From !khure

Project A5: A restudy of Madagascar-Africa breakup and evolution of Lemurs

French pi: J.J. Jaeger (with J. Dyment)
South African pi: J. Masters (with M. de Wit)

RSA Participants: Phd Student John Decker and students from Univ Fort Hare



Lemurs are living representative of one of the most primitive strata of primate evolution. They are represented by two distinct groups, the Lemurs and the Loris which split more than 37 Ma ago. Both share many primitive Primates characters but also a uniquely derived character, the tooth comb. This dental speciality is the result of the procumbent inclination of lower incisors and canines which become also peg-like, designing what is commonly called the tooth comb. Their modern geographic distribution is restricted to Madagascar for the Lemurs and to the old world tropics (Tropical Africa & Asia) for the Loris. Their fossil record is very scanty, only fossil Loris being known from the late Eocene to Pleistocene of Africa and from the Miocene of Indo-Pakistan. No fossil of Lemurs older than Holocene is presently known. Thus, the geographic origin of Malagasy Lemurs remains one of the most resistant biogeographic puzzles. Surprisingly, recent discoveries of Eocene Primates in Africa did not contribute to solve that problem. In addition, the peopling of Madagascar is not only restricted to that of Lemurs. Other endemic Madagascar mammals indicate that, according to our knowledge about the history of mammals, 4 or 5 successive waves of peopling occurred during the Tertiary, either from the Indian plate, from Africa or from both places! In order to gain a better understanding of the problem of geographic origin of Lemurs, we propose to address the problem using several distinct methods:

  1. Reinvestigating marine geophysical and DSDP data in the Mozambique Channel and vicinities, reassessing ancient and recent plate kinematics of this part of the Indian Ocean.
  2. Prospecting the Paleogene deposits of Madagascar.
  3. Prospecting the Paleogene deposits of South Africa and Mozambique.
For the first phase of the project, only parts 1 and 2 will be developed here.

Part 1: Reinvestigating marine geophysical and DSDP data in the Mozambique Channel and vicinities, reassessing ancient and recent plate kinematics of this part of the Indian Ocean.

The colonization of Madagascar by lemurs and others mammals from Africa requires either the crossing of land bridge or the transportation on natural rafts. The second hypothesis, also known as the “sweepstakes model” (Simpson, 1952), has been suggested to be very unlikely on the base of statistical calculation (Stankiewicz et al., 2006). Therefore, the “episodic emergence of inter-channel islands along the Davie Fracture Zone during the end of the Mesozoic and early Cenozoic [may have] played a significant role in facilitating this colonization process […]. Alternatively, perhaps Africa was not the source of all of the Malagasy mammals, including the lemurs” (Stankiewicz et al., 2006).

Considering the present-day bathymetry around Madagascar, there are only a few possible candidates for even discontinuous, temporary land bridges: We consider the Davie Ridge, west of Madagascar, and volcanic banks, northeast of Madagascar, to have possibly played such a role during the long and complex history of the western Indian Ocean.

The most likely candidate is the Davie Ridge, a sub meridian structure that separates the Somali Basin, north of Madagascar, from the Mozambique Basin, southwest of Madagascar. The Davie Ridge is initially a major fracture zone which recorded the southward motion of Madagascar separating from north-eastern Africa between 170 and 118 Ma, it is still seismically active and may have been reactivated when the East African Rift initiated, at ~30 Ma (Mougenot et al., 1986a, b). This complex history may translate in two ways: either the Davie Ridge was a prominent, continuous structure that remained at least partly emerged in the Mesozoic and early Cenozoic, then was progressively submerged by the subsidence of an aging crust and possibly dissected by the nascent extensional plate boundary between Somalia and Nubia – the lemur would have crossed the remains of that land bridge shortly before its total submersion; or the Davie Ridge was already a shallow submarine structure during early Cenozoic and was first rejuvenated by the initiation of the new plate boundary, either tectonically (although a significant amount of compression seems unlikely) or, more probably, by some amount of volcanism – in this case the lemur would have crossed the Mozambique Channel passing from volcano to volcano – then eroded, faulted, and affected by subsidence. Both scenarios are plausible in the context of the Somalia-Nubia plate boundary, as suggested on land by the dominantly tectonic Western Branch and the volcanic Eastern Branch of the East African Rift (e.g. Chorowicz, 2006).

We therefore propose to address this problem through a geophysical investigation of the structure of the Davie Ridge and the surrounding basins, first from available seismic data and, if required, through the acquisition of new seismic data, in order to decipher major normal faults, recent volcanism, and possible erosional surface that would mark emersion. Conversely, a detailed characterization of both the initial transform motion along the Davie Ridge at the time of its formation, in the Mesozoic, and the more recent motion of Somalia with respect to Nubia along the East African Rift and its seaward extension are also essential to discriminate among these scenarios.

