States comprised four vast inclined planes. Perhaps the natural structure would be more easily and more correctly comprehended by regarding this great physical region as comprising the two distinct mountain systems; the Appalachian on the east, and Chippewayan on the west, and then view the adjacent and dependent slopes or inclined planes as only parts of the mountain systems.

The Appalachian system, as defined on the general maps of the United States, is divided into two great sections by the tide vale of Hudson's river. This very remarkable natural water separation has induced many to consider the two mountain regions on each side, as distinct systems, but we shall soon see adequate element to demonstrate the connexion, although the Atlantic tides are rolled through the gaps which intervene. The Hudson forms, however, a demarcation altogether sufficient to warrant its adoption in classifying the natural features of the United States, and to justify separating, in our description, the Appalachian system into two sections, that of the southwest, and that of northeast.

The southwestern section rises very prominent from the Hudson valley, at two distant points. At the distance of very nearly one degree of latitude north from the outlet of the Hudson into the Atlantic ocean, and between N. Lat. 41° 20', and 41° 30' the tides are borne into the interior of the continent between enormous walls of native rock, rising to from 1200 to 1500 feet above the high water level. To the primitive mariners who discerned the Hudson, this passage, now appropriate ly called the Highlands, must have appeared on a distant view, as an impassable barrier, in place of an opening to the concealed regions beyond. But clearing this narrow, and deep vale, the distant ridges of mountains seem to lie scattered on the far distance, until advancing about one degree of latitude farther north, the masses again rise to between 3000 and 3500 feet. This very striking group now called the Catskill, or Catsberg chain, discharging the Delaware river to the southwest and the Schoharie river to the north, seem, when first viewed ascending the Hudson, as if the chains at their bases rose from the stream, though at the nearest ten miles distant. The oceanic level of the Hudson gives full elevation of view and raises the peaks of the Catsbergs to an imposing height. Amongst the objects brought before the voyager for the first time up the Hudson the mountains arrest far most attention. Those parts of the northeast section of the Appalachian system are seen distinctly, but on the edge of the horizon, whilst the Highlands and Catsbergs protrude their masses into the very presence of the observer.

If we assume the extreme of the southern side of the Highlands, and the northern of the Catsbergs, as the breadth, the Appalachian system contiguous to the channel of the Hudson is something about 90 miles wide. A line, however, parallel to the Hudson, cuts the system rather obliquely, but even when traversed at right angles from Highlands, the system is upwards of 80 miles wide.

From the Hudson valley the chains, in no place clearly defined, range nearly southwestward into

the basin of the Delaware, but with diminished height. Before we advance farther with the general description, we may pause to remove, by stating a few facts, a very general error as respects the connexion of the rivers with the mountains of the United States.

It has been already shown, that the Appalachian chains are not the superlative of hills, though such is the received opinion on the subject. The mountains are considered as the dividing ridge of the waters. So far from this being the fact, the Mohawk branch of Hudson, the higher branches of Delaware, and Susquehannah river all rise to the westward of the Appalachian nucleus. It is, therefore, in a qualified sense that the mountain masses can be considered as the most elevated part of the physical section to which they belong. We shall have another and more appropriate occasion to notice this important feature of the geography of the United States, and now proceed to survey the mountain system under view.

Taking the extremes of the northwestern section of the Appalachian system, the course is between S.W. and S.W by W., but between these extremes there are very considerable inflections. For instance, leaving the Hudson basin, the chains incline to the westward, and still curving in that direction, traverse the Delaware basin and enter that of Susquehannah, ranging something W. of S. W. by W. A curve is again made in the opposite direction in passing over the latter, and when reaching the Potomac basin, the course of the chains is very little west of south. Passing over the northern side of the Potomac basin, the system once more deflects, and with the basins of the Potomac, Rappahannoc, York, James, Roanoke, Yadkin, and Santee, to the southeast, and the great confluents of Ohio, the Monongahela, Little Kenhawa, Great Kenhawa, Sandy, Licking, Kentucky, Cumberland, and Tennessee, on the northwest.

