Abstract Sir Archibald Geikie (1835–1924), one of the towering figures in the history of British geology, maintained a long professional relationship and correspondence with pioneering American geologists of the nineteenth century, including James Dwight Dana, Clarence Dutton, Ferdinand Vandeveer Hayden and Grove Karl Gilbert. Geikie made two trips to the USA. Geikie’s first trip, accompanied by his former student at Edinburgh Henry Drummond, took place in August–November 1879 for field excursions in the American ‘Far West’, including Colorado, Utah, Wyoming, Idaho and Montana, where he could find ample evidence for the dominant role of fluvial erosion in the denudation of continents (a school of thought called ‘fluvialism’ or ‘erosionalism’). These geological excursions also resulted in nine articles by Geikie published during 1880–82 as well as a large number of landscape paintings and sketches now preserved at the Haslemere Educational Museum in England. Geikie’s second trip from April to May 1897 was due to an invitation to deliver a series of lectures on ‘The Founders of Geology’ at Johns Hopkins University; these lectures became the basis for his classic book of the same title. Geikie kept an active interest in geological mapping and discoveries in the USA, as evident from his numerous references to American geology in his 1882 Text-book of Geology . Geikie’s American connections demonstrate that geology, although primarily a field-based regional science, did not evolve in isolation in various countries during the nineteenth century, but that there was lively exchange and synergy in geological research and findings among geologists working in Europe, the Americas and Asia.

Grove Karl Gilbert served with the U.S. Geological Survey for 39 yr, from its inception in 1879 until his death in 1918. Thanks to his reputation as an explorer and to his friendship with John Wesley Powell, Gilbert occupied many administrative positions—chief of the Division of the Great Basin, chief of the Appalachian Division, chief geologist (1888–1892), and head of the section on physiographic geology. Gilbert was responsible for establishing the first hydraulic laboratory in the Survey, served on numerous committees to set cartographic nomenclature and style, supervised a series of correlation essays, represented the Survey at some international congresses, advised Powell on practically all matters relating to Survey administration, oversaw bibliographic compilations, edited manuscripts, and, in general, set an example for scientific “investigators” who did not wish to be teachers. Identifying his career with the Survey brought both rewards and costs. Gilbert’s opportunities for research dried up; nearly all studies were done on his own time and at his own expense. His major work of this period was Lake Bonneville , and he tried to follow it with similar works on the Great Lakes and lunar maria. But lacking time and support for full studies, he investigated instead the processes of scientific thought itself in several important methodological essays. The Survey nonetheless gave this nonacademician a responsible job, brought him into a social environment which he enjoyed, and ultimately returned him to the field at age 62 for some of his finest work, the hydraulic mining studies in California.

Grove Karl Gilbert (1843–1918) began his work in professional geology in 1869 as “local assistant” on the newly established Second Geological Survey of Ohio (the “Newberry Survey”). He worked without salary but received $50 per month for expenses. Gilbert investigated Williams, Fulton, and Lucas Counties, the three most northwestern counties adjacent to the Michigan State line. The area extended from Indiana on the west to Toledo at the western end of Lake Erie. His originality, analytical power, and clear verbal and graphic exposition, which were the distinguishing characteristics of his work throughout his career, are completely exhibited in this his earliest work, published in nine articles and reports from 1871 to 1874. Gilbert’s maps of glacial geology of Williams and Fulton Counties are noteworthy for their detail and accuracy. He was the first to discover and map the major end moraines of the Maumee basin, from east to west (youngest to oldest): the Defiance, Fort Wayne, and Wabash Moraines. Gilbert found that the drift sheets of the moraines were actually multiple and that the uppermost drift was draped over a core of older material, a concept not to be widely recognized until almost 50 years later! Gilbert discovered the continuity and mapped the beach ridges now called, from highest to lowest, Maumee, Whittlesey, and Warren. After distinguished work in the West from 1871 to 1881, Gilbert returned to the East and, among other projects, began to study the glacial geology of New York, especially Niagara Falls and the Lake Ontario raised beaches. In his classic study of the origin and retreat of Niagara Falls, he concluded that the Horseshoe Falls had retreated at the rate of 5 ± 1 ft/yr. Gilbert’s other work in the Ontario basin was on the uplifted shorelines of Lake Ontario, the character and amount of their later tilting, and the surface features and drainage channels of the Rochester region. He was impressed by the boulder pavements in the till and by the evidence of glacial and postglacial folding and faulting, on which he published several papers.

