IN TTTE SOLAR SYSTEM. 103 



The knowledge of this coincidence lends an interest to the quasi prediction of the in- 

 telligent Olbers, which before it wanted ; and its fulfilment, should it occur, will certainly 

 give stability to the theory which induced the search for and discovery of the cycle of 

 solar rotation. 



Theophancs, one of the Byzantine historians, records that, in November of the year 

 472, the sky appeared to be on fire over the city of Constantinople, with coruscations of 

 flying meteors.* 



Deducting the year 472 from 1799, gives us 1327, which is one year more than 39 cycles 

 of solar rotation. The fact is, therefore, only interesting to us, as it awakens the suspicion 

 that the historian who recorded an event occurring three centuries before his own time, 

 may have erred in the date by a single year. 



The considerations which induced me, when investigating the influence of ether upon 

 the seasons, to conclude that the mean day between the dog days and the Indian summer, 

 should be more advanced in the year than the 8th September, have a similar bearing upon 

 the subject of periodical shooting stars. 



Between the 10th August and the 12th November, the mean day is the 26th September, 

 which is eighteen days after the 8th September. It is a fair subject of inquiry, What 

 is the significancy of these eighteen days % 



Without being able to offer an opinion for which I can furnish a basis, I respectfully 

 submit that, under the influence of the tangential force, acting upon the metallic stream 

 after leaving the sun, the eighteen days represent the time required for the vaporous metals 

 to reach the earth's orbit. 



The investigation of the relations between the period of the sun's rotation and the 

 sidereal year, not only teaches us that there is a very great probability of a display 

 of shooting stars on the 12th November, 18(57, but further, that there is a lesser proba- 

 bility of a smaller exhibition on the anniversaries of that day in the years 1864 and 1870, 

 and for the following reasons: 



The number of days in three years is 1095.795. This is just sixteen hours and fifty-five 

 minutes less than forty-three rotations of the sun. 



In a whole cycle of solar rotation, the years which show the greatest probability of a 

 meteoric display, after the termini of the cycle, are the third year before, and the third 

 year after those epochs. 



It is not likely that the inhabitants of the countries which witness the displays at the 

 termini of the cycles, will also sec those which may occur at the triennial anniversaries ; 

 as the longitudes of the places where the three phenomena will be visible are different. 



On examining Figure 6, showing the relation between the earth's orbit and the ascend- 



* Gallery of Nature, p. 138. 



