Sun-spot Areas and Diurnal Temperature-Ranges. 303 



equalities contain distinctly two maxima and two minima. There are 

 doubtless solitary exceptions where the two former Inequalities exhibit 

 two oscillations, and the latter an approach to a single oscillation only, 

 but these are exceptions, and may we think fairly be supposed to be 

 caused by the presence of some irregularity which we have not been 

 able to eliminate. We consider this constancy of type to be favour- 

 able to the view of a physical connexion between these phenomena. 

 The argument may, perhaps, be put in this way. We are not entitled 

 to expect in all cases terrestrial periods of one maximum and one 

 minimum to correspond with solar periods of one maximum and one 

 minimum, inasmuch as in another instance, namely, that of atmospheric 

 pressure, we have two maxima and two minima more or less distinct 

 for one solar day. But we are entitled to expect, on the supposition 

 that there is a physical connexion between the phenomena we are 

 investigating, that they should exhibit a constancy of type. This 

 constancy is, therefore, a thing which we might expect on the sup- 

 position of a connexion, but a thing unlikely to occur on the sup- 

 position that there is no true connexion between these various 

 Inequalities. 



17. Comparison in Phase. Not only might we expect on the hypo- 

 thesis of a physical connexion between them that there should be a 

 constancy of type in these various phenomena, but we are further- 

 more entitled to look for a definite relation in phase between the 

 corresponding celestial and terrestrial Inequalities. 



A comparison in phase seems calculated to afford better evidence of 

 a connexion than a comparison in length of period. Referring for 

 example to Table II, we find a well-marked maximum Inequality 

 which occurs at +19 for sun-spots, and +20 for Toronto temperature- 

 ranges. Had there been no other maximum than this, and had the 

 sums rapidly fallen off on either side of this point, we should naturally 

 have imagined that this nearness was a strong proof of a physical 

 connexion, but there are evidently several apparent Inequalities, and 

 these are somewhat close together. 



Thus while there is a certain kind of evidence in the fact that for 

 each sun-spot maximum Inequality we have a Toronto temperature- 

 range Inequality not far distant, yet this evidence is rather derived 

 from the circumstance that for each solar there is a terrestrial 

 maximum than from the circumstance that these occur close 

 together ; for if there be a sufficient number of such couplets they 

 must occur close together. 



It thus appears that viewed in this way the evidence of a connexion 

 derived from nearness of period becomes the less striking the greater 

 the number of periods. The truth is that if there were only a single 

 Inequality in each record a comparatively short series would enable 

 us to detect a connexion, but if there are many Inequalities close 



