RECORD: Pearson, Paul N. 1996. Charles Darwin on the origin and diversity of igneous rocks. Earth Sciences History 15, no. 1, pp. 49-67.

REVISION HISTORY: Transcribed (single key) by AEL Data 9.2012, corrections by John van Wyhe. RN2

NOTE: See the record for this item in the Freeman Bibliographical Database by entering its Identifier here. Reproduced with permission of Earth Sciences History.

[page] 49



Department of Geology, University of Bristol, Wills Memorial Building, Queens Road, Bristol BS8 1RJ, United Kingdom


Charles Darwin provided one of the first detailed explanations for the diversity of igneous rocks. Building on many observations made during the Beagle voyage, Darwin hypothesized that density differences among crystals within a mass of partially molten rock would result in their physical separation by sinking and floating. Such a process, he proposed, could be responsible for the separation of compositionally distinct lavas from a single source. This idea, in modified form, lies at the heart of the modern science of igneous petrology. Darwin also speculated that partial melting of rocks in the deeper regions of the Earth's crust could produce basaltic melts. However, due to his lack of knowledge of the melting points of the silicate minerals, and his misinterpretation of a puzzling field locality at Bahia in Brazil, he wrongly believed granitic gneiss to be the progenitor of these basalts. Despite this error, Darwin's igneous speculations show a characteristic blend of detailed observation and broader theorizing. Most interesting of all, striking analogies can be found between Darwin's igneous work and his theory of natural selection, which he developed at about the same time.

"One of the most obvious facts about igneous rocks is that they are extremely variable both in mineralogy and chemical composition. This leads petrologists to think automatically in evolutionary terms, like zoologists and botanists. Yet it is not immediately obvious that inanimate materials have the capacity to evolve, until one contemplates the variety of igneous rocks and asks how the individual types may have come to be created. One is obliged to postulate either that they were all created different, or that some processes exist which have the capacity to generate variety."

(K.G. Cox, J.D. Bell and R.J. Pankhurst,
The Interpretation of Igneous Rocks, 1979, p. 2.)


During the voyage of HMS Beagle (1831-1836), Charles Darwin made numerous geological observations on volcanic and plutonic areas around the world. He also collected thousands of rock specimens, many of them igneous, which were studied after his return to England with the help of the mineralogist W.H. Miller (1801-1880). On first reading, much of Darwin's published igneous work appears to be little more than a series of lengthy descriptions of various localities and rock types, more or less in the order they were visited. However, as is characteristic of his later speculations in many other fields, the meticulously collected and collated facts provide the foundation for some bold theorizing.

Two theoretical questions dominated Darwin's thinking on igneous geology during the voyage, as eventually spelled out in the remarkable summary chapter (Chapter 6) of his book on Volcanic Islands.1 The first and most complex theme is the relationship between volcanism and the uplift of mountain chains. This will not be dealt with here because it is linked to a variety of other apects of Darwin's geology and would require a dedicated contribution to fully discuss it. In the second theme, Darwin sought to understand the processes by which the various types of igneous rocks could have been formed. This provides the subject of the present essay.

As is typical of his approach to other scientific questions, Darwin was not content merely to subdivide, distinguish and classify the igneous rocks, but was keen to investigate their similarities and consider the relationships among them. He examined the transitions between different rock types as observed in nature and speculated on processes that might occur when a rock was in a molten or semi-molten state. Building on earlier suggestions, most notably by Leopold Von Buch (1774-1853), George Poulett Scrope (1797-1876) and Charles Daubeny (1795-1867) (Figure 1), Darwin suggested that the silica-rich and silica-poor lavas could be derived from a single source liquid, and that the settling out of crystals from the melt was the mechanism by which this separation was effected. This concept of crystal differentiation is central to modern igneous petrology. Although this achievement has tended to be neglected by historians of science and Darwin's many biographers, it was discussed by the igneous petrologist J.P. Iddings in 18922 and has been briefly mentioned in several textbooks on petrology since then.3-7

In this paper, Darwin's igneous theories are investigated in relation to the state of knowledge in the first half of the nineteenth century. As is explained below, his ideas on igneous separation formed part of a wider scheme for chemical cycling in the Earth's crust in the tradition of James Hutton and Charles Lyell, some aspects of which seem remarkably prescient in the light of modern knowledge. It should be stressed, however,

Earth Sciences History, v. 15, no. 1, 1996, p. 49-67

[page] 50

Figure 1. Portraits of igneous theorists from the first half of the nineteenth century. Top left: Charles Robert Darwin (1809-1882), Top right: Christian Leopold Von Buch (1774-1853), Bottom left: George Poulett Scrope (né Thomson) (1797-1876), Bottom right: Charles Giles Brindle Daubeny (1795-1867). The portrait of Daubeny is reproduced by kind permission of the President and Fellows of Magdalen College, Oxford (ref: MCA F. XIII, fo. 43).

[page] 51

that Darwin was by no means entirely correct in his observations or deductions. Furthernore, his theories were largely ignored by most his contemporaries. Because modern igneous petrology developed from later theoretical and empirical studies, as collated in Bowen's classic textbook8 (in which Darwin is not even mentioned), it is doubtful whether his ideas influenced the subsequent development of the science. However, they do provide a fascinating example of the type of interests and the mode of reasoning that led him, almost concurrently, to develop the theory of evolution by natural selection, a theme which is discussed futher in the final section of this paper.


Darwin's igneous theories were outlined in his Volcanic Islands volume of 1844 and provide the focus of what is otherwise an obscure book. They were never expanded on at any length, probably because Dawin appreciated that they were speculative and at the time only poorly supported by experimental and field data. Unpublished comments and notes provide a much fuller picture of the role of igneous petrology in Darwin's theorizing and are used extensively in this contribution to supplement his published account.

Many of Darwin's letters, journals and notebooks have been published and provide extremely useful sources to augment his actual publications.9-17 His field notebooks and extensive geological diary from the Beagle voyage have not been transcribed, and so were viewed at the Darwin Museum in Downe, Kent, and the University Library, Cambridge, respectively. The latter manuscript in particular is extremely valuable in being an extensive record of Darwin's observations and developing theories as the voyage progressed. In addition, various interesting marginal annotations can be found in Darwin's own books, some of which are at Cambridge and some at Down House. It has also been possible to study Darwin's rock collection from the Beagle voyage18 in the Department of Earth Sciences, Cambridge (an informative museum display of some of these rocks has been arranged there by Dr Graham Chinner of that department). This collection has been particularly useful in elucidating details of the terminology used by Darwin. Many thin sections were made of the Darwin rock collection around the turn of the century, and notes were made on some of them by the petrologist Alfred Harker (1859-1934)19-20 Harker's studies do not appear to have been historically motivated but rather he utilized the Beagle rock collection for continuing petrological research. These thin sections are also available for study in the Sedgwick Museum, Cambridge. They were not, of course, available to Darwin, but have yielded important additional insights on the true relationships among the rocks he collected.


Geology itself began to emerge as a distinct discipline following the great advances in chemistry in the eighteenth century. An early contentious issue was the nature of basalt.21-22 As has often been recounted, some early authorities regarded all rocks as sediments from the chemical soup of a universal ocean that once enveloped the earth, but others adopted the view that basalt and some other crystalline rocks must have been formed by cooling from a molten state.23 In the case of tabular bodies such as dykes and sills, it was debated whether they were fissures infiltrated by sediment from above, or injected from below by molten rock. Darwin himself was taught both these views as a student at Edinburgh in the late 1820s, by Robert Jameson (1774-1854) and Thomas Charles Hope respectively.24 However, Darwin was unimpressed by Jameson's by then outmoded views and the igneous (i.e. molten) origin of the volcanic rocks was never questioned by him.

A livelier debate persisted regarding the nature of granite and its significance in the Earth's crust. The "plutonists" of the previous century, most notably James Hutton (1726-1797), believed it to be igneous and cited several instances of granite veins penetrating the surrounding rock at localities such as Glen Tilt and North Glen Sannox, Arran.25-28 Hutton's adversary, Richard Kirwan (1733-1812) objected that the crystals in granite appear to impress one another equally, and thus could not have crystallized sequentially.29-30 Hutton's theories were slow to be accepted, and many early nineteenth century authorities such as Charles Daubeny reiterated Kirwan's objections31-32 To Daubeny, granite was the "skeleton of the planet" which had formed in a primeval era under conditions very different from those of the present. Granite was usually depicted as a "primary" rock in cross sections of the crust, often implying that it is the oldest or most fundamental rock type and the ultimate geological basement. A splendid example of this view is William Buckland's idealized section of the Earth from his "Bridgewater Treatise,"33 part of which is reproduced in Figure 2A. Even the Huttonian John MacCulloch (1773-1835) went so far as to regard basalt as "the granite of the newer strata,"344 while other writers, such as Henry De la Beche (1796-1855), although convinced of the igneous and intrusive nature of granite, supposed that there had been "a greater prevalence of granitic rocks over the trappean [i.e. basaltic] at the earliest periods."35 Thus, the continuing debate on granite involved shades of opinion, and was intimately related to the contentious question of whether Earth history was progressive or "steady-state." Darwin enthusiastically followed the lead of Charles Lyell (1797-1875)36 (Figure 2B) in opposing the progressionist ideas which were prevalent at the time, and believed that granite was igneous and capable of being formed at depth in the present day at essentially the same rate as in the geological past. It is within this Lyellian context that Darwin's igneous ideas are best understood.

