POROREEO IRE SEI SM a re pe ee ee ye

a

ane

Aen

“a

CHAPTER 5

DISCUSSION
5.1

DISCUSSION OF ERRORS
The statistiesl errors in.the counting rates have been minirized

vbenever possible by reading the activities at specified tires from

smooth decoy curves.
The omission of a self-absorption - selfascattering correction in the reduction of the analytical data for Sr 9, 2x95

pal40, and Celhas introduced additional errors.

The maximum

possible veriation, or spread, in the data that this omission can

produce is $.5 per cent for Sr89, 4.6 per cent for Zr95, 2.7 pvr cent
for Bal40, and 10 per cent for Celt44¥, These maximum percent2ces were

calculated by taking the difference between the extreme self-cbserptione
self~scattering correctionsfor 211 the samples and dividing by the

°

average of these two corrections.

The self-absorption-self-src.ttering

corrections were estimezted from the data presented by Engels-+reir et

al.18/

In all cases the errors among the four aliquots of any ore

sample are much smaller than the maximum possible error presented here,
These two sources of errors are combined in the estinate of tha
totel error for each analysis. These estirates are reported as syresds
in Table 5.1. ° As in the calculation of the maxirum possible error tie
to self-absorption - self-scattering effects, the
spread = highest valua ~ lowest value

average velue

x 100

In general the spreads are approximately twice the stendard ‘eviattions,
In the calculations of the S-~values the spreads are comround.1 ‘y tre
equation

where
‘

2/2
SR = (sf2. + 332. + 332, + si)
Sp = Spread of the R-value

Sy, S2, 33, and Sy = the spreads of the individual analyses for the

mclides in question (Cf, definition of ?-valse
section 4,1)
35

SECRET — RESTRICTED DATA

Select target paragraph3