1. The widely accepted theory that calcium participates in the coagulation mechanism in the form of Ca(++) and acts as a catalyst is not in accord with several important experimental findings: (a) The anticoagulant action of sodium oxalate is much slower than the precipitation of ionized calcium as the oxalate salt. (b) Sodium citrate begins to depress prothrombin activity at a concentration at which ionized calcium is still present. The inability of tricalcium phosphate to adsorb prothrombin from citrated plasma indicates that citrate forms a complex with prothrombin and it is postulated that prothrombin is thereby inactivated. (c) In plasma which is decalcified, i.e. in which the Ca(++) is markedly reduced, the labile factor of prothrombin rapidly decreases. A concentration of 0.01 M sodium citrate sufficient to inhibit coagulation does not depress Ca(++) enough to cause diminution of the labile factor, whereas when the concentration is increased to 0.02 M the labile factor decreases as rapidly as in oxalated plasma. 2. It is postulated that calcium functions in coagulation not as Ca(++) but as combined with a component which is part of the prothrombin complex that is not adsorbed by tricalcium phosphate. A concentration of sodium citrate just sufficient to inhibit coagulation is not enough to remove calcium from the essential prothrombin component. The primary anticoagulant action of sodium citrate is therefore not decalcification but antiprothrombic. 3. It has been shown that citrated plasma is basically different from oxalated plasma in several important aspects. Unless cognizance is taken of these differences, serious errors and misinterpretations of experimental findings may be made.