STOCHASTIC SIMULATION OF RADIONUCLIDE UPTAKE
817
tribution of 14] from local fallout on plants was prescribed as log-
normal. (Whether lognormal distributions are characteristic of mate-
rials deposited in stratospheric fallout is not known, but Osburn® has
pointed out that gross beta activity from stratospheric fallout in Colorado alpine environments is lognormally distributed). Did the use of
lognormal distributions of ‘I on plants give rise to the asymmetrical
distributions predicted in consumer populations, or is the frequency
distribution of a radioisotope in a population largely independent of
the distribution of the substance in the diets of the consumers? As of
now, these questions cannot be resolved. If the first of the alternatives
is true, then non-Gaussian distributions of radionuclides in consumer
populations may simply reflect the distribution of such substances in
the environment.
=
tees
es
Ahrens®’ has maintained that itive:‘elements (e.g., thorium) are
lognormally distributed in granitie rocks. Rogers and Adams*® have
confirmed these observations and,have developéd a model that predicts
lognormal distributions of roek materials present in low concentrations. Thus there is a theoretical basis for-expectingnaturally occur-
ring radioelements (such as thorium and **Ra) to be lognormally
distributed within homogeneous geological bodies. ‘However, there is
no obvious relation between the Rogers—Adams model, developed in
terms of the fractional crystallization or diffusion of elements in
rock, and the deposition of fallout radionuclides.
Regardless of Whether skewed frequency distributions of radioelements in consumers arise because ofthe distribution of these sub-
stances in their diets or in spite of it, there is one point pertinent to
both arguments. The Rogers—Adams mode}-applies. to distributions of
elements within homogeneous bodies. Thorium is lognormally distributed in samples from the
ynway granite of New Hampshire, but
if samples are included fromadyeinifig granites the new distribution is
no longer lognormal.
In biology,
the population is anajogous to the
homogeneous body of the Peslsgiets. Hence, if different populations
are combined to create a large composite distribution, there is no way
to predict its nature. The, distribution may be of almost any form.
Kulp et al. give such a distribution (Ref. 39, Fig. 4D, p. 1253); and, as
pointed out by the authors, it “...is clearly not normal, nor does it
correspond closely to
de ormal pattern.” The distribution of *°Sr
in the world population 3fy be.useful in some endeavors, butit is
much easier to attac Dn fiigical significance to distributions of radionuclides in consu ss@@roeh 2 localized area. For example, the dis-
tribution of gy. ing pele from New York City (Ref. 39, Fig. 4A,
p. 1253) appeareapproxif™ltely lognormal.
A. final problem is ‘whether or not the form of some of the observed distributions is a technical artifact. If values at the low end of
a distribution are particularly susceptible to error or are rejected