1873.] 



in determining the best Course for a Ship. 



269 



tion of the isochrone. If A, B, and C be the probable errors of these three 

 respectively, then the probable error of the navigator's estimate will be 

 VA 2 4-B 2 -fC 2 . My desire is to reduce this large item to a simple A, 

 which may itself be minimized until it ceases to be of any practical in- 

 convenience. It is no new thing that statistics should require discussion 

 and elaborate calculation before they can be turned to account and be 

 made the familiar basis of vast commercial undertakings. Classified lists 

 of ages at death are no more fitted to be the immediate guides of those 

 who grant annuities or engage to pay reversions, than are the crude 

 meteorological statistics of the ocean to be the immediate guides of the 

 navigator. 



It is probable that most vessels may practically admit of division into 

 some moderate number of classes, and that it would suffice to calculate 

 isochrones for each of these classes ; but in any case the number of cal- 

 culations must be very large, because they would differ not only for the 

 class of ship and the particular destination, but also for the season of the 

 year when the voyage was made. It is therefore important that even 

 individual ships should be enabled to have isochrones drawn for their 

 especial use at a trifling cost. I will now show how this may be effected 

 by mechanical means. 



The drawing I give (fig. 3) is only a diagram to explain the principle of 

 the machine ; the framework is left out, and the proportions are somewhat 

 varied for the convenience of illustration. Also, for simplicity of explana- 

 tion, I have supposed the machine constructed to apply to no more than 

 4 points of the compass, and it is represented as adjusted to work out the 

 example already given in Table III. 



A long tray, open at the top and almost wholly open along the front, 

 has 8 grooves, into which pieces of zinc, thin wood, or even stout card- 

 board may be dropped, much like the glass plates in a photographer's 

 box. In fig. 3 we only see the base of this tray, A A, and the pieces of 

 zinc standing upon it in the position in which they would be held by the 

 grooves. The zinc plates have curved edges in front, which refer to the 

 sailing qualities of the particular ship under consideration ; these are cut 

 out from the data in Table II., each plate corresponding to the column 

 whose heading it bears. The ordinate of the curve is proportionate to 

 the force of the wind, and the abscissa to the distance sailed in one day 

 on the specified course (0, 8, 16, or 24) with that force of wind. There 

 are grooves cut in A A, one under each zinc plate, and there are tramwavs, 

 T, upon which A A may be set, in gear with those grooves. In the figure 

 it is so set that is opposite to n, $ to e, 16 to s, and 24 to w ; but if it 

 were lifted up and laid one groove more to the left, (on the right hand) 

 would be opposite to w, 8 to n, 16 to e, and 24 to s. Similarly, by setting 

 it two or three grooves to the left, the other possible variations would be 

 gone through. The slides n, e, s, and w refer to the course of the ship 



