Inferred, Then Seen, Then Understood

The companion was found by arithmetic before anyone looked at it. In the 1840s Friedrich Bessel, measuring the motion of Sirius across the sky with great precision, found that it did not move in a straight line but wavered, and concluded that it must be orbiting a common centre with something unseen and massive. That is a striking inference to publish about an object nobody could detect. It was observed in 1862 by Alvan Graham Clark while testing a new telescope, appearing as a faint point almost lost in the glare of its companion. For half a century after that it was simply a very dim star, and the puzzle was why an object of substantial mass should be so faint.

The answer came from its light. Spectra obtained in the early twentieth century showed the companion to be hot and white rather than cool and dim, which meant the faintness could only be explained by an extremely small surface, and therefore by a density unlike anything then known. It became the defining example of the class of collapsed stellar remnants now called white dwarfs, an object about the mass of the sun compressed to roughly the size of the earth, and later measurements of the shift in its light provided an early test of general relativity.