"'That a star should act as time-warner it was necessary that it should have so nearly the same amplitude as the sun that it could be seen from the adytum through the eastern door, if it was to give warning at its rising, or to have a similar but reversed amplitude towards the west, if its heliacal setting was to be observed; and it follows that in the choice of the festival day and the corresponding orientation, on these principles, both the amplitude of the sun at its rising and that of the star eastwards or westwards, as the case might be, would have to be considered in connection with one another. "'From what has been said it is obvious that in the intra-solstitial temples the list of available bright stars and constellations is in the first instance limited to those which lie within a few degrees of the ecliptic, and it will be found that in the list above given and those which follow, if we omit Eleusis, where the conditions were exceptional, all but one of the stars are found in the zodiacal constellations. A very great limit is imposed, in the second place, by one of the conditions being the heliacal rising or setting of those stars from which the selection has to be made. So that, when both these combined limitations are taken into account, it becomes improbable to the greatest degree that in every instance of intra-solstitial temples of early foundation of which I have accurate particulars, being twenty-eight in number and varying in their orientation from 21° N. to 18° 25′ S. of the true east, there should be found a bright heliacal star or constellation in the right position at dates not in themselves improbable unless the temples had been so oriented as to secure this combination.' "I have just been looking into the number of possible stars which could have been used, _i.e._ within the limits of the greatest distance from the ecliptic that could have been utilised. "The stars which could have been utilised in addition to the seven which serve for nearly thirty temples are ten only, viz.:-- Aldebaran. Pollux. β Arietis. β Tauri. α and β Capricorni as a group. β Libræ. α Libræ. α Leonis. γ Leonis. β Leonis. "If the orientations had been placed at random, would not our thirty temples have made many misses in aiming at these seventeen stars, it being necessary also to hit exactly the heliacal margin? And would they have secured anything like a due archæological sequence? * * * * * "Another point is this:-- "Whenever a star less than first magnitude is used (Pleiades only excepted) it has been necessary, to secure coincidence, to give it several more degrees of sun depression than in the cases of Spica and Antares."