January 2, 1891.] 



SCIENCE. 



II 



not accompanied by a considerable fall of rain " (Loomis's eighth 

 paper). 



As early as 1876 Hmn found, from the observations on the 

 alpine peaks, that the highest temperature in the upper air oc- 

 curred with the highest pressure, and explained it as due to the 

 dynamic heating of descending air. 



In 1886 Dechevrens showed that on the European peaks Pic du 

 Midi and Puy de Dome, and on Pike's Peak in the United States, 

 the lowest temperature occurred with the lowest pressure, which 

 was exactly the opposite of observations at sea level. He also 

 gave an example of simultaneous observations at the base and 

 summit of the Puy de Dome during a low and during a high 

 pressui'e, as shown by the barometer at both stations. At the 

 base the temperature was highest with the low pressure, but at 

 the summit the lowest pressure and temperature occurred together 

 {American Meteorological Journal, August, 1886). 



In the American Meteorological Journal for May, 1886, Mr. 

 Dewey stated that trom thirty-four pairs of observations during 

 the winter months of 1872 and 1873 he found the average differ- 

 ence of temperature between Burlington, Vt., and the top of 

 Mount Washington to be 6.6° F. when the latter was within a 

 liundred miles of the centre of an anticyclone. The normaldiffer- 

 ence between the two stations is 19". In the different quadrants 

 ■of the anticyclone he found the following differences : north, 9°; 

 east, iO°; south, 4.5°; west, 12.2°; avei-age, 9°. He found the 

 average difference two degrees greater in cyclones. Hazen's re- 

 sults for Mount Washington and Burlington, however, differ from 

 these (Americcm Meteorological Journal, October, 1887), so that 

 further comparisons are needed. 



In a footnote to an article on the origin and development of 

 storms in the American Meteorological Journal. September, 1886, 

 I cited the following reasons for thinking that warmer air is not 

 the essential condition of storm-formation : ' ' Storms sometimes 

 originate along the eastern Rocky Mountain slope when the tem- 

 perature of the air is lower there tlian in any piart of the United 

 States (for an example see the Signal Service charts of Jan. 

 19 and 20, 1886), and storms appear to orginate in this region as 

 often in the night as in the day." 



Very recently Hann has investigated the temperature observa- 

 tions at numerous stations in the Alps during the passage of sev- 

 eral cyclones (Meteor ologische Zeitschrift, September, 1890), and 

 has concluded that the temperature of the air-column as a whole 

 is lower in cyclones than that of the surrounding air. Hann's in- 

 ^ estigations may not be conclusive for reasons stated by Ferrel, 

 ,but they certainly add a link to the chain of evidence. 



As a result of their investigations, Loomis and Hann both de- 

 cided that cyclones were largely the result of mechanical causes. 

 Loomis concluded that they were originated by the conflicting 

 winds between two or more anticyclones, and Hann suggests that 

 they are whirls originating in the upper air. 



Now, I think Ferrel, in his recent letter to tcience, uninten- 

 tionally did Davis an injustice by suggesting that Davis had sud- 

 denly altered his opinion merely because Hann advanced these 

 views. Davis has for years been the leading exponent in this 

 country of the dynamical healing of the air in anticyclones, and 

 during recent yeais I have several times spoken with him about 

 the mechanical origin of cyclones; and, if he is now inclined to 

 give these views more weight, it is because this last link in the 

 chain of evidence has convinced him of the necessity of reconsid- 

 ering the condensation theory. 



I have for several years been convinced that mechanical action 

 had much to do with the origin and development of cyclones, and 

 as working hypotheses in making weather-predictions have care- 

 fully watched the following conditions as favorable for the pro- 

 duction of cyclones : 1. The central region between approaching 

 anticyclones. 2. The region where lower air-currents set in nearly 

 opposed in direction to upper air-currents, so as to favor the pro- 

 duction of a whirl. This latter condition is most frequently 

 brought about in the United States when colder winds, moving 

 from the noi'th-west near the earth's surface, set in to the south 

 or south-west of an area of high temperature or very high pressure, 

 which give rise to upper currents moving from the south. This 

 was the condition preceding tha origin of the very violent storm 



of March 13, 1888. 3. The deflection of air-currents by a long, 

 tall range of mountains, such as the Rookies. I have several 

 times predicted the origin of cyclones under these conditions. One 

 of these was on April 19, 1883. 



