10. The stream entering at b rises in a swamp on section 16, and is joined by a branch in section 15, which rises in section 10. There are numer

ous springs along its course, and it is 8 feet wide and a foot deep, at its mouth discharging as much water into Lake Itasca as the outlet of Elk Lake does. The inlet at c is a small brook, 2 feet wide and a foot deep, that rises in a swamp less than a quarter of a mile from the lake.

This brings me to the largest feeder of the lake, the one entering at d. It is 16 feet wide and 24 feet deep at the place where it enters into Itasca, and is the stream mentioned by Nicollet, in his report of his explorations in 1836, as "the one remarkable above the others, inasmuch as its course is longer and its waters more abundant; so that, in obedience to the geographical rule that the sources of a river are those that are most distant from its mouth, this creek is truly the infant Mississippi; the others below, its feeders and tributaries." The exploration of this stream was the most complicated and difficult of our undertakings, and it was with considerable difficulty that we were able to identify the three lakes which Nicollet describes; but while on the ground, and after the most careful study of the problem, we came to the conclusion that Nicollet's three lakes were those marked on the map as A, B, and C. At first sight, it would seem, from Nicollet's description, that these could not be the ones he referred to; and I have given much study to the points involved, endeavoring to reconcile his descriptions with some other theory. We followed the stream to the first laké at the edge of the hills and through the swamps; and the course of the brook is two miles in length, and seemed like four. Distances on the ground double up very fast when one follows crooked streams, as you will remember when you compare the length of the stream between Elk Lake and Lake Itasca (1,040 feet) with the actual distance between the two lakes (350 feet). If we add to the actual length of the course of the stream from the lake A to its outlet at d, which is in reality 2 miles, the difficulties that Nicollet encountered in wading through the tamarack marsh, we can easily believe that this is the course which he describes as two or three miles' in length. His report makes the distance between the first and second lakes comparatively short, and that between the second and third lakes still shorter, so that there is no other lake which answers the description for the third or higher lake but the one marked C. however, is not the source, at the present time at least, from which Nicollet's stream draws its principal supply of water; and to find that source, after considerable exploration, we were obliged to

This,

go to a lake which has its head in the north-western quarter of section 34. This is the utmost source and fountain head of the water flowing north into Lake Itasca. The lake itself is fed by numerous springs along its borders, and its surface is 92 feet above the level of Lake Itasca. The small inlet from the lake marked I was dry when we visited it, but water runs through it in the wet season. The hills south rise from 20 to 160 feet high, and water has never flowed over them northward. It might be interesting to know how far it flows under them. It is certain that it does, but there is no way to trace its course or distance. All the streams in this part of the basin rise in springs in tamarack swamps, which undoubtedly are fed by water percolating under the hills from lakes and swamps beyond; and no doubt the group of lakes, U, V, W, and X, in the southern part of sections 33, 34, and 35, which spread out to a considerable extent in sections 3, 4, and 5 of the townships next south, are the reservoirs which feed a number of these springs. Beginning with the lake marked H, it spreads northward nearly half a mile. At its northern end the water flows out of this lake in a stream 1 feet wide and 1 foot deep, and, running west about 200 feet, empties into a small lake about 2 acres in extent, marked G. This lake connects with another of the same size about 20 feet to the west of it.

At the time we were there, both ponds were full of moss and bogs, and apparently almost dried up, the abundant inflow of water running out by underground passages as fast as it came in but both lakes show that at some seasons of the year they contain 4 feet more of water, caused by the increased flow in the springtime and in the rainy season. At this time the underground passages are not large enough to carry the water off, and so it accumulates and the ponds fill up. Apparently they once had a surface outlet which is now closed by a beaver dam. The water flowing from the two lakes feeds the two springs numbered 3 and 5. Proceeding to the spring marked 5, we find the water bubbling up and flowing away in a rapid, lively stream, in a direction generally northward. It is fed by springs along its course until it reaches the extreme south-western corner of section 22, where it is 24 feet wide and 8 inches deep, and discharges into a small pond of about 5 acres in extent. This pond is the most remarkable one in the course of the stream; it has no surface outlet, and, from the formation of the land about it, apparently has never been any larger than it now is; but, with the large volume of water flowing into it, we perceive that it must, of course, have a steady

