Shown by Recording Tide-gauges. 135 



undulations at Halifax and fail to affect Pictou and vice versa. 

 This discrepancy is still more marked when it is considered 

 that Pictou is almost exactly N.E. from Halifax, and it is on 

 the poleward current, moving N.E., that Mr. Denison relies 

 for the production of the atmospheric billows in question. (It 

 is to be understood here that I am not discussing the existence 

 of atmospheric billows but the effects attributed to them.) 



On the other hand, the close correspondence between Atlan- 

 tic (including Bay of Fundy) ports, in contrast with the lack of 

 correspondence between Atlantic and Gulf ports, is strongly in 

 favor of the view that the disturbance at any port is usually 

 due to the transmission by water of distant disturbances of the 

 general body of water caused by storms. 



Secondary Undulations and barometric records. 



Any connection between atmospheric disturbance and second- 

 ary undulation should, of course, be shown by a comparison 

 of tidal and barometric records. I have carefully compared 

 the tracings of secondary undulation at St. John for the sum- 

 mer of 1896 with tracings of the St. John barograms for the 

 same date. These latter I owe to the kindness of Mr. Bell 

 Dawson. They were made as carefully as possible with a hard, 

 sharp-pointed pencil. The dates are June 8, 9, 10, 21, 22, 23 ; 

 July 8, 25 ; Sept. 1, 5, 6, 7, 18, 19, 20, 22, 23— the only occa- 

 sions, I believe, on which marked secondary undulations 

 occurred at St. John during the summer of 1896. A careful 

 comparison of both series of records shows that (1) disturb- 

 ances of atmospheric pressure usually occur about the times of 

 secondaiw undulations, and, in general, the stronger the atmos- 

 pheric disturbances the stronger the water undulations, but 

 (2) the atmospheric disturbances rarely if ever begin and end 

 at the same time as the water undulations ; they usually precede 

 but sometimes seem to follow the latter, and the secondary 

 undulations may continue for several hours (e. g. 12 hours on 

 Sept. 20) after the barometric disturbances have ceased ; (3) in 

 no case is there any suggestion of regular periodicity in the 

 barometric disturbances such as there is in the water undula- 

 tions. 



The utmost therefore that it seems to me possible to deduce 

 is that secondary undulations usually accompany storms near by 

 or at a distance, but that the period of the oscillations at any 

 place is wholly unaccounted for by atmospheric conditions. 

 (This is also in accord with an extensive series of comparisons 

 that I made, of the Atlantic Pilot Charts and the Canadian 

 Weather Maps on the one hand and the records of secondary 

 undulations on the other — into the details of which it seems 

 superfluous to enter.) 



