lO 



SCIENCE. 



[Vol. XVII. No. 413 



the plate, both upon the ah- vvhich is retarded and stopped in front 

 of the plate, and also upon that behind the plate. 



It is doubtful whether a descending current in the open air of 

 more than two metres per second could be found anywhere in the 

 whole atmosphere. This, we have seen, would increase the baro- 

 metric pressure 0.0194 of a millimetre, a quantity which could not 

 be detected by the most delicate and accurate barometer. It is 

 seen, therefore, how very improbable is Dr. Hann's theory of the 

 •cause of high-pressure areas. 



Dr. Haun lays great stress upon the efficiency of the steep 

 gradients of the upper part of the atmosphere, in the middle and 

 higher latitudes, in producing both cyclones and high pressure 

 areas. But the forces arising from these gradients are almost 

 completely counteracted by the deflecting forces of the earth's 

 I'otation in connection with the eastwardly moving currents in 

 these latitudes, the velocities of which increase with increase of 

 altitude very nearly in the same . proportion as the steepness 

 of the gradients Although the steepness of these gradients 

 at high altitudes, especially in the southern hemisphere, is con- 

 siderable when considered with reference to gravity simply, 

 yet, if all the forces are taken into account, there is no part 

 of the atmosphere in the middle latitudes where the gradients 

 are smaller, the velocity of the easterly motion being such as to 

 not quite counteract the force from the gradients, and to leave a 

 residual force simply which is sufficient to counteract the frictional 

 resistance in these high altitudes, which is very small. It would 

 be just as reasonable to maintain that there is a strong tendency 

 in the water of the ocean to rush toward the poles, because there 

 are steep gradients, considered with reference to the earth's attrac- 

 tion only, and leaving out of consideration that the centrifugal 

 force arising from the earth's rotation counteracts this tendency, 

 as to maintain that the air in these high altitudes has a strong ' 

 tendency to rush toward the poles. Wm. Ferrel. 



Martinsburg, W.Va., Dec. 22. 



Recent Investigation on the Causes of Cyclones and 

 Anticyclones. 



If I were required to name the man who impressed mo as the 

 most profound meteorological writer whom I had read, I should 

 without hesitation say Professor Ferrel. 



The most of us are qualitative meteorologists: he may be called 

 a quantitative meteorologist. Not content with mere general 

 statements of causes and forces, he attempts to determine the ex- 

 act value of each one, and by rigid mathematical formulas to 

 •determine if they are sufficient to account for the given results. 



This represents a high, if not the highest, development of a 

 scientific mind. For this reason I would hesitate to dissent from 

 Professor Ferrel's conclusions more than from any writer I know ; 

 but he has himself, in his recent letter to Science, severely criti- 

 cised the supposed blind following of authority, and, if there were 

 needed any excuse, I would give this as the reason for presenting 

 the views opposed to those of Ferrel. 



There are two methods of arriving at results. The one is by 

 ■deduction, in which the thinker, starting from axioms, well de- 

 termined constants, or general laws, works out the results which 

 must follow. The other is by induction, in which the thinker 

 starts from observation, or separate individual facts, and arrives 

 at general laws. Both methods are necessary; and most thinkers 

 of to-day will admit that no theory of natural phenomena is com- 

 plete until the results of deductive reasoning correspond to the 

 results of inductive reasoning, or vice versa. 



Now, Ferrel is essentially a deductive reasoner. It is necessary 

 in such reasoning that the fundamentals, or physical constants 

 from which one starts, should be correctly determined. In Fer- 

 rel's and Marvin's replies to Hazen in Science and in the American 

 Meteorological Journal, I believe it has been shown that the con- 

 stants forming- the basis of the calculation in Ferrel's condensa- 

 tion theory of cyclones were satisfactorily determined. Starting 

 with these, and following Espy, he has shown, that, given a 

 ■warmer body of air, or a rapid vertical decrease of temperature 

 over a considerable area, the causes are adequate to iffitiate and 

 maintain a cyclone. 



