One remarkable feature of the storm at this point, which will strike the meteorologist upon a bare inspection of the plot, is, that the current which destroyed the house should have been so far in advance of the axis as to carry fragments 30 rods to the west, before it curved into and mingled with the axis.

A second feature is that such unparalleled violence in the overthrow of the house, should have been exhibited by the outskirts of the storm. When the house fell, the axis could not have been less than 40 rods from it. Usually in narrow tornados, a point 40 rods from the axis is out of the range of violence.

What effect the contour of the ground may have had in producing these results, I leave meteorologists to judge. The portion of the tornado which struck the house, can be easily traced up the valley, which lies near the right hand border of the storms path. It passed up the valley and when near the head, plunged down the slope, and bridging over the bottom with fallen trees, ascended the opposite side, and gradually curved round into the axis. Portions of the house which were first swept around westward into the axis, were subsequently carried by it towards the east. A tea spoon was found E. 30° S. from the house, half a mile distant. A part of an eve-trough near the same place.

A piece of a chest easily identified, was carried N. E. two miles, and fell a mile north of the track of the storm. If the tornado had the form of an inverted cone, like the miniature whirls so often seen, then the transportation of this stick upwards and outwards is easily explained. Other explanations might indeed be given; but they are scarcely necessary unless the rotary action be denied.

Five miles east of the Graham place, the tornado descended a high and steep range of hills which enclose the Whitewater river on the west, prostrating as usual the forest in its descent.

In crossing the valley, it passed through the south part of Harrison. Fortunately the track of destructive violence was here narrowed to two or three hundred feet

The roof of a large church was raised perpendicularly into the air, and carried over the steeple towards the east. A tin ball two feet in diameter on the top of the spire was nearly torn from the iron rod which passed through it. There was no bruise upon it to indicate that it had been struck by a hard body.

The roof of a carpenter's shop was taken off and the posts broken just below the upper story. The lower story was thrown down, and the upper fell back upon the fragments without further injury. Two workmen were in the upper part, and were not aware of their descent till the door, which led to a flight of steps on the outside of the building was opened.

It was supposed by some intelligent observers that the effects produced upon the shop and some other buildings were indicative

of a violent explosive action from within. With one exception I found no evidence of anything more than the usual ascending current in the axis of tornadoes. The windows of the shop were driven in before the roof went off: the gable end of the church. fell inside.

Where both walls gave way, one as far as I could ascertain fell in and the other out.

The wall of a small House was sprung out on the north, so that the ceiling was detached. A lady who was in the house at the time stated that the south door was first driven in with great violence.

The exception referred to was a brick house: A portion of the south wall fell outwards, but whether a current of air had found its way to the inside through a door or window could not be ascertained.

Five miles east of Harrison the tornado struck a dense forest. On account of the level surface and the open country around, the position was deemed a favorable one for giving a truthful record of the mode of the storms action. Here as at the Graham place, the track was divided into sections of three rods each, and the bearings of the fallen trees taken with a compass. These bearings are carefully laid down in the plot. The original division of this part of the State into sections of a mile square by the United States, afforded a means of determining accurately the course of the storm.

In crossing five sections, the tornado diverged to the north one and a half miles.

The following table gives the bearings of the last survey; and is inserted for the benefit of those who may be desirous of examining critically the law of storms. The order of the bearings is the same as in the tables belonging to the Brandon storm,-from right to left across the track.

The progressive velocity of the tornado could not be ascertained accurately. It was about ten minutes in passing from Harrison to New Haven, a distance of five miles. This would give thirty miles per hour. This estimate cannot be far from the truth; from some circumstances I am inclined to believe that it is a little too high.

A few general remarks will close this article.

.

