I now pass on from Merrivale as an example of those avenues the direction of which lies somewhere in the E.-W. direction. Others which I have not seen, given by Rowe, are at Assacombe, Drizzlecombe and Trowlesworthy; to these Mr. Worth adds Harter or Har Tor (or Black Tor). The avenues which lie nearly N. and S. are more numerous. Rowe gives the following:--Fernworthy, Challacombe, Trowlesworthy, Stalldon Moor, Battendon, Hook Lake, and Tristis Rock. Of these I have visited the first two, as well as one on Shovel Down not named by Rowe, and the next two I have studied on the 6-inch Ordnance map. _Fernworthy_ (lat. 50° 38′).--Here are two avenues, one with azimuth N. 15° 45′ E., hills 1° 15′. There is a sighting stone at the N. end. We appear to be dealing with Arcturus as clock-star 1610 B.C. This is about the date of the erection of the N. avenue at Merrivale. The second avenue has its sighting stone built into a wall at the south end. Looking south along the avenue, the conditions are azimuth S. 8° 42′ W., hills 3° 30′. Both these avenues are aligned on points within, but _not_ at the centre of, the circle. _Challacombe_ (lat. 50° 36′).--This is a case of a triple avenue, probably the remains of eight rows, in a depression between two hills, Challacombe Down and Warrington. There is no circle. The azimuth is 23° 37′ N.W. or S.E., according to direction. The northern end has been destroyed by an old stream work; there is no blocking stone to the south on either of the remaining avenues, but one large menhir terminates one row of stones. The others may have been removed. So it is probable that the alignment was to the north. If so, we are dealing with the setting of Arcturus, warning the summer solstice sunrise in 1860 B.C. To the S. the hills are 4° 48′, to the N. 4° 50′. To this result some importance must be attached, first, because it brings us into presence of the cult of the solstitial year, secondly, because it shows us that the system most in vogue in Brittany was introduced in relation to that year. In Brittany, as I have before shown, the complicated alignments, there are 11 parallel rows at Le Ménac (p. 99) (there _were_ 8 parallel rows at Challacombe), were set up to watch the May and August sunrises, and the solstitial alignments came afterwards. The Brittany May alignments, therefore, were probably used long before 1860 B.C., the date we have found for Challacombe, where not the sunrise but the setting star which gave warning of it was observed. [Illustration: FIG. 45.--The remains of the eight rows of the Challacombe Avenue. Looking North of East. Terminal Menhir on the extreme right.]