Molecular Ecology Resources (2011) 11, 620628

doi: 10.1111/j.1755-0998.2011.03001.x

Molecular scatology: how to improve prey DNA detection success in avian faeces?
JOHANNES OEHM, ANITA JUEN, KARIN NAGILLER, SIGRID NEUHAUSER* and M I C H A E L TRAUGOTT Mountain Agriculture Research Unit, Institute of Ecology, University of Innsbruck, Technikerstrasse 25, 6020 Innsbruck, Austria

Abstract
The analysis of prey DNA in faeces is a non-invasive approach to examine the diet of birds. However, it is poorly known how gut transition time, environmental factors and laboratory treatments such as storage conditions or DNA extraction procedures affect the detection success of prey DNA. Here, we examined several of these factors using faeces from carrion crows fed with insect larvae. Faeces produced between 30 min and 4 h post-feeding tested positive for insect DNA, representing the gut transition time. Prey detection was not only possible in fresh but also in 5-day-old faeces. The type of surface the faeces were placed on for these 5 days, however, affected prey DNA detection success: samples placed on soil provided the lowest rate of positives compared to faeces left on leaves, on branches and within plastic tubes. Exposing faeces to sunlight and rain signicantly lowered prey DNA detection rates (17% and 68% positives in exposed and protected samples, respectively). Storing faeces in ethanol or in the freezer did not affect molecular prey detection. Extracting DNA directly from larger pieces of faecal pellets resulted in signicantly higher prey detection rates than when using small amounts of homogenized faeces. A cetyltrimethyl ammonium bromide-based DNA extraction protocol yielded signicantly higher DNA detection rates (60%) than three commercial kits, however, for small amounts of homogenized faeces only. Our results suggest that collecting faeces from smooth, clean and non-absorbing surfaces, protected from sunlight and rain, improves DNA detection success in avian faeces. Keywords: Corvus corone corone, diet analysis, feces, Melolontha melolontha, prey detection, Tenebrio molitor, transition time, trophic interactions Received 12 November 2010; revision received 20 January 2011; accepted 26 January 2011

Introduction
Examining the dietary choices of birds is needed to understand their biology and the role of birds in food webs as well as to protect and manage them. To investigate the diet of birds under natural conditions, the stomach content has been analysed after killing and dissecting them (Miller & McEwen 1995; Scribner & Bowman 1998). Gut content analysis where the consumer has to be killed, however, is limited because of ethical reasons. Hence, non-lethal approaches have been used as well, including the use of neck-collars (Exnerova et al. 2003), ushing the crop or stomach (Hull 1999; Moorman et al. 2007), forced regurgitation (Valera et al. 1997) or collecting prey remains from boluses and faeces to retrieve dietary
Correspondence: Dr Michael Traugott, Fax: ++43 512 507 2817; E-mail: Michael.Traugott@uibk.ac.at *Present address: Institute of Microbiology, University of Innsbruck, Technikerstrasse 25, 6020 Innsbruck, Austria.

remnants for morphological identication (Xavier et al. 2005; Gamez-Virues et al. 2007). Although the latter approach is non-invasive, identication of the prey remains is often impossible as soft-bodied prey, such as insect larvae, rarely provides morphologically discernable remains. Moreover, although stomach ushing may be practicable for large and ratite birds (e.g. penguins), catching and treating passerine birds this way is much more difcult and might be harmful to them. Molecular approaches of prey detection (King et al. 2008) are a versatile tool to examine birds feeding ecology, for example, Jarman et al. (2002) identied krill species from both stomach contents and faeces of Adelie penguins (Pygoscelis adeliae), while Deagle et al. (2007, 2010) investigated the diet of two other penguin species (Eudyptes chrysolophus and Eudyptula minor) by either cloning and sequencing of PCR products or pyrosequencing 16S mtDNA of prey in faeces. Similarly, Sutherland (2000) identied invertebrate prey DNA from faeces of ve

2011 Blackwell Publishing Ltd

I M P R O V I N G P R E Y D N A D E T E C T I O N I N A V I A N F A E C E S 621 passerine bird species (Sturnus vulgaris, Emberiza citronella, Parus major, Parus caeruleus and Hirundo rustica) using restriction digest, cloning and sequencing. Hence, DNA-based analysis of faecal samples provides a clear advantage to study the diets of birds: (i) it is non-lethal and non-invasive and (ii) it allows identifying the dietary remains on a species-specic level. However, currently, little information is available examining DNA-based faecal analysis on a more experimental basis. Here, we present experiments, conducted with carrion crows and insect prey. First, we examined the minimum and maximum time after prey ingestion during which insect prey DNA can be amplied from faecal material. Then, we investigated how environmental conditions, such as sunlight and rain or different surfaces the faeces were placed on, affected prey DNA detection rates. Finally, sample storage and DNA extraction protocols were compared with regard to prey DNA detection success. second time, as the rst extracts tested negative in a PCR with general metazoan primers.

