Introduction

Tropical forests have received considerable attention recently, stimulated in part by the high rate at which they are being modified or completely destroyed. Emphasis has been given to their potential effect on the global carbon balance and on biogenic emissions (García-Méndez et al., 1991; Matson & Vitousek, Chapter 16). Most studies have focused on tropical forests growing in wet or humid climates which account for 58% of tropical forests (Brown & Lugo, 1982), while seasonal forests growing in drier climates, which represent the remaining 42%, have been very little studied (Murphy & Lugo, 1986a). Tropical dry forests represented nearly 20% of the total biomass of forests in the world in the early 1970s (Persson, 1974). But due to deforestation, and extensive use of the area for intensive agriculture, pastures and shifting cultivation among other uses, the present area is considerably reduced.

Constraints on soil biological activity and their effect on ecosystem production, soil organic matter formation and nutrient cycling have been little studied in dry forests (Anderson & Flannagan, 1989). Most of the work on this topic has been carried out in India with considerably less information available from dry forests of the neotropics.

In this review I will discuss how water availability primarily constrains soil biological activity in tropical dry forests, in order to highlight its importance in the dynamics of organic matter and nutrients in the ecosystem.

Climatic characteristics of tropical dry forests

Strong rainfall seasonality is the overriding macro-determinant in dry forests. Seasonality and, most important, the duration and intensity of dry and wet periods, exert a strong influence on the biological activity of both the above- and below-ground parts of the ecosystem.

Traditionally, the task of increasing crop productivity has been seen as the role of crop breeders and agronomists, drawing respectively upon the sciences of plant genetics and plant physiology. From an agricultural science point of view, weeds of crops have been seen simply as a nuisance to be removed using the best available technology to hand. Much of weed science has therefore been devoted to the provision of ‘tools’ for weed removal. Only in relatively recent times has weed management been viewed as a problem to which ecological principles can be applied.

For historical reasons, then, the study of weeds has been divided between two groups of scientists. Agronomists and horticulturalists have seen the presence of weeds as a pragmatic problem to be solved (how can the weeds be killed?). Plant ecologists, on the other hand, have seen cropped land as somewhat unnatural, human-managed habitats and weeds as particular organisms that are able to exploit such habitats (i.e. as academic curiosities). Books considering weeds have followed this dichotomy, although there has been more emphasis on the practice of weed control than on weed ecology. Indeed, despite the fact that much of the temperate regions and an increasing area of the tropical regions of the world are farmed or managed in some way, it is surprising that plant ecologists have given agroecosystems relatively so little attention. As with other areas of pest control, the agroecosystem–natural ecosystem dichotomy has persisted to the detriment of weed science. However, with increasing concern over the preservation of biodiversity, many ecologists are becoming interested in alien invasions threatening more natural habitats (‘environmental weeds’).

Introduction

Dry tropical forest vegetation in Thailand and adjacent parts of Southeast Asia exhibits structures and ecological processes very different from those characteristic of neotropical dry forests. Unlike the Pacific coast of Central America and México where dry forests are virtually all deciduous, evergreen forest types are widespread in the dry forest climatic regime of Southeast Asia. Evergreen dry forests in areas with 1200–1500 mm y−1 precipitation and deciduous forest in areas with up to 2300 mm y−1 are surprising occurrences in relation to climate–vegetation seen in dry neotropical forests. Many of these regional differences can be attributed to the poor nutrient status and low water-holding capacity of the shallow and infertile latosols and lithosols which predominate in Southeast Asia. Many other factors – historical, abiotic and genetic – are no doubt involved as well. In this review, we provide a broad biogeographic survey of the geography and dry forest communities of Thailand as an introduction to this region.

The literature on forest vegetation in Southeast Asia has been largely unavailable in Western libraries, thus limiting ecological interest in this important region. International concerns over global climate change and loss of biodiversity, however, have led to renewed interest in the structure and function of forest vegetation in Thailand (Round, 1988; Elliot, Maxwell & Beaver, 1989) as elsewhere. Also, future studies contrasting paleotropical and neotropical dry forests will undoubtedly lead to new perspectives on old problems, and to a much better understanding of the interactive nature of nutrient availability and seasonality in soil moisture which lead to broad patterns of forest dominance by evergreen or deciduous species.

