Predator-induced plasticity in anuran larvae

Predator-induced plasticity in anuran larvae

Av Martin Johansson

Nedanstående text är ett universitetsarbete som gjordes i kursen "Biologisk litteraturstudie 7.5hp", vilken går ut på att skriva en rapport av vetenskaplig natur om något biologiskt. Hela arbetet (litteratursök, bearbetning, skrivande osv.) motsvara heltidsstudier under 4,5 veckor. Då detta är en review följer den inte riktigt standardmallen för sådana rapporter i form av rubriker. Detta är alltså ingen ny forskning eller egen studie som gjorts, utan snarare en (behövlig) sammanfattning av vad som redan gjorts i ämnet, med ett försök till att se över det hela kritiskt.

Arbetet i sig går ut på att ta reda på vad som har gjorts (och inte gjorts) vad gäller forskning kring fenotypisk plasticitet hos groddyngel, mer specifikt när de möter ett rovdjur. Fenotypisk plasticitet innebär att samma genotyp kan ge upphov till flera olika fenotyper beroende på omgivningen. Ett klassiskt exempel är frön med samma genotyp vilket blir ett träd om det planteras södra Sverige, men en buske om det planteras på ett fjäll. Jag kommer fram till att grodyngel uppvisar en otroligt fenotypisk plasticitet som respons mot rovdjur, på flera olika sätt, där både beteende och morfologi ändras drastiskt hos individer som växer upp i närheten av rovdjur.

Jag vet att artikeln är lång och bitvis tung att ta sig igenom, men för den herpetologiskt intresserade (och andra grodfanatiker) tror jag bestämt att man kan lära sig mycket från denna artikel. I övrigt hoppas jag att artikel kan fungera som inspiration till vad man kan göra för herpetologiskt arbeten i skollivet.

Slutligen vill jag påpeka att detta är ett just ett skolarbete, vilket innebär att det kan finnas konstiga meningar, stavfel och sådana otäckheter. Sedan denna review skrevs kan det tillkomma nya studier i ämnet, detta får ni också ha överseende med.


Predators greatly affect the behavior of their prey, and in some cases, they even change the morphology of their prey. This have been found to be true in several amphibians, and plenty of research have been conducted in order to investigate it fully. This report was conducted in order to review the current knowledge base in inducible defenses in anuran larvae, as well as examining what flaws in methodology that current exists. This was done by examining and summarize several articles, where a large focus were placed on species diversity, in order to determine where this kind of plasticity can be found, and how it is expressed. Inducible defenses as a form of plasticity is found to be ubiquitous in anuran larvae, where morphological, behavioral, and life-history traits changes as a response to predators. This report found that induced defenses are both prey- and predator-specific, creating a variety of different expressed phenotypes when faced with predators. Reversibility is also found in several studies, as well as the ability to gradually change traits in response to fluctuating threat-levels. Finally, the report discusses possible faults in methodology as well as give suggestion to future research in the subject of inducible defenses in anurans.

Keywords: Anuran larvae, induced defense, phenotypic plasticity, predator, development.

Table of content

  1. Introduction and background
  2. Materials and method
  3. How is the plasticity expressed
    1. Morphology
    2. Development and growth
    3. Behavior
  4. What kind of cues trigger the defenses?
  5. Are induced defenses gradual?
  6. Are induced defenses reversible?
  7. Does the induced phenotype affect individuals after metamorphosis?
  8. How widespread is this phenomenon amongst anurans?
  9. Conclusions
  10. References

In nature, there are countless of strategies used by prey to evade predation, ranging from behavioral tactics as simple as hiding, to complex morphological strategies as elaborate camouflage, sharp spines or toxins. However, there are survival strategies that might not attract as much attention in media as bright warning colors and camouflage which render species virtually invisible, but are nonetheless sophisticated adaptations for surviving predators. Phenotypic plasticity expressed as induced defenses is one of these survival strategies, and it is widespread among many taxa and can be expressed in many different ways (Wells 2007). Perhaps the most famous example is the rotifers, where offspring to individuals exposed to predators develops spines to avoid predation (Halbach 1971). Pattern of which this spectacular kind of plasticity is expressed can be described with norms of reaction. Where plasticity in species explains which trait changes to the environment, a norm of reaction describes exactly how the expressed phenotype changes as a function of variation in the environment. Plastic responses to predators are tightly connected to tradeoffs in lifehistory variables, since when the ability to evade predation increases, some other life-history trait decreases, otherwise the induced phenotype should be fixed, and no plasticity would be required.

Since the increased interest in phenotypic plasticity and induced defenses began in the 1990, scientist began to observe how amphibian larvae behave when faced with predators, since they truly are a perfect group of vertebrates to conduct such study on. Amphibians are the only group of vertebrates with such extreme changes in life-history, as they change dramatically in morphology from egg to adult frogs, and even change habitat from aquatic to terrestrial. They also typically have aquatic free-living larvae, which makes them more likely to encounter chemicals from nearby predators, making it far easier to ensure that predators affect the larvae, without actually putting the studied individuals at risk. Finally, amphibians larvae are unusually easy to study from a practical point of view in terms of feeding, caring, and monitoring, which is an immensely important part if several studies are to be made. Many studies have indeed been conducted since the 1990?s, with a variety of different anuran species and different results. Where one species changes morphologically in one way to reduce predation, another species might change in the exact opposite way to the same threat. Other studies have shown that the same anuran species develops differently depending on what predator it is exposed to, meaning that anurans are even more plastic than first suggested. This review is created in order to review and discuss several biases in methodology as well as update the knowledge base in this area, as no up-to-date overview exists.

