Ar-Ar Geochron Table

Sample	Location	Total Gas	Plateau		Isochron
OWY-12	Lower Saddle Butte	452.8 ± 94.1	173.0 ± 144.8		n/a
OWY-13	Upper Saddle Butte	8.31 ± 0.62 Ma	7.03 ± 1.00		n/a
OWY-22	Upper West Crater	69.86 ± 19.15	37.60 ± 21.01		7.0 ± 8.5
OWY-23	Lower West Crater (?)	292 ± 39	182 ± 42		120 ± 130
OWY-35	Upper AM-PM	301 ± 24.3	247.6 ± 24.9		179 ± 21
OWY-36	Lower AM-PM	194 ± 27	n/a		n/a

Here is a blurb of related text I received from UNLV:

Nevada Isotope Geochronology Laboratory - Sample Descriptions

General Comments:

Isochrons are the most desirable treatment of ⁴⁰Ar/³⁹Ar data. This is because the isochron actually defines the isotopic composition of the initial argon in the sample (non-radiogenic argon). Ages calculated for an age spectrum are referred to as "apparent ages" because they are calculated assuming the initial argon is atmospheric in composition - thus, if there is excess argon (⁴⁰Ar/³⁶Ar > 295.5) the age will be overestimated. Isochrons have their measure of reliability, known as the mean square of weighted deviates (MSWD) which is a statistical goodness of fit parameter. If it is greater than a certain value (which changes depending on the number of points, see Wendt and Carl, 1991, the statistical distribution of the mean squared weighted deviation, Chem. Geol., v. 86, p. 275-285) then there is more scatter than can be explained by analytical errors and it is not a statistically valid isochron. If we provide an isochron it means that the statistical test is valid, if not then no valid isochron was obtained. Also, there are issues of number of data points defining the isochron - the more the better. Four points should be considered a bare minimum for statistical reasons, three points is getting to be a real concern. This can be understood simply by considering two points - a perfectly fit straight line can be put through any two points, so completely accidental data can have a perfect line fit. It follows that with three points there is less of a chance of an accidental line fit, but it is still a very real possibility (especially if analytical errors are fairly large), this possibility gets exponentially smaller as the number of points defining the line (isochron) goes up, thus more points = a more reliable isochron.

If there is no isochron, then a plateau age is next in preference. This is because a sample that gives ages which are analytically indistinguishable from step to step is exhibiting what is known as "ideal" behavior, which suggests it has a simple geologic history, e.g., rapid cooling as a basalt lava, followed by no reheating or alteration, both of which may produce disturbed (discordant) age spectra. A reliable plateau is 3 or more consecutive steps which are indistinguishable in age at the 2 sigma level and comprise >50% of the total ³⁹Ar released. The lack of an isochron or a plateau does not mean the sample provides no useful information, but their presence gives greater confidence in the ages obtained and requires less subjective interpretation.

Of course, you must consider that we run samples such as this "blind" in that we do not know the geologic relations of the samples, either when we analyze them, or when we provide these general interpretations. The geologic constraints must always be considered when interpreting isotopic ages; if any discrepancies arise feel free to discuss them with us, as it can in some cases make a difference in how age data are interpreted. All analytical errors are 1σ.