Effects of Emerging Infectious Diseases on Amphibians: A Review of Experimental Studies

Numerous factors are contributing to the loss of biodiversity. These include complex effects of multiple abiotic and biotic stressors that may drive population losses. These losses are especially illustrated by amphibians, whose populations are declining worldwide. The causes of amphibian population declines are multifaceted and context-dependent. One major factor affecting amphibian populations is emerging infectious disease. Several pathogens and their associated diseases are especially significant contributors to amphibian population declines. These include the fungi Batrachochytrium dendrobatidis and B. salamandrivorans, and ranaviruses. In this review, we assess the effects of these three pathogens on amphibian hosts as found through experimental studies. Such studies offer valuable insights to the causal factors underpinning broad patterns reported through observational studies. We summarize key findings from experimental studies in the laboratory, in mesocosms, and from the field. We also summarize experiments that explore the interactive effects of these pathogens with other contributors of amphibian population declines. Though well-designed experimental studies are critical for understanding the impacts of disease, inconsistencies in experimental methodologies limit our ability to form comparisons and conclusions. Studies of the three pathogens we focus on show that host susceptibility varies with such factors as species, host age, life history stage, population and biotic (e.g., presence of competitors, predators) and abiotic conditions (e.g., temperature, presence of contaminants), as well as the strain and dose of the pathogen, to which hosts are exposed. Our findings suggest the importance of implementing standard protocols and reporting for experimental studies of amphibian disease.


Introduction
Rapid rates of biodiversity loss have supported the notion that the Earth is heading toward a sixth major extinction event [1][2][3].Current species extinction rates are higher than pre-human background rates, suggesting this biodiversity crisis is largely attributed to anthropogenic changes [1][2][3][4][5][6].Although numerous species from all taxonomic groups are affected, amphibians are at the forefront of this crisis [3,7,8].Their populations are declining more rapidly than those of birds or mammals [8].Like other groups, amphibians are affected by multiple factors contributing to population declines [9].These include habitat destruction, contaminants, climate change, over-harvesting, invasive species, predation, and infectious diseases, all of which may work independently or synergistically to affect amphibian populations [9][10][11][12] (Figure 1).Some of the research we summarize below focused on how a particular pathogen alone affects a host, whereas some studies addressed how a pathogen may be affected by other variables that may interact with pathogens.tion rates of biodiversity loss have supported the notion that the Earth is heading toward a sixth major extinction event [1][2][3].Current species igher than pre-human background rates, suggesting this biodiversity crisis is largely attributed to anthropogenic changes [1][2][3][4][5][6].species from all taxonomic groups are affected, amphibians are at the forefront of this crisis [3,7,8].Their populations are declining mo of birds or mammals [8].Like other groups, amphibians are affected by multiple factors contributing to population declines [9].The truction, contaminants, climate change, over-harvesting, invasive species, predation, and infectious diseases, all of which may work inde tically to affect amphibian populations [9][10][11][12] (Figure 1).Some of the research we summarize below focused on how a particular patho st, whereas some studies addressed how a pathogen may be affected by other variables that may interact with pathogens.
1. Potential abiotic and biotic factors that may influence host-pathogen dynamics in amphibian disease systems.
g the major threats to amphibians are emerging infectious diseases (EIDs).Several prominent pathogens and associated EIDs affect a s worldwide.Batrachochytrium dendrobatidis (hereafter referred to as Bd) is a pathogenic fungus that causes amphibian chytridiomycos Among the major threats to amphibians are emerging infectious diseases (EIDs).Several prominent pathogens and associated EIDs affect amphibian populations worldwide.Batrachochytrium dendrobatidis (hereafter referred to as Bd) is a pathogenic fungus that causes amphibian chytridiomycosis [13][14][15].This disease can cause population declines, local extinctions and contribute to species extinctions [8,16,17].A related yet highly divergent fungal pathogen that also causes amphibian chytridiomycosis, Batrachochytrium salamandrivorans (hereafter referred to as Bsal), is a newly discovered pathogen primarily infecting salamanders [18].Iridoviruses of the genus Ranavirus (hereafter referred to as Rv) have been implicated in declines and mass mortalities of amphibians [19][20][21][22][23]. Teacher et al. [22] stated that populations can respond differently to the virus and emergence can be transient, catastrophic, or persistent with recurrent mortality events.Although amphibians are hosts to an assortment of pathogens/parasites, including bacteria, viruses, fungi, water molds and helminths [13,[24][25][26][27], we focus on Bd, Bsal and Rv, given accumulating evidence of their potentially devastating effects on amphibian populations worldwide.In particular, we focus on reviewing the literature that report the results of experiments (manipulation of key variables [28]) conducted with Bd, Bsal, and Rv concentrating on papers that used live amphibian hosts.Given the complexity of these host-pathogen systems, experimental approaches are crucial for disentangling potential mechanisms driving patterns of transmission and examining variation in lethal and sublethal effects due to host species, host life-history traits, pathogen strain, host populations, and environmental conditions.
Prior to 2009, relatively few studies of amphibian diseases employed standard experimental designs [28] (Figure 2).Since 2009, there has been a surge in the use of experiments to determine how diseases affect amphibians.Experimental design, methods, and interpretation vary; thus, it is useful to summarize these aspects to assess generality.One problem with experimental work on amphibian diseases has been the lack of standardization in experimental methods.Here, we present a synthesis of experimental studies and attempt to address some of the issues regarding the lack of standardization and difficulties in generalizing about the dynamics of the host-pathogen systems we focus on.
Diversity 2018, 10, x FOR PEER REVIEW 3 of 48 diseases has been the lack of standardization in experimental methods.Here, we present a synthesis of experimental studies and attempt to address some of the issues regarding the lack of standardization and difficulties in generalizing about the dynamics of the host-pathogen systems we focus on.First described by Longcore et al. [29], Bd is a fungal species in the phylum Chytridiomycota that has multiple hosts on every continent where amphibians exist [15,16] and has been associated with numerous population declines and some extinctions [30][31][32].Recent evidence suggests that that the source of Bd was traced to the Korean peninsula, where one lineage, BdASIA-1, exhibits the genetic hallmarks of an ancestral population that seeded the panzootic emergence [33].O'Hanlon et al. [33] date the emergence of Bd to the early 20 th century, coinciding with the global expansion of commercial trade in amphibians.
Bd has a complex life cycle that consists of a free-living infectious aquatic zoospore stage and a non-motile zoosporangium stage.Motile zoospores are chemically attracted to keratin in amphibian host, such as keratinized larval jaw sheaths or keratinized epidermal layers of adult amphibian skin [34,35].Infection can lead to hyperkeratosis and hyperplasia of the dermal layer, erosions and ulcerations of the skin, and disruption of the epidermal cell cycle [30,[34][35][36][37].The inability to regulate ions through the skin may lead to cardiac arrest [38].Clinical signs of chytridiomycosis include lethargy, lack of appetite, abnormal posture, loss of righting reflex, cutaneous erythema, and increased skin sloughing [37].However, not all infected animals are symptomatic when infected.Once within the host, the zoosporangia mature and develop pathogenic zoospores that are released outside the host into the aquatic environment.

