of the 26th mile good bottoms, but here another slate spur would require walling for a mile. From the 27th to the end of the 34th mile no difficulty of moment is met with, but here we encounter another of these spurs, which would require walling for 1 miles. From this to the end of the 50th mile there would be about another mile of walling along cliffs; the rest is generally through good bottom-lands.

This has brought us to Mauk's Mills, opposite which walling for another mile along a steep hillside.

There are very good rich bottoms from this to the end of the 55th mile, where it be comes necessary to cross to the eastern bank, the western being one series of precipit ous cliffs and steeps from the top of the mountain to the water's edge.

For 3 miles, or to the end of the 58th mile, we have good bottoms. This brings us to the perpendicular limestone cliffs above Honeyville.

From the 58th to a little beyond the 76th mile, the river winds its course between the rugged spurs of the Peaked Mountain and Blue Ridge, which here converge and interlock. Numerous steep limestone cliffs project into the river on either side. They are, however, generally short on eastern side, which is that occupied by the survey. The uplands within this distance are very poor, and even the broad bottoms that sometimes occupy a large part of the numerous bends are composed of sand, with but little fertility, and deeply furrowed in places by the river. Kemper's Falls are passed in the above distance.

From the 76th to near the end of the 90th mile, the country opens into fine broad and rich bottom or flat lands, that extend in places to a width of 2 miles.

The survey continued up the eastern side of the river to Port Republic; a work of any kind should however cross to the western side near the 90th mile. This is to avoid the steep, sideling ground opposite Mr. Gilmer's fine bottoms, which would require walling in the river for 1 miles. The water is deep, and the bottom treacherous. Beyond this, on either side, the broad bottom continues to the forks.

The surface of the river at Port Republic was found to be 796.83 feet above the Potomac, or 1,039.26 feet above tide; the rise from the forks 586.35 feet, which is 6.109 feet per mile.

It being the middle of July, and the waters unusually low, I gauged the river be tween Port Republic and New Haven, above where it receives the South River. The surface of the water measured 266 feet. This was divided into 27 sections, and repeated trials with a float having a deep arm were made within each for periods of from 10 to 15 seconds. Four-fifths of the mean result collectively gave 128.56 cubic feet per second as the prism of water passing, which may be regarded as the minimum quantity.

From which it follows, that exclusive of the supply from the South River, there would be ample water to pass a boat every two minutes through locks as large as those of the Chesapeake and Ohio Canal.

THE NORTH AND MIDDLE RIVERS.

The North and Middle Rivers will be described when I conclude my reports. The mode of improvement that may be adopted on the principal branch will determine the character of its tributaries.

THE CANAL

In conformity to law, I am progressing with the plan and estimate of the cost of a canal to extend from Harper's Ferry to Port Republic.

THE RAILROAD.

I have carefully revised that portion of the line that was run by me under the late principal engineer, and leaving the river at Stony Creek, have extended it through Woodstock to Winchester. I am not prepared at present to report on the subject, as the estimates are, or ought to be, predicated on actual calculations, which require time, and vary with the different plans of construction. I have only to remark, that though I had great difficulties to encounter, I became perfectly satisfied of the practicability of the scheme as I progressed with the surveys.

REMARKS.

From the facts set forth, it is clear that the choice rests between a canal or other improvement of the river from Harper's Ferry to Port Republic, a distance of 150 miles, or a railroad from Harper's Ferry to Staunton, along the river of about 175 miles, or of a railroad through the valley from Staunton to Winchester of about 105 miles.

This latter route has, in my opinion, other claims to attention than its shortness. It passes through and connects the four county towns, Staunton, Harrisonburg, Woodstock, and Winchester, as well as passing through or near the other towns, and a most populous district, the products of which would have for a considerable distance to be transported across the Peaked and Massanutten Mountains to reach a river improvement. The whole distance from Staunton, through Winchester, to the Potomac would be about 142 miles.

Respectfully submitted.

JAMES HERRON,
Civil Engineer.