The relationship between granite and gneiss was often confused throughout the nineteenth century. This

[page] 52

resulted from uncertainty about the causes of schistosity and gneissic foliation as is evident, for example, in the writing of Henry Boase.37 During his voyage, however, Darwin developed a reasonably advanced understanding of the difference between granite and gneiss based on his own observations. He is credited with producing one of the first clear accounts of the relationship between cleavage and schistosity arising mainly from his observations in South America.38 In his published works he upheld a fairly clear distinction between granite and gneiss, unlike many of his contemporaries. Once again, he became closely allied to Lyell on this issue, as their correspondence shows.39 Nevertheless, in accordance with the terminology of the day, Darwin used the names "granitic" and "primary" interchangeably to refer to tracts of land that were either plutonic or metamorphic, or both, and did not use Lyell's preferred term "hypogene."

Another contentious topic among geologists in the early nineteenth century was the study of volcanic landforms and products. Questions of particular interest were the nature of volcanic heat, the structure of volcanic cones, the degree of fluidity of lava in eruption, the relationship between vulcanicity and earthquakes, and the tracing out of geographical lines of volcanic activity. With respect to volcanic cones, some workers such as Leopold von Buch and Alexander von Humboldt (1769-1859) distinguished two types of volcano, which they termed "craters of elevation" and "craters of eruption."40 They believed that some volcanoes (such as at the peak at Tenerife) had not been formed by successive outpourings of lava, but rather had been upthrust by vertical pressure forming an anticlinal dome. Scrope41-42 and Lyell43 were strongly opposed to the idea of elevation-craters44 which soon became linked to Élie de Beaumont's (1798-1874) theories of catastrophic mountain formation. 45 After the Beagle voyage, Darwin was better disposed toward the theory and adopted it in modified form to explain the geological structures of Mauritius and St Jago. However, his correspondence shows that he later regretted making a contribution to the discussion, probably because he did not want to be associated with the catastrophist school.

The sciences of petrology and mineralogy were already closely allied to chemistry in the early nineteenth century. 46 The minerals in igneous rocks were identified using the handlens, goniometer (to measure the characteristic cleavage angles) and magnet. A great proliferation of mineral names were in use and the major groups of silicate minerals, such as the pyroxenes and feldspars, were clearly recognized as distinct families sharing common properties, although the concept of solid solution was not yet developed. Authors such as Charles Daubeny with a strong background in chemistry appreciated the nature of atomic co-ordination between the different components of a melt, and that (for example) quartz only formed when silica was in excess of that required to combine with other components in feldspar. A classification of the volcanic rocks themselves began to emerge in the mid- to late eighteenth century. By the early nineteenth century a relatively stable nomenclature had developed, such as would be broadly recognizable to a modern geologist, as described by Scrope.47 It should be appreciated, however, that rock names were applied with much less sophistication than they are by today's specialists and the external appearance of a rock was often of the greatest importance in assigning a name. Names commonly applied to volcanic rocks in Darwin's time include hornstone, pearlstone, pitchstone, greenstone, graystone, claystone, clinkstone (or phonolite), obsidian, porphyry, basalt, trachyte and domite. The term "rhyolite" was not introduced until later in the nineteenth century, for quartz-bearing trachytes.

Despite this superficial variety, most geologists agreed that all the lavas belonged to two principal families or "series." In general, the darkest lavas were allied to basalt, and the lighter lavas were allied to trachyte. Lyell gives simple guidelines for the use of these terms; "When [feldspar] is in great excess, lavas are called trachytic; when augite (or pyroxene) predominates, they are termed basaltic."48 Darwin himself defined trachyte as a "harsh, generally rather pale-coloured lava, with crystals of glassy feldspar." 49 Because it is primarily the silica content that controls the frequency of the various minerals, the division of volcanic rocks into basaltic and trachytic tended to reflect this important factor. Thus, both obsidian and pitch-stone, which were known to be exceptionally rich in silica, although dark in colour, were grouped with trachyte.

A straightforward classification of igneous rocks (as opposed to a nomenclature) along modern lines was not developed until the later nineteenth century. Competing schemes in vogue in the early nineteenth century seem somewhat arbitrary to modern eyes. Most modern classifications focus on the content of silica in

Figure 2. Comparison of idealized geological cross-sections of the Earth's crust according to William Buckland's Geology and Mineralogy Considered with Reference to Natural Theology (1836; Top) and Charles Lyell's Elements of Geology (1838; Bottom), illustrating differing interpretations of Earth structure and history in the first half of the nineteenth century. Buckland's cross-section, of which only a small part is illustrated here, was closer to orthodox opinion at the time Darwin sailed on the Beagle. It shows granite as the "primary" rock underpinning all others and, by implication, pre-dating the sedimentary rocks. Elsewhere on the section, intrusive granite veins are shown cutting the "primary" granite and later rocks but they are on a comparatively insignificant scale. Volcanoes are fed by dykes which cut the granite, but their ultimate origin deep in the Earth is unclear. Most dykes feed modern or post-Tertiary volcanoes, implying that basaltic volcanism is more prevalent at the present than in the geological past. Note also that in the centre of the figure, trachyte is shown underlying an extinct (basaltic) volcano. Lyell's cross—section was an intentional rebuttal of Buckland's (who was his old teacher at Oxford), and illustrates his "theory of the contemporaneous origin of the four great classes of rocks" (aqueous, volcanic, metamorphic and plutonic). In Lyell's section, the volcanic rocks are shown as derived from the plutonic rocks. In the original hand-colored figure, the relationship between the plutonic granite and the higher-level intrusive dykes and volcanic rocks is fudged as a gradual transition from red to purple. It was this part of Lyell's scheme that Darwin improved upon.

[page] 53

[page] 54

a rock, with the amount of alkali metals and the grain size also of importance. This scheme is used because it makes the most sense in terms of the genetic relationships among different rock types, which in turn reflects the processes which give rise to that diversity. Before Darwin's time the processes were only poorly known and hence the classifications available were rudimentary, and often confusing. Fortunately, as pointed out by Alfred Harker,50 Darwin's preference for a relatively simple terminology makes his own intended meanings relatively easy to decipher.


Darwin had been fascinated by chemistry since his early childhood.,51 As he recalled in his autobiography, he had been nicknamed "Gas" as a boy because of his chemistry experiments and he collected minerals. A greater rigor was imparted in his education at Edinburgh and Cambridge which well qualified him for examining the geology of the Beagle voyage.52

It is clear from Darwin's correspondence that in these early days he regarded geology as his principal science. After his return to England, his articles and books quickly propelled him to a place among the geological élite53 and he served a term as a Secretary of the Geological Society of London. Notable articles on Darwin's geology have been written by Archibald Geikie,54 John Wesley Judd55-57 and several more recent authors.58-64 Although today Darwin is most remembered for his "soft rock" contributions to geology, especially his palaeontological discoveries and theory on coral reefs, his notes reveal that "hard rock" issues were at least as important to him.

As has been often recounted, Darwin's interest was fuelled in particular by his reading of the first volume of Lyell's Principles of Geology. The initial landfall made by the Beagle was at St Jago (San Tiago) in the Cape Verde Islands65 where, as he recalled in his autobiography, he became convinced by the "wonderful superiority of Lyell's manner of treating geology,"66 a statement which is at least partially supported by his notes. Darwin immediately set about examining the lava flows and volcanic cones of the island and was very impressed by the evidence for gradual and almost uniform uplift in the manner described by Lyell. He also thought he had found evidence for a great marine incursion and "a part of the long disputed Diluvium"67 although it is not clear to what extent he might have equated this with the Biblical deluge. This claim was crossed out later in the voyage with the comment: "I have drawn my pen through parts which appear absurd." Nevertheless, at St Jago Darwin's original observations so excited him that he resolved to write a book on the geology of the places visited during the voyage. As it turned out, three books and a series of technical papers were eventually written on the Beagle geology alone.

Any geologist, when visiting a volcanically active area, cannot help but be impressed at the power of erosion in rapidly reshaping the landscape. This is particularly the case for marine erosion around volcanic islands. As Darwin marvelled in one of his later notebooks: "If man could raise such a bulwark to the ocean, who would ever suppose that its age was limited? Who could suppose such trifling means could efface and obliterate so grand a work?."68 And yet, he was immediately able to satisfy himself that many eroded volcanoes, although no longer active, were of extremely recent geological origin. For example, at St Jago he found a limestone bed containing the shells of living species stratigraphically beneath an old eroded volcanic vent, which threw the enormity of geological time into perspective: "To what a remote age does this [limestone bed] in all probability call us back & yet we find the shells themselves and their habits the same as exist in the present sea."69 Darwin also had a keen appreciation of natural beauty, as many passages in his famous Journal of Researches clearly attest. The power of the new geological science, when combined with the ambition of a young person out to make his mark as an author, and, no doubt, the excitement of the voyage and the beauty and grandeur of the landscape, is the combination of circumstances that accounts for the great affinity and enthusiasm Darwin immediately professed for geology.

