I88b\] The Harmonic Analyser at the Meteorological Office. 383 



mainly worked out by Prof. Stokes and Mr. de la Rue. Plans were 

 obtained from two firms of mechanical engineers, and those of 

 Mr. Munro being ultimately adopted, the construction of the instru- 

 ment was entrusted to him. It was considered sufficient to limit the 

 action of the machine so as to extend only to the determination of 

 the mean, and the coefficients as far as those of the third order, in the 

 expression 



E=a+a ] costf + ^i sin + <x 2 cos 2# + & 2 sin2#-|-&3 cos 39 

 + 6 3 sin 'SO + &c, 



and to obtain these it was necessary to have seven sets of spheres, 

 disks, and cylinders. 



A description of the machine, as actually constructed, was published 

 in "Engineering" for December 17th, 1880, and we are indebted to 

 the proprietors of that journal for permission to reproduce the 

 engravings which illustrate that description, as well as a portion of 

 the text, which we now proceed to quote : — : 



"The machine is shown in the accompanying engravings, figs. 1 to 

 3 ; figs. 2 and 3 showing details of the arrangements of the ball, disk, 

 and cylinder. In principle it is, of course, precisely similar to its 

 predecessor — differing from it only in constructive details intended to 

 secure stability, and accuracy in its movements. Instead of being 

 largely made of wood, as was the case with the model, it is entirely 

 of metal ; the cast-iron frames carrying the disks being secured to a 

 firm iron bed supported by two substantial uprights ; the disks them- 

 selves are of gun-metal, and the spheres of steel carefully turned, and 

 nickel-plated to prevent rusting; the horizontal bar carrying the forts 

 for moving the spheres is also of steel and plated, and is carried above 

 the disks upon five iron uprights or guides. The forks are provided 

 with adjusting screws allowing of very accurate centering of the 

 spheres upon the disks, and adjusting screws are likewise provided 

 for the frames carrying the recording cylinders, by which their 

 parallelism to the faces of the disks can be rigidly secured. The 

 spheres are not touched by the forks themselves, but by the flat faces 

 of two screws passing through their lower extremities, and in this 

 way a ready means of preventing looseness or tightness of the spheres 

 in the forks is provided. 



" Each recording cylinder has attached to it a counter for register- 

 ing the number of its complete revolutions, and to secure a maximum 

 of freedom in their movements the spindles of the cylinders, as well 

 as the slides carrying the racks for giving motion to the disks, and the 

 horizontal steel bar, are all made to run upon friction rollers; the slides 

 have in addition counterpoises attached to them to prevent error from 

 backlash. 



" The motion of the shaft at the rear of the machine is communi- 



VOL. XL. 2 D 



