I now pass from a general description of the circles to the azimuths of the sight-lines already referred to, so far as they can be determined from the published Ordnance maps. To investigate them as completely as possible without local observations in the first instance, I begged Colonel Johnston, R.E., C.B., the Director-General of the Ordnance Survey, to send me the 25-inch maps of the site giving the exact azimuth of the side lines. This he obligingly did, and I have to express my great indebtedness to him. In Fig. 41 I show the sight-lines from the south and north Circles as determined by the stones and barrows marked on the map. The sight-lines on Arcturus are from the centres of the three circles in succession. I shall point out later the significance of the fact that the November alignments are from the south, the solstitial ones from the north Circle. [Illustration: FIG. 41.--The Sight-lines at the Hurlers.] Of the various sight-lines found, those to which I wish to direct attention in the first instance, and which led me to the others, are approximately, reading the azimuths to the nearest degree, Lat. 50° 31′ N. Az. S. circle to central circle N. 12° E. Central to N. circle N. 15° E. N. circle to tumulus N. 19° E. In a preliminary inquiry in anticipation of the necessary local observations with a theodolite, I assumed hills half a degree high, for the reason given on p. 112. We have the following declinations approximately:-- Dec. N. 38¹⁄₂° „ 38° „ 37° Here, then, we have declinations to work on, but declinations of what star? To endeavour to answer this question I studied the declinations of the three brightest stars in the northern heavens, having approximately the declinations in question some time or other during the period 0 to 2500 B.C. Vega is ruled out as its declination was too high. The remaining stars Capella and Arcturus may have been observed so far as the declinations go. For time limits we have:-- Dec. N. Capella. Arcturus. 38¹⁄₂° 500 B.C. 1600 B.C. 36° 1050 „ 1150 „ Now there is no question as to which of these two stars we have to deal with, for the northern circle is evidently less ancient than the others, for some of the stones are squared and the others are less irregular than those in the S. circle. This being so, the approximate dates of the use of the three circles at the Hurlers can be derived. They are, with the above assumption:-- B.C. Southern circle aligning Arcturus over centre of central circle 1600 Central „ „ „ N. circle 1500 Northern „ „ „ tumulus 1300