From St Jago onwards, Darwin continued to be impressed by Lyell's approach which stressed both cyclicity and the cumulative effect of gradual change in explaining geological phenomena. In the tradition of Hutton and Lyell, Darwin came to adopt a mechanistic view of the Earth in which igneous and tectonic cycles were paramount. He was less involved in the stratigraphic research for which early nineteenth century geology is now mostly remembered. But by circumnavigating the globe, Darwin achieved a rare global view and a keen appreciation of the significance of scale in space as well as time. In addition to experiencing earthquakes, he witnessed a spectacular nighttime volcanic eruption of Mount Orsono in South America.70 Throughout the voyage, he was always particularly interested in volcanic and tectonic phenomena as his extensive manuscript notes attest (Table 1). In particular, it is interesting to note that during his five week stay in the Galápagos archipelago, Darwin wrote a total of 109 manuscript pages on volcanic observations (in addition to field notes), and only 37 on zoology. He shipped back a large number of rock specimens from the islands. These observations provide a new twist to Sulloway's71-72 debunking of the legend that Darwin was immediately converted to evolution on encountering the Galápagos fauna—he was clearly preoccupied with volcanic geology at the time.

In describing volcanic landscapes, he was keen to trace a continuity from newly-formed and pristine lava cones through to eroded sheets and plugs (Figure 3). His geological writings are full of such instances of inter-gradation between classes of objects. If anything, this was the essence of his scientific method—finding intermediates was the first step toward recognizing a

[page] 55

Table 1. Key to the igneous and related observations in Darwin's "Diary of observations on the geology of the places visited during the voyage" (DAR 32-33) and "Notes on the geology of the places visited during the journey" (DAR 34-38) which are deposited in the Manuscripts department of Cambridge University Library. Although the geological journal forms a single series it was divided on archiving into the above two categories. The former are numbered from pp. 1-274, the latter from pp. 1-960. The notes have been arranged in approximately chronological order but several parts are out of sequence, including the first few pages. Some miscellaneous material has also been included. Most of the geological journal was written up on the ship after a particular place had been visited. On the reverse side of most pages are notes which are reference to the main text. Darwin found igneous rocks in many of the places he visited. The following is intended to indicate some of the variety he encountered.

Reference Place and Date Topics Discussed
DAR 32.1
3-5 Bahia, 1836 Gneiss and associated dykes of Hornblende rock
15-20 Quail Island, Jan. 1832 General description, including volcanic rocks
21-36 St Jago, Jan. 1832 Volcanic rocks, etc.
37-38 St Paul's Rock, Feb. 1832 Serpentine rocks
39-40 Fernando Noronha, 1832 Topography, volcanic rocks, etc.
41-48 Bahia, Feb.-March 1832 Gneiss and associated dykes
49-50 Abrohlos Islands General geology, including trap rocks
54-60 Rio, June 1832 Gneiss and greenstone dykes
DAR 32.2
83-84 Monte Video, Aug. 1832 Appendix on granitic rocks of district
85-122 Tierra del Fuego, Feb. 1833 General descriptions, including trap dykes
DAR 33
153-164 Maldonado, June 1833 General geology, including granite and trap
229-244 Port Desire, Jan. 1834 General geology, including porphyry and pitchstones
DAR 34
135-154 Santa Cruz, May 1834 South American geology, including lava flows
153-156 Port Famine, June 1834 Geology, including greenstone dykes
DAR 35.1
206-217 Chiloe, July 1834 Geology, including basalts and volcanic sediments
218-226 Valparaiso, July 1834 Geology, including trap and porphyry dykes
233-258 Chonos, 1834 Geology, including various dykes
288-299 Chiloe, Nov. 1834 General description, including volcanic rocks and trachyte
DAR 35.2
371-418 Valparaiso, 1834 Gneiss, porphyry, greenstone dykes, etc.
DAR 36.2
438-451 Chili Earthquakes and vulcanicity
466-501 Mendoza, April 1835 Geology of Portillo Pass, including syenite, greenstone, etc.
502-549 Mendoza, April 1835 Geology of Uspallata Pass, including dykes, granite, etc.
550-610 Journeys, May 1835 Mineral veins, granite, greenstone, etc.
DAR 37.1
611-648 Copiapo, 1835 General geology, including granite and dykes
688-703 Lima, July 1835 General description, including igneous injections
DAR 37.2
716-723 Galapagos Petrological notes
746-795 Galapagos, Oct. 1835 General description, especially lava flows and craters
798-801 Tahiti, Nov. 1835 General geology, including lavas
802-811 New Zealand, 1835 Geology, including trappean rocks and lava flows
DAR 38.1
812-836 New S. Wales, Jan. 1830 Geology, including granite and trap
837-857 Hobart, Feb. 1836 Geology, including basalt, greenstone and granite
864-881 K. George's Mar. 1836 General geology, including granite and basalt
882-891 Mauritius, May 1836 Basalts and trachytes, topography, craters
DAR 38.2
902-919 Cape of Good Hope, 1836 Granite, metamorphism, and trappean dykes
920-935 St Helena Volcanic geology, craters, etc.
936-953 Ascension Mineralogy of volcanic rocks
954-956 Bahia, 1836 Rough notes on gneiss and dykes
957-960 Terceira, Azores Volcanic geology and landforms

[page] 56

Figure 3. Darwin's illustration of Kicker Rock in Chatham Island, Galápagos (redrawn), which he recorded as being 400 ft high and interpreted as an eroded remnant of a volcanic neck. Similar structures, but in an even more eroded state, are common in central Scotland and had been visited by Darwin in his student days in Edinburgh. Darwin and Lyell were keen to demonstrate that there was no great difference between modern and ancient volcanoes and their products.

common origin for disparate but related phenomena. Although his elegant theory of coral reefs is the best-known example of this approach, his writings on the crystalline rocks also contain many instances. He traced every stage in the decomposition of fresh minerals to their weathered products (something particularly necessary for fieldwork in tropical climates), and from obviously sedimentary rocks, including limestones and mudstones, to their counterparts that had been thoroughly metamorphosed by heat. The most significant question that this method led him to confront, was that of the relationships between the different families of volcanic rocks.


Scrope's Considerations of Volcanos and Daubeny's Active and Extinct Volcanos were both present in the library aboard the Beagle73 and, with Lyell's Principles of Geology, would have been Darwin's main sources on the subject. As discussed above, in these and other works of the time most lavas were regarded as belonging to two great families, the "trachytic" and "basaltic" series. Lavas that are intermediate between the two extremes had occasionally been described by various geologists, and indeed the name "graystone" had been suggested for them by Scrope.74 The great variety of igneous rocks is in a sense paradoxical, because the various types are often found closely associated but the same range of variation is encountered in many places the world over. It can be conjectured that, as the voyage progressed, this suggested to Darwin's intuitive mind a uniformity of action in the volcanic process, the nature of which was unknown.

Near the beginning of the voyage, at St Jago, Darwin found that basalt and trachyte could be found in close proximity. His geological notes show that he recognized the island as being dominantly basaltic, but he also found "cones of a trachytic rock" which consisted of "numerous large crystals of vitreous Feldspar in a base of the same."75 At this early stage he began to speculate on igneous processes, finding that the basaltic dykes of the island "differ solely from neighbouring rock [i.e. eruptive basalt] in being more compact & crystalline & lacking a prismatic form." He wondered: "Are not these differences explained by supposing the dykes have cooled more slowly?."76

Later in the voyage, Darwin found several other examples where basalt and trachyte were closely associated, or had even flowed from the same vent (for example, at Ascension Island, and in the Azores). He was aware of the parallels between these instances and the classic localities of the Auvergne that had been well-studied since the previous century. In that area, basalt tends to overlie trachyte, and some previous authorities, including Alexander von Humboldt (whom Darwin greatly admired) had come to the conclusion that trachyte was the characteristic volcanic effusion of a distant geological epoch, whereas basalt was the lava most commonly produced at the present day (see also Figure 2A). Although this concept continued to be advocated through much of the nineteenth century by various continental authors, Scrope and Lyell were strongly opposed to it: it was irreconcilable with their new philosophy of geological uniformity. Nevertheless, it seems that Darwin accepted as a generalization from his own observations that basaltic streams do often flow over trachyte, which he described this as their "usual order of superposition."77 As we shall see, he was eventually able to provide a satisfying non-progressionist explanation for the phenomenon.

In March 1835, the Beagle was traversing the coast of Chile. At Concepcion, Darwin described a great many dykes of different compositions from the same area. He wrote: "It may be doubted perhaps whether all the above dykes are of the same age, or whether it is possible that two dykes, so similarly circumstanced [i.e., in orientation and thickness], the one composed of Dolerite and the other a white Feldpathic stone, could have flowed from the same fluid mass."78 He even speculated that a trachytic rock encountered soon after was "perhaps a Greenstone dyke, its mineralogical nature altered from the removal of pressure & which poured forth its contents,"79 a remark which seems to imply the separation of a dyke's feldspar crystals from its molten basaltic groundmass. Shortly after making these observations, Darwin made an adventurous journey across the Andes to the town of Mendoza in the plains beyond. High up in the Portillo Pass he found lavas which he described as "dark-greyish, harsh rocks, intermediate in character between trachyte and basalt"80 which might have implied to him that the diverse volcanic rocks are indeed related through intermediate kinds. On the return trip from Mendoza, in the Uspallata Pass, Darwin was again puzzled by finding several different generations of injected rocks of different compositions all occurring in the same vicinity. This time, he was able to show by cross-cutting relationships that each type of rock was injected during a different phase in the geological evo-

[page] 57

Figure 4. A selection of Darwin's hand specimens from the Galápagos Islands. Many of Darwin's volcanic rocks have large feldspar crystals set in a finer grained matrix.

lution of the area. He wrote in his notes: "How are we to understand the variety of melted matter beneath the same part of the surface … From extended observations I must reurge this fact.—I may point out the analogy of a volcano which pours out from the same source at different periods very different lavas." But he also added: "I cannot believe that true dykes entirely different in their constituents can have proceeded from the same source."81

Darwin was to reconsider the question of whether the two main families of volcanic rocks (the basalts and trachytes) ultimately have a common subterraneous origin. If the two rock types were related, the problem was one of chemical segregation—how could different magmas, one rich and the other poor in silica, be derived from a single source? Scrope had published some interesting speculations on this subject in his Considerations of Volcanos: "We may imagine the production of basalt to have been caused by exposure, within the vent of a volcano, of an intumescent mass of granite to reconsolidation, effected by the augmentation of temperature, and consequent expansion, of its lower beds. In these parts the extreme heat may be supposed to volatalize the mica and other ferruginous minerals while the intense pressure would separate them in a gaseous state from the felspar, thus leaving a felspathose lava with very little iron in one part of the chimney, and occasionally the crystallization of a highly ferruginous lava in another."82 It does not seem that Darwin took any special notice of this passage (unusually for him he had not underlined or scored the text in his own copy). Although it is a somewhat fanciful theory, involving a gaseous separation of the mineral phases, the idea clearly involves the natural separation of basalt from trachyte. It may have set Darwin thinking on the subject.