The goal with this article is to give an overview of what is currently known about induced defenses in anuran larvae. At the end, a brief summarizing conclusion will be provided which will discuss the many results in this area. What research has been conducted, as well as what aspects that have not been studied, will be examined, while specifically answering the following questions:

The primary method used for finding relevant literature was searching the scientific database "Web of Science", with the keywords "tadpole development" together with "response to predators". Since time was an important limiting factor in this article, some constraints served the purpose of removing literature irrelevant to addressing the purpose of this review. Multifactorial designs had a lower priority since the aim was to summarize and review the current knowledge level of predators? direct effect on larvae. Articles from the begining of 1990 (and older) also had a low priority since many of these results could be found in later studies which replicated the experiments. Moreover, articles older than 1990 were hard to access, since they were seldom available in digital format. Lastly, active selection based on acquiring as many different species as possible were made, preferable with as many different families as possible, in order to provide a systematical variation in the reviewed species. This review also have a general focus on investigating species and their families, where mostly omitting genera, for two reasons. Firstly, genera still have a high level of uncertainty in anural systematics, which can be observed in several species, like Rana sylvatica, which have been move around frequently between Lithobates and Rana, with suggestions that the entire genus Lithobates is removed all together (Wells 2007, Pauly et al. 2009, Pyron and Wiens 2011). Similar relationship between Anaxyrus and Bufo is also found. This article will depend on recent phylogenetic studies (Pauly et al. 2009, Pyron and Wiens 2011), which classify all species previously in Lithobates in Rana, and all Anaxyrus in Bufo. Furthermore, only a few families among the reviewed species actually have representatives from different genera. This decision is made since outdated systematic can skew conclusions regarding where and how plasticity is found in anurans.

This article frequently uses the suborders Archaeobatrachia, Mesobatrachia and Neobatrachia. Although recent studies do suggest a removal of both Archaeobatrachia and Mesobatrachia, since molecular studies show a more gradient phylogeny than previously considered, the species within these groups still have fundamentally different anatomy from each other and species within the former Mesobatrachia suborder are still considered to be a evolutionary link between Archaeo- and Neobatrachia (Wells 2007, Pauly et al. 2009, Pyron et al. 2011.)

Perhaps the most spectacular part of the anuran defense strategy is their ability to alter their morphology when facing lethal predators. 36 species have been examined in regards to alteration in morphology, representing twelve families, and all three suborders of Anura (Table 1). The majority of these, 86%, showed some form of morphological plasticity when confronted with predators, with representatives from every family except Scaphiopodidae. Not all species changes the same traits as one another and not in the same way either (Relyea 2011, Nunes et al. 2014). However, there is one response in morphology that is found in the majority of examined species; an increased height of the tail fin (from a lateral view). This was observed in 26 species and thereby found in every morphologically responsive family except Bufonidae (Table 1) Out of the four examined bufonids, Bufo bufo and Bufo arabicus did not alter the tail height at all, while Bufo americanus actually decreased the height of the tail fin (Lardner 2000, Barry 2011, Relyea 2001 a, Nunes et al. 2014). However, the pattern is not quite as simple whether a species increase the tail fin height or not. Hyla versiocolor did increase tail fin depth when exposed to dragonfly larvae, but decreased the height when match with a species of predatory water bug. Similar patterns of predator-specific alterations were found in several other species as well. R. sylvatica did increase both tail fin height and tail muscle height when exposed to native fish predator, but the presence of dragonfly larvae only increased tail fin height (Relyea 2001 a). Elachistocleis bicolor on the other hand, reduced its body size when faced with a water bug, but increased tail fin height when reared with predatory fish (Gómez and Kehr 2012). These are good examples of predator-specific plasticity in anuran larvae, but the complexity does not end there. Prey-specific plasticity, as opposed to predator-specific plasticity, plays a great roll as well. This means that different types of anurans respond differently to the same threat. For example, shallower and longer tails seems to be a common response to fish, as it supposedly increases swimming speed (and thereby increase the ability to avoid predation by a pursuing predator). This is found in B. americanus, among other species (Relyea 2001 a). However, E. bicolor and R. sylvatica both increase tail fin height and tail muscle height, the direct opposite of the common response to pursuing predators (Relyea 2001 a and Gómez and Kehr 2012). This is likely to be the result of how they evade predators. They are both fairly stationary larvae, which means that they are not well suited to evade predators by swimming great distances. Instead, they rely on high acceleration to avoid predation, which is commonly how ambush predators, like water bugs or dragonfly larvae, are evaded. Acceleration is increased by higher tail fin depth and higher tail muscles, thereby increasing the method which they are already morphologically adapted to (Wells 2007). Another way of avoiding predation of ambush predators is to only increase the tail fin height, while letting the muscle height remain the same. This creates added non-vital tissue which acts as bait, causing the predators to strike there instead of at vital parts. Since the induced extra-tissue easily breaks, the predator loses its grip and the larva is involuntarily released.

