26 detonation. Little, if any, correlation exists between the probe values and the actual levels of radioactivity present in sea water 4s determined from samples taken curing the Marsh expedition (Fig. 7). The lack of sensitivity might possibly have been due to faulty "a" and "B" pattertes in the probe assembly although this is not likely. The instrument was sensitive to a J source inserted into the calibration port several times throughout the duration of the trip. Although the instrument, in its present form, cannot be used for detecting low levels of contamination the following modifications might be made to increase its sensitivity and usefulness. 1. Incorporate suitable voltege stabilization circuits in the filament, B+, and high voltage supplies (the ship- line supply fluctuates from 80 v to 135 v). 2. Replace the present crystal and photomultiplier with a 7-inch diameter unit. 3. Rebuild the water tank to minimize vibration effects on the scintillation probe and inorease the volume of water sround the probe, 4k, Inoorporate intermediate ranges into the control box with the following ranges: 5, 30, 50, 300, 500 and 3000 microroentgens. Fish Levels of radioactivity in fish tissues are given in Table 4, They are remarkably uniform in the three skipjack taken near station 23; the liver has the highest level of activity, 186 a /efg a at