Having made literally hundreds of pages of observations relating to the igneous rocks of South America, Darwin may have had these problems in mind when the Beagle departed the mainland on 7th September 1835 bound for the Galápagos Archipelago. Here he enjoyed yet another splendid opportunity to study well-exposed volcanic rocks. It so happens that many of the rocks Darwin collected from the Galápagos have large crystals (mostly feldspars) set in a finegrained matrix, giving the so-called "porphyritic" texture (Figure 4). In the works available to Darwin, such crystals were often interpreted as remnants left over from some pre-existing rock which had either been partially melted or somehow caught up in the melt.83 However, Darwin's own notes show that he tended toward another explanation.

On examining the rocks of James Island (Isla San Salvador), Darwin was intrigued to find that the lavas "in the lower parts of the Isd are very cellular and the imbedded Crystals of glassy feldspar very large & abundant," but that "In the higher central part the rock generally is more compact, the base blackish grey with scarcely any crystals."84 He also found some flows "interesting from containing very many small generally angular, fragments of altered rocks which clearly have been Granites and Syenites."85 In this he was mistaken, because the xenoliths in Darwin's hand

[page] 58

specimens are a light-coloured altered gabbro. Nevertheless, he went on to speculate: "One is led to suspect that all such Crystals [within the volcanic rock] proceed from the Granite & that they are not produced in the liquid lava" (i.e., that they are xenocrysts), but he declared himself "unwilling to take up this opinion." Instead he speculated: "In the fused mass, when at an intense heat, does not the quartz & a small portion of the other ingredients form the Crystals of glassy Feldspar.—May not these Crystallize at a temperature when the rest of the matter is fluid? This will explain the imbedded & extraneous appearance of the Crystals."86

As Darwin rightly suspected, a molten rock is a complex mixture of chemicals in which, over an appreciable temperature range, particular components aggregate into crystals while the remainder stays fluid. He was probably aware of the experimental work of Sir James Hall and others which had suggested this fact.87 Although he did not explicitly make the point, such incomplete crystallization amounts to chemical segregation, because the chemical composition of a crystal is closely constrained by its structure, and is unlikely to be the same as that of the melt from in it is formed. This phenomenon was the first clue to the problem of the differentiation of the lavas but, on its own, it cannot explain the permanent separation of melts of different composition. Ultimately, even the molten part of the rock will cool and solidify around the crystals producing the "base" or groundmass of the porphyritic texture.

Elsewhere on James Island, Darwin found a key locality that reflected in miniature his general observation that the lower lavas on the island tend to contain larger and more numerous crystals of feldspar. This, as he describes in his notes, is "a small bay close to Albemarle Isd [now Isla Isabela] … which would appear to be [a] slip out of [the] centre of an old crater, both sides being worn into precipitous cliffs by the action of the sea." In the middle of the eroded crater he found "a nearly horizontal bed of lava which thins out at its edges and appears to have filled up formerly the basin of the crater.—It is about 200 ft thick—The stone is a compact blackish or greenish with but few Cryst. of glassy Feldspar." However, on examining the lower surface of this flow he found "to the thickness of about 2 ft., where it lies on the Volcanic detritus is an igneo—cemented Mass of Fragments; these consist of the same substance, slightly & finely cellular & containing larger & more numerous Cryst. of glassy Feldspar than the compacter kinds."88 Unfortunately, he did not collect hand specimens from this locality but the description appears to be that of an igneous cumulate, and Darwin believed that the crystals had sunk to the bottom of the flow.

On the long return leg of the voyage, Darwin speculated freely in his "Red Notebook" about various subjects that interested him, apparently consciously seeking synthetic theories to promote on his return. He repeatedly considered chemical cycling in general and the role of volcanoes in particular. In style, these jottings are similar to the "Transmutation" notebook series in which he developed his evolutionary theory (which in fact run on from the latter pages of the "Red Notebook") and are equally difficult to interpret. In a strikingly Huttonian passage he wrote: "In the endless cycle of revolutions by [the] actions of rivers currents & sea breakers all mineral masses must have a tendency to mingle." But this mingling must be counterbalanced by processes which separate out the chemical constituents. Thus, "the sea would separate quartzose sand from the finer matter resulting from decomposition of feldspar and other mineral containing alumina." The separation of calcium was more problematic: "Is it washed from various solid rocks by the action of springs or more probably by some unknown Volcanic process?"89 A little later he wrote: "Volcanos must be considered as chemical retorts" and continued "neglecting the first production of trachyte, look at sulphur, salt, lime … how comes it they do not flow out together? How are they eliminated?"90

Although he seems to have put his volcanic work aside immediately after the return, a further spur to his theorizing was his reading of Leopold Von Buch's Description Physique des Isles Canaries (which was newly translated into French and had not been present in the Beagle library). In this work, Von Buch describes an obsidian flow at Tenerife, in which "the feldspar appears to become more abundant in the lava proportional with depth in the flow, and in the deeper parts becomes so abundant that the lava often resembles a primary rock."91 This presumably reminded Darwin of the James Island crater for he has scored this section in his own copy of the work. Von Buch then goes on to state, in a sentence also scored by Darwin: "the experiments of M. de Drée, in which he has formed various lavas in a crucible, have proved that in such a fluid mass the crystals of feldspar tend to sink to the bottom."92

Von Buch drew no conclusions from this experiment regarding chemical segregation, but Darwin referred back to his Galápagos notes and scribbled some comments on the back of some of them. He wondered: "Could the Basalt forming Paste to Trachyte, be softer & so basalt forms with less heat than Trachyte," but he also noted that this view "Does not explain Olivine." He also wrote: "It will be safe to state that the felspathic material might easily be sorted & the basalt [derived] from [the] base." He was also concerned about explaining the common order of eruption of these lavas and noted that "Von Buch is very strong about trachyte being first ejected."93 These comments are cross-referenced with similar notes in the "A" notebook, and were written in the same ink and scrappy handwriting. It is safe to conclude, therefore, that they were written some time in late 1837 or 1838 when the "A" notebook was open and Darwin was drafting the initial manuscript for his big geology book (which eventually became three separate books). The comments indicate that at this time he was developing the

[page] 59

Figure 5. 'Pattinson pots,' apparatus for the industrial purification of lead by repeated rounds of fusion and crystallization. Molten lead was cooled in a crucible, while stirring, until crystals began to form. On assaying, such crystals were proved to be of a purer composition than the remaining liquid. They sunk to the bottom and were removed from the melt using perforated ladles. Darwin was aware of the Pattinson process and noted the similarity with his theory of igneous fractionation in nature. Redrawn from H.F. Collins, The Metallurgy of Lead (Charles Griffin: London, 1910).

igneous theory that was eventually published in the Volcanic Islands volume. The events of this period are discussed further in the final section of this paper.

Darwin does not appear to have found any further particulars relating to the experiments of Monsieur Drée, and simply re-quotes him on the authority of Von Buch (who does not give a source). However, in Volcanic Islands, in which the theory eventually appeared, Darwin drew an interesting analogy with "a valuable, practical discovery" made by a Mr H.L. Pattinson (1796–1858) in the field of metallurgy and presented to the Newcastle meeting of the British Association for the Advancement of Science in 1838.94 In the Pattinson process, silver impurities are separated from lead by repeated rounds of fusion and crystallization (Figure 5). Darwin realized that this was similar to the process as that which was occurring on a much larger scale in bodies of molten rock.

As is typical of Darwin's method, he was confident enough of his observations to build upon them a broad and sweeping generalization. As he wrote in Volcanic Islands, the gravitational settling of crystals within a semi-molten body of rock was "worthy of further consideration, as throwing light on the separation of the trachytic and basaltic series of lavas."95 Within the earth, where great masses of molten rock were being formed, intruded, and gradually cooled, the settling out of crystals could be occurring on a huge scale. This process could account for the variety in igneous rocks that had so puzzled geologists:

In a body of liquified rock, left for some time without any violent disturbance, we might expect, in accordance with the above facts, that if one of the constituent minerals became aggregated into crystals or granules, or had been enveloped in this state from some previously existing mass, such crystals or granules would rise or sink, according to their specific gravity. Now we have plain evidence of crystals being embedded in many lavas, whilst the paste or basis has continued fluid… Lavas are chiefly composed of three varieties of feldspar, varying in specific gravity from 2.4 to 2.74; of hornblende and augite, varying from 3.0 to 3.4; or olivine, varying from 3.3. to 3.4; and lastly, of oxides of iron, with specific gravities from 4.8 to 5.2. Hence crystals of feldspar, enveloped in a mass of liquified, but not highly vesicular lava, would tend to rise to the upper parts; and crystals or granules of the other minerals, thus enveloped, would tend to sink. We ought not, however, to expect any perfect degree of separation in such viscid materials.96

An apparent inconsistency in Darwin's account is exposed here, because at the James Island crater he refers to feldspar crystals having sunk, not risen, in a basaltic lava. He explained this difficulty by proposing that the many bubbles of gas enveloped in the lower part of that flow might have lowered the specific gravity of the melt sufficiently for the feldspar to sink. In general, however, Darwin believed that the feldspar would rise relative to the denser and less siliceous minerals.

