ROOF. 



the ufe of the painter, there being a fpace of twelve feet 

 between it and the dome. This machine, it is faid, was 

 original of its kind. 



Roof, Rotative, or Rotatory, in AJlronomy, is one which 

 is made to turn round in a horizontal direction, fo that its 

 openings or doors may be direfted to any azimuthal line in 

 the concave expanfe of the heavenly regions. In a tranfit 

 room, where a meridian line is the only one that is wanted, 

 a fixed roof, with trap-doors opening at different altitudes 

 above one another, fufficiently anfwers the purpofe of the 

 aftronomcr ; but in an obfervatory, where an inftrument 

 that meafures azimuths as well as altitudes is placed ; or 

 where a good telefcope, of either the achromatic refracting, 

 or reflecting conilruftion, is ufed, for viewing the ttars and 

 planets in various parts of the heavens on the fame evening, 

 a rotative roof is indifpenfably requifite. The conftruftion 

 which has been molt generally adopted, and which we fhall 

 firll defcribe, was contrived by the ingenious Mr. Smeaton, 

 one of whofe papers, of the date of 1788, now lying before 

 us, was defigned for the late Mr. Aubert's obfervatory at 

 Highbury, and will fupply us with materials fufficient for 

 the purpofe of illuftration. In Plate XXII. of AJlronomical 

 Injlruments, Jg. 1. reprefents a feftion of the whole roof 

 along that plane, which pafl'es through the doors, or open- 

 ing, that divides it into two equal portions ; Jig. 2. is the 

 plan of the fame, and ferves to illuftrate^. 1 ; in both thefe, 

 a, a, are the rafters, or door-cheeks, covered by two oblong 

 doors meeting at the apex, and fo contrived by clofo fitting, 

 as to exclude the rain and fnow when nicely made. An en- 

 larged feftion acrofs one of thefe doors is feen in_^. 3, and 

 will be explained prefently. The frame of the roof is co- 

 vered with alternate boards of deal of a triangular fhape, 

 tongued into one another at the edges, and when perfectly 

 put together with ftrong glue, and feveral times painted 

 over, will be light, and will lalt feveral years ; but if a coat 

 of copper, particularly of tinned copper of Mr. Wyatt's 

 patent manufactory, be fuperadded, the weight will not be 

 confiderable, and the durability will be enfured. A fquare 

 foot of Mr. Wyatt's copper of the thinneft kind weighs a 

 pound, and therefore a given number of fquare feet will be 

 covered by fo many pounds weight. An edge plate, formed 

 into a circle, terminates the eaves, and conceals a ftrong an- 

 nular fupport, which rells on ten rollers, five of which are 

 feen in Jig. 1, three as circles, and two at the extremities of 

 the roof in feftion as fmall parallelograms. All the ten 

 rollers are reprefented in the ring feen in fg. 4, which is 

 united by five pairs of hinges that connect the five equal 

 portions thereof, as feen on the plane of the ring ; while 

 each of thefe five portions contain two rollers feen in feftion, 

 and one handle projecting inward?. The ufe of the hinges 

 is to allow each of the five portions of the ring to yield to 

 the preilurc of the fuperincumbent roof, in order that all 

 the rollers may be made to aft in every pofition of the roof, 

 in the event of any of the parts becoming diftorted or par- 

 tially worn. This ring, which receives the pivots of the 

 rollers, lies intermediate between the roof and fixed circle 

 that bears the whole weight, and is reprefented by the letters 

 b,l, mjg. 1, the fixed circle or bed being denoted by the 

 letters c, c ; both which, in the feftion, appear as ftraight 

 lines, though they are circles or rings, as above defcribed, 

 and the fixed one, c c, reds on ftrong upright pieces, d, d, 

 properly braced by frame-work, which it is not neceffary to 

 (hew. From this fhort defcription the reader will now con- 

 ceive that the roof is feparated from the bed c c, by exaftly 

 the whole diameter of the rollers, which therefore ought to 

 be turned nicely to the fame dimenfions, in order to preferve 

 the horizontal pofition of the ring, and to make the rollers 



