Patten, B. C. [Ed.].1975. Systems analysis and simulation in ecology, volume 3. Academic Press, Inc., New York, xv + 601 p. $39.50.

Although simulation of ecosystems by the technique known as systems analysis modeling has been becoming more popular over the last few years, I have never been convinced that it is the only route, or even the optimal route, to predictive theories about ecosystems. Nevertheless, I had the uneasy feeling that there must be something in it because a lot of brilliant ecologists have been practicing and preaching systems analysis very fervently. For years I thought I should learn something about the subject, but kept putting off what I knew would be a difficult job until the arrival of Systems analysis and simulation in ecology, uolume 3, on my desk gave me the incentive to begin. However, I soon discovered that the job was more difficult than I had imagined. The book was simply bewildering. Even the sections that looked easy were difficult to read. I once reached the point of pleading with an unsympathetic editor to let me keep the book without writing a review. (By that time there were too many undeletable expletives ball-penned throughout the text for me to return it. Furthermore, the beautifully designed dust jacket on the bookshelf deceived students into thinking that I had at last progressed beyond reading Liebig and Darwin. ) A completely negative response from the editor forced me to open the book again and after alienating myself from the rest of the family and suffering two imaginary, terminal illnesses, I finally got down to work. It was pleasant to discover that Part II is easier to read than Part I. What really disconcerts the nonmodeler is Part I, which comprises seven chapters on the U.S.A. biome models. Chapter 1 prepares us for great things because it tells us that the objectives of the biome program were simply to solve all the problems that ecologists have been working on for generations. In the second chapter our wild enthusiasm quickly changes to total bewilderment, because we have entered an unfamiliar world. It is a biology with a new language-a new paradigm. In the old days, despite our professional fragmentation, ecologists could understand and be interested in the work of geneticists, embryologists, molecular biologists, and others although we were inadequately trained to make original contributions to these fields. Now, it seems that a branch of our own discipline is beyond some of us because the chapters on biome modeling left me bemused. Eventually the source of the difficulty became clear. Systems analysis modelers have a totally different publication paradigm. Whereas we (the oldies) tend to publish the results of experiments we have actually done, they (the modelers ) seem to be publishing the equivalent of experiments they intend to do. Perhaps a more appropriate analogy would be with a theorist publishing the incomplete sketch of a theory he hopes, one day, to complete. Having made this discovery, I found it much easier to read on and learn what Part I contained. In fact, it contains little of interest for the oldies. Chapter 2 on the grassland biome was mainly a discourse on the mechanics, choice of language, and mathematics of modeling. It gave simplified descriptions of a few submodels in the form of crude flow charts, and a few example outputs, but no comparison of predicted with observed values. Chapter 3, on the eastern deciduous forest biome, is similar, but presents even less information about the model. It reads a little like an advertising brochure for a product that has not been invented. Chapter 4 on the desert biome and Chapter 5 on the tundra biome describe models that are a little further developed, but again a reader cannot learn about modeling from them or evaluate the models. For example, the desert group is reported to have progressed to the point where they have become able “to set up a firm base for modeling, and to chart a line of advance.” (P. 93) Chapter 6 on the coniferous forest biome is very interesting if you skip the introduction. More than the other chapters, this one attempts to explore the human and philosophical problems that face a large modeling group. By now I do not need to tell you that this model is never exposed either. Chapter 7 is a critique and comparison of biome system modeling and it comes as an anticlimax. The author obviously had to grope for something to say. What is there to say about a theory that has not been published and which can always be defended as outdated by a more recent, but still undescribed model? Consequently, Watt’s critique seems to be suggesting that the baby, if it exists, should be taught to run before it can crawl. Part II has five chapters on models of freshwaterestuarine ecosystems. These describe models that appear to be more complete but much more restricted in scope. For example, Chapter 8 deals with only two components, algae and flies in a hot spring, a highly simplified system. Chapter 9 presents a second generation model of the community in a small woodland stream. Chapter 10, on a cove in Lake Texoma, is by far the bulkiest and in several ways the most interesting chapter. First, it contains more information than the others about the sources of biological information and biological ideas that gave rise to the system of compartmentalization and interactions put into the model. Second, it is unique in giving enough detail about the model that someone could actually run it and check its predictions. Finally, it contains a truly