AND NEIGHBOURING SEAS DURING 1897. 217 



5 

 daily at differeut stations, and the factor then becomes simply - ; or 



-rx 5 X Eesultant Pressure , • , - ,, , r ■^ j -n. j 



D = , which gives us the number oi miles drifted. 



The accuracy of tliis equivalent can only be determined by the 

 frequency with which its results accord with those of direct experiment. 

 It depends On an assumption which may invite criticism, viz., that the 

 velocity of the drift varies as the pressure of the wind, and not directly 

 as its velocity. This pressure-ratio is confessedly only an approximation 

 to the true law of drift, but I consider it to be nearer the truth than the 

 velocity-ratio adopted by Mohn, which assumes the existence of a level 

 sea for winds of all velocities. 



§ 6. Application of Factor to cases of Drift. — The use of drift-bottles 

 for estimating the velocity of drift is attended by a possible source of 

 serious error, viz., the difficulty of determining the length of time 

 between the actual stranding of a bottle and its subsequent recovery. 

 In some cases this source of error is removed by the statements made 

 by the finders of the bottles, but in most cases it must always remain 

 as a condition to be taken into account. On the whole, however, 

 I believe that the maximum error due to this cause is inconsiderable 

 except when the bottles have been recovered in unfrequented parts 

 of the coast. Most bottles are picked up in the neighbourhood of 

 towns and fishing ports where the shore is much frequented, and in 

 other regions the coastguardsmen are always on the look-out. The 

 conspicuousness of our bottles must also tend to reduce the error due 

 to this cause. Altogether I should estimate the average error as amount- 

 ing to not more than 12 hours in summer and 24 hours in winter, 

 which is inconsiderable except for short journeys. I would particularly 

 point to the records of batches LII. and LIII. in support of these 

 remarks. The simultaneity with which so many of these bottles were 

 recovered by different individuals after drifting for more than a fort- 

 night is well worthy of note, and we have had still more striking 

 examples of the same thing this year. 



The following table shows the results of an application of the drift- 

 factor to certain cases of drift recorded in Table I. Open water and an 

 absence of deflecting currents are assumed. The actual direction and 

 distance of drift are compared side by side with the estimated direction 

 and distance of drift, and these have been calculated from the winds 

 prevailing at the time by the employment of the factor and the method 

 already described. The wind records, except those of Kousdon and 

 Falmouth, were extracted from the Daily Weather Eeports of the 

 Meteorological Office. 



