Epidermal club cells do not protect fathead minnows against trematode cercariae: a test of the anti-parasite hypothesis

Transcription

1 Biological Journal of the Linnean Society, 2009, 98, With 2 figures Epidermal club cells do not protect fathead minnows against trematode cercariae: a test of the anti-parasite hypothesis CLAYTON T. JAMES 1, BRIAN D. WISENDEN 2 and CAMERON P. GOATER 1 * 1 Department of Biological Sciences, University of Lethbridge, 4401 University Drive, Lethbridge, Alberta, T1K 3M4, Canada 2 Biosciences Department, Minnesota State University Moorhead, th Avenue South, Moorhead, MN 56560, USA Received 22 April 2009; accepted for publication 16 June 2009bij_1327 Epidermal club cells of fishes in the superorder ostariophysi have puzzled evolutionary biologists because they were historically linked to chemical alarm signalling and relied on group selectionist explanations. Alternative hypotheses to explain the existence of these cells include the possibility of an anti-pathogenic or anti-parasitic function. If this is so, individual fish should invest in increased numbers of club cells after exposure to parasites, and club cells should contain components that reduce the infectivity of skin-penetrating larvae. Infectivity of cercariae of the trematode Ornithodiplostomum sp. was significantly reduced when exposed to the skin extract of fathead minnows (an ostariophysan), but also to skin extract of mollies (a non-ostariophysan that lacks club cells), respectively, compared to controls. Moreover, club cell density was not affected by exposure to cercariae. Taken together, these results are inconsistent with an anti-parasite function for these cells and instead suggest a generic role in response to injury.. ADDITIONAL KEYWORDS: alarm substance anti-predation Ornithodiplostomum Ostariophysi infectivity Pimephales. INTRODUCTION Given the extent and variety of costs that parasites impose on individual hosts, parasite-mediated selection should favour host defences that limit exposure to infective stages, or limit their negative effects (Poulin, 2007). The detection and/or avoidance of parasites appears to be limited to conspicuous and pathogenic infective stages, especially for aquatic hosts (Wisenden, Goater & James, 2009). Furthermore, immune defences are expressed at the expense of growth and reproduction, often with important fitness consequences (Zuk & Stoehr, 2002). Thus, natural selection may favour anti-parasite strategies that defend against parasites at the point of host *Corresponding author. cam.goater@uleth.ca Present address: Alberta Conservation Association, Box , Peace River, Alberta, T8S 1T4, Canada parasite contact or entry into the host, reducing the associated costs of parasites at the site of infection. In fish, the epidermis represents an initial site for complex immune responses against water borne parasites and pathogens (Whitear, 1986; Kearn, 1999). Immunological components, such as lymphocytes, macrophages, and several types of granulocytes found in fish mucus (produced by goblet cells located in the epidermis), have been shown to reduce the infectivity, growth, and reproduction of various parasites at the site of infection (Shephard, 1994). Another cellular component of fish epidermis that has been linked to host immunity comprise the epidermal club cells of fishes in the superorder ostariophysi (Chivers et al., 2007; Halbgewachs et al., 2009). Epidermal club cells that contain the chemical alarm cue known as Schrekstoff are a defining characteristic of the ostariophysi (Pfeiffer, 1977). It is now well established that damaged skin serves as a 884

2 ANTI-PARASITE FUNCTION OF CLUB CELLS 885 mechanism for risk assessment by fish (Chivers & Smith, 1998). These cells pose a metabolic cost (Wisenden & Smith, 1997), although the mechanism by which individuals that invest in these cells recover fitness benefits is not clear because the cells release their contents only when ruptured by a predator. The lack of obvious fitness benefits to justify investment in epidermal club cells poses a challenge for evolutionary theory (Williams, 1992). One idea that neatly resolves the evolutionary enigma is the possibility that epidermal club cells contain anti-pathogenic or anti-parasitic agents (Smith, 1992; Magurran, Irving & Henderson, 1996). Recently, Chivers et al. (2007) provided several lines of empirical evidence in support of the anti-parasite hypothesis. Ostariophysan fish are frequently infected with the resting stage of aquatic trematodes, known as metacercariae. These encyst on and in a range of species-specific tissues, typically awaiting ingestion by an appropriate final host. Two general arguments lead to the prediction that club cells may deter trematode infections of fish. First, most cercariae of fishes actively