I am informed that in very early days the Oregon Steam Navigation Company operated a line of steamers from Portland to the obstructions in the Columbia River at "The Cascades," now known as Cascade Locks, Oregon (about 43 miles above the mouth of the Willamette River), there transporting cargoes overland by a portage road, 6 miles long, around these obstructions. From Cascade Locks up to the town of The Dalles, about 45 miles, steamers were again used in what was then called the Middle River.

At The Dalles a second portage road, on the Oregon shore, about 13 miles long, transferred freight and passengers around the obstructions between Three-mile Rapids and Celilo to the "upper river," and there again vessels were taken for the Upper Columbia and Snake rivers. Later, upon the completion of the Portland-Spokane rail line of the Oregon Railroad and Navigation Company, practically paralleling the Columbia and Snake rivers between Portland and Riparia, Wash., the use by boats of the Columbia above Celilo and of the Snake below Riparia was materially diminished, so that at the present day the only navigation of either the Columbia or the Snake River above Celilo of any special importance is carried on along that portion of the Snake between Riparia and Lewiston, 73 miles, where a daily line of steamboats is still running, connecting with the Oregon Railroad and Navigation Company's rail line at Riparia and with the Spokane-Lewiston rail line of the Northern Pacific Railroad Company at Lewiston. The need for better commercial connection between the interior country and the coast has, however, grown materially in recent years on account of the great increase in the acreage of the wheat-producing regions and other tilled land and on account of the discoveries of gold, copper, and other valuable minerals in Idaho and eastern Oregon and Washington. The natural rapid development of the interior country has given the question of transportation facilities a decided importance. The Columbia River was formerly obstructed at "The Cascades," about 43 miles above the mouth of the Willamette, by rapids over which vessels could not pass safely, and for many years the Government has been engaged in constructing a canal around these rapids, which canal, although in some respects still incomplete, was opened to navigation in November, 1896, so that the river may now be considered open to navigation from its mouth up to The Dalles, Oreg., a distance of 186 miles. The only remaining obstruction in the Columbia River below Priest Rapids that is insuperable at the present time for navigation is, as was before referred to, in the portion lying between The Dalles, Oreg., and Celilo, a stretch of about 13 miles, where all continuous navigation is interrupted at all stages.

As before stated, the river above Celilo is navigable for light-draft steamboats drawing, loaded, not over 4 feet of water, although this navigation, especially at and near low-water stages, has always been carried on with some danger and difficulty, owing to the rapids and rocky bars that are encountered more or less frequently in the Columbia above Celilo and in the Snake between Ainsworth and Riparia. Improvement by the Government has been carried on for many years in these portions of the Columbia and Snake, materially lessening the difficulties of navigation.

Although the work of improving the Columbia from The Dalles to Priest Rapids and the Snake from Ainsworth to Lewiston so that navigation for light-draft steamboats may be provided during the

greater portion of each year would be one of considerable magnitude, it would not involve any insurmountable difficulties from an engineering point of view. It would be mainly a question of expense.

The opening of the Columbia at "The Cascades" in 1896 by the partial completion of the canal and locks around the obstructions at that place has given an added reason for some improvement of the Columbia in the 13 miles between The Dalles and Celilo, so that the regions above The Dalles may be reached by an open river and thus permit these regions to get the benefit of water rates for freight.

Obstructions.-Much if not all of the areas of Idaho, Washington, and Oregon are believed to have been covered in some former geological age by successive layers of hot lava, which in cooling have formed the basaltic cliffs and the igneous rocks of various degrees of hardness now forming the greater part of the surface of these regions. These layers of basalt in some localities are said to be 4,000 feet thick. The Columbia River between The Dalles and Celilo has apparently been slowly cutting its way for centuries through what appear to be thick layers of hard basalt and andesite. The total low-water discharge of the river as computed from observations taken near this place during the survey just completed is 108,000 cubic feet per second, and the extreme high-water discharge is likewise computed to be 1,600,000 cubic feet per second. This discharge must all pass through the narrow and crooked channels worn in the volcanic rock between The Dalles and Celilo. These conditions combine to form complete obstructions to navigation of a very unusual nature.

