A total of 2128 calcitic and phosphatic shells, mainly brachiopods with some conodonts and belemnites, were measured for their δ¹⁸O, δ¹³C and ⁸⁷Sr/⁸⁶Sr values. The dataset covers the Cambrian to Cretaceous time interval. Where possible, these samples were collected at high temporal resolution, up to 0.7 Ma (one biozone), from the stratotype sections of all continents but Antarctica and from many sedimentary basins. Paleogeographically, the samples are mostly from paleotropical domains. The scanning electron microscopy (SEM), petrography, cathodoluminescence and trace element results of the studied calcitic shells and the conodont alteration index (CAI) data of the phosphatic shells are consistent with an excellent preservation of the ultrastructure of the analyzed material. These datasets are complemented by extensive literature compilations of Phanerozoic low-Mg calcitic, aragonitic and phosphatic isotope data for analogous skeletons. The oxygen isotope signal exhibits a long-term increase of δ¹⁸O from a mean value of about - 8‰ (PDB) in the Cambrian to a present mean value of about 0‰ (PDB). Superimposed on the general trend are shorter-term oscillations with their apexes coincident with cold episodes and glaciations. The carbon isotope signal shows a similar climb during the Paleozoic, an inflexion in the Permian, followed by an abrupt drop and subsequent fluctuations around the modern value. The ⁸⁷Sr/⁸⁶Sr ratios differ from the earlier published curves in their greater detail and in less dispersion of the data. The means of the observed isotope signals for ⁸⁷Sr/⁸⁶Sr, δ¹⁸O, δ¹³C and the less complete δ³⁴S (sulfate) are strongly interrelated at any geologically reasonable (1 to 40 Ma) time resolution. All correlations are valid at the 95% level of confidence, with the most valid at the 99% level. Factor analysis indicates that the ⁸⁷Sr/⁸⁶Sr, δ¹⁸O, δ¹³C and δ³⁴S isotope systems are driven by three factors. The first factor links oxygen and strontium isotopic evolution, the second ⁸⁷Sr/⁸⁶Sr and δ³⁴S, and the third one the δ¹³C and δ³⁴S. These three factors explain up to 79% of the total variance. We tentatively identify the first two factors as tectonic, and the third one as a (biologically mediated) redox linkage of the sulfur and carbon cycles. On geological timescales (≥ 1 Ma), we are therefore dealing with a unified exogenic (litho-, hydro-, atmo-, biosphere) system driven by tectonics via its control of (bio)geochemical cycles.