1873.] in determining the best Course for a Ship. 273 



priate to the occasion, most likely out of a stock which he would keep 

 ready to hand ; and applying it to his map, he would note the direction in 

 which it will cut the largest number of isochrones towards the port. If 

 we take cases where the isochrones for the occasion are (1) N MB and 

 (2) WB, the first being with a N. and the second with a W. wind, 

 then, if the isochrones towards the port be as in the series h v Jc 2 , &c, 

 the course to be steered in both instances is A B .; but if they are as in 

 the series i v i 2 , &c, the course in case (1) will be A N, and in case (2) 

 A W, quite independently of the direction in which the port may happen 

 to lie. 



There is much to say about the proper method of discussing the crude 

 statistics derived from ocean districts artificially bounded by lines of 

 latitude and longitude in order to obtain the most probable meteoro- 

 logical values, but I will only allude to them here. First, homogeneous 

 districts and periods of time have to be made out ; secondly, the crude 

 observations in each subdivision of those districts have to be dis- 

 cussed in connexion with those made at adjacent subdivisions in the 

 same district ; and, thirdly, they have to be discussed in reference to 

 those made in preceding and succeeding periods of time. There is no 

 doubt but that labour spent in these discussions would after a time 

 become more remunerative than the same amount of labour in accumu- 

 lating fresh observations. 



I submit (fig. 6) a series of 8-hour isochrones computed to 8 points from the 



Eig. 6. 



crude observations taken in each " 2-degree square " of the ocean between 



