BIOGRAPHICAL MEMOIRS

Herbert George Andrewartha 1907-1992

This memoir was originally published in Historical Records of Australian Science, vol.9, no.3, 1993.

Abstract

Herbert George Andrewartha was born in Perth on 21 December 1907, the second of three children of George and Elsie Andrewartha. His father was a primary-school teacher, later a headmaster. The family moved often to country towns in Western Australia where his father was posted, but they maintained a base in a small farm at Gosnells, about 40 km from Perth. At the end of his primary schooling, Andrewartha was awarded a scholarship to Perth Modern School. He took up residence at the Gosnells home and lived alone. He travelled 5 km on horseback and 40 km by train each day to school. After several years, when his father retired to work the farm, his parents and two sisters returned. Andrewartha remained closely attached to his family throughout his life. His younger sister, Bon, died soon after marrying but his elder sister, Ethel, survived him. He was always very proud of her self-taught musical ability, taking an interest in the choral groups she led and for which she composed. He himself was tone deaf!

On completing his secondary schooling in 1924, Andrewartha was awarded a cadetship with the Western Australian Department of Agriculture and enrolled at the University of Western Australia, where in due course he took his Bachelor's degree in Agriculture. During his secondary schooling, and at University, he took little part in the social or sporting life because of the long journeys each day, but he played cricket and football for the local Gosnells teams, and tennis on a court made of clay from termite mounds - perhaps this led to his interest in insects. In any event, he continued to play tennis on his own grass court until he was laid low by a stroke a few years after his retirement in 1972.

After graduation, Andrewartha began work as Assistant Entomologist in the Department of Agriculture under the supervision of L.J. Newman, who had recently come to Western Australia from the Burnley School of Horticulture in Melbourne. He commenced a study of a weevil, Otiorrhynchus cribricollis, that infested fruit trees and this established his interest in the detailed analysis of life-cycles and behaviour of insects and in the phenomenon of 'diapause', in which an individual ceases morphological and reproductive development for a period and usually becomes highly resistant to inclement weather.

In 1933, Andrewartha took up a post in Melbourne with the CSIR as Assistant Research Officer and became part of a group of biologists studying the biology of the apple thrips, Thrips imaginis, which at that time was devastating apple crops in southern Australia. He worked in the School of Agriculture and Forestry at the University of Melbourne under the direction of Professor (later Sir) Samuel Wadham and was able to complete a thesis for which he was awarded the degree of Master of Agricultural Science. Through this work he met, worked with and married Hattie Vevers Steele, another young biologist.

In 1935, Andrewartha and his new wife moved to Adelaide, to the Waite Agricultural Research Institute, to join Professor James Davidson who was himself engaged in studies on T. imaginis. Andrewartha took some part in these studies, as did his wife for a time, but his main duties were the study of the plague grasshopper, Austroicetes cruciata, which was destroying field crops in South Australia and Western Australia in those days.

The plague grasshopper lived in the marginal farming areas of South Australia, lands where the mean annual rainfall was only just or not quite sufficient for growing wheat and the variability was great. Thus drought was frequent, and during the Great Depression much of the land had become badly eroded due to excessive cropping. Studies on the grasshopper entailed long journeys, often over very rough country, in a specially equipped utility truck – Andrewartha always carried two lengths of roof guttering-iron to help him negotiate landhills! His farm background and his innate ability to improvise enabled him to negotiate these trips without serious mishap, with his wife as diary-keeper and cook. However, on one occasion, when they did not arrive back in Adelaide as expected, Davidson asked the Australian Broadcasting Commission to send out a request for their whereabouts, which they did during a broadcast of a test-match, much to Andrewartha's chagrin! The broadcast stated that Andrewartha and his wife were last seen at Mount Hopeless!

In 1946 Professor Davidson died suddenly leaving unanalysed fifteen years of records of the numbers of thrips. Davidson and Andrewartha had discussed these data in detail over the years and had consulted E.A. Cornish of CSIR's Mathematical Statistics Section on the appropriate statistical analysis. So it was that Andrewartha was in a position to direct the very laborious statistical procedures, to appraise the results and to write the two papers that appeared under their joint names. These papers were to cause much controversy and were to be the forerunners of Andrewartha's general theory of population ecology.

