Pelham Town

Till

Pia. 32. - Pelham Lake section. A generalized section from Swift Eiver to Fort Elver at East Street, drawn through Pelham, showing the different outlets of the Pelham lakes.

this heavy deposit covers the whole southern slope of the basin. From any point high up on this slope, as on the road to the well-known mineral locality, the asbestos mine, one sees massive accumulations of sand, much of it very fine and all well sorted, which rise in a series of terraces of great regularit}'', with broad, flat surfaces and flat scarps, to a height of 1,000 feet on the east side and 830 feet on the north and west sides of the basin, and are almost wholly wanting on the south side. Westward, the highest terraces end abruptly when they come to the entrance of the basin. At a lower level, 500 feet above sea, they seem to stretch, in the portal terrace, right across this entrance, like a great earthen dam - the narrow notch which the brook has cut being scarcely visible - and they dip steeply down into the basin toward the east in a series of beautiful terraces, and on tlie opposite side descend in a series of terrace scarps and irregular slopes to the level of the normal high terrace of the Connecticut River, at 290 to 295 feet. Looking across to the northern horizon, one sees two slight, broad depressions in the line Avhich joins the eastern and western bounding ridges - by Avhich the two roads pass north from Pelham to Shutesbury - and these mark the southern termini of two valleys by which the waters which deposited the sands entered the basin. Their elevation at the southern end, where they open into the basin and whence the terrace sands extend southwardly, is 820 feet above sea, and they run far north, rising slowly and showing abundant traces of the passage of the waters in their shape and in the tails of sand which lie in the lee of projecting rocks.

But the most remarkable deposit of all is a great ridge (k) of yellow sand (see PI. XIII), 40 to 50 feet high, which starts from the mouth of the eastern of these channels and stretches down the slope of the basin southward with sinuous course, bending at last westward and skirting the brook and running for a long distance out upon the till of the valley bottom, from which it is as sharp)ly demarcated as a new railway embankment thrown across a grassy field. This ridge has sharp slopes on either side, and ends abruptly far in. advance of the remaining terrace sands. Much of it is a rather coarse sand, or rather a sand with many pebbles, and rarely a great bowlder is embedded in it. It drops by great steps, so that one is at first uncertain whether to consider it an esker in the sense now current, or to think it a section of the ordinary terrace sands, from which streams cutting back into the mass on either side - their waters being held up to the sand level by the subjacent till - have removed so much of the loose material that this long ridge remains as an index of the former greater extension of the sands toward the center of the valley.

The entire freedom of the broad bottom of the basin from sand or clay, and the great improbability that any such deposit has ever been present and been so entirely removed that no trace or indication of its presence or of any erosion by which it can have been removed is discernible, make it far more probable that the first explanation is the true one, and that it is a deposit in a temporary ice channel, dropped by the melting of the ice on the steep slope down which it now winds like a great snake.

Indeed, its uniform size for so long a distance scarcely admits of any other explanation. This is represented on PI. XXXV, C, and its end appears in the section, fig. 32, p. 578.

The next striking peculiarity is that, while the grand terraces are heaped high on the east, north, and west of the basin, and across its western portal, the flat bottom is, and has always been, bare till, and scarcely a trace of sand can be seen upon the south slopes - nowhere enough to mark the water stand. Finally, the highest terrace on the east, at 1,000 feet, has no counterpart on the other side.

I have expressed upon the map (PL XXXV, C) the explanation which seems to me most plausible, so far as this was possible. In the first place, the highest sands on the east of the basin, at 1,000 feet, seem to have been deposited by waters coming down the eastern of the two northern valleys when the basin was still nearly filled with ice, and, as these sands are on the same level as the lowest portion of the ridge to the east, the waters would seem to have escaped east into the West Branch Valley. This lowest ground is just north of the section line. The ice barrier (b 8, PI. XXXV, C) placed on these lake beds (1 p^ PI. XXXV, C) may well have been somewhat farther west, as the lower-level waters have worn into and terraced these sands on the west.

Some temporary posture of the ice turned the waters of this eastern stream otit across its surface in a course directed toward the portal, and the sand filling this channel sank to form the great esker as the ice melted. With the retreat of the ice from the basin the asbestos mine valley on its south rim was set free at a level of 830 feet above the sea, furnishing a permanent waste weir for its waters south into the Belchertown Lake, along the course described on page 576, and as the ice still filled the whole Connecticut Valley opposite, it completed the barrier across the portal on the west. The lake basin was then rapidly filled by sands pushed south as great deltas from the two northern valleys, and the waters coming down from the north between the ice and the west slope of Hygeia entered the basin at the portal and sent a third delta into the basin, thus completing the terrace on the western side. The life of the lake at this stage was a very brief one, and when the deltas had advanced halfway across the bottom the ice barrier (b^) failed at the portal and the waters escaped, breaching the portal terrace and moving south by a channel, still well marked, which runs south from the Orient House ceUar. The duration of the waters was so brief that httlo or notliiny was deposited ui)ou the till over the center of the basin, or so little that it has been removed by wind and rain. Yet, starting from this flat bowlder-covered bottom of the basin, one toils up more than a mile over the slope of fine sand of the Shutesbury road to the top of the delta at the mouth of the western valley, and on the other side one can step from an ice-bowlder onto the steep sand slope of the esker, so sharp is the boundary.

If one stands on the south slope of the valley and examines the great sand rampart already described, which is tlii-own across the portal, it seems still intact as when the ice left it. The narrow notch which the brook has cut deeply tln-ough it is barely visible. The terrace surfaces slope 5° eastward into the basin across the portal, as they do in their northward prolongation where they abut on the ridge of Hygeia to the west.

Fortunately the ditch for the main of the Amherst waterworks ran from the west across the flat where the ice rested at the entrance of the portal (Pelham City), giving a complete section of the semimorainic beds that rested on the ice with all their irregularity. It continued past the Orient, exposing the passage beds to the fine della sands, and, passing high up above the brook into the notch which this brook has cut into the portal or entrance terrace, it continued along its southern slope through the whole delta deposit and far out into the central portion of the basin. At first, and nearest the ice, the beds dipped west, and these may be "backset" beds, as Prof W. M. Davis would say,^ or may have taken this posture as a result of the melting of the ice beneath and their sinking westwardly. For the most part the beds dip strongly east into the basin and show that the current came from the west - that is, from the ice.

I append a detailed description of the beds, written when I had no clear view of the meaning of the whole. A describes the till-covered flat outside the portal; B, the sands and gravels deposited against and on the retaining wall of ice and confused by its melting, which occur in decreasing amount eastward; C, the finer eastward-dipping delta sands to their ending in the center of the lake. The section runs parallel to and a little north of the section given (fig. 32, p. 578) where the gneiss ridge southwest of the

Orient cellar has dropped down and is covered with till, while the sands at the Orient cellar drop by a sharp slope to those of the highest terrace of the Hadley Lake.

A. - At the first house in the west village of Pelham, about 320 feet above the sea level, the last trace of the highest bench of the Amherst basin (the Hadley Lake) disappeared, and through the village, eastward into the notch, the coarse till made the surface to the point where the Shutesbury road branches off, as has been described on page 581.

