ROOF. 



a carriage, we may fay that it rides on, or is carried by its 

 roller, or wheel, and that with the fame velocity with which 

 the roller or wheel moves forwards, by virtue of its rota- 

 tions. But, in the next place, let the rollers be confidered 

 as attached to the ring in fig. 4, which has as many holes 

 cut through it as there are rollers for them to pafs through ; 

 and let us conceive this ring to be put in motion, as it is 

 placed on its circular bed cc, without any reference to the 

 roof. I;i this fituation, the ring rides on, or is carried by 

 the rollers with their velocity, as the roof or carriage was 

 on our firll luppofition. Let now the roof reft on wedgec 

 placed on the ring, but fo a6 net to touch the rollers, and it 

 will alfo ride as before, with the velocity of the rollers ; 

 but remove the wedges, and let the roof fall upon the upper 

 edges of the rollers, and it will now have two motions com- 

 bired : for, firil, it will ride in common with the ring on 

 which it reds, in the way we have defcribed ; and, fe- 

 condly, it will be urged by the rollers acting on its annular 

 face, as fo many toothlefs pinions, with a fimilar velocity 

 with which it rides on the rollers, along with the ring ; or, 

 in other words, the roof will have double the "velocity of the 

 ring on which the rollers are fixed. To undertland this ef- 

 fect more clearly, let the reader fuppofe himfelf in an open 

 carriage. While he fits ftill, he is carried with the velocity 

 of his vehicle ; but if he fhould venture to place his foot on 

 the circumference of one of the wheels, in getting out, 

 while the vehicle is in motion, he will run the rifle of being 

 tolled forward by the faid moving wheel with an increnfed 

 •velocity, that will convince him of the danger of inch an ex- 

 periment. This effect is familiar to all, but the caufe is not 

 generally confidered, perhaps but little underftood. 



In order that the adaptation of all the parts that are con- 

 nected with the rollers may be the more evident, we have 

 given an enlarged feclion of one corner of the roof, near the 

 eaves, within the ring of Jig. 4, which, therefore, may be 

 confidered as a feparate figure, in which the fame letters 

 reprefent the fame parts as in fig. 1 : namely, a is one of 

 the rafters, forming one of the door-cheeks, projecting 

 above the covering of the roof; b b, the roller, feen both in 

 feCtion and plan ; c, a feftion of the fixed annular bed ; g, a 

 fedtion of the moveable annular part of the roof, on which 

 the eaves lie ; and g b, an iron cranked bar, fcrewed to g 

 above, and therefore moving with the roof, but in fuch 

 way, that the crooked part at b lies jull below an annular 

 piece i, attached to the bed c ; fo that it' any blaft of wind, 

 or other external force, fhould move the root, it will be fe- 

 cured from ovcrfetting, by the crank-bar at h catching the 

 faid fixed piece i ; and it four or more of thefe crank-bars 

 be ufed, the roof will be fee u re in all directions. 



The objection which we have Itated, as applying to the 

 Cont'..uCtion of Smeaton's roof, lias very lately been ob- 

 viated by the Rev. \V. Pearfon of Kail Sheen, Surrey, 

 who hasjull finifhed a rotative dome, to receive a large and 

 very accurate circular inltrument for mealuring azimuths, 

 as well as altitudes, by Troughton ; which inllruinent was 

 originally ordered by the Ruffian Academy, bul was eoun- 

 termanded bv realm, . ommitted by tin- French 



in Rufiia, at the time that the conftru&ton "I it was in 

 hand. T: dome ,- 5, 



whi re the circular rafters are teen at fix. 



