Oecologia

, Volume 144, Issue 4, pp 647–658

Resolving temporal variation in vertebrate diets using naturally occurring stable isotopes

Authors

    • Department of ZoologyStockholm University
  • A. Angerbjörn
    • Department of ZoologyStockholm University
Stable Isotopes Issue

DOI: 10.1007/s00442-005-0118-0

Cite this article as:
Dalerum, F. & Angerbjörn, A. Oecologia (2005) 144: 647. doi:10.1007/s00442-005-0118-0

Abstract

Assessments of temporal variation in diets are important for our understanding of the ecology of many vertebrates. Ratios of naturally occurring stable isotopes in animal tissues are a combination of the source elements and tissue specific fractionation processes, and can thus reveal dietary information. We review three different approaches that have been used to resolve temporal diet variation through analysis of stable isotopes. The most straightforward approach is to compare samples from the same type of tissue that has been sampled over time. This approach is suited to address either long or short-term dietary variation, depending on sample regime and which tissue that is sampled. Second, one can compare tissues with different metabolic rates. Since the elements in a given tissue have been assimilating during time spans specific to its metabolic rate, tissues with different metabolic rates will reflect dietary records over different periods. Third, comparisons of sections from tissues with progressive growth, such as hair, feathers, claws and teeth, will reveal temporal variation since these tissues will retain isotopic values in a chronological order. These latter two approaches are mainly suited to address questions regarding intermediate and short-term dietary variation. Knowledge of tissue specific metabolic rates, which determine the molecular turnover for a specific tissue, is of central importance for all these comparisons. Estimates of isotopic fractionation between source and measured target are important if specific hypotheses regarding the source elements are addressed. Estimates of isotopic fractionation, or at least of differences in fractionation between tissues, are necessary if different tissues are compared. We urge for more laboratory experiments aimed at improving our understanding of differential assimilation of dietary components, isotopic fractionation and metabolic routing. We further encourage more studies on reptiles and amphibians, and generally more studies utilizing multiple tissues with different turnover rates.

Keywords

δ13Cδ15NEcophysiologySeasonalityFeeding ecology

1 Introduction

Many vertebrate species show a strong temporal variation in their diet, either as a seasonal variation or as a long-term effect over many years. Seasonal variation in food sources can be crucial for understanding population dynamics (e.g., Reid et al. 1997) and long-term trends in diet could be important to assess ecological effects of environmental change such as global warming (Brown et al. 2001). From a conservation standpoint, management of endangered populations relies on sound information of ecological conditions for the population under concern. Knowledge of the temporal and spatial utilization of food resources is in this context often a cornerstone for successful management (Fuller and Sievert 2001).

Traditional ways of analyzing diets, including identification of content in collected faeces, identification of content in stomachs from diseased animals, or direct observations of feeding habits, often fail to adequately resolve temporal patterns in diet use since these techniques reflect small and nonrandom samples where pseudoreplication is a reoccurring problem that is difficult to account for (Reynolds and Aebisher 1991; Deb 1997; Darimont and Rimchen 2002). Measurements of stable isotopes in animal tissues can, however, be a powerful alternative since isotopes reflect average dietary records and thus eliminate some of the shortcomings of traditional dietary studies (DeNiro and Epstein 1978, 1981). Traditionally, carbon and nitrogen isotopes have been the main elements used in dietary analyses (see reviews in Peterson and Fry 1987; Gannes et al. 1998; Hobson 1999; Kelly 2000), while oxygen, hydrogen and sulphur has been extensively used to track animal movements and to study climatic changes (Ayliffe et al. 1992; Genoni et al. 1998; Hobson 1999; Rubenstein and Hobson 2004).

