top of page

I.C. Russell's Reconnaissance in Central Washington, 1892

J.S. Diller, the Geologist in charge of the Cascade Division for the U.S. Geological Survey, dispatched a young Israel C. Russell to central Washington in the Spring of 1892 to assess the region's potential for artesian water. Westward expansion, largely through homesteading and irrigated farming, was a priority of the U.S. Government at the time. USGS was the agency charged with irrigability assessments across the West, their work in no small part defined the pathways of pre-WWII settlement. Russell's traverse through central Washington (March 22 - June 30, 1892) covered an enormous number of miles - from Grand Coulee to Waterville to the head of Lake Chelan, past Wenatchee and White Bluffs, and several back-and-forths across ridges of the Yakima Folds. His report was published as USGS Bulletin 108 (Russell 1893). Interesting quotes from the's report are highlighted below. My additions are in square brackets.

Russell's field area covered in three months during 1982.


Russell, like other geologists of the late 1800s, worked at a broad scale. Formation names familiar to us today were just beginning to be formalized. Russell named the Columbia River Basalt, the Ellensburg Formation, and others. In Russell's time, Eastern Washington's geology consisted of about six units that have since been refined, reorganized, and renamed as dozens of different formations.

'Crystalline rocks' - Mesozoic and Paleozoic granites and metamorphic rocks of the Okanogan Highlands and Cascades. Rocks that we recognize today as accreted terrains and old arc assemblages.

'Kittitas system' - Pre-Miocene (Lower Tertiary or 'Paleogene') sedimentary and volcanic rocks that rest unconformably atop eroded crystalline basement and associated metamorphosed sedimentary units, and underlie the Columbia River Basalts. Contains tilted sandstones and shales with coal reserves exposed near Wenatchee, the Teanaway dikes, and graben volcaniclastics. Russell coined the term Kittitas system, which lives on in synthem literature (Cheney 2016).

'Columbia lavas' - Miocene Columbia River Basalt and various interfingering/interbedded sediments sourced from west and north, all of which we now call the Ellensburg Fm. The Ellensburg includes a few thick airfall ash beds ("lapilli"), quartzitic gravels, diatomite beds, and the Gingko-type petrified logs, and is equivalent to Latah Fm interbeds found to the east (clayey, lacustrine, fossiliferous beds of Pardee and Bryan 1925),

'John Day system' - Includes undeformed, unconsolidated, white tuffaceous, post-CRB sediments believed at the time to have been deposited in large lakes. Most of these sediments are what we today call Neogene continental basin-fill sediments of the Miocene Ellensburg/Latah and Pliocene Ringold Fm. The so-called 'John Day lacustral system' was not well delineated or understood in Russell's day. Light-colored, quiet water strata exposed at White Bluffs (Ringold Fm), beneath Eureka Flat (Ringold Fm), Wenas Creek (Ellensburg Fm), Thorp (Ellensburg Fm), and Naches (Ellensburg Fm) were considered coeval with similar-appearing strata in The Dalles area, along the Snake River at Hagerman, ID, and in the Painted Hills along the John Day River of central Oregon. It was generally thought that following the cessation of CRB eruptive activity the area between the Cascades and Rockies was covered by a vast lake system (Lake John Day). The post-basalt lakes received great quantities of volcanic ash from eruptive centers far from the basalt plateau. Lots of fossil work was being done in Tertiary lake strata at the time, but a lack of precise age control resulted in some correlation errors between large exposures of the region - Miocene or Pliocene? In an 1897 report on the geology of the Palouse-Snake River region, Russell admits, "The connection between [central Washington and the John Day Valley] has not been actually traced, however, and the correlation...must be considered as provisional". Age relationships were eventually sorted by later workers (Farooqui, 1981), but questions surrounding post-Miocene drainages and inter-basin connections still linger (see Lydia Staisch's work). To further add to the confusion, certain Ice Age flood deposits were taken for 'John Day' (i.e., Touchet Beds exposed at Zillah, WA). The term 'John Day system' has largely been abandoned and the regional lake concept revised. Big Pliocene lakes were certainly a thing in teh West, but they mostly developed separately in isolated subbasins. Once in a while they spilled, carving canyons between them (i.e., Hells Canyon at ~4 Ma).

