Understanding Food Chains and Food Webs, 1700–1970


Some ecological ideas developed gradually and only gained coherence and details after they had become commonplace. The history of two interrelated ideas, food chains and food webs, is an example of a gradual, cumulative history. Here is a brief survey of these concepts from about 1700 to 1970 (Fig. 1).

Figure 1.

The earliest identified food chains seem to have concerned hyper-parasitism (Egerton 2005, 2006a), which students of insects discovered in the later 1600s. But it was entrepreneurial naturalist Richard Bradley (Egerton 2006b) who generalized the concept (Bradley 1718, part 3:60–61):

Insects which prey upon others are not without some others of lesser Rank to feed upon them likewise, and so to Infinity; for that there are Beings subsisting, which are not commonly visible may be easily demonstrated…in a Microscope.

This was turned into verse by Jonathan Swift (Fig. 2) in 1733 (lines 341–344).

Figure 2.

So, Nat'ralists observe, a Flea Hath smaller Fleas that on him prey, And these have smaller yet to bite ‘em, And so proceed ad infinitum.

In Bradley's account, food chains illustrated the balance of nature (Egerton 1973:333–335). Swift was a prominent literary figure who had a general interest in science, but his lines on fleas were meant as a swipe at lesser poets. Neither Bradley nor Swift provided an illustration, so we can help them out with this one from Alfred Elliott's 1957 Zoology textbook (Fig. 3), though Elliott borrowed the idea for the central figures of fleas from Robert Hegner's 1938 book, Big Fleas Have Little Fleas, or Who's Who Among the Protozoa.

Figure 3.

Carl Linnaeus, in an ecologically important essay, “The Economy of Nature” (Linnaeus [Latin] 1749; [English] 1775:114, 1977), briefly itemized the stages of two food chains, one terrestrial, one aquatic (Egerton 2007a). There are likely other naturalists between Linnaeus and Darwin who reported on food chains, but attracted little notice. Darwin may be the first to report a food web, occasioned by the Beagle's stopover at the rather barren island of St. Paul on 16 February 1832. He gave its location as 0° 58′ north latitude and 29° 15' west longitude, and 540 miles from America. He found only two species of birds, the booby (a gannet) and the Noddy Tern. The latter built a simple nest with seaweed. Then follows his food web (Darwin 1839:10):

By the side of many of these nests a small flying-fish was placed; which, I suppose, had been brought by the male bird for its partner…quickly a large and active crab (Craspus), which inhabits the crevices of the rock, stole the fish from the side of the nest, as soon as we had disturbed the birds. Not a single plant, not even a lichen, grows on this island; yet it is inhabited by several insects and spiders. The following list completes, I believe, the terrestrial fauna: a species of Feronia and an acarus, which must have come here as parasites on the birds; a small brown moth, belonging to a genus that feeds on feathers; a staphylinus (Quedius) and a woodlouse from beneath the dung; and lastly, numerous spiders, which I suppose prey on these small attendants on, and scavengers of the waterfowl.

After reading this account, Rear-Admiral William Symonds told Darwin that he had seen at St. Paul crabs drag young birds from nests and eat them. Darwin added his information to this passage in the second edition (1845) of his book on the voyage of the Beagle (Edwards 1985:34).

In The Origin of Species (1859:73–74), Darwin reported the most famous example of a food chain in the scientific literature (Fig. 4). It is in chapter 3 on the “Struggle for existence,” and involves humble bees (called “bumble bees” in America) pollinating red clover; though some bees were eaten by field mice, the mice, in turn, were kept in check by domestic cats. Darwin speculated that if it were not for the cats, the mice would decimate the bees, and the clover would go unpollinated, since only humble bees pollinate clover. A later, unknown commentator extended this chain further (Milne and Milne 1966:6) by suggesting that old maids commonly kept cats, that clover-fed cattle were eaten by British seamen who protected the British Empire, and that if it were not for old maids, the British Empire would fall! In other words, Darwin's food chain became a biological version of Englishman George Herbert's well-known admonition (1640):

Figure 4.

For want of a nail the shoe is lost,

For want of a shoe the horse is lost,

For want of a horse the rider is lost.