The inuence of the surface underlying the faeces on prey DNA detection
For this and all following experiments, four carrion crows were fed with larvae of Tenebrio molitor (Coleoptera: Tenebrionidae). Feeding experiments were conducted between 8th and 15th August 2007 in aviaries of the Institute of Ecology at the University of Innsbruck. Each bird was starved overnight and fed with 30 mealworms in the morning. One and 2 h after this initial feeding event, each bird received another 30 mealworms (i.e. 90 mealworms per bird in 3 h). One hour after each feeding event, all droppings produced within 1 h were collected with plastic spoons and transferred individually into collection tubes (one dropping per tube). After the three mealworm feeding events, grains (Avena sativa, Zea mays), fruits (Malus sp., Vitis vinifera), cheese (from Bos taurus) and cat food (including Bos taurus and Gallus gallus) were offered to the crows as standard food. To facilitate faecal collection and to ensure that only droppings produced within the feeding experiments were taken, the basement of each aviary was covered with a thin plastic foil every morning before the feeding experiments started. One hundred freshly collected droppings were transferred onto four different surfaces: open plastic tubes, maple leaves, willow branches and potting soil. Ten protected (under cover of roof) and 15 unprotected droppings per surface were exposed to direct sunlight and rain. The droppings were left on the surfaces for 5 days (17 August 200721 August 2007), and a data logger recorded air temperature and humidity. Thereafter, all remaining pellets were scratched from the surfaces with plastic spoons. DNA was extracted following the CTAB protocol (Juen & Traugott 2005).

Materials and methods Gut transition time of insect prey

To determine how early and how long after feeding the DNA of the prey is detectable in faecal pellets of birds, two carrion crows (Corvus corone corone) were offered larvae of the cockchafer, Melolontha melolontha (Coleoptera: Scarabaeidae). In addition, standard food (approximately 2060% of the diet) was provided that consisted of cattle heart, mealworms (Tenebrio spp.), as well as a mixture of insects, meat and curd. The experiments were carried out between 20th June and 1st July 2004 in an aviary at the Alpenzoo Innsbruck. Each feeding experiment lasted between 4 and 6.5 h and started with a feeding period of 0.52.25 h where each bird was fed with 717 secondinstar cockchafer larvae. The birds droppings were collected continuously throughout the experimental time; all droppings produced from a bird within 1 h were pooled and stored in a 20-mL collection tube. Within two of four feeding experiments, additional faecal pellets were obtained 24 h after the birds had been feeding on the cockchafer larvae: all faecal pellets produced within 30 min by the individual birds were collected. The faecal samples were cooled immediately upon collection and subsequently stored at )28 C. Additionally, samples from the standard food sources and saliva samples from each bird were collected and the DNA extracted. For DNA extraction, three subsamples of approximately 0.1 g were taken from each faecal sample. The subsamples were extracted following a cetyltrimethyl ammonium bromide (CTAB) protocol and puried using a silica-based purication kit as described in Juen & Traugott (2005). Eight faecal samples were extracted a

The inuence of storage conditions on prey DNA detection

Upon collection, 20 faecal samples were stored in ethanol (70%). After 14 months, these droppings were homogenized using a Vortex (Scientic Industries), and aliquots of 900 lL were incubated overnight at )28 C. To test whether the addition of salt improves DNA precipitation, additional aliquots of 900 lL of each sample were incubated overnight after adding 600 lL NH4Ac (5 M). After spinning the samples at 4 C and 21250 g for 30 min, the supernatant was removed and the remaining pellet was subjected to the CTAB protocol to extract the DNA (Juen & Traugott 2005). The success of prey DNA detection was compared between the samples kept in ethanol

