Marine bony fish are now known as primary producers of calcium carbonate. Furthermore, within the shallow sub-tropical platform settings of the Bahamas, this production process has been shown to occur at rates relevant to carbonate sediment production budgets. Fish excrete these carbonates as loosely aggregated pellets which, post-excretion, exhibit a range of distinctive crystal morphologies and have mineralogies ranging from low (0-4mol% MgCO ) to high (4-40mol% MgCO ) Mg-calcites, aragonite and amorphous carbonate phases. Here we provide the first quantitative assessment of the size fractions of the carbonates produced by a range of tropical fish species, and document the extent of post-excretion carbonate pellet break down under a range of physical agitation conditions. Specifically, we document the morphologies and size fractions of: i) intact pellets at the point of excretion; ii) intact pellets after agitation in seawater; and iii) the particles released from pellets post-disaggregation. Results indicate that fish-derived pellets initially fall within the very fine to very coarse sand fractions. Exposure to conditions of moderate seawater agitation for 30days results in significant pellet diminution; 66% of initial pellet mass being released as individual particles, whilst 34% is retained as partially intact pellets that are smaller (fine sand-grade) and more rounded than initial pellets. In contrast, pellets exposed to very gently agitated conditions for up to 200days show little change. Where pellet disaggregation does occur, particles are commonly released as individual clay- and silt-grade crystals. However, some morphotypes (e.g. polycrystalline spheres) can be intergrown and are released as strongly cohesive particle clusters falling within the coarse silt to fine sand fractions. Only very vigorous agitation may disaggregate such particles, resulting in the release of their component clay-grade crystals. We conclude that fish-derived carbonates may thus contribute not only to the mud-fraction of marine carbonates, but also to the fine sand fraction as intergrown particles, and to the fine to coarse sand fractions as intact and partially intact pellets. These experimental data indicate that hydrodynamic regimes local to sites of excretion will influence the generation of carbonates with different size fraction ranges. Rapid pellet disaggregation is more likely in high energy settings, hypothesised to result in redistribution of liberated mud-grade particles to lower energy platform-top settings and/or off-platform. In contrast, pellets excreted in lower energy settings are more likely to be preserved intact, and are thus proposed as a previously unrecognised source of pelletal and peloidal carbonate sediments. •Break-up of carbonate pellets produced by marine bony fish is investigated. •Pellets can remain intact and may contribute to sedimentary pellets and peloids. •Rapid pellet break-up releases component particles in agitated settings. •Particle size ranges from clay to fine sand grade: different depositional fates. •Relevant to surface sediments in shallow sub-tropical marine provinces. 3 3

Carbonate mud is a major constituent of recent marine carbonate sediments and of ancient limestones, which contain unique records of changes in ocean chemistry and climate shifts in the geological past. However, the origin of carbonate mud is controversial and often problematic to resolve. Here we show that tropical marine fish produce and excrete various forms of precipitated (nonskeletal) calcium carbonate from their guts ("low" and "high" Mg-calcite and aragonite), but that very fine-grained (mostly < 2 μm) high Mg-calcite crystallites (i.e. > 4 mole % MgCO(3)) are their dominant excretory product. Crystallites from fish are morphologically diverse and species-specific, but all are unique relative to previously known biogenic and abiotic sources of carbonate within open marine systems. Using site specific fish biomass and carbonate excretion rate data we estimate that fish produce ∼6.1 × 10(6) kg CaCO(3)/year across the Bahamian archipelago, all as mud-grade (the < 63 μm fraction) carbonate and thus as a potential sediment constituent. Estimated contributions from fish to total carbonate mud production average ∼14% overall, and exceed 70% in specific habitats. Critically, we also document the widespread presence of these distinctive fish-derived carbonates in the finest sediment fractions from all habitat types in the Bahamas, demonstrating that these carbonates have direct relevance to contemporary carbonate sediment budgets. Fish thus represent a hitherto unrecognized but significant source of fine-grained carbonate sediment, the discovery of which has direct application to the conceptual ideas of how marine carbonate factories function both today and in the past.