G. K. Gilbert’s report on the Henry Mountains is classic for its contribution to knowledge of igneous structures, especially laccoliths, for its contributions to the understanding of geomorphic processes, and as an example of excellent technique in geologic reporting and writing. Present-day Ph.D. candidates and many of their faculty would do well to adopt Gilbert’s technique. His accomplishments are especially impressive when viewed in the perspective of the status of geologic knowledge at the time he did his work and the hazards accompanying his field work. His conclusions, seemingly elementary today, were received with skepticism by many of his contemporaries and were not fully accepted for about a quarter of a century. Modern surveys have confirmed his principal conclusions.

While exploring with the Wheeler Surveys in 1871 and 1872, Grove Karl Gilbert recognized and demonstrated that faulting, rather than folding, dominated mountain building in the basin ranges. Widely accepted now, this concept was challenged by Spurr and others. Gilbert recognized further that relief had been produced incrementally along these range-bounding faults, evidenced by “piedmont scarps,” which he first noted and named. Piedmont scarps, in turn, he recognized as evidence of earthquakes, and in 1883 he warned the citizens of Utah that the absence of such scarps along one segment of the front of the Wasatch Range strongly suggested that a large earthquake might eventually occur there. That warning was the first paper on earthquakes authored by a member of the fledgling U.S. Geological Survey. Gilbert’s studies of the San Francisco earthquake of 1906 stand out to present-day investigators as his principal contribution to the knowledge of earthquakes. His photographs, diagrams, and descriptions of the behavior of the San Andreas fault during that earthquake are data that have been used repeatedly during the 1960s and 1970s. But almost unknown to investigators during those decades was Gilbert’s paper “Earthquake Forecasting,” published in 1909. It was the only paper listed in the Bibliography of North American Geology between 1785 and 1922 about earthquake forecasting or predictions. The issues and concepts in Gilbert’s paper—earthquake prediction, earthquake engineering, land use, risk evaluation, and insurance—anticipated many elements of the Earthquake Hazard Reduction Act of 1977. It was not the quality or originality of a particular work of Gilbert that governed its impact on subsequent studies, but rather the existence or nonexistence of a scientific audience, suitably attuned to the subject at hand and sufficiently knowledgeable to perceive and be influenced by his work.

Gilbert’s reports on Lake Bonneville are, like his Henry Mountains report, classic contributions to geology, but the two studies are very different. The Henry Mountains study involved only 2 months of field work, and the report was completed and published within months after completion of the field work. The Lake Bonneville reports were based on many seasons of field work, first with the Wheeler Survey, then the Powell Survey, and finally the U.S. Geological Survey. Writing and publishing his final report on Lake Bonneville were delayed. He recognized three main stages of the Pleistocene lake: (1) an early stage represented by what is now known as the Alpine Formation, a major interruption in lake history represented by an unconformity between the early and late lake deposits; (2) another rise of the lake to its highest level, the Bonneville shoreline, overflow of the lake into the Snake River via a gap at Red Rock Pass; (3) drop of water level about 300 ft as the outlet was eroded downward to a bedrock lip, and stillstand of the water at that level (Provo stage) to produce what is known as the Provo Formation. The lake history ended with another 300-ft drop to a poorly developed shoreline (Stansbury stage). The subsequent drop in level of about 300 ft to the level of the Great Salt Lake probably should be considered post-Bonneville history, although Gilbert was vague about the terminal stage. Gilbert clearly recognized that the lake basin and its islands and peninsular mountains were formed by Tertiary diastrophism. Pre-Bonneville erosion of the mountains produced huge alluvial cones around the mountain bases and partly filled the basins. Lake Bonneville was formed after the alluvial cones and after most of the faulting and volcanism. Faulting, volcanism, and deposition of alluvial cones were renewed during and since the formation of the lake. Two major contributions to structural geology include recognition of repeated displacements on faults along the Wasatch Front and doming of the lake basin as a result of isostatic rise due to unloading of the crust as the lake desiccated. Both structural contributions have been amply confirmed by modern surveys.

