From John Beckman:

Re: Bars may kill spiral galaxies (Nov. 8): I am an astrophysicist working at the IAC (Astronomical Institute of the Canaries), and one of my specialities is the structure of galaxies.

It has been known for some time that although the fraction of spiral galaxies observed at optical wavelengths which have identified bars is close to 50%, this fraction rises to over 65% when observations are made in the infrared, and with very careful analysis of the two-dimensional light structure an extra fraction of galaxies is found to have bars, whether the observations are made at any wavelength (i. e. at optical wavelengths this rises to well over 50% and in the infrared to over 70%). In other words although the original classification by Edwin Hubble was into "normal" and "barred" galaxies, in fact barred galaxies are more "normal" than unbarred galaxies. This is point number one. Point number two is that the reason why a higher fraction appears barred in the infrared (and the same would hold true between blue and red, i. e. more galaxies would appear barred when observed in the red) is that not only does the light from bluer galaxies come from more specific zones, i. e. star forming zones, and does not exactly reflect the underlying mass distribution, so that a bar in the total stellar population could easily be masked, bluer galaxies also contain more interstellar dust, whose distribution follows that of the interstellar gas rather than that of the stars, and is in any case much more patchy, much less smooth. Here again the underlying mass distribution of stars can be partly masked. If the same blue galaxies were observed in the infrared a higher fraction of them would show bars, because the dust absorbs much less in the infrared, and the contribution of star forming regions to the light is also relatively less in the infrared than in the blue. So the effectdescribed in the article may well not be due to bars being less common in blue spirals, but that they are much less easy to detect. This would be particularly true for the type of work done in Galaxy Zoo, where visual detection is the research tool, and the images are not especially deep. I should add that I greatly respect the work of Galaxy Zoo and do not want to raise any kind of general criticism of it.