In America, this admonition is attributed to Ben Franklin, who borrowed it without acknowledgement for Poor Richard's Almanac (1757). But getting back to Darwin's food chain, in 1947 W. L. McAtee pointed out that Darwin's food chain dynamic lacks full validity, since we now know that honey bees also pollinate red clover and that humble bees often appropriate mouse holes, so humble bees and mice have an ambiguous relationship. In Darwin's defense, he heavily depended on H. W. Newman's 1850–1851 study “On the habits of the Bombinatrices” (Darwin 1975:183).

The next discussion of note for our purposes is from a remarkable German zoologist, Karl Semper. In 1877, he gave 12 lectures at the Lowell Institute in Boston, published simultaneously in English and German editions in 1881. The English title is Animal Life as Affected by the Natural Conditions of Existence. This book was the first detailed synthesis of animal ecology. In a discussion of the food of herbivores and carnivores (Semper 1881:51–52), he pointed out that when herbivores transform vegetation into flesh, there is a loss of mass due to oxidation of organic material, and that the same is true when carnivores transform the flesh of their prey into their own flesh. To illustrate this, he arbitrarily assumed a 10 to 1 ratio of food to flesh. One thousand units of plant food could only support 100 units of a herbivore, and those 100 units of herbivore could only support 10 units of a carnivore. Although his book has 106 illustrations, this generalized food chain was not illustrated. However W. E. Pequegnat's diagram (Fig. 5) from Scientific American (1958:86) captures Semper's concept, even to the point of using a 10 to 1 ratio. Semper wrote at a time when there was little quantified thinking in natural history. He had first trained as an engineer and then as a physiologist (Mayr 1975), and that background came to the fore in this discussion. Although his book was widely read, apparently no one carried this line of quantitative thinking any further in the 1880s or 1890s.

Figure 5.

We are used to seeing food chains or webs diagrammed. The advantages are obvious: they provide a visual panorama of detailed information. The early history of such diagrams is elusive. The bibliography on food chains and webs that Allee, Emerson, Park, Park, and Schmidt compiled (1949:514) can assist in the search. However, they did not discover the earliest ones now known, published in 1880 by Lorenzo Camerano, which are reprinted in an English translation of his article (1994:377–378). Since Camerano's two diagrams do not resemble any known from later zoologists, it seems likely that he did not have much, if any, influence on later diagrams. Joel Cohen (1994:353–355) suggests that Camerano was influenced by diagrams for other purposes in books by Darwin and by Hermann Helmholtz, though Camerano's diagrams do not resemble theirs. Like Semper's, Camerano's food webs are generalized rather than specific.

The earliest specific food web I have found (Fig. 6) is on “The boll weevil complex,” published in 1912 by Pierce, Cushman, and Hood in a USDA Bulletin. Their motive was to promote bowl weevil eradication—by encouraging its predators and parasites. Theirs may not have been the first specific diagram published, because others appeared about the same time in different biological specializations, where it is unlikely that the members of one specialization were reading the literature of other specializations.

Figure Figs. 6, 7, and 8.

Figure Figs. 6, 7, and 8.

The following year, University of Illinois animal ecologist Victor E. Shelford (Fig. 7; photo, Croker 1991) published Animal Communities in Temperate America as Illustrated in the Chicago Region, which contained diagrams of both aquatic (Fig. 8) and land food webs (Fig. 9). There is no reason to suspect that he was influenced by the boll weevil diagram of 1912. Shelford used both of his diagrams to show how the community tends toward equilibrium, although the terrestrial community was more complex than the aquatic community, and consequently its equilibrium was more precarious. Shelford became a leading American animal ecologist (Croker 1991); his book was reprinted in 1937 and 1977.

Figure 9.

The earliest known food web diagram for a marine community was drawn by Danish fishery biologist Johannes Petersen (Fig. 10) in “A preliminary result of the investigations on the valuation of the sea”1915. He studied the Kattegat region of shallow water between eastern Denmark and Sweden (Fig. 11), an area with maximum length of 150 miles and maximum width of 90 miles. Significantly, he attempted to establish the annual productivity for this region, and his diagram indicates the thousands of tons of each group of organisms, with both a number and a proportioned rectangle (Fig. 12). In the text he stated that the eel-grass (Zostera marina) figure of 24,000,000 tons represents only the amount produced in the summer, and that the annual production is twice that. Presumably, all the other figures are annual production and not just summer production. The tons of plaice and cod are the actual commercial catch of those fish from International Fishery Statistics for 1910, and that was possibly true also for the tons of herring given, though he did not say so. The numbers given for other animals seem to be estimates. Although he indicated on his diagram that herring fed on plankton, he thought plankton was much less important than Zostera as a foundation for this food web. He concluded that the Kattegat had a “very unfavourable proportion between producers and consumers”(Petersen 1915:32). What he meant by this seems to be indicated by the following sentence in which he stated that carp ponds have “even without artificial feeding, given a yield of fish per hectare several times greater than that of the Kattegat.” Petersen reproduced the same diagram with minor alterations in his final report, “The sea bottom and its production of fish-food” (1918:23). His colleague, H. Blegvad, also used rectangles in his diagram of “Food of fish and principal animals in Nyborg Fjord” (1916:24) (Fig. 13) but without attempting to represent precise quantities. However, he did give quantitative data in the text of his article, which provided some sense of the quantities of organisms involved at each level.

