Sorry! the second half of my post was missing

The amounts of the other isotopes are then easily calculated from their reported ratios wrt

^{206}Pb

^{207}Pb = 112.9 and

^{208}Pb = 34.2

^{204}Pb is negligible

You then use the amounts of

^{206}Pb and

^{207}Pb to calculate the amount of

^{238}U and

^{235}U from the reported

^{206}Pb/

^{238}U and

^{207}Pb/

^{235}U ratios and obtain a value of 111.76 for

^{238}U and 0.81 for

^{235}U. Mistake no. 2. This is also wrong!

The Pb and the Pb/U isotopic ratios are reported in atomic ratios whereas (although this is not stated in the table). The actual quantities of Pb and U are reported in weight ratios. To convert from one to another the atomic weights of each isotope must be taken into account. Thus the ratio of

^{206}Pb/

^{238}U in atomic units must be multiplied by the ratio of their atomic weights ie (206/238)=0.866 to obtain the weight ratio.

Similarly for

^{207}Pb and

^{238}U. The conversion factor is (207/235) = 0.884.

Using the values from Wilde, row 1, Table 1.

^{206}Pb/

^{238}U = 0.928 in atomic units

^{206}Pb/

^{238}U = 0.928 * 0.866 = 0.803 in weight units, and the value for

^{207}Pb/

^{235}U becomes 69.5 * 0.884 = 61.44

So

^{238}U =

^{206}Pb/0.928 = 207.86/0.928 = 258.85 and

^{235}U =

^{207}Pb/61.44 = 1.84

As there are only three naturally occuring isotopes of Uranium the amount of

^{234}U can be calculated from the simple difference between U

_{Total} less the sum (

^{238}U +

^{235}U) i.e.258 - (258.85 + 1.84) = -2.69. The slight error is due to the errors in the absolute concentrations of U and Pb caused by uncertainties in the sample weight - something that I imagine is common to all SHRIMP analyses of zircons - and large uncertainties in the concentration of U and Pb, which can, apparently, be up to as much as 20%! As these values are not used in the calculations of age in the concordia method these errors have no bearing on the result and presumably is the reason why no errors are listed in the Table. The Pb/Pb and Pb/U ratios used for age calculation are, however, extremely accurate as Wilde's Table 1 shows.

So your more accurate method is based on a simple algebraic mistake and a missing conversion factor. This hardly inspires confidence in the accuracy of your other model.

Best wishes

Chris