224 SPARKS FROM A GEOLOGIST'S HAMMER. 



are of equal importance with the means. It signifies 

 little that the growing season begins in March, if liabil- 

 ity to killing frosts continues to the middle of May, as in 

 Tennessee. A mean October temperature of 60 is com; 

 paratively valueless after a September freeze. The mean 

 temperature of a season may be mild, or even delightful, 

 at the same time that one or two days have brought de- 

 structive cold. One killing frost is as bad as a dozen, 

 for vegetation has but one life to destroy. It is the 

 liability to these exceptional temperatures wdiich we must 

 know before forming final judgment on the adaptability 

 of a district for a particular crop. A winter which aver- 

 ages mild may be marked, like the climate of Saint Louis, 

 by one, two or three mornings destructive to everything 

 which would triumphantly survive all the rest of the 

 season. Every fruit-raiser knows that it is not the aver- 

 age weather of winter or spring which endangers his buds 

 or his trees. It is the one or two nights of the whole 

 season which brings him apprehension, especially if ac- 

 companied by high wind. It is of no consequence that 

 the winter mean of Saint Louis is 33 and that of Grand 

 Haven 21, or of Traverse City 24, if the thermometer 

 falls sometimes 22 below zero at Saint Louis and never 

 sinks more than 16 below zero at Grand Haven or 

 Traverse City. It is precisely against these exceptional 

 extremes that the great lakes exert their most striking 

 influence. 



There are two ways to consider extremes of climate. 

 We may consider the mean minimum of a locality, or its 

 extreme minimum, for a series of years. There is a low- 

 est point reached by the thermometer at each locality 

 every winter. DijQferent winters may vary greatly in 



