concerning the superstitious rites of this mysterious people, it seems natural to conclude that in each of these cases a human victim was sacrificed as part of the funeral ceremonies, doubtless as a special tribute of respect to a person of distinction.

All the skeletons in this mound, except one, appeared to have been buried in a horizontal position with the face upwards. The exception was the skeleton of the aged female found in the grave, which lay on its side; but this may have been owing to the fact that the body had been bent together, perhaps in consequence of age. The skeletons which had received a regular interment all had their heads toward the east, but no such definite position has been noticed in the remains found in other mounds, As the grave had the same direction, this can hardly have been unintentional, although it may have been determined by the position of the ridge upon which the mound stood. The layer of charcoal, not unfrequently found in sepulchral mounds, was wanting in this instance, as was also the evidence, usually afforded by the same substance, that the fire, which consumed the human remains, had been suddenly extinguished by a covering of earth. Possibly the former, as well as other objects of interest, were contained in the outer portion of the mound, which was not examined, although usually everything deposited by the mound-builders was placed near the center; and hence our explorations were chiefly confined to that part.

Such is a brief and incomplete description of one of the ancient mounds of the West, of which at least ten thousand are known to exist in the single state of Ohio, and countless numbers elsewhere in the valleys of the Mississippi and its tributaries. These structures are the only remaining memorials of a race whose history has been buried with them, and from these alone can we hope to learn who this people were, and whence they came. The Indians of this country, although retaining no tradition of this more ancient population, regarded their works with great veneration; but the present possessors of the soil have, in general, little of this feeling, and hence hundreds of these monuments of the past are annually swept away by the plow, and their contents irretrievably lost. A few pioneers in American archæ, ology have, indeed, rescued much that is valuable, but the work is hardly commenced; and a careful and systematic investigation of these various monuments would not only add greatly to our knowledge of this interesting people, but doubtless also help to solve the question of the antiquity of man on this continent, and, perhaps, that more important one of the unity of the hu

man race.

New Haven, Ct., Feb. 1866.

ART. II. On the production of Thermo-electric currents by percussion; by O. N. ROOD, Prof. of Physics in Columbia College.

THE production of thermo-electric currents by friction was observed by P. Erman in 1845,' but I do not know that the subject of the present article has ever been examined with any care. For the purpose of studying the thermo-electric currents produced by percussion, the apparatus represented in the figure

was devised: it consists of a vertical brass wire a, stretched in the manner indicated; on it a brass ball weighing 17 oz. slides freely, the wire passing through one of its diameters. The ball can be raised to the height of from 1 to 5 inches by a string connected with the brass plate p. At the proper moment the ball can be allowed to fall: this is effected by passing the bent end of the rod r through one of the five holes in the brass plate p; by turning this rod through an angle of 90° the ball is set free and falls. The rod is fastened at such a height that when its bent end is in the highest of the five holes, the distance between the lower surface of the ball and the anvil below, is one inch. The holes in brass plate again are exactly one inch apart, so that the experimenter can easily, without altering the appa

Arch. de l'El., v. 477; Inst. No. 614, p. $55.

ratus, obtain at will, a fall of 1, 2, 3, 4 or 5 inches successively, by raising the ball by the string, and using in turn each of the five holes in the brass plate. By this means the production of accidental thermo-electric currents from the heat of the hands is avoided, as the string and bent rod enable the observer to make the necessary adjustments from some distance.

The ball falls on a thermo-electric couple t, consisting of a compound wire of German silver and iron soldered together, or better, of a compound plate of the same metals, the juncture being soldered, as when plutes are used, the couple suffers but little injury from the repeated falls of the ball. In the selection of these two metals for the couple, a suggestion of Poggendorff is followed, who showed that they give a strong current when their juncture is heated. The couple is so arranged that the ball strikes just on the juncture of the two metals, and there by means of the heat developed, produces a thermo-electric current. The juncture of the two metals was generally insulated by silk, &c., to prevent the heat from being immediately conducted off. The two farther ends of the couple were fastened by the binding screws ss, which were in metallic connection with a delicate galvanometer.

Below the couple is the brass anvil A.

