COMPENSATION IN FISH POPULATIONS SUBJECT TO CATASTROPHIC IMPACT Pamela J, Mansfield, David J. Jude, Philip J. Schneeberger , and James A. Wojcik Under Contract to the Electric Power Research Institute Palo Alto, California Special Report No. 117 Great Lakes Research Division The University of Michigan Ann Ar;bor, Michigan 1985 ACKNOWLEDGMENTS We wish to thank several people for their contributions to this report. Steven L. Kohler wrote the first draft of the definition of compensation and mechanisms, and wrote part of the annotated bibliography. Philip Hirt and Mary Sweeney rendered valuable assistance by, respectively, conducting much of the literature search and entering text; both made many trips to the library for us. Peter Gable drafted Figure 1. We thank John Forney, Bernard Skud, Jim Diana, John Palmer, and Jack Mattice for their reviews of the draft; the first two in particular contributed many useful critical comments. Henry Regier and Paul Rago provided several thought -provoking ideas in the early stages of our work. John Blake, as technical advisor, guided us by questions and reminders concerning the direction and focus of the product. We had many stimulating discussions with Robert Otto about our methods of literature search, and how to maximize success of our report and make it most useful to EPRI and to other researchers. Several researchers and biologists were particularly cooperative in providing or discussing unpublished data: Robert Siler, Duke Power Company, Huntersville, NC; Richard McLean, Oak Ridge National Laboratory, TN; Dave Thomas, Charles Main Inc., Boston, MA; David Michaud, Wisconsin Electric Power Company, Milwaukee, WI; and Burt Johansen, Public Service Electric and Gas Company, NJ. 11 TABLE OF CONTENTS ACKNOWLEDGMENTS ii LIST OF TABLES V LIST OF FIGURES vi INTRODUCTION 1 Purpose and Objectives 1 Definition of Compensation and Mechanisms 1 Catastrophic Impact 8 METHODS 13 Literature Search 13 Organization of Results 13 CASE HISTORIES 18 Introduction 18 Alewife 18 Threadfin Shad 26 Gizzard Shad 33 Lake Whitefish 41 Pacific Salmon 46 Rainbow Smelt 51 Yellow Perch 55 Walleye • 67 Freshwater Drum 73 SYNTHESIS ' 81 Life History Strategies 81 Resilient and Stable Species 83 Community Effects 83 Adverse Effects of Compensation 85 Effects of Catastrophe Type 86 Effects of Geographic Area 87 RESEARCH RECOMMENDATIONS 89 Introduction 89 Clupeids 90 Pacific Salmon 91 Yellow Perch 91 Freshwater Drum 91 Follow-up Studies 92 Expansion of Ongoing Monitoring Programs 93 111 Relative Importance of Recommended Research 94 REFERENCES ' 96 APPENDICES 116 Appendix 1. Annotated Bibliography 117 Appendix 2. Author - Date List of Literature 160 IV LIST OF TABLES Table 1. Characteristics of some catastrophic impacts 10 Table 2. Keywords used for computer search done by Science Applications, Incorporated 14 Table 3. Published studies of catastrophic mortalities to fish populations with information on subsequent compensation used in this report 16 Table 4. Characteristics of species chosen for this report 18 Table 5. Abundance of alewife in Lake Michigan, 1959-1967 21 Table 6. Growth increments (mm) by age-group for alewives before and after catastrophic mortalities at various locations 23 Table 7. First-year growth (mm total length) of threadfin shad. . • . 30 Table 8. Mean total lengths and length ranges (mm) for each age-group of gizzard shad from several locations 35 Table 9. Grand mean lengths (upper line) and increments (lower line) in mm at time of annulus formation for gizzard shad at various locations 37 Table 10. Average 'total length (mm) and weight (g) of pre- (1941) and post-mortality (1944-1945) rainibow smelt from Green Bay, Lake Michigan. Adapted from Van Oosten (1947) 54 Table 11. Growth and age at maturity for freshwater drum at various locations 76 Table 12. Unpublished or unanalyzed data sets 95 LIST OF FIGURES Figure 1. Scheme of the life history of a fish, showing when various compensatory mechanisms are most likely to operate Figure 2. Resilience versus stability, as defined by Rolling (1973), of the species investigated for this report 84 VI INTRODUCTION PURPOSE AND OBJECTIVES Our primary purposes in this study were to review literature, critically evaluating case histories, to determine the mechanisms and extent of compenjsation by various fish species subject to catastrophic impacts, and make recomm€mdations for future research. We searched for patterns and relationships among fish species and the compensatory mechanisms they exhibited, with the objective of functionally grouping species. By grouping species, the results could be applied to fish populations for which no data are available. There are relatively few studies where mass fish mortalities have b€5en followed by documentation of subsequent compensation. Thus, we sometimes pieced together evidence of compensation by comparing impacted populations with stable populations elsewhere. To gain a complete picture of compensation by the species we investigated, it was necessary to include data from chronically exploited or impacted populations as well as those catastrophically impacted. An outgrowth of our literature review was the identification of data gaps. Researchers presented with an impacted fish population or a mass mortality have chosen different ways of documenting it: counting dead fish, tracking the subsequent increase in numbers or biomass, recording growth rates, or sometimes recording fecundity or age-at-maturity changes. Some potential compensatory mechanisms which are difficult to quantify, such as age-specific survival or competition, have been ignored or merely stated as the probable reasons for increased abundance or growth rates. Thus, the full spectrum of potential compensatory response is not known for most fish populations, and many data gaps were identified. Our final objective, then, was to consider the importance and difficulty of obtaining information on previously unquantified compensatory mechanisms and to recommend subjects for further study. DEFINITION OF COMPENSATION AND MECHANISMS Introduction In simplest terms, compensation is any process which leads to a population increasing when at low densities and decreasing when at high densities. There are several levels of organization at which compensation can be considered: ecosystem, community, species, population. Data are usually insufficient to separate stocks, or to ascertain the limits of one population. Therefore, most of our discussion involves a species and assumes the data refer to one integrated population. Although we emphasize compensation at the population level, we include both internal and external mechanisms at that level of organization. Where possible, distinctions are drawn between immigration from neighboring populations and wi thin-population compensation, i.e., increased survival. Immigration presents a problem in analysis for oceans or rivers, where a localized catastrophe usually results in fish moving in from neighboring areas, not compensation by survivors. The best demonstrations of compensation occur when the entire species undergoes a catastrophic mortality in a bounded system such as a lake. Population Regulation and Compensation Many processes can affect the size of populations. Some factors, such as weather and environmental disturbances (e.g.^ earthquakes, volcanoes) typically affect population size in a density-independent manner. That is, the proportion of individuals affected does not vary with population size. Other processes affect population size in a density-dependent manner (i.e., the proportion of individuals affected varies with population density) . Although both density-independent and density-dependent mechanisms may act to influence population size at any point in time or over some time frame, it is generally agreed that density-dependent processes must operate for a population to maintain some long-term "equilibrium" density (Ricklefs 1983). Density-dependent processes affecting population size can be compensatory or depensatory. Compensatory mechanisms cause an increase in death rate and/or a decrease in birth rate as population size increases and a decrease in death rate and/or increase in birth rate as population size decreases. Depensatory processes tend to decrease birth rate and/or increase death rate as population size declines. Obviously, if only depensatory mechanisms were operating in a population, the population would eventually go extinct. Compensatory mechanisms must operate for a population to have long- term stability. The regulation of fish population size by compensatory processes has received considerable theoretical speculation but relatively little empirical study. The following discussion of compensatory mechanisms is based on reviews by Solomon (1949), Ricker (1954), Cushing (1974, 1977), Gulland (1977), Jensen (1981) and, with particular emphasis on responses of populations to perturbations, by McFadden (1977) and Goodyear (1980). In the following section we will briefly describe the compensatory mechanisms that are thought to be most important in affecting the size of fish populations. Some mechanisms may operate throughout a fish's life cycle (survival, growth) while others primarily operate at a certain stage (Fig. 1). In general, the process that forms the underlying foundation for the operation of most compensatory mechanisms is how per capita resource availability changes with population density. Food and space (e.g., spawning sites, feeding territories) are frequently cited as resources that can limit fish population size. In the following discussion, we will be concerned with how changes in resource availability affect processes (compensatory mechanisms) which, in turn, affect population size. Compensatory Mechanisms Age-specific Fecundity — Per capita food availability increases as population density declines. This leads to increases in individual growth rates and causes individuals at a given age to be larger than they would be if food was in shorter supply. As INTRASPECIFIC COMPETITION DISEASE PARASITISM AGONISTIC BEHAVIOR INTRASPECIFIC COMPETITION OTHER SPECIES INTRASPECIFIC COMPETITION DISEASE PARASITISM FIG, 1. Scheme of the life history of a fish, showing when various compensatory mechanisms are most likely to operate. size and fecundity are positively correlated in fish, this leads to an increase in egg production per female as population size decreases (Goodyear 1980; Jensen 1981), which can affect the balance between birth and death rates in a compensatory fashion. For most fishes, the density-dependence of growth declines with age, so for those species, density-dependent fecundity cannot control populations in itself over wide variations in density (Gushing 1977). Age at Iteturity — T3iis is a related mechanism which operates by influencing the total number of eggs produced per female per lifetime rather than the number of eggs produced per female per reproductive bout. More rapidly growing fish often mature earlier than slower growing fish (with the differences in growth rates resulting from differences in resource availability) . Females that mature earlier can produce more eggs in an average lifetime than females that mature later, which leads to an increase in birth rate as population size declines (Goodyear 1980; Jensen 1981). Earlier reproduction contributes more to "r", the intrinsic rate of increase, than does an equal amount of offspring produced later in life (Birch 1948) . Size-specific Fecundity — Reduced food availability can affect the number of eggs produced per female independent of any effects of ration on growth rate. A decline in daily ration with increasing population size can reduce the energy available per female for egg production (Goodyear 1980). While this mechanism may be important in growing populations, it is likely of less importance in declining populations since, at a given size, there are physical limitations on the number of eggs that can be produced. Sex Ratio — Reduced density may cause more females than males to be produced in the population, or high density may cause more males, as sex is determined pheno typically in at least some fishes (Brown 1970; Torblaa and Westman 1980). Exactly how density determines sex differentiation is unknown. Torblaa and Westman (1980) speculated that high density could induce stress through crowding, influencing differentiation toward maleness; or, that crowding could cause disproportionate female mortality after differentiation. Changes in age composition, growth, and abundance were concurrent with sex ratio changes in Lake Michigan bloaters, Conegonus hoyi (Brown 1970). Rapid growth could trigger mechanisms that produce more female offspring (Jensen 1981). Multiple Spawning During One Year — This mechanism, though not well documented, may be a compensatory response to low densities for some fish. Egg diameter frequencies and presence of fish larvae in late summer after a catastrophic mortality appeared to indicate that threadfin shad (Donosoma petenense) sometimes spawn twice in 1 year (Ki Iambi and Baglin 1969; Johnson 1971). The second seasonal complement of eggs may be resorbed at higher fish densities (Johnson 1971). Multiple spawns are, of course, less likely in temperate latitudes than in warmer climates. Egg Deposition — In the preceding mechanisms, the total number of eggs produced in the population was affected by population size. It is also possible for population density to affect the number of eggs actually spawned independent of the number of eggs produced. Spawning may be suppressed at high population densities (i.e., some females may not spawn at all - Jensen 1981) due to competition for spawning sites (e.g., salmon - Larkin 1977) or by the excretion of hormone-like repressive factors by members of the population (Goodyear 1980) . Change in Size of Spawning Area — Pacific sardines (Sardinops sag ax) and northern anchovies (EngnauJis mordax) distribute their eggs over a large area at times of high population abundance, then contract their spawning area toward more productive nearshore sites when abundance is low. As a result, at low population levels fish larvae are more concentrated inshore where a high density of food is most likely. Although this maximizes the probability of spawning success, it also increases variability in year-class strength. Conversely, large populations achieve good survival for a part of their range by spawning over a large area, and year-class strength varies less (Radovich 1979). All of the mechanisms considered above concern means of increasing birth rate as population size declines and vice versa. The remaining mechanisms that will be discussed affect mortality rate in a density-dependent manner. Predation — There are two general ways that predation can result in density-dependent mortality in a population. One way results from an interaction between density-dependent growth and size-selective predation. When growth rate is density-dependent, it takes members of a cohort longer to reach a given size as population size increases. If a population exhibiting density-dependent growth is subject to size-selective predation, individuals will be subject to predation longer at high population densities than at low densities. This produces density-dependent mortality since the proportion of the initial cohort size that is eaten increases as the duration of exposure to predation increases (Goodyear 1980). Predation is usually a more important controlling mechanism in early life than later, because as a fish grows its potential predators become less numerous (Cushing 1977). To demonstrate that this mechanism is operating in a population requires that both density-dependent growth and size-selective predation be shown to occur. Note that this mechanism does not require that the predator exhibit a positive numerical response to the density of its prey. The second general predation mechanism does require the predator to show a numerical response to 'prey density or a change in feeding behavior with changes in prey density to result in compensatory mortality. If the predator: (1) aggregates in areas of high prey density, (2) feeds more selectively on the pr&Y species as prey density increases, or (3) increases in abundance as prey demsity increases, the proportion of the prey population eaten may increase as its population density increases (Goodyear 1980). Cannibcilism — Cannibalism may be an important compensatory mechanism among piscivorous fishes and may be considered as a special case of predation. In populations where age-1 and older fish prey on young of the year (YOY), abundance of age-1 and older fish can strongly affect year-class strength, providing an effective means of regulating population size (Goodyear 1980). Disease, Parasitism — These mechanisms operate in similar ways, and show similarities to predation. At high population densities, transmission of bacteria, viruses, or parasites throughout the population is more likely to occur than at low densities, if high densities increase the number of intraspecific contacts. As high populations deplete food supply, increased food-seeking movements coupled with decreased resistance would promote the spread of disease organisms (Solomon 1949). The disease organisms, parasites, and predators whose density-dependent action is most pronounced are those which concentrate on one species for prey or host. Parasites, diseases, or predators may be reduced greatly when the host or prey population is at low density, so that the host can increase again for a time before control is again exerted (Solomon 1949). Incubation Mortality — Eggs may survive better at low densities than at high densities. Oxygen demand may be limiting at high egg densities (Johnson 1965), and waste product buildup may reach toxic proportions (Hunter 1959; Foerster 1968). Closely spaced eggs are also more prone to fungal infection (LeCren 1965). Interference Competition; Agonistic Behavior — Aggressive competition for feeding territories can lead to compensatory mortality. Individuals of species that are territorial and are unable to establish territories may be more susceptible to starvation or predation than individuals that successfully establish territories. If space for feeding territories is limiting, the proportion of individuals using less favorable territories, or unable to establish territories at all, will increase as initial population size increases and, consequently, the mortality rate will increase as population size increases (Goodyear 1980). Starvation — Mortality resulting from starvation in early life history stages is considered to be a potential, though not well documented, means of population regulation. Starvation mortality can be compensatory only if the population, and not some other factor, is responsible for reducing its food supply (Goodyear 1980). Starvation is believed to be most important just after yolk- sac absorption, termed the "critical period" (Hjort 1914), but may occur at other life stages as well. While food may be limited and reduced by the population, many fish do not actually die from lack of food but are weakened, so that they are more vulnerable to other causes of mortality. Mortality Resulting from Stress — In many populations, the probalulity of an individual surviving stressful conditions (e.g., overwintering conditions) increases as individual size increases. Consequently, if growth rates of individuals in the population are density-dependent, the proportion of individuals that survive stressful conditions increases as population size decreases (Goodyear 1980) . Rosenthal and Alderdice (1976) discussed the effect of sub-lethal stressors on marine eggs and fish larvae. They said that it was important to identify stressors but it was more complex to identify the consequences and ecological implications to fish. Temperature and salinity affect the size of Pacific herring (CJupea hanengus palJasf); no jaws develop at < 5°C. Radiation causes eye malformations, and cadmium causes the death of eggs. Studies provide insight and recognition of basic pathways a fish may follow to compensate for environmental stress. The scope for compensation is limited in the basic mechanisms involved and changes as development proceeds from the egg onward. The basic reaction to chemical insult at the egg stage occurs at the biochemical level — decoupling of oxidative phosphorylation and blocking production of ATP. This can cause the dedifferentiation of cells and retard development of eggs. Low temperature and low dissolved oxygen can cause stress and lengthen the time to hatching. DDT exposure causes weaker swimming which can affect food gathering and the ability to escape predators. Exposure to pollutants can cause difficulty in maintaining position in the water column, causing decreased feeding and increased predation. Many effects channel back to a few compensatory responses. Most sub-lethal effects are biochemical. Fish appear to trade certainty of death at one stage for decreased survival at the next. Integrative Mechanisms — Growth and condition factor are two easily measured parameters which relate to several of the aforementioned compensatory mechanisms. Increases in growth rate or condition are taken as evidence that compensation is occurring - evidence that decreases in competition had permitted increased growth, and greater size-at-age would then increase fecundity or decrease predation, cannibalism, mortality from stress, or age at maturity. Because competition, predation, cannibalism, and stress mortality are difficult to quantify, density-dependent growth in length or weight is often treated (in the literature and in this report) as a compensatory mechanism in itself. Density-dependent survival is a general term for a compensatory process consisting of several potential mechanisms: predation, cannibalism, disease, parasitism, starvation, and stress mortality. Similar to the growth rate- condition factor situation, survival is sometimes treated as one compensatory mechanism because its components may be very difficult to separate or measure. Many fish populations are believed to be stabilized by density-dependent mortality before recruitment (Gushing 1977). General Concepts of Compensatory Response Compensatory processes and population regulation differ when a community or an interacting system of stocks or subspecies , rather than one population, is involved. One species may compensate better than, or instead of, another species. A different set of processes may control populations after a perturbation than controlled them before (Goodyear 1980), which is sometimes due to community changes, sometimes environmental changes. The new processes prevent an impacted population from recovering its former level of abundance and cause it to be stable at a new decreased level. Holling (1973) termed such discrete equilibria "domains of attraction." He cited the example of Great Lakes fisheries: a period of intense exploitation with high harvests followed by a sudden, precipitous drop in populations. Although fishing pressure was then relaxed, populations did not recover to former levels of abundance; this was not entirely attributable to environmental change (Holling 1973). Thus, it is apparent that extent and mechanisms of compensation after catastrophic impacts may not be of the same order as compensation for a chronically exploited species. Failures to recover from impact are part of our results and help determine the limits of compensatory reserve. Compensatory reserve is defined quite precisely by Goodyear (1977) as the difference between a stock's existing compensation ratio (which is related to fecundity and survival of exploited and unexploited populations) and the critical limit of the compensation ratio beyond which the population tends toward extinction. Compensatory reserve is related to resiliency of a population, its ability to withstand additional stress. Resilience is decreased by added mortality, making a stock more vulnerable to chance destruction by environmental change or fluctuations in mortality factors already present (Ricker 1977; Goodyear 1980). Evidence of compensation (e.g., high growth rates or fecundity) may therefore indicate a population with already-reduced resilience (the cause may be unknown), so that a higher exploitation rate is unwise. More complex systems, those with greater numbers of species, tend to have less variable, more resilient populations (Solomon 1949). Holling (1973) uses resilience in a broader sense: "a measure of the persistence of systems and of their ability to absorb change and disturbance and still maintain the same relationships between populations or state variables." Holling also defines stability as the ability of a system to return to equilibrium after a disturbance. Under these definitions, a highly unstable system, one which fluctuates widely between domains of attraction, would also be resilient if it always maintains the same relationships (e.g., spruce budworm forest community) . Great Lakes fisheries provide an example of stable but non-resilient systems (Holling 1973). In our report, we will use the narrower definition of resilience as the ability of a population to withstand stress without catastrophic collapse. CATASTROPHIC IMPACT For the purposes of this investigation, we defined a catastrophic impact as an event which removes some limitation(s) on a fish population, such as food or space. Thus we included new species introductions, as well as mass fish mortalitieSr as examples of catastrophes. (The catastrophic event as defin€5d here applies only to the species introduced, not to other species already present in the system.) In either case, resources are suddenly relatively more abundant than before, or species may be released from competitive or predatory pressure. In some instances, introductions illustrate species expanding numerically when they fill a "vacant niche," or take advantage of an unused food resource, exemplifying compensatory ability after a short-term event. Catastrophic events have various means of impacting fish populations (Table 1). These events may kill only a particular age-group; e.g., drawdown may kill demersal eggs or relatively immobile larvae inhabiting shallow water, while catastrophic overfishing removes adult fish. Such changes in the age structure of the population have implications for its potential compensatory response. Similarly, events with lingering effects such as toxic spills, where the substance may remain in the system for years before decomposing or flushing out, may inhibit or prevent a compensatory response. Other species may take advantage of an impact to one species, and compensate at the expense of the impacted species. Thus changes in community composition, or a different domain of stability (Holling 1973; Goodyear 1980), may result from a catastrophic impact. These implications will be considered in species accounts and applied to groups of species in our discussion. Catastrophe theory (Zeeman 1976; Jones and Walters 1976; Walters et al. 1980) offers a concept to understand relationships between fish stocks and their controlling factors. Exploited fish populations may be operating close to a "cliff edge" so that slight increases in fishing mortality or natural mortality such as predation would cause a stock collapse. This has occurred in the Atlantic herring {CI u pea harengus hanengus) fishery. The catastrophe theory concept was developed by Walters et al. (1980) for lake trout (SaheJfnus namaycush) . Fishing and sea lamprey (Petnomyzon maninus) predation decrease the domain of stability of spawner biomass; i.e., the upper equilibrium biomass moves closer to the lower critical biomass. When these biomasses reach the same level, the stock collapses. With decreased fishing effort, lamprey control, and some stocking, the fishery recovered, as predicted by the model. Catastrophe theory is consistent with the existence of critical depensation, where an infinitesimal increase in fishing effort leads to an unstable state followed by extinction. This would result when mortality increases as stock levels decline (which may be true for schooling fishes) so that a depensatory relationship exists between stock and recruitment at low stock levels (Clark 1974). When fishing has caused a population to lose its resilience (Holling 1973), random fluctuations in reproductive success or in the environment may precipitate a stock collapse (Clark 1974). Critical depensation would also occur when factors that are normally density-dependent (e.g., predation, parasitism) begin to act more intensely as the subject population decreases and cause a catastrophic decline. An example is the predicted collapse of Lake Michigan alewives (AJosa pseudo/ianengus) due to continued stocking of salmonid predators (Stewart et al. 1981), which continued targeting alewives even when alewif e populations began to decline x: u •i in (I) •H in a O m 0) rH .Q (1) 4J in • c in 0) rd O Qi U ••^^ u Oi ^ o o U iH 4-> rH in O ro in U 4J U 0) ro o s in -H O -W in a u o •H i^ in in •H (d U 4J 0) m +j u u (d M-i v* o x: in O 0) rH • 6 rH fO < M H ra in 4-1 u 0) ^-4 IM 0) en •H i o O 4J tp in a> rd 4J +J rd rd U o 4J 4J c c rd in rd •H 0) a a-H •H 4J M in -H •M (^ •p u a 0) g 0) M-i e a tn in in M-i 0) (U 0) rd u 0- in >i N >i o >i iH X) •H in in in in in en S a 0) S §:. in 0) rd >t >i >i c c c rd rd rd § § rd M D 4-i rd rd u •P 0) •H x: jc XI 4J in (d -H -H x: 0) : -p in u H •H o in 0) •H U 0) cu in • in c o •H ■p u ■s •2 0) I 0) x: 4-» rd 0) in •« 0) rH X: rH Oi-H o ^ 04 M XJ rd 0) in P S c: u H o ^ +j ■^ in x: en D O in M 0) rd in Ti OJ -H _ D^ 4J 5 E rd TJ cu a) x: 4J -H =J fO M-l C/J u ^ C •H 4J o u T3 C 0) CO x: >i o u u O 4J Oi in (D as as as U u o J §•'8 O *J u C W D ro x: J3 D in U in -H D D «M g O m in "8 /-N O O •8 /-*\ in x: in u:: S in c C ro in 0) •H •H •H i >< Q) «M TJ O TJ 4-» TJ +J 'O O >«./ "w •H \^ •H v-^ x: r-l M x: 0) 4J 0) ■H Q) •H 0) 4-i ^ iH X) X3 X3 X3 X3 rH >1 CO >i ro >i ro >i ro § 0) x: § 4-» in § 4-f in g 0) x: o o c c 2 ^ 9 in in 3 in 0) * >i D ro ro •H -H in U Cn u 4J 0) C 0) fH >• >t cu 9 a 9 § § in o w TJ 0) >i a> (0 iH iH rH rH ro ro ro ro i-i u M Ui a a D 9 4J 4J 4-1 4J ro ro ro ro a C C C x: x: x: x: in in in in •H •H •H •H «M M-4 *M H..4 CP c o c •H -M O x: •H 0) in d> 4J in •H C ro a (M •H rH fO M TS 9 fH 0) ro Q4rH > 0) o O o »H Q4 U c c: 0) -H >i 'O ro C s ^ ^ c 0) c: -H c in CL-H >, 4J •H 0) 4J M (D •H •H 4^ TJ U in 4J +J •H 1 ro 4-> ro 0) ro in >i ro a) Q4 ■g rc{ ■P in in in (^ •H 4J C -H Q u ro in fi 0) Q u 04 0.1 C 0) Q4 0) 0^ oT Q ro TJ 11 (J. F. Kitchell, University of Wisconsin - Madison, personal communication; Great Lakes Research Division unpublished data) • Other catastrophic impacts, such as experimental manipulations, toxic spills, and species introductions, are generally not described by catastrophe theory because these events are not a continual part of the system (as fishing or predation are), and their "big effects" are from big causes, not small causes as expressed by Ricker (1963); there is no discontinuity or "cliff edge" in the behavior of the system. Predation, overfishing, competition, disease, and parasitism are all potential catastrophes which, however, usually do not cause a catastrophic impact; when they do, it is a special case. 12 METHODS LITEM.TURE SEARCH We began our search of the primary literature by examining recent (1979- 1984) abstract listings and indexes: Biological Abstracts, Sport Fisheries Abstracts, Current References in Fish Research (Cvancara), and Science Citation Index. The most productive subject keywords included population studies, mortality, introductions, species names, and causes of mass mortalities such as toxicity, winterkill, and pond experiments. Several useful papers were found in "Compensation in Fish Populations: A Partial Bibliography" compiled by T. Gregory of Science Applications, Incorporated. Many compensation-related studies were discovered in literature cited sections of reference works and classic studies. The most fruitful of these were Ricker (1954), Beverton and Holt (1957), Gerking (1967), LeCren et al. (1977), Gulland (1977), McFadden (1977), Clepper (1979), and Goodyear (1980). Science Applications, Incorporated, conducted a computer search for us using keywords we supplied (Table 2). This search resulted in a list of 139 papers, about 20 of which were pertinent to our subject. We checked literature cited sections from all papers obtained from all of the above sources, resulting in a branching network of literature. Approximately 800 papers were viewed from the primary literature, about 300 of which proved to be useful for this project. Gray literature was investigated primarily through personal contacts. We spoke to industry, academic, and research biologists in attempts to obtain unpublished data and learn of obscure studies. Many data sets exist which could complement published studies or stand alone in demonstrating compensation; however, many are not summarized and are in too raw a form for us to use them given our limited time and resources. Any such data sets we discovered are mentioned in species accounts for future reference. We visited the libraries of two government agencies in our area, the Great Lakes Fishery Laboratory of U.S. Fish and Wildlife Service and the Institute for Fisheries Research of the Michigan Department of Natural Resources, in an effort to obtain unpublished studies. GRGANIIZATIGN OF RESULTS One of the results of this project was the annotated bibliography. Appendix 1. Selected papers were critically reviewed for appropriate methods (sampling gear, experimental design, data analysis) and applicability of the data to compensation studies. For each entry, the bibliography includes the citation, keywords, abstract, selected data, and our review. Keywords were selectied to facilitate use with an editing computer program so papers on a particular subject could be retrieved. Keywords included fish species, geographic region, type of catastrophic event, and compensation mechanism(s) . 13 TABLE 2. Keywords used for computer search done by Science Applications, Incorporated. All combinations of these With these fish kills toxic spills winterkill storms temperature or thermal shock catastrophe population collapse overfishing die-off competitive interference drawdown experimental fish kills fish toxicants drought red tide El Nino salinity change disease predation parasitism competition compensation compensatory growth compensatory mortality compensatory survival density-dependent growth density-dependent mortality density-dependent survival fecundity sex ratio sexual maturity carrying capacity abundance growth mortality survival cannibalism condition factor disease predation parasitism competition Any literature useful for quantifying compensation after a catastrophe was read and assimilated. Table 3 presents all studies used in this report where catastrophes were followed by data documenting compensation. For this report, a number of separate species accounts were written (see CASE HISTORIES), although some species were grouped. Species are addressed in taxonomic order. Examples of catastrophic mortalities are summarized. Compensatory mechanisms are listed and, whenever possible, quantified. Cases where populations were decreased beyond their ability to recover were also included to demonstrate the limits of compensation. We have reported extreme values of growth, fecundity, and some other life history parameters, as such values may reflect compensatory responses. Given that a species exhibits a compensatory response, a catastrophe that drastically lowers population density probably induces a greater response than a small impact; thus extreme values may result from extreme impact. The literature often demonstrates the operation of compensatory mechanisms which are difficult to quantify or not measured for the species in question - competition, predation, cannibalism on larvae, stress mortality, early survival. We report these qualitatively when no quantitative data are available. 14 The SYNTHESIS section includes an overview of the results ^ discussion of studies on other species, and general topics • We looked for patterns in compensatory responses in an attempt to functionally group species, to make generalizations and predictions about species for which data are not available. Summaries of data gaps are deferred to the SYNTHESIS section. 15 TABLE 3. Published studies of catastrophic mortalities to fish populations with information on subsequent compensation used in this report. Category and catastrophe Author ( s ) and date Species Experimental Drawdovm Pond or lake manipulations, experimental fishing Toxicant treatments Toxic spill or acidification Species introductions or new impoundment Johnson 1970, 1971 Bagenal 1977 Healey 1980 Johnson 1977 Lawler 1965 LeCren et al. 1977 Priegel 1971 Ricker and Gottschalk 1940 Rose 1957 Rose and Moen 1952 Swingle 1949 Beckman 1941, 1943 Grice 1959 Higley and Bond 1973 Huish 1957 Phillippy 1964 Torblaa and Westman 1980 Beamish 1976 Ryan and Harvey 1980 Svardson 1976 Aim 1946 Bodola 1966 Bryant and Houser 1968 Eraser 1978 Jackson 1957 Ki Iambi and Baglin 1969 Mahon and Ferguson 1981 Miller 1957 Threadfin shad Eurasian perch, northern pike Lake whitefish Yellow perch, walleye, white sucker Yellow perch, northern pike Eurasian perch Freshwater drum Yellow perch, walleye, others Gizzard shad Yellow perch, walleye, freshwater drum Gizzard shad Rock bass Yellow perch, others Chinook, coho salmon Gizzard shad Threadfin shad, gizzard shad Sea lamprey Yellow perch Yellow perch Yellow perch Yellow perch Gizzard shad Threadfin shad Yellow perch Gizzard shad Threadfin shad White sucker, minnows Alewife Patriarche and Campbell 1957 Gizzard shad 16 TABLE 3. Continued. Category and catastrophe Author (s) and date Species Smith 1968 Wells 1977 Weather-related catastrophes Winterkill Beckman 1950 Bergstedt and 0' Gorman 1983 Kohler and Ney 1981 Siler et al. in press Brovm 1968, 1972 Elrod et al. 1980 Hanson and Waters 1974 none Abrupt temperature changes Storms and floods UpwellingSf drought, red tide, El Nino, estuary salt and freshwater mixing Overfishing leading to Anderson and Smith 1971 population collapse Brown 1970 Parsons 1970 Density-dependent agents acting in depensatory way Disease Parasitism Predation Competition Bardach 1951 Craig 1980 Van Oosten 1947 Bur rough aiid Kennedy 1979 Forney 1971 none Alewife Alewife, yellow perch Bluegill, pumpkinseed, yellow perch Alewife Alewife Threadfin shad Alewife Alewife Brook trout Lake herring Bloater Walleye Yellow perch Eurasian perch Rainbow smelt Roach Yellow perch 17 CASE HISTORIES INTRODUCTION We chose to concentrate our efforts on a limited group of fish species in order to maximize our results and be certain of thorough analysis for the species selected. We used several criteria in choosing species s (1) various life history strategies, (2) geographic locations, and (3) taxonomic placements were to be represented (Table 4); and (4) species chosen were those for which information was known to be available, including species already known to exhibit mass mortalities or wide fluctuations in abundance. TABLE 4. Characteristics of species chosen for this report. F = fish (piscivorous); P = planktivorous; = other (combination or benthos); r and K = r- and K-selected species. Life Geographic location Species history Food of represented in or group strategies adults this report Family Alewife r P Great Lakes, Virginia Clupeidae Threadfin shad r P Southeast , Southwest Clupeidae Gizzard shad r P Great Lakes, Southwest, Southeast Clupeidae Lake'whitefish K Great Lakes, Canada, far North Salmonidae Pacific salmon K F Northwest Salmonidae Rainbow smelt r Great Lakes Osmeridae Yellow perch r Great Lakes, Midwest, New York, England Percidae Walleye K F Great Lakes, Midwest, New York, Canada Percidae Freshwater drum r Great Lakes, Midwest, South Central Sciaenidae ALEWIFE Introduction The alewife (AJosa pseudoharengus) is native to Atlantic coastal waters of the United States, but enters fresh water to spawn and has established landlocked populations in several areas. Also, the alewife has been purposely introduced in lakes and reservoirs in order to exploit unoccupied niches as a pelagic planktivore and to prqvide forage for sport fishes. Alewives inhabiting coastal waters, inland lakes, and the Great Lakes have different 18 population and growth characteristics, relationships with other species, and vulnerability to mortality factors, all of which may affect compensation by alewife. Catastrophic mortalities have not been documented for coastal populations of alewife. There are four possible reasons: (1) fishing mortality has not been excessive or has been compensated for rapidly, (2) alewives evolved in salt water, so are not subject to osmotic stress there, as oppos€5d to alewives in fresh water, (3) water temperature fluctuations are more common and of greater magnitude in bodies of fresh water than in the sea, and (4) alewives may be more controlled by predation in the sea, leading to better growth and condition. Therefore discussion will focus on freshwater alewife populations and the mass mortalities common to them. Alewives in Lake Ontario may live up to 10 years (Bergstedt and 0* Gorman 1983). Less than 1% of alewives trawled from Lake Michigan in 1964-1970 were more than 6 years old (Brown 1972), although annuli may not form when growth is suppressed (Bergstedt and O'Gorman 1983), so some alewives may be older than their scales indicate. Fecundity in Lake Michigan, 1964-1965, was 11,000 to 22,000 eggs, increasing with length (Norden 1967), but anadromous alewives produce 60,000 to 100,000 eggs per female (Belding 1921 cited in Norden 1967). Survival is only about 1% during the first 3 months after hatching (Mansfield and Jude in press). Freshwater alewives mature at age 1, 2, or 3 in various waters (Nigro 1980). The high fecundity, early maturity, and mortality at early life stages of the alewife imply high potential for increase and high compensatory capability. Catastrophic Mortalities Great Lakes — Alewives became abundant in Lake Ontario in the late 1800s and suffered a number of die-off s from 1876 to 1928 (Pritchard 1929). Although the Welland Canal was completed in 1829, alewife did not colonize the rest of the Great Lakes immediately. In each of the upper four lakes, increases in alewife abundance did not occur until predator abundance declined: for Lake Erie, walleye (Sti zostedfon vitneum vitneum), blue pike (S. vitneum gJaucuni), and sauger (S, canadense) declined in the late 1920s? and in the other lakes, lake trout declined in the 1940s and 1950s (Smith 1970). This appears to exemplify compensation after release from predation, but data are insufficient to quantify predation during those years. Supporting the predator-alewife relationship. Lake Michigan alewives have declined in abundance since about 1976. This is probably attributable to continued salmon stocking and apparent preference of salmon for alewife prey, even when other prey are more abundant - a depensatory relationship (Stewart et al. 1981; J. F. Kitchell, University of Wisconsin - Madison, personal communication; Great Lakes Research Division unpublished data) . The intrinsic rate of increase for alewife is demonstrated by the explosive increase in Lake Huron populations during the 1950s (Miller 1957) cind in Lake Michigan populations during the early 1960s (Smith 1968; Table 5). The increases could have been accomplished through several mechanisms, including increased survival of young alewives, decreased competition, predation or cannibalism, or increased growth, which are all 19 interrelated. Colonization of Lakes Michigan and Huron by alewife is analogous to a population increase after a catastrophic impact; in both cases, limiting factors are decreased or absent, allowing exponential increase of the population until it approaches carrying capacity. Capacity for alewife in Lake Michigan was reached or exceeded during the mid-1960s, as indicated by an increase in average age of stock between 1965 and 1966, failure of the large 1964 year class to increase adult stocks in 1966 (Smith 1968), slow growth of adults, and the mass mortality in 1967 (Brown 1968, 1972). Several billion alewives, about 70% of Lake Michigan's population, died in the 1967 mortality (Wells and McLain 1972). Alewives compensated after the mortality by increasing growth in length and weight, and increasing survival, shown by increase in average age (Brown 1972). Growth changes occurred primarily in age-classes 3 through 5, and possibly later. Ages 3 through 5 grew 7.1 to 23.7% faster during 1967 than the 1962-1969 means (Brown 1972). Average weight increased 16-26% immediately after the die-off. Females increased length more than males, but males put on relatively more weight than females (Brown 1972), possibly suggesting that fecundity did not increase greatly. Mean age was 2.6-3.0 yr in 1965-1966, increasing to 3.2-3.8 in 1968- 1970, reflecting improved survival (Brown 1972). Ages of these alewives could have been incorrect because suppressed growth in turn may suppress annulus formation (Bergstedt and O'Gorman 1983). Poor condition prior to the die-off, because of high density, caused alewives to be more vulnerable to low- temperature stress, and conversely, improved condition after the die-off increased alewife resistance (Brown 1972). Another example of density- dependence is the apparent inverse relationship between recruitment each year and abundance of adults during fall of previous years; however, variability obscures some of this relationship (Brown 1972). The Wisconsin Department of Natural Resources, Wisconsin Electric Power Company, and U.S. Fish and Wildlife Service have further unpublished data on Lake Michigan alewife populations. A die-off similar in magnitude to the 1967 Lake Michigan die-off occurred in Lake Ontario during the winter of 1976-1977. As many as 75% of the Lake Ontario alewives died, mostly the youngest age-groups (Elrod et al. 1980). The 1976 year class was virtually eliminated, and a strong year class was produced during 1977. The abundant 1977 year class appeared in 1979 spring trawl catches, where 67% of the alewives were age 2, compared to 7% and 0% in 1972 and 1978 catches respectively. Alewife adults increased sevenfold from 1978 through 1981 (Bergstedt and O'Gorman 1983). Condition factor and growth showed compensation during this period, both being poor the year before the die-off and increasing markedly during 1977, declining again in 1978. During 1977, age-groups 2, 3, and 4 increased in length by increments which were 2-3 times the calculated increments of the previous year (Table 6). Alewife of age-groups 4-6 were generally longer (up to 11 mm) in 1979 than those of the same ages in 1978 because of the large growth increment in 1977. Elrod et al. (1980) had expected a small die-off in spring 1979, but a substantial mortality occurred. They believed the hot, calm weather precipitated the die-off. If so, it is an example of density-independent factors exacerbating mortality and relates to other studies suggesting a combination of density-dependent and density-independent contributions to 20 rH D> m V) ^ -H t7» u c •s M -H c: 0) TJ in s ^ 3 O »H O u jC (0 (0 § II OS -H x: 0) ro •-3 0) •H 0) o 0) u c: in M .J 3 S o o 0) en ««: 0) o A •M C M rH (d 0) 3 TJ i3 (0 0) U pQ ^% S S? •s < 2 ^ ^ (C m C C in i m 0) •H c TJ •O II 3 0) c: c C>J O x: 3 •fH vo >< 4J ^ o\ ro rd w Q) m u inr^^inrooor-rorOfHino r^r^inininfOvooo^vochfH LOO^iH^fOfOiHOroCJOJ fHiHracMinvofncTicNfOin rH H iH iH H vo 04 »H ra O 00 00 in ro in rH r- o CM in m ro "i* vo i-i fH i> in CM r* vo ^ O a\ ^ ^ vo 00 00 r- H» CM m m in 'O ro rH ^ ^ CM r-l H OOO^VOrHCMVO^O cMr*invofor^^roo\ O m rH o\ iH r* ro fo i^ o o o O CM CM CO O O O O iH ^ o c^ in ro o o rH rH vo o\ r- o CO Ch o^ cj\OrHcMro^invor*-ooo\o invovovovovovovovovovor*" o^CT\CJ\o^o^o^o^o^o\o\o^o\ (0 TJ •8 x: in •H rH JQ 3 a c 3 in 0) •H M 0) x: in •H b rH CO •H U U 0) B • • o /-*^ /-N ^ CM CO r- vo O 3 cd r* x: 0) ^ 4J u o •H 3 u s CQ 03 U1 (0 JH 21 catastrophic alewife mortalities. Winter mortalities again increased in 1981- 1982, but were small in 1982-1983, which was a very mild winter (Bergstedt and 0* Gorman 1983)* Unpublished data by the same researchers show record numbers of alewives in Lake Ontario in J.983, followed by a die-off of about 50% that winter, the largest decline since 1977. During fall 1983, adult alewives were in worse condition than they had been in spring, suggesting a die-off was imminent. Alewives 150-179 mm long averaged 29.7 g in fall 1983, compared with 32.5 g in 1982 to spring 1983 and 41.1 g in 1977 (Bergstedt and 0* Gorman 1983). In addition, water temperature during mid-April 1984 was the second lowest recorded during 1979-1984. In summary. Lake Ontario alewife data show compensation in growth and condition factor, and probably also survival of early life stages as evidenced by large numbers of age-0 alewives. Small Freshwater Systems — Alewives were introduced to Claytor Lake, an impoundment in Virginia, in 1968-1969. Virginia is near the southern edge of the alewife' s native coastal range (Scott and Grossman 1973), and this lake contained the first southeastern landlocked population (Kohler 1980). Die-off s occurred regularly during the winter, but were apparently not a large fraction of the population. Growth rates during 1971 were fairly rapid compared with Great Lakes alewife populations (Boaze and Lackey 1974; Table 6). A massive die-off occurred during winter 1977-1978 (Kohler and Ney 1981). Although percent mortality was not estimated, no alewives were collected during intensive sampling in spring 1978 (until mid- June), implying a substantial proportion of the population was lost. Most alewives collected in 1979 were age 1, showing a strong year class was produced despite the reduced stock (Kohler and Ney 1981). Alewife growth, when compared between 1976-1977 and 1978, increased significantly for age- groups 2 and 4. Kohler and Ney (1981) discussed reduced intraspecific competition as the reason for growth compensation and presented zooplankton data as evidence. Alewives selected zooplankters >1 mm, and large zooplankton species were more abundant in summer 1978, just after the die-off, than in summer 1979 (Kohler 1980). Alewives as small as 35 mm, young of the year, selected large zooplankton (Nigro 1980). These data are similar to results obtained for Lake Michigan by Wells (1970) and for Connecticut lakes by Brooks and Dodson (1965) showing alewives change the size structure of zooplankton populations. Kohler and Ney (1981) did not present condition data, but their results concerning growth and strength of the first year class after catastrophe agreed with results of Great Lakes studies. This study demonstrated the extreme resilience of an alewife population that was drastically reduced. Causes and Effects of Mortalities — Proposed contributing factors to mass mortalities of alewife include warm temperatures, cold temperatures, sudden temperature changes, fungus iSapnolegnfa) , iodine deficiency, electrolyte imbalance, and high densities of alewife (Graham 1956; Brown 1968; Smith 1968; Colby 1971, 1973). Sapnolegnia infected only 20% of alewives sampled during the 1967 die-off, and was not believed to be an important factor (Brown 1968). Christie (1973) stated that 22 TAIJLE 6, Growth increments (mm) by age-group for alewives before and after catastrophic mortalities at various locations. Location, source and time Age-group 1 2 3 4 Landlocked, freshwater populations Lake Michigan Brovm (1972)1 1965 1966 1967 1968 1969 97/97 98/99 99/100 92/94 -2.2/-2.3* 2.6/-0.7 3.5/1.9 2.4/4.1 -5.9/0.4 0/-2.6 3.9/6.2 7.1/9.2 -3.2/-13.8 -7.7/-16.4 -10.3/-15.0 -10.3/-7.0 9.2/19.0 9.2/18.0 8.0/5.0 -23. -10. 