Operating the Canal

Water to fill the canal and to refill the empty locks as boats passed from one level to another was the most obvious but not the major requirement. Although the Commissioners of 1811 warned that "more [water] will filter through the sides and bottom of a canal than those of a river, which are generally saturated" (p. 18), neither they nor the Commissioners of 1816 made any estimate of the quantity of water required to make up for the leakage and evaporative losses. Nor did a review of British experience (Sutcliffe, 1816) emphasize the matter of leakage, discussing water requirements mainly for lockages. The amount of water required for the operation of the Erie Canal proved to be far greater than first provided and additional supplies were sought over the years. Annual maintenance, repairs, and improvements kept the canal viable, and the cost of these operations totaled to a sum that equaled the base investments in the building of the canal.

LOCKAGE

To fill the canal at the beginning of each navigation season was a single operation. But each time a barge moves through a lock, water moves from the upper to the lower level or pound. The same or equivalent volume of water can be used in a succession or flight of locks provided that the pounds are long enough to store-to impound-the lockage volume. (See Glossary for explanation of hydraulic terms.) Lockage volume, therefore, for each flight from summit to a sag or low point equals the prism of the largest lock times the number of boat passages, if all in one direction, or one-half the number of boats if the up-and-down passages alternate. Since the same volume of water can be used successively in each lock between summit and trough in the profile, and since the largest lock need only clear the largest vessel, the volumes of water required for lockage need not be great. The higher the lifts in each single lock (on the Erie, lifts averaged 8 feet), the fewer are the number of locks that are needed and hence the shorter the transit time, but at the expense of greater volume of water for lockages. The amount of water for lockage was easily calculated. A typical lock chamber (see fig. 7) contained about 10,000 cubic feet. A total of 20,000 lockages alternating in direction represented a busy 220-day open-water season. These numbers are equivalent to a water flow of only about 10 cubic feet per second in each tier of locks. However, the demands at a summit for lockage could be critical, as water must be discharged in two directions. Consider two barges following each other ascending to the summit. Boat 1 leaves the top lock full as it enters the summit pound. After pass- ing through the summit, it leaves the top lock on the down side empty. Following boat 2 then approaches the summit and finds the lock full (as it was left by boat 1), empties it, and then refills it, drawing off the necessary water from the summit pound. At the other end of the summit, boat 2 finds the first lock down to be empty, as boat I left it, and therefore fills it, goes down, and leaves it empty. Boat 2 therefore drew two lockfuls of water from the summit. A succession of boats in one direction could overdraw the available supply at the summit. At times, therefore, barges were detained at the summit, so that locks were filled or emptied only when occupied by a barge. Time was saved at the cost of water by a process called swelling, that is, flushing "down" boats out of the locks by discharge of a small flow through the upper gate paddles. (Rafter, 1905, p. 792.) Although swelling was frowned upon when water was in short supply, it was sometimes necessary at just such times to produce a wave or swell in the lower pound to get grounded barges moving. Water usage in locks separated by long pounds was minimized when up and down barges alternated. When the pound reaches were short, or as in the combined locks at Lockport where the lock chambers were in a tier without any intervening pounds or storage, then water use was minimized by having boats follow one another in the same direction. The tier of combined locks was built in parallel; an up staircase and a down staircase. Here, of course, one lock chamber of water is used for each up and each down boat. If, however, alternating passages were used on a combined lock, then the whole set of chambers would need to be filled or emptied for each passage. The conspicuous leakage at the timber miter gates was not normally a serious problem, save at the last or bottom lock in a tier. A good flow of water was needed down the canal and what did not leak would otherwise spill over by by-pass weir at the lock. Depending on details of construction and fit, the quantities inbound were highly variable; reported figures for gates in the old 40 by 4 foot canal with 8-foot lift (12-ft bottom gate) ranged between 380 and 1,000 cubic feet per minute (6 and 17 cfs). To give some idea of the looseness of the gates, these rates are those associated with orifices of 0.5 and 1.2 square feet, respectively, at a 6-foot average head on a 12-foot bottom gate. Yet the gate leakage was sufficient, as will 'II be evident, for only 3.8 to 10 miles of canal seepage losses, and therefore it rarely, if ever, approached the flow required down the canal. Although considered a problem when the canal was low, lock leakage was rarely a basic cause of low water, as water retained to maintain levels in an upper pound would be at the expense of lower levels down the canal. .