Batrachochytrium salamandrivorans
The recent isolation and characterization of the fungal pathogen, Bsal may explain some amphibian population declines.For instance, the drastic decline of fire salamanders, Salamandra salamandra, in the Netherlands, Germany, and Belgium, has been linked to Bsal [39][40][41].A study conducted by Martel et al. [42] proposed Bsal originated in East Asia and coexisted with salamanders there for millions of years.The introduction of Bsal to Europe is hypothesized to have occurred due to

Batrachochytrium dendrobatidis
First described by Longcore et al. [29], Bd is a fungal species in the phylum Chytridiomycota that has multiple hosts on every continent where amphibians exist [15,16] and has been associated with numerous population declines and some extinctions [30][31][32].Recent evidence suggests that that the source of Bd was traced to the Korean peninsula, where one lineage, BdASIA-1, exhibits the genetic hallmarks of an ancestral population that seeded the panzootic emergence [33].O'Hanlon et al. [33] date the emergence of Bd to the early 20 th century, coinciding with the global expansion of commercial trade in amphibians.
Bd has a complex life cycle that consists of a free-living infectious aquatic zoospore stage and a non-motile zoosporangium stage.Motile zoospores are chemically attracted to keratin in amphibian host, such as keratinized larval jaw sheaths or keratinized epidermal layers of adult amphibian skin [34,35].Infection can lead to hyperkeratosis and hyperplasia of the dermal layer, erosions and ulcerations of the skin, and disruption of the epidermal cell cycle [30,[34][35][36][37].The inability to regulate ions through the skin may lead to cardiac arrest [38].Clinical signs of chytridiomycosis include lethargy, lack of appetite, abnormal posture, loss of righting reflex, cutaneous erythema, and increased skin sloughing [37].However, not all infected animals are symptomatic when infected.Once within the host, the zoosporangia mature and develop pathogenic zoospores that are released outside the host into the aquatic environment.

Batrachochytrium salamandrivorans
The recent isolation and characterization of the fungal pathogen, Bsal may explain some amphibian population declines.For instance, the drastic decline of fire salamanders, Salamandra salamandra, in the Netherlands, Germany, and Belgium, has been linked to Bsal [39][40][41].A study conducted by Martel et al. [42] proposed Bsal originated in East Asia and coexisted with salamanders there for millions of years.The introduction of Bsal to Europe is hypothesized to have occurred due to a lack of biosecurity in the international pet trade [42].Although Bd and Bsal infections result in lethal skin erosion, the pathogenic mechanism of Bsal is not well understood.Bsal produces motile zoospores, contain colonial thalli, and produce germination tubes in vitro [18].Studies have assessed the presence of Bsal in various amphibian populations in North America (e.g., [43][44][45]) and China [46] utilizing several methods (phalanges histology, nested PCR, qPCR and duplex qPCR), but its presence has yet to be confirmed in those populations.Given its high lethality, increased field surveillance of these naïve populations will be critical to contain the potential spread of this newly isolated pathogen, particularly in North America, a global biodiversity hotspot for salamanders [47][48][49][50].

Ranavirus
Rvs are a group of large double-stranded DNA viruses in the family Iridoviridae with fish, reptile, and amphibian hosts [51].The first Rv were isolated from Lithobates pipiens in 1965 [52].The Global Ranavirus Reporting System (https://mantle.io/grrs/map),created to aid in tracking Rv occurrences and studies, shows Rv to be fairly widespread in Canada and the US west of the Rocky Mountains.This tool is intended to facilitate communication among researchers concerning Rv detection and to accelerate research and management of the disease threat.
The genus Rv is composed of 6 identified viral species, three of which infect amphibians (Ambystoma tigrinum virus (ATV), Bohle iridovirus (BIV), and Frog Virus 3 (FV3)) [51].Although the effects of Rv are well documented, little is known about the genetic basis for virulence across isolates [53].FV3 and ATV infect many amphibian species, but these isolates are most virulent within the anurans and urodelans, respectively, from which they were isolated [54].Laboratory experiments have shown that introduced Rv isolates may be significantly more virulent than endemic strains [55].
Amphibians become infected with Rv by physical contact, dermal exposure to contaminated water, or direct ingestion of virions [56,57].Infection can occur in as short as a one second of direct contact with an infected individual of the same species [56] or 3 h of contact with contaminated water [58].Empirical studies confirming its potential effects in amphibians are limited [56,[59][60][61].Fish can also be infected with Rv, but susceptibility to Rv in fishes appears to be low, though there is potential for fish to transfer Rv to amphibians in habitats where they overlap [62,63].
Rvs infections can cause cell apoptosis and tissue necrosis within a few hours [51,64].Common indicators of Rv infection include erratic swimming, lethargy, erythema, skin sloughing, loss of pigmentation, lordosis (excessive inward curvature of the spine), and ulcerations [65,66].Lesions and hemorrhages associated with fatal cases of Rv occur in internal organs, particularly the liver, kidney, intestine, spleen, and reproductive organs [25,67,68].However, the precise mechanisms of Rv dissemination within the host are relatively unclear, especially at the earliest stages of infection.A recent study demonstrated that FV3 infection is capable of altering the blood brain barrier in Xenopus laevis tadpoles eventually, leading to Rv dissemination into the central nervous system [69].Death can occur without external signs of infection [70].

Methods
The effects of Bd, Bsal, and Rv found in experimental studies are summarized in Table 1.Our search was conducted via the Web of Science and supplemented with a Google Scholar search using the keywords "Batrachochytrium dendrobatidis + amphibians", "Batrachochytrium salamandrivorans + amphibians", and "Ranavirus + amphibians", respectively.Duplicates and non-experimental studies were removed and the remaining studies were documented.Studies that examined interactive effects (i.e., pesticide + pathogen) were included, but only the effect of the pathogen independently was reported.The Bd search (1999-2017) resulted in 1207 hits, of which 110 were experimental studies.The Bsal search resulted in 41 hits, of which 5 were experimental studies.The Rv search (1992-2017) yielded 269 hits, of which 33 were experimental studies.If one publication examined multiple species or host life stages, each species and life stage was reported separately (Figure 3).Publications were compiled using the search strings "Batrachochytrium dendrobatidis + amphibians" and "ranavirus and amphibians" in the Web of Science database, from which duplicates and articles that were unrelated were removed.The Bd search yielded a total of 1207 hits and the Rv search yielded 269 hits.