To the PRESIDENT AND DIRECTORS OF THE BOARD OF PUBLIC WORKS.

Report No. 2 of Mr. James Herron, civil engineer, on the surveys of the Shenandoah River and Valley.

SHENANDOAH CANAL.

Before proceeding to estimate the cost of a canal, it is necessary to determine the size capable of affording the requisite facilities to the trade, if the latter be known, and also if it be a branch or continuation of another work already complete or in construction. It is clear that in determining on the plan of our locks, we should make them accord in some measure with the plan of the locks on the previous work, as through them our boats will have to pass to market. The Chesapeake and Ohio Canal, with which the Shenandoah Canal could form a junction at Harper's Ferry, is at that point 60 miles long, from tide-water at Georgetown. Throughout this distance it is intended to hold 6 feet water in depth, and to be generally.60 feet wide on the surface. The rise from Georgetown to Harper's Ferry is overcome by 31 cut-stone locks, each 148 feet long, 15 feet wide, and 100 feet long in the chamber between the gates; with a few exceptions, they are of 8 feet lift, which is very nearly the average they give for the whole distance. In the third annual report of the company it is stated that 27 of these locks, when finished, will have cost $260,30, which, on the supposition that they are of 8 feet lift, is at the rate of 1,207.47 per foot, or $9,659.75 for each lock. Most, if not all, of these locks were constructed with side culverts; latterly, the culverts have been omitted in the construction, and the lock is filled through valves in the upper gates, which, for that purpose, are made to extend to within a foot of the bottom of the lock, a necessary precaution to prevent swamping the boat. I have understood that when constructed on the latter plan, a lock of 8 feet lift cost about $9,400. Locks of the foregoing dimensions would evidently be too expensive, if not too large, for a canal adapted to the wants of the Shenandoah Valley. The following plan will, I trust, be found to possess such dimensions as a due regard to the economy of time and expense of water on the Chesapeake and Ohio Canal would require; and while adequate to any trade that can reasonably be expected on the Shenandoah, from the less quantity of masonry and other materials used, and greater simplicity in construction, they will not cost one-half what the large locks would; they will also be shown to bear such proportions to the other dimensions of the canal for which they are intended as the scientific deductions of French engineers, and the practical precedents of English works, would justify.

Width between the walls 7 feet 9 inches in the clear.

Chambers 80 feet long between the gates.

Gates single, and to extend across the lock.

The lock to be filled and emptied by means of valves in the gates.

The total length will be about 106 feet, and to have at all times 4 feet water on the flooring. The breast may be constructed of timber, as it is always submerged. It should be placed immediately under the plank grooves, about 2 feet above the recess of the upper gate. The flooring is also contemplated as constructed of timber uuless when founded on rock.

Locks of these dimensions will pass boats capable of carrying from 21 to 30 tons. They should be navigated in pairs, as two of them can be passed at a time through a lock of the Chesapeake and Ohio Canal. If constructed of cut stone laid in hydraulic lime, they will cost about $4,550. Rubble range work, or good brick-work, with cut coping and hollow quoins, will cost about $3,500. I have introduced the latter in the estimates. Brick is much used in the hydraulic works of England and Holland, and Mr. Strickland says in his reports, pages 8, 9, and 10, that locks thus constructed are now good, after a use of upwards of 60 years. And from the following extract it will be seen that the dimensions I have given are as nearly those of some of the English canals as the consideration of passing two boats at a time through the locks of the Chesapeake and Ohio Canal would admit of.

"The lock chambers of many of the most profitable canals in England are narrow and long, particularly those of the Grand Trunk and Birmingham Canals. These are 75 feet long and 8 feet wide, with vertical and parallel sides, built of brick-work 4 feet in thickness, and coped with stone." (See page 10 and plates 15, 15, and 29.)

Mr. Strickland also remarks that the boats carry 30 tons, and are 7 feet wide and 70 feet long. But there is evidently some discrepancy in his statements as regards the length of the boat when compared with the length of the lock. For a boat 70 feet long the lock chamber should be about 79 feet from sill to sill, or else how could the lower gate, which extends 9 feet within the chamber, be shut or opened with a 70-feet boat in the lock?