CANAL LEAKAGE

The volume of water for lockage as minor compared to thevolume that seeped out of the canal bed and banks. In the level situation (see fig. 7) the , bankments (berm and towpath) would be formed of excavated materials- that is, there would be a balance between cut and fill. With a freeboard of 2 or 3 feet between: top of the banks and the water surface and a water depth of 4 feet, the surface of the water would stand above the original land surface and would abut the fill material. ln side-hill emplacement (see fig. 8) the contact between water and fill on the downhill side would be even. greater. Leakage was noted as a problem on the first sections built. The 25-mile reach from the "Nose" to Schenectady was found to leak so badly in 1822 that it was drained and lined with clay (1823 As. Jour. 46). As leakage was still a problem, measurements of this loss in selected reaches were made 1824. (An account of methods of flow measurement is given in the section "Measurement of Water Flow.") Measurements of total or aggregate loss of water from the canal were continued because leakage was a continuous problem. The results, as compiled by Rafter (1905) are given in table 2. These data indicate that the rate of loss ranged from a low of 2.89 inches per day in the Clyde level to more than 10 inches in the reach along the Mohawk River from Amsterdam to Schenectady, with a general average of about 8 inches per day. The low figure for the Clyde level probably cor. responds to the fact that it is the sag point on the profile that includes the Montezuma-Cayuga Marsh, a region of high water table. It might be expected that siltation or swelling of clay particles of the bed and banks would in time decrease the rate of seepage, yet the data seem to indicate no progressive change. A comparison of the 1824 with 1841 and 1847 measurements in the western - division (say, Brockport-Rochester, 1824, with Lockport to Pittsford, 1841) shows a 30-percent reduction, as noted by Rafter 1897, p. 178). However, slope-area measurements'(by hydraulic formula) for the Lockport-Rochester reach in 1877 (Searles, 1877) some years after that channel had been enlarged, indicated a rate of water loss of 7.5 inches per day, not much less than that reported in 1824. These data on water losses were obtained by measuring rates flow at the end of long reaches of the canal between feeders, and ascribing the differences to a loss of water. This loss would include net leakage and evaporation. Of these losses, leakage is the dominant factor, for evaporation would be 0.2 inches per day even in midsummer.

Table 2 _____________________________________________________________________________ Locality Date Length of reach Canal losses Observed (miles) (Inches) _____________________________________________________________________________

Brockport to Ninemile Creek 1824 ---- 8.18 Brockport to Cayuga 1824 75 8.28 Brockport to Rochester 1824 20 8.58 Amsterdam to Schnectady 1824 18 10.22 General 1838 --- 8.18 Lodi to Little Falls 1841 61.8 4.74 Palmyra level 1841 8.3 8.89 Clyde level 1841 27.7 2.89 Both levels 1841 36.0 6.88 Lockport to Pittsford 1841 69.0 5.97 Lockport to Pitlock 1847 122.0 7.0 ------------------------------------------------------------------------------------------------------------------------------------

It might be expected that siltation or swelling of clay particles of the bed and banks would in time decrease the rate of seepage, yet the data seem to indicate no progressive change . A comparison of the 1824 with 1841 and 1847 measurements in the western division (say Brockport-Rochester, 1824 with Lockport to Pittsford, 1841) shows a 30% reduction, as noted by Rafter. However, slope area measurements for the Lockport-Rochester reach in 1877 some years after that channel had been enlarged, indicated a rate of water loss of 7.5 inches per day, not much less than that reported in 1824. These data were cited for many years and appear among the data on loss of water from American canals compiled by Woodward (1930, p. 1578), all of which relate to 19th century measurements. In that list, water loss from the Erie is within the range of experience of other canals, albeit a bit on the high side. A seepage loss of 8 inches per day for a 40-foot wide canal amounts to about 100 cubic feet per minute (1.7 cfs) per mile, as reported by Blake as early as 1823 (see section "Measurement of Water Flow"). With a canal section of 136 square feet, the amount of water required to replenish the losses is the same as if the canal were refilled every 5 days.

DIVERSIONS OF WATER FOR POWER

The Erie Canal combined a source of mechanical energy with means for transport of supplies and products. Numerous useful differences in head were created either between the canal and a natural drainage course in -,which case power generation depended on surplus canal waters, or between a feeder and the canal, in which case the tailrace discharged into the canal. In fact, each lock contained the required characteristics for a waterpower site although the ordinary single lock with a 7- to 9-foot head offered only low horsepower (hp). The combines at Lockport offered great possibilities for there was not only the 60 feet of head on the canal itself, but also the difference in level between the canal and the bed of the creek that passed under the canal. (See fig. 6.) Beginning with sawmills, manufacturing uses of waterpower increased at Lockport, particularly after the canal was enlarged in the 1840's, increasing the diversion of water from Lake Erie. In the 1890's nearly 400 horsepower was developed using flow diverted from the canal and returned to the canal below the locks. An additional 2,600 hp was developed using surplus waters that would otherwise overflow a waste weir into Eighteen mile Creek that crossed under the canal a short distance to the east of the foot of the locks. (See fig. 6.) A measurement of the diversion to Eighteenmile Creek from the canal made in 1887 was reported by Rafter (1897, p. 214) to be about 19,000 cfm or 320 cfs. Other water powers were developed along the canal, as at Black Rock, Medina, Syracuse, and Little Falls. Although, beginning with the idea of using "surplus waters," waterpower began to exercise a claim on diversions of the water from the canal and thereby to diminish navigable depths. Waterpower along the canal became a matter of public controversy (1870 As. Doc. 133) as being extraneous to the primary objective of the canal-navigation. Later critics, such as Rafter (1897, p. 213), saw the possibility of combined use that would optimize total benefits and recommended a clear statement of policy to achieve it. But nonelectric waterpower (that which is directly connected to factory machinery) was soon to be technologically obsolete.