At this time, Andrewartha made his first journey overseas, visiting England and the United States, where he met many of the influential scientists working with insects and in ecology. On his return, he wrote a review of the importance of diapause in enabling insects both to synchronise their life-cycles with the changing seasons and to persist in adverse conditions, and thus of its ecological significance. He also put forward a theory of the physiological nature of diapause but, unlike most of his work, this was based on inadequate data and did not stand up to critical analysis - a cautionary tale of entering into fields where one is insufficiently informed. We see similar instances when people, distinguished in other fields, pronounce on matters ecological.

Davidson and Andrewartha had been deeply engaged in an attempt to account for the often violent fluctuations in the numbers of individuals in populations. After Davidson's death, Andrewartha joined forces with his former graduate student, L.C. Birch, in a determined effort to construct a theory that would illuminate and unify the multifarious studies that had been published. The result was The Distribution and Abundance of Animals, which appeared in 1954. It was this book that established their reputations in the vanguard of population ecologists.

Soon after the publication of this book, Andrewartha moved from the Waite Institute to the Zoology Department of the University of Adelaide to become Reader-in-Charge of a small Unit of Animal Ecology, set up by Professor W.P. Rogers with financial support from CSIRO. Although Andrewartha while at the Waite Institute had taken a close interest in the work of the younger members of staff, and of the few postgraduate students in the Entomology Department, and had done some stints of lecturing, especially in statistics, he had not been in a position actively to supervise students nor to develop a course in ecology. So, on his move to Zoology, he began to set up a final-year unit in Animal Ecology and to gather round him a group of honours and postgraduate students. The undergraduate work was a considerable challenge because it is fair to say that no course in experimental ecology existed at that time, and Andrewartha was determined to introduce his students to the study of the distribution and abundance of animals through laboratory and field experiments that could be achieved within the restrictions of undergraduate time-tables. He also set out to design each exercise to be amenable to an appropriate statistical analysis, something with which zoology departments, at that time, were not at all familiar. His lectures were, of course, based on The Distribution and Abundance of Animals, but this was much too formidable a book for undergraduates, and contained a good deal of material that he did not consider relevant to his course. The result was the publication in 1961 of his Introduction to the Study of Animal Populations, a book containing a section on the theoretical treatment of populations and another consisting of a series of experiments, complete with representative data (much of which had been collected in his classes) and the detailed appropriate statistical analyses set out and explained. This course continued to attract large numbers of students until his retirement.

In 1962, Professor Rogers resigned from the Chair of Zoology and Andrewartha was appointed in his place. This was near the beginning of the great increase in funds available for postgraduate study, and so it was that during his tenure of the Chair, a large and flourishing school of research students and postdoctoral workers was built up. His agricultural background and his strong belief that good fundamental studies could always be done using animals of economic importance, either beneficial or detrimental, led most of his students into studies on abundant animals of importance in agriculture, medicine or households. A steady stream of PhD's emerged from the Department to go out into academia, industry or government research with thorough backgrounds in the theory and practice of experimental ecology, on which ideas like sustainable agriculture and forestry and conservation of natural resources depend.

The Distribution and Abundance of Animals had emphasised the importance, in trying to elucidate the causes of fluctuations in numbers, of a knowledge of certain aspects of the physiology of the animal being studied, and also of its behaviour. The Department of Zoology was fortunate in having strong groups in the early days in parasite physiology and biochemistry and in reproductive physiology, which helped broaden Andrewartha's understanding of these topics, to the benefit of his ability as a supervisor, and of his own thinking.

Andrewartha was a strong personality with strong convictions. He had very definite views about the supremacy of science as a way of thinking. This coloured much of his conversation with students and colleagues. The big difference of opinion he had with the so-called density-dependent school of regulation of animal numbers was due, he felt, to a misunderstanding on their part of the role of induction and deduction in science. He never accepted the premises that the density-dependent school regarded as fact. This misunderstanding was so important that he included a substantial statement on induction, deduction, hypothesis formation and theory in the book, The Ecological Web, that he later wrote with Birch.

Because of his strongly stated views, some people regarded him as somewhat abrasive or even pig-headed. Others found his position a wonderful anvil on which to hammer out their own views. The climax of the dispute between the density-dependent school and supporters of Andrewartha's views came at the 22nd Cold Spring Harbor Symposium in 1957 on 'Population Studies: Animal Ecology and Demography'. The symposium was largely organized by Theodosius Dobzhansky for the purpose of reconciling the opposing factions. The public dispute between Andrewartha and A.J. Nicholson became highly charged and personal, much to the dismay of the Americans in particular. Dobzhansky went away more confused than ever about these argumentative Australians.

