B. TERRESTRIAL PHYSICS AND METEOROLOGY. 115 



that the probability of rain increases from one o'clock in the 

 mornimr to a maximum at seven o'clock in the morning, then 

 sinks to a minimum at two o'clock in the afternoon, and rises 

 again to a maximum at midnight. The diurnal period is 

 thus almost opposite to that which obtains under the trop- 

 ics, where in the afternoon, at the hour at which the temper- 

 ature is at its maximum, and at which the clouds are, on the 

 average, the highest above the earth, it rains most frequent' 

 ly. 19 C, VIL, 234. 



PERIODICITY OF SEVERE WINTERS. 



A memoir by Renoir, the distinguished French meteorolo- 

 gist, upon the periodicity of severe winters, although publish- 

 ed many years ago, has been recently quoted in defense of 

 the opinion that such periodicity actually exists. Accord- 

 ing to Renou, rigorous winters return about every forty-one 

 years. They are arranged in groups, generally composed 

 of a central winter, and four or five others disposed on either 

 side of it, within a space of twenty years. Mixed with these 

 years are others, also, of unusual warmth, in such a manner 

 that the mean cold of the season is not sensibly altered. 

 The period of forty-one years seems to be that which corre- 

 sponds to the maxima of the solar spots at the same season of 

 the year. A central cold winter arrives eighteen months 

 after the maximum of spots has coincided with the warmest 

 season of the year. The severe winters seem to alternate 

 between the northern and southern hemispheres of the earth. 

 13-5,135. 



ORIGIN OP THE CENTIGRADE THERMOMETER SCALE. 



According to the historical notes contained in the meteor- 

 ological observations of Lafou, president of the Meteorolog- 

 ical Commission of Lvons, the first thermometer that was 

 ever seen at Lyons was brought thither in February, 1736, 

 by Duhamel, to Father Duclos, director of the observatory 

 founded by the Jesuits in the chapel of their college. This 

 thermometer was constructed with alcohol, according to the 

 principles of Reaumur, and was used for some time. A mem- 

 ber of the Academy of Lyons, named Christin, replaced the 

 alcohol by mercury, as had, indeed, previously been done by 

 Fahrenheit in 1724, and by Dr. Sauvage at Montpelier in 