B. - Here the till goes under sands and continues about horizontal or rising slightly for a long distance eastward, so that the increased elevation of the surface is to be referred wholly to the thickening of the overlying sands.

Beginning at the Shutesbury road the sands are, for a distance, thin, and contain here and there a large stone - the largest, 1 foot in diameter. They then thicken to 70 feet (fig. 32), and hold this thickness past the Orient House cellar. Where they begin to thicken - at the house west of the Orient - clean-washed, dry, whitish sand appears in layers from 6 to 12 inches thick, dipping 10° W., alternating with layers of washed gravel 2 feet thick containing many pebbles up to 6 inches, and rarely one a foot, in diameter. The sand layers are evenly stratified and show no finer structure. Sometimes a sand layer waves up and down, and the gravel layer above thickens to fill the depression. A little farther east one layer of fine, well-washed gravel - the pebbles averaging 1 inch - grades westward into sand and eastward into 6 -inch gravel. All along in front of the Orient House the cutting was mostly in sand, showing most beautiful and contorted flow-and-plunge structure, the dip of the laminae being 20 to 30° E., as if urged by a rapid current from the west.

Much of the way, however, the whole is thrown into great confusion. Layers of sand a foot thick, with fine false bedding, stand directly on tlieir heads or are variously twisted in the gravel, or the sand cuts off verticallyagainst the gravel, and vice versa. In one case a band of gravel 2 feet wide was intruded into the sand like a dike; in another a mass of sand 8 feet wide occupied the same position in the gravel, as if the water had worn a channel in the frozen gravel and filled it with its own sands.

Leaving the road at the Orient House the ditch passed across the flat field in front of the house, the beds growing finer in grain and the disturbances gradually diminishing.

The Pelham Lake And Esker

Up to tliis point the section, after reaching' the sand, has crossed the j)l;iiu of the Orient House (shown in fig. 32, p. 578), a pUiin whose surface has been produced by a later erosion. The sands exposed in the cutting, however, are a part of the entrance terrace, and their irregularity seems due to their dei)()sition against or upon a shifting barrier of ice.

C. - From here the section continues at the level of this plain along the north slope of the notch cut by the brook in the portal terrace, about 40 feet below the surface and 60 feet above the brook, and for a long distance runs in clean-washed, very fine white sand, laid horizontally in broad, flat lenticular masses, 1 to 6 inches thick and many feet long, with clayey boundaries which projected on slight weathering. It preserves everywhere its original delicate structure undistui'bed.

Eastward the hill is cut down by erosion and the ditch sinks on its side, showing these fine sands to be in great force and to rest upon till, through which the ditch passes a short distance and rises along the side of a second hill and continues in the sands. There was exposed the following section :

1. Below, a very fine white sand, in layers 4 to 6 inches thick, which ran with very slight undulations for 20 feet or more, and, thinning out, were replaced by others. This was exposed in a thickness of 1 to 2 feet.

2. Above, for 2 feet, was the same fine sand, but showing a most delicate and beautiful flow-and-plunge structure, the laminae dipping 20° E. Above, this sand is limited by an imdulating surface of erosion upon which rests -

3. Two feet of coarser sand, slightly reddish, with sharp regular cross bedding, which dips 30° E. By the weight of the sliding bank above, this has been compressed into curious corrugations.

The ditch rises and sinks in the steep hillside, and the lower horizontal sands (1) can be traced for 40 rods eastward and are present in considerable thickness. The coarser sands (2) extend probably to the top of the hill, about 30 feet.

A specimen of the sand taken from the lowest bed (1) was, when dried, like the finest corundum flour, and consisted of sharp, transparent quartz grains 0.03 to 0.04"'" in diameter, with here and there a scale of wine-yellow biotite.

The flow-and-plunge stnicture of the upper portion of the fine sand, dipping to the east, indicates a current coming from the west, and to this current we may attribute the erosion which prepared the surface upon which the coarser sands (3) were deposited, and the eastern dip in these indicates also that the current continued to flow eastward, or toward the center of the lake.

The Hadley Lake Drainage

The last of the considerable deposits (m t) at hig-h level on the eastern side of the valley dependent upon the obstructed drainage which attended the retreat of the ice occupies the long, narrow valley along the west border of the Belchertown quadrangle, which runs north by the village of Leverett, along Pond Brook, and past the entrance to "Rattlesnake Gutter," and extends south across the town line into Shutesbury, skirting Mount Boreas on the east and ending just south of this mountain, where, south of 42° 25', the north branch of Fort River, which occupies the lake's former outlet, passes into a canyon to reach the open valley above East Street village, in Amherst. This lake was early filled with sands, and the waters carried their surplus into the valley, contributing to form the abnormally abundant sands of the high terrace of the main valley a mile north of East Street village. One follows the heavy sands from Boreas northward, filling the valley clear across, until, just on the Leverett line, they are carried away entirely by Roaring Brook, which comes out of the mountains on the east and has worn out a deep circular basin in the lake deposits, cutting them clear across, and escapes through a deep, narrow transverse valley into the valley of the Connecticut. The high lake beds extend a little way down this valley and stop abruptly, and it is clear that the lake could have existed only so long as the ice remained to dam this outlet. Two miles north of Leverett Center and opposite Rattlesnake Gutter, there is a passage connecting it with the deep valley which runs down the east side of Mount Toby, and by which the waters which entered this latter valley were diverted into the Leverett Lake.

The lay of the beds in this Mount Toby Valley is very peculiar and interesting; a deep water-cut canyon runs the whole length of the valley between the conglomera,te and the crystalline rocks, and as a brook runs south in its southern part and another north in its northern portion, while its center is without flowing water, it is clear that it was not formed by the present streams.

In the iitirtidu of the valley south of the entrance of the side valley troni the Leverett Lake this canyon is bordered by rough ledges and till. North of this the sands coining out of the side valley fill the main valley from side to side Avith a great volume of coarse sands, cut only by the nortlnvard prolongation of the canyon already noted. These sands preserve for a long way a flat surface, but as they approach the north end of the Mount Toby Valley they become gradually kettle-holed on a grand scale, and as the valley widens the sands widen also and preserve their lieight, 400 feet above sea, until they rest against the northernmost spur of Mount Toby on the west and extend up along the mountain side on the east, alDOve the notch by which Locks Brook comes out of the mountains, and a remnant still runs up into this notch. They end abruptly along a broad curve (b", PI. XXXV, C), concave to northward, and sink down by a slope as steep as sand will take to the level of the high terrace (336 feet above tide) which formed the shore flats of the Montague Lake (see p. 615).

It seems to me clear that the ice that filled the Montague basin pressed into the north end of the Mount Toby Valley, compelling the waters of Locks Brook (or Sawmill Brook) to find their way southwardly down this valley and clogging the northern portion of the valley with heavy sands; and so far south as the kettle-holes extend, so far south the snout of the ice was projected into the valley and heavily covered by the sands - the kettle-holes beiiig a measure of the portion which still remained unmelted beneath the sands when the further recession of the ice allowed Locks Brook to run directly westward into the Connecticut, while the great concave slope which bounds these sands (b^^, PI. XXXV, C) on the north marks tlie shore line of the thicker ice, against which the sands were piled.

