January 17, 1895J 



NA TURE 



273 



these statistics have been prepared. How far these records are 

 absolutely correct is beside the preie.it question ; all the data 

 have been obtained with the same instruments, and are strictly 

 comparab e between themselves. 



It has been the custom at the observatory to treat as storms 

 or gales, all oc:a-ion5 in which the Robinson anem:)meter re- 

 corded the horizontal velocity. of the atmosphere as equal to or 

 exceeding fifty miles an hour. The total number of instances 

 in the twenty-eight years under consideration is 321, or an 

 average slightly below one a month. Iiwill frequently happen 

 in stormy periods that the critical velocity will be registered, 

 and be followed by a partial lull in the storm, to be succeeded 

 by gusts of greater force. If these periods of comparative calm 

 have lasted for about twenty-four hours, the disturbance would 

 count as two storms, although probably both are parts of the 

 same atmospheric disturbance. 



The greatest number recorded in a year is in 186S, when no 

 less than twenty-eight gales were reported ; and the leait occurs 

 in 1880, when there were only two, and these neither long in 

 continuance nor great in violence. It is not without signifi- 

 cance — though, of course, it is not intended to insist upon the 

 coincidence — '.hat when the number of storms is plotted as an 

 ordinate with the time for abscissa, a rough curve can be drawn 

 among them giving maxima at practically equal intervals of five 

 years, from 1868, with minima at intermediate dates. Or, if 

 we take the sumsof the maximum years 186S, 1873, 1S78, &c., 

 it is found that in the six we have records of eighty three storms ; 

 while in the six years of minimum, 1871, 1875, 18S1, &:., we 

 have only thirty-seven. • 



It is scarcely necessary to refer to the direction in which 

 storms approach the observatory, after w'jat has been said of its 

 geographical position. Roughly, they are all from the west, 

 with slight deflec.ions to the north and south. As a matter 

 of fact, only five have deviated from this rule, and they have 

 been either east or south-east, anl have been comparatively 

 slight in their character. 



The time scale on which the velocities havo been registered is 

 not a very open one. The recording drum mjves through 

 rather less than one inch in an hour, and the habit has been to 

 read and record the distance travelled from the commencement 

 of one hour to the beginning of the next. The maximum hourly 

 velocity is, therefore, not to be understood as the greatest in 

 any sixty consecutive minutes, but as the greate';t in one wh^le 

 hour as marked by the clock. \\\ this sense the following table, 

 which exhibits the number of times the greatest hourly velocity 

 has exceeded noticeable amounts, is to be understood : — 



Here again, curiously enough, we find some slight evidence 

 in favour of a five-year period. Not only do the two instances 

 of the greatest velocity recorded in the observatory occur in 

 years already noticed as those of maximum disturb ince, but 

 the average velocity of all storms in the years 186S, 1S73, &j., 

 is 592 miles, as ompared with 57'9 miles in the years of 

 minimum number. 



Intimately connected with velocity, though probably a less 

 accurate measure of the true force of a storm, is the pressure 

 recorded per square foot. Here Bidston has long held a record 

 for the British Isles, having placed to its credit a pressure of 

 ninety pounds on M.irch g, 1871. Tiie accuracy of this measure 

 has often been questioned, and probably it is too gieat owing to 

 the momentum in the moving parts of the machmery, but it is 

 certainly the record of a far greater pressure than has ever been 

 witnessed since. Considering over what small areas these 

 excessive pressures are exercised, and the great variation that 

 exists from moment to moment in the velocity of the wind when 

 a storm is raging, it is not an impossible amount, but it would 

 certainly be misleading to c mclude that such a pressure was a 

 measure of the force of the wind a fe.v feet, or even a few inches, 

 away from the pressure-plate. The hourly velocity on the 

 occasion when this pressure was registered has frequently been 

 exceeded, without reproducing similar pressures. Herein is 

 represented tlie great difficulty in determining a simple relation 

 between pressure and velocity, or, rather, the square of the 

 velocity ; for one may regard Hutton's law of wind-pressure on 



NO. 13 16, VOL. 51] 



a given obstructing surface as satisfactorily proved. The fact of 

 such accidents as that referred to destroys any value that can be 

 drawn from averages, but as a mere matter of fi^'ures, it appears 

 that the average m iximum hourly velocity for all storms is 

 584 miles, and the mean of the maximu.n pressures 37 '6 lbs. 