I have found the following conditions favorable to the increase 

 of energy in cyclones : 1. The meeting of cyclones moving from 

 nearly opposite directions; 3. The closing-up of a long trough of 

 low pressure by the pressure increasing at both ends; 3. Cyclones, 

 being mainly controlled in their movements by upper air-currents, 

 are sometimes carried by these toward areas of denser air near the 

 earth's surface, and under these conditions tend to increase in 

 energy. Examples of violent storms, developed, as I think, by 

 these mechanical methods, will be found on the following dates: 

 Oct. 14, 1886; Jan. 9, 1889; and Jan. 9,1886. 



The immense gain that would come from being able to antici- 

 pate this class of storms may be inferred from the fact that not 

 one of those I have mentioned in this paper was heralded by our 

 Weather Service in time to be of any use, though the amount of 

 damage done was enormous. 



The views I hold are, that differences of pressure result from 

 differences of temperature over immense areas, as between equa- 

 tor and pole, ocean and continent. This distribution of pressure 

 is modified by the effect of the earth's rotation, and is continuous- 

 ly varying with the changes in temperature of the air. 



The smaller cyclones and anticyclones of our weather-maps are 

 partly or chiefly brought about by the mechanical action of 

 counter-currents in the manner previously explained, though 

 greatly modified by local differences of temperature and density 

 within the cyclone: in other words, they are caused by forces 

 originating outside then- field of origin instead of within it, as 

 supposed by Ferrel. 



General rains are chiefly the result, and not the cause, of ascend- 

 ing currents of air. Differences of pressure in the upper air have 

 a very important bearing on the origin and development of cy- 

 clones. Well-defined areas of low pressure, accompanied by pre- 

 cipitation and an inward tendency of the upper wind, occasion- 

 ally exist in the upper atmosphere without being indicated by the 

 bai-ometric pressure at the earth's surface. 



I have held most of these views for several years, as will be 

 found by my review of Loomis in the American Meteorological 

 Journal, and by two articles in Nature on the origin of anticy- 

 clones, and the cause of precipitation {Nature, vol. xxxvi. 1887, 

 and vol. xxxviii. July, 1888), and have hoped to make some 

 quantitative estimates of the forces and supposed causes; but I 

 have not had the time, and fear I have not the ability to do so. 



I trust Professor Ferrel will not dismiss these as vague hypoth- 

 eses uiiworthy of notice, but will tell us (1) whether the method 

 suggested by Loomis is insufficient to generate a cyclonic whirl 

 according to mechanical principles ; (2) whether conflicting air- 

 currents can be supposed to have sufficient inertia to aid in pro- 

 ducing a whirl, as, for instance, when denser air sets rapidly in- 

 ward from both ends of a long trough of low pressure ; and (3) 

 whether such cyclones as that of Jan. 20. 1886, which originated 

 near the longitude of Denver, where the temperature was lower 

 than in any other part of the United States, when the observations 

 on Pike's Peak showed no vertical decrease at all between the 

 summit and base of the mountain, and when there was no appre- 

 ciable precipitation within a thousand miles of the place of origin, 

 could be explained by any reasonable assumption of a higher mean 

 temperature of the air-column within the field of the cyclone. 



H. Helm Clayton. 



Blue Hill Observatory, Dec. ii9. 



BOOK-REVIEWS. 



Handbook of Problems in Direct Fire. By jAiiES M. Inqalls. 

 New York, Wiley. 8°. $4. 

 This book, which is believed to be the first of its kind ever pub- 

 lished, shows the close attention now given to what maybe called 

 the scientific side of modern warfare, or, rather, of preparation 

 for war. It is devoted wholly to problems in gunnery involving 

 the use of ordinary service charges of powder and angles of ele- 

 vation for the guns not exceeding 15°, which is the definition of 



