Aims. A thorough study of the effects of mass loss on internal and surface abundances of A and F stars is carried out in order to constrain mass loss rates for these stars, as well as further elucidate some of the processes which compete with atomic diffusion. Methods. Self-consistent stellar evolution models of 1.3 to 2.5 M stars including atomic diffusion and radiative accelerations for all species within the OPAL opacity database were computed with mass loss and compared to observations as well as previous calculations with turbulent mixing. Results. Models with unseparated mass loss rates between 5 × 10 −14 and 10 −13 M yr −1 reproduce observations for many cluster AmFm stars as well as Sirius A and o Leonis. These models also explain cool Fm stars, but not the Hyades lithium gap. Like turbulent mixing, these mass loss rates reduce surface abundance anomalies; however, their effects are very different with respect to internal abundances. For most of the main-sequence lifetime of an A or F star, surface abundances in the presence of such mass loss depend on separation which takes place between log ΔM/M * = −6 and −5. Conclusions. The current observational constraints do not allow us to conclude that mass loss is to be preferred over turbulent mixing (induced by rotation or otherwise) in order to explain the AmFm phenomenon. Internal concentration variations which could be detectable through asteroseismic tests should provide further information. If atomic diffusion coupled with mass loss are to explain the Hyades Li gap, the wind would need to be separated.

Aims. Abundances of 15 chemical elements have been derived for 28 F and 16 A stars members of the Hyades open cluster in orderto set constraints on self-consistent evolutionary models that include radiative and turbulent diffusion. Methods. A spectral synthesis, iterative procedure was applied to derive the abundances from selected high-quality lines in highresolution, high-signal-to-noise spectra obtained with SOPHIE and AURELIE at the Observatoire de Haute Provence. Conclusions. While part of the discrepancies between derived and predicted abundances could come from non-LTE effects, including competing processes such as rotational mixing and/or mass loss seems necessary in order to improve the agreement between the observed and predicted abundance patterns.

Context. Abundance anomalies observed in a fraction of A and B stars of both Pop I and II are apparently related to internal particle transport. Aims. Using available constraints from Sirius A, we wish to determine how well evolutionary models including atomic diffusion can explain observed abundance anomalies when either turbulence or mass loss is used as the main competitor to atomic diffusion. Methods. Complete stellar evolution models, including the effects of atomic diffusion and radiative accelerations, have been computed from the zero age main-sequence of 2.1 M stars for metallicities of Z 0 = 0.01 ± 0.001, and shown to closely reproduce the observed parameters of Sirius A. Surface abundances were predicted for three values of the mass loss rate and four values of the mixed surface zone. Results. A mixed mass of ∼10 −6 M or a mass loss rate of 10 −13 M /yr were determined through comparison with observations. There are 17 abundances that were determined observationally and that are included in our calculations. Up to 15 of them can be predicted to within 2σ; three of the four determined upper limits are compatible. Conclusions. While the abundance anomalies can be reproduced slightly better using turbulence as the process competing with atomic diffusion, mass loss probably ought to be preferred since the mass loss rate required to fit abundance anomalies is compatible with the observationally determined rate. A mass loss rate within a factor of 2 of 10 −13 M /yr is preferred. This restricts the range of the directly observed mass loss rate.