Arthur Holmes's Principles of Physical Geology


  1. Arthur Holmes
  2. The Principles
  3. Continental Drift and Plate Tectonics
  4. Overthrusting
  5. Earth Genesis
  6. References

Arthur Holmes

Arthur Holmes (1890-1965) is one of the most creative and important geologists of all time, ranking with the pioneers of the science like Hutton, Lyell, Dana and Suess. He calibrated the geologic time scale with the use of radioactivity, clarified the origins of granite and emphasized the important role of rheidity, and set the stage for plate tectonics by suggesting mantle convection and marshalling the evidence for continental drift. Fortunately, he left a superior textbook of Physical Geology, with which this article is concerned.

Dr Cherry Lewis has written a biography of Holmes, and links to two short web biographies are given in the References. Holmes was born in Gateshead, on the Tyne, and was educated on a scholarship at Imperial Collge, where he studied physics and geology, graduating in geology in 1910. In need of funds, he went prospecting in Mozambique in 1911, where he found malaria. On his return to England, he studied for his doctorate, and in 1913 published his first work on radioactive dating, where he used the U/Pb method to date a sample to 1,600 million years before the present. At the time, geologists thought 100 my might be excessive, and Kelvin's recent estimates of 20 my or so were current. He was not the first to suggest the use of radioactivity, but he was the first to carry it out. The discovery of isotopes in 1913 later led to great improvement. He was chief geologist for an oil company in Burma, but received no pay when the company failed.

In 1924 Durham University set up a geology department, and Holmes was appointed professor. Long interested in continental drift, he suggested mantle convection currents as a mechanism in 1928. After his first wife, Maggie, had died in 1938, he married the remarkable Doris Reynolds, an extremely competent geologist, in 1939. Reynolds, educated at Bedford College, London had studied the Newry Complex and Slieve Gullion in Northern Ireland to clarify the origin of granites while a research assistant at Queen's College. This work meshed very well with Holmes's ideas, and is quoted extensively in his book. In October 1943 he was appointed Regius Professor of Geology and Mineralogy in Edinburgh University, where he stayed for the remainder of his career.

This article will not mention Holmes's brilliant work on calibrating the geological time scale in years by radiochronology, since it is very well known and appreciated.

The Principles

The title suggests, probably intentionally, the title of Charles Lyell's influential 1830 Principles of Geology, which was an exposition of Huttonian uniformitarianism, surveying existing geological processes to show that they were fully capable of producing all the effects reflected in the rocks, without postulating unusual or irrational events, and also, but cautiously, the inference that geological time is immense. It was not, at first, a textbook of geology, but became one as a result of later enlargment and revision, up to the final edition of 1875.

Homes firmly supports the views of Hutton and Lyell, but remarks that the term "uniformitarianism" can be fundamentally misunderstood as requiring that conditions in the past be exactly the same as at present. He prefers the term actualism, meaning that the same physical laws governed in the past as they do at present, and that all postulated processes are completely concordant with the actual properties of matter and energy. From beginning to end, this concept guides the book's reasoning. Holmes is quite strict when it comes to conferring the dignity of a "theory" on a collection of suppositions, guesses and presumptions. A "theory" must not only be capable of proof, the proof must actually have been given. To Holmes, a theory is the way something works, not a guess about the way something works.

The Principles (as Holmes's text will be called for brevity) is most emphatically a textbook, with the material carefully and logically chosen and arranged. The first edition of 21 chapters and 509 pages is divided into three parts: a preliminary survey that includes enough of the elements of mineralogy, petrology and palaeontology to make the exposition understandable; a study of external processes, including the effects of water, ice and wind, of weathering, rivers and seas, and the important role played by life; and a study of internal processes, such as earthquakes and earth structure, volcanism, orogeny and finally continental drift. There are many illustrations, both photographs and line drawings, and the most interesting regions of the earth are introduced as examples. There are many references, but no problems or exercises. Any adequate instructor should be able to furnish a wealth of exercises and other activities to provide added interest and to illuminate the material, and demonstrate its general application. This should involve reading of classic references, an activity too often omitted, as well as a little field work.

The contents show what Holmes means by the term Physical Geology. There is a concentration on general principles, but with concrete illustrations. The subject is not divided geographically, but topically. Historical Geology is not really a complement, but a combination of palaeontology with physical geology that concentrates on specific regions. It can hardly be properly treated without bringing in orogeny and plate tectonics. The first edition remains an excellent text in general geology, a wonderful resource for anyone wanting to understand geology who may not have been introduced to it before. With the references, it is excellent for self-study, and is contained in a reasonable compass. Copies can still be obtained on the internet for only a few dollars.

