In order to study the relationship between substorm occurrence and magnetospheric compression caused by solar wind discontinuities, 125 storm sudden commencements (ssc's) observed during the 4 years 1967-1970 were examined by using ground magnetograms, AE indices, and magnetic field data obtained by Explorer 33, 34, and 35 and OGO 5. Statistical and case studies confirm that geogmagnetic activity and interplanetary magnetic field conditions before the ssc's are important factors for the subsequent triggering of substorms and that the triggering probability tends to increase with increase in sec magnitude. Negative bays at auroral latitudes followed 43% of the ssc and si (sudden impulse) events examined. In approximately 90% of these cases the AE indices showed appreciable activity (AE>100&ggr;) during an interval of 15 min before the ssc. When the previous AE is small (<&ggr;), the probability of triggering is very low (&agr;8%) and is almost independent of ssc magnitude. An examination of interplanetary magnetic field data also revealed that the north-south component was negative or decreasing over a period of 30 min before the passage of the discontinuity when negative bays immediately followed the ssc. These results suggest that preconditions are required for triggering of a substorm by magnetospheric compression. In other word, the triggering probability is high when the magnetosphere is in a metastable or favorable state for substorm occurrence, as is expected during the growth phase. When a substorm is in progress, a further enhancement of the auroral electrojet immediately follows a large-scale magnetospheric compression. A careful case study is made of several apparent exceptions to this conclusion. It is found that an unusually large enhancement of an SqP-like current system sometimes occurs in association with the southward interplanetary magnetic field behind the shock. Magnetic perturbations caused by this polar cap current system can cause extremely large disturbances at auroral zone stations. Since the sense of these perturbations is the same as that of perturbations due to substorm expansions in the morning and afternoon auroral zone, it is possible to confuse the two. Reports of the apparent triggering of substorm expansions without the expected preconditions can usually be attributed to this confusion.