The instrumental group delay dθ/dω is considered here. First, these delays were calculated for three different recording systems that were used in a precise travel-time monitoring experiment where the delays varied between 10 and 40 msec for the high frequencies of the seismograms involved. A technique is demonstrated by which these delays may be readily accounted for and by which instrumental malfunctions can be readily detected.
Second, two of these systems are also currently used for the recording of short-period teleseisms; at the 1-sec period, the group delays are from 0.3 to 0.4 sec, which is significant and must be accounted for. This is particularly important when these systems are used in connection with data from other systems that have different delays, such as the World-Wide Seismograph Station Network and Canadian Seismograph Network stations. Neglecting these delays will create serious problems in seismological tomography and earthquake catalogs.
Third, for long-period phases recorded by the SRO-type instruments, the delays for the 10- to 20-sec periods are 6 to 12 sec; again, these are significant and must be accounted for.