Table 1. List of all species currently investigated. Sorted by family which is indicated by cell color. Yes = trait altered. No = Trait remain unchanged. Unknown = not been examined. Morph. = morphology, Dev. = development, Dist. = distribution, AP = Arabian Peninsula, AU = Australia, CA = Central America, E = Europe, NA = North America, SA = South America. Species summarized from indicated articles: 1 = Nunes et al. 2014, 2 = Lardner 2000, 3 = Vorndran et al. 2002, 4 = Relyea 2001 a, 5 = Barry 2011, 6 = Gallie et al. 2001, 7 = Bernard and Fordyce 2003, 8 = Chivers et al. 1999, 9 = Touchon and Warketin 2011, 10 = McCollum and Van Buskirk 1996, 11 = LaFiandra and Babbitt 2004, 12= Schoeppner and Relyea 2005, 13 = Relyea and Hoverman 2003, 14 = Smith and Van Buskirk 1995, 15 = Gómez and Kehr 2013, 16 = Gómez and Kehr 2011, 17 = Kraft et al. 2005, 18 = Lane and Mahony 2002, 19 = Gómez and Kehr 2012, 20 = Relyea 2007, 21 = Van Buskirk and Arioli 2002, 22 = Van Buskirk and Saxer 2001, 23 = Larner 1998, 24 = Kiesecker et al. 2002, 25 = Hossie and Murray 2012, 26 = Balaa and Blouin-Demers 2013, 27 = Babbitt 2001, 28 = Collier et al. 2011, 29 = Relyea 2004, 30 = Dayton and Fitzgerald 2011, 31 = Orizaola et al. 2012, 32 = Laurila et al. 1998.

Species Behavior Morph. Dev. Growth Dist.
Alytes cisternasii 1 Unknown Yes Increased No E
Bombina bombina 3 Unknown Yes Reduced Reduced E
Bombina variegata 3 Unknown Yes Reduced Reduced E
Bufo americanus 4, 6 Yes Yes No No NA
Bufo arabicus 5 Unknown Yes Unknown Increased AP
Bufo boreas 7, 8 Unknown Unknown Increased No NA
Bufo bufo 1, 2, 32 Yes Yes Increased Reduced E
Epidalea calamita 2 Unknown No No No E
Discoglossus galganoi 1 Unknown Yes Reduced Reduced E
Dendropsophus ebraccatus 9 Unknown Yes Reduced Reduced CA
Hyla arborea 1, 2 Unknown Yes Increased Increased E
Hyla chryoscelis 10 Yes Yes Unknown Unknown NA
Hyla femoralis 11 Unknown Yes Reduced Reduced NA
Hyla meridionalis 1 Unknown Yes No Increased E
Hyla versicolor 4, 12, 13 Yes Yes No No NA
Pseudacris crucifer 14 No No No No NA
Pseudacris triseriata 14 Yes Yes Increased Increased NA
Scinax nasicus 15 Unknown No Increased Reduced SA
Physalaemus albonotatus 16 Unknown Yes Increased Increased SA
Limnodynastes peronii 17 Unknown Yes Reduced Unknown AU
Limnodynastes tasmaniensis 18 Yes Unknown No No AU
Elachistocleis bicolor 19 Unknown Yes No No SA
Crinia signifera 18 Yes Unknown No No AU
Pelobates cultripes 1 Unknown Yes Increased Increased E
Pelobates fuscus 2 Unknown No No No E
Pelodytes ibericus 1 Unknown Yes No Increased E
Pelophylax kl. esculentus 20 Unknown Yes Reduced Unknown E
Pelophylax lessonae 21 Yes Yes Reduced Unknown E
Pelophylax perezi 1 Unknown Yes No No E
Pelophylax ridibundus 22 Unknown Yes Reduced Unknown E
Rana arvalis 2, 23 Unknown Yes Increased No E
Rana aurora 24 Unknown Unknown Increased Reduced NA
Rana catebeianus 4 Yes Yes No No NA
Rana clamitans 4 Yes Yes No No NA
Rana dalmatina 2 Unknown Yes No No E
Rana pipiens 4, 25, 26 Yes Yes No No NA
Rana sphenocephala 27, 28 Unknown Yes No Reduced NA
Rana sylvatica 4, 29 Yes Yes No No NA
Rana temporaria 2, 31, 32 Yes Yes Increased Reduced E
Scaphiopus couchii 30 No No No No NA

There is one trait which so far seems unique to a single species. Kraft et al. (2005) showed that Limnodynastes peronii develops a large hump of tissue dorsally, located where the notochord is most vulnerable from external force, creating a shield of tissue against predators. The uniqueness should be treated with caution, since the second limnodynastid species and the other Australian species were not investigated in regards to morphology (Lane and Mahony 2002).