1st. This storm was remarkable for the occasional exhibitions of extreme violence. The Graham place afforded the most striking examples. The house was not merely thrown down, but shivered. A tea spoon was carried half a mile; and a piece of a stove thirty rods. A heavy waggon was carried a few rods and dashed to fragments against a tree. A piece of scantling twelve feet long and three by four inches, was taken thirty rods east, and driven into the ground three and a half feet. Four men tried in vain to pull it out; and the writer in connection with Mr. Graham, succeeded only after digging around it to the depth of nearly three feet. The earth thus penetrated was, with the exception of the first six inches, a stiff, yellow clay. What renders this incident more striking, was the fact that the end of the scantling was not pointed. The cross section presented a resisting surface of eight square inches, and the whole amount of earth displaced was equal to 336 cubic inches. According to experiments made at the New York Navy Yard, a shot 6-8 inches in diameter would penetrate earth five feet nine inches. A shot of three inches in diameter, nearly the resisting surface of the scantling, would under similar circumstances penetrate three feet nine inches. This is but three inches deeper than the scantling was driven. While the weight of the timber in question would have been greater than the weight of a three inch shot; on the other hand the form of the end was not so favorable for penetration. What portion of its velocity was due to falling cannot be told with accuracy. As it fell within thirty rods of the building from

which it was taken, it could not have ascended to a great beight. It entered the ground at an angle of about 45°; and if from this we estimate the velocity acquired from falling, to equal one-half of the whole there is still left a velocity of 500 feet per second due to the wind. The effects of such a wind would be fearful indeed it would move at the rate of 340 miles per hour.

2nd. The involute form of the curve described by the wind as it approached the axis appears more marked in this than in the Brandon storm. The plots of both Surveys indicate this, more especially that at the Graham place. At the same time the cycloidal curve is easily made out. The reverse action of the loop is clearly exhibited at the Graham place; the direct and reverse at Wakefield's. Most of the prostrations at Wakefield's on the right of the axis, were not made by the front of the storm, but took place when about one-half of the cycloid had passed over. Some of the trees were turned outward, showing that they stood till struck by the heel of the storm.

3d. The plunging action of the tornado was a remarkable feature. The ground west of the Whitewater was favorable for observing this, as it was intersected by deep ravines. Ascending slopes were touched lightly in general, or not at all; while the descending were often swept with fearful violence. I am not aware that this peculiarity has been mentioned by other observers. The explanation does not seem difficult; but too much space has already been occupied to state it here.

In conclusion the writer would take the liberty to suggest to observers, that he has found it important to carry his observations beyond the track of greatest violence. Though no trees nor houses may be thrown down, yet valuable evidence to show the mode of action can oftentimes be obtained.

Mr. Laird's house in the vicinity of the Graham place, was on the left of the axis, but too far from it to suffer any injury. The wind was violent but left none of the ordinary marks which could determine its direction. Mr. Laird however stated to the writer that the wind first blew in the south door, and two men were unable to shut it. A moment afterwards, and the north door which had been locked, was violently driven in. The direct and reverse stroke of the loop seem pointed out here. The action of a tornado along the axis only, affords but confused data to elucidate the laws which govern it.

ART. XV.-On the Geographical Distribution of Crustacea; by JAMES D. DANA.

(Continued from vols. xvi, xviii, and xix.)

HAVING in the preceding pages on the geographical distribution of Crustacea, treated of their distribution according to zones of temperature, I now take up the other branch of the subject.The Distribution of Crustacea according to Geographical

Provinces.

In presenting a series of tables in which the distribution of the Genera is given, I divide the surface of the globe, for marine zoological geography, into three sections, the Occidental, the Afri co-European, and the Oriental; the first, including the east and west coasts of America and adjoining islands; the second, the eastern side of the Atlantic Ocean, the coasts of Europe, and also of Africa as far as the Cape of Good Hope; the third, embracing the Indian Ocean, and its coasts and islands, the East Indies and the Pacific Ocean, with its coasts and islands, exclusive of the western coast of America and the neighboring islands. The total number of species in each is given in a separate column.

I make further groupings or subdivisions, by which the several portions of these great regions are distinguished. These general tables are not here copied from the author's Report, and particular explanations therefore need not be given.

The following is an abstract of some of the results:

The division A, includes the Atlantic and Pacific coasts and islands of America; B, the European and West African coasts and islands, from Cape Horn to Greenland inclusive; and C, the coasts and islands of the Indian and Pacific Oceans (America excluded).*

35 (2)

243 (3)

*The discrepancies between the enumeration here and the summaries of the preceding tables, arise from species omitted in one or both, on account of the uncertainty of their localities.

tia, means that 1 of the 24 is identical with a species under A; and 16, that 1 of the 63, is identical with a species under B. So, below.