evidenced by its appearance in 1960, 18 months after the cessation of large scale nuclear tests. Such maxima, in the absence of recent nuclear tests, are attributed to the disruption of the northern hemisphere tropopause (in the case in point) in middle latitudes during the spring months, resulting in an influx of high altitude debris, much of it of polar origin, into the lower atmosphere. In addition, the mass transfer of stratospheric debris into the lower atmosphere occurs as a result of the rising tropopause during spring and early summer in middle and polar latitudes. An attempt to study the magnitude of this phenomenon has been made using naturally oc- curing Be! in surface air. 1961 is shown in Figure 10. The concentration of Be! in surface air for the period August 1959 through January The level of Be! goes through a maximum during the spring months, and indicates a higher value during the summer months than is seen during fall and winter. also shown for comparison. 137. The concentration of Cs 3 is Since Be! is being produced at a constant rate in the atmosphere (dependent upon the latitude and altitude) the air concentration of this nuclide should be a measure of production rate. This is not strictly true at low altitude, however, because removal by precipitation as well as by dry deposition prevents the establishment of true equilibrium between production and radioactive decay. relate the observed concentration of Be! to precipitation. Hence it is necessary to A method of doing this is shown in Figure 11 where the reciprocal of Be" in atoms per liter is plotted as a function of the fraction of the annual rainfall occurring in the month previous to sampling.° Using the expression Q = AN, where Q is the rate of Be’ production in atoms/liter/day, A is the rate of Be’ removal per day and is the sum of the radioactive decay constant and the precipitation removal factor, and N is the concentration of Be! in atoms/liter. Writing this expression as 1/N = A/Q, at zero precipitation it becomes simply 1/N = A/Q, and thus the R = 0 intercept in Figure 11 gives a measure of Q. Note that the data points in Figure 11 appear to group into two distinct sets, implying two distinct production rates, the higher one occurring during the summer months. The increase in production rate in the summer months is in accord with that expected using Lal's figures for Be! production for 55°N geomagnetic latitude assuming that the winter tropopause height is 35, 000 feet and rises to 45-50, 000 feet during the summer. Concurrently it is to be supposed that a portion of the cs!37) as well as other fission activity, seen in surface air following the spring maximum, has been transferred into the lower atmosphere as a result of the rising tropopause. Another feature of fission product activity in surface air, in addition to the spring maximum, is the occurrence of an annual minimum during the fall and winter months as shown in Figure 10. It is presumed that activity seen during this period is a measure of the leakage of debris through the tropopause and represents a steady state condition as opposed to the period of tropopause disruption and mass transfer. If this be so, then comparison of the concentration of csi?! in surface air between successive minima should provide a measure of stratospheric residence time. The mean stratospheric residence time found by comparing the cst?" levels in the fall 1958 with those in the fall 1959 is 14 months; a mean residence time of 23 months is found between fall 1959 and fall 1960. It is interesting to note that roughly one-half of the annual fallout deposition (as derived from the product of mean monthly air concentration and monthly precipitation) comes about as a consequence of the spring break and the rising tropopause. Hence were it not for these latter two mechanisms, the mean stratospheric residence time would be twice as long. The lengthening of the residence time observed between 1958-59 and 1959-60 is in large part due to the almost total removal of Soviet October debris from the stratosphere by the end of 1959 as indicated in Figure 9. Furthermore, the mean residence time of wi 8! labeled Hardtack debris increased from 8 to 10 months in 1958-59 to in excess of 15 months in 1959-60. Such an in- crease in residence time is to be expected as the lower altitude, and thus more readily removable, debris 126