Abstract The Congo deep-sea fan was surveyed recently by a unique set of data that includes as of today more than 30 piston cores and 25,000 km of 2D high resolution seismic acquisition. During the different cruises, the following tools were used simultaneously: an EM12 dual Simrad swath bathymetric and sonar system, a 3.5 kHz mud penetrator, an oceanographic seismic system (6 traces), and a multichannel seismic streamer (96 traces) with G.I. air gun source. The data set covers an area in excess of 200,000 km 2 from the canyon head to the most recent depositional lobes 700 km downslope. In this four year project cosponsored by ELF and IFREMER, the prime objective was to gain comprehension of the construction and the lateral evolution of the recent fan. As a result, we were able to build a chronostratigraphic and architectural model at the fan scale. The Congo Canyon geometry has been enhanced and has confirmed the lack of any tributaries downdip from the continental shelf. Furthermore, no real large scale slumped blocks have observed at the base of the steep submarine canyon slopes. Therefore, widening of the canyon seems to have occurred through progressive sliding and creeping of the bordering Congolese and Angola platform into the canyon valley. The recent Congo fan growth results from the stacking and overlapping of at least 3 main stratigraphic units. Each unit is composed of numerous channel-levee complexes laterally switching path through time by avulsion. Previous interpretation of “distributaries” at the toe of the Congo submarine canyon is invalidated as this surveys shows that they are in fact abandoned channel-levee complexes. It has also been confirmed that in this particular mud-rich submarine fan setting, only one channel is active at a time. The present day “active” fan channel as well as the canyon are eroding into previously deposited sediments. In the levees, individual sub-kilometer scale morphological features have been mapped, and sedimentary and/ or structural processes responsible for such geometries have been reviewed. The present day depositional lobe is more than 700 km away from the Congo River mouth. It occupies an area of around 300 km 2 and shows clustered lobes, 20 km long, prograding down slope. It is now clear that multibeam coverage of the lower part of the canyon, over complex lobe, and channel avulsion zones shows strong similarities with seismic attribute maps of West Africa’s structurally controlled turbiditic channel complexes.

Map of Congo Estuary and Submarine Canyon The compilation of this map (Pl. 2) has been made from the sources acknowledged in the title. Unfortunately it is not possible to print both soundings and contours in the area covered by Carta Hidrografica do Rio Zaire because of the large number of soundings plotted on this excellent Portuguese chart, which is printed on a natural scale of 1:50,000. A section of the submarine canyon in the vicinity of Banana Point, where some of the sharp bends exist, is shown in Figure 10 with all the soundings. This also shows plainly the evidence of the dissected character of the side walls of the submarine canyon. In this case, as well as in the canyons of the Atlantic Continental Slope of the United States, there is undoubtedly more dissection than the available soundings prove. A copy of Carta Hidrografica do Rio Zaire was obtained only a few months before the senior author's death, and as the remainder of the data for the Congo Estuary map had been assembled, and all the known sources explored, it was necessary to add this excellent piece of hydrographic work, contoured by the junior author, to the partially completed sheets. Between the eastern limit of the Portuguese chart and the eastern limit of the map, the submarine configuration is taken from the British Admiralty Chart No. 638. The shallow channels in the upper river undoubtedly change in this region from year to year, so the information in this . . .