Results
Results from experimental studies are summarized below.We presented general trends across studies according to the response variable (e.g., physiology, behavior) and/or source of response variation (e.g., life stage, virus strain).We then focused on interactive effects and summarize the experimental work with each pathogen in combination with natural or anthropogenic environmental stressors.Below, we provide a summary of patterns and gaps in the accumulated experimental work on host-pathogen dynamics of Bd, Bsal, and Rv and their amphibian hosts.Specific results of experimental studies are detailed in Table 1 and data summarizing the number of papers published, survivorship and life stages are summarized in Figures 4-6.Publications were compiled using the search strings "Batrachochytrium dendrobatidis + amphibians" and "ranavirus and amphibians" in the Web of Science database, from which duplicates and articles that were unrelated were removed.The Bd search yielded a total of 1207 hits and the Rv search yielded 269 hits.

Results
Results from experimental studies are summarized below.We presented general trends across studies according to the response variable (e.g., physiology, behavior) and/or source of response variation (e.g., life stage, virus strain).We then focused on interactive effects and summarize the experimental work with each pathogen in combination with natural or anthropogenic environmental stressors.Below, we provide a summary of patterns and gaps in the accumulated experimental work on host-pathogen dynamics of Bd, Bsal, and Rv and their amphibian hosts.Specific results of experimental studies are detailed in Table 1 and data summarizing the number of papers published, survivorship and life stages are summarized in Figures 4-6.
The number of experimental studies conducted on hosts at different life stages varied, with most studies of Bd conducted in hosts after metamorphosis and most studies of Rv conducted with larvae (Figure 4).The only experimental studies we found on Bsal were conducted with post-metamorphic hosts (Figure 4).Experimental studies and survival showed clear differences with host life stage (Figures 5 and 6).Moreover, the dose of pathogen administered during susceptibility experiments is also important in interpreting results (Figure 7).
Comparative strain experiments along with observational amphibian surveys are useful in investigating the relationships between host population trends and Bd virulence variation.For example, Piovia-Scott et al. [92] linked an observed Rana cascadae population decline to a known, highly infectious, and lethal Bd strain through multiple lines of analyses.In one experiment, adult Rana cascadae, exposed to the Bd strain cultured from a site undergoing a host population decline, had significantly lower survival rates, compared to those exposed to a strain from a site with a stable host population [92].This Bd strain also displayed greater immunotoxicity in experimental assays [92].Exposure to endemic vs. novel strains can also affect host survival.Doddington et al. [93] found survival differences in captive-bred Alytes muletensis experimentally exposed to two Bd strains, a local Mallorcan strain (TF5a1) or a hypervirulent Bd-GPL strain (UKTvB).Toads exposed to the Bd-GPL strain had higher mortality than individuals exposed to the Mallorcan strain or control group [93].
Diversity 2018, 10, x FOR PEER REVIEW 6 of 48 The number of experimental studies conducted on hosts at different life stages varied, with most studies of Bd conducted in hosts after metamorphosis and most studies of Rv conducted with larvae (Figure 4).The only experimental studies we found on Bsal were conducted with post-metamorphic hosts (Figure 4).Experimental studies and survival showed clear differences with host life stage (Figures 5 and 6).Moreover, the dose of pathogen administered during susceptibility experiments is also important in interpreting results (Figure 7).
Comparative strain experiments along with observational amphibian surveys are useful in investigating the relationships between host population trends and Bd virulence variation.For example, Piovia-Scott et al. [92] linked an observed Rana cascadae population decline to a known, highly infectious, and lethal Bd strain through multiple lines of analyses.In one experiment, adult Rana cascadae, exposed to the Bd strain cultured from a site undergoing a host population decline, had significantly lower survival rates, compared to those exposed to a strain from a site with a stable host population [92].This Bd strain also displayed greater immunotoxicity in experimental assays [92].Exposure to endemic vs. novel strains can also affect host survival.Doddington et al. [93] found survival differences in captive-bred Alytes muletensis experimentally exposed to two Bd strains, a local Mallorcan strain (TF5a1) or a hypervirulent Bd-GPL strain (UKTvB).Toads exposed to the Bd-GPL strain had higher mortality than individuals exposed to the Mallorcan strain or control group [93].Differences in methodology can complicate our interpretation of the results from comparative strain experiments.For example, Bd dosage, site of strain isolation, and strain passaging history can influence outcomes of strain experiments [15,[86][87][88][94][95][96].
Accumulating evidence suggests that some host species vary in their susceptibility to Bd.Some species can persist with infection [97] and others experience mortality rapidly after Bd exposure [86,[97][98][99][100].Variation in skin composition, including keratin abundance, distribution, and thickness, may affect the depth, of the zoospore-produced germination tube which can affect the severity of infection among amphibian hosts [35,101].Differences in the ability of amphibian species to mount sufficient endocrinological responses, particularly stress responses, may also play a role [102][103][104][105]. Furthermore, habitat preference may influence host susceptibility to infection [106,107].Future  Differences in methodology can complicate our interpretation of the results from comparative strain experiments.For example, Bd dosage, site of strain isolation, and strain passaging history can influence outcomes of strain experiments [15,[86][87][88][94][95][96].
Accumulating evidence suggests that some host species vary in their susceptibility to Bd.Some species can persist with infection [97] and others experience mortality rapidly after Bd exposure [86,[97][98][99][100].Variation in skin composition, including keratin abundance, distribution, and thickness, may affect the depth, of the zoospore-produced germination tube which can affect the severity of infection among amphibian hosts [35,101].Differences in the ability of amphibian species to mount sufficient endocrinological responses, particularly stress responses, may also play a role [102][103][104][105]. Furthermore, habitat preference may influence host susceptibility to infection [106,107].Future Differences in methodology can complicate our interpretation of the results from comparative strain experiments.For example, Bd dosage, site of strain isolation, and strain passaging history can influence outcomes of strain experiments [15,[86][87][88][94][95][96].
Accumulating evidence suggests that some host species vary in their susceptibility to Bd.Some species can persist with infection [97] and others experience mortality rapidly after Bd exposure [86,[97][98][99][100].Variation in skin composition, including keratin abundance, distribution, and thickness, may affect the depth, of the zoospore-produced germination tube which can affect the severity of infection among amphibian hosts [35,101].Differences in the ability of amphibian species to mount sufficient endocrinological responses, particularly stress responses, may also play a role [102][103][104][105]. Furthermore, habitat preference may influence host susceptibility to infection [106,107].Future research should consider amphibian life-history traits, particularly of species that do not seem to be susceptible to Bd infection, to better understand differences in host susceptibility and will be useful to target species, which may act as reservoirs for the pathogen.research should consider amphibian life-history traits, particularly of species that do not seem to be susceptible to Bd infection, to better understand differences in host susceptibility and will be useful to target species, which may act as reservoirs for the pathogen.
Figure 7.The effect of Bd dose (in log zoospores) on survival.These data are direct counts from Table 1.Experiments that use multiple dose levels or multiple strains were excluded.Reduced survival means mortality of hosts exposed to Batrachochytrium was significantly higher than control mortality.
Here, we display the minimum, first quartile, median, third quartile, and maximum zoospore dose regarding host survival.
An important driver of host-pathogen interactions is host behavior [72,108,109].Basking, for example, may be an indication of disease infection in amphibians [110][111][112].Altered thermoregulatory behavior (i.e., behavioral fever) may aid in clearing Bd infection.However, fever behavior depends on species and life stage [108,113].Additionally, it has been suggested that aggregation behaviors can increase Bd prevalence.Thus, schooling species may be more at risk than amphibian species with solitary life styles [109].This prediction depends strongly on the assumption that infected hosts shed infectious zoospores.Recent work shows that spillover infection does not occur in all hosts, suggesting that aspects of life history (e.g., body size) and behavioral interactions (e.g., interspecific competition) between hosts may drive infection severity in host communities [114].Infected tadpoles have demonstrated altered activity levels, which may be an important indicator of anti-predator behavior [72,115].While reduced activity can make tadpoles less visible and thus less at risk for predation, sluggish behavior can hinder an individual's ability to escape a predation event.Han et al. [115] observed Bd-infected toad tadpoles seeking refuge more often than other species tested.Parris et al. [72] demonstrated that when tadpoles were exposed to only visual predation cues, uninfected individuals positioned themselves farther from the predator than infected animals.Carey et al. [99] observed that post-metamorphic toads exposed to Bd were holding their bodies out of water more than unexposed individuals.In one study, frogs that had never been exposed to Bd displayed no significant avoidance or attraction to the pathogen, whereas previously infected frogs associated with pathogen-free frogs a majority of the time [83].This indication of potentially learned behavioral avoidance to Bd and perhaps other pathogens warrants further exploration.
Differences in Bd susceptibility are dependent on amphibian life stage, with juveniles and adults usually being more susceptible than embryos and larvae, most likely due to increased keratin distribution and abundance after the larval stage [80,116].Bd infection in tadpoles rarely results in mortality (see [15,86,98], but has generally been related to reduced foraging efficiency and food intake  1. Experiments that use multiple dose levels or multiple strains were excluded.Reduced survival means mortality of hosts exposed to Batrachochytrium was significantly higher than control mortality.Here, we display the minimum, first quartile, median, third quartile, and maximum zoospore dose regarding host survival. An important driver of host-pathogen interactions is host behavior [72,108,109].Basking, for example, may be an indication of disease infection in amphibians [110][111][112].Altered thermoregulatory behavior (i.e., behavioral fever) may aid in clearing Bd infection.However, fever behavior depends on species and life stage [108,113].Additionally, it has been suggested that aggregation behaviors can increase Bd prevalence.Thus, schooling species may be more at risk than amphibian species with solitary life styles [109].This prediction depends strongly on the assumption that infected hosts shed infectious zoospores.Recent work shows that spillover infection does not occur in all hosts, suggesting that aspects of life history (e.g., body size) and behavioral interactions (e.g., interspecific competition) between hosts may drive infection severity in host communities [114].Infected tadpoles have demonstrated altered activity levels, which may be an important indicator of anti-predator behavior [72,115].While reduced activity can make tadpoles less visible and thus less at risk for predation, sluggish behavior can hinder an individual's ability to escape a predation event.Han et al. [115] observed Bd-infected toad tadpoles seeking refuge more often than other species tested.Parris et al. [72] demonstrated that when tadpoles were exposed to only visual predation cues, uninfected individuals positioned themselves farther from the predator than infected animals.Carey et al. [99] observed that post-metamorphic toads exposed to Bd were holding their bodies out of water more than unexposed individuals.In one study, frogs that had never been exposed to Bd displayed no significant avoidance or attraction to the pathogen, whereas previously infected frogs associated with pathogen-free frogs a majority of the time [83].This indication of potentially learned behavioral avoidance to Bd and perhaps other pathogens warrants further exploration.
Differences in Bd susceptibility are dependent on amphibian life stage, with juveniles and adults usually being more susceptible than embryos and larvae, most likely due to increased keratin distribution and abundance after the larval stage [80,116].Bd infection in tadpoles rarely results in mortality (see [15,86,98], but has generally been related to reduced foraging efficiency and food intake in larvae [117][118][119][120].In post-metamorphic amphibians, Bd infection is manifested in the keratinized epidermis; thus, the effects of foraging efficiency are dependent on the locality of infection.For example, in adult salamanders (Plethodon cinereus), Bd-infected individuals displayed increased feeding behaviors in comparison with uninfected individuals, a behavioral modification that has been suggested as a strategy to offset the costs associated with immune activation [121].
Body size may also be a factor in host susceptibility to pathogens [122].Experiments have shown that individual size may be an influential factor in Bd susceptibility [116].Garner et al. [79] showed that smaller toads (Anaxyrus boreas) were more prone to Bd-induced mortality compared with larger individuals.
Experiments on host-Bd interactions have addressed physiological stress responses.In both field and laboratory investigations, Bd significantly elevated physiological stress hormone (corticosterone) levels in amphibian hosts of multiple species [102][103][104]123], though there is no evidence that exposure to endogenous corticosterone alters amphibian susceptibility to Bd [104].Different strains of Bd elicit significantly distinctive hormonal stress responses from their hosts, with more virulent strains resulting in higher corticosterone levels [123].New methodologies, such as a non-invasive stress hormone assay [102], enhance the value of field studies coupled with experimental laboratory investigations on physiological stress response.The dynamics between stress response and chronic disease manifestation warrant further exploration.

