Delayed AfterDepolarisations (DAD) are depolarisations that occur after full repolarisation and have been related to sodium and/or calcium overload in the cell. They are triggered by “spontaneous release” of calcium from the sarcoplasmic reticulum (SR) leading to calcium oscillations and via the electrogenic sodium-calcium exchanger (NCX) to changes in membrane potential (1). Due to the variety of experimental findings the trigger of the release is still under debate: SR overload (2) vs. dyadic calcium (3). We investigate the underlying mechanism of DAD induction using mathematical atrial and ventricular cell models of human, dog, rabbit, guinea-pig, rat and mouse (31 models). To compare inducibility of calcium oscillations, each model is reduced to its “calcium subsystem”; trans-membrane potential, intracellular sodium ([Na]i) and potassium concentrations are fixed at their end-diastolic values, all trans-membrane currents but the NCX are set to zero. Intracellular calcium handling is unchanged. We find that in all models showing Calcium Oscillations in the SubSystem (CaOSS) the same parameter changes also produce DADs in the full model, which in the case of the Noble’98 model (4) lead to action potentials (Fig 1). We do not find parameter combinations that initiate DADs but do not show CaOSS. Thus the initiation of DADs in the full model is concurrent with the appearance of CaOSS. Reduced models are tested for CaOSS by increasing [Na]i (increases [Ca]i via NCX) and the sarcoplasmic reticulum calcium ATPase (SERCA) current (reflecting increased sympathetic tone). All 19 reduced models that show CaOSS have i) Markov formulations for RyR, ii) the RyR modulated by either dyadic subspace (Cass) or cytoplasmic (Cacyt) calcium and iii) a SERCA formulation dependent on both Cacyt and CaSR (all but two). To investigate which of these model properties are essential for the production of CaOSS (and thus DADs) we perform in-silico experiments removing the dependence of RyR on Cass/Cacyt, or the dependence of SERCA on CaSR. The change in RyR abolishes CaOSS, whereas SERCA modulation changes just their frequency and amplitude. Modulation of RyR by CaSR is included in some models, but can be removed without preventing CaOSS – indicating that SR overload modulates DADs, but is not the key mechanism. Our results confirm that the main trigger for CaOSS and DADs is an increased level of calcium within the cell, which is also the common feature of the experimental results of calcium-induced DADs. The “spontaneous release” of calcium from the SR occurs by the same mechanism as calcium induced calcium release, i.e., the modulation of the RyR opening by calcium in the dyadic subspace. Future work will investigate the correlation between Markov formulations of RyR and DAD appearance.