Cold-leg small-break loss-of-coolant accident experiments were conducted for break areas ranging from 0.5 to 10% of the scaled cold-leg flow area using the Large-Scale Test Facility (LSTF). The LSTF is a volumetrically scaled simulator of a Westinghouse-type pressurized water reactor. For all the experiments, the core collapsed liquid level was temporarily depressed when liquid in the primary loop U-bend (crossover leg) was being cleared by steam. For scaled break areas <2.5%, the minimum core liquid level was equal to the lowest elevation of the crossover leg. For break areas >5%, the minimum core level was even lower because differential pressures created by the residual liquid holdup in the steam generator (SG) upflow side affected the core liquid level adversely. This influence of SG liquid holdup on the minimum core liquid level was larger for larger break sizes within the range of these experiments; thus, a more severe core level depression was seen for larger break sizes. Also, for the same break size, the core level depression was more severe when higher core power values were used for the simulation of the postscram core power decay. The RELAP5/MOD2 code reasonably well predicted the major phenomena observed in the experiments; however, several shortcomings were found in interfacial drag calculation for the SG U-tube inlet and the hot-leg outlet to the SG inlet plenum and core.