Andrewartha's greatest influence was with his students, particularly his graduate students. They quickly warmed to his sincerity, humour and kind attention. Those whom he set on their life's path remained steadfast colleagues of his for the rest of their careers. They felt they had a mission to fulfil in making better understood the ecological thinking that had been so important for themselves. There was also a substantial group of ecologists, particularly in the Netherlands and the USA, who found in Andrewartha's theories concepts that guided their own research. Professor Daniel Simberloff, in writing the citation for the eminent Ecologist Award of the Ecological Society of America for 1988 to Andrewartha and Birch, said:

Both men have worked specially with insect populations, but their insights have informed our field to the extent that the "Andrewartha-Birch school" connotes a widely recognized viewpoint and suggests a distinctive research protocol. The Distribution and Abundance of Animals was the landmark synthesis of field population ecology that inspired a generation widely credited with constructing modern ecology...Their field and laboratory efforts were the starting points for a more general theory of population ecology that culminated in The Ecological Web...Andrewartha and Birch have been consistent skeptics, continually confronting fashionable models with hard-won field data on specific organisms...Their skepticism implies neither hostility to theory nor failure of their own viewpoint to evolve...Largely because of their books, Australian systems and Australian ecological research are part of the common vocabulary of ecologists throughout the world.

Andrewartha was ably supported by his wife Vevers until she died some years before his death. She was an entomologist and for some years worked on the embryology of the grasshopper Austroicetes cruciata when Andrewartha had this species as his main ecological project. She was always the gracious hostess at the many evening discussions of Honours and graduate students at their home in Netherby, near the Waite Institute. Together they worked long hours in their very large garden and looking after their grass tennis court, which was usually the venue of tennis games for friends throughout the tennis season.

Long before automatic garden watering systems had been invented, Andrewartha invented his own for their large garden. It was very much a 'Heath-Robinson' sort of contraption but it worked. Andrewartha was skilled at inventing all sorts of contraptions for the laboratory and field work as well. Scarcely any bit of laboratory equipment would come into the laboratory that 'Andy' did not alter and improve in some way. He designed the fine set of temperature controlled cabinets in his laboratory. On one occasion, when he was visiting Thomas Park at the University of Chicago, the Parks' toilet flushing system stopped working. Much to Tom's surprise, Andy got to work. It was going again in no time. 'Do all Australians do these sorts of things?' asked the somewhat abashed American host.

Especially in his younger days in Adelaide Andrewartha took a close interest in left-wing political matters. He was also active in the professional organizations associated with science and agriculture. He was a committee member of the Federation of Scientific and Technical Workers, a group interested in the working conditions of scientists and technicians, and he was active in and at one time president of the local branch of the Australian Institute of Agricultural Science. His presidential address to the branch was a polemic about the laissez-faire system of farming, which he thought had been largely responsible for the devastating soil erosion that had occurred during the droughts of the 1930s. Later he resigned from the Society because he disagreed with a new policy of advocating studies in Agriculture as a profession, when the Society took no responsibility for the future employment of those who were influenced by its advocacy.

He was active in the conservation movement in its early days, being elected president of the Nature Conservation Society of South Australia, and was, for six years, chairman of the National Parks and Wildlife Advisory Council of South Australia.

Andrewartha shared (with L.C. Birch) the David Syme Prize of the University of Melbourne, the Clarke Medal of the Royal Society of New South Wales and the Verco Medal of the Royal Society of South Australia. He was elected a Fellow of the Australian Academy of Science in 1961. In 1987 he was awarded the Gold Medal of the Australian Ecological Society and in the following year was named (together with L.C. Birch) 'Eminent Ecologist of the Year' by the Ecological Society of America.

Andrewartha suffered a serious stroke in 1975 which left his left arm paralysed. However, with great fortitude and patience and helped by his inventive genius, he largely overcame the disabilities of this stroke and wrote his last book with Birch while still recovering. He died on 27 January 1992 at the age of 84, after a long illness that followed a broken knee sustained in a fall at his home. He is survived by his son Graeme and daughter Susan Dutch.

