"4. They are of a spectral type approximately solar; no Orion, Sirian or Arcturian stars having been found among them. "5. They seem to be found in greater numbers in certain parts of the sky, notably in the Milky Way, but exhibit no tendency to form clusters. "6. All those stars whose radial velocities have been studied have been found to be binaries whose period of orbital revolution coincides with that of their light change. "7. The orbits, so far as determined, are all small, _a_ sin _i_ being 2,000,000 kilometres or less. "8. Their maximum light synchronizes with their maximum velocity of approach, and minimum light with maximum velocity of recession. "9. No case has been found in which the spectrum of more than one component has been bright enough to be recorded in the spectrograms."[347] It is very difficult to find an hypothesis which will explain satisfactorily _all_ these characteristics, and attempts in this direction have not proved very successful. Mr. J. C. Duncan suggests the action of an absorbing atmosphere surrounding the component stars. On March 30, 1612, Scheiner saw a star near Jupiter. It was at first equal in brightness to Jupiter's satellites. It gradually faded, and on April 8 of the same year it was only seen with much difficulty in a very clear sky. "After that date it was never seen again, although carefully looked for under favourable conditions." An attempted identification of Scheiner's star was made in recent years by Winnecke. He found that its position, as indicated by Scheiner, agrees with that of the Bonn _Durchmusterung_ star 15°, 2083 (8½ magnitude). This star is not a known variable. Winnecke watched it for 17 years, but found no variation of light. From Scheiner's recorded observations his star seems to have reached the 6th magnitude, which is considerably brighter than the _Durchmusterung_ star watched by Winnecke.[348] With reference to the colours of the stars, the supposed change of colour in Sirius from red to white is well known, and will be considered in the chapter on the Constellations. The bright star Arcturus has also been suspected of variation in colour. About the middle of the nineteenth century Dr. Julius Schmidt, of Athens, the well-known observer of variable stars, thought it one of the reddest stars in the sky, especially in the year 1841, when he found its colour comparable with that of the planet Mars.[349] In 1852, however, he was surprised to find it yellow and devoid of any reddish tinge; in colour it was lighter than that of Capella. In 1863, Mr. Jacob Ennis found it "decidedly orange." Ptolemy and Al-Sufi called it red.