wine, which begins to boil on being exposed to the sun for not over half an hour. The vapor of alcohol is then condensed in a worm. The brandy thus obtained is very agreeable in flavor, no matter what kind of wine is used. It possesses an aroma resembling that of Kirschwasser. "It suffices," adds M. Mouchot, "to fill the kettle with water, and then to interpose between it and the worm a receptacle containing sweet-smelling leaves and flowers, in order to obtain all the essences yielded by distillation."
The Florida Cockroach new to American Science.—It is somewhat remarkable that, in certain parts of Florida, living is made almost impossible from the presence, in amazing numbers, of a cockroach not known North. The queer thing is that, while this pest has been long known in Florida, the fact has escaped the knowledge of scientific men. Mrs. Treat lately sent specimens to Prof. S. H. Scudder, the orthopterist, who was surprised to see them, and pronounces them the Periplaneta Australasia of Fabricius.
Meteorological.—A sixth paper by Prof. Loomis is published in the American Journal of Science for January, giving "results derived from an examination of the observations of the United States Signal Service." The object of this important series of papers is to generalize results, using as data the vast amount of observations made in all parts of the United States.
In this paper, Prof. Loomis considers first the period of unusual heat which occurred in June, 1873. The thermometer rose to 108° at one point—Fort Sully—and to 95° and 100° at other points, for several days in succession, indicating a temperature 20° above the mean for the month. The heated area was north of latitude 39°, and east of the Rocky Mountains, and advanced slowly eastward to Western New York.
It appears that the heat was over a well-defined area, which was also an area of depressed barometer. There was also a gentle movement of air from the south into that area, which accounts for some of the excess of heat; but the region where it arose—Colorado, Montana, and contiguous districts—was excessively dry. No northern winds occurred to cool the air, and Prof. Loomis thinks the great excess of heat may be attributed to the hot south winds already referred to, and, secondly, to the accumulated effects of the sun's radiation.
In the second part of the paper the movements, form, and distribution of rain areas south of latitude 36° are considered. When two or more inches of rain falls within eight continuous hours, we have a "great rainfall." It appears that such rainfalls do not usually continue more than eight hours, and only very rarely do they continue twenty-four hours, either at one station or at successive stations.
It is shown that, on the Gulf and Atlantic border, the great rainfalls are twice as frequent on the coast as at 200 miles inland from it. A cause assigned is the rising of the air from the ocean as it impinges upon the land, and the consequent condensation of its vapor. This movement of the air assumes a cycloidal direction, as was found to be the case in a great number of instances, the motion being from right to left, in the direction contrary to that of the hands of a watch. "Hence, every great rain-storm should be accompanied by an inward and cycloidal motion of the air."
In the distribution of fifty-two cases of great rainfall by seasons, it was found that forty occurred in summer and autumn to twelve in winter and spring. Northward of latitude 36° the difference was still greater, being as five to one. It is thus shown that great rainfalls are most frequent when the sun's heat is greatest, and the air contains most vapor.
The hours of the day have a direct relation to great rainfalls. Thus, they most frequently occur before 4.35 p. m., and seldomest at 11 p. m.; only eight out of fifty-two instances are reported by the night observations made at 11 p. m.
The area of greatest rainfall is found to be within that of the cycloidal movement of air, but not at the centre of low pressure. It is almost invariably eastward from it, sometimes more than 250 miles. Thus, a storm-area, as previously shown by Prof. Loomis, usually assumes an oblong shape, the long radius of which is ahead of the
