I append a reproduction of one of M. Gaillard’s plans, which will give an idea of his use of the index menhir. It shows the alignments at Le Ménec, lat. 47¹⁄₂° (Fig. 26). The line A--Soleil runs across the stone alignments and is fixed from A by the menhir B, but there does not seem any good reason for selecting B except that it appears to fall in the line of the solstitial azimuth according to M. Gaillard. But if we take this azimuth as N. 54° E., then we find the alignments to have an azimuth roughly of N. 66° E., which gives us the amplitude of 24° N. marking the place of sunrise at the beginning of the May and November years, and the alignments may have dealt principally with those times of the year. I esteem it a most fortunate thing that while I have been casting about as to the best way of getting more accurate data, Lieutenant Devoir, of the French Navy and therefore fully equipped with all the astronomical knowledge necessary; who resides at Brest and has been studying the prehistoric monuments in his neighbourhood for many years, has been good enough to give me the results of his work in that region, in which the problems seem to be simpler than further south; for while in the vicinity of Carnac the menhirs were erected in groups numbering five or six thousand, near Brest, lat. 48¹⁄₂°, they are much more restricted in number. I am much indebted to him for permission to use and publish his results. [Illustration: FIG. 26.--Alignments at Le Ménec.] Lieutenant Devoir, by his many well-planned and approximately accurate observations, has put the solstitial orientation beyond question, and, further, has made important observations which prove that the May and August sunrises were also provided for in the systems of alignments. I give the following extracts from his letter:-- [Illustration: FIG. 27.--Menhir (A) on Melon Island.] “It is about twelve years ago that I remarked in the west part of the Department of Morbihan (near Lorient) the parallelism of the lines marked out by monuments of all sorts, and frequently oriented to the N.E., or rather between N. 50° E. and N. 55° E. I had ascertained, moreover, the existence of lines perpendicular to the first named, the right angle being very well measured. “The plans, which refer to the cantons of Ploudalmézeau and of St. Renan (district of Brest) and of Crozon (district of Chateaulin), have been made on a plane-table; the orientations are exact to one or two degrees.