WATER SUPPLY TO THE CANAL

Water for lockage and to replenish leakage along the way was brought into the canal by gravity diversions from rivers and creeks. A dam, typically of rock and brush, was built across a stream to a depth sufficient to divert water into the feeder canal that led into the navigation canal. The amount that could be diverted depended not only on the amount of water in the streams but on how much leaked through the dams which were usually tightened or improved during periods of water shortage in the canals. In order to minimize disruption of canal navigation by floods, the feeder canals contained a guard gate that was closed during period of high water on the rivers. Diversion dams and guard gates were built of timber, brush, earth, and stone-local materials which permitted an ad hoe response to experience, and favored an easy accommodation to the environment. In 1825, when the canal as a whole began operations, water supply was obtained from 12 sources including Lake Erie, as listed in the section on "Feeders, Locks, and Stream Crossings." On the original canal the total quantity of water used for a tier from summit to sag averaged 1.7 cfs per mile of canal for leakage (0.5 cfs/mi in the region of the Montezuma-Cayuga Marsh) plus 25 efs for the water let out at the bottom lock. For the original canal system as a whole, therefore, the following quantities of water were needed:

Lake Erie to Montezuma (145 mile leakage plus 25 cfs)----------- 300 cfs Clyde to Jordan level (25 miles , plus 25 cfs)------------------------ 35 cfs Jordan level to Syracuse (12 miles, plus 25 cfs)----------------- 30 cfs Syracuse to Rome summit (60 miles, plus 25 cfs)------------------ 125 cfs Rome summit to Hudson River (110 miles, plus 25 cfs)----------- 215 cfs

By 1862 the total flow to the enlarged (70 x 7 ft) canal had increased to 1,900 cfs when the number of feeders had about doubled (see section "Feeders, Locks, and Stream Crossings.") There were 40 feeders in operation in 1891. The quantities of water handled by the Erie rank it as the major water project of the 19th century.

The increase in the rate of intake from 700 efs to 1,900 cfs may be largely accounted for by the increase in the leakage and related losses. These quantities would be approximately proportional to the product of the surface width of the canal by the square root of the depth of water in the canal. As between the 7OX7-foot section and the original 4OX4-foot section, the ratio of losses would be 2.3 to 1, nearly as much as the increase in flows noted above.

RESERVOIRS

The first reservoirs (1830 As - Doc. 47, p. 32) built on the canal system were those constructed in the 1830's to supplement the low flows of the small streams that were used as feeders at the high summit level of the Chenango Canal, a lateral to the Erie Canal at Utica, which led up the valley of Oriskany Creek (see section "Hydrologic Data and Analyses"). Although the Chenango Canal was later abandoned, the reservoirs were maintained to augment the low seasonal flows of Oriskany Creek, an Erie Canal feeder (see section "Feeders, Locks, and Stream Crossings"). Lakes, however,. were used for the major part of the storage capacity by putting control works at the outlets. Total storage capacity by decades is given in table 3 (1892 As. Doc. 75, p. 72). A large amount of capacity came into operation in the 1850's as a result of the construction of the Forestport feeder to the Black River Canal which discharged into the Rome summit level.

Some idea of the magnitude of the total volume can be obtained by expressing it in terms of the number of days of operation. Considering the water intake was about 2,000 cfs, the capacity built by 1890 amounts to a reserve sufficient for about 55 days operation.

Table 3.-Total reservoir capacity on Erie canal feeders, by decades

Up to Capacity (millions of cubic feet) 1840 ------------------------------------------------ 1,275 1850 ----------------------------------------------- 3,470 1860 ----------------------------------------------- 6,290 1870 ------------------------------------------------ 9,090 1880 ------------------------------------------------ 9,460 1890 ------------------------------------------------ 9,900

WATER SHORTAGES

Leakage, lockage, and the diversions for power over the years added to the need for water. The potential supply of water was adequate as the canal commissioners correctly judged in 1818"With a country of from fifteen to sixty miles wide, stretching its whole length, and abounding with lakes and streams, which all seek their natural discharge by crossing it, no deficiency of water can ever be apprehended" (1819 As. Jour. 42nd sess., p. 207). New York State is favored with a humid climate and copious water resources, and the low flow from 7,000 square miles (equivalent to a strip along the canal about 20 miles wide) would indeed supply the 700 cfs needed, and 19,000 square miles (55 miles wide) to supply the 1,900 efs used by 1862. The problem was to develop the supply and to convey the water down the canal to maintain navigable depths.