2011 Blackwell Publishing Ltd

622 J . O E H M E T A L . and 48 frozen ones; the latter were homogenized and extracted with the CTAB protocol (Juen & Traugott 2005). PCRs were carried out in 10-lL assays, containing 1.5 lL of extracted DNA, 0.2 mM dNTPs (GeneCraft), 1 lM of each Tenebrio primer, 3 mM MgCl2, 1 PCR buffer, 1 lg bovine serum albumin (BSA; Applichem) and 0.75 U Taq DNA polymerase (GeneCraft). PCR cycling conditions consisted of 2 min at 94 C, followed by 40 cycles of 20 s at 94 C, 30 s at 66 C, 45 s at 72 C and 3 min at 72 C. All samples that tested negatively for prey DNA were tested with the universal metazoan primers of Folmer et al. (1994) and the minibarcode-primer (Meusnier et al. 2008) to check for false-negative results. PCR mix and cycling conditions were the same as described for detection of mealworm DNA, except for the 48 C annealing temperature. DNA extracts from faecal samples left on different surfaces where no DNA could be amplied, neither by prey-specic nor by universal primers, were spiked with DNA from Agriotes spp. (Coleoptera: Elateridae) to test for potential inhibition of the PCR. The 10-lL reaction mix contained 1.5 lL of extracted faecal DNA, 1 lL of Agriotes DNA, 0.2 mM dNTPs (GeneCraft), 1 lM of each of the universal primers (Folmer et al. 1994), 3 mM MgCl2, 1 PCR buffer, 1 lg BSA (Applichem) and 0.75 U Taq DNA polymerase (GeneCraft). Cycling conditions were the same as described earlier for PCR with universal primers. Within all PCRs, positive (diluted DNA of the target prey species) and several negative (molecular grade water instead of DNA) controls were included to check for amplication success and DNA contamination, respectively. The PCR products were separated electrophoretically on ethidium bromide-stained agarose gels and visualized using a UV-transilluminator. The specicity of the Melolontha and Tenebrio assays was tested against DNA samples from the birds and from standard food sources.

The inuence of DNA extraction methods on prey DNA detection

DNA from 96 faecal samples was extracted using four different protocols to evaluate extraction-specic differences in prey DNA detection success: a CTAB-based method (Juen & Traugott 2005) and three silica-based kits, UltraClean Fecal DNA Kit (MoBio Laboratories, Carlsbad CA, USA), ExtractMaster Fecal DNA Extraction Kit (Epicentre Biotechnologies, Madison WI, USA) and QIAamp DNA Stool Mini Kit (Qiagen, Hilden, Germany). The mass of each dropping was determined to the nearest 0.01 mg in 48 samples. Thereafter, droppings were dissolved in PBS buffer in a ratio of 1:1.8 (w:v) and homogenized, and 20-lL aliquots were taken. Within each DNA extraction method, 48 aliquots were extracted (small amount of homogenized samples). The remaining 48 faecal samples were divided into four equal parts, and each part was extracted by one of the four DNA extraction protocols (larger pieces of faeces).

Diagnostic PCR
Detection of cockchafer DNA was achieved using the Melolontha-specic primers M-387f and M-387r, targeting the mitochondrial cytochrome subunit one gene and amplifying a fragment of 387 bp (Juen & Traugott 2005). Each PCR was carried out in a 20-lL assay, containing 6 lL of extracted DNA, 0.2 mM dNTPs (GeneCraft), 1 lM of each primer, 3 mM MgCl2, 1 PCR buffer and 0.95 U HotStarTaq DNA Polymerase (Qiagen). PCR cycling conditions consisted of 15 min at 95 C, followed by 40 cycles of 30 s at 94 C, 120 s at 64 C, 100 s at 72 C and nally 10 min at 72 C. To check for false-negative results, all samples that did not produce a prey amplicon were tested in a separate PCR using the general metazoan primers CO1-S38 (5-ttt tca act aac cat aaa gat att gga ac-3) and CO1-A36 (5-taa act tct ggg tga cca aaa aat ca-3). The PCR mix and the cycling conditions were the same as described earlier, except for the annealing temperature that was lowered to 52 C. Detection of mealworm DNA was achieved using Tenebrio-specic primers. Part of the cytochrome subunit one gene was sequenced from three mealworms (GenBank accession numbers HQ891143 HQ891145) using the primers of Folmer et al. (1994). These sequences were used to design the primer pair Ten-mol-S210 (5-tac cgt tat tcg tat gag cag tag-3) and Ten-mol-A212 (5-cgc tgg gtc aaa gaa gga t-3) which amplied a 128-bp product.