With the higher valley of Tennessee, the Appalachian chains undergo a noted change in their relative courses. From the Hudson basin to the higher Tennessee valley, the chains which compose the system deflect in a general conformity to each other, but with the latter physical section, the two most prominent chains, Blue Ridge on the southeast, and Cumberland mountains on the northwest, first diverge from, and thence incline upon each other, enclosing one of the most remarkable mountain vallies of the earth. This singular valley will be more particularly noticed in the sequel, but we may here observe, that with its southwestern extreme, the distinctive chains of the Appalachian system no longer appear, and the basins of Mobile, and Appalachicola, commence the declivity or slope of the Gulf of Mexico.

Considering the Mobile basin on the southwestward and the Hudson basin on the northeastward, as the limits of the southwestern section of the Appalachian system, the general course does not materially differ from N. E. and S.W., nor does the length vary essentially from 900 miles. The relative bearing and length are easy to determine, however, when compared with the breadth of the sys

tem. An attentive observer, comparing the objects in nature with any of our maps, must very soon perceive, that the chains of the Appalachian system, southwest from the Hudson, are too much restricted on both sides. Again, comparing the courses of the rivers with the ranges of the mountains, will enforce the conviction that the mountain structure prevails and exerts influence over the water channels far from the extreme chains marked on our maps. With these remarks we proceed to examine the chains specifically, premising that as no connected and scientific survey of this mountain system, much of what will be given under the section of chains may be regarded as theory to stimulate to future inquiry.

It may seem, indeed, the excess of hypothesis to consider the falls of the rivers as the outer chain of the Appalachian system, but if we trace a line from the perpendicular rocks on the Hudson, through the lower falls of Passaic, Raritan, Delaware, Schuylkill, Christiana, Susquehannah, Gunpowder, Patapsco, Patuxent, Potomac, Rappahannoc, Mattapony, Pamunky, James, Appamattox, Nottaway, Mehexin, Roanoke, &c., and continue a similar demarcation to the Mississippi, and then compare its inflections with those of the Blue Ridge, the correspondence will appear in a remarkable manner striking. The series of falls in the Atlantic rivers, southwest from the Hudson outlet, is in no place passed by the ocean tide, but these cataracts do not, however, regularly bound the tides, as they are not in every river channel reached by the ocean swell. A remark on the great defect of our maps is here irresistible. On the very expensive maps of Pennsylvania, Virginia, North Carolina, and South Carolina, the limit of ocean tides is alike wanting, and what renders the omission the more to be regretted, is the great importance of such data, in the commercial and physical history of the Atlantic rivers. In fact, the still more obvious if not equally important feature, the river falls are also omitted, but the points are in most instances marked by the rise of cities, and very prominent landmarks are determined by the cities or towns of New York, Paterson, New Brunswick, Trenton, Philadelphia, Wilmington, Baltimore, Georgetown, Fredericksburg, Richmond, Petersburg, Weldon, Fayetteville, Columbia, Milledgeville, and others of lesser note. By drawing a marked line through these points on a map, a great physical limit is obtained, separating the seasand alluvial region from the hilly, or more correctly, the mountainous interior. The most careful view of external nature, cannot in every place trace the line of separation, but if that line is crossed at right angles, or nearly parallel to the general course of the rivers, the features of the country so completely change as to demonstrate the essential distinction between the components of formation above and below the river falls.

The sea-sand alluvial region has a very prominent termination on the northeast, at the mouth of the Hudson, or more correctly at that of the Rariian. The Nevesink hills are detached and belong strictly to the chain which forms Long Island, the

inner verge of the sea-sand alluvion meets the primitive rock in a line very nearly coinciding with southwest and northeast, from the lower falls of Raritan to the lower falls of Delaware, leaving the peninsula now forming the southern and rather larger section of New Jersey on the sea-sand phy. sical tract. Here occurs a phenomenon, nowhere else to be found on the Atlantic slope of North America. That is the rise of a stream on the seasand tract, with its course towards and over the primitive; this very remarkable case is afforded by Millstone river; which, rising on the sea-sand alluvion, flows directly from the opposite Atlantic coast, and towards the Appalachian chains, crosses the primitive ledge, and unites with the Raritan about ten miles above the head of tide water. Millstone, it may, however, be premised, is one of a series of small rivers and creeks which flow northwestward by west, having the Raritan and Delaware as recipients; but, as has already been noticed, the Millstone is the only stream which has its source on the alluvion, and final discharge within the outer margin of the primitive.