16 16 4/-14.5 4/-10.5 .9/23.7 .9/19.7 0/15.8 Lake Michigan Norden (1967)^ 95 46 19 13 Lake Ontario Elrod et al. (1980) 1975 1976 1977 1978 91 ND 91 ND 36 37 ND 56 13 13 35 ND 9 8 31 8 8 7 28 11 Lake Ontario Graham (1956) ^ 77 57 22 5 23 Clay tor Lake, Virginia Kohler and Ney (1980)1 1976-1977 1978 163 160 19 57 18 25 7 19 Seneca Lake, New York Odell (1934)1 68 76 7 17 4 Lake Hopatcong, New Jersey Gross (1959)2 76 32 42 23 TABLE 6. Continued. Ag^-group Location, source and time 1 2 Landlocked, freshwater populations Cayuga Lake, New York Rothschild (1965)2 102 19 Echo Lake, Maine Lackey (1970) ^ 114 61 Anadromous populations Hamilton Reservoir, Rhode Island Richkus (1975) 1970 106 1971 91 1972 102 Long Pond, Maine Havey (1961) 135 86 53 28 13 1 Back-calculations from scale analyses. 2 Percent deviation of increments from the 1962-1969 means. Male/ Female. 3 Length-frequency analyses. severity of mortality was probably related to water temperature, and alewife abundance may not necessarily cause die-off s, but die-off s would not be noticeable during periods of low density. Thus alewife die-off s are most likely density-independent, except at periods of such extremely high density that condition factor comes into play. Poor condition results in increased susceptibility to low temperature stress, but cold temperatures alone may cause alewives to stop feeding and die (Colby 1973). Any attempt to predict occurrence and extent of alewife mortalities must consider water temperature, which fluctuates somewhat erratically in the Great Lakes. Temperature-related mortalities should not affect the subsequent ability of alewives to compensate. Age structures of alewife populations are changed by die-off s. Usually spring and early summer die-off s consist primarily of adult alewives, and fall die-off s are of young-of-the-year or yearling alewives (Pritchard 1929; Graham 1956). This may be explained by the tendency 24 of age-groups to inhabit different water depths and strata (Wells 1968), becoming exposed to temperature gradients at different times. However, age structure changes should not affect an alewife*s ability to compensate, because sexual maturity comes at an early age (males may spawn at age 1 and females at age 2 - Pritchard 1929) . Thus even if all alewives age 2 or more died, spawning could occur the following year and the population could recover quickly due to the alewife's great resilience. We did not find a case of mass mortality from which an alewife population did not recover. So far, although Lake Michigan alewife populations are being held low by depensatory salmonid predation, alewife reproduction is still occurring (Great Lakes Research Division unpublished data) . Therefore the potential still exists for alewives to ret)Ound in Lake Michigan. Compen.satory Mechanisms Few compensatory mechanisms have been documented for alewife, although we believe some undocumented mechanisms to be operating. Those that have not been documented are age-specific fecundity, size-specific fecundity, sex ratio, egg deposition, predation, disease, parasitism, and interference competition. Fecundity probably is density-dependent for alewife, because the alewife shows density-dependent growth (although perhaps only under extreme changes) and larger fish have larger gonads, producing more eggs (Norden 1967). However, the only data found for fecundity related to density showed a downward trend in number of eggs per gram body weight for a chronically impacted Lake Michigan population during 1973-1975 (Gerking 1976). There were no data relating fecundity to age. The limited predation data and circumstantial evidence available indicate, for Lake Michigan at least, predation is density-independent or depensatory. Although competition has not been documented or quantified, the changes in Lake Michigan zooplankton and the alewife' s density-dependent growth suggest that competition may well be an important regulator and affect survival. Improving early survival is a potent mechanism for quickly increasing abundance, which alewives have been shown to do. Age at Maturity — Alewives matured earlier in Claytor Lake (most at age 1) than in Lake Michig«an (age 2), probably because of faster growth (Nigro 1980). Therefore, the de]isity-dependent growth usually observed after a mass mortality should result in a lower average age at maturity. Cannibcalism — Few studies have documented cannibalism by alewife. During 1971 and 1972 in Lake Michigan, alewife larvae were about 20% of the volume of adult alewife stomachs (Rhodes et al. 1974). Alewife abundance was fairly high during those years. In Claytor Lake during the 2 years after the die-off, only 5 out of 524 alewives collected had consumed larval alewives (Kohler and Ney 1980). 25 Some adult Lake Michigan alewife we examined during fall^ 1973-1982, contained alewife larvae >15 mm (Great Lakes Research Division unpublished data). These results represent some slight evidence for density-dependent cannibalism. Since cannibalism is documented for alewife, presumably it would be more frequent at high densities than at low densities, and possibly would be compensatory. Starvation and Mortality Resulting from Stress — These two mechanisms work interdependent ly for alewife, as demonstrated by the Great Lakes case histories. Although actual starving to death is not documented for alewife, density-dependent condition factor implies food limitations. Annual die-off s often correspond with seasonal energy lows (Flath and Diana 1985). A high proportion of alewife mortality in the 1960s was probably due to overwintering and spawning stress combined with insufficient lipid storage (i.e., feeding) in the fall (Flath and Diana 1985), possibly due to intraspecific competition (Brown 1972). The conclusion of most researchers is that poor condition due to limited food makes alewives more vulnerable to low temperature stress, and conversely, alewives grow faster after a mass mortality because of decreased intraspecific competition for food (Brown 1972; Colby 1973; Kohler and Ney 1981; Bergstedt and 0* Gorman 1983). This of course is an interpretation of the data, but it appears to fit better than other explanations such as fungus infection or electrolyte imbalance, which may be results of stress rather than causes. Summary Landlocked alewife populations are more prone to wide fluctuations and catastrophic mortalities than anadromous populations. After catastrophic mortalities, freshwater alewife populations usually demonstrate improved growth and condition of adults and increased abundance of young. The data and case histories suggest that at low densities, decreased intraspecific competition is the mechanism bringing about increased growth and survival. Although predation was not documented as compensatory in these case histories, it may be in some systems where predators select other abundant prey when alewife are at low densities. THREADFIN SHAD The threadfin shad (Doposoma petenense) was originally native to the east coast of Mexico and, without man's intervention, spread to U.S. Gulf Coast states by 1950. River impoundments apparently stabilized conditions enough for threadfin shad to invade parts of the Tennessee River from the Mississippi, and acclimatization or selection for cold-resistance permitted further expansion. Simultaneously, in the 1950s and 1960s, threadfin shad were stocked widely in reservoirs in Virginia, Georgia, Pennsylvania, California, Nevada, Arizona, Kansas, and New Mexico, although not successfully in all these places (Minckley and Krumholz 1960; Burns 1966; Beers and 26 McConnLell 1966) • Particular success and impressive invasions have occurred in Arizona and California, and the species has become established in the lower Ohio River, Threadfin shad are susceptible to cold kill, with a usual lower temperature limit of 7°C (Parsons and Kimsey 1954; McGee et al. 1977). Winter mortalities of threadfin shad are often demonstrated by increased power plant impingement (McGee et al. 1977; unpub. data, Duke Power Company, North Carolina) . The invasion abilities and recovery from winterkill of threadfin shad suggest it is often able to compensate quickly after a catastrophic event. About 200 yearling threadfin shad were introduced during May in Watauga Reservoir, Tennessee, and spawned soon afterward. By fall, large numbers of YOY came through powerhouse turbines, and many schools were observed in the reservoir (Parsons and Kimsey 1954). Even more impressive invasion abilities were demonstrated in the Colorado River. Two plantings totalling 1,020 fish were made in Lake Havasu. Within 18 months threadfin shad had populated all of the Colorado River from Davis Dam south to the Mexican border (about 370 km of river), including the Salton Sea and related irrigation canals (another 150 km - Cole et al. 1958 cited in Burns 1966). Large winterkills of threadfin shad are common in Lake Norman, North Carolina, since North Carolina is somewhat out of the shad's native range. However, some threadfin shad, mostly YOY, overwinter in heated discharges there, permitting repopulation each summer. During 1979-1982, population levels fluctuated between 200-600 million in summer and 2-30 million in spring in Lake Norman (Siler et al. in press). Up to 8 million threadfin shad per year have been impinged at Marshall Steam Station on Lake Norman, mostly during winter mortalities, and each year the population recovers (Siler et al. in press; unpub. data, Duke Power Co.). This demonstrates the potential explosive increase of this species. Similarly, in Montrose Lake, Missouri, winter survival of the threadfin shad population is thought to depend on a thermal plume (Pf lieger 1975). Although mostly YOY threadfin shad were observed to survive winterkills in Lake Norman, larger individuals appeared to survive cold better in Watts Bar Reservoir, Tennessee (Adams et al. 1982). These shad used a relatively warm spring-fed cove as a thermal refuge. As in Lake Norman, the population recovered year after year from catastrophic winterkills, although changes in size frequency and gill net catch indicated a possible decline in abundance (McLean et al. 1981). Age at Maturity Threadfin shad live no more than 4 years in central Arizona reservoirs (Johnson 1971), and few reach 2 years in the Tennessee River (Parsons and Kimsey 1954). However, age at maturity fluctuates widely within that short lifespcin. Age at first spawning is variously reported as, occasionally, young of the year (Berry et al. 1956 cited in Johnson 1971; Kimsey et al. 1957; McConnell and Gerdes 1964), 1 year (Parsons and Kimsey 1954; Johnson 1971), or as high as 3 years (Kilambi and Baglin 1969). Yearlings in Arizona reservoirs spawnecl later in the season than older fish (Johnson 1971). Late larvae may be subject to more predation or stress because of smaller size than early cohorts, but timing of food abundance relative to hatching is also important 27 in determining early survival and was not investigated. Johnson (1971) found YOY as large as some spawning fish (>49 mm), but the gonads of the YOY were not developed. Kimsey et al. (1957) and McConnell and Gerdes (1964) did not have proof of YOY spawning, but felt that late-occurring larvae and large- sized YOY were evidence for this. Gonad maturation indices showed multiple spawns during a year by individual fish, which would account for late- occurring larvae in that population (Johnson 1971). The only case of apparent compensatory change in age at maturity was given by Kilambi and Baglin (1969). In Beaver Reservoir, Arkansas-Missouri, where threadfin shad were introduced 3 years previously, they matured at 1-2 years, while in Bull Shoals Reservoir where fish were stocked 6 years before, they matured at 2-3 years. Population abundance was not given in this paper, but the Beaver Reservoir population may not yet have finished expanding to carrying capacity, so that its age at maturity reacted in a density-dependent manner. Bull Shoals Reservoir was also 11 years older than Beaver Reservoir, suggesting a potential difference in the biological community, and thus in food supply which could adversely affect shad. Age of threadfin shad may be underestimated because of timing, or complete lack of annulus formation during a season. Threadfin shad in two Arizona reservoirs formed annuli usually from March to July, depending on age- group or size, but during one year, a year class with a history of poor growth had not added an annulus by October (Johnson 1970). Fecundity Threadfin shad fecundity in two Arkansas-Missouri reservoirs increased geometrically with length and linearly with weight (Kilambi and Baglin 1969) ; thus age-specific fecundity may be a compensatory mechanism if growth is density-dependent. A lOO-mm female may contain from 6,700 to 12,400 eggs (Burns 1966). Kilambi and Baglin (1969) found a wide range of fecundities, from about 2,300 eggs for an 83-ram fish to 25,000 eggs for a 152-mm fish. Arizona threadfin shad 79-119 mm had 923 to 8,540 eggs (Johnson 1971). The latter two studies used a different minimum size for an egg to be considered mature, accounting for Johnson's somewhat lower estimates. However, Johnson did not have many large fish (>120 mm) to compare with those of Kilambi and Baglin. Johnson (1971) found gonadal aberrations of three types, all more common in larger fish: failure of one ovary to produce eggs, no recruitment ova or primary oocytes present after spawning (thus no preparation for future spawning), and gonads filled with fatty tissue. About 77% of the threadfin shad larger than 99 mm SL (131 mm TL) had fatty gonads. Thus, in some waters at least, the increase in fecundity with size is limited, perhaps a type of senescence. Threadfin shad, particularly fish over 1 year old, may be capable of spawning twice a year, as demonstrated by egg diameter frequencies and presence of larvae in late summer (Kilambi and Baglin 1969; Johnson 1971). After Saguaro Lake, Arizona, was drained and refilled, the shad population re- established itself the next season, and larvae appeared in late summer. Two 28 spawns per year thus may be a density-dependent response. The second seasonal complcsment of eggs was apparently resorbed in other Arizona lakes, and in Saguaro Lake during other years (Johnson 1971). Incubcttion Mortality Threadfin shad eggs fertilized and incubated at densities l/iam^ (eggs touching each other or in layers) all showed hatching success less than 15%. After about 40 h, viable eggs developed pigmented eyes, but only eggs in the top layer reached this stage of development. Fungal growth was frequent and denser at higher egg densities. In a large population, threadfin shad may deposit eggs in large masses which, as shown in McLean's experiments, may be subject to decreased hatching success. Thus, density-dependent egg mortality may be a compensatory mechanism for threadfin shad (McLean et al. 1981). Growth The threadfin shad does not attain a large size. Maximum size in Louisiana was about 200 mm (Lambou 1965), while fish in Arizona reservoirs attained 160-175 mm maximum (McConnell and Gerdes 1964; Beers and McConnell 1966; Johnson 1970). Threadfin shad in salt water may reach 236 to 257 mm (Kimsey et al. 1957; Finucane 1965, cited in Johnson 1970). Duke Power Company has unpublished growth data for Lake Norman threadfin shad. First- year growth varies widely (Table 7). Mean temperatures and time of spring warming appeared to control growth of all age-classes in the Arizona lakes studied (Johnson 1970), with one notable exception: a case of apparent density-dependent growth in Saguaro Lake. Growth of threadfin shad in the three reservoirs studied had been similar until Saguaro was drained and refilled in 1966. Growth of the 1967 year-class far exceeded the average because of an extended growing season, not because growth per month increased. Modal length of YOY was in the 50-60 mm TL range in all three lakes in July, but growth in Saguaro Lake continued through winter. YOY attained modal lengths of 73 mm by November and 88 mm by March 1968, while spring modal lengths stayed below 66 mm in the other two lakes. Johnson (1970) attributed the increased growth to lack of intraspecific competition. In these lakes YOY generally segregate from older threadfin shad for several months. Later in the year, YOY begin to inhabit the same areas as older fish and YOY growth stops about that time. During 1967 in Saguaro Lake, the few adult shad that had survived drainage were not a competitive threat to their offspring, thus YOY growth continued through the time when it would normally be halted, probably by competition for food. Growth of both the 1967 and 1968 year classes stopped at a "normal" time in 1968, thus stabilization was approached in 1 year. However, few 1968 YOY were collected, possibly indicating competitive suppression by the 1967 year-class. Johnson (1970) also pointed out that year-class strength and growth did not always act in a density- depend4°C water, thus it floats at the surface where shad may congregate. Gizzard shad die-off s are sometimes attributed to warm 33 temperatures or a difference between carbon dioxide tension of surface waters and the CO2 partial pressure of the air (Wickliff 1953; Miller 1960). However, the die-off s in question, at Erie, Pennsylvania and Norris Lake, also occurred during winter, suggesting that cold stress may have had some effect. There are several other possible causes of mass gizzard shad mortalities. During drawdowns of Elephant Butte Lake, New Mexico, extensive gizzard shad mortalities were attributed to nitrogen bubbles caused by bacteria, Aenobacten aenogenes. Aenobacten is normally harmless to fish, but causes a gas-bubble disease when fish are subject to stress such as crowding (Jester and Jensen 1972). Parasitism or fungus may cause mortalities, or may be a secondary infection after shad are weakened (Miller 1960). Gizzard shad were apparently avoiding the strong current of the Detroit River when they ran up an intake pipe of the Parke-Davis Company, Detroit, Michigan, and 2-3 tons were killed by the plant (unpub. report, Inst. Fisheries Research, MDNR, Ann Arbor, MI). Mass spawning mortalities were reported in two Florida lakes by Moody (1957). Gizzard shad finished spawning in May in Lake Panasoffkee, and the largest fish disappeared from the catch. In May 12% were over 356 mm TL, while in June and July only 1% were >356 mm (Moody 1957). Although Moody reports a large concurrent reduction in numbers, catch data do not appear to be adjusted for number of samples taken, which were fewer in June than May. Spills of fly ash slurry and sulfuric acid occurred a few years apart in the Clinch River near Carbo, Virginia, decimating fish populations. The slurry reduced index catches by 79% in numbers and 72% in weight, while the acid killed about 5,300 fish. Effects were noted for 120 km and 21 km downstream respectively. Gizzard shad comprised 19.6% of the catch before the first kill, and 54.5% immediately post-kill. The increased percentage suggests better ability to withstand or recover from toxic spills (and, perhaps, compensate) than the other species present, primarily suckers, sunfishes, and minnows (Raleigh et al. 1978). Gizzard shad have the ability to rapidly invade new environments, demonstrated by their increase in abundance and dominance in Clearwater Lake, a new reservoir in Missouri (Patriarche and Campbell 1957), their increase in the Great Lakes, and in many other lakes where they have "over populated" (Miller 1960) . Their potential for increase, sometimes at the expense of other species, appears to be strong. Growth Growth of gizzard shad is highly variable among sites (Tables 8, 9). The fastest growth rate was reported by Bodola (1966), for Lake Erie gizzard shad, 258.7 mm to the time of first annulus formation and 107.1 mm to the second. Subsequent yearly increments are comparable to or smaller than those at other sites. However, the first 2 years' growth keeps Lake Erie shad larger than shad at other sites except for Upper Spavinaw Lake, Florida, where the third- year increment is 130 mm (Tables 8, 9). The slowest first-year growth, 75.8 mm, is reported for Conchas Lake, but larger subsequent increments bring those fish to a size similar to fish at other sites (Table 9). In many cases, growth rate and length of life have an inverse relationship (Table 9) . 34 TABLE 8. Mean total lengths and length ranges (mm) for each age-group of gizzard shad from several locations. Location, time and source Age-group Lake Beulah, Florida (Huish 1957)"^ March 1955 227 178-267 March 1956 118 251 89-165 229-304 July 1956 101 76-127 July 1957 133 238 114-178 216-254 Elephant Butte Lake, New Mexico (Jester and Jensen 1972)^'^ 1962-1963 108 173 1964 108 155 221 1965 96 172 188 250 1966 90 162 213 218 279 1965-1966« 86 128 156 185 214 1970 94 150 183 219 254 Clearwater Lake, Missouri (Patriarche and Campbell 1957)^ 1947 (124) (36) (30) (15)' 1948 163 216 249 272 297 312 (53) (33) (23) (25) (15) 1949 152 216 249 272 289 (64) (33) (25) (15) 1950 163 249 (86) 279 (30) 304 (25) 1951 152 292 (140) 315 (23) Chickamauga Reservoir, Tennessee (Eschmeyer et al. 1944) Fall 1942 90 190 April 1942 216 35 TABLE 8. Continued. Age-group and source 12 3 4 5 Upper Spavinaw Lake, Oklahoma (Jackson 1957) 1952-1954 196 307 437 Lower Spavinaw Lake, Oklahoma (Jackson 1957) 1952-1954 165 224 264 297 325 Missouri River, Nebraska (Hesse et al. 1976) 1971-1975 139 285 Calculated from length-frequency tables; fish not aged. 2 Large mortality due to drawdown in 1964. Lengths represent entire growth season; back-calculated lengths at time of annulus formation the following year. 4 Unpublished data, R.R. Patterson, New Mexico State University. Length increments in parentheses. We found several examples of density-dependent growth for gizzard shad. Although mass mortalities in Elephant Butte Reservoir were not reflected in sudden growth increases, predation pressure did result in increased growth. As catch per unit effort of gizzard shad declined 62% over 7 years, mean weight of the fish increased in 3 years from 57 to 85 g, then leveled off at 74-76 g in following years (Jester and Jensen 1972) . Lengths during 1965-1966 were less than lengths in 1970 (Table 8), although the 1965-1966 data were from a different study (perhaps different methods), and back-calculated growths do not show the same trend. In 1948, the first year of impoundment of Clearwater Lake, Missouri, gizzard shad produced a large year-class that dominated catches from 1948 to 1952. First-year growth (age 0) was 163 mm in 1948 and was not surpassed in subsequent years (Patriarche and Campbell 1957; Table 8). However, 1949 growth of the 1948 year-class (53-mm increment) was well below average, probably a density-dependent response. Growth increments in following years were close to the average. 36 «-4 lA cin :3 c: (13 a) a (0 CM dj c C 01 ^ 1-3 u c 0) (d O M fH 0) >»-*' 4J cn cii 01 c 0) a 12 U CE4 U C • •W tfl c TJ O C -H (Q 4J (0 ^ u 0) o C »H •H 3 U O a 03 3 > w (0 4J TT D> (fl C x: 0) w iH T3 c U m (Ct (U N a N •HI •o 0> c (0 M M o o ou3ioio 00 o» N CO lO^v^lOOOiOlO d lA t» CM CI lA CO <0 O to O CO f-4 CI CI CO lA N CO CO CI CI e» CO CI CO (A CI CO e» OO e» lA lA CI d to 0> f-< «o CI ^ O O CO CO lA lA V CO ^ ^ CO d C4 CO <0COOOIOCOCOU9IOOO OCOe^lOlAlAOvi^OOOOiAOO CO 00 ^ ^ 0» e* C* lA lA CI 00 lA f-4 lA OO CO CI CO CO o o 0> CO C4 CI CIlAOCOlAUSOCIvi^CO oDcie* e^'^cod^iA CI F^ CO d CO C0 00OC«lOdOlAeiOOlAOOOlA(0a0OU9lAO^eOtAiAlAOOd<0 d<0OlA«-4OeoOdlA cotAcode-coiAS^d •N d d CO »M CO 5: U3 2 o d 2 CO lA f t* o f ^ o d !«• t* o> CO 00 CO 9. 00 w4 00 o CO CO ^ lA CO CO CO CO (A O CO lA CO CO CO ^ d d CO d d CO lA^AOaOOiAlA«AlAlAOOOlAOOO«HiAlAlAO^ COlAlAOiAOlAe^O oco^•dcoaodd_ lAusd'^rtocoOoco 1^ ^ d f-4 d ^•^^•lAlAlA^otA ^IAlAU3^0tAC<*C*toOiA(OlAC»C*OC«CO(DttCl df^d d C0«-4d d ^ d 9H ^ CO ^^okoooooooe S«^U3lAlAlAOOO A 00 00 t* C* O O O M) U3 O O O CI CI C> (• O O «A o e» ^ 00 OO '^^^'^•'^•'^•'^■vfci^OOJOJlAlA pi4fia4^4^p-«v-«^lr-4^4fiNdd 00 00 o d OOtAtAlAlACOCO WOiOOO^-^OOOOlAiA ^^»M*-**-»»i^e«idfii^»-^p^»Ndd •3 • * S ® 00^ S >j J2; ,j ^ » d 2 .^ TJ J g' © 2 ^ 01 ^^ as A OS •a 5 s s -g 3 s H iS S* £ 37 Gizzard shad growth was more rapid in a new reservoir, Upper Spavinaw Lake, than in a 29-year-old reservoir, Lower Spavinaw Lake, Oklahoma (Jackson 1957; Table 8). The mean (1952-1954) growth increments for the second and third year were particularly large compared to other sites. Jackson (1957) states that growth of the 1952 year class (the year of impoundment) was very poor in 1954, indicated by time of annulus formation, and he believed intra- year-class competition was responsible. However, he presented only mean sizes over 3 years, so density-dependent growth could not be quantified. Lake Beulah, Florida, was selectively rotenoned to remove shad four times over 3 years. The four poisonings removed 10,555, 3,589, 688, and 477 kg of gizzard shad respectively, and 411 kg of threadfin shad in the first poisoning only (Huish 1957). Samples of gizzard shad from the poisonings indicated increased reproduction and growth (Table 8) • No yearlings were collected during the first poisoning. Sizes of yearlings increased greatly from the second to fourth poisonings, and sizes of YOY increased from the third to fourth (Table 8). Condition Gizzard shad coefficient of condition in the Missouri River varied from 0.849 in April to 1.081 in October, generally increasing from spring to fall (Hesse et al. 1976). Gizzard shad in Elephant Butte Lake had a mean coefficient of condition of 0.867 for 1965-1970. Condition was highest in 1965 (0.965) and lowest in 1967 (0.814) and was not directly related to population changes as shown by catch per effort (Jester and Jensen 1972). The 1964 mortality due to drawdown may have had a density-dependent effect on the condition of fish collected in 1965, but 1964-1965 catch per unit effort was higher than any subsequent year to 1970. Jester and Jensen (1972) noted that mean condition in Elephant Butte Lake was lower than any other water, including Foots Pond, Indiana (Lagler and Applegate 1942), Herrington Lake, Kentucky (Turner 1953), the Spavinaw Lakes (Jackson 1957), and Lake Erie (Bodola 1966). No density-dependent relationships were apparent at those sites, except possibly at the Spavinaw Lakes where gizzard shad in the new reservoir achieved a given weight (225 g) at a shorter length (284 mm) than those in the 29-year-old reservoir (292 mm - Jackson 1957). Fecundity and Age at Maturity Unlike alewife or threadfin shad, the gizzard shad shows declining fecundity with age, and with size beyond 500-600 g. Bodola (1966) found precocious age-1 fish from Lake Erie to have the fewest eggs, 22,400-96,560, while age-2 fish had the most, 211,380-543,910 eggs (mean = 378,990) at fish sizes of 282-305 mm and 524-593 g. Fish of ages 3 to 6 showed declining fecundity, although Bodola (1966) took data from only 13 fish. Average fecundity of three 3-year-olds was 344,780 eggs; of two 4-year-olds, 308,750; and one 6-year-old had 215,330 eggs. The 3- to 6-year-old fish all were >322 mm and >700 g (Bodola 1966). A few 3-year-old shad were immature. Gizzard shad in Elephant Butte Lake showed a similar declining trend in fecundity with 38 age. Shad matured at 2 or 3 years. Twenty-six age-3 fish had fecundities from 28,545 to 70,874 (mean 40,500), while three age-4 fish had 27,815 to 31,396 eggs (mean 29,884). The one 6-year-old shad had 58,467 eggs, but its ovarian weight was only about 6% of total body weight, in contrast to 9-13% for 3-year-olds. Regressions using size and age of fish were all negative, indicating declining fecundity. The highest correlation coefficient was obtained for a regression including length, weight, and fecundity (Jester and Jensen 1972). Declining fecundity with age and size indicates fecundity is probatdy not a compensatory mechanism for gizzard shad, at least not in the same way as other species where attaining a larger size permits more eggs to be produced. Reproductive senescence may occur for gizzard shad as for threadfin? gonads of old fish may contain more connective or fatty tissue. The only comparison that can be drawn is the greater fecundity of Lake Erie shad than those from Elephant Butte Lake. The other notable difference is the faster growth and larger ultimate size of Lake Erie gizzard shad. Habitat, including food supply, influences growth and egg production. We have no evidence that the fecundity difference between the two populations is density- dependent. Age at maturity has not been shown to be density-dependent. Egg Deposition Gizzard shad in good condition in Melvern Reservoir, Kansas, spawned earlier in the season than those in poorer condition. Early spawners may have spawned more than once during the season (Willis 1984). This could be a density-dependent response, but complete data have not been published. Competition Like threadfin shad, gizzard shad are believed to compete with centrarchids . In pond experiments with largemouth bass and bluegill, the average weight of adult bluegill was 79 and 62 g in two ponds with gizzard shad and 121 g in a pond without shad (Swingle 1949). Additionally, during the second year in the pond with all three species, there was little gizzard shad reproduction, probably due to high densities of large shad, bluegills, or both (Swingle 1949). A comparison of two similar lakes in Oklahoma showed that five out of the six most abundant species (all centrarchids) in a lake without gizzard shad present, displayed faster growth, better condition, and a greater average length than the same species in a similar lake with gizzard shad (Jenkins 1957). The exception was largemouth bass. This effect was most notable for black crappie and redear sunfish (Lepomis mi CPOlophus) . Black crappie reached 81 mm in 1 year and 231 mm in 4 years in the lake with gizzard shad, and 114 mm in 1 year and 272 mm in 4 years in the lake without shad. For redear sunfish, 1-year and 6-year lengths were 109 and 221 mm respectively in the shad lake, and 142 and 257 mm in the lake with no shad. Coefficients of condition were 12-19% lower in the shad lake for redear sunfish, and 10-28% lower for black crappie. Average lengths were 8 mm greater for black crappie 39 and 18 nun greater for redear sunfish in the lake with no shad (Jenkins 1957). These differences could not be attributed to physical, chemical , or biological characteristics of the lakes other than presence of gizzard shad. Other species believed to compete with gizzard shad are common carp (Cypninus canpfo) and threadfin shad. When gizzard shad were selectively removed from Black Hawk Lake, Iowa, as the percent by weight of shad declined from 98.1 to 2.8%, the percent by weight of carp increased from 1.8 to 96.3% (Rose 1957 cited in Miller 1960). This suggests that gizzard shad had been successfully out-competing common carp until the removal operation. Threadfin and gizzard shad may also compete, as shown by the data of Huish (1957) and Phillippy (1964) - see THREADFIN SHAD, Competition . There are several case histories that show evidence of intraspecif ic competition for gizzard shad, primarily where removal or new introductions of shad resulted in faster growth or better recruitment than usual. After removal of 10,000 adult gizzard shad from Black Hawk Lake, Iowa in 1950-1951 and stocking of predators in 1951, about 18,400 young shad per hectare were removed in fall 1951 (Rose 1957 cited in Miller 1960). Increased survival of young due to decreased competition was probably responsible for this successful recruitment. Similarly, increased abundance and size of YOY gizzard shad in Lake Beulah was probably due to removal of many shad (Huish 1957). In two of the six Florida lakes that were selectively rotenoned to remove shad, numbers of shad increased while biomass stayed the same or did not increase in proportion to numbers (Phillippy 1964). A total of 70,777 kg of gizzard shad was removed from Lake Panasoffkee, Florida in 1 year. There was no apparent trend in catch per unit effort of gizzard shad during that time, although catch was expected to decline (Moody 1957). Relatively constant catches' indicate gizzard shad apparently compensated for the losses, but the mechanism is unknown. In Swingle's (1949) pond experiments, gizzard shad reproduced abundantly in the first year, but apparently did not in the second year when fish densities were high. The aforementioned growth rate and condition differences of gizzard shad in Lower and Upper Spavinaw Lakes probably reflect intraspecif ic competition differences between an old reservoir with an established population and a new reservoir with an expanding population (Jackson 1957). Clearwater Lake, Missouri, also demonstrated this phenomenon; during the first year of impoundment a large year class of gizzard shad was produced, which dominated gill net catches for 4 years thereafter, but showed slow growth in its second year. Seven years after impoundment, the fish populations were stabilized and gizzard shad age distributions had become more balanced (Patriarche and Campbell 1957). Predation Striped bass were introduced into Keystone Reservoir, Oklahoma in 1965 and began natural reproduction by 1970. Standing crop of gizzard shad fluctuated between 290.6 and 2,067.2 kg/hectare during 1971-1979, but there was no downward trend, despite gizzard shad representing 83% by volume of 40 striped bass food (Combs 1980) • Thus gizzard shad must have compensated for predation mortality. Gizzard shad >178 mm declined from 27% to 1% of the total catch during 1971 to 1979. Shad eaten by striped bass were 25-241 mm with a mode of 76 mm. Thus the reduction in large shad was probably not entirely a result of predation; some environmental factor may have contributed (Combs 1980). The 1953 year -class in Lower Spavinaw Lake was greatly reduced and was outnumbered by yearling 1952 fish in a fall 1953 sample. Extreme low water in 1953 concentrated the population and may have permitted predators to take an increased proportion of the gizzard shad YOY (Jackson 1957). Removal of vegetation appeared to induce fluctuation of shad populations in some lakes (Bailey 1978), possibly because shad became more vulnerable to predation or the food base changed. Summary The gizzard shad is a resilient species , undergoing wide natural fluctuations in abundance and invading new habitats rapidly. There is some evidence for density-dependent growth and egg deposition, and for population regulation by competition and predation. Fecundity and age-at-maturity are apparently not density-dependent, and data are not available for other mechanisms. Like alewife and threadfin shad, the strongest compensatory responses occurred after catastrophes when populations showed increased abundcmce of young, probably partly due to decreased competition permitting increased survival. An added dimension for gizzard shad is their faster growth and much larger ultimate size, which take them out of the size range for prey sooner than alewife or threadfin? thus gizzard shad predation mortality may be related to density-dependent growth. Such a compensatory response may be what permits gizzard shad to out-compete threadfin shad when both are selectively removed from a lake. LAKE WHITEFISH Introduction The lake whitefish (Conegonus cJupeafonmi s) is distributed throughout Canada, the Laurentian Great Lakes, and fresh waters of the northern United States (Scott and Grossman 1973). Lake Erie represents the southern extremity of the whitefish range (Hartman 1972; Leach and Nepszy 1976). This species is a valuable food fish and is commercially exploited throughout its range. Lake whitefish are cold stenotherms which require silt-free spawning grounds for successful reproduction and cold, well-oxygenated, hypolimnetic waters for summer habitat (Leach and Nepszy 1976). Under natural conditions, lake whitefish are long-lived (up to 28 years - Scott and Grossman 1973), produce multiple spawns, and are thereby buffered against times of unfavorable environmental conditions (Christie 1972). 41 Factors (all of which are in some way attributable to man) which have been detrimental to whitefish populations (especially in the Great Lakes) during the last 150 years include overfishing, environmental deterioration, predation by the sea lamprey, and introduction of other species (Spangler 1970; Berst and Spangler 1972; Christie 1972; Hartman 1972; Lawrie and Rahrer 1972; Wells and McLain 1972; Regier and Hartman 1973). Some studies have shown resilience of whitefish populations and demonstrated their compensatory potential; however, in some systems the population has collapsed despite fishing pressure which was less intense than in other lakes with successful fisheries (Bell et al. 1977). Whitefish year-class variation can be great in amplitude and apparently random in frequency making it difficult to determine whether a decline in stocks is caused by overharvesting, cyclic phenomena, or environmental deterioration (Christie 1968). Some aspects of lake whitefish biology make lake whitefish subject to severe fluctuations under natural conditions, and increase their susceptibility to man-related stress factors. Specifically, lake whitefish spawning success is highly variable (especially in the southern part of their range), and depends on a fairly narrow range of temperatures during egg development and hatching. Year-class success was not related to size of spawning population, number of planted hatchery-raised fry, or such factors as turbidity, wind, precipitation, or water level (Lawler 1965). A climatic warming trend in the Great Lakes region over the past 50 years (Beeton 1961) may push lake whitefish north of Lake Erie, the present southern limit to their distribution. Lake whitefish and other coregonines form discrete populations and during spawning, less than 5 % of the fish will stray to other, even closely adjacent, spawning grounds (Cucin and Regier 1965; Christie 1972; Colby et al. 1972). This means that a depleted stock may be replaced only very slowly by adjacent stocks and perhaps not at all if the original stock is reduced substantially below a critical level (Colby et al. 1972). The reproduction of whitefish in the littoral zone makes them vulnerable to thermal pollution, nutrient loading, siltation, increased levels of chlorinated hydrocarbons, and heavy metals, as physically or chemically toxic effluents can be concentrated by prevailing onshore winds or other physical means (Berst and Spangler 1972). Documentation exists showing that compensatory mechanisms are exhibited in lake whitefish populations although in many cases, such a myriad of influences is working on a population that it is difficult to distinguish which stimuli are bringing about which responses. Few studies document lake whitefish response to catastrophic events. Lake whitefish fisheries have collapsed and slowly recovered in the Great Lakes (Christie 1968; Kenyon 1978) and Lesser Slave Lake (Bell et al. 1977), but few data on compensation have resulted. The following presents the responses of lake whitefish to mortality, organized by compcmsatory mechanisms as we have defined them. We also include some information on other Conegonus species. 