G. K. Gilbert’s major works relating to gravity and isostasy consisted of three separate efforts: (1) a study of the deformation of the Lake Bonneville shorelines, (2) participation in the first gravity profile across North America, and (3) a U.S. Geological Survey Professional Paper on the interpretation of gravity anomalies. With the Lake Bonneville data, Gilbert established that the strength of the Earth’s crust in the Great Basin could not support the load of water in Lake Bonneville and that the basin subsided and rebounded as the water load was first applied and then removed. As a participant in the first major gravity survey in the United States, he experimented with techniques of interpreting gravity observations and speculated on the significance of these sparse data. As one of his last major scientific undertakings, he wrote an essay calling for a more flexible approach to the interpretation of regional gravity variations and outlining his thoughts on the deep structure of the Earth. Although each of these three efforts made a significant scientific contribution, the conclusions he reached in each are not totally consistent with one another, and he attempted neither to reconcile the inconsistencies nor to develop a unified theory of isostasy. Apparently by hypothesizing in each instance, he hoped to stimulate the geodesists to a broader view in their examination of the data. Gilbert’s extraordinary talents are apparent in his work in gravity and isostasy, but he does not appear to have had a great impact on the geodesists who had dominated the development of these disciplines for many years. However, his conclusions that major deformations of the crust reflect “horizontal movements of the upper rocks (lithosphere) without corresponding movements in the nucleus and thereby imply mobility in an intervening layer (asthenosphere)…” and that the driving tectonic forces stem from a “primordial heterogeneity of the earth which gives diversity to the flow of heat energy. . . .” (Gilbert, 1914, p. 34, 35) indicate a remarkable insight into large-scale tectonic processes.

Sketches made by G. K. Gilbert and based on telescopic observations of the Moon look amazingly similar to photographs obtained 75 yr later by spacecraft. He was very successful in correlating lunar surface features with counterparts on Earth. His observations and experiments led him to the conclusion that most lunar craters are the product of impact. After establishing this, he studied the Coon (Meteor) Crater of Arizona. He did not have as much success applying what he had learned from the Moon to the terrestrial case. He conducted a topographic study of the crater to check whether there was an added volume due to the incoming projectile. An overestimation of the size of the meteorite and neglect of the possibility of its fusion, evaporation, and ejection forced him to rule out an impact origin for this crater. In his observations on lunar features, Gilbert had expressed the basic elements of a lunar stratigraphic system. His discussion of crater rays, and particularly of the “sculpture” that surrounds the Imbrium basin, greatly influenced the thinking of lunar geologists of our day. Coupled with his recognition of the importance of crater density and overlap relationships, he can be easily considered the father of lunar stratigraphy. Today there is a crater on the Moon bearing the name of Gilbert in commemoration of his many contributions to geology.

Gilbert’s studies of underground water were overshadowed by his magnificent reports on geologic structure and landforms. In fact, his 1896 publication on the underground water of the Arkansas Valley in eastern Colorado was called by W. M. Davis “for the most part a straightforward geological account of the successive strata.” Gilbert’s 1896 report devoted equal space to stratigraphy and to underground water. His discussion of artesian water was based largely on Chamberlin’s excellent paper on that subject, a decade previously; he briefly acknowledged his debt to Chamberlin for the section dealing with the general occurrence of artesian water, and acknowledged more completely his debt to his U.S. Geological Survey colleague, F. H. Newell. Gilbert’s reports of 1896 and 1897 are largely practical ones on where and how to explore for artesian water, and they include maps showing areas where water can be expected in wells at different depths. Plagued by too little information, he was courageous enough to put lines on a map, although he admitted that the data “are too imperfect to fix the lines definitely except at a few points.” Thus, he said, “I confess that I have drawn this map with much reluctance.” He recommended experimental borings, to gather the additional information needed before putting down a well. Gilbert intended his text and maps to be read by local residents (nongeologists) in their search for artesian water supplies, and this is seen as a major contribution of his work on the underground water of the Arkansas Valley.

