"As a halo is never seen except when the sky is hazy, it indicates that moisture is accumulating in the atmosphere which will form clouds, and usually result in a storm. But the popular notion that the number of bright stars visible within the circle indicates the number of days before the storm will occur, is without any foundation whatever, and the belief is almost too absurd to be refuted. In whatever part of the sky a lunar halo is seen, one or more bright stars are always sure to be noticed inside the luminous ring, and the number visible depends entirely upon the position of the moon. Moreover, when the sky within the circle is examined with even a small telescope, hundreds of stars are visible where only one, or perhaps two or three, are perceived with the naked eye." It is possible to have five Sundays in February (the year must of course be a "leap year"). This occurred in the year 1880, Sunday falling on February 1, 8, 15, 22, and 29. But this will not happen again till the year 1920. No century year (such as 1900, 2000, etc.) could possibly have five Sundays in February, and the Rev. Richard Campbell, who investigated this matter, finds the following sequence of years in which five Sundays occur in February: 1604, 1632, 1660, 1688, 1728, 1756, 1784, 1824, 1852, 1880, 1920, 1948, 1976.[61] In an article on "The Last Day and Year of the Century: Remarks on Time Reckoning," in _Nature_, September 10, 1896, Mr. W. T. Lynn, the eminent astronomer, says, "The late Astronomer Royal, Sir George Airy, once received a letter requesting him to settle a dispute which had arisen in some local debating society, as to which would be the first day of the next century. His reply was, 'A very little consideration will suffice to show that the first day of the twentieth century will be January 1, 1901.' Simple as the matter seems, the fact that it is occasionally brought into question shows that there is some little difficulty connected with it. Probably, however, this is in a great measure due to the circumstance that the actual figures are changed on January 1, 1900, the day preceding being December 31, 1899. A century is a very definite word for an interval respecting which there is no possible room for mistake or difference of opinion. But the date of its ending depends upon that of its beginning. Our double system of backward and forward reckoning leads to a good deal of inconvenience. Our reckoning supposes (what we know was not the case, but as an era the date does equally well) that Christ was born at the end of B.C. 1. At the end of A.D. 1, therefore, one year had elapsed from the event, at the end of A.D. 100, one century, and at the end of 1900, nineteen centuries.... It is clear, then, that the year, as we call it, is an ordinal number, and that 1900 years from the birth of Christ (reckoning as we do from B.C. 1) will not be completed until the end of December 31