Smith And Worthy 2006

Comparative Biochemistry and Physiology, Part B 143 (2006) 486 – 499 www.elsevier.com/locate/cbpb

Stratification and intra- and inter-specific differences in fatty acid composition of common dolphin (Delphinus sp.) blubber: Implications for dietary analysis Heather R. Smith ⁎, Graham A.J. Worthy 1 Physiological Ecology and Bioenergetics Laboratory, Texas A&M University at Galveston, Galveston, TX 77551, USA Received 31 May 2005; received in revised form 19 October 2005; accepted 31 December 2005 Available online 24 February 2006

Abstract Sixty-five fatty acids were quantified in the blubber of common dolphins (Delphinus delphis, D. capensis) incidentally caught off the coast of southern California. Dolphins were grouped by sex, reproductive status and species, and a blubber sample was collected at a mid-lateral site located caudal to the trailing edge of the dorsal fin. Samples were divided horizontally into inner, middle and outer layers and gradients in fatty acid content (mass percent) were observed across the depth of the blubber. Levels of monounsaturated fatty acids were greatest in the outer layer, whereas levels of saturated and polyunsaturated fatty acids were greatest in the inner layer. Degree of stratification was greatest in sexually mature dolphins. Blubber of sexually immature, but physically mature, male dolphins was also highly stratified, suggesting that this difference may be attributed to differences in diet. Classification and regression tree analysis resulted in the fewest misclassifications when dolphins were grouped by species, possibly indicating that these closely related animals forage on different prey species. Dietary-derived fatty acids were typically selected as splitting criteria in classification and regression tree analyses, suggesting that the observed differences in fatty acid composition between the various groups of dolphins may be attributed to differences in diet. © 2006 Elsevier Inc. All rights reserved. Keywords: Blubber; Cetacean; Diet; Dolphin; Fatty acid; Feeding habits; Layer; Stratification

1. Introduction Blubber is a biochemically dynamic tissue, in which fatty acid (FA) composition is potentially influenced by diet. Marine mammals rely on their blubber layer for thermoregulation, streamlining, buoyancy, and energy storage (Parry, 1949; Worthy and Edwards, 1990; Pabst et al., 1999), and researchers have begun to take advantage of its potential as a record of dietary intake (Iverson et al., 1997b; Walton et al., 2000; Hooker

⁎ Corresponding author. Washington Cooperative Fish and Wildlife Research Unit, School of Aquatic and Fishery Sciences, University of Washington, Box 355020, Seattle, WA 98195-5020, USA. Tel.: +1 206 221 5453; fax: +1 206 616 9012. E-mail address: [email protected] (H.R. Smith). 1 Present address: Physiological Ecology and Bioenergetics Laboratory, Department of Biology, University of Central Florida, Orlando, FL, 328162368, USA. 1096-4959/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cbpb.2005.12.025

et al., 2001). In order to effectively use FA composition of blubber in diet analysis techniques, the biochemical structure of this tissue must be considered. Biochemical stratification, or layering, has been observed in the blubber of Pacific walrus (Odobenus rosmarus divergens) (West et al., 1979b), polar bears (Ursus maritimus) (GrahlNielsen et al., 2003), and several species of phocid seals (Fredheim et al., 1995; Best et al., 2003). Other studies, however, have shown that phocid blubber is homogeneous and therefore not layered (Jangaard and Ke, 1968; Käkelä and Hyvärinen, 1993). In contrast, biochemical stratification has consistently been observed in baleen whale blubber (Ackman et al., 1965, 1975a,b; Lockyer et al., 1984; Olsen and Grahl-Nielsen, 2003). Biochemical layering has also been noted in the blubber of toothed whales such as bottlenose dolphins (Tursiops truncatus) (Shoda et al., 1993; Samuel and Worthy, 2004), harbor porpoise (Phocoena phocoena) (Koopman et al., 1996), killer whales (Orcinus orca) (Worthy et al., 2003; Krahn et al., 2004) and numerous other odontocetes (Koopman, 2001).

