32 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



DIURNAL FLUCTUATIONS OP PAIN. 



This brings us to a consideration of perhaps the most difficult and complex i^art of our study. 

 The persistency of the type, for more than eighteen years, of these hourly beginnings of pain rea- 

 ders the curve in Fig. F of peculiar interest. We have asserted in jirevious pajjers that there are 

 two subjective conditions which attend attacks of pain. The first one to attract our attention is 

 the drowsy tendency to sleep preceding the attack, and this is equally true when the onset of pain 

 is in the morning. I have no theory to account for this at all, but it seems to have a functional 

 intimacy with the pain and was noted early in the history of this case more than twenty yeai"s 

 ago. The other condition is the eft'ect of eating, which emphasizes a case already on and may pre- 

 maturely develop an attack which otherwise would have been delayed several hours. 



In the daily pain curve we find three principal maxima, the first at 11 a. m., the second at 2 

 p. m., and the third at 7 p. m., with two subordinate ones at 9 a. m. and 4 p. m. Eating may pos- 

 sibly have something to do with the rise at 9 a. m., 2 p. m., and 7 p. m., but seems to bear no rela- 

 tion to the 11 a. m. maximum with a subordinate minimum at 10 a. m. intervening between it and 

 the 9 a. m. 



The part of the curve at 6, 7, and 8 a. m. has almost an inverse counterpart in 8, 9, and 10 

 p. m.; but to take the curve from 9 a. m. to 7 p. m., inclusive (leaving night out for reason pre- 

 viously given), we find one very important minimum at 12 and 1 p. m., then a subordinate mini- 

 mum at 3 p. m., very like the 10 a. m., followed at 5 and 6 p. m. with a distinctly marked minimum. 



We have placed on this pain curve two other curves constructed fi'om Prof. Draper's record 

 of sixteen years, from 1869 to 1884, inclusive, nearly all of which time is comprehended in the 

 eighteen years' pain curve. The broken line is constructed on ordinates the lengths of which are 

 the respective measui-es of the percentage of the number of times a storm began to rain on each 

 hour. The unbroken line is similarly constructed and gives the percentage of depth of rain for 

 each storm. The mean of each for the twenty-four hours is almost identical, but the mean of the 

 number of storms for the period from 9 a. m. to 7 p. m., inclusive, is greater than the mean for the 

 whole time, and the mean of the rain depth for same time is less than mean for the whole twenty- 

 four hours. 



While the absolute values of the rain depths and number of storms on the hours of greatest pain 

 do not show a full correspondence, it is evident at once that there is a_ relative value between the storm 

 and rain curves and the pain, and a parallelism of the two to the latter. 



In all discussion of this subject to this time we have used only instrumental records, such as 

 the synoptic charts, thermograms, etc., but we are aware that the physical effects produced by 

 the temperature on the southeast side of a cyclone is very different from the same temperature on 

 the west side. We have found that the eastern edge of the nem-algic crescent lies more than six 

 hundred miles, or about twenty-three hours, to the east of the center of the "lows" as an average; 

 but no two cyclones or rain conditions are exactly alike and have different kinds of rain, and when 

 we consider that there are two atmospheres, one of a dry gaseous air and another one distinct of 

 aqueous vapor, with the latter never thoroughly intermixed with the former and never in a state 

 of equilibrium, we partially realize how impossible it is for any given storm to get a complete 

 record by instruments and of those which are the best pain producers, for without apparent cause 

 they vary much in this respect. No local hygi'ometric reading can give anything more than a 

 vagaie approximation of the quantity of vapor or its mixture in a vertical column of air. 



A tyi)ical neuralgic atmosphere anticipating a cyclone may be a dry air with a diminishing 

 light, variable northerly to westerly winds, with a pale, delicate, bluish whiteness in the atmos- 

 phere, more marked as one's attention is di'awn to the horizon and especially in the morning. 

 The whiteness may increase for two days with an increasing" discomfort when exposed to the 

 peculiar penetrating power of the sun's rays, but during all this typical condition we have nothing 

 which has measured the changing aspect of the air nor its changing diathermaneity; nor the 

 absolute vapor, except at surface, nor its first faint speck of light cLrrus in the milky whiteness, 

 nor the distinctive character of the cyclone heat in the southeast quadrant, nor the direction of 

 the upper and cross currents of wind, nor above all the modifications of the diurnal and "semi- 

 diurnal" effect superimposed on the cyclone movement. It therefore seems probable that in the sue- 