penetrate the host s epidermis. Although cercariae vary widely in size across species, most are large relative to club cells and, as such, likely pass through or disrupt these cells during the penetration process. Second, the putative active chemical in ostariophysan club cells that elicits alarm reactions is a nitrogen oxide (NO) side group, such as the one contained in hypoxanthine- 3-N-oxide (Pfeiffer et al., 1985; Brown et al., 2000). NO plays an important role in host resistance to the free-living stages of some helminth parasites (James, 1991). For example, NO has been shown to inhibit metabolic activity of the schistosomula of Schistosoma mansoni (James & Hibbs, 1990). It is conceivable that NO in club cells may play a similar anti-cercarial role in the epidermis of fish. In experimental studies, Chivers et al. (2007) showed that minnows exposed to cercariae had a higher density of epidermal club cells compared to unexposed controls. Furthermore, Poulin, Marcogliese & McLaughlin (1999) showed that rainbow trout exposed to trematode cercariae induced a fright reaction in naïve, unexposed conspecifics. This finding implies that trematode cercariae rupture club cells when penetrating the host epidermis and ultimately come in contact with the contents of these cells. However, whether these results are generalizable to other host/parasite systems remains unknown. In the present study, we tested two predictions of the anti-parasite hypothesis. First, on the basis of the study by Chivers et al. (2007), we predicted that exposure to cercariae should result in increased investment in epidermal club cells. Second, if club cells contain anti-parasite properties, we predicted that immersion of cercariae in skin extract should reduce cercariae infectivity. MATERIAL AND METHODS HOST/PARASITE SYSTEM Interactions between fathead minnows and metacercariae of trematodes in the genus Ornithodiplostomum have been studied extensively in one of our laboratories (C.P.G.). One advantage of this minnow/ trematode interaction is that the life-cycles of Ornithodiplostomum spp. can be maintained in the laboratory using the F 1 generation of field-collected snails (Physa spp., first intermediate host), uninfected fathead minnows (second intermediate host), and domestic chickens (surrogate final host) as experimental hosts (Sandland & Goater, 2000; James et al., 2008). In these experiments, we exposed parasite-naïve fathead minnows to cercariae of Ornithodiplostomum sp. This unidentified species, and its congener Ornithodiplostomum ptychocheilus, are the two most common and abundant trematodes in populations of fathead minnows in Alberta, typically infecting all fish in a population with up to several hundred encysted metacercariae (Sandland, Goater & Danlychuk, 2001). Subsequent to penetration of the epidermis of minnows, metacercariae of Ornithodiplostomum sp. encyst in the liver and body cavity, where they remain for the remainder of the life of the host, or until they are ingested by appropriate fisheating birds. We used Ornithodiplostomum sp. in the present study partly as a result of the availability of experimentally infected Physa gyrina that consistently released large numbers of cercariae, and also because our earlier work had indicated a strong negative affect on growth of experimentally-infected minnows (James et al., 2008). EXPERIMENT 1: EFFECT OF FISH SKIN EXTRACT ON CERCARIAE INFECTIVITY Fathead minnows (30 days old) used in the experiment were obtained from the US EPA National Health and Environmental Effects Research Laboratory (Duluth, MN, USA) on 19 July The fish were held in four separate 190-L tanks for 3 days before the experiment and fed ad libitum on Tetramin flake food. The overall aim of the present study was to evaluate the infectivity of known-aged cercariae that had been immersed in one of: (1) water containing blank dechlorinated tap water (control); (2) water containing minnow skin extract (containing ostariophysan club cells and other skin components); or (3) water containing molly skin extract (generic fish skin

3 886 C. T. JAMES ET AL. lacking ostariophysan club cells). Although the epidermis of some non-ostariophysan fishes such as Poecilids are known to contain alarm substances (Smith, 1982), and some non-ostariophysans also contain club cells (Chivers et al., 2007), the present study was designed to control for the anti-parasite properties of generic fish skin that are not attributable to ostariophysan club cells. Ornithodiplostomum sp. cercariae were obtained in accordance with the methods of Sandland & Goater (2000) for O. ptychocheilus. Eight, 1-day-old chickens were fed metacercariae from the viscera of wild fathead minnows collected in June from Gold Spring Lake, southern Alberta, Canada ( N, W). Trematode eggs were collected