'Glacial records' - Russell did not recognize Ice Age flood features in Eastern Washington for what they were. He was there a bit too early. Bretz showed up in the 1920s. Bretz and Russell likely knew each other through their publications and perhaps met at meetings, but I know of no collaboration between these two field men. They certainly covered much of the same ground (on foot) and I surmise they would have shared considerable respect for one another, a common trait among field geologists. Many features that Bretz would later identify as scabland were noted by Russell, but not without some unease. Hints of doubt creep into Russell's reports from the field. Certain phrases are unusually terse for the loquacious writer, "Columbia river was dammed by these glaciers and escaped southward through Grand Coulee". Several times he all but apologizes for the accelerated pace of his investigation - inadequately detailed with "meager" observations, especially along the former glacial margin. I suspect his boss, J.S. Diller, wanted the f-ing job done quick. In the end, all of the scabland features Russell observed were attributed to work by fluvial processes of a more vigorous Columbia River during a wetter Pleistocene. His ideas persisted into the 1920s (Leighton, 1919), when Thomas Large and J Harlan Bretz began to be challenge them.

'Lake Lewis' - Symons mapped the shoreline of 'Ancient Lake Lewis' in 1882 as part of a navigability assessment of the Columbia River. He believed glaciers from the north terminated in a large Pleistocene-age proglacial lake that spanned most of south-central Washington. Symons used elevation readings, slackwater deposits, and ice-rafted erratics to delineate the shoreline, but in places included some older sediments (Ringold Fm). Today, Lake Lewis describes the maximum upstream limit of slackwater ponding of glacial floodwaters against Wallula Gap (~366m shoreline elevation). Its a composite shoreline, in places knife sharp but more often fuzzy.

Leutinent Thomas W. Symons, one of few professionals to survey the Columbia Plain prior to Russell, described his reconnaissance mapping in a straightforward terms befitting a ship's captain, "I have endeavored to outline this ancient lake as far as practicable, and propose for it the name of Lake Lewis, after Capt. Merriweather Lewis, the leader of the exploring party which first saw any of the headwaters of the Columbia."

Lake Lewis defined by the 380m elevation by Bretz (1919) and a map I created from digital elevation data using the 400m contour as the shoreline. The single elevation contour does a lot of work in predicting where flood deposits occur in the Scablands. Also, note the strong valley connections to the north and east.

The U.S. Geological Survey was created in 1879. Israel Russell joined his first survey in 1880 and would go on to become one of the great American field geologists - an explorer, a writer, a teacher. He was short, quiet, and tough. A good horseman and a keen observer. Prior to his arrival in central Washington, Russell (2nd from right in photo) lead a boundary survey crew through difficult terrain in Alaska's Yakutat-Mount St. Elias region.


"The main object of the present paper is to direct the attention of the people of central Washington to the features in the geology of their country which bear on the possibility of obtaining subsurface water under pressure. The better those interested in drilling wells are acquainted with the geology of the country in which they work, the more likely are they to avoid carrying on their search in unpromising localities." (p. 14)

"The conclusions reached are in general not favorable to the view that a large portion of the central part of the state can be irrigated by artesian water, but it is hoped that by pointing out the unfavorable conditions a considerable waste both of capital and energy in drilling useless wells may be avoided." (p. 13)

"The exceptions are Moxee valley, where the existence of a small artesian water supply has already been demon-

strated, and possibly the eastern part of the Yakima Indian reservation [where the Wapato Irrigation Project has since been built], where, as suggested in a previous page, the conditions are such as to justify furtlier inquiry."