The second edition, which Holmes just lived to see, drops the three-fold division explicity, but the 31 chapters are still arranged in the same logical order. It is pleasant to see the many uncertainties and questions left from the first edition resolved by the twenty years' research that intervened. Holmes is not ashamed to admit being wrong, but this happens to be a very rare occurrence. In most cases, he was uncannily correct and far-seeing. With the extensions, the book now runs to 1250 pages of text, so it is not as compact as the first edition, and takes much longer to digest. However, all the additions are valuable and important. There are now many more illustrations, one of the treasures of the book.

The text remains a celebration of actualism, and it seems to build continually toward the final chapters on earth history. The final chapter is entitled "Continental Drift and Palaeomagnetism." 1965 was the threshold of plate tectonics, and Holmes comes up to the threshold, but does not step over. Nevertheless, it is obvious that he realized his long persistence was finally crowned by success, and that geologists were accepting that the continents moved as a result of mantle convection, and that a "theory" was about to be confirmed by proof. It is very interesting to compare the two editions on this subject. Doris Reynolds published a third edition later, in which plate tectonics was finally present, but we will not consider it here.

I was disappointed to find that one small lapse in the 1st edition had not been corrected in the second. It is in connecton with water waves, when, at the bottom of p. 283, Holmes says, "Owing to friction the diameters of the orbits rapidly diminish with depth until at a depth of the same order as the wave length they become negligible." Of course they do, but not because of friction. They are surface waves, and by nature the amplitude decreases exponentially with distance from the free surface, as exp(-4πd/λ), without any friction. The disappointment is not so much with the author as with the readers who should have brought it to his attention.

Continental Drift and Plate Tectonics

In the 17th century, it was noted that the Americas would fit rather nicely against Europe and Africa if the Atlantic Ocean were absent. This could either be a remarkable coincidence or a logical consequence. Even after Taylor in the U.S. and Wegener in Germany proposed that the continents had actually moved, and that it was logical consequence, not coincidence, based on a large body of evidence on geological and biological correlations across oceans, geologists rejected the proposal, principally because no reasonable mechanism was adduced, and other claims of the proponents were preposterous. However, this episode created renewed interest in continental drift, in Holmes as well as in others, especially those who had pondered the complex structure of the Alps.

Some of the strongest evidence came from Carboniferous glaciation in South America, Africa, India and Australia. If these continental areas were moved and reassembled near the South Pole, then the glacial evidence became consistent as a result of a single continental ice sheet. This assembled continent was called Gondwanaland. The Caledonian orogeny in Scotland and Norway seemed to be continued in Newfoundland, the Maritimes, and across the northern part of the Appalachians, as if the Atlantic had not been there. The northern continents could be assembled into Laurasia, and finally Gondwanaland and Laurasia could be assembled into Pangaea, a term introduced by Wegener. Pangaea itself broke up by the late Mesozoic, when the Atlantic ocean began to be form, and India left Gondwanaland and ploughed into Asia, while Australia pressed into the Pacific. This history was clear to many people long before plate tectonics, but there was no suitable mechanism for it.

American geologists, who did not worry a great deal about the Alps, or India, or Australia, firmly held to the fixity of continent and ocean, each somehow existing in their current configurations back to the origin of the earth. Many others shared these views, but few geologists were as universally convinced. American geologists had long held pretty but dogmatic views of geological processes that have proved to be without physical foundation, and often without utility as well. Many of these concepts were so beautiful that they were almost universally accepted, and little account was taken of criticism. These views did not interfere greatly with geological research, except when continental drift, orogeny or igneous rocks were concerned, and educational efforts to induce students to believe them implicitly were very successful.

The most important of these theories was that of James Hall, New York State Geologist, who in 1857 conceived of a long, oval basin that would fill with sediment, being pushed down by the weight of the sediment so it could accommodate more and more. When this basin was "full" it would somehow crumple and slump into complex folds that, when elevated, would erode into mountains. Or, at least into the Appalachians, with which he was principally concerned. James D. Dana named such basins geosynclines in 1873, and Hall's concept became "the" theory of mountains, with, as Dana remarked, the actual making of the mountains left out. This theory completely disregarded the requirements of isostasy, the nature of sediments and rocks, and the actual structure of most mountains. Geologists would identify these geosynclines, and automatically assume that they controlled the forces of orogeny. This theory was as preposterous as the wild theories of Taylor and Wegener, but appeared in every geology text since then. Even Holmes cannot avoid mentioning the geosyncline, though he seems to regard it as a convenient term for a thick basin of sediments, with no direct association with mountain-building.

It was fortunate for Hall's theory that the eastern flank of the Appalachains was deeply buried under later sediments. The Appalachian geosyncline was full of sediments that came from the east in large quantity, not from the sparse rocks of the continental platform to the west. However, there is no land there any more. In the south, the Ouachita geosyncline was filled with dark shales and cherty limestones unlike anything to the north on the platform: it all had to come from high land to the south. There is no land there any more. These are just two examples of a problem that arose with nearly all geosynclines that had later formed mountains. When you read a study in which this occurs, there is usually little more than a plaintive note or two that the source of the sediments cannot be located. We now know that the sediments in the case of the Applalachians came from Africa, and in the Ouachitas, from South America, both of which departed the scene long ago. The geosynclinal concept is one that should vanish from geology texts, but it is deeply embedded in the souls of geologists.

