Paul Dennis posted this in the comments. It's always nice to have some opinion from an expert, and I thought it was striking enough to turn into a head post.

The Iain Stewart article posted on the Met Office website is both misleading, highly selective and largely in error in its presentation of data. Misleading in the presentation of graphics that supposedly represent the timings of mass extinction events in relation to atmospheric CO2 levels. The plot shows extinction events occuring at local maxima in CO2 levels e.g. late Cambrian, Ordovician and Devonian amongst others. The reality is that our knowledge of past CO2 levels during these times is incomplete at best and certainly lacks the resolution in both time and CO2 level to highlight local maxima.

The article is highly selective in suggesting that there is a correlation between earth surface temperature and CO2 levels. Any examination of the data will show that again our knowledge of Earth surface temperatures and CO2 levels are not good enough to make such statements. There is abundant geological evidence for extended periods of wide spread glaciation in many periods of the geologic past that are thought also to be periods of high CO2 levels. For example, during the late Ordovician when there is a mass extinction event. This indicates a decoupling of CO2 and temperature. The isotopic evidence has been interpreted as indicating a widespread decoupling between surface temperature and CO2 (see Veizer et al., 2000, Evidence for decoupling of atmospheric CO2 and global climate during the Phanerozoic eon, Nature, vol. 408, 698-701). Work is ongoing using clumped isotopes to try and refine the temperature estimates.

The article is disingenuous in suggesting that there is 'little patience' for talk of cosmic rays. In my experience very few geologists have studied cosmic rays in terms of their possible effects on climate. Those that understand anything about cosmic rays have often come from a background of radiocarbon dating, or in-situ cosmogenic isotope studies, or from a geophyiscal perspective (e.g. Vincent Courtillot and others) and have a much more open mind to these new ideas. By far the majority of geologists have little knowledge of, or understanding of cosmic rays.

The errors in the article are straightforward and should not have been made and perhaps belies Stewart's point of view. He writes of Details lost in time:

While the recent history of carbon dioxide levels comes from modern instruments, or before that from annual coral or tree growth rings or snow layers, the distant geological history of CO2 marches back in steps of 10 million years. In other words, the further back you peer in time, the more the subtle intricacies of climate are smeared out.

Here he is explicitly linking climate and CO2 levels. Perhaps it is a Freudian slip. The truth is past CO2 levels are not determined from annual coral, tree growth rings or snow layers. He subtly links CO2 and climate in these 2 sentences!

Stewart is writing from a pre-determined position and doesn't represent or understand the views of most geologists. He certainly doesn't represent my views.

[Update 7/7/12 The original article has disappeared from the Met Office site. I've replaced the link with a Wayback Machine copy]