In no part of the other 385 miles of the Columbia River between its mouth in the Pacific Ocean and Priest Rapids is anything like such a condition met with. These obstructions between The Dalles and Celilo consist of rapids, falls, and contracted channels, and may be divided into four parts or sections: First and uppermost, Celilo Falls, at Celilo, which has a fall of 50 feet, approximately, at low water, in a distance of about 3,000 feet, where the bed of the river is suddenly broken off and filled with rocky islands and ledges, over and between which the water rushes with great fierceness (see photograph No. 1). There is at one place at Celilo a sheer fall of 21 feet, approximately, making it. absolutely impossible for any vessel to pass this point with safety at anything like low-water stages. Below Celilo Falls the next obstruction consists of what is known as Ten-mile Rapids (see photographs Nos. 2 and 3), where the whole volume of the river at low water passes through a narrow channel about 2,700 feet long, hemmed in on both sides by high banks of basaltic rock, nearly vertical, and from 200 to 300 feet apart. Behind one of these banks, however (on the Washington or north side), at stages about 20 feet above low water, there is a high-water channel, so that when the surface of the river is raised at this point to the extent of about 20 feet it spreads over a greater area, thus increasing its area of cross section and decreasing its slope and velocity.

A

Just below Ten-mile Rapids are a number of rocks which, although dangerous in their present condition for any use of the river by boats at this point, could, when found necessary, be removed at comparatively small expense, so as to cause but little trouble to vessels. short distance farther down the river the third obstructions, and by all means the most serious of all, are what is known as Five-mile Rapids or the "Dalles" (see photographs Nos. 4, 5, 6, and 7), this word being

said to mean "trough" or "mill race." Here, for the distance of about 9,000 feet, the total low-water discharge of the Columbia is forced between high banks of solid rock nearly vertical and in some places not over 160 feet apart, the low-water surface having a fall of 11.2 feet in this distance. The banks are solid basalt rock, precipitous, and from 40 to 60 feet high, and the bottom is a succession of pockets 40 to 60 feet deep. On account of the choked condition of the channel and the quantity of water forced through it the velocities are high and cross currents and eddies so serious that large vessels have never attempted to go up through these rapids even at most favorable stages. The remark is often made that this section of the river is "turned up on edge," and the conditions justify the use of such a description. This narrow portion of the river ends, at its lower end, in a whirlpool of great depth, known as "Big Eddy" (see photograph No. 13), where the river makes a bend at right angles to its course just above.

Along the Oregon or south side of this narrow portion of Five-mile Rapids for much of its length are found two natural high-water channels or canals (see photograph No. 7 for general view of both canals and photographs Nos. 8, 9, 10, 11, and 12 for details of the natural canal lying next south of the river) lying nearly parallel to the river. These natural channels offer advantageous sites for canals, and it is expected to utilize the one lying next south of the river in the system of canals proposed herein. About 1 mile below Big Eddy the fourth obstruction is met with, known as Three-mile Rapids (see photograph No. 14), where the channel is considerably obstructed and contracted at several places by islands and by rocky points extending from both the Oregon and Washington shores into the river. This last obstruction is the only one that seems to lend itself with any ease to improvement. It is believed that the removal of small portions of rock at several places along Three-mile Rapids will also remove the dangerous features of these rapids.

The condition of the four above-named obstructions has been described with regard to low-water stages. The problem of improvement, already extremely difficult on account of the natural conditions existing at low water, is vastly complicated by the change that takes place in this vicinity during high-water stages each summer. The total fall in the 9 miles of river from above Celilo Falls to quiet water below Big Eddy during the low-water stages is, approximately, 81 feet. The greater part of this fall during low water takes place at the upper end, at and near Celilo Falls. The fall throughout the remainder of the distance is not absolutely prohibitive of the use of it at extreme low water by vessels, provided the channel be somewhat widened and otherwise improved. As the water level rises, however, these conditions are reversed; the surface slope at Five-mile Rapids increases and that at Celilo Falls decreases, so that at high water the water surface at Celilo Falls has a slope of only about 3 feet to the mile instead of about 50 feet, whereas the fall at Five-mile Rapids increases from a total of 11.2 feet to the maximum of 35 feet, or from about 6.6 feet to the mile to 21 feet to the mile. These conditions are largely caused by the conformation of the banks above the bench that forms the low-water banks, causing at all stages a gorging of the water. (See profile map of water surface.1)

1 Not printed.

Further detailed descriptions of the Columbia River and the obstructions between Celilo and The Dalles are contained in the Annual Report of the Chief of Engineers, United States Army, for 1882, Part III, pages 2690-2713; in House Ex. Doc. No. 73, Fiftieth Congress, second session, and in Senate Ex. Doc. No. 7, Fifty-third Congress, first

session.