The deep oceans are an even greater paradox. Fortunately for those that preached about them, the ocean floor was largely unknown until recently. If the continents and the oceans were permanent, then nothing went on in the abyss except the slow accumulation of red mud and other sediments over billions of years, with no geosynclines and, therefore, no mountains, just a featureless plain. The laying of Atlantic cables in the 1850's, and the Challenger expedition of 1872-1876, had shown the existence of great relief, troughs and trenches, and only a thin layer of sediment, but this troubling evidence was hardly mentioned in texts, and certainly did not trouble the supporters of fixed continents. In the 1920's, Vening Meinesz made gravity surveys in submarines along the Banda Arc, demonstrating the existence of large negative anomalies over these trenches. A. L. du Toit published Our Wandering Continents in 1937. He, in particular, made Gondwanaland practically a certainty. There was still, however, no believable mechanism for continental drift.

Harry Hammond Hess (1906-1969), a Professor of Geology at Princeton University, and a submarine commander in World War II, used ultrasonic sounding to map the ocean's floor much more accurately than it had been done before, and his and subsequent work have revealed a wonder: the ocean floor is filled with trenches, faults, rises, mountains and all kinds of irregularity. He identified rises, such as the East Pacific Rise, and the Mid-Atlantic ridge, as spreading centres, where Arthur Holmes's convection currents rose to make new ocean crust. It had long been obvious that the trenches were areas in which the ocean floor was diving beneath continental or oceanic crust and disappearing into the mantle. Putting Holmes's convection currents and Hess's spreading centres together, the idea that the continents were riding on the conveyor belts of rigid upper mantle and crust became clear. This, finally, is plate tectonics, in which continental drift has found a mechanism.

Holmes does not quite recognize this fully, since he remarks that the westward movement of South America since the Mesozoic has caused the Andes to rise as the continent ploughs into the Pacific, but that there was folding long before this time. He would have seen clearly that the Nazca plate had long been subducting beneath the continent, if plate tectonics already existed. The astounding thing, which Holmes clearly recognized, was that the ocean floors were not old, but quite young. In fact, the oldest ocean floor is about Jurassic in age, as has recently been clearly shown. He could not yet furnish an explanation of the fact, but fact it was. The structure of the ocean floor, as now known, is clear and obvious evidence of continental drift. On page 1022 (2nd ed.) is a diagram of a subduction zone precisely like those now known to exist, for the purpose of explaining the origin of andesitic lavas.

The work of John Tuzo Wilson (1908-1993) in determining the age of Pacific seamounts, and showing that they become older the farther they are from the East Pacific Rise, and demonstrating the existence of transform faults originating at spreading centres, gave further evidence. Finally, the magnetic stripe patterns of normal and reversed geomagnetic fields studied by F. J. Vine and D. H. Matthews (1963), which could be correlated at every spreading centre, convinced all those who would ever be convinced of the reality of seafloor spreading and continental drift. The series of events beginning with a hot spot developing into a divergent (constructive) plate boundary, followed by spreading, the creation of a subduction zone or destructive plate boundary, and eventual closing and collision, followed by isostatic uplift and peneplanation, is called the Wilson Cycle in his honour.

Geology is full of "cycles" that are not cycles at all, because they do not "go around." The thing about a cycle, such as the hydrologic cycle, or the carbon cycle, is that, like a wheel, it returns to its initial state and continues again and again. Most geological cycles are just stages from some initial condition to some final condition. A good example is Davis's beautiful but largely meaningless cycle of erosion, through youth, maturity and old age. This provides some descriptive terms, but is devoid of any deeper meaning. As Holmes points out, the earth is seldom stable enough, particularly in the present post-glacial times, for the Davis cycle ever to work out in an orderly manner. The assumption that erosion decreases slopes from youth through maturity to old age is simply wrong in all but a few special cases. Holmes points out that recession of a constant slope is far more likely. I personally noticed in the semi-arid country with which I have been very familiar that nearly vertical cliff slopes were maintained in all stages of erosion from youth to old age, when there were just a few buttes scattered around at the end, a while before I found out that this had actually been noticed. My traditional geology texts gave no hint, just sketches of the hypothetical Davis cycle of erosion. The "orogenic cycle" is so much not a cycle that it should scarcely need mention.

