Vll 



position it was useless for the purpose for which it was intended, and 

 asking that my observations as to the error, and calculations of its 

 amount, might be referred to a practical astronomer for verification. 

 The request was referred to Admiral Bowden-Smith, who replied that 

 directions had been given for the Dart to comply with it. Captain 

 Purey-Cust then came to the Observatory, and I placed everything at 

 his disposal. After making a great number of observations he prepared 

 the following report to the Commander-in-Chief of H.M. ships, 

 Australia : — 



Difficulty in Ascertaining Correct Time with Transit Instruments at 

 Observatory, Hobart.— H.M.S. Dart, in Frederick Henry Bay, April 5, 

 1894. Sir, — In accordance with your memo of February 21 as to the 

 above, I have the honour to report as follows:— 1. Transit Instrument — 

 The instrument in use is of 2iu. aperture, and 22Jin. focal length ; it has 

 none of the usual accessories of modern transit instruments, such as 

 micrometer, tangent screw for sideway motion of the eyepiece, reversing 

 apparatus, etc., which greatly facilitate and add bo the excellence and 

 accuracy of the observations obtained ; still, in the hands of a carefnl 

 observer no doubt an accuracy sufficient for the purpose required, viz,, 

 the rating of chronometers, is obtained. Mr. Kingsmill has lately 

 caused to be added a Bohenbergen eyepiece and amalgamated mercury 

 trough ; this has greatly facilitated the speedy and correct adjustment 

 for collimation and level. The amalgamated form of mercury trough is 

 especially essential on account of the vibration set up by the trams, 

 which pass on either side of the Observatory at a distance of under 

 quarter mile on an average of four times every quarter of an hour. 2, 

 Clocks. — There are two clocks — one, a sidereal in the transit hut, is 

 exposed to every variation of temperature, and in consequence has an 

 ever-fluctuating rate ; the other, a mean solar clock in the Observatory 

 building, has to be corrected every day at noon to exact mean time for 

 dropping the ball, and can therefore be said to have no known rate. 

 Consequently in the event of cloudy weather no dependence can be 

 placed on either of them, and the time is regulated by a single box 

 chronometer kept in the Observatory building. At present there are 

 no ready means of accurately comparing the clock in the transit hut 

 used for observation with either the chronometer or the mean solar 

 clock. At small expense a clock face might be fitted close to the 

 mean solar clock electrically connected with the sidereal clock, affording 

 a simple and extremely accurate method of comparison between the 

 two. This is the method usually employed in all observatories. A 

 good standard clock is much needed ; this could be placed in the cellar 

 for the sake of uniformity of temperature, and electrically connected as 

 above with the room in which is the mean solar clock. 3. Error in 

 Position of the Meridian Mark on Mount JNelson.— In order to obtain 

 this the distance of the meridian mark from the transit instrument was 

 first ascertained by triangulating between the two, using certain data 

 obtained from a local triangulation by Mr. Mault, for a base line. 

 This distance was found to" be 6,479ft. A wooden scale was then 

 erected horizontally across the meridian mark, and graduated in arc so 

 as to form a distant micrometer when viewed through the telescope of 

 the transit instrument ; this enabled the azimuth error, as found by 

 observation, to be actually shown in lineal distance at the meridian 

 mark. Observations were made as opportunities offered on various 

 nights of, pairs of stars differing considerably in declination, of polar 

 stars and pairs of circumpolar stars at opposite culminations, with the 

 resulting mean deviation of 64" of arc, showing the present position of 

 the meridian mark to be 2ft. to the westward of the true meridian. 

 This corresponds very nearly with the error that Mr. Kingsmill had 

 previously estimated and allowed for. The probable error of the 

 observations by the method of the least squares is + - 2"6 ; this cor- 