"Plans for taking water from Spokane river and conducting it westward to the thirsty land of Lincoln and Douglas counties has also been suggested. These projects, however, involve the expenditure of large sums of money in order to carry them into effect, even if engineers report them practicable." (p. 83)


"The region traversed embraced about 10,000 square miles." (p. 11)

"...this region has received the name of the "Great Plain of the Columbia" [term borrowed from Symons 1882] more familiarly it is known as the "Big Bend country," from the fact that it is embraced in a great western curve of the Columbia river." (p. 17)

"The great plateau region east of the Columbia, where the relief is low, trees are absent, streams are few and feeble, but deep canyons cut through slightly disturbed rocks, indicate a former period of more abundant precipitation. The land is desert-like, but produces luxuriaut bunch grass." (p. 19)

"The region was previously almost entirely unknown geologically, and for this reason, at least, it is thought that the results reached will be of interest. (p. 14)

"As the region explored is one of a few within the United States concerning which there is but little information to be obtained from books," (p. 12)

"The Northern Pacific railroad crosses the southern portion of the area examined, and the Great Northern railroad was being built across its central part during the season of our visit." (p. 14)

A portion of Russell's central-Washington study area map showing travel routes in red lines.


As the region it occupies is drained almost entirely by the Columbia river, I venture to name it the "Columbia lava."

"The Columbia lava is not one vast flow, but is composed of many separate flows, sometimes separated by land surfaces, which frequently contain the stumps of large trees, or by sheets of lapilli...The sheets of which it is composed overlap and supplement each other, so as to form one continuous but highly compound [Miocene-age Ellensburg] system. No single sheet can be traced over the entire field...individual flows may be followed continuously for a score or more of miles.

The [Ellensburg-affinity] layers of sediment interbedded in the basalt have been baked and considerably altered by the heat of the inclosing rocks. (p. 24)

"The records seen in many sections of lava and lake beds show that the period of extensive [Miocene] volcanic overflow ended in a [mostly Pliocene] lacustral period, during which the region...of the Cascades eastward to the mountains of Idaho was occupied by a vast lake of Miocene age [the Tertiary 'John Day' system], in which many hundreds of feet of sediments were laid down...Over large areas nearly the entire series has been washed away, leaving the lava as the surface rock. Remnants of the John Day beds occur throughout our field of study, with the exception of the granitic areas at the north, but are best exposed in the White bluffs of the Columbia [Pliocene Ringold Fm], and in Wenas and Naches valleys [Miocene-Pliocene Ellensburg Fm]. (p. 22)

"The foregoing account of the John Day system is somewhat at variance with previous descriptions of what is supposed to be a southern extension of the same formation. In describing the deposits of a Miocene lake in northern Oregon Prof. O.C. Marsh says, 'The typical localities of this Miocene basin are along the John Day river, and this name may very properly be used to designate the lake basin.' " (p. 24)

Miocene-age Ellensburg Formation sediments, which Russell and others of his time considered part of the "John Day" system, interfinger with and overlie flows of the Columbia River Basalts. Russell gave them the name "Columbia lava". The sketch above is an early, but clear depiction of the relationship between sediments and lava flows along the east flank of the Cascades. Sedimentary interbeds contain water and natural gas as well as evidence of landscape development (erosion, deposition, soil development) that occurred between lava flows erupted during a remarkably brief period around 16 million years ago (Kasbohm et al., 2018; 2023).


"The evidence that the main drainage lines of central Washington were established before the present relief was initiated, is cumulative an d abundant. The rivers began to flow when Lake John Day was drained, and had their courses determined by the slope of the surface of the bottom of the old lake." (p. 97)

"There is one deep canyon-like valley in the lake beds, however, which starts near the east end of Saddle mountain and leads to the Columbia. This is clearly an ancient drainage channel, but its individual history has not been traced." (p. 97)

In his 1897 report on the Palouse-Snake River region, Russell notes, "Several considerations lead to the conclusion that the [Snake River] canyon was excavated to its present depth in late Tertiary times, and was filled in with gravel during the Glacial period. If this inference is correct, the amount of excavation done since the Glacial period is trifling in comparison with the work performed by the river previous..."

'Sketch map of Yakima Canyon, showing anticlinal axes' by G.O. Smith (1903).


"After Lake John Day was drained and its sediments and the rocks on which they rest uplifted into mountains and deeply eroded, there came a climatic change which admitted of the existence of glaciers in the Cascade and Rocky

mountains. A great glacier then filled the valley now occupied by Lake Chelan; another descended Okanogan valley...Columbia river was dammed by these glaciers and escaped southward through Grand Coulee." (p.

