... a car produces 1/3 of its pollution before it hits the showroom....

Not 1/3.  According to the ILCA, it's 10%.

A simple carbon tax will roll the cost in automatically.  That's why it's so important.

10% is way too low

Dirty from cradle to grave

10% is right, unless you accept Whitelegg's assumptions (including that the vehicle only runs 81,000 miles in its lifetime).  ILEA assumes 160,000 miles, which is reasonable for the USA (I sold my Taurus with about that much on the odometer).  Their assumptions about economy are about the same (21.8 MPG for ILEA, 23.4 for your cite) so that doesn't account for it.

If you'd paid attention, you would have noticed that your link talks of things like "cubic meters of polluted air", a rather elastic measurement (if you concentrate or dilute the emissions, you can make the numbers into whatever you want them to be).  Last, it's a newspaper article duplicated on a personal site.  You should be citing - and reading - the original source; there's no telling what the reporter decided to leave out.

Of course, you're hoisting yourself with your own petard there.  The ILEA page you cite appears to be simply reporting the results of a 1998 Carnegie Mellon study.  That isn't exactly "citing - and reading - the original source."  

Frankly, it looks as though the difference is largely down to the amount that the vehicle is driven and the Heidelberg research including the eventual disposal, which the Carnegie Mellon research left out.  Europeans drive less so that manufacture is a bigger percentage, Americans drive more so manufacture is a smaller percentage.

No, it's much more than mileage; if you assume an 81,000 mile lifespan the manufacturing fraction only rises to ~20%.  Something else is needed to account for another multiplier of roughly 1.6.
On the mileage thing brand new totaled vehicle wrecks bring the average down, as do any wrecks.