The waters running southward through the canyon already described cut a deep and narrow channel in the jointed quartzite, where they seem to have formed rapids, and at one place a distinct waterfall. They left no deposits in this narrow part, but found it in flood time an insufficient outlet and turned eastward through the side valley into the Leverett Lake, clogging this with abundant sands. At the south end of this Mount Toby canyon there remains on either side of its widened mouth a delta deposit of rounded bowlders, 6 to 8 inches in size, from which all finer material is removed, as a witness of the violence of the current, and farther south the broad, level South Leverett plain (1 s h), west of Leverett railwaj^ station.

gives evidence of the volume of the sands which were carried through this narrow gorge to form the massive delta thrust into the main valley which now remains as a most important portion of the shore bench of the Connecticut Lake (see p. 639).

The Locks Pond Lake

Following up Locks Brook into the mountains, we find high sands bordering it with every widening of the narrow valley at all levels until we reach the top of the hills and come upon the broad basin of finely sorted sands surrounding Locks Pond. These I have already connected with the line of sands which can be traced southward through Pelham to Palmer. It is clear, thus, that the ice retreated down this valley, but that its shape did not favor the formation of extended deposits, and its position as a deep transverse gorge extending quite across the block of hills between the Connecticut and Swift River valleys was such that it intercepted all southward currents so soon as it was free from ice, and it is curious to see how by de^dous ways it deposited its burden, now in the Leverett Lake, now in the broad delta at the south end of the Mount Toby Valley, now clogging up the northern end of the latter, while at the last it has contributed very little to the filling up of the Montague basin, into which it now enters from the mountains.

Locks Pond now lies in the midst of a broad accumulation of fine sands (1 p^) which, followed eastward by the road to the Mineral Springs House, ends abruptly on the verge of the steep descent to the Swift River Valley, and this seems to have been at one time an outlet for the lake and to have controlled the height of its waters.

Moody Corners

The manner in which the Belchertown notch was occupied by the Pelham River, and in which, by the expansion of this river into the Dwight's station lake, its waters came to occupy also the next pass west - the Bay Road Pass - is detailed in the section on p. 577. These events were the prelude to the complete occupancy of the valley by the lake waters, but at earlier times, immediately following the emergence of the range from the ice, the passes were used as transient watercourses, though no line of esker ridges extends north or south from any one of them.

Notches Through The Holyokb Eange The Granby Road Lake

At the second pass west of the Belchertown ponds, occupied l)y the little-used road from Andierst to Granby, there expands in the center of the pass a broad, flat plain of stratified sands (m t) at a level of 410 feet above sea. The western half is well preserved. The eastern half has been deeply notched by the waters of a spring-fed brook which escape toward the north. North and south the road goes down over till to the lower and later sands, but toward the south the watercourse by which the ovei-flow passed into the basin to the south is well marked by thin layers of sand and gravel.

The Notch

In the middle of the east-west portion of the range a pass 463 feet above sea level has traces of coarse-bedded sands in its bottom, and is continued south in a deep canyon cut in the sandstones and underlying diabase, down the side of which the road goes. This canyon I imagine to have been cut by a torrent coming off the ice to the north and through the notch, or at least to have been occupied and enlarged by such a stream. (See p. 510 and PL XI, p. 610.)

The Low Place And Moody Corners Lake

Farther west and just east of the Holyoke House is another pass, which is, however, tmiied east by the great mass of the Black Rocks diabase, and the waters coming through this pass in the same way seem to have supplied the sands which filled up a small lake (m t) that extended east and west between the two diabase ridges north of Moody Corners. This lake stood at the height of 314 feet and drained from its west end southwardly across the eastern tongue of the Black Rock dike where it is narrowest, and it was filled with sands to great depth, the earlier sandstone having been very deeply scooped out here by the ice.

The Holyoke Notch

The same flood waters continued farther west, and passed at this high level between the ice and Moimt Holyoke into the Springfield Lake, forming the gorge terrace of Dry Brook Hill described on page 661.

The continued melting of the ice at last restricted it to the Connecticut Valley, and melting back from the steep rim of the valley on the east it formed a great waterway all along the eastern side of the Amherst basin, which continued south through the Belchertown notch along the eastern edge of the Springfield basin. In it was deposited a great body of sand and gravel (m t), occupying the position of a lateral moraine of the Connecticut Valley glacier, but ha^dng rather the origin and structure of an esker.-^ It was a great temporary river bed, its eastern bank being the mountain side, its western in part the low front ranges of the gneiss, ridges of till, and the eastern end of the Holyoke range, but for the most part the eastern rim of the great ice mass which still filled the valley and formed for much of the way the bottom as well as the western bank of the stream. The heavy sands deposited by this stream, where they rested upon a rock bottom, are still flat-topped and rest against the rock on the east, with all the peculiarities of a river-bottom deposit, at a height of 50 to 60 feet above the highest terrace of the Connecticut Lake which followed. This is true on the West Pelham plain, and southward along the eastern part of the deposit, while its western portion seems to have rested on the ice, and as the ice melted this was dropped to lower and lower levels, its bedding being much shifted and confused by the process until it came to rest in a great series of kettle-holed sands stretching down to and below the level of the highest normal terrace of the Connecticut. This latter also has abundant kettle-holes, from which I conclude that the deeply buried remnants of the ice had not wholly disappeared when the great lake assumed its place and entered upon its work of carving out its terrace flat in these kame

The normal terrace or bench of the Connecticut Lake is determined by its agreement in level with the highest terrace on the other side of the valley, where great deltas of the finest material, most delicately stratified, mark the highest level of the waters, and where on^ passes from these directly onto rock or till without crossing the complicated series of bedded deposits on the eastern side which I have described.

'It is tlie " moraine terrace " of President Hitchcock, an extremely apiiosite name, showing that he had very clearly grasped the peculiarities of its formation. Surface Geology, page 33, 1860.

The excavations of the Central Kaih-oad on either side of Dwight's station and tlu-ough tlie Belchertowu notch gave me abundant opportunity to study the anatomy of these sands, and especially the peculiarities of the kettle-holes, and this material, with matter derived from other portions of the valley, I have brought together in Chapter XIX, page 665.

I have given the same color to all the kame-like sands (1) which stretch southward at the foot of the eastern rim of the valley just above the highest normal terrace, (2) which extend along the north and west slope of the Holvoke range, and (3) which rise in the central parts of the valley above the level of the terrace flat. The first and most interesting series, to which I have given for convenience of reference the name of the Pelham River, from the place where its remnants are best preserved, is continuously traceable from North Amherst through the Belchertown notch, and from this point great disconnected patches of entirely similar sands and gravels occur at the foot of the eastern valley rim south across the State ; and while one can not assert that they were laid down exactly contemporaneously in the bed of a single glacial river, the fact that they maintain just the same slope as the high terrace makes that the most simple supposition. At all events, their common origin - for they were all deposited between the ice and the valley rim - is sufficiently probable to justify a common color for them all.