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

While it is evident that stratification exists, the number of distinct layers within cetacean blubber remains unclear. Koopman et al. (1996) found differences in FA composition in porpoise blubber that was divided into inner and outer layers, and suggested the existence of a continuous gradient in composition. Ackman et al. (1965, 1975a) noted differences between inner and outer layers in mysticete whales. Evidence for the existence of three distinct layers in bottlenose dolphins has been suggested based on histological (Cowan and Worthy, 1991), toxicological (Shoda et al., 1993), and biochemical (Samuel and Worthy, 2004) evidence. This stratification has been attributed to different levels of metabolic activity within layers (Ackman et al., 1975a,b; Koopman et al., 1996; Koopman, 2001). Differences in FA composition within and between species have been attributed to diet (West et al., 1979a; Borobia et al., 1995; Smith et al., 1996; Iverson et al., 1997a,b). Intraspecific differences have also been attributed to age (Koopman et al., 1996; Koopman, 2001), body site (Koopman et al., 1996), season (Samuel and Worthy, 2004), reproductive status (Samuel and Worthy, 2004) and starvation (Koopman, 2001). Interspecific differences have been attributed to phylogeny (Koopman, 2001) and thermal regime (Koopman, 2001; Worthy et al., 2003). With few exceptions (West et al., 1979a,b; Koopman, 2001; Samuel and Worthy, 2004), the possible influences of sex and reproductive status on FA composition within a single species have received little attention. Examination of FA stratification in marine mammal blubber is confounded by the type of ester prepared from lipid extracted from the tissue of interest. Methyl esters are commonly used to quantify medium- and long-chain FA in order to make dietary inferences (Smith et al., 1997; Walton et al., 2000; Iverson et al., 2004). These FAs have been used as “biomarkers” in diet studies as they are generally passed from the prey to predator adipose storage tissues without modification, whereas ingested shortchain FAs are generally oxidized as a source of energy (Pond, 1998). Unlike methyl esters, butyl esters allow for the quantification of short-chain fatty acids (Christie, 1972), and have been used in studies where levels of short-chain FA such as isovaleric acid (iso 5 : 0) are of interest (Koopman et al., 1996, 2003). Common dolphins (Delphinus sp.) are relatively small, toothed whales (Suborder Odontoceti) that are easily identified at sea by their black or dark grey V-shaped saddle. Adult animals weigh an average of 80 kg, and are sexually dimorphic with males being larger than females. Common dolphins are distributed worldwide in temperate, tropical and subtropical oceans. They are found along most coasts over the continental shelf, and are commonly associated with prominent bathymetric features such as seamounts and ridges (Evans, 1994). The distribution of common dolphins has also been correlated with the distribution of preferred prey species (Young and Cockcroft, 1994), water temperature, and possible competitive exclusion by spinner (Stenella longirostris) and spotted (S. attenuata) dolphins (Evans, 1975). Delphinus sp. forage on numerous species of fish, cephalopods and crustaceans throughout the water column (Fitch and Brownell, 1968; Jones, 1981; Evans, 1994; Silva, 1999). Stomach content analysis has indicated that the relative proportions of prey types vary with several factors

487

including season and age of individual (Evans, 1975; Young and Cockcroft, 1994; Chou et al., 1995; Silva, 1999). Prior to 1994, common dolphins off the southern coast of California were grouped into short- and long-beaked populations. These sympatric populations have not been observed to form mixed herds (Evans, 1975), and were later recognized as distinct species which could be differentiated based on aspects of color pattern, external morphology and skeletal characters (Heyning and Perrin, 1994). Rosel et al. (1994) examined mitochondrial DNA sequences and found no DNA haplotypes common to both forms, providing further support for the existence of two species. These two species of common dolphins are typically parapatrically distributed, with D. delphis (short-beaked) found in pelagic waters, and D. capensis (long-beaked) in nearshore waters less than 100 fathoms deep (Evans, 1975). As noted by Heyning and Perrin (1994), these two species must be exploiting the environment differently as their distributions overlap off the southern coast of California. It is reasonable to expect that foraging strategies, and therefore diet, influenced by differences in body size and skull morphology, would differ between these two species. There is little published data supporting this hypothesis, however, as previous diet composition studies have failed to distinguish between or compare the diets of these two species. As with other cetaceans, common dolphins spend much of their time away from shore and below the surface, making direct observation of dietary intake near impossible. Other dietary analysis techniques such as scat and stomach content analyses are of limited use with cetaceans as their scats are virtually impossible to collect, they do not haul-out, and stomach contents are typically only available from a small segment of the population such as stranded animals or those associating with and incidentally caught in fishing nets. Fatty acid analysis shows great promise for gaining insight into the diet of freeranging cetaceans as only a small piece of blubber is required for analysis. However, in order to design appropriate sampling protocols for this type of work, factors that may influence the fatty acid composition of cetacean blubber must be considered. The overall aim of this research was to examine the fatty acid composition of common dolphin blubber in the context of making dietary inferences for these species. The first objective of this study was to determine if Delphinus blubber could be divided into biochemically distinct layers. The second objective was to examine the intraspecific differences in blubber FA composition with respect to sex and reproductive status. The final objective was to determine if D. delphis could be distinguished from D. capensis on the basis of blubber FA composition. 2. Materials and methods 2.1. Sample collection Blubber samples (n = 84) and accompanying biological data (species, sex, body size, reproductive condition) were made available by the Southwest Fisheries Science Center (SWFSC)