42 Compensatory Mechanisms Growth — Evidence for compensatory growth by lake whitefish when population densities were lowered was found by several investigators. In two similar Alberta lakes, Bidgood (1973) documented a higher growth rate of lake whitefish in the lake which had the lower whitefish density. Cucin and Regier (1965) inferred that periods of greater growth rates of lake whitefish in southern Georgian Bay were due to periods of decreased population density. Both in a survey of Northwest Territories lakes and in an experimental situation, Healey (1975, 1980) found that exploited populations of lake whitefish exhibited greater growth than unexploited populations. Lake whitefish growth rate increased 12 mm per year on the average when 30% of the population was removed for 2 consecutive years, which amounted to about 50% of the total population (Healey 1980). We consider this size population decrease to be a catastrophic event. Other investigators reported less definitive or even conflicting results on the subject of compensatory growth, but inconsistencies are probably attributable to other interacting factors and stresses. Miller (1947) examined two lakes in Alberta and found that growth rate of the lake whitefish population of one lake increased when rate of exploitation increased, but the converse was not true: in the other lake, a period of reduced exploitation did not decrease growth rate. In Lake Ontario, Christie (1972) found no evidence of compensatory shifts in growth, recruitment, or survival when fishing pressure had brought abundance levels of whitefish well below carrying capacity. Lake whitefish in Lake Erie grew rapidly during the 'late 1920s when they were abundant but heavily fished (Kenyon 1978). The Lake Erie stock collapsed, but data from the 1970s show slower growth despite low population levels; slow growth was attributed to habitat degradation. However, weights of whitefish at given lengths were significantly greater in the 1970s than the late 1920s (Kenyon 1978), perhaps showing compensation in condition factor. Survival — Few studies addressed the question of changes in survival of lake whitefish in response to changes in population density, but two that did indicated that this factor does not operate as a compensatory mechanism for whitefish. In Lake Winnipeg, the lake whitefish population was subjected to overfishing, yet depensatory survival was exhibited such that as the population fell, survival rate decreased from 34 to 12% (Davidoff et al. 1973). In Lake Ontario, a decrease in abundance to well below carrying capacity did not affect survival rates (Christie 1972). In the population experimentally reduced by 50%, Healey (1980) found increased abundance of age- and age-1 lake whitefish. Abundance of these year classes was still evident 4 years after the experiment, showing improved recruitment which no doubt was related to survival during juvenile stages. 43 Fecundity — Bell et al. (1977) found that fecundity was size-specific and not age- specific for lake whitefish in Lesser Slave Lake, Canada, so if growth works as a compensatory mechanism for this population, then fecundity would act as an interacting mechanism. Healey (1980) contended that whitefish fecundity may increase when populations are exploited. Sex Ratio — Exploitation of whitefish in Lesser Slave Lake brought about fluctuations in sex ratio as males matured earlier and suffered greater mortality (i.e., were exposed to the fishery earlier) than females (Bell et al. 1977). This does not demonstrate compensation. In two Alberta lakes from which whitefish were exploited at different rates, the sex ratio was proportionate, 1:1 (Bidgood 1973), but it was not clear what factors influenced this situation. Bloaters collected in Lake Michigan, however, were 99% female after heavy exploitation (Brown 1970). Although gear bias accounted for some of this female predominance, the sex ratio was undoubtedly unbalanced. Age at Maturity — In Lesser Slave Lake, Bell et al. (1977) documented that mature lake whitefish of a given age were considerably larger than immature fish of the same age, implying that maturity was size- not age-specific. Thus in conjunction with compensatory growth, age at maturity could act as a compensatory mechanism. Healey (1975) thought that heavily exploited whitefish stocks matured at both a younger age and possibly a smaller size, which would be evidence for age at maturity acting independently of growth as a mechanism of compensation. The evidence is not definitive however, and some studies indicated that age at maturity could not be considered compensatory. For whitefish in two Alberta lakes and in Lake Winnipeg, maturity seemed to be age-rather than size-dependent (Rybicki and Doan 1966; Bidgood 1973). Cannibalism — Bidgood (1973) documented that lake whitefish eggs comprised a major portion of whitefish diet from October through January in a lake that was overpopulated by whitefish; however, during December and January eggs comprised only a small portion of the diet of a less dense whitefish population in a nearby lake. Disease — No references were found relating disease to whitefish population density. 44 Predation — Predation by sea lamprey has had a major effect on the decline of lake whitefish populations in Lake Superior (Lawrie and Rahrer 1972), Lake Michigan (Wells and McLain 1972), Lake Huron (Spangler 1970; Berst and Spangler 1972), and Lake Ontario (Christie 1972). Lamprey predation was not a major factor in the decline of whitefish populations in Lake Erie, perhaps because of a relative lack of streams suitable for lamprey spawning in the Lake Erie area (Hartman 1972). Control of sea lamprey in Lake Huron brought about increased survival, an increase in modal age, and a slight broadening of population age structure of lake whitefish (Spangler and Collins 1980). According to Christie (1972), the stress exerted on prey stocks by lamprey increases as prey become numerically reduced, thus lamprey predation is depensatory. This apparently occurs, at least in part, because lamprey may turn to ever smaller prey as older, larger individuals become eliminated (by predation, fishing, or natural mortality) from the prey population, and the collapsing age structure of fish becomes incapable of adequate reproduction to maintain the stock. As an indirect effect on whitefish populations, destruction by sea lamprey of major predators of the Great Lakes (lake trout and burbot. Lota lota) allowed the flourishing of other species (alewife and rainbow smelt, Osmenus mondax) which had been controlled by the predators but which subsequently multiplied unchecked and competed with whitefish for food or preyed on whitefish young (Berst and Spangler 1972). Parasitism — Stress can make individuals more vulnerable to parasitism which in turn can cause further stress and affect density-dependent responses of fish. For example, Colby et al. (1972) found that in oligotrophic Canadian lakes, eutrophication sometimes resulted in an increased incidence of infestation of the parasite Taenia which is suspected to cause slow growth and high mortality of YOY coregonines. Intra- and Interspecific Competition — Compensatory mortality can result if young whitefish are suppressed by the adult stock. This may occur if older fish force young into marginal areas or habitats (Johnson 1976) or inhibit their normal exploratory and foraging activities (Healey 1980). Of two lake whitefish populations in different lakes, the population which was subjected to a relatively greater intensity of intra- and interspecific competition for food existed on a relatively inferior diet in terms of both quantity and quality (Bidgood 1973). Dietary differences were reflected by differences in growth rates of individuals of the two populations, although age of maturity and fecundity were about the same for whitefish in either lake. 45 Whitefish populations sometimes produce an exceptionally strong year- class (Lawler 1965). Healey (1980) propounded that the presence of such a year-class may cause a decrease in the growth rate of the whole population presumably because competition for food becomes greater. Looking at all Cisco species in Lake Ontario, Christie (1972) noted a sequence of progressive loss from largest to smallest forms. The total biomass of ciscoes harvested by the fishery stayed close to the same through this period of loss however, because as bigger forms disappeared, smaller forms exhibited compensatory increases in reproduction, growth, and productivity. Species in competition can be affected by indirect causes. Colby et al. (1972) gave the example of a lake where eutrophication caused oxygen depletion at the mud-water interface which changed the bottom fauna from animals such as Mysis, Pontoponeia, and mayfly larvae to chironomids and dligochaetes . Lake whitefish could out-compete other fish species for the food items available under the former conditions, but percids, centrarchids, and cyprinids all out-compete whitefish under the changed conditions. Another example is given by Miller (1947) who described how a competitor species of lake whitefish could thrive if the older age-groups of whitefish were subjected to increasing fishing pressure while no such pressure was put on the competitor species. Condition factor and egg deposition — No references were found relating these subjects to compensation. Summary Lake whitefish exhibit the compensatory mechanisms of growth, size- specific fecundity, age at maturity, cannibalism of eggs, survival, and inter- and intraspecific competition. Under natural conditions these mechanisms help maintain whitefish populations. Complications arise because of the complex web of direct and indirect stresses caused mostly by man and his activities. Such stresses cause imbalances and make whitefish populations more vulnerable to other stresses, in the end having a snowball effect from which recovery by compensatory mechanisms may not be possible. PACIFIC SALMON Introduction Population dynamics of the five Oncorhynchus species native to North American waters have been well studied because of the salmon's economic importance. Larkin (1977) gives a good review of their life histories and 46 mechanisms of population regulation. The sockeye salmon {0 . nenka) in particular has much published information available; thorough treatments of various aspects are in Johnson (1965) and Foerster (1968). Although the native range of the Oncophynchus species is the northern Pacific, several species have been stocked in fresh water and successfully spawned as landlocked populations. Large numbers of coho {0 . kisutch) and Chinook salmon {0. tsfiawytscfia) have been stocked in the Great Lakes, and some natural recruitment has occurred although stocking continues. Pink salmon (/?, gopbuscha) were released accidentally in Lake Superior and have spread to all the Great Lakes, with spawning populations common in northern Lakes Michigan and Huron in addition to Superior. Some life history patterns, and thus compensation potential, are different for landlocked salmon populations. Catastrophic mortalities of salmon are often related to spawning migrations upriver. Obstacles to migration prevent fish from reaching their spawning grounds, thus eliminating that year class from the stream. Obstacles may be landslides, rockfalls, or dams. Low water can be a hindrance to migration. Delays in migration may cause the fish not to spawn, even if they finally do reach the spawning grounds (Larkin 1977). Nitrogen super saturation below Columbia River dams resulted in pre-spawning mortality of summer-run Chinook and sockeye salmon. Mortalities of chinook were estimated as high as 59% below the Wells Dam in 1967 (Meekin and Allen 1974). Similarly, migration of juvenile salmon downstream to the sea, or smolting, is affected by obstacles. Dams on the Snake River delayed migrations of juvenile chinook salmon and caused direct mortality by turbine passage and gas super saturation, and indirectly caused mortality by slow migration through reservoirs and exposure to predation during low-flow years (Raymond 1979). Spawning migrations in rivers and streams by salmon have led man to pay particular attention to early life stages, in contrast to other species for which the literature is not so well established. The requirements and mortality of salmon eggs have been found to have some density-dependent relationships . Spawning Site Competition and Egg Deposition Salmon excavate redds when spawning so that percolation through the gravel, and thus oxygen supply to the eggs, is enhanced. The first arrivals after migration choose the most favorable sites. At greater densities, some fish are forced to less favorable redds and there is more chance of late arrivals digging up the redds of those who came earlier (Foerster 1968; Larkin 1977). Such results clearly do not harm the adults but have a direct effect on survival of eggs. Percent survival of sockeye salmon eggs in Lake Dalnee, Kamchatka ranged from 1% when 52,800-73,100 females were present, to 72-79% when 9,300-38,000 females were present (Krogius and Krokhin 1948; Krogius 1951, tjoth cited in Foerster 1968). Krogius (1951) attributed the 1% survival to digging up of early redds by later spawners. Semko (1954, cited in Johnson 1965 and in Foerster 1968) found sockeye salmon spawning at a high density retained. 6% of their eggs and lost 37%, thus 57% were successfully deposited 47 in the redds. Egg retention may be as low as 0.5% for sockeye (Foerster 1968). Egg retention was not density-dependent in Hooknose Creek, British Columbia (Hunter 1959). Incubation Mortality Chum salmon {0 . keta) eggs have an oxygen demand of 0.00013-0.0003 mg/ egg/h at temperatures of 0.1-8.2®C. Measurements of oxygen within gravel interstices in Nile Creek, Vancouver Island, showed that oxygen was insufficient for eggs in some parts of the stream where dead eggs had been found (Wickett 1954). Droughts, floods, and particularly silting can increase egg mortality (LeCren 1965). Floods could dislodge eggs, permitting predation, while droughts or silting prevent oxygenated waters from reaching the eggs. These are basically density-independent forms of mortality. However, at high densities eggs are more subject to oxygen depletion or fungus infection (Johnson 1965; LeCren 1965), or buildup of carbon dioxide (Foerster 1968) or nitrogenous wastes (Hunter 1959). Egg losses during incubation are greater at high densities, whether due to dislodging, predation or oxygen depletion (Foerster 1968). Accumulation of adult carcasses may increase oxygen depletion as well, due to decomposition (Foerster 1968). Losses during incubation may range from 70% to 100% (Larkin 1977). Population density in Hooknose Creek affected pink and chum salmon survival from egg to fry, while temperature, stream discharge, and sex ratio did not affect survival perceptibly. Egg and alevin mortality was 70-80% at low density (<500 eggs/ m2) and >90% at densities >1,500 eggs/m^ (Hunter 1959). Compensatory mortality during spawning and incubation is considered to be one of the important population regulators for pink and chum salmon (Neave 1953). Predation Predation on salmon fry is one of the most significant mortality factors, but is generally not compensatory. Because the predators inhabit the streams year-round, they are not dependent on salmon fry for food and their numbers are not limited by salmon fry. Thus the predators take a fixed quantity of fry as prey, regardless of fry densities, a depensatory effect (Neave 1953; Hunter 1959; Johnson 1965; Larkin 1977). When predators in a western Canada stream were destroyed, survival of fry was much higher (Lack 1954). The only density-dependent aspect of predation on fry takes place after fry establish feeding territories. At high densities (sometimes due to low water levels), fry unable to establish territories are displaced downstream and may be consumed by predators (Larkin 1977). Adult salmon in the ocean are preyed upon by sharks, seals, and killer whales, and of course by man in the sea and rivers, but predation on adult salmon has not been demonstrated to be density- dependent (Larkin 1977). 48 Cannibalism Many stocks of pink salmon exhibit alternate-year cycles of numerical dominance. Since nearly all Pacific Coast pink salmon spawn at age 2, either the odid-year spawning line is more abundant than the even-year spawners, or vice versa. One mechanism suggested for perpetuating the dominance of one line is cannibalism. Age-0 pink salmon may be abundant in coastal waters at the time the piscivorous age-l+ pink salmon pass through on their way to spawning streams. However, there is no evidence that the older fish alter their feeding habits according to abundance of the younger fish; therefore cannibalism of pink salmon is probably depensatory (Ricker 1962). Sockeye salmon may remain in fresh water for several years (Foerster 1968) , thus are potentially cannibalistic , but this is not a documented form of compensation for them. Competition and Growth Young salmon of some species establish feeding territories during stream residence and competition for food becomes a density-dependent influence on growth. Young pink salmon survived better at low than high densities, believed due to food competition (Pritchard 1948 cited in Lack 1954). Young sockeye salmon inhabit lakes rather than establishing stream territories. Some studies suggest that food abundance (zooplankton) influences growth rate of sockeye, but analysis is complicated by increased survival; more fish are present to take advantage of the favorable food supply (Johnson 1965). Rock slides and continuous heavy fishing severely reduced sockeye salmon populations in the Fraser and Skeena Rivers. Subsequently the abundance of young salmon was very low, but they apparently did not grow at an increased rate, nor did competitor fish populations increase noticeably (Foerster 1968). Thus compensation after the catastrophe was not demonstrated. Coho and chinook salmon reared in Happy Valley Reservoir, Oregon, experienced catastrophic mortalities due to rotenone and subsequently poor water quality which decreased their condition, making them easy prey for the surviving older salmon. Both species displayed density-dependent growth in length and condition after the mortalities, shown by comparing growth rates before and after rotenoning. Also, the first year class after the rotenone treatment grew faster in its first year than the next year class, at least partially due to the difference in fish density between the 2 years (Higley and Bond 1973). Juvenile coho salmon in a limited-food laboratory experiment grew faster in the absence of juvenile steelhead trout {Sal mo gafndneni), given the same food ration and total fish densities (Laarman 1969), indicating steelhead trout have a competitive advantage over coho. Kokanee salmon (landlocked sockeye) grew faster in Kootenay Lake after introduction of mysids, and while eutrophication was proceeding (Northcote 1972). 49 There is some evidence for density-dependent growth of salmon during their life at sea. Seven sockeye salmon stocks showed significant decreases in growth rate, up to 10-22% for two stocks, when large numbers were present. The most noticeable effects were early in ocean life and were believed to be due to competition for food. Total sockeye abundance in the Gulf of Alaska was at least as important as within-stock abundance in determining body size (Peterman 1984). Increased food abundance in the ocean contributes to faster growth of sockeye (Foerster 1968). Other salmon species apparently have not shown density-dependent growth at sea. Abundance of juvenile migrants may be associated with smaller size of returning adults, but the evidence is conflicting (Larkin 1977). Mortality at sea is less when smolt size is larger (Larkin 1977), thus density-dependent growth in streams may help determine later mortality of fish. Similarly, high rates of marine mortality for sockeye (probably due to predation) are associated with large smolt productions, suggesting a density- dependent response (Larkin 1977). For pink salmon, however, mortality at sea does not appear to be compensatory over the range of observed population sizes (Larkin 1977). When Snake River chinook salmon juveniles were greatly reduced in abundance by dam effects, record-low numbers of adults returned in spawning runs (Raymond 1979) . This demonstrates a lack of compensation during sea life; mortality was not density-dependent. Age at Maturity Fast-growing sockeye salmon migrate to sea 1 year before slower -growing ones; also, faster-growing sockeye mature and return to spawn sdoner. The range for age at spawning is 3 to 8 years (Foerster 1968). Heredity, however, exerts more influence on age at maturity than environmental (i.e., food) factors. Fecundity Fecundity of sockeye salmon increases with size (Johnson 1965), thus density-dependent growth would result in higher fecundity. However, the range in size and thus egg production in south Kamchatka was within 12-18% of the mean over long periods of time (Krogius 1961 cited in Johnson 1965), thus fecundity would not be a strong population regulator. Other populations of sockeye have demonstrated no density-dependent effect on fecundity (Johnson 1965). Summary Density-dependent incubation mortality related to spawning site competition, and density-dependent juvenile growth and survival due to (usually) intraspecific competition for food, appear to be important compensatory mechanisms for Pacific salmon. There is also some evidence for compensatory egg deposition, age at maturity, and under certain conditions, 50 fecundity and predation. Cannibalism is prevented for some species by the spatial separation of young and adults, and in the species where it does occur, it may not be compensatory. RAINBOW SMELT Introduction The smelts (Osmeridae) are anadromous fishes having an irregular but wide geogrciphical distribution throughout the Northern Hemisphere (Kendall 1927). The rainbow smelt, Osmerus mondax (Mitchill), is an introduced, landlocked form which is common in the Great Lakes. Regier et al. (1969) believed Lake Erie smelt populations had relatively good self -regulating mechanisms, preventing wide fluctuations in abundance. The following information concerning rainbow smelt of western Lake Superior is provided in a study by Bailey (1969): (a) The growth rate of smelt was variable; a large range of lengths for fish in each age-group resulted in length being a poor index of age. (b) Female smelt were found to grow faster than males after the second year, (c) Both sexes made their best annual growth in length during their second year of life; largest weight increases came in their third year, (d) All 1-year-old smelt were immature; whereas, for 2-year-old fish 40.7% of the males and 17.7% of the females were mature; all smelt greater than 2 years old were mature, (e) No smelt less than 127 mm were mature but all individuals in the 150-155 mm length interval were mature, (f ) Spawning males were generally shorter than spawning females and average length of both sexes decreased as the spawning season progressed, (g) The sex ratio and mean age of spawning-run smelt also varied according to the time within the spawning season, (h) Males dominated the spawning populations at the start of the run, the sex ratio was near 50:50 at the peak, and females may have been dominant as the spawning season closed, (i) Fishermen generally began to harvest smelt at the beginning of the spawning season and fishing effort declined with time as markets and appetites became glutted; as such, the males which initiated the spawning run suffered greater mortality than females; males were not caught beyond age-group 5 and females strongly dominated age-groups 4-7. (j) The average number of eggs contained in the ovaries of 10 smelt between 185 and 224 mm was 31,338. Community Interactions Present-day rainbow smelt stocks in the Great Lakes resulted from the planting of over 16 million eggs in Crystal Lake, Benzie County, Michigan, in April 1912 (Van Oosten 1937b). Smelt spread progressively from Crystal Lake to all of the Great Lakes, but this process was inhibited by time lags involved in locating suitable habitats (large bays of shallow-to-moderate depth) in the lakes, and due to resistance to colonization in the form of predation pressure (Christie 1974). According to Eck and Wells (1983) rainbow smelt comprised 18% of the 3,037 metric tons of forage fish consumed by lake 51 trout in Lake Michigan from 15 May to 1 December 1979. Christie et al. (1972) reported that offshore predators preyed on adult smelt while predators in the littoral zone ate juvenile smelt. Release from predation pressure resulted from the decline of predator populations (caused by overfishing and lamprey predation) and allowed smelt populations to increase. Eutrophication of the Great Lakes also may have favored smelt over native species. Apparently eutrophication causes destabilization or shifts in food bases or habitat characteristics such that shifts in species dominance are ongoing. Compared with native species, an invader under unstable situations faces much less organized resistance and has a better chance of succeeding provided it can accommodate to the new environment (Christie et al. 1972). The ability of smelt to accommodate to new environments is illustrated by their flexibility in reproduction requirements. Normally being stream-run spawners, smelt adapted to reproducing on limestone ledges of points (Christie 1972). Also documented is the ability of smelt to compete with or prey on other species. Smelt are widely implicated in the decline of coregonines (Christie 1972, 1974; Christie et al. 1972; Crowder 1980; Wells and McLain 1972) and in particular have been known to prey on lake herring (Conegonus antedii) larvae and probably other species (Christie 1972; Selgeby et al. 1978; Crowder 1980). Smelt are not always the superior competitor however; explosions of alewife populations may have caused declines in smelt populations in Lakes Michigan and Huron. The Wisconsin Department of Natural Resources and Wisconsin Electric Power Company have unpublished data on Lake Michigan smelt populations . Another factor which may relate to the pollution resistance of smelt was reported by Christie (1974). The sessile eggs of smelt rest on short pedicels and as such are free from the bottom where pollution or the effects of pollution may be concentrated. Mortality Factors Rainbow smelt in the Great Lakes are fished extensively by both commercial and sport fishermen. It is mainly by these fisheries that the abundance of smelt populations is measured or followed. The commercial fishery consisted mainly of gill nets set under the ice and pound nets. Sportsmen harvest rainbow smelt by using dip nets or shore seines during the spring spawning runs. Natural fluctuations of rainbow smelt populations occur due to many reasons. Mass mortalities and subsequent recoveries of rainbow smelt populations have been cited in the literature, but most such occurrences seem poorly understood, and did not have follow-up studies. Kendall (1927) reported on the sudden simultaneous deaths of thousands of rainbow smelt in Lake Champlain both in 1882 and 1883. These mortalities were attributed to various phenomena including the aurora borealis, electric storms, and windstorms but no explanation adequately accounted for the die-off s. In Lake Ontario, rainbow smelt stocks declined from 1952 to a low in 1965 but have since recovered substantially (Christie 1972). The decline was thought to 52 have been related to the amount of spring run-off, but subsequent observations made this hypothesis doubtful, and in any case, other unknown factors were probably involved (Christie 1972). Compensatory Mechanisms Growth — The most massive rainbow smelt mortality occurred in Lakes Huron and Michigan during fall and winter, 1942-1943 (Van Oosten 1947). Because fishery statistics were available from before, during, and after the mortality, we have a good chance to evaluate the compensatory potential of smelt. The mass mortality was attributed to a bacterial or viral disease which killed smelt of all ages and sizes and both sexes, starting in Lake Huron and spreading to Lake Michigan. Just prior to the mortality, fishery statistics indicated that smelt populations were very abundant, perhaps more abundant than ever before. The disease decimated the smelt populations and virtually wiped out the fishery for a few years. The commercial gill net catch dropped 61% and the pound net catch dropped 93% between the 1939-1942 means and the 1943 catch, and catch was almost nil during 1944 (Van Oosten 1947). Some signs of recovery were indicated in 1945, but smelt abundance was thought to have been less than 10% of the "pre-mortality" years. Further recovery was observed in subsequent years and stocks re-attained full abundance in Lake Huron by 1950; recovery was rather slow in Lake Michigan but a high yield was obtained in 1958 ((Zhristie 1974). Compensatory growth of smelt in Green Bay, Lake Michigan, was observed as fish in 1944 and 1945 were significantly longer and heavier than fish of corresponding age captured in the same region in 1941 (Table 10). After the mortality, smelt populations recovered more slowly and fluctuated less than alewife did after their 1967 mortality (Brown 1972; Christie 1974). Fecundity — Some evidence for compensatory fecundity of rainbow smelt was reported by Gerking (1976). When a Lake Michigan population of rainbow smelt was subject to impingement by the Zion Power Plant, fecundity per gram body weight increased for fish of all length classes from 140 to 239 mm. The most dramatic difference was shown by 160-169-mm fish, which had mean fecundities of 443 eggs per gram body weight in 1973, 540 eggs/g in 1974 and 927 eggs/g in 1975 (Gerking 1976). Cannibcilism — Yearling and older rainbow smelt in the Great Lakes region prey on conspecific juveniles, sometimes with greater frequency than they consume other fish species (Creaser 1929; Hale 1960; Gordon 1961; Price 1963; 0' Gorman 1974; Great Lakes Res. Div. unpub. data). Although the data do not document increased cannibalism at higher smelt densities, generally cannibalism is likely to be more intense when many young fish of that species are available, thus cannibalism may be compensatory for smelt. Lake Erie rainbow smelt 53 TABLE 10 • Average total length (mm) and weight (g) of pre- (1941) and post -mortality (1944-1945) rainbow smelt from Green Bay, Lake Michigan. Adapted from Van Oosten (1947) Item Age-group Males Females Date 1 2 3 1 2 3 4 Mar 1941 N 35 11 1 36 12 1 len 170 188 135 175 201 193 wt 28 40 14 31 51 40 Mar 1944 N 2 25 1 1 19 2 len 163 185* 216 160 196* 244 wt 28 42* 74 25 54* 96 Jan 1945 N 1 25 5 2 18 13 2 len 124 193* 201 140 213** 221* 267 wt 14 57** 62 17 76** 85* 139 Significantly greater than pre-mortality (p < 0.05) ^ Significantly greater than 1944 (p < 0.01). yearlings and YOY are found near the thermocline throughout summer, where cannibalism probably occurs and may have accounted for the alternate-year cycle of abundance observed during the 1960s (Regier et al. 1969). Summary Although rainbow smelt populations in the Great Lakes and elsewhere have experienced catastrophic mortalities and recoveries, little attention has been given to the mechanisms of recovery. Smelt from Green Bay, Lake Michigan, compensated after the 1942-1943 mortality by increased growth in length and weight, probably due to decreased intraspecific competition. One study showed fecundity to be compensatory, independently of growth rate. Cannibalism is potentially a strong compensatory mechanism for smelt. 54 YELLOW PERCH Introdluction MacLean and Magnuson (1977) state that percids evolved in a simple predator-prey complex in North America. Predation, competition, and other species interactions cause resource partitioning in fish communities which decreases species interactions. The cool-water niche of percids separates them spatially from stenotherins and eurytherms. Yellow perch (Penca fJavescens) and Eurasian perch (P. fJuviatflis) combined have an almost circumpolar distribution in fresh waters of the northern hemisphere. They have been one of the most intensely studied freshwater species because of their importance to sport and commercial fisheries. Yellow perch inhabit waters of moderate temperatures from large lakes to ponds or quiet rivers, and reach greatest abundance in fertile waters with large plankton blooms and rich bottom fauna. They inhabit dystrophic waters where stunting is common (Deelder 1951) to oligotrophic lakes, where they effectively compete with salmonids for available food (Fraser 1978). In more favorable habitats, populations of yellow perch are usually very stable, if they are not undergoing some severe stress, such as over exploitation, competition from an exotic species, or pollution. In most populations, yellow perch maintain their stability through the dual regulatory mechanisms of cannibalism and predation. Among density-independent factors, water temperature at the time of hatching and early recruitment has been consistently cited as important in determining year-class strength. This species has great potential for contributing to our understanding of compensatory mechanisms because of the vast amount of research that has been done and the extensive data sets that exist. Future research can build on the conceptual foundation that already exists for many of the life processes of this species. We found examples of a large number of different agents which sometimes acted catastrophically to severely depress yellow perch populations. Disease caused a 98% reduction in Eurasian perch in Lake Windermere, and a severe reduction was also recorded for Lake Mendota yellow perch; compensatory increases in growth and fecundity were noted. A die-off of yellow perch in Lake Norman, North Carolina during summer was also noted (J.R. Siler, Duke Power Co., Huntersville, N.C., pers. comm.). Man was responsible for the removal of large proportions of the perch populations in Lake Windermere (Bagenal 1970; Craig 1980) for use during the war, and overfishing reduced populations to low levels in Saginaw Bay, Lake Huron (Eshenroder 1977). Predation by alewife on larval yellow perch was hypothesized by Wells (1977) and Jude and Tesar (1985) to have depressed yellow perch stocks in Lake Michigam. Predation by walleyes in Oneida Lake acted as the agent of depensatory mortality for YOY yellow perch (Forney 1971) . The response of yellow perch to changes in competing species (rough fish removal operations - Grice 1957; Johnson 1977) was also studied. The impact of major changes in the environment, such as acidification and eutrophication, will also be discussed so that specific incidences of catastrophic declines can be 55 under stocxi in reference to the quality of the environment for yellow perch. Clady (1976) has established for Oneida Lake that a series of minor factors, rather than one catastrophic event, established the strength of the final YOY cohort produced. More detail is given under each of the sections to follow. Growth Many of the mechanisms of compensation are eventually manifested in growth by fish. Growth is also much easier to measure than, for example, competition. Thus, this section contains many examples of changes in growth which were caused by several different mechanisms. If growth was the dominant response, we report it here. LeCren et al. (1977) studied the numbers, weight, and year-class strength of Eurasian perch in Lake Windermere, England during 1941-1966. They removed large numbers of yellow perch and northern pike (Esox Jucfus) in an effort to reduce competition and predation on salmonids in the lake. Eurasian perch did not exhibit growth increases in the first few years subsequent to harvesting (Craig 1980) . Lack of compensation by the perch population was attributed to (1) Eurasian perch were also intensively harvested, and (2) competing species occupied the "niche" of the depressed perch. An exceptional combination of events, both physical and biological, was necessary before Eurasian perch were able to produce a strong year class. Boiling's (1973) concept of "domains of attraction" may apply here, as perch declined from a high population and some equilibrium status with its predator (northern pike) and competitors in Lake Windermere, to a lowered state where populations remained depressed from 1941 to 1955, before a strong year class was formed. In Lake Windermere, cannibalism by the 1955 cohort caused increased mortality for the 1959 cohort in the south basin (LeCren et al. 1977). Northern pike predation and perch cannibalism regulated populations through the 1960s and 1970s. Northern pike were able to keep the perch populations depressed because perch were already low from the war fishery. To form a successful year class in Windermere, several favorable events needed to occur. Removal of perch through trapping caused a decline in the average age of the fish and a three-fold increase in the mean weight of the perch. This* was translated into a three-fold increase in fecundity. The population in 1956 was 6% of the 1941 level so there was only a partial compensation in numbers for the decline. More perch (over two times more) fed on fish in later years (1956) than in earlier years. LeCren (1958) also found that removal of old perch via traps had no effect on YOY and yearling growth since these fish ate zooplankton, while larger fish ate benthos and fish. Food eaten by Eurasian perch in Windermere showed that adults normally did not eat many YOY perch. However, LeCren et al. (1977) found that culling of northern pike from the lake was very important, since normally northern pike ate age-2+ Eurasian perch, but when this size perch was scarce they switched to YOY and yearlings. Mergansers in Lake Windermere also ate large quantities (up to 3.6 tonnes/year). These predators consorted to depress perch populations so that a good year class was not formed for 13 years (1955) 56 after the initial reduction of perch and northern pike in 1941. There was an interesting effect of ice cover in Lake Windermere on populations of Eurasian perch « During 1947 and 1963, years of prolonged ice cover, there was a failure of the year class when females failed to spawn. Year-class strength was positively correlated with: (1) warm years, and (2) low biomass of perch present, and negatively correlated with: (1) low temperatures, (2) increased biomass of perch, (3) low ratio of spawning males to females, (4) northern pike predation, (5) merganser predation, and (6) extended period of ice cover. They concluded that the 1941 removal of perch and northern pike from Windermere caused both species to remain in a depressed state despite removal of the fishing pressure. In the north basin, there was a more drastic perch removal and northern pike predation was more severe. As a result, numbers of perch continued to fall. The first good year class was formed in 1959, some 18 years after the first removals of Eurasian perch. In the south basin, fewer perch were removed, northern pike predation was less, and a strong year class was formed much earlier (in 1949, and a very strong one in 1955) than in the north basin where predation was much higher and populations of perch more depressed. Weather was particularly important, since years of warm water temperature were associated with strong year classes. Grice (1959) found that yellow perch exhibited compensatory growth increases after severe reduction in competing species in a study of Massachusetts ponds. Rapid growth was attributed to reduced competition for food because of lower population densities. A similar response was observed in renovation of Lake Okoboji, Iowa in 1940-1951 (Rose and Moen 1952). In Oneida Lake during 1965-1971, cannibalism was not decisive in limiting yellow perch year-class strength (Tarby 1974). Bagenal (1982) noted that Eurasian perch in Lake Windermere produced a large cohort in 1955 and cannibalism, along with pike predation, combined to suppress other cohorts. In Oneida Lake, year classes of yellow perch which were most abundant as young often fared poorly as yearlings (Forney 1980). This within-cohort compensatory mortality was probably a consequence of density-dependent growth mediated by changes in zooplankton composition. Aim (1946) put Eurasian perch into perch-free lakes and into lakes with an established perch population; he found that those in the perch-free environment grew much faster. Aim (1952) also showed that cannibalism could suppress recruitment in small lakes where perch was the only species. Pond studies of yellow perch without competing species showed growth of young perch was compensatory (Schneider 1973). These ponds were fertile (higher survival rates) and without perch predators, and contrast with more natural diverse systems (Oneida Lake - Forney 1971) where predation was the principal mortality factor, and it was often depensatory. Schneider concluded that extrinsic mechanisms, such as predation or competition by other species, were necessary to stabilize perch populations so large fish are produced. Overfishing removes large proportions of the adult population of yellow perch and is another category of catastrophic loss. Spangler et al. (1977) discussed the impact of overfishing on Great Lakes stocks and found there were typical responses, such as changes in variability of recruitment, compensatory increases in growth rate, and reductions in age of first spawning. Forney 57 (1971) with walleye in Oneida Lake and Schmitz and Hetfeld (1965) with muskellunge {EsQX masguinongy) in a Wisconsin lake both showed that predation removed enough yellow perch from the population to cause depensatory mortality and increased growth of remaining perch. Sometimes density-independent processes can impact yellow perch year- class strength. Clady (1976) found wind was negatively correlated with early survival, which varied from 1.6-18.4%. Physical destruction of eggs and greater mortality of prolarvae in low temperatures were cited to account for low survival during some years. However, he felt a number of minor mortality factors, rather than a catastrophic event, controlled survival. Another environmental change, gradual acidification (Beamish 1976; Ryan and Harvey 1980), caused yellow perch to first lose competitors, then as perch populations became increasingly stressed, reduced densities of age-groups 1-3 occurred; survivors responded with increased growth. Older fish had high mortality and severe osmotic stress. Environment also impacts yellow perch and their compensatory response. For example, in Lake Memphremagog, Quebec, two basins, one enriched and the other not, had perch populations which grew at the same rate, but more were produced in the enriched one. Here, since perch were growing near maxima for the species, the increased energy was channeled into increased numbers, rather than increased growth. Kelso and Bagenal (1977) in their treatise on unperturbed systems, stated these environments were harsh and perch there were characterized by low growth and variable to low recruitment. An infestation of Eurasian milfoil (Borawa et al. 1978) favored yellow perch which had increased populations to the detriment of some other species. Finally, Leach et al- (1977), studying the response of percids to eutrophication, found they generally benefit since competitors and predators are less tolerant and food supply increases. In other studies. Wells (1977) addressed the changes in yellow perch populations in Lake Michigan. He found that yellow perch declined abruptly in the mid-1960s about the same time as the alewife invasion. The commercial fishery also hastened the decline. Reduced densities of yellow perch allowed increased growth, which resulted in the commercial fishery increasing their harvest even more. After the alewife die-off in 1967, yellow perch populations increased, but this was only a small effect. Recent papers (Wells 1985; Jude and Tesar 1985) document a sharp decline in the numbers of alewife in Lake Michigan, which was predicted by Stewart et al. (1981). They warned of this decline because of the increased stocking of salmonids. In response to the alewife decline, yellow perch populations have increased dramatically, which was hypothesized by Wells (1977) to occur because of reduced predation by alewife on yellow perch larvae. Wells (1985) chronicled this yellow perch expansion in more recent years and found the 1983 year class was probably the largest ever produced in Lake Michigan. In addition, a decline in growth of yellow perch was recorded as populations apparently hit their peak and food became limiting. Yellow perch of the 1983 year class reached an average total length in the southeastern part of the lake of only 100 mm at the end of 2 58 years; in the 1970s this length was about 160 mm. By weight, these same fish produced in the 1970s were four times heavier than those produced in 1983 after 2 years of growth. Leach and Nepszy (1976) examined yellow perch trends in Lake Erie. They found that large year classes of yellow perch and walleye were produced in 1959, 1962, 1965, and 1970 which were strongly related to water temperature. They noticed no major growth rate changes since the 1950s. Yellow perch was the dominant percid in catches from Lake Erie in the 1970s. Nelson and Walburg (1977) found that growth of yellow perch in their first summer in two Missouri reservoirs was inversely proportional to abundance. Mills and Forney (1981) found that since 1968, YOY yellow perch growth in Oneida Lake was inversely correlated with density. An increased density of YOY caused the main food (Daphnfa) to decline which in turn acted as a mechanism to depress YOY numbers, presumably through increased starvation. Density-dependent growth was mediated by Daphnia populations and year-class strength. Schmitz and Hetfeld (1965) studied muskellunge predation in a Wisconsin lake and found that muskellunge ate mostly yellow perch > 150 mm which reduced yellow perch densities and resulted in increased growth for age-1 and 2 yellow perch. A growth increase was not observed for YOY because they ate zooplankton. Competition is suspected to be an important compensatory mechanism, but it is difficult to quantify. In a study on yellow perch in Wilson Lake, Minnesota, Johnson (1977) found that yellow perch and white sucker (CatOStomus commensoni) were competing for the same resource. Abundance, growth rate, and condition factor of yellow perch increased after 85% of the weight of the white sucker standing crop was removed. Yellow perch numbers then increased, and density-dependent mechanisms (probably food limitation) operated to cause growth and condition of yellow perch to decline in later years. Fraser (1978) studied the impact of the introduction of yellow perch into oligotrophic lakes stocked with three species of salmonids. In 1962 only salmonids (brook trout, SalveJinus fontinalis; splake, 5. fontinaJis x namaycus/!