandenburghi (Volcan Alcedo on Isabela), and vicina (Volcan Cerro Azul on Isabela). Variances from Van Denburgh (1914) include substitution of nigrita (Duméril and Bibron, 1835) for porteri (Rothschild, 1903), over which it has priority, and the dropping of the form wallacei (Rothschild, 1902) (Rabida or Jervis Island; Van Denburgh, 1914) which is based on a single specimen considered an artificial introduction (Snow, 1964; MacFarland et al., 1974a; Pritchard, 1996a).
The 12 species differ in a number of morphological characters such as the color, thickness, and overall shape of the carapace, maximum size, and the lengths of the neck and limbs; however, there is much variation within each taxon, and the differences between some taxa are small (MacFarland et al., 1974a). The major differences occur between those with domed carapaces (nigrita and vandenburghi) and the saddlebacked species (abingdoni, becki, ephippium, hoodensis, and especially phantastica), in which the carapace is strongly flared and upturned above the neck, an adaptation which allows the tortoise to stretch its neck and head upward to a greater extent, allowing feeding on high vegetation or for territorial displays. The saddlebacked shell probably evolved from the domed form, and Fritts (1983) has reported some differences between the saddlebacked species that suggest this type of shell may have evolved more than once. Pritchard (1979) presents good profile drawings of the various saddlebacked species. In an electrophoretic study of blood proteins from seven island species, Marlow and Patton (1981) found that genetic similarity is unrelated to carapace shape, suggesting that the shape has been independently derived several times on separate islands.
Most natural history data reported for Galápagos tortoises has been published under the blanket name "elephantopus", and can only be properly applied to species if the island of origin is given. Unfortunately, this has invalidated several good reproductive studies of captive tortoises. In addition, the only karyotypes described, 2n = 52: 26 macrochromosomes, 26 microchromosomes (Goldstein and Lin, 1972; Benirschke et al., 1976; Dowler and Bickham, 1982), and a DNA fingerprinting study (Ryder et al., 1989) were reported as "elephantopus" with no locality data.

Undescribed Taxon
A yet undescribed, very small population of saddleback-type tortoises occurs in the dry northwestern part of Santa Cruz Island (Geochelone sp (Sta Cruz Is); Geochelone sp (Sta Cruz Is) 2). Thomas H. Fritts (pers. comm.) reports on these:

"The animals from Cerro Montura, northern Santa Cruz (Indefatigable) Island are from an enigmatic population of tortoises only poorly understood. They may represent an undescribed species or merely a secondary population of a saddleback tortoise species from another island. They are obviously not G. nigrita, the tortoise that occurs in wet habitats on Santa Cruz Island.
The presence of an unknown number of these saddleback tortoises on northern Santa Cruz has been documented on several occasions since Dr. David Snow, then director of the Charles Darwin Research Station, investigated the population in the early 1960s. Ironically, Rollo Beck collected a saddlebacked tortoise in the same area in 1901, but apparently the population even at that time was sufficiently small that no one recognized the possibility of a tortoise population in the dry zones of Santa Cruz in subsequent years. The status of this enigmatic population awaits detailed surveys of the large area potentially inhabited and the results of genetic comparisons of this population with other tortoises in Galápagos."

Dr. Eddie Louis of the Henry Doorly Zoo in Omaha, Nebraska, and Drs. Jeff Powell and James Gibbs at Yale University are currently performing those genetic studies, which seem to indicate that this population may be introduced from another island (Linda J. Cayot, pers. comm.). Pritchard (1996a) presented a picture of a juvenile (p. 45), so apparently this population does successfully reproduce.

Remark
Pritchard (1996a) uses the name Testudo nigra Quoy and Gaimard, 1824b as the senior synonym of Geochelone elephantopus (Harlan, 1827), the namesake of the Galápagos tortoise complex. He first proposed this name change in 1984, on the basis of priority, in a footnote to a paper on the Pinta Island tortoise published in Noticias de Galapágos (39: 23), and it has been used as late as 1994 by David. However, still another name precedes Testudo nigra: Testudo californiana Quoy and Gaimard, 1824a. Pritchard (1996a) states that the name californiana is a nomen oblitum, but it is no more so than the name nigra (which he resurrected in 1984) since neither had been used for over a century, while the name elephantopus has been in popular use for about 40 years. Zug (1997) published a note covering these points, and the unresolved status of major nomenclatural issues in the Galápagos tortoises. Pritchard's (1997) reply indicates that the controversy continues.

Species identification
Jump to the key of the genus: Page 138:Genus Geochelone, or directly jump to the species of the [l][m]Text Key[/m][r]Page 149[/r]Page 149:Galapagos-complex[l][m]Text Key[/m][r]Page 65: Genus Acanthochelys[/r] (G. elephantopus/[l][m]Glossary[/m][r]niger[/r]nigra)