488

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

Table 1 Species, sex and sexual maturity of Delphinus delphis and D. capensis blubber samples used in fatty acid analysis Species

Sex

Sexual maturity

D. delphis

Male (58)

Mature (48) Immature (10) Mature (10)

Female (17)

D. capensis

Male (4) Female (5)

Immature (7) Mature (2) Immature (2) Mature (3) Immature (2)

Notes Long a (5), Short (5) PL b (1), PnL (2), nPL (3), nPnL (4)

nPL (1), nPnL (2)

Numbers in parentheses indicate sample sizes. a Long (i.e., physically mature) animals had a mean body length of 174 ± 2 cm, short (i.e., physically immature) animals had a mean body length of 137 ± 2 cm (mean ± SEM). b PL = pregnant and lactating; PnL = pregnant, not lactating; nPL = not pregnant, lactating; nPnL = not pregnant, not lactating.

of the National Marine Fisheries Service (NMFS), and the Los Angeles County Museum of Natural History (LACM) (Table 1). Samples were collected from common dolphins caught incidental to commercial fisheries operating along the coast of southern California from 1990 to 2000. NMFS observers aboard gillnet vessels collected blubber samples from dead, by-caught common dolphins on the left or right lateral side of the body, just caudal to the trailing edge of the dorsal fin (Jefferson et al., 1994). The entire blubber layer was collected such that the epidermis and underlying muscle were still attached to the blubber. Blubber samples were wrapped in foil or placed in whirl-pak bags, and frozen until later analysis. Skin and gonads were also collected, and geographic location, sex, and total body length were recorded (Perrin et al., 1976; Jefferson et al., 1994). Reproductive condition was determined using a variety of criteria. Males were classified as sexually mature if the weight of the right testis without epididymis was greater than 200 g (Ferrero and Walker, 1995). Reproductive tracts of female dolphins were examined in the laboratory to determine state of sexual maturity and pregnancy (Akin et al., 1993). Females were classified as sexually mature if either corpus albicans or corpus lutea were present in either ovary (Perrin et al., 1976; Akin et al., 1993; Ferrero and Walker, 1995). Mammary glands were palpated to determine if the female was lactating (Jefferson et al., 1994). NMFS observers performed the test for lactation in the field if the whole carcass could not be brought back to the laboratory. Dolphins were grouped by species, sex, and level of sexual maturity for this study. Sexually mature females were further grouped according to state of pregnancy and lactation. Sexually immature male D. delphis could easily be divided into two groups on the basis of body length. “Long” animals had a mean body length of 174 ± 2 cm and were termed physically mature, while “short” animals had a mean body length of 137 ± 2 cm and were termed physically immature. 2.2. Sample analysis To reduce the loss of lipid during sample preparation, blubber was cut while still partially frozen. Freezer burn, identified

by a dark yellowish color, was trimmed from the edges of the sample and discarded. With no visible layering apparent, each sample was divided into five layers of approximately equal thickness. The outer (adjacent to dermis), middle and inner (adjacent to muscle) layers were retained, while the two transitional layers were discarded (Fig. 1). Duplicate subsamples (ca. 0.5 g) were taken from each of the three layers. Lipid was extracted from blubber according to Folch et al. (1957) and Iverson et al. (2001a,b). Fatty acid methyl esters (FAME) were prepared from the lipid extract using the Hilditch reagent (0.5 N H2SO4 in methanol) according to Iverson et al. (1992). FAME were analyzed using temperature-programmed gas-liquid chromatography (GLC). A Perkin Elmer Autosystem XL gas chromatograph (capillary injector, FID) linked to a computerized integration system (Turbochrom 4 software, PE Nelson) was fitted with a 30 m × 0.25 mm i.d. column coated with 50% cyanopropyl methylpolysiloxane (0.25 μm film thickness; liquid phase DB-23; J and W Scientific Inc., Folsom, CA). Helium was used as the carrier gas. The temperature program employed was developed by Iverson et al. (1992) as further modified by S.J. Iverson (personal communication). Identifications of FAME peaks were determined using commercially prepared standards (Nu-Chek Prep., Elysian, MN) and known reference mixtures. Reference mixtures contained the full suite of FA quantified, and peaks were identified using silver nitrate (argentation) chromatography and gas-chromatography/mass spectrometry according to Iverson et al. (1997b, 2001a,b). All chromatograms were individually examined to ensure that peak names were applied correctly. Individual FAs were converted to mass percent of total FA and unknowns by applying theoretical response factors relative to 18 : 0 (Ackman, 1972, 1991). FAs were named using the IUPAC shorthand nomenclature of C : Dn − x, where C is the number of carbons in the FA, D is the number of double bonds, and n − x denotes the position of the first double bond as numbered from the terminal methyl end of the FA. epidermis O — outer layer