The deposits of the Pelham River begin just south of North Amherst and swing round east with the curvature of the rocky slope to the point where the stream received the waters of the Leverett Lake (p. 584). The Pelham Lake drained into this stream, breaching the last great terrace which had stretched across the mouth of the basin, and the stream itself wore deeply eastward into the soft material of this terrace, forming the flat on which the Orient House stood (fig. 32, p. 578), and ran south from here, bounded for a long way on the west by the gneiss ridge which extends south from the west village of Pelham, and washed the side of the mountain on the east to a height of 400 feet. It is a perfect water course, which farther south lacks a western boundary, it having here rested against the ice (b^°, PI. XXXV, C, D), and where it passes into Belchertown a massive di-umlin forms its western rim, and it now ends in a great delta thrust out into the depression in which Dwight's station lies. As one stands on the eastern end of the Holyoke range and looks north, one sees this delta . resting against the mountain on the east and against the low dam of the drift hill on the west, while at a much lower level the high terrace of the Connecticut Lake swings round the drift hill on the west and south. This was the last chapter in the history of the stream. This delta was caused by a breaching of the stream at this point by the melting back of the ice to leave a small temporary lake in its course, in the angle in which Dwight's station now lies, for the stream during its earlier stage seems to have flowed aci'oss this depression upon the surface of the ice, and its heavy sands are continuous at the pro2Der level along the flank of the Pelham Hills east of Dwight's station, across from the delta to the sands of the Belchertown Pass, and a section of them is figured and described in the section on kettle-holes (p. 665). Through the Belchertown Pass it threw down the abundant sands and gravels which stretch from wall to wall of the pass and extend through its entire length. Their greatest height is 337 feet, though many kettle-holes, some of the largest size, disguise the original level of the sands. The three Belchertown ponds occupy three of these depressions.

The railroad cuttings showed most confused and tortuous stratification, abrupt alternation from fine sand in great mass to the coarsest gravel, great bodies of fine sand standing with the bedding almost vertical, as if they had been undei-mined when frozen, and kettle-holes, some partly and some wholly filled up by later sands, as if the ice beneath had melted away while the floods were still in progress.

We may at this point imagine, for the sake of clearness, the following stages in the retreat of the ice from the Springfield Lake basin, or the Granby basin, which is a part of the former, premising that the ice would disappear south of the mountain much earlier than north: (1) When the ice had only melted away from the eastern rim, so as to make a waterway continuous south from the Belchertown notch along the foot of the eastern valley rim; (2) when the ice had melted away from the south face of the Holyoke range ; (3) when the ice had melted back from the north face of the same range for a small distance. We imagine the ice to still fill the whole valley of the Connecticut and to prevail over the western hills, but to have disappeared from the eastern.

1. In the first case the waters passing through the Belchertown notch would have continued southward, and we find, after a brief interruption, coarse kettle-holed sands, which commence at P. Chandler's, opposite the lower pond, at 326 feet above sea, and are quite continuous across Granby, often developed as liiu-s of kaiue ridges or as reticulated ridges, in part bounded on tlie west by a line of drumlins, as opposite the Belchertown poorhouse, in part sloping down directl}' to the high lake ten-ace.

( )n passing the north line of Ludlow a great area of till, leaning on the hi"-h rock border on the east, projects west across the path of the watercourse we are following. Across this area there is a channel of proper height to have been the continuation of the same, but I foinid here only indistinct traces of water action. South of this, however, before one reaches the villaoe of Ludlow, is a broad area, which extends southward beneath the village and to the Chicopee River, of tlie same high sands, which have come for the most part down Broad Brook, and partly also down the Chicopee River from the Belchertown Lake. The same sands appear south of the Chicopee and extend in less amount across Wilbraham.

2. With the melting back of the ice from the southern slope of the Holyoke range the waters passing through the Belchertown notch Avould be deflected westward and southwestward, along the south foot of the range, to fill up the deeply eroded area that extends past Moody Corners clear to the Connecticut, which has since been partly reexcavated by Bachelors and Elmers brooks; and as the ice retreated still more the waters would carry their load of sand and gravel directly into the wide Granby basin to build up the broad plain surrounding Forge Pond. The present condition of the gravels extending south from the notch makes it certain that this was the last course of the waters. The coarse gravels extending through the notch where they surround the third pond are still very coarse, pebbles 6 inches in diameter being abundant. Here the gorge expands and the gravels extend across the widening basin, growing gradually finer. Halfway to the east line of the town (Granby) they are 4-mch gravels; at the town line, 2-incli gravels; where the wood road cuts deeply into them north of Moody Corners and in South Hadley they are exposed for 35 feet as fine, well-bedded sands.

3. By the melting back of the ice from the north side of the Holyoke range the river expanded lake-like along its northern foot and, aided by waters coming directly across the ice, a great body of sand was rapidly carried in here. This extends west just beyond the "notch road" to South Hadley and stretches through the much lower notch near the east end of the range, and is continuous with the sands in the Belchertown notch.

4. Finally the waters expanded so considerably that the Pelham River emptied into the lake thus formed, building the delta already described above Dwight's station, and soon disappearing.

The Sunny Valley Lake

The only other large accumulation of glacial sands on the eastern side of the river occupied the Sunny Valley in the northern part of Warwick, and extended across into Winchester, New Hampshire, and was drained by the Valley Brook into the Perchee Brook and thus into the Connecticut. This lies for the most part outside the limits set in this work.

The sands which rest against the northern slope of the Holyoke range extend west to a point a little beyond the notch road, and are conspicuously absent from the rest of the north face of the range farther west. This may indicate the distance west to which the waters penetrated from the Dwight's station lake, or enlargement of the Pelham River (see p. 589), or these sands may have extended farther west and have been swept away by the later lake waters. Sands at the same level above the high terrace begin again on the west side of the river, just east of the old road to the Nonotuck Mountain House, and extend thence along the whole face of the Mount Tom range in Northampton and Easthampton, in a great mass of ridgy sands and gravels, in which the high terrace flat of the lake is cut. Similar sands at the same level also cover White Loaf, in Southampton, and the ridge east, which rise as islands in the broad sand flats of the Hampden plains. They are not given on the map. They had clearly a common origin, having been swept in between the mouiitain and the ice, or off the ice onto these islands after the ice had uncovered them, and they stood out like nunataks above it. They have served a common purpose in the later economy of the valley, as they furnished, I have no doubt, a large portion of the material carried south by the two channels on either side of White Loaf, and spread as coarse gi-avels around Hampden ponds, which dwindle farther south in the broad plain to the fine sands of "Poverty Plain," here, of course, reenforced by the abundant contributions of the Westfield rivers.

Ohaptee Xviii The Granville Lake

As the ice retreated northwesterly, still sending great lobes down the Connecticut and down the Westfield River, it abandoned the high valley which occupies the whole middle of the town of Gi-anville, while its Connecticut lobe still closed the outlet of this valley at the northeast corner of the town and the gap of Munns Brook, which is cut so curiously through the middle of the eastern rim of the valley. This rim is caused by the greater durability of the vertical schists of which it is made, which strike north and south and form an impassable barrier along the whole eastern side of the town, except that it is cut asunder by this deep notch in its middle, down which a road once ran.

