To Unst, in Shetland, the northern extremity of the arc just mentioned, sergeant Steel now repaired, and ascertained the existence of a disturbance at Saxavord, but in the contrary direction. This was fully established by taking a similar series of observations at the Gerth of Scaw, near Lambaness, and on the small uninhabited island of Balta. The relative position of these stations he fixed astronomically as to latitude, and geodetically by triangulation and levelling from the mean level of the sea, which involved observations with regard to the ebb and flow of the tides. By the series of observations so far made, it was clearly proved, that the latitude of a place could not be measured with the degree of certainty formerly supposed, and that though astronomers may profess to give seconds, tenths, and even hundredths of a second of their latitude, yet the real truth is, that the record may often be _several whole seconds in error_. The discovery, now confirmed by sergeant Steel’s inflexible accuracy, is likely to produce some interesting discussion in the scientific world, and has already been made the subject of an article in the ‘Philosophical Journal of Science’ for April, 1853, embraced in a review of Captain Yolland’s Sector Volume.[59] After passing a station at Cowhythe Hill, in Banffshire, to verify the sector operations of 1813, and which object was satisfactorily attained, the sergeant fixed his observatory at Southampton, where, in carrying on the duty, he made various experiments to ascertain the cause of apparent errors. In taking the usual readings of the telescope micrometers, the value of the zenith point, derived from each double observation of a star, varied sensibly. To determine this more accurately, by ascertaining the true value of the divisions of the screw, and correcting the error involved in the reduction of the whole of the observation, he adopted the method of making two distinct observations of the same star without reversal of the revolving frame, in the manner described in the Sector Volume, page xxvii, and so excellent was this method considered, that the value of the screw thus obtained, was finally applied to all the observations.