x — discarded layer ~15 mm

M — middle layer x — discarded layer

I — inner layer

muscle

Fig. 1. Division of blubber into layers prior to fatty acid composition analysis. Horizontal lines indicate where blubber was cut with a scalpel. Inner, middle, and outer layers were retained for analysis, whereas the two transitional layers, indicated with an “x”, were discarded.

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

2.3. Data analysis Mean values for individual FA were calculated from duplicate subsamples for each layer for each individual dolphin. Except as noted, and with the exception of 16 : 4n − 3 and 18 : 3n − 6, these values were used in all subsequent calculations and analyses. FA 16 : 4n − 3 and 18 : 3n − 6 often co-eluted with adjacent peaks, preventing accurate measurement, and were thus omitted. Values for total saturated FA (∑SFA), monounsaturated FA (∑MUFA), n − 3 polyunsaturated FA (∑(n − 3) PUFA), n − 6 polyunsaturated FA (∑(n − 6)PUFA), and the ratio ∑(n − 3)PUFA / ∑(n − 6)PUFA were calculated for each layer for each individual dolphin. Mean values for individual FA, summed FA and the ∑(n − 3)PUFA / ∑(n − 6)PUFA ratio were calculated for each layer for each group of dolphins. 2.4. Blubber stratification Stratification of FA in common dolphin blubber was primarily examined in sexually mature, male D. delphis, as this group of dolphins had the largest sample size (n = 48). Differences in FA composition between the inner, middle, and outer layers were examined using multivariate and univariate approaches. A subset of 15 FA commonly present in quantities N 1% by mass was often analyzed in place of the entire data set to facilitate the use of statistical methods sensitive to the number of variables being examined: 14 : 0, 14 : 1n − 9, 14 : 1n − 7, 14 : 1n−5, 16:0, 16:1n−9, 16:1n−7, 18:0, 18:1n−9, 18:1n−7, 18 : 2n − 6, 20 : 1n − 9, 20 : 5n − 3, 22 : 5n − 3, and 22 : 6n − 3. Mass percent data were arcsin transformed (Steele and Torrie, 1980) prior to most analyses. 2.4.1. Multivariate approach Principal components analysis (PCA) was used to explore the possibility that sexually mature, male D. delphis blubber may be divided into three layers on the basis of FA composition. Principal components (PC) were generated from the subset of 15 arcsin transformed FA listed above, and PC with eigenvalues ≥ 1 were retained for interpretation. The varimax rotation was applied to enhance interpretation of the PC. In an attempt to reduce the effects of a priori variable selection, PC were also generated from an additional subset of 23 arcsin transformed FA commonly present in quantities N 0.5% by mass: 12 : 0, 14 : 0, 14 : 1n − 9, 14 : 1n − 7, 14 : 1n − 5, iso 15 : 0, 16 : 0, 16 : 1n − 9, 16:1n −7, 16 :2n− 4, 17 :1,18:0, 18 :1n− 9, 18 :1n − 7, 18 :2n − 4, 18:2n−6, 18:3n−3, 20:1n−9, 20:4n−6, 20:4n−3, 20:5n−3, 22 : 1n − 11, 22 : 5n − 3, and 22 : 6n − 3. Scatter plots of PC1 vs. PC2 were constructed for each PCA completed. As the PC could be roughly interpreted as the summary variables calculated earlier (e.g., ∑SFA), these summary variables were used in the discriminant analysis (DA) instead of the PC as they encompassed nearly all the original data and did not require a priori variable selection. Discriminant analysis (DA) was employed to determine the extent to which the layers could be distinguished from one another using ∑SFA, ∑MUFA, ∑(n − 3)PUFA, and ∑(n − 6)PUFA as discriminating variables. Jack-knifed classification rates were computed,