This broad valley, which extends across the whole length of the town and a long way into Connecticut, was filled by a great body of sands, now finely terraced down by brooks which i-un out of the basin on the north, east, and south (g P, PL XXXV, B). Its height is plainly deteiTuined by its southern outlet, where the brook has a rocky bottom, and it is clear that when the sands filled the basin the northeast and the east outlets (b^) must have been closed, since, on being opened, the brooks which occupy them cut down deeply through the sands before they reached the rocky bottom of these outlets, proving that the latter were preexistent and deep enough to have kept the waters at a much lower level if they had been open.

The North Granville Lake

Another lake of great extent stretches from Granville into Blandford, suiTOunding Cobble Mountain (g P). Its coarse sands reach a gi-eat depth. It was drained by the setting free of the South Branch of the

Mon Xxix 38 593

"Westfield River tlirougli the wildest gorge in Massachusetts, deeply rock cut, and abounding in very large potholes. One is inclined to surmise that some portion of this erosion may date from Glacial or eai'ly post-Glacial time.

The Westhampton Lake

A much more extensive and interesting lake than the one mentioned in the last paragraph occupied, at a much lower level, all the eastern part of Westhampton, about the headwaters of the Manhan River (g P, PL XXXV, B). It wound sinuously among great islands of granite, and extended south into Easthampton, where it passed over a rocky sluiceway, which fixed its level, into the headwaters of the South Branch of the Manlian. From Loudville south into Southampton, around the east flank of Great Mountain, the sands of this lake extend out freely into the Connecticut Valley, without rocky support on the east, and end in a terrace scarp, below which a broad slope of till extends down 170 feet to the normal high ten-ace of the Connecticut Lake. It is plain that the ice of the Connecticut River glacier furnished the bank of the lake within these limits (b®). This was at its greatest size. Subsequently it seems to have persisted, with diminished boundaries, and to have received the overflow from the Williamsburg Lake, farther north, when the ice had retreated north so as to set this free, and so to have for a long time I'etained connection with a long and complicated series of watercourses, which extended back north to the valley of the Deei'field River, and ceased to be occupied when this valley was abandoned by the ice.

So soon, however, as the ice set free the gorge at Loudville the lake was tapped in its upper portion, and the greater part of its area is now drained through this channel. Two watersheds, however, developed in this area, one at the south, setting off a portion to be drained by the South Branch of the Manhan, and another at the north, which drains through the Roberts Meadow Brook into the Mill River. These streams, especially the middle ones, have cut a fine set of upland tei-races in the sands of the ancient lake, and as one sees these brooks leave their upland meadows and plunge into deep gorges on their way to the valley below, one is inclined to ask how far these gorges were preexistent and how far they have been cut by the streams since the time of the lake. In some cases the disappearance of the lake may have been due to such a cutting down rather than to the removal of an ice barrier or the sudden sealing up of the sources of the glacial waters. I have for the most pai-t used the latter as a working- hypothesis. In several places, as in the Counecticut gorge above the mouth of Millers River, much erosiou of the rocks has taken place since the time of the lakes.

The Williamsburg Lake

The Mill River at Williamsburg village has cut its terraces in a great body of coarse sand, whose flat surface is 33 feet above the stream. TTntil the ice had so far melted back that the Deerfield River was open, the overflow of the Ashfield and Conway lakes, described below, reached this lake by way of Mill River and Joe Wright's brook, respectively, though, as the waters left in their passage through these narrow gorges no deposits to attest their former presence, there is on the map an apparent break in the continuity of the deposits between them, of considerable extent in the case of Mill River. The proof of this former continuity is given in the description of the other lakes.

The sands of this lake (g l\ PL XXXV , A) can be followed down South street to the south line of the town, where they divide, one band going south to join the Westhampton Lake, the other southeast in Northampton to where it widened into the Roberts Meadow Lake, whose waters regained the valley of the Mill River at Leeds, and also passed south into the valley of the Connecticut, down the deep, empty gorge west of Roberts Hills. This gorge has, as in so many other cases, a brook heading in its bottom and running north, and another brook heading farther south and running south, while the gorge is continuous and of uniform size and depth from the Roberts Meadow basin to the open Connecticut Valley. It seems to me a product of subglacial drainage or of the obstructed post-Grlacial drainage I am here tracing.

The Beaver Brook Lake Above Leeds

This small lake lay encircled by high hills in the east part of Williamsburg and was drained by a deep gorge through which the brook above named now flows to join the Mill River at Leeds.

The Deerfield River Lakes

The obstructed drainage south of the Deerfield River (PI. XXXV, A) was most curious and complex. That it was, mutatis mutandis, the counterpart of the drainage south from the Millers River (p. 573) comes out very clearly. The ice melted back across the high, irregular area south of the river with a front rudely east and west, or a little north of east and south of west, and the deep transverse valley of the river runs northwest, while deep longitudinal valleys extend south from it. The ice thus set free the lower portions of the Deerfield Valley first, but the Connecticut River glacier dammed its mouth for a long time, thus turning the waters south across Conway to form the Conway Lake, south of Bardwells Ferry, whose waters escaped south across Hampshire County to enter the valley of the Connecticut by way of the southward ramifications of the Williamsburg Lake.

At the same time the ice had set free the Deerfield up nearly to the mouth of Bear River in Conway, and still sent a lobe up this valley for some distance, from the south end of which the waters escaped that formed the broad Bear River Lake, seen in the middle of PI. XXXV, A, which drained into the Ashfield Lake. This latter was supplied by a lobe of the glacier which, west of Shelbiu-ne Falls, followed up Clessons Brook to beyond Buckland Center (b"). This lake drained east into the Conway Lake arid sovith by way of the headwaters of the Mill River into the Williamsburg Lake. Still farther west the waters gathering from the ice which fill ed the Deerfield Valley below escaped sotitheastwardly from the extreme northeastern corner of Hawley and followed Clessons Brook down as the preceding current followed it ujd, and the two currents met at the sharp bend of the stream at Buckland Four Comers and joined to form the Ashfield Lake mentioned above. Another current from the ice going south up the valley of Chickley River formed the lake which occupied the middle of Hawley. For this lake I could discover no southward outlet, but no doubt one existed.

Three points will be noticed in regard to this drainage : (1) The waters moved up old valleys and filled them to a level determined by passes far to the south, over which the waters continued into another drainage basin ; (2) after the retreat of the ice, brooks heading up near these divides ran north, carrying back north a good portion of the gravels which had been cai'ried south ; (3) it is only near the mouth of the Deerfield River that the deposits of these lakes extend north to the river itself; farther west they begin some distance south and the valleys between their beginnings and the river are empty and bowlder strewn, while on the east side of the Connecticut the opposite order holds, viz: at the head of Millers River the corresponding deposits extend up to the river and beyond it to the north, in its middle course up to the river and in its lower parts not quite up to the stream. In the tollowiiig- more minute description of these lakes more stress is hiid upon the shiftiugs of the ice front during the formation of each hike and on the character of its deposits.