Batrachochytrium salamandrivorans
Due to its recent discovery, there are few experimental studies documenting the effects of Bsal on amphibian hosts (Table 1b).Bsal primarily affects newts and salamanders rather than anurans.The common midwife toad (Alytes obstetricans), a species susceptible to Bd, did not experience any clinical signs of Bsal infection [18].Further, Martel et al. [42] showed that ten anurans tested were resistant to skin invasion, infection, and disease signs when exposed to a dose of 5000 zoospores of Bsal.Studies conducted with Bsal on potential urodelan hosts demonstrated that responses varied across species and within the same genus.Bsal induced lethal effects on Lissotriton italicus, the Italian newt, whereas no infection or disease signs were documented in L. helveticus [42].The results of Bsal-host experiments show that Bd and Bsal differ in how they show the effects of exposure to these pathogens [18,42].Experimentally infected fire salamanders, Salamandra salamandra, experienced ataxia, a rarely reported sign in experimental studies with Bd.The study also identified three potential reservoir species, the Japanese fire belly newt (Cynops pyrrhogaster), the Chuxiong fire-bellied newt (Hypselotriton cyanurus), and the Tam Dao salamander (Paramesotriton deloustali), as individuals of these species were able to persist with or clear infection in some capacity [42].
Bsal transmission dynamics are not yet well documented.In a study examining transmission between infected and naïve hosts, Martel et al. [18] found that two days of shared housing in salamanders resulted in infection and mortality of formerly naïve hosts within one month.All experimental work done regarding Bsal has used only one pathogen isolate, a small range of doses, and few source populations for each species tested (Table 1b).Because experiments conducted on Bd-host dynamics show that responses are heavily dependent on species, population, pathogen isolate, temperature, and exposure dose, future research should consider how these factors influence infection dynamics in the Bsal system.