489

and differences in mean canonical scores between layers were examined using ANOVAs with α = 0.025 (0.05 / 2 = 0.025). Post hoc comparisons were made using Scheffé's test. A scatter plot of the mean canonical scores for the two DF was constructed. 2.4.2. Univariate approach To test for differences between the three layers, a MANOVA was performed on the data using all 15 arcsin transformed FA. Differences in mass percent of individual FA were then examined using individual ANOVAs with α = 0.0033. The Bonferroni method of modifying the alpha value (α = 0.05 / 15 = 0.0033) to reduce the risk of committing a Type I error was employed (Johnson and Wichern, 1998). Scheffé's test (α = 0.05) was used post hoc to examine the nature of the differences for each FA as indicated by the ANOVAs. The summary variables ∑SFA, ∑MUFA, ∑PUFA, ∑(n − 3)PUFA and ∑(n − 6)PUFA (all arcsin transformed), and ∑(n − 3)PUFA / ∑(n − 6)PUFA were tested for layering differences using MANOVAs, ANOVAs with modified alpha values, and Scheffé's test. 2.4.3. Classification and regression tree analysis approach The ability to distinguish among the three layers on the basis of FA composition in mature, male D. delphis blubber was examined using classification and regression tree (CART) analysis (Smith et al., 1997; Iverson et al., 1997b; Brown et al., 1999; Walton et al., 2000; Samuel and Worthy, 2004). All FA data, including summary variables were used in CART analysis. The classification and regression tree was generated using the TREES function in SYSTAT (version 10 for Windows, SPSS Inc., Chicago, IL). The default stopping criteria were used with the exception of the minimum proportion reduction in error, which was set at 1%. 2.4.4. Blubber stratification in other dolphin groups Small sample sizes for the other groups of dolphins precluded similar statistical analysis. The existence of FA stratification in these other groups was examined visually using bar graphs of the 15 most abundant FA. 2.5. Degree of stratification Koopman (2001) developed a stratification index (SI) in order to compare overall FA stratification in blubber among groups of odontocetes. Koopman's SI was calculated as the sum of the absolute values of the differences in mass percent between the inner and outer blubber layers for the 16 most abundant FA. SI values were similarily calculated for Delphinus in this study, using the mass percent values for the 15 FAs commonly present in quantities N1% by mass. Mean SI values were then calculated for each dolphin group. Mean SI values were compared for male D. delphis using an ANOVA with α = 0.05. Scheffé's test was used post hoc to examine differences among groups. Mean SI values were compared visually among female D. delphis, male D. capensis, and female D. capensis as small sample sizes precluded statistical testing.

490

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

2.6. Intraspecific differences in FA composition The ability to distinguish between blubber samples on the basis of potential intraspecific differences in FA composition was examined using CART analysis as described previously. Analyses were conducted separately on each layer for both short- and long-beaked common dolphins. Sex was used as the grouping variable for sexually immature and mature animals, as well as for both mature and immature dolphins combined. Sexual maturity was used as the grouping variable for males and females (mature vs. immature), and reproductive status (i.e., pregnant — P, lactating — L; not pregnant — nP, not lactating — nL) was used for sexually mature females alone (PL, nPL, PnL, nPnL). Sexually mature D. delphis females were also simply categorized as pregnant or non-pregnant, and lactating or non-lactating, and CART analysis was conducted on these two additional data sets. In addition to the modified stopping criteria listed earlier, minimum group size was altered to reflect group sizes smaller than five. 2.7. Interspecific differences in FA composition Potential interspecific differences in FA composition were also examined using CART analysis. Separate analyses were conducted for sexually mature males, sexually immature males, nPL females, nPnL females, and sexually immature females. Data from each layer were analyzed separately, and minimum group size was altered to reflect groups smaller than five. 3. Results Sixty-five individual FA (excluding 16 : 4n − 3 and 18 : 3n − 6) were regularly identified in all blubber samples examined. The majority of these FAs were present in quantities ≤0.5% by mass and only 15 FAs were commonly present in quantities ≥ 1% by mass: 14 : 0, 14 : 1n − 9, 14 : 1n − 7, 14 : 1n − 5, 16 : 0, 16 : 1n − 9, 16 : 1n − 7, 18 : 0, 18 : 1n − 9, 18 : 1n − 7, 18 : 2n − 6, 20 : 1n − 9, 20 : 5n − 3, 22 : 5n − 3, and 22 : 6n − 3. In mature male D. delphis, these 15 FAs accounted for total mean values ± SEM of 87.23 ± 0.20%, 87.39 ± 0.19%, and 86.13 ± 0.28% by mass in inner, middle and outer layers, respectively (Table 2). ∑MUFA accounted for the greatest proportion of all FA measured in all dolphin groups examined. In all dolphin groups combined, ∑MUFA ranged from 52.68 ± 0.71% by mass in the inner layer to 69.42 ± 0.67% by mass in the outer layer (mean ± SEM). ∑SFA, ∑(n − 3)PUFA, and ∑(n − 6)PUFA each accounted for less than 30% by mass of all FA and unknowns in all samples. In contrast to ∑MUFA, ∑SFA was greatest in the inner layer (22.85 ± 0.28% by mass), and least in the outer layer (13.19 ± 0.41% by mass). ∑(n − 3)PUFA and ∑(n − 6)PUFA were also greatest in the inner layer, and least in the outer layer. ∑(n − 3) PUFA ranged from 18.10 ± 0.55% by mass in the inner layer to 9.68 ± 0.47% by mass in the outer layer, while ∑(n − 6) was fairly consistent across all blubber layers, ranging from 3.13 ± 0.05% by mass in the inner layer to 2.56 ± 0.05% by mass in the outer layer. ∑(n − 3)PUFA / ∑(n − 6)PUFA ratios ranged from 5.70 ± 0.11 in the inner layer to 3.64 ± 0.10 in the outer layer (Table 3).