Previous examinations and surveys.-The first investigation of the obstructions between Celilo and The Dalles by the Government, made with a view to their improvement, was made in 1874 by Maj. N. Michler, Corps of Engineers, whose report is to be found in the Annual Report of the Chief of Engineers, United States Army, for 1875, Part II, page 787. The report of this examination was favorable for improvement, Major Michler reporting as follows:

The importance of the proposed improvements can not be overestimated. The difficult problem of engineering which must be solved will, at least at one of the localities, The Dalles (Five-mile Rapids), at first study appear almost insurmountable. * *

*

Later investigations were made by the Government under the direction of Congress in 1879, and a detailed and comprehensive survey, with a project for improvement, was made by Capt. (now Major) C. F. Powell, Corps of Engineers (report printed in Annual Report of the Chief of Engineers for 1882, Part III, p. 2690), which provided for use of the river at low-water stages from the town of The Dalles, Oreg., up as far as the foot of Celilo Falls by an open river without locks. The passage of Celilo Falls was provided for by a canal and locks on the Washington bank. This project required the excavation of much rock in the open-river portion in widening and straightening the channel from The Dalles up to Celilo Falls. The estimate of cost of this project was $7,674,495.51. A supplementary project, permitting the use of the river during high stages, was included, requiring $2,842,848.20 additional, or a total of $10,517,343.71 for improvement for all boating stages. Major Powell in his report stated as follows:

The Columbia is the great river on the Pacific coast. In volume and commercial value it is second only to the Mississippi. Its banks are more stable, its waters are more clear, its ice blockades are much less in duration than on the great waterway of the East. Unlike the Mississippi, the Columbia seeks the ocean on a line parallel to trade channels and not at right angles to them.

The high cost of this plan of improvement, however, probably induced Congress to order further investigation, and in 1888 a Board of Engineer Officers was convened to report upon the feasibility of a plan selected by Congress of overcoming these obstructions by means of a boat railway, by which vessels were to be removed entirely from the water by vertical hydraulic lifts, the one at one end 77 feet high and the one at the other end 62 feet high, the boats to be transported on a car over a broad-gauge track, about 9 miles long, around the obstructions, and then returned to the river by the lift at the other end. This report, submitted in 1890, was favorable, and the cost was estimated at $2,860,356.35, or nearly $3,000,000. A portage road around the obstructions, constructed and operated by the Government free of cost to shippers, was recommended by the Board for immediate relief, the boat railway not to be put in until requirements of navigation demanded (House Ex. Doc. No. 64, Fifty-first Congress, first session1).

Also printed in the Annual Report of the Chief of Engineers for 1890, page 3028.

In 1892 a second Board was convened to reconsider these plans. It submitted a report under date of April 12, 1893 (printed in Senate Ex. Doc. No. 7, Fifty-third Congress, first session), recommending the construction of a free portage road first and the construction of a canal with locks as a later and more permanent improvement, to be built when found necessary to meet the demands of navigation. A portion of this Board, however, submitted a minority report, recommending a boat railway, practically as embodied in the previous report of 1890. Congress, apparently acting on the recommendation of this minority report, appropriated, on August 18, 1894, $100,000, and on June 3, 1896, $150,000 for the construction of this boat railway without making any provision for the construction of a free portage railway. Preliminary work connected with the boat railway was started promptly, the acquisition by the United States of right of way being the first step. Condemnation proceedings were instituted in some cases, and in others the land was acquired by direct purchase. Preliminary plans have been prepared, the ground surveyed, and the line of the proposed railway carefully laid out. Actual construction of the boat railway has been delayed, however, pending the time when full right of way, which has not yet all been obtained, could be secured, further action by Congress being first necessary in order to effect an exchange of land belonging to the United States for land belonging to the Oregon Railroad and Navigation Company.

Many objections have, however, arisen among river men to any scheme for improvement of this section of the river by which their vessels are to be removed from the river and carried about 9 miles overland on a railway, and others have objected on the ground that the estimated first cost and the difficulties of maintenance and operation of a boat railway would likely be found greater in practice, no such boat railway having ever been successfully operated, the plan being as yet only experimental. It is also maintained by some that any device that removes a vessel from its proper element and carries it on a car over a track 9 miles long, which necessarily has both vertical and horizontal curves, would, in the case of the light-draft river vessels used in shallow streams, likely expose them to serious twisting and probable damage, especially if not carefully loaded. It has even been asserted that those especially interested in delaying as much as possible any adequate improvement of this part of the river for personal reasons were outwardly much in favor of a boat railway, as they believed the plan impracticable, and therefore not likely to be successfully completed.

Present survey. This locality having been surveyed carefully in 1879-80 under Maj. C. F. Powell, Corps of Engineers, the present survey has had for its main object obtaining certain further, more definite information and checking the old maps for any changes that might have taken place since that date. The time available for this present survey being too short for the observation of gauge readings and of changes in water levels necessarily extending over a large part of a year, and this data being already at hand as obtained in the survey of 1880, with no special need of duplication, the water-level observations of this previous survey have been used for this project, modified in some minor particulars by the observations of the present survey. A careful topographical survey of the proposed canal sites was made

'Also printed in the Annual Report of the Chief of Engineers for 1894, page 2664.