More recently, Prof. W. M. Flinders Petrie,[13] whose plan of the stones is a valuable contribution to the study of Stonehenge, was led by his measures of the orientation to a date very greatly in the opposite direction, but, owing to an error in his application of the change of obliquity, clearly a mistaken one. The chief astronomical evidence in favour of the solar temple theory lies in the fact that the “avenue,” as it is called, formed by two ancient earthen banks, extends for a considerable distance from the structure, in the general direction of the sunrise at the summer solstice, precisely in the same way as in Egypt a long avenue of sphinxes indicates the principal outlook of a temple. These earthen banks defining the avenue do not exist alone. As will be seen from the sketch plan (Fig. 15), there is a general common line of direction for the avenue and the principal axis of the structure; and the general design of the building, together with the position and shape of the naos, indicates a close connection of the whole temple structure with the direction of the avenue. There may have been other pylon and screen equivalents as in other ancient temples, which have disappeared, the object being to confine the illumination to a small part of the naos. There can be little doubt, also, that the temple was originally roofed in, and that the sun’s first ray, suddenly shining into the darkness, formed a fundamental part of the cultus. With regard to the question of the roof, however, the above suggestion, I now find, is not new, the view having been held by no less an authority than Dr. Thurnham, who apparently was led to it by the representations of the Scandinavian temples as covered and enclosed structures. Since the actual observation of sunrise was doubtless made within the sanctuary itself, we seem justified in taking the orientation of the axis to be the same as that of the avenue, and since in the present state of the S.W. trilithon the direction of the avenue can probably be determined with greater accuracy than that of the temple axis itself, the estimate of date must be based upon the orientation of the avenue. Further evidence will be given, however, to show that the direction of the axis of the temple, so far as it can now be determined, is sufficiently accordant with the direction of the avenue. The orientation of this avenue may be examined upon the same principles that have been found successful in the case of Greek and Egyptian temples--that is, on the assumption that Stonehenge was a solar temple, and that the greatest function took place at sunrise on the longest day of the year. This not only had a religious motive; it had also the economic value of marking officially and distinctly that time of the year and the beginning of an annual period.