Data analysis
We determined the minimum and the maximum time interval between consumption and defecation during which detection of the ingested Melolontha in the droppings was possible. Below, these two time intervals are referred to as the minimum and maximum gut transition time. As the duration of feeding differed within the rst set of experiments, we calculated the time intervals after consumption of the rst and the last Melolontha larva for each pooled faecal sample set. For these calculations, we used the medium time of collection for the pooled droppings. The resulting two values for each pooled faecal sample set were plotted against each other and used to estimate minimum and maximum transition time of detectable Melolontha DNA. Mean air temperature was compared between protected and unprotected sample locations using

2011 Blackwell Publishing Ltd

I M P R O V I N G P R E Y D N A D E T E C T I O N I N A V I A N F A E C E S 623 MannWhitney U test. Differences in DNA detectability were statistically compared by chi-squared tests. Fischers exact test was used if not all conditions were fullled for chi-squared test; in this case, only the P-value is reported. All tests were carried out with SPSS 12.1 (SPSS Inc., Chicago, IL, USA). Within the comparison of the homogenized samples among the four DNA extraction protocols, a distinction was made between samples that did not amplify with the specic insect primers but which tested positive with the general primers (termed other ampliable DNA) and those samples where neither specic nor universal primers gave an amplicon (termed no ampliable DNA).

The effect of sun and rain as well as surface on prey DNA detection success
Within the 5-day-experimental period, relative air humidity ranged between 69.1 and 98.3% (mean 90.5% 7.13 SD). Air temperature ranged between 12.6 and 20.2 C (mean 15.1 C 1.71 SD) and 11.4 and 23.2 C (mean 15.3 C 2.67 SD) for faecal samples protected from and exposed to sunlight and rain, respectively. Although no signicant difference in mean air temperature was found between the two treatments (U = 6244, P = 0.245), midday temperatures were approximately 45 C higher for exposed than for protected samples on the 18th and 19th of August. At these 2 days, the exposed samples also received rain. From the faecal samples placed on leaves, branches, soil and plastic tubes, 56% (14 of 25), 68% (17 of 25), 72% (18 of 25) and 100% (25 of 25), respectively, could be recollected after exposure. Some samples were lost because the droppings were either getting mouldy, dried-in or were washed off the surfaces (Fig. 2). Prey DNA detection success was signicantly lower in exposed droppings (7 of 40, 17.5%) than in protected ones (23 of 34, 68%; v2=19.2; P < 0.001). Among the protected droppings, those from soil showed the lowest prey DNA detection rate (two of six, 33%) compared to 89% (eight of nine) of faecal samples testing positive that were placed on leaves (Fig. 3). Among the exposed droppings, none of the ones left on soil scored positive for prey DNA whereas in 13% (one of eight), 20% (one of ve) and 33%

Results Specicity of PCR assays

Both primer pairs that were designed to amplify DNA from M. melolontha and T. molitor proved to be specic, as only with the target DNA amplication products of the expected length were generated.

Gut transition time of insect prey

Detection of cockchafer DNA in droppings was possible as soon as 0.5 h (minimum gut transition time) and as late as 4 h after feeding (maximum gut transition time). Droppings produced later did not result in ampliable DNA of M. melolontha (Fig. 1).

24 h 330

Time (min) since feeding on last grub

300 270 240 210 180 150 120 90 60 30

Fig. 1 Detection of cockchafer DNA in faecal samples of carrion crows in relation to time since feeding: Small symbols represent a single faecal sample, large circles two samples obtained within the same time frame after feeding. Positive detection of cockchafer DNA is indicated by lled symbols for the single samples and lled half-circles for each of the two samples represented in the larger circles. Vertical and horizontal dotted lines indicate the minimum and maximum gut transition times, respectively.

0
30 60 90

120 0 30 60 90 120 150 180 210 240 270 300 330 360 24 h

Time (min) since feeding on first grub

2011 Blackwell Publishing Ltd

624 J . O E H M E T A L .
(a1) (a2)
Fig. 2 Faeces of carrion crows that were placed on either soil (a), leafs (b), branches (c) or plastic tubes (d) before (1) and after (2) receiving direct sunlight and rain for 5 days.

(b1)

(b2)

(c1)

(c2)

(d1)

(d2)

(5 of 15) of the faeces recovered from branches, leaves and plastic tubes, respectively, an amplication of prey DNA was possible (Fig. 3). All samples scoring negative for prey DNA in the surface experiment were tested using universal metazoan primers to examine whether they contained any ampliable DNA at all. PCR products could be generated only from protected droppings placed on soil (17% of all samples negative with specic primers; Fig. 3). The DNA extracts that scored negative with both the prey-specic and the universal primers were tested for PCR inhibition by spiking them with click beetle DNA. Signicantly more samples obtained from soil (77.8%) showed inhibition compared to samples from leaves (42.9%; v2=4.9; P = 0.027), plastic tubes (40%; v2=6.1; P = 0.014) and branches (29.4%; v2=8.2; P = 0.004). DNA extracts from

exposed droppings showed signicantly higher inhibition rates (65%) than the protected samples (26.5%; v2=10.9; P = 0.001).