Ranavirus
Experimental studies have shed light onto the comprehensive effects of Rv on amphibians worldwide (Figure 3; Table 1c).Experimental Rv mortality is influenced by a variety of factors most notably, exposure method.Ingestion of Rv infected carcasses result in infection transmission and reduced survival [57,124].Exposure to Rv via water induced variable rates of mortality, with most studies showing slower rates of mortality when transmission occurred via water, compared to when it occurred via ingestion [70,125].Hoverman et al. [126] found that infection and mortality rates were greater for tadpoles that were orally inoculated with Rv compared to those exposed via water bath.Aggressive interactions may serve as an efficient transmission route of Rv [56].Cannibalistic behavior may be harmful to the individual exemplifying the behavior because of disease transmission, but an experimental study showed cannibalism can result in decreased contact rates between naive and infected individuals in the population [56].Additionally, experiments have suggested that necrophagy may serve as a common route of Rv transmission, shifting transmission from density-dependent to frequency-dependent [56,57,124,127,128].
Temperature influences Rv infectivity and survival rates in hosts [129,130].When exposed to the Rv, ATV, larval Ambystoma tigrinum salamanders experienced higher survival rates when exposed at 26 • C than those exposed at 18 • C and 10 • C with virus titer being higher in cooler temperatures, and viral replication rates were higher at higher temperatures [130].Similarly, Echaubard et al. [129] found that the probability of Rv infection increased at lower temperatures (14 • C), but that the effects were isolate and species-dependent.
It is critical to take a comparative approach to experimentally investigate species variation in susceptibility with regards to Rv. Understanding the relative susceptibility of hosts to a pathogen is important for predicting host-pathogen dynamics.Coevolution between Rvs and their hosts has been hypothesized to be a driving force behind host variation of susceptibility [131].Hoverman et al. [132] discovered a wide range of lethal effects among 19 larval amphibian species, which resulted in mortality rates spanning from 0 to 100%.Their study showed that anurans in the family Ranidae were typically more susceptible to Rv than the other five families tested.
Previous experimental work has demonstrated infection and virulence variation among isolates and Rv species [54,125,132,133] though phenotypic variation among Rv isolates is not well understood.Schock et al. [54] determined that FV3 and ATV Rv species vary in their ecology and restriction endonuclease profiles, even though they have identical major capsid protein (MCP) gene sequences.Their results further emphasize the importance of characterizing isolates beyond MCP sequence analysis.Cunningham et al. [125] detected differences in tissue trophism and pathology between two strains of FV3-like Rvs in common frogs (Rana temporaria).Schock et al. [133] revealed that ATV strains differed in virulence, but this was dependent upon the origin of the salamander host.Similarly, Hoverman et al. [132] showed that mortality rates were ~50% greater with a Rv isolate obtained from an American bullfrog (Lithobates catesbeianus) culture facility compared to FV3.These results highlight the importance of controlled experimental studies to elucidate patterns of differential host susceptibility with regards to Rv isolates and species.
Experimental and observational field studies have shown that late-stage larvae that are nearing metamorphosis are the most susceptible to lethal effects of Rv infection [60,61,105,134,135].When exposed to ATV, metamorphosed Ambystoma tigrinum larvae were five times less likely to be infected than those that remained at the larval stage [70].Experimental studies suggest that the effects of Rv are more lethal to larvae than any other host life stage.In an experimental study examining seven amphibian species at various developmental stages, Haislip et al. [136] observed that mortality and infection prevalence were greatest at the hatchling and larval stages in four of the species tested compared with frogs undergoing metamorphosis, and that the embryo was the least susceptible stage, possibly due to the eggs protective membranous properties.Similarly to what has been observed with Bd infections, life-stage variation in susceptibility has been attributed to changes that occur in the hypothalamic-pituitary-interrenal axis (the central stress response system) around the time of metamorphosis, which helps to mediate the immune system [137].Host gene expression variation may contribute to life-stage differences in susceptibility.Andino et al. [134] found that larvae experienced greater infection rates and possessed lower and delayed expression of inflammation associated antiviral genes.It has been suggested that impacts of epizootic events may be underestimated due to increased difficulty of detecting mass mortality of hatchings and larvae in the field [136].
Though few studies have examined host physiological responses to Rv, these studies are important in assessing species-specific impacts of infection.Warne et al. [105] demonstrated tadpoles infected with an FV3-like isolate had higher corticosterone relative to controls.In a study examining immune function, Maniero et al. [138] demonstrated that Xenopus laevis frogs develop an effective and persistent humoral immunity after exposure to FV3.

Interactive Effects of Disease, Anthropogenic, and Natural Stressors
Anthropogenic and natural environmental stressors can exacerbate the effects of emerging wildlife diseases [14].Though the impact of one factor may be particularly devastating to amphibians in certain regions, considering simultaneous effects of several factors may be more realistic because amphibians, like other organisms, are exposed to many abiotic and biotic factors simultaneously [9,139].Host-pathogen relationships in amphibians are mediated by, for example, climate, contaminants, disease, predation, and competition [9,15,79,140] (Figure 1).These factors display a high degree of spatial and temporal variation and can result in complex local interactions that are often poorly understood [9].Realistic insight can be gained by taking a population-specific approach in assessing the variables involved and overall status of a population using long-term field data [141].Experimental approaches can be particularly helpful in disentangling the mechanisms of interacting variables.Gaining a comprehensive understanding of how environmental factors may influence infection and pathology is critical to amphibian conservation.