Figure Figs. 10 and 11.

Figure Figs. 10 and 11.

Figure 12.
Figure 13.

In the same year as Blegvad, the American zoologist Harold Sellers Colton published what Jonathan A. D. Fisher calls (2005:145) “possibly the first intertidal marine food web ever illustrated.” It is in Colton's article on a carnivorous snail, Thais lapillus (now Nucella lapillus), and shows both which animals the snail eats and which animals eat the snail (Fig. 14). Colton did not indicate what inspired his diagram, but his brief bibliography does include Shelford's book (1913). Fisher did not find references in the later relevant literature to Colton's two articles on this snail (probably due in part to Colton's leaving marine biology for archeology [Miller 1991]), so we do not know of any influence that his diagram exerted.

Figure 14.

Charles Elton (Fig. 15) helped make such diagrams commonplace. He went on an Oxford University Arctic expedition in 1921 to Spitsbergen and took along Shelford's book as a possible model for his own study (Elton 1966:33). However, Elton soon realized that the community he studied had a different dynamic than Shelford's aquatic and terrestrial ones. Elton was impressed by the transfer of food from sea to land, which is reflected in his diagram (Fig. 16) published in 1923. Although V. S. Summerhayes is listed as the senior author of their joint study, since he was a botanist, we can assume that Elton developed this diagram, in which plants are not emphasized. Two years later, in 1925, Elton published this much simpler Canadian food web (Fig. 17), which includes information on the lengths of animals. In 1924, English fishery biologist A. C. Hardy published a diagram (Fig. 18) on food consumed by herring at different stages of development. It bears no similarity to any diagrams previously shown, and it seems likely that he either was inspired by some unidentified example from the fisheries literature, or that he independently developed his diagram. Be that as it may, in 1927 Elton published his classic textbook, Animal Ecology, which reprinted and explained these last three diagrams by himself and Hardy. In that book Elton also introduced (1927:55) the terms “food chain” and “food cycle.” Widespread use of his book popularized the use of food web diagrams. In both Hardy's diagram and in Elton's for 1925, more information was conveyed than merely which animal ate which food. Hardy's additional information was on the age of herring in relation to food, and Elton's was on the size of the consumer in relation to food. Elton also popularized the idea of a food pyramid (1927:68–70), which concept had been implied by Semper.

Figure 15.
Figure 16.
Figure 17.
Figure 18.

In 1926 Germany's leading limnologist, August Thienemann (Fig. 19) published this unique food web of lakes (Fig. 20). His 50-page article on nutrient cycles in lakes introduced into limnology the terms “producers,” “consumers,” (though Petersen [1915], quoted above, had used both terms in marine biology) and “reducers.”

Figure Figs. 19 and 20.

Figure Figs. 19 and 20.

Thienemann's 1926 paper and two of his other papers influenced an American postdoctoral student, Raymond Lindeman, who produced one of the most influential diagrams in the history of ecology (Fig. 21), though few if any ecologists have published similar diagrams. It appeared in his posthumous paper, “The trophic–dynamic aspect of ecology” (1942). Like Thienemann's diagram, Lindeman's is a generalized food web, but both men had hard specific data backing up their concepts. In that respect theirs were similar to diagrams by Shelford, Elton, and Hardy, which illustrated specific food webs, and unlike Semper's generalized food web, which was an educated guess. In the caption to his diagram Lindeman indicated that it was similar to one he had published the previous year. A comparison of his two diagrams indicates what he learned in his year at Yale University working under Evelyn Hutchinson (Cook 1977). The 1941 diagram is identical to the 1942 diagram except it lacks the symbols for trophic levels along the side. Lindeman (1942:159) used Thienemann's terms “producers” and “consumers,” but suggested substituting the term “decomposers” for Thienemann's term, “reducers,” to signify that the indicated process was not just chemical, but also biological.

