CONTEIBUTIONS TO METEOROLOGY. 47 



formed; that is, tlioy drift, in a sense similar to that in which waves, eddies, &c., Cormed on the 

 surface of a river, drift with the current. They advance as the water of the river advances, and 

 in the same direction. IJut we liave found tliat the average direction of movement of areas of low 

 barometer does not generally correspond with the average directiou of the wind for the same 

 'region. This is seen not only in the case of tropical storms l)nt also in storms of the middle 

 latitudes. Near the West India Islands the average direction of storm tracks, wliile the storms 

 are moving westward, ditfcrs about ;50 degrees from the average direction of the wind for the same 

 season of the year. In the Cliina 8ea the average directiou of storm tracks is nearly at right 

 angles with the average directiou of the wind, and the average directiou is nearly the same 

 during those months in wlii(^h the prevalent wind is from the southwest as during those moutlis 

 in which the prevalent wind is from the uortheast. In the western part of the Atlantic Ocean, 

 near latitude 50°, the average directiou of storm paths is about 30° more northeiiy than that of 

 the average wiiul, and in the eastern part of the Atlantic Ocean, near latitude 5.5°, it is almost 

 ;jO° more southerly. In the northwestern part of the United States, between the Rocky Mount- 

 ains and the meridian of 90° from Greenwich, we find places where the average direction of storm 

 tracks is io° more northerly than that of the wind, and other places where it is 20° more southerly 

 than that of the wind. 



65. Bnt it uiay be claimed that the progress of storm areas is not determined entirely by the 

 average movement of the atmosphere, but by that movement which is taking jilace at the date of 

 the storm. I have endeavored in the preceding 'pages to investigate this question, and to present 

 the ev-idence for the above hyjiothesis in the most lavorable light, but if we scan the evidence 

 critically we must conclude that it is entirely unsatisfactory. ' If we claim that the progressive 

 movement of a storm area is due to the progressive movement of the general mass of the 

 atmosphere in which it is formed it seems necessary to admit that a njass of the atmosphere of 

 considerably greater extent than the storm area is advancing in the same direction and at the 

 same rate as the storm advauces. In order to decide whether such is the fact we need only consult 

 a well-constructed weather map of sufiBcient dimensions to include not merely a storm area but a 

 considerable margin beyond it. The storm represented on Plate III had an average diameter of 

 2,500 miles, and during the twenty-four hours succeeding the date of the map it advanced about 

 350 miles towards the northeast. If the movement of this storm area was due to a general diift 

 of the atmosphere then this drift must have included not merely the area within the isobar 30 

 inches, but al.so the adjacent areas of high ]iressure which clung persistently to the low area. 

 This map seems to be too limited to furnish the required information in a form which is entirely 

 satisfactory, and it is desirable to have similar maps for several successive days. The Signal 

 Service maps afford abundant materials for the proposed purpose, and Hoft'meyer's maps are still 

 better, since they include a much larger portion of the earth's surface. If we open Hottineyer's 

 Atlas anywhere at random we shall not find the mass of the atmosphere in the rear of a storm 

 moving forward in the same directiou as that iu which the storm advances. Plates VIII and IX 

 accompanying this pami)hlet are thought to be decisive on this point. The storm maps of the 

 United States furnish similar testimony. Plates I and II show that the general movement of the 

 atmosphere in the rear of a storm is not iu the same direction as that in which the storm center 

 advances, and the evidence would be still clearer if the maps included a larger area. A slight 

 examination of the United States weather maps, or of Hotlmeyer's charts, must satisfy any one 

 that the general mass of the atmosphere surrounding a great storm is not advancing in the same 

 directiou as that in which the storm center advances. 



66. If we follow the progressive movement of a great storm from day to day by means of 

 maps representing the phenomena at intervals not greater than eight hours we shall find that in 

 front of the storm the air appears to be drawn iu towards the center, by which means the pressure 

 on the front side of the storm is diminished. The air thus drawn in towards the center rises to a 

 considerable elevation above the surface of the earth, and its vapor is condensed. In the rear of 

 the storm the exterior air rushes in and restores the pressure on that side, and as the result of this 

 double process tBe i)oiut of least barometric pressure is carried foiward. This movement of the 

 exterior air in the rear of a storm is not necessarily in the same direction as that iu which the 