The Deerfield River And Its Tributaries On The North

This stream hes so nearly in tlie direction of the melting of the ice that it was a main channel for the exit of its waters, and was not itself encumbered, except that once the readvance of the ice in the Connecticut Valley threw a dam across its mouth. The same is true in the main of its northern branches.

The deep, tortuous, and most picturesque valley of the Deerfield widens slightly in Cliarlemont and Shelburne, and here a considerable body of sand was gathered, and in several places pretty series of intermediate teiTaces have been cut in this deposit of the flood period and above the present flood plain of the river. Exactly as in Russell the Westfield River encircled a great hill in midstream, so the flooded Deerfield surrounded a great hill in Shelburne, and the railroad makes a short cut through the abandoned waterway.

Below Shelburne Falls the river runs in a deep canyon till it reaches the Connecticut, and it is hard to say whether the great height of the Chaiiemont-Shelburne beds is due entirely to the setting back of the waters above this deep, narrow gorge or partly to possible ice dams. In all this distance there are only traces of terraces referable to the flood period or to any subsequent time, but the stream runs deep in a rocky gorge.

It is characteristic also of the streams that enter the Deei-field from the north that either they were open waterways during the melting of the ice - and they are therefore now deeply sunk in empty valleys with traces of their high flood terraces in coarse gravel beds left as remnants in sheltered places and with narrow and interrupted flood plains - or they seem to have been occupied by the ice during the height of the flood, at least in their upper reaches, and so are wholly empty of anything except till above the low level of recent flooding. This seems to have been the case with the Deerfield itself above the tunnel entrance, and fine terminal moraines (1 m) have been thrown across the stream at several places during the recession of this last lobe of the ice.

The curious "delta terraces" which appear where the tributaries on the north side meet the main stream are discussed elsewhere (see p. 605).

The ti'ibutaries on the south side have for the most part cut their waythrough heavy glacial lake deposits, and their facies is thus extremely different from the north-side tiibutaries, as the former are bordered in their upper courses by broad, heavy deposits of high-level sands and gravels in which they have cut their teiTaces, while on the north the streams come down in narrow gorges, and only in Coleraine does one widen to form any considerable meadow. This in part explains why Conway and Ashfield are more flourishing villages than the northern tier of towns from Leyden to Monroe.

The Conway Lake

Just at the Conway station the train crosses at a dizzy height the South River where it enters the Deerfield. The station is about at the level at which the waters stood at their highest flood after they had obtained free passage to the Connecticut. The Deerfield now runs in its rocky bed nearly 200 feet below, and the road from the station toils iip over heavy sands another hundred feet to a broad, level area of sand, bounded on the south, east, and north by the South and Deerfield rivers, which, when the ice obstruction (b^^, PI. XXXV, A) below was removed, cut down through these sands into the rock so quickly that they did not wear back at all into the sands. A great triangle between the rivers, bounded on the west by the road from the South River to the Deerfield, and a mile on a side, is occupied by this great body of sands (g P) ; in its eastern and larger part it is quite horizontal, at about 460 feet above sea, while in its western part the sand rises nearly 100 feet higher and is covered by till, as if, at an earlier stage of the lake, the waters had stood above this higher level and brought up the sands to that level, and then oscillations of the ice brought in a covering of till over part of the area and determined then or later a lower level for the lake, down to which its sands were terraced. The area requires more study than I could give it. The sands at this lower level are continuous as a broad band south, up the valley of the South River, widening over the area of the village of Conway and receiving there a body of sands which extend up the South River through Burkville, in that part of its valley which runs northwest.

The South River enters the town running east, and holds this direction a mile (the latter part is an empty valley); then it bends south a mile to Conway, and in this part it was occupied by a glacial current which moved south to meet at Conway village the current moving south up the valley of South River from the station. It seems to me probable that when the ice fii'st melted nut of the great triangle at the station it still filled the deej) valley at and north of Conway village, while the Connecticut glacier danuned up the mouth of the Deei-field and held the waters up to a height of ])('rhaps 550 feet. After this the ice abandoned the South River Valley as far \\-est as the first bend in the river mentioned, above a mile from the west line of the town. Then the waters, passing down the two lobes of this valh'A- and joining where the village now is, filled the valley to the height of the ground where the academy stands, and all the area to the south (now reemptied by erosion), and escaped southwardly by the narrow pass by which the main road goes south into Whately. As one goes south from Conway he sees the broad, level gravel plains, 100 feet above the village, which surround the town and extend south, passing continuously and at the same level into the fiat, plainly waterworn bottom of the narrow canyon, and it is clear that this canyon has fixed the level of the lake. We follow this watercourse southward easily and find it expanding into a lake around the headwaters of Roaring Brook, and branching to follow this valley down a mile southeast, ending abru^Dtly. The Connecticut glacier would seem to have still clogged this valley, and I have so represented it (b^^, PI. XXXV, A). The main stream channel continues south into Whately and, reaching the headwaters of West Brook, widens considerably and then contracts at West Whately, where it bends sharply west into Williamsburg. It is quite remarkable that it should thus bend, for the open West Brook Valley turns here toward the east, and, after 50 rods of empty valley way, expands on either side of the brook's bed ; and 200 feet below the sands of the above watercourse a great area of heavy sorted gravels begins, which continues southeast with the brook for a mile, widening to above a half mile and ending abruptly with a great scarp which overlooks a broad, sandless valley. Remnants of ice on the east of the main-valley watercourse must have kept it out of the Roaring Brook Valley, as above, and also deflected it at West Whately (b^^), and the ice, being breached, let the waters tlirough in a flood, which dropped this great volume of gravel as it expanded into the open valley after the manner discussed under the section concerning high-level deltas (p. 605), or the ice may have erected a later barrier lower down the valley (b^"). In Williamsburg the stream expanded into another basin, now a broad sand plain, and escaped south through a curious narrow canyon, in the bottom of which Wright's brook now rises from a large spring. Down the course of this brook the waters passed to join those of the Williamsburg Lake.

The Bear River Lake

The fullness with which all the evidence concerning the formation of the deposits in this Ijasin can be traced leaves little to be desired. The basin is a broad one, extending, with its greatest dimension east and west, across the line between Ashfield and Conway (middle of PI. XXXV, A). It is surrounded on all sides by high ground, except a narrow passage at the northeast corner, and sends tlu-ee great lobes northward among the hills. In its southern part it is filled to a great height by an enormous volume of coarse sand, especially in the area around the cemetery at School No. 2. The height to which these sands were brought up was determined by the height of the col at the southwest portion of the basin by which the waters escaped into the Ashfield Lake along the road which runs south to Ashfield Plains. These sands were brought in while the ice was retreating from the valley and still filled its northern parts, as in all the northern portions of the basin the sands occupy a lower level, and it is plain that they were never heaped up to the height of the south-side sands and then eroded. It is therefore probable that the north side of the basin was still filled with ice when the latter were brought in, as they must have come from the north.

