Aim: The loss of dispersal on islands hypothesis (LDIH) posits that wind-dispersed plants should exhibit reduced dispersal potential, particularly if island populations are old. In this study, we tested this hypothesis using a detailed phylogeographical framework across different geographical scales. Location: Mainland and island areas of the Atlantic and Mediterranean regions, including Macaronesia (Canary Islands and Cape Verde) and Mediterranean islands in the strait of Sicily. Methods: Forty-five populations of Periploca laevigata, a wind-dispersed shrub, were sampled. Plastid and nuclear microsatellite data were used to reconstruct spatiotemporal patterns of island colonization, and estimates of seed terminal velocity used as a surrogate for dispersal ability under both field and common garden conditions.Results: Our findings did not provide evidence of loss of dispersability in any island lineage. In all of the regions considered, dispersal ability was similar on island and mainland populations, or higher on islands. Contrary to LDIH expectations, lineages inferred as the oldest (western Canaries and Cape Verde) converged towards the most dispersive seed phenotype. This pattern was supported by data obtained under common garden conditions. Within the western Canarian lineage, successful dispersal was shown to be very rare among islands and extensive within islands, but dispersability did not vary significantly from older to more recent sublineages. Considering all the study islands, we found a strong, positive correlation between dispersal ability and estimates of within-island habitat availability. Main conclusions:This study suggests that dispersal ability can be favoured on islands, possibly because traits enhancing wind dispersal are positively selected when habitat availability is high. Our results challenge broad generalizations of the LDIH, but we discuss how overlooking species 0 phylogeographical history may give rise to misleading conclusions. K E Y W O R D S anemochory, dispersal ability, island colonization, Macaronesia, parallel evolution, seed dispersal

Abstract. We study the problem of sinking particles in a realistic oceanic flow, with major energetic structures in the mesoscale, focussing on the range of particle sizes and densities appropriate for marine biogenic particles. Our aim is to evaluate the relevance of theoretical results of finite size particle dynamics in their applications in the oceanographic context. By using a simplified equation of motion of small particles in a mesoscale simulation of the oceanic velocity field, we estimate the influence of physical processes such as the Coriolis force and the inertia of the particles, and we conclude that they represent negligible corrections to the most important terms, which are passive motion with the velocity of the flow, and a constant added vertical velocity due to gravity. Even if within this approximation three-dimensional clustering of particles can not occur, two-dimensional cuts or projections of the evolving three-dimensional density can display inhomogeneities similar to the ones observed in sinking ocean particles.

In an incompressible flow, fluid density remains invariant along fluid element trajectories. This implies that the spatial distribution of non-interacting noninertial particles in such flows cannot develop density inhomogeneities beyond those that are already introduced in the initial condition. However, in certain practical situations, density is measured or accumulated on (hyper-) surfaces of dimensionality lower than the full dimensionality of the flow in which the particles move. An example is the observation of particle distributions sedimented on the floor of the ocean. In such cases, even if the initial distribution of noninertial particles is uniform within a finite support in an incompressible flow, advection in the flow will give rise to inhomogeneities in the observed density. In this paper we analytically derive, in the framework of an initially homogeneous particle sheet sedimenting towards a bottom surface, the relationship between the geometry of the flow and the emerging distribution. From a physical point of view, we identify the two processes that generate inhomogeneities to be the stretching within the sheet, and the projection of the deformed sheet onto the target surface. We point out that an extreme form of inhomogeneity, caustics, can develop for sheets. We exemplify our geometrical results with simulations of particle advection in a simple kinematic flow, study the dependence on various parameters involved, and illustrate that the basic mechanisms work similarly if the initial (homogeneous) distribution occupies a more general region of finite extension rather than a sheet. a) gabor@ifisc.uib-csic.es 1 Sedimentation of small particles in complex flows is an outstanding problem in science and technology. In particular, the sinking of biogenic particles from the marine surface to the bottom is a fundamental process of the biological carbon pump, playing a key role in the global carbon cycle. A complete understanding of this problem is still lacking. It has been recently shown that despite fluid incompressibility, sedimenting particles moving as noninertial particles in the ocean on large scales show density inhomogeneities when accumulated on some bottom surface. Here, we analytically derive the relation between the geometry of the flow and the emerging distribution for an initially homogeneous sheet of particles. From a physical point of view, we identify the two processes that generate inhomogeneities to be the stretching within the sheet, and the projection of the deformed sheet onto the target surface. We point out conditions under which an extreme form of inhomogeneity, caustics, can develop.

Marine resources stewardships are progressively becoming more receptive to an effective incorporation of both ecosystem and environmental complexities into the analytical frameworks of fisheries assessment. Understanding and predicting marine fish production for spatially and demographically complex populations in changing environmental conditions is however still a difficult task. Indeed, fisheries assessment is mostly based on deterministic models that lack realistic parameterizations of the intricate biological and physical processes shaping recruitment, a cornerstone in population dynamics. We use here a large metapopulation of a harvested fish, the European hake (Merluccius merluccius), managed across transnational boundaries in the northwestern Mediterranean, to model fish recruitment dynamics in terms of physics‐dependent drivers related to dispersal and survival. The connectivity among nearby subpopulations is evaluated by simulating multi‐annual Lagrangian indices of larval retention, imports, and self‐recruitment. Along with a proxy of the regional hydroclimate influencing early life stages survival, we then statistically determine the relative contribution of dispersal and hydroclimate for recruitment across contiguous management units. We show that inter‐annual variability of recruitment is well reproduced by hydroclimatic influences and synthetic connectivity estimates. Self‐recruitment (i.e., the ratio of retained locally produced larvae to the total number of incoming larvae) is the most powerful metric as it integrates the roles of retained local recruits and immigrants from surrounding subpopulations and is able to capture circulation patterns affecting recruitment at the scale of management units. We also reveal that the climatic impact on recruitment is spatially structured at regional scale due to contrasting biophysical processes not related to dispersal. Self‐recruitment calculated for each management unit explains between 19% and 32.9% of the variance of recruitment variability, that is much larger than the one explained by spawning stock biomass alone, supporting an increase of consideration of connectivity processes into stocks assessment. By acknowledging the structural and ecological complexity of marine populations, this study provides the scientific basis to link spatial management and temporal assessment within large marine metapopulations. Our results suggest that fisheries management could be improved by combining information of physical oceanography (from observing systems and operational models), opening new opportunities such as the development of short‐term projections and dynamic spatial management.