The question now is, do the investigations of inductive meteor- 

 ologists sustain these views ? 



In order to study the results which follow rapid vertical de- 

 crease of temperature in the atmosphere, Loomis "selected from 

 the volumes of the published observations of the Signal Service 

 (November, 1873, to January, 1875, and from January, 1877, to 

 May, 1877) all of the cases in which the temperature at Pike's 

 Peak was 40" lower than at Denver." With this difference be- 

 tween them, the air would theoretically be in unstable equilib- 

 rium. ' ' The number of these cases in twenty months of observa- 

 tion was 343. Only 39 of these cases occurred during the seven 

 winter months of observation, and they occurred most frequently 

 during the months of May. . . . The facts appear to show 

 that at the dates given there were seldom any extraordinary dis- 

 turbances on Pike's Peak. In two cases hail was reported, in 

 four cases sleet and in fifteen cases either rain or snow. These 

 facts seem to indicate an occasional uprising, but it is remarkable 

 that so few such cases occurred; and it will be noticed that a 

 difference of temperature of at least 45° between Pike's Peak and 

 Denver often continued from day to day for Jong periods. ... I 

 think we may hence infer that dry air, even when greatly heated, 

 has but little ascensional force" (Loomis's "Contributions to 

 Meteorology," 13th paper, in American Journal of Arts and 

 Sciences). 



Loomis also found that heavy rainfall was not necessarily pro- 

 ductive of cj clones. In his sixth paper, after examining a large 

 number of cases, he says, " We conclude, therefore, that great 

 rainfalls do not generally continue over eight hours, and very 

 rarely do they continue for twenty-four hours, either as experi- 

 enced at one station, or in succession at different places." He 

 arrives at the same conclusion in his seventh and seventeenth 

 papers, and adds, " The forces which impart that movement to 

 the ah- which is requisite to an abundant precipitation of vapor, 

 instead of deriving increased force from a great fall of rain, 

 rapidly expend themselves, and become exhausted." 



Furthermore, after examining a large number of areas of low 

 barometric pressure with which there was little or no rain, he 

 says, " There seems to be no room for doubt that barometric 

 minima sometimes form with little or no rain, and continue with- 

 out any considerable rain for eight hours, and sometimes for 

 twenty -four hours or longer; ... so that it seems safe to 

 conclude that rainfall is not essential to the formation of areas of 

 low barometer, and is not the principal cause of their formation 

 or of their progressive motion." 



" In order to determine the circumstances under which storms 

 originate and ultimately acquire their full intensity," Loomis 

 selected thirty-six cases from the Signal Service weather-maps in 

 which the storm appeared to develop in the United States, and, as 

 a result of a study of these, says, "The first stage in the develop- 

 ment of each of these storms was an area several hundred miles 

 in diameter, over which the height of the barometer differed but 

 little from thirty inches, with an area of high barometer both 

 on the east and west sides, and at a distance of about 1,000 

 miles. In the few cases in which a high barometer is not reported 

 on both sides of the origin, it is because the area of observation 

 is not sufficiently extended. The mean value of the barometer 

 on the east side was 30.43 inches, and the mean distance 1.033 

 miles; on the west side the values were 30.31 inches and 977 

 miles. . . . On Hoffmeyer's storm-charts we frequently find 

 three areas of high barometer surrounding an area of low barom- 

 eter. These areas of high barometer are regarded as one of the 

 causes, and generally the most important cause, of the storm 

 which succeeds. . . . Since the air presses in on all sides 

 towards this area of low barometer, the area tends to assume an 

 oval form, which may become sensibly circular if the winds are 

 very violent, and the centrifugal force resulting from this revolv- 

 ing motion causes a still further reduction of the barometer. . . . 

 Rain is one of the circumstances which increases the foi-ce of a 

 storm, and it invariably attends storms when they have attained 

 considerable violence. . . . Some rain was invariably reported 

 whenever the barometer fell below 29.4 inches, and generally 

 there was some rain reported whenever the barometer fell below 

 29.0 inches. I have found no storm of great violence which was 



