removed from Eniwetok befo. 2 the failure was detécted. Three spare accelerometers. which were installed on the roof and rear and side walls of the recording shelter, Station 603, to re- spond to vertical, radial, and tangential cornponents of acceleration, were connected to the information channels assignedto Station 650.04. 1.5 1.5.1 DATA ANALYSIS Accelerations Acceleration-time data were obtained from all component instruments at five stations. No data were recovered from Station 650.05 on Aomon as a result of failure of the recorder tapetransport mechanism.’ Only the initial portion of the recorded data at Station 650.01 on Bogon was useful because incidence of the air shock at the recorder shelter, Station 601, reactivated the cal-step timer with consequent recording of spurious signals. Pertinent sections of the acceleration-time data are presented in Appendix A, Figs. A.1 to A.11. Each of these figures includes, in addition, velocity-time and displacement-time data derived by integration. Ground-transmitted signals and those produced by local incidence of air shock are separated by appreciable time intervals for all stations beyond Engebi and are plotted as separate parameter-time graphs. The data from local incidence of the air shock at Station 650.02 are reproduced on an expanded timescale for clarity. Acceleration-time data from Station 650.01 are compared in Fig. A.1 with data (dashed curves) from a recorder channel which monitored the carrier power supply to the gauges. The carrier power was apparently stable until 1.436 sec after zero time, when reactivation of the cal-step timer is thought to have occurred. Signals recorded on the accelerometertraces after this time follow details of the carrier monitor record. Data from Station 650.01 are good until 1.436 sec and false thereafter. A vertical dashed line on each parameter-time curve for this station indicates the end of the valid portion of the curve. The ground-transmitted acceleration presented for Station 650.06 on Parry in Fig. A.11 does not representthe initial arrival which occurred at about 7.35 sec, but a later, stronger signal, probably the first pulse reflected from basement rock. Weaker reflected pulses arrived later, but they are not included because the signal-to-noise ratio was low as a result of poor set-range estimation. Arrival times, peak accelerations, and acceleration frequencies comprise the information available directly from the recorded data. Air-overpressure data from Project 6.1 stations‘ are also pertinent to study of ground motion since most of the acceleration data show a readily distinguishable air-shock induced signal. Data from Station 650.01 in which the motion is derived indistinguishably from both sources are the only exceptions observed. Data from Bokon,Station 603, are compared with overpressure data from Aitsu because no air-pressure records were obtained from Bokon. Arrival times for both air overpressure and ground acceleration are comparedin Table 1.1. Zero time on the records from Parry was derived from the cal-step signal at —15 sec (Ref. 4) because no zero time signal was recordedat that site. Absolute times on Parry records are considered good only to about 0.1 sec. Interval timing on these records is, however, as good as that for the other stations since the timing-channel frequency was recorded satisfactorily. Overpressure arrival was determined to have been 83.75 sec after zero time. Ground-acceleration data were adjusted to assumed simultaneous arrival of overpressure and air-shock induced ground motion. Arrival times plotted as a function of ground range show the anticipated branched curve (Fig. 1.5). Arrival of the air-shock induced accelerations corresponds closely with the curve for arrival of air overpressure. This curve indicates an initial propagation velocity greater than 6000 ft/sec, decreasing beyond 20,000 ft to an apparent velocity of about 1230 ft/sec. Ground-transmitted accelerations are propagated with a velocity of nearly 18,000 ft/sec beyond Station 650.01 and with a velocity of 6300 ft/seo out to that station. The latter pattern is consistent with seismic refraction under the conditions shown by deep drilling at Elugelab and Parry—a deep interface between massive basalt and overlying water-filled sand containing 19