The

The series of water courses, which commence, advancing from north to south, with the lower confluents of the Raritan, and which flow inland, is not terminated by Delaware Bay, but is continued on the Chesapeake and Delaware peninsula to its extreme southern point, Cape Charles. The Manalapan and Millstone flowing into Raritan, and the Assanpink, Rancocus, Pensauken, Coopers, Oldmans, Salem, and other creeks entering the Delaware, are followed beyond that estuary by the Elk, Sassafras, Chester, Choptank, Nantikoke, Pocomoke, and numerous smaller streams which, rising on the sea-sand alluvion, find their discharge in Chesapeake Bay.

In the physical features of the small rivers originating on the sea-sand alluvion, and which have the Delaware bay as a common recipient, and those of similar origin on the Chesapeake and Delaware peninsula, there are two marked distinctions. The northwestern slope of the New Jersey peninsula does not exceed a mean breadth of 15 miles, and the Millstone and Rancocus, the longest of the streams it produces, both fall short of thirty miles direct course; the direction of these numerous but short water courses almost invariably between N. W., and N. W. by W. directly at right angles to the great opposing primitive ledge. But passing the Delaware bay, and traversing the Chesapeake and Delaware peninsula, we discover the rivers have changed their courses, and now flow south of S.W., parallel to the Appalachian chains. A critical survey of the Atlantic slope of the United States, superinduces the induction, that Long Island Sound, Delaware bay, and Chesapeake bay, are specifically similar to Albemarle, and Pamtico Sounds; and, that Long Island, the peninsula between Delaware river and the Atlantic Ocean; that again, between Delaware bay, the Atlantic Ocean, and Chesapeake bay, the low capes outside of Albemarle and Pamtico sounds, and finally, the peninsula of Florida, were all protuberances of specifically similar nature.

Long Island is separated from the continent by a sound, through which the water yet passes to considerable depth, and when compared, there are strong reasons to conclude that such was once the case with the Delaware and Raritan; with the Delaware and Chesapeake, and with the Chesapeake and Albemarle Sound. Beyond Long Island, a similar theory will apply to Cape Cod, and, exterior to the limits of the United States, to the peninsula of Nova Scotia.

Calling in the aid of comparative geography, to illustrate the particular features of the United States, and seeking another part of the earth most similar in respect to position relatively, our attention is arrested by southeastern Asia. By actual calculation, the coast of North America from Cape Tehuantepec, crossing the intermediate land, and thence along the coast of the United States, deflects from the meridian 47° 44', whilst the entire south-eastern coast of Asia is within an inconsiderable fraction of N.E. and S.W.; the two coasts only differing in direction 24°. To the remark, how very nearly parallel are these two coasts at a distance of under 6880 statute miles, we may add that in relation to the respective continents, and to the general course of atmospheric currents, the North American and Asiatic coasts, they are the two objects of such immense extent on earth which coincide in so many points of resemblance. The islands and gulfs of south-eastern Asia are more like the contiguous continent, on a wider scale, than are the corresponding features of the United States Atlantic coast. The application of the resemblance here. adverted to, will be seen under the head of climate; we now return to our particular subject.

The sea-sand alluvial region, falling, as we have shown, from the primitive ledge over which the rivers are precipitated, is narrow and interrupted by the ocean reaching the primitive to the north eastward from the Hudson; but thence rapidly widening to the southwestward, is traversed by the Delaware, Chesapeake, and lesser bays, and though penetrated by the tides to the margin of the primitive, the ocean and interior mountains seem to recede from each other, leaving between the Chesapeake and Gulf of Mexico an immense plain of above 600 miles in length, and upwards of 100 miles mean breadth.

The remains of shell banks along the inner margin of this great, and but very slightly inclined plain, fully establish the fact, that over its whole surface the ocean once rolled, and by a slow recession left most of the surface dry land. The bays and sounds are, however, remains of the former shallow ocean border. We are naturally, or more correctly by custom, led to estimate the depth of water on coasts, relative to the draught of ships, and to designate bays, gulfs and river mouths and channels deep, if that depth exceeds the utmost in which the largest vessels float, but philosophically speaking, a depth of even 40 or 50 feet is really shallow when compared with that of any large sea, much less an ocean. But, on the scale we usually estimate water depth, the rivers and sounds of the United States, contained in the great alluvial physiVOL. XVIII, PART I.

cal section south-west from the Chesapeake, are shallow, one river only, the St. Mary's, admitting the entrance of the largest vessels; and in many instances of apparently spacious harbours, the smallest coasters are navigated with difficulty. If the Atlantic Ocean is still receding, the result, in a series of ages, must produce radical changes, and many expensive experiments have been made to determine a much less important problem. When in Europe the abasement of the Baltic Sea was first suggested, the fact was rejected with something like horror, but the only safe test of truth finally put the contest at rest, and compelled the adoption of the result with all its consequences. This very interesting part of geographical history will be again touched under the head of river basins; we now return to our review of the great and general outlines of the physical sections.