.live ot the opening, a a, ad, which i« nearly nine 



;, through which the I I "t the inltrument, 



placed on the tup of the pillar A, is direct' d to a:. \ put I 

 the heavens. This f •micircular opening, which divides the 

 dome into tWO equal portions, or quarter:, i.t an <>l. 

 fphere, is covered by five doors, two on eacli tide, and one 

 at the fummit ; but in fuch a vay, that a fmall door on one 



fide faces a large one on the other, and the top door takes 

 in five or fix inches more on one fide than on the other, fo 

 that there is no altitude to which the telefcope may not be 

 direfted, by reverfing the pofition of the dome, i. e. by 

 turning it juft half round, when necefiary ; and yet the 

 fittings are fo contrived, that any one of the doors may be 

 opened, while the ret! are (hut. The two quarters of the 

 oblong fphere are connected by brafs bars at the points 

 where the doors meet, and grooves are made on the edges 

 of the rafters, that form the door-cheeks ; fo that any rain, 

 that may be driven into the tides or ends of the doors, runs 

 down thefe grooves to the eaves, and difappears, without 

 being admitted within. We have not thought it necefiary 

 to (hew thefe doors, which turn back on hinges, placed at 

 their edges, upon the roof, which is covered firft with tri- 

 angular flips of deal about three-eighths thick, and then 

 witli tinned copper from Wyatt's manufactory ; which 

 metal being eafily feamed, requires not the aid of many 

 nails. The ornament at the top, being light, turns back 

 with the top-door, and lies alfo on the roof, while obferva- 

 tions are made in or near the zenith. The rotative part of 

 this roof is of the fimpleit conftruction that can well be 

 imagined, and yet is fafe, and more free from friition than 

 Smeaton's, though made with the utmotl care. The ring 

 of rollers is here entirely left out, and three large balls of 

 lignum vita; are fubftituted in a detached tlate, that is, 

 without any fixed axis of motion, or pivots ; and a four- 

 fided circular tunnel is fo formed of wood, that thefe balls, 

 by their fimilarity of dimenlions, keep their relative dittances 

 from one another, and aft as friction-rollers both in a ver- 

 tical and horizontal direction ; nay, even in both at once, 

 whenever they happen to come in contact with cither of the 

 fides of the faid tunnel. The balls, which are 4^ inches in 

 diameter each, are (hewn in fit;. 5. at oppolite diameters of 

 the dome, near the eaves, a> though four were uted, for the 

 fake of exhibiting their mode cf pofition and aCtion ; but 

 three only are much better, becaufe in every lituation of 

 their revolution they are certain to bear equal (hares of the 

 whole weight, when the bed is horizontal ; and there- 

 fore none of them will refufe to move, when the dome is 

 pufhed round ; but fhould they happen to gain ground 

 of one another in moving forwards, by any difference in their 

 diameters, they will refume their original fituations refpec- 

 tively during the backward motion of the dome. Fig. 6. 

 is an enlarged fcale of the tunnel and of one of the balls, in 

 which A is a portion of the roof, terminated by the annular 

 piece of deal B, which is lined with the annular piece ol 

 wainfcot C, to both which the covering-piece D above 

 til-- eaves is fcrewed ; E is a piece of wainfcot forming the 

 inner fide of the tunnel, and G is a continuation of one of 

 the upright planks attached to the annular bed H, and forms 

 the exterior fide of the fame. All the four fides ot the 

 fquare tunnel, which we called circular, becaufe it iurrounds 

 the circular erection, are either of waintcot, or lined witli 

 wainfcot, to come in contaift with the lignum vita: which 

 compstfes the balls j though the rings of deal, Band I. might 



have been in contact with the balls it they had been hard 

 enough ; but the W< ight of a dome, of nearly ten teet dia- 

 meter, covered with copper, would, in that cite, have pro- 

 duced hollow places, thai 1 ight have impeded the unilorm 



motion of the balls. The piece J is one ol the fupporters 

 of the Ixd I, to which the deal planki G are 1 nd 



alfo the external pilaiterr.tli.it 01 am il the . tion 



of the ftrudture. 1" this dorai the telocit) ■■'■ the fu] 

 llruCture, as in Smet the balls 1 



and at the! are left at liberty to turn m ai ion, where 



an obftacle to motion is pi •tented to any ot them, they in- 

 y llautly 



