ly cut but pleasant valley, till it comes to a point where the chalk crosses its course in a semicircular range, and seems as if it would stop its further progress with a dam nearly four hundred feet high. The river has, however, conquered this wall by boring under it a tunnel fifty feet high and half as wide, through which it rushes in a very lively torrent. In the course of a little over a hundred yards, it passes a chimney-shaped shaft, which extends to the whole height of the mountain and presents an opening more than thirty feet in diameter at the surface. After another hundred yards the stream crosses the floor of a doline (or sink-hole) four hundred feet broad, and then, after crossing a narrow ledge, enters the great doline of St. Canzian. Here the steep, frequently impending rocks on three sides form a gigantic kettle, the western wall of which falls perpendicularly more than five hundred feet. On the southern side a turf-covered slope descends toward the bed of the river, to end abruptly in a precipice of nearly two hundred and fifty feet. Having twice bored the hills for relatively short distances, the Recca continues its course till it meets the rock-wall a third time and excavates a third subterranean channel, this time of thirty-five kilometres, or twenty-two miles. This is the Recca Cave proper, and from it the stream emerges near San Giovanni di Duino into the important river, though a short one, the Timavo, the mystery of the origin of which has been solved by this tracing of the course of its main affluent.
Scottish and Irish Crannogs.—Dr. Robert Munro, in his "Ancient Scottish Lake-Dwellings or Crannogs," draws a parallel between the island-fortifications of the western Celts and the lake-dwellings of Switzerland, and then suggests a connection of development between the crannog and the moated castle of the middle ages. "Crannog" is a Gaelic term, from crann, a mast or tree, and seems to point to the fact that wooden piles or tree-trunks formed an important part in the structure. While the crannogs have several features in common with the Swiss pile-dwellings, they exhibit also some important points of difference, whereas the Irish and Scottish structures are essentially similar. The latter were really fortified islands, sometimes natural, but generally artificial. When complete and in use, they would present the appearance of small islands surrounded by strong palisades for defense, with buildings of various kinds on their surface, dug-out canoes ready for use, and in some cases a causeway or gangway communicating with the shore. They were certainly built with great skill, and with a solidity of which the endurance of parts of them to the present time is the best evidence. Stone weapons have been found in the crannogs, but the bulk of the remains they yield are of bronze and iron, and some of the coins and pottery point to Roman influences. It is generally admitted that even the Irish crannogs are long subsequent in date to the earlier Swiss lake-dwellings. The crannogs, moreover, continued much longer in use than the corresponding lake-dwellings in Switzerland; those of Ireland down to the seventeenth century, those of Scotland to a century or two earlier. They were evidently used mainly for defense. In the more northern and wilder parts of Scotland the wooden structures gave way to stone castles, and in the end, as Dr. Munro points out, instead of the castle being brought to the water, the water was brought to the castle in the shape of a moat. It is certainly possible that some individual castles may be the direct representatives of former crannogs, but it would be very hard to prove that there has been, as Dr. Munro seems inclined to think, any general connection of the kind between the two structures.
Effects of Gases on Insects.—Mr. L. P. Gratacap reports, in the "American Naturalist," respecting experiments he has made upon the power of different insects to live in various gases. The Colorado beetle proved the hardiest of them; it was killed outright in the vapor of prussic acid, which it, however, stood longer than any other insect experimented with, while it was paralyzed for a time in illuminating gas, and died after two hours' imprisonment in nitrous oxide. The effects of oxygen were not very marked; hydrogen produced lethargy in flies, and was bad for snapping beetles, moths, and a wasp; carbonic acid killed flies at once, and threw Colorado beetles on their