170 THE PHYSICAL GEOGRAPHY OF THE SEA. 



exact position. The polar calms are no more a point than the 

 equatorial calms are a line ; and, considering that these poles are 

 areas or discs, not points, it is a little curious that philosophers in 

 different parts of the world, using different data, and following up 

 investigation each through a separate and independent system of 

 research, and each aiming at the solution of different problems, 

 should nevertheless agree in assigning very nearly the same po- 

 sition to them all. Are these three poles grouped together by 

 chance or by some physical cause ? By the latter, undoubtedly. 

 Ilere, then, we have another of those gossamer-like clews, that 

 sometimes seem almost palpable enough for the mind, in its hap- 

 piest mood, to lay hold of, and follow up to the very portals of 

 knowledge, where we pause and linger, fondly hoping that the 

 chambers of hidden things may be thrown open, and that we may 

 be permitted to behold and contemplate the mysteries of the winds, 

 the frost, and the trembling needle. In the polar calms there is 

 (§ 215) an ascent of air; if an ascent, a diminution of pressure 

 and an expansion ; and if expansion, a decrease of temperature. 

 Therefore we have palpably enough, a connecting link here be- 

 tween the polar calms and the polar place of maximum cold. 

 Thus we establish a relation between the pole of the winds and 

 the pole of cold, with evident indications that there is also a phys- 

 ical connection between these and the magnetic pole. Here the 

 outcroppings of a relation between magnetism and the circulation 

 of the atmosphere again appear. 



862. Thousands of observations, made by mariners and record- 

 The barometer in the ^^ ^^ their abstract logs, havc enabled us to de- 

 wind bands. tcrmine approximately the mean height of the ba- 



rometer for the various bands (§ 352) at sea. Between the paral- 

 lels of S6° S. and 50° N., Lieut. Andrau, of the Dutch Navy, has 

 collected from the abstract logs at the Meteorological Institute of 

 Utrecht no less than 83,834 observations on the height of the ba- 

 rometer in the following bands : 



