oped. Thus, the problem of institut- ing MPAs in waters outside national jurisdictions involves complex legal issues, as well as scientific ones. Within the North Atlantic, for example, several special areas should be considered as potential MPAs: 8 Great Meteor Seamount in the central North Atlantic (around 30°N, 28"W), where various eco- logical investigations have been conducted (see the following article). * Hydrothermal vent areas on the Mid-Atlantic Ridge between 14"N and 38"N, sites of many recent vent studies. The extensively studied areas of: 1. The Porcupine Abyssal Plain station (around 48"N, 16"W), where British and European (MAST-funded) programmes have concentrated much of their activity during recent years, and 2. The BIOTRANS/BIO-C-FLUX study site (around 47"N, 2OoW), where ecological long-term investi- gations have been located since 1984, and which became the central JGOFS station in 1989. These latter two sites should be joined with the Porcupine Seabight area (a British long-term study site in Irish national waters) to establish (as already proposed), the European Deep-sea Transect (EURODEST), as a basis for long-term studies on ecological variability and climate change. Studies of deep-sea com- munities require substantial invest- ment of technical, financial and personnel resources; it is vital that such long-term datasets are not rendered useless by uncontrolled use of the sea floor close to such special scientific study sites. The recent group meeting proposed a multidisciplinary workshop. However, funds need to be raised for the support of participants before progress can be expected. Hjalmar Thiel Hamburg Tony Koslow Hobart and Monterey ~nteract~~ns around ocean islands Franz Uiblein and Fernando Bordes Oceanic islands of volcanic origin - e.g. the Canaries in the eastern central Atlantic -typically have a narrow shelf and a steep slope down to the surrounding sea-floor (cf. Figure 1 ). One consequence of this particular topography is that coastal areas are space-limited and oceanic conditions can be encoun- tered only a few miles offshore. Such a 'boundary' situation is of particular ecological interest, as there are manifold trophic interac- tions between various ecological groups of neritic, oceanic, benthic and benthopelagic organisms, all in close proximity to one another. layers at 400-700 m during both day and night, and also at less than 150m depth at night (cf. Figure 1). During their nocturnal residence in the epipelagic realm, vertical migrators may be carried towards the shelf by upwelling events or horizontal drift, and then become 'trapped' in shallow waters before they have a chance to migrate downwards. Such trapping effects can be observed above deeper sea- beds (300-400m) and in non- migratory midwater fishes (e.g. Argyropelecus aculeatus, Figure 2(b) overleaf) too, as revealed by recent collections on the la tea; of Atlantic suggest intensive trophic interactions between mesopelagic animals and the bottom fauna in such environments. For instance, several so-called 'pseudoceanic' Recent studies of fish community structure and fisheries-related problems in the eastern central species (e.g. ~ i a ~ h u s adenomus, Figure 2(a) overleaf) are known to impinge on steep slopes and feed there on benthopelagic slope- dwelling prey. the Great Meteor ~eamou;lt (see map opposite) during cruise M42/3 with the German RV Meteor. In several cruises undertaken around the Canary Islands, financed bv the local government (Gobierno The studies around the Canaries further showed that many larvae and juveniles of slope-dwelling Figure 1 Typical night-time echo- sounder record off the shelf of Fuerteventura in the Canaries, showing both shallow and deep scattering layers. The shallow scattering layer continues towards the shelf as a result of spatial interaction between migratory meso- pelagic fauna and neritic shelf fauna; the deep scattering layer (at - 400-650 m) Ocean Challenge, Vol. 9, No.2 15 dk canarias):rnesopelagic fishes have been found to form a major component of dense scattering impinges on tLe sleep slope. (The maximum depth covered by the transect is about 850m; the length of the section is about 5 km.) (1999)