"A great glacier flowed southward down the broad Okanogan valley, and reaching the Columbia, not only turned the river from its course but was of sufficient volume to fill its canyon there 2,400 feet deep, and to cross it without apparently being deflected from its general southern course. After crossing the canyon [onto the Waterville Plateau] it continued southward on the Great Plain of the Columbia, grinding down and scoring the basaltic rock over which it passed and scattering thousands of huge bowlders over the plateau. The southern limit of this ice invasion was in the neighborhood of Coulee city, and its extension eastward was limited by Grand Coulee. At the south it ended in a great lake, known as Lake Lewis, briefly described below, on which icebergs floated and carried bowlders far and wide over the region to the south. These bowlders of granite, gneiss, basalt, etc., came from the north, some of them perhaps from Canada. They are the latest of all geological records in many portions of our field, and the one that, perhaps, claims the greatest amount of popular attention." (p. 25)

"Lake Lewis. The glaciers from the north ended in a large lake, the northern shore of which crossed the central part of Douglas county, and is known as Lake Lewis. Icebergs floating on this lake carried their freight of bowlders over the Great Plain of the Columbia and into many valleys opening from it." (p. 11)

Lake Lewis was first recognized so far as I am aware by Lieut. T. W. Symons [Symons 1882] and named by him in honor of Capt. Merriweather Lewis... (p. 26)

"On the map referred to the southern shore of Lake Lewis is represented as following a general east and west course through

Walla Walla and Wallula. Present information concerning the topography of this region indicates that the [shoreline] boundary is much farther south than is indicated by Symons..." (p. 26)

"...if we had a good contour map of the country we could trace with a considerable degree of accuracy the outline of its shore. In the absence of such a map we can only say that its water extended far south into Oregon and probably far enough east to cross the present Washington-Idaho boundary. The lake probably owed its existence to a dam at the Dalles - perhaps a glacier then obstructed the drainage or else there was a subsidence sufficient to allow the ocean to enter the great central valley between the Cascades and Rocky mountains [seawater incursion up the Columbia Gorge was an idea later pursued by Allison(?)]."

"No deltas built by streams flowing into the lake have been recognized, unless the heavy gravel deposits filling the canyon of the Columbia to the depth of 700 feet in the neighborhood of Lake Chelan, be considered as of this nature." (p. 27)

"Lake Lewis existed at the time of the greatest extension of the glaciers on neighboring mountains and was fed by their melting. It was a Pleistocene water body, and had many contemporaries in the now desert valley of the arid region between the Eocky mountains and Sierra Nevada." (p. 28)

"The region embraced in the present reconnoisance only touches the southern part of the glaciated area, and an extended study of the ice records will have to be postponed." (p. 25)

The Great Terrace near Chelan, WA. USGS photo 1929.

The Great Terrace at southern end of Okanogan Valley (Flint 1935).


"In journeying along the Columbia and in ascending Lake Chelan, observations were made on the main features in the geology of the southern border of Okanogan county, which it may be well to record, since so little is known of that promising region." (p. 76)

"...we notice that there are many terraces on each side of the river. The most remarkable of these, and one of the finest

examples of terrace structure that can be found anywhere, is a level-topped shelf formed of gravel and water-worn bowlders, the surface of which is 700 feet above the [pre-dam] Columbia. This truly remarkable terrace...several hundred feet broad and runs back into lateral gorges...the sides of the main canyon are deeply scored by lateral drainage before the gravel forming the terrace was deposited. On the west side of the valley there are other fragments of the same deposits...built against the steep slope, and has the same level as the great terrace on the east side of the river." (p. 78)

"The terrace gravels extend into the valley of Lake Chelan and form conspicuous terraces about its lower end. For many miles both up and down the Columbia other fragments of the same level topped deposit occur, always forming striking features in the landscape, owing to the marked contrast of their smooth horizontal lines with the vertical line due to the erosion of rills and creeks." (p. 78)

"Beside the great terrace...there are many others but less conspicuous horizontal lines on the sides of the Columbia canyon. Some of these below the horizon of the main terrace are stream terraces, made by the river in lowering its bed."

