Before the original workers abandoned their work they seem to have made several attempts to set out a slightly reduced obelisk, which would avoid all serious cracks by reducing the length and thickness of the original design. This is seen in the lines γ δ ε ζ and the transverse lines ι κ λ μ. The last four lines are so faint that they can only be seen just after sunrise or before sunset, and it is not clear with which of the longitudinal lines they connect. On the south side the lines are quite clear, but on the north side there seem to have been more lines even than those shewn on plate I. These lines γ δ ε ζ, do not lie at equal distances from either of the two centre lines η and θ. There is no doubt in my mind that the original obelisk was to have had a straight-edged pyramidion, as the rough edge of the boundary trench lies evenly on either side of the centre line η. (10) The outline for another, and most probably later, obelisk is set out on a new centre line θ, and keeps closely to the north edge of the original design, avoiding the series of fissures on the south. Just before sunset, the tentative outlines for the curved pyramidion can be traced; plate III, no. 3, taken at that time, shews these lines. In this, the right-hand curve appears to engage with a line to the north of that engaged by the curve next to it; this is only an effect of light and they really engage as shewn on plate I. I have outlined the original design and the later scheme, in red. Though the lines setting out these obelisks are easily traceable, the colouring is mine, and is only intended to shew up the lines more clearly. Since there are some actual red lines on the obelisk, another colour would have been preferable; considerations of cost in printing have limited the plate to two colours. In outlining the later design in red, I have chosen the larger pyramidion, as the shape decided on. There is no proof of this, but the proportions are, to me, more effective, and there is no reason for abandoning the odd metre of difference, as the stone here is perfectly sound. Taking the longer pyramidion, we have the following dimensions for the obelisk: metres. Total length 32.10 Pyramidion height 5.31 Pyramidion base 2.02 Obelisk base 3.15 {9} (11) I had intended to give a diagram of the better-known obelisks superimposed. I found it, however, almost impossible to get the sides of the obelisks distinct one from another without making the scale inconveniently large. I give, therefore, a table shewing the principal dimensions of ten examples. Those marked with an asterisk are scaled off photographs, making slight allowances for foreshortening.