Plate tectonics indeed marks a revolution in geology, a paradigm change as Thomas Kuhn would say. There is a great amount of complexity, but the basic reason for orogeny is now known, if not understood in detail. The Appalachians and the Ouachitas are the crumpling caused by the assembling of Pangaea. The Rockies are the result of light rock carried beneath the continent by subduction, giving the Miocene uplift as a result of isostasy. There are no real geosynclines in the Rockies, despite valiant efforts to find them. The deep basin of Cretaceous rocks in front of the Rockies was not folded. The whole western part of the continent, from Utah west, is a Mélange carried in on plates subducting to the west, an example of collage tectonics, something previously completely unexpected. As the Rockies rose, the Great Basin was stretched as the deep light rock was shifted eastward, so it subsided from its previously lofty position in the Cretaceous (where it provided sediment whose provanance was problematical) across what would later be the Rockies. These are American examples, but Alpine, African, and Asian examples abound. The Himalayas, for example, have a double thickness of sialic continental crust, 70 km rather than the usual 35 km, giving the gravity low and the high altitude, a result of isostatic adjustment.

When looking at a current map of crustal plates, remember that the plates are not permanent things, but can split and join depending on the mantle currents beneath them. The Australian plate, for example, once split moving the portion containing New Zealand, and the shallow epicontinental seas surrounding it, to the east, and then healed again. Microplates have also been identified. These may be joined to other plates, or may disappear by subduction, in the course of time, while new fragments break off of existing plates. Sialic crust carried down by subduction may reappear as andesitic magma, newly activated by the water that is also carried down. The origin of volcanic island arcs is then easily explained, and even the depths of earthquakes in the region. Diamonds, created in the rheid part of the mantle (150 km deep), often show carbon isotope ratios that indicate that the carbon came from biological sources, from carbon-containing sediments carried down by subduction. All volcanic phenomena are the result of active fluids that carry heat and reactivity up from the mantle, producing magma in the crust. Holmes and Reynolds were instrumental in clearing up this connection as well, which we shall now consider.

The book by Bailey Willis (1857-1949) of Stanford University, Structural Geology, listed in the References is an excellent example of geological thinking, especially American, in the 20th century. The book was published in 1929, but its views were current among American geologists and many European ones, up to about 1945. For small-scale structures and field work, it is an excellent text, well worth reading. It touches on more general aspects, however, such as the nature of rock flow, structure of the earth, and even the Planetary and Nebular hypotheses of earth genesis (perhaps the most worthless and meaningless of all geological investigations). Willis was a widely-travelled geologist. He had seen the Andes and the Alps, and had a breadth of knowledge beyond the average of his colleagues. Much more is said about his views on rock flow below.

Willis is interested in crustal mechanics, and gives many of the block diagrams so familiar in geologic texts. These diagrams have forces and movements indicated, but the source of the forces, and what was present before an object moved to occupy its place, are never shown. On p. 81 the famous "keystone" hypothesis of graben (rift valley) formation is shown, where the keystone drops into heaven knows what, and the surrounding rocks are in tension. Somewhat more reasonable diagrams are shown on the next page, where the graben apparently sinks into molten rock (according to Willis, rock can only flow when actually melted). Even the compressional case is shown, which was favored by Holmes in the first edition of the Principles. This is one of the rare cases where Holmes's conclusions were later found to be in error. Compare Chap. XIX in the 1st edition to Chap. XXIX in the 2nd edition. The African Rift Valley is now considered to be a failed constructive plate boundary. The plate boundary went out the Gulf of Aden instead.


Glaciers and salt domes are evidence that what may seem solid may not be. Both ice and rock salt can be fractured with a hammer blow, yet both can create flow structures as if they were viscous fluids. They are not the only such substances; pitch and glass have also been observed to flow. The essential parameter is time. Ice and salt are solid over short time intervals, but liquid over longer ones, perhaps days in the case of ice, and years in the case of salt. This can be expressed as a large coefficient of viscosity, and there is evidence that many such materials are Newtonian fluids, and cannot support any shearing stress for an infinite time. Ice, indeed, is crystalline, and the mechanism of its flow is understood. Such materials were called rheids by S. Warren Carey, and Holmes embraces this term.

Rocks are, in fact, rheids. At the temperature and pressure of the mantle, rocks are sufficiently rheid that convection currents can be maintained. Carey's rigidity is 1000 times the ratio of the viscosity to the shear modulus (in cgs units), and is expressed in time. The rheidity of ice is 12 days, of gypsum one or two years, of salt a few years, and some thousands of years for mantle material (see p. 204 of the 2nd ed.). This information was necessary for Holmes to realize that his hypothesis of mantle convection was possible. Plate tectonics makes it all but certain. This reflects actualism at its best: rheidity is a property of matter, but not one evident in the normal course of events.

To a typical American geologist, anything that exhibits flow structures must have been a fluid at one time, and anything that does not, has always been a solid. The fact that both longitudinal and transverse earthquake waves propagate through the mantle, as they would through a solid, proves to him that the mantle is solid and rigid. However, it only means that the time scale of the shearing motion is insufficient to show rheid behavior. S waves are perfectly consistent with convection currents. Another application of this reasoning is to rocks like gneisses and migmatites, which show flow structure. Reynolds proved conclusively that this did not imply that the rocks were ever liquid, but that the observed structures could be produced by the active fluids causing the metamorphism. These fluids could bring in potassium that had been leached out of the sediments that formed the gneiss, and eventually produce granophyre or granite.