There are some alterations in morphology, which seem to be constricted to certain groups of anurans. Bufonids seem to be more adapted than any other family to alter their body shape (Relyea 2001 a, Nunes et al. 2014). When reared with grasping predators like water bugs, bufonids reduce their body size, effectively minimizing the risk of being captured (Barry 2011). On the other hand, when exposed to gap size limited predators such as caudates, the bufonids excel at increasing their body size in every dimension, making it impossible for the predator to eat the larvae (Gallie et al. 2001, Relyea 2001 a).

Since five species did not show any kind of plasticity at all, it is impossible not to wonder what might cause this. Since plasticity is found in a wide range of families spread over fundamentally different branches of the anuran phylogeny, it is clear that lack of plasticity is not caused by phylogenetic restraints (Table 1). It is more likely that habitat, more specifically the breeding ponds, is the cause. Scaphiopus couchii shows no sign of plasticity what so ever (Dayton and Fitzgerald 2011), and breeds in the most extreme forms of ephemeral waters, where it hatches after 30 hours, and metamorphose within nine days (Wells 2007). This extreme rate of development is crucial since the ponds do not last much longer than approximately ten days. Therefore, one can assume that there simply is no room for plasticity, since it is already fine-tuned to such an extreme environment, and a tradeoff cannot be made since it is not possible to allocate resources from metamorphosing as quickly as possible. Two other ephemeral breeders, Epidalea calamita and Scinax nasicus, both face the same problem (although not as extreme), which is likely the cause for their inability to adapt as well (Lardner 2000, Gómez and Kehr 2013). The environment restriction is further backed up by the responses in Bombina bombina and Bombina variegata, which are genetically very similar, although B. variegata was more plastic in the altered traits (Vorndran et al. 2002). This is likely due, once again, to their breeding ponds, where B. variegate breeds in semi-permanent ponds and B. bombina breeds in permanent ponds, which causes a higher variation in predation pressure, and therefore benefits more from a higher plasticity than B. bombina. Pelobates fuscus did not show any sign of plasticity, but the author does claim that this is probably due to faults in the methodology, caused by the immense size of the larvae (Lardner 2000).

Even though it seems like increased tail height is by far the most common trait in morphological changes, it is important to consider the variation in predators. Dragonfly larvae are by far the most commonly used predator in studies conducted so far, followed by larvae of diving beetles and water bugs. Very few studies used species of fish or caudates. This might create a bias towards increased tail fin height, since the predators mostly use hunts in largely the same way.

Changes in life-history traits such as development and growth have been frequently examined in studies on inducible defenses in anuran larvae (Table 1). Older models predicted that a predator presence will cause amphibians to metamorphose earlier and at a smaller size, due to increased perceived mortality. In other words, it was thought that the response was to leave the dangerous environment as quickly as possible, and they would achieve that by allocating resources from growth to an accelerated development process, resulting in shorter larval period and smaller size at metamorphosis (Werner 1986, Relyea 2007). Examining several studies which investigated both development and growth shows that the majority (88%) of species do not follow the prediction of accelerated development and decelerated growth. As seen in Table 1, almost half of the examined species have unaltered development, and slightly more than half have an unaltered growth rate (Relyea 2001 a, Lane and Mahony 2002, Stamper et al. 2009, Barry 2011, Dayton and Fitzgerald 2011, Touchon and Warkentin 2011, Gómez and Kehr 2012, Gómez and Kehr 2013, Nunes et al. 2014). In fact, 24% of examined species including, but not restricted to both species of bombinatorids (Vorndran et al. 2002), several ranids (Van buskirk and Saxer 2001, Van Buskirk and Arioli 2002, Relyea 2007) and L. peronii (Kraft et al. 2005) show a slower rate of development, as opposed to older predictions. Only 29% appear to have an accelerated development rate in response to predators, (Smith and Van Buskirk 1995, Lardner 2000, Kiesecker et al. 2002, Laurila et al. 1998, Chivers et al. 1999, Gómez and Kehr 2011, Nunes et al. 2014). Out of these 11 species, only Rana aurora, Rana temporaria, B. bufo and S. nasicus follows the old prediction of an decreased growth rate in addition to the increased development rate. Three species had no change in growth rate, and the remaining 4 increased their growth rate as well. This clearly shows that old predictions are not only wrong, but greatly underestimated the complexity of induced defenses in anurans.

Changes in behavior have been found in almost every examined species so far (Smith and Van Buskirk 1995, McCollum and Buskirk 1996, Relyea 2001 a, Lane and Mahony 2002, Van Buskirk och Arioli 2002, Relyea 2007, Gómez and Kehr 2011, Orizaola et al. 2012, Gómez and Kehr 2013) with only two exceptions; S. couchii (Dayton and Fitzgerald 2011) and Pseudacris crucifer (Smith and Van Buskirk 1995), which showed no signs of behavioral plasticity. Every other species chose to reside on the opposite side of the enclosure, relative to the caged predator. Most species also reduce their feeding activity to minimize the risk of being exposed to the predator; however this is not the case with Limnodynastes tasmaniensis and Crinia signifera, who just avoided the predator-side, but foraged the same amount as before (Lane and Mahony 2002).