Scientific research

Andrewartha's first work was a study of aestivation of adults of the beetle Otiorrhynchus cribricollis in Western Australia. This became part of a later interest in the dormant phases of stages in the life-histories of insects. With CSIR in Melbourne he began work on the ecology of Thrips imaginis, a pest of some fruits. This work was to continue with Professor James Davidson as one of a number of projects when he moved to the Waite Institute in Adelaide. However, his first major research programme at the Waite Institute was on the population ecology of the plague grasshopper, Austroicetes cruciata, on which he published some dozen papers. In the 1960s he developed a programme in conjunction with J. Munro and N.L. Richardson on the use of sterile males to control populations of the Queensland fruitfly Dacus tryoni, which was a possible threat to fruit growing in the southern States if it became established there.

Two general principles motivated his ecological research. In the first place he was dedicated to doing something useful as well as adding to basic scientific information about the ecology of animals. Curiosity alone was not a sufficient reason for him to embark on a research programme. All the animals he worked with were of agricultural importance. He was convinced that success in helping the 'useful arts', as he called them, depended upon the quality of basic scientific research. The useful arts that depend on the science of ecology include the control of pests, the conservation of wildlife and the management of game. All these activities have a common goal, to control or to manipulate the reproduction and the life-expectancy of the animals in a natural population so that the density of the population remains between certain pre-determined limits.

Secondly, all his research was to contribute to a general theory of population ecology. Population ecology was a term coined by Thomas Park in Chicago. Andrewartha preferred to say that he was working on the principles that governed the distribution and abundance of animals. This phase comes directly from Charles Elton's classic Animal Ecology published in 1927. It became the title of his first book (with L.C. Birch) in 1954. Whether he used the phrase population ecology or the distribution and abundance of animals, Andrewartha made a clear distinction between the study of the ecology of a single species and community ecology. He was never persuaded that community ecology would contribute substantially to the science of ecology.

As he pursued his interests on the principles governing the distribution and abundance of animals, he was to find that in general the best work was done on animals that were either pests or were animals managed as game. This is borne out in the preponderance of examples from these areas in the book The Distribution and Abundance of Animals. It was simply true that the best work was in these areas. Much of it was done in Australia.

Andrewartha's long-term study on the ecology of Thrips imaginis was done in collaboration with Professor James Davidson. Davidson died in 1945, but their major findings on Thrips were published in two classical papers in the Journal of Animal Ecology in 1948. These were to be the focus of much debate amongst ecologists for years to come. The work on both Austroicetes and Thrips led Andrewartha and Davidson to develop a theory of population ecology that was at variance with the conventional wisdom of the day. Following Davidson's death, Andrewartha collaborated with a former student of his, L.C. Birch (who was no longer in Adelaide but at the University of Sydney), in further developing a theory of population ecology. This theory, together with supporting data and arguments, was published in two books jointly authored by Andrewartha and Birch, The Distribution and Abundance of Animals in 1954 and The Ecological Web in 1984.

The ecology of Thrips imaginis

The basic data for the ecology of Thrips was a daily estimate of their numbers in the rose garden of the Director's residence at the Waite Institute for fourteen years, together with daily meteorological data recorded in the meteorological station at the Waite Institute, just by the rose garden. Fourteen years is an unusually long time for the study of any population of animals, but such a long time turned out to be critically important in the analysis of the data. The method of partial regression was used to analyse the data in attempting to explain what were the main components of the environment that could account for the day-to-day and season-to-season changes. This was the first time partial regression had been used in population ecology. The regressions included the following independent variates: rainfall and temperature for each of the three days preceding the sampling of the population, the maximum temperature and rainfall for the day immediately before the day the population was sampled. These independent variates were chosen on the basis of what was known about the biology of Thrips. Food was not included in the analysis as it was never in short supply.

Some 78 per cent of the variance was accounted for by four quantities that were calculated entirely from meteorological records. This left virtually no chance of finding any other systematic cause for variation, since 22 per cent is quite a small residuum to be left as due to random sampling errors. All the variation was attributed to causes unrelated to the density of Thrips. Not only did Andrewartha find no 'density-dependent factor', but he claimed there was no room for one. This was the aspect of the analysis that created much discussion and controversy, since the dominant school of population ecology claimed that the numbers in natural populations can be regulated only by density-dependent factors. In their absence, the numbers were supposed to go on increasing without limit or to decrease to extinction. Neither of these things happen in the case of Thrips, since the seasons subject Thrips to a period that is favourable for increase in numbers to be followed by a season that is very unfavourable. One important reason why Thrips does not become extinct in the area studied is the heterogeneity of the places where it may live. It may become extinct in one locality only for this to be colonized again from a locality where it has not become extinct.