Table 2 Fatty acids (FAs) commonly present in quantities ≥0.5% by mass and summary variables for inner, middle, and outer blubber layers of sexually mature, male, D. delphis (n = 48) FA

Inner

Middle

Outer

Saturated fatty acids 12 : 0 14 : 0 Iso 15 16 : 0 18 : 0 ∑SFA

0.21 ± 0.01 4.30 ± 0.07 0.25 ± 0.01 13.89 ± 0.38 2.50 ± 0.09 22.90 ± 0.41

0.36 ± 0.01 3.77 ± 0.09 0.40 ± 0.02 7.68 ± 0.38 1.28 ± 0.07 15.13 ± 0.49

0.66 ± 0.02 3.02 ± 0.08 0.78 ± 0.04 4.06 ± 0.18 0.77 ± 0.03 11.18 ± 0.27

Monounsaturated fatty acids 14 : 1n − 9 14 : 1n − 7 14 : 1n − 5 16 : 1n − 9 16 : 1n − 7 17 : 1 18 : 1n − 9 18 : 1n − 7 20 : 1n − 9 22 : 1n − 11 ∑MUFA

0.16 ± 0.01 0.13 ± 0.01 0.60 ± 0.06 0.78 ± 0.04 10.80 ± 0.36 0.60 ± 0.01 31.38 ± 0.61 2.71 ± 0.04 3.13 ± 0.10 0.77 ± 0.05 53.37 ± 0.89

0.66 ± 0.06 0.61 ± 0.05 1.95 ± 0.13 1.92 ± 0.10 17.61 ± 0.54 0.73 ± 0.02 32.15 ± 0.53 2.42 ± 0.05 2.35 ± 0.08 0.45 ± 0.03 62.78 ± 0.91

2.25 ± 0.13 1.59 ± 0.07 3.89 ± 0.13 3.86 ± 0.13 24.03 ± 0.45 0.80 ± 0.02 30.79 ± 0.41 1.79 ± 0.06 1.13 ± 0.07 0.15 ± 0.02 71.81 ± 0.60

Polyunsaturated fatty acids 16 : 2n − 4 18 : 2n − 6 18 : 3n − 3 20 : 4n − 6 20 : 4n − 3 20 : 5n − 3 22 : 5n − 3 22 : 6n − 3 ∑(n − 3)PUFA ∑(n − 6)PUFA ∑(n − 3)PUFA / ∑(n − 6)PUFA ∑(15 FA)

0.55 ± 0.02 1.18 ± 0.02 0.47 ± 0.02 0.71 ± 0.03 0.57 ± 0.02 2.68 ± 0.18 3.26 ± 0.18 9.74 ± 0.35 17.41 ± 0.69 3.12 ± 0.06 5.49 ± 0.14 87.23 ± 0.20

0.56 ± 0.02 1.23 ± 0.02 0.49 ± 0.01 0.77 ± 0.02 0.54 ± 0.02 3.01 ± 0.14 2.49 ± 0.15 8.27 ± 0.30 15.40 ± 0.57 3.06 ± 0.05 4.97 ± 0.12 87.39 ± 0.19

0.51 ± 0.02 1.19 ± 0.02 0.70 ± 0.06 0.65 ± 0.02 0.36 ± 0.02 2.15 ± 0.10 1.11 ± 0.11 4.51 ± 0.30 9.20 ± 0.47 2.55 ± 0.06 3.51 ± 0.11 86.13 ± 0.28

∑(15 FAs) is the sum of all FAs commonly present in quantities ≥1% by mass. Values are mean ± SEM, mass percent of total FAs and unknowns.

