66 ANNUAL RECORD OF SCIENCE AND INDUSTRY. 



Such a period has been argued from observations of tempera- 

 ture, terrestrial magnetism, auroras, atmospheric electricity, 

 and, finally, the recurrence of cyclones in the Indian Ocean. 

 The most extended labor on this subject is, however, that of 

 Dr. Koppen, of the Physical Observatory of St. Petersburg, 

 who has, with great labor, brought together an immense 

 mass of observations of the temperature in order to establish 

 his conclusions upon irrefragable bases. In his opinion, in 

 which doubtless all coincide, it is absolutely necessary that 

 observations should be gathered together from all regions of 

 the world, and that our conclusions be not based upon a 

 single series made at any station, or over any country. 

 There seems to have been no observations of temperature 

 made in any portion of the world that have not been used by 

 him, so that his conclusion may be said to embody all that 

 can be deduced from the present state of observational me- 

 teorology. Arranging the stations according to meteorolog- 

 ical zones the tropics and subtropics, the warmer temperate 

 zone, the colder temperate zone, and the cold zone he throws 

 the mean temperature for each year and each zone into the 

 graphic form of a curve, which can then be directly com- 

 pared at a glance with the curve of sun-spots as deduced by 

 Wolf from all known observations of the sun. At the very 

 first one is struck with the great agreement of these curves. 

 In the torrid zone the maximum of heat occurs from six to 

 eighteen months before the spot maximum. To the north 

 of the tropics the maximum of temperature occurs still later 

 than the minimum of spots, being retarded even as much as 

 three years. The regularity and magnitude of the variations 

 of temperature are most beautifully displayed within the 

 tropics, and diminish as we proceed thence toward the poles. 

 The length of the period between the maximum tempera- 

 ture varies, as also does that of the sun's spots, thus, as 

 the interval between the minimum and maximum of spots 

 is almost always shorter than the interval between the 

 maximum and minimum, so does the temperature follow a 

 precisely corresponding change. The parallelism in the 

 series of numbers is so G^reat that there no longer remains 

 the slightest chance of a mere accidental coincidence be- 

 tween these apparently independent variations. The two 

 phenomena evidently are connected, but in what manner can 