A certain amount of heat is developed at the junction of the couple by a given fall of the ball; if now the couple were left in contact with the ball and anvil after the fall, this heat would be rapidly conducted away; it therefore became necessary to contrive, first, an apparatus for raising the ball instantly after its fall out of contact with the couple, and second, some arrangement for raising the couple at the same instant out of contact with the anvil. The former of these ends is accomplished by the lever L, the shorter arm of which is cut out so that when it is pressed down by the spring B, it rests on the anvil over the couple, and is out of the reach of the falling ball. As the sound of the concussion is heard, the long end of the lever is quickly pressed down, and fastened by turning the bent wire at w. The lever thus raises the ball inch above the couple, and the latter itself acting at the same instant as a spring, raises itself by its own elasticity above the anvil. The wires from the binding screws were connected with an apparatus for breaking the circuit, in which small cups of mercury were used. This portion of the apparatus was placed on a table; the galvanometer, however, on a shelf attached to the wall of the room with brass nails, it being found that iron nails exercised a considerable effect on the astatic needle. When thus arranged, and observed with the telescope, the steadiness of the needle was not sensibly affected by a person walking about the room.

The upper needle of the galvanometer was provided with a very fine glass rod, which served as an index, the breadth of the rod being only half of that of the divisions on the galvanometer circle. The end of the glass rod was blackened to render it plainly visible. Directly over the needle, a mirror silvered by Liebig's process was placed at an angle of 45°; the index was observed with aid of this mirror and a small telescope magnify. ing five diameters; in this manner could be estimated.

The falling apparatus was enclosed by wooden screens, also the apparatus for breaking the circuit and the galvanometer. If these precautions are neglected accidental currents are constantly circulating in the wires employed, and no reliable results can be obtained. It is farther necessary after exchanging the couple or handling the binding screws, to allow the apparatus to remain at rest for two or three hours, so that the currents may subside; it is also necessary to select for observation, those intervals of time when the temperature of the room is constant. I may remark, finally, that in spite of all these precautions it is rarely the case that very feeble and nearly constant accidental currents are wholly absent.

The galvanometer was made by Duboscq; after balancing the magnetism of the needles it was found that the copper wire of the coil was so magnetic that the needles took up a position 30°35° on either side of the zero point. I re-wound the frame with American wire, when the needle readily returned to the true zero; upon, however, bringing the two needles very nearly into the same plane, and carrying forward their astasie, the same difficulty was again experienced, when another sample of Amer ican wire was tried with a result which was but little better.

All of these samples when tested in the apparatus used for experiments on diamagnetism, were evidently magnetic, the French sample being strongly so. The difficulty was evaded by bending the needles slightly out of the true plane, when they took up a position nearly east and west, and returned with certainty to the true zero. In this state of inferior sensitiveness one simple oscillation consumed 18 seconds. There were suf ficient indications to show that owing to the magnetism of the coil the needle was more sensitive to currents when standing at 10°-15° than when at 0°; it accordingly became necessary to calibrate the instrument with care. This was done by one of the methods described by Melloni and quoted by Tyndall, (Heat considered as a mode of motion, p. 370).

For degrees under 10° the constant currents employed in the calibration were produced by a small thermo-electric pile with one of its faces turned toward the exterior colder wall of the room, while the other face was directed toward an interior wall.

These, as it were natural sources of heat gave very constant currents, and by partially closing one of the caps of the pile, any desired deviation between 0° and 10° could readily be obtained.

It was found that for about 6° the deviation of the needle was directly proportional to the strength of the current; for degrees beyond this, it was necessary to construct a curve embody. ing the corrections obtained experimentally. The ratio between the first and final deviation up to 30° was also obtained; it was constant for 6°. These latter determinations were important, as after the first deviation the needle, owing to conduction in the couple, slowly sinks to 0°, and only then comes to rest. I was not able to measure with exactitude the time required for currents produced by falls of the ball from different distances to subside, the imperfect results obtained showed that it varied be tween 1 minutes up to 3 minutes, according to the distance fallen by the ball. It having been found then in the calibration experiments, that the force of the current was proportional to the deviation up to 6°, and farther, that the first deviation was proportional to the final deviation for the same number of degrees, in the results given below, where the first deviation was below 6°, the observations actually obtained and unreduced will be given, but where the first deviation exceeded 6° the reduced results will be found.

As the total amount of heat produced by the fall of a body is divided between the falling body and that arresting its motion, it is evident that if the mass of the latter be small compared with that of the falling body, its temperature will, owing to this fact, be correspondingly high; and if the arresting body be a thermo-electric element of small mass, a proportionately large deviation of the galvanometer needle will be produced. If, however, the couple at the moment of the percussion and afterwards, be allowed to be in metallic contact with the metallic ball, the temperature of the couple will by conduction be rapidly reduced to that of the metallic ball, so that the deviation of the needle will be very small, and the phenomena complicated. To illustrate this I give, in table 1, the small and irregular deviations which were produced under these circumstances; the ball, couple, and anvil all remaining in metallic contact after the fall.

In table 1, a newly prepared compound plate similar to that

used in table 4 was employed.