Receding westward from the Atlantic Ocean along the zone, between the 30th and 31st degrees of north latitude, the alluvial plain is perpetuated over the peninsula of Florida, and thence westward along the northern coast of the gulf of Mexico, through 7 degrees of longitude, or about 450 statute miles to the estuary of Pascagoula river. Here the vast Delta of the Mississippi protrudes the alluvion to the 29th degree of North Lat., gradually, however, again receding to the northward, between 29o and 30° North Lat. passes the Sabine, stretching beyond the limits of the United States.

There is one feature of the alluvial zone north of the Gulf of Mexico, which demands particular notice, as it is connected with a phenomenon very little known. Between Vacassansa Bay of the western coast of the Gulf of Mexico and the mouth of the Mississippi river, extends an elliptical bay, the longer axis of which is North Lat. 29o. This bay curves to North Lat. 30° 23', between the mouths of the rivers Appalachicola and Pascagoula, and along this part of the coast of the Gulf of Mexico the alluvial plain is narrow, and confined, indeed, to the small counter-bays of St. Joseph's, St. Andrew's, Choctaw, Pensacola, Mobile, and Pascagoula. To the westward from the outlet of Pascagoula, the coast entirely changes character. The Delta of the Mississippi is protruded upwards of eighty minutes of latitude directly into the Gulf of Mexico; and again, to the westward of the Delta to the Sabine, the alluvial zone is spread, as already noticed, about one degree of latitude south of its extent in that direction, east of the Delta.

The writer of this article personally examined the coast of Louisiana and Florida, and, in the progress of that survey, found that the debris, or wood floated out of the Mississippi river, was invariably borne westward from the outlets of that stream. These fragments of timber, or not unfrequently whole trees of vast size, lie scattered along the gulf margin, and afford the only fuel on that dreary coast. The northern shore of the great bay between the Delta and West Florida is covered with a dense forest. This forest is continued on very nearly the same geographical zone to the Sabine, whilst on, and directly westward of the Delta, the much most extensive part of the land surface is denuded of tim 2 R

ber, presenting an immense series of grassy plains, with partial clumps or very thin selvedges of trees along the streams.

It may be placed amongst the most remarkable facts in the hydrography of the earth, that a canal or series of canals, and bay or sound navigation, could be constructed from the mouth of Hudson river to that of Sabine, with a rise of very little, if any, above 200 feet, and seldom more than from 10 to 20 feet; a series which, if formed, would constitute the most important and productive chain of inland navigation which the surface of our planet admits.

Over the long and wide alluvial plain we have glanced upon, flow the numerous rivers having their origin on the Appalachian nucleus in its rear. Those rivers we now proceed to notice, advancing from north-east to south-west; giving the name of the basin to the main stream, which terminates in the oceanic recipient.

Hudson river, as connecting two physical sections, comes appropriately under either head, but when carefully examined on a map, in connexion with the minor basins of Passaic and Raritan, the basins of Hudson and Delaware are more intimately blended than is the former with the contiguous basins of St. Lawrence, Connecticut, and Housatonick. Taken in its utmost extent, the Hudson basin presents a marked anomaly amongst the Atlantic rivers. By the main channel, the ocean tides traverse the primitive chains, and reach the base of the central secondary. The Hudson is formed by two branches; the northern, or Hudson proper, and the western, or Mohawk.