Pathogens Climate and Atmospheric Change
Climate change and associated atmospheric changes may alter disease dynamics by fostering conditions more or less hospitable for pathogens and their hosts.For example, different outcomes have been reported regarding the interaction of ultraviolet-B (UV-B) radiation and pathogens.A modeling approach by Williamson [142] suggests that the selective absorption of ultraviolet radiation by dissolved organic matter (DOM) decreases the valuable ecosystem service wherein sunlight inactivates waterborne pathogens.In controlled experiments, Overholt et al. [143] showed that low levels of UVR (as well as longer-wavelength light) sharply reduced the infectivity of parasitic fungal spores, but did not affect host (Daphnia) susceptibility to infection.However, a field experiment showed that fluctuations in water depth were associated with increased UV-B radiation, which resulted in greater sensitivity to the pathogenic water mold, Saprolegnia [139].Experimental studies regarding the effects of UV-B radiation and Rv are absent from the literature.However, decreased pond depth has been associated with increased Rv prevalence [63], which suggests the possibility that water depth and UV-B penetration may affect Rv-host dynamics, as Kiesecker et al. [139] showed for Saprolegnia-amphibian interactions.In a laboratory experiment, no interaction was found with increased UV-B radiation and Bd [144,145].However, Ortiz-Santaliestra et al. [146] showed that Bd loads were significantly lower in tadpoles exposed to environmental UV-B intensities than in tadpoles not exposed to the radiation.Another field experiment showed that ultraviolet radiation (UVR) killed the free-living infectious stage of Bd.However, permanent ponds with more UVR exposure had higher infection prevalence [147].The authors suggested that UVR reduced the density of Bd predators and that permanent sites fostered multi-season host larvae that fueled parasite production.
Global climate change appears to increase temperature variability, which can mediate disease dynamics.Bosch et al. [148] documented rising temperatures are linked to the occurrence of chytridiomycosis.Fluctuating temperature regimes have had negative effects on survival and development of amphibians in the presence of Bd [149][150][151], while higher temperatures often resulted in higher host survival rates [78,152].Raffel et al. [150] demonstrated that Bd growth and infection-induced mortality on newts, Notophthalmus viridescens, was greater following a shift to a new cooler temperature, but this was dependent on increased soil moisture.Host thermal acclimation is context-dependent and can serve as a key mediator of climate-disease dynamics.Recent models based on the Intergovernmental Panel on Climate Change (IPCC) suggest that Bd will shift into higher latitudes and altitudes due to increased environmental suitability in regions under predicted climate change [153].Specifically, these models predicted a broad expansion of areas suitable for establishment of Bd on amphibian hosts in temperate zones of the Northern Hemisphere.Thus, novel amphibian hosts may be susceptible to predictable shifts in Bd.

Pathogens and Contaminants
Many contaminants break down quickly in the environment, yet exposure can have major carry over effects, and the effects of interactions between multiple contaminants and between contaminants and disease cannot be well understood without experimentation [154,155].Contaminant exposure may contribute to amphibian population declines directly or indirectly [9,[156][157][158].However, research on the interactive effects of contaminants and pathogens remains inconclusive.Some studies examining this interaction investigate if pesticides and contaminants play a role in decreasing amphibian immune response, rendering amphibians more susceptible to infectious disease [159][160][161].However, few experimental studies support this hypothesis [118,[162][163][164][165][166][167][168][169][170][171][172].Rohr et al. [173] found that early-life exposure to atrazine decreased survival post-metamorphosis when combined with Bd in Osteopilus septentrionalis.Likewise, Buck et al. [163] demonstrated that exposure to pesticides in tadpoles resulted in higher Bd loads and increased mortality in post-metamorphic individuals from three species, but not for two other species.A possible reason for findings with little or no interactive effects may be that certain compounds can inhibit or diminish the growth or integrity of Bd, as was demonstrated outside of the host species [162,167,170].Thus, contaminants may have direct negative effects on both amphibian hosts and Bd, which can lead to no differences in infection across a range of contamination.
The use of pesticides has been associated with increased Rv prevalence in the field [63].Forson and Storfer [174] revealed that ecologically relevant levels of the pesticide atrazine and the fertilizer sodium nitrate significantly decreased Ambystoma tigrinum larvae peripheral leukocyte levels and that larvae exposed to atrazine significantly increased susceptibility to ATV.Furthermore, Kerby and Storfer [175] showed that atrazine and Rv exposure marginally decreased survival in larvae of the same species.Conversely, Forson and Storfer [174] revealed Ambystoma macrodactylum larvae exposed to atrazine and ATV had lower mortality levels and ATV infectivity compared to larvae exposed to ATV alone, suggesting atrazine may compromise virus integrity.Additional research is needed to assess the impacts of pesticides and fertilizers and their metabolites on Rv viability and amphibian physiology.Contaminants are becoming increasingly widespread with over 50% of detected insecticide concentrations exceeding regulatory thresholds [176].Thus, the importance of researching the interrelationships between contaminants and disease in amphibian disease should not be overlooked.Experiments designed to identify mechanisms that are generalizable across classes of pesticides will also enable better management and conservation planning, as known contaminants are phased out and new ones are introduced to market.

Pathogens and Community Composition
Higher biodiversity may influence disease risk through a variety of mechanisms.The dilution effect hypothesizes that greater biodiversity in an assemblage decreases disease risk, but this is somewhat controversial [177][178][179].Olson et al. [16] reported a negative association between Bd occurrence and species richness.Some experimental evidence supports the dilution effect in the Bd-host system.Greater species diversity of larvae resulted in lower Bd zoospore abundance [100,[180][181][182]. Searle et al. [100] demonstrated that the experimental addition of Rana cascadae tadpoles to tanks with larval toads (Anaxyrus boreas) decreased the infection risk for toad larvae, which may be due to differing feeding strategies and life-history traits between species.
Venesky et al. [183] showed that some tadpoles can filter feed Bd zoospores.Moreover, experiments have shown that zooplankton, such as Daphnia, can consume Bd zoospores, significantly reducing infection probabilities in tadpoles [184][185][186].Additionally, species "reservoirs" may be important for community-level Bd dynamics.For example, evidence suggests the Pacific treefrog, Pseudacris regilla, may act as a Bd reservoir; P. regilla thrive and occupy 100% of study sites where a sympatric species has been extirpated by Bd [101].
Predation can interact with infection in varying ways.The healthy herd hypothesis states that predators may decrease infection prevalence by decreasing overall population size of potential hosts and through selective predation upon infected individuals [187][188][189].Several hypotheses regarding predator/prey dynamics and disease remain untested regarding disease and amphibians.For example, is selective predation occurring, or alternatively, are predators capable of avoiding infected prey?Han et al. [115] experimentally demonstrated the potential of non-selective predation occurring in the predator/prey interactions in the Bd system.Salamander predators consumed Bd-infected and uninfected tadpoles at the same frequency and predation risk among prey was not altered by Bd infection.This area warrants further exploration as predation behavior may have significant impact on outcomes in amphibian disease systems.The presence of a predator resulted in decreased infection loads in wood frog (Lithobates sylvaticus) larvae [190] and has resulted in increased developmental rates [162,191].Effects of predation in combination with Rv remain inconclusive.Dragonfly predator cues have resulted in decreased survival in combination with Rv exposure [192].However, Haislip et al. [193] found no evidence that Rv exposure in combination with predator cues increased mortality across four species of larval anurans.
In addition to predator presence, other aspects of community composition can play an influential role in disease dynamics.When reared in higher densities, amphibians metamorphose at smaller body masses than when reared individually [194,195].Furthermore, when these higher densities were combined with the presence of Bd, larvae also experienced a delayed time of metamorphosis [194,195].Increased densities have also been associated with the increased likelihood of Bd infection [196], but other experimental studies have not observed this association [100].These results are in direct contrast with the effects of density with regards to Rv.At higher densities of larvae and in the presence of Rv, the rate of metamorphosis was documented to be three times faster and the probability of mortality was five times lower than in the controls [197].However, even though higher densities lead to higher contact rates, transmission of Rv rapidly saturates as density increases [198].