from chick faeces 5 days post-infection. The resulting miricidia were pipetted into small vials containing laboratory-reared juvenile pond snails, Physa gyrina, measuring 3 5 mm in maximum length. Snails began to release cercariae at approximately 28 days postexposure. To obtain cercariae, 12 infected snails were placed into a flask for 3 h. The numbers of cercariae in three, 1-mL aliquots were counted and averaged to estimate the total numbers of cercariae released over 3 h. We used this estimate to evaluate the volume of water containing 100 cercariae. Skin extract was prepared from wild adult female fathead minnows (mean ± SE = 5.74 ± 0.08 cm, N = 55) collected from Deming Lake (MN, USA) ( N, W) on 11 July Mollies were obtained from a commercial supplier and, upon arrival in the laboratory, were immediately euthanized to prepare skin extract. Each fish was euthanized by cervical dislocation (MSUM IACUC protocol 07-R-009-N-N-C/D). Skin fillets were removed from both sides of the fish, measured for area and then immersed in 100 ml of chilled dechlorinated tap water. An area of of minnow skin (72.15 cm 2 ) was collected and homogenized using a hand blender and filtered through polyester fibre. This solution was diluted to 720 ml with dechlorinated water. Similarly, we collected cm 2 of molly skin, blended it with a hand blender and diluted with dechlorinated tap water to 700 ml. Thus, the concentration of both skin extract solutions comprised 2 cm 2 of skin per 20-mL dose. Twenty-mililitre aliquots of each type of skin extract and dechlorinated water (control treatment) were pipetted separately into 118-mL plastic containers (N = 35 containers per treatment) with sealable lids and frozen at -20 C until needed. The experiment was set up as a single factor analysis of variance (ANOVA) with three levels of cue type: water control, fathead minnow extract, and molly skin extract. On the day of exposure, the containers were removed from the freezer and thawed. One hundred cercariae were pipetted directly into each container and, 30 min later, minnows were added. Containers were arranged so that all treatments were evenly distributed spatially and minnows were assigned at random to containers. After exposure, all minnows were removed from the Petri plates, and separated into 37-L tanks based on treatment. Minnows were fed Tetramin flake food daily for 5 weeks under a 12 : 12 dark/light cycle to allow metacercariae time to develop into the encysted stage (Sandland & Goater, 2000). Minnows were euthanized with an overdose of methane tricaine sulphonate (MS222), dissected, and the numbers of metacercariae in the body cavity counted using standard methods. EXPERIMENT 2: EFFECT OF CERCARIAE EXPOSURE ON THE DENSITY OF EPIDERMAL CLUB CELLS AND EPIDERMAL THICKNESS This experiment tested the prediction that individual fish should invest in more club cells following exposure to cercariae. The source of fish, snails, and cercariae was the same as that described for Experiment 1. The exposure protocol was also the same. A total of 90, 30-day-old minnows (average length ± SE, ± 0.22 mm) were divided into three groups of 30. One group (high-dose) was exposed three times to batches of 50, 2-h-old cercariae every 2 days commencing on 30 July 2007 and ending on 3 August A second group (low dose) was exposed on the same days to ten cercariae; the third group (control) received water only. Fish were assigned at random to Petri dishes that contained 40 ml of water with zero, ten or 50 2-h-old cercariae. The exposure period was 2 h. On the last day of the experiment (i.e. 3 days after the third and final exposure), 20 fish from each treatment were euthanized and a section of tissue from the nape region was removed from each minnow, preserved in 10% buffered formalin and prepared for histological examination. The remaining ten fish in each treatment were maintained on flake food for 4 weeks to allow assessment of the success of metacercariae encystment. Tissue samples were sent to the North Dakota State University Veterinary Diagnostic Laboratory for histological preparation. For each epidermal tissue sample, three paraffin-imbedded sections (7 mm thick) were removed and stained with Periodic Acid Schiff s reagent and then counterstained with Lillie s haematoxylin. Three sections of each tissue sample were evaluated for the number of club cells and epidermal thickness. Epidermal thickness was determined as the distance between the basement membrane of the epidermis and the outer surface of the epithelium (Wisenden & Smith, 1997). The three measurements of club cell density and epidermal