In this paper, we review three different approaches to use stable isotopes to resolve temporal diet variation. First, the most straightforward approach is to simply compare samples from the same type of tissue that has been sampled over time. Second, one can compare tissues with different rates of molecular turnover, which means that they will reflect dietary records over different time periods (Hobson 1993). Third, it is possible to compare different sections from a tissue that has progressive growth, such as hair, feathers, claws and teeth, since these tissues will retain isotopic values in a chronological order (Bearhop et al. 2003; Schwertl et al. 2003). From a temporal perspective, the first approach is suitable for addressing questions regarding both long and short-term variation while the latter two can be used to resolve short-term dietary patterns.

One basic principle behind all three approaches is tissue specific isotopic turnover. Most animal tissues show a turnover time, although some are metabolically inert, e.g., hair and feathers. Further, some organisms such as most reptiles and fish have continuous growth, where incorporation of new protein into tissues is a combination of turnover and growth. Analyses of stable isotopes for ecological applications in animals have used several different tissues of both organic and inorganic origin such as different proteins (Tieszen et al. 1983), lipids (Howland et al. 2003), and hydroxyapatite in bones or teeth (Ambrose and Norr 1993). However, most studies have focused on atoms bound to protein molecules (Tieszen and Boutton 1989). Since elemental turnover seem to be related to protein turnover, isotope values measured in animal tissues will reflect the dietary record for the time the protein has been assimilating (Kurle and Worthy 2002). Protein turnover is related to the specific metabolic rate of each tissue, so that tissues with fast metabolic rates have fast protein turnover (Welle 1999). This means that tissues with different turnover rates reflect average dietary records over different time windows (Tieszen et al. 1983; Hobson and Clark 1992a).

In bioaccumulation of lighter elements into biological tissues, the isotope ratio of the accumulated element is often somewhat different compared to the source element, a process called isotopic fractionation. This is a combination of kinetic and chemical processes that generate a differential preference for light versus heavy isotopes in the metabolic steps that incorporate elements into biological tissue (Peterson and Fry 1987). Each tissue (or actually each metabolic process) has its own fractionation value for each element. Since different proteins have separate metabolic pathways, most proteins probably have different fractionation values. In tissues with several types of protein molecules, the overall fractionation value is thus a mean value for all separate fractionation coefficients combined. This also expands to other tissue components, such as lipids, and we emphasize that measurements, therefore, should be made on as pure protein as possible, with lipids and other nonprotein tissue components extracted. Thus, if values from different tissues are compared, they need to be normalized to account for differences caused by fractionation alone. There is, however, an ongoing discussion whether amino acids directly are routed to tissue protein or if they are broken down and resynthesized in different tissues (Ambrose and Norr 1993). For example, amino acids can be built from glycerol, which in turn has been built from dietary sources (Keeling and Nelson 2001). Thus, it is important to emphasize that the fractionation observed between a source element and target tissue is the combined fractionation for all metabolic steps that has preceded assimilation of the measured target tissue.

We have confined the discussion to wild vertebrates and stable isotopes of C and N in proteins. However, many of the general principles could easily be expanded to other taxonomic groups and other light elements as well. The studies were identified from a systematic search on Web of Science, including a combination of the search term “isotope” and any of the terms ”temporal”, “diet” or “seasonal”. This database contains the majority of journals within social and biological sciences since 1985. We also incorporated other studies of which we were aware of although not identified by the literature search. Although a large number of archaeological and anthropological studies have used stable isotope to trace the diets of historic and prehistoric humans (e.g., Chisholm et al. 1982; Schoeninger et al. 1983; Ambrose and Krigbaum 2003), the focus on temporal diet patterns is mostly found among studies on wild vertebrates (but see White and Schwarcz 1994). Therefore, we have excluded archaeological and anthropological studies from our review.

2 Resolving temporal patterns in diet

In total, we identified 64 studies that had used stable carbon and/or nitrogen isotopes to investigate seasonal diet patterns in wild vertebrates. Although the database used to identify published work only goes back to 1985, most application of stable isotope techniques has occurred during the past 10 years. We consequently feel that although this is not the total number of published studies where stable isotopes has been used, these studies will be representative for how the technique has been used across taxa and different tissues.