Introduction

The history of a plant community is reconstructed from remains preserved in a fragmentary fossil record. The completeness of the reconstructed community is, in part, a function of whether the community grew under conditions favorable to the preservation of macro- and microfossils. Macrofossil assemblages (leaves, fruits, seeds, wood) generally record plants growing near the site of deposition, and these afford opportunities to determine the general vegetation type and the paleoclimate by comparisons with modern analogs and by the use of leaf physiognomy. Microfossils (pollen, spores, trichomes, cuticles, phytoliths, microscopic organisms) provide a record of the regional vegetation, and also include species often not represented by macrofossils, such as annual, suffrutescent and herbaceous plants. Each methodology has its own set of strengths, weaknesses, practitioners and advocates, but the most complete history of plant communities is produced when both macro- and microfossil floras are available.

Northern Latin America

In the case of the tropical dry forest in northern Latin America, the reconstruction of its history is made more challenging by the facts that (1) dry environments have fewer sites of deposition, and less water for the transport of remains to these sites, and (2) there are very few well-preserved Tertiary macrofossil floras of significant size or diversity known for northern Latin America. An exception is the Oligocene San Sebastian flora from Puerto Rico, but it has not been studied or revised since Hollick's (1928) original publication.

Abstract

The nature of wind damage as a risk to managed forests is reviewed, based primarily upon experience in Britain. Risk assessment is required to guide site selection, choice of techniques to counter the damage, and plans to respond to damage. Assessment may simply rank sites but prediction of timing of damage is also needed for many decisions. Subjective assessments can be unreliable where there is little history of forest management in that locality or when the damage is determined by many factors. Objective assessments are preferred but are not easy to derive. The existing British windthrow hazard classification is readily applied to sites and can be mapped. However, both its treatment of the wind climate and its estimates of timing of damage are deterministic. The potential to improve the system is discussed using a conceptual model of the assessment of risk which highlights a number of important research issues. Wind damage occurs as a result of an interaction between a variable wind climate and changing tree vulnerability. Vulnerability includes persistent, progressive and episodic components. There can be substantial variability in the frequency and magnitude of damaging storms which is not adequately represented by the conventional separation into storms causing endemic and catastrophic damage. The change in vulnerability with time can have a very marked effect on the frequency of damage. Recent validation has indicated that the estimates of damage based on the hazard classification of damage tend to be pessimistic, and that there is more variability in rate of windthrow than the classification allows. The variability requires a stochastic model and necessitates flexible forest planning; the latter conflicts with many pressures on forest managers.

Introduction

Why biological control?

The past 15 years have witnessed a dramatic increase in research related to biological control. The current perception that biocontrol will have an important role in commercial agriculture in the future contrasts markedly with previously-held views that biocontrol agents perform too inconsistently, or are too narrow in their spectrum of activity, as compared with chemical pesticides, to be commercially feasible on a large scale. Renewed interest in biological control is in part a response to widespread concern about the potential negative impact of chemical pesticides on public health and the environment. Furthermore, the techniques of molecular biology have revolutionized the field by facilitating the identification of the molecular basis of pathogen suppression and by providing the means for construction of ‘superior’ biocontrol agents. New biocontrol agents resulting from recent intensive research are slowly becoming available to agriculture, and the trend should accelerate throughout this decade. One example is Gliocladium virens, which is being marketed in potting-mix to control Pythium and Rhizoctonia (see Lumsden and Walter, Chapter 25).

This chapter deals with the potential benefits and risks from the introduction of biocontrol agents for the control of root diseases of wheat, as well as the impediments to the application of this technology in commercial agriculture. The focus of the chapter is on biological control of take-all of wheat by fluorescent Pseudomonas spp. because it is a model system for the study of the molecular basis of pathogen suppression, root colonization by introduced bacteria and field application of biocontrol agents.