The answer to whether amphibians alter the behavior or not in the presence of a predator is often not as simple as a "yes or no". Several species of larval anurans change behavioral traits (predator avoidance and feeding activity) differently depending on what kind of predator it reacts to. For example, B. americanus does not actively avoid caged predators, unless the predator is a dragonfly larva. At the same time, the feeding activity of B. americanus is greatly reduced by several other kinds of predators, including the dragonfly larva, but not when the predator in questions is a caudate (Notophthalmus viridescens) (Relyea 2001 a). This kind of predator-specific response is consistent throughout the studies, which have examined behavioral traits in response to different kinds of predators (Relyea 2001 a, Gómez and Kehr 2011, Gómez and Kehr 2013). However, when comparing the number of studies which investigated behavior in contrast to how many that observed morphological changes, it is obvious that there is a great gap in knowledge about behavioral responses in different species (Table 1).

Amphibian larvae innately (Gallie et al. 2001) use waterborne chemicals to detect and react to predators in their surroundings (Van Buskirk and Arioli 2002, Schoeppner and Relyea 2005, Wells 2007, Collier et al. 2011, Hettyey et al. 2010, Balaa and Blouin-Demers 2012). There are two groups of chemicals that trigger and determine the extent of change in larval anurans, named kairomones and alarm cues. Kairomones are chemical cues emitted from the actual predator itself, whereas alarm cues are released by prey when being consumed by a predator (Balaa and Blouin-Demers 2012). Van Buskirk and Arioli (2002) found evidence for different kind of defenses being induced whether the larvae were exposed to either chemical alone, where alarm cues affected behavioral changes, and kairomones affected morphological changes. Their interpretation of this is that morphological traits take time to alter, and are therefore induced directly when a predator is present. Behavior, on the other hand, is instantaneously changeable, and the change can therefore be postponed until there is a real threat, i.e. the predator actually consumes prey in the vicinity. However, this simple pattern was later challenged when Schoeppner and Relyea (2005) claimed that both alarm cues and kairomones triggers behavioral change, but a mix of both is required for any morphological changes to occur. Schoeppner and Relyea (2005) did not actually include a "kairomones only"-treatment, only "alarm cue only" and "both", which mean that they can?t say for sure that Van Buskirk and Arioli (2002) were actually wrong in their previous statement. Even though they did use fairly related species (both ranid species), this could also be an effect of different species using different strategies to detect and react to pray. Collier et al. (2008) later found out that even visual cues matters, revealing even more complexity in how larval anurans detect predators.

The origin of the alarm cues is another important factor in anuran inducible defenses, as least in the two species that has been studied in this matter. H. versicolor showed great plasticity in behavior and morphology when exposed to alarm cues and kairomones from predators which fed on different species of amphibians (H. versiocolor, P. crucifer, R. sylvatica, Rana pipiens, Ambystoma maculatum), while the larvae showed no difference when exposed to alarm cues from predators which ate different kind of invertebrates (Scheoppner and Relyea 2005). A similar pattern were found in Rana sphenocephala, where alarm cues from conspecifics altered growth the most, and alarm cues from mosquito larvae resulted in intermediate change, relative to the control-treatment (Collier et al. 2008). Nunes et al. (2014) also found that larval anurans appear more plastic when exposed to native predators, in contrast to invasive predators, which the anurans have shared a very short evolutionary history with. This shows that the phylogenetic relationship of consumed prey does have some effect on the alarm cue and how larvae reacts to it, but only to a small extent. The larvae did not change differently depending on which species of amphibians the alarm cue came from, which might tell us that the phylogenetic is not of great importance. The authors claim that this is indeed proof that phylogenetic relationship of prey is of great importance, but that seems like an exaggerated claim, since the tested species are of little systematical variations, and the results show no difference between amphibians. Though it would be interesting to see a similar study with a greater spread of anuran species, Schoeppner and Relyea did use a caudate species (A. maculatum). Caudata are the sister group to all other living anurans, which makes is more distantly related to H. versicolor than any other anuran would be. This creates the assumption that phylogenetic distance of prey has little effect, other than if the prey is amphibian or not. Lastly, it would be interesting to see a study which exposes anuran larva to different alarm cues generate by amphibians species which differ in habitat and/or geographical distribution, rather than pure phylogeny.