Overall, we found that most studies focused on birds (n=20) and mammals (n=27), and that an intermediate number focused on fish (n=16; Fig. 1). However, we found only one study that used stable isotopes to investigate temporal diet patterns in reptiles, and no study at all that used the technique on amphibians. Most studies compared samples from the same tissue collected over time and the most commonly measured tissues were blood, muscle and feathers (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00442-005-0118-0/MediaObjects/442_2005_118_Fig1_HTML.gif
Fig. 1

Compilation of 64 studies that were found to have used stable carbon and nitrogen isotopes to resolve temporal variation in diets of wild vertebrates using three different approaches: (1) comparisons of samples from the same tissue sampled over time, (2) comparisons of samples from tissues with different metabolic rate, (3) comparisons between segments of tissues with progressive growth. No studies were found to have used stable isotopes to investigate temporal patterns in feeding ecology of amphibians. References for studies underlying the figure, including tissues measured, are given in Appendix 1

https://static-content.springer.com/image/art%3A10.1007%2Fs00442-005-0118-0/MediaObjects/442_2005_118_Fig2_HTML.gif
Fig. 2

Tissues used for resolving temporal variation in diet of wild vertebrates using either: (1) comparisons of samples from the same tissue sampled over time, (2) comparisons of samples from tissues with different metabolic rate, (3) comparisons between segments of tissues with progressive growth. References for studies underlying the figure, including tissues measured, are given in Appendix 1

2.1 Comparisons of the same tissue sampled over time

The straightforward approach to use stable isotopes to assess temporal variation in diet is to repeatedly collect samples of the same tissue and compare these across suitable time intervals. This could reveal both an annual or seasonal variation, depending on the time interval between the samples and the turnover rate of the measured tissue. This is similar to traditional ways of analyzing diets, when records of other types of dietary data are compared over time (e.g., regurgitated droppings—Reid et al. 1997; faecal droppings—Elmhagen et al. 2001; stomach content—Helldin 2000). Especially for birds and mammals, this was the most common approach to address temporal variation in diet in the reviewed studies (Fig. 1).

The most common tissues for such comparisons were bone, muscle, blood and feathers (Fig. 2). Bone collagen has a substantial elemental turnover rate. It is, therefore, mainly suitable for long-term trends in dietary patterns since short-term temporal variation will be averaged out. Most other measured tissues, on the other hand, have a relatively shorter turnover rate and can thus be used to reveal seasonal patterns (Table 1). Some tissues, for instance blood and most inert tissues such as claws, hair and feathers, have the advantage that they can be sampled repeatedly on the same individual, thus allowing for a temporal analysis on an individual basis (e.g., Hildebrand et al. 1996; Ben-David et al. 1997; Ainley et al. 2003). Hair and feathers can also be sampled noninvasively, i.e., without any actual handling of the studied animals. Feathers were indeed the most utilized tissue in birds. However, there was not an equal use of mammalian hair (Fig. 2), possibly due to the fact that feathers are easier to collect. Feathers also have the advantage of a relatively distinct growing phase, which for instance has been used to compare diet before and after moult (Cherel et al. 2000). For mammalian hair, we lack precise knowledge about their growth rate, although several recent studies have explored both growth rates of hair and potential element pools from which hair has been built up (Schwertl et al. 2003; Sponheimer et al. 2003; Ayliffe et al. 2004; Greaves et al. 2004).
Table 1

Element specific isotopic turnover of C and N for various tissues in mammals and birds

Tissue

Species

Half life (Days)

Source

13C

15N

Mammals

Muscle

Gerbil

27.6

 

Tieszen et al. (1983)

Liver

Gerbil

6.4

 

Tieszen et al. (1983)

Whole blood

Lesser long-nosed bat

120

 

Voight et al. (2003)

