mains very nearly the same as the pulse is propagated to distances. This is partially because the distribution of energy with frequency near the source was different for Shot 3 and Shot 2, The curves also show attenuation with distance at any frequency within the range analyzed. It is apparent that the higher frequencies are attenuated relatively more. It is interesting to compare Figure 13, which shows spectral intensity values as received for Shot 2, and Figure 14, which shows comparable values with the distance factor removed. Under the conditions of the Castle tests, as distance increases, attenuation due to distance becomes relatively less important. Yie d_ Predo t Frequency. There appears to be an approxi- mate relationship between yield and frequency at which peak energy occurs, and there is some theoretical justification for this relationship (Reference 13). Plotted as Figure 16 is a theoretical curve from Reference 13, with points indicating frequencies with maximum energy, as calculated from close-in waveforms, for some of the Upshot~Knothole detonations, Castle Shots 2 and 3, and estimated frequencies for Castle Shots 4, 5, and 6. Narrow-Band Records. Narrowband (approximately 200 cps) record- ings were also made during the Castle series. This technique may be necessary in making measurements under operational conditions, Values of spectral intensity were also calculated for field strength narrowband values received for Shot 2 (Reference 11). Figure 13 has narrow band points indicated by crosses and it will be noted that in most cases, the agreement with the corresponding broad-band spectral curve is quite good. In some cases, the trigger level was somewhat high, which resulted in missing the initial portion of the waveform. Other reasons for discrepancies were variation in bandwidth with signal level and aging of tubes. ' Peripheral Lightning Flashes. Fast—frame moving picture photo- graphy (3,000 or more frames per second) of Ivy Mike show what appears to be lightning flashes between the natural cloud cover and the sea on the periphery of the fireball. This phenomena starts at about 5 msec after the beginning of the nuclear reaction and continues for about 75 msec or more. These visible flashes were also evident on the highspeed photographic film recordings of the Castle series. High-gain receiving equipment, with a bandwidth of about 1 ke to 1 Me, long-wire antenna, and a tape recorder were used at Enivetok in an attempt to detect these flashes. No signals attributable to the discharges were noted. Loy-FrequencyTransmiters. An analysis of field strength mea- surements of low-frequency transmitters has been carried out by NBS. wee ee (References 4,12) At widely-separated locations during Castle, NBS and DRL monitored NSS, NLK, NPM and GBR to obtain field strength measurements. Narrow-band equipment was used; essentially the same as that used for recording the narrow-band field strengths from mclear detona- tions. Measurements were made at the same time at each location, once each hour. Five selected plots of field strength versus time from _. 33

Select target paragraph3