238 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1948 
200 
180 
Mode,3350 my. 
160 
Frequency 
120 
100 
sot ——— Average,3200 my. 
60 
40 
20 
2000 2500 3000 
Average(without J),3290 my: 
£o in Millions of years 
3500 4000 
Ficure 6.—Histogram representing the frequency distribution of 1,419 solutions for ¢, 
within ranges of 100 million years. The frequencies of 162 solutions involving the ab- 
normal Joplin lead samples (Nos. 9 and 11) are indicated by the dotted areas. The 
histogram for the 1,257 solutions (without the aberrant Joplin results) is outlined boldly. 
younger lead samples. ‘The intersec- 
tions provide over 200 solutions, with 
a marked concentration about 3,300-— 
3,400 million years. Although there 
are naturally a few aberrant results, 
these are distributed more or less 
evenly about the mean, suggesting that 
no systematic or cumulative error is 
involved in the data. 
Additional sets of curves can be con- 
structed for pairs based on No. 23 
(Cerussite, Broken Hill, N. S. W., 
1,200 million years), No. 22 (Galena, 
same locality and age), and No. 21 
(Galena, Quebec, 800 million years). 
From the intersections of all five sets 
of curves and of the corresponding yt 
curves, 1,257 solutions have been ob- 
tained for f., x, and y. The average 
value of ¢, is 3,290 million years, but a 
histogram of the ¢, values (fig. 6) re- 
veals a well-marked mode at 3,350 
million years, which is found to be 
somewhat better than the average 
value when tested by a least-squares 
method (Holmes, 1947a, p. 127). The 
spread of the results is comparable with 
that of figure 5, and is not surprising 
in view of the many possibilities of 
error. Careful consideration of the 
latter suggests that on the whole they 
tend to counterbalance rather than to 
become cumulative. The dotted fre- 
quencies on the left of figure 6 illus- 