4 ANTI-PARASITE FUNCTION OF CLUB CELLS 887 Metacercariae recovered Club cell density per mm Con Low High 10 0 Water Molly Minnow Epidermal thickness (mm) Figure 1. Mean ± SE number of encysted metacercariae in minnows exposed to: (1) cercariae immersed in dechlorinated water control; (2) cercariae immersed in skin extract of non-ostariophysan mollies; and (3) cercariae immersed in skin extract of the ostariophysan fathead minnow. Identical letters above bars indicate no significant difference (Tukey test, P < 0.05). thickness were averaged for each section to provide an estimate of an individual s investment in club cells. The observer was blind to all treatments. STATISTICAL ANALYSIS Data were tested for normality prior to analyses using Kilmogorov Smirnov tests. Differences among host lengths for all experiments were compared using one-way ANOVAs. The effect of cue type on metacercariae intensity in the first experiment was analysed using an analysis of covariance (ANCOVA) with minnow length as a covariate. For the second experiment, alarm cell density and epidermal thickness were compared with one-way ANCOVAs with host length as the covariate. Where appropriate, post-hoc comparisons of means between two samples were performed using Tukey tests. RESULTS Eighty-nine percent of 105 minnows in the first experiment survived to 4 weeks post-infection, with minnows represented for each treatment. All surviving minnows were infected with encysted metacercariae. Mean host length did not differ among the three treatment groups (one-way ANOVA: F 2, 88 = 0.61; P = 0.55). The mean number of metacercariae was significantly affected by cue type (ANOVA: F 2,78 = , P < 0.001, Fig. 1). Water controls had significantly more encysted metacercariae than either Figure 2. Number of epidermal club cells per millimeter in the nape region as a function of epidermal thickness for fathead minnows exposed to: (1) 0 cercariae; (2) 30 cercariae; and (3) 150 cercariae. Con, control. of the skin extract treatments, although there was no difference between the two skin extract groups (Tukey test, P < 0.05). The average total length of minnows in the second experiment did not differ among groups (one-way ANOVA: F 1,57 = 0.39, P = 0.68) and host length was not significantly correlated with club cell density (F 1,56 = 2.35, P = 0.13) or epidermal thickness (F 1,56 = 3.39, P = 0.07) and was dropped from the analyses. All fish exposed to cercariae were infected with metacercariae after the 4-week development period (Fig. 2). Cercarial dose did not significantly affect mean club cell density (one-way ANCOVA: F 2,56 = 0.37, P = 0.691) or epidermal thickness (F 2,56 = 0.19, P = 0.828). There was no effect of cue treatment on club cell density when epidermal thickness was used as a covariate (ANCOVA cue type: F 2,54 = 1.36, P = 0.266; epidermal thickness: F 1,54 = 52.98, P < 0.001, cue type epidermal thickness: F 2,54 = 1.18, P = 0.316; Fig. 2). DISCUSSION The results from both experiments were inconsistent with the predictions of the anti-parasite hypothesis. Although immersion of cercariae in water containing skin components of fathead minnows led to markedly reduced infectivity, a similar reduction occurred for cercariae immersed in skin components of nonostariophysan mollies. Thus, a component of fish skin extract reduces cercarial infectivity, although that component is likely not associated with the contents of club cells. Furthermore, fathead minnows did not alter their investment in club cells after repeated exposure to Ornithodiplostomum sp. cercariae over

5 888 C. T. JAMES ET AL. 6 days. These results indicate that club cells do not provide specific protection against these skinpenetrating parasites. Diplostomules of Ornithodiplostomum sp. traverse the epidermis to reach their ultimate site of encystment. Thus, they must come into contact with various components of the epidermis in addition to club cells. Populations of lymphocytes, macrophages, and granulocytes have been identified in the fish epidermis (Whitear, 1986). These secretions, alone and in combination, could play a direct role in reducing the infectivity of Ornithodiplostomum sp. diplostomules. Mucus or goblet cells are also known to secrete a variety of anti-parasite compounds, such as lysozymes, lectins, and proteolytic enzymes (Buchmann & Brescianni, 1998), which have been shown to protect fish from parasitic fungi, bacteria, and monogenean trematodes. Fathead minnow skin contains large numbers of mucus cells (Wisenden & Smith, 1997), although their functional role as a defense against parasites has not been evaluated. Because all fish possess goblet cells within their epidermis (Shephard, 1994), components in fish mucus most likely serve nonspecific anti-parasitic functions that may interfere with cercariae during attachment to the host or penetration through the epidermis (Haas, 1994). An alternative explanation for the observed reduction in cercarial infectivity is that a component of skin