Figure 2. Model output.

Model output across the experimental range of CO₂, with empirical data superimposed. The black horizontal dashed line shows the expected composition of abiogenic calcite. Universal constants used to generate model output are constrained by the points outlined in bold, here re-sampled 1,000 times from distributions defined by the associated analytical uncertainties. The data from this study also include additional G. oceanica data (calcite only; green filled circles) and organic carbon isotopic compositions for C. pelagicus from samples from a previous study⁹. See Supplementary Data 1 for values. Additional literature data are from dilute batch experiments manipulated by varying DIC⁹ (calcite carbon isotopic compositions only: C. leptoporus—dark blue upwards triangles; C. pelagicus—blue squares; E. huxleyi—yellow downwards triangles; and P. placolithoides—orange diamonds), and by varying pH (ref. ⁷⁵; E. huxleyi—organic carbon isotopic compositions only: small yellow filled circles). Further data are superimposed in Supplementary Fig. 3. Data points are all the average of two replicates, with the ends of bars representing the value of each replicate. Solid lines and shaded envelopes represent respectively the mean and standard deviation of isotopic compositions predicted by the model for the 1,000 repeated calibrations. Values for compartmental pH are taken from published measurements³³. Different species are represented by a representative set of parameters, which for ease of illustration are held constant with varying [CO₂]. Representative values are taken from across a range of sources in the literature and from our own unpublished data (cell radius, division rate, PIC:POC are as follows: E. huxleyi – 2.3 μm, 1 day⁻¹, 0.5; G. oceanica - 3.5 μm, 1 day⁻¹, 1.1; C. pelagicus – 8 μm, 0.9 day⁻¹, 1.2; C. leptoporus – 6 μm, 0.65 day⁻¹, 2.2; and P. placolithoides – 7 μm, 0.8 day⁻¹, 0.3). Strong co-variance between cell size or growth rate with CO₂ would not be well represented in this projection. Note: for the pH manipulated model output uptake is enhanced at low CO₂, manifest as a more rapid isotopic enrichment at low CO₂ (steep yellow model curve). All species curves are a similar shape controlled by a number of trade-offs, and with the point of inflexion and maximum δ¹³C_calcite determined by varying combinations of cell size, PIC:POC and growth rate. There are some regions of parameter space that are theoretically possible, but may not be observed in reality due to poor growing conditions. Regions of model space not populated by data should be treated as hypothetical.