OJ287 has been trumpeted as an example of true periodicity in a quasar. But is it? 

We have already seen how, if OJ287 is really periodic, after a century of observations we should be able to measure the interval between its outbursts better. Could the problem be more sinister though?

Let us suppose that the light curve of OJ287 shows a period of exactly 11.6 years. That means that if we slice up the light curve of OJ287 into pieces 11.6 years long and stack them on top of each other, the outburst should happen in exactly the same place every time. If the period is a little longer than that they will appear to happen slightly earlier each time; if shorter, they will happen slightly later. As Figure 3 shows, they line up just nicely if you assume that the period is 11.85 years. So, what is the problem?
 

 

Figure 3 – The same light curve as in Figure 1 but with the data lined up assuming that there is a period of 11.85 years in the outbursts. We can see how the big maxima line up, although there is no information on them prior to 1948 as there are too few observations to cover the light curve well enough before then. The vertical scale is in what astronomers call flux units rather than magnitudes; the advantage of these is that, unlike magnitudes, twice as big does, genuinely mean twice as bright. Image prepared by the author.

The problem is that you can only get them to line up in 1948, 1960, 1983 and 1994 and, just possibly, 1972, but there are too few data to be certain. Before 1948 there is no evidence of the period whatsoever, mainly because nobody was looking at the right times, although in 1912 people were looking just about the time when OJ287 should have been especially bright and it was not (that could be down to bad luck and the maximum being brief and just missed). Of course there are two maxima separated by a little over a year. The second maximum lines up neatly in 1995, 1984, 1937, (probably in) 1925 and in 1913, but it is not there in 1901 and 1973. In other words, there is just enough doubt to make me wonder if OJ287 is just fooling us. We will see in 2006 and 2007.