Another candidate for a discontinuous, temporary land bridge is the set of bathymetric highs that separate Seychelles from Madagascar, including Bulldog Bank, Farquhar Island, Providence Reef, and further north the Amirante Bank. Most of these seamounts are shallower than 200 m and may have emerged during periods of regression in the Cenozoic. The Seychelles micro continent and India have been connected up to magnetic anomaly 27 (62 Ma), the older magnetic anomaly found on the Carlsberg Ridge flanks (Dyment, 1998; Chaubey et al., 2003; Royer et al., 2003). But another land connection, formed by the Deccan-Reunion hotspot track, has probably existed up to 35 Ma, when the Saya de Malha and Nazareth Banks (both parts of the Mascarene plateau) and the Chagos Bank (part of the Chagos-Laccadive Ridge) broke apart and the Central Indian Ridge established in its present configuration.

We propose to investigate this possibility by looking at the sparse geophysical data available on the bathymetric highs between Madagascar and Seychelles on one hand, on the Mascarene Plateau and Chagos-Laccadive Bank on the other hand. Conversely, we plan to refine the paleogeographic reconstructions obtained by the French party in the framework of other projects (CEFIPRA 1999-2003 and 2006-2009) with a special attention to volcanic plateau, seamounts, and other bathymetric highs.

A last possible candidate for a land bridge could be along the Comoros Islands, a chain of volcanic islands which may result from a hotspot plume. The chain is 5-0 Ma old, with traces of volcanic 10 Ma old in northern Madagascar (Emerick and Duncan, 1982) and shallow banks – probably former islands – connecting Madagascar to the Comoros. The most recent and active volcano (Karthala) is located on the westernmost island, Grande Comore. The distance between Grande Comore and the eastern coast of Africa is about 280 km, but this distance can be reduced by a factor of two if the northern end of the Davie Ridge was emerged. The Comoros land bridge is therefore a variation of the Davie Ridge land bridge, which may account only for recent colonization (i.e. Holocene) – before this time, the westernmost islands were not built yet! Again, beyond the work planned to substantiate the Davie Ridge land bridge hypothesis, we plan to revisit the sparse geophysical data available in the southern Somali Basin, around the Comoros Islands.

To summarize, the proposed work is twofold:

  • reinvestigating available marine geophysical (mostly seismic and DSDP data) on and in the vicinity of the potential land bridges, in order to find evidences of tectonic and/or volcanic activity and identify possible erosional surfaces which may mark the emersion of these bathymetric high;
  • reassessing ancient and recent plate kinematics of this part of the Indian Ocean, through the compilation and re-interpretation of available magnetic anomaly profiles and the use of paleogeographic reconstruction software
providing uncertainty intervals.

In term of the data required by the project, we will use the so-called “international data” (compiled by and available from the US National Geophysical Data Center - also a World Data Center - in Boulder, Colorado). This general set of single beam bathymetry, magnetics, and gravity data will be complemented by more specific geophysical data collected by the French research vessels in the last 40 years as part of the exploration of the Indian Ocean. Among these data, the magnetics and (single beam) bathymetric data have substantiated the pioneer work of Jacques Ségoufin and Philippe Patriat at Institut de Physique du Globe de Paris in the early eighties (Segoufin and Patriat, 19xx; Patriat and Segoufin, 1988). Although both have retired (Philippe Patriat is still visiting the lab occasionally), their data base remains available to our investigation and has been completed by more recent ship tracks. Conversely, several old seismic profiles (4-6 traces) acquired in the Somali Basin under the guidance of Roland Schlich are available in digital format from the seismic repository at the Institut de Physique du Globe de Strasbourg. As for the denser seismic profiles acquired over the Davie Ridge by French research teams, mostly Geoscience Azur in Nice, the digital tapes are no longer available but the paper sections have been scanned and are available to our investigations. Additional data may also be available in SISMER, the French marine geophysical data repository, in Brest, which will be consulted in this respect. Moreover, data have been acquired by the French hydrographic office in the Mozambique Channel around the “Iles Eparses” (Tromelin, Glorieuses, Juan Nova, Europa and Bassas de India), a French Territory; we have already obtained multibeam bathymetric data from them but should investigate if more is available in this area! The marine data will be completed by the gravity anomaly grids derived from satellite altimetry.

Part 2: Prospecting the Tertiary deposits of Madagascar in order to find Paleogene mammalian localities.

