At a 650 m deep site in the axis of Capbreton Canyon an 18-cm-thick turbidite was deposited in December 1999. During subsequent campaigns, an almost monospecific fauna of the benthic foraminifer Technitella melo, considered as a pioneer species, was found in May 2000. In 2001 this fauna had disappeared and was replaced by an exceptionally rich fauna strongly dominated by the opportunistic species Bolivina subaenariensis. We present sedimentological, radionuclide and foraminiferal data of new cores, sampled in 2005 and 2011, taken with the aim to study the further evolution of the benthic ecosystem. Cores sampled in 2005 show that in the canyon axis a new, ca. 5 cm thick, turbidite has been deposited. The live benthic foraminiferal faunas were much poorer than in 2001, but still had a high dominance and low diversity, although less extreme than in 2001. We conclude that in the canyon axis, benthic foraminiferal faunas remain in an early stage of ecosystem colonization. It appears that the very thick 1999 turbidite marks an exceptional event. The uncommonly rich faunas observed in 2001 could be a response to the concentration of organic-rich material in the fine-grained top of this deposit. In 2011, cores were sampled at a slightly different site, on the lower canyon flank. The sedimentary sequence here is marked by the absence of coarse turbidite layers, although some levels show slightly increased grain size, and lower 210Pbxs activities, indicative of an admixture with advected older sediments. Live foraminiferal faunas are much more equilibrated, as shown by their higher diversity, lower dominance, and deeper penetration into the sediment. All these characteristics are indicative of a much more stable ecosystem. Dead faunas are present throughout the core, indicating that the levels with slightly elevated grain size are not typical turbidites resulting from hyperpycnal currents (which are characterized by levels barren of foraminifera) but denote other, more long-term sedimentary processes leading to the advection of older material, such as bottom nepheloid layers, or repetitive fine-grained turbidite deposits due to small-scale slumping. The comparison of live and dead fauna shows that at both sites, the foraminiferal turnover rates are fairly low. At the lower canyon flank site sampled in 2011, the foraminiferal faunas are renewed every 1.5–2 years. Such a fairly long foraminiferal lifespan corresponds to earlier estimates, but is surprisingly high for the opportunistic taxa that dominate the faunal assemblages in these unstable and food-enriched submarine canyon settings.