The effect of sample storage on prey DNA detection success

In DNA extracts of bird droppings frozen at )28 C, 29 samples (60.4%) were positive for Tenebrio DNA. In extracts of faecal samples stored in ethanol, 15 samples (75%) scored positive for prey DNA. No signicant differences in DNA detection success (P = 1) were found between the two kinds of storage. Similarly, addition of NH4Ac to ethanol-stored faecal samples prior to DNA extraction did not affect prey detection (P = 1): 14 of 20 samples scored positive for Tenebrio DNA.

2011 Blackwell Publishing Ltd

I M P R O V I N G P R E Y D N A D E T E C T I O N I N A V I A N F A E C E S 625
Other amplifiable DNA Prey DNA No amplifiable DNA
Prey DNA No DNA amplifiable
100

Other amplifiable DNA

(a) 100
80

80

% Samples

% Samples

60

60

40

40

c
20

20

0
Soil
Leaves Branches Tube

b
CTAB n = 48 Epicentre n = 47 MoBio n = 47 Qiagen n = 48

Surface
(b) 100
80

DNA extraction method

Fig. 4 Prey DNA detection success in homogenized faeces of carrion crows fed with Tenebrio molitor using four methods for DNA extraction. From each faecal sample, 20-lL aliquots were used per extraction method. In case no DNA amplication with prey-specic primers could be obtained, DNA extracts were tested with universal primers and counted within other ampliable DNA if a PCR product was obtained. Samples that tested negative with both prey-specic and general primers were classied as no DNA ampliable. Different letters indicate signicant differences between extraction methods for the detection of prey DNA only (P < 0.05).

% Samples

60

40

20

0
Soil

Leaves

Branches

Tube

Surface
Fig. 3 Amplication success of Tenebrio molitor DNA in carrion crow faeces lying on different surfaces for 5 days. Faecal samples were either protected from (a) or exposed to (b) direct sunlight and rain. Number of samples analysed (protected exposed): soil 6 12, leaves 9 5, branches 9 8, plastic tubes 10 15.

The effect of DNA extraction on prey DNA detection success

In the Tenebrio-specic PCR, 60% of the samples where small amounts of homogenized droppings were used tested positive for mealworm DNA when extracted by the CTAB protocol. Contrary, prey DNA detection success was signicantly lower when this type of sample was extracted with the kits of Epicentre (6%; v2=31.0, P < 0.001), MoBio (32%; v2=7.8, P = 0.005) and Qiagen (33%; v2=7.1, P = 0.008), respectively (Fig. 4). When larger pieces of faeces were extracted, however, prey DNA detection rates were quite similar between the four methods (CTAB: 83%; Epicentre: 85%; MoBio: 87%; Qiagen: 90%). All samples where small amounts of homogenized droppings were taken and which were negative for meal-

worm DNA were tested with universal metazoan primers to check whether they contained other ampliable DNA. In 47% and 40% of all extracts performed by the MoBio and Qiagen kit, respectively, DNA could be amplied with the universal metazoan primers only, whereas this was true for only 2% and 0% of the samples extracted with the CTAB-based protocol and kit of Epicentre, respectively (Fig. 4). Hence, within the homogenized samples, the largest proportion of samples that contained no prey or other ampliable DNA was found in the kit of Epicentre (94%).

Discussion
Insect prey DNA could be amplied from carrion crow droppings collected 30240 min after prey ingestion only. Comparing these prey DNA transition times with hard part gut transition and retention times shows that these can vary considerably among bird species and food types, for example, yellow-rumped warblers (Dendroica coronata) have mean retention times of 46, 62 and 114 min for fruit-, insect-, and seed-based diets, respectively (Ak & Karasov 1995). The transit times in nectarivorous Palestine sunbirds (Nectarinia osea) range from 10 to 60 min depending on the sugar concentration (Markman et al.

