66 
POST SCRIPT.—In response to enquiries the following information 
has been kindly sent by the United States Weather Office:— 
“ The tornado at St. Louis, Mo., on January I2th, 1890, passed within 
about 2,800 feet of the position of the barograph. That at Owensboro, 
Ky., on March 27th, 1890, passed about 7,000 feet from the barograph, 
By a typographical error, “ Science” gave a tornado at Cincinnati, Ohio, 
on March 27th, 1890, which should have been a thunderstoi-m, though 
intimately connected with the Louisville, Ky., tornado. A copy of the 
original sheet at St. Louis, Mo., for January 12th, is sent with this. 
The original sheet at Owensboro, Ky., unfortunately for tins country, 
was given by its owner to Prof. A. Buchan, of Edinburgh, Scotland, of 
whom it is probable 3 mu can get a tracing.—A. W. Greeley, Chief 
Signal Officer, February 25th, 1891.” 
The request here suggested was made, but nothing further has 
been received. 
The explanation of the Cincinnati curve perhaps points to a 
connection between the upward “kick” and gusts of wind due to 
thunderstorms, when the latter are the predominating phenomena. 
Such “kicks” have been noted when no special wind-rush occurred. 
Possibly they are due to swirls not reaching down to the earth’s surface, 
of the kind which maj' be noticed causing such commotion among the 
thunder clouds. The association here with the Louisville tornado lends 
support to the suggestion. 
The Seaton Carew curve was due to a whirl without a thunderstorm, 
at St. J.ouis and Louisville this was subordinate. We may notice how, 
in all five tracings, the effect occurred during a rapid fall, soon followed 
by a rise. 
As the St. Louis curve shows several interesting details, not given 
by the rough figure in “ Science,” it is the more unfoitunate that we 
have not at hand a more accurate tracing of the Owensboro record. 
“ Science ” represents the former b}-^ a stead}^ descent of one-tenth inch 
in hours, suddenly changed to a vertical, i.e. instantaneous, rise of 
rather more; after which the curve bends sharp over, falls about 
one-fortieth inch in the next two hours and ends with s^'iuptoms uf a 
rise. 
In the exacter tracing, for the whole of the da^", we find a 
tremulous line from midnight to mid-daj% which indicates unsteady 
pressure. This, until 5 a.m., rises and then falls between 29'42 and 
29’44. B}" noon it is down to 29T8. Then follows a rapid fall, with 
steady line, until 5-15 p.m. A jerk at 2 39 and again at 9-15, appears to 
be accidental. At 5-15 the tracing lies at 28'87; the next quarter 
hour saw it reach 28'82, this sudden drop being due, possibly, to the 
suggested suction. In the next five minutes there is a rise to 28‘93 or 
OTL in., instantly followed by a drop of nearly 0‘03 and a second 
recovery of 0 02 by 6 p.m.; then follows a seiiond slight oscillation 
before 6-30 and a slight drop to 28’89 until 8. Here begins a rapid 
rise so that the record at 10 p.m. is 29’06, and at midnight 29T6 
inches.—J. E. C. 
