Thus the shadow method, while difficult of application to give a true north-and-south horizontal line, would fail utterly to give material indications of the sun's elevation on particular days, without which it would be impossible to obtain in this manner any material indications of the position of the celestial pole.
A natural resource, under these circumstances--at least a natural resource for astronomers who could afford to adopt the plan--would be to build up masses of masonry, in which there should be tubular holes or tunnellings pointing in certain required directions. In one sense the contrivance would be clumsy, for a tunnelling once constructed, would not admit of any change of position, nor even allow of any save very limited changes in the direction of the line of view through them. In fact, the more effective a tunnelling would be in determining any particular direction, the less scope, of course, would it afford for any change in the direction of a line of sight along it. So that the astronomical architect would have to limit the use of this particular method to those cases in which great accuracy in obtaining a direction line and great rigidity in the material indication of that line's position were essential or at least exceedingly desirable. Again, in some cases presently to be noticed, he would require, not a tubing directed to some special fixed point in the sky, but an opening commanding some special range of view. Yet again it would be manifestly well for him to retain, whenever possible, the power of using the shadow method in observing the sun and moon; for this method in the case of bodies varying their position on the celestial sphere, not merely with respect to the cardinal points, would be of great value. Its value would be enhanced if the shadows could be formed by objects and received on surfaces holding a permanent position.
We begin to see some of the requirements of an astronomical building such as we have supposed the earlier observers to plan.
First, such a building must be large, to give suitable length to the direction lines, whether along edges of the building or along tubular passages or tunnellings within it. Secondly, it must be massive in order that these edges and passages might have the necessary stability and permanence. Thirdly, it must be of a form contributing to such stability, and as height above surrounding objects (even hills lying at considerable distances) would be a desirable feature, it would be proper to have the mass of masonry growing smaller from the base upwards. Fourthly, it must have its sides carefully oriented, so that it must have either a square or oblong base with two sides lying exactly north and south, and the other two lying exactly east and west. Fifthly, it must have the direction of the pole of the heavens either actually indicated by a tunnelling of some sort pointed directly polewards, or else inferable from a tunnelling pointing upon a suitable star close to the true pole of the heavens.