This article (http://www.google.com/hostednews/ap/article/ALeqM5in1T5nvGNc...) mentions previous results and that they were thought to be due to measurement error, but claims that these readings had a speed of c + 60 nS, with a measurement margin of error of +-10 nS.