Changed phenotypes as a response to the presence of predators comes with a cost, such as reduced feeding activity (Relyea 2002). Some kind of negative side-effect is expected, since the predator-induced phenotype needs to have a disadvantage relative to the non-predator phenotype, when the predator is absent. If no tradeoff is present, the change should not be a plastic trait, but rather fixed. A study did find that reduced feeding activity as a response to predators does not actually reduce the amount of ingested food, presumably since the physiology of the intestines changes to more efficiently process food (Steiner 2007). However, given the uncertainness of the study, and the fact that this trait should always give a higher fitness in both the presence and absence of predators, this result is probably a result of a lacking methodology. Knowing that an induce defense comes at a cost, the question whether larvae can adapt on a smaller temporal scale to variation in predator threat-levels, presents itself. This has only been tested a few times, in a few species, with a few species of predators. However, the studies have shown that both R. temporaria and R. pipiens indeed have reversible defenses when the predator is removed (Orizaola et al. 2012, Hossie and Murray 2012). Orizaola et al. (2012) even found that the time it takes to reverse traits were equal to the amount of time it took to induce the defense. In R. temporaria, behavioral traits only took one hour to alter (and alter back), where morphological traits unsurprisingly needed more time, approximately one week. R. pipiens, on the other hand, needed two weeks to both alter and reverse morphological traits. A plausible explanation for this is that the individuals of R. temporaria where collected from high latitude ponds, where the numbers of days with high temperature are low, meaning the larvae need to be able to adapt quickly to environmental hazards in order to reach metamorphosis. It was further determined that the change in both behavior and morphology are fully reversible, meaning that individuals with reversed traits were indistinguishable from the control-treatments. Even more impressive, besides from a small increase in development time, no negative effect on life-history traits or any other traits were found due to the reversing on traits. Hossie and Murray (2012) also examined how different temporal patterns of exposure to predators affect changes in both morphology and behavior, finding once again that the larvae do indeed respond to changes in predator pressure. When larvae were exposed to predators in a biweekly pattern, they showed intermediate change in morphology compared to constant exposure and no exposure, showing that they follow the pressure-level, but with an expected time lag.

As previously stated, the phenotype induced from predators is only beneficial when a predator is actually present, and an unnecessary hindrance when a predator is absent. Reversibility of traits is therefore an important aspect of adapting to the environment, developing anti-predatory traits when necessary, and de-develop the traits when they are not needed anymore. However, there is another important aspect when adapting to predators; the actual level of danger the predator poses. Since anuran larvae continually adjust to whether a predator is present or not, they might be able to adjust to the actual number of predators in the vicinity. Pelophylax lessonae and R. temporaria do indeed show a gradual change to perceived predation risk in several morphological traits, such as shape of the body, body length, tail shape and tail fin height, where the change is greater with a greater number of predators (Van Buskirk and Arioli 2002, Orizaola et al. 2012). The authors did find one trait, tail length, which appears to not be gradual. A second study, using R. sylvatica, found largely the same results, although they did unfortunately not measure tail length, making it harder to hypothesize whether the lack of gradual tail length is something unique to P. lessonae or not (Relyea 2004).

Galliet et al. (2001) found B. americanus increases body width and length until the gap-size limited predator (N. viridescens) cannot swallow it. After that cutoff level is reached, B. americanus larvae resume their normal development. This further increases anurans? capability to separate different predators, and adjust accordingly.

While inducible defenses during the larval stage have been studied extensively throughout the years, few studies have investigated whether exposure to predators during larval stages cross over to the post-metamorphic juveniles. Even though the studies are rare, some results address this question. Three studies examined the post-metamorphic morphology in three different species. R. sylvatica juveniles changed in regard to leg length, where predatorexposed larvae developed slightly longer legs after metamorphosis (Relyea 2001 b). Meanwhile, another study showed that Pelophylax ridibundus developed shorter and more muscular legs, when the larvae were exposed to predators (Van Buskirk and Saxer 2001). The third article stated that juveniles of R. temporaria developed thinner bodies and femurs, but this morphological change only appeared 12 weeks post-metamorphosis (Stamper et al. 2009).

The difference in morphological changes in leg structures between R. sylvatica and P. ridibundus are in line with one would expect of those two species, if one were to assume that juveniles adapts morphologically to predators. R. sylvatica is a mainly terrestrial species, where longer legs can increase jump distance, which is this species? primary method of escaping predators. P. ridibundus, on the other hand, is an almost exclusively aquatic species, where broader legs enhance swimming ability (Wells 2007). The authors explain the decrease in leg length as a byproduct of reduced development or growth during the larval stage. However, they did not actually measure growth or development (Van Buskirk and Sazer 2001).

Other studies examined other aspects than pure morphological changes and found various results. Growth rate and survival of juvenile H. versicolor did not change at all in contrast to whether the larvae were exposed to predators or not (Relyea and Hoverman 2003). The efficiency of swimming, as a proxy for predator prevention, was examined in R. temporaria, where juveniles who were exposed to predators during their larval stage swam at a slower speed than non-predator larvae. They also swam shorter distances in response to being threatened (Stamper et al 2009). The author regarded this reduced survival capability as a trade-off for the increased development rate during the larval stage, perhaps by a reduce amount of mitochondria. This difference did disappear after another four weeks, meaning that disadvantages from the tradeoff was in this case not permanent. Finally, Bernard and Fordyce (2003) showed that juvenile Bufo boreas, a species known to change both morphology and development as a response to predators (Chivers et al. 1999), had a threefold concentration of bufadienolides (toxins) in their skin in if the larvae were raised in the presence of predators. However, the concentration was not elevated at all in the larvae. This seems like a strange adaptation, since the metamorphosed individual is primarily terrestrial. Since it has been established that most species can adapt to different kinds of predators, an elevated toxin-concentration as metamorphosed toads seems like an expensive (and perhaps unnecessary) response to a threat that might not exist when the terrestrial stage of their lives begin. This is especially true when taking into account that the study found that juveniles which as larvae lived with predators where consumed at a faster rate than juvenile which did not encounter predators. This is presumably due to the added cost of higher quantities of toxin, which came at the expense of general stamina (and therefore the ability to struggle when being eaten). Although, it is important to note that other toxins might have increased in the larvae, but the authors did not measure toxins other than bufadienolides.