extract interferes with the normal host-finding, hostcontacting, and/or migration behaviors of cercariae and diplostomules. In vivo and in vitro experimental studies involving a closely-related trematode, Diplostomum spathaceum, indicate that the process of cercariae penetration and diplostomule migration within fish intermediate hosts is highly complex (Haas et al., 2007). For example, migration of individual diplostomules from the point of penetration into subdermal tissues, and ultimately into peripheral veins, requires navigation along concentration gradients of chloride ions, glucose, and glucosamine. We expect that similar complex migratory mechanisms occur for Ornithodiplostomum sp. diplostomules as they traverse the epidermis. Thus, the skin extracts used in the present study may have disoriented cercariae and/or diplostomules by interfering with the detection of navigation cues that would be present in the intact epidermis. The results obtained in Experiment 2 contrast those previously reported by Chivers et al. (2007), involving cercariae of another species of trematode. In their study, the density of epidermal club cells in fathead minnows repeatedly exposed to cercariae of the turtle trematode Teleorchis was significantly higher than those exposed to water controls. Several factors are known to affect the rate of production of club cells, which, alone or in combination, could explain these contrasting results. Minor differences in the time course of the experiment, in host age (Carreau-Green et al., 2008), and in host condition (Wisenden & Smith, 1997) could have contributed to the lack of a skin response in the present study. James et al. (2008) showed that fathead minnows infected with Ornithodiplostomum sp. lost weight relative to controls, suggesting a metabolic cost as the cercariae develop and metamorphose into fullyencysted metacercariae. Loss of body condition as a result of developing metacercariae also occurred in European minnows that were experimentally exposed to cercariae of the brain-encysting trematode, Diplostomum phoxini (Ballabeni & Ward, 1993). A developmental cost of this type is unlikely for Teleorchis-exposed minnows because the metacercariae of these reptile trematodes are unlikely to develop beyond the penetration phase. One intriguing possibility for the contrasting results obtained in the two studies is that club cell proliferation, and perhaps other forms of epidermal immunity, may be most detectable in hosts exposed to novel pathogens and parasites. Supportive evidence for this hypothesis would include the identification of mechanisms used by specialist parasites such as Ornithodiplostomum sp. that migrate through the epidermis without direct damage to club cells. Migrating schistostomules of Schistosoma spp. navigate between cells in the epidermis of experimentally exposed mice on their way to blood vessels located within the dermis (McKerrow & Slater, 2002). By contrast, the schistosomules of Trichobilharzia, a common blood fluke of water birds, cause localized cellular and tissue damage in the epidermis of nonhost mammals that provokes a generalized host immune response leading to cercarial dermatitis, or swimmers itch (Bahgat & Ruppel, 2002). One extension of the novelty hypothesis is that strong and consistent alarm cell responses may only exist for those parasites or pathogens that remain for extended periods in the epidermis of fish, such as some pathogenic water molds (e.g. Saprolegnia) or those metacercariae that cause black spot in the epidermis of their intermediate hosts (Chivers et al., 2007). The data obtained in the present study add to the hypothesis that epidermal club cells of the ostariophysi are part of a complex, generalized response system that confronts pathogens and parasites such as cercariae (especially dead ones in the case or Teleorchis; Chivers et al., 2007), fungal hyphae (Saprolegnia; Chivers et al., 2007), or bacteria in the case of secondary bacterial infection after epidermal damage from mechanical abrasion or exposure to ultraviolet light. Thus, club cells in both ostariophysian and perciformes (perch and darters) fishes likely serve this function (Chivers et al., 2007). The data

