21 



tlie western edge of Lake Micliigan, and tlie line dividing Indiana and Illinois, 

 descending, as low as Tennessee. The frost of September appears to have 

 followed the same direction. It was severe on the IStli in the northwestern 

 States ; but in the northeastern — in Maine, New Hampshire — on the 22d, and 

 in New York on the 23d and 24th. The progress of both seems to have been 

 more rapid from Minnesota into Kentucky than in an eastern direction. At 

 Bardstown, Kentucky, there Avas a heavy frost on the night of the 18th of 

 September, but more severe the ensuing night. It did not travel westward to 

 any gi-eat extent, for it was not general in ]Missouri, and was light in Kansas. 

 In Missoiu-i it spread over the low timbered valleys, but left untouched the open 

 and higher prairies. It followed the lines of the summer rains in that State, so 

 that where fell the rains, there fell the frost, and as vegetation was advanced by 

 the rains, the injuries from frost were light in that State. An observer in 

 Illinois writes that the frost of August 30th " played some very curious freaks 

 unknown in meteorology, and in defiance of all the known laws of that science." 

 In all the extensive correspondence that has reached us of the character and 

 effects of these frosts, we see nothing remarkable, except their unknown cause. 

 The winds, prevailing a few days before the 30tli of August, seem to have blown 

 south and eastward the warmer air then lying over the northern States, creating 

 a vacuum that brought down the cold air of the north, from a point nearly 

 directly north of the middle of Lake Superior, without having any great lateral 

 extension. The INIississippi valley is open both in the north and south, and 

 hence is subject to great changes of temperature. What produced these winds 1 

 Whether the heat was lowered by an extensive and connected scries of hail and 

 rain storms which fell about that time, as suggested by a correspondent, or 

 whether it was by magnetic action, cannot be known until more extensive ob- 

 servations are made. Eut we have seen no phenomenon of these frosts that 

 cannot be explained by our remarks on the frost of August in the report for that 

 month. Heated and cold air, like warm and cold currents of water, do not 

 readily commingle. The colder air, could we see it, would be found flowing 

 down the hill-sides like water, oh account of its greater weight and density, and 

 settling inlow places, over wide districts of country, as in Missouri, as it moves 

 from the north to the south. A correspondent writes that its destruction was 

 greatest at the foot of hill-sides ; that a ridge to the northwest, which stopped 

 the cold air as it came from the north, shielded the crops on the southeast side 

 of it; and correspondents from Geauga county, Ohio, and Onondaga county. 

 New York, state that the highlands in those counties were untouched. The 

 elevation of the table lands of the first is 1,200 feet, and of the second 1,743 feet 

 above the sea level. 



But cold air, when forced by winds into a body of warm air, is divided, pre- 

 senting alternate portions of cold and warm air, which do not readily mingle on 

 account of their different pressures, and hence those phenomena so much noticed 

 by our correspondents, of one field being destroyed, whilst another alongside of 

 it escaped ; of certain rows of corn or tobacco being killed, whilst others in the 

 same field escaped. Sandy soils having greater heat, and more readily radiating 

 it, better protect against frost than the colder and more dense clay soils. But a 

 more complete account of these frosts will be found under the meteorological 

 division of this report, and we have given this notice of them here, that the 

 injuries sustained by the fdl crops might be better understood. 



IV/teat. — This report gives the final returns of all our summer crops, and the 

 view presented is highly favorable. Although the wheat crop was injured in 

 Michigan three-tenths, or 30 per cent. ; in Maine, New Hampshire, Vermont, and 

 Minnesota, tico-tenths ; in Ohio and Illinois, one and a third tenths; and in New- 

 York, Kentucky, Maryland, Wisconsin, and Iowa, one-tenth, yet the aggregate 

 crop of 1863 is greater than that of 1862. In 1862 it was 189,993,500 bushels, 

 and in 1863, 191,068,239 bushels, being an increase of 1,074.739 bushels. 