Palla’s long-tongued bat

113–126

 

Voight et al. (2003)

Blood cells

American black bear

40

40

Hildebrand et al. (1996)

Plasma/Serum

American black bear

>4

>4

Hildebrand et al. (1996)

Harbor seal

0.7

 

Lesange et al. (2002)

Gray seal

0.7

 

Lesange et al. (2002)

Harp seal

1.0

 

Lesange et al. (2002)

Hair

Gerbil

47.5

 

Tieszen et al. (1983)

Lesser long-nosed bat

537

 

Voight et al. (2003)

Domestic cattle

 

19

Schwertl et al. (2003)

Domestic horse

0.5, 4.3, 136

 

Ayliffe et al. (2004)

Brain

Gerbil

28.2

 

Tieszen et al. (1983)

Wing membrane

Lesser long-nosed bat

134

 

Voight et al. (2003)

Palla’s long-tongued bat

102

 

Voight et al. (2003)

Fat

Gerbil

15.6

 

Tieszen et al. (1983)

Birds

Collagen

Quail

173.3

 

Hobson and Clark (1992a)

Muscle

Quail

12.4

 

Hobson and Clark (1992a)

Liver

Quail

2.6

 

Hobson and Clark (1992a)

Whole blood

Quail

11.4

 

Hobson and Clark (1992a)

Great skua

14

15.7

Bearhop et al. (2002)

Garden warbler

11

5.0–5.7

Hobson and Bairlein (2003)

Yellow rumped warbler

4–6

7.45–27.7

Pearson et al. (2003)

Dunlin

11.2

10

Evans-Ogden et al. (2004)

Blood cells

American crow

29.8

 

Hobson and Clark (1993)

Plasma

American crow

2.9

 

Hobson and Clark (1993)

Yellow rumped warbler

0.4–0.7

0.7–1.7

Pearson et al. (2003)

Several authors have recognized the potential of using stored specimens to create longitudinal time series of isotope data (e.g., Szepanski et al. 1999; Hiron et al. 2001), or to compare diets of the same populations over very long-time intervals (Ainley et al. 2003). Museum collections or natural collections can thus be used as archives to test hypotheses about climatic change or other long-term processes (Ayliffe et al. 1992). For both mammals and birds, bone has been the most utilized tissue for these analyses, but other tissues, such as bird feathers and fish scales, have shown to be equally useful (Thompson et al. 1995; Pruell et al. 2003). This approach has been used to investigate long-term dietary change in a variety of species, from grizzly bears (Hildebrand et al. 1996) to striped bass (Pruell et al. 2003), and highlights the scientific values of long-term biological archives.

In these analyses, comparisons should preferably be made on samples of the same tissue, in order to avoid confounding effects that could be related to comparing different tissues (see below). This also means that inferences can be drawn from raw isotope data, without inclusion of fractionation coefficients, unless specific questions related to the isotopic composition of the sources are addressed. However, results from studies of tissues with fast molecular turnover, for instance plasma (with a turnover as short as only one day for many species, see Table 1), and with small sample sizes, should be treated somewhat with caution, since individual measures of these tissues are strongly influenced by the most recent dietary activities, i.e., similar to what has been discussed as a disadvantage for analyses of stomach and gut contents (Korschgren 1994).

2.2 Comparisons of tissues with different metabolic rates

Comparisons of isotopic signatures across tissues with different metabolic rates can potentially reveal information of temporal variation in diet, since each tissue assimilate elements over time spans specific to its metabolic rate. If we assume a species with two distinct seasonal diets, samples from a tissue with a fast turnover would reflect the dietary record from the season when it was sampled whereas samples from a tissue with sufficiently slow turnover would reflect dietary records over several seasons. Several authors, starting with Hobson (1993), have utilized this by comparing a suite of tissues to reveal dietary patterns on several different time scales. The most common tissues in these combinations are blood, liver, muscle and bone (Appendix 1). However, despite the novelty in this approach, we feel a general lack of studies that have taken full advantages of it to test specific hypotheses regarding temporal change in diets.