} and rainbow trout, SaJmo gaindneni) were stocked in the lake. Yellow perch entered in 1967 and showed an immediate burst of growth. Thereafter, the yellow perch population increased drastically and their growth declined. Salmonid production and growth declined and food eaten by salmonids shifted to smaller items. Both inter- and intraspecific competition occurred. Similar consequences were observed by Aim (1946) who studied stunted perch in Swedish lakes. A natural reduction in perch resulted in increased growth rates. Removal of Eurasian perch from Lake Windermere also caused the growth rate of remaining fish to increase (Bagenal 1970). In general, a decline in the numbers of yellow perch results in increased growth of adults (Rutlecige and Barron 1972). Thorpe (1977), in an effort to study stunted yellow perch populations, found that with increased food, Gaffibusia, growth also increased. 59 Kelso and Bagenal (1977) reviewed percids in unperturbed ecosystems. Growth in these lakes was low, and recruitment was not regular, because these systems were usually in harsh environments. Temperature stood out as a major controlling factor. They felt that genetic factors did not constrain yellow perch growth, since fish could be transplanted from these harsh systems and grow well in a favorable environment. They concluded that these systems had low productivity and that stability of biomass appeared to exist for yellow perch . Aim (1946) noted that strong year classes restricted recruitment through cannibalism while the age-group was passing through the population. Craig (1980) noted this effect for the 1955 cohort of Lake Windermere Eurasian perch. This mechanism appears to exist for yellow perch. Ryan and Harvey (1980) studied the response of yellow perch in acid lakes in Ontario. Yellow perch in these lakes generally had low population densities; those of age 1-3 showed increased growth, while fish of age 4-9 had depressed growth. They concluded that yellow perch growth may serve as an indicator of environmental stress. There was less successful recruitment of yellow perch in more acid lakes. The mechanisms proposed to explain the increased growth of age 1-3 fish were less inter- and intraspecific competition because of lowered densities of yellow perch in these lakes. Those fish greater than 3 years old had depressed growth, which was due to increased maintenance costs for older fish in the acid medium. This osmotic stress was due to loss of sodium. In addition, there was less prey available to older fish, as the acid-intolerant groups such as minnows declined. El-Zarka (1959) examined the Saginaw Bay yellow perch fishery and found that growth declined in 1943-55 compared with 1929-30 growth rates. He concluded that the growth decline in the 1940s was due to a lack of space because, he concluded, there was plenty of food. The mean age in trap net catches increased from 3.8 years in 1929 to 4.3 years in 1955 and as well, growth decreased. Modal length declined from 8.5-8.9 to 6.5-6.9 inches. Sexual maturity occurred at a smaller size in 1955; females were mature at 5- 7.5 inches, while males matured at 7-7.4 inches. Ricker (1963), in his paper on big effects from small causes, pointed out that a 5% increase in catch can cause a 12-15 age-group stock to decline. These stocks depend on having a large number of multiple age-groups eligible to spawn to overcome the vagaries of the weather and other physical factors that can adversely affect stocks. A population that produces greatest sustained yield at 75% is close to the knife edge, and if use increases it may collapse. We saw examples of this with the Saginaw Bay fishery, where exploitation reduced the larger females to the point that recruitment failure resulted (Eshenroder 1977). Similar effects were noted for Lake Erie (Nepszy 1977). 60 Age at Maturity Spangler et al. (1977), in a review paper on percids, concluded that percids in response to exploitation exhibited a change in the variability of recruitment, increased growth rate, and a decline in age at maturity. Less tractable changes included changes in the genetic stock and changes in interspecific relationships. Changes in genetic stocks due to overfishing were not documented, but must have occurred in the fishing-up process. These types of changes, especially early reproduction, can be deleterious, since as Roff (1984) pointed out, reproduction can cause decreased growth or an increase in mortality. Therefore populations that are impacted so that there is a decrease in the age at maturity may experience early mortality as an additional effect. Yellow perch in Lake Erie have also exhibited a decline in age at maturity as the time of first spawning went from age 3 in 1927-1937 (Jobes 1952) to age 2 during more intense exploitation in 1960-1966 (Spangler et al. 1977). In Lake Windermere, Craig et al. (1979) found that, after the die-off that killed 98% of the perch, age at maturity declined. Eurasian perch in 1978 matured at age 2; whereas, before the die-off they matured at age 3. This was due to increased food per fish which resulted in increased growth and reproduction. Sex Ratio Eshenroder (1977), studying the effects of several factors on the yellow perch of Saginaw Bay,' Lake Huron, found as fishing pressure increased in Saginaw Bay there was an increase in growth 'rate, a decline in the average age in the catch, females became scarce because they grew faster and were harvested earlier than males, and these changes resulted in low recruitment because of low brood stock. Temperature was important in determining year- class strength. In 1966, the size limit on the commercial fishery was removed, more fish were harvested, and immediately there was an increased growth rate in remaining perch, a shift toward younger fish, and a decline in the abundance of female spawners. Trawl samples revealed that only 11% of the catch was female, a shift in the sex ratio that had significant ramifications for recruitment. A decline in the walleye populations in the 1940s and an increase in the numbers of alewives and rainbow smelt affected the abundance of yellow perch the most. Walleye restrained yellow perch populations through predation, allowing good growth; when walleyes declined, yellow perch numbers increased. Males were more abundant (62%) in Saginaw Bay in 1955, at high population levels, than in 1929 (25%), at low abundance (El-Zarka 1959). In 1941, the year removals were initiated in Lake Windermere, the sex ratio was two males to one female and the male Eurasian perch biomass was 1.25 times the female biomass (LeCren et al. 1977). In 1944, after removals, the ratio was 0.7 males per female and the male biomass was also reduced to 0.17- 0.26 times the female biomass. 61 Predation Forney (1971) showed that depensatory mortality can occur in yellow perch populations in Oneida Lake. From trawl catches he found that there was a decline in catch from August to October which was more severe for weak than strong year classes. The decline was due to walleyes which ate a fixed number of prey items regardless of the size of the yellow perch population. Nepszy (1977) has concluded from studies of Lake Erie yellow perch that increased exploitation has caused irregular recruitment and no strong year classes were formed. Predation on yellow perch by rainbow smelt and walleye was cited as a possible contributing factor. Lawler (1965) studied Hemming Lake in Ontario and removed a large quantity of northern pike. The density of yellow perch immediately increased. In Oneida Lake, New York, YOY yellow perch are preyed on by walleyes which have a stable population and are usually the main determinants of year-class strength in yellow perch (Forney 1971). In Windermere, Eurasian perch are preyed on through age 4 by northern pike and the predator population is unstable. Cannibalism Eschmeyer (1936) recorded some of the characteristics of stunted yellow perch populations in northern Michigan inland lakes. He found that stunted fish preyed heavily on YOY and yearlings and that older fish died of starvation. Most fish were 2-year-old fish. Stunted fish grew slowly; 75% were age 2 and the sex ratio was 74 females to 100 males. Most females were older fish and the young fish were males; smaller fish ate insects, while older fish starved to death. Tarby (1974), in studies of Oneida Lake yellow perch, found them to be cannibalistic. However, during the period of study the growth rate of young perch appeared independent of density. This finding suggests that cannibalism was unrelated to perch abundance (density- independent) . In fact, Forney (1971) found that during their first year, young perch experienced depensatory mortality, which was attributed to walleye predation. In a related paper, Smyly (1952) found yellow perch fry were cannibalistic when the zooplankton crop was depleted in June-August; however the effect on the populations was undetermined. Ilina (1973, cited in Spangler et al. 1977), in experiments in Russia, showed how some density- dependent mechanisms worked for perch. The proportions of cannibals (YOY) varied according to the environment. There were usually three groups of YOY perch that formed, given a starting group of all the same size and age. This included a group that ate plankton, one that ate benthos, and one that was cannibalistic. When the temperatures under which the experiment was conducted were cool and within a narrow range, few cannibals resulted. However when the water was warm, fish grew faster, and more cannibals were produced. 62 A generalization about yellow perch seemed to emerge from the studies examined. When yellow perch were studied in ponds or lakes where they were the only species present, cannibalism acted as a strong density-dependent regulator of the populations (Aim 1946; Schneider 1973). In contrast, in more diverse systems with stable predator-prey systems, predation was the most important agent regulating perch populations, and cannibalism was density- independent (Forney 1971; Craig 1980). Interference Competition Ricker and Gottschalk (1940) studied the impact of the removal of rough fish from Bass Lake, Indiana. They removed 40 metric tons of common carp, 18 metric tons of carpsuckers {Carpi odes spp.), and 5 metric tons of buffalos (Ictfobus spp.). Their removal caused the abundance of game fish including yellow perch to increase. Borawa et al. (1978) studied the fish populations of a lake before and after an infestation of Eurasian milfoil in Currituck Sound, North Carolina. After the invasion, yellow perch and ictalurids increased the most. MacLean and Magnuson (1977) found that predation and competition were the most important driving forces within Penca and in their interactions with other species as well. Percids are temperate-zone mesotherms and are usually spatially segregated from other fish. Thus species interactions are minimized with regard to food and temperature because of their segregation from other species. Predation is intense on and by yellow perch. Interaction with other species is a major complicating feature in diverse ecosystems, and is no less so for yellow perch.' Yellow perch have been impacted by exotics and competitors and in turn have acted as strong competitors. There is ample literature discussing the impact of alewif e on yellow perch populations in the Great Lakes and in inland lakes (Kohler and Ney 1980). Wells (1977) discussed the impact of alewif e on Lake Michigan yellow perch, populations of which were severely depressed. Predation by alewif e on perch larvae was the probable cause. Eshenroder (1977) discussed a similar pattern for yellow perch in Saginaw Bay, Lake Huron, but overfishing and walleye population changes were also important agents of change. In fact he argued that populations of rainbow smelt and alewif e were high during periods of abundance of yellow perch. The major difference between the Lake Michigan and Saginaw Bay cases may be the abundant vegetation and large tributaries running into the bay which may have provided refugia for larval yellow perch until they grew large enough to escape alewif e predation and competition. Changes in dominance of the rudd (Scandinius enythnopthalms) and roach (Rut 7 Jus nutiJus) in an English lake (Burrough et al. 1979) provide another example of the depressing influence of a competing species. Competition for food at early life stages was the cause for low Eurasian perch survival in the presence of high populations of roach (Burrough et al. 1979). In Wilson Lake, Minnesota, white suckers were removed, which resulted in a concomitant increase in yellow perch growth (Johnson 1977). The reverse case occurred 63 when yellow perch were inadvertently introduced into oligotrophic lakes stocked with salmonids - salmonid growth declined and yellow perch numbers and biomass increased dramatically (Fraser 1978) . Acidification of lakes also provides another interesting sequence of events involving the very insensitive yellow perch, and provides more insight into the impact of competing species on yellow perch and compensatory responses to release of this pressure. Beamish (1976) and Ryan and Harvey (1980) presented a hypothesis backed with strong evidence and literature on how fish respond to increasing acidification of lake habitat. As a lake increases in acidity, a pattern of loss of competing species occurs, which favors yellow perch growth and survival. Then, under severe acid conditions, good growth of younger age-groups (1-3) is observed, but older fish (>3) experience physiological stress and starvation. Starvation Deelder (1951) studied stunted yellow perch in Holland. Yellow perch exhibited good growth to 14 cm, then there was no growth. Eschmeyer (1936) also examined stunting in Michigan lakes and also found little growth by larger individuals. Both investigators attributed this to starvation of these larger fish which had inadequate prey. Aim (1946) also found similar patterns in the stunted fish he studied in Sweden. Some populations grew very little but lived for unusually long periods of time, up to 18 years. Stress Clady (1976) studied the importance of catastrophic events on yellow perch eggs and larvae in Oneida Lake. He found that survival from the egg to the larval stage was 1.6 to 18.4% and that survival was positively correlated with temperature and negatively with wind. Year-class size was determined by a complex of minor factors rather than by catastrophic events. Low temperature adversely affected prolarval yellow perch. Eshenroder (1977) found that the decline of yellow perch in Saginaw Bay was related to eutrophication. Changes which occurred in response to increased eutrophication included: (1) switch to alternate prey, (2) change in distribution when the environment was degraded, and (3) reproductive behavior adaptations. Eutrophication affected the selection and succession of species. Disease Craig et al. (1979) reported on a disease that impacted the Eurasian perch of Lake Windermere. Disease killed 98% of the yellow perch in 1976. Fungus and epidermal lesions were noted on fish in January 1976; in September 1976 all attempts to catch adult yellow perch failed. This resulted in increased fecundity (almost fourfold) as remaining fish compensated for the 64 loss of most of the population. For example^ from 1944 to 1960, fecundity stayed, about the same, 13,000 eggs/fish, but in 1977 after the die-off, fecundity increased to 47,000 eggs/fish. Craig et al. (1979) determined that numbers of perch in the lake prior to wartime harvesting (1941) were considerably higher than populations in 1967 to 1976. However, biomass values were very similar between the two time periods indicating the population was at a high level and was trying to achieve stability of biomass. The precipitous decline in 1976 due to disease may thus be density related. Disease has caused catastrophic declines in Lake Mendota yellow perch (Bardach 1951) and Lake Windermere Eurasian perch (Craig et al. 1979). Compensatory increases in growth occurred because of reduced intraspecific competition and cannibalism. Parasites Leach et al. (1977) examined the response of yellow perch to eutrophication. Yellow perch have a curve of response to increased eutrophication where production first increases, then declines. The sequence is predictable and can be exacerbated by overfishing and habitat changes. As eutrophication increases, yield of yellow perch may not increase since nutrients sometimes are channeled through unharvested food chains. Yellow perch biomass and growth increase to threshold levels then decline. Parasites increase with increased nutrient enrichment. In Lake Windermere, Eurasian perch showed increased growth due to eutrophication. Parasitism also increased with eutrophication there since the incidence of Di pJostomum sp., a parasite which causes hematoma of internal organs, increased dramatically and caused an early summer mortality of adults (Craig et al. 1979). Parasites increased because of more vegetation which provided more habitat for snails, intermediate hosts for the parasites. Summary A strong characteristic of yellow perch populations is self regulation, which is clearly and abundantly documented for well balanced, and even highly exploited populations. In many cases, cannibalism acts as a coarse control of populations, allowing increased growth of remaining fish and ensuring a greater chance of survival. Predation acts as a fine control on remaining fish, which at times can cause depensatory mortality in YOY (Forney 1971). In Oneida Lake, cannibalism was determined to be density-independent (Tarby 1974) while Aim (1952) found that it was the principal mechanism of establishing dominance in dystrophic lakes, where perch was the dominant species in a low diversity system. The most catastrophic decline in perch populations occurred in Lake Windermere, England, where over 98% of the Eurasian perch were lost to disease in 1976. This decline prompted compensatory increases in growth, fecundity (fourfold), and a decrease in the age at maturity of these fish. A myxosporidian disease, reduction in a competing species, and increased fishing 65 pressure on yellow perch in Lake Mendota, Wisconsin caused a reduction in cannibalism on YOY and a compensatory increase in growth in remaining yellow perch (Bardach 1951) • Removal of large numbers of Eurasian perch from Lake Windermere in 1941-1947, to provide food for English people during the war, depressed perch population numbers for many years. Northern pike predation also contributed to depressed perch recruitment. Predation by walleyes (Forney 1971) and muskellunge (Schmitz and Hetfeld 1965) caused depensatory mortality in yellow perch resulting in increased growth among remaining fish. In another case (Lake Windermere - Craig 1980), removal of Eurasian perch via trap nets did not elicit a compensatory growth increase because perch continued to be cropped off and competing species occupied the "niche" of the yellow perch. Predation also entered into Lake Michigan yellow perch-alewife interactions. Wells (1977, 1985) and Jude and Tesar (1985) postulated that alewife eat the larvae of yellow perch, thus reducing their population levels. An alewife die-off in 1967 and a collapse in the alewife population in 1982 have resulted in an immediate resurgence of the yellow perch population, first by increasing abundance. The most recent collapse of alewife allowed yellow perch to reach a peak in 1983, and reduced growth resulted as density-dependent processes apparently took effect. Competition, though difficult to document, was suspected in the case of the rudd, Eurasian perch, and roach in England. Competition for food at early life stages caused low survival of perch in the presence of high populations of the roach (Burrough et al. 1979). In a comparable case in a Minnesota lake, a competing species, white sucker, was reduced substantially and a concomitant growth increase in yellow perch followed (Johnson 1977). Overfishing has reduced populations of yellow perch in several lakes (Eshenroder 1977; Spangler et al. 1977). In most cases responses were generalized: changes in variability of recruitment, compensatory increases in growth rate, and a reduction in the age at first spawning. Some major changes in the environment have provided evidence of compensation in yellow perch populations. Acidification of lakes has generally favored yellow perch survival and growth, since perch are one of the most insensitive species (Beamish 1976; Ryan and Harvey 1980). At a pH of 4.5, good growth of younger age-groups (1-3) occurs, but fish >3 years old experience physiological stress and starvation. Eutrophication generally favors yellow perch (Leach et al. 1977). In the case of Lake Memphremagog, Quebec, there were two different basins, one enriched and the other less so. Perch did not grow faster in the enriched basin, but instead produced higher numbers of fish than in the less productive basin (Nakashima and Leggett 1975). Lastly, Clady (1976) found the pattern of early survival of larval yellow perch was due to a number of minor mortality factors, rather than one catastrophic event. Nielsen (1980) concluded that there were complex patterns and causes of mortality for yellow perch in Oneida Lake, which were also observed in the Lake Windermere case (LeCren et al. 1977). These interactions indicate that a 66 simple mcxiel cannot explain or predict the response of percids to an altered environment or altered fish community^ or forecast the structure of percid communities across broad habitat or geographical ranges. WALLEYE Introduction The distribution of walleye (Stf zostedion vitneum vitreutn) is limited to the fresh waters of North America (Scott and Grossman 1973) • This species has been extensively studied and is of great interest to man both as a commercial and sport species. Factors which have been documented in the literature as having adverse effects on walleye populations may be categorized broadly as: (a) over exploitation, (b) pollution, (c) loss of spawning habitat (destruction of habitat or limitation of access to sites), and (d) climate. Overexploitation contributed to depletion of stocks of various populations (Car lander 1945; Anthony and Jorgensen 1977), which caused truncation or alteration of age distribution (Busch et al. 1975; Anthony and Jorgensen 1977; Shuter and Koonce 1977), increased growth (Anthony and Jorgensen 1977; Gaboury and Patalas 1984), and decreased CPUE with changes in recruitment to fisheries (Anthony and Jorgensen 1977). The blue pike (Stizostedfon vitneum glaucum) of western Lake Erie failed to compensate and became extinct. The primary drain on the population was overfishing, and desegregation through introgression with the walleye (Regier et al. 1969) may have been the ultimate cause of extinction. Pollution affects walleye populations directly by limiting reproductive success. Construction activity, agricultural runoff, and other factors contributed to long-term ecological changes which affected walleye spawning grounds in Toronto Bay, Lake Ontario (Whillans 1979). Pollution and sedimentation also affected spawning areas in Lake Erie (Busch et al. 1975). In examining effects on fish of acid rain. Beamish (1976) found walleye to be among the more sensitive species; walleye stopped reproducing when pH dropped to around 5.5-6.0. From the 1950s to the mid-1970s, walleye in western Lake Erie decreased severely in abundance in large part because of loss of suitable spawning habitat (Busch et al. 1975). Walleye numbers were so reduced in the late 1950s that density-dependent processes were not influential and abiotic density-independent factors influenced walleye production. Busch et al. (1975) observed that an uninterrupted series of years with adverse environmental conditions could ultimately destroy the population. 67 Climate is important mostly as it affects spawning success of walleye. Thermal and wind conditions, water currents, and turbulence are among important climate-related factors noted by Busch et al. (1975). In Toronto Bay, Lake Ontario, a catastrophic event led to the decline and virtual extinction of a walleye population (Whillans 1979). A protective peninsula was breached by a storm producing unfavorable habitat for walleye and other species. Construction activities and agricultural runoff also contributed to ruination of local walleye spawning grounds at a time when exploitation of walleye was at a peak. Along with long-term ecological changes, these events brought about a decline in the walleye population from which they did not recover. Introgression and appearance of a sterile hybrid were also reported (Whillans 1979). For this report, information on walleye was gleaned from the literature and organized according to how it related to compensatory mechanisms. Though not always generated from or pertaining directly to catastrophic events, such information as a whole helps define the compensatory potential of walleye. Growth Many authors (Wolfert 1969; Johnson 1977; Shuter and Koonce 1977; Colby et al. 1979; Shuter et al. 1979) have recorded an inverse relationship between growth rate and population density. Shuter and Koonce (1977) observed that this relationship for walleye in Lake Erie was somewhat offset by a proportional relationship between growth rate and forage density. Growth rate of Lake Erie walleye increased markedly during the 1950s when the fishery intensified greatly; the average weight of age-3 fish increased from 0.4 kg in 1952 to 1.3 kg in 1960 (Regier et al. 1969). The Lake Erie walleye fishery virtually collapsed in the late 1950s. Comparing 1943-1954 with 1955-1962, average length of walleye in commercial catches increased from 366 to 406 mm, weight from 440 to 731 g, and condition factor from 0.90 to 1.09; simultaneously, average age decreased from 2.7 to 2.0 years. Average length of age-1 walleye in 1958 was 366 mm, whereas the previous high was 348 mm, in 1949. Also, average weights of age-2 walleye, after fluctuating between 294 and 437 g in 1943-1954, increased from 518 g in 1955-1956 to 986 g in 1960 (Parsons 1970). These data demonstrate growth compensation after a catastrophic reduction in density. Forney (1984) however, found that catastrophic mortality of walleye in Chautauqua Lake, New York, did not elicit a detectable increase in growth. In Oneida Lake, New York, growth of certain walleye year classes appeared to have been independent of numerical abundance of an individual year class because competition within a year class was not markedly greater than competition among year classes (Forney 1965). A study undertaken in Cross Lake, Manitoba revealed that of two discrete walleye populations, the one which was subjected to a higher level of exploitation exhibited a significantly (P < 0.05) faster growth rate (Gaboury and Patalas 1984). Comparing data from Lake Nipissing, Ontario during 1967- 1968 and 1972-1975, Anthony and Jorgensen (1977) interpreted greater mean lengths of walleye specimens in every age-group recorded during the latter 68 years as an indication that overfishing had occurred. In Rainy Lake, Ontario, mean total length of age-4 walleye increased significantly from 37.3 cm in 1959 to 41.6 cm in 1965, at the same time as CPUE (expressed in kg of walleye per km of gill net per day) decreased from 32 to 20 (Chevalier 1977). These data show density-dependent growth. Differences in growth rate according to sex and age of walleye have been documented. Car lander (1945) found that females grew slightly more rapidly than males after the third or fourth year in Lake of the Woods, Minnesota. Females also outgrew males in Oneida Lake, New York according to Forney (1965) who further speculated that growth of YOY and older walleye was probably determined by different factors. Swenson and Smith (1976) indicated that growth compensation by YOY walleye occurs when food availability changes from limited to unrestricted within a growing season. Implications of different growth rates include earlier maturity of rapidly growing fish, and a relatively high percentage of non-ripening females during years of poor growth (Forney 1965; Colby et al. 1979). Growth was also an important factor in terms of size at maturity (Wolfert 1969), duration of predation and cannibalism on YOY (Forney 1976; Colby et al. 1979), recruitment (Forney 1977), and intraspecific competition (Forney 1974, 1977; Shuter and Koonce 1977). Walleye were more susceptible to angling during years of poor growth when food availability was low (Forney 1965). Survival In a study by Forney (1976), survival of walleye from the egg to prolarval stage was very low but consistent; differences in size of year- classes were not attributable to variability in early mortality. Little evidence was found to indicate compensatory mortality during the first year though the possibility existed that it was masked by density-independent mortality. Forney (1984) did not find an immediate improvement in early survival after the catastrophic walleye mortality in Chautauqua Lake. However, reproductive success, probably through improved early survival, increased during the 1940s in Lake Erie while commercial walleye catch was increasing (Regier et al. 1969). According to Busch et al. (1975), reproductive success of walleye in Lake Erie was most clearly related to rates of water warming during spawning and incubation periods. AlDiotic factors which have negative effects on early survival included strong winds, temperature reversals, turbulence, and physical displacement of eggs from reefs to unsuitable mud bottoms (Busch et al. 1975). As for older fish, Colby et al. (1979) noted a tendency for late-maturing walleye to have a longer life span than early maturers. Forney (1977) related fluctuations in walleye survival with changes in prey abundance. Johnson (1977) proposed that there may have been an initial increase in survival of young walleye immediately following an elimination of white suckers from a lake in Minnesota. Following two extremely weak year 69 classes, increased survival of walleye (as evidenced by a strong year class) came about in Lake Erie at a time when overall abundance of walleye and other predators was low while forage species abundance was up (Parsons 1971). Evidence for a relationship between recruitment and relative size of brood stock was conflicting. Shuter and Koonce (1977) found that recruitment increased with increasing numbers of reproductive stock, but Busch et al. (1975) indicated no correlation between the two factors. In Cross Lake, Manitoba, Gaboury and Patalas (1984) studied two distinct walleye populations from two basins. While walleye abundance and catch per unit effort increased in one basin and decreased in the other, survival remained unchanged. Fecundity Colby et al. (1979) found that walleye fecundity increased with age, and Wolfert (1969) established that there was a linear relationship between fecundity and weight and between log fecundity and log length, but other factors merit consideration. If low temperatures persist during spawning season and spawning is delayed beyond a certain point, female walleye may resorb their eggs (Busch et al. 1975). In addition, years of relatively low food availability and poor growth translated into higher than normal proportions of non-ripening females (Colby et al. 1979). Forney (1977) inferred that when interspecific competition was great, female walleye matured at small sizes and had small egg complements which, if age at maturity remained the same, would presumably result in a decrease in "r", their intrinsic rate of increase. Sex Ratio Changes Based on data for only 1 month of sampling, walleye in Lake of the Woods, Minnesota, exhibited sex ratio changes with age, as males were apparently shorter lived than females (Carlander 1945). The sex ratios (male : female) for various ages were: age 4 = 1.32 : 1, age 5 = 1.11 : 1, and age 6 = 0.64 : 1. Sex ratio was not shown to be density-dependent. Age and Size at Maturity Throughout the literature male walleyes were documented as having matured at younger ages and smaller sizes than females (Carlander 1945; Forney 1965; Wolfert 1969; Forney 1976; Anthony and Jorgensen 1977). For exploited populations this may cause problems as females may grow to be recruited into a fishery earlier than males even though females mature at later ages (Anthony and Jorgensen 1977). Mature fish were consistently larger than immatures of the same year class (Forney 1965; Wolfert 1969). When population size decreased, average age at maturity decreased (Shuter and Koonce 1977), and for walleye from Oneida Lake and Lake Erie respectively, Forney (1965) and Wolfert (1969) noted a trend toward earlier maturation among more rapidly growing fish. However, two walleye populations in Cross Lake, Manitoba had different 70 growth rates but a similar mean age at onset of maturity for both sexes combined (Gaboury and Patalas 1984). Comparing populations from two basins in Lake Erie, Wolfert (1969) found that females from the western basin matured at shorter lengths and younger ages than those from the eastern basin, but males from the western basin matured at longer lengths and greater ages than those from the eastern basin. Cannitelism Data from two studies indicated that walleye were not cannibalistic (Swenson and Smith 1976; Johnson 1977) though the former authors recognized that this interpretation may have been invalid due to inadequate sampling. In other studies cannibalism was well documented. Parsons (1971) found that walleye accounted for 2% of the diet of walleye in Lake Erie, Colby et al. (1979) documented several instances of cannibalism, and the phenomenon in Oneida Lake was studied in depth by Forney (1974, 1976, 1977). According to Forney, intensity and duration of cannibalism acted as a density-independent factor contributing to variation in year class size and as a compensatory factor limiting population size. Because young yellow perch make up the main forage base in this system, they acted as a buffer controlling walleye cannibalism and indirectly regulating walleye population size. Chevalier (1973) inferred from the relative timing of mortality and incidence of cannibalism that cannibalism was the principal source of walleye mortality in Oneida Lake. He speculated that because cannibalism can occur late in the first year of life when most of the mortality has already occurred, cannibalism may be significant in regulating year-class size. Young walleyes in stomachs were significantly smaller than those in trawl catches, suggesting that adult walleyes were selecting the smaller individuals (Chevalier 1973). Cannibalism, like predation, thus may cause growth rates of young fish to appear to increase by removing slower-growing fish. Disease and Parasitism Parasites and diseases affecting walleye were tabulated and described by Colby et al. (1979). It was noted that incidence and spread of disease was greater during spawning season when fish were congregated at greater than normal densities. Predation By preying on eggs, white suckers were noted as having a potentially large impact on walleye in Lake Nipissing, Ontario; predators on young walleye (northern pike and yellow perch) were thought to have aggravated effects on an already stressed walleye population (Anthony and Jorgensen 1977). Density- independent mortality of walleye during early stages of development has been attributed to factors including predation, inadequate food, or interactions of these variables with climatic factors (Forney 1976). 71 Predation on walleye may increase during years when poor growth is exhibited by YOY walleye (Colby et al. 1979) or when alternate prey are less available. After the removal of suckers from a lake in Minnesota, the percentage of walleye in the diet of northern pike increased (Johnson 1977) • Some fish have a variety of interactions with walleye on different levels. In the Great Lakes, rainbow smelt preyed on pelagic larval walleye, competed with these larvae for food, but also provided forage for adults (Schneider and Leach 1977). Regier et al. (1969) suggested that abundant rainbow smelt in Lake Erie during the 1950s suppressed or prevented reproduction of walleye and blue pike. Inter- and Intraspecific Competition Interspecific competition with walleye was described by several authors. In Lake Winnebago, walleye fry competed for food with freshwater drum (ApJodinotus gnunniens) fry, and various authors implicated alewife as being serious competitors with walleye fry for food items, or predators on walleye larvae, in northern Green Bay, Lake Michigan, and in Lake Erie (Schneider and Leach 1977; Colby et al. 1979). Removal of suckers from a Minnesota lake allowed walleye YOY to make greater use of small organisms (e.g., crustaceans and small insects) which had been exploited mainly by adult yellow perch but which became more available when these perch switched to more varied diets available after sucker removal (Johnson 1977). Large numbers of common carp, quillback (Carpi odes cypnfnus), and buffalo were removed from Bass Lake, Indiana, over 2 years. Average catch of walleye increased from 2.3 to 5.8 per seine haul over those 2 years (Ricker and Gottschalk 1940), perhaps due to a combination of decreased interspecific competition and improved water quality. In Lake of the Woods, Minnesota, yellow perch and saugers were identified as competitors with walleye (Carlander 1945), and in Lake Nipissing, Ontario, white suckers, northern pike, and yellow perch were competitors with and predators on walleye (Anthony and Jorgensen 1977). Inter-specific competition may affect growth and survival of walleye. Intraspecific competition can be significant in limiting food availability, food consumption, and growth of walleye (Swenson and Smith 1976), but in most systems both intra- and interspecific competition must be considered. An excellent illustration of this point is provided by Forney (1977) describing Oneida Lake, New York: between 1900 and 1974, a sequence of events led to a decline of American eel and esocids (competitive predators with walleye), and an increase in the walleye population. The mean size of walleye in the lake stayed approximately the same, as the decrease in interspecific competition was apparently balanced by the increase in intraspecific competition. Variability in prey abundance was an important factor in walleye growth and recruitment. Years of high prey abundance enhanced growth of young walleye and reduced cannibalism, favoring development of strong year classes. This led to increased intraspecific competition for food which caused prey abundance and growth rates to decrease, cannibalism to increase, and survival of succeeding year classes to be suppressed. This 72 scenario repeated over the years adjusted walleye growth to prey production, although time lags and environmental perturbations induced instability in annual growth and recruitment. Removal of buffalo, common carp, and freshwater drum from East Okoboji Lake, Iowa, did not immediately result in an increase in walleye abundance. Howev€ir, after a winterkill in the eleventh year of rough fish removal, abundance of YOY walleye suddenly increased from a 10-year average of 10 YOY walleye per seine haul to 196 per haul (Rose and Moen 1952). This increase in recruitment probably reflects improved early survival due to decreases in both inter- and intraspecific competition. Summary The walleye does not fluctuate widely in abundance as do some species, e.g., clupeids. However, walleye populations in some areas have suffered catastrophic mortalities from overfishing, pollution, and weather-related causes. Many walleye populations have demonstrated compensation through growth, and both inter- and intraspecific competition operate to control walleye abundance. Survival, age at maturity, cannibalism, and predation are compensatory mechanisms which appear to operate in some populations of walleye and not in other populations. Both predators and cannibals of young walleye have been shown to select the smaller individuals, causing mean growth rate of juveniles to appear faster. FRESHWATER DRUM Introduction The freshwater drum or sheepshead (ApJodinotus grunniens) has a large latitudinal range from Hudson Bay south to Guatemala. In the United States it is found from Pennsylvania west through the Great Lakes and Mississippi drainages to Montana and Nebraska (Scott and Grossman 1973). Freshwater drum normally may live up to 13 years with a weight up to 8.2 kg and a length of 750 mm (Wrenn 1968). It is not a favored game or commercial species due to its tough and coarse flesh, especially in larger individuals, although it has been harvested commercially in Lake Erie, Lake Winnebago, and the Mississippi River (Wirth 1958; Butler 1965; Leach and Nepszy 1976). Catastrophic Mortalities and Compensation Through Recruitment Only one study reported significant natural mortality to a population of freshwater drum. Dechtiar (1972) noted "considerable" mortality of drum had occurred along with yellow perch in Lake Erie in 1963. This observation was based upon the high numbers of dead fish which appeared along the Ohio and Ontario shorelines during June and July. Some evidence, though not conclusive, for the cause of the kill was a heavy infestation of the protozoan 73 parasite, Ichthophthypius muJtifiJfis. Although parasitism, which is common for drum (Dechtiar 1972), could be density-dependent, no evidence suggested that this mortality of drum was related to drum density. A rather large yellow perch year class from 1962 may have triggered the parasitic infestation which could have subsequently increased parasitism upon the drum population. Fewer than 5% of the dead fish observed were freshwater drum and no assessment of overall impact on the drum population in Lake Erie in subsequent years was made (Kenneth Muth, U.S. Fish and Wildlife Service, pers. comm.). However, commercial fishing data for Lake Erie (Baldwin et al. 1979) showed no significant change in the total catch of drum caught in 1963 compared to that of previous years. Thus, assuming constant fishing effort in the early 1960s, the mortality of drum in 1963 probably represented a small proportion of the population, at least among larger size fish most susceptible to commercial fishing. A case of unusually high mortality for drum where some compensatory responses may have occurred was a commercial harvest of drum at Lake Winnebago, Wisconsin between 1955 to 1966 (Priegel 1971). The purpose of this harvest was to control the drum population because it was believed to compete with more desirable game species in the lake. During this operation at least 680,400 kg of drum were removed annually from Lake Winnebago (a 55,772 hectare lake with an average depth of 4.7 meters). The fishing effort was monitored during the study and fish were collected to assess changes in the condition, age, and size composition of the drum population during the harvest period. Results were that the total catch of drum ranged from 1,540,000 to 2,000,000 kg annually from 1955 to 1959. However, the catch varied from 680,000 to 1,360,000 kg (1,000,000 kg average) from 1960 to to 1966. This was said by Priegel (1971) to have indicated a substantial decline in the drum population after 1960. However, his interpretation should be questioned since based on the catch per unit effort data presented for various fishing gear used, no significant change in either the number or biomass of drum taken was evident. The use of catch per unit effort data to interpret a change in a fish population would eliminate bias of the total catch results caused by variable fishing intensity. The age composition of the Lake Winnebago drum population did change during the commercial harvest period. For example, samples of drum in 1955 and 1960 showed the majority of fish were of age-group 5 or older. By 1963, however, 55% of the drum were in age-group 4 and by 1965 most of the fish were age 1 or 2. Approximately 60% of the fish taken after 1962 were less than 356 mm long whereas in prior years at least 50% of the fish caught were over 356 mm. These results indicated a gradual reduction in the age and size composition of the drum population. If the overall population numbers and biomass did not change in Lake Winnebago while the percent frequency of smaller and younger drum increased, then the absolute number of younger drum increased. This would indicate an increase in recruitment had taken place. The extent to which numbers of young drum increased in the lake was not clear from this study since the percent frequencies for age- and length-groups given above were not from fish samples 74 expressed as catch per unit effort • Recruitment apparently increased, although it was not well quantified, and was a density-dependent response to a substantial mortality of the larger freshwater drum, indicating compensation had occurred perhaps as a decrease in mortality of drum larvae. However, the exact compensatory mechanism is not known since no data were obtained on such parameters as larval fish mortality, age or size at maturity, fecundity, or causes of mortality of freshwater drum* We speculate that decreased intraspecific competition allowed improved survival of young drum. An increase in the condition factor of the drum population may have also occurred. However, the condition factor fluctuated over the study period but did not show any consistent increase for any size groups of freshwater drum. Growth rate changes for drum were not analyzed by Priegel (1971). Large numbers of rough fish, including freshwater drum, were removed by seine and trap from East Okoboji Lake, Iowa. Abundance of drum, both adult and YOY, declined over the 11 years of fish removal, until a winterkill in the last year of the study. Drum YOY had averaged 387 per seine haul in the first 3 years of the removal, then dropped in abundance to 1-3 per haul in years 8 to 10. In year 11, after the winterkill, 217 YOY drum per haul were collected, and YOY of walleye and yellow perch also increased (Rose and Moen 1952). For drum, the increase in recruitment was probably due to decreased interspecific competition. Intraspecific competition was probably not a factor in recruitment success of drum in this case, since numbers of YOY drum had continually declined as drum were removed. Only when other species (e.g., walleye and yellow perch) experienced winterkill were drum able to compensate by increased recruitment, probably through improved early survival. Potential Compensatory Mechanisms Various studies have been done with freshwater drum on age frequencies and growth rates in lakes, rivers, and reservoirs (Van Oosten 1937a; Schoffman 1941; Butler and Smith 1949; Daiber 1953; Houser 1960; Swedburg 1965; Wrenn 1968; Priegel 1969). Decreasing latitude and thus generally warmer temperatures apparently accounted for some increase in growth rate of drum (Schoffman 1941; Daiber 1953). A growth response to temperature would be a density-independent phenomenon if temperature alone was a limiting factor. However, as seen in Table 11 where growth rate is compared for various sites arranged from north to south, latitude alone would not account for differences in growth rate. Although food data were lacking for the various studies of freshwater drum, most likely food availability (a potentially density- dependent factor) also would be involved in determining growth rate (Swedburg 1965). The growth of drum also varied with age of the individual, with some populations having the most rapid growth early (Schoffman 1941; Butler and Smith 1949; Nelson 1974) and some late in life (Van Oosten 1937a). The rate of growth appeared to differ only slightly between the sexes (Van Oosten 1937a; Butler and Smith 1949; Nelson 1974). Growth rate variation also had little or no relationship to water level fluctuation in the Mississippi (Butler and Smith 1949) or the Lake Oahe Reservoir on the Missouri River (Nelson 1974). 