Is the Universe Birefringent?

Borge Nodland and John Ralston claimed in Physical Review Letters (78, 3043, 1997) to have detected rotation in the polarization of radio waves from distant galaxies, as if the Universe were birefringent (sort of, but not really, like a giant piece of icelandic spar). Daniel Eisenstein and I published a brief comment explaining an error in this paper. We concluded that there is at this point no demonstrated evidence that this effect is real.

Our comment, which was published in Physical Review Letters (vol. 79, p. 1957), is available from this site in either HTML or PostScript format. The PostScript format will look nicer and be easier to read; the HTML is provided for those whose Web browsers don't know how to handle PostScript.

If you really don't like the HTML version, but your machine won't let you view the PostScript, you can always download a free PostScript previewer. If you have trouble getting these files, or if you want more information, send me e-mail.

Here are some related sources of information:

• Nodland and Ralston's paper is available from the astro-ph preprint archive as astro-ph/9704196.
• In addition to the above paper, Nodland and Ralston have written another description of their results.
• Dr. Nodland has written a less technical description of their findings.
• Nodland and Ralston have distributed a response detailing why they disagree with our comment. Our reply to this document is available from this site, in either HTML or PostScript format.
• Ralston and Nodland have written an update on the status of their claim.
• Jain and Ralston have performed an analysis of some different data to search for a correlation between polarization angle and position on the sky. They claim a statistically significant detection.
• Sean Carroll and George Field have written a paper in which they analyze the data set used by Nodland and Ralston. They find no statistically significant evidence of birefringence. Incidentally, Carroll and Field, together with Roman Jackiw, originally collected this data set and used it to search for effects like the one Nodland and Ralston claim to have found. See Phys. Rev. D 41:1231 (1990) for details.
• Carroll has written an on-line summary of the key results of the above paper.
• J.P. Leahy has written a comment in which high-resolution data on distant radio sources is used to set very stringent limits on the possible amount of birefringence. According to Leahy, the maximum possible amount of birefringence is some 30 times lower than the level claimed by Nodland and Ralston.
• Nodland and Ralston have responded to Leahy's comment. Leahy has written a reply to this response and appended it to his original comment.
• A paper by Wardle, Perley, and Cohen uses similar techniques to Leahy's and reaches similarly strong conclusions. A press release describing the paper is available from the National Radio Astronomy Observatory.
• Loredo et al. use Bayesian statistical techniques to analyze the Nodland-Ralston data. They do not detect birefringence; in fact, they set an upper limit several times smaller than the Nodland-Ralston claim.
• There have been at least seven theoretical papers, by Moffat, Obukhov et al., Dobado and Maroto, Bracewell and Eshleman, Ralston et al., Kühne, and Mansouri and Nozari attempting to explain the Nodland-Ralston effect. (Yes, it is the same Ralston; Nodland is a coauthor. Incidentally, Obukhov et al.'s model is, contrary to their assertion, strongly ruled out by observations of the microwave background. In an act of shameless self-promotion, I urge you to check out a paper I wrote to see why.)
• Ned Wright did some calculations to assess the statistical significance of the Nodland-Ralston result and concluded that the effect was not statistically significant. He summarizes his conclusions in his cosmology tutorial. (Click on the "old news of the universe" link.)

Back to Ted Bunn's home page.