Andrewartha's conclusions about Thrips went against the conventional wisdom that dominated population ecology at the time. The two conclusions that the numbers of Thrips can be explained by the succession of good and bad seasons and by the heterogeneity of places where they lived, became central to the general theory of population ecology that Andrewartha was developing.

The ecology of Austroicetes cruciata

The ecology of this plague grasshopper was studied by Andrewartha, from 1935 to 1942 together with James Davidson, and with L.C. Birch. The life-cycle is characterized by an intense obligate diapause or state of dormancy in the egg stage. Consequently there is only one generation each year. Since diapause in the egg was critical to understanding the ecology of the grasshopper, Andrewartha studied the influences that determined the stage of onset of diapause and the influences that ended diapause. Soon after being laid, the embryos inside the egg developed up to an early stage, after which obligate diapause ensued. The embryo remained in that state during the winter. A particular sequence of cold temperature was responsible for the resumption of development at the end of the winter.

In this study, the numbers of grasshoppers were not determined precisely as was the case with Thrips. Qualitative or subjective estimates were made. It became clear quite early in the study that dryness during the spring was critical in determining the number of grasshoppers that survived in any year. In some seasons the drought was so severe that grasshopper populations crashed and only a succession of good years to follow enabled numbers to become abundant again. During drought years the stocks of food ran out, not because they were eaten by grasshoppers but because drought made the grass useless as food for grasshoppers. The supply of food was unrelated to the number of grasshoppers.

The rate of increase of grasshoppers was not determined by density-dependent factors, yet their numbers did not go on increasing to the limit of their resources of food. Calamity in the form of drought overtook the grasshoppers before this could happen. This sort of calamity occurred with a certain frequency that was calculated from the meteorological records. Grasshoppers became extinct in many localities during a dry season. However, some always survived in more favourable places. It was from these places that new colonies could be set up in subsequent years. Grasshoppers never became extinct over the whole range of their distribution. The explanation of numbers of Austroicetes was strikingly similar to the case with Thrips, and once again it challenged conventional wisdom about what determined the numbers of an animal.

Austroicetes was distributed in a belt of country that had been cleared of its native vegetation of plants such as saltbushes, which are not food for grasshoppers. However, the grasses that took the place of the original vegetation were suitable food for grasshoppers. Much of this country was marginal wheat country, and much of it had been abandoned for growing wheat. There was then a simple ecological solution to getting rid of grasshoppers, and that was to return the country to its original saltbushes. However, that turned out to be unacceptable politically and perhaps economically.

Diapause

Andrewartha's study of diapause in the eggs of Austroicetes led him to an intensive study of diapause in general amongst insects, as is indicated by the long chapter on the subject in The Distribution and Abundance of Animals. His study was the most complete that had ever been made up to that time. An important contribution of this study was his discovery that diapause in the egg stage characterized the life-history of plague grasshoppers around the world. These are grasshoppers that do not migrate. They cope with the hostile season by going into diapause. Locusts on the other hand were thought not to have a diapause at any stage and to cope with the hostile season by migrating huge distances. Andrewartha was able to study this on the spot in South Australia where the locust Chortoiocetes terminifera had its 'outbreak' areas in the northern desert of the state, from which the population at long intervals spilled over in huge migratory swarms to the southern areas of Australia. Most of the locusts did not survive long in these places, but some did. Andrewartha reasoned that eventually some of their descendents managed to get back to the dry outbreak areas where populations were permanent. This helped them to be permanent inhabitants of these places. Andrewartha assembled a great deal of data indicating that with plague grasshoppers, diapause in the egg enabled them to survive in the inhospitable season, and that with locusts, migration from inhospitable areas enabled them to remain extant. More recent work has shown this concept not to be as clear-cut as Andrewartha thought, though the two mechanisms of survival are well recognized.