When the ice had completely left this basin it marked its next halting place with beautiful clearness. The water maintained its former level after this retreat, this being conditioned by the col in the southwest corner. It carved a broad bench in the till around the north border of the lake and brought in great bodies of sand over its bottom, but not enough to fill it. If these sands are followed to the northeast corner of the basin, they are found to extend a distance along a narrow rock-bottomed valley, scarcely covering the rocky floor. The rock bottom of the valley then sinks rapidly, while the sands widen somewhat and continue at the old level, thus filling the deepened valley with beds of very great thickness. At a certain point in this valley they end abruptly (b"), and one goes down by a great lobed scarp to the deep valley bottom, and for 100 rods downstream (northwardly) the steep valley sides are strewn with glacial bowlders to the water's edge. These sands are grooved in the middle by the Bear

The Ashfield Lake

It is plain tliat the ice dammed this valley and that the sands were hoai)ed up against it, and that it then retreated and left the sands to cave into the o-reat northward-facing scarp after the changes in the ice farther up the Ueertield Kiver had opened up other channels of escape for the waters.

The gi-eat bodies of flat sands in the middle of Ashfield (middle of PL XXXV, A) have naturally, in this extremely hilly country, given the village the name of Ashfield Plains. The Ashfield Lake is represented by a peculiar body of sand surrounding a great rocky hill which overlooks the village. At South Ashfield it turned west and di-ained down a long valley to the east into Conway, and from this point I was uncertain as to its course. The valley it has followed to this point (South Eiver Valley) runs east into the Conway Lake. It is empty of sands for a mile, and then begin deposits which are continuous into the Conway Lake.

The Ashfield beds seem to turn south just at the town line, up a branch of the Soxxth Eiver, whose valley they fill for a long way soxxth, to the soxxthwest corner of the Greenfield qxxadrangle, and then repeat the opei'ation already described at the nox'th end of Bear River Lake (p. 600). The valley in which we are followixig up the deposits ends ixi a cxxl-desac, but is continxxed soxxthward at a mxxch higher level by two valleys, one of which laxns through a corner of Plainfield and along the east of Moores Hill, in Groshen, iix a naiTOw canyon, and thence down the steep slope into Williamsbui-g Lake, and the other more directly soxxth, by City Pond, into the valley of Mill River, and into the same lake. I have represented the deposits in disconnected patches in both these courses, because much of the way the valleys are so naxTOw that all traces of these eax'lier occupaxits have been swept out by the wild floods of the present brooks. It is, of course, probable that the Williamsburg Lake came into existence as soon as the ice melted back from it, and that, as the ice retreated, the two courses I have last traced were long used by waters at various stages of this retreat. I am here, however, following oxxt the last occupation before the opening of the Deei-field River.

It is also quite possible that the Ashfield Lake did drain into the Conway Lake, flowing back sonih to fill the cul-de-sac described above without overflowing to the south, and that for a mile east along the South River Valley it left no deposits. My attention was not closely directed to the point when on the ground, and my opinion was formed from a view of the entire area from - the top of the highest hill after I had gone carefully over the whole region.

The Buckland Lake

If we follow the outer contours of the Ashfield Lake where the sands border on the rocks, we shall find them converging just north of Great Pond upon a narrow, rocky canyon, and it is plain that the waters came through this passage for a long time and with great force, bringing the sands which extend south from its mouth. The drainage of Great Pftnd is southward, but a small rise of its waters would send it north through this gorge. On entering this gorge one expects it to rise among the hills and terminate as a mountain glen, and expects to find the brook which has brought down the great volume of sand, but a short distance north the valley widens somewhat and sinks 300 feet with great suddenness, so that it has been very difiicult to carry the road down to its bottom. One sees immediately that the ice must have filled this deep valley (b^^, PI. XXXV, A) when the waters swept across its back and through the narrow gorge bearing the great volume of sands which now form the Ashfield Plains, for otherwise the deep valley to the north must have been filled first. That it was not filled and then reeroded is certain from its bare, rocky, and bowldercovered sides and from the abundant openings made by the new road carried down to the valley bottom. Taking this road, we go down sharply to the valley bottom over till, and along the bottom for a short way also over till, when we come suddenly upon a great bank of fine, well-bedded sands, about 33 feet high, with a slope as regular as an earthwork, facing us (i. e., facing south), and extending right across the valley and resting against its walls. It is like a dam, only breached at the center by a brook which runs north, and we seem to be in the bottom of an abandoned mill pond. Climbing to the top of the slope, Ave find it is the southern termination of a great body of sand which once filled the A'-alley from this point north across Buckland to Buckland Center, and which, though now largely eroded, can be ciisilv t'nllowL'd to this point uiul not farther. It maintains a level about 200 feet below the Ashfield Lake and 100 feet above the Deerfield River terraces.

It is clear that after filling this lower valley for a long time and allowing the waters to transport the great body of sand into the Ashfield Lake the ice retreated north to Buckland Center and stood there for a time (b"), maintaining a lake of great depth, which still drained through the narrow canyon and across the sands of the Ashfield Plains, producing the exceptionally large amount of erosion of these sands, while the sands advancing along its own bottom were checked when the waters passed into the narrow gorge, and were dropped so suddenly that a steep, submerged delta front was formed. Exactly this has been happening now for several years at Millers Falls, where, to improve the railroad, the river above has been turned into a new course for a distance through deep sands, and has thus a great amount of material at its disposal, and a broad, flat bar extending across stream is creeping down into the deep water above the dam, presenting a sharply sloping delta front to the obstructing dam, as do the sands here to the obstructing gorge.

At their south end these lower sands are at first fine grained, well sorted, cross bedded, and undisturbed; northward they are soon changed to coarser sand, the surface becomes pitted with kettle-holes, and the sands grow coarser and become coarse gravel. At last glacial bowlders are intermixed and the sand is twisted and tortuoiis in stratification, and it seems almost to grade into till, as if bowlders were cari'ied south with masses of ice by the waters and mingled with the sands, or as if the ice itself had advanced with many oscillations and disturbed the sands (b").

The ice here postulated (b") was a lobe sent southward from ice which then filled the Deerfield River Valley and much of the high ground north of the river. Another lobe projecting southward in a much shorter valley, and one rising very rapidly to the high level, produced another considerable accumulation of sand in the extreme northeast corner of Hawley, at the headwaters of Ruddock Brook, which passes down the valley of Clessons Brook to join the sand described already as extending south from Buckland Center up the same brook. They join at the south line of the town (at Buckland Four Corners), and run across Ashfield to the delta front described above.

Basin Of The Deerfield River

On page 573 I have traced the southern boundary of the ice at its last marked halt across the eastern half of Franklin County, and a similar line may be drawn across the western half of the county by conriecting- the ice barriers which formed the northern limit of the glacial lakes south of the Deerfield and kept the deep valleys tributary to this river empty in their lower (northern) reaches across from Coleraine to Monroe.