With regard to the dimensions of the canal the Chevalier Du Buat, lieutenant-colonel of the royal corps of engineers, found by experiment that the cross-section of the boat should bear a certain ratio to the cross-section of the canal to enable the boat to be drawn through the canal with as little exertion of the moving power as would be required to draw it on an open expanse of water. From this experiment he deduced the following formula, which appears to stand unrivaled in England as well as France,* and was also employed by the United States board of internal improvement, when projecting the plan of the Chesapeake and Ohio Canal.

R will represent the resistance; the resistance on an indefinite expanse of water being 1.

c. The area of the cross-section of the canal,

b. The area of the section of the boat;

and K, a constant quantity determined by experiments to be = = 8.46.

Du Buat also concluded that a canal cannot be considered as an indefinite expanse of water unless the width is 4.46 times the width of the boat. The proper width for the boats of the Shenandoah canal would be 74 feet; when fully loaded the mean section immersed will be about 16 feet; the canal should, therefore, according to the rule, be 32 feet wide and have a cross-section of 103 feet. The United States board did not consider it necessary to adopt the dimensions required by the rule to the full extent; and when we take into consideration that the boats will not always travel so fully loaded, it may appear that there will not be sufficient advantage gained in traction to compensate for the increased cost incurred in construction, which will more fully appear by filling the formula with the following dimensions, which I consider well adapted to practical purposes.

Width on the surface of the water 30 feet, on the bottom 18 feet, and 4 feet in depth, which gives a cross-section of 96 feet, being six times that of the boat, and a width 4.13 times the width of the boat.

Then

So that the resistance experienced by such a boat on a 30-feet canal, to what would be experienced in open sea, is in the ratio of 105 to 100, being only 5 pounds greater in the 100 pounds traction on the canal than it would be in open sea. Moreover, the tow-path and berm bank are calculated to be 2 feet above the water, but the canal might be filled so that they would be only 18 inches above, which would increase the width of surface to 31 feet, and the sectional area to 111. The latter would probably be the better plan, as a boat loaded with 24 tons would then experience no greater resistance on the canal than in open sea; or with a load of 30 tons the traction would be increased in the ratio of 110 pounds to the 100 pounds. Now, as it is found that a horse is sufficiently loaded when he exerts a force of 20 pounds, in drawing a boat ou a canal at the rate of 24 miles an hour, and as this is the force he has to exert when the boat is loaded with 24 tons, it follows that he would be overloaded with the 30-tons boat, exclusive of the resistance peculiar to the canal.

*See the American edition of Brewster's Encyclopedia, article Hydrodynamics. page 886, correcting a typographical error in the formula as above, or see Gregore's Mathematics for practical men, London edition of 1825, page 366.

Mr. Smeaton states that 22 tons burthen at from 2 to 24 miles per hour is the work of a horse on a canal. And Mr. Bevan informs us that the horses on the Grand Junction Canal usually travel 26 miles per day, and draw a boat containing about 24 tons at the rate of about 2.45 miles per hour; the empty boat being nearly 9 tons, the total mass moved is about 33 tons, and the average force of traction 80 pounds.

The increase of load is estimated thus: When the boat and load together weigh 33 tons, and the force of traction has been ascertained to be 80 pounds or the 1.924 part of the mass moved, then 30 tons load and 9 tons boat, together 39 tons, the 1.924 part of which is 94.5 pounds, the power required to draw the latter boat in open sea. And as the resistance increases on the canal, even with the section of 111 feet, in the ratio of 110 to 100, then 100: 110 :: 94.5: 103.95-so that it will require 104 pounds traction to drag a boat loaded with 30 tons through the canal. We cannot, therefore, calculate on a greater load than 24 tons without injury to the horse, as the diagonal pull and constant stress of 80 pounds for 10 hours in a day are quite equal to the powers of an average horse.