The Queensland fruitfly Dacus tryoni

In most parts of Australia, with the exception of the central and western deserts and southern Tasmania, the climate is hospitable for the Queensland fruitfly. Its distribution, however, is much more restricted. This is probably due to the absence of a succession of fruits. Before fruit trees were cultivated in Australia, Dacus was restricted to tropical and sub-tropical forests. With the cultivation of fruit trees, it spread south as far as eastern Victoria. Occasionally, infested fruit has been found on trees further west and around Adelaide. In temperate areas, the chance of a female finding a mate becomes quite small during two periods. The over-wintering population of adults is quite small and is widely dispersed in a few sheltered localities. In early spring, the first generation that arises from the over-wintering adults is also small and dispersed. Moreover, most of the newly emerged adults leave the locality in which they originated.

Because of these facts, Andrewartha surmised that early spring would be a strategic time to liberate sterile flies with the objective of flooding the population with sterile males and so reducing the chance of females finding fertile males, a technique for the control of populations that had recently been proposed, and tested, by E.F. Knipling of the United States Department of Agriculture. This appealed to Andrewartha as a possible method of control, particularly in areas that might become newly colonized by the insect such as Victoria and South Australia.

Andrewartha, J. Munro and N.L. Richardson worked out a strategy for the control of fruitflies by releasing huge numbers of sterile males. Many millions of flies were bred in Sydney and pilot experiments were done in towns in western New South Wales. The work had to be terminated before it could be demonstrated that the method was economically feasible, though initial results were encouraging. The work is important as it was the first attempt to use this technique in Australia and showed its feasibility, if sufficient funds could be found to breed and release the huge numbers of insects needed.

General theory of the distribution and abundance of animals

The general theory of the distribution and abundance of animals was based on the original field studies of Andrewartha, Davidson and Birch, together with the analysis of many case-histories of the ecology of animals, many of them from Australia. In the course of these investigations, Andrewartha spent much time working through the data of other investigators who had accumulated a great deal of information but had never put it together in the form of a general theory. This applied, for example, to the work in the CSIRO on the European rabbit in Australia, on waterfowl and on the Australian magpie. In working on these case-stories and many others, Andrewartha not only made the work in question more accessible to ecologists but put it together in the context of a general theory. The results of this work were published in the two books with L.C. Birch, The Distribution and Abundance of Animals and The Ecological Web.

The central proposition of the theory is that the numbers of an animal (and therefore its distribution and abundance) depend upon its chance to survive and reproduce, which in turn depends on the animal's environment. It follows that the ecologist needs to have precise ways of defining and measuring the chance to survive and the chance to reproduce, and secondly to have a precise definition of environment and its components. The precise meaning of and measurement of the chance to survive and reproduce were established in The Distribution and Abundance of Animals. The meaning of environment was developed in the same book but became much more refined thirty years later in The Ecological Web. Indeed, this latter work presented the results of many years of discussion between Andrewartha, Birch, T.O. Browning and B.S. Niven. It was a unique contribution and basically different from what most ecologists had been working with.

The problem was how to define environment and how to split it up into components that did not overlap. The environment is not just anything that happens to be around the animal. It is anything that influences an individual's chance to survive and/or to reproduce. The word 'individual' is critical. In this theory there is no such thing as the environment of a population, for every individual may exert an influence on every other individual. The concept of the environment of the individual contrasts with the then current concept of the environment of the population. The importance of defining environment in terms of the individual took a long time to catch on and is still rejected by many ecologists. In these deliberations Andrewartha aimed to discover a precise and completely unambiguous way of defining each component of environment. This had never been done before. Terms such as 'biotic' and 'physical' were hopelessly imprecise. It was for this reason that he welcomed the contribution of Susan Niven, who used the symbolism of formal logic for this purpose. Her somewhat complex formal procedure was included as an Appendix in The Ecological Web.

The definition of the components of environment is summarised in Table 1.01 in The Ecological Web. This is a 2´2 table with four compartments depending on (a) the reaction, either negative or positive, of the component to the animal and (b) the response of the animal, either negative or positive, to the component. This gives four 'directly acting' components: resources, mates, malentities and predators. Anything that influences the activity of the directly-acting components is not in this centrum of directly acting components but is part of a web of influences extending outwards from the centrum. 'Malentity' is a rather new idea, and 'resource' has a more restricted meaning than is general in ecology.

The Ecological Web introduced the idea of the 'envirogram'. This is a diagram for the animal that shows the directly-acting components as the proximate causes of the condition of the animal, and the distal components as a web extending outwards from the directly-acting components. It shows any interactions there may be between components, and is basically simple to look at. This contrasts with the complicated flow diagrams that ecologists had up to then constructed.

