Follett, P. A. & Duan, J. J. Nontarget Effects of Biological Control. , editors. 2000 . Kluwer Academic Publishers , Dordrecht, The Netherlands . 316 pp. $140.00 (hardcover). ISBN 0-7923-7725-7 . Biological control is a subject that should interest conservation biologists because it potentially can affect biodiversity in both positive and negative ways. Biological control is increasingly being used against invasive weeds that threaten natural communities. For instance, biological-control programs against invasive weeds currently are underway in at least three United Nations World Heritage Sites, against Acacia species in South Africa's Cape Fynbos, Mimosa pigra in Kakadu National Park in northern Australia, and Melaleuca quinquenervia and Old World climbing fern (Lygodium microphyllum) in the Florida Everglades. In contrast, biological control has rarely been used against invasive insects that threaten native biota. One recent project has tried to control an invasive Central American weevil (Metamasius callizona) that attacks many native bromeliads in Florida. Traditionally seen as the green alternative to pesticide use, biological control is now recognized to have environmental risks, and both improved practice and regulation are needed (Strong & Pemberton 2000). The actual, probable, and possible negative effects of introduced biological-control agents have become the subject of much debate. Nontarget Effects of Biological Control is part of this debate and more. It is an outgrowth of a 1997 Entomology Society of America symposium, Biological Control for a Small Planet: Non-Target Effects of Biological Control. Thirty-three authors contribute 17 chapters, and although there is a strong North American focus, the editors have added contributors from New Zealand, Australia, and the United Kingdom. From the table of contents onward, it is apparent that biological control is not a single discipline but a loose collection of disciplines involved primarily in the biological control of insects and weeds. Because native plants and insects have been and are valued differently by society and biologists alike, the practice and regulation of biological control of plants and insects differ markedly with regard to what is considered to be appropriate ecological risks, although considerable variation exists among the people in these disciplines. The debate over the safety of biological control is often characterized as opposing camps of biological-control defenders and ecologist-conservationists critics, a perspective voiced by the editors in their preface. Although some chapters in this excellent book confirm this view, the majority attempt to understand the complex ecologies of natural enemy groups and their actual or possible interactions with nontarget organisms. Four introductory chapters offer perspectives on safety that, more or less, fit the biocontrol defender-ecologist critic model. Ehler calls for a sensible balance between economic reality and conservation ethics and is concerned that without such a balance biological control itself will become an endangered discipline. Messing discusses the restrictive regulatory climate in Hawaii that has almost paralyzed the practice of biological control and gives as an example the requirement for host-specificity testing on a wasp belonging to the Aphididae, which he wishes to use against an exotic pest aphid. This wasp family has evolved with and is completely restricted to aphids, an insect group absent from Hawaii's native fauna. Messing's argument about an excessive focus on biological-control safety fails with the example of the rejection of potentially useful natural enemies of leafy spurge because of safety concerns. As the person who tested and rejected these “potentially useful natural enemies” because of their unacceptably broad host ranges, I am happy to report that leafy spurge is being controlled in most parts of its U.S. range by the safer specialists (Aphthona flea beetles) that were petitioned for release. Lockwood reviews possible indirect ecosystem effects of the project to biologically control the native rangeland grasshopper which he helped stop. He argues that reductionist western science causes us to focus on nontarget species instead of ecosystem processes that may be affected by biological-control introductions. Stiling and Simberloff analyze the literature to show that biological control agents have host ranges that usually include more than one species, that they often attack native hosts, and that they cause damage to nontarget species more commonly than is generally believed. Overall this is a worthwhile critique, but at times it is wide of the mark. The authors' selective use of the literature to derive statistics for rates of attack on native species by biological-control agents unnecessarily overstates reality. For instance, Louda et al.'s (1997 ) data on the thistle weevil's (Rhinocyllus conicus) use of native Cirsium indicate that all five of the Cirsium species examined were attacked by the weevil. The problem is that their study deals only with thistles attacked by the weevil, and it is well-known that this weevil attacks only thistles that flower during the univoltine weevil's egg-laying period. Currently, 22 of approximately 90 Cirsium species native to the United States are known