Despite the continual addition of feeders and the impoundment of water in reservoirs, shortages of water led to delayed traffic because shallow water increased hydraulic drag on the barges or because barges had to lay by to await other barges to share the lockage volume-that is, an up barge on finding a lock full would have to await a down barge before the lock was emptied. The annual reports of the Canal Commissioners and later the State Engineers repeatedly refer to water shortage - particularly in dry spells, and report efforts to shore up the supply to the canal by such measures as tightening the diversion dams. A short historical summary of water shortages is given in the State Engineer's report for 1883 (1884 Sen. -Doc. 9, p. 24).

Even as late as 1891, by which time there were 40 feeders, the State Engineer reported (1892 As. Doe. 75, p. 127) "During the latter half of the month of August and the whole month of September, the water in the Mohawk River was very low, and it was with difficulty that navigation could be maintained I ' "' And again (p. 196), "This fact gives additional force to the remarks in previous reports on the necessity for additional sources of supply and additional storage reservoirs"a startling admission after 65 years of trying to cope with water shortages. The scheme of adding a new feeder as river flows receded during dry spells never seemed to measure up to the job. Rafter came to the conclusion that the fault lay with inadequate data. "In the absence of systematic information as to the yield of streams, the general tendency has been to overrate the summer overflow, with the result of shortage frequently at points where the supply was believed to be ample."

To judge the adequacy of the streams to meet the requirements, the canal engineers measured their flows during the late summer and autumn seasons when the streams were usually at their annual minima. These measurements were made during the year or so of planning that preceded construction. As is now ell established from continuous records of streamflow, rivers are highly variable in flow from year to year as well as seasonally. A single set of measurements by themselves cannot give an indication of the low-flow regimen of the river. As a demonstration, assume that measurements made during the a 3-year period of planning and design succeed in determining the lowest flow period. How does that flow compare with the experience of say the next 50 years? The current records of the flow at 5 different long-term gaging stations show that the low flows during any 3-year sample would have a fourfold range and, on the average, would be from 2 to 3 times greater than the lowest flow in the record. This comparison indicates that the isolated measurement gave optimistic indications of low flows. It was this kind of experience that beset the canal during the 19th century.

HYDRAULIC INEFFICIENCIES, TRAFFIC DELAY

When the canal opened, 30-ton barges made the 363-mile round trip between Albany and Buffalo in 16 days (equivalent to a barge speed of 2 mph, allowing 10 minutes per lockage). By 1841 the average time for a found trip for 70-ton barges had become 22 days, a third longer than originally (1842, As. Doe. 24; p. 15), subtracting significantly from the, advantages of the larger payloads. The chief villains that decreased efficiency were inadequate hydraulic capacity, flow blockage, and hydraulic drag. Losses in efficiency also resulted from the introduction of larger barges that caused collisions when passing. Unscheduled barge traffic was the cause of waiting-line delays at locks. The adverse economic consequences of all these inefficiencies were doubtless considerable as they led to pressures to enlarge the facilities, rather than measures to abate them at the source.

HYDRAULIC CAPACITY

Besides low flows in the streams that were diverted into feeders, shallow depth for navigation also occurred as a result of insufficient capacity of the canal prism to convey the necessary water down the canal, particularly when the hydraulic capacity was decreased by the presence of barges.

As the Commissioners of 1811 had anticipated, some slope was needed to convey water to maintain navigable depths. Moreover, slope could save lockage, provided that velocity induced by the slope was not excessive. The original canal was built without benefit of formal hydraulic computations to assure that these conditions were met minimally, let alone optimally. In Great Britain a slope or fall of about 1 or 2 inches per mile in the pound reaches was about right, but that would depend on the rate, of channel losses and the spacing of feeders. If losses are low, water conveyance is low and a flat slope would be satisfactory. In the Erie, water sufficient for about 35 miles of canal could be conveyed with a slope of 1 inch per mile and for about 50 miles with a slope of 2 inches per mile.,

The western division of the canal offered a classic hydraulic problem which is treated in some detail in the section "Hydraulic Computations.11 In concept, that division-from Lake Erie to the sag point in the profile in the lake district (fig. 4)-had a CODIOUS source of water in Lake Erie. However, the original canal could not meet the objective of conveying a supply of water from Lake Erie because of a lack of hydraulic capacity. The grade of the canal from the lake to the mountain ridge was held at 1 inch per mile to avoid the added rock excavation that would have been required to increase the slope and therefore the flow. Because a slope this flat would convey only enough water from the lake for about 35 miles of canal, a feeder was introduced in 1823 from Oak Orchard Creek which was crossed by the canal 40 miles along its length from Lake Erie. The upper Tonawanda Creek was diverted into the north flowing Oak Orchard Creek at a point where there is only a low saddle between them. (See fig. 2.)