75 m C O •H -P o o P o • •H 03 }-^ 4J 03 fd > T3 4J O T3 03 O ^ -H 1 ^ u 03 as 3 O ^ V4 CO O 0) 0) rH < 03 s E 'O .^ ro cr o ex 0) o o o c 0) 03 o o: en en en en en -^ 0) ^ 0) -^ 0) -. o --. aoQ i 03 03 >i m VO r*- 00 O -- 03 vfi GPN 03 o 03 c o en H 0) z jQ 0) c c 0) 03 ON O) CD 04 ^J w ►J w >J >-. 03 V4 03 T? C 03 en m3 u D jQ rH 03 c 03 cn u 3 jQ o o 03 <3\ VC OC — H ,^-^ «-*» cn cn en en cn ^ i 0) >i VO 03 m c 0) 0) 4J •rH cn u o Cxa o (D c -itf 03 03 > en cn u u 03 cn (^ (D u 0) >i 03 >i 'sr > VD WD CM LO en ^ o 03 ^ 03 E O x: 03 }H u o ^ en 00 oi: ON V4 FIG. 2. Resilience versus stability, as defined by Holling (1973), of the species investigated for this report. Species on branches close to one another are relatively similar in their resilience or stability. There are several examples of community shifts between populations of yellow perch and other 'species. Eshenroder (1977) traced part of the yellow perch decline and subsequent dominance changes in Saginaw Bay to. the invasion of the exotic alewif e, a phenomenon also noted by Wells (1977). The mechanism for this decline in Lake Michigan was hypothesized to be predation on larval yellow perch by alewif e. Another example is given by Burrough et al. (1979) for the Eurasian perch in England. Here interactions between rudd, a cyprinid, and the roach, caused perch populations to decline when roach populations assumed dominance. They suggested competition for food at younger stages as the mechanism. Svardson (1976) studied a lake with roach and yellow perch present. With increased acidification, roach populations declined and yellow perch, which are less sensitive to acid stress, immediately had increased growth, again suggesting competition for food as the depressing mechanism. Bardach (1951), in studies of Lake Mendota, saw increased growth of yellow perch concomitant with a decline in lake herring populations. Yellow perch have also disrupted communities. Fraser (1978) cites the case of salmonids being stocked in a Ontario oligotrophic lake and subsequent entry of yellow perch. He found that growth of the salmonids declined while perch growth improved. Lastly, there is a case where habitat alteration (infestation of Eurasian milfoil) caused a dramatic increase of yellow perch in a North Carolina lake at the expense of several other species (Borawa et al. 1978). 84 Selective removal of certain species may change community composition, but not necessarily in the way desired. Examples of this are the increase in gizzard shad when gizzard and threadfin shad were removed (Phillippy 1964), the increase in common carp when gizzard shad were removed (Rose 1957), and increased growth of thinned species (rough fish and panfish) rather than sport fish (Grice 1957). ADVERSE EFFECTS OF COMPENSATION While traditional fishery management theory holds that thinning a population to increase size of fish is desirable, there are disadvantages pertaining to a population which is compensating. An overview of many species shows a positive relationship between growth rate and mortality, and a negative relationship between those parameters and age at maturity. The energy costs of reproduction decrease energy available for growth and probably weaken the fish, increasing mortality (Roff 1984). Thus a population which is compensating for a catastrophic mortality by increasing spawning effort (increasing fecundity or decreasing age at maturity) may be increasingly vulneratDle to other mortality factors such as disease or predation. An example of increased growth rates combined with early death is the gizzard shad in Lake Erie and Florida (Table 9) . Since age at maturity has not been demonstrated to be size-specific for gizzard shad, we do not know yet whether growth cind mortality are directly related, or indirectly through reproduction. Evidence is also presented by Aim (1952) who found that strong year classes of Eurasian perch appear rarely, but they dominate for 10-15 years for both sexes in dystrophic lakes in Sweden. Growth is poor and individuals small, but longevity was extended. These are uncomplicated ecosystems, with only a few northern pike present. In eutrophic lakes, strong year classes occur more often and they do not dominate for a long period. Growth is good and perch attain large sizes. Schneberger (1935) and Harkness (1922) (cited in Aim 1952) pointed out that yellow perch which grew slowly attained a greater age. Aim (1952) concluded that natural mortality was the same in stunted and well balanced perch populations; however, in the latter lakes, intensive fishing, predation and intra- and interspecific competition increased mortality in the perch population. Age of a spawning fish affects the quality of its eggs. Eggs from the youngest and oldest fish tend to have less yolk (Ricker 1977). There was a direct relationship between parental age of Eurasian smelts and whitefishes (female and male) and survival of larvae to the tail-bud stage, with some decreased egg survival from the very oldest parents (Korovina 1961 cited in Ricker 1977). Early maturing freshwater drum females had smaller eggs (Daiber 1953). Atlantic herring showed some tendency for egg weight to increase with age of the female. Some stocks showed a 2-10% increase in mean egg weight from one age-group to the next (Hempel and Blaxter 1967), and some mean egg weights increased 20-30% between recruits and older fish (Blaxter and Hempel 1963). The stocks which showed a statistically significant relationship between egg weight and length of female were the faster-growing stocks of smaller fish, while egg size and size of female were not correlated for stocks 85 of older, slower -growing fish (Hempel and Blaxter 1967). Bigger eggs produced longer larvae with heavier yolks and better survival when starved (Blaxter and Hempel 1963). Thus a fish population which is compensating by lowering age at maturity may be sacrificing egg quality and survival of larval fish, even though total number of eggs produced in a lifetime increases. Not only may lower age at maturity sacrifice egg quality, but also young fish may not exhibit appropriate spawning behavior or optimal seasonal timing (due to physiological constraints or behavior), resulting in failure of fertilization or poor survival. Thus increased spawning effort early in life may be wasted. Classical ecological theory indicates that the replacement of stocks with low reproductive capacity by stocks of high reproductive capacity may indicate environmental disturbance (Smith 1978), thus an increase in spawning effort by a population may be a result, and a sign, of increased mortality. Fish populations may be capable of rebounding after a mass mortality, but mortality which removes part of the population may also remove a portion of its genetic variability. One of the least tractable and potentially most damaging effects of overfishing is changes in genetic stocks (Spangler et al. 1977). Thus the health and flexibility of the population and its ability to adapt to new stresses may be impaired. This is a hidden cost to compensation which cannot easily be measured. Loss of adaptability may mean that a population which recovers from a mass mortality one time cannot recover from another of a similar magnitude, particularly if the cause is different. Survivors of a cold kill may not be the same fish best adapted to survive a toxic spill. A population which is compensating is probably not operating at optimal levels of growth and reproduction. Such a population is less resilient because its compensatory reserve has been reduced. It may be more vulnerable to environmental change or random factors and so be more likely to collapse entirely (Ricker 1954, 1977; Goodyear 1980). Extreme values of parameters related to compensatory mechanisms should be perceived as danger signals for the population's continued existence. EFFECT OF CATASTROPHE TYPE The type of catastrophic event may affect subsequent compensation of fish. Catastrophes affecting habitat may prevent compensation altogether. Examples of this include toxic spills and acid precipitation, where the substance that kills fish lingers in the system. If early survival is usually density-dependent, and a density-independent effect of a toxic substance is to decrease early survival, the population would be unable to compensate by that mechanism. We did not find good data for this specific a premise. There were, however, cases such as lake acidification where all fish were killed, and continuing acidification would prevent any recolonization and attempts to study compensation. There have been other ramifications of habitat alteration such as an infestation of Eurasian milfoil, which afterwards favored yellow perch survival and populations increased (Borawa et al. 1978). Increased 86 acidification of lakes has led to a series of events for yellow perch. First competing species are eliminated, releasing competitive pressure and allowing perch population increases (Svardson 1976 cited in Burrough et al. 1979). As acidification increases further, adult reproductive failure occurs as adults apparently have more difficulty with calcium metabolism (Ryan and Harvey 1980; Beamish 1976). Thus, populations of younger fish (age 1-3) had reduced densities, less cannibalism from stressed and dying age-4-9 fish, and thercjfore grew better. Certain events may affect some age-groups of fish more than others as was observed for yellow perch subjected to increased acidification. Stock collapse due to overfishing generally removes older fish first, and recruitment often then responds positively. Drawdown may directly kill young fish in the littoral zone, or may concentrate fish so that forage fish are subject to increased predation. The fishes we studiecl each had more than one compensatory mechanism available, thus were able to respond even if an option was removed. For example, forage fishes whose populations were reduced to a few adults produced large year classes, presumably through increased growth and survival of young, when age at maturity or fecundity compensation did not have time to occur. Elimination of one year class during its first year usually first served to increase growth of age-1 fish, then caused high recruitment the next year. EFFECT OF GEOGRAPHIC AREA Fishes from the fringes of their range could compensate in different ways than fishes in their normal range. Life history parameters may vary with latitude; for example, growth and reproductive effort of the lake herring decrease with more northerly latitude, probably due to the shorter growing season (Morin et al. 1982). Kelso and Bagenal (1977) examined yellow perch in unperturbed systems, which were rarely found and studied. They concluded perch inhabited harsh environments in these systems and that growth and recruitment were low and variable. Unperturbed systems usually had a wide range of age-groups present and a dominant year class was expressed for a longer period of time than in more perturbed systems. This natural variation in life history parameters implies that some compensatory mechanisms would operate more readily in certain parts of a fish's range. However, data gaps present a difficulty in answering this question. For no species is the documentation of compensatory response complete. If species X increases its fecundity in the south and increases its survival in the north, we usually do not know if it also increases fecundity in the north and survival in the south. Differences in a life history parameter between the fringes and center of a range may be attributable to the rest of the community, genetic stocks, or random factors. One frequently seen effect of geographic area is the exacerbation of catastrophes. Populations in a harsh environment, such as gizzard shad in Lake Erie, may not be very healthy, so more are affected by a new mortality factor. Related to this are failed introductions, such as threadfin shad in Kansas and parts of Virginia and Ohio. These species, introduced beyond their 87 native range or expanding their range naturally, are vulnerable to cold kill; the farther north, the more likely that none will survive. Thermal plumes in North Carolina and Missouri allow threadfin shad to persist in lakes beyond their native range, where normally all would die of cold. In a different vein, systam scale has an effect on compensation. Large systems containing numerous stocks, including oceans, rivers, and the Great Lakes, are unlikely to undergo a catastrophe affecting all stocks of a given species. A localized event opens the way for immigration from adjoining stocks or populations. Immigration may prevent detection of compensation, or may actually prevent survivors of the catastrophe from compensating because fish densities do not remain low for long. We found many examples of this, particularly in rivers, where fish kills are easily observed. Catastrophic overfishing is sometimes an exception, because fishermen may range widely and affect all stocks of a species, depleting the species so that immigration becomes insignificantly small. Man, as a predator, can act in a depensatory manner because prey have had no time to evolve compensatory mechanisms strong enough to maintain their populations (Radovich 1979). 88 RESEARCH RECOMMENDATIONS INTRODUCTION In this section we recommend future research which, we believe, will significantly improve understanding of compensatory mechanisms. Our criteria for choosing research recommendations were: (1) to fill in the most important data gaps, improving our general knowledge of compensation, (2) efficiency, i.e., obtaining maximum information for minimum effort or expense, and (3) feasibility of successfully obtaining results, or amenability of the parameter to research and measurement. We have proposed further research on several species, such as clupeids and yellow perch, for which compensation is already partially understood, with the purpose of completing our knowledge and possibly generalizing to similar species. Study of mechanisms affecting, and rates of, early survival are emphasized for several species because, although many researchers believe compensation frequently occurs in early life, early survival is not well studied empirically. Recent techniques of aging fish larvae now allow us to calculate mortality, pinpoint "critical periods," and perhaps correlate times of mortality with environmentcil factors or community interactions to discover the important mechanisms, compensatory or not, controlling early survival. We have also attempted to derive methods of obtaining useful results from accidental catastrophes, and of obtaining maximum information from existing monitoring programs,. See RELATIVE IMPORTANCE OF RECOMMENDED RESEARCH for a ranking of these programs. In future studies of compensation, researchers should address the unit or level at which compensation operates. We have assumecl through most of this report that previous studies concerned one integrated population, an aggregation of stocks, which was affected as a unit by a catastrophic mortality and subsequent compensation. However, many of these studies may have actually represented some other scenario such as extinction of one stock and its replacement by another, perhaps with slightly different values of life history parameters. In such cases compensatory mechanisms would not be operating in the same fashion as they appear to. The most relevant results would be obtained from a system where the species under investigation was of one integrated population, and this is more likely to occur in a pond or small lake than in a larger system. Power plant cooling ponds and public water supply reservoirs offer opportunities for experimental manipulation of populations. System scale for such research trades off controllability - too large a system and the variables and logistics become immense - versus reality: too small a system may have too few species or too little physical variety to resemble an actual ecosystem. Also, large systems are complicated by the above-mentioned stock versus population problem. Many cooling ponds and reservoirs are appropriate sizes for population experiments, with natural species complexes sometimes available. Exploitation usually can be controlled completely by the utility. Such ponds and reservoirs exist in many geographic regions, permitting a variety of species and communities to be studied. 89 CLUPEIDS The most notable data gap for alewife, threadfin shad, and gizzard shad is survival of age-0 fish at various densities. Increased survival of young, thought to be an important result of compensation for these species, could and should be quantified and its variations delimited. We suggest a site such as Lake Norman, North Carolina, where threadfin shad are subject to repeated winter kills and respond by building up large populations each year. This site would be compared with a more southerly lake, perhaps in Florida, where mass mortalities are rare. Length-frequencies for larvae and juveniles at each site could be adjusted using otoliths to reflect actual age of the fish, and densities at each age used to calculate survival. Nutritional condition of the larvae should be compared between sites (methods in Theilacker 1978) for evidence of competition or adequacy of food supply, and adult fishes* stomachs checked in an attempt to quantify predation. If as we suspect, survival of larval and juvenile fish is greater where mass mortalities are common, it would demonstrate the existence of compensatory mechanisms, and the difference between sites should help quantify compensatory reserve. A widely fluctuating forage population undergoing depensatory predation is available for study: the Lake Michigan alewife. Population abundance indices, growth, larval fish densities, and other data are available at the Great Lakes Fishery Laboratory of U.S. Fish and Wildlife Service, back into the 1960s, and from our Cook Nuclear Plant and Campbell Plant studies at the Great Lakes Research Division for 1973-1982 and 1977-1981 respectively. The decline of the alewife in the last few years offers a unique opportunity for further study. Ongoing food habits studies for salmon (Michigan and Wisconsin Sea Grant) should be correlated with population abundance to see if depensatory predation is continuing. Also we recommend collection of growth and survival data to compare with earlier years when alewife was abundant in Lake Michigan. A relatively small amount of additional field collection would be needed, as various agencies mentioned above have already done a great deal of the work. We believe this is one of the better, most easily accomplished opportunities to quantify some parameters for a large system. Interesting results could be obtained if stocking practices in Lake Michigan were changed. A large-scale manipulation might be performed by stocking many more predators (e.g., lake trout or salmon) than usual and measuring whether survival, growth, or any other factor changes for alewife. Compensatory response may already be stretched for alewife; increasing stocking may finally drive alewife populations to low levels or zero. Increase of stocking might thus test compensatory capacity of alewife. Naturally, state Departments of Natural Resources and U.S. Fish and Wildlife Service would have to be convinced to allow this. One point in favor in their view might be improvement in fishing for the sport angler. 90 PACIFIC SALMON Compensatory mechanisms are well understood for Pacific salmon in comparison with most other species. Egg mortality is known to have density- dependent aspects and to be fairly important in population regulation. One aspect of the compensatory reserve of Pacific salmon could be tested by experime^nts to determine egg mortality at different densities. Such experimcants have been done for pink and chum salmon (Hunter 1959), and should be done in more detail and for other species as well. We envision laboratory experiments or in-stream enclosure experiments with different egg densities and different substrates to measure mortality and test the rate of dislodging due to spawning in marginal habitat or digging up and reburying of eggs. Sources of mortality - oxygen depletion, fungus infection, waste product buildup, dewatering, predation - and their relative importance should be discovered. YELLOW PERCH In order for a research effort to be useful to EPRI, it should be generic in nature, with results generally applicable. Growth is one of the final integrators of a vast plethora of interacting factors. It should be conceptually possible to develop a set of growth equations or a model for a number of regions across the U.S. Yellow perch may be a particularly good possibility for this effort, since large and extensive data sets on growth and other vital functions are available in the literature. These data would be used to generate maxima standards for various regions, r which would give a upper limit for yellow perch growth. A- utility could then determine current growth conditions in the population impacted or of interest and compare these with the growth maxima for that region to quantitatively derive an estimate of compensatory capacity. More restricted studies could be done both to test the model and to provide parameters for it. Some method for handling predation and cannibalism may have to be introduced into the equations. Pond experiments could be conceived with a factorial design using as factors: predation, harvest levels (0, 25, 50, 75%), age-group harvested (larvae, YOY, adults), competition (with bluegills, cyprinids), and possibly nutrients. Data from these experiments could shed considerable light on compensatory capacity and interactions of an impacted perch community as well as generate data on growth and response of various age-groups to catastrophic loss. FRESHWATER DRUM Research on freshwater drum and its compensatory potential should focus upon the factors affecting larval fish mortality and the subsequent relationships to year-class strength. High fecundity normally found for drum would suggest a small change in the larval fish mortality rate could have a large effect on recruitment. Therefore, a catastrophic event impacting drum and the subsequent recovery of the population would depend largely on any 91 change in the survival of the eggs and larvae • To first determine whether any change in larval drum mortality influences recruitment and year-class strength, a means to eliminate large numbers of larvae is needed. By varying the magnitude of this loss of larvae, one could th^n observe the relative impact upon the recruitment and year-class strength in later years. Such an experiment might be done with a population at a reservoir such as described by Walburg (1971). Here flushing of larvae over a dam accounted for substantial losses of drum larvae from the reservoir. Such losses could be then controlled by adjusting the passage of water through the dam or at least measured in the discharge. The subsequent impact on drum of variable losses of larvae could be followed in year-class strength. One problem with using reservoirs would be the potential immigration of larvae from upriver which could be significant. Such immigrants would have to be taken into account in assessing losses of larvae to the reservoir drum population. Determining the mechanisms of larval fish mortality for freshwater drum in natural populations and whether they were density-dependent would be complicated if several of these mechanisms (i.e., disease or parasitism, predation, starvation, stress mortality) operated simultaneously. One approach would be to monitor all the potential causes of mortality, including potential density-independent ones such as temperature, and correlate their variation to changes in larval mortality rate. This approach would probably be the most relevant in describing the relative effects of several factors. However, this method would require good measures for disease, parasitism, predator consumption of drum larvae, larval fish feeding in relation to food supply, and indications of stress. Such measures are not necessarily easily made. Furthermore, larval fish mortality would also need to be monitored for several years to account for variation in density-dependent and density- independent parameters. Larval fish mortality rates would be quantified by sampling with plankton nets to give a length-frequency relationship (see Cada and Hergenrader 1980) along with larval otolith samples to determine their age. Since subtle changes in mortality rates could have a major effect on recruitment, high precision in sample estimates of larval abundance as well as potential causes of mortality would be needed. Large lakes tend to be more variable environments with greater mobility of fish and thus more difficult to obtain high precision in sampling. Therefore, we suggest a small lake or pond with a known drum population could be best used to conduct the analysis suggested here. Such a situation might also offer the advantage of controlling to some degree some of the potential causes of larval fish mortality, such as predation, and thus give a better idea of the impacts of various population regulating factors. FOLLOW-UP STUDIES We found many examples in the literature of mass fish mortalities where no follow-up data were collected; these were more numerous than studies with subsequent data. We recommend an opportunistic approach. A fund could be established to take advantage of unfortunate accidents resulting in fish kills. When a mortality occurs, a small-scale study could begin, perhaps performed by university professors and students to limit costs. Fish could be 92 sampled at fairly wide intervals immediately after the mortality and when the population had recovered, a year or two later. Easily collected data would include length, weight, age, fecundity, sex ratio, and age at maturity. These parameters could be compared between the immediate post-mortality and later times. 1/^ile detailed quantitative studies would not be carried out this way, the result would still be more useful than much previous work which only presents amounts of dead fish. Examples of studies that should have been continueci, and perhaps still should be, are: Alcorn et al. (1980), Kushlan (1974), Turnpenny and Williams (1981), Crunkilton et al. (1980), Bury (1972), Johnston and Cheverie (1980), and Olmsted and Cloutman (1974); all are included in the annotated bibliography. Appendix 1. Various agencies are conducting bioassay and toxicity research, and accidental introduction of toxic materials into our waters occurs often enough to provide much information without the need for EPRI to put high priority on an experimental approach to compensation after toxic spills. To implement such a program, EPRI would first need to contact state agencies to which toxic spills must be reported, and ask to be notified when such spills occur. Industry biologists and various newsletters or bulletins may lead to knowledge of winterkills, disease epidemics, or other catastrophic mortalities. We recommend that EPRI set up a small coiiranittee of several people, lx>th from the utility industry and outside it, who would (1) set up, in advance, a contact list of people who could conduct the sampling, (2) authorize the sampling, and (3) set up guidelines for sampling methods, design, and analysis, including quality control and data set comparability. The committee or another consultant may be required to integrate results of many such small studies into the larger picture of compensatory response. EXPANSION OF ONGOING MONITORING PROGRAMS Becciuse of regulatory and monitoring requirements, the power industry has a number of data sets which could be used to demonstrate compensation given the manpower to analyze the data. Many of these data sets are of limited scope and have the same problems, but they could still be used to determine some life history parameters, particularly if some improvements or additions are made to sampling and analysis. Population abundance or densities of adult and larval fish and numbers of fish impinged or entrained over time are tools for helping demonstrate the compensatory reserve of fish populations. Frequently during monitoring studies, biologists collect other data such as sizes of fish, from which growth rates, age structure, or survival could be determined with a relatively small amount of extra effort. Other examples of neglectecl parameters are fecundity, age at maturity, predation/cannibalism, and sex ratio. We recommend that EPRI contribute funding to improve industry monitoring studies. Measurement of a few more parameters from fish that already must be collected would help fill out the picture of compensatory responses in these populations. On a smaller scale, we recommend that EPRI at least assign or fund a few people to interpret already-existing data sets. Many industry biologists are aware of or have contributed to extensive population estimates and 93 descriptions of fish distributions, but have had no time to suiranarize or interpret the masses of data available. There are several examples from people we spoke to for this project (Table 12). Age and growth, survival, age at maturity, and fecundity data, combined with their existing work pn biomass, numbers, and distribution of fish, would be very useful in quantifying compensatory reserve. RELATIVE IMPORTANCE OF RECOMMENDED RESEARCH We have ranked research programs we recommended according to three criteria: (1) value of the information to our understanding of compensatory response to a catastrophe, (2) relatively large amount of information obtained for a relatively small effort, and (3) feasibility of successfully obtaining results. The resultant ranking is: 1. Comparison of threadfin shad between Lake Norman and more southern lake 2. Yellow perch pond experiments 3. Depensatory predation on Lake Michigan alewife 4. Follow-up studies 5. Expansion of monitoring programs 6. Yellow perch growth models 7. Freshwater drum early life history 8. Pacific salmon egg experiments 9. Large-scale predator stocking in Lake Michigan 94 TABLE 12* Unpublished or unanalyzed data sets. Species Data or life history parameters Organization having data Alevdfe Alewife Threadfin shad Bloater Rainbow smelt Rainbow smelt Common carp Age and growth Relative abundance Age and growth Mortality estimates Biomass Abundance: decimation of population and rebound Age and growth Plankton (food) Impingement Age and growth CPUE Age and growth Relative abundance Length-frequency, CPUE Renova t ion/r emoval (Tons removed) Wisconsin Electric Power Co, U.S. Fish and Wildlife Service Duke Power Company Yellow perch Summer kill Bay anchovy (Ancfioa mi t chilli) Bay anchovy Bay anchovy Bay anchovy Maturity, competition, age composition, sex ratio, abundance, mortality, fecundity, spawning frequency Abundance Impingement Abundance Winterkill Wisconsin Dept. of Natural Resources Wisconsin Electric Power Co. Wisconsin Dept. of Natural Resources Wisconsin Dept. of Natural Resources Duke Power Company Public Service Electric and Gas Co. Baltimore Gas and Electric Co. Lawler, Matusky and Skelly Engineers New Jersey Central Power and Light Co, 95 REFERENCES Adams, S^M., R.B. McLean, and M.M. Huffman. 1982* Structuring of a predator population through temperature-mediated effects on prey availability. Can. J. Fish. Aquat. Sci. 39:1175-1184. Agersborg, H.P.K. 1930. The influence of temperature on fish. Ecology 11:136-144. Alcorn, D.J., L.E. Fancher, and J.G. Moss. 1980. Harbor seal and fish populations before and after a sewage spill in south San Francisco Bay. Calif. Fish and Game 66:238-239. Aim, G. 1946. Reasons for the occurrence of stunted fish-populations. Rep. Swedish State Inst. Freshwater Fish. Res. Drottningholm 25. Aim, G. 1952. Year class fluctuations and span of life in perch. Rep. Inst. Freshwater Res. Drottningholm 33:17-38. Anderson, E.D., and L.L. Smith, Jr. 1971. Factors affecting abundance of lake herring (Conegonus antedii Lesueur) in western Lake Superior. Trans. Amer. Fish. Soc. 100:691-707. Anthony, D.D., and C.R. Jorgensen. 1977. Factors in the declining contributions of walleye (Sti zostedion vitneum vftneum) to the fishery of Lake Nipissing, Ontario, 1960-1976. J. Fish. Res. Board Can. 34:1703- 1709. Bagenal, T.B. 1970. An historical review of the fisheries investigations of the Freshwater Biological Association, mainly at the Windermere Laboratory. J. Fish. Biol. 2:83-101. Bagenal, T.B. 1977. Effects of fisheries on Eurasian perch (Pencd fluviatj'lis) in Windermere. J. Fish. Res. Board Can. 34:1764-1768. Bagenal, T.B. 1982. Experimental manipulations of the fish populations in Windermere . Hydrobiologia 86 : 201-206 . Bailey, M.M. 1969. Age, growth, maturity, and sex composition of the American smelt, Oswenus mondax (Mitchill), of western Lake Superior. Trans. Amer. Fish. Soc. 93:382-395. Bailey, W.M. 1978. A comparison of fish populations before and after extensive grass carp stocking. Trans. Amer. Fish. Soc. 107:181-206. Baldwin, N.S., R.W. Saalfeld, M.A. Ross, and H.J. Buettner. 1979. Commercial fish production in the Great Lakes. Great Lakes Fish. Comm. Tech. Rep. 3:187 p. Bardach, J.E. 1951. Changes in the yellow perch population of Lake Mendota, Wisconsin, between 1916 and 1948. Ecol. 32:719-728. 96 Beamish, R.J, 1976. Acidification of lakes in Canada by acid precipitation and the resulting effects on fishes. In Proc. Internat. Symp. Acid Precip. Forest Ecosystem. USDA Forest Ser. Gen. Tech. Rep. NE-23. 477 pp. Beckman, W.C. 1941. Increased growth rate of rock biass, AmblopJites nupestnis (Rafinesque), following reduction in the density of the population. Trans. Amer. Fish. Soc. 70:143-148. Beckman, W.C. 1943. Further studies on the increased growth rate of the rock bass AmblopJites nupestnis (Rafinesque), following the reduction in density of the population. Trans. Amer. Fish. Soc. 72:72-78. Beckman, W.C. 1950. Changes in growth rates of fishes following reduction in population densities by winterkill. Trans. Amer,, Fish. Soc. 78:82-90. Beers, G.D., and W.J. McConnell. 1966. Some effects of threadfin shad introduction on black crappie diet and condition. J. Arizona Acad. Sci. 4:71-74. Beeton, A.M. 1961. Environmental changes in Lake Erie. Trans. Amer. Fish. Soc. 90:153-159. Bell, G., P. Handford, and C. Dietz. 1977. Dynamics of an exploited population of lake whitefish (Conegonus cJupeafonmis) . J. Fish. Res. Board Can. 34:942-953. Bergstedt, R., and R. O'Gorman. 1983. Lake Ontario fish stocks: surveillance and status of fish populations. Progress in Fishery Research 1983. Great Lakes Fish. Lab., U.S. Fish and Wildl. Serv., Ann Arbor, MI: 15-16. Berry, F.H., M.T. Huish, and H. Moody. 1956. Spawning mortality of the threadfin shad, Donosotna petenense (Gunther), in Florida. Copeia 1956:192. Berst, A.H., and G.R. Spangler. 1972. Lake Huron: effects of exploitation, introductions, and eutrophication on the salmonid community. J. Fish. Res. Board Can. 29:877-887. Beverton, R.J.H., and S.H. Holt. 1957. On the dynamics of exploited fish populations. Fishery Invest. Lond. Ser. 2(19). Bidgood, B.F. 1973. Divergent growth in two lake whitefish (Conegonus clupeafonmi S) populations. J. Fish. Res. Board Can. 30:1683-1696. Birch, L.C. 1948. The intrinsic rate of natural increase of an insect population. J. Animal Ecol. 17: 15-26 « Blaxter, J.H.S., and G. Hempel. 1963. The influence of egg size on herring larvae (CJupea hanengus). J. Cons. Perm. Explor. Mer 28:211-240. 97 Boaze, J.L,, and R.T. Lackey. 1974. Age, growth, and utilization of landlocked alewives in Claytor Lake Virginia. Prog. Fish-Cult. 36:163- 164. Bodola, A. 1966. Life history of the gizzard shad, Donosoma cepedianum (LeSueur), in western Lake Erie. U.S. Fish and Wildl. Serv. Fish. Bull. 65:391-425. Borawa, J.C., J.H. Kerby, M.T. Huish, and A.W. Mullis. 1978. Currituck Sound fish populations before and after infestation by Eurasian Water -Milfoil. Proc. Ann. Conf . S.E. Assoc. Fish and Wildl. Agencies 32:520-528. Brooks, J.L., and S.I. Dodson. 1965. Predation, body size, and composition of plankton. Science 150:28-35. Cited in C.C. Kohler (1980). Brown, E.H., Jr. 1968. Population characteristics and physical condition of alewives, AJosa pseudohanenguSf in a massive dieoff in Lake Michigan, 1967. Great Lakes Fishery Commission, Technical Report No. 13, 20 pp. Brown, E.H., Jr. 1970. Extreme female predominance in the bloater (Coregonus hoyi) of Lake Michigan in the 1960's. pp. 501-514 in: C.C. Lindsey and C.S. Woods (eds.). Biology of coregonid fishes. Univ. Manitoba Press, Winnipeg, Manit. Brown, E.H., Jr. 1972. Population biology of alewives, Alosa pseudofianengus, in Lake Michigan, 1949-70. J. Fish. Res. Board Can. 29:477-500. Bryant, H.E., and A. Houser. 1968. Growth of threadfin shad in Bull Shoals Reservoir. Proc. Ann. Conf. S.E. Assoc. Game and Fish Comm. 22:275-283. Burns, J.W. 1966. Threadfin shad. pp. 481-488 in: A. Calhoun (ed.) Inland fisheries management. California Dept. of Fish and Game, Sacramento. Burrough, R.J., and C.R. Kennedy. 1979. The occurrence and natural alleviation of stunting in a population of roach, RutiJus nutilus (L.). J. Fish Biol. 15:93-109. Burrough, R.J., P.R. Bregazzi, and C.R. Kennedy. 1979. Interspecific dominance amongst three species of coarse fish in Slapton Ley, Devon. J. Fish Biol. 15:535-544. Bury, R.B. 1972. The effects of diesel fuel on a stream fauna. Calif. Fish and Game 58:291-295. Busch, W.-D.N., R.L. Scholl, and W.L. Hartman. 1975. Environmental factors affecting the strength of walleye (Stfzostedi on vitneum vitneum) year- classes in western Lake Erie, 1960-1970. J. Fish. Res. Board Can. 32:1733-1743. 98 Butler, R.L. 1965, Freshwater drum, ApJodinotus gnunniens, in the navigational impoundments of the upper Mississippi River. Trans* Amer. Fish. Soc. 94:339-349. Butler, R.L., and L.L. Smith Jr. 1949. The age and rate of growth of the sheepshead, ApJodinotus gnunniens Rafinesque, in the upper Mississippi River navigation pools. Trans. Amer. Fish. Soc. 79:43-54. Cada, G.F., and G.L. Hergenrader. 1980. Natural mortality rates of freshwater drum larvae in the Missouri River. Trans. Amer. Fish. Soc. 109:479-483. Car lander, K.D. 1945. Age, growth, sexual maturity, and population fluctuations of the yellow pike-perch, Sti zostedion vitneutn vitneum (Mitchill), with reference to the commercial fisheries. Lake of the Woods, Minnesota. Trans. Amer. Fish. Soc. 73:90-107. Chevalier, J.R. 1973. Cannibalism as a factor in first year survival of walleye in Oneida Lake. Trans. Amer. Fish. Soc. 102:739-744. Chevalier, J.R. 1977. Changes in walleye (Sti zostedion vitneum vitneum) population in Rainy Lake and factors in abundance, 1924-75. J. Fish. Res. Board Can. 34:1696-1702. Christie, W.J. 1968. Possible influences of fishing in the decline of Great Lakes fish stocks. Proc. Conf. Great Lakes Res. 11: 31-38. Christie, W.J. 1972. Lake Ontario: effects of exploitation, introductions, and eutrophication on the salmonid community. J. Fish. Res. Board Can. 29:913-929. Christie, W.J. 1973. A review of the changes in the fish species composition of Lake Ontario. Great Lakes Fishery Comm. Tech. Rep. No. 23. 65 pp. Christie, W.J. 1974. Changes in the fish species composition of the Great Lakes. J. Fish. Res. Board Can. 31:827-854. Christie, W.J., J.M. Fraser, and S.J. Nepszy. 1972. Effects of species introductions on salmonid communities in oligotrophic lakes. J. Fish. Res. Board Can. 29:969-973. Clady, M.D. 1976. Influence of temperature and wind on the survival of early stages of yellow perch, Penca fJavescens. J. Fish. Res. Board Can. 33:1887-1893. Clark, C. 1974. Possible effects of schooling on the dynamics of exploited fish populations. J. Cons. Cons. Int. Explor. Mer 36:7-14. Clepper, H. (ed.). 1979. Predator-prey systems in fisheries management. Sport Fisheries Institute. 99 Colby, P.J. 1971. Alewife dieoffs: why do they occur? Limnos 4:18-27. Colby, P.J. 1973. Response of the alewives, Alosa pseudohanengus, to environmental change, pp. 163-198 in: W. Chavin (ed.). Responses of fish to environmental changes. C.C. Thomas, Springfield, XL. Colby, P.J., G.R. Spangler, D.A. Hurley, and A.M. McCombie. 1972. Effects of eutrophication on salmonid communities in oligotrophic lakes. J. Fish. Res. Board Can. 29:975-983. Colby, P.J., R.E. McNicol, and R.A. Ryder. 1979. Synopsis of biological data on the walleye Stizostedion v. vitneuiu (Mitchill 1818). FAO Fisheries Synopsis No. 119. 133 pp. Cole, C.F., J. Trenary, and S. Finkelstein. 1958. Experimental introduction of threadfin shad. Ark. Game and Fish Comm., Ann. Rep. Cited in Burns (1966). Combs, D.L. 1980. Fish population changes in Keystone Reservoir fourteen years after striped bass introductions. Proc. Ann. Conf . S.E. Assoc. Fish, and Wildl. Agencies 34:167-174. Cox, D.K., and C.C. Coutant. 1976. Acute cold-shock resistance of gizzard shad. pp. 159-161 in: G.W. Esch and R.W. McFarlane, eds. Thermal ecology II. Energy Research and Devt. Admin., Tech. Info. Ctr., U.S. Dept. Comm. Springfield, VA. Craig, J.F. 1980. Growth and production of the 1955 to 1972 cohorts of perch, Perca fJuviatiJis L., in Windermere. J. Anim. Ecol. 49:291-316. Craig, J.F., C. Kipling, E.D. LeCren, and J.C. McCormack. 1979. Estimates of the numbers, biomass and year-class strengths of perch Perca fluvfatilfs L. in Windermere from 1967 to 1977 and some comparisons with earlier years. J. Anim. Ecol. 48:315-325. Creaser, C.W. 1929. The food of yearling smelt from Michigan. Papers Mich. Acad. Sci. Arts Letters 10:427-431. Crecco, V., T. Savoy, and L. Gunn. 1983. Daily mortality rates of larval and juvenile American shad {Alosa sapfdf ssima) in the Connecticut River with changes in year-class strength. Can. J. Fish. Aquat. Sci. 40:1719-1728. Crowder, L.B. 1980. Alewife, rainbow smelt and native fishes in Lake Michigan: competition or predation? Env. Biol. Fish. 5:225-233. Crunkilton, R.L., J.M. Czarnezki, and L. Trial. 