"Still another class of terraces, both numerous and conspicuous, have been formed as moraines on the sides of the glacier that once filled the canyon up to an elevation of 1,200 feet above the river as it flows today. The moraine terraces are of older date than the great terrace...and about the entrance of Chelan valley have been partially buried by it." (p. 79)

"In enbayments along the sides of the main canyon and back of the ridges of stone and bowlders left by the ancient glaciers, there are flat areas which have been filled in with fine material, washed from higher levels. These plains have in some instance been cut by small streams flowing across them, thus adding other horizontal lines to the complex topography of the canyon walls." (p. 79)

Glacial erratics on the Waterville Plateau. My photo.


"On the plateau surface near Waterville, there are rounded hills with broad flat-bottomed valleys which mark the courses of former drainage lines..." (p. 85)

"Waterville is a thriving agricultural town without either railroad or steamboat connections." (p. 86)

"Just when our ride was becoming monotonous, our attention was attracted by a new feature in the topography of the prairie.

In front of us rose an irregular undulating ridge...[with]...hundreds of bowlders, some of them larger than the cabins of the settlers to be seen here and there among them. Actual measurements show that many of the blocks are between 50 and 60 feet in their various diameters. Beyond the first line of bowlders resting on the ridge others...could be seen to the east as far as the eye could reach...piled in heaps, but usually they are separated by a few rods of grassy meadow. The greater part of the bowlders and all of the larger ones are black basalt...there are also occasional bowlders of granite, gneiss, etc., of smaller size, but frequently measuring 8 or 10 feet in diameter." (p. 86)

The evidence of ice action was not only recorded by the bowlders, but in the surfaces of outcropping layers of basalt which were planed down and scored by lines running southeast, conforming with the trend of the long lines of bowlders and showing the direction in which the ice invasion moved...The greater part of the glacier continued on southeastward, however, and reached to about the present site of Coulee City, where it ended in Lake Lewis. The eastern border of the glacier after reaching the basaltic plateau of Douglas county was determined by the Grand Coulee, which was cut before the ice invasion." (p. 87)

"The great glacier that descended Okanogan valley was of sufficient thickness to ride over the cliff 2,400 feet high which crossed its path and to carry away huge masses of its hard capping layer and leave them scattered over the plain beyond for a distance of 50 miles." (p. 89)

"Bowlders are scattered broadcast over the plain, but without much regularity or order; there are, however, no conspicuous moraines and only a slight accumulation of what might be called ground moraine. The hollows in the rocks are filled mostly with angular granitic sands." (p. 89)

"Bowlders and all other evidences of an ice invasion are absent to the east of Grand Coulee. The country there assumes

the characteristics of the region about WatervilIe..." (p. 87)


"No more impressive scene can be found in the Big Bend country than is presented by the great cliffs of black basalt below Coulee city [Dry Falls], over which the Columbia once poured, but where now all is hushed and desert shrubs grow in the ancient channel." (p. 92)

"The history of the Grand Coulee and of its southern prolongation is the same as is recorded in Crab creek bottom, Wake Cum Tux and Moses coulees, and Foster creek canyons..." (p. 91)

"The [boulder gravel] deposits which obstruct the drainage below Soap lake appear to have been laid down in the Pleistocene lake, but in this connection my observations are meager." (p. 91)

During a part of [the Pleistocene] at least, when the Columbia flowed through Grand Coulee, it emptied into Lake Lewis, at the present site of Coulee city, but a careful study of levels is required before the details in the history of the river and of the lake can be satisfactorily made out. (p. 92)


"Moses coulee, like the gorge below Coulee city, begins abruptly and has small lakes at its head. The beginning of this deep depression in the basaltic plateau was not visited during the present reconnoissance; all the information we have concerning it is derived from Symons's [1882] report and from conversations with ranchmen. The fact that it begins at about the southern limit reached by the ice invasion from the north, suggests that it may at one time have been flooded by water from the melting ice." (p. 92)

I.C. Russell, who began his career under G.K. Gilbert, would eventually move indoors and away from field work, becoming a professor at the University of Michigan and serving as President of the Geological Society of America. He passed in 1906.


Last 50 Posts
All Posts by Month
    bottom of page