the Mariner'' s Compafs is fi/fpended. 427 



the vertex of tHe conical cavity. It is neceffary that it Qiould be above the cent|:e of gra- 

 vity ^ but this dillancc muft be fo fmall as that the libration of the needle whcA.pne end-is 

 deprefled fhall be very flow, and yet fpeedy enough to recover the horizontal portion in a 

 reafonably fliort time. In fadt, the whole of the lleadinefs of the compafs and its box ap- 

 pears to depend on this principle of flow vibration. For, if a needle perform its vertical 

 vibration in eight feconds, it will be very little difturbed by an alternate action that lafts but 

 a fecond or two. 



The greater number of workmen have imagined that the agitation of the compafs is com- 

 "mumcated by fri£lion at the points or edges of fufpenfion, and have accordingly exerted 

 their ingenuity to diminifla this friftion, by contrivances fimilar to that of a conical cap 

 balanced on a point, and itfelf affording another point to fupportthe needle. But it is very, 

 readily pi'oved by experiment, as well as argument, that the greatefl difturbance of the' 

 needle is produced by the quantity pf horizontal progreflive motion, and not by the merc; 

 inclination or angular motion. A corripafs-needle fupportcd on a fimple point will fufler. 

 very little agitation from any angular motion or moderate deviation from perpendicularity 

 in the pin ; but it will inllantly begin to vibrate if moved horizontally. Thus the common 

 experiment of tilting the compafs-box in all pofitions while its centre remains immoveable 

 is fallacious, and there are very few compafTes indeed which will bear to be Aided back- 

 wards and forwards upon a table. 



It appears therefore that the fteadinefs of a needle which vibrates flowly. Is the confc- 

 quence not only of the length of time it allows for alternate aflions to operate, and deftroy 

 each other ; but alfo of the difficulty with which it yields to fuch imfjreflions. If the centres 

 of fufpenfion and of gravity in the needle were coincident, no angular motion would be 

 produced by any aiSlion of the pin, excepting by the efFe£ls of fritStion ; and the angular 

 motion produced in other cafes will be lefs, the fliorter the diftance between thefe two 

 centres, or the lever by which it is propagated. 



The fimple fufpenfion of the needle on a point has been applied to the compafs-box, for 

 which it is little fuited, not only becaufe of the wear upon fo fmall a furface, but alfo becaufe 

 it admits the box to traverfc horizontally; an efteifl which is inconvenient, and cannot be 

 remedied by any means not calculated in fome refpe£t to increafe the effedls of agitation. 

 The method moft generally received, and in fa£t the befl adapted to this inftrument, arc the 

 gimbals. 



This well-known contrivance confiftsof an hoop fupported upon two pins diametrically 

 oppofite each other, and iffuing from the external furface of the ring in fucha direftion that 

 both lie in the fame diametrical line. When the hoop is fufpended on thefc pins, it is at 

 liberty to turn freely round the diameter of which they conftitute the prolongation. The 

 notches or holes of fupport are difpofed horizontally. The compafs-box itfelf is placed in 

 a Gmllar ring with two projecting pivots ; and thefe pivots arc iiifertcd in holes made in tlic 

 former ring at an equal diftance from each of its pivots. If, therefore, we fuppofc the 

 whole to be left at liberty, the compafs-box may vibrate upon the diametral line of the 

 outer ring, and alfo upon a line formed by its own pivots, at right angle? to that diametral 

 line. The confe(iy,ciici; of this arrangement is, that the centre of gravity of the compafs- 

 box wiU difpnfe itfelf immediately beneath the intcrfcftion of both lines on which it js at 



3 I 2 libctty 



