We estimate the magnesium stable isotopic composition (δ26Mg) of the major compartments involved in the biomineralisation process of euryhaline bivalve, the manila clam Ruditapes philippinarum. Our aim is to identify the fractionation processes associated with Mg uptake and its cycling/transport in the bivalve organism, in order to better assess the controlling factors of the Mg isotopic records in bivalve shells. δ26Mg were determined in seawater, in hemolymph, extrapallial fluid (EPF), soft tissues and aragonitic shell of adult clams collected along the Auray River estuary (Gulf of Morbihan, France) at two sites showing contrasted salinity regimes. The large overall δ26Mg variations (4.16‰) demonstrate that significant mass-dependent Mg isotopic fractionations occur during Mg transfer from seawater to the aragonitic shell. Soft tissues span a range of fractionation factors relative to seawater (Δ26Mgsoft tissue-seawater) of 0.42 ± 0.12‰ to 0.76 ± 0.12‰, and show evidence for biological isotopic fractionation of Mg. Hemolymph and EPF are on average isotopically close to seawater (Δ26Mghemolymph-seawater = -0.20 ± 0.27‰; 2 sd; n = 5 and Δ26MgEPF-seawater = -0.23 ± 0.25‰; 2 sd; n = 5) indicating (1) a predominant seawater origin for Mg in the intercellular medium and (2) a relatively passive transfer route through the bivalve organism into the calcifying fluid. The lightest isotopic composition is found in shell, with δ26Mg ranging from -1.89 ± 0.07‰ to -4.22 ± 0.06‰. This range is the largest in the dataset and is proposed to result from a combination of abiotic and biologically-driven fractionation processes. Abiotic control includes fractionation during precipitation of aragonite and accounts for Δ26Mgaragonite-seawater ≈ 1000 ln αaragonite-seawater = -1.13 ± 0.28‰ at 20 °C based on literature data. Deviations from inorganic precipitate (expressed as Δ26MgPhysiol) appear particularly variable in the clam shell, ranging from 0.03‰ to -2.20‰, which indicates that bivalve shell formation can proceed either under fractionation similar to inorganically-precipitated aragonite or under variable physiological influences. These physiological isotopic effects may be consistent with a regulation of dissolved Mg content in hemolymph and/or EPF due to Mg incorporation into soft tissue and/or Mg fixation by organic macromolecules. Using closed- and open-system models we estimate that Δ26MgPhysiol can be satisfactorily resolved with a remaining Mg fraction in hemolymph and/or EPF of 74% down to 2%. However, this feature is not reflected in our hemolymph and EPF data and may indicate that regulation processes and isotopic fractionation may take place in self-contained spaces located close to calcification sites. The potential role of the shell organic matrix, which may host non-lattice-bound Mg in the shell, is also discussed but remains difficult to assess with our data. Regarding the large physiological effects, the δ26Mg record in the Manila clam shell offers limited potential as a proxy of temperature or seawater Mg isotopic composition. In contrast, the sensitivity of its δ26Mg to the salinity regime may offer an interesting tool to track changes in clam biological activity in estuarine environments.