aft fmoothly, when the roof is pufhed round by any force 

 internally applied ; but unlefs fome provifion is made, more 

 than we have yet defcribed, the roof might be pufhed from 

 its bed, and be upfet ; to avoid this accident a groove is 

 made along the upper face of the annular bed c c, in which 

 the rollers move, and while all the parts of this contrivance 

 remain unaltered by the weather, the aftion of the rollers is 

 uniform, and the conftruftion anfwers its purpofe in a fatif- 

 faftory manner : but in many inftances, we learn that the 

 work of turning a large roof, thus conftrufted, becomes 

 laborious, in confequence of either the ring of rollers, or 

 the groove in the bed, taking an elliptical fhape, in a 

 fmall degree, by partial fhrinking or fwelling in different 

 ftates of the atmofphere ; in which cafe the rollers bind 

 in the groove, and require mechanical force to move them, 

 in certain parts of the revolution ; which force will of 

 courfe ilrain the joints, and ultimately deltroy the union of 

 the different parts of the ftrufture. It is, therefore, of im- 

 portance that an attention be paid to the grain of the wood, 

 as well as to its being feafoned, before the upper covering 

 be attached ; which precaution will prevent the bad effeft 

 of external moifture. At e, iu Jg. 3. is one of the hinges 

 of the doors, attached firit to the fide of the door, and then 

 to a piece of wood that lies over the roof to receive the 

 fcrews, without penetrating the roof itfelf; and the crank- 

 fhape of the feftion of the door will explain how it clafps 

 the door-cheeks, fo as to exclude the admiffion of rain or 

 fnow, fo long as the materials do not warp or decay. The 

 handle-piece, f, pufhes the door back, and brings it clofe 

 again, by the aid of a ftaff with a hook at the end, that 

 takes hold of the ring at the lower extremity of the 

 handle f, when the roof is too far elevated for the reach of 

 the human hand. During the perufal of this account of 

 the mechanifm of Smeaton's rotatory roof, it may have oc- 

 curred to the reader, that the rollers might have been at- 

 tached to the annular portion of the roof, which now refts 

 on the upper extremities of the rollers, and that confe- 

 quently the ring, that now carries the faid rollers, and that 

 is reprefented by Jg. 4, might have been omitted. But the 

 inventor had a good praftical reafon for the introduftion of 

 this ring, which may not yet have occurred to our reader ; 

 and it might perhaps puzzle him to difcover it, without our 

 afiiftance. Smeaton knew very well that the roof with 

 rollers, or fmall wheels under it, would move in a groove' 

 made in the bed to receive them ; but he knew, moreover, 

 that, to move in a circular groove, the diameters of the 

 rollers could not be great, without (licking fait againft the 

 fides of the faid groove ; and, alfo, that the pivots of the 

 rollers mull be Jlrong, to fupport the whole roof. The re- 

 lative diameters of the rollers, and of their pivots, on which 

 the quantity of friftion depends, would not consequently 

 produce, under thefe circumftances, the eafy motion of the 

 roof that conilitutes a leading objeft of the altronomer : the 

 ring was, therefore, introduced, for the purpofe of di- 

 minifliiniT the friftion one-half, without altering the neceffary 

 dimenfions of the rollers, and of their pivots ; which effect 

 may he thus explained. In the firll place, fuppofing the 

 roof to be moving forwards on rollers attached to it, like a 

 carriage on wheels ; in this cafe, the velocity of the roof 

 and of the rollers, or of the carriage and wheels, would be 

 the fame ; and the horizontal diitance pailed over in a given 

 number of revolutions of the roller, or wheels, would be 

 meafured by the circumference of the roller, or wheel, mul- 

 tiplied by its number of revolutions ; and the mechanical 

 diminution of friftion would be as the diameter of the roller 

 to that of the pivot of the axio, or of the wheel to that of 

 its axle. When a motion is thus produced in a roof, or in 



a carriage, 



