312 



Prof. B. Stewart and W. Dodgson. [Nov. 20, 



11. The results of Table VITI are exhibited graphically in Plates 2 

 and 3. 



It will be seen that there is a striking likeness between the corre- 

 sponding inequalities of the various observatories. Of these Kew is 

 perhaps most like the mean, while Utrecht follows next in similarity ; 

 Toronto is least like, but even here the similarity is very apparent. 



12. Before leaving this division of our subject we have one remark 

 to make. We do not suppose that certain solar inequalities only affect 

 directly some particular region of the earth from which they are 

 propagated, say from west to east. It seems to us more probable that 

 any solar inequality affects directly all the various stations. But at 

 the same time there may be a progress from west to east of the weather 

 variations caused by these inequalities. Suppose for instance that 

 one of these has the period of 24 days. The temperature-range at 

 Kew will, we may imagine, be directly influenced by the inequality. 

 At the same time it may very probably be influenced by weather 

 changes proceeding from west to east, that have also the same period. 

 These two sources of inequality will be superimposed on each other 

 at Kew, and since each has the period of 24 days, their resultant 

 will have the same period, only in such a case the phase at Kew 

 may be imagined to occur later in time than the corresponding- 

 phase of a station to the west, and earlier in time than the corre- 

 sponding phase of one to the east of Kew. 



13. Let us now proceed to discuss our magnetic results. In our 

 last report we applied our method to the Kew declination-ranges, 

 disturbances being excluded. We have since had reason to question 

 the propriety as far as our purpose is concerned, of excluding such 

 disturbances. For these disturbances are known from the researches 

 of Sabine and others to be due to the state of the sun's surface, and 

 Broun in his recent researches has rendered it highly probable that 

 such disturbances indicate short periods of solar inequality as well as 

 the long period of eleven years. In fine, the disturbed as well as the 

 undisturbed ranges are probably influenced by the same solar periods. 

 Now the effect of excluding disturbances will be to cut off the greatest 

 and least values, and thus greatly to diminish the range of inequality. 

 A slight glance at Table IV of our previous communication will render 

 this evident. If we add up the 27 inequalities (from —7*0 to +6*0) 

 of this table in which disturbances are excluded, we obtain only 46,040 

 as against 98,247, which last is the corresponding sum deduced from 

 the Kew temperature ranges. 



We have, therefore, been induced to apply our method to the Kew 

 declination-ranges, including disturbances. Subjecting these to the 

 treatment already described, and using always the same sixteen years, 

 we obtain the following table : — 