6 ANTI-PARASITE FUNCTION OF CLUB CELLS 889 obtained in the present study signal the need for additional studies on the suite of epidermal barriers and defences against parasite attack. ACKNOWLEDGEMENTS Funding for this research was provided by an operating grant to C.P.G. from the Natural Sciences and Engineering Research Council of Canada and faculty research grants to B.D.W. from the Minnesota State University Moorhead College of Social and Natural Sciences. REFERENCES Bahgat M, Ruppel A Biochemical comparison of the serine protease (elastase) activities in cercarial secretions from Trichobilharzia ocellata and Schistosoma mansoni. Parasitology Research 88: Ballabeni P, Ward PI Local adaptation of the trematode Diplostomum phoxini to the European minnow, Phoxinus phoxinus, its second intermediate host. Functional Ecology 7: Brown GE, Adrian JC, Smyth E, Leet H, Brennan S Ostariophysan alarm pheromones: laboratory and field tests of the functional significance of nitrogen oxides. Journal of Chemical Ecology 26: Buchmann K, Brescianni J Microenvironment of Gyrodactylus derjvini on rainbow trout Oncorhynchus mykiss: association between mucous cell density in skin and site selection. Parasitology Research 84: Carreau-Green ND, Mirza RA, Martinez ML, Pyle GG The ontogeny of chemically mediated antipredator responses of fathead minnows Pimephales promelas. Journal of Fish Biology 73: Chivers DP, Smith RJF Chemical alarm signalling in aquatic predator-prey systems: a review and prospectus. Écoscience 5: Chivers DP, Wisenden BD, Hindman CJ, Michalak TA, Kusch RC, Kaminskyj SGW, Jack KL, Ferrari MCO, Pollock RJ, Halbgewachs CF, Pollock MS, Alemadi S, James CT, Savaloja RK, Goater CP, Corwin A, Mirza RS, Kiesecker JM, Brown GE, Krone PH, Blaustein AR, Mathis A Epidermal alarm substance cells of fishes maintained by non-alarm functions: possible defence against pathogens, parasites and UVB radiation. Proceedings of the Royal Society of London Series B, Biological Sciences 274: Haas W Physiological analyses of host-finding behaviour in trematode cercariae: adaptations for transmission success. Parasitology 109: S15 S29. Haas W, Wulff C, Grabe K, Meyer V, Haeberlein S Navigation within host tissues: cues for orientation of Diplostomum spathaeceum (Trematoda) in fish towards veins, head, and eye. Parasitology 134: Halbgewachs CF, Marchant TA, Kusch RC, Chivers DP Epidermal club cells and the innate immune system of minnows. Biological Journal of the Linnean Society 98: James SL The effector function of nitrogen oxides in host defense against parasites. Experimental Parasitology 73: James SL, Hibbs JB The role of nitrogen oxides as effector molecules of parasitic killing. Parasitology Today 6: James CT, Noyes KJ, Stumbo AD, Wisenden BD, Goater CP Cost of exposure to trematode cercariae and learned recognition and avoidance of parasite risk by fathead minnows, Pimephales promelas. Journal of Fish Biology 73: Kearn GC The survival of monogenean (Platyhelminth) parasites on fish skin. Parasitology 119: McKerrow JH, Slater J Invasion of skin by Schistosome cercariae. Trends in Parasitology 18: Magurran AE, Irving PW, Henderson PA Is there a fish alarm pheromone? A wild study and critique. Proceedings of the Royal Society of London Series B, Biological Sciences 263: Pfeiffer W The distribution of fright reaction and alarm substance cells in fishes. Copeia 1977: Pfeiffer W, Riegelbauer G, Meier G, Scheibler B Effect of hypoxanthine-3(n)-oxide and hypoxanthine-1(n)- oxide on central nervous excitation of the black tetra Gymnocorymbus ternetzi (Characidae, Ostariophysi, Pisces) indicated by dorsal light response. Journal of Chemical Ecology 11: Poulin R Evolutionary ecology of parasites, 2nd edn. Princeton, NJ: Princeton University Press. Poulin R, Marcogliese DJ, McLaughlin JD Skinpenetrating parasites and the release of alarm substances in juvenile rainbow trout. Journal of Fish Biology 55: Sandland GJ, Goater CP Development and intensity dependence of Ornithodiplostomum ptychocheilus metacercariae in fathead minnows (Pimephales promelas). Journal of Parasitology 86: Sandland GJ, Goater CP, Danlychuk AJ Population dynamics of Ornithodiplostomum ptychocheilus metacercariae in fathead minnows (Pimephales promelas) from four northern-alberta lakes. Journal of Parasitology 87: Shephard KL Functions of fish mucus. Reviews in Fish Biology and Fisheries 4: Smith RJF Reaction of Percina nigrofasciata, Ammocrypti beani, and Etheostoma swaini (Percidae, Pisces) to conspecific and intergeneric skin extracts. Canadian Journal of Zoology 17: Smith RJF Alarm signals in fishes. Reviews in Fish Biology and Fisheries 2: Whitear M The skin of fishes including cyclostomes epidermis. In: Bereiter-Hahn J, Matoltsy AG, Richards KS, eds. Biology of the integument. Heidelberg: Springer-Verlag, Williams GC Natural selection: domains, levels, and challenges. New York, NY: Oxford University Press.

7 890 C. T. JAMES ET AL. Wisenden BD, Smith RJF The effect of physical condition and shoalmate familiarity on the proliferation of alarm substance cells in the epidermis of fathead minnows. Journal of Fish Biology 50: Wisenden BD, Goater CP, James CT Behavioral defenses against parasites and pathogens. In: Zaccone C, Perriere A, Mathis A, Kapoor G, eds. Fish defenses, Vol. 2. Enfield: Science Publishers, Zuk M, Stoehr AM Immune defense and life history. American Naturalist 160: 9 22.