July 2, 1920] 



SCIENCE 



17 



formed chronicler and a sympathetic inter- 

 preter of Pasteur, it is doubtless true that 

 Smith represents a type of mind and work- 

 manship much more like that of Pasteur him- 

 self. He, like Pasteur, has fought the 

 pioneer's battles and keenly relished the fray. 

 One who knows this is prepared to find every- 

 where, in the introduction, in the translator's 

 notes which are scattered through the book 

 and especially in the annotations at the close 

 a certain flavor quite other than Duclaux 

 himself could impart. 



The translation is avowedly addressed to 

 the younger generations of American scien- 

 tists who are liable to forget the dramatic con- 

 flicts through which were won the ways to the 

 higher and freer conceptions which they have 

 inherited. For this reason the book (and the 

 translation better than the original) should 

 serve admirably as a reading book about 

 which a seminary may be conducted with stu- 

 dents desiring to trace the history of biolog- 

 ical thought during the last century. The 

 reviewer proposes so to use it. For such pur- 

 poses there is much gained through the addi- 

 tion by the translators at the close of the book 

 of an annotated list of all persons to whom 

 reference has been made in the text. Here 

 as elsewhere there is evidence of those inti- 

 mate, highly individualistic, personal touches 

 through which Duclaux and Smith in com- 

 bination have served to reflect so much of 

 the dominant individualism of Pasteur. 



L. E. Jones 



University of Wisconsin 



NOTES ON METEOROLOGY 



THE DISTRIBUTION OF MAXIMUM ELOODS 



In an interesting paper read before the 

 American Meteorological Society at its New 

 York meeting, January 3, 1920, Professor A. 

 J. Henry* analyzed the flood records of many 

 of the streams of America and Europe with a 

 view to determining their time and space dis- 

 tribution, and, if possible, any systematic or 

 cyclic recurrence. 



* ' ' The Distributiott of Maxamum Floods, ' ' ihid., 

 pp. 861-866. 



For the purpose of his discussion. Professor 

 Henry has used the " average annual flood," 

 which is defined as "the arithmetical mean 

 of the annual floods for a number of years"; 

 also the " maximum flood " which is that 

 caused by excessive run-off when the rainfall 

 is sufficient to raise the ground storage to a 

 high level and thus saturate the soil, or when 

 a warm rain falls on a snow cover; and 

 finally, the "absolute maximum flood" which 

 is the greatest reported for a given station. 

 It is important to know what the absolute 

 maximum flood magnitude is for a given 

 place, and to know whether that maximum 

 has been reached. Says Professor Henry: 



It can not 1)6 too strongly emphasized that the 

 occurrence of the absolute maximum flood is usually 

 conditioned upon the synchronism of certain cli- 

 matic events which in themselves have no fixed law 

 of occurrence. Very intense rainstorms are seldom 

 long continued and of great extent. The heavy 

 summer showers that occur in the United States 

 being limited in area may cause an extraordinary 

 flood in a small watershed, and doubtless many 

 such floods occur in some part of the country an- 

 nually. These extreme floods in small streams are 

 completely absorbed as soon as they reach the 

 trunk stream. 



To the end of determining any relation 

 between the absolute maximum and the 

 average, the ratios of these two values have 

 been tabulated for 45 of the principal rivers 

 of the United States. The agreement of the 

 various ratios is, as the author remarks, 

 " better than was expected," amounting in 

 general from 1.3 to 1.5; that is, the absolute 

 maximum flood was 1.3 to 1.5 times the 

 annual average. There are a number of 

 ratios of greater magnitude, but, in most 

 cases, this is accounted for either in the 

 nature of the watershed, or the local condi- 

 tions surroimding the gaging. Small ratios 

 are found at places where overflow takes place 

 easily and the river may greatly increase its 

 cross-section. 



There appears to be no cyclic distribution 

 of floods but "that the dominating control is 

 rainfall, and since there may be one, two or 

 even three years of excessive rainfall, it 

 follows that great floods may likewise occur 