and sufficient outlet underground. This we found to be toward the west, where it bursts forth in an immense spring or pool, marked 2, in the extreme south-eastern quarter of section 21. The lowest point on the hill between the pond and the spring is 12 feet above the level of the pond; and the water, dropping underground, bubbles up in the swamp 200 feet away and 33 feet below that level. You will notice that the stream thus passes underground from section 22 into section 21, and is therefore invisible to one following up the course of the section line, a fact which will be referred to again in a latter portion of this report. Proceeding from the spring marked 2, the water flows in a northwesterly direction, and empties into the lake marked B, - the second one of Nicollet's chain of lakes. The outlet of this lake is on the west side, a stream 3 feet wide and a foot deep, which is joined at a short distance by another from the south. Following up the stream, which joins the main one on section 21, we find it rises on section 28 at a spring marked 3, evidently fed by an underground passage from the pond F. These streams are re-enforced throughout their course by springs which ooze from the bases of the hills that line the tamarack swamps; so that, when the creek leaves lake A, it flows with a brisk current 12 feet wide and 1 foot deep, which is further re-enforced by numerous springs all the way to Lake Itasca. At the point of its discharge into the lake, it is a broad, well-defined stream, 16 feet wide, and 2 feet deep at its deepest point. Lake A is ten feet above the level of Lake Itasca. Recurring to the subject of Nicollet's three lakes, I recall the fact that Nicollet states, that, at a small distance from the heights where the head waters originate, they unite to form a small lake, from which the Mississippi issues with a breadth of 14 feet and a depth of 1 foot. At no great distance, however," so Nicollet says, "this rivulet uniting with other streamlets, supplies a second minor lake," so we were obliged to look for the upper of the three lakes at a reasonably short distance from the lake B. If the spring, numbered 2, would fill the bill as a lakelet, it would meet all the other requirements of the case perfectly. The only alternative seemed to me to be the lake marked C. At present the outlet of this lake is obstructed by two beaver dams, and no water flows from it except what little may percolate under these obstructions. Its principal feeder, marked m, rises in a spring in section 27, and is also nearly dry, but there is a small amount of water flowing through its channel. I leave it to you, or to future explorers, to settle the question as between the spring 2 and the pond C.

There are four small streams flowing into Elk

66

Lake. The first one rises in a spring, the outlet of which flows into a small pond 50 feet in diameter in the north-western quarter of section 34. It leaves this pond a brooklet 6 inches wide and 2 inches deep, and flows with a rapid current to the centre of section 37, where it is joined by another and larger branch coming from a tamarack swamp in the south-eastern quarter of section 27. At the point where it flows into Elk Lake it is 2 feet wide and 6 inches deep. The elevation of the source of this stream at the spring marked 10 is 88 feet above Elk Lake and 89 above Lake Itasca. The largest stream flowing into Elk Lake rises in the north-western quarter of section 26 in a spring marked 13. This is joined, at a short distance from its source, by another branch, which is supplied by a small lake in section 26, marked N. The outlet of this lake is by an underground current, it being closed by a beaver dam; but water has flowed out by a surface outlet at some period, perhaps at the time of Nicollet's visit. Where the main stream enters Elk Lake it is 3 feet wide and a foot deep. This lakelet N in section 26, and its outlet, were to me among the most interesting things found in this region. To my mind they prove conclusively that Nicollet not only explored Elk Lake, but also its feeders. Referring to the copy of his larger map, which you sent me, I find just such a lake laid down at the head of a small stream flowing into Elk Lake from the south-east. This is the most important feeder of Elk Lake, just as Nicollet indicates it to be. The other two streams flowing into Elk Lake are quite small, and originate as shown on the map. We found a dry channel between the lake M and Elk Lake. No water was flowing from this lake, although it probably does discharge some water in the spring and when the water is high. In measuring the amount of water supplied by the various tributaries of Lake Itasca, we found the three streams discharging at b, d, and e, furnishing practically all the perennial water-supply of the south-western arm of the lake; and of this I would estimate that Nicollet's creek furnishes, and the other two, each about 1.

THE WORK OF THE GOVERNMENT SURVEY.

It was an important part of our task to observe the posts and blazings left by the government surveyors, and we carefully ran the main lines with the view of detecting any errors that they might have made. In this part of their work, and also in meandering of the two lakes, our examination proved their work to be correct in every material point. A singular mistake, however, on the government plat, is easily accounted for. The course of the stream from lake H until it crosses the south

line of section 22 is substantially correct as laid down on the government map: but, when they ran the line between sections 21 and 22, this stream was not crossed again, and they naturally supposed it ran due north through the western edge of section 22, and that the stream flowing out of section 21 into 22 was a branch running into the main stream; whereas this is the main stream, which, passing westward under their feet into section 21 by an outlet which they did not see because it was underground, takes its course through the eastern part of section 21, and crosses into section 22 again at the point where the government surveyors had indicated a feeder to the main stream. The two small lakes C and D on section 22, and the two A and B on section 21, would not be crossed by a section line: hence they were not indicated by the surveyors. At a point where the section line between sections 21 and 28 crosses the branch of the spring flowing out of section 28, the course of the stream is through a boggy swamp, and it would hardly be noticed as the stream without going a considerable distance north or south of the section line: hence it is not shown on the government maps. but in place of it is shown a marsh. In all other respects the work of the government surveyors is well done. Their business was to establish section corners, blaze lines between the sections, note all lakes intercepted by the section lines, meander lakes of more than 40 acres in extent, note streams crossed and indicate their apparent direction, etc. Trifling errors will creep into their work; but, when we take into consideration the difficulties they have to contend with, it is not to be wondered at.