Knowledge of tissue specific turnover and fractionation is thus fundamental for a meaningful interpretation of the results from these types of comparisons. Since metabolic rates are related to body size (Schmidt-Nielsen 1984), the elemental turnover of a specific tissue will also be allometrically dependent. If turnover rates are unknown for the species in focus, we thus suggest a use of turnover rates obtained from another species as close in body size, taxonomy and ecology as the measured species as possible. In contrast to comparisons of the same tissue, comparisons of raw isotope data are in this context not very meaningful, since any differences between tissues could be caused by fractionation alone. Despite a large number of studies done on isotopic fractionation in different species and on different tissues (Table 2), reliable values of isotopic fractionation is often a limiting factor in these types of analyses. Due to a large inter-specific variation in fractionation, even within specific tissues (Table 2; see also Vanderklift and Ponsard 2003), we strongly suggest that efforts should be made to use species-specific fractionation values as far as possible when comparing values from different tissues. In studies that exclusively address hypotheses regarding seasonal change in diet, it would actually be enough to derive the fractionation difference between tissues, rather than the specific fractionation for each tissue. Although the importance of tissue specific fractionation coefficients has been recognized, they have typically not been applied when comparisons have been made between different tissues (e.g., Thompson and Furness 1995).
Table 2

Fractionation coefficients for 13C and 15N in different protein tissues and different species of mammals and birds

Tissue

Species

Δ 13C

Δ 15N

Source

Mammals

Collagen

Pig

2.98–3.7

 

Howland et al. (2003)

Mouse

3.5–4.4

 

DeNiro and Epstein (1981)

Gerbil

1.0–1.3

 

Tieszen and Boutton (1989)

Rat

3.8

 

Ambrose and Norr (1993)

Mouse

0.6–8.1

 

Tieszen and Fagre (1993)

Pig

2.9

 

Howland et al. (2003)

Rat

5.4

 

Jim et al. (2004)

Muscle

Gerbil

>1.0

 

Tieszen et al. (1983)

Seala

1.3

2.4

Hobson et al. (1996)

Red fox

1.1

3.6

Roth and Hobson (2000)

Domestic cattle

1.8

 

De Smet et al. (2004)

Liver

Gerbil

>1.0

 

Tieszen et al. (1983)

Seala

0.6

3.1

Hobson et al. (1996)

Red fox

0.4

3.6

Roth and Hobson (2000)

Domestic cattle

0.5

 

De Smet et al. (2004)

Whole blood

Red fox

0.7

2.6

Roth and Hobson (2000)

Domestic cattle

1.0

 

De Smet et al. (2004)

Red blood cells

Harbor sealb

1.1–1.6

1.5–2.0

Lesange et al. (2002)

Gray sealb

1.2–1.7

1.6–1.7

Lesange et al. (2002)

Harp sealc

1.4–1.9

1.4–2.0

Lesange et al. (2002)

Serum/Plasma

Red fox

0.6

4.2

Roth and Hobson (2000)

American black bear

0.4–4.5

 

Hildebrand et al. (1996)

Harbour sealb

0.6–0.8

2.7–3.5

Lesange et al. (2002)

Grey sealb

0.5–1.0

2.9–3.4

Lesange et al. (2002)

Harp sealc

0.5–1.0

3.0–3.6

Lesange et al. (2002)

Domestic cattle

1.7

 

De Smet et al. (2004)

Hair

Horse

1.8

 

Jonson et al. (1981)

Gerbil

1.0

 

Tieszen et al. (1983)

Seala

2.8

3

Hobson et al. (1996)

Red fox

2.6

3.4

Roth and Hobson (2000)

Harbor seal

2.3

2.3

Lesange et al. (2002)

Cattle

2.7

 

