Introduction

Recent developments in our knowledge about the natural distribution of stable isotopes in marine food webs have provided us with a useful tool to study movements and the trophic level at which seabirds feed, which can be applied when more direct studies are impossible (e.g., Cherel et al. 2005a, 2006; Quillfeldt et al. 2005; Bearhop et al. 2006; Gladbach et al. 2007). In particular, carbon stable isotope ratios can reflect foraging locations (reviewed in Rubenstein and Hobson 2004), while nitrogen isotope ratios become enriched in 15N with trophic level by approximately 3.0–5.0‰ (Minagawa and Wada 1984, Owens 1987) and δ15N can thus indicate trophic position and be used to infer dietary composition (e.g., Hobson and Welch 1992; Hobson et al. 2000; Forero et al. 2005). Depending on the tissue chosen, dietary information spanning different temporal scales can also be obtained (Hobson and Clark 1992a).

The stable isotope approach is especially useful for studies of diet and foraging areas outside the breeding season, when pelagic birds are not usually accessible for sampling using traditional methods (Barrett et al. 2007). A number of studies have made use of the fact that the molting period is in this time, and thus, feathers grown by adult seabirds reflect the diet in the interbreeding period (e.g., Nisbet et al. 2002; Quillfeldt et al. 2005; Cherel et al. 2006; Gladbach et al. 2007). To assess diet and foraging areas during the breeding season, on the other hand, several tissues can be sampled non-destructively, including blood and feather samples of chicks and adults. However, due to the specific metabolic processes involved in tissue synthesis, different tissues may differ in their isotopic enrichment factor relative to the diet (e.g., Hobson and Clark 1992a).

As noted by several authors (Bearhop et al. 2000a; Cherel et al. 2005a, b), differences in isotopic discrimination among tissues are a potential problem when comparing different tissue types. Temporal studies comparing the blood (representing the breeding season) and feathers (representing the interbreeding season) of adult seabirds are dependent on the assumption that isotope analyses of the two tissues provide directly comparable dietary proxies. However, most researchers working in the field of isotope ecology are aware that the enrichment factors of blood and feathers may differ when endogenous reserves are used as precursors during feather synthesis (Cherel et al. 2005a) or when blood contains 15N-depleted uric acid (Bearhop et al. 2000b). Thus, most studies have limited their analysis to either blood or feathers, but have assumed both tissues to represent a difference of one trophic level to diet. A few studies have directly compared feather and blood isotope values (e.g., Podlesak et al. 2005; Bearhop et al. 2006), mainly on the assumption that the influence of tissue specific fractionation is small compared to the effect they aimed to measure.

The aim of this study is to compare the carbon and nitrogen stable isotope values of blood and simultaneously growing feathers of five species of Procellariiformes in order to test for differences in isotopic tissue discrimination between tissues. We also hope that by testing the basis of assumptions that are often made in dietary studies regarding the equivalence of different tissues, we can help clarify if blood or feather is the optimal tissue to sample. We also consider the limitations of extrapolating different-tissue stable isotope comparisons within temporal studies or between studies using feather and studies using blood.

Materials and methods

Study site, species and sampling

Species, abbreviations used in Fig. 1 and sampling locations are listed in Table 1, for sample sizes see Table 2.

Fig. 1
figure 1

Carbon and nitrogen stable isotopes (means ± SE) of seabird blood and growing feathers, representing similar time periods. For sample sizes see Table 2

Full size image
Table 1 Species, abbreviations used in Fig. 1 and sampling locations of the Procellariiformes included in the present study
Full size table
Table 2 Carbon stable isotopes (means ± SE) of seabird blood and growing feathers, representing similar time periods, and individual differences between the simultaneously sampled tissues (means ± SE)
Full size table

Great shearwaters Puffinus gravis, Atlantic Yellow-nosed albatrosses Thalassarche chlororhynchos and Spectacled petrels Procellaria conspicillata were sampled in the Southwestern Atlantic Ocean off the Brazilian coast, onboard fishing vessels from February to June 2006. Birds scavenging discards were attracted to the vessel and trapped with a cast net as described by Bugoni et al. (2008b). Blood was collected from the tarsal vein using a 2-ml disposable syringe and needle. Growing body feathers, i.e., still with blood in the calamus, from the breast and dorsum were collected and stored dry in sealed plastic bags. This sample includes immature as well as breeding and non-breeding adults.