Introduction

A species is described as an invader when it colonizes and persists in an ecosystem in which it has never been before (Mooney and Drake, 1989). Invasions by insects seem to be rather well documented, but that is a false impression. It is only for two categories of insect species that documentation of successful invasions is complete: (i) for pest species (usually herbivores); and (it) for natural enemies (predators and parasitoids of insect pests, or phytophagous insects used in weed control). This information relates mainly to agro-ecosystems and few or no data are available for insect invasions in other ecosystems. This might mean that such ecosystems are not frequently invaded, that the effect of invaders is not dramatic (but see Howarth, 1991; also, see Hopper, Chapter 6 and Andow et al, Chapter 10), or that these other ecosystems are considered economically so unimportant that they receive no attention. The first view seems to be supported by the general literature on colonization, where the massive restructuring of natural ecosystems is usually regarded as the main cause for making invasions possible. Complex natural systems are replaced by a few plants and animals that are of direct use to humans (Mooney and Drake, 1986a).

This paper is structured as follows: first, some general facts about insect invasions are stated and then several case studies are provided to illustrate problems created by insect invasions. Following this the characteristics of the insect invaders and the systems which are invaded are discussed. Finally, some specific questions related to the predictability of the effects of insect invasions are answered.

Introduction

Introductions of parasitic wasps have often led to economic control of introduced insect pests (Clausen, 1978). Biological control practitioners argue that such introductions have fewer and less severe detrimental impacts than alternative controls (DeBach and Rosen, 1992). However, others have challenged this viewpoint (Howarth, 1983, 1991) with the argument that introduced biological control agents can drive non-target populations extinct, especially if these are already rare. Others have responded that the impact on conservation of insects species is unlikely if introductions are done carefully (Samways, 1988). Unfortunately, this controversy has given rise to more heat than light, being plagued by excessive generalization, lack of hard data on adverse impacts, and lack of concrete recommendations about how to reduce the risks of introductions. The lack of data on adverse impacts arises, in part, because many biological control projects lack the funds for evaluating the impact of the introduced agent on the target species, let alone on non-target species. The low level of funding, relative to the complexity of the problems involved, also means that methods to reduce risks will have to rely heavily on extant knowledge, if biological control introductions are to continue. In this chapter, I will address the following questions: How can one predict the potential impacts on threatened and endangered insect species from introductions of parasitic Hymenoptera for the control of insect pests in the continental United States?; and, What are the potential impacts from selected introductions? I have chosen a narrow scope of impacts and types of introductions to avoid overgeneralization.

Introduction

The use of chemical pesticides has resulted in a considerable increase in crop yields throughout the world. There is, however, increasing public pressure against the use of chemical pesticides due to their adverse effects on human health and the environment. The very concept of using chemical pesticides has been challenged by the green movement, which is constantly gaining momentum. World-wide, annual agrochemical sales amount to approximately $US 25 billion (Powell & Jutsum, 1993). Biocontrol agents and biopesticides currently contribute only about 0.5% to the total (Meneley, 1990; Anonymous, 1992). Products based on Bacillus thuringiensis account for 90-95% of biopesticide sales (Feitelson et at., 1992). Bacillus thuringiensis is a biologically produced pesticide, but the active ingredient is not a living organism.

The economics of pesticide use can be considered at three levels: the producer/distributor, the user and society. A pesticide must eventually bring profit for the producer and it must give ample economic return to the user. From society's point of view, the economic benefits from use of a pesticide must clearly offset the adverse effects (Pimentelera/., 1991).

There are scant historical data on the economics of biopesticides other than for Bacillus thuringiensis-based products. Most of the data are proprietary. In this chapter, I discuss the economic aspects of biopesticides from the viewpoint of a company commercializing biopesticides (see Cullen and Whitten, Chapter 26 for a discussion of the economics of biocontrol from a research perspective). The scope is limited to microbial biopesticides. Special emphasis is given to the risks involved in various steps of the development and commercialization process.

Cypripedium

A circumboreal genus, Cypripedium occurs in bogs, swamps and woodland, especially in forest margins and clearings, usually not in deep shade. C. calceolus is virtually confined to calcareous soils, the topsoil being characteristically friable with a high humus content (Fuchs & Ziegenspeck, 1926a). C. candidum grows in black humus saturated with lime, the soil having a pH of 7.4 (Stoutamire, 1990). A stand of C. reginae in Canada was found growing in soil with about 20% more calcium than the average for the region (Harvais, 1980). Although most of the species thus prefer a neutral or alkaline substrate (Wherry, 1918; Curtis, 1943), C. acaule grows in acid soil (pH 3.5–5.4: Wherry, 1918; Curtis, 1943; Stoutamire, 1963).