After stating how crystal settling could account for the separation of compositionally distinct volcanic rocks, Darwin then proceeded to show how almost all of his disparate observations on lavas, as reported rather dryly in the earlier pages of the Volcanic Islands volume, could be reconciled with the theory. For example, it would explain why trachyte, which he quotes as having a specific gravity of about 2.45, separates out at the top of a chamber of molten rock and tends to be erupted first, followed by basalt with a specific gravity of about 3.0. This would account for the "usual order of superposition" of these lavas as seen at Ascension and elsewhere without recourse to the unwelcome progressionist notion of trachytes being the ancient equivalents of basalts. On the other hand, in volcanic mountains of "lofty and great dimensions," differentiation might take place inside the body of the volcano itself, explaining why trachyte flows from the higher vents, and the denser basalt, almost contemporaneously, from the base. And it would also explain why obsidian, which has a lower specific gravity even than feldspar, is erupted only from the quot;highest orifices" of volcanoes.

Many of these generalizations still hold today, and attest to the great success of this part of Darwin's theory. However, it cannot be claimed that he had a complete understanding of the processes he described. Darwin's ideas form a large part of the modern theory of "fractional crystallization" in igneous rocks, the importance of which for generating variety can hardly be overstated. However, Darwin had no appreciation of the distinction between eutectic and non-eutectic mixtures, which is essential for a complete understanding of the mechansims involved, and only a vague appreciation of the most likely compositional changes that crystal settling might produce. Nevertheless, modern petrologists agree that fractional crystallization is an important process within volcanoes and at greater depths, and it is the chief mechanism which gives rise to diversity among the igneous rocks. Gravity settling is regarded as one of the most important processes by which this separation is effected.97 Such a process is a reasonable explanation for the various examples Darwin quoted, although new collecting and detailed chemical analysis would be required to confirm any particular case such as the James Island crater, or Von Buch's obsidian flow at Tenerife.

[page] 60


In parallel with his theory for the separation of trachytes from basalt, Darwin pondered the origin of molten rock within the Earth. All nineteenth century authors seem to have assumed that granite was the fundamental material from which the volcanic rocks were produced. The idea can be traced back to the mid-eighteenth century and early researches in the Auvergne, where granite and gneiss underlie the volcanic district. Granitic lumps can commonly be found carried within the basalt streams there, which seemed to prove that they were formed by melting of the granite. Darwin himself thought he had observed the same phenomenon at both Ascension and in the Galápagos (wrongly, as it turned out). This presented a great problem to the chemically-minded Darwin, because granite and granitoid gneiss are silica-rich rocks, whereas basalt is silica-poor.

Charles Daubeny, the principal British volcanologist in Darwin's time, advocated a simple scenario by which granite could be related to the volcanic rocks which, if anything, can be taken as the standard theory of the day. Melting of granite by subterranean heat, Daubeny hypothesized, would produce trachyte, which he knew to be of a similar if not identical chemical composition to granite. Daubeny thought that basalt must in turn be produced from this primary trachyte by some natural process, although he had no specific theory of how that might occur beyond the "continuation of heat" and the "admixture of other matters" of unspecified composition.98 Darwin could not find this scheme attractive because, by his reckoning, trachyte was itself a segregation product formed during crystallization (see above). Nor does it appear to "balance" chemically, because there is no obvious way in which the silica-poor rocks such as basalt can be derived from trachyte.

During the Beagle voyage, Darwin saw several "primary" districts where the geological underpinning of granite and gneiss had been laid bare by erosion. In all such areas he visited, he found concentrations of basaltic (otherwise known as "trappean" or "greenstone") dykes. Near the start of the voyage, at Rio de Janeiro in June 1832, Darwin was sufficiently impressed by the greenstone dykes he had seen cutting the gneisses and mica slates of the district that he wondered whether they had been formed by "some chemical alteration of the primitive crust of the globe."99 He encountered further examples of dykes cutting "primary" rocks in Tierra del Fuego. Here he noted their similarity in appearance to the neighbouring metamorphic clay slate formation and wondered whether fusion of the latter might give rise to greenstone.100 A vast number of basaltic dykes were encountered cutting the mica slates of the Chonos archipelago and elsewhere on the coast of Chile, some of which were on a grand scale. Darwin was repeatedly impressed by their general parallelism to the axis of the granitic mountains which seemed to indicate that they might be related in some way.101 Further examples of trappean dykes cutting granites were found at King George's Sound, Australia. Here, "In one part they were in such close proximity, that the Greenstone in quantity exceeded the partitions of granite"102 (a photograph and discussion of this locality are given by Armstrong).103 He had also made similar observations in South Africa. All these occurrences are reminiscent of the dykes in the idealized geological sections of Buckland and others (see Figure 2A) which cut the "primary" districts. Perhaps some process of segregation had occurred at depth in these regions, akin to that which he had devised to account for the various lavas, which could give rise to the basaltic magmas in the first place.

An important outcrop which puzzled Darwin on two occasions during the voyage of the Beagle is in the vicinity of Bahia (Salvador) in Brazil. Although it is a somewhat intricate locality, a detailed consideration of Darwin's observations there goes a long way to providing an understanding why he developed his particular theory for the origin of basalt. His first visit to Bahia was near the start of the voyage in February and March of 1832. His second was three and a half years later in August 1836 when the Beagle unexpectedly returned to Brazil after circumnavigating. His original descriptions from these visits survive in his unpublished geological notes, and include some rough field sketches. These notes are particularly important in deciphering Darwin's initial impressions of these rocks (Figure 6), because, just as in the case of the James Island crater, he was to base much subsequent theorizing on this locality and eventually reinterpreted what he had seen in the light of a distinct conceptual model.

Darwin's notes show that on his first visit to Bahia he immediately recognized that it was a region of granitic gneiss—his first encounter with this high-grade metamorphic rock in the field. He was particularly puzzled by "angular pieces with very defined edges of Hornblendic rock" that were surrounded on all sides by gneiss. Darwin's sketches and descriptions appear to represent so-called "mafic pods" that have been pulled apart during metamorphism (as is frequently found in such rocks). Darwin also noted that this basaltic rock also "occurs in dykes varying in thickness from 4 inches to as many yards" in which "the hornblende only differed from the entangled pieces in not being so much crystallized." He wrote that

Figure 6. A sample of Darwin's "Notes on the Geology of the Beagle Voyage. "This is the first of several pages which record observations made during his first visit to Bahia in February-March 1832. Numbers in the margin refer to hand specimens. Thin sections of some of these have subsequently been made by Alfred Harker. See text for a discussion of Darwin's observations at Bahia. By permission of the syndicates of Cambridge University Library.

[page] 61

[page] 62

Figure 7. Thin-sections of two of Darwin's so-called "Hornblende rocks" from Bahia. Right: The "Hornblende rock" which comprizes the broken-up pods (Darwin's specimens 319, 320), described in Darwin's specimen list as "Hornblende rock entangled in gneiss," is a high-grade metamorphic rock (plagioclase amphibolite). The texture is one of interlocking equidimensional crystals of amphibole and feldspar, as occurs in high-temperature metamorphism. Left: The "Hornblende rock" which comprizes the dykes "of black heavy basaltic stone" (Darwin's specimens 319, 3833, 3837) is an unmetamorphosed dolerite, with a characteristically igneous texture of unorientated feldspar needles in a finer matrix.

"The double case of the Hornblendic rock penetrating the surrounding mass & itself in angular pieces being caught up in the gneiss appears to me very curious." He therefore wondered if both the gneiss and hornblende rock were "poured forth at nearly the same time & in different spots either one of the two to have hardened first & subsequently to have been broken up or penetrated by the other" (which, incidentally, illustrates that at this early stage in the voyage, Darwin did not fully appreciate the distinction between igneous and metamorphic processes).104

On his second visit, Darwin had another look at this peculiar locality. Although he was much more experienced at fieldwork and now appreciated the distinction between granite and gneiss, the rocks still proved difficult to interpret. He confirmed many of his 1832 observations and also found a peculiar dyke that had been broken up by faults and appeared to prove to him that "most or all of the fragments [i.e. the mafic pods] have been derived from the breaking up of the dyke." This observation once again seemed to prove that a "body of igneous matter" had been injected into a granitic gneiss which entrained it, semi-fluid, in a complex manner. Around the dyke he found numerous curvilinear threads of hornblende rock which seemed to penetrate the gneiss. He marvelled that only "a substance fluid as water could be injected into such minute fissures." Darwin supposed that the "threads proceed from dykes" but he was uncertain, and wrote that "an examination of the mineralogical nature of the dikes and fragments will settle the question of [their] reciprocal position."105

Back in Cambridge, an examination of the hand specimens by Miller failed to shed much further light on the question. A study of Alfred Harker's thin sections of Darwin's rocks (an option which was, of course, not available at the time) helps solve the puzzle (Figure 7). The "hornblende rock" that comprizes the broken up pods is clearly a high-grade metamorphic rock and must be an integral part of the gneiss. On the other hand, the "hornblende rock" that comprizes the dykes is an unmetamorphosed dolerite. This rock has not experienced the extremes of temperature and pressure of the former and so the dykes must have been intruded long after the metamorphism occurred. Because of their similarity in hand specimen and chemical composition (as ascertained by the blowpipe), Darwin failed to appreciate the fundamental distinction between these two rock-types.