Figure 21.

In 1943, a year after Lindeman's 1942 diagram appeared in the journal Ecology, Harvard marine ecologist George Clarke published this conventional food web (Fig. 22), but three years later, after he had studied Lindeman's diagram and its explanation, Clarke published his diagram (Fig. 23) in Ecological Monographs, of a marine food web that emphasizes productivity and human removal of material. It also shows Clarke's concern for the rate of production at each trophic level.

Figure 22.
Figure 23.

The Odum brothers, Eugene and Howard Thomas, carried Lindeman's thinking further. The Atomic Energy Committee became interested in radiation ecology (Kwa 1989:48), and Eugene Odum (Fig. 24; photo, Craige 2001) developed a program at the University of Georgia to study food chains at the Savannah River Research Facility to trace radioactive pollution (Craige 2001). By injecting plant stems with radioactive phosphorus-32, he and his colleagues traced it up the food chain to leafhoppers, beetles, and spiders (Kwa 1989:58). About 1957 the programs at Oak Ridge and Savannah River converged, with both programs using radioactive tracers to measure the flow of materials up the food chain (Kwa 1989:66). In the second edition of Eugene Odum's famous textbook, Fundamentals of Ecology (1959:47), there is a 1949 diagram of a food chain (Fig. 25). When I saw it, I assumed that Lindeman's influence had flowed across the Atlantic in just a few years, but when I compared it with British ecologist Erichsen Jones' own diagram, I discovered what Odum meant when he wrote that his diagram was “redrawn” from the one by Jones: Odum added the labels to the left of the diagram as a pedagogical aid. Howard Thomas Odum (Fig. 26; photo from Katherine Ewel) received his graduate training under Hutchinson at Yale. (In 1954 he taught me freshman zoology at Duke.) In 1956, he produced a diagram (Fig. 27) of matter and energy flow, in steady-state flowing-water communities in Florida. At that point, the reader could still understand the diagram without special training. However, H. T. Odum continued developing his thinking along the lines of systems ecology and used symbols from electrical engineering. By 1971 he published esoteric diagrams (Fig. 28) that integrate humans into the biotic community. This was an important step towards founding several applied ecological sciences (Mitsch 1994, Hall 1995, Egerton 2007b).

Figure 24.
Figure 25.
Figure 26.
Figure 27.
Figure 28.

Other ecologists developed food chain and food web concepts in another direction. In 1948, D. E. Howell reported finding DDT in human fat, and by 1949 biologists were reporting that fish feeding on insects killed by DDT were also being killed (Hoffmann and Surber 1949, Langford 1949). Rachel Carson (Fig. 29) publicized the discovery of insecticides traveling up the food chain in ever-increasing concentrations in her best-selling book, Silent Spring (1962:110–111), as did Robert Rudd in his less-read book, Pesticides and the Living Landscape 1964. Carson did not provide diagrams, and the ones Rudd used were quite simple. Here are four (Fig. 30) of the seven diagrams in his book. DDT was the most notorious insecticide, and in 1967 George Woodwell published a diagram (Fig. 31) in Scientific American showing increased concentrations of DDT as it progressed up the food chain. By 1970, Clive Edwards constructed a much more detailed food web (Fig. 32), showing DDT pathway and concentrations from the time of spraying DDT into the air, all the way up the food chain until it became concentrated in predatory birds, mammals, and humans.

Figure 29.
Figure Figs. 30, 31, and 32.

Figure Figs. 30, 31, and 32.

From simple narratives around 1700, food chain and food web concepts have been developed into progressively more sophisticated vehicles for conveying ecological ideas (Polis et al. 2004, de Ruiter et al. 2005). Lorenzo Camerano's two 1880 diagrams of food webs had no known influence, but after the visual stimulus of diagrams became established in the early 1900s, many ecologists found creative ways to express visually their discoveries concerning food chains and webs.

Figure 1.


This is a revised version of a talk given at the ESA Annual Meeting in August 2006 in Memphis, Tennessee. For comments preceding the talk, I thank Robert P. McIntosh, Professor Emeritus of Biology, University of Notre Dame (now in Florida). For several references used in the revision, I thank Jonathan A. D. Fisher, Department of Biology, University of Pennsylvania, Philadelphia.