In all species the adult plants have a horizontal rhizome and well-developed, dark green foliage leaves, and maintain this foliage from spring to autumn. They overwinter below ground. The vegetative period before flowering is assumed to be long: up to 10 years or more in American species (Curtis, 1943) and 6–10 years in C. calceolus (Kober, 1972; Fast, 1985).

In C. arietinum, C. candidum and C. calceolus var.pubescens the tips of 2 to 3-year-old roots may form buds as a means of vegetative propagation (Curtis, 1943).

Life history

C. calceolus presumably germinates in spring (Fuchs & Ziegenspeck, 1926a), although Irmisch (1853) observed very small protocorms of C. calceolus in December. C. acaule begins to germinate in spring and early summer (Curtis, 1943). The seeds usually lie fairly deep (2–5 cm), typically occurring below a layer of living mosses and partially decomposed plant debris on top of the humous mineral soil (Fuchs & Ziegenspeck, 1926a; Curtis, 1943).

Abstract

Two experiments were carried out with Sitka spruce and European larch grown under intermittent wind in tunnels. In the first, both species were grown to determine whether there are any specific effects of wind on root growth, as opposed to effects on growth in general. In the second experiment, larch seedlings had their tap root removed 20 mm below the soil surface to mimic the formation of a shallow root plate, such as develops when larch is planted on seasonally waterlogged peat in much of the United Kingdom. In both experiments, lateral roots were counted and their orientation relative to the tap root recorded. In the wind-stressed trees in both experiments there was an increase of almost 60% in the number of large roots on the windward side of the trees and of 45% on the leeward side of larch trees compared with the number of roots growing at right angles to the direction of wind. In the first experiment, the sum of the cross-sectional area (Σ CSA) of lateral root bases was greater on the windward side of the tree compared with the other sides in both species. In the second experiment, Σ CSA of lateral root bases of wind-stressed larches was greatest on the leeward side whereas the control plants had a larger Σ CSA in the regions perpendicular to the wind direction. It appears that wind action stimulates the diameter growth of those roots most important for anchorage, but has a smaller effect on root development than other factors such as uneven nutrient supply.

Abstract

Mathematical models of the transfer of wind momentum into forest canopies and the subsequent bending stresses in individual tree stems are presented. The windspeed profiles within and above canopies are predicted for conditions where the extreme windspeed likely to occur in a 50 year period in south-west Scotland is used as a reference. The stresses predicted by the models in these conditions are compared for Sitka spruce (Picea sitchensis (Bong.) Carr.) growing in an unthinned stand (3800 stems ha−1) and in a stand recently thinned by removing half the trees. The mechanical model of the trees includes the responses to dynamic as well as static bending forces. The calculated variation of stress along the stems of six trees of different dimensions is compared and found to have a maximum near the stem base in some trees but a nearly constant pattern for several metres in other trees. The maximum value of stress is about 10 MPa for trees in the unthinned model and about 20 MPa for the recently thinned model. These stresses are compared with the estimates from published studies of the strengths of the root systems and of the stem wood. The increases in stress with thinning are shown to reflect qualitative observations in British plantations that wind damage is more likely and that this will be in the form of overturning rather than stem snap. It is concluded that the models should be a good basis for further work on estimating the risk of wind damage to plantation trees.

Abstract

Catastrophic wind events impact forests over the entire globe. Although recent examples of hurricanes in the Caribbean have led to intense examination of the impacts on, and recovery of, forests, these research efforts have largely been in isolation. Little has been done to compare the impacts of storms of varying intensity on different ecosystems. Therefore, we can not as yet fit catastrophic wind events into a general model of forest disturbance and recovery. Papers examining the impacts of 26 different wind events (cyclonic storms, tornadoes and gales) on 27 different forests are reviewed. Hurricane damage is measured as numbers or percentage: stem damage, canopy damage, biomass or stand volume loss, or mortality. The populations sampled varied from minimum stem size of 2 cm to 20 cm in diameter. Sampling methodology included small circular plots, transects, large gridded plots and remote sensing of the landscape. Plots were established 10 days to 3 years after the wind event. The implications of these different quantification systems are examined using data from a large gridded plot established in the Luquillo Mountains of Puerto Rico to study the impacts of Hurricane Hugo. Intensity of disturbance, measured as percentages of different damage types, varied depending on the minimum stem size used in the analysis. Damage to individual species also varied depending on the variable used to quantify it. Clearly, a standard measure of wind damage is needed to facilitate comparisons of the impacts of different storms on different forests. I suggest a damage measure that includes both mortality and structural loss as measured by decrease in basal area.

