Coulee City to Ellensburg?

The below letter describes a through route from Coulee City to Ellensburg that never was completed. The drawn map does not survive, to my knowledge. The route was partially built as the Adrian cutoff, from Coulee City to Crab Creek.

C.C. Van Arsdol, Assistant Engineer, Northern Pacific Railroad, Tacoma, Wash., to
Charles S. Bihler, Division Engineer, Northern Pacific Railroad, Tacoma, February 4, 1900.

I beg to hand you herewith a sketch map, and profile of grades from
reconnaissance of proposed lines from Coulee City, via Crab Creek and
Johnson Canon to Ellensburg; and from Lind via Lind Coulee to connection
with above line from Coulee City.

From Coulee City it would be impracticable to follow the Grand
Coulee, which a few miles below Coulee City falls abruptly about four
hundred feet with no ground for supporting. The bench on the west side
disappears about opposite of the head of the lake, and the bench or
table on the east for a distance of three or four miles back from the
coulee, is broken by frequent short side coulees.

The route I have outlined follows a comparatively even scab rock
country between the breaks of the Grand Coulee and the breaks of Crab
Creek to the east; supporting down on easy slopes to Crab Creek at a
point a few miles east of Adrian (on the G.N.Ry); thence via Willow Lake
Coulee, and smooth table lands to Lind Coulee. From this point I have
indicated two lines, one via Lind Coulee and Crab Creek; the other
supporting up the south slope of Lind Coulee to a saddle between Lind
Coulee and the head of Dry Coulee, and down Dry Coulee to Crab Creek.
The distance and approximate cost would be about equal by these two
lines, abut taken in connection with the proposed line from Lind, the
latter would give about 12 or 14 miles shorter line between Lind and
Ellensburg. From the junction of Dry Coulee and Crab Creek the line
would follow the smooth valley of Crab Creek, then crossing of the
Columbia about 3/4 mile below Crab Creek; thence up the Columbia Valley
to Johnson Canon and via Johnson Canon to Ellensburg.

There is a possible alternate route between the Columbia and
Ellensburg, via Harrison Canon and the heads of Squaw Creek. This
however has a summit 150 feet higher than Johnson Canon and apparently
less favorable slopes for development or support, on the Columbia side.
East of the Columbia there is a coulee about 6 miles north and
parallel with the lower Crab Creek, which may prove feasible if a
suitable connection can be found between the head of this coulee and the
mouth of Lind Coulee. This point I have not examined.

At the crossing of the Columbia indicated on the map a rock ledge 4
to 6 feet below high water, projects out to the waters surface. The
channel at low water is about 1200 feet wide, and at high water about
1800 feet wide. The western bank is coarse gravel and boulders
indicating that the bedrock may extend across or nearly across the
channel.

Following is a statement of approximate distances.
055 miles -- Coulee City to common point at head of Dry Coulee
070 miles -- Head of Dry Coulee to Ellensburg
006 miles -- Development Johnson Canon for 2 1/2 percent grade
west bound and 1 1/2 percent grade east bound
131 miles -- Total Coulee City to Ellensburg

021 miles -- Lind to head Dry Coulee
070 miles -- Head Dry Coulee to Ellensburg
006 miles -- Development of Johnson Canon
097 miles -- Total Lind to Ellensburg

Following is approximate cost of grading and bridging.
340,000 -- Coulee City to head Dry Coulee
559,750 -- Head of Dry Coulee to Ellensburg exclusive of
Columbia River Bridge
210,000 -- Columbia River Bridge

073,500 -- Lind to head of Dry Coulee

[1,183,250 total]

The above estimate is based upon the use in Johnson Canon of 1 1/2
percent grade west bound and 1 1/2 percent grade east bound. If a
lighter west bound grade were adopted, the mileage would be
proportionately increased and the cost per mile for grading the lower 8
miles of Johnson Canon would be considerably heavier. Increasing or
decreasing the rate of grade west of the summit of Johnson Canon would
proportionately increase or decrease the distance, but would not very
materially affect the cost per mile on that portion of the line.