Identifying prey

A survey of the literature suggests that the molecular identification of prey remains is becoming an increasingly popular tool. If we know what animals are feeding on, we will have a better understanding of ecology at both the species and community levels because feeding is relevant to predator--prey and host-parasite interactions, food chains, intra-and interspecific competition, and niche partitioning. Identifying prey is sometimes possible through observations of predator-prey interactions, although this approach is often impractical for small, secretive or wideranging species. Prey remains from gut contents or faeces can sometimes be identified from fragments of bone, carapace, seeds, feathers, scales or other resistant parts, although the capture, ingestion and digestion of prey often leaves characters too mutilated for identification. Furthermore, soft prey items such as slugs or invertebrate eggs are unlikely to leave any identifiable remains. Molecular identification of prey remains therefore presents a useful alternative because it does not require morphological features to be preserved.

Provided that appropriate primers are available, prey items can often be identified by amplifying DNA from composite samples such as faeces, gut contents or even the entire predator, and then matching the characterized DNA to sequences or allele sizes from existing DNA databases. Deep-sea marine invertebrates provide a good example of animals whose feeding habits cannot be observed easily and whose prey are unlikely to be recognizable on the basis of morphology once they have been partially digested. In a recent study, Blankenship and Yayanos (2005) used universal cytochrome c oxidase I primers to amplify copies of this mitochondrial gene from the stomach contents of the deep-sea crustaceans Scopelocheirus schellenbergi and Eurythenes gryllus from the Tonga Trench. They found that both species fed on a wide range of prey, not all of which would have been carrion. This was an unexpected result because both species were previously believed to feed exclusively as scavengers, but the identity of some of their prey items means that they also may be predators.

Studies such as this are extremely useful because they represent a novel way of quantifying the prey items of relatively inaccessible species. However, although the technique is relatively straightforward, a note of caution is in order. Potential problems associated with degraded DNA, inhibitors and contamination of gut or faecal material are just some of the reasons why the molecular identification of

Figure 6.13 A coyote (Canis latrans). Food preferences in this omnivorous species can be identified from faecal genotypes. Author's photograph

prey can be a rather demanding task that requires meticulous attention to all aspects of sample collection, preservation and laboratory work.

Individual food preferences

One interesting outcome of using molecular techniques to identify prey remains is that it can help us to understand which prey items individual predators are choosing. In one study, the food habits of coyotes (Canis latrans; Figure 6.13) in the California Santa Monica Mountains were investigated by researchers who collected and genotyped 115 coyote faeces (Fedriani and Kohn, 2001). Sequence data were used to identify which species the prey items belonged to, and microsatellite data were used to generate individual coyote genotypes that would tell the researchers which coyote had left each faecal sample. This is possible because faeces contain some cells from the gut of the animal that left them, and if variable markers such as microsatellites are used then it is possible to generate individual-specific genetic profiles that will link each sample to an individual animal (see Box 6.5). Sequence data were obtained from up to 11 faeces from each coyote, and from these it was possible to calculate the percentage of each animal's diet that was made up of small mammals, other vertebrates, invertebrates, fruit and rubbish. Although the majority of coyotes fulfilled expectations by taking small mammals as their primary food source, 18 per cent of the sampled coyotes had an alternative primary food source (Figure 6.14), which is a substantial portion of the population whose food preferences would have been overlooked if all prey items had been pooled.

Individual coyotes

Figure 6.14 A breakdown of individual coyote diets. Microsatellite data were used to assign faeces to individual coyotes, and mitochondrial sequence data were used to identify faecal prey remains. Adapted from Fedriani, and Kohn (2001)

Individual coyotes

Figure 6.14 A breakdown of individual coyote diets. Microsatellite data were used to assign faeces to individual coyotes, and mitochondrial sequence data were used to identify faecal prey remains. Adapted from Fedriani, and Kohn (2001)

6.5 Probability of identity

Matching samples such as faeces to individual animals is possible only if the molecular markers being used are sufficiently variable to generate a different genotype for each individual within a population. We can test for this by calculating the probability of identity (PI), which is the likelihood that two individuals chosen at random from the population will have the same genotype. A low probability of identity means that two or more individuals are unlikely to share the same genotype, in which case we would be confident that a sample has been linked to the correct individual. PI values are most commonly based on microsatellite loci, and before to calculate PI we need to know the frequency of each relevant allele within the population because high-frequency alleles are much more likely to be found in multiple individuals than are low-frequency alleles. One of the simplest ways to calculate the likelihood of two individuals in a population sharing the same genotype was developed for a study of black bears using the following equation (Paetkau and Strobeck, 1994):

where pi and pj are the frequencies of the ith and jth alleles at each locus in a given population. The probability of identity is revisited in the context of wildlife forensics in Chapter 8.