J. Barrell had named the asthenosphere in 1914 that upper layer of the mantle between 30 and 800 miles (48-1230 km) where a lowered thermal conductivity caused higher temperatures and local melting due to the heat rising from lower in the earth. This, of course, is a physical impossibility, since it is always hotter where the heat comes from than where it flows to. Such details did not worry geologists. A Zone of Flow in this region was hypothesized, but Bailey Willis strongly warned that this did not imply anything like, well, actual flow. His words are (loc. cit., p. 449): "[zone of flow] is not a well-chosen name, for it conveys the idea that the rock yields easily, like a liquid, which is not the fact. It is necessary to emphasize the point to avoid the misconceptions that follow from the idea of fluidity or mobility. So long as it has not been melted crystalline rock remains elastic and rigid no matter to what degree overloaded and confined. That is, it retains the characteristics of a solid." Willis recognizes four ways in which rock could flow: mashing, shearing, recrystallizing and melting. He had to countenance some rock flow, as was identified by the great Alpine geologists Albert Heim and Edouard Suess in the 1870's. Nevertheless, this shows the resistance with which Holmes's convection was met, and incidentally the difficulty of arguing with even excellent and well-travelled geologists like Willis who were dogmatic and not too well versed in physics.

There are large granitic bodies created at the cores of what were orogenic belts called batholiths. Bailey Willis had no doubt that they were granitic intrusions of magma fractionated from basalt at greater depths, though correctly noting that they are later than the orogenic belts in which they lie. Remnants of the wall, called roof pendants, or separated xenoliths, were found in the upper parts of the batholith, and the theory was formed that the hot magma quarried and dissolved the country rock, a process called magmatic stoping. Anything that appeared to have flowed, had to have been liquid when it did so. Many basaltic sills and dikes clearly insinuated themselves into joints and bedding planes as liquids, displacing strata accordingly. However, there are apparent intrusions, mostly or entirely acidic, in which the rocks are not pushed apart, as if the intrusion dissolved space for itself. These, clearly, were not the result of the penetration of liquid lava. Batholiths are often associated with mineralization, as in Cornwall or the Colorado Front Range, and many other places. The Sierra Nevada in California has a huge batholith at its core.

It is now clear, in the clear light shone on the problem by Holmes and Reynolds, that the processes are far different from, and much more varied than, simple intrusion of magma. Active fluids from the mantle find ready employment in the deep roots of mountains, and create the granite from the metamorphic rocks they find there, carrying in the necessary absent components. The xenoliths have not been stoped, but are the remnant of country rock that has not yet been granitized. The sialic rocks of the crust can, under suitable circumstances, complete an actual cycle from granite to sediments to metamorphics and back to granite in the natural course of events. The origin of batholiths, and indeed of acidic lava in general, was a great problem. The great hope that basaltic lava and mantle rock in general would somehow fractionate off acidic lava was completely at variance with the actual behavior of matter.

Most intrusions, dikes and sills, and extrusions, lava sheets, have solidified from fluid basaltic lava that is easily made in the upper mantle by the addition of active fluids to what is around between the Moho and the surface. Basalt is called basic, while ultrabasic rocks, typical of the mantle (olivine, peridiotite) have less feldspar. It melts in the vicinity of 1000°C, and is quite fluid. It extrudes in the familar pahoehoe and aa forms, is often vesicular from the expansion of gas bubbles, and thick sheets cool with the characteristic columnar structure. The composition ranges from olivine basalt, of typical mantle type, to other types with increasing amounts of feldspar, up to andesite. Holmes explains the origin of andesite in 2nd ed., p 1022. It does not usually make central volcanoes, but when it does, they are domed shield volcanoes formed by successive quiet flows, as in Hawaii. Andesite volcanoes alternate lava flows with pyroclastics, making layered stratovolcanoes and subsidiary cinder cones.

Acid volcanic activity is essentially different. A lava containing much silica not only melts at a higher temperature, but is always thick and viscous. The volcanic activity is characterized by explosions and the creation of pyroclasts, from fine ash to nuées ardentes and bombs, accompanied by blocks of already solid rock blown from the vent. Cinder cones, heaps of pyroclastics, are built, and then destroyed in a violent explosion. In extreme cases, so much material is ejected in the explosion that a caldera is formed as the relict material sinks into the space evacuated by the ejected material. Sheets of rock that look like rhyolite lava (an oxymoron) are really welded tuffs of fine pyroclasts, that formed mobile clouds of colloidal fragments in a hot, gaseous medium. There is every stage between rock that seems to be rhyolite to coarse tuffs, which can exist in huge amounts. These eruptions occur further from a subduction zone, where the active fluids can rise through silicic rocks, or above a hot spot, as at Yellowstone. The essential thing is the hot, active fluids (mainly water) that cause the event. Reservoirs of acidic lava are probably nonexistent, as are extensive acidic lava flows.

