Mathematics – Logic
Scientific paper
Apr 1961
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=1961gecoa..23..116s&link_type=abstract
Geochimica et Cosmochimica Acta, vol. 23, Issue 1-2, pp.116-134
Mathematics
Logic
8
Scientific paper
Recent articles on element distributions have given insufficient attention to geological aspects of the problems, and have neglected to consider the effects of sampling, sample reduction and analysis error on the interpretation of frequency distributions. No single probability function can be expected to suit all elements. These and other considerations load to the proposal of five rules (given below) concerning trace element distributions which include ' proposed law (1954a) in a revised form. The first three rules are based on theoretical considerations and require further experimental proof except in the case of certain ore deposits whose lognormal distribution is well established. It is recommended, however, that the lognormal law be used as the basis for discrimination tests in all cases. 1. (1) The frequency distribution of concentrations of a camouflaged trace element in a homogeneous rock formed by a chemical equilibrium process can be considered as continuous probability functions with limits zero and + . 2. (2) The best probability function to use is the logrionnal law, which is justified both by practical utility and theoretical considerations. 3. (3) The previous rules also apply to those trace and minor elements which are largely concentrated in accessory or ore minerals of simple constitution. Many economic deposits are in this category. The Poisson function may also be convenient in this case, if the discontinuous distribution of mineral grains is considered. 4. (4) The geochemical variance of a group of data is less than the measured variance, which should be corrected by subtracting manipulation variance (sampling, sample reduction, analysis). 5. (5) The coefficient of variation c g (= s.d./mean) of a small body of geochemical data provides a criterion for the choice between a normal and a lognormal law for subsequent statistical tests. When c g >> 0.2 either law will predict similar probabilities, but when c g >> 0.2 the lognonnal law is more realistic. It is important to correct the observed relative variance c 2 (as in rule 4) to obtain c g 2 , otherwise the choice will be biassed towards a lognormal law.
No associations
LandOfFree
Element distribution laws in geochemistry does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Element distribution laws in geochemistry, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Element distribution laws in geochemistry will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1805588