Later in the experience of the canal (see section "Hydraulic Computations"), it was also recognized that a uniform canal cross section (1842, As. Doc. 24, p. 14-15) was not appropriate for long pound reaches. Considerable difficulties because of insufficient water to float boats were reported to occur in the long levels in the western division and in the long summit level at Rome. There was insufficient hydraulic capacity at the upper end to convey the water needed to maintain navigable depths in the lower end of the reach. Hydraulic capacity needed to be proportioned to the flow and distance from the supply.

Beginning in the 1840's with increasing traffic there were many reports of flow blockage by boats on the canal. The hydraulic capacity of the 40 x 4-foot canal would be reduced by more than half when part of the section was occupied by a barge and more so when passing boats wedged together. Further, the blockage effect was even greater when the canal was only partially full, and thus, the effect was compounded down the canal. For example, it was noted (1842 As . Doc. 24, p. 14) that after a break in the canal banks had been repaired the time required to refill the canal was lengthened because the flow down the canal was obstructed by 215 boats. State Engineer McAlpine (1854 Sen. Doc. 50, p. 77 and 80) reported that when the canal was crowded with boats there was difficulty in "sending forward the requested quantity of water to keep up the levels and supply the locks." Sand bars had a similar effect (1845 As - Doc. 28, p. 52; Rafter, 1905, p. 850). In addition aquatic vegetation, chiefly Potamogeton, growing in the canal impeded the flow; even when cut the stubble added considerable resistance to the flow of water. (1850 Sen. Doc. 41, p. 8).

HYDRAULIC DRAG

Barges were initially of the narrow type-7 feet wide, 3 ½, feet draft, and 60 to 70 feet long, and carrying 30 tons, or 1,000 bushels of wheat. Increasing traffic soon led to introduction of broad 70-ton barges, limited only by the 15-foot width of the locks, and canal enlargements were planned in 1832 after only 7 years of operation.

As previously mentioned, it appeared to be a simple matter to estimate canal width and depth, when given the size of a barge: one needs only to provide clearance on bed and banks for a pair of passing barges. But the problem of hydraulic drag soon emerged as it had in Europe. A vessel that moves with little clearance above the bed or about its sides is retarded by increased hydraulic drag. The moving vessel creates a reverse flow of water that must take place in the confined space between the barge and the bed and banks. And because of the narrow space, the velocity gradient between barge and channel is steepened, which adds to the shear resistance. Hence a barge in a canal moves more slowly for a given motive power than the same barge in a water body of considerable extent.

B. Franklin in 1768 (Willcox, 1972 p. 115-118), after noting the drag encountered by vessels when rowed over shallow water, made tank experiments to demonstrate and measure it. The work was done in England and as his report was published among his observations on electricity, it was probably not known to the canal designers. It was not until a decade after the completion of the canal, the hydraulic elements of design became evident from research that had been published by DuBuat in 1786. Thus, in 1835 the problem was examined by John B. Jervis, Holmes Htithins.n, and Nathan S. Roberts. (Their reports are reprinted in 1863 A,. Doc. 8, p. 198 et seq.) Hutchinson (1835, As. Doc. 143, p. 42), referred to the experiments of DuBuat to the effect that the cross -section of the canal ought to be, with moderate velocity, "6 and 46/100 times" the cross section of the boat, and the water life 41/2 times the breadth of the boat. Hutchinson included the following data from DuBuat (order changed)

Ratio of sectional area Observed resistance

of canal and boat (in a water body of indefinite extent)

2.106 2.11

2.476 1.90

3.192 1.33

4.212 1.33

Apparently these data define a smooth curve, which extended, indicates a resistance ratio of 1.0 for a section ratio of 6.46, which accounts for the above recommendation. Jervis ascribed to the old Erie Canal a 40 percent excess in resistance; Hutchinson made it 59 percent, referring to the above table.

"The resistance in a fluid of indefinite extent being equal to 1, would give to the present Erie Canal boats of the largest class a resistance of 1.59 or 59 percent more than in a fluid of indefinite extent." Hutchinson stated that for the largest boats then (1835) on the Erie (14 feet wide, and 3.5 feet draft) "to move to the greatest advantage, as on an indefinite extent of water, the canal should be 63 feet wide at the surface, 5.11 feet deep, and have a cross-section of 271.32 feet." (sic) For boats 14 feet wide but of 4 feet in draft, the canal should be 63 feet wide at the surface and 6.08 feet deep.

Jervis recommended a canal section 42 feet wide at the bottom, 70 feet at the surface and 7 feet deep, for a barge 13 ½, feet wide and 41/2 feet draft. Nathan S. Roberts on January 17, 1835 (quoted in 1863 As - Doe. 8, p. 201) proposed a section 48 feet at the surface, 30 feet wide at the bottom, and 6 feet deep, effectively adding more to the depth than to width.