At a later stage, Darwin reflected on his Bahia observations. Wrongly believing the two "hornblende rocks" to be one and the same, it dawned on him that the locality might provide a clue to the process of chemical segregation. Referring to the filaments of hornblende rock around the dyke in Volcanic Islands, he candidly admitted to have originally "doubted whether such hair-like and curvilinear threads could have been injected [from the dyke]; and I now suspect that … they were its feeders."106 Thus, he hypothesized that the threads were tiny veins of melt that had been squeezed out of the gneiss at high temperature and had flowed together to fill the dyke. Once again he was willing to turn the specific observation into a grand generalization. He proposed that metamorphism

[page] 63

in the deeper regions of the earth's crust might on the one hand produce the so-called "primary" rocks such as gneiss, and, on the oher hand, result in partial melting and squeezing out of a basaltic liquid as he thought had occurred at Bahia. The chemical composition of granite, which was produced by melting of gneiss in such regions, was, therefore, explained as opposite and complementary to basalt. In this scheme, granite was an igneous rock produced from the melting of the refractory remnants left over from the initial melting and squeezing out of the more fusible hornblende, augite and other basic minerals under great heat and pressure. The frequently encountered "greenstone" dykes that criss-cross the "primary" districts were the conduits by which the first-formed basaltic liquid was drained away "into deep and unseen abysses," afterwards perhaps to be brought to the surface "under the form either of injected masses of greenstone and augite porphyry, or of basaltic eruptions."107 As Darwin wrote to Lyell, if such "separation of the molten elements" did not occur, "how else could the basaltic dykes come into great granitic districts such as those of Brazil?."108

To Darwin, these igneous processes were central to the crustal cycle of chemical segregation and recombination. In his 1837-8 "A" Notebook, he called this scheme "my theory of changes of granites into trachytes."109 Tellingly, the only other idea he referred to as "my theory" at that time was natural selection itself. Darwin's scenario for the origin of basalt can clearly be identified as one of "partial melting," which is now widely appreciated to be another important mechanism by which the compositions of igneous rocks are determined. It is a reasonable supposition that if heat is applied to a rock, (or if the pressure is reduced) and if the rock does not melt entirely at one temperature, the most easily fusible minerals would dictate the chemistry of the liquid evolved. Unlike some of his contemporaries such as Daubeny, Darwin appreciated that the chemistry of the liquid evolved might be wholly different from that of the rock from which it was produced.

The theory does, however, have a fundamental flaw. A key assumption is that the basic minerals such as hornblende and augite are the ones with the lowest melting points. In this, Darwin was following the universally accepted view of his time. For example, he scored passages in relevent works that make this claim.110 It was not until the turn of the twentieth century that it generally became appreciated that most of the common silicate minerals that have the lowest melting point are generally richer in silica. Mineralogists of the time had not conducted sufficient experiments on the silicate mineral systems to determine their melting properties. It seems that the low viscosity of basalt on eruption was erroneously equated with ease of melting. We now know that partial melting of granite or granitic gneiss cannot produce a melt of basaltic composition. Thus, Darwin's interpretation of the Bahia locality cannot be correct, and his model of igneous cycling falls down at this point.

Nevertheless, Darwin's scheme was a fairly comprehensive one and certainly a respectable attempt for the time. It was thoroughly Lyellian in outlook, in that it sought to account for the chemical variability of igneous rocks in a "steady-state" manner without accepting that any of the variation in mineralogical composition could be due simply to the fact that particular rocks were formed in particular geological periods. It left open the problem of the origin of concentrated ore deposits and mineral veins, of which Darwin had seen a great variety in South America, and there is much evidence in his notebooks that this became his geological priority after the return to England. For example, his brief excursion to Salisbury Crags in Edinburgh in June 1838 seems to have been to study the "veins of segregation" there. However, ill health, marriage and growing preoccupation with species prevented him from pursuing these lines of investigation in the field.


Charles Darwin's igneous theories are interesting in that they tell us much about the scope and limitations of "hard rock" geology in the early nineteenth century. They also provide good examples of the scientific method employed by Darwin, who is, after all, one of the principal figures of the history of science. More fundamentally, however, it is possible to draw a limited analogy between his theory for the separation of igneous rocks by crystal separation and evolution by natural selection. Indeed, some evolutionary theorists might even be persuaded that crystal separation is a kind of natural selection, albeit under very limited conditions. It is, therefore, conceivable that it was one of the many background influences that led Darwin towards his most famous idea. Before developing this argument, however, the close correspondence in time in the formulation of the two theories should be explained.

By the time the Beagle had departed the Galápagos in October 1835, Darwin had made observations and collections that ultimately proved crucial to both his igneous and evolutionary theories. Although doubts about the fixity of species occurred to him during the voyage, it is well-known that he was to work on evolution only after his return to England. Similarly, Darwin's geology notes show very clearly that he had in mind some kind of scheme for igneous segregation in the last months of the voyage but it is difficult to discern how developed it was.

His first priority on his return, after a hectic round of visits to relatives and scientists alike, was to write a general account of his travels for a popular audience (the Journal of Researches). This he abstracted from his diaries and letters home, and so managed to complete by September 1837, although its eventual appearance in print was much delayed. At the same time he also began to deliver various specialist papers to the Geological Society and elsewhere promoting his

[page] 64

Lyellian agenda, but none specifically on the igneous rocks. He apparently worked on volcanic geology initially between October 1837 and June 1838.111 It was in this period that he read Von Buch's Isles Canaries, and may have first encountered Pattinson's process for the purification of lead—both obvious influences on his igneous theory. At this time he also made several relevant entries in his "A" notebook and reworked the Beagle geology notes. In a jocular letter to a friend dated May 1838, Darwin hinted at his preoccupation with the igneous question. The possibility of his getting married, he wrote, "always drives granite and trap out of my head in the most unphilosophical manner!"112 The final appearance of the volcanic theory in print was much delayed, however. Darwin put the question to one side again until 1843-4 when, in poor health, he finally compiled the Volcanic Islands volume for publication.

This sequence of events is not dissimilar to the well-studied chronology for Darwin's discovery of natural selection. His meeting with the ornithologist John Gould in March 1837, in which he was told that the famous Galápagos finches were not one but several species, has been regarded as a significant spur for the development of his theory of evolution. After this, he soon began to make jottings on species in his notebooks. By July 1837 he was enthusiastic enough to open a separate notebook devoted to "Transmutation" (the "B" notebook) that ran concurrently to the geological "A" notebook. At this time, however, Darwin did not have a clear idea of the mechanism by which evolution might occur. But by the spring and summer of 1838, his notes (by now the "C" and "D" notebooks) show that he had become concerned with the powers of selection in shaping species.113 Thus, as far as can be ascertained, Darwin's initial ideas on selection occurred in the same period as he was actively working out his theory of crystal separation.

Darwin's evolutionary speculations began to preoccupy him and to some extent displaced his geological work. In a letter to Lyell in September 1838 he referred to both his interests in consecutive sentences: "I wish with all my heart that my geology book will be out soon.—the volcanic chapter, will, I think, contain some new facts.—I have sadly been tempted to be idle, that is as far as pure geology is concerned, by the delightful number of new views, on the classification and affinities and instincts of animals, bearing on the question of species." It was later that month that he read Malthus's Essay on Population and his theory of evolution by natural selection began to take on its final shape.114 As is well known, the evolutionary theory was also much delayed—the Origin of Species was not written until 1858-9.

The most obvious similarity in the two theories, beyond the timing, is that Darwin was invoking simple, naturally occurring mechanisms for generating variety. As the petrologist Alfred Harker wrote in relation to the igneous rocks (in another context): "The only practical alternative to magmatic differentiation, as accounting for the observed facts, is the doctrine of countless special creations."115 It should be remembered that Darwin was enthusiastic about the transmutation of species well before he settled on the mechanism of natural selection. Thus, it is not unreasonable to propose that his success at finding the common origins of diverse geological objects, including among the categories of volcanic rocks, may have predisposed him toward pondering the inter-relation of species.

To anyone not well acquainted with the igneous rocks this might seem like stretching a point. However, the language of modern igneous petrology strongly reflects the "evolutionary" nature of the subject. N.L. Bowen's classic textbook of igneous petrology is tellingly entitled The Evolution of The Igneous Rocks.116 The compositional path taken by a liquid during fractional crystallization as it becomes "more evolved" has become known as the "liquid line of descent." The quotation at the head of this article, which is drawn from a modern textbook on the subject, also serves to emphasize the point.