Another part of the duty of the government surveyors is to indicate the names of streams and bodies of water, and, in case no modern name has been given to them, to retain the Indian name or its English translation. Following this rule, the name of Lake Itasca, being generally accepted, was retained.

With regard to the name of Elk Lake, Mr. Hall, who was the chief of the surveying party of 1875, recently told me that when he was surveying township 143 N., range 36, he met an Indian trapper at Lake Itasca, who had made this region his trapping-ground for years. He asked him the Indian name of Lake Itasca and Elk Lake, and the Indian gave him the name of 'Omushkos,' or Elk,' for the lake in section 22, and another name, which Mr. Hall has forgotten, for Itasca. As Lake Itasca had a name already, he simply recommended to the surveyor-general the name Elk Lake' for the other body. But the Indians are by no means agreed upon the designa

land shown as tamarack on the plat of land between Elk Lake and Lake Itasca, and back the water up to the narrow strip of high land on the outskirt of Elk Lake, thus bringing the lakes within 80 feet of each other. Whether this was the case when Nicollet was there, I will not attempt to answer. His map would seem to indicate that it was, by the fact that he shows the two lakes so closely connected, Elk Lake so much larger than it is, and the two arms of Lake Itasca so much out of proportion with their present outline. But this can readily be accounted for on other grounds. The shores of the south-eastern arm are abrupt and bluffy, while the shores of the south-western arm are low and swampy. This makes the south-western arm look wider than it is, and the south-eastern arm narrower than it is. The shores of Elk Lake are also abrupt and lined with bluffs, and to one looking south across it, it does not look half as large as it does to one standing on the hills south of it and looking north. Distances across water are always deceiving. The view from different points of Lake Itasca might be sketched by a dozen different parties, and no two sketches would look alike. My impression is, that Nicollet sketched the south-eastern arm of Lake Itasca from some point on its western shore and Elk Lake, and the south-western arm of Itasca from the knoll between the lakes; and when we take into consideration how insignificant is the distance between the two lakes, compared to the total length of both, it can readily be understood why he has shown them as though Elk Lake were a bay instead of a separate body of water. From the nature of the springs which feed the principal stream emptying into Lake Itasca, it is evident that very few changes have taken place in that part of the basin since Nicollet was there, and very few will take place in the next fifty years. The springs that feed it are supplied by underground currents and reservoirs from the lakes and the Height of Land, and, as they cannot be drained, no amount of settlement or clearing will change them. They are among the permanent features of the country. Lake Itasca of to-day is the same in its main features that it was when Nicollet was there, and for a hundred years before. Its level may have been a little higher, the surface of Elk Lake may have been a little lower, Itasca may have spread out over some acres more of marsh, Elk Lake may have been somewhat smaller in its surface extent; thus they may have come more nearly together, and nearer to being one continuous body of water. But the main features of this remarkable basin will remain the same for generations to come, and Lake Itasca will be then, as it is now, the first important reservoir of all the

springs that feed the head waters of the Mississippi River.

Our meteorological observations were taken with an effort at system; but it is sufficient, perhaps, to say that the atmospheric temperature varied from 20 to 70 degrees during the five days that we were at Lake Itasca, and that we had the extremes of clear weather and invigorating atmosphere, and of desolate, soaking rain. The severest storm overtook us when we were within 5 miles of Lake Itasca, and we passed a most unenviable night in an improvised camp. We took the temperature of the water in Elk Lake and Lake Itasca when the temperature of the atmosphere was 51° F., the temperature of the water being 46°. The temperature of the water in the second lake on Nicollet's creek was 42°.

Among the mishaps which invariably attend such explorations, were two that are worthy of note, the loss of my revolver, and the leaving behind, unaccountably, of my copy of the Nautical almanac. I had intended taking the latitude of the northern end of Elk Lake, and aiso establishing a meridian and noting the exact variation between the true and magnetic meridian; but when I got on the ground, of course this was impossible without my tables. Still worse luck followed the observations with the barometer. I had arranged with Sergeant Lyon, of the U.S. signal service at St. Paul, to take simultaneous readings of the barometer. The instruments were adjusted together when we set out for Itasca, but, when we got back to St. Paul, mine read 200 feet higher than his. As there was no way of determining when this change occurred, all that work was of no account. As our first observations were taken at 6 A.M., and the last at 10 P.M., they involved considerable sacrifice of rest, which I am sorry yielded so little result.

