would decrease as device yield decreased. He also indicated fractionation would increase with depth, that is, air bursts would be less fractionated than surface bursts which would be less fractionated than sub-surface bursts. From Freiling's studies it could be cautiously expected that the high yield surface burst creating the BRAVO fallout caused a moderate to high degree of fractionation which occurred moderately to extensively throughout the debris. For the coral surface burst, Freiling observed that the ratio of Zr-95 to Sr-89 activity could be chosen as a representative measure of the overall degree of fractionation between refractory and volatile elements. This ratio was observed twice and had a value of 5 for a deep water surface burst of megaton range and a value of 100 for a coral surface burst (Fr61). The unfractionated value for this ratio on day 26 post detonation for thermonuclear neutron fission of U-238 was calculated to be 1.6 from data given by Crocker (Cr65). From the average of Yamatera and Tsuzuki data, the calculated ratio of Zr-95 activity to Sr-89 activity measured on day 26 was 4.8. This measured value for the degree of fractionation was characteristic of a deep water surface burst of the megaton range, moderately but not highly fractionated. This moder- ate fractionation probably occurred to extensively throughout the fallout because of the large yield and surface location of the device (Fr61). The effect of fractionation on decay rate is very complex and simple observation of overall radioactive decay does not yield significant information. Even 0, the decay rate from widely distributed samples obtained out to 300 miles away from the BRAVO detonation site were similar as were the decay rates from activity on different size fallout granules collected at the same site (0c68). These facts alone may not be used to indicate the same fractionation was common to all granule sizes. In fact, small granules traveled