This axis is supported horizontally in such a way that it points as nearly as may be exactly east and west. The telescope itself being square with the axis, always points in a north-and-south direction. It is possible to rotate the telescope about its axis so as to reach all parts of the sky that are directly north or south of the observatory. In the field of view of the telescope certain very fine threads are mounted so as to form a little cross. As the telescope is rotated this cross traces out, as it were, a great circle on the sky; and this great circle is called the astronomical meridian. Now we are in possession of certain star-tables, computed from the combined observations of astronomers in the last 150 years. These tables tell us the exact moment of time when any star is on the meridian. To discover, therefore, whether our clock is right on any given night, it is merely necessary to watch a star with the telescope, and note the exact instant by the clock when it reaches the little cross in the field of view. Knowing from the astronomical tables the time when the star ought to have been on the meridian, and having observed the clock time when it is actually there, the difference is, of course, the error of the clock. The result can be checked by observations of other stars, and the slight personal errors of observation can be rendered harmless by taking the mean from several stars. By an hour's work on a fine night it is possible to fix the clock error quite easily within the one-twentieth part of a second. We have not space to enter into the interesting details of the methods by which the astronomical transit is accurately set in the right position, and how any slight residual error in its setting can be eliminated from our results by certain processes of computation. It must suffice to say that practically all time determinations in the observatory depend substantially upon the procedure outlined above. The observatory clock having been once set right by observations of the sky, its error can be re-determined every few days quite easily. Thus even the small irregularities of its nearly perfect mechanism can be prevented from accumulating until they might reach a harmful magnitude. But we obtain in this way only a correct standard of time within the observatory itself. How can this be made available for the general public? The problem is quite simple with the aid of the electric telegraph. We shall give a brief account of the methods now in use in New York City, and these may be taken as essentially representative of those employed elsewhere.