488 Scientific Proceedings, Royal Dublin Society. 



Some of the earliest successful domes were made on this prin- 

 ciple, and worked very much better than some of the elaborate 

 systems of wheels or rollers. Figs. 5« and 56 show a slight 

 modification of fig. 5. 55 is the best ; because even if the dome 

 slightly alters its form, it does not compel the ball to roll up 

 one side of the groove, as in the other case ; and if this form was 

 carefully made there is no reason why it should not work well for 

 moderate sized domes. When, however, it is necessary from the 

 size of the dome to employ more than three balls one is apt to get 

 loose occasionally and run out of its place and give much trouble 

 to replace it. 



This form of revolving gear got into discredit mainly because 

 the wall-plates and sole-plates of domes were generally made of 

 wood, and when the dome remained idle for many months the balls 

 sunk into the wood, and left hollows here and there which soon 

 reduced the efl&ciency. In order to obviate this, the form fig. 6 

 was proposed and carried out in many cases. This consisted of a 

 set (any number according to size of dome) of rollers rolling on a 

 cast-iron wall-plate, and on which rolled a cast iron sole-plate 

 attached to the dome. These rollers were coupled together by a 

 light framework to prevent any danger of their getting displaced. 

 This form is better in theory than in practice. If the rollers be 

 made wide, there is the loss from twisting spoken of in referring 

 to figs. 2 and 3. Consequently the rollers must be made narrow, 

 or at least bearing only toward their centres ; and then it is found 

 that wind pressure has a great effect on the dome, for the rollers 

 have no base, and rock on the narrow central bearing. This causes 

 strain on the connecting framework, and the system has not been 

 found to work as well as might be expected, and has been little 

 used. 



Somewhere about the year 1870 my father, desiring to utilize 

 to the full the undoubted advantages of the live roller system, 

 designed the form fig. 7. In this case each roller consists of 

 three parts, but except for sake of convenience in manufacture, 

 these might all be cast and made in one piece, as they are bolted 

 on one spindle and revolve together. An outer taper roller, and 

 an inner taper and grooved roller revolve on two projecting rails 

 on a cast-iron wall-plate, and a planed cast-iron rail on dome 

 rolls on the centre roller of the three. By planing the rails and 