The definition of environment, the analysis of the components and the envirogram were the subject of enthusiastic acceptance by some ecologists, but the object of much criticism by others. One problem was that you had to know quite a lot about your animal before you could correctly decide which were the four directly-acting components and which were the indirectly acting ones. Also, to get there required a lot of careful thought. While this turned out to be stimulating for some students, it proved too strenuous for others. Some critics so misunderstood the meaning of the envirogram that they incorrectly said it failed to show interactions. It is too early to say to what extent this analysis will be accepted by population ecologists, though there are some fine examples of its application by others, particularly in the ecology of birds, and Susan Niven has used it in her studies of a wide range of invertebrates and vertebrates.

A second main plank of the general theory is that populations are multipartite, that is to say the population is subdivided into local populations where environments differ widely. Moreover, each local population may be genetically different. The extent of dispersal from one local population to another is very important in understanding the dynamics of the whole process and its outcome. In The Ecological Web the term used for this was 'spreading the risk', which was proposed by P.J. den Boer of the Netherlands to indicate the way in which heterogeneity between local populations could confer stability and persistence in the natural population. The concept of the multipartite population and spreading of risk was illustrated by diagrams in both The Distribution and Abundance of Animals and The Ecological Web, each diagram representing a different way in which spreading of risk operates.

Andrewartha considered that the general theory of population ecology set out in these two books replaced the conventional wisdom that populations have to be regulated by density-dependent factors such as competition. He was happy to use the word competition when one genotype was shown to be superior to another, but in other contexts he felt the word confused or obscured what was happening. This again was quite a radical idea that was in opposition to conventional ecological thinking.

The general theory of population ecology that Andrewartha developed with his colleagues held that any component of environment could play a part in determining numbers, preventing numbers from increasing indefinitely and preventing extinction of the natural population as a whole. Each component of environment could be assigned a probability, none was completely deterministic. This was a probabilistic as contrasted with a deterministic approach. It was clearly in conflict with the widely held theory that populations could only be controlled by density-dependent factors and represented by deterministic models.

Unfortunately, the theory was widely misconstrued by some of its critics who supposed that it claimed that so-called density-dependent factors played no part in the regulation of animal numbers. On the contrary, according to the theory these took their place with all other components, as should have been clear from the examples given in which predators were critical in determining numbers. The arguments that ensued resulted in Andrewartha writing quite a lot about the role of induction, deduction and theoretical models in scientific method (e.g. pages 187-191 in The Ecological Web).

Andrewartha was convinced that the study of community ecology had not contributed to understanding the causes of the distribution and abundance of animals. He hoped that population ecology (the study of single species) would eventually displace studies in community ecology.

Andrewartha held that the first requirement of an ecologist was to be a good naturalist. You have to know something about your organisms before you can start to make theories about them. He was not opposed to theory but he was opposed to theories that made gross abstractions from the real world and seemed to depend little if at all on data from the field, going only to the field or the literature to find results that seemed to conform to the conclusions of the theoretical model. In his opinion the theoretical models of his day left out of account the idea of the multipartite population and the heterogeneity of the environment. This made them useless except for theoretical exercises in mathematics.

Andrewartha's contribution to population ecology was largely spread through the two books written with Birch. The Distribution and Abundance of Animals was widely used in ecology courses, particularly in the USA, and influenced a whole generation of ecologists. The influence of The Ecological Web has yet to be assessed. Andrewartha paid great tribute to Charles Elton, who put animal ecology on the map in the United Kingdom, particularly for his workmanlike approach to what ecology was about. The two differed greatly on the value of community ecology. Andrewartha had great respect for the work directed by Harry Smith at the Citrus Experiment Station at Riverside in California and used many of these studies in his writings. He was influenced in later years by the work of den Boer and his students in the Netherlands.

There is a sense in which Andrewartha's contribution to population ecology was ahead of its time. The conventional wisdom was against it. The sudden efflorescence of highly abstract theoretical models based on the ability of a computer to follow complex interactions tended to engulf the literature in the '70s and '80s, and biosphere studies emphasised community ecology until its practitioners realized that this did not answer the questions that were being asked. Through all this, there is beginning to emerge a new valuation of the long-term field study and the necessity of studying the multipartite population. One day this approach may become more dominant than it is at present. When that happens, we can thank Andrewartha for setting the direction with such skill and persistence.