Sponheimer et al. (2003)

Goat

3.2

 

Sponheimer et al. (2003)

Alpaca

3.2

 

Sponheimer et al. (2003)

Llama

3.5

 

Sponheimer et al. (2003)

Rabbit

3.4

 

Sponheimer et al. (2003)

Forest hog

3.1

 

Cerling and Viehl (2004)

Domestic horse

0.15–0.44d

 

Ayliffe et al. (2004)

Domestic cattle

2.6

 

De Smet et al. (2004)

Brain

Gerbil

>1.0

 

Tieszen et al. (1983)

Skin

Seala

2.8

2.3

Hobson et al. (1996)

Whiskers

Seala

3.2

2.8

Hobson et al. (1996)

Nail

Seala

2.8

2.3

Hobson et al. (1996)

Lung

Seala

1.8

2.3

Hobson et al. (1996)

Heart

Seala

1.2

3.1

Hobson et al. (1996)

Kidney

Seala

1.3

2.7

Hobson et al. (1996)

Kidney fat

Domestic cattle

−3.3

 

De Smet et al. (2004)

Spleen

Seala

1.3

2.1

Hobson et al. (1996)

Lipids

Pig

−2.4

 

Howland et al. (2003)

Apatite

Pig

10.2

 

Howland et al. (2003)

Rat

9.5

 

Jim et al. (2004)

Birds

Collagen

Domestic chicken

0.8

1.5

Hobson and Clark (1992b)

Japanese quail

2.7

2.5

Hobson and Clark (1992b)

Ring-billed gull

2.6

3.1

Hobson and Clark (1992b)

Muscle

Domestic chicken

0.3

0.2

Hobson and Clark (1992b)

Japanese quail

1.1

1.0

Hobson and Clark (1992b)

Ring-billed gull

0.3

1.4

Hobson and Clark (1992b)

Dunlin

1.9

3.1

Evans-Ogden et al. (2004)

Liver

Domestic chicken

0.4

1.7

Hobson and Clark (1992b)

Japanese quail

0.2

2.3

Hobson and Clark (1992b)

Ring-billed gull

0.4

2.7

Hobson and Clark (1992b)

Dunlin

1.1

4.0

Evans-Ogden et al. (2004)

Whole blood

Japanese quail

1.2

2.2

Hobson and Clark (1992b)

Ring-billed gull

0.3

3.1

Hobson and Clark (1992b)

Great skuae

4.3–7.1 (1.1–2.3)f

2.8–4.2 (2.6–4.0)f

Bearhop et al. (2002)

Garden warbler

1.7

2.4

Hobson and Bairlein (2003)

Yellow-rumped warblerg

−1.2–2.2

1.7–2.7

Pearson et al. (2003)

Dunlin

1.3

2.9

Evans-Ogden et al. (2004)

Red blood cells

Dunlin

1.5

3

Evans-Ogden et al. (2004)

Serum/Plasma

American crow

0.5

 

Hobson and Clark (1993)

Yellow-rumped warblerg

−1.5–0.6

2.5–3.0

Pearson et al. (2003)

Dunlin

0.5

3.3

Evans-Ogden et al. (2004)

Feather

Domestic chicken

−0.4

1.1

Hobson and Clark (1992b)

Japanese quail

1.4

1.6

Hobson and Clark (1992b)

Ring-billed gull

0.2

3.0

Hobson and Clark (1992b)

Fish eating birds

3.3

4.4

Mizutani et al. (1992)

Great skuae

2.1–2.2 (5.3–7.0)f

4.6–5.0 (4.4–4.8)f

Bearhop et al. (2002)

Garden warbler

2.7

4.0

Hobson and Bairlein (2003)

Yellow-rumped warblerg

1.9–4.3

3.2–3.5

Pearson et al. (2003)

Kidney

Dunlin

1.3

4

Evans-Ogden et al. (2004)