The biogeographic puzzle of Madagascar lemurs origins has led to numerous speculations which can hardly been resumed here. The traditional view consider than Lemurs originated in Africa and migrated from Africa to Madagascar sometimes during the Paleocene or the Eocene. This scenario relies on the fact that Lorisidae are the sister group of Lemurs, and are found in Africa since the Late Middle Eocene Fayum Locality (37 Ma; Seiffert et al., 2006) and that the molecular clock indicates a divergence age between Loris and lemurs of between 62 and 58Ma. However, the analysis of the fossil primate record of the lower Eocene of North Africa indicates that no Strepsirrhine with teeth comb did exist at that time in North Africa. The referred Strepsirrhines from Tunisia and Algeria belong to primitive groups of Strepsirrhines which could be ancestral of Lemurs, but in that case tooth comb would have been developed after 50 Ma and before 37 Ma. Another critical point is that the first Loris from Fayum are recorded in the same level as the earliest African anthropoids (Jaeger et al., 1999) which are clearly immigrants from Asia (as are rodents, anthracotheres and other contemporaneous taxa from the same locality). These Loris could therefore as well originated from Asia and enter into Africa at the same time as Anthropoid primates and therefore hold no information relative to the geographic origins of lemurs. A scenario concerning an Indian origin can equally well be supported (Marivaux et al., 2001) because of the absence of fossil remains in Madagascar, older than Holocene.

In order to gain some information about the zoogeographic history of Madagascar, we propose to organize field work to search for Paleogene fossil mammals. We have already a good record of such kind of field work which was successful in many other places as Morocco, India, Pakistan, a.s.o. The main originality of our project is to avoid unfossiliferous terrestrial deposits and to search the marine neritic deposits of the Paleocene and Eocene of Madagascar. Sections provided by ancient geological mapping work (Besairie, 1972.) indicate that some marine layers were deposited under low water depth, allowing the possible occurrence of small channels made by palaeorivers. Microconglomerates from these small channels may contain reworked micromammal remains as in many other world areas. The marine carbonate rich sedimentary environment offers an excellent protection for carbonate apatite (fossil bones and teeth), increasing the chances for fossil preservation. But carbonate matrix may have to be dissolved in diluted acid solution, which is a time consuming process. Paleocene and Eocene marine outcrops are numerous in Madagascar. They occur on the western coast, both in the North (Mahajanga Basin) and in the SouthWest of the continent- island (Morondava Basin). For logistic constraints, a first survey should be organized in the south, near the city of Majunga. Sections of Paleogene marine sediments should be surveyed using gully made by rivers cutting these deposits in an East-West direction.


  • Besairie,H. Géologie de Madagascar 1 : Les Terrains Sédimentaires Annales Geologiques de Madagascar, 1972, XXXV
  • Emerick, C. M. & Duncan, R. A., 1982, Age progressive volcanism in the Comoro Archipelago, western Indian Ocean and implications for Somali plate tectonics.-Earth Planet. Sci. Lett. 60, 415–428.
  • Jaeger, J.-J., Thein, T., Benammi, M., Chaimanee, Y., Soe, A. N., Lwin, T., Tun, T., Wai, S., and Ducrocq, S. (1999): A new primate from the middle Eocene of Myanmar and the Asian early origin of the anthropoids. Science, 286, 528-530.
  • Marivaux, L., Welcomme, J.-L., Antoine, P.-O., Metais, G., Baloch, I. M., Benammi, M., Chaimanee, Y., Ducrocq, S., and Jaeger, J.-J. (2001). A fossil lemur from the Oligocene of Pakistan. Science 294, 587-591.
  • Masters, J.C., Lovegrove, B.G. & M.J. de Wit. 2007. Eyes wide shut: can hypometabolism really explain the primate colonization of Madagascar? J. Biogeogr. 34:21-37.
  • Mougenot, D., M. Recq, P. Virlogeux and C. Lepvrier, 1986a, Seaward extension of the East-African Rift, Nature 321 (1986), pp. 599–603.
  • Mougenot, D., P. Virlogeux, J.R. Vanney and J. Malod, 1986b, La marge continentale au Nord du Mozambique: résultats préliminaires de la campagne MD40/MACAMO, Bulletin Société géologique France 8 (1986) (II, 3), pp. 419–422.
  • Stankiewicz, J., Thiart, C., Masters, J.C. & M.J. de Wit. 2006. Did lemurs have sweepstake tickets? An exploration of Simpson's model for the colonization of Madagascar by mammals. J. Biogeogr. 33: 221-235.
  • Yoder, A.D. & M. D. Nowak. 2006. Has vicariance or dispersal been the predominant biogeographical force in Madagascar? Only time will tell. Annu. Rev. Ecol. Syst. 37: 405-31.