When totaling the numbers of examined anuran species in the reviewed studies, it is clear that the subject overall lacks the knowledge to actually determine whether induced defenses are something all species are capable of. Only 40 species, divided over twelve families, have currently been examined. When this is compared to the approximately 5500 described species of anurans, distributed over approximately 50 families, it is clear that a systematical widespread examination is missing (Pyron and Wiens 2011). It is also important to consider the systematical placement and phylogenetic relationships of the studied species as well. As illustrated below (Table 2), the majority of examined species and families all belong to the suborder Neobatrachia. While this is an obvious bias towards neobatrachian species in absolute numbers, the percentage actually reveals that Neobatrachia is in fact the least studied group of anurans. In addition to the low percentage of studied species in Neobatrachia, the majority of them are confined within three families (Bufonidae, Hylidae and Ranidae). Indeed, 38% of the investigated species within Neobatrachia all come from Ranidae, where nine of them are placed in the same genus, Rana (Pauly et al. 2009, Pyron and Wiens 2011). The last 4 ranids are all from Pelophylax, which are closely related to Rana (Pyron and Wiens 2011). 25% are hylids, where the majority is in the Hyla genus. Lastly, the variation in geographical distribution is also lacking. 45% of all species are distributed in Europe and 35% in North America, and the remaining 20% are divided on the Arabian Peninsula, Australia, Central- and South America (Table 1).

These aspects put together create little phylogenetic variation in the studies of anurans pasticity. However, some conclusions can be made from these studies. Even though a majority of species expressed plasticity in morphology, life-traits or development rate, some did indeed only alter their behavior. However, two species (P. crucifer and S. couchii) showed no signs of change in behavior, morphology, or life-traits, and therefore remained unchanged during exposure to predators (Smith and Van Buskirk 1995, Dayton and Fitzgerald 2011). E. calamita showed no form of plasticity as well, but behavior have never been studied in this species (Lardner et al. 2000, Nunes et al. 2014). These three species are from different families (Hylidae, Scaphiopodidae and Bufonidae respectively) and indeed even different suborders (Neobatrachia and Mesobatrachia). Since several other bufonids (including the very closely related B. bufo) do have the ability to alter several kinds of traits in response to predators, it is unlikely that there is a strict phylogenetic restriction that causes this in E. calamita (Chivers et al. 1999, Lardner 2000, Relyea 2001 a, Barry 2011, Nunez et al. 2014). The same can be said about P. crucifer, since several hylids have induced defenses (Smith and Van Buskirk 1995, McCollum and Van Buskirk 1996, Relyea 2001 a, LaFiandra and Babbitt 2004, Touchen and Warketin 2011, Gómez and Kehr 2013, Nunes et al. 2014). It is harder to draw the same conclusion about S. couchii, since no other Schaphiopoid have been studied. However, both S. couchii and E. calamita share a common trait, even though they are separated greatly through both phylogeny and geographical distribution (Wells 2007, Pyron and Weins 2011). They are both breeding in very ephemeral ponds, where the desertliving S. couchii does so in extremely ephemeral ponds, not uncommonly lasting only nine days before drying out (Wells 2007). This very short time to hatch, develop, and metamorphose, might create a non-tolerance for plasticity, since they are already optimized to be able to metamorphose in such small amount of time. Predator-induced changes in these tadpoles are probably enough to lower the chance to make it out of the pond, and have therefore been selected against during their evolutionary history.

Table 2. Summary of number of studied species and families in relation to total amount of described species and families. Note that due to constantly changing systematic and classification, the number of described species in both Neobatrachia and Mesobatrachia are approximations.

Archeobatrachia Mesobatrachia Neobatrachia
Species Families Species Families Species Families
Studied 4 4 4 3 32 6
Described 27 5 168 6 5500 50
Percentage 15% 80% 2,4% 50% 0,6% 4%

With the current data set, it is impossible to say for sure whether most anuran species have induced defenses or not. Since approximately 75% of all families lack studied species, it is impossible to say if certain families have lost this ability all together. It is even possible that entire branches of the anuran tree have lost this ability all together. However, since it appears in all three suborders, it is unlikely to be a new trait to emerge in anurans. It is also possible, and supported by the studies done so far, that lack of induced defenses in response to predators are not due to phylogeny, but rather the environment of certain species. According to the studies made so far, 86% of all species do possess the ability to somehow change in response to predators, but the sample size is, of course, to small and to biased in regard to variation, to apply this number to the entire order Anura.

Publication bias is factor which could be of great importance in determining where inducible defenses are present in anurans. Potentially, the number of examined anuran species could be higher, but if no difference were found between the predator and control-treatment, it is very unlikely that such a study would be published, unless it is published as a response to another study. This is very problematic indeed, since a "no response"-result is not the same a "no result".