For comprehensive reviews of Δ13C and Δ15N, see Ambrose and Norr (1993), Venderklift and Ponsard (2003)

aHarp seal (Pagophilus groenlandicus), harbour seal (Phoca vitulina) and ringed seal (Phoca hispida) combined

bRange of average values from different facilities

cRange of average values for three different diets

dRange of average values for three hypothetical isotope pools

eRange of average values for two diets, sprat and beef

fCalculated on diet including lipids

gRange of average values for different diets

2.3 Comparisons of sections from tissues with progressive growth

A cross section of a metabolically active tissue, such as muscle or internal organs, will contain a well mixed blend of old and newly synthesized protein. Some types of tissues, however, have progressive growth where the large bulk of the tissue is metabolically inactive (Welle 1999). Cross segments of these types of tissues will, thus, contain protein synthesized at different times and retain isotopic records in a chronological order reflecting time of assimilation (Jones et al. 1981). Tissues with this type of growth include teeth, guard hair, claws, feathers, fish scales, otoliths and to some extent bone.

Comparing isotope signatures of segments from these types of tissues has been used to assess temporal diet variation in a variety of different species. Darimont and Rimchen (2002) used section of guard hairs from wolves (Canis lupus) to explore seasonal use of salmon, while Schell et al. (1989) used sections of baleen in bowhead whales (Balena mysticetus) to explore seasonal use of different feeding locations. Teeth can also reveal temporal information, as illustrated by Hobson and Sease (1998) for pinnipeds and by Wiedeman et al. (1999) for bovids. In a similar way, Barnett (1994) used sections of hoofs to investigate seasonal grazing habits of caribou (Rangifer tarandus) and otoliths have been extensively studied to investigate temporal variation in fish (Appendix 1).

This approach can be more fine-tuned than the previous two and is capable of exploring short-term variations in diets. Since the comparisons are made across samples from the same tissue, there is no need to include coefficients for isotopic fractionation, unless the source signatures specifically are sought.

3 Discussion

Our review shows that the three alternatives to get a time perspective on diet through stable isotopes can be fruitful for addressing a wide range of temporally related hypotheses for a diverse range of species. For a shorter time perspective, such as seasonal differences, the methods of using different tissues with different molecular turnover or to use different section of inert tissues are suitable. However, the method of using the same tissue at different sampling occasions is suitable for both short and long-term specific questions. Taxonomically, mammals, birds and fish are well represented, but the lack of data for reptiles and amphibians is unfortunate, since the methods are equally suitable for these taxonomic groups as well. There are prospects for an increase in studies measuring isotopes in bone in reptiles, birds and fish, and generally an increase in studies using combinations of tissues to address specific hypotheses regarding seasonal dietary variation.

Although analyses of stable isotopes can reveal ecological patterns related to temporal diets, there are limitations to the methods. Gannes et al. (1997) recognized three processes that complicate dietary reconstruction from stable isotopes; (a) dietary components can be assimilated with different efficiencies, (b) isotopic fractionation can alter isotope values in tissues relative to the source, and (c) metabolic routing can disproportionally distribute the source element among different tissues (e.g., Schwarcz 1991). Gannes et al. (1997) pushed for an increase in laboratory experiments to improve our understanding of these underlying processes. Several years later, our knowledge of how these processes affect patterns in stable isotope ratios in animal tissues is still highly inadequate. We consequently argue that there still is an urgent need for controlled studies aimed at improving our understanding of how these different processes affect the abundance of stable isotopes in animal tissues.

