An Interview with Philip Grime

University of Sheffield's Philip Grime: Strategic Advances in Plant Ecology

"Plant ecology is multidimensional," says Philip Grime of the University of Sheffield, U.K. "In seeking 'strategies' we are looking for universals rather than concerning ourselves with the peculiar particulars of each species."

Formulating generalizations in ecology is particularly difficult, not only because of the complexity of the natural world but also because of the sheer variability of plants, animals, and microbes. This is a land without laws, but one with no shortage of hypotheses. Among those trying to unravel patterns in the ecological wasteland, whose wanderings have taken him over the uncharted territory that lies between physiology and community ecology, is Philip Grime of the University of Sheffield, England. Much of his work has concentrated upon the detailed study of individual species of plants, screening populations and mapping their precise requirements and tolerances. He has now accumulated an extensive database covering a broad range of species and has published many widely cited papers that set out novel ways of classifying plants according to their "strategies"--a controversial term that covers a syndrome of ecological characters fitting a species to a particular mode of life. Grime's 1979 book Plant Strategies and Vegetation Processes has been cited more than 1,200 times. Under Grime's influence the world of plant ecology has become somewhat triangular, with ruderals, competitors, and stress tolerators occupying strategic corners.

Having spent much time in analysis, the time has now come for synthesis--the fitting together of the individual pieces (species) to make a whole (ecosystem). Hence Grime is setting out along new (and still uncharted) tracks involving the building of communities and is seeking to sort out the rules that underlie their composition--what he likes to call the Periodic Table that ecology has lacked.

At Sheffield, Grime is currently Director of the Natural Environmental Research Council Unit of Comparative Plant Ecology. He is very much a home-grown product, having taken his first degree and his Ph.D. at the university, and having spent the bulk of his research career there, with the exception of a spell at the Connecticut Agricultural Experiment Station in the early 1960s.

The Unit he directs consists of about 30 members of academic and technical staff, graduate students, and visitors, and is perhaps best known internationally for its Integrated Screening Program, in which a suite of plant species is being tested under a considerable array of physiological growth conditions in the laboratory. Grime recently spoke with Science Watch's ecology correspondent, Peter Moore, in London.

What is distinctive about your contribution to ecology?

Grime: I have tried to maintain a balance between demography and ecophysiology--to keep the approach as broad as possible at a time in research when the tendency is for research to become narrower in many fields.

Does this mean that you are more interested in synthesis than in reductionism?

Grime: No. One of the reasons I have stayed in Sheffield a long time is to build up a team around me who have expertise in all areas of ecology, so that we can analyze in the greatest possible detail the nature and requirements of individual species or populations. This has been the approach of hard science.But once there is a sufficient database relating species to environment, based upon rigorous comparative laboratory studies, one can then develop a more synthetic approach and hopefully a more successfully predictive approach to vegetation and eventually to the ecosystem.

The use of ecological and physiological screening of individual species (the Integrated Screening Program) as an approach to the study of vegetation is one you have made your own at Sheffield. Where did you first get this idea?

Grime: It was Roy Clapham, in his presidential address to the British Ecological Society in the 1950s, who first put forward the idea of "screening" plant species, though he did not use the term at that time. The idea was to use the controlled conditions of the laboratory for the comparison of species in isolation from their field environment. But of course you have to choose which species to use in the first place. Our approach has been to ensure that we have common species that are representative of a wide range of vegetation types and have a wide spread of life histories. We currently use 43 test species. The process of selection thus begins in the field; testing proceeds in the laboratory; and the hope is that eventually we can make predictions about their behavior back in the field. The laboratory analyses should also allow us to predict the behavior of the synthetic ecosystem on the basis of its component parts (and this is an area in which both Sheffield and Imperial College London are involved). John Lawton at Imperial has gone inside, as it were, using controlled chambers in which he can observe his microcosms, while we have stayed outside and have constructed our experimental ecosystems outdoors in transparent, ventilated boxes, or even in open-field conditions.

Are you trying to make ecology into a hard science?

Grime: Yes, I am. I have long been saying that ecology lacks a Periodic Table. To continue the chemical metaphor, we have often been content to observe at the molecular level when we need to begin our studies at the atomic level. When we began this work, we looked at the information in the literature and found that there simply was not enough useful information about plant species. There was plenty about morphology and anatomy, but not about resource use and physiology. So the laboratory screening process was the next step. We set this up around 1986-7, having discussed just what things we wanted to know about plants and how we could measure them in a simple, repeatable way. We made some wrong choices, and we have had to add later some tests that we had not initially considered.
One question that concerned us was that of "trade-offs." Presumably the laws of physics and chemistry are placing some constraints upon adaptive radiation. If a species develops along one line it may not be able to develop along another. Gaining fitness in one area may involve sacrificing fitness in another. Are there, therefore, design constraints that are common to all organisms?

Highly Cited Papers by Philip Grime
Published Since 1977
(Ranked by total citations, with citations updated through July 1998)

Rank	Paper	Cites through 1995*	Cites through July 98
1	J.P. Grime, "Evidence for existence of three primary strategies in plants and its relevance to ecological and evolutionary theory," Am. Natur., 111(982):1169-94, 1977.	406	519
2	K. Thompson, J.P. Grime, "Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats," J. Ecol., 67( 3 ):893-921, 1979.	243	323
3	J.P. Grime, et al., "A comparative study of germination characteristics in a local flora," J. Ecol., 69( 3):1017-59 , 1981.	165	211
4	J.P. Grime, M.A. Mowforth, "Variation in genome size: an ecological interpretation," Nature, 299(5879):151-3, 1982.	66	83

SOURCE: ISI's Science Citation Report
*citations reported with original interview

This implies that, having looked at the individual species (the atoms, in your analogy), you then seek natural groupings--functional types?