Coinfection Dynamics
Infection by multiple pathogens is common for most wild animals [199], though experimental evidence of coinfection patterns in amphibians remain sparse.Several studies have investigated coinfection dynamics in amphibian hosts in the field and have found that coinfections in amphibians is common [132,[200][201][202].However, there are few experimental studies of coinfection dynamics in amphibians.Romansic et al. [74] experimentally investigated the effects of three pathogens: Bd, the trematode Ribeiroia sp., and the water mold, Achyla flagellata, which resulted in little evidence for interactive effects.Wuerthner et al. [203] found that prior infection with trematode parasites (Echinoparyphium sp.) reduced ranavirus loads and increased survival of Rv-infected frogs.Thus, the interrelationships of coinfection could be explored further via experiments.

Host, Isolate, and Geographic Biases
Uneven sampling of host species is considered to be a source of bias when interpreting the dynamics of host-parasite systems [204].There are 7728 amphibian species described [205], yet our analysis of experimental studies documenting the effects of these pathogens have only reported effects for <1% of species across these pathogens (0.01% of species with regard to Bd, 0.005% of species for Bsal, and 0.005% of species with regard to Rv).Of the species studied in these disease systems, there is a high degree of interspecific variation in disease susceptibility [80,86,97,98,100,132].Furthermore, responses can vary based on strain, population, and host life-stage [54,56,70,88,98,124,133,[206][207][208].Additionally, a distinct disparity exists in species-studied and geographic regions (Figure 8).Much of the research has focused primarily on host species located in Europe, North America, and Australia.However, Bd and Rv have global distribution and effects, yet far less is known about infection in hosts from Africa, Asia, and South America.For Bsal, experiments have only been conducted with an isolate from Europe, and most studies have used a dose of 5000 zoospores, a low dose in comparison to studies on Bd [80].Similarly, the bulk of the studies examining Rv pathogen-host dynamics are largely biased toward those in North America, with a minority of studies coming out of Europe, Africa, and Australia (Table 1).These biases are likely due to the number of researchers in these regions, institution locality, and access to collaborators, species, isolates, feasibility and cost.

Non-Standard Methods and Reporting
Experimentation is advantageous because it is repeatable, and well-designed studies can provide unequivocal results [209,210].However, there are limitations on experimental work, as is illustrated in amphibian disease ecology.One problem with experimental work on amphibian diseases has been the lack of standardization in experimental methods.Kilpatrick et al. [87] highlighted the importance of standardizing and reporting all relevant infection protocols within and between species when conducting laboratory studies regarding Bd and its host species.This includes how individuals are collected for experiments, how they are reared, the developmental stage in which they are tested, the population origin, inoculation and exposure protocols, and strains of pathogen being used.For instance, reporting and standardizing the zoospore exposure concentration (total number of zoospores per mL of water in total volume of water) in experimental procedures would make relative species comparisons among experiments more useful.Developmental stage should always be reported as this can also confound the interpretation of results.Additionally, whether hosts are reared from eggs or caught as larvae, juveniles, or adults, or even bought from supply houses can dramatically alter the results of experiments and their interpretation.Our analysis shows that, 27%, 12%, and 23% of experiments examining Bd, Bsal, and Rv, respectively, were using animals not reared from eggs, even though rearing amphibians from eggs ensures that individuals have not previously been infected with Bd or Bsal.Even when tested for current infection prior to an experiment, wild-caught individuals have different ecological histories and may have a more or less robust immune system depending upon whether they were previously exposed to a particular pathogen [86].Field surveillance shows that amphibian parasites, such as echinostomes, are widespread [211,212] and essentially many, if not all individuals, collected from the wild will inevitably possess trematodes.The potential influence of these parasites on amphibian immunological response poses a serious problem for experiments that use individuals, not reared as eggs.
We emphasized the importance of utilizing subjects raised from the embryo stage in experimental investigations.Because of lack of standardization, each experiment must be taken independently and applied to those specific individuals at the reported experimental conditions.When protocols are standardized, we can more easily generalize effects of Bd and Rv on hosts, as has been accomplished in several studies [80,97,98,100,132].However, even in experimental studies that have standardized methods, interpretation of results must be in context with, for example, the knowledge that the results of susceptibility to a particular pathogen may vary with host age, life history stage, population, the presence of abiotic factors (e.g., contaminants), biotic factors (e.g., competitors, predators), pathogen strain etc.
Experimental studies using different methods for the same host species illustrate the difficulties in making generalizations of how specific pathogens affect a host.For example, western toads (Anaxyrus boreas) have been investigated in a number of experimental studies (Table 1a).These studies used different Bd strains, different Bd doses and different life stages and the results of how the host was affected differed among the studies.For example, some studies showed reduced survival after exposure to Bd, whereas others did not.Even experiments by the same investigators [108,115] on western toads showed certain differences in how toads responded to Bd.In these studies, western toads were examined at the same life stage, but each study used different Bd strains and different Bd doses.
Small differences in experimental methods and design can lead to different results, highlighting the importance of standardized experimental protocols.Importantly, under controlled environmental conditions, observed effects after pathogen exposure can be attributed to intrinsic biological factors of the host, rather than environmental differences [206].  .Experimental studies published on Bd and Rv with respect to amphibian host genus and geographic range.Methods to generate the number of studies were produced in the same fashion as explained in Table 1.N indicates the number of studies for a particular region.

Conclusions
The initial sounding of the alarm for amphibian population declines in the 1990s [213] prompted a multitude of interdisciplinary investigations focused on understanding the causes of the declines.As part of this interdisciplinary approach, field observations along with well-designed experiments have helped us more fully understand the dynamics of amphibian population declines [214].Because disease is one of the key factors contributing to amphibian population declines, experiments have been especially useful in aiding our understanding of amphibian host-pathogen dynamics.Welldesigned experiments are useful tools that can provide unambiguous answers to specific questions about host-pathogen interactions.Several types of experiments have been employed.Field experiments are useful in mimicking natural conditions, but are not always feasible when investigating disease.Laboratory and mesocosm experiments have been used successfully to examine a variety of ecological processes [209,210], including various aspects of amphibian  1. N indicates the number of studies for a particular region.