The Hudson has its remote source interlocking with those of the Racket river branch of St. Lawrence, and with those of the Saranac and Sable rivers flowing into Lake Champlain, at about N.Lat. 44°, Lon. 3o E. from W.C. Flowing south southeastward until only separated by a narrow ridge of five miles from Lake George, the Hudson deflects to the south, and at about 50 miles comparative course from the source, receives the Sacondago river from the west. Both those streams rise in the valleys of a ridge or rather chain of mountains, the continuation of the Catsbergs, as we shall see in the sequel. Below their junction the course of the united waters is to the eastward 12 miles direct, but a much greater distance following the channel, to where it is precipitated down a mountain ledge called Glenn Falls, and enters the very peculiar valley formed by Hudson river, Lake Champlain, and Sorel or Richelieu river. There is only one other known valley of the earth having specifically similar features with that we are reviewing. Scotand is divided into two unequal parts by a vale, extending from the Atlantic Ocean on the northwest side of the island, in a direct course nearly N.E. into Murray Frith or bay, of the German Sea, length 120 miles. This valley is well designated by the term glen, signifiying a confined vale between steep and impending hills or mountains. The Caledonian Canal now unites the small rivers and lochs or lakes of the Scottish Glen, and admits the navigation of frigates across the island of Britain.

Naturally, the highest intermediate summit between the Atlantic Ocean and German Sea was only about 70 feet in this interesting valley.

The North American Glen, on a much larger scale than that of Scotland, has the southern extreme in the outlet of the Hudson into the Atlantic Ocean, between Staten and Long Islands, and its northern termination in the St. Lawrence, at the mouth of Richelieu, or, as called in the lower part of its course, Chambly river. The tide part of the Hudson, about 160 miles in length, though usually denominated a river, is really a bay, in the strict meaning of the latter term. This bay is merely long and narrow, and in those circumstances only does it differ from other bays. Rivers and bays differ materially in the opposing curves of their sides respectively. Rivers have opposing sides, with usually very nearly coinciding curves; on the contrary, bays have entering and retiring curves, with no apparent compliance in their inflections. Under the preceding distinction, the Hudson, below the junction of Mohawk and Hudson proper, is a bay.

Immediately below Glenn's Falls, the Hudson, as stated above, enters the Great Vale, and abruptly turns to the southward, and, continuing that course 35 miles, unites with the Mohawk a very little above the head of the tides. The intermediate distance of 20 miles from the great bend of Hudson, below Glenn's Falls, to the head of Lake Champlain, at the mouth of Paulet river, contains the highest summit level, 140 feet oceanic elevation between the tides of St. Lawrence and Hudson. The mind of the voyager on Lake Champlain, in view of massy mountain scenery, can scarcely admit the proof afforded by mathematical admeasurement to convince him that the fine sheet of water on which he is sailing is really only 87 feet above the level of the Atlantic Ocean, and that so small a comparative descent would bring him to the tide level in St. Lawrence.

The whole length of the Great Glen of North America is within a trifling fraction of 375 miles in length, extending from north latitude 40° 34′ to 46° 2′, and differing so little from the meridian as not to deflect one degree of longitude in 328 minutes of latitude. But what in a very striking manner gives interest to both the American and Scottish glens, is the little deviation of either from a straight line between the extremes. The Hudson bay and river, Lake Champlain and Richelieu river, forming the chain of the American Glen, as do Loch Linnhe, Loch Eil, Loch Lochi, Loch Oich, Loch Ness, and Murray Frith, that of Scotland.But to return to our survey.

The fall in the Hudson waters, from Glenn's Falls to tide water, is 104 feet, which has been overcome by locks on the Hudson and Champlain Canal.

Mohawk river, or the northwestern constituent of the Hudson, has its remote sources interlocking with those of Black river and Oswego, flowing into Lake Ontario, and those of the Chenango branch of Susquehannah. The Mohawk valley lies very nearly at right angles to that of Hudson, and pre

senting the second lowest gorge in the Appalachian system between the Atlantic Ocean and Canadian inland sea. The higher sources of Mohawk flow, by a very rapid descent, to the southward about 20 miles, to an alluvial summit level of only 421 feet above tide water in Hudson Bay; the Mohawk summit level being 281 feet higher that of the Hudson and Champlain.