Most fundamentally, however, the specific mechanisms that Darwin suggested for magmatic differentiation and organic evolution have striking similarities. The former theory, simply stated, is that the crystallization and removal of a mineral from a body of molten rock inevitably causes a chemical change in the remaining melt. In effect, the composition of the melt is "driven away" from the composition of the mineral that is crystallizing out and being removed, or segregated. For example, if a particularly magnesium-rich mineral (such as olivine) were to be removed from a melt, the remainder must be left depleted in magnesium and relatively enriched in all the other chemical components present. This was the key to how very different-looking rocks are formed. It is the subtractive nature of this process that was Darwin's special insight and distinguishes his theory from those of his competitors (Charles Daubeny, by contrast, referred to the "successive additions of lime and magnesia" [my italics] which connect trachyte to basalt). Subtraction is also the essence of natural selection. Thus, if a particular animal displays a disadvantageous trait, it is likely to be removed, by death, from the population. The remainder of the population, inevitably, will be less likely to exhibit that trait. Similarly, any characteristic found in the species that the dead individual does not possess must be increased in the general population. Evolutionary change is thus directed away from the genetic composition of the organism selected, and the amount of change is proportional to the selection pressure, just as the direction of magmatic differentiation is away from the composition of a crystallizing mineral and proportional to the degree of fractionation.

It has often been said that organic evolution is a logical necessity following from the postulates of variation, inheritance and the struggle for life.117 Similarly, magmatic differentiation follows inevitably from several factors, most important among which are the existence of rocks in a semi-molten state, the observed

[page] 65

chemical and density differences between minerals, and the influence of gravity. Thus, at the core of both processes lies a natural "algorithm," in the sense proposed by Dennett.118 The algorithm which underlies crystal separation is an important one to geologists, because it is at work the world over, and indeed on other planetary bodies.

Despite these parallels, it is obvious that crystal separation does not amount to "evolution" in the scientifically accepted sense: at least it lacks some important elements that are necessary for organic evolution. In living things, reproduction perpetuates selected traits indefinitely. In this way, Darwin realized, populations eventually become so modified that organisms appear to be designed so as to promote their own survival. In the world of igneous rocks there is no reproduction and hence, no such adaptation can occur. Heredity, not selection, has been identified as the "great divide" that separates the processes of "evolution" in the living and non-living worlds.119 Darwin himself was well aware that his theory of evolution was a modular one, with selection, heredity, and variation as separate, but equally essential components.

Some evolutionists have acknowledged that "natural selection" can occur in the absence of heredity. For example, Dawkins makes a distinction between "single step" selection, which does not involve heredity, and "cumulative selection" which does. The aggregation of crystals from solution is one example cited by Dawkins of "single step" natural selection.120 Density segregation, as occurs in crystallization differentiation, can be thought of as a particularly good example of this. It is one of many processes of sifting and sorting that impart order to the non-living world and are, in a limited way, a kind of natural selection. Darwin himself never acknowledged the common ground shared by his two theories. This lack of a "smoking gun" does not, however, preclude the possibility that the ideas were intimately cross-fertilized in his mind, or at least were products of the same view of nature and method of creative thinking.


I would like to thank the staff of the University Library, Cambridge, the Down House Museum, Downe and the Sedgwick Museum, Cambridge, for assistance in locating source material, and Mr George Pember Darwin for permission to reproduce extracts from Charles Darwin's unpublished notes. Dr Graham Chinner, Professor Steve Sparks and Dr Keith Cox kindly read the manuscript from the point of view of the modern science of igneous petrology and each made very useful comments. Professor Martin Rudwick commented on an earlier draft from the point of view of the history of geology. The paper benefitted greatly from reviews by Dr Sandra Herbert and Dr Hatten S. Yoder Jr.


1. Charles Darwin, Geological Observations on the Volcanic Islands Visited During the Voyage of HMS Beagle, Together with some Brief Notes on the Geology of Australia and the Cape of Good Hope, Being the Second Part of the Geology of the Voyage of the Beagle, Under the Command of Capt. Fitzroy R.N., During the Years 1832 to 1836, (London: Smith Elder, 1844).

2. Joseph P. Iddings, "The Origin of Igneous Rocks," Philosophical Society of Washington, Bulletin, 1892, 12:89-214.

3. Alfred Harker, The Natural History of Igneous Rocks, (London: Hafner, 1909).

4. R.A. Daly, Igneous Rocks and their Origin, (New York: McGraw-Hill, 1914).

5. T.F.W. Barth, Theoretical Petrology, (New York: John Wiley, 1952).

6. L.R. Wager and G.M. Brown, Layered Igneous Rocks, (San Francisco: W.H. Freeman, 1967).

7. E.A.K. Middlemost, Magmas and Magmatic Rocks, (London: Longman, 1985).

8. Norman L. Bowen, The Evolution of the Igneous Rocks, (Princeton: Princeton University Press, 1928).

9. Nora Barlow (ed.). The Autobiography of Charles Darwin 1809-1882, With Original Omissions Restored, (London: 1958).

10. Sandra Herbert (ed.), The Red Notebook of Charles Darwin, Bulletin of the British Museum (Natural History) Historical Series, Vol. 7 (London: Cornell U. Press, 1980).

11. PH. Barrett, P.J. Gautrey, S. Herbert, D. Kohn, and S. Smith, Charles Darwin's Notebooks 1836-1844, (Cambridge: Cambridge University Press, 1987).

12. F. Burkhardt and S. Smith (eds), A Calendar of the Correspondence of Charles Darwin, 1821-1882, (New York: Garland, 1985a).

13. F. Burkhardt and S. Smith (eds), The Correspondence of Charles Darwin, Vol. 1, 1821-1836, (Cambridge: Cambridge University Press, 1985b).

14. F. Burkhardt and S. Smith (eds), The Correspondence of Charles Darwin, Vol. 2, 1837-1843, (Cambridge: Cambridge University Press, 1986).

15. F. Burkhardt and S. Smith (eds), The Correspondence of Charles Darwin, Vol. 3, 1844-1846, (Cambridge: Cambridge University Press, 1987).

16. F. Burkhardt and S. Smith (eds). The Correspondence of Charles Darwin, Vol. 4, 1847-1850, (Cambridge: Cambridge University Press, 1988).

17. M.A. DiGregorio and N.W. Gill, Charles Darwin's Marginalia, (New York: Garland, 1990).

18. D. M. Porter, "The Beagle collector and his collections," in D. Kohn (ed.), The Darwinian Heritage, (Princeton: Princeton University Press, 1985), pp. 9-34.

19. Alfred Harker, "Notes on the rocks of the 'Beagle' collection—I," Geological Magazine, 1907, 4:100—106.

20. L.J. Chubb and C. Richardson, "Geology of Galapagos, Cocos, and Easter Islands," Bernice P. Bishop Museum, Bulletin, 1933, 110,1-67.

21. Kenneth L. Taylor, "Nicolas Desmarest and Geology in the Eighteenth Century," in C.J. Schneer (ed.), Toward a History of Geology: Proceedings of the New Hampshire Inter-Disciplinary Conference on the History of Geology, September 7-12, 1967, (Cambridge, Mass: MIT Press, 1969, p.339-356).

22. Rachel Laudan, From Mineralogy to Geology: The Foundations of a Science, 1650-1830, (Chicago and London: University of Chicago Press, 1987).

23. Anthony Hallam, Great Geological Controversies, 2nd ed. (Oxford: Oxford University Press, 1989).

24 James A. Secord, "The discovery of a vocation—Darwin's early geology," British Journal for the History of Science, 1991, 24: 133-157.

25. James Hutton, "Theory of the Earth, or an investigation of the laws observable in the composition, dissolution, and restora-

[page] 66

tion of land upon the globe," Transactions of the Royal Society of Edinburgh, 1788, 1: 209-234.

26. Edward Battersby Bailey, James Hutton—the Founder of Modern Geology, (Amsterdam: Elsevier, 1967).

27. Robert H. Dott, Jr., "James Hutton and the concept of a dynamic Earth," in C.J. Schneer (ed.), Toward a History of Geology, September 7-12, 1967 (Cambridge, Mass: 1969, pp. 122-142.

28. Dennis R. Dean, James Hutton and the History of Geology, (Ithaca and London: Cornell University Press, 1992)

29. Richard Kirwan, "Observations on the proofs of the Huttonian Theory of the Earth, adduced by Sir James Hall, Bart.," [Nicholson 's] Journal of Natural Philosophy, 1800, 4:97-102 and 153-158.

30. Cyril Stanley Smith, "Porcelain and Plutonism," in C.J. Schneer (ed.), Toward a History of Geology, pp. 317-338.

31. Charles Daubeny, A Description of the Active and Extinct Volcanos; With Remarks on their Origins, their Chemical Phenomena, and the Character of their Products, as Determined by the Conditions of the Earth During the Periods of their Formation, (London: 1826).

32. Charles Daubeny, A Description of Active and Extinct Volcanos, of Earthquakes, and of Thermal Springs; with remarks on the Causes of their Phaenomena, the Character of their Respective Products, and Their Influence on the Past and Present Condition of the Globe, (2nd ed.) (London: 1848).

33. William Buckland, Geology and Mineralogy Considered with Reference to Natural Theology, ("Bridgewater Treatise") (London: Pickering, 1836).

34. John MacCulloch, "General view of the origin, character, and disposition of unstratified rocks and veins," Quarterly Journal of Science, Literature and the Arts, 1826, 22,1-28.

35. Henry T. De La Beche, A Geological Manual, 3rd ed. (London, 1833).

36. Charles Lyell, Principles of Geology, Being an Attempt to Explain the Former Changes of the Earth's Surface by Causes Now in Operation, (London: John Murray, 1830-1833).

37. Henry S. Boase, A Treatise on Primary Geology; Being an Examination both Practical and Theoretical of the Older Formations, (London, 1834).

