must be known to convert the volume to mass of air and due to uncertainties in the absolute altitude of the collection, the mass of air
collected is not known to better than + 15 per cent below 70,000 feet
and + 25 per cent above 70,000 feet.
Thus the error in determining the mass of air containing the tritium
is the predominant error.
Combining all the sources of error,
it is estimated that a single
determination of the tritium concentration should be determinable to
+ 4 per cent with a somewhat greater accuracy at the lower levels of the
stratosphere,
5.
5.1
RESULTS
Early Results
A few early results were obtained in the period between June 1955
and August 1956; all samples being collected in the neighborhood of
Minneapolis, Minnesota.
These early results when added to excess carbon~l4 and related
to fissions through krypton-35 measurements in the same samples, give
values for excess neutrons per fission in good agreement with the
_weighted average number of excess neutrons calculated from the known
performance of the Castle devices.
We ray, therefore, > heave sone confi-
dence in these early data even though the tracing vas performed by adding
deuterium oxide to the collection bag before launching the balloon.
These carly results are recorded here in Table 5.1.1 to show that
the more recent results of 1957-1958 are not essentially lower or higher
than the concentrations in the stratosphere a year to two years after