In a state of nature, previous to the construction of the great western canal, in seasons of high floods, the waters were so nearly poised on the level near where Rome now stands, that part of the flow passed into Lake Oneida, by Wood creek, and thence into Lake Ontario, by Oswego river. The discovery of this remarkable feature in the topography of the table land between the Mohawk and Oneida valleys was one of the incidents which led, and naturally lead, to the original design of the Hudson and Erie canals; and the existence of such a depression in the table land between the tide, deep, and narrow valley or bay of Hudson and the middle part of the St. Lawrence basin, afforded an opening to inland navigation very rarely found in the mountain systems of the earth. It may also be remarked, that this plateau or table land is alluvial, and that in the construction of the Hudson and Erie canal, the most extensive masses of rock encountered were found below the summit level; and that feature in the local line of the great canal of New York leads on to the remark, that the mountain chains which compose that section of the Appalachian system, southwest from the Hudson, do not ever generally, much less uniformly, constitute the dividing ridges between the sources of the rivers of the Atlantic slope and those which flow into the Mississippi basin. This relative structure of our mountains, and the courses of our rivers, so contrary to common opinion, demands and deserves particular notice and description, since without understanding this perhaps peculiar structure, no adequate idea can be obtained of the true configuration of the United States.

Departing from the Hudson Bay, and without regarding the mountain chains or ridges, the table land rises slowly, and between the sources of Susquehannah and those of the confluents of Lake Ontario, one or more valleys exist where the summit level between the Atlantic tides, in Chesapeake Bay, and the surface of Lake Ontario falls under nine hundred feet, but passing the basin of Susquehannah, the mountain table land rises more rapidly, and also more uniformly. With some but partial interruptions, the rivers flow, and the mountains lie on a vast inclined plain from the valley of the west branch of Susquehannah southwestwardly, to where the sources of the Yadkin, Kenhawa, and Watauga branch of Tennessee separate. From the latter point, occupied by Ashe county, North Carolina, the great plain declines in the contrary direction, and if the distance is continued to tide water in Alabama and Tombigbee rivers, the length of descent in the two plains is very nearly equal.

If we draw a line on a good map along the dividing ridge between the sources of the rivers, we shall by the operation discover the fact, that the

mountains in no one instance determine the recipient; but, on the contrary, the rivers in most cases rise on plains higher than the bases of the chains, and so far from flowing from, they are discharged through the mountains. The Mohawk, though a confluent of the Atlantic, has nearly the whole of its course to the interior of the Appalachian system. The west branch, and most of the minor confluents of the east branch of Susquehannah rise also within the Appalachian chains. The Potomac rises westward of the central chain, and a similar remark applies to James and Roanoke rivers. Passing the basin of James river, we discover the reverse, as, in a distance of 250 miles from the extreme sources of Roanoke to those of Savannah and Chattahooche rivers, the confluents of Ohio have their most remote sources in Blue Ridge. The entire force of these remarks will be more distinctly seen under the respective heads of the rivers; but we may remark in this place, that the relative aspect of the rivers and mountains of the United States, almost enforces the theory, that the latter were formed subsequently, and have been broken at various points by the streams, to the courses of which these great masses of rock formed impediments. By tracing the real line of separation between the confluents of the two great recipients, the reader will be enabled to more distinctly review the individual basins.

Commencing this river line at the summit level between the valleys of the Mohawk and Oneida, it curves first S.W. and thence W. and N.W. by an elliptic indenting into the basin of Susquehannah, and out of which flow the higher sources of Onondago or Oswego river. This curve, forming the extreme northern boundary of the basin of Susquehannah, is about 120 miles in length, and affords another and striking instance of the conformity of the line we are surveying, with the curves of the adjacent recipient, as the curve which separates the basin of Susquehannah from the Valley of Onondago, complies in its general inflections with the south-eastern shores of Lake Ontario.

Sweeping round the sources of the Tioga branch of Susquehannah, which it separates from those of Genessee, the latter a confluent of Lake Ontario, the river line now inflects to a general southern course 65 miles, to that elevated table land from whence flow northwardly the higher sources of Genessee; north-westward the extreme source of Alleghany, and semicircling round from east to south, and on to west, the numerous fountains of Cowanesque, Pine Creek, Sinnamahoning, and other confluents of Susquehannah. From this plateau the watercourses flow in every direction, like radii, from a common centre. From it, the nearest tide water is the mouth of Susquehannah, at a direct distance of 185 miles; but the head of tides round from Albany in the Hudson, including the entire course of the Hudson, the head of tides in Delaware, Schuylkill, and Potomac, all fall within 210 miles, and of course within 25 miles of being equi-distant from that great crown which stands at the extreme north-eastern source of the immense basin of the Mississippi and the north-western source of the less