The only remaining point, in which the example furnished by the Eddystone and Bell Rock Lighthouses has been at all materially departed from, is (as has already been hinted at by an unavoidable anticipation) the mode of uniting the different parts of the masonry together. In both these Towers the stones were dovetailed throughout the buildings, chiefly (at least in the case of the Bell Rock where the foundation was so much below the tide) with the view of preventing the sea from washing away the courses which might be left exposed to the winter storms before the weight of the superstructure had been brought to bear upon them. In the upper part of the Bell Rock my father also introduced a kind of band joggle, which consists of a flat ribband of stone raised upon the upper bed of one course, and fitting into a corresponding groove cut in the under bed of the course above; and this system of tying the adjoining courses together also forms a chief feature in his design for a Lighthouse on the Wolf Rock.[16] When the great pressure of the superstructure of these Towers, however, and the effect of the mortar are considered, there seems little probability of one course being dislocated, in defiance of the friction resulting from the weight of the column. An impulse sufficient to produce such an effect would tend to overset the whole superstructure from mere deficiency in weight, and in this case the joggle would have little effect. But if joggles be thought necessary for this purpose, the ribband form certainly produces a better arrangement than that of the cubic joggles employed by Smeaton for connecting the adjoining courses of his building together, as the sectional strength of these scattered square joggles is very small compared to the effect of a shock which could be supposed capable of moving the whole mass of a Tower. In the lower parts of the Skerryvore Tower, I entirely dispensed with dovetailing and _joggles between the courses_, and thus avoided much expensive dressing of materials. The stones were retained in their places during the early progress of the work, chiefly by common diamond joggles, and the courses were temporarily united to each other by wooden treenails, like those used in the Eddystone and Bell Rock. These treenails had split ends, with small wedges of hardwood loosely inserted, which being forced against the bottom of the holes in the course below, into which the treenails were driven, expanded their lower ends until they pressed against the sides of the holes; while their tops were made tight by similar wedges driven into them with a mallet. I have, however, adopted the ribband-joggle in the higher part of the Tower, where the walls begin to get thin in the very same manner as at the Bell Rock, where it was used, partly that it might counteract any tendency to a _spreading_ outwards of the stones, and partly that it might operate as a kind of _false joint_ to exclude the water which, when pressed with great