However. in the following year Jervis and Mills (1836 As. Doc. 99, p. 282-283) expressed the opinion that to accommodate the growing trade, a canal 80 feet by 8 feet would be most suitable for the enlargement (see also 1863 As. Doc. 8, p. 202).

DuBuat's rules are equivalent algebraically to a specification that barge draft not exceed 70 percent of the canal depth. So much was known in 1835. Yet the practice of building barges to occupy 85 percent of the depth was continued in the enlargement-3.5 foot draft in a 4-foot channel of the original; 6-foot draft in a 7-foot channel in the enlargement. For such drafts resistance was at least 50 percent greater than for DuBuat's recommended 70 percents.

Drag was particularly sensitive to shortages of water that made difficult the maintenance of full depths. For example, when water depth decreased from 4.0 to 3.9 feet, then drag on a vessel drawing 3.5 feet was increased by about 25 percent.

CANAL SLOPE, AND CURRENT

The section "Feeders, Locks, and Stream Crossings" lists the locks on the original Erie Canal as obtained from various published sources. Assuming that the lifts as given are correct, the net difference in levels between the Hudson River and Lake Erie, accounted for by the locks, totals 542 feet as against a difference in elevation of 572 feet between these two levels. This arithmetic leaves 30 feet or 1 inch per mile to be accounted for by friction slope in the reaches between the locks. This slope in a well-formed canal, 4 feet deep, would produce a mean current (west to east) of about 0.5 foot per second or about 0.35 mile per hour.

It was early understood that a current in the west to east direction could be a net advantage, since the heavy tonnage was carried in that direction. In principle, optimal advantage exists only when the downstream current is one-half the difference between the speed of a fully loaded barge and a lightly loaded barge when drawn through still water. However, the advantage is small and later as the boats were heavily loaded in both directions, the current was a distinct disadvantage. As Garrity recollected (1966), the westbound trip over the 62-mile level from Rochester to Lockport took 50 hours, or a speed of only 11 1/4 mph.

WEDGING

As wider barges were introduced (see table 4), collisions in passing became more frequent as State Engineer McAlpine re-

TABLE 4

Period Dimensions (feet)

Width Length Draft Capacity(tons)

1818-30 7 61 3 ½ 30

1830-50 12 75 3 ½ 75

1850-62 15 90 3 ½ 100

1862-99 17 ½ 98 6 240

ported in 1854 (Sen. Doc. 60, P. 77) "the wedging of boats is daily occurrence." By impeding the flow of water, the wedging of boats tended to ground those boats below the wedge. This experience was apparently not sufficient to deter the Commissioner from adopting 18,-foot Wide links in the enlarged canal despite Jervis' recommendation that they be held to a 16-foot width in order to limit the width of barges. Jervis (1877) had this to say of the Commissioners' decision which was impelled by a desire to float barges capable of carrying a load of 250 tons:

No one appreciated more than 1, the importance of substantive and durable works, but unfortunately at that time, there was so high an estimate of the value of the canal, that the ideas of men were extravagant, and advocated work of an expensive character, that was in no way substantial or useful. This included a more or less expensive policy, that increased the cost of the work much beyond the necessary. I had recommended, that the chambers of locks for a canal 7 feet deep by 70 feet top width, should be 16 feet wide and 155 feet between the gates. The Canal Board, on the petition of navigators increased the width to 18 feet which I regard an error.

DELAYS AT LOCKS

Waiting lines formed at locks, despite the fact that total capacity for service was well in excess of the demand. The waiting lines occurred simply because barges were dispatched at intervals during the day that suited the convenience of shippers and because slow barges tended to impede those behind. Both factors tended to increase grouping along the canal and concurrent arrivals at the locks.

Some data on detention and service times at locks (1844 As. Doc. 16, table P. 27) provide a classic example. With an average rate of arrival of 100 boats per day at a lock, well within service capacity, 40 percent needed to wait for service, yet the lock was idle 16 out of the 24 hours. Costs in time and temper were high and proper queue discipline became difficult to preserve.

Although unscheduled traffic was the cause, the response of the canal authorities was to enlarge the facilities. The improvement of the canal (1836-62) included the doubling of locks (in parallel) first in the heavily traveled and heavily locked section between Albany and Syracuse, and ultimately throughout.

The flood hazard to canals, as generally to linear transport systems including, roads and railroads, was of two kinds: that due to rivers paralleled by the canal and that due to streams that are crossed by the canal-the cross drainage. Neglect of the first proved catastrophic on the, later built Chesapeake and Ohio Canal (Sanderlin, 1946) that paralleled the Potomac River for most of its length; neglect of cross drainage proved to be an expensive nuisance on the Erie. In the absence of record keeping, it is fairly 'difficult to judge the flood potential of a. stream when it is in its low-water state, with sufficient confidence to justify what might then seem to be an inordinate construction cost. The builders of the Erie Canal were fortunate at least in that respect for, as previously mentioned (P. 16 ), severe floods in 1817 before construction began provided clear and persuasive evidence of the flood danger.

