Disturbances due to Tides and Waves. 155 



that C is determined less by the local tidal flow than by the 

 general flow in that part of the English Channel. A dif- 

 ference of from 2 to 3 hours is known to exist between high 

 or low water at Dartmouth and the turn of the tide at the 

 Homestone Buoy (fig. 3), and the observed difference of 

 phase may perhaps thus be accounted for. 



B. Experiments with Drifting Electrodes. 



Two electrodes were towed in tandem behind a vessel, the 

 near electrode being about 60 yards and the distant electrode 

 160 yards astern. In this way the electrodes were con- 

 veniently extended along an approximately horizontal line 

 at a distance of about 100 yards apart ; and the orientation 

 of this line could be readily altered by changing the course 

 of the vessel. The vessel steered a course which may be 

 termed a " tidal square." This simply means that she 

 followed an approximately square course relatively to the 

 water such that two sides of the square were parallel to the 

 direction of tidal flow ; owing to side drift the actual course 

 would be as shown in fig. 7 B (PI. II.). The electrode cables 

 were connected to a Unipivot galvanometer, specially balanced 

 for use at sea, which served as the voltmeter. Readings 

 were taken at intervals of 5 seconds throughout the run 

 and were plotted as shown in fig. 7 A, the deflexion being- 

 plotted as positive when it indicated the distant electrode to 

 be at a higher potential than the near electrode. As a rule 

 at least one side of the square was repeated as a test of the 

 constancy of the electrodes. 



The direction and strength of the tide were stated by the 

 navigating officer, generally by observation of the wash past 

 a buoy. 



Examining the records reproduced in PL II. figs. 7, 8, & 9, 

 we observe that though the reading is by no means constant 

 even on a straight course, yet a distinct change of zero 

 occurs as the vessel takes up a new course. The records 

 were analysed as follows. Let the mean readings of the 

 galvanometer (reduced to millivolts) be d\ with the tide, 

 d 2 across the tide from right to left, d 3 against the tide, 

 d± across the tide from left to right. Then the transverse 

 component of the apparent potential slope in the sea from 

 right to left is (d 2 — d±)/2 millivolts per 100 yards, whilst the 

 component in the direction of flow is (d l — d 3 )/2 millivolts 

 per 100 yards. From these rectangular components the 

 actual direction and magnitude of the apparent slope e 2 were 

 obtained. The results thus obtained for 11 tidal squares are 



