404 



SCIENCE. 



[Vol,. II., No. »). 



ican association for the advancement of science, 

 or he would have found the tornado of Aug. 9, 

 1851, in Connecticut, recorded, and that of 

 Maj^ 3, 1868, at Shanghai, 111. Nor has he 

 searched through the state agricultural reports, 

 where he would have found that of June 3, 1860, 

 in Illinois, and doubtless others. Again, Niles's 

 American register gives one at Keene, N.H., 

 on Jul3^ 25, 1807, and at Knoxville, Tenn., on 

 Maj- 25, 1808. The Philosophical transactions 

 would have j-ielded him one in New England, 

 Jul3' 10, 1760; and several others could have 

 been picked up inBlodgett, Piddington, and in 

 the American journal of science. Even that 

 of May 22, 1873, in Illinois and Iowa, reported 

 in the publications of his service for 1873, 

 seems to have escaped his attention. 



As average results like those deduced hj 

 Mr. Finley depend for their value on the num- 

 ber of individual cases taken into considera- 

 tion, would it not have been wise for him to 

 have collated those occurring in other coun- 

 tries, so far as they were accessible without 

 difficult}' ? Peltier would have yielded him 

 quite a crop, some of which, by the waj-, come 

 curiously near home. Other text-books would 

 have given other European ones ; and Chinese 

 and African ones have been described, the 

 latter frequentl}-. Tornadoes are bj- no means 

 exclusivel}' American ; and by a comparison 

 with those in the other countries their essential 

 features could be more easilj' sifted out, and 

 the incidental ones given their proper promi- 

 nence. 



In the pamphlet, ' Tornadoes, their special 

 characteristics and dangers,' the author classi- 

 fies the rotatory storms. It was in the pursuit 

 of a classilication of storms, that he first had 

 his attention called to the insufflciencj- of our 

 knowledge of this species. Tornadoes are here 

 described in some detail, and numerous direc- 

 tions given for the protection of life and 

 property on their occurrence. It is the best 

 description of the storm known to the writer. 



Mr. Finlej' considers the tornado a much 

 hetter-defined species than is likelj- to be 

 acknowledged b}' meteorologists generally. 

 Right names are extremel}- useful, but we must 

 not permit them to conceal any underlying 

 unit}'. By his anxiety to get a clear species, 

 the author shuts out the light which he might 

 get from the study of storms of so similar 

 character that one is compelled to believe that 

 their differences are due only to difference in 

 surroundings. Thus water-spouts are onl}- tor- 

 nadoeson the water, with circumstances remark- 

 abl}' favorable for observation. They occur 

 not infrequently on the Great Lakes, and the 



change from tornado to water-spout has been 

 observed more than once. Judging from the 

 only description known to me of the riband 

 storms of British North America {Cosmos, 2d 

 series, iii. 274, 275), thev are also somewhat 

 modified tornadoes. And while the name 

 cloudburst refers rather to a single feature of 

 subordiuate meteorological importance, the 

 phenomenon is probabl}' often of tornado char- 

 acter. Indeed, leaving out of account eddies, 

 which it is not, the tornado differs only quan- 

 titatively from the other members of that list 

 of storms which begins with the formation of 

 a cumulus cloud, passes on to thuuderstorms 

 and hailstorms, and culminates in the ' low- 

 centre,' the hurricane, and the typhoon. They 

 all find their origin in the transformations of 

 water ; and to overlook the relations thej' have 

 to each other, is to refuse assistance in a prob- 

 lem well-nigh insoluble with that assistance. 



It is expressly stated (see p. 4 of the last- 

 mentioned pamphlet) , that the g^'ratory motion 

 is alwaj'S from right to left. The writer would 

 point out the exceeding difficulties which sur- 

 round the determination of this point. Some 

 of the early observers saw only indications of 

 a radial inpour, and in the descriptions of 

 tornadoes one frequently finds dextral whirls 

 mentioned. In so small a storm, the earth's 

 rotation would surely have no appreciable dis- 

 turbing effect ; and that, in a difference of lati- 

 tude of only a few rods, it should originate 

 velocities of a hundred or more miles an hour, 

 is so unlikely that it need hardl}- be considered. 

 Furthermore on p. 7 the author admits varia- 

 tions in the g\Tation of the tornado's other 

 self, — the water-spout. So, while unwilling 

 to differ from so experienced an observer on 

 such a point, both the records and general con- 

 siderations lead the writer to think that the 

 direction of gyration maj' be indiflferentlj- dex- 

 tral or sinistral. 



There is one possible feature of tornadoes 

 which has not yet been definitel}' proven, but 

 of which we ought now to be able to ascertain 

 the truth or falseness by an investigation like 

 that just discussed; viz., Are tornadoes dis- 

 posed to return on the same path ? The writer 

 spent his childhood in northern Illinois, where 

 heavy hail and other tornado-like storms are 

 not rare. He remembers several instances of 

 their following the exact path of their prede- 

 cessors. Professor Whitfield {Amer. joiirn. 

 sc, 3d series, ii. 99) says in regard to south- 

 ern tornadoes, "It is not an established fact, 

 but it is commonl}' believed, and with some 

 reason, that the tornado does, in the course of 

 j-ears, return along its beaten path, and that 