Conclusions
The initial sounding of the alarm for amphibian population declines in the 1990s [213] prompted a multitude of interdisciplinary investigations focused on understanding the causes of the declines.As part of this interdisciplinary approach, field observations along with well-designed experiments have helped us more fully understand the dynamics of amphibian population declines [214].Because disease is one of the key factors contributing to amphibian population declines, experiments have been especially useful in aiding our understanding of amphibian host-pathogen dynamics.
Well-designed experiments are useful tools that can provide unambiguous answers to specific questions about host-pathogen interactions.Several types of experiments have been employed.Field experiments are useful in mimicking natural conditions, but are not always feasible when investigating disease.Laboratory and mesocosm experiments have been used successfully to examine a variety of ecological processes [209,210], including various aspects of amphibian population declines [214] and amphibian-pathogen dynamics (Table 1).
Studies of the three pathogens we focused on show that (1) host susceptibility varies with such factors as species, host age, life history stage, population and various ecological conditions including biotic (e.g., presence of competitors, predators) and abiotic conditions (e.g., temperature, presence of contaminants); (2) host susceptibility also depends upon the strain of the pathogen, to which they are exposed.The number of experimental studies of the three pathogens conducted on hosts at different life stages varied (Figure 4).Experimental studies and host survival showed clear differences with host life stage (Figures 5 and 6).Moreover, the dose of pathogen administered during susceptibility experiments is also important in interpreting results (Figure 7).
The issues we discussed in this paper illustrate some of the difficulties of standardizing experimental methods and interpreting and comparing results from studies that use different methods.As a baseline for standardization of experiments and to help interpret and compare the results of different experimental studies we recommend several protocols: (1) Collecting newly laid eggs and rearing them from larva through metamorphosis for experimentation lowers the likelihood that animals used in experiments were exposed to pathogens in the field; (2) the developmental stage, age, snout-vent length and mass of experimental animals should be reported; (3) abiotic conditions (e.g., temperature, humidity) during experimentation in the laboratory or field (mesocosm) should be recorded; (4) the duration of the study should be reported; (5) in susceptibility experiments, the method of exposure of hosts to the pathogen should be detailed.Important information would include dose parameters such as units used (e.g., #zoospores per unit volume); (6) explanation of the procedures used to quantify pathogen load should be reported in detail (e.g., qPCR); (7) the strain and if possible the origin of the strain of pathogen should be reported.Moreover, the age of the strain should be reported if possible because strain virulence may change while in culture; (8) treatments should be described fully and the number of individuals exposed to each treatment, including controls, should be reported.Many but not all studies include the parameters we listed above.Moreover, our list was not an exhaustive one but we feel that experiments reporting those parameters would aid researchers in interpreting and comparing results of different experimental studies.
We suggest future studies examine differences in susceptibility at the species and population levels as well as those that investigate strain variability, using controlled experiments.Controlled experimental studies examining differences in susceptibility to pathogens can aid in our understanding of the dynamics of epizootic outbreaks.Standardizing experimental methods is an essential component of investigating the role of pathogens in amphibian population declines.Moreover, studies that focus on a single cause contributing to amphibian population declines may underestimate the roles of multiple factors working simultaneously to cause both direct and indirect effects.Developing a mechanistic understanding of how biotic and abiotic factors can drive disease dynamics will allow us to better predict outbreaks and better manage and alleviate consequences associated with emerging infectious diseases [215].
Table 1.An overview of the effects of Bd (a), Bsal (b), and Rv (c) on amphibian species based on experimental studies.Publications were compiled using the search strings "Batrachochytrium dendrobatidis and amphibians", "ranavirus and amphibians" and "Batrachochytrium salamandrivorans and amphibians" in the Web of Science database from which duplicates and articles that were unrelated were removed.If one publication examined multiple species or host life stages, each species and life stage was reported separately.We have included each species International Union for Conservation of Nature (IUCN) Red List Status (http://www.iucnredlist.org),a widely recognized mechanism for assessing conservation status.Species of Least Concern (LC), Near Threatened (NT), Vulnerable (VU), Endangered (EN), and Critically Endangered (CE).na = not available.Reduced survival means mortality of hosts exposed to a pathogen was significantly higher than hosts in controls that were not exposed to a pathogen.* animals were not reared from eggs.** animals were not reared from eggs but were verified as Bd or Rv negative before the start of the experiment.*** collection information unavailable.Increased pathogen skin burden within two weeks of exposure, higher pathogen burden in deceased frogs, decrease in pathogen loads over time [245] Lithobates sylvaticus LC JEL 404, JEL 423 10 6 -10 7 zoospores and 10 5 -10 6 zoosporangia Larvae Reduced survival, no differences in growth or time to metamorphosis [86] JEL 404, JEL 423 10 6 -10 7 zoospores and 10 5 -10

Figure 1 .
Figure 1.Potential abiotic and biotic factors that may influence host-pathogen dynamics in amphibian disease systems.

Figure 3 .
Figure 3. Trends in all articles published on Bd (top) and Rv (bottom) in the literature over time.Publications were compiled using the search strings "Batrachochytrium dendrobatidis + amphibians" and "ranavirus and amphibians" in the Web of Science database, from which duplicates and articles that were unrelated were removed.The Bd search yielded a total of 1207 hits and the Rv search yielded 269 hits.

Figure 3 .
Figure 3. Trends in all articles published on Bd (top) and Rv (bottom) in the literature over time.Publications were compiled using the search strings "Batrachochytrium dendrobatidis + amphibians" and "ranavirus and amphibians" in the Web of Science database, from which duplicates and articles that were unrelated were removed.The Bd search yielded a total of 1207 hits and the Rv search yielded 269 hits.

Figure 4 .
Figure 4.The number of experimental studies conducted at a single life stage.Obtained from direct counts from Table 1.

Figure 4 .
Figure 4.The number of experimental studies conducted at a single life stage.Obtained from direct counts from Table 1.

Figure 5 .
Figure 5. Effects on survival in experimental studies.These data are direct counts from Table1.

Figure 6 .
Figure 6.Percentages of experiments showing reduced survival at a single life stage.These data are percentages from Table 1 (Experiments showing reduced survival/total # of experiments with survival as an endpoint).

Figure 7 .
Figure7.The effect of Bd dose (in log zoospores) on survival.These data are direct counts from Table1.Experiments that use multiple dose levels or multiple strains were excluded.Reduced survival means mortality of hosts exposed to Batrachochytrium was significantly higher than control mortality.Here, we display the minimum, first quartile, median, third quartile, and maximum zoospore dose regarding host survival.

Figure 8
Figure 8. Experimental studies published on Bd and Rv with respect to amphibian host genus and geographic range.Methods to generate the number of studies were produced in the same fashion as explained in Table1.N indicates the number of studies for a particular region.

Figure 8 .
Figure 8. Experimental studies published on Bd and Rv with respect to amphibian host genus and geographic range.Methods to generate the number of studies were produced in the same fashion as explained in Table1.N indicates the number of studies for a particular region.

.
Diversity 2018, 10, x FOR PEER REVIEW 7 of 48Figure 5. Effects on survival in experimental studies.These data are direct counts from

Table 1 .
% of exp. with reduced survivalFigure 6. Percentages of experiments showing reduced survival at a single life stage.These data are percentages from

Table 1 (
Experiments showing reduced survival/total # of experiments with survival as an endpoint).

Table 1 .
Diversity 2018, 10, x FOR PEER REVIEW 7 of 48 Figure 5. Effects on survival in experimental studies.These data are direct counts from

Table 1 .
% of exp. with reduced survivalFigure 6. Percentages of experiments showing reduced survival at a single life stage.These data are percentages from

Table 1 (
Experiments showing reduced survival/total # of experiments with survival as an endpoint).