1980. Severe gas bubble disease in a warmwater fishery in the midwestern United States. Trans. Amer. Fish. Soc. 109:725-733. Cucin, D., and H.A. Regier. 1965. Dynamics and exploitation of lake whitefish in southern Georgian Bay. J. Fish. Res. Board Can. 23:221-274. 100 Gushing, D.H. 1974. The possible density-dependence of larval mortality and adult mortality in fishes, pp. 103-111 in: J.H.S. Blaxter (ed.). The early life history of fish. Springer-Verlag, N.Y. 765 pp. Gushing, D.H. 1977. The problems of stock and recruitment, pp. 116-133 in: J. A. Gulland, ed. Fish population dynamics. Wiley & Sons, London. Daiber, F.G. 1953. Notes on the spawning population of the freshwater drum (Aplodi notus gnunniens Rafinesque) in Western Lake Erie. Amer. Midi. Nat. 50:159-171. Davidoff, E.B., R.W. Rybicki, and K.H. Doan. 1973. Ghanges in the population of lake whitefish (Conegonus clupeafonmi s) in Lake Winnipeg from 1944 to 1969. J. Fish. Res. Board Gan. 30:1667-1682. Davis, J. 1959. Management of channel catfish in Kansas. Univ. Kansas, Mus. Nat. Hist. Misc. Publ. No. 21. 56 pp. Dechtiar, A.O. 1972. New parasite records for Lake Erie fish. Great Lakes Fish. Gomra. Tech. Rep. 17:20 p. Deelder, G.L. 1951. A contribution to the knowledge of the stunted growth of perch (Perca fJuvfatiJis L.) in Holland. Hydrobiologia 3:357-378. Eck, G.W., and L. Wells. 1983. Biology, population structure, and estimated forage requirements of lake trout in Lake Michigan. Technical Paper 111. U.S. Fish and Wildl. Serv. Elrod, J.H., R. O'Gorman, R. Bergsfedt, and G.P. Schneider. 1980. Status of major forage fish stocks, U.S. waters of Lake Ontario, pp. 155-163 in: Great Lakes Fish. Gorom., Lk. Ontario Gommittee Meeting, Appendix 13. El-Zarka, S. 1959. Fluctuations in the population of yellow perch, Penca flavescens (Mitchill), in Saginaw Bay, Lake Huron. U.S. Fish Wildl. Serv. Fish. Bull. 59:365-415. Eschmeyer, R.W. 1936. Some characteristics of a population of stunted perch. Papers Mich. Acad. Sci., Arts and Letters 22:613-628. Eschmeyer, R.W., R.H. Stroud, and A.M. Jones. 1944. Studies of the fish population on the shoal area of a TVA mainstream reservoir. Rep. Reelfoot Lake Bio. Sta. 8:70-122. Eshenroder, R.L. 1977. Effects of intensified fishing, species changes, and spring water temperatures on yellow perch, Penca fJavescens, in Saginaw Bay« J. Fish. Res. Board Gan. 34:1830-1838. Finucane, J.H. 1965. Threadfin shad in Tampa Bay, Florida. Quart. J. Florida Acad. Sci. 28:267-270. Gited in Johnson (1970). 101 Flath, L.E., and J.S. Diana. 1985. Seasonal energy dynamics of the alewife in southeastern Lake Michigan. Trans. Amer. Fish. Soc. 114:328-337. Foerster, R.E. 1968. The sockeye salmon, Onconfiynchus nerka. Bull. 162. Fish. Res. Board Canada/Ottawa. 422 p. Forney, J.L. 1965. Factors affecting growth and maturity in a walleye population. N.Y. Fish & Game J. 12:217-232. Forney, J.L. 1971. Development of dominant year classes in a yellow perch population. Trans. Amer. Fish. Soc. 100:739-749. Forney, J.L. 1974. Interactions between yellow perch abundance, walleye predation, and survival of alternate prey in Oneida Lake, New York. Trans. Amer. Fish. Soc. 103:15-24. Forney, J.L. 1976. Year-class formation in the walleye (Stizostedion vitreum vitneum) population of Oneida Lake, New York, 1966-73. J. Fish. Res. Board Can. 33:783-792. Forney, J.L. 1977. Evidence of inter- and intraspecific competition as factors regulating walleye (Stf zostedion vitneuw vitneum) biomass in Oneida Lake, New York. J. Fish. Res. Board Can. 34:1812-1820. Forney, J.L. 1980. Evolution of a management strategy for the walleye in Oneida Lake, New York. N.Y. Fish Game J. 27:105-141. Forney, J.L. 1984. Status of fish stocks in Chautauqua Lake - 1983. Warmwater Fisheries Unit Job IV-b. 36 pp. Fraser, J.M. 1978. The effect of competition with yellow perch on the survival and growth of planted brook trout, splake, and rainbow trout in a small Ontario lake. Trans. Amer. Fish. Soc. 107: 505-517. Gaboury, M.N., and J.W. Patalas. 1984. Influence of water level drawdown on the fish populations of Cross Lake, Manitoba. Can. J. Fish. Aquat. Sci. 41:118-125. Gammon, J.R. 1971. The response of fish populations in the Wabash River to heated effluents. Proc. 3rd Nat. Symp. Radioecology: 513-523. Gerking, S.D., ed. 1967. The biological basis of freshwater fish production. Blackwell Scientific Publications, Oxford. 495 pp. Gerking, S.D. 1976. Report on fecundity in the Waukegan-Zion sampling area of Lake Michigan for Commonwealth Edison. In: Nalco Envir. Sci., Compilation of reports relating to entrainment and impingement studies at Zion Generating Station, 1976, vol. II. 102 Gocxiyear, C.P. 1977. Assessing the impact of power plant mortality on the compensatory reserve of fish populations, pp. 186-195 in: W. Van Winkle, ed. Assessing the effects of power-plant-induced mortality on fish populations. Pergamon Press, New York. Goodyear, C.P. 1980. Compensation in fish populations, pp. 253-280 in Hocutt, C.H. and J.R. Stauffer, Jr. (eds.) Biological monitoring of fish. D.C. Heath & Co., Lexington, Massachusetts. 416 pp. Gordon, W.G. 1961. Food of the American smelt in Saginaw Bay, Lake Huron. Trans. Am.er. Fish. Soc. 90:439-443. Graham, J.J. 1956. Observations on the alewife PomoJobus pseudofianengus (Wilson), in fresh water. Univ. Toronto Biol. Ser. 62, Publ. Ont. Fish. Res. Lab. 74. 43 pp. Great Lakes Fishery Laboratory. 1970. Physical and ecological effects of waste heat on Lake Michigan. U.S. Fish and Wildl. Serv. 101 pp. Grice, F. 1957. Effect of removal of panfish and trashfish by fyke nets upon fish populations of some Massachusetts ponds. Trans. Amer. Fish. Soc. 87:108-115. Grice, F. 1959. Elasticity of growth of yellow perch, chain pickerel, and largemouth bass in some reclaimed Massachusetts waters. Trans. Amer. Fish. Soc. 88:332-335. Gross, R.W.' 1959. A study of the alewife AJosa pseudo/ianengus (Wilson) in some New Jersey lakes, with special reference to Lake Hopatcong. M.S. thesis, Rutgers Univ., New Brunswick, NJ. Gullaiid, J. A. 1965. Survival of the youngest stages of fish and its relation to year-class strength. Spec. Pubis, int. Comm. N.W. Atlantic Fish. 6:365-371. Gulland, J. A., ed. 1977. Fish population dynamics. J. Wiley & Sons, London. 372 pp. Hale, J. 1960. Some aspects of the life history of the smelt (Osmenus tnondax) in western Lake Superior. Minn. Fish Game Invest., Fish Invest. Rep. No. 204, p. 25-41. Hanson, D.L., and T.F. Waters. 1974. Recovery of standing crop and production rate of a brook trout population in a flood-damaged stream. Trans. Amer. Fish. Soc. 103:431-439. Harkness, W.J.K. 1922. The rate of growth of the yellow perch in Lake Erie. Univ. Toronto Stud. (Biol. Ser.) Publ* Ontario Fish. Res. Lab. 20:88-95. 103 Hartmaiii W.L. 1972. Lake Erie: effects of exploitation, environmental changes, and new species on the fishery resources. J. Fish. Res. Board Can. 29:899-912. Havey, K.A. 1961. Restoration of anadromous alewives at Long Pond, Maine. Trans. Amer. Fish. Soc. 90:281-286. Healey, M.C. 1975. Dynamics of exploited whitefish populations and their management with special reference to the Northwest Territories. J. Fish. Res. Board Can. 32:427-448. Healey, M.C. 1980. Growth and recruitment in experimentally exploited lake whitefish (Conegonus cJupeafonmis) populations. Can. J. Fish. Aquat. Sci. 37:255-267. Hempel, G., and J.H.S. Blaxter. 1967. Egg weight in Atlantic herring {CI u pea fianengus L.). J. Cons. Perm. int. Explor. Mer 31:170-195. Hesse, L.W., C.R. Wallace, and L. Lehman. 1976. Fishes of the channelized Missouri. Nebraska Game & Parks Comm. Tech. Ser. No. 4. 61 pp. Higley, D.L., and C.E. Bond. 1973. Ecology and production of juvenile spring Chinook salmon, Oncophynchus ts/jawytscha, in a eutrophic reservoir. U.S. Natl. Mar. Fish. Serv. Bull. 71:877-891. Hjort, J. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp. P. -v. Reun. Cons. Int. Explor. Mer 20:1-228. Holling, C.S. 1973. Resilience and stability of ecological systems. Ann. Rev. Ecol. Syst. 4:1-23. Houser, A. 1960. Growth of freshwater drum in Oklahoma. Oklahoma Fish. Res. Lab. Rep. 78:13 p. Huish, M.T. 1957. Studies of gizzard shad reduction at Lake Beulah, Florida. Proc. Ann. Conf. S.E. Assoc. Game & Fish. Comm. 11:66-70. Hulsey, A.H. 1956. Effects of a fall and winter drawdown on a flood control lake. Proc. Ann. Conf. S.E. Assoc. Game & Fish. Comm. 10:285-289. Hunter, J.G. 1959. Survival and production of pink and chum salmon in a coastal stream. J. Fish. Res. Board Can. 16:835-886. Jackson, S.W., Jr. 1957. Comparison of the age and growth of four fishes from Lower and Upper Spavinaw Lakes, Oklahoma. Proc. Ann. Conf. S.E. Assoc. Game Se Fish Comm. 11:232-249. Jenkins, R.M. 1957. The effect of gizzard shad on the fish population of a small Oklahoma lake. Trans. Amer. Fish. Soc. 85:58-74. 104 Jensen, A.L. 1981. Population regulation in lake whitefish, Conegonus cJupeafonmfs (Mitchill). J. Fish Biol. 19:557-574. Jester, D.B., and B.L. Jensen. 1972. Life history and ecology of the gizzard shad, Dorosoma cepedianum (LeSueur) with reference to Elephant Butte Lake. New Mexico State Univ. Agric. Exper. Sta. Res. Rep. 218. 56 pp. Jobes, F.W. 1952. Age, growth and production of yellow perch in Lake Erie. U.S. Fish Wildl. Serv. Fish. Bull. 52:205-266. Johnson, F.H. 1977. Responses of walleye (Stizostedion vitneum vitneum) and yellow perch (Penca fJavescens) populations to reanoval of white sucker (Catostomus comersoni) from a Minnesota lake, 1966. J. Fish. Res. Board Can. 34:1633-1642. Johnson, J.E. 1970. Age, growth and population dynamics of threadfin shad, Donosoma petenense (Giinther), in central Arizona reservoirs. Trans. Amer. Fish. Soc. 99:739-753. Johnson, J.E. 1971. Maturity and fecundity of threadfin shad, DoPOSoma petenense (Giinther), in central Arizona reservoirs. Trans. Amer. Fish. Soc. 100:74-85. Johnson, L. 1976. Ecology of Arctic populations of lake trout, SaJveJinus namaycush, lake whitefish, Conegonus cJupeafonmis, Arctic char, SaheJinus aJpinus, and associated species in unexploited lakes of the Canadian Northwest Territories. J. Fish. Res. Board Can. 33:2459-2488. Johnson, W.E. 1965. On mechanisms of self -regulation of population abundance in OncoPhynchus nepka. Mitt. Inst. Verein. Theor. Agnew. Limnol. 13:66- 87. Johnston, C.E., and J.C. Cheverie. 1980. Repopulation of a coastal stream by brook trout and rainbow trout after endosulfan poisoning. Prog. Fish- Cult. 42:107-110. Jones, D.D., and C.J. Walters. 1976. Catastrophe theory and fisheries regulation. J. Fish. Res. Board. Can. 33:2829-2833. Jude, D.J., and F.J. Tesar. 1985. Recent changes in the inshore forage fish of Lake Michigan. Can. J. Fish. Aquat. Sci. 42:1154-1157. Kelso, J.R.M., and T.B. Bagenal. 1977. Percids in unperturbed ecosystems. J. Fish. Res. Board Can. 34:1959-1963. Kendall, W.C. 1927. The smelts. Bull. U.S. Bur. Fish. 42:217-375. Kenyon, R.B. 1978. Growth changes in a population of whitefish in Lake Erie. N.Y. Fish Game J. 25:129-139. 105 Kilambi, R.V., and R^E, Baglin, Jr. 1969. Fecundity of the threadfin shad, Donosoma petenense, in Beaver and Bull Shoals Reservoirs. Trans. Amer. Fish. Soc. 98:320-322. Kimsey, J.B., R.H. Hagy, and G.W. McGairanon. 1957. Progress report of the Mississippi threadfin shad, DoPOSOtna petenensfs atchafaylae in the Colorado River for 1956. Calif. Fish & Game, Inland Fish Admin. Rep. 57- 23. 48 pp. Cited in Johnson (1971). Kohler, C.C. 1980. Trophic ecology and impact of alewife in Claytor Lake, Virginia. Ph.D. dissertation. Virginia Polytechnic Inst, and State Univ., Blacksburg, VA. 192 pp. Kohler, C.C, and J.J. Ney. 1980. Piscivority in a land-locked alewife {Alosa pseudohanengus) population, can. J. Fish. Aquat. Sci. 37: 1314- 1317. Kohler, C.C, and J.J. Ney. 1981. Consequences of an alewife die-off to fish and zooplankton in a reservoir. Trans. Amer. Fish. Soc. 110: 360-369. Kushlan, J. A. 1974. Effects of a natural fish kill on the water quality, plankton, and fish population of a pond in the Big Cypress Swamp, Florida. Trans. Amer. Fish. Soc. 103:235-243. Laarman, P.W. 1969. Effects of a limited food supply on growth rates of coho salmon and steelhead trout. Trans. Amer. Fish. Soc. 98:393-397. Lack, D. 1954. The natural regulation of animal numbers. Clarendon, Oxford. 343 p. Lackey, R.T. 1970. Observations on newly introduced landlocked alewives in Maine. N.Y. Fish & Game J. 17:110-116. Lagler, K.F., and V.C Applegate. 1942. Age and growth of the gizzard shad, Donosoma cepedianum (LeSueur), with a discussion of its value as a buffer and as forage of game fishes. Invest. Ind. Lakes Se Streams 2:99-110. Lambou, V.W. 1965. Observations on size distribution and spawning behavior of threadfin shad. Trans. Amer. Fish. Soc. 94:385-386. Larkin, P. A. 1977. Pacific salmon, pp. 156-181 in: J. A. Gulland (ed.). Fish population dynamics. John Wiley Se Sons. New York. 372 pp. Lawler, G.H. 1965. Fluctuations in the success of year-classes of whitefish populations with special reference to Lake Erie. J. Fish. Res. Board Can. 22:1197-1227. Lawrie, A.H., and J.F. Rahrer. 1972. Lake Superior: effects of exploitation and introductions on the salmonid community. J. Fish. Res. Board Can. 29:765-776. 106 Leach, J.H., and S.J. Nepszy. 1976. The fish community in Lake Erie. J. Fish. Res. Board Can. 33:622-638. Leach, ^J.H., M.G. Johnson, J.R.M. Kelso, J. Hartmann, W. Numann, and B. Entz. 1977. Responses of percid fishes and their habitats to eutrophication . J. Fish. Res. Board Can. 34:1964-1971. LeCren, E.D. 1958. Observations on the growth of perch (Penca fluviatilis L.) over twenty-two years with special reference to the effects of temperature and changes in population density. J. Anim. Ecol. 27:287- 334. LeCren, E.D. 1965. Some factors regulating the size of populations of freshwater fish. Mitt. Inst. Vereino Theor. Agnew. Limnol. 13:88-105. LeCren, E.D., C. Kipling, and J.C. McCormack. 1977. A study of the numbers of perch (Penca fluviatilis L.) in Windermere from 1941 to 1966. J. Anim. Ecol. 46:281-307. MacLean, J., and J.J. Magnuson. 1977. Species interactions in percid communities. J. Fish. Res. Board Can. 34:1941-1951. Mahon, R., and M. Ferguson. 1981. Invasion of a new reservoir by fishes: species composition, growth, and condition. Canadian Field-Naturalist 95:272-275. Mansfield, P.J., and D.J. Jude. In press. Alewife {Alosa pseudofianengus) survival during the first growth season in southeastern Lake Michigan. Can.. J. Fish. Aquat. Sci. 43. Marzolf, R.C. 1957. The reproduction of channel catfish in Missouri ponds. J. Wildl. Mgt. 21:22-28. McCammon, G.W., and C. von Geldern, Jr. 1979. Predator-prey systems in large reservoirs, pp. 431-442 in: H. Clepper, ed. Prcada tor -prey systems in fisheries management. Sport Fishing Inst., Washington, D.C. McConnell, W.J., and J.H. Gerdes. 1964. Threadfin shad, Donosoma petenense, as food of yearling centrarchids . Calif. Fish and Game 50:170-175. McFadden, J.T. 1977. An argument supporting the reality of compensation in fish populations and a plea to let them exercise it. pp. 153-183 in: W. Van Winkle, ed. Assessing the effect of power-plant induced mortality on fish populations. Pergamon Press, New York. McGee, M.V., J.S. Griffith, and R. McLean. 1977. Prcsy selection by sauger in Watts Bar Reservoir, Tennessee, as affected by cold-induced mortality of threadfin shad. Proc. Ann. Conf . S.E. Assoc. Fish Se Wildl. Agencies 31:404-411. 107 McLean, R^B,, P.T. Singley, and D. Lcxige. 1981. Threadfin shad impingement: population response. ORNL/NUREG/TM-339, Oak Ridge Natl. Lab., Oak Ridge, TN. 63 pp. Meekin, T.K., and R.L. Allen. 1974. Summer chinook and sockeye salmon mortality in the upper Columbia River and its relationship to nitrogen super saturation. Wash. Dept. of Fisheries, Tech. Rep. 12: 127-153. Miller, R.B. 1947. The effects of different intensities of fishing on the whitefish populations of two Alberta lakes. J. Wildlife Mgt. 11:289- 301. Miller, R.R. 1957. Origin and dispersal of the alewife AJosa pseudo/janengus, and the gizzard shad, Doposoma cepedianuui, in the Great Lakes. Trans. Amer. Fish. Soc. 86:97-111. Miller, R.R. 1960. Systematics and biology of the gizzard shad Donosoma cepedfanum and related fishes. U. S. Fish & Wildl. Serv. Fish. Bull. 173:370-392. Mills, E.L., and J.L. Forney. 1981. Energetics, food consumption and growth of young yellow perch in Oneida Lake, New York. Trans. Amer. Fish. Soc. 110:479-488. Minckley, W.S., and L. Krumholtz. 1960. Natural hybridization between the clupeid genera Do no soma and SfgnaJosa with a report on the distribution of S. petenense (Giinther). Zoologica 45:171-182. Moody, H.L. 1957. A fisheries study of Lake Panasoffkee, Florida. Quart. J. Florida Acad. Sci. 20:21-88. Morin, R., J.J. Dodson, and G. Power. 1982. Life history variations of anadromous cisco (Conegonus antedfi), lake whitefish (C. clupeafonmi s) , and round whitefish (Pnosopfuff! cyJfndnaceum) populations of eastern James-Hudson Bay. Can. J. Fish. Aquat. Sci. 39:958-967. Nakashima, B.S., and W.C. Leggett. 1975. Yellow perch (Penca flavescens) biomass responses to different levels of phytoplankton and benthic biomass in Lake Memphremagog, Quebec-Vermont. J. Fish Res. Board Can. 32:1785-1797. Neave, F. 1953. Principles affecting the size of pink and chum salmon populations in British Columbia. J. Fish. Res. Board Can. 9:450-491. Nelson, W.R. 1974. Age, growth, and maturity of thirteen species of fish from Lake Oahe during the early years of impoundment, 1963-68. U.S. Fish and Wildl. Tech. Pap. 77:29 pp. 108 Nelson, W.R., and C.H. Walburg. 1977. Population dynamics of yellow perch {Penca flavescens), sauger {Stizostedion canadense), and walleye (S, vitneum vitneum) in four main stem Missouri River reservoirs. J. Fish. Res. Board Can. 34:1748-1763. Nepszy, S.J. 1977. Changes in percid populations and species interactions in Lake Erie. J. Fish. Res. Board Can. 34:1861-1868. Nielsen, L.A. 1980. Effect of walleye (Stf zostedi on vitneum vitneum) predation on juvenile mortality and recruitment of yellow perch (Penca flavescens) in Oneida Lake, New York. Can. J. Fish. Aquat. Sci. 37:11- 19. Nigro, A. A. 1980. Reproduction and early life history of a landlocked population of alewife (AJosa pseudohanengus) in Claytor Lake, Virginia. M. S. thesis, Virginia Polytechnic Inst, and State Univ., Blacksburg, VA. 85 pp. NordcMi, C.R. 1967. Age, growth, and fecundity of the alewife, Alosa pseudohanengus (Wilson), in Lake Michigan. Trans. Amer. Fish. Soc. 96:387-393. Northcote, T.G. 1972. Kootenay Lake: man's effects on the salmonid community. J. Fish. Res. Board Can. 29:861-865. Odell, T.T. 1934. The life history and ecological relationships of the alewife (PomoJobus pseudohanengus [Wilson]) in Seneca Lake, New York. Trans. Amer. Fish. Soc. 64:118-126. O'Gorman, R. 1974. Predation by rainbow smelt (Osmenus mo nd ax) on young-of- the-year alewives (AJosa pseudohanengus) in the Great Lakes. Prog. Fish- Cult. 36:223-224. Olmsted, L.L., and D.G. Cloutman. 1974. Repopulation after a fish kill in Mud Creek, Washington County, Arkansas following pesticide pollution. Trans. Amer. Fish. Soc. 103:79-87. Parsons, J.W. 1970. Walleye fishery of Lake Erie in 1943-62 with emphasis on contributions of the 1942-61 year-classes. J. Fish. Res. Board Can. 27:1475-1489. Parsons, J.W. 1971. Selective food preferences of walleyes of the 1959 year class in Lake Erie. Trans. Amer. Fish. Soc. 100:474-485. Parsons, J.W., and J.B. Kimsey. 1954. A report on the Mississippi threadfin shad. Prog. Fish-Cult. 16:179-181. Patriarche, M.H., and R.S. Campbell. 1957. The development of the fish population in a new flood-control reservoir in Missouri, 1948 to 1954. Trans. Amer. Fish. Soc. 87:240-258. 109 Peterman, R.M. 1984. Density-dependent growth in early ocean life of sockeye salmon (Oncopfiyncfius nenka). Can. J. Fish. Aquat. Sci. 41:1825-1829. Pflieger, W.L. 1975. The fishes of Missouri. Mo. Dept. Conserv. 343 pp. Phillippy^ C.L. 1964. Results of selective shad treatments in six central Florida lakes. Proc. Ann. Conf . S.E. Assoc. Game & Fish. Comm. 18:198- 212. Price, J.W. 1963. A study of the food habits of some Lake Erie fish. Bull. Ohio Biol. Surv., New Ser. 2. 89 pp. Priegelr G.R. 1969. Age and rate of growth of the freshwater drum in Lake Winnebago, Wisconsin. Trans. Amer. Fish. Soc. 98:116-118. Priegel, G.R. 1971. Evaluation of intensive freshwater drum removal in Lake Winnebago, Wisconsin, 1955-1966. Tech. Bull. Wis. Dept. Nat. Resour. 47:28 p. Pritchard, A.L. 1929. The alewife (PomoJobus pseudofianengus) in Lake Ontario. Univ. Toronto Stud., Publ. Ont. Fish. Res. Lab. 38:39-54. Radovich, J. 1979. Managing pelagic schooling prey species, pp. 365-375 in: H. Clepper, ed. Predator-prey systems in fisheries management. Sport Fishing Inst., Washington, D.C. Raleigh, R.F., D.H. Bennett, L.O. Mohn, and O.E. Maughan. 1978. Changes in fish stocks after major fish kills in the Clinch River near St. Paul, Virginia. Amer. Midi. Nat. 99:1-9. Raymond, H.L. 1979. Effects of dams and impoundments on migrations of juvenile chinook salmon and steelhead from the Snake River, 1966 to 1975. Trans. Amer. Fish. Soc. 108:505-529. Regier, H.A., and W.L. Hartman. 1973. Lake Erie's fish community: 150 yr of cultural stress. Science 180:1248-1255. Regier, H.A., V.C. Applegate, and R.A. Ryder. 1969. The ecology and management of the walleye in western Lake Erie. Great Lakes Fish. Comm. Tech. Rep. 15. 101 pp. Rhodes, R.J., D.A. Webb, and T.S. McComish. 1974. Cannibalism by the adult alewife (AJosa pseudofianengus) in southern Lake Michigan. Proc. 17th Conf. Great Lakes Research, Intern. Assoc. Great Lakes Res. :593-595. Richkus, W.A. 1975. Migratory behavior and growth of juvenile anadromous alewives AJosa pseudofianengus, in a Rhode Island drainage. Trans. Amer. Fish. Soc. 104:483-493. Ricker, W.E. 1954. Stock and recruitment. J. Fish. Res. Board Can. 11:559- 623. 110 Ricker, W.E, 1962. Regulation of the abundance of pink salmon populations. pp« 155-201 in: N.J. Wilimovsky, ed. Symposium on pink salmon. Univ. British Columbia, Vancouver, B.C., Canada. Ricker, W.E. 1963. Big effects from small causes: Two examples from fish population dynamics. J. Fish. Res. Board Can. 20:257-264. Ricker, W.E. 1977. The historical development, p. 1-26 7/7 : J. A. Gulland, ed. Fish population dynamics. J. Wiley & Sons. London. Ricker, W.E., and J. Gottschalk. 1940. An experiment in removing coarse fish from a lake. Trans. Amer. Fish. Soc. 70:382-390. Ricklefs, R.E. 1983. The economy of nature: a textlDOok in basic ecology. 2nd ed. Chiron Press, N.Y. 510 pp. Roff , D.A. 1984. The evolution of life history parameters in teleosts. Can. J. Fish. Aquat. Sci. 41:989-1000. Rose, E.T. 1957. Results of fish management at Blaclchawk Lake. Iowa Conserv. Comm. Quart. Biol. Rep. 9:6-11. Cited in Miller (1960). Rose, E.T., and T. Moen. 1952. The increase in game-fish populations in East Okoboji Lake, Iowa, following intensive removal of rough fish. Trans. Amer. Fish. Soc. 82:104-114. Rosenthal, H., and D.F. Alderdice. 1976. Sublethal effects of environmental stressors, natural and pollutional, on marine fish eggs and larvae. J. Fish. Res. Board Can. 33:2047-2065. Rothschild, B.J. 1965. Aspects of the population dynamics of the alewife, AJOSB pseudohanengus (Wilson), in Cayuga Lake, New York. Amer. Midi. Nat. 74:479-496. Rutledge, W.P., and J.C. Barron. 1972. The effects of the removal of stunted white crappie, Pomoxis annularis Rafinesque, on the remaining crappie population of Meridan State Park Lake, Basque County, Texas. Texas Parks Wildl. Dept. Tech. Bull. No. 12. 41 pp. Ryan, P.M., and H.H. Harvey. 1980. Growth responses of yellow perch, Penca fJavescens (Mitchill), to lake acidification in the La Cloche mountain lakes of Ontario. Environ. Biol. Fishes 5:97-108. Rybicki, R.W., and K.H. Doan. 1966. Changes in the population of Lake Winnipeg whitefish. Fish. Branch, Man. Dept. Mines Nat. Resour., Winnipeg. 16 pp. Saville, A., and D. Schnack. 1981. Overview - some thoughts on the current status of studies of fish egg and larval distribution and abundance. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178:153-157. Ill Schmitz, W.R,, and R^E, Hetfeld, 1965. Predation by introduced muskellunge on perch and bass, ll:yr 8-9. Wisconsin Academy of Sciences , Arts and Letters 54:273-282. Schneberger, E. 1935. Growth of the yellow perch (Penca fJavescens Mitchill) in Nebish, Silver and Weber Lakes, Vilas County, Wisconsin. Trans. Wis. Acad. Sci., Arts StLetters 29:103-130. Schneider, J.C. 1973. Density dependent growth and mortality of yellow perch in ponds. Mich. Dept. Nat. Res., Fish. Res. Rep. 1795. 18 pp. Schneider, J.C, and J.H. Leach. 1977. Walleye {Sti ZQStedi on vitneum vitneum) fluctuations in the Great Lakes and possible causes, 1800-1975. J. Fish. Res. Board Can. 34:1878-1889. Schoffman, R.J. 1941. Age and growth of the drum in Reelfoot Lake. J. Tenn. Acad. Sci. 16:100-110. Scott, W.B., and E.J. Grossman. 1973. Freshwater fishes of Canada. Bull. 184. Fish. Res. Board Can. Ottawa, Ontario. 966 pp. Selgeby, J.H., W.R. MacCallum, and D.V. Swedberg. 1978. Predation by rainbow smelt (Osmerus mopdax) on lake herring (Conegonus antedii) in western Lake Superior. J. Fish. Res. Board Can. 35:1457-1463. Shuter, B.J., and J.F. Koonce. 1977. A dynamic model of the western Lake Erie walleye (Sti zostedfon vitneum vitneum) population, j. Fish. Res. Board Can. 34:1972-1982. Shuter, B.J., J.F. Koonce, and H.A. Regier. 1979. Modeling the western Lake Erie walleye population: A feasibility study. Great Lakes Fish. Comm. Tech. Rep. 32. 33 pp. Siler, J.R., W.J. Foris, and M.C. Mclnerny. in press. Spatial heterogeneity in fish parameters within a reservoir, in*. G. E. Hall, ed. Reservoir fisheries management - strategies for the 80' s, a symposium on managing reservoir fishery resources. Lexington, KY. Smith, P.E. 1978. Biological effects of ocean variability: time and space scales of biological response. Rapp. P.-v Reun. Cons. int. Explor. Mer. 173:117-127. Smith, S.H. 1968. Species succession and fishery exploitation in the Great Lakes. J. Fish. Res. Board Can. 25:667-693. Smith, S.H. 1970. Species interactions of the alewife in the Great Lakes. Trans. Amer. Fish. Soc. 99:754-765. Smyly, W.J. P. 1952. Observations on the food of the fry of perch {Penca fluviatiJis) in Windermere. Proc. Zool. Soc. Lond. 122:407-416. 112 Solomon, M.E. 1949. The natural control of animal populations. J. Anim. Ecol. 18:1-35. Spangler, G.R. 1970. Factors of mortality in an exploited population of whitefish, Conegonus cJupeaformfs, in northern Lake Huron, pp. 515-529. in: C. C. Lindsey and C. S. Wood. Biology of coregonid fishes. University of Manitoba Press, Winnipeg, Manit. Spangler, G.R., and J.J. Collins. 1980. Response of lake whitefish {Conegonus cJupeafonmis) to the control of sea Icimprey (Petnomyzon manfnus) in Lake Huron. Can. J. Fish. Aquat. Sci. 37:2039-2046. Spangler, G.R., N.R. Payne, J.E. Thorpe, J.M. Byrne, H.A. Regier, and W.J. Christie. 1977. Responses of percids to exploitation. J. Fish Res. Board Can. 34:1983-1988. Stevens, R.E. 1964. Investigation of fish populations in Lakes Moultrie, Marion, Wateree, and Catawba, South Carolina. So. Carolina Wildl. Res. Dept., D-J Federal Air Proj. F-l-R-11, Combined Quart. - Ann. Prog. Rep. and Creel Census Completion Rep. Cited in Burns (1966). Stewart, D.J., J.F. Kitchell, and L.B. Crowder. 1981, Forage fishes and their salmonid predators in Lake Michigan. Trans. Amer. Fish. Soc. 110:751- 763. Svardson, G. 1976. Interspecific population dominance in fish communities of Scandinavian lakes. Rep. Inst. Freshwat. Res. Drottningholm 55:144-171. Swedberg, D.V. 1965. Age and rate of growth of freshwater drum, Lewis and Clark Lake, Missouri River. Proc. S. Dakota Acad. Sci. 44:160-168. Swedberg, D.V., and C.H. Walburg. 1970. Spawning and early life history of the freshwater drum in Lewis and Clark Lake, Missouri River. Trans. Amer. Fish. Soc. 99:560-570. Swenson, W.A., and L.L. Smith, Jr. 1976. Influence of food competition, predation and cannibalism on walleye (Sti zostedfon vitneum vftneum) and sauger (5. canadense) populations in Lake of the Woods, Minnesota. J. Fish. Res. Board Can. 33:1946-1954. Swingle, H.S. 1949. Experiments with combinations of largemouth black bass, bluegills, and minnows in ponds. Trans. Amer. Fish. Soc. 76:46-62. Tarby, M. 1974. Characteristics of yellow perch cannibalism in Oneida Lake and the relation to first year survival. Trans. Amer. Fish. Soc. 103: 462- 471. Theilacker, G.H. 1978. Effect of starvation on the histological and morphological characteristics of jack mackerel, Tnacfiurus syirifnetni cus, larvae. Fish, Bull. 76:403-414. 113 Thorpe, J. 1977. Synopsis of biological data on the perch Penca fJuvfatiJis Linnaeus, 1758 and Penca fJavescens Mitchill, 1814. Food and Agric. Organ. United Nations, Fishery Synopsis 113. Torblaa, R.L., and R.W. Westman. 1980. Ecological impacts of lampricide treatments on sea lamprey {PetPOWyzon maninus) ammocoetes and metamorphosed individuals. Can. J. Fish. Aquat. Sci. 37:1835-1850. Turner, W.R. 1953. The age and growth of the gizzard shad, Donosoma cepedianum (LeSueur), in Herrington Lake, Kentucky. Kent. Fish. Bull. 13. 14 p. Turnpenny, A.W.H., and R. Williams. 1981. Factors affecting the recovery of fish populations in an industrial river. Environ. Pollution Ser. A: 26:39-58. Van Oosten, J. 1937a. The age and growth of the Lake Erie sheepshead, Aplodinotus gnunniens Rafinesque. Pap. Mich. Acad. Sci., Arts and Lett. 23:651-668. Van Oosten, J. 1937b. The dispersal of smelt, Osmenus mondax (Mitchill) in the Great Lakes region. Trans. Amer. Fish. Soc. 66:160-171. Van Oosten, J. 1947. Mortality of smelt Osmenus mondax (Mitchill), in Lakes Huron and Michigan during the fall and winter of 1942-1943. Trans. Amer. Fish. Soc. 74:310-337. Walburg, C.H. 1971. Loss of young fish in reservoir discharge and year-class survival, Lewis and Clark Lake, Missouri "River, p. 441-448 //?: G.E. Hall, ed. Reservoir fisheries and limnology. G.E. Hall, ed. Amer. Fish. Soc. Spec. Publ. 8. Walters, C.J., G. Steer, and G. Spangler. 1980. Responses of lake trout (SalveJinus namaycush) to harvesting, stocking, and lamprey reduction. Can. J. Fish. Aquat. Sci. 37:2133-2145. Wells, L. 1968. Seasonal depth distribution of fish in southeastern Lake Michigan. U. S. Fish, and Wildl. Serv. Fish. Bull. 67:1-15. Wells, L. 1970. Effects of alewife predation on zooplankton populations in Lake Michigan. Limnol. and Oceanogr. 15:556-565. Wells, L. 1977. Changes in yellow perch (Penca fJavescens) populations of Lake Michigan, 1954-75. J. Fish. Res. Board Can. 34:1821-1829. Wells, L. 1985. Changes in Lake Michigan's prey populations with increasing salmonid abundance, 1962 to 1984. Paper presented to Great Lakes Fish. Comm., Lake Michigan Committee Meeting, Ann Arbor, MI, 20 March, 1985. 114 WellSr L., and A.L. McLain. 1912. Lake Michigan: effects of exploitation, introductions, and eutrophication on the salmonid community. J. Fish. Res. Board Can. 29:889-898. Whillans, T.H. 1979. Historic transformations of fish communities in three Great Lakes Bays. J. Great Lakes Res. 5:195-215. Wickett, W.P. 1954. The oxygen supply to salmon eggs in spawning beds. J. Fish. Res. Board Can. 11:933-953. Wickliff, E.L. 1953. Gizzard shad rides again. Ohio Conserv. Bull. 17:23- 25. Willis, D.W. 1984. Reproductive biology of gizzard shad in Melvern Reservoir, Kansas. Presented at 46th Midwest Fish Se Wildlife Conf., Indianapolis, Ind. Wirth, T.L. 1958. Lake Winnebago freshwater drum. Wisconsin Cons. Bull. 23:30-32. Wolfert, D.R. 1969. Maturity and fecundity of walleyes from the eastern and western basins of Lake Erie. J. Fish. Res. Board Can. 26:1877-1888. Wrenn, W.B. 1968. Life history aspects of smallmouth buffalo and freshwater drum in Wheeler Reservoir, Alabama. Ann. Conf. £>. E. Assoc. Game Fish. Comm. 22:479-495. Zeeman, E.C. 1976. Catastrophe theory. Sci. Amer. 234:65-83. 115 APPENDIX 1. ANNOTATED BIBLIOGRAPHY 116 PAPERS INCLUDED IN ANNOTATED BIBLIOGRAPHY Authors, date Review Number Alcorn, D. J., L. E. Fancher, and J. G. Moss 1980 24 Anderson, E. D., and L. L. Smith, Jr. 1971 31 Bagenal, T. B, 1977 40 Bardcich, J. E. 1951 51 Beamish, R. J., and H. H. Harvey 1972 26 Beckman, W. C. 1941 1 Beckman, W. C. 1943 2 Beckman, W. C. 1950 3 Boaze, J. L,, and R. T. Lackey 1974 36 Bohnsack, J. A. 1983 38 Burrough, R. J., P. R. Bregazzi, and C. R. Kennedy 1979 14 Burrough, R. J., and C. R. Kennedy 1979 13 Bury, R. B. 1972 25 Butler, R. L. 1965 47 Cada, G. F., and G. L. Hergenrader 1980 50 Chevalier, J. R. 1977 39 Clady, M. D. 1976 61 Craig, J.F., C, Kipling, E.D. LeCren, and J.C. McCormsick 1979-60 Crunkllton, R. L., J. M. Czarnezki, and L. Trial 1980 34 Gushing, D. H. 1983 33 Deelder, C. L. 1951 59 Elwood, J. W., and T. F. Waters 1969 12 Eschmeyer, P. H. 1957 27 Eshenroder, R. L, 1977 62 Fraser, J. M. 1978 53 Grice, F. 1959 54 Gunning, G. E., and T. M. Berra 1968 16 Hanson, D. L., and T. F. Waters 1974 15 Healey, M. C. 1975 10 Healey, M. C. 1980 9 Higley, D. L., and C. E. Bond 1973 32 Holden, A. V. 1973 37 Holling, C. S. 1973 11 Jensen, A. L. 1981 43 Johnson, F. H. 1977 19 Johnston, C. E., and J. C. Cheverie 1980 17 Jones, D« D., and C. J. Walters 1976 44 Keleher, J. J. 1972 23 Kelso, J. R. M., and T. B. Bagenal 1977 55 Kenyon, R. B. 1978 42 Kneib, R. T. 1981 4 Kohler, C. C, and J. J. Ney 1981 22 KushLan, J, A, 1974 30 Mahon, R., and M. Ferguson 1981 5 McFadden, J. T. 1977 46 Moring, J. R. 1981 25 Mortensen, E. 1977 28 117 Nakashimar B. S,, and W. C. Leggett 1975 56 Olmsted, L. L., and D. G. Cloutman 1974 18 Priegel, G. R. 1971 49 Raleigh, R.F., D.H. Bennett, L.O, Mohn, and O.E. Maughan 1978 41 Ricker, W. £•, and J. Gottschalk 1940 6 Rose, E. T., and T, Moen 1952 29 Ross, S. T., and J, A, Baker 1983 7 Ryan, P. M., and H. H, Harvey 1980 57 Spangler, G. R., and J. J. Collins 1980 8 Swedburg, D. V., and C. H. Walburg 1970 48 Tarby, M. J, 1974 58 Torblaa, R. L,, and R. W. Westman 1980 20 Turnpenny, A. W. H., and R. Williams 1981 21 Walters, C.J., G. Steer, and G. Spangler 1980 45 Wells, L. 1977 52 118 Beckman, W.C. 1941. Increased growth rate of rock bass, Ambloplites nupestrfs (Rafinesque), following reduction in the density of the population. Trans. Amer. Fish. Soc. 70:143-148. Keywords: GROWTH, ROCK BASS, MIDWEST, EXPERIMENT, ROTENONE Abstract : The fish of the south basin of Standard Lake were poisoned with rotenone. Samples of the rock bass population were taken at the time of the poisoning, in September, 1937. During 1939 and 1940 further collections were made in order to determine the effect of the population reduction. An increase! growth rate, too great to be accounted for by any normal growth fluctuation, had occurred in fish of all ages. Data and Review : Beckman poisoned 1/2 of the lake, which had stunted fish populations. He allowed the fish to freely move between basins and sampled rock bass 2 and 3 years after poisoning. There was a significant increase in growth rate observed after removal of 20,192 fish (3,155 fish/hectare; 24.5 kilograms/hectare) by poisoning (rotenone). Growth was measured as change in length. This was a clear demonstration of compensation by an increase in growth rate. No data were given on fecundity, post-poisoning population size, growth, etc. Beckman, W.C. 1943. Further studies on the increased growth rate of the rock bass AmbJopIftes nupestrfs (Raf inesque) , following the reduction in density of the population. Trans. Amer. Fish. Soc. 72:72-78. Keywords: GROWTH, ROCK BASS, MIDWEST, EXPERIMENT, ROTENONE, CONDITION FACTOR, AGE-SPECIFIC GROWTH Abstract : An increased growth rate, too great to be accounted for by any normal growth fluctuation, was shown to have occurred in all age-groups of rock bass, following the reduction in the population by poisoning the fish of the south basin of Booth (formerly called Standard) Lake, Michigan. Additional information was obtained in 1941 and 1942 which made it possible to analyze the change in growth in weight as well as length. Not only did the length and weight increase, but the mean coefficient of condition, K, also increased. Data and Review : Continuation of Beckman (1941); two more years of data. There was little difference in the yearly growth increment (length) pre- and post -poisoning for age 1 fish; all other ages showed a significantly greater post -poisoning growth (increments nearly two times greater). The growth rate was approximately 3-5 g/yr before poisoning, with no appreciable variation with age. The growth rate after poisoning was approximately 11-64 g/yr and increased, with age. The coefficient of condition was 3.07 pre-poisoning and 119 4.10 post-poisoning. Virtually the same paper as Beckman (1941), with a more extensive data set. Collections one whole year were done by fly fishing, which may bias size results. Beckman, W.C. 1950. Changes in growth rates of fishes following reduction in population densities by winterkill. Trans. Amer. Fish. Soc. 78:82-90. Keywords: GROWTH, WINTERKILL, BLUEGILL, LARGEMOUTH BASS, YELLOW PERCH, PUMPKINSEED, MIDWEST Abstract: Studies were made on four lakes in southern Michigan to determine the effect of the reduction in numbers of fish by winterkill upon the rate of growth of fishes. Earlier investigations on another lake had demonstrated that a severe reduction in population density had a marked effect on the rate of growth of the survivors. At the time of the winterkill in 1945 scale samples and counts of dead fish were obtained. Samples have been collected each year since the kill. An increase in the rate of growth too great to be accounted for by annual variations followed the winterkill in all but one species from one lake. The increase in growth rate was not maintained, however, for longer than a year in most species. Data and Review : Beckman compared the growth (length) of four game fish species before and after a major winterkill in four lakes. The bluegill population in one lake was considered stunted prior to winterkill. All other populations were considered to be growing at about state average prior to winterkill. For five of the six populations studied, growth increased strongly the first year after winterkill, but then generally declined the second year after winterkill. In one lake, bluegill growth increased 35.6% in the first year after winterkill, over an average of the seven years before the winterkill. In another lake, bluegill growth increased 19.5% the first year and 29.0% the second year. Largemouth bass showed no increase in growth. Yellow perch growth in one lake increased 61.3% the first year, and decreased the second year. In another lake yellow perch increased 35.0% the first year and 32.7% the second year. Pumpkinseed growth was up 79.2% the first year, 15.5% the second year, and 39.1% the third year. (All percentages are compared to a seven year average before the kill.) In the last two cases (perch and pumpkinseed), bluegills were entirely killed off from that lake, which may account for the high growth rate of pumpkinseeds . Chubsuckers and yellow and brown bullheads were more abundant after winterkill than before. They may have been out-competing centrarchids and perch, and thus account for decreased growth rate of centrarchids and perch after a year or two. Rough fish showed less mortality from winterkill. Fish were collected by netting (no specifics given), hook and line, and subsequent minor winterkills. Methods were vague. 120 Kneib, R.T. 1981. Size-specific effects of density on the growth, fecundity and mortality of the fish FunduJus hetePOClitus in an intertidal salt marsh. Mar. Ecol. Prog. Ser. 6:203-212. Keywords: GROWTH, FECUNDITY, MORTALITY, MUMMICHOG, SIZE-SPECIFIC EFFECTS, DENSITY-DEPENDENT, EXPERIMENT, EAST Abstract : Three size classes of Fundi! J us hetenocJitus (L.) (small, 40-50 mm total length; medium, 50-69 mm TL; large, >70 mm TL) were maintained on an intertidal salt marsh in separate enclosures (10 m^) for 10 months at 3 densities (1, 2 and 4 ind./m^). During the final 6 weeks of the experiment (March-April 1979) the effects of density on individual growth as well as mortality and fecundity of the enclosed populations were assessed. Although fish density and growth were inversely related in all size classes, the growth rate of small F. hetenocJitus remained positive even after 6 weeks at the highest experimental density. Medium size fish grew when enclosed at densities of 1 and 2 ind./m^ but not at 4 ind./m^. The growth of large fish was positive only in cages containing 1 ind./m*. Over the range of experimental treatments, fecundity of medium size F. hetenoclitus exhibited the strongest negative response to high density. However, the gonadosomatic index for large fish was lower than for either of the other size classes at all experimental densities, suggesting that fecundity of large individuals may be severely impaired even below the lowest density (1 ind./m^) used in the experiment. Density and fish losses (mortality + escape) were directly related only in the small and large fish size classes.. After 6 weeks at a density of 4/m2, 77% of the small fish and 60% of the large fish were lost compared to only 32% of the medium size individuals. Escape accounted for an estimated 13.2% of all losses and was independent of density. Total losses reflected a density-dependent trend in small and large fish mortality. Data and Review : Kneib studied size-specific effects of density on individual growth, fecundity, and mortality of mummichog {FunduJus hetenocJitus) in field enclosures in an intertidal marsh. Major findings: 1) Growth (change in length) was inversely related to density. This was probably due to food limitation at high densities. 2) There was an inverse relationship between density and proportion of medium- sized females with ripe oocytes. 3) Significant density-dependent mortality was observed for small and large fish. The same trend was observed for medium-sized, fish but was not significant . This paper does not deal with catastrophic effects, but gives a good overview of density-dependent growth, fecundity, and mortality for mummichog. The methods are well documented. 121 Mahon, R., and M. Ferguson. 1981. Invasion of a new reservoir by fishes: species composition, growth, and condition. Canadian Field-Naturalist 95(3):272-275. Keywords: BLUNTNOSE MINNOW, COMMON SHINER, CREEK CHUB, WHITE SUCKER, GROWTH, CONDITION FACTOR, IMPOUNDMENT, CANADA Abstract : After the first summer of inundation, the ichthyofauna of the reservoir known as Guelph Lake (impoundment of Speed River) was compared to that of the Speed River before impoundment. Several species that had been rare in the river showed marked increases in relative abundance in the reservoir. The bluntnose minnow (PfffiephaJes notatus) was most abundant among 19 species collected. For young-of-the-year of three species, bluntnose minnow, common shiner (NotPOpis connutus), and creek chub (SemotiJus atnomacuJatus) , which were abundant enough for comparison, growth was faster in the reservoir than in the river. For these species and white sucker (Catostomus comensoni), the average condition (weight at a given length) was significantly higher in the reservoir. Data and Review : The authors suggested that increase in growth may be due to increased food abundance but have no corroborating data. Mean SL of young-of-the-year (in mm, presumably) in September: Species Impoundment River common shiner 36.4 29.4 creek chub 54.7 40.9 bluntnose minnow 33.3 27.6 Ratio of observed weight (from impoundment) to expected weight (from river): common shiner, 1.32; white sucker, 1.28; bluntnose minnow, 1.13; creek chub, 1.13. Sampled by 6-mm circular mesh seine, 18 stations. Ricker, W.E., and J. Gottschalk. 1940. An experiment in removing coarse fish from a lake. Trans. Amer. Fish. Soc. 70:382-390. Keywords: WALLEYE, WHITE BASS, SMALLMOUTH BASS, LARGEMOUTH BASS, BLUEGILL, PUMPKINSEED, YELLOW PERCH, BLACK CRAPPIE, EXPERIMENT, MIDWEST Abstract and Review : Ricker extensively seined a 1,600-acre lake over a 2-year period (1935-1936) to remove rough fish (carp, quillback, buffalo, gar). Within one year after seining was initiated, there was a large increase in abundance of game fish accompanied by a decrease in turbidity and an increase 122 in nuacrophytes. Also observed was an increase in longnose and shortnose gar, species that were removed by seining; the authors attributed this, in part, to greater similarity in feeding habits between gar and game fish than with rough fish,. This paper does not permit quantification of compensation in response to catastrophic reduction in densities. Interspecific competition may be the mechanism accounting for the increase in abundance. Ross, S.T., and J. A. Baker. 1983. The response of fishes to periodic spring floocls in a southeastern stream. Amer. Midi. Nat. 109(1) :1-14. Keywords: SOUTH, FLOODING Abstract ; Movement of fishes onto a fringing floodplain was studied by seining and trapping during five spring floods. We collected 26 species from the inundated floodplain; the known channel fauna is 42 species. Species numerically dominant on the floodplain were FunduJus olivaceus, F . notti , Gambusia afffnis, Notpopis weJaka, HI. texanus, N. noseipfnnfs, Lepomfs macpochipus, Z. cyanelJus and L. marginatus. a flood-exploitative species. Not no pis texanuSf had higher population abundance during 3 high-flow years than in 3 low-flow years, suggesting that spring f loocling may exert significant control over fish community structure. Data and Review ; It was shown that abundances of some fish species (termed "flood exploiters") are positively correlated with spring flow. Increases in abundance apparently were not due to compensation; rather^ flood exploiters apparently used abundant floodplain resources prior to spawning to increase energy available for reproduction. Spangler, G.R., and J.J. Collins. 