3.1. Blubber stratification 3.1.1. Multivariate approach In the first PCA (using 15 FAs), two PC were retained and interpreted. Cumulatively, they accounted for 83.58% of the total variance. In the second PCA (using 23 FAs), five PC, cumulatively accounting for 90.44% of the total variance, were retained. In general, the SFA and MUFA loaded heavily on PC1, and the PUFA loaded heavily on PC2, PC3, PC4 or PC5. Scatter plots of the first 2 PC showed that the data points tended to group together according to layer (Fig. 2). Discriminant analysis resulted in the three layers being classified correctly 83% of the time (jack-knifed classification) (Table 4). The middle layer was misclassified more frequently than either the inner or outer layer (Table 4). Two significant discriminant functions (DF) were produced (Wilks' lambda = 0.1830, P b 0.001). The first mean canonical score was significantly different for all three layers; the inner layer had the highest score, whereas the outer layer had the lowest score. The second mean canonical score only differed significantly

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

491

Table 3 Fatty acid composition summary variables for all common dolphin groups

Delphinus delphis Male, s. mature c n = 48 Male, s. mature, strand n=2 Male, s. immature, long n=5 Male, s. immature, short n=5 Female, s. mature, PL a n=1 Female, s. mature, PnL n=2 Female, s. mature, nPL n=3 Female, s. mature, nPnL n=4 Female, s. immature n=7

Delphinus capensis Male, s. mature n=2 Male, s. immature n=2 Female, s. mature, nPL n=1 Female, s. mature, nPnL n=2 Female, s. immature n=2 All dolphins n = 86

Layer

SFA

MUFA

(n − 3) PUFA

(n − 6) PUFA

(n − 3) PUFA / (n − 6)PUFA

Ib M O I M O I M O I M O I M O I M O I M O I M O I M O

22.90 ± 0.41 15.13 ± 0.49 11.18 ± 0.27 23.57 ± 0.33 16.79 ± 1.94 12.55 ± 1.87 21.21 ± 1.15 13.57 ± 0.51 13.49 ± 0.65 21.85 ± 0.92 18.80 ± 0.65 20.34 ± 0.91 25.19 17.27 12.75 24.85 ± 0.40 14.88 ± 0.11 11.80 ± 0.44 23.91 ± 0.44 15.59 ± 0.59 14.56 ± 2.69 22.45 ± 2.30 15.04 ± 1.84 11.05 ± 0.80 23.04 ± 0.44 17.62 ± 0.76 18.29 ± 1.21

53.37 ± 0.89 62.78 ± 0.91 71.81 ± 0.60 45.73 ± 0.61 51.26 ± 3.74 62.54 ± 4.16 58.30 ± 2.26 68.46 ± 1.63 72.33 ± 0.95 57.38 ± 1.20 62.97 ± 1.21 66.02 ± 1.03 58.50 64.93 74.07 50.85 ± 5.64 64.93 ± 4.23 72.91 ± 0.62 53.32 ± 2.82 65.47 ± 0.97 65.34 ± 7.33 52.76 ± 4.61 62.14 ± 5.41 71.88 ± 2.09 53.84 ± 1.19 64.89 ± 1.04 66.70 ± 1.51

17.41 ± 0.69 15.40 ± 0.57 9.20 ± 0.47 23.80 ± 0.80 24.86 ± 2.03 17.37 ± 2.72 14.68 ± 1.89 11.51 ± 1.10 6.52 ± 0.34 15.41 ± 0.93 12.71 ± 0.74 7.65 ± 0.56 11.69 12.73 6.40 18.60 ± 4.85 14.19 ± 3.90 7.60 ± 1.20 16.61 ± 2.71 12.41 ± 1.24 12.33 ± 4.87 18.48 ± 3.15 16.24 ± 3.37 9.18 ± 2.34 17.11 ± 1.01 11.46 ± 1.17 8.04 ± 1.50

3.12 ± 0.06 3.06 ± 0.05 2.55 ± 0.06 3.64 ± 0.10 3.72 ± 0.01 3.24 ± 0.23 2.82 ± 0.13 2.70 ± 0.09 2.18 ± 0.06 2.88 ± 0.08 2.61 ± 0.04 2.17 ± 0.09 2.32 2.57 2.01 2.81 ± 0.25 2.80 ± 0.34 2.36 ± 0.22 3.01 ± 0.24 2.89 ± 0.13 2.85 ± 0.33 3.08 ± 0.25 3.03 ± 0.29 2.48 ± 0.29 2.95 ± 0.11 2.56 ± 0.09 2.22 ± 0.12

