E. C. WINKLER 



The rainwater was collected at weekly intervals, the filter paper from the 

 bottom of the beaker was removed and the water was filtered. The two 

 pieces of filter paper were then ashed and the ash was weighed and counted 

 for 20 min. This was the 'fall-out' material. 



One litre of the filtered rainwater was evaporated, the residue rubbed off 

 the evaporating dish on to a piece of filter paper which was then ashed and 

 counted for 50 min. This was the 'rained out' material. 



The gummed paper was changed daily except over weekends. A week's 

 samples were then ashed together, weighed and counted for 20 min. This 

 was compared for only 5 weeks. 



All samples were recounted 3 days after collection, as, according to 

 Tajima and Doke^ the natural radio-activity due to radon and thorium 

 would have become negligible by this time. 



RESULTS 



The results have been graphed as net counts per min for each week of 

 collection, which gives a relative, if not an absolute, picture with the counting 

 system used. The background count was approximately 8 counts per min. 



In Figure 2 are plotted the activities of cheesecloth, 'fall-out' and rainwater 

 samples. 



Peaks of unusual activity were found as follows : 



(i) For airborne particles, 10 counts per min was the activity of a 

 normal week's collection. 



On the 6th June, 1957 500 counts/min were recorded 



On the 24th October, 1957 300 counts/min were recorded 



And in July and August, 1958 100 counts/min were recorded 



(ii) For fall-out particles, 10 counts per min was again the activity of 

 a normal week's collection of dust. 



On the 24th October, 1957 1000 counts/min were recorded 



On the 7th November, 1957 200 counts/min were recorded 



And in August and September, 1958 40 counts/min were recorded 



(Hi) For the rainwater the counting rates are harder to interpret and 

 these are replotted on a larger scale in Figure 3. If we examine the 

 activities plotted on a larger scale this will become more obvious. 

 The counts per min per litre of rainwater have been graphed 

 (continuous line). Since the collecting area is known and the total 

 weekly rainfall can be obtained from the Bureau of Meteorology, 

 the counting rate per total weekly rainfall can be calculated, 

 assuming that the litre of water counted was of average activity for 

 the week. These are plotted as the broken line. 



As can be seen, certain peaks occur on one graph which have no corres- 

 ponding peak on the other graph. When the rainfall is low, the counting 

 rate per litre of water would tend to be high — a marked example of this 

 being seen for 21st January, 1958, when only 250 ml. of water were collected 

 and the counting rate was extrapolated to give that for one litre. When the 

 rainfall is high, the counting rate per weekly rainfall would tend to be high 

 also. 



For these reasons, it is probable that the significant peaks are those which 



129 