2011 Blackwell Publishing Ltd

626 J . O E H M E T A L . 2006). Barnea et al. (1990) found the rst seeds of Solanum spp. in droppings of bulbus (Pycnonotus xanthopygos) and blackbirds (Turdus merula) 5 min after ingestion. Considerably longer transit times of up to 10.7 and 12.7 h have been reported for the folivore hoatzin (Opisthocomus hoazin) and the frugivorous hornbills (Ceratogymna spp.), respectively (Grajal & Parra 1995; Whitney et al. 1998). In addition, other factors such as age, sex and satiation status of the bird might inuence both the minimum and the maximum transition time (Markman et al. 2006). These ndings suggest that DNA gut transition times are probably to be also affected by diet composition, bird species and the physiological status of the consumer, which needs to be taken into account for eld studies. In general, gut transition times in birds seem to be within the range of a few minutes up to several hours, meaning that prey DNA in fresh faeces will closely be related to the current availability of specic food sources. Recent data presented by Deagle et al. (2010), however, indicate that DNA of sh consumed at least 4 days ago can be amplied form penguin faeces, using highly sensitive pyrosequencing techniques. The present results show that detection of prey DNA is not just possible from fresh faeces, but also from 5-dayold droppings that were protected from direct sunlight and rain. A marked decrease in prey detection success, however, was observed when the droppings were exposed to direct sunlight and rain for 5 days. These results conrm ndings of Deagle et al. (2005) who could not amplify sh DNA from Stellar sea lion scat samples that were exposed to outdoor conditions for 57 days. The DNA in these faeces is probably degraded both by UV radiation and enzymatic processes or consumed by microbes (Lindahl 1993). However, if faecal material is protected from direct sunlight and kept dry, it can effectively archive food DNA as demonstrated by the current study. This can hold true even for extended times postdefecation, as reported in ancient DNA work where coprolites yielded short DNA fragments of the extinct ground sloth Nothroptheriops shastensis along with DNA from plants the sloth had consumed (Poinar et al. 1998; Hofreiter et al. 2000). We found that also the type of surface, on which the bird droppings were kept for 5 days, affected prey DNA detection success that was lowest in those droppings that were left on soil. Extracellular DNA is quickly degraded in soil by entering the microbial food chain or may persist but be unavailable for further analyses by adsorption of DNA to humic acids, soil minerals (e.g. clay) or other soil components (Levy-Booth et al. 2007). We hypothesize that bacteria and fungi (e.g. coprophilous zygomycetes), which are ubiquitous in soil, rapidly utilized the carrion crow droppings as a substrate and thus contributed to an accelerated breakdown of prey DNA. This hypothesis is supported by the observation that soilbased droppings got mouldy within a few days which was not observed for droppings left on the other surfaces (Fig. 2). The soil-resting samples also showed the highest levels of PCR inhibition compared to samples from the other surfaces. We consider humic material extracted with the faeces for having contributed to these inhibitory effects (Tebbe & Vahjen 1993; Wilson 1997; Juen & Traugott 2006). Moreover, it has been hypothesized that within decomposing faeces, substances of inhibitory nature are produced (Deagle et al. 2005). Within the protected droppings, prey DNA detection success was similar for faeces placed on leaves, branches and plastic tubes. Among the exposed droppings, prey DNA was signicantly more often detected in those samples that were stored in plastic tubes. The tubes, although being transparent and uncovered, probably protected the DNA from excessive UV radiation and thus enhanced prey detection rates. Obtaining high-quality DNA is a central step in studies that utilize molecular approaches to unravel trophic interactions (King et al. 2008). These types of studies usually deal with large sample numbers (e.g. Traugott et al. 2008; Deagle et al. 2009); hence, rapid and cost-effective DNA extraction protocols are needed. For small amounts of homogenized samples, the CTAB-based protocol allowed in signicant more samples mealworm DNA to be amplied as the three extraction kits did. However, the kits of MoBio and Qiagen surpassed the CTAB-based protocol when the DNA extracts were tested for other ampliable DNA (using universal metazoan primers in samples where no prey DNA could be detected). We ascribe the high recovery rate of prey DNA in the CTABbased protocol to the fact that it had been optimized to the detection of prey DNA which usually is present in tiny amounts of short-sized fragments. The commercial kits, however, are probably quite effective in extracting long DNA fragments from the faecal material, putatively DNA from the crows themselves. In vertebrate faeces, a considerable fraction of DNA can originate from cells of the intestinal mucosa of the defecating animal and not from prey (Albaugh et al. 1992; Deagle et al. 2006; but see Roeder et al. 2004). When pieces of faecal pellets were directly subjected to DNA extraction, prey DNA detection success ranged between 80% and 90% in the four methods compared here. We ascribe this to the fact that more faecal material was used here for DNA extraction compared to the much smaller amounts used in homogenized samples. It is essential to rapidly stop DNA degradation in the sample when collecting faecal material in the eld. We found that detection success of prey DNA was similar in frozen and ethanol-stored faeces, conrming ndings by Sutherland (2000). Wasser et al. (1997), comparing