As previously mentioned, isotopic fractionation is fundamental for the ability to interpret stable isotope data. It is also the most studied processes of the three recognized by Gannes et al. (1997) and the only one that commonly is attempted to account for in dietary reconstructions. However, fractionation has been suggested to vary even within a given tissue, probably partly depending on gross intake of dietary protein (Sealy et al. 1987). Some recent studies support this, thus further complicating inferences about dietary intake from stable isotopes (Bearhop et al. 2002; Lesange et al. 2002; Pearson et al. 2003; Sponheimer et al. 2003; De Smet et al. 2004). Although there is an increasing body of literature presenting experimental measurements of fractionation (Table 2), both in different tissues and in different species, we find it surprising that there are currently very few studies that present isotopic fractionation values for bone collagen (e.g., Ambrose and Norr 1993). Bone collagen is an important protein from several aspects. It is a single large protein with a relatively well-known metabolism (Welle 1999). Further, it has a slow metabolic rate, and thus reflects dietary records over a long-time period, which for medium and large size animals can reflect years or even decennia (Chisholm et al. 1982). Since collagen can be extracted from old bones, it is also an ideal tissue to investigate long-term dietary changes, for instance, from a climatic perspective (Ambrose 1990).

Metabolic routing, i.e., differential distribution of source molecules to different metabolic pathways, can also confound interpretations of stable isotope data (Howland et al. 2003). For instance, if ingested molecules are stored in reservoirs before being assimilated into new tissue, there will be a time lag between ingestion of elemental sources and their incorporation in new tissue. Such time lags would affect temporal isotope variation between and within tissues. Further, when molecules are stored in reserves, for example in adipose tissue, the actual fractionation will be the combined fractionation of source to storage tissue and from storage tissue to target. In a recent study on domestic horse, carbon incorporated into tail hair was attributed to three different carbon pools, each with specific turnover and fractionation (Ayliffe et al. 2004). This suggests that potential implications of these processes should be considered when interpreting isotope data, even from data sets from single tissues.

The physiological state of an animal will also affect the isotope ratios in its tissues. Starving and water stressed animals have been shown to be isotopically enriched in nitrogen compared to nonstressed animals, mainly due to recycling of nitrogen (Ambrose and DeNiro 1986; Selay 1987; Ambrose 1993; Hobson et al. 1996). The same effect can be seen in nursing offspring and hibernating species (Nelson et al. 1998; Lidén and Angerbjörn 1999). Age has, in similar ways, been suggested to change isotope ratios (Ponsard and Averbuch 1999; Keeling and Nelson 2001; Witt and Ayliffe 2001), although the precise mechanism in which this occurs in different species still is poorly understood. However, it is most likely linked to specific metabolism during growth (Lidén and Angerbjörn 1999). For animals with continuous growth, age dependent change in isotope is further complicated by different metabolic pathways that incorporate molecules during turnover and growth.

Finally, a temporal pattern in measured stable isotopes could be caused by variation in the animals’ dietary sources rather than by dietary variation. For instance, a strong seasonal pattern of stable isotopes in a predator could be the result of three sources of temporal variation; (a) an actual diet shift of the predator, with a switch between prey with distinct isotopic signatures, (b) isotopic variation in the prey caused by seasonal variation in its diet or physiology, without any dietary variation in the predator, or (c) isotopic variation in the preys diet, with a constant diet both for the predator and the prey. Similarly, any temporal dietary variation between food sources with undistinguishable isotope signatures will not be detected by stable isotope analyses. To avoid misleading conclusions because of this, it is always useful to try to understand isotopic variation not only in the focal organism, but also in organisms one or a few trophic levels below.

To conclude, analyses of stable isotopes can be a very useful tool in assessing temporal dietary patterns of many animals. So far, the approach of using the same tissue from different occasions is the most common one. The two alternatives provide interesting perspectives with other time scales and noninvasive sampling. Taxonomically, we encourage more studies on amphibians and reptiles. However, we stress that all sources of isotopic variation, including ecological, physiological, biochemical and physical, must be considered when attempting to interpret stable isotope data. Without a basic understanding of how these factors interact in their effects on isotopic variation, any attempt of reconstruct dietary patterns from stable isotopes is likely to generate misleading conclusions.

Copyright information

© Springer-Verlag 2005