July i8, 1890.] 



SCIENCE. 



ported, that within two or three days after the appearance of 

 a spot on the eastern limb of the sun, or by the rotation of 

 the sun, storms break forth or greatly increase in violence 

 on the earth, and that it would be possible to use this fact in 

 making predictions of violent storms. It will be easily seen 

 that this view, if true, is of the extremest importance. No 

 attempt is made to explain why it is that after two or three 

 days this spot influence dies down, while the spots are still 

 in full view of the earth for ten days. Since electric energy 

 is transmitted at once from the sun, why should it not begin 

 its action at once upon the earth? We must remember that 

 this action is not a direct one, but electricity must act first 

 upon the clouds or atmosphere, and possibly these upon the 

 earth in turn, before the storm is produced or influenced; so 

 that there need not be a direct connection between the two. 

 The reason that this hypothesis has gained in favor has been 

 already explained, and lies in ignoring discordances and 

 emphasizing coincidences. Our curve of tornado activity 

 furnishes a most extraordinary means of determining this 

 specific eflFect of spots, if there be such. 



The Greenwich, India, and Mauritius photographs of spots 

 show exactly the appearance of each spot by solar rotation, 

 and its area, or, as we may say, its relative intensity. We 

 can easily determine, then, the activity of spots on the eastern 

 limb of the sun during any day. It will be seen that this 

 process is ^ery dissimilar to the one adopted in obtaining 

 the annual intensity ; for that used the spotted area over the 

 whole sun each day, while this method uses the area of a 

 spot for three days only, and immediately after its first ap- 

 pearance by rotation. The years 1874 to 1886 were used, and 

 the tornado months March to September. Curves were con- 

 structed showing both the spot and tornado activity for each 

 day of the above period. An examination of the curves 

 showed, (1) 46 tornadoes coincident with spots, (2) 156 spots 

 without tornadoes, and (3) 393 tornadoes without spots: 

 46 : 593, or 8 per cent, which is insignificant. Next there 

 were compared three successive days of spots, and the same 

 days of tornadoes, during the extremely abundant tornado 

 years 1882 to 1885 and the months April to August. It will 

 be seen that in this comparison every thing was in favor of 

 the tornado and spot- rotation hypothesis. There were 43 

 coincidences, 30 spot groups without tornadoes, and 79 tor- 

 nado groups without spots: 43:109, or twenty-eight percent. 

 It is plain that this hypothesis breaks down completely un- 

 der this investigation. 



General Influence of Spots. 



On observing the sun-spots carefully, we find that there 

 are marked periods during which the spot activity increases 

 and diminishes. In order to compare such periods with tor- 

 nado activity, curves of both were drawn ; and it was found 

 that while a rise in the tornado curve occurred during the 

 greatest activity of spots, yet the converse of this proposition 

 was not true, for there might be a rise in the spot curve 

 without a response from the tornado curve. This may have 

 been due, as has already been suggested, by the masking of 

 the spot eifect at the earth's surface. The evidence, while 

 showing a tendency to increased tornado action during an 

 abundance of spots in at least one «leven-year period, does 

 not show any marked specific action or relation between any 

 definite spot phenomenon and a corresponding response by 



storms on the earth. The subject is exceedingly complex, 

 and merits further study, especially in the line of a removal 

 of any outside influences which would tend to mask the in- 

 fluence of spots at the earth. 



Predictions. 



It has been said that one of the best tests of the advance- 

 ment of a science is its power to make predictions. Unfor- 

 tunately, weather science has easily lent itself to all classes 

 of wise or simple persons, and has suffered by their ignorant 

 attempts at foisting upon it crude and imperfect forecasts of 

 the future weather. An interest attends some of these ef- 

 forts, however, from their curious and incongruous medley. 

 Witness, for example, the famous lines of Dr. Jenuer, be- 

 ginning 



' ' The ass begins to bray, 

 We shall have rain to-day, ' ' 



now so universally quoted. Many of these signs have a real 

 significance: for example, the very old saying, "When it is 

 evening, ye say, 'It will be fair weather, for the heaven is 

 red ;' and in the morning, 'It will be foul weather to-day, 

 for the heaven is red and lowering.' " It has been suggested 

 already that much may be learned from cloud appearances 

 to assist in determining the probability of a tornado. A 

 peculiar livid and greenish color is often seen, or ragged 

 and angry-looking clouds in the west announce the greater 

 disturbance. The appearance of lightning and heavy thun- 

 der usually precedes tlie tornado, though of course both these 

 may be followed by an ordinary thunder-storm. The loud 

 indescribable roar is unmistakable as a precursor of the fun- 

 nel-cloud. 



The question is often asked, however, ''Can there be a 

 prediction of a tornado a day, week, year, or even a century, 

 in advance?" This problem has been most exhaustively 

 studied by so-called weather prophets, and the public have 

 been not a little mystified by the varying claims put forth 

 by each prophet, and especially by the extraordinary success 

 in such predictions that these men insist they have had. 

 The scope of this paper will not permit more than an outline 

 of these theories ; but we may lay down certain rules that 

 should guide us in estimating the value of such jiredictions, 

 and in putting our faith in them. The moon, the planets, 

 and the sun have been the most potent factors in these theo- 

 ries of storm-formation. 



The Moon's Influence. 



The comparative nearness of this body, and the fact that 

 its phases occur in about seven days, which is approximately 

 the interval for the occurrence of storms, have made it one 

 of the most popular influences for weather changes. Soon 

 after the death of the elder Herschel there appeared a sin- 

 gular set of weather tables ascribed to him, and purporting 

 to give weather predictions according to the age of the moon. 

 These tables have been scattered broadcast over this country. 

 It is needless to remark that Herschel had nothing to do 

 with them, as has been shown by his son. The moon is an 

 inert mass, and can have no influence on our weather, ex- 

 cept, as we have just seen, it has a very slight tendency to 

 drive away clouds. Observations have shown that the tide 

 produced by the moon in our atmosphere is only four-thou- 

 sandths of an inch of mercury. 