The figures given in the first part of this report for the elevation of the crest of the Height of Land are therefore necessarily only approximate, as the variation in my aneroid barometer destroyed the value of my observations, on which I largely depended for this part of my work. The heights noted for elevations between the lakes and for the springs and streams were obtained by the drainage-level, and these may be relied upon as practically correct.

I considered it very fortunate that our trip was made just at the end of a long spell of dry weather such as has hardly been known in Minnesota for years. This enabled us to judge of the sources of water-supply that are perennial in their flow, as distinguished from the surface drainage in the spring and in the rainy seasons. The rain of the night before we reached the lake was

not enough materially to disturb these conditions.

The last thing we did before leaving our camp between the lakes was to erect on the top of the little knoll, in plain view from both lakes and from Schoolcraft Island on the north, a monument to the memory of Nicollet. on which was inscribed the following: "To the memory of J. N. Nicollet, who discovered the source of the Mississippi River, August 29, 1836." This was done after fully exploring the country for miles around; and our little party of three was fully satisfied that fifty years ago Nicollet had discovered all there was to discover of the sources of the Mississippi; and that if he had lived to complete his report on The sources of the Mississippi and the North Red rivers,' and to give to the world his unpublished map, there would have been no chance for any Glazier to confuse the geographical world, or to play tricks upon the learned societies of two continents. We found our work difficult enough, though we were only a day's ride from civilization and the railroad, and though the whole township had been marked off and blazed at every turn by the government surveyors. What, then, must have been the heroism of the invalid devotee of science, who buried himself for months in the unbroken wilderness, and gave his life to the exploration of the frontiers of his adopted country!

I have done my work without any prejudice or bias, and determined only upon finding out and stating the truth in regard to the sources of the great river of our continent whose exploration has commanded the service of so many worthy men in every period of our history.

As a preparation for the survey, I had read every thing I was able to gather on the subject, and I took with me tracings of all the maps of the region. either published or to be found in the government departments. The work has been done by actual survey, and in such a way that I believe it will bear investigation by any surveyor who wishes to check it.

Minneapolis, Minn., Dec. 7, 1886.

HOPEWELL CLARKE.

THE BRITISH SCHOOL AT ATHENS. A WRITER in the Athenaeum states that the managing committee have now drawn up and will immediately issue a series of rules and regulations for this school. Its objects are declared to include, 1°, the study of Greek art and architecture in their remains of every period; 2°, the study of inscriptions; 3°, the exploration of ancient sites; 4°, the tracing of ancient roads and

routes of traffic; and, further, the study of every period of Greek language and literature from the earliest age to the present day. The students of the school will fall under the following heads : 1o. Holders of travelling fellowships, studentships, or scholarships at any university of the United Kingdom or of the British colonies; 2°. Travelling students sent out by the Royal academy, the Royal institute of British architects, or other similar bodies; 3°. Other persons who shall satisfy the managing committee that they are duly qualified to be admitted to the privileges of the school. Students attached to the school will be expected to pursue some definite course of study or research in a department of Hellenic studies, and to write in each season a report upon their work. Such reports are to be submitted to the director, and may be published by the managing committee if and as they think proper. Intending students are required to apply to the secretary, Mr. George Macmillan, 29 Bedford Street, Covent Garden, London. No person will be enrolled as a student who does not intend to reside at least three months in Greek lands. Students will have a right to use the library of the school free of charge. So far as the accommodation of the house permits, they will (after the first year) be admitted to reside at the school building, paying at a fixed rate for board and lodging. The managing committee may from time to time elect as honorary members of the school any persons actively engaged in study or exploration in Greek lands.

The director is to deliver at least six free public lectures at Athens during the season, and at the end of each season he is to report to the managing committee upon the studies pursued during the season by himself and each student. A sub-committee has been appointed to purchase books for the library so far as funds will allow. Presents of books or pamphlets will be gratefully received and acknowledged by the honorable secretary.

THE NATURAL METHOD' OF LANGUAGETEACHING.

No single word has created so great a confusion of thought as the word 'natural.' Its bare etymological meaning is plain enough; but its application is confined by the bounds of no dictionary, and its sense is as mutable as the shifting sands of the seashore. No other word has so often been used by writers as the convenient vehicle of their own individuality. 'Natural' is often simply what one desires from his own particular view to be natural. It is necessary, accordingly, always carefully to scrutinize its use, and thus to discover