1980. Response of lake whitefish (Conegonus cJupeafonmis) to the control of sea lamprey (Petnomyzon marinus) in Lake Huron. Can. J. Fish. Aquat. Sci. 37:2039-2046. Keywords; LAKE WHITEFISH, SEA LAMPREY, GREAT LAKES, CANADA, SURVIVAL Abstract: Changes in population parameters of three stocks of whitefish in Lake Huron following control of sea lamprey have varied between the major basins of the lake. In the North Channel and main basin, whitefish survival rates increased significantly, from values of 23 to 37% and 16 to 24%, respectively. In both areas the abundance of lamprey declined while whitefish increased. In Georgian Bay, whitefish populations did not reflect changes attributable to lamprey control, possibly because of low lamprey abundance. During the control period, whitefish survival rates declined from 57 to 39%, abundance declined and growth rates increased. 123 Data and Review ; Population parameters of three whitefish stocks in Lake Huron were examined before and after lamprey control. In two basins where lamprey were abundant, whitefish increased after lamprey abundance declined. Whitefish survival rates increased greatly during this period. Total annual survival equaled 0.228 and 0.162 before lamprey control in two areas; 0.366 and 0.240 after, respectively. The paper did not show growth compensation. The results are consistent with basic predator-prey theory and not strongly applicable to compensation in response to catastrophic reduction in population size. The data were from commercial gill net catches. Healey, M.C. 1980. Growth and recruitment in experimentally exploited lake whitefish (Conegonus cJupeaform's) populations. Can. J. Fish. Aquat. Sci. 37:255-267. Keywords: LAKE WHITEFISH, GROWTH, AGE-SPECIFIC SURVIVAL, CANADA, EXPERIMENT Abstract: From 1971 to 1978 Healey conducted an exploitation experiment on lake whitefish in four Northwest Territories lakes. Significant increases in size-at-age were recorded in the exploited lakes, and the degree and persistence of the increase were proportional to the intensity of exploitation. Good recruitment was stimulated on the heavily exploited lake, and to a lesser degree in the moderately exploited lake. Analyses of growth and year-class strength in the unexploited control lake and the lightly exploited lake indicated that natural variation in growth and recruitment, on a similar scale to that stimulated by exploitation, had occurred in the past. The compensatory responses of lake whitefish populations to exploitation appear to be a logical consequence of the natural population regulating mechanisms for the species. Results indicate that in unexploited lakes, recruitment and growth of young fish is regulated by the established population of mature fish. Data and Review : This article also applies to non-catastrophic compensation. There were three degrees of exploitation in three lakes. The highest exploitation was approximately 50% of the population. The timing of samples was not explained much. Harvest samples were taken "summer and autumn"; Figure 2 (catch curve) shows the start of harvesting in 1973 and the end in 1974. The article does not say how many sampling times were in between, or what N is for each curve. Healey had a good control lake; he paid attention to control data. He held survey sampling of the control lake to 100 fish per year. Drygeese Lake was exploited at a maximum rate. At 30% exploitation per year for two years (total population reduced about 50%), seven of thirteen age classes showed significant increases in length. The average increase (all ages) was 12 mm (0-3 years after exploitation). After four years, four ages still were significantly longer than pre-exploitation. At Drygeese Lake, catches began to recover in the first year after exploitation. Mature fish were almost eliminated; ages 0+ and 1+ were very abundant during exploitation, and the abundance peak was still evident four years later. The discussion was 124 good; predation, catchability of young fish, and the influence of adults on young via behavior were discussed. Increased growth rate appeared to be the most important compensatory mechanism for lake whitefish. 10 Healey, M.C. 1975. Dynamics of exploited whitefish populations and their management with special reference to the Northwest Territories. J. Fish. Res. Board Can. 32:427-448. Keywords: LAKE WHITEFISH, GROWTH, CANADA, GREAT LAKES, AGE AT MATURITY Abstract : Available data on mortality, growth, reproduction, and stock size in exploited and unexploited populations of lake whitefish (Conegonus cJupeafonmf s) are reviewed with a view to understanding the dynamics of exploited populations and improving their management. Natural mortality ranged from about 0.20 to 0.80 in unexploited populations. In exploited populations total mortality was as high as 0.94. Unexploited populations showcid a wide range of growth rates. Growth rate increased with increasing exploitation, and growth rate in all heavily exploiteci populations was similar to the most rapid growth rate shown by unexploited stocks. Heavily exploited whit€»fish matured at a younger age and possibly also at a smaller size than those which were unexploited. Limited data on stock size suggest that although total population size declines under heavy exploitation, the vulnerable population remains of similar size. It is concluded that whitefish respond to fluctuations in population size through compensatory changes in growth rate. The age difference between growth rate in a population and maximum growth rate is a measure of its scope for compensating for increased mortality. Populations with slow growth rate and low mortality should, therefore, have the best fishery potential, while those with high growth rate and high mortality have a low fishery potential. Further, it is possible to judge the fishery potential of a population or its •stage of exploitation from relatively simple measurements of mortality, growth, age structure, and maturity. Data and Review : This is a literature review. Most of the examples are for unexploited or constantly exploited populations, and are not catastrophic. Mortality was so variable for lake whitefish that evidence of compensatory mortality was not found. The fishery collapse example in Miller (1947) was discussed here and will also be reviewed. There was much evidence for compensatory growth. Figure 4, page 437 shows growth rates of many populations exploited to different degrees. Limited data suggested that age at maturity is younger and size at maturity smaller for exploited populations. Healey believed fecundity is probably not compensatory in lake whitefish. 125 11 Holling, C.S. 1973. Resilience and stability of ecological systems. Ann. Rev. Ecol. Systematics 4:1-23. Keywords: REVIEW, MODELING, STABILITY, RESILIENCE, LAKE TROUT Abstract : Rolling discusses traditional models and the factors that make them different from real ecosystems. Examples of different system behaviors are shown. More than one discrete domain of attraction (local stability, spiraling in from certain starting points) may exist for a given system. Freshwater aquatic systems are approximately self-contained, homogeneous, with small climatic fluctuations. In such systems, fishing pressure, environmental change, or appearance of exotics may set conditions for population collapse by reducing resilience of the system, as with lake trout in Lake Michigan. Stability is the ability of a system to return to equilibrium after a temporary disturbance. Resilience is a measure of the persistence of the system and its ability to absorb change or disturbance and still maintain the same relationships between populations or state variables. An unstable system may be highly resilient. Data and Review : Few data are given, as this is a review paper. The paper gives good explanations of traditional models and their shortcomings. It discusses Great Lakes examples in terms of the theory involved, characterizing them as stable, non-resilient systems which can easily go "over the edge" to another domain of attraction (e.g. lake trout-bloater complex to alewife- smelt-perch or exotics). The paper cites Smith (1968) and Beeton (1969). Lake Michigan was affected by fishing pressure, eutrophication, physical changes, a foreign predator (sea lamprey), and foreign competitors (alewife and carp) . A precondition for lake trout collapse was set by the harvesting of fish, which reduced resilience of the system so when lamprey appeared (unexpected event), the population collapsed. If it had not been the lamprey, something else would have destroyed lake trout populations: climate change, environmental change, or species composition change. It should be possible to determine, from lake trout history, the boundary population below which they cannot compensate. 12 Elwood, J.W., and T.F. Waters. 1969. Effects of floods on food consumption and production rates of a stream brook trout population. Trans. Amer. Fish. Soc. 98:253-262. Keywords: BROOK TROUT, FLOODING, MIDWEST, GROWTH Abstract: Food consumption and production rates were estimated for a stream population of brook trout, SaJveJinus fontinaJis Mitchill, over a two-year period (1965-1966) in which four severe floods occurred. Two year classes were nearly eliminated as producing components of the population. Standing 126 crops of older age-groups were reduced as a result of a decrease in the stre460 mm was slightly higher 3-6 years after sucker removal (K=0.91) than before sucker removal (0.87), but declined 7 years after removal. Walleye <460 mm showed no increase in condition factor after sucker removal. A strong walleye year class was produced 4 years after sucker removal. This paper demonstrates compensation but does not give the detailed data necessary to quantify all mechanisms mentioned. Johnson (1975), cited in this paper, may give more data. Depth or substrate of stations may affect catch for these species, but sampling stations are not described, so they cannot be evaluatecl. Otherwise the methods seem appropriate. The conclusions may not have gen€5ral applicability to more complex systems; this lake has few species. 131 20 Torblaa, R.L., and R.W. Westman, 1980* Ecological impacts of lampricide treatments on sea lamprey (Petnomyzon maninus) ammocoetes and metamorphosed individuals. Can. J. Fish. Aquat. Sci. 37:1835-1850. Keywords: SEA LAMPREY, GREAT LAKES, POISONING, GROWTH, SEX RATIO Abstract : Abundance of ammocoetes and transformed individuals has declined in most watersheds with a few exceptions where density-dependent factors may have been influential. Increased growth was related to reduced density. Sex compositions of larval and metamorphosing populations were highly variable during initial lampricide treatments. Females predominated in some streams, males in others. Streams once dominated by males now favor females in residual and reestablished populations of larvae. The shift to femaleness in the larval populations has precipitated a similar shift in adult sea lamprey populations of the upper Great Lakes. Data and Review : In Great Lakes tributaries, sea lamprey ammocoetes of year- classes established after treatment with TFM or Baker 73 grew faster than succeeding year-classes; as density increased, growth rate declined. No original growth data are given for ammocoetes. Greater growth of transformed sea lampreys following TFM treatment was also observed; mean length increased from 143 mm during the first treatment to 157 mm during the second treatment. Most growth information in this paper is apparently from Purvis (1979), cited in this paper. For Lakes Superior and Huron tributaries, where males predominated before treatment, percentage males decreased significantly after 2 treatment (x test, P<0.001), showing sex ratio compensation (more females when densities lower). In tributaries where females dominated before treatment, sex ratios usually remained about the same. Lake Ontario tributaries did not show as clear results. Streams that were predominately male before treatment all had high densities of larvae, also indicating density-dependence of sex ratio. Percent males decreased from 72 to 15% in the Michipicoten River during 5 treatments, one of the more striking examples. Many sex ratio data are given in tables and figures for adults and ammocoetes. 21 Turnpenny, A.W.H., and R. Williams. 1981. Factors affecting the recovery of fish populations in an industrial river. Environmental Pollution (Series A) 26:39-58. Keywords: BROWN TROUT, BRITAIN, GROWTH, CONDITION FACTOR, TOXICS, POLLUTION Abstract : The River Ebbw Fawr, an industrial river of southeast Wales, was investigated over a three-year period to follow the re-establishment of fish populations as a result of pollution control measures at coal washeries and a steelworks on the river. Five freshwater fish species became established in parts of the river during the study period (1974-77). The brown trout SaJmo 132 tnutta I, was the first to enter. Good growth and condition factors among the recolonizing brown trout stock suggest that a sport fishery could be developed on the river. Data and Review ; Little useful information is given on compensation. This paper follows reestablishment of fish populations in a river after effective pollution control was established. It shows that recolonizing fish have high growth rates and high condition factor in comparison with fish in similar, non-polluted streams. No natural reproduction was taking place in the river (excessive silt in gravel). No data were given on fish populations prior to and/or during pollution. 22 Kohlcsr, C.C., and J.J. Ney. 1981. Consequences of an alewife die-off to fish and zooplankton in a reservoir. Trans. Amer. Fish. Soc. 110:360-369. Keywords: ALEWIFE, SOUTHEAST, INTERSPECIFIC AND INTRASPECIFIC COMPETITION, WINTERKILL, GROWTH Abstract : Alewife AlosB pseudohanengus was the principal pelagic forage species in Claytor Lake, Virginia, prior to a severe die-off in the winter of 1977-1978. Due to successful reproduction in 1978, abundance of alewives rebounded sharply in 1979. Alewife-induced shifts in zooplankton composition toward smaller forms represent a potentially significant adverse impact on cohabiting planktivores, including young-of-the-year sport fishes. Growth of adult alewives and young-of-the-year walleyes and white bass was significantly greater in 1978 than in the 2 years preceding the die-off, suggesting that both intra- and interspecific trophic competition operate when alewife density is high. Annual growth of age-2 walleyes and white ba.ss did not differ substantially before and after the die-off; impact of the population collapse may have been mitigated by high reproductive success and rapid first-year growth of alewives in 1978. These findings indicate that alewife size- selective planktivory may have a more enduring and potentially severe impact on the reservoir fishery than does instability in the alewife population. Data and Review : Kohler and Ney examined age-specific growth rates of alewives in a reservoir before and after a die-off. The growth rate (annual increment in total length) was significantly greater after the die-off for age 2 and 4 alewives and marginally greater (P=0.10) for age 3 fish; growth rates for age 1 fish were nearly identical before and after the die-off. Age structure of the population before and immediately after the die-off was not given. 23 Keleher, J.J. 1972. Great Slave Lake: effects of exploitation on the salmonid community. J. Fish. Res. Bd. Canada 29:741-753. Keywords: LAKE TROUT, LAKE WHITEFISH, CISCO, CANADA 133 Abstract : A 20-year investigation commenced at the inception of the Great Slave Lake commercial fishery, and it is inferred that exploitation has been the sole cause of change in the fish stocks. Lake whitefish, Conegonus cJupeafonmis, and l^ke trout, SalveJinus namaycusfi, have made up 94% of the commercial landings of 54 million kg during the period. The annual catch of these two species rose from 0.7 million kg in 1945 to a maximum of 4 million kg in 1949 then gradually deceased to about the 2 million kg level. The average size of commercially caught fish has declined to about 0.9 kg for whitefish and 2.2 kg for trout. Information on catch per unit effort for other species indicates no evident explosions in their numbers, though ciscoes, Leuci chthys spp., appear to have increased in abundance. Data and Review : Keleher reviewed the 20-year history of commercial fishing exploitation on fish stocks in Great Slave Lake. Virtually no information is given on compensation. 24 Alcorn, D.J., L.E. Fancher, and J.G. Moss. 1980. Harbor seal and fish populations - before and after a sewage spill in south San Francisco Bay. Calif. Fish and Game 66:238-240. Keywords: PACIFIC COAST, SPILL Abstract: A sewage spill occurred 4 through 29 September 1979, at the San Jose-Santa Clara Water Pollution Control Plant, during which 15 billion liters of partially treated sewage flowed into Artesian Slough. Obvious short-term effects of the sewage spill on nearby Mowry Slough seal and fish populations appear to be minimal or absent. Long-term effects are possible and monthly monitoring of both seal and fish populations continues. Data and Review : The sewage spill had no detectable impact on fish or harbor seal populations. No information is given on compensation. 25 Moring, J.R. 1981. Changes in populations of reticulate sculpins (Cottus penpJexus) after clear-cut logging as indicated by downstream migrants. Amer« Midi. Nat. 105:204-207. Keywords: RETICULATE SCULPIN, NORTHWEST, LOGGING, GROWTH Abstract : Downstream movements of reticulate sculpins Cottus penplexus were monitored for 7 years before and 7 years after clear-cut logging of a watershed of a tributary of the Alsea River, Oregon. Short-term decreases in populations, as indicated by reduced numbers of migrating fish, persisted for 5 years. During this period, the average length of the sculpins remained significantly higher then before, possibly indicating a reduction in competition. Timing of downstream movements did not change. 134 Data and Review ; Moring compared the number and size (total length) of reticulate sculpins moving downstream annually in two small streams before and after the entire watershed of one of the streams had been clearcut. The sculpin population was greatly reduced in the clearcut stream; the two youngest year classes were apparently eliminated immeiiately after clearcutting (Krohn 1968 cited in this paper) • Population size (as evidenced by number fish trapped moving downstream) remained low in the clearcut stream for 7 years post-logging, and the average total length of fish in the population increased post-logging in the clearcut stream, but not in the control stream. This size increase could be due to compensatory growth (as the author suggested) or simply due to a change in the size structure of the population. Since small fish were most strongly affected by the logging, the latter explanation seems most parsimonious. Age-specific growth rates would be needed to determine if this is a good example of compensation. 26 Beamish, R.J., and H.H. Harvey. 1972. Acidification of the La Cloche Mountain Lakes, Ontario, and resulting fish mortalities. J. Fish. Res. Bd. Canada 29:1131-1143. Keywords: LAKE TROUT, LAKE HERRING, WHITE SUCKER, YELLOW PERCH, CANADA, ACIDIFICATION, REPRODUCTIVE FAILURE Abstract : The loss of populations of lake trout (SaJvelfnus namaycusfi), lake herring (Conegonus antedii), white suckers (Catostomus comensonf), and other fishes in Lumsden Lake was attributed to increasing Icsvels of acidity within the lake. An absence of fishes was also observed in nearby lakes. In some lakes, acid levels have increased more than one hundredfold in the last decade. The increases in acidity appear to result from acid fallout in rain and snow. Data and Review : This paper documents virtual elimination of game and non-game fish populations in Lumsden Lake over a ten-year period (1961-1971) in which all of the lakes in the La Cloche Mountain Lakes becanie highly acidic (pH between 4 and 5 in most lakes). Yellow perch were eliminated. The study does not provide any information on compensation. Acidification is severe and ongoing, thus fish cannot recover or compensate. 27 Eschmeyer, P.H. 1957. The near extinction of lake trout in Lake Michigan. Trans. Amer. Fish. Soc. 85:102-119. Keywords: LAKE TROUT, GREAT LAKES, PREDATION, OVERFISHING Abstract: After the collapse of the commercial fishery for lake trout (SaJveJfnus namaycush) in Lake Michigan in the late 1940s, the further decline of the population was traced by records of numbers of small lake trout (mostly 135 28 to 41 cm in total length and 3 to 5 years old) caught in small-mesh nets of the chub (Conegonus spp.) fishery. By 1951 the estimated abundance of these lake trout in lower Lake Michigan was only about 4% of their abundance prior to the invasion of the sea lamprey {P^etPomyzon maninus). This remnant of the population declined severely in subsequent years to a point near extinction by 1955. In April-July 1955 only eight lake trout were caught in 1.7 million m of gill net. Lake trout were brought near extinction by lethal attacks of the sea lamprey and by the near or perhaps complete failure of natural reproduction in 1949 and subsequently. Data and Review ; This paper describes the decline of lake trout in Lake Michigan through 1955. The only information remotely related to compensation concerns the size distribution of lake trout taken in chub nets (small-mesh nets) between 1951 and 1955 in southeastern Lake Michigan; percent of legal- sized lake trout in chub nets increased from 5% (n=l,113) in 1951 to 75% (n=4) in 1955. The author argued that this is evidence for reproductive failure (and it most likely is), but compensatory growth could also have occurred. 28 Mortensen, E. 1977. Density-dependent mortality of trout fry (Sal/no tnutta L.) and its relationship to the management of small streams. J. Fish Biol. 11:613-617. Keywords: BROWN TROUT, EUROPE, STREAM MODIFICATION, DENSITY -DE PENDENT MORTALITY Abstract : The natural mortality rate of trout fry and the impact of weed cutting and cleaning out of weeds, branches and twigs from the stream-bed upon the mortality rate of trout fry was studied on several reaches in small streams. Above an initial density of 3.1 fry/m* natural mortality rate was found to be density-dependent in the first few months of life. In reaches where weed cutting and cleaning out of the stream-bed had taken place the mortality rate was higher than that found under natural conditions in undisturbed reaches. Data and Review : Populations of brown trout were followed for the first several months after emergence in several small Danish streams • Fry mortality in undisturbed stream reaches was compared to two reaches (in different streams) where weeds were cut and weeds, branches, twigs, and larger stones were removed from the stream bed. In undisturbed reaches, fry mortality was density-independent at low initial densities (<3 fry/m^) and density-dependent at initial densities >3 f ry/m^ . Daily instantaneous mortality rates were considerably greater in disturbed reaches (-0.018) than in undisturbed reaches (-0.012) at comparable initial fry densities. Since these trout populations appeared to be regulated by density-dependent mechanisms, follow-up studies in disturbed streams would be interesting in showing if and how stream trout populations compensate for increased fry mortality. This paper shows density- dependent mortality, but unrelated to the catastrophe. 136 29 Rose,r E.T., and T. Moen. 1952. The increase in game-fish populations in East Okoboji Lake, Iowa, following intensive removal of rough fish. Trans. Amer. Fish. Soc. 82:104-114. Keywords: YELLOW PERCH, WALLEYE, CENTRARCHIDS, BLACK BULLHEAD, MIDWEST, WINTERKILL, COMPETITION, RECRUITMENT Abstract and Review : An effort was made to remove rough fish (largely bigmouth buffalo Ictiobus cypni nellus, carp, and freshwater drum) from East Okoboji Lake, Iowa, from 1940 to 1951 using seines and traps. Game fish populations (catch per seine haul using a 762-m, 6.4-cm-mesh drag seine) were monitored throughout the 12-year rough fish removal period. Gaiiae fish (total of eight species) per seine haul increased from an average of 89 (27 kg) in 1940 through 1945 to an average of 1,100 (254 kg) in 1947 through 1950. [A winter kill in 1950-1951 greatly reduced catch per unit effort in the spring-summer of 1951 for both game and rough fish. But catch of YOY walleye and yellow perch increased noticeably after the winterkill: to 196 and 5,383 respectively compared to average of 7 and 97 for 1940-1945.] Not all game fish increased in response to the rough fish removals. CPUE of young-of-the-year and older northern pike, walleye, largemouth bass, and white bass did not change significantly over the 12-year period (except for walleye after the winterkill). Yellow perch, crappie, bluegill, and black bullheads showed significant increases in young-of-the-year and older fish in the years following 1945. The species that did not respond to the rough fish removal are all relatively large piscivores. These species probably do not compete with the rough fish for resources (especially food), which helps to explain why their densities did not change in response to rough fish removal. Release from competition is the most likely explanation for the increase in the other game fish species, but improvements in habitat quality (e.g., reestablishment of weed beds and the elimination of blue-green algal blooms) and, perhaps, concomitant changes in carrying capacity undoubtedly also played a role. 30 Kushlan, J. A. 1974. Effects of a natural fish kill on the water quality, plankton, and fish population of a pond in the Big Cypress Swamp, Florida. Trans. Amer. Fish. Soc. 103:235-243. Keywords: DEWATERING, SOUTHEAST, OXYGEN Abstract : A naturally occurring fish kill took place during the southern Florida dry season when receding water levels forced aquatic organisms out of shallow swamps and into a small pond. The fish kill had no effect on the long-term limnological characteristics of the pond, as water quality returned to normal range within 2 months and phytoplankton disappeared within 1 month after the fish kill. The species of fish in the pond showed a range of susceptibility to low oxygen concentration and other aspects of the fish kill. 137 Data and Review : This paper documents the occurrence of a massive natural fish kill in a Florida pond at the end of the dry season in 1970, Fish occurring in Big Cypress Swamp became concentrated in isolated ponds due to declining water levels during the dry season. The fish kill in the pond studied occurred when dissolved oxygen concentrations fell to nearly ppm (<1 ppm at 1 cm depth; ppm throughout the rest of the water column). This particular study has little value concerning compensation because it is only concerned with which species were eliminated during the kill, which species survived, and how some species were able to survive the low DO concentrations. A follow-up study, if available, could contain valuable information as this was certainly a good case of a catastrophic reduction in fish populations (99.4% of the total fish population in the pond died during the fish kill). 31 Anderson, E.D., and L.L. Smith, Jr. 1971. Factors affecting abundance of lake herring (Conegonus antedii Lesueur) in western Lake Superior. Trans. Amer. Fish. Soc. 100:691-707. Keywords: LAKE HERRING, GREAT LAKES, OVERFISHING, INTERSPECIFIC COMPETITION, GROOTH, CONDITION FACTOR, FECUNDITY Abstract: A field and laboratory study was conducted in 1965-1968 to investigate factors affecting lake herring (Conegonus antedii Lesueur) abundance in western Lake Superior. Commercial landings in November-December in the Duluth-Superior area dropped from 502,650 kg in 1954 to 900 kg in 1966. Wisconsin catches also declined markedly. Abundance and fishing intensity indices decreased with landings. Commercial catches and abundance of chubs (Conegonus spp.) and -smelt (Osmenus mondax) increased simultaneously with the herring reduction. Average age of adult herring in commercial catches increased from 3.9 years in 1950-1959 to 5.1 in 1966-1968 but showed no significant difference between areas. Percentage of females did not show a trend from 1960 to 1968; however, the 1968 percentage of 85.0 was the greatest since 1962. Percentage of females increased with age. Average length and weight increased from 287 mm and 148 g in 1950-1959 to 340 mm and 301 g in 1966-1968. Average number of eggs per gram of fish increased from 30 in 1950- 1954 to 39 and 36 in the Duluth-Superior and Apostle Islands areas, respectively, in 1966-1967. Smelt was the most abundant species in both areas and more abundant at Duluth-Superior. Correlations between fishing intensity and herring abundance suggest that overfishing was not the major factor in the decline of the species. Predation was not found to be a serious influence on herring. Food of herring, bloaters (Conegonus fioyi), and smelt was quite similar during early stages. Food similarities and highly significant negative correlations between herring and bloater abundance and herring and smelt abundance suggest that competition for food at the critical larval stage was likely the major influence in the lake herring decline. Data and Review : Lake herring abundance in western Lake Superior declined drastically between 1954 and 1966 due to overfishing and interspecific competition. Anderson and Smith believed competition for food between lake 138 herring, chub (bloater) and smelt was largely responsible for the decline; competition was likely most intense between larvae and may have helped to account for declining recruitment while fecundity is increasing. Based on the data,r the authors* conclusion that competition is largely responsible for the decline is certainly debatable. Rather, overfishing of lake herring may have permitted the other species to increase. Mean age of the population (3.9 yr in 1950-1959) increased as population size decreased (5.1 yr in 1966-1968) indicating a decline in recruitment over the period. Size of herring increased as population size decreased. 1950-1954 mean length 287 mm; mean weight 148 g. 1956-1961 mean length 305 mm; mean weight 207 g. 1966-1968 mean length 340 mm; mean weight 301 g. Condition factor = 1950-54, 0.63; 1956-61, 0.73; 1966-68, 0.77. Fecundity increased as population size decreased. 32 Higley, D.L., and C.E. Bond. 1973. Ecology and production of juvenile spring Chinook salmon, Oncopfiynchus ts/iawytscAa, in a eutrophic reservoir. Fishery Bulletin 71:877-891. Keywords: CHINOOK SALMON, NORTHWEST, ROTENONE, PREDATION, SUMMERKILL, INTRODUCTION, GROWTH, CONDITION FACTOR, COMPETITION Abstract : Juvenile spring chinook salmon, Onconfiyncfius ts/iawytscfia, were reared in a central Oregon reservoir of 7.5 hectare. In 1961, epilimnion conditions of high pH, high temperatures, decreasing oxygen concentrations, and possibly algal toxins caused condition loss and deaths among salmon. Fry planted in 1961 (75,300) and in 1962 (150,000) suffered first-summer mortalities in excess of 80%, primarily due to predation by older salmon. Summer growth was rapid, but dependent on population densities. Coho salmon, 0, kisutcht and chinook salmon remaining from a 1959 plant averaged 280 and 215 g after 30 mo. The 1961 year class averaged 62 g at 10 mo and 89 g at 22 mo. The 1962 year class averaged 22 g at 9 mo. Average condition factor values rose above 1.20 in the summer. Net production by the 1961 year class was 159 kg/hectare in 1961 and 35.5 kg/hectare in 1962. The 1962 year class produced 170 kg/hectare in 1962. Apparently competition from the 1962 year class caused the 1961 year class to feed more on littoral and terrestrial forms and to grow and produce less in their second year. Data and Review : Higley and Bond studied the growth of chinook salmon fry introduced to a eutrophic reservoir in Oregon that had been treated with rotenone in 1960 to remove all salmonids. Fry were introduced in Feb., 1961 and M55 cm. Perch populations (abundance and biomass) were greatly reduced in both basins due to the wartime fishery (2 to 18-yr-old fish). North: 58 metric tons in 1941, 10 metric tons in 1949 South: 80 metric tons in 1941, approximately 32 metric tons in 1949 There were concomitant changes in perch growth rate and sex ratio, but these changes are not well documented in this paper [see LeCren. 1958. J. Anim. Ecol. 27:287-331]. Perch populations increased in both basins in the mid- to late 1950s and reached peak biomass levels in 1961 and 1962 that exceeded 1941 biomass; populations steadily declined after 1962. Pike abundance increased from about 1,500 fish in 1945 to around 4,000 in 1969. Pike biomass remained relatively constant at around 3 metric tons from 1944 to 1969. Constant pike abundance (in terms of biomass) was due to changes in age distribution and growth rate. Pike were dominated by the 5-9 yr age-group in 1945. By 1975, dominance had shifted to fish <5 yrs old. Fish >9 yrs comprised about 10% of the population in 1945 but only <1% in 1975. Growth rates of pike in all age- groups increased between 1945 and 1975. For example, mean weights (kg) of age 7 fish were: 1944-45 males = 2.79 females = 5.73 1952-55 males =2.95 females =5.99 1972-75 males = 3.20 females = 7.26 144 Increas€ss in mean weight were greatest for younger fish. Compensation (in terms of population biomass), caused by pike selectively removing older fish, occurred through increased numbers of younger fish and increased growth rates in all age-groups. Factors responsible for the increase in numbers of younger fish were not discussed. The increase in pike numbers was apparently sufficient to maintain the perch population at low levels due to the increased predation rate by pike on perch. 41 Raleigh, R.F., D.H. Bennett, L.O. Mohn, and O.E. Maughan. 1978. Changes in fish stocks after major fish kills in the Clinch River near St. Paul, Virginia. Amer. Midi. Nat. 99:1-9. Keywords: GIZZARD SHAD, SUCKERS, MINNOWS, SOUTHEAST, TOXIC SPILL Abstract : Fish populations were sampled during 1973 and 1974 in a section of the Clinch River, near St. Paul, Virginia, that was subjected to chemically induced fish kills in 1967 and 1970. Species composition and relative abundance were compared with those from prekill, postkill and postkill recovery periods. A slight decrease in number of species and a change in relative abundance were found. Our data support the findings of others that general recovery following fish kills is relatively rsipid. Data and Review : Chemically-induced fish kills occurrcid in the Clinch River in 1967 and 1970. The 1967 kill reduced fish species richness from 15 species (plus two species groups - minnows and darters) to seven species (plus minnows). Fish abundance was reduced by 79% and virtually all b'enthic invertebrates were eliminated for the first 5-7 km below the chemical spill and greatly reduced for the next 120 km. The stream was restocked in 1968 with channel catfish, muskellunge, redbreast sunfish (Lepom's aunitus), rock bass, smallmouth bass (MicPoptenus doJomieui) , and walleye. Muskellunge and redbreast sunfish were not present or were very rare in the stream prior to the 1967 kill. The 1970 kill was less severe, killing 5,300 fish, but effects on species composition were not documented. The river and several tributaries were extensively sampled by electroshocking between 1973 and 1974. The major findings were: 1) There were 47 species collected postkill, 25 of which were minnows and darters (1973-1974), and 15 species (aside from minnows and darters) collected prekill (1964-1965). The differences could largely be attributed to the restocking program and to differences in collecting techniques and efficiencies between pre- and postkill sampling. Eleven species in the 1973-1974 sampling were represented by only one individual. Therefore, species richness was probably similar in pre- and postkill periods. 2) Relative abundances of species showed major shifts between pre- and postkill periods. Golden redhorse (Moxostoma enytfinunufn) and gizzard shad were the most abundant species in prekill samples. Postkill 145 samples were dominated by shorthead redhorse {M. macnoJepidotufn) , golden redhorse, carp, minnows (collectively), and northern hog sucker (ffypenteJi ufij nigricans). 3) Colonization of defaunated areas appeared to largely come from the river proper rather than from the nearest tributary. However, 30 of the 57 species present in 1973-1974 could have re-invaded from tributaries. This study effectively showed that, in streams, immigration is a powerful means of recovering from catastrophic reductions in abundance. Without additional follow-up studies, it is difficult to interpret the shifts in community composition (i.e., relative abundances of species). It is possible that the 1973-1974 community had not yet reached an "equilibrium" and that, given sufficient time without disturbance, the community composition may return to prekill status. 42 Kenyon, R.B. 1978. Growth changes in a population of whitefish in Lake Erie. New York Fish and Game Journal 25:129-139. Keywords: LAKE WHITEFISH, GREAT LAKES, OVERFISHING, GROWTH Abstract: Experimental data collected from lake whitefish in the Pennsylvania waters of Lake Erie from 1971 to 1974 were used to describe the population since its drastic decline after the late 1940s. Growth was calculated and compared with that of the whitefish population of the late 1920s when the stocks were considered abundant. The lower rate of growth in recent years was deemed significant. A significant increase in weight per unit of length between the two periods was indicated by length-weight regressions (log transformations). Probable explanations for these changes are discussed, focusing primarily on climatology, the trophic state of Lake Erie and the zoogeography and abundance of whitefish. Data and Review : Kenyon compared growth rates of Lake Erie lake whitefish (Conegonus Clupeafonmis) between 1927-1930 and 1971-1974. The 1927-1930 samples were taken while whitefish were still abundant in Lake Erie and they were heavily fished. The population declined markedly in the 1940s and was still very low during the 1971-1974 sampling. Growth rates were greater in 1927-1930 than in 1971-1974, largely due to slower growth of older age-groups in the 1970s. The author interpreted these differences as the result of compensation. He argued that growth rates were high in 1927-1930 because the population was responding to the fishing pressure. However, compensatory growth would also be expected in the 1970s because of the low population densities, but he contended that growth rates are low because of poor habitat quality. Consequently, it is difficult to make conclusions about compensatory growth by comparing the 1920s and 1970s populations because an additional variable, habitat quality, confounds the comparison. 146 43 Jensen,, A^L. 1981. Population regulation in lake whitefish, Conegonus cJupeafontnis ,(Mitchill). J, Fish Biol. 19:557-573. Keywords: LAKE WHITEFISH, GREAT LAKES, MODEL, AGE AT IMATURITY, SURVIVAL, GROWTH, AGE-SPECIFIC FECUNDITY Abstract : Mechanisms that might compensate for changes in mortality are well known but documentation of their operation and determination of their relative capacities is difficult. In this study the potential of lake whitefish to compensate for fishing mortality is quantified using a conventional fishery stock assessment model. The mechanisms examined are proportion of females, fecundity, survival of eggs and larvae, decrease in age at maturity, and growth. The change in each of these parameters necessary to maintain constant recruitment with increasing fishing pressure is determined. Decrease in age of maturity and increased survival of larvae and eggs have the greatest potential for compensation. The total capacity for compensation appears large even when the mechanisms are considered alone. A review of several lake whitefish population studies indicates a large variation in the vital statistics. There is a close relation between total mortality and growth. There is also a close relation between growth and age at maturity. The analyses indicate that increased fishing results in increased growth which then results in a lower age at maturity. This feedback mechanism enables lake whitefish to respond to changes in environmental conditions and on average to maintain a birth rate in balance with mortality. Data and Review : Jensen applied the Beverton and Holt stock assessment model to lake whitefish populations to investigate which mechanisms might allow populations to maintain constant recruitment with increasing fishing pressure. Parameter estimates for the model were taken from studies of lake whitefish in northern Lake Michigan (see Jensen 1979 cited herein).. Changes in sex ratio and fecundity were shown to have little potential to compensate for exploitation. Jensen reviewed data from several lake whitefish populations to determine which mechanisms appeared to be operating in nature. In all populations, exploitation was associated with faster growth and a decreased age at maturity. Model results showed that increase in growth rate alone cannot be an important compensatory mechanism; however, increased growth rate was related to other factors - increased fecundity and lower age at maturity. In the exploited populations, there was a significant relationship between growth rate and age at maturity (higher growth rate - lower age at maturity) . Also there was a positive relationship between total mortality rate and growth rate. Jensen concluded that increased mortality (on mature individuals) reduced intraspecific competition, making more resources available per individual. This resulted in increased growth rate which in turn, resulted in decreased age at maturity and perhaps increased age-specific fecundity. 147 44 Jones, D.D,, and C.J. Walters. 1976. Catastrophe theory and fisheries regulation. J. Fish. Res. Board Can. 33:2829-2833. Keywords: MODELING, CATASTROPHE THEORY, OVERFISHING Abstract : Catastrophe theory is a mathematical approach for analyzing systems that show abrupt changes. This paper attempts to apply the theory to fishery development and collapse, where the key variables are assumed to be stock size, level of economic investment or fleet size, and technological efficiency of the fishery gear. The analysis produces no new predictions about uncontrolled development and its consequences, but it appears to provide a useful way of exploring the implications of regulatory policies involving taxation, development subsidies, and technological improvement. Review : This paper gives a clear statement of catastrophe theory, showing development of stock size - fleet size - fishing efficiency relationships and construction of a catastrophe manifold, a graphical 3-dimensional surface. The folded manifold is a "cusp catastrophe," where a small change in fleet size or efficiency results in stock collapse. This is applied to the Antarctic fin whale fishery. The paper states that Pacific salmon are another example, due to the advent of purse seining (increased efficiency), but no details are given. The authors suggested regulatory schemes based on the theory. 45 Walters, C.J., G. Steer, and G. Spangler. 1980. Responses of lake trout {Salveli nus namaycush) to harvesting, stocking, and lamprey reduction. Can. J. Fish. Aquat. Sci. 37:2133-2145. Keywords: LAKE TROUT, GREAT LAKES, PREDATION, OVERFISHING, CATASTROPHE THEORY, AGE-SPECIFIC SURVIVAL, GROWTH, AGE AT MATURITY, MODELING. Abstract: Sustained yields, declines, and recovery of lake trout (SalveJinus namaycush) can be explained by a simple model that hypothesizes normal population regulation through density dependent body growth, coupled with depensatory lamprey mortality. The model indicates that either lamprey or fishing alone could have caused the Lake Superior decline, though they apparently operated in concert. The presence of depensatory lamprey mortality leads to a "cliff edge" in the system's dynamics, such that catastrophic changes may be repeated in the future. It is not unlikely that Lake Superior is on the verge of a second collapse. Review: The authors used data from many lake trout studies as parameters for a model demonstrating lake trout decline and recovery. Catastrophe theory is used to demonstrate large changes in stock biomass resulting from small changes in fishing effort and lamprey abundance when stock is too close to the "cliff edge". Many useful references are given. The authors related model to 148 policy. A good overview is given of lake trout population dynamics in Lakes Michigan and Superior, despite simplifications in the model (e.g. same reproductive success assumed for planted and native fish). Early warning indicators for lake trout are 4ecrease in native yearling ratio, decrease in abundance of large fish, and increase in body weight of fish < age 8. Fishing and lamprey predation decrease the domain of stability of spawner biomass; i.e., upper equilibrium biomass moves closer to lower critical biomass. 