 This would require the factor for maltiplying the square of the 

 velocity in miles, to obtain the pressure per square foot, to be 

 greater than one-tenth, which is evidently anl necessarily 

 erroneous, since we are comparing the accidental momjntary 

 pressure with the average velocity obtained throughout the 

 entire hour. Taking the extreme pressures for what they are 

 worth, the numbers come out as follows : — 



Pressure in lbs. on sq. ft. Number of instances. 



20-30 132 



30-40 HI 



40-50 50 



50-60 ... • II 



60-70 12 



On two occasions the pressure registered was greater than 

 70 lbs. to the square foot, and on three did not reach twenty. 

 Since the extreme pressure, are in a sense accidental, a ad do 

 not represent with any accuracy the force of a storm, it does no; 

 seem desirable to determine the relative pressures in whu have 

 been called years of maxima n and minimum storm occurrence. 

 But the general features are again borne out. Tne expla la'.ion 

 here is probably that the greater the number of storms the 

 greater the chance of finding a high pressure. It seems more 

 profitable to inquire what is the av.-rage length of a storm, how 

 long may a violent disturbance ht expected to last. This 

 question is unfortunately complicate 1 by the fact already alluded 

 to, that a storm may subside for a few hours and then reappear 

 with its original violence. If the interruption is only for an 

 hour or two, as already explained, the depression is considered 

 as one, but only those hours are counted in which the registered 

 velocity exceeds fifty miles. There are only two instances in 

 which this amount of violence has been maintained for thirty 

 consecutive hours, viz. in February, 186S, and again last Feb- 

 ruary. The total number of stormy hours in thj twenty-eight 

 year, is 1732, giving an average of 5 '4 hours for each storm. 

 Our local disturbances are therefore not of long duration. 



An examination of the dates when gusty weather is most pre- 

 valent, goes neither to substantiate the ancient myth of the equi- 

 noctial gales, nor to uphold the evil supremacy which has long 

 been assigned to the winds of March Since no attempt has 

 been made to equalise the lengths of the months, February has 

 been somewhat unfairly treated in the following table ; but 

 notwithstanding this handicap, it possesses the unenviable 

 privilege of compressing within its shortened limits more tem- 

 pestuous weather than any other month. 



... vT r . No. of stormy 



Month. No. ofstjrms. . ,,^ ' 



hours. 



January ... .. 47 ... ... 260 



February ... ... 42 ... ... 2S1 



March 47 238 



April ... 14 63 



May ... 7 27 



June 3 19 



July 7 21 



August 17 65 



September ... ... i6 ... ... 77 



October ... ... 26 ... ,. 180 



November ... 44 ... 254 



December .. .. 51 ... ... 247 



Total 321 1732 



Of course one would expect to find the greatest number of 

 storms in the winter ; but that three-fourths of the whole should 

 be compressed within five months of the year, is a greater dis- 

 proportion than was expected. Th:; variations of the barometer 

 during these storms, and the dependence of this variation upon 

 the direction of the wind, are of considerable interest, but may 

 not very well be entered upon here while treating simply of 

 numerical statistics. William E. Plummer. 



Peculiarities of Psychical Research. 



Mr. Di.\'in asks in his first letter: Coald an abnormil dis- 

 tribution of the cards aflfect the result, if certain precautions 

 were taken? In his second letter he says there was nothing in 

 his first letter to indicate that he under-estimated the import- 