An original hypothesis, arrived at apparently from pure geological thought without refrence to the actual properties of magma, was that large pools of basaltic magma fractionally crystallized out the silica and feldspar to make acidic lava, which could then form batholiths or volcanoes. There was no explanation of how the increased temperature would be produced, or how the thick magma could be encouraged to move in a sprightly manner. N. S. Bowen, in his classical work of the 1920's, put this supposition to rest, but then there was no theory at all of the production of acidic lava or the origin of batholiths. Willis had even proposed, in some desperation, that volume change on metamorphism was responsible for the creation of orogenic pressures.

Holmes emphasizes at every turn the importance of active fluids in volcanic events, and he probably does not overstate the case. They seem to be responsible for granite, batholiths, volcanoes, geysers, hot springs and diatremes. The word "diatreme" comes from dia-trhma, "through" and "perforation." These are the kimberlite and lherzolite pipes in which diamonds are sometimes found. Once the importance of active mantle fluids are realized, they become much less of a mystery. Indeed, diatremes originate in the mantle or at least in mantle rock, and bring up samples from that mysterious world. The similar term "diapir" from dia-peirein, "to pierce through" is reserved for a piercement in which rheid flow is important. A salt dome is the usual example, but mud diapirs, gypsum diapirs and even diapiric structures in massive rock also occur. Holmes treats diapirs at some length, suggesting that batholiths could be diapirs as well (they probably are not, which he knows), and giving examples of structures that are probably diapiric. He emphasizes that there is no change of volume in a diapir. The rise of the penetrating rheid is balanced by the flow of the country rock into the space vacated. These things are very well illustrated in salt domes. It is amusing that they were first found in a country where the geologists would be least likely to appreciate such paradoxical structures. Nevertheless, American geologists did very well indeed in studying and interpreting these structures, because of their great economic importance.

In Willis's 1929 book on structural geology, salt domes are not mentioned, though they had recently been of great importance in American geology, and neither are diatremes, which he had inspected in Africa. Geophysical methods are barely acknowledged, and gravity measurements are not even mentioned as a tool to decipher deep structure, though Vening Meinesz, the "Diving Dutchman," was already reporting remarkable results from submarines. In Principles, 1st ed., Holmes mentions the crustal thickness of 30 miles, as stated in Willis, which was supposed to be the depth at which the shear strength of rock was overcome by pressure, so that no voids could exist. Holmes, however, puts it at the depth where the rocks become homogeneous, rather than exhibiting the crustal variety. In the second edition, the Mohorovicic discontinuity at about 35 km on continents and 12 km in the oceans is taken as the base of the crust. The astonishing thing is that Mohorovicic discovered this discontinuity, the Moho, in 1909! Geologists did not let mere observation get in the way of their dogma. Holmes is notable, and very unusual, in his breadth of understanding and ability to form new ideas. Although Willis praises T. C. Chamberlin's "Method of Multiple Working Hypotheses" (loc. cit., p. 342), he quite inadvertently ignores it, making his reasoning fit the dogma. Standard geological ideas were not far different in 1944 than in 1929, and all American geological texts repeated them dutifully even later after many doubts had arisen, making the Principles even more remarkable.

Holmes had great hopes for the ambitious Mohole project (2nd ed., p. 858) that began drilling in 1961 off the southwestern coast of Mexico. Reaching the Moho at 5 km should be easy for drills that had gone down 7.6 km in the Anadarko Basin in the search for oil. The drill passed through oceanic sediment, and penetrated 44 feet of basalt before it broke down. Russian efforts were also made off the Kuril Islands about the same time, with about the same results. American government mismanagement, funnelling of public money to business friends, and cold war excitement brought an unfortunate end to the project in 1966. Far more money has since been wasted on the more absurd aspects of the "space" program. One scientifically meaningless shuttle launch could pay for a mohole several times over. However, we do get samples from the upper mantle in diatremes, so we are not completely out of contact with the Moho. Ocean drilling was resumed in 1968 with the Glomar Challenger, as a cover for CIA operations to raise a Soviet submarine, and has since provided a great deal of information about ocean sediments, but the more costly Mohole does not seem to have been drilled yet.


"Overthrusting" is a term that implies a mechanism. Much interesting geology is associated with "overthrusts" in which a sheet of older rock lies on younger rock, with a surface of discontinuity, the "thrust plane," separating them, over which the older rocks have glided. The Moine thrust in the Scottish Highlands, where Precambrian schists rest on Cambrian sandstone; the Alps, where huge sheets or nappes have slid on on another in a monstrous tangle; the Himalayas, where (according to the older view) Tibet has been pushed over the Indian Gangetic plains; and even in northwestern Wyoming, in the Overthrust Belt of the Rockies; are but a few examples.