5.49 ± 0.14 4.97 ± 0.12 3.51 ± 0.11 6.55 ± 0.41 6.68 ± 0.54 5.33 ± 0.46 5.14 ± 0.43 4.22 ± 0.29 2.98 ± 0.09 5.33 ± 0.20 4.85 ± 0.23 3.52 ± 0.16 5.05 4.95 3.18 6.51 ± 1.14 4.96 ± 0.78 3.20 ± 0.22 5.44 ± 0.50 4.28 ± 0.31 4.08 ± 1.13 5.89 ± 0.64 5.18 ± 0.61 3.51 ± 0.56 5.79 ± 0.21 4.44 ± 0.35 3.50 ± 0.47

I M O I M O I M O I M O I M O I M O

21.75 ± 1.56 17.44 ± 3.64 12.35 ± 2.09 22.81 ± 0.79 17.36 ± 1.07 15.00 ± 1.21 25.90 22.02 16.63 21.49 ± 0.47 19.42 ± 0.41 19.13 ± 2.65 23.81 ± 0.80 21.18 ± 1.11 20.86 ± 0.61 22.85 0.28 15.95 0.35 13.19 0.41

47.20 ± 0.46 55.62 ± 6.51 67.39 ± 8.10 39.82 ± 0.22 54.31 ± 3.76 62.29 ± 3.75 38.57 42.75 53.68 45.74 ± 3.99 49.82 ± 6.72 57.45 ± 3.19 43.44 ± 0.22 49.95 ± 0.52 54.24 ± 1.25 52.68 0.71 61.97 0.77 69.42 0.67

22.15 ± 1.42 19.17 ± 1.54 11.90 ± 5.00 28.54 ± 0.87 19.18 ± 5.51 12.14 ± 6.28 27.73 27.50 21.18 25.33 ± 2.82 22.86 ± 5.69 14.30 ± 5.44 25.27 ± 0.43 21.88 ± 0.27 17.72 ± 1.79 18.10 0.55 15.43 0.52 9.68 0.47

3.68 ± 0.50 3.47 ± 0.45 3.13 ± 0.74 3.90 ± 0.13 3.67 ± 0.26 3.18 ± 0.46 3.41 3.50 3.58 3.58 ± 0.06 3.57 ± 0.13 3.24 ± 0.42 3.80 ± 0.03 3.52 ± 0.02 3.04 ± 0.12 3.13 0.05 3.02 0.05 2.56 0.05

6.20 ± 1.24 5.56 ± 0.28 3.63 ± 0.74 7.33 ± 0.02 5.15 ± 1.13 3.61 ± 1.46 8.13 7.87 5.92 7.05 ± 0.66 6.36 ± 1.36 4.27 ± 1.12 6.65 ± 0.17 6.21 ± 0.12 5.82 ± 0.36 5.70 0.11 5.02 0.10 3.64 0.10

Values are means ± SEM, mass percent of total fatty acids and unknowns. a PL = pregnant and lactating; PnL = pregnant, not lactating; nPL = lactating, not pregnant; nPnL = not pregnant, not lactating. Long (i.e., physically mature) animals had a mean body length of 174 ± 2 cm, short (i.e., physically immature) animals had a mean body length of 137 ± 2 cm (mean ± SEM). b I = Inner, M = middle, O = outer. c s. mature = sexually mature; s. immature = sexually immature.

between the inner and middle, and middle and outer layers (ANOVA, both P b 0.001; Scheffé's test with α b 0.05) (Table 4). The scatter plot of mean canonical scores showed that the DF could be used to differentiate between the three layers; however, there was some overlap between groups. The scatter plot also demonstrated that the first DF was much more effective than the second DF at separating the three groups (Fig. 3).

3.1.2. Univariate approach FA composition was found to differ significantly between the three layers when all 15 FAs in the subset were considered together (MANOVA, P b 0.001). Twelve of the 15 FAs examined were present in significantly different concentrations in all three layers (ANOVA, all P b 0.001; Scheffé's test with α b 0.05). Levels of 20 : 5n − 3 differed significantly only between the inner and outer, and middle and outer layers. No

492

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

a)

b) 2

PC2

1 0 -1 -2

3

I M I M I M O M I MI O O I I M M II I O M MM O M M I I I M M M M OOO M O M M I II I O OO O O M III M MO M M MO O M M M O O OO O O I III I M M O O M O OO O OMM IM I I O OM O M M I II I OO O M I I OO O O O M M I II I OO M O M III M OO I M M I I O I M I I

-3 -2

-1

0

1

2 1

PC2

3

0 -1 -2

O O

II I II I I I M I I III OOO O I I I M M MI II OO OO OO MO O O OO O M OO MM IMI O M O M M MM I MM II I OO M O O O O O O M O I I O OO OO OO O O M M MI MMM I M OM M IM MM