The other dimensions I have to propose for the Shenandoah Canal are, a tow-path 10 feet wide, berm bank in open formation or in embankment 8 feet wide and in cutting 5 feet.

I have omitted surf berms, as I find that they would very much increase the expense, and I conceive it would be better to obtain the desired width of surface by increasing the interior slope of the banks. In many cases the width of the canal will be much increased at a less expense than it could be made otherwise; as in some cases of single bank formation, the materials for which have either to be hauled or are scraped up between it and the natural slope which forms the other bank.

To enable me to form an opinion of the cost of the canal, I have supposed the position of the different works in the first 30 miles to have been determined on, as per abstract of estimates; from which it will be seen that I estimate this distance at $253,927.87, which gives an average of $9,464.26 per mile. The fall overcome in this distance is 160.61 feet; the fall to be overcome from this place up to Port Republic, including a feeder-dam at that place, is 648.22 feet. The lockage, as estimated, is at the rate of $437.50 per foot, which with the addition of 10 per cent. gives $77,000 for the lockage on the first 30 miles, leaving an average of $6,897.59 per mile for the other works on the canal. The distance from Harper's Ferry to Port Republic, as the canal would have to run, is 149 miles, from which, deducting 30 miles, leaves 119 to be estimated for. On the supposition that the average cost per mile, exclusive of the increased amount of lockage, will not be greater above than it is within the first 30 miles, the whole distance above of 119 miles would, at $6,897.594 per mile, amount to the sum of $824,262.60; but for about 60 miles of the distance between the forks and Port Republic, the canal formation will cost about one-third more in consequence of the number of large culverts that will be required to pass mountain ravines, the increased quantity of walling along cliffs, and the less favorable nature of the soil in general. Making the necessary addition of these objects of $137,951.89, we have for the cost of the canal formation $962,214.49; and 648.22 feet of lockage at $481.25 per foot, including an addition of 10 per cent. on the net estimate, amounts to $311,955.87.

At about every 20 miles a feeder-dam, river-lock, and stop-gates will be required, say six in the above distance. These works will cost on an average about $9,000, and will therefore amount to $54,000.

In the plan of an independent canal, three aqueducts across the Shenandoah would be required, and one across the Potomac, to unite with the Chesapeake and Ohio Canal. These structures, if built of stone, would of themselves amount to upwards of half a million of dollars, and as I feel satisfied that the navigation will experience no great inconvenience by substituting tow-bridges and crossing the river in the ponds occasioned by the feeder-dams which will occur about the proper places for crossing, I have omitted the one and introduced the other in the estimate.

RECAPITULATION.

First 30 miles from Harper's Ferry.

119 miles of canal formation, at $6,897.594

One-third additional for 60 miles of the above..

648.22 feet of lockage, at $481.25 per foot...

Six feeder-dams and river-locks, at $9,000

....

Three tow-bridges and an addition to the bridge over the Potomac

Total......

$283,927 87

824,262 60

137,951 89

311,955 87

54,000 00

17,000 00

1,629, 098 23

This sum, however, is exclusive of damage to land and commutation for fencing; the latter items will amount to about $1,000 a mile, which for 1494 miles amounts to $149,500, making the total sum of $1,778,595.23, which is at the rate of $11,896.98 per

mile.

To pay the interest on the first cost, and to pay for renewals, and the ordinary repairs of the canal, also to pay lock-keepers, toll-gatherers, and superintendents, a certain amount of trade should pass annually on the canal, the tolls on which should decrease as the trade increased, and the maximum toll that may be charged will depend on what it costs per ton for transportation.

I have shown that a boat adapted to the canal will be loaded with 24 tons. It is not to be expected, however, that boats will be able to obtain loading to this amount both going to market and returning, nor to be steadily traveling every day throughout the year, yet the expenses will go on; what with the delays incidental to loading and unloading and delays for repairs, as also the difference between the ascending and descending trade, we cannot rate the average load higher than one-third of this, or 8 tons per day for 312 working days in a year. The boat will be drawn by one horse at the rate of 2 miles an hour for ten hours in the day, and will require the attendance of two men and a boy; they will consequently travel 25 miles a day, which, when the delays at the locks are taken into consideration, is quite as much as we can expect.