Trindade petrels Pterodroma arminjoniana were sampled in breeding colonies of Trindade Island (c. 20°30′S–29°19′W), where there is a population estimated at 1,130 breeding pairs (Luigi et al. 2008). Blood and growing feathers were collected from December 2006 to April 2007 and preserved in the same way described above. Four adults were sampled during chick feeding stage, whereas 14 were prospecting birds, but both groups were sampled over the same time period despite differences in breeding stage as the species nests throughout the year (Luigi et al. 2008). During the prospecting period (i.e., before the pre-laying exodus, which lasts usually 2 months), adults are frequently in final moult stage, with some growing body feathers still present (Luigi et al. 2008). Chicks were sampled following the same procedure, and full grown or growing body feathers of the definitive plumage were collected. One chick was 40-days-old and others were 70–100 days old, so close to fledging, which occurs about the 95th to 100th day (Luigi et al. 2008). The half-life of isotope turnover of avian whole blood has been determined to be 11.4 days for quail (Hobson and Clark 1992a), and so represents the same time window as growth of new body feathers, as these feathers are 20–30 mm long and growth rate of feathers in birds varies from 2.5 to 10 mm per day (Langston and Rohwer 1996).

Feather and blood samples from chicks of Thin-billed prions Pachyptila belcheri were collected as part of ongoing studies of their breeding biology at New Island Nature Reserve, Falkland Islands (e.g., Quillfeldt et al. 2003, 2006, 2007a, b, c). The present study includes samples collected from 40-day-old chicks during the breeding seasons 2004–2005 (N = 7) and 2005–2006 (N = 10). Birds were captured by hand and blood samples (0.2–0.4 ml) were collected from the brachial vein using heparinized capillaries within 2 min from burrow opening to the end of blood sampling. Blood samples were immediately transferred to 0.5-ml tubes and kept on ice until centrifugation. Plasma (used for hormone analyses, e.g., Quillfeldt et al. 2006, 2007c) and blood cells (used in the present analysis) were stored frozen in separate 0.5-ml tubes at −20°C. Undertail covert feathers of chicks were collected shortly before fledging by gentle pulling and placed in individual sealed plastic bags. The first tail feathers of chicks including the tail coverts start to emerge after 3 weeks of age (e.g., Strange 1980; “at 22 days the sheaths of the rectrices can be felt protruding”), therefore the distal parts of these feathers represent the middle of the nestling growth period of 50–56 days. The half-life of isotope turnover of avian red blood cells was 29.8 days in crows Corvus brachyrhynchos (Hobson and Clark 1993) and 10.9 days for Yellow-rumped warblers Dendroica coronata (Podlesak et al. 2005). The red blood cell samples collected from Thin-billed prion chicks therefore represented the diet ingested ca. 2–4 weeks before the sampling at 20 days, i.e., at the same time as the analyzed feather parts.

Sample preparation and stable isotope analysis

Feathers were cut into small fragments using stainless steel scissors, and whole blood and red blood cells (RBC) were freeze-dried and ground. Carbon and nitrogen isotope analyses were carried out on 0.65–0.7 mg aliquots of homogenized powder, weighed into tin cups.

Carbon and nitrogen isotope ratios were measured simultaneously by continuous-flow isotope ratio mass spectrometry (CF-IRMS) using a Costech Elemental Analyser (EA) linked to a Thermo Finnigan Delta Plus XP Mass Spectrometer. Two laboratory standards were analyzed for every ten unknown samples, allowing any instrument drift over a typical 14-h run to be corrected. Stable isotope ratios were expressed in δ notation as parts per thousand (‰) deviation from the international standards V-Pee Dee belemnite (carbon) and AIR (nitrogen), according to the following equation δX = [(Rsample/Rstandard) − 1] × 1000 where X is 15N or 13C and R is the corresponding ratio 15N/14N or 13C/12C. Based on internal standards (tryptophan), the analytical precision (± 1 SD) was estimated as ±0.18‰ and ±0.17‰ for δ15N and δ13C, respectively.

Data analysis

Statistical tests were performed in Sigma Stat 2.03. All tests used pair-wise data for each individual, carried out separately for each species and age group. Sample sizes are given in Table 2.

Results

In all species and age groups, the feathers were enriched in 15N and 13C compared with blood (Table 2; Fig. 1). The differences ranged between 0.4 and 1.5‰ for δ15N and between 0.9 and 1.6‰ for δ13C (Table 2; Fig. 1). In all species except Trindade petrels, the δ15N and δ13C values of feathers and blood samples were highly correlated in individuals (Table 2, Table 3).

Table 3 Nitrogen stable isotopes (means ± SE) of seabird blood and growing feathers, representing similar time periods, and individual differences between the simultaneously sampled tissues (means ± SE)
Full size table

Discussion

Our study shows differences between the carbon and nitrogen stable isotopes of blood and simultaneously growing feathers of five species of Procellariiformes, representing five genera. We found that feathers were consistently enriched in 15N and 13C compared with blood in all species.

This is in line with previous studies that reported differences in isotope values between blood and feathers (e.g., Hobson and Clark 1992b; Bearhop et al. 2000a; Cherel et al. 2005a, b). However, the present approach is novel in that it used simultaneously collected samples from wild birds in their natural habitat, and exhibiting their natural foraging behaviors. In contrast, previous studies did not sample the same individuals simultaneously (e.g., Cherel et al. 2005a; Bearhop et al. 2000a, for simultaneously sampled hydrogen see Langin et al. 2007). Further, studies with simultaneous sampling, to our knowledge, have been carried out only in captive birds thus far (Hobson and Clark 1992b; Cherel et al. 2005b, but see Bugoni et al. 2008a).

