From G. H. Darwin to Leonard Darwin 29 October 1872

Dear Pouts,

I think I’ve hit on the contrivance for our machine. Let one wheel of survey^s. mach. be fastened to screw shaft & the other to the clock; on the indicating part fix a wheel or disc—thus with a no. of pins on it & let *P [DIAGRAM HERE] *Q these pins be metal on one side & wood or bone on the other. Take two pieces of watch spring & fix them in the direction of two radii of this disc, so that the disc in turning round (either way) just catches *P [DIAGRAM HERE] *Q the tips & passes by deflecting them. Then as the wheel goes one way there will be metallic connection & as it goes the other there will not—for in one case metal comes on metal & the other against wood. Let the springs be so arranged that when one is just deflected to the slipping point the other has just begun to be deflected, then as the wheel goes in direction of metal v. metal there will always be metallic contact & of course when it goes in the other direction there will never be so. Let the metallic contact or either spring & a peg complete the circuit of the electro-magnet & shut off the Steam. It is clear that if the screw has been racing & then goes at exactly the rate of the clock the circuit will be complete until the rate has fallen below the rate of the clock, & if the screw has been going slower than the clock that the contact will not be made until the Screw is going faster than the clock. Thus then is a small want of sensitiveness, but not of a bad kind for we may make the clock go a little slower than the screw is required to go. We might easily have 60 pins—then if excess of 4 turns of screw over that of clock makes one turn of [illeg] disc 15° of excess of turn of screw over that of clock corresponds to distance between two pegs—so that the greatest poss. delay in mak^g the contact