to be developmental hosts of the weevil ( Pemberton 2000), by far the most serious nontarget use known in the biological control of weeds, but still well short of a 100% attack rate. An important clarification is needed with regard to the example of an unpredictable host shift that host-specificity testing failed to predict. “For example the leaf beetle Chrysolina quadrigemina, which was imported in the 1940s into northern California to control Klamath weed, began attacking an horticultural plant, Hypericum calycinum, a host shift which may have involved formation of a new host race” ( Ehler 1991). Ehler (1991) cites Andres (1985), who clearly indicates that H. calycinum and all Hypericum species tested were developmental hosts of the beetle. Preliminary post-release laboratory rearing tests, done during the 1980s, suggest only increased survival of the beetle on H. calycinum. Klamath weed (H. perforatum) is one of the many Hypericum hosts of the beetle in its native Europe. “Predicting Risk from Biological Control Introductions,” the chapter by Barratt, Ferguson, Goldson, Phillips, and Hannah, is one of six dealing with the biological control of insect pests by parasitoids and predators. This chapter uses laboratory host-specificity testing coupled with field surveys to examine the laboratory and field host ranges of the two Microctonus wasps, introduced for the control of two exotic weevil pests of forage plants. The chapter authors found good correlations between laboratory and field host ranges for both wasps, one of which has a much broader diet than the other. This finding represents some of the strongest evidence to date demonstrating the clear value of host-specificity testing of agents for the biological control of insects. Interesting retrospective studies on introduced parasitoids of a pest pentatomid bug in Hawaii by Follett, Johnson, and Jones, and by Duan and Messing on introduced parasitoids of fruit fly pests, examine the actual and possible parasitoid drift onto a native pentatomid bug and native fruit flies. The indigenous koa bug is highly parasitized at lower elevation sites but escapes in higher elevations. No native fruit flies, which are associated with the flowers of various Asteraceae, were attacked by the introduced parasitoids. Orr, Garcia-Salazar, and Landis examine the lab and field host range of the egg parasite Trichogramma brassicae, a wasp mass reared and released against crop pests in many parts of the world. In the laboratory, the wasp attacked eggs of all 25 tested moths, belonging to 11 families. In the field, sentinel egg baits of all offered species in the preferred cornfield habitat were attacked, but none of these sentinel egg baits placed in adjacent woodland were attacked. This generalist parasitoid's strong habitat preference and poor dispersal ability appeared to limit its potential to harm nontarget species. Obrycki, Elliot, and Giles examine ladybird beetles (Coccinellidae) imported for aphid biological control and state that present knowledge of coccinellid ecology does not allow for prediction of the interactions and effects of these insects. They conclude that the benefits of introducing aphidophagous coccinellids may not outweigh the potential risk to native coccinellids. Memmott's informative chapter on the use of food-web analysis to examine nontarget effects of biological-control agents is followed by chapters on the biological control of weeds. McEvoy and Coombs's fine, thought-packed chapter argues effectively that the biological control of weeds often resembles a lottery in which large numbers of natural enemies are released in hopes of establishing an effective agent. They call for the use of fewer but more carefully evaluated agents to produce more effective biocontrols with fewer negative effects. Withers, McFadyen, and Marohasy use a clearwing moth, imported for parthenium weed control, to demonstrate Australian research and importation protocols. Louda updates her important findings on the direct effects of the thistle weevil on nontarget native Cirsium and the weevil's displacement of the thistle's native fruit-fly herbivore. Nichols follows with an analysis of biological control programs for pest thistles and makes suggestions for improvement and conflict resolution. The book ends with three interesting chapters on attempting to understand nontarget issues related to the use of microorganisms for the biological control of insect pests: entomopathogenic fungi ( Hajek and Butler), Bacillus thuringiensis kurstaki ( Btk) bacteria (Miller), and nematodes ( Barbercheck and Millar). It has been documented that Btk kills many nontarget Lepidoptera. Although insect-killing nematodes are generalists, they are not thought to cause much nontarget damage because of their poor dispersal through the soil and limited persistence after application. Fungal diseases of insects vary greatly in breadth of host range and nontarget use, characteristics made less predictable by a lack of knowledge of epizootiology of many of these pathogens. I strongly recommend this book to life scientists and agriculturalists interested in pests, invasive species, and biodiversity. Nontarget Effects of Biological Control should be a must-read for biological-control researchers and those involved in regulation of the technology.