The most likely explanation for increased 15N and 13C isotope enrichment in feathers compared to blood are metabolic differences. These may be found in the different protein sources (several proteins in blood, keratin in feathers) and differences would also appear if endogenous reserves are used as precursors during feather synthesis (e.g., Cherel et al. 2005a). For example, a large difference between 15N enrichment factors of whole blood and feathers may indicate that most amino acids available for keratin synthesis are provided by endogenous reserves. Conversely, almost identical and low 15N enrichment factors in whole blood and feathers may suggest a dietary origin of feather amino acids (Cherel et al. 2005a).

The presence of uric acid, which is 15N depleted (Petersen and Fry 1987), has also been suggested as a cause of isotopic differences between blood and feathers. Bearhop et al. (2000b) found that uric acid removal in Great skua Stercorarius skua blood increased the measured δ15N by 1.2‰, while δ13C was not affected. As Cherel et al. (2005a) pointed out, the effect of uric acid is likely to be minimal because of the small amount of uric acid in blood, compared to protein (Garcia-Rodriguez et al. 1987; Boismenu et al. 1992). Higher values might be found, however, in times of high protein turnover such as during growth in chicks (Bearhop et al. 2000b and references therein). In the present study, both the nitrogen and carbon isotope ratios were lower in blood than in feathers, indicating that uric acid is not the only source for the observed difference between tissues. In Thin-billed prions, blood cells rather that whole blood were analyzed, and as uric acid is present in the plasma, it should not interfere with the measurements in this case. In fact, we found that the difference between δ15N of blood and feathers was smallest in Thin-billed prions (0.4‰), and the difference was statistically significant (P = 0.021) rather that highly significant (P < 0.001) as in all other species. Thus, when the separation of plasma is possible, this may be a good way to exclude uric acid as an error source and allow better comparison between blood and feathers from the same species.

In four of five species in this study, δ15N and δ13C values of feathers and blood samples from individual birds were highly correlated. There was one exception of this pattern. In Trindade petrels, we observed a correlation only for adults and only for nitrogen. It is tempting to suggest that this may have resulted from a large variation in foraging areas over a short time, as the Trindade petrel is the only tropical species in our dataset. In tropical species, foraging occurs over vast oceanic areas, and thus there may be more short-term variation in diet, as well as individual metabolic differences (e.g., in the use of internal stores during feather growth). Foraging trips by Trindade petrel parents vary from hours to weeks (Luigi et al. 2008), which could suggest that chicks use both diet and endogenous stores for tissue formation, depending on feeding frequencies. Thus the lack of intertissue correlations of carbon and nitrogen isotope values could reveal individual differences according to nutritional status.

In the present study, the difference between blood and simultaneously grown feathers was remarkably similar across five species of Procellariiformes representing five genera (Fig. 1), despite differences in ecology and geographical regions. It was also in the range of the difference found for captive King Aptenodytes patagonicus and Rockhopper Eudyptes chrysocome penguins, where Cherel et al. (2005b) found difference of 1.5 and 1.7‰ for δ15N, and between 0.9 and 1.1 ‰ for δ13C.

In deciding which tissue type is more reliable, both tissue types have error sources. In blood, variable content of uric acid and lipids is of concern, but this may be avoided if red blood cells can be used rather than whole blood, which may also be beneficial in many studies because red blood cells sample a much longer time window than plasma (e.g., Podlesak et al. 2005). In feathers, the variable use of endogenous reserves during feather growth has to be taken into account, especially of certain amino acids such as sulphur-containing amino acids (cysteine and methionine) that are abundant in keratin but may be scarce in recently ingested food, such that feather stable isotopes may also be biased. For chicks the role of stored reserves is not usually a major concern, although this needs to be confirmed in species undergoing mass recession during the growth of the feathers, as is the case in Procellariiformes. The best solution may be the analysis of both blood and growing feather so that the two together give a better indication than either tissue alone.

We suggest that future studies consider the following: (i) Sample both tissues if possible to overcome limitations of each. (ii) Use similar tissues to compare temporal patterns when possible, e.g., compare moult feathers with regrowths in the breeding season. (iii) RBC may be preferable to whole blood or plasma, by avoiding uric acid effects on nitrogen isotope ratios, (iv) Correlation between isotopic values of different tissues of similar age should be addressed, which could provide arithmetic corrections and make comparisons across tissues viable. (v) More research is needed on metabolic differences between chick and adults, as well as studies addressing differences in chick and adult diets in the nesting period. (vi) Avoid preserving tissues with preservatives containing the stable isotopes to be analyzed, such as ethanol or heparin for determination of SI of carbon in blood.