Willis distinguishes competent strata, such as strong limestones, that can transmit a compressive stress (and perhaps even a tensile stress, to judge from some of his diagrams), from incompetent strata, like shale, that crush and crumble when stressed. An overthrust occurs when competent strata are driven forward by a compressive stress, perhaps first to fold into an anticline and a syncline, which overturn and break, after which the top sheet rides over the bottom until the force decides enough is enough. The origin of the force is never indentified, except with the rest of the mysterious orogenic forces that come from nowhere. Such events do indeed occur on a small scale, over perhaps a few miles. They can never occur over large distances, such as in the examples mentioned above, since rocks are not that strong, and the forces cannot be found to cause them. Pushing on a tablecloth only wrinkles it where you are pushing, not on the other side of the table, unless the cloth slides as a whole on the tabletop.

Of course, the motion is a relative one, and the lower sheet could just as well be pushed beneath the upper, making an underthrust, but any such distinction is devoid of meaning unless assumptions are made about what has moved and what has remained still. In plate tectonics, subduction boundaries can be thought of as underthrusts, but the concept really has no meaning, since the competent beds are the whole crust, and the origin of the force is traction on the sole of the crust. This explanation, of course, does not extend to tectonic over- or underthrusts, which must have a different cause.

A force is always available to cause the movement of a sheet, and that is gravity. It now seems probable that all major thrust sheets (and the name is now inappropriate, since they are in no sense "thrust") simply slide downhill. When you tilt the table, the tablecloth will slide and crumple into a heap. Holmes explains carefully how the objection of friction on the thrust plane is overcome by the effects of fluids, even ground water, creating a slippery slide for the nappe. This new replacement for overthrusting is called gravity tectonics. The effects of gravity in causing landslides and avalanches is well appreciated; this only extends it to these much larger cases. In the Wyoming Overthrust Belt, the detached nappes have piled up to the east, today moving from the lower elevations of the Great Basin to the heights of the Rocky Mountain uplift. When they were formed, however, the Great Basin was high, the source of sediments that spread over the plains as far as the Missouri River, while the Rocky Mountains were a low epicontinental sea, not far above sea level. There was plenty of gravitational potential energy available to support the gravity tectonics. Geologists clung tenaciously to rocks that moved apparently on their own volition. Bailey Willis uses phrases like "warping ensued" or "development of powerful compressive forces" or even "Atlantic basin was widened," with no hint of a reason. The pressure to fold the Appalachians "came from the southeast," along with the sediments whose source was invisible. Geologists bore this disappointment with fortitude, steeling themselves against the temptation to ask why. Arthur Holmes always asked why.

Earth Genesis

Principles does not treat the problem of earth genesis, but Bailey Willis devotes many pages to it in Structural Geology. This subject was of considerable interest in the early part of the 20th century, when Laplace's Nebular Theory was shown to be impossible. Willis's treatment includes the internal constitution of the earth, as well as his views on rock flow; none of this is easy to understand, and its bearing on structural geology is not clear. Willis always seems worried about forces, but their sources are never identified. He states explicitly that the continents are probably in the same positions as at the origin of the earth, never even alluding to any suggestion that they are not. The subject of earth genesis is a peculiar one, since it refers to a single, irreproducible event. Similar events could occur in other planetary systems, but this is beyond our powers of observation. The earth is also strange in having a moon that is large in comparison with its primary. There are some current speculations, dealing more specifically with lunar genesis than earth genesis, that are much like the earlier, and equally hard to maintain with a straight face.

The new theories of earth genesis of the early 20th century are now realized to be inconsistent with stellar structure and dynamics. They are mainly variations on the Planitesimal Theory of Moulton and Chamberlain, which in its fundamentals is similar to modern views. In this theory, the planets are formed from material drawn out of the sun that agglomerates as cold particles that grow to form the planets. It is very difficult to imagine any mechanism that leads to the agglomeration of cold particles. There is always the difficulty of disposing of the relative kinetic energy so that the particles cohere. Swarms of particles, as in the rings of saturn, tend to remain stable swarms of particles. The solar system seems to have been swept reasonably clear of debris, as well.

Before the nature of solar prominences and sunspots was known (especially to geologists) it was thought that the same forces driving prominences outward could disperse planetary raw material. Willis saw through this supposition, though the extreme difficulty of raising matter in a gravitational field had probably not been fully appreciated by him. The favored mechanism became a near-miss with a large, massive star that would pull out the necessary matter and provide the required angular momentum. There was, naturally, no evidence of such a star. Moulton and Chamberlain supposed the sun to be much as it is now at the origin of the solar system, since the amount of mass necessary to form the plants would be a negligible fraction of the mass of the sun. The Planitesimal Theory would result in an earth that was initially cold, and idea that was not disagreeable to Willis.