Abstract

Two instruments are described for assessing the trunks of standing trees for defects which affect their strength. One, the Metriguard Stress Wave Timer, measures the time taken for a sound wave to travel through the trunk; defects reduce the velocity of the wave. The other, called the Fractometer, measures the stiffness, strength and static fracture energy of a core removed with an increment borer.

Introduction

Two devices for the assessment of wood quality in standing trees are described. The first is a modification and an adaptation of an existing device, namely a Metriguard Stress Wave Timer, and the second is a new tool for measuring the strength of core samples. These two devices support a method of assessment known as Visual Tree Assessment (VTA) which was developed at the Karlsruhe Nuclear Research Centre (Mattheck & Breloer, 1993). VTA is based on the observation that a tree which contains a defect (crack, hollow, etc.) will repair itself by attachment of more wood at the weakened cross-section in order to restore the state of even load distribution (Mattheck, 1991). Figure 13.1 shows some typical symptoms and related defects. If these symptoms are visible in trees that could pose a hazard to the safety of the public, further investigation is justified. Also, in commercial forestry it may be necessary to check the wood quality of a stand as the basis of management decisions. This requires an assessment of the size of the internal defect, and a measure of the strength of the affected wood – tasks that can be achieved with the devices described below.

Introduction

The application of biological control agents is an established if under-exploited method for the control of agricultural pests and diseases. There are considerable benefits to be gained from this approach, however caution is necessary in the introduction of biocontrol agents because of possible effects on non-target organisms. In some situations the biocontrol organism itself may become a pest.

There are numerous examples of experimental biocontrol treatments for fungal and bacterial plant root diseases, though on a practical level there are less examples than for control of insect pests. Plant root diseases can often be controlled by management practices such as rotation with a non-host crop or by several years of crop monoculture, a practice which can lead to the build-up of a disease-suppressive soil (Hornby, 1983). Microbially mediated disease suppression can also occur in soil-less potting mixes (Hoitink and Fahy, 1986). Fungicides (e.g. Baytan for control of take-all) are used for control of disease, but success has been limited especially when applied to field crops. Biological control agents used to control soil-borne root diseases are most commonly soil bacteria or fungi.

On the whole, there are relatively few publications dealing with risk assessment in the literature on biocontrol. Recent papers include that by de Jong et al (1990). The need for risk assessment is now much greater, particularly as genetic manipulation is beginning to be used more often in biological control (Jones and Kerr, 1989; Ryder and Jones, 1991). The use of recombinant DNA technology brings its own risks and benefits. Given that there are both risks and benefits for pest control methods, procedures for their estimation are desirable.

Abstract

The determination of the wind regime in a potential or existing forest area is an important aspect of any site assessment. Wind affects the growth rate of the trees and determines the occurrence of windthrow in the later years of the development of a forest. Consequently the prediction of tree growth, and the forecast of the financial viability of any forest project, are dependent upon an accurate assessment of the windspeed. Many forestry locations have sparse windspeed records due to difficulties of siting equipment in remote areas. In Britain, methods of windspeed assessment for such purposes have included the use of tatter flags and, more recently, numerical and statistical modelling techniques. An attempt is made here to relate windspeed to the topographic and geographic characteristics of the site in order to avoid the necessity for long-term on-site wind measurements. The variables related to windspeed are altitude, which has a recognised relationship with windspeed, topex (a measure of the exposure of a site), roughness length (related to the height of the surface elements), United Kingdom Ordnance Survey grid position, and distance to the coast. Initially, each geographic variable is related individually to annual mean windspeed at a set of 21 sites, taken from both the Forestry Commission and United Kingdom Meteorological Office networks. Subsequently, the variables are related in multiple regression equations to the annual mean windspeeds of 1989 and these equations are then used to predict the windspeeds in 1990. A test of the method has been made on an independent site not used in the previous stages. The results are encouraging.