Prior to 2001, only four species of anurans had been investigated with regards to morphological and life-history responses to predators, where three of the species adapted to the predators through an increased tail fin height (McCollum and Van Buskirk 1996, Van Buskirk and Relyea 1998, Relyea 2001 a) and one species showed no plasticity in morphology (Smith and Van Buskirk 1995). It later became clear that an increased tail fin height is only one of many adaptations to evade predators. Tail length, tail muscle height and body proportions are traits frequently altered as well. It is also apparent that anurans are capable of determine what kind of predator it is exposed to, and react accordingly, where ambush predators generally induce traits which increase acceleration, while pursuing predators alter the morphology which determine swimming speed. Exposure to gap-size limited predators, such as caudates, often results in wider and higher bodies in anurans, making it harder to swallow the larvae. Additionally, the plasticity itself seems plastic, since the morphological change also depend on original behavior and anatomy of the species. Development and growth rate show a variety of combination in response to predators, making it hard to decipher overall patterns. Where some studies suggest that morphological traits alters as a response to kairomones and behavior reacts to alarm cues (Van Buskirk and Arioli 2002), others studies suggest that the relationship between kairomones, alarm cues and even visual cues are far more complicated than originally thought (Schoeppner and Relyea 2005). Indeed, the equation is still not entirely solved (Hettyey et al. 2010)

Anuran larvae do not only adapt to different predators depending on their capture method, anurans also adapt to the probability of actually being caught, changing gradually as more and more predators populates their environment. However, scientists are still unsure whether all traits shows this kind of gradualness or not, since different results have been found in different studies (Van Buskirk and Ariolo 2002, Relyea 2004). Perhaps the gradualness of different morphological traits are different from one anuran species to another, something that can only be answered by studying this specific aspect of plasticity in more species. The same can be said about the reversibility of induced defenses, which is very obvious in the few species that has been studied, but then again, very few have indeed been studied in this manner (Steiner 2007, Orizaola et al. 2012, Hossie and Murray 2012).

The changes caused by predators in the anuran larval stage seem to have limited carry-over effect to the adult life stage. Toxin levels seem to increase drastically, but so far the carry-over effect seem to be limited to minor changes in limb-morphology, with no long-lasting effect on growth or development (Relyea 2001 b, Van Buskirk and Sazer 2001, Relyea and Hoverman 2003, Bernard and Fordyce 2003, Stamper et al. 2009). This is surprising, since there should be some trade-off in inducing plastic changes, otherwise the induced phenotype should be fixed, since it only provides advantages with no apparent cost. Perhaps the drastic metamorphose decouple the two life-stages to such a degree that development and survival are mostly unrelated between life-stages.

After reviewing 40 species, it is clear that inducible defenses are ubiquitous among anurans (Table 1). There appears to be some patterns, and so far they only seem environmentally determined rather than of phylogenetic nature. Low variation in systematic and geographic distribution of examined species makes it hard to make further conclusions about the commonness of anuran inducible defenses.

There are several approaches future research could take in order to further broaden the insights on the plasticity of anuran defenses. A wide study on basic morphological traits with several species from so far unexplored families should be conducted. If such a study would include many cross-road species such as Ascaphidae and Leiopelmatidae, together being the sister group to all other living anurans, important conclusions could be made regarding the origin and evolution of their extraordinary plasticity. Similar approaches with geographic distribution and habitat as main focus would also provide useful information about how different environmental cues affect the ability to adapt to predators. Since the vast majority of studies conducted today used dragonfly larvae (Anax and Aeshna), and the rest have only examined four other types of predators, it would be interesting to see a greater variety in the choice of predator. Other methodological variables should be more carefully examined when conducting further studies. Where some use mesocosm (e.g. Relyea and Hoverman 2003), representing a fairly natural habitat, others use a strict lab environment with aquariums (e.g. Relyea 2001 a). Feeding is also varied from project to project, ranging from no additional feeding (only found in mesocosm-studies) and ad libitum feeding to a constant amount of food added with a certain interval. Such factors could affect traits such as development and growth and perhaps partially explain some patterns regarding a reduced feeding activity combined with an unchanged or increased growth rate.

Finally, future research will have to take into account several rarely considered variables. There appears to be a cross-over effect from anuran egg to larvae, when the eggs are exposed to predator-kairomoes. This could skew results, since differences found could be due to the fact that eggs were exposed to predators before the study (and collecting of eggs) began (Saglio and Mandrillon 2006). Studies have also shown that there appears to be an interaction effect between the effect of water temperature and the effect of predators (Touchon and Warketin 2011), as well as the effects of density and the effects of predators (Gómez and Kehr 2013). Larvae within the same species which originate from ponds of different predation pressure have different timing on development (Lardner 1998). Such easily overlooked effects could very well undermine future research, as results could be skewed unknowingly by the norm of reaction from other factors than predators, and perhaps this has already happened with current studies. Nonetheless, more research is required if inducible defenses in anuran larvae are to be fully understood.

Artikel skriven av: MartinJ
Publicerades 2015-02-16
Visningar: 4265st


Du behöver vara inloggad för att se och skriva kommentarer
Skriv kommentar