Directly opposite the bamer on Dry Hill, in Montague, on the east of the Connecticut River, is the great barrier above Bardwells Ferry (b^^) on the west, and these may be looked upon as two synchronous halting places of the ice front. Between these the lobe of the ice, which at the same time extended down the Connecticut Valley, may have been thrust into the clays south of Deerfield to produce the disturbances at the Wapping cutting figured on PI. XVIII (p. 694). West of Bardwells are the Bear River barrier (b^^), that at Buckland Four Corners (b"), that on Ruddock Brook (b^^), and that above West Hawley (b^^).

The moraine across the valley of the Deerfield above the mou.th of Hoosac Tunnel (d) lies in the continuation of the curved line which joins these ice barriers, and as these barriers on both sides of the river represent a time when the ice halted for an unusual time, and as they lie along a single curved line, one may assume that they represent a single and exceptionally long halting place of the ice.

Glacial Lakes North Of The Deerfield

Tlie Hawley Lake requires no special mention (see PI. XXXV, A), and the hia-h-level sands at the mouth of the tributaries of the Deerfield on the north are described below. Besides these there is an interesting lake of small dimensions west of Shelburne Center, extending to the river and formed when the Deei-field Valley glacial lobe projected beyond the present mouth of Shiice Brook, throwing back its waters into a lake which drained over a rocky sluice i-unning east from near the cemetery toward the sawmill south of Shelburne Center.

High-Level Deltas

One finds few other traces of obstructed drainage marked by stratified deposits of any extent across the high ground to the north line of the State.

When the reservoir of Mill River broke through the dam at Williamsburg, in 1874, the waters spread out fan-like after their first plunge in the area, a few rods below the dam, and, rapidly losing momentum, they were in eff"ect suddenly overloaded and deposited immediately a portion of the sand they were transporting, in an extended flat-topped layer 1 to 3 feet thick, pushed forward in broad lobe-like projections and bordered downstream by a sharp terrace slope of 30°.

All the streams which come down from the high grounds on the west side of the basin in Westhampton, Northampton, Goshen, Williamsburg, and Whately have here and there in their course torrent deposits of a size all out of proportion to their present dimensions, especially where the streams, after running- tlu'ough narrow channels, deboiich into broad, level portions of their valleys.

I am inclined tQ refer all these deposits to occasional violent floods when the ice was melting in the upper part of the drainage area of the brooks where they are found, and think they may have been made much as were the smaller terraces described above by the flooded Mill River in modem times. At the same time, each one of them may have been formed while a barrier of ice still filled the main valley and blocked up the mouths of these east-west valleys, forming glacial lakes like those described above.

There is one of these deltas on the upper portion of the Sawmill Brook, on the road from West Farms to Westhampton. It hangs in a remarkable way over the broad valley, into which the stream passes here from between the hills, its downstream slope being 40°, and its broad, flat sui-face 40 to 50 feet above the brook. It is brought out flush with the surface of the hills on either side of the brook, as if it had been built up against a wall of ice resting against these hills and filling the valleys below.

Two other deltas are found in the upper waters of the Mill River in Goshen, one at the first road crossing above the reservoir, where the brook comes out of a narrow gorge in granitic rocks. This is pushed out into the valley, with broad, flat surface and steep downstream slope, and the stream has now cut a deep channel in it to its base. The other occurs about a mile below the reservoir on the same brook and east of Hubbard's ledge. This is much larger than the one higher up, and its surface will include several hundred acres.

It is a very curious circumstance that every brook coming into either of the transverse valleys of the Deerfield and Westfield rivers from the north is flanked at its mouth by a distinct terrace, generally triangular from the flaring of the valley, 65 to 100 feet above the stream, and now divided down its center by the deep cutting of the brook. At the villages of Charlemont, Zoar, Orange, and Huntington are fine examples. The explanation that they were caused by the ice lobe coming down the valley and being thrown across the mouth of the side stream is nowhere excluded by anything I have seen, but it seemed to me possible that they might owe their origin to sudden floods of overladen waters into the open valley in the manner described above.

The features at Charlemont admit of an easy interpretation upon this supposition; opposite the entrance of the tributary, and on the south side of the Deerfield River, the high rocky border of the river is set back in a large semicircle, and the south half of this semicircle is still occupied by a great body of sand and gravel, whose level surface slopes south as if in continuation of the slope of the delta ten-ace on the other side of the main stream and flanking the tributary. Indeed, if one could restore in imagination what must, on this hypothesis, have been removed by the main stream and by the tributary itself, the great body of sand would form an alluvial fan extending right across the valley of the Deerfield into the great cirque described above, the southern portion of which fan has since been separated from the rest by the erosion of the Deerfield. However, a southward-sloping terrace on the south side of the Deerfield would not be an impossibility, and the rock of the region is so monotonous that it gives no clew to the source of the sand. These terraces were called delta terraces by President Hitchcock.^

In order not to multiply colors, I have colored these delta terraces with the same shade as that which would be applied to them if they were "glacial lakes" - that is, sands deposited by the obstructed drainage during the retreat of the ice.

Flood Deposits Of Tub Westfield Kivee River

From tlie j^oint where it leaves the gates of the mountain on the west lino of Westtiekl, the Westfield River is bordered by high-level, coarse beds alono- the narro^^' A^alley sides, which widen somewhat where lateral valleys come in, as at Russell, Huntington, and Chester. The valley narrows above

Chester, becomes a canyon between Becket and Washington, widens broadly across Hinsdale, and joins the Housatonic Valley in Dalton with increased width. The canyon is a low water-parting. The whole central part of Hinsdale is deeply covered by the stratified beds of a glacial lake which received its waters from the Housatonic Valley while the ice clogged the lower portion of that valley, and discharged them through the canyon of the Westfield. It thus follows that the Westfield Valley was for a time the recipient of the deflected drainage of the tipper Housatonic after the ice had disappeared from its own headwaters, and the many bowlders and pebbles of Cheshire quartzite found down the valley i-iq. 34 Sand bowlders crushed by the ice while frozen, from just south of

An extremely interesting section was opened by the Boston and Albany Railroad, in 1896, just east of Russell, which throws light on the way in which the terraced beds in the valley of the Westfield were built up, as shown in fig. 33.

terminal moraine, the large angular bowlders being abundant. In the upper portion of this moraine are curious separate areas of stratified sand, discontinuous and much twisted in the unstratified mass of the moraine. One of these is shown in fig. 34. These seem to be plainly parts of a stratum of sand washed upon the broad, flat moraine and then, while frozen, broken into blocks by the farther advance of the ice and mixed with the other bowlders of the till. This shows the presence of a lobe of the ice, moving as a valley glacier down the Westfield Valley and halting at this point.

The second matter which is well illustrated is the long-continued, steady, torrential flow, and the high level of the Westfield River during the time immediately following, while the stream was receiving the waters from the melting ice.

On the downstream side (the right of fig. 33) the current quickly filled up the area in the lee of the moraine with strongly cross-bedded sands of medium grain, and above this extends for many rods a bed, 20 feet thick, of well-sorted and well-rounded 6-inch gravel, in perfectly horizontal beds. Thei'e is rarely a pebble above 8 inches across, and almost everything below 2 inches across is washed out of the bed. One gets here another side of the activity of the strong stream which brought the great volume of sands to build up the broad plains of Westfield and Southwick.