While I don't have much to say about the paper itself, it is again time to appreciate people who comment on the arXiv blog. The post in question:
http://www.technologyreview.com/blog/arxiv/25420/
The comment in question:
Wrong domain for graph
If you wanted to prove that there's a 27 Million year peak, wouldn't it make sense to present the frequency domain data instead of the time domain data? If there were a peak every 27 Million years, then it would be clear from a frequency domain graph.
And the evidence bsdimp did not read the paper itself?
