In connection with the possible astronomical uses of these temples, I find that when one of the temples has been built, the horizon has always been very carefully left open; there has always been a possibility of vision along the collimating axis prolonged. Lines of sphinxes have been broken to ensure this;[44] at Medînet-Habû, on the opposite side of the river to Karnak, we have outside this great temple a model of a Syrian fort. If we prolong the line of the temple from the middle of the Naos through the systems of pylons, we find that in the model of the fort an opening was left, so that the vision from the sanctuary of the temple was left absolutely free to command the horizon. It may be said that that cannot be true of Karnak, because we see on the general plan that one of the temples, with an azimuth of 72½°N., had its collimating axis blocked by numerous buildings. That is true; but when one comes to examine into the date of these buildings, as I propose to do in a subsequent chapter, it is found that they are all very late; whereas there is evidence that the temple in question was one of the first, if not the very first, of the temples built at Thebes. * * * * * (3) To determine the declinations to which the various amplitudes correspond. In this direction I have made use of the German Catalogue of star places from 1800 A.D. to 2000 B.C., the places for dates beyond this, and for southern stars, having been calculated chiefly by my son, Mr. W. J. S. Lockyer, B.A. Some places for Sirius and Canopus have been obligingly placed at my disposal by Mr. Hind, and approximate values obtained by the use of a precessional globe constructed for me by Mr. Newton. This globe differs considerably from that previously contrived by M. Biot, about which I was ignorant when I began the work, and enables right ascensions and declinations, but especially the latter, to be determined with a fair amount of accuracy for forty-eight equidistant points occupied by the pole of the equator round the pole of the ecliptic (assumed to be fixed) in the precessional revolution. Some simple astronomical considerations may here come to our help. If the north polar distance of a star is increasing--that is, if a star is increasing its distance from the north pole--its declination if north or south will be decreased or increased respectively, and the orientation of the temple would be gradually becoming more and more parallel to an E. and W. line; if the declination north or south of the star be increasing, then the orientation of the temple would have to be likewise increased. The change in the orientation, therefore, gives us information towards determining in which quarter of the heavens each particular star might have been. * * * * *