Recently, free-drifting float experiments revealed for the first time how WMDW newly formed in the Gulf of Lions was advected several hundreds of kilometres away by sub-mesoscale coherent vor- tices, SCVs (Testor and Gascard, 2003). The trajec- tories of some SCVs were observed to cross the western Mediterranean Sea from the Gulf of Lions down to the Algerian coasts over a period of 5 months. During their journey, SCVs are able to transfer newly formed WMDW properties to the ambient waters. The upper part of SCVs would be warmed by LIW, whereas the bottom part of SCVs would be cooled by old WMDW. In some cases (Testor and Gascard, 2003) the interaction of an LIW eddy with an SCV has been observed. These eddy-eddy interactions are thought to influence the large-scale spreading of intermediate and deep waters. Moreover, the advection of WMDW by SCVs implies a role in the large-scale thermohaline circulation and in the ventilation of the deep western Mediterranean Sea.
The Channel of Sardinia is a key place for WMDW circulation due to its bathymetry. On their way from the Algerian basin, those waters flowing above 2000 m depth will be able to flow into the Tyrrhenian Sea, whereas those flowing deeper than 2000 m will be forced to flow along the western continental slope of Sardinia.
DEEP-SEA FAUNA AND BENTHIC FLUXES
Information on Mediterranean biogeography can be extracted from non-quantitative historical stud- ies, but considerable caution is needed. During his investigations in the Aegean Sea, Forbes (1844) reported that life did not exist deeper than 300 fath- oms (= ca 550 m), and used the term “azoic zone” to describe his finding. This led to a misunderstanding which was not resolved until the beginning of the 20th century. Although Marenzeller (1893) reported that species occur deeper in the Levant than else- where in the Mediterranean, his records were later considered suspect, an artifact resulting from a «sys- tematic mistake in the depth measurements» that «needs to be cleared up in the future» (Fredj and Laubier, 1985).
The present-day Mediterranean deep water fauna is less closely related to the Atlantic bathyal than was the case in the Pleistocene (Barrier et al., 1989).
This disparity is attributed in part to the shallow Gibraltar sill that bars the deep water of the Atlantic Ocean from entering the Mediterranean, and to the Mediterranean outflow that bars the entry of the deep water Atlantic fauna into the Mediterranean (Salas, 1996). The onset of warmer climates led to the demise of many cold stenothermic and stenoha- line species and to the eventual impoverishment of the bathybenthos. In addition, the extreme oligotro- phy of the Levantine Sea prevented settlement by members of the Atlantic bathyal that would have been able to cross the shallow Gibraltar Straits and the Siculo-Tunisian sill (< 400 m) (Péres, 1985). The recurring stagnant (dysoxic and anoxic) Quater- nary episodes resulted in a reduction, or even the extinction of deep bottom-living fauna unable to avoid annihilation by adapting to a shallower depth: Van Harten (1987) reported that «several species of deep-water ostracodes that are still common in the western Mediterranean became extinct in the eastern Mediterranean basin at the onset of early Holocene S1 sapropel deposition». Bacescu (1985) believed that the bathyal bottoms of the Levant are still «unfavourable», or even, «azoic», after the last sapropelic event, dated between 9000 and 6000 years BP, following the suggestion (George and Menzies, 1968) «that sufficient time has not elapsed to allow colonisation of the deep-sea floor».
Bouchet and Taviani (1992), in a very provocative paper, suggested that much of the Mediterranean deep-sea fauna is made-up of non-reproducing pseudopopulations that have entered the Mediter- ranean as meroplankton with the Atlantic inflow at Gibraltar. However, the populations of the most com- mon benthic molluscs at depths greater than 1000 m in the Levantine Sea are composed of both adult and juvenile specimens. Moreover, gravid benthic deca- pod crustaceans and fish have been collected repeat- edly from the depths of the Levantine Sea (Galil and Goren, 1994; Goren and Galil, 1997; Fishelson and Galil, 2001). Though much reduced in diversity and richness compared with the deep sea fauna of the western and central basins of the Mediterranean, the Levantine bathy-benthos appears to be composed of autochthonous, self-sustaining populations of oppor- tunistic, eurybathic species that have settled there fol- lowing the last sapropelic event.
Despite the thorough review of Fredj and Laubier (1985) regarding qualitative aspects of the
MEDITERRANEAN DEEP-SEA BIOLOGY 19