.
et
Foghensnncahl
SS 8
o
Ee)
evar
would be required
(including one repeat for an initiator test),
unless some combinations of the five items are possible.
Cer-
tain combinations are possible, at sacrifice of definiteness, but
it will be shown by the detailed considerations of the specific
plan that the minimum number of shots is three,
each one of
which has to be "good", im a sense to be defined.
When it is
realized that atomic weapons are subject to fiuctuations in behavior beyond human contro],
thin.
cven a three-shot program appears
A test plan invoiving & small number of shots has to
ignore such statistical considerations and rely on a degree
of compensation provided by the fact that an unfavorable fluctuation,
resulting in low yield,
formation,
alpha,
can still yield behavior in-
especially from other measured quantities such as
the rate of rise of the reaction rate.
In Pact,
this “¢
generally true regardless of the cause of a low yield and emphasizes that the test program is a series of experiments
on weapon types rather than simply a confirmation or disproof
of behavior predicted from the performance of the first FM type.
/
Dok
Gk
Finally,
considerations of production indicate that it
would be very desirable,
o*
a
if possible,
to investigate the efrect
a decrease in the permissible lower limit of initiator Siren
gth.
It is agreed, however, hat fissitonable material is more