Jeans and Jeffreys modified the theory by assuming the sun to be a giant that was tidally disrupted by the collision. They supposed that this meant that the earth could form from molten matter, not cold planitesimals, a more traditionally comfortable idea. Jeffreys presumed that solidified crust would be heavier than the liquid rock beneath it, and so would sink, and the earth would solidify from the centre up in kind of honeycombed structure. It is hardly needless to say that none of this is consistent with observation, since the density of the earth increases with depth. The principles of stellar structure are also violated.

Jeffreys was correct in deducing that radioactive heating was only important in the earth's crust, in the first 13 to 16 km. Holmes had brought this out in a 1915 paper in Geological Magazine where he supposed that radioactive elements had risen through the liquid earth to concentrate at the surface. 80% of the heat that flows due to the surface gradient of about 1°C for every 33 m is due to crustal radioactivity, principally in continental sial. Only the remaining 20% is furnished from the mantle. The surface gradient implies that the temperature reaches 1050°C, the melting point of basalt, at a depth of about 35 km, which seems very reasonable. The temperature deeper in the earth is not known accurately; it must increase with depth to support thermal convection, but the source of energy is unknown. The theories of earth genesis have so far not shone any light on the problem. Perhaps the earth is living off of the stored heat in the core, which we would be unhappy to conclude.

Willis suggests that geomagnetism is evidence that the earth's core is iron, because, as everyone knows, iron is magnetic. However, the temperature is well above the Curie point under normal conditions, and the pressure, as well as apparent liquidity, would not seem to favor ferromagnetism. Anyway, everyone believes the geomagnetic field is produced by conduction currents. Here is a subject in which surprises are possible, but that the earth is a permanent magnet is probably not supportable, if only because of the secular variations.

A modern speculation on the origin of the moon by Robin M. Canup of the Southwestern Research Institute postulates a "glancing collision with a smaller planetary body" close to the time of the formation of the earth. This body is destroyed in the collision, and its fragments form a ring of or disc which in "10 months" coalesces to form the moon. Of course, it is all based on computer simulation, which you can't argue with, and has been rather widely acclaimed (Dr Canup has just been awarded the Urey Prize). Why don't saturn's rings coalesce? Some people claim geological consequences for this encounter, but I do not see what they could be. I wonder why the fragments did not continue to move along the body's orbit, and how they were captured. I wonder how the moon's angular momentum about the earth came to be, though the glancing collision seems designed for this purpose, as well as for causing the earth to rotate, another claim. Making the earth rotate requires very much more angular momentum than a glancing collision could supply. There seem to be many problems with this bold assertion, and there were no witnesses of the event. Remarkable claims demand remarkable proof, which seems absent here. The low eccentricity of the earth's orbit is one argument against collisions with considerable bodies. There is also no means of capturing such a large satellite as the moon if it arrived from outside. In all such speculations, one is free to choose the initial conditions and write your programs so that the desired result is obtained.


Arthur Holmes, Principles of Physical Geology (Edinburgh: Thomas Nelson and Sons, 1944 and New York: Ronald Press, 1945).

Arthur Holmes, Principles of Physical Geology, 2nd ed. (New York: Ronald Press, 1965).

Cherry Lewis, The Dating Game (Cambridge: Cambridge University Press, 2000).

Brief Biography of Arthur Holmes.

Brief Biography. of Arthur Holmes.

Bailey Willis and Robin Willis, Geologic Structures, 2nd ed. (New York: McGraw-Hill, 1929). This is a very good text on structural geology. The rather rambling account of earth genesis and internal constitution is confined to the final section of the book.

J. Gilluly, A. C. Waters and A. O. Woodford, Principles of Geology, 3rd ed. (San Francisco: W. H. Freeman, 1968). This is one of the best American classic geology texts, though crammed with misprints and minor errors. Holmes is mentioned only in the footnotes to the geologic time scale, and plate tectonics is not mentioned at all, though the magnetic stripe anomalies are illustrated. Continental drift is more or less ridiculed. There is no advance in earth theory on Bailey Willis's ideas of 40 years earlier, despite the impending victory of plate tectonics. In 1968 all the fundamental ideas were already known and widely accepted.

For earth science books, I recommend Mt. Eden Books, P.O. Box 1014, Cedar Ridge, CA 95924.

A web site with a good explanation of plate tectonics and the Wilson Cycle is Fichter's Course. Many such sites that you will find are "forbidden," but do not be put off. Anyone who will not share information on the web is not worth looking at anyway (this is much truer than it might appear). Professor Fichter deserves our gratitude for sharing with us.

My information on Robin Canup's hypothesis came from BBC news via the web, and has apparently been the subject of a TV program. Full information is available at R. M. Canup.

Return to Geology Index

Composed by J. B. Calvert
Created 17 February 2003
Last revised 22 February 2003