MKMOIKS OF THE NATIONAL ACAHK.MY OK SCI KNCKS. 'J«» 



Tho aurora of tlio iiij;lit of April H! and 17 was oiu' of tlui most brilliant oftlio year. At tin- 

 same tinit« in England "aina^'notic stonn of nnnsnal inlcnsily ra;,'fd fnnn inidiii;(lit of tlie Hllh tn 

 midnight of tlio ITtli." There were no cyclones within neuralgie diHtance during this time. 



On the 2d of October occurivd au important one, and was visible from the lake region a<Toss 

 the Atlantic to En<iland, when a magnetic stitrm was registered at the Uoyal Observatory. A 

 eyclono passed at extreme limit of neuralgie distance (Ui the 2d and .'td. 



Corresponding to the neuralgit- attack of December lo and It! there was an extensive and 

 brilliant aur<)ra. There was no cyclone within neuralgic distani-e during this attack. There are 

 other magnetic storms, with accompanying aurora, which further illustrate the relationship here 

 considered. Such storms afb tho.se of October 2, 1882, February 24, 18.s;$, April 3, 1883, July 2!», 

 1883, November 2, 1883, and others of less note. 



COMPARISON OF WEATHEU ELEMENTS WITH THE I'AIN CUKVE. 



All that we have written in the two former studies .stands substantially conlirmed liy the 

 extension of the record seven years more. 



In Fig. C, paper of 1883, we presented a relationship betwec^n the pain, pressure, and temi)era- 

 ture ciu-ves. A contiiiuation of the record for the additional tinu' maintains the relationship there 

 expressed, perhaps not so accurately, for the pain has during this latter period decreased iit the 

 annual rate of fiftyeight hours instead of forty hours, as stated in last paper for a i)re\ious i)eriod; 

 and also for four years of this additional period the pain curve was cou.structed farther away from 

 the storm track, which would have the effect of making relationship less marked. 



The barometric and other curves as presented in Fig. D (])aper of 1883) stand unafl'ected, but 

 it is proposed to make a closer monthly comparison of the various meteorological elements with 

 the pain curve extending through many years. For this purpose we have constructed the curves 

 as sliown in Fig. E. From this we are enabled to present Table A, in which the column with + 

 sign shows number of times the weather curves were found parallel with the pain curve, and the — 

 sign indicates the number of times they were divergent. 



It will be seen that the pressure curve during the one hundred and sixty-eight months, or 

 foiuteen years, has 56 per cent of the time parallel with the pain and 44 per cent not coinciding 

 with it. The mean hourly wind velocity for eleven years stands 51 per cent parallel to 4!) per 

 cent not so, which indicates that force of wind ni.ay be eliminated as a factor of pain. In the next 

 column the " number of days of rain" for same time shows a stronger jjarallelism than the pressure; 

 so, also, does relative humidity, depth of rain and ozone. We add the ozone curve with some 

 hesitation on account of the well known difficulties in making a correct record of it. This, as most 

 of these are, is from the Central Park Observatory under Trof. Daniel Draper. The pain curve, 

 for most part, was constructed more than 20 miles away from place of these records. The 

 temperature curve, the hours-ofsunshine curve, and, to some extent, the elasticforce-of-aqueous 

 vapor curve, which bears nearly an inverse relation to relative humidity, are large in their negative 

 values and operate to diminish pain, while all the others, excepting wind, but including number of 

 storms, increase anumnt of i)ain. 



The storm curve, which bears the same relation to pain as the pres.snre curve, is con.structed for 

 the most part on Lieut. Finley's tables, and partly from annual and monthly reports of the Signal 

 Bureau. 



To demonstrate more closely the relationship between pain and the number of storms, we have 

 taken twelve months scattered through the whole period from 1.S75 to 1885, inclusive, which gave 

 the highest pain readings. The result is as presented in Table B, in which the + sign gives the 

 value above the mean for the month, and the — sign the value bcloir mean for same. The + 

 ami — values of pressure are almost equal for the twelve months, but for same period there were 

 live and a half days of rain in excess of mean and 4 inches of excess of rain, a slight excess of 

 relative humidity, and, as we should exiu'ct, a negative value in elastic force of vapor. A marked 

 feature, however, of this table is the low temperature, being a deficiency for the whole twelve 

 months of over 14° ( — 14.20), or an average deticiency of over 1° a mouth. Another very marked 



