voith reference to the WiirVwind Theory ofStm-ms. 361 



It is an error to allege that I have " admitted" a fact such 

 as is here stated. On the contrary, in careful explorations 

 made on foot, through an aggregate extent of more than fifty 

 miles of the tracks of various tornadoes, I have never met with 

 such " a fact," or combination of facts, as Dr. Hare describes. 

 In all the cases 1 have met with in which trees have fallen 

 one upon another, if their tops pointed in opposite or nearly 

 opposite directions, these directions have never been parallel 

 to the course pursued by the tornado, but always in directions 

 more or less transverse to the same ; and I consider the op- 

 posing allegation as one of the chief errors of my opponents. 



The trees which have fallen in directions which are more 

 or less backward from the course pursued by the tornado, are 

 almost invariably found on the left side of the track exterior 

 to the line of its axis. But few of these point directly back- 

 ward, and still fewer can be found near the axis, as the hypo- 

 thesis of my opponents requires. Of the trees found with 

 their tops pointing directly for'ward, or nearly so, a small 

 number have been seen on or near the right margin of the 

 track, with appearances which showed them to lie as they 

 first fell ; a fact which seems equally fatal to their hypothesis. 

 Some trees, along and near the line of the axis, are, however, 

 found pointing in this onward direction, and much stress is 

 laid on this fact by one of my opponents : but it appears, on 

 examination, that in all these cases the trees have been torn or 

 ttoisted from the transverse position in -jshich they first fell \ 

 owing, as I infer, to the more violent force exhibited at and 

 immediately behind the centre of the whirl, or at the point 

 which may not inaptly be termed the heel of the voj-tex"^. 



It is true, however, as I have " admitted," that when trees 

 are found to have fallen one upon another, the top of the up- 

 permost tree points in a direction more onward than the one 

 beneath ; as is seen by the diagrams and schedules of Professor 

 Bache, and as may be inferred, perhaps, from the sketches 

 given by Professors Olmstead and Loomis f : and it is equally 

 true, that this fact no more favours the hypothesis of a di- 

 rectly inward motion, than that of a whirlwind being, as an 

 abstract deduction, " reconcilable" with either. The proper 

 generalization of this class of facts l> have attempted to give 

 in my paper on the New Brunswick tornado; which is, "that 

 the uppermost or last fallen of these trees points most [or 

 more] nearly to the course pursued by the tornado ;"/. f. 

 more nearly than the underlying tree which fell first ; diver- 



• See this Journal, January 1841, p. 20-29, and map. 



t See Siiliman's American Journal, vol. xxxiii. p. 369 ; vol. xxxvii. p. 343. 



