SEA-FISHERIES LABORATORY. 127 



ON IRREGULARITIES IN SEA-TEMPERATURE DUE 

 TO TIDAL OSCILLATORY STREAMS. 



By R. J. Daniel, B.Sc. 



It has been suggested by Professor D'Arcy Thompson* 

 that the annual temperature variation, at a fixed position in 

 the sea, can be represented by a Fourier Series. There ought 

 to be one principal component in such a series, and this ought 

 to account for the greater part of the variation above and 

 below the annual mean temperature. Then there ought to 

 be other components of J, J, J, and so on, of a whole year, each 

 of which also accounts for some of the variation from the mean 

 annual value. Some, of course, of these minor components 

 may be expected to have small amplitudes ; so small that 

 they may be neglected. An harmonic analysis of the annual 

 temperature fluctuation, at any fixed position, ought, therefore, 

 to give us a series of equations, each of which will represent 

 an harmonic oscillation, and the phase and amplitude of each 

 oscillation would be given by the analysis. If, then, the various 

 components be added algebraically the summation ought to 

 reproduce the observed annual temperature oscillation. 



Of course, in making such an analysis, we ought to assume 

 the periods of the various components and then evaluate 

 their phases and amplitudes. That supposes a theory (or 

 explanation, rather) of the causes of variation in sea-tempera- 

 ture. Obviously the principal cause is the annual change of 

 declination of the sun. In the case investigated in this paper 

 the other causes are tidal oscillatory streams, by means of which 

 heat is transferred from the sea to the land, or vice versa. 

 These oscillations are assumed to have (1) a semilunar period 

 (the neap and spring tides), and (2) a semi-annual period 



* Kept. II Northern area, North Sea Fisheries Investigations, cd. 

 3358, 1907, p. 185. 