MOHAWK RIVER

Potentially the most hazardous location of the canal was the 110 miles that paralleled the Mohawk River. Destructive drainage here could potentially destroy the canal as a viable artery of commerce. That the canal was not subjected to general disruption as it might have been, proved that the indications of the flood of 1817 were apparently as satisfactory as they were assumed to be, even though the canal did not escape flood damage during the 19th century. Floods in 1821, 1832, 1833, 1846, and 1866 were reported to have reached the canal at one or more places, namely: November 1821 "raised the waters in the canal in some places above its banks" at Little Falls (1823 As - Jour. 46th sess., p. 504) ; March 13, 1832, at Schenectady- (breaks in the north bank of the canal) ice and flood damages to feeder dams across the Mohawk in March and April 1832; flood of May 1833 "covered the Erie Canal, in the valley of the Mohawk, to an uncommon extent, and for a few days partially interrupted the navigation." (1834 As Doc. 55, p. 45.) Feb. 1842, extensive injuries to banks (1843 As Doc. 25, p. 67) March 1846-Mohawk River ice laden, pours into bad break at Schenectady; June 1866-break 5 miles west of Schenectady 300 feet of towing path swept into Mohawk River (record not clear as to origin of washout-whether main river or side stream).

In their report for 1821, the Canal Commissioners (dated 27Feb. 1822, p. 23) refer to their route down the Mohawk river valley which they attempted to put

"in all places above the flood of the river; and avoid on the other hand, as far as was predictable, the sides of steep banks where the soil is liable to slip, and the canal to be otherwise injured by the torrents from the hills. The correctness of this location was tested by the great flood of November last (i.e. 1821), which suddenly raised the Mohawk river to an unusual height, was not observed anywhere to approach within many feet of the top of the banks, or to do any injury to the works which were completed."

Basil Hall (1829) who traveled on the canal along the lower Mohawk River noted (p. 119) "our perpendicular height above the stream may have been 30 to 40 feet." Based on trace s of abandoned sections on the topographic map's,'author's calculations show that the canal banks were placed about 20 feet above the river bed in the upper reach, 25 feet in the middle reach, and 35 feet in the lower reaches.

Except in the vicinity of Schenectady, the canal banks were as high or higher than the maximum levels reached by the Mohawk River during the 70 years or so of regular record keeping on that river. The exception at Schenectady was caused by an ice jam in 1914. lt must be concluded that the engineers succeeded in locating the canal if not, as claimed, "above the floods of the river," at least beyond the reach of ordinary floods.

CROSS DRAINAGE

Continuing with their report of the flood of November 1821, the Commissioners add (report dated 27 Feb. 1822, p. 23)

"On the land side, [i.e., cross drainage] more damage was sustained ; the flood from the hills filled the canal and in some places broke down the new and unfinished bank, destroyed the wing of the dams and injured several culverts."

Thus it was that the numerous small creeks that caused the major damage were a potential hazard early recognized by the Commissioners in the following language:

" To secure our work from injury, by floods and freshets, that will often suddenly collect, from the extensive land drain, and the abundant waters, above alluded to, we have been compelled to make numerous culverts and waste weirs. The office of a culvert is, to pass waters, not wanted for navigation, under the canal; that of a waste weir, to discharge the extra waters, that may be in it"

The object was clear, but it was not achieved. Disruption of traffic and costly repair, vexed the canal throughout its period of operations. A canal washout usually suspended traffic for a period of 7 to 15 days, as neither detours nor temporary service could be arranged. Even traffic on pounds not directly involved had to be suspended, for in order to. save water the locks could not be opened.

The problem began with choice of grade elevations for the canal as these fixed the clearance over streambeds. Raising the local elevation to improve clearance was not feasible on a canal as on a highway or even on a railroad. The adopted alternative was to excavate river and streambeds "for the purpose of free and safe discharge of water under aqueducts and culverts" (Hutchinson, 1834). This procedure could not work. River channels in nature are adjusted in elevation, slope, depth, and width to carry their loads of water and sediment. Any excavation of the bed can only be temporary as the stream sediment will return the bed to the original profile and choke the culvert.