2011 Blackwell Publishing Ltd

I M P R O V I N G P R E Y D N A D E T E C T I O N I N A V I A N F A E C E S 627 freezing and storage of mammal faeces in 70% ethanol, found amplication of DNA from the defecating animal in frozen samples to be better. Frantzen et al. (1998), however, could not discover any differences between these two storage methods. The latter authors assumed that the type of preservation of mammal faeces has little effect on DNA extraction and amplication success when DNA fragments smaller than 200 bp are targeted. These ndings suggest storing the droppings of birds in ethanol as a cheap and feasible method to preserve them for further molecular analysis. Within the current experiments, we found that 5-dayold droppings, protected from direct sunlight and rain, are probably to contain usable molecular dietary information. This suggests that faeces collected from sites where birds rest or gather can be used for the detection of prey DNA. Moreover, it circumvents the need to catch birds and to get them to defecate or to collect birds droppings immediately upon defecation, both being tedious and time-consuming and considerably hampering largescale eld studies. The type of surface, where the droppings are placed at, plays a crucial role as well: droppings deposited on soil are less suited for molecular analysis of prey remains, whereas smooth, clean and non-absorbing surfaces such as leaves preserve the faecal DNA better. Similar to our feeding experiments, plastic foil might be laid out at rookeries where birds gather to sleep, breed or rest, to collect faeces. From this surface, the faeces can be easily transferred to reaction tubes lled with high-concentration ethanol. The foil can be exchanged after each sampling bout to reduce the risk of sample contamination and degradation. However, further work is needed to determine how fast prey DNA breaks down in avian faeces depending on the type of surface and exposition. Commercially available DNA kits can then be used for extracting the prey DNA from individual droppings. In case very small and or degraded droppings have to be processed, a CTAB-based extraction protocol is recommended to maximize the chances of prey DNA detection.
Albaugh GP, Iyengar V, Lohani A, Malayeri M, Bala S, Nair PP (1992) Isolation of exfoliated colonic epithelial-cells, a novel, noninvasive approach to the study of cellular markers. International Journal of Cancer, 52, 347350. Barnea A, Yomtov Y, Friedman J (1990) Differential germination of 2 closely related species of Solanum in response to bird ingestion. Oikos, 57, 222228. Deagle BE, Tollit DJ, Jarman SN, Hindell MA, Trites AW, Gales NJ (2005) Molecular scatology as a tool to study diet: analysis of prey DNA in scats from captive Steller sea lions. Molecular Ecology, 14, 18311842. Deagle BE, Eveson JP, Jarman SN (2006) Quantication of damage in DNA recovered from highly degraded samples: a case study on DNA in faeces. Frontiers in Zoology, 3, 11. Deagle BE, Gales NJ, Jarman SN, Robinson S, Trebilco R, Hindell MA (2007) Studying Seabird diet through genetic analysis of faeces: a case study on macaroni penguins (Eudyptes chrysolophus). PLoSOne, 2, e831. Deagle BE, Kirkwood R, Jarman SN (2009) Analysis of Australian fur seal diet by pyrosequencing prey DNA in faeces. Molecular Ecology, 18, 20222038. Deagle BE, Chiaradia A, McInnes J, Jarman SN (2010) Pyrosequencing faecal DNA to determine diet of little penguins: is what goes in what comes out? Conservation Genetics, 11, 20392048. Exnerova A, Stys P, Kristin A, Volf O, Pudil M (2003) Birds as predators of true bugs (Heteroptera) in different habitats. Biologia, 58, 253264. Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplication of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3, 294299. Frantzen MAJ, Silk JB, Ferguson JWH, Wayne RK, Kohn MH (1998) Empirical evaluation of preservation methods for faecal DNA. Molecular Ecology, 7, 14231428. Gamez-Virues S, Bonifacio RS, Gurr GM, Kinross C, Raman A, Nicol HI (2007) Arthropod prey of shelterbelt-associated birds: linking faecal samples with biological control of agricultural pests. Australian Journal of Entomology, 46, 325331. Grajal A, Parra O (1995) Passage rates of digesta markers in the gut of the hoatzin, a folivorous bird with foregut fermentation. Condor, 97, 675 683. Hofreiter M, Poinar HN, Spaulding WG et al. (2000) A molecular analysis of ground sloth diet through the last glaciations. Molecular Ecology, 9, 19751984. Hull CL (1999) Comparison of the diets of breeding royal (Eudyptes schlegeli) and rockhopper (Eudyptes chrysocome) penguins on Macquarie Island over 3 years. Journal of Zoology, 247, 507529. Jarman SN, Gales NJ, Tierney M, Gill PC, Elliott NG (2002) A DNA-based method for identication of krill species and its application to analysing the diet of marine vertebrate predators. Molecular Ecology, 11, 26792690. Juen A, Traugott M (2005) Detecting predation and scavenging by DNA gut-content analysis: a case study using a soil insect predator-prey system. Oecologia, 142, 344352. Juen A, Traugott M (2006) Amplication facilitators and multiplex PCR: Tools to overcome PCR-inhibition in DNA-gut-content analysis of soilliving invertebrates. Soil Biology & Biochemistry, 38, 18721879. King RA, Read DS, Traugott M, Symondson WOC (2008) Molecular analysis of predation: a review of best practice for DNA-based approaches. Molecular Ecology, 17, 947963. Levy-Booth DJ, Campbell RG, Gulden RH et al. (2007) Cycling of extracellular DNA in the soil environment. Soil Biology & Biochemistry, 39, 29772991. Lindahl T (1993) Instability and decay of the primary structure of DNA. Nature, 362, 709715. Markman S, Tadmor-Melamed H, Arieli A, Izhaki I (2006) Sex differences in food intake and digestive constraints in a nectarivorous bird. Journal of Experimental Biology, 209, 10581063. Meusnier I, Singer GAC, Landry JF, Hickey DA, Hebert PDN, Hajibabaei M (2008) A universal DNA mini-barcode for biodiversity analysis. Bmc Genomics, 9, 214.