Gilbert interpreted rhythmic spacing in limestone units in the Upper Cretaceous of Colorado as a geological response to a planetary cause—the 21,000 yr precession of the equinoxes. In this light, he judged the Upper Cretaceous marine sequence of Colorado to have been deposited over a span of about 21 m.y.—a figure that seems remarkably close to that yielded by modern radiometric geochronology, 24 to 35 m.y. A reexamination of three of Gilbert’s four short rhythmic sequences, using the available radiometric data of Obradovich and Cobban, in conjunction with a model of subsidence and sedimentation, yields bedding rhythms in the 18,000 to 22,000-yr range, and seems to confirm Gilbert’s hypothesis. Most rhythmic Cretaceous sequences in other parts of the world also yield bedding rhythms close to the precessional period, according to the Obradovich-Cobban time scale. The alternative Van Hinte time scale, however, yields a wider scatter of values, and suggests that only some of the rhythms are related to the precession, others seeming to be closer to the 41,000-yr period of obliquity. The equinoctial precession can affect geology only when acting in conjunction with the eccentricity of the Earth’s orbit, which waxes and wanes irregularly with a mean period of 93,000 yr. One should therefore expect precessionally caused rhythms to occur in sets averaging 4.5. That limestone-shale bedding rhythms occur in sets was shown long ago by Schwarzacher for upper Paleozoic and Mesozoic sequences. His observation that the mean number of rhythms per set lies between 5 and 6 suggests that the orbital parameters may have changed. At least two of Gilbert’s four Cretaceous sequences of Colorado are bundled in this fashion. Thus, Gilbert’s suggestion that bedding rhythms provide a basis for geochronology takes on new interest—not to compete with radiometry in the rough calibration of Earth history, but as a refinement. It may also provide a means of tracing the evolution of the Earth’s orbital behavior.

In the latter part of the nineteenth century, G. K. Gilbert began a study of the origin of the lakeshore features of ancestral Lake Bonneville. By means of hypothesis and observation, he used features of the shorelines of the Great Lakes and the Atlantic and Pacific Oceans to form modern-ancient analogs. Studies of present coastal processes and the geometry and internal structure of the Lake Bonneville shorelines lead to the hypothesis that the littoral transport mechanism was dominant in the formation of lagoon-barrier coastal systems. His works on barrier evolution have stood the test of time. Although some of the world’s barrier shorelines have evolved by other processes, most of them appear to fit Gilbert’s hypothesis for barrier evolution. Furthermore, the process of littoral transport appears to be of great importance in modification or alteration of coastal barrier landforms no matter what their origin.

G. K. Gilbert, as a member of the Harriman Alaska Expedition of 1899, studied and described nearly 40 glaciers, many of which reached the sea and produced icebergs. Gilbert’s maps and photographs from marked locations are still being used to record glacier fluctuations, as at Columbia Glacier. Noting that some termini were stable or advancing but that others were retreating rapidly, he suggested that a general change in climate, perhaps related to a change in ocean temperature, might cause such local differences in behavior. This conclusion was remarkably prescient, but it is now known that terminus stability is also involved. Gilbert’s discussions of the processes of glacier flow adjustment to an uneven bed, glacial erosion (including erosion below sea level), and variations in the rate of iceberg calving are remarkably modern and relate to one of the most important problems in glaciology today—the role of a water layer in coupling a glacier to its bed.

The laboratory experiments on sediment transport conducted by G. K. Gilbert differed importantly in technique from such studies of more recent date. Gilbert’s flume was level and could not be altered in slope. Sediment was introduced at the upper end at a predetermined rate and by deposition built a bed gradient sufficient to transport the introduced load. The adjustment of slope in Gilbert’s flume has contributed to the idea widely held by geologists that a river achieves equilibrium by adjusting its slope to provide just the velocity required for the transportation of the supplied load. In fact, slope adjusts but little to a change in amount of introduced sediment load. The adjustment takes place principally among other hydraulic factors: width, depth, velocity, bed forms, channel pattern, and pool-riffle sequence. Gilbert sensed this complicated adjustment process, but its details are as yet only partially known in quantitative terms.

The scientific origins of Gilbert’s geomorphic model are explored, and the distinctive character of that model is compared with that of the rival model of W. M. Davis, with special importance being given to the role of negative feedback by Gilbert. The early development of Gilbert’s ideas (1871, 1875a, 1876) are treated, culminating in a detailed analysis of his classic work on the Henry Mountains (1877). His later work on coastal geomorphology, notably with reference to Lake Bonneville (1881, 1890) and on glacial geomorphology, showed a continuing preoccupation with the relations between form and process. An analysis of geomorphic work since World War II highlights the overriding importance of Gilbert’s recent influence, particularly over the application of systems philosophy to the environmental sciences.

An examination of G. K. Gilbert’s methodological works leaves some doubt as to the role he meant to assign to analogies in the quest for geological knowledge. The drawing of analogies can have only limited utility in a historical discipline, such as geology, in which it is supposed that each phenomenon encountered is assignable, at least in principle, to a kind of phenomenon identified by an extant physical theory.