In the building and in the rebuilding throughout the 19th century, emphasis was put upon structural solidity, as, for example, this report on construction:

"Stone culverts of different sizes, all to be arched and placed upon permanent foundations, and more than half of which are now finished with great solidity and beauty"

Yet the design of a device to render the simple service of a culvert is now known to be Complex, involving not only the structural details but their hydraulic properties and the flood discharges that are likely to be experienced. Without this information, repair and enlargement became surrogates for design. When a culvert washed out, the tendency would be to rebuild it larger so far as was possible. Several washed-out sections in the abandoned canal were observed by the author where the bed of the canal was but 1 or 2 feet above the bed of the stream crossing. Such culverts could only be widened, but widening would be hydraulically less effective than adding to depth. Even so, a practice of enlarged replacement could not cope with the problem. Stream crossings averaged about one per mile; thus the 360 mile canal contained that number of exposures to risk. If, say, as many as 25 culverts washed out in a 5-year period, and were enlarged, there would still remain 335 culverts. A century would not be long enough to remove the difficulty.

Any glamour in a culvert is hidden. Out of sight to the boatman on the canal as it is today to the motorist on a highway, it serves to carry the flow of a stream under the canal or road. Although each culvert may represent a very modest investment, their very great number and the disruptive effect of a single failure upon the entire transport system warrants a very great attention indeed.

OPERATIONS, MAINTENANCE, AND REPAIRS

A grasp of the operational problems and of the magnitude of the work needed to maintain the canal may be obtained from a comparison of costs. The base investment in the construction of the canal came to $39 million ($7.1 million for the original and $31.8 million for the enlargement (Whitford, 1906, p. 1366)). By 1882 total costs including operations, improvements (such as added feeders and locks), maintenance, and repairs came to $78.8 million (see table 1), indicating that added costs duplicated the original investment. Looking only at annual costs, the total for operations, maintenance, and ordinary repairs on the original canal came to about $280,000 or 4 percent of the $7 million investment cost; on the enlarged canal these averaged about $700,000 per year or 2 percent per year of the $32 million investment. Costs of operations, maintenance, and repairs on modern hydraulic public works are usually 1 percent or less.

Water required for lockage, although the most obvious to the planners of the canal, proved to be relatively minor compared with the amounts of water that were required to compensate for the leakage through the bed and banks of the canal. The total water intake of 700 cfs in the original canal increased to 1,900 cfs by 1862 as the canal was enlarged. The total quantities of water taken into the canal made it the largest hydraulic undertaking of the century in the United States. The diversion of water for the generation of power in factories attracted to the canal as a source of hydraulic power and means for the transport of goods added to the water requirements. Although new feeders and reservoirs to extend the supply were built throughout its history, these efforts to cope with water shortages were never fully successful. The primary cause of the persistent deficiencies in supply was the method used to estimate the flow of the streams available during extended dry spells. Dependence on spot, ad hoc measurements of flow consistently overestimated the supply.

Water shortages in the canal were also created by a lack of hydraulic capacity to convey the water needed to maintain full depths over long distances between feeders. Failure to maintain depth in the canal increased greatly the hydraulic drag encountered by the moving barges and lengthened the travel times.

A decrease of the bottom clearance normally 0.5 foot, by only 0.1 foot added to the drag by 25 percent. The hydraulic and traffic problems of the canal were exacerbated by the practice to build and operate barges as large as possible, which then led to enlargement of the canal. Unlike the railroads, the canal was not viewed as a unified traffic system, a fact that contributed to its second place in that competition.

What might have been known beforehand of some of the water problems of canals is contained in a critique of British canals that was published in 1816 (Sutcliffe, 1816). Almost a polemic, this report found fault with economics and engineering nearly everywhere in Great Britain. As to economics, Sutcliffe concluded (p. 74), "If the average of profit and loss, of all canals that have been projected and executed within the last twenty-five years were accurately taken, I am inclined to believe, that the balance would be found greatly against them."["The last twenty-five years" would refer to the final or second generation of British canals.] As to engineering, he offered his critical judgment on construction and on inadequacy of water supply, noting (p. 77), "It is difficult to say, whether the estimates of canal engineers for supplying them with water, or their estimates for executing, are the more erroneous," following the observation (p. 73) that "many of them are little better than so many dry ditches." The earliest reference to Suteliffe's treatise found in the Erie Canal literature was some 20 years after publication as part of the newly learned hydrologic computations for the enlargements.

The major flood problem was caused by cross drainage-the small creeks that crossed under the canal in culverts. Washouts of culverts were a never-ending source of sporadic disruption of traffic, each occurrence being of 1 or 2 weeks duration and adding greatly to shipping costs. Repairs and replacement could not cope with the problem created by deficiency in information about the flood potentials of the small streams.

A fortunate occurrence of a severe flood in 1817 when construction began provided such clear and present evidence of the flood potentials of the Mohawk River, which the canal followed for 100 miles, as to compel putting the canal at a high level in difficult terrain. Otherwise, there can be little doubt that the Erie Canal would have been of no more use for transport than the Chesapeake and Ohio Canal, which was put out of service repeatedly by -floods on the Potomac River.

The canal survived its persistent operational problems, which were the inevitable consequences of moving ahead on a grand scheme without having obtained and -analyzed some of the pertinent data.

| Environmental Aspects of the Erie Canal | The Canal and the Environment |