Acknowledgements
We are grateful to the Alpenzoo Innsbruck for providing carrion crows and aviaries to conduct the feeding experiments. The University of Innsbruck nancially supported this project by a Hypo Tirol Bank grant and a PhD studentship for Johannes Oehm. We also thank three anonymous referees for their comments that improved the manuscript.

References
Ak D, Karasov WH (1995) The trade-offs between digestion rate and efciency in warblers and their ecological implications. Ecology, 76, 2247 2257.

2011 Blackwell Publishing Ltd

628 J . O E H M E T A L .
Miller CK, McEwen LC (1995) Diet of nesting Savanna sparrows in interior Alaska. Journal of Field Ornithology, 66, 152158. Moorman CE, Bowen LT, Kilgo JC et al. (2007) Seasonal diets of insectivorous birds using canopy gaps in a bottomland forest. Journal of Field Ornithology, 78, 1120. Poinar HN, Hofreiter M, Spaulding WG et al. (1998) Molecular coproscopy: dung and diet of the extinct ground sloth Nothrotheriops shastensis. Science, 281, 402406. Roeder AD, Archer FI, Poinar HN, Morin PA (2004) A novel method for collection and preservation of faeces for genetic studies. Molecular Ecology Notes, 4, 761764. Scribner KT, Bowman TD (1998) Microsatellites identify depredated waterfowl remains from glaucous gull stomachs. Molecular Ecology, 7, 14011405. Sutherland RM (2000) Molecular Analysis of Avian Diets. PhD Thesis, University of Oxford, UK. Tebbe CC, Vahjen W (1993) Interference of humic acids and DNA extracted directly from soil in detection and transformation of recombinant-DNA from bacteria and a yeast. Applied and Environmental Microbiology, 59, 26572665. Traugott M, Symondson WOC (2008) Molecular analysis of predation on parasitized hosts. Bulletin of Entomological Research, 98, 223231. Traugott M, Bell JR, Broad GR, Powell W, van Veen FJF, Vollhart IMG, Symondson WOC (2008) Endoparasitism in cereal aphids: molecular analysis of a whole parasitoid community. Molecular Ecology, 17, 3928 3938. Valera F, Gutierrez JE, Barrios R (1997) Effectiveness, biases and mortality in the use of apomorphine for determining the diet of granivorous passerines. Condor, 99, 765772. Wasser SK, Houston CS, Koehler GM, Cadd GG, Fain SR (1997) Techniques for application of faecal DNA methods to eld studies of Ursids. Molecular Ecology, 6, 10911097. Whitney KD, Fogiel MK, Lamperti AM et al. (1998) Seed dispersal by Ceratogymna hornbills in the Dja Reserve, Cameroon. Journal of Tropical Ecology, 14, 351371. Wilson IG (1997) Inhibition and facilitation of nucleic acid amplication. Applied and Environmental Microbiology, 63, 37413751. Xavier JC, Croxall JP, Cresswell KA (2005) Boluses: An effective method for assessing the proportions of cephalopods in the diet of albatrosses. Auk, 122, 11821190.

2011 Blackwell Publishing Ltd