The daily expenses will be about 50 cents apiece for the wages of the men, 25 cents for the wages of the boy, 50 cents for the interest, decrease of value, feeding, shoeing, &c., of the horse, and 25 cents for the wear and tear, interest, &c., of the boat and towrope; or, in all, $2 per day.

We have, therefore, 8 tons transported 25 miles for $2, which is at the rate of one cent per ton per mile,* or counting ten barrels of flour to the ton, the boat-owner's expenses would be less than 15 cents per barrel from Port Republic to Harper's Ferry. We will suppose that this includes the boat-owner's profit, as on a canal each farmer may become his own carrier if he finds it to his interest to be so. I have, therefore, only to consider the expenses he will probably incur; that those may not exceed 75 cents on the barrel of flour from Port Republic to Harper's Ferry, the tolls should not exceed 4 cents per ton per mile, which, together with the transportation, is half a cent a mile for each barrel of flour.

The sum that must be raised annually by tolls to pay for repairs, wages of lockkeepers, and other agents will depend on the number and connection of the locks and the general stability of the works. I have thought that this would be more satisfactory if predicated on the actual expenses accruing for a number of years on two similar works situate within the State. I accordingly examined the reports of the second auditor made to the house of delegates on the receipts and expenses of the James River Company, from which I find that exclusive of the extension and enlargement of the lower canal, the ordinary expense of repairs, collection, lock-keepers, &c., have for seven years past, that is, from 1826 to 1832, inclusive, averaged $11,316.80 per annum, which for 30 miles, the length of the canal, is at the rate of $377.224 per mile. And ou the Blue Ridge Canal, from the date of its opening on the 1st February, 1827, to the 31st December, 1832, being five years and eleven months, the average annual expenses have been $2,672.95, which for 7 miles, the length of the canal, is $381.854 per mile per annum. The mean expense per mile on the two canals is $379.54.

The average cost of constructing the Shenandoah Canal per mile, I have estimated at $11,896.98; the annual interest on this sum at 5 per cent. is $594.85, to which add for the annual repairs, expense of collection, lock-keepers, &c., $379.54, and we have $974.39 to be raised annually by tolls on each mile of the canal. At 4 cents per ton per mile, we would require an annual trade of nearly 24,350 tons; or as there are 312 working days in a year, about 78 tons daily, which would require the passage of from 3 to 4 loaded boats, or, with the returning boats, say 8 daily. That the caual is ample to ac commodate any trade that can reasonably be expected will appear from the consideration that 150 boats can be passed with ease through each lock in the course of 10 hours, which, on the supposition that only one-half of them are loaded, is at the rate of 1,800 tons daily, and that there is water sufficient for a much greater amount of trade than even the latter sum, is evident from the fact that the lockage for the latter amount would not use 21 cubic feet per second even on the supposition that the trade was all in the one direction, and that each boat used a lockful of water; the current in the canal would then be only about 12 feet per minute. But I have stated in my former report that we could calculate on 128 cubic feet per second from one branch of the river at Port Republic.

With regard to improving the rivers by locks and dams, I beg leave to refer to my

*Sir Henry Parnell, in his treatise on roads, published in London, 1833, says that horses on canals draw 25 tons of goods over a space of 16 miles in a day at the rate of 23 miles an hour, and that the actual expense is half-penny per ton per mile. A balfpenny is less than a cent, or, more exactly, it is the .925 part of a cent. He also says, this has been the regular charge on the Ellismere Canal for some years, and is now introduced on the Oxford canals and several others." (See page 109.)

+ Horses are not used on either of these canals, hence there is less wear of the towpath; but on the lower canal there is a considerable expense occasioned by an obliga tion the company is under of keeping up fencing, which the upper canal is, I believe, entirely exempt from.

Twelve feet per minute is scarcely more than the seventh of a mile an hour.