Probability of identity can be a useful aid to tracking individual animals that are rare or difficult to recapture. In one study, researchers wished to follow the movements of individual wolves (Canis lupus) from packs that had recolonized the Italian Alps after a century-long absence, and they were able to do this by matching faecal samples to individuals on the basis of multi-locus microsatellite genotypes (Lucchini et al., 2002). In another study, a population of the highly endangered hairy-nosed wombat (Ladiorhinus drefftii) was censused accurately by determining individual genotypes from hairs that were collected at the entrances to their burrows, a method that presented no risk to the individual wombats (Sloane et al., 2000).

Biological control

A practical application of the molecular identification of prey was illustrated by a study of Linyphiid spiders. These are important aphid control agents in arable fields, although they supplement their diet with other invertebrates of higher nutritional value such as Collembola. Sequences from mitochondrial cyto-chrome oxidase I were used to determine which collembola species the spiders were consuming, and it turned out that a high percentage of spiders ate the collembolan Isotoma anglicana even though this was relatively scarce in the field. On the other hand the most common collembolan, Lepidocyrtus cyaneus, was seldom eaten by the spiders. The spiders therefore were clearly choosing their prey items on the basis of more than simply abundance. A practical outcome of this study is the knowledge that increasing the population of I. anglicana could indirectly help to keep the aphid population under control by boosting the lynyphiid spider population (Agusti et al., 2003). Other examples of the molecular identification of prey from arthropod guts are given in Table 6.7.

Predation and conservation

Predator--prey interactions can also be important in conservation biology. The coccinellid beetle Halmus chalybeus was introduced into the Hawaiian island Kaua'i in 1894 as a biological control of homopteran pests, but the recent invasion of these beetles into the Alaka'i swamp is now a cause for concern because many coccinellid beetles are generalist predators. In a recent study, DNA analysis was used to identify their various prey items, which turned out to include some species of considerable conservation value, notably endemic moths of the genera Scotor-ythra and Eupithecia (Sheppard et al., 2004).

Table 6.7 Some of the prey items that have been identified from PCR amplification of arthropod gut contents using various molecular markers



Molecular marker


Mirid bug

Cotton bollworm


Agusti, De Vicente

(Dicyphus tamaninii)


and Gabarra



Carabid beetles

Mosquitoes (Culex

Esterase genes

Zaidi et al. (1999)

(Pterostichus cupreus)



Ladybird beetles

Cereal aphid

Cytochrome oxidase

Chen et al. (2000)



II (mitochondrial)

convergens) and


lacewings (Chrysoperla


Lady beetle

European corn

ITS 1 (internal

Hoogendoorn and


borer (Ostrinia

transcribed spacer

Heimpel (2001)



1 of nuclear


Mite (Anystis

Apple-grass aphid





dehydrogenase 1

Fleming and



Murchie (2003)

Ground beetle

Slugs (Arion


Harper et al.


intermedius and

oxidase I and



Deroceras reticulatum),

12S rRNA

grass snail (Vallonia


pulchella), aphids

(Aphis fabae,


padi), beetles

(Sitona sp.)

In the previous example an introduced species was jeopardizing native species through predation, but the feeding behaviour of introduced species may also threaten native species through competition for the same food sources. Biologists were concerned that a social wasp (Vespula germanica) that was introduced into Australia in the 1970s may be outcompeting a native wasp (Polistes humilis). They therefore sequenced a portion of the 16S rDNA mitochondrial gene from both species' prey items, and used these sequences to determine which invertebrate orders were being preyed upon by each species. They found that the native wasp P. humilis fed almost exclusively on Lepidoptera, whereas the introduced wasp V. germanica fed on a wide range of invertebrates including Lepidoptera, Hemiptera, Diptera, Orthoptera and Odonata. These results show that P. humilis is a specialist whereas V germanica is a generalist; furthermore, comparisons with earlier studies revealed variations in the diet of V. germanica, suggesting that it is an opportunist that feeds on whatever prey is readily available. Generalism and opportunism are two characteristics that help invasive species to spread rapidly across their new terrain and give them the upper hand when competing with native species for food and other resources (Kasper et al., 2004).

A final example of the relevance of feeding behaviour to conservation biology comes from a study in which microsatellite markers were used to identify which waterfowl chicks the glaucous gull (Larus hyperboreus) was eating in Alaska. Of particular concern was the possibility that the gulls were contributing to the decline of emperor geese (Chen canagica) and spectacled eiders (Somateria fischeri) by eating their young. An examination of the gut contents revealed no evidence of predation on the spectacled eider, although 26 per cent of all gulls examined had eaten emperor geese (Scribner and Bowman, 1998). Other species eaten were the white-fronted goose (Anser albifrons) and the cackling Canada goose (Branta canadensis minima). Glaucous gulls seem to feed preferentially on goslings, although they do not appear to discriminate between the different species of geese because they feed on them in proportion to their availability (Bowman, Stehn and Scribner, 2004).

Was this article helpful?

0 0

Post a comment