38. John W. Judd, "Critical Introduction" to Charles Darwin's South America, (London: Ward Lock, 1890).

39. See, for example, Lyell to Darwin, 29th August 1837 (Trancribed in Burkhardt and Smith, eds., Correspondence).

40. L. von Buch, 'Uber die Zusammerensetzung der basaltischen Insels und uber Erhebungs-Cratere', Abhandlungen der Koniglichen Akademie der Wissenschaften, Berlin, 1820, 51: 86.

41. George Poulett Scrope, Considerations of Volcanos, the Probable Causes of their Phenomena, the Laws which Determine their March, the Deposition of their Products, and their Connexion with the Present State and Past History of the Globe: Leading to the Establishment of a New Theory of the Earth, (London: 1825).

42. George Poulett Scrope, Memoir on the Geology of Central France; Including the Volcanic Formations of the Auvergne, the Velay, and the Vivarais, (London: 1827).

43. Lyell, Principles.

44. Martin J.S. Rudwick, "Lyell on Etna, and the Antiquity of the Earth," in C.J. Schneer (ed.), Toward a History of Geology, pp. 317-338.

45. Secord, "Vocation."

46. John G. Burke, "Mineral classification in the early nineteenth century," in C.J. Schneer (ed.) Toward a History of Geology, pp. 62-77.

47. George Poulett Scrope, "Descriptive arrangement of volcanic rocks," Quarterly Journal of Science and the Arts, 1827, 21, 216-299.

48. Lyell, Principles, p. 396.

49. Charles Darwin, "Geology: Section VI," in J.F.W Herschel (Editor), A Manual of Scientific Enquiry: Prepared for the use of Her Majesty's Navy and Adapted for Travellers in General, (London: John Murray, 1849).

50. Harker, "Beagle collection."

51. Silvan S. Schweber, "The wider British context in Darwin's theorizing," in D. Kohn (Editor), The Darwinian Heritage, (Princeton: Princeton University Press, 1985, p. 35-39).

52. Secord, "Vocation."

53. Martin J.S. Rudwick, The Great Devonian Controversy, (Chicago: Chicago University Press, 1985).

54. Archibald Geikie, Charles Darwin as Geologist, (Cambridge: Cambridge University Press, 1909).

55. John W. Judd, "Critical Introduction" to Charles Darwin's Coral Reefs, (London: Ward Lock, 1890).

56. John W. Judd, "Critical Introduction" to Charles Darwin's Volcanic Islands, (London: Ward Lock, 1890).

57. John W Judd, "Critical Introduction" to Charles Darwin's South America, (London: Ward Lock, 1890).

58. D.R. Stoddart, "Darwin, Lyell, and the geological significance of coral reefs," British Journal for the History of Science, 1976, 9, 199-218.

59. Sandra Herbert, "Remembering Charles Darwin as geologist," in R. Chapman and C.T Duval (eds.), Charles Darwin: A Centennial Commemorative, (Wellington: Nova Pacifica, 1982), pp. 231-258.

60. Sandra Herbert, "Darwin the young geologist," in D. Kohn (ed.). The Darwinian Heritage, (Princeton: Princeton University Press, 1985), pp. 483-510.

61. Martin J.S. Rudwick, "Darwin and the world of geology," in D. Kohn (ed.), The Darwinian Heritage, (Princeton: Princeton University Press, 1985), pp. 511-518.

62. Sandra Herbert, "Darwin as a Geologist," Scientific American, 1986, 245: 94-107.

63. Secord, "Vocation."

64. Patrick Armstrong, Darwin's Desolate Islands: a Naturalist in the Falklands, 1833 and 1834, (Chippenham, 1992).

65. Secord, "Vocation."

66. Barlow (ed.), Autobiography, p. 77.

67. Cambridge University Library, Manuscripts department, DAR 32.1:20.

68. Down House "Red" Notebook, p. 109 (transcribed by S. Herbert in Barrett and others, Charles Darwin's Notebooks, p. 56).

69. Cambridge University Library, Manuscripts department, DAR 32.1:34.

70. Charles Darwin. Journal and Remarks, 1832-1836, Vol. 3. of R. Fitzroy (ed.), Narrative of the Surveying Voyages of His Majesty's Ships 'Adventure' and 'Beagle'; Between the Years 1826 and 1836, (London, 1839), p. 291.

71. Frank J. Sulloway, "Darwin and his finches: the evolution of a legend," Journal of the History of Biology, 1982, 15:1-53.

72. Frank J. Sulloway, "Darwin's conversion: the Beagle voyage and its aftermath," Journal of the History of Biology, 1982, 15:325-396.

73. Burkhardt and Smith, Correspondence, Vol 1, Appendix IV, p. 533.

74. Scrope, "Volcanic Rocks."

75. Cambridge University Library, Manuscripts department, DAR 32.1:26.

76. Cambridge University Library, Manuscripts department, DAR 32.1:24.

77. Darwin, Volcanic Islands, p. 42.

[page] 67

78. Cambridge University Library, Manuscripts department, DAR 35.2:363.

79. Cambridge University Library, Manuscripts department, DAR 35.2:391.

80. Charles Darwin, Geological Observations on South America, Being the Third Part of the Geology of the Voyage of the Beagle Under the Command of Capt. Fitzroy, R.N., During the Years 1832-1836, (London: Smith Elder, 1846), pp. 184-185.

81. Cambridge University Library, Manuscripts department, DAR 36.2:532.

82. Scrope, Considerations, p. 146.

83. Charles Lyell, Elements of Geology, (London: John Murray, 1838), p. 475.

84. Cambridge University Library, Manuscripts department, DAR 37.2:770.

85. Cambridge University Library, Manuscripts department, DAR 37.2:775.

86. Cambridge University Library, Manuscripts department, DAR 37.2:776.

87. Laudan, Mineralogy.

88. Cambridge University Library, Manuscripts department, DAR 37.2:772. (A similar passage which probably represents an earlier draft can be found in DAR 37.2:716-721)

89. Down House, "Red" Notebook, p. 21 (transcribed by Barrett and others, Notebooks, p. 46).

90. Down House, "Red" Notebook, p. 78 (transcribed by Barrett and others. Notebooks, p. 26).

91. Christian Leopold Von Buch, Description Physique des Isles Canaries, (Paris: 1836), p. 190. ("Le feldspath parait augmenter dans la lave, a mésure qu'on considère des parties plus profondes dans le courant, et il y devient tellement abondant que cette lave resemble souvant à une roche primitive").

92. Von Buch, Isles Canaries, p. 191. ("Les expériences de M. de Drée, dans lesquelles il a fait fondre diverses laves dans un creuset, ont prouve que dans une telle masse fluide, les cristaux de feldspath devaient tendre à se precipiter au fond").

93. Cambridge University Library, Manuscripts department, verso of DAR 37.2:794-5.

94. H.L. Pattinson, "On a new process for the extraction of silver from lead," Report of the Eighth Meeting of the British Association for the Advancement of Science, Vol vii, Transactions of the Sections, 1838, 50-55.

95. Darwin, Volcanic Islands, p. 118.

96. Darwin, Volcanic Islands, p. 119-120.

97. Keith G. Cox and Clive Mitchell, "Importance of crystal settling in the differentiation of Deccan Trap basaltic magmas," Nature, 1989, 335:447-449.

98. Daubeny, Volcanos.

99. Cambridge University Library, Manuscripts department, DAR 32.1:58.

100. Cambridge University Library, Manuscripts department, DAR 32.2:102.

101. Cambridge University Library, Manuscripts department, DAR 35.1:254.

102. Cambridge University Library, Manuscripts department, DAR 38.1:865.

103. Patrick Armstrong, Charles Darwin in Western Australia, (Nedlands: University of Western Australia Press: 1985).

104. Cambridge University Library, Manuscripts department, DAR 32.1:3-5; DAR 32.1:41-48.

105. Cambridge University Library, Manuscripts department DAR 38.2:956.

106. Darwin, Volcanic Islands, p. 123.

107. Darwin, Volcanic Islands, p. 124.

108. Darwin to Lyell, September 1849 (Transcribed by Burkhardt and Smith (editors). Correspondence, Vol. V, p. 46).

109. Cambridge University Library, "A" Notebook, p. 35 (Transcribed by Barrett and others. Notebooks, p. 94.

110. J. Fournet, "Filons de l'Arbrelsle," L'Institut, 1837, 5 (218): 246-249.

111. Barrett and others, Notebooks.

112. Correspondence, Vol. VII (Supplement), p. 468.

113. Barrett and others, Notebooks.

114. Dov Ospovat, The Development of Darwin's Theory: Natural History, Natural Theology, and Natural Selection, 1838-1859, (Cambridge: Cambridge University Press, 1981).

115. Harker, Petrology, p. 310.

116. Bowen, Igneous Rocks.

117. Julian Huxley, Evolution: The Modern Synthesis. (London: George Allen and Unwin, 1942), p. 14.

118. Daniel C. Dennett, Darwin's Dangerous Idea: Evolution and the Meanings of Life. (London: Allen Lane The Penguin Press, 1995), pp. 52-60.

119. Graham Cairns-Smith, Genetic Takeover, (Cambridge: Cambridge University Press, 1982).

120. Richard Dawkins, The Blind Watchmaker (Oxford: Oxford University Press, p. 94).

Return to homepage

Citation: John van Wyhe, ed. 2002-. The Complete Work of Charles Darwin Online. (

File last updated 25 September, 2022