Grime: Yes, this is so. But there is some misconception here. Even if we can detect functional types, and we believe we can, it does not imply that we can predict everything about their ecology. Plant ecology is multidimensional, and in seeking "strategies" we are looking for universals rather than concerning ourselves with the peculiar particulars of each species. We are examining whether there are physiological and ecological traits that are locked together, and our application of multivariate analysis to the data emerging from the screening program has confirmed that within our 43 screened species there are several sets of about 20 attributes which we call "Axis 1." This shows tight covariance between resource capture and loss above and below ground and appears to entrain a variety of vital attributes of plants. Hence, demand for all the major nutrients varies in common; fast-growers are rich in nitrogen, phosphorus, and other minerals, while slow-growers are poor. Leaves of fast-growers have strong herbivore deterrents and decompose slowly, and so on. But the main conclusion to emerge from all this is that the most critical generalities about plant dynamics relate to mineral nutrients. And what is even more interesting is that work from other parts of the world and on other groups of organisms, such as White's in Australia on animals, appears to be drawing to a similar conclusion. So we are looking here at a principle that crosses the trophic levels and geographic boundaries and may be the key to ecosystem studies.
If mineral nutrients are shown to be the most influential currency of ecosystem functions, then Axis 1 offers tremendous opportunities for prediction, even at ecosystem and landscape scale, also in studies of resistance and resilience of ecosystems, in successional studies, and so on. And there is plenty of room for other axes of variation, and endless opportunity for fine-tuning among species along these axes.

Your laboratory work has taken you deeply into the realms of theoretical ecology, but what of the practical applications of the Integrated Screening Program? Have you, for example, been able to use the information you have gained in the study of the impact of climate change on vegetation?

Grime: Concern with climate change has come at a time when my interests have been moving back towards the field. We need to test the predictive models built from our laboratory screening, and the climate-change debate has provided the motivation and the funding for this development. We have a limestone grassland field site at Buxton in northern England where we have enlisted the aid of engineers to provide us with the means of elevating CO2 and temperatures near the ground surface, applying late frosts, enhancing UV-B radiation, and also adjusting summer rainfall in an extensive plot experiment. So our lab work has led to predictions that we can now test in the field.
In addition to manipulating natural ecosystems we have been concerned since the 1980s with testing our predictions from the laboratory by synthesizing communities and ecosystems. Recently we have concentrated on building low-cost, outdoor microcosms for this purpose. This approach has several advantages over laboratory systems. We are far better able to replicate; the lighting systems are more natural; the day length is not a problem; and so on. We are particularly interested at the moment in co-evolution between plants, herbivores, and even carnivores in high-productivity and low-productivity environments. We hope that this will all link back to our strategy theories.

But this is a big jump, from the artificial simplicity of the lab out into the complexity of the field.

   Grime: Yes, but we have devised a protocol for this development. The lab work provides us with the data on which we base predictions. We take them to the field for testing and often, of course, they don't work! So what's wrong with the prediction? The problem then comes back to the lab for a new cycle of lab work and modeling. This is the way that science has to work, and ecology is at last becoming scientific in this sense.
   We are also now able to model in sophisticated ways, using cellular automata to produce predictions of the behavior of multi-species assemblages. The raw data from the screening experiments, of course, have provided us with a knowledge of what attributes and relationships are important and need to be included in these models. These are the elements of our Periodic Table.
   Field testing also takes us away from a traditionally agricultural type of approach where we are looking at the response of our "crop" species (usually fast-growing, short-lived, and with a simple life cycle) in a heavily controlled environment. In the field, we have to deal with perennial species over long time scales. Regeneration may be very rare. Unpredictable events, such as severe drought or frost, may be very influential. In these respects, the long-term monitoring studies of grasslands that our collaborator Arthur Willis has been conducting at Bibury for the last 37 years are worth their weight in gold. For example, he has been able to detect a correlation between changes in the Gulf Stream position from year to year with the productivity of certain grassland species, and this in turn seems to correlate with coincident signals detected in marine and freshwater plankton. We are now looking at historical agricultural returns for England in the light of this work to see whether these also correlate.

There is one other area of ecology that you have worked in and which we have not yet touched upon: namely, your work on the size of genomes in plants.

Grime: This work began to emerge in the 1970s (following the research of M.D. Bennett) when it became apparent that plants with big cells had a large genome and slow rates of cell division. We have described a suite of associated characteristics, some of which are of ecological significance. So, for example, if we set up an experiment using large-genome and small-genome plant species and subject them to, say, warmer conditions, it is the small-genome species that respond fastest and leave all the large-genome ones behind. Similarly, if you look at the long-term data from Bibury, then the highest coefficients of variation of growth from year to year are in the small-genome types. With their short cell-cycles, these are the most responsive. The big problem for the cell biologists has always been to explain the selective advantage of having a large genome and a large cell, and this is where ecological insights can help.

Are you saying that there is environmental selection for genome size, or that the selection operates on some other, correlated trait?

Grime: Selection, I think, operates on cell size. The key observation is that large-celled species (with large genomes) appear to be able to grow by cell expansion at times when it is too cold for rapid cell division. The ability of a species to pack tissue away efficiently that can expand rapidly in a cold season is clearly advantageous. If there is a temporal separation of cell division and cell expansion, one can see potential advantages in that. You may have optimum cell division temperatures in one season and moisture available to permit cell expansion in another, as in the Mediterranean climate. In many geophytes, for example, there is hardly any temporal overlap between cell division and expansion. Many large-genome species are geophytes and are found in Mediterranean-climate areas.
There was a time, some fifteen years ago, when cell biologists were very interested in this topic, but the line seems to have gone dead! We ecologists need some help here.