46 McFadden, J.T. 1977. An argument supporting the reality of compensation in fish populations and a plea to let them exercise it, p. 153-183. In W. Van Winkle [ed.] Proceedings of the conference on assessing the effects of power- plant-induced mortality on fish populations. Pergamon Press Inc., New York. 380 p. Keywords: REVIEW Data and Review : An iconoclastic, eclectic review of the literature on compensation in fish populations. McFadden reviewed what is meant by compensation, gave a historical review of the development of the concept, and then argued for the widespread occurrence of compensation in fish populations based on a discussion of density-dependent and density-independent models of population growth and Ricker's stock-recruitment models. [Table 1 (pp. 169- 170) gives 17 examples of compensation.] He concluded with a section dealing with the applicability of fisheries dynamics principles to power plant impacts. He argued (based on Ricker 1954) that the principles of fisheries management can be applied to power plant impacts (largely directed at early life history stages) and that "fish populations can readily sustain 'exploitation' by power plants at levels comparable to those experienced in commcsrcial and sport fisheries." 47 Butl€5r, R.L. 1965. Freshwater drum, (ApJodfnotus gnunniens) , in the navigational impoundments of the upper Mississippi River. Trans. Amer. Fish. Soc. 94:339-349. Keywords: FRESHWATER DRUM, MORTALITY RATES, YEAR-CLASS STRENGTH, MIDWEST Abstract : Scale samples from fish taken during the yecirs 1944 through 1948 were used to analyze the freshwater drum fishery of the Upper Mississippi. Total and fishing mortality rates were calculated for heavily and lightly fished populations in the impoundment system of 26 dams. Mortality rates of drum from the northern section of the river were compared with those from the southern section. Weak year classes were formed during 1942, 1943, and 1945, and strong year classes were formed during 1941 and 1944. Strength of year class was directly related to air and water temperature during May but not to water stage during the spawning season. 149 Data and Review : Fluctuations in drum year-class strength were related to air and water temperature. An argument was given that larva mortality rates are dependent on changes of temperature and thus ultimately determine year-class strength. However, no supporting data on larva mortality or their response to temperature were provided. 48 Swedburg, D.V., and C.H. Walburg. 1970. Spawning and early life history of the freshwater drum in Lewis and Clark Lake, Missouri River. Trans. Amer. Fish. Soc. 99:560-570. Keywords: FRESHWATER DRUM, FECUNDITY, FEEDING, YEAR-CLASS STRENGTH, MIDWEST Abstract : The early life history of freshwater drum, (ApJodi notus gnunniens), was studied in 1965-1967 to determine factors influencing year-class strength in a main stem Missouri River reservoir. Fish spawning occurred over a period of 6-7 weeks in June and July when water temperatures reached 18®C. Fecundity of fish 307 to 386 mm long and 6-9 years old ranged from 34,000 to 66,500 ova. Ova reared in the laboratory at a water temperature of 23 ®C hatched in 27 h and the prolarvae averaged 3.2 mm long. The prolarval stage was complete 45 h after hatching and postlarvae averaged 4.4 mm long. Adult characters developed at an approximate length of 15 mm. Diet of fish 6-15 mm long was Daphnia and Cyclops. Bottom fauna became important in the diet of fish longer than 20 ram. Fish movement between channel and floodplain was related to changes in diet. Growth, year-class strength, and mortality varied among years. Strong year classes were associated with warm summer (June-August) water temperatures in the spawning and nursery area and weak year classes with cold temperatures. Size of year class was determined before fish attained a length of 25 mm. Data and Review : Much of this paper provided background information on fecundity, larval development, larval and juvenile food, movement of larvae, and factors influencing year-class strength. A relationship was drawn between year-class strength of drum over three years and summer water temperatures (as cumulative degree days). Warm temperatures coincided with strong year-classes whereas cold temperatures coincided with weak year-classes. Mortality curves for age-0 fish indicated the differences in year class strength could not be accounted for by different mortality rates over the three years for fish between 25 and 94 mm long. The authors, therefore, argued that much of the mortality that determined year class strength must have been for fish less than 25 mm long. However, they were unable to actually measure such mortality and thus could not support this claim with actual data. 150 49 Priegel, G.R. 1971. Evaluation of intensive freshwater drum removal in Lake Winnebago, Wisconsin, 1955-1966. Tech. Bull. Number 47, Wise. Dept. Nat. Res., Madison, Wisconsin. 29 p. Keywords: FRESHWATER DRUM, EXPERIMENTAL FISHING, CONDITION, RECRUITMENT, INTER- AND INTRASPECIFIC COMPETITION, MIDWEST Abstract : An intensive freshwater drum removal program was undertaken on Lake Winnebago with assistance of qualified commercial fishermen in 1955. A 12- year evaluation study, 1955-1966, was initiated at the start of the intensive removal program to evaluate the effects of freshwater drum removal on the freshwater drum population and other species. From 1955 through 1966, 15 million kg of freshwater drum were removed from Lake Winnebago. Initial heavy removal, 5 million kg of freshwater drum in 1955-57, resulted in a decided change in the condition of freshwater drum; however, the improvement in condition that was evident early in the study tended to be lost later. The commercial harvest was intensive enough to crop off the larger and older freshwater drum by 1962 so that the harvest after 1962 was composed of smaller and younger fish even though rate of growth remained the same. Selected trap net sets during April-June and trawling during the summer and autumn were effective methods of removing freshwater drum. To keep the freshwater drum population at an optimum size in Lake Winnebago, 1.1-1.3 million kg should be removed annually. Since the freshwater drum is very prolific fish and occupies a favorable habitat, the population would soon consist of many slow- growing, old-age fish, which was the situation before the program began in 1955, if maximum effort is not maintained. There were benefits to the sport fishery that appeared to result from the freshwater drum removal program; however, no positive correlation could be demonstrated.. There was no indication that commercial removal of freshwater drum was detrimental to any game or panfish population. Data and Review : After intensive removal of freshwater drum over a 12 year period by commercial fishermen in Lake Winnebago, the change in the numbers or biomass of the population appeared to be negligible based on the catch-per- unit effort data provided. The author claimed, however, a substantial decline in the drum population occurred but based his conclusion upon total catch data which would be biased due to varying fishing intensity. As CPUE remained steady in both numbers and biomass, and age composition shifted to younger fish, recruitment must have improved, possibly due to decreased intraspecific competition. The condition factor of drum fluctuated over the study period with no consistent trend. Figures and text on condition factor in this paper do not agree, making it difficult to evaluate the author's interpretation. The suspicion here was that compensation must have occurred to maintain the drum population despite intensive removal of adults. However, the mechanism for this could not be ascertained based on the data presented. CPUE of white bass (MoPOne chnysops), black crappie, and sauger rose strikingly after drum removal. Sauger were most abundant in 1960 and 1964, dropping back down to pre-experiment levels in 1961-1962, whereas white bass and black crappie peaked in 1961-1962 and then dropped. This apparent inverse relationship, coupled with the changes in freshwater drum condition, suggests that some complex competitive community interactions were occurring. 151 50 Cadar G.F., and G.L. Hergenrader. 1980. Natural mortality rates of freshwater drum larvae in the Missouri River. Trans. Amer. Fish. Soc. 109:479-483. Keywords: FRESHWATER DRUM, LARVA MORTALITY, MIDWEST Abstract : Instantaneous total mortality rates for planktonic larval freshwater drums (ApJodf notus gnunniens) in a channelized stretch of the Missouri River were estimated from analyses of cumulative catch curves. Mortality rates, which ranged from 0.11 to 0.21 per day, were significantly greater in 1976 than in 1974. Basing our interpretation on the underrepresentation of eggs and early larvae in the channelized river samples, we hypothesize that favorable environments upstream represent a more significant source of recruitment of larvae to the freshwater drum population than the channelized river . Data and Review s A method was described to effectively assess larva mortality rates for a natural population of drum based of length-frequency data of larvae caught in plankton nets. Consideration was given to sampling inefficiency for some size groups of larvae, so mortality rates were estimated only for fish between 5 and 11 mm long. Results of sampling drum for two years on the Missouri River showed significant differences in mortality between years but not within a year for two separate sampling stations. No causes for the difference were given nor was the impact on year-class strength known. The lack of larvae less than 5 mm long was probably due to spawning and early development having taken place upstream and not in the sampling area considered. Thus, later larval stages taken in the samples were drifting downriver from these spawning areas. In rivers, compensation would be more difficult to detect due to extensive immigration from upstream. 51 Bardach, J.E. 1951. Changes in the yellow perch populations of Lake Mendota, Wisconsin, between 1916 and 1948. Ecology 32:719-728. Keywords: YELLOW PERCH, MIDWEST, DISEASE, COMPETITION, GROWTH, CANNIBALISM Abstract: Data from 1916-1948 on yellow perch from Lake Mendota, Wisconsin were collated and discussed. There was a general decrease in the number of yellow perch over the last 50 years and a concomitant increase in their size (doubled) . The decline in abundance and growth improvement were related to disease (myxosporidian of the genus MyxoboJus) , reduction in the numbers of a competing species (lake herring), and an increase in fishing pressure. The disease also reduced cannibalism on younger perch because larger yellow perch were the most affected by the disease. The most noticeable acceleration in growth rate occurred in 1939 when the disease reached epizootic proportions. From 1916 to 1947 there was an approximate 4.3-fold decrease in catch/effort. 152 The average total length increased from 162 to 243 mm and mean weight went from 50 to 180 g or 3.6 times its former level. Condition factor improved from 1.18 in 1916 to 1.25 in 1948. Data and Review ; Bardach's paper generally supports some of the compensatory responses seen in other studies. All the conclusions are based on circumstantial evidence and co-occurring events and are not strongly supported, but believable. 52 Wells, L. 1977. Changes in yellow perch (Penca fJavescens) populations of Lake Michigan, 1954-75. J. Fish. Res. Board Can. 34:1821-1829. Keywords: YELLOW PERCH, LAKE MICHIGAN, ALEWIFE, COMPETITION, PREDATION, OVERFISHING Abstract: In the early and mid-1960s yellow perch in Lake Michigan declined, which was related to the alewife invasion. An intensive commercial fishery hastened the decline. High production in the commercial fishery was related to increased growth of remaining perch (growth compensation) . Data and Review : This is a good discussion of the impact of an exotic species on yellow perch. The mechanism for the impact was speculated to be predation on yellow perch larvae by alewife. Catches of yellow perch over 1962-1975 are given. An excellent table (3) showing growth of perch is given. For example, female yellow perch of age 4 were 189 mm in pre-alewife days (1954), then after alewife appeared in the 1960s, age-4 yellow perch averaged 203 mm (1963), 222 mm (1964), 267 mm (1970) and 241 mm (1975). 53 Fraser, J.M. 1978. The effect of competition with yellow perch on the survival and growth of planted brook trout, splake, and rainbow trout in a small Ontario lake. Trans. Amer. Fish. Soc. 107:505-517. Keywords: ONTARIO, YELLOW PERCH, COMPETITION, GROWTH, SURVIVAL Abstract : Brook and rainbow trout and splake were stocked in a lake for 12 yr. Yellow perch were introduced after the first 6 yr. Mean return per kilogram of trout stocked decreased from 3.3-6.8 kg to 0.4-0.9 kg after perch were introduced. Yellow perch dominated the fish community (20-30 kg/ha) and caused a change in the food eaten by salmonids which resulted in a 50% decline in growth rates. Salmonids could not compete with the yellow perch. Data and Review : This study is a classic example of competition between yellow perch and salmonids. Good data are given on biomass of yellow perch and of salmonids, before and after perch were stocked. Evidence of repression of the following year class of yellow perch by the 1969 year class, possibly due to 153 competition^ is presented in Fig. 4. Some growth data are also presented. Yellow perch grew well to age 2, then barely grew at all. Members of the 1967 year class of yellow perch reached 240 mm in 26 mo, surpassing growth rates for most waters. Growth ceased after 240 mm. Yellow perch growth rate declined after they became established. The only negative interaction between species was believed to be competition for food, as salmonids did not eat any yellow perch or vice versa. 54 Grice, F. 1959. Elasticity of growth of yellow perch, chain pickerel, and largemouth bass in some reclaimed Massachusetts waters. Trans. Amer. Fish. Soc. 88:332-335. Keywords: YELLOW PERCH, NORTHEAST, GROWTH, LARGEMOUTH BASS, COMPETITION, ROTENONE Abstract: Massachusetts ponds and lakes were reclaimed with rotenone. Growth rates of several species, including yellow perch, increased after reclamation, probably due to reduced competition for food. Rate of growth was more dependent on density than length of growing season. Data and Review : This is another good example of the response of yellow perch and several other species to decreased inter- and intraspecific competition, primarily for food. Growth in some of these lakes was as rapid as any previously observed, demonstrating the maximum capacity for growth for these species. Some excellent growth data are given in Table I on a wide range of lakes and ponds. 55 Kelso, J.R.M., and T.B. Bagenal. 1977. Percids in unperturbed ecosystems. J. Fish. Res. Board Can. 34:1959-1963. Keywords: YELLOW PERCH, GROWTH, SURVIVAL, RECRUITMENT, REVIEW Abstract: A review of yellow perch in unperturbed systems showed these areas to be harsh environments and rarely studied, with growth being low and variable. Temperature stands out as a major factor affecting yellow perch growth, but a general growth rate applicable to all unperturbed populations cannot be identified. Particular year classes dominated a population in lakes undergoing light to no exploitation, but domination by a year class was less marked in perturbed systems (e.g., Windermere). Unperturbed systems usually have a wide range of age-groups present. Survival ranges from 27 to 69%. Conservation of biomass was a consistent feature of yellow perch populations. 154 Data and Review ; This is a literature review giving a good overview of unperturbed systems • Such systems are valued as good hypothesis testing opportunities • Growth, survival, and recruitment are compared between perturbed and unperturbed systems. 56 Nakashima, B.S., and W.C. Leggett. 1975. Yellow perch {Pence flavescens) biomass responses to different levels of phytoplankton and benthic biomass in Lake Memphremagog, Quebec-Vermont. J. Fish. Res. Board Can. 32:1785-1797. Keywords: YELLOW PERCH, COMPETITION, GROWTH, MORTALITTiT, NORTHEAST Abstract : Perch biomass in Lake Memphremagog, Quebec was 2.5 times greater in the more productive south than in the less productive north basin. A large perch population existed in the south basin; growth was similar in both basins. This study provided empirical support that with abundant food supply, population size and not growth is enhanced. Data and Review : The data and conclusions from this study provide empirical verification of the compensation phenomenon, that when a fish is near maximum growth, its ability to compensate through growth is rtsduced. In this study, perch were growing at or near their maximum levels for Quebec; therefore, with increased eutrophication and an abundant supply of food, growth did not increase (as often is the case when perch are below msiximum) . Greater food supplies in the south basin increased survival of young perch. YOY perch were 5-10 times more abundant in the south than the north tesin, while older perch were only about two times more abundant, so compensation occurred in the early life stages and tapered off later. Growth and mortality rates are given. Summer and fall predation played a small part in regulating YOY populations. Cannibalism rarely occurred. Rainbow smelt were the main forage fish of several species, including yellow perch. 57 Ryan, P.M., and H.H. Harvey. 1980. Responses of yellow perch, Penca flavescens (Mitchill), to lake acidification in the La Cloche mountain lakes of Ontario. Env. Biol. Fish. 5:97-108. Keywords: YELLOW PERCH, CANADA, GROWTH, RECRUITMENT, INTRASPECIFIC COMPETITION, SURVIVAL, ACIDIFICATION Abstract : Yellow perch in acidified lakes of Ontario had reduced densities. Perch of age-group 1-3 responded with increased growth, while those aged 4-9 had decreased growth. Thus, growth rate was proposed as an indicator of environmental stress. 155 Data and Review : This paper presents growth data on 1,667 yellow perch from 38 Ontario lakes subject to acidification and offers the premise that perch growth rates, which are subject to complicating density-iiependent factors such as predation and competition, can act as indicators of water quality and environmental stress • Extensive data are given on all lakes for yellow perch growth each year. Mean length of yellow perch at a given age vs. lake pH relationships indicated that growth changes were occurring in an accelerating manner with decreasing lake pH. The more acid the lakes the less successful was recruitment. These results support the hypothesis that the mechanism of increased growth rate of the younger age-groups at high acid levels was through lowered population densities due to recruitment or reproductive failure. Improvements in growth of the younger age-groups can be attributed to increases in the available food supply per individual as mortality rates increased. These losses may have diminished interspecific or intraspecific competition. Perch were among the most acid tolerant of 30 species investigated. Thus young perch in lakes in the 4.5-5.5 pH range may experience greatly reduced competition for invertebrate food, reduced population density, failure of recruitment of year classes, elimination of acid-sensitive individuals from populations and random loss via episodic acid events in spring. Increased maintenance requirements for older fish cause mortality and, if mortality increases beyond a certain point, reduced recruitment and sometimes population extinction. 58 Tarby, M.J. 1974. Characteristics of yellow perch cannibalism in Oneida Lake and the relation to first year survival. Trans. Aroer. Fish. Soc. 103:462-471. Keywords: YELLOW PERCH, CANNIBALISM, NORTHEAST, YOY SURVIVAL, PREDATION Abstract : Yellow perch in Oneida Lake, New York, were studied from 1965 to 1971 to identify the factors affecting the intensity of cannibalism and to evaluate the effect of cannibalism on the abundance of young perch. Wide variations occurred in the annual intensity of cannibalism. Cannibalism was most intense in August and operated as a density-independent factor. Walleye predation on yellow perch YOY caused depensatory mortality of young perch, implying that cannibalism was not decisive in limiting the strength of year classes in 1965-71 in Oneida Lake. Data and Review : This is a very good paper with well documented data on cannibalism. Tarby used multiple regression analysis to identify length and abundance of young perch as the the most important factors controlling the intensity of cannibalism. The positive relationship between the incidence of perch cannibalism and the abundance of young suggested that cannibalism operates as a compensatory mortality factor. Tarby believed intuitively that high densities of YOY yellow perch should operate as a density-dependent compensatory factor and cause decreased growth in August; however, growth remained unchanged over the period of most intense cannibalism. This implied that cannibalism was density-independent and unrelated to young perch abundance. Forney (1971) attributed most mortality of young perch during the 156 demersal stage to walleye predation. Total mortality of yellow perch from August to the following May generally exceeded 96%; cannibalism (mostly during August) was a very small component of this total mortality. By the time cannibalism peaked, walleyes had already consumed two-thirds of the perch year class, reducing the potential impact of cannibalism on the abundance of young perch. Cannibalism by yellow perch along with walleye predation can in certain years suppress yellow perch YOY abundance. 59 Deelder, C.L. 1951. A contribution to the knowledge of the stunted growth of perch (Penca fluviatilis L.) in Holland. Hydrobiologia 3:357-378. Keywords: YELLOW PERCH, EUROPE, GROWTH, STARVATION Abstract: Studies of stunting of yellow perch in Holl80% mortality which occurred in each year. The low year-to-year variation in survival suggests that temperature and wind influenced year-class size through a complex of many relatively minor mortality factors, rather than through one catastrophic event. Data and Review : Data are given on the number of mature yellow perch in Oneida Lake from 1969 to 1975 as well as estimates of fecundity, number of larvae produced, and survival to the 8-mm stage. The multi-age character of the adult stock and the low and constant mortality experienced by adult perch contributed to stability in the number of eggs produced each year. The multi- stock character of perch minimized the effect of differences in numerical size of year classes. Mortality of prolarvae was presumed to be low. Clady *s important conclusion was that many relatively small mortality factors combined to produce the high mortality eventually experienced by YOY yellow perch. 62 Eshenroder, R.L. 1977. Effects of intensified fishing, species changes, and spring water temperatures on yellow perch, Penca fJavescens, in Saginaw Bay. J. Fish. Res. Board Can. 34:1830-1838. Keywords: YELLOW PERCH, GREAT LAKES, OVERFISHING, GROWTH, COMPETITION, ALEWIFE, RAINBOW SMELT, WALLEYE 158 Abstract: The yellow perch population in Saginaw Bay is discussed with regard to historical changes and more recent (1960-1970s) fluctuations. Intensified effort and size limit removal resulted in exceptional landings of yellow perch from Saginaw Bay. Effort concentrated in the outer bay, then shifted to the inner bay. Yellow perch growth rate increased, and older age-groups were no longer prominent in the catches. Females became relatively scarce shortly after attaining vulnerability to fishing gear. The low recruitment of the late 1960s and early 1970s may have been related to rciduced brood stocks. In the 1940s yellow perch were most affected by loss of the walleye; predation by walleye apparently held down yellow perch populations, resulting in good perch growth. When the walleye population collapsed in the 1940s , yellow perch abundance increased and growth slowed. Rainbow smelt and alewife affected perch by competition and predation in the 1950s, but overfishing had the most intense effect in the 1960s-1970s. Spring temperature had an important role in repraiuctive success of yellow perch. Data and Review ; This is a very good discussion of the behavior of yellow perch stocks in Saginaw Bay. Commercial fishing effort and landings are given from 1880 to 1976, demonstrating how the fishery affected perch stocks after removal of the size limit. Older fish were removed from the population, females more than males; growth increased, and recruitment was lowered. Eshenroder presented a table of growth data covering 1929 to 1970 by year of life, plus data on percentage of fish in each age-group. Temperature and year-class strength showed a strong positive relationship. 159 APPENDIX 2 Author -Date List of Literature Checked for Information on Compensation Numbers at ends of lines refer to order of placement in Annotated Bibliography, Appendix 1 Key to Codes X = no copy made; no information on compensation or catastrophe C = contains information on compensation; N = no information on compensation S = catastrophic event occurred; N = no catastrophic event Y = contains data for quantifying compensation; N = no data R = reviewed; N = not reviewed 160 "^ CM 2 a 2 z u z z z Of > t/> (J o z o (13 "^ CM O C^ 2! >- > > c/") c/) to t/) 2: cj cj o 222Cr)0)(nO)C3)0)O)OTO)O)QC7)O)0)O)O5CDCDO)O)0)O)0)CnOT0)O)0)0)0)0)O)0)0)CDO)OiO^ o c o E E 13 C -H L. (0 -M O (U u ' J- P3 0) r • • c U- -D I • T3 T3 UJ C C (0 (0 u. a • X • -0 Q Q ► Q . . . • O) - Q Q) £- • • "D t. E - 0) ^ ^i: O O JC - — "D -U £- O) C T3 C - »C5 E£-W<0Oa)X-^ '- 0) E (0 C t/) > o • c U. U (0 > o • 13 C • •> (0 Q • C5 Ui • O (0 - w 3 a c o O CD :- U) "D TJ — c c < < < c (0 o Q -Z Q - • - C LL. > o • C L. C5 - o +^ c JZ w - o ■H E r -H C C- W -M < < < < > L • IL t. 0) > • o • X •D • C c •D O C OT rtJ T3 L. 5 -^ O CO c (0 • O -J t/) . -00 00 0) 2 :^ w • X -D 00 . . (/) 00 (0 •^ c r (0 ■M E (0 0) o • 00 < o - -3 • C B E B (0 (0 (0 z :^ zoo . .^ ^ ^ ^ ^ - c q: a C -H (0 (0 (C (0 > >-r- — —cccccuos • - -H3a)©a)a) — — ■D.-r- D>(Q(Qn}(0(Q(0(0(0(Q <- ^ "0 ^ (0 (0 O "O "D ■ o • X M- i^ T3 • C -D awe X (0 z >^ o < c a 0) > O) - C CO o - c ■H C (0 0) .^ ^ 0)0)0) 00 00 00 ■ Q Q. > - • Q >CJ E (0 - (0 0) 0) 0) 00 00 00 • :5 ^ 0) 0) - O) 0) • L. £- < O 0) 0) £_ +-• "H 0) 0) 0) ■D C C CCC 0) 0) 00 00 00 161 z > to z u tp 00 CO CO z a > a > z z Z Z (/> z z z z z z z z z > > z z to I/) Z Z o o o z ^ n in T- ^ CM CNNR CSYR NSNR X X XXXX XXXX X X XX X X XX JQ (0 Ocococo^*c^lcoco^.c^lOcoco^-OcO'^■^^^flcoa)'5tOl^•^lnc^loo^*^pooc^l^co•^cococoocNOGf)0)cncD^)0)0)0)c^lln t^t^^*oo^)^-^^coln^o^)r^lnln • ^ (0 < s • £- +J T3 C D C (0 O (0 to 0^ • • TJ < -J C "0 < < < "D C C - - - Q) 01 O O (U T3 5 +^ * « »+J4-»4J,— 00 • CO-H-H-HWL.t.L.O'-il JL £- L- I_ £> > > >X5 C3) i- QQCDQQfflODmQQmcaOQCD ^ to c X c o o L. £. 00 OQ c c 3 3 o o 00 OQ — - £_ 0) (0 <0 C O) (0 3 C « O D 3 i- L- £_ 00 00 00 0) c Q.«-i E »-• 0) t-i 0) 0) • O D L. • SZ O -D K X O) C T3 ■ < a flJ E "0 • U c !!<: z "D • C • C O CO UJ (Q • C CX X O O U) (0 • -T3 T3 Q C ^ ^ . (0 c c ^ > > > u t- L. t. D 00 00 00 00 X3 Q C • (0 13 > C • O (0 OT X W . 2 • o • a o a -J . . . (D -D >^ CD - O T3 +- * r- O •- T3 ^ - ja C "D - - IL t- 3 3 D ZI 3 00 OQ 00 00 03 O ■D c (0 > ^ ."0 (0—0 N C H N C (0 • ca>^ 2 • 00 tt (0 "D "3 :ic: (0 (0 (0 c c c i. i^ L. D D D 00 OQ 00 a a • T3 CL C • L. (0 • ■M -D-D • — Q • • I 00 — . . o C3) 0)a o o -r t. t. > u L. £. £. (0 D 3 3 3 00 00 00 00 162 O ID 0) (0 Of (^ z z z z z z z z z z 222222222222222CDO)(Dcno)0)0)<»o)cncno)0)cncno)OTQSSSSSSS > N (A a z C (0 £- > E a • c (0 • £- £. U — — W 4-» -H 35 3 3 03 CQ OQ C. CD >T3 > fl3 0) £. -J c: a) ■ c» Q) • i: Ui T3 -i C (0 • o Z "CJ c: "D (0 (« 0) (D E • • CO u.. < -D -O C £- 3 ' 0) » O 5 — «J H O ^-^ "D -- -- D (0 (Q (0 CQ U o a o a a o :^ ^ Q O 0) c to t. (0 £. EEC (Q (0 (0 CJ O O L. L. 0) (D C C (0 (0 (. L. (0 (G C L. 0) 0) ■D X3 C C (0 (0 (0 (0 Q) O C C (0 (0 C- £_ (0 (Q O CJ (0 - I • -o I Z Q) C • z r 0. a • 15 I -M C (/) (0 • c O • (0 < • 5 ^ >^ T3 C3 C . . - (0 C Q Q £- r . o - - C (Q (0 <0 - - E E »- -H c a a £- +-• flj (0 (0 (0 (0 i: ii r O O CJ o o :» ? o a . . 0) c c " - -^ (0 (0 c »- — E E -^ (0 (0 a a > > > (0 (Q (0 i: r r r i: CJ u CJ o CJ -3 I (U 0* -3 ^ 3 3E OQ 3: 3; 0" - S 2 S UL C) O c (0 c • • (0 Q Z 36 T3 C CO (0 -^ - — J^ O O CIO i. i, . CD 3 I (0 -J • K lU ^-«4J+J4-'4-»-HX: OTto{ntowtno>>>^j)ij)C •*- — -^ -^ — — £- X3 "0 XJ £- t. t. t:^t-C-t.t.3<0(0m(0aj(0 OCJUOOCJOCJCJCJCJOO CJ T3 - Q. C • • (0 T3 to ^ CJ C • O £-"0 • to -D to E — «2 t. 5 to • • L ■ $ O--D-CQ+^-D0-D-D i3 • X I C -^ -H 0:5 - (0 • • • a -•CQLucjQ.acLCL- - W C O «| -CC0+- J:^-'-0atfl0Ct.>.>>,> - t. > E a D)— (D n n n n ucj(jcj CJ 163 o rr n CO 03 0) in cc z > > Z (/) z z z z z C/) z a z z z z z z z Of z z z z > t/) Tr-r-t^OOnt^o)OOOtDO)Oinion^ooooa)^innoc>icnr)a>'r-^n'^(jDCNcooocoo)'^Ococ»5'«-oo)Trooot*-t>- 030)0)0)CDO>o)0)0)cncno)(7)(no)0)0)(na)o>o)0)0)0)(DO)(ncDO)a)0)0)CDO)(n(ncDa^(no>Q(no)(no)0)0)0)^ c +> -^ (D Q) ^ — CJ o o OQ ■ (0 O - O TO O C - « Z O O • 5 in • "O Z C (0 0) n c E C O O . OQ • • ' • CO CD < UJ UJ . T3 • - • -O C O 0) L. £- flj C 0) 0) »»- - O a a Q) 0) 0) T3 o o a a a £- o o o o o o c o (0 . — 4J . . C C U. U. H o o CJ • • -3 "0 c - - © (0 O) O) u ^ ^ c X (0 (0 (0 O iL ^ &. o o o o 0) (0 > • (0 s OQ o :3C (0 • z CJ T3 • c J (0 T3 • C < (0 • O X "H o O > - o • 3 N Q) C L. «J N 0) 2 5) 0) • a o o o o c o (0 Q) I. £. H -< "D U • • C OQ -J • (0 • X • > • ■ "O £- CSC T3 X < £- C • (0 • (0 • 5 -J C Q > O • - - V Q • ->£-»- 03 O 0) 0) -r- - C O C T5 ^ C •'- O «H 5 C — r 0) o D o (n £-£.£.£-3 3 u o o o u o CQ • • c X X « O O X O (0 •«- £- ja o >x: a -H 3 • (0 3 o o a X3 c • (0 OQ - • s : o < — C Z (0 >5 •D £- • (0 QC X T5 • C GC (0 T3 C ■ <0 C T3 •«- C -H (0 C to C •^ S O — i^ X 0) • QC Z OQ QC . . a 3^ TJ ■ "0 13 C t/) C C (0 (0 (0 : -D X O) 0)u. c c r r --' (0 (0 jC 3 3 (0 O U Q £.£.0 0) ® 0) T3 0) U 13 -^ -»- " — > > (Q (Q CO (Q O Q O Q 10 C C Q) O O •^ S W > «J (Q 0) 0) Q O O • • < C3 a X . . . -D • • 5 < -D Q -35 >^ - - c c -0 C - £- t. O •- C O C (D 0) CO D O C K -O (0 X :$ • 3 X in X -5 C • <0 < E £- C O Q) O - > C 3 " • > - £- O (0 (0 "D O < (Q (0 r j^ O T3 to £- - - - C - (0 TJ T3 t. £. -O'-NOOQ) C £. (0 (DX(Q>>C£.E®a}^EO)i 0)-.- 0£.t. 33 >UJQOX5 O)-'-- — OOOOOOQOLULUUJUJUJUJUJLUUJ > c 0) (0 (0 164 CM r^ CM ^ w (0 oc (X. QC z z >- (/) in z u z o CO in an >• «/) u z z z o ^*ooo)^Ci)co^»cM^-lD^* • '- C- o c i. (0 CD • • •r -D o a - ' X3 > C C C O T3 O O O O O 0) ■H +^ 3 X2 "O '- >- ^ £ C UJ UJ LU UJ LU :5 :» ^ I - .... (0 a J- L. ;.{.£. i. (1) 0) 0) O 0) 0) "O > > > > > o 0) Q) 0) 0) £. e E E s e c r r r jz r q> o o o o o r en (0 (0 (0 (/} (/) III III i^f in ^^ ^ If I c (0 I c , D) 0) c (0 0) (D 0) c O E E (0 (0 (0 3 — -^ c .- - 0)— '- <0 -^ - S 3 3 • TJ T3 -3 C C (0 (0 r r c - (0 c o (0 3 01 c U. U. iX. U. U. I V c (0 c q: -D 0) 0) Z (0 ' c c c -J • C5 C C • O O (0 ^ 0) 0) -D Q -J . L. • w - _- Q 0) -J UJ i > r I o E •^-J-J C-J C-J-J-J-J I h- i: r . . o • o • • • • 10) i(na)(Da)(D0C}a)a)a)*- u^^(UO »£.-h (UCCCCCCCCC-H -cccwc>>w lt.I.£.£.£.Cl.L.l.£.L.X(Q(0(0<0(DQ}a)O iOOOOOOOOOOOO£-£-t.t-£-£-L.£. 0) C3) a a (c> .^ .^ V' ^ ^ .^ •0 "0 tri c c (0 (0 T3 c • • • (0 LU Ui LLi (0 o (0 ca L Cl Q) (0 • tn 5 03 -D • I u c • • T5 z c (0 :» 2B 3 " " - (Q 4-» +-• -H 4J OT m w -^ 0—. 1- t. £- D I > <0 t. — c T3 3 — O 3 O D> E .^ X3 "O E 3 (0 (Q (Q u- O O CJ (0 C- O) "D Q) (Q (U O to Q T3 a c o V a c CO T3 c • (0 < • C- X <0 - o L. Of E « ui-a C L. (0 (0 (3 (3 c (0 a • • • -D H • • u. il CJ . . . ' CJ K -D > . - ► (Q • 0) (0 3 0) D) O) 0) (0 to 1. t. "D W 3 O L. - Q Q r • • 0)0)0) c c c ^ JS^ -5^ C £- J- O CJ CJ 165 z z ID 52 > tt z z z (/) t/) cj z >- t/) o a. >• z a z z > z z en Z Z X X CM CO > CJ (nQCDO)Qcn(no)CDO)CDa)(n(DO)0)ooG)0)CDcn(Dcno)G)0)CDO)(nco(no)0)CDa)0)0)CDO)0)0)0)(no)0)0)Q Ql 0. Ql U- • • • -D -D o CJ a w LU • . (0 - t- £- £- <0 C U- U- C 0) (D Q) E 13 - - '-"O'D'D'DJCCOaJOO f-^000fi3(D(D"--^ •r-0000£-£-£-t-£- CJOC3C5C3CJC!JC5Ci5CU C5 C5 ■D C3) £. C C 0) CB -^ H — C -H -- C C 3 3 3 CD CJ C3 I C O X3 -D C IL (Q O O • C L. O Q) C C 3 3 O (3 >0Q • • £- • -3 (0 . 0) W C (D (0 C T3 (0 (0 I I - e c c > (0 "(DO) > O £. E E 0) 0) (0 (D +^ -H > > JC — £- t. IL L. m to (0 (0 (Q (n (Q (Q X X X X X X C5 c (D r a 0) t/> • • C < < (0 r r • (/5 +» "H a •(DO) Q >«: :£ u- CJ (0 c r 3 L. 0) 0.0 0^ r . 4- T3 < 3 C (0 (0 • O IX. O • (0 "D T3 C C C -^ 03 <0 '- a u- DC O r- o 3 ^ c 0) U) 3 (0 X X c > > w — 0) Q} 0) 0) CD > > > > N (0 (0 (0 (0 (0 X X X X X S to Q C o - - ■> tn > > > 0) 0) £- 0) 0) 0) Ji '- Q) r- r- ^ +- C "O (C (0 «J -D -^ C CU (1) 0) 0) CD Q) X X X X X X Q t. (!) • "O LL. (0 - C > (U £- 0) C C- 0) 0) X X o (n c c i^ O O (D V -M 3 O) C3) C C - .,- .- Q) c c en C- JL w 0)0)0) XXX c O V to c E O O OQ C r o K • to UJ (Q Q) < • Q O " 0) - (0 (0 •- 0) 0) n- .- .,- X X X X 166 CO ^ ^ CO z o z z z z z z o Z Q: Q£ z > z Z (/) t/) o o z z en z to^^"«-0•'-co^cococo^-cooo■^ln■^t^t^c^t^oo^Pc^lO)^•cot^<»l^tOlnt^•^"«^'^c^lO)c^l^O'^lnO'^r^t^OcocolncDe^^ coco^^nlnlnlf)t^lnt^u)^*t^t^t^^>coa)lDlnI^ln^lO^»^*lDcolor^l^ln^^oot^t^r^lDln^ocp^-t^^'^^cocoI^t^^*r^^ CDO)0)(nO)QO)CnO)CDa)0)GI)CnO)ODO)0)0)CDO)CnO)CDCT>(7>(na)(J)0)0)0)CJ)(nO)(J}CDCnO)(nC})CDO)(J)C^ q: a a i 0)" XIX 13) £- C Q) 3 C -J -H •H • Q) iJL D 00 d • c • CO X C- C > g i: "D o c OT 3 X UL r Q. o a (C ■ • q: q: q: E • i^ C E t. > - J!^ 0) (0 0) 0) — o W +^ -H »- r- O '^ C C i. >v a to a. Of ex 3t Q a a 5 -D -D I < < Q c • - » • - to a y- c - CCCOTW--- i^i^CCCOC OOOCCCCt-L -. OOO-HO a)(O(/)-^'*-a)00(D.^c') - "(otocotou) oooccccOTw.~i3rri:ri:i:i: 3333T3W<0(0<0(D(Da)a)(Ua)-'-0 0000000 xxxXh-i.-*-d-d-d-d-d-d-d-d-d-3-:3-d-d-d-d-3-5-d Q Q Q c • a Q O Q C u. •H » • • • U) V) (n (0 tn U) . c c 0) (D 0) r r c c c c c 3 -D "D -D -3 -D -D -D 167 CM If) oc q: 2 2 Z Z (X. > C/) o ^ CM Ol DC z a z > > > > z t/) t/) t/7 u o u U g 1 o)cno)cn(no)o>o)o>o)cno)0)0)(no)0)0)0)0)0)c»(no)0)0)0)0)CDGnGDO)0)0)0)cno)cn UJ D cn o 0) E • (0 Of 3 X3 <0 C • Q. (0 O * X 1. (DON 13 (0 flj -D >^ i*: - T3 C (0 (D 3 0) (0 • OQ -D L. r o 0) (0 © a! o o CO to 0) 0) o — £- t- — (0 3 n- U i: i: o o • • D J/) JZ • u a 3t • T3 (/) C "D (0 C • i (0 ^ • ■D . . (0 E • • • (0 < TD -D -5 - - -36 L. 1. S. f 0) (D 0) * 0) O) Oi > C C C "D -.- .^ ^ 0) a a a c e £ E c Qi (D C -D -D -J • • T3 < £- • (0 > > »- L. CD ... • (0 • O CJ O 3 X • q: q: £^ a ... ■ ^ ^ c 0)0)0 -^30£E cc>;-OT«+' C C C - 0) 0) r- .^ M- X5 0> (0 »- r- — .^ c (0 . C J^ (0 L. E 0) ai (0 r • u X • T3 O C <0 .- C - 0) L. 4-« to 0) V TJ -- ceo ^ :^ 2^ c r o > O T3 JC Z C ^ to -^ E "D • ty^ O "D 1- UJ O c ao. (0 o +^ T3 • to C Z U -^ CO • r • > o o o ► - - (0 t. C £- C (U Q) 0) .- '-'-£> r r (0 o o O J- L. ^ 2^ :^ ^ < < to • •^ -J to 3 O • E -I Z (0 N C * ^ -^ C O -M to H JC r- E ji r 3 > to t- ^ 3 :^ >^ ^ < 3 c c (0 to '- E r £. to (0 3 to 0) • •H X to • 0)T3 H 0) C — (0 •H a £- a • 0) < u. 3B O 13 • • a c ^ > to > t. £- 3 0) 0) O ^ — - i3 O 0) O) E (0 CO to to "D • C X to u 13 r C -H (0 .- E • t/1 (0 £. • 0) QQ "D •»- t/) r c • to • CO oc E tr o • r • • • < a I- a h- a T3 to T3 T3 0) t. •»- t. £- t- - a O O O O c M- f— M- «4- E •'- - 0)0)0) O)-^ Ji^ ^ C C C C :- t. t. CO to to to to to to ^ ^ ^ L. L. £. CO CO CO 168 z z z z z z z o X X X X X X X z z z z z z z z > X^X X X xxxxxxxxxxxxx r^^^c>JO)U5 0c^lWCD•^ooc^lc^llnc^ln^^^)^^^tOoocoCD•^co0^c^lO)ln0^t^<»'•-lnuoO^-•^^DCNO(00•^c^lK■«-^c^l ^*^*^?^*(jJ^*t^^«•^lI^ln^)^p^pt^t^^*'^*r^oo^*cD^*^Dcooo^*•^*^o^*^*oot^lOc^colnw^*^•a)^*a 0)0)0)0)03(DO}QO)0)0)0)0>CDO)0)a)(n(n(nO)(nO)(n(nO)0)CDO)0>0)QO)(»CD(DCDO)(nO)0)0)0)CDO)CDQ • •»• N • c • T- . 0) • ^» ^ • » • ^ ▼■ • y » w# ' UJ . 0) C ■0 Q) c 0) c (0 0. •5 N c • (Q (/) c tx. • "D E . to 0) o 3 c o Z 0) 3B (0 ^ 3B f- T3 C u. c J^ ^ q: (0 C 13 (0 (0 C D o "0 . E E (0 (Q c £. c- E U • (0 •H (0 c v o u (Q (0 (0 c c to • I O c 0) L. z z • r a ■H r •"" ^"^ £_ • o (0 -3 iJ ' •3 (0 (0 • T3 o 6 • X 2 • ■M 0) . ! 0-3 • S • r t- • • "D (U "o -i -3 • to < • © C > • +J ^ T3 • • c - c ^ c " S3. "0 T3 ■ Q) C • •D Z -H • • c • (Q Ui CO • 0) o o L. £- 5^ c £ s.^ • (D (0 a c c £. . JL c > £. L. »— N - - L. (0 C 13 (0 T3 C . 0) - (D CO • Q) « i: (0 r r O •• - W C ^ Q) • to C t. 10 C -^ • 3 • (0 to > ^ • ^~ (0 (0 I 5> 0) a c »- <0 0) r- (0 J^ < 2 C to 01 "D . (0 QC Oi. c c » (Q r 0) • E I - 4-» . C I u. y. 3t a • - Z C t- • a • »- 3 C Z > • O • CQ I • -< (0 "D • • Q • T3 • • > L. X -D UJ tC • -D • (i: • to • • 3 • X • < •»- •^ "3 -D • • • 2 Q < . C K • C T3 • . . i Q • < 0) • Z ^ • ^ K O • . . 1 • (0 C "3 O < • £_ . U- • "D "D • T3 X U- . X • (0 T3 . © ■ h- "D • U) C3 . • 00 C Z • • o • -H ■«■' T3 • • • C X u. • V ' ■0 "D • . . c • (J • O _ ^ < • (0 • T3 -D • • £- L. -COS • (Q • c • - • t. t • (0 ■ "2 -: U. C U -3 {/) "0 • 0) ai 0^ (0 •3 « . . (0 • • « « • " • - C 2 -T3 • Q. . (0 C • c ' n £\ - X5 • 2 ■ V O (Q 'C- ■D - (0 in • "3 • • (0 •OO'W'C "3C"3 IA)«-D"3 < < • "•••QQQQQQQQQ -O<0 • • • C -UJ UJ • • •aa. ^a)<(0(o x: x} uj E •«- . • C • • • u- ^^ .:.:.;•. ^^ • • -j « > ^a -tn • -a . • .. .^(o . .£.-D-D-3(oa -aa • • LUUJUJUJUJUJUJliJUJ ^U.-' ^ • ' C >T £-£., E-3-30£: uu-^u-u. acL--o<<< aa-.wc5CJ--o-' 3 33 O ..-3 3 0) •"-•-"'-" "^ "3 - • (/) ^ C- E T3 -H T3 " . -3 •jir— - • • ccca'D --^-^ r-" r-' f:^ ,'*«:«:-: S^ 5^ E 5^ ? ^ ^ ^ ^ -h o. - ► « • - o »- w o r o C/)+^4-'f-CC© -000 C>-H4-» - cc -CL00(D(D0(DQ)a)(i)a)aj(i) - .4^ ww . •>wwva)C3)oo ^L(0(Q(aa)cco(O(o "(oocDtDco ■r-.^(oa)©y---t.£-£-j-£.£.t.j-£-rra) -ccot3wccw-c3- •>> •'<>5 -ia^^CC>U-H5E WOUOUUOT3D)0) UiSZ -^ '^ C C L. (oto(0(Q(Qn3(Q(oa)(D(D(Da)(ua)a)OQ)(Da)a)a)^)coo)0)^•^cg^ln^•^^^*cnTfcMlr)0)0^^0^•coOOOI^^•Or^^tDco■^^*00)^Doooooo^Dcot^O) 0>CDQO)0)0)QG)0)0)0)0)0>QO)(nO)0)CDCDa)a90)0)0)a)CDO)CDO)CnQO)0)0)0)QO>0)(nO)Cn(nO)0)QO^ (0 V L. (0 JZ o +^ > a O u. I X Q. • -D " -D U. • LU I to ■o C "0 li- I (0 c (0 T3 ' • • c -D -D QC (0 • c z • c a (0 (0 W -H • O - I a: a. 0) •H C -H O 0) in r 0) •D • C C/) < ■o • c T3 . i, L. C i. 0) 0) (0 0) a a T3 O O L. C O O (0 o o a X O 0) . . o »- -J -J o < ■o • C Ui < (0 > > Q) -^ (0 (0 (0 (Q s s s s c o ^ o o •H N — £- £- "H (0 (Q (Q - - c > > o ^ 0) E D)- E C 3 •'- •^ (0 t. •H O O 3 • I- 0) E • (0 C o C (0 (0 ^ - o < i- T3 • C -J 0) I a • < T3 C • C o E E (0 o (0 o o u o c c Q) 0) T3 "0 (0 (0 O O c c (0 (0 u. u. O O c c T3 "D T3 V (Q (0 u. u. O O -T3 0) o (D 0) O -J O O 5 -D >^ • (0 >Z OQ CO 0^ E -H • • • • I- • (0 -^ uj a a a C3 -D E 0) . . . . C - •£_£-L.£-^ ■D a o cc V Q) c ^ 0) c o • o • •*- o cr 5 • H- -- ■D > Q. O . c . ^ > -3 0) X -a 0) -H EX — '-—Ctv-ECOOO •^•^•^■-oooooo r- • L. (0 Q^ Ul 0) ► • w C O) c (0 C 0) ^ C £. s s s c • (Q CD • £- • QQ 0) O -H • Q - -D OL > Q • •> .- (D 0) - 4^ »- — N ■H > > (0 O O O £- z s z s ►H »H C3 C5 O CO c w (0 3 O '- tC "0 r- J. £_ 3 3 3 > > > a a £- i. i^ L. 3 3 3 z s s 170 (0 in > z o z z X X X X X X X X X ino)otot)30)Or^^^tpi^r^inoco^t^OO)cocNt*»c>ico^Ocoh»^(poo'^coTrtDO-'-tDooc>ii^''-00'^^ooo)^0)c») (nCDO)0)OI)0>0)GDCDCDO)0)0)(nO)0)0)0)0)0)CDO)0)0)0)CDQO)CI)QO)CDO)0)QC>0)COO)0>0)0)Q (U c (0 03 O :j c • O 0) (/) c. J^ I/) 03 C (0 :«: -D - . . Q) a: Of! ^ TJ "D m c c Q tfJ (0 • ecu. a) 0) > > - 4-' -H O t/> (/) 3 (0 • • X CI o O CI o CO a —»-•.- c c 0) L C JC 3 I- (0 • C (Q C CO • -D E • — -D 0) I T3 C • (0 • I E - (0 o — I L. • 3 I p- I (Q 3 O (0 - E - - C r (u (0 Q) CO 12 0) 0) z z L. £. £. Q} 0) J^ ^ ^ 0) 0) 0) XI 13 13 0) 0) z z z •^ 3 15 •■5 3 T3 C c CO > •^ c E C in 0) u c CO • -D 13 I • • C a -D CO • < ^ to -3 ^ - c c w — o o en -- Cfl w 0) ^ ,- ^ 0) 0) 0) 0) c (0 o * . . r- 0) C >-D »- - ON W CO - £ ^ '- a > o o 0) CU 0) — -^ z z z z z • (/) -CD < • < - .a -^ c • •'- o -< ^ en c r r- o - w O CO o — ^ »- t- o O 0) 0)J^ z z z z c > o Q -H (0 • C3) < D • O O O • C O CO I c o - to - "O — 3 (D I O z • o c — CO OH e 4-» • 3 O O O -D • "D (5 C CO « »-* C CD o r. ' Q O I -H •o — c ^ . CO 0) - > o CO :^ E T5 -^ C CO -0 c • CO to - — 0) O '- ^ n f- o z z O CD 0) . o O coo CD ^ r "D -H -H (. £. C z z z 03 to CD • E uj CD CO "0 K • - 5 > - - 13 L. f- - CO CU 3 r- -^ U "D 0)-- O O O O >2 CO • -J C!3 T3 a to o E -H • • • :* 3t < "3 -D to O O - S E - - a-^ — to to E • Q c £- to a"3 - £-t-* --cocDaa TJ cocoj*: -a)CE+^-^-^ '1- •^•^^>o-*-t-cor-^> U z z a z z z Z Z «/) zz. z 01 z z 0) t^ CM OC q: > >- CO z o z in > O XX X XXX X XX XX X X XX X XXX n n n n coo)00)'^ooo)lD'^lnO•^^^oot^O)c^c»^^)0)c^lcn^coo)OT^OCnO)0)QO)0)0)0)0)0)O^QCDO)0)a)0)(^ CQ CD in ui ui t. £- O 0) 0) > 3 5 0) O O £. 0. Q. Q. 13 C (0 C 0) (0 O -H 3 0) (0 -H O (U o a — -J a • a • < X Q o :^ "0 • T3 c > c (0 (0 (0 -3 O • T5 +- C E ® (0 (0 0) 0} 0) Q) T3 — (0 (D r 0) o • a Q£ a r li.I3U) 3<<<<<<<<<'-~3UJUJUiLiJLUUJlAJUJUJ '00 '- -i a i X i I X x X X X 3 a 3s aJ : : 3 3 ^ 3 3 3 3 (0 (0 (0 > r i: ^ o o o o 13 r r- >» > >* (0 CO (0 £.0)CCL.CL.;.L.£.L.(.LO(OiLl.L.L.L.£.L.£.L.L. (l)-r-OO(l>(l}(l)(D(D(l)(l)CD(DC(l}(D0(l>(Da)D)a)D)0)0)0)CnCJ)0)COUOUOOOOOOO (a U O O 3 tX tX QL < > E 0) — > O t. u <0 - X (0 z • X ■D c • <0 X • -D (/> C (0 -H C/> S • O • Q£ a O. a a c a a (0 3 3 > (X CL CL 111 z z z z z z z z > z (J z z z u z z z z ooo)a)'^cocDO'^ooocN'*tinoiir)inh-ocnh*r*0)'^iotnt^ooc>inc»)c>ioo-^Oc>ia)0)t*-Ot^o)0)0)0)ir)iDniniD« O)O)O)a9O)09O)O)CDGDO)O)CDO>O)(n0)O)0)0^CnCDO)O)CDO)CDO)0)O)O)O)CDO)O>CDO)0>CDO)O) < ^ •0 I c • (0 .J • TD • 3 C3 O X 0) o o Q. (0 c • r CJ o c/> c • (a a T3 ? X C CQ QQ (0 < I • • ( 3 ? E 3 •0 o c • f- LU lU UJ • £- - £- 0) -^ > (0 a q: (A) (0 - (0 13 ^ c I/) CO CO W - I- £- £- L. 0) Q) 0) • r JC JZ (0 (0 (0 u o o c/) (/:»(/) to (/> I/) (D 0) r jc u o to (/7 «5 . W N W -H O - -- E u o (D 0) . Ui Ui Ui L. S. £. £. 0) 0) 0) 0) "0 13 U J3 - 0) 0) 0) 0) C C C C x: r r r o u o u 10 en (/) I/) u U 0) "^ -J E • • (0 . . ■D "D "D a o < £-£-£. C T3 X 0) 0) 0) (0 '- T3 "D -0 E 0) - •^ •- -^ M- -^ Ji 0) (D 0) M- ««. (J C C C O O 3 r r r r r x: o o o o o u t/) tn I/) t/) to I/) > C 2 o • (0 o • CQ > c C -3 (0 T3 C (0 • TJ C O ■ Q. flJ q: i •3 O Z XJ -0 < Q) (0 « a a £. (U 0) 0) ji i: r (/> t/) 10 C CO 0) D -H .- ■H 0) Q) - X < -" - C CQ ► 2 J 2 2 J - 3 1- (0 C L. C7 0) E - O Q)._^^ o corrrxir 3-0 (1)'- E E — -^- — -^ i:-----^EEEE£ t/>C/1(/)t/)t/)tO(/)l/)(/)C/)l/) 0) (D 3 3 i: r c/) tn 173 z z z z z z o > z u z > C/) u z z z z 00 ^ z 1 z 1 01 1 t/) 2 o u z 00 in z a z > z z XX X X X XXXXXX XX xxxx jQ (0 cDOO^ooOc>ic>ir^ooo)OTOOr^Oc>ir)(^^oiif)c>iooOcoco'^"^"^cr)oot*»CDO)if)Oiooioina>c3)'^"^OtD^ooo)coo lX)(i)^)(plD^<*^^^^^«.^*(JP^lnt^^*oo^*^'lnln^*^*co^*oo^*^*•^oo^»^*cocn':fr^^«.oo(^ O)O)O)QCDO)O)O)QO)CDQO)O)O)O)CDO)O)O)O)CDQ(nQO)O>QCDO)O)0)O)O)CDQO)CD(nO)O)O)QO)O)CDQO)O)O)Q(^ L.-0 i. O I X C < C -J I 3S -J X ■♦-• • -H • S-Ji/)t/)t/>t/)X(/)-J xf r r r r r r r r E E E E e B C/7 t/> to E E E (/) U) t/) S > ' c c - S- o o r Q) E E •H -0 o o •»- >^ r- E C O O i/l t/> c/) (/) • c (I) o c o > • ■ TJ a z c 0) * OJ • • o 0) . . O (D (3 -D 5 T5 • C (. (0 "3 (U ►< < 0)0 ? c - O <0 £- t- X 0) U - X I- i. i^ L. 0)0)0 O) D O) C C C (0 (0 (0 a a a CO to l/> L. Of O Q • O (/) (1) - O 0) -D-D - £- 1- - 1. .> . 0) r- (I) •H 4J D) 0) •'- ■H +^ C 0)0. L- t. t. £- • a a a a-n to to (/) to (/) o • a • o Q c o • 0) • "3 t/) 0) ► > w •^ to 0) c O. O)-- £- 0) c <0 • (0 (0 0) ■H -H "H +^ t/) (/)(/) t/) c • jC < o -D • c - o c E O 0) (0 O) c 0) 0) U 0) • o O 5 •0 • c o: 13 JZ X en u o 0) • L. t. £- ^ 0) -O 3 0) z jQ «♦- r — J- • ^J (0 (0 u -^ 3 ^ E -to • • > X uj • . . o U X u- . -J •O "O • c c a -0 (0 (Q C - (0 o 0) c u 1- t. (0 (0 • 13 -D C (0 •o c • (0 (X, > > -J t. . • • >«i I 0) < to to . . . J3 CO O • Q Q X O • • • -I .>^.. .^..QQ a^xx coo-saoi-o ... 0- 3 -a)D)C-'--0 W+^ C:-lL(0>CO)0)Ji CJ-t-t- ->T3T3(0O3:3Ea)O'-'-E a)5DI3C£.00(aCT3X5a)a)CCC£- a)Q)Oooo£.t.c(0Q)Q)a)0)0)-'- — -M •H4^-H4-'4-'4-'-H+J-H4-'335335333N tOt/)t/)t/)t/)t/)t/)t/)tOt/)tOt/)t/)t/)t/)tOt/)t/)t/)l/> 0) C K (0 Q • E - ^ t. 0) r c ^ c U 3 C • X •,- o • • •H a Q • L. ' -J - - (0 - c >j^ t. 0) <0 J3 0) t. E «^ C i- (0 «*- (0 0) r •'- -^ K I- H- >- H 174 o ^• CM CM a. Z q: > Z z i/i z (/> (J z z in 0^ z t/) u z z z CM in z q: 1 > 1 tn u o 1 z z z X X X XXXX XXXX XXXX XX xxxx 13 0)0)J)CDO)0)CDO>0)0)QO)0)0)0)0)0)0)0)0)0)CDQCDO)CDO)QO)0)Q 0) (D > (0 o C C C (0 (0 (Q '- E o D m £. 3c • (/) 5 • -o • • c -o 3 (0 C • (0 -3 • 3f • - K a — LU - W - W C (0 - C -^ (0 £- O ^ — 0) E aJD E E E C (Q -00 K K H- I- I a t. i: c O £- • c o < •• c c > (1) 0) C +^ 4-» c w to 0) o o Q.O O c £- C C 13 (0 <0 h- > > C C Z. 0) 0) O tn OT »- o o 0) o o > c. c c. (0 (0 (0 > > > "D C 0) 4-» • ' CO s 13 . 0) - • : 3 0) Q -> • — < C C Ji ^ 0) Q) 3 Q U c c c (Q (0 (0 > > > o 0) • C O • c uj (0 r- CO *%• e o <0 I ^■^ E r o r- (0 — ( D (0 <0 >- Q. ' -s < 3 uan z 3 . • • I- Q Q < C C C C m (0 (Q (0 i: i: i: E O) O) O) o D 3 D O (0 (Q (Q £. > > > > ♦-» II 3 • • T3 -D-D Q) - C C - - O 0) CO (0 O) C (D 0) O) O)— — (0 (0 (Q (0 3 :? 3 3 c (0 • I UJ c c • o L. £. CO 0) L. jQ r- (U — UiTi - c c 1 CO OJ ail. • tn t: cj T3 c C CO • CO o -D -D . t- o 0) (0 co ^ "D CO to CO * 3 ^ - £- L. L. L. (0 H CD CD CD CD r * yj +j 4^ +j C to -H r- »- ^ — C- £- +J CO CO CQ to (0 CO (0 3 3: 3B 3 :» 3 3 c/) "D CO : c O I o i V c CO > O CD c 1o -J c o •- Z CO • o -J 73 • C < C to •0 c CO -J 1 1 • n CD Q O 13 • c - to •D—a-.J-l-l -TJCOtO CD E 0) • L. C C -4-»(0E - - - tocoto £-tOjiOtOtOtOtOO)— »- CD C-H O'-r-'-^-Hr- — — •.-.^-^1^-.-,-,- CO — — tDCDtDQ) CD CDtDCDCDrr 3333 3K3:»3:^:^3 3 3 C C 3 to to > CD CD ::<. C (0 5 3 " - » - 0) "^ CD CD 0) CD &. r- 4J 4-» +j +^ J^ - £1 O r - — :5 5 3 r r r :* 3 26 175 ^h.a)0) 13 • Q I O 3B C C Ql -M -I C O . 0) O LU -a -H - to •H • C £- 0) > O (U js: 0) w -M o '- — c 5 5 5 5 - O 3 3t • T5 • -I C O (0 - > • 0) X3 ► «♦- 73 TJ --0 3: :2 ^ ^ OQ QQ •H 3: 3t O O — Q) O •^ a V O • O) O H- ^ o • -J 3 T3 • C • liJ (0 C3 XJ • •0 C Q C (0 (0 E • (0 • • (/) — u. -J — u . . a ^ . 0^ U ? Ill +- +^ ^ w c c c r r w C3) to c C O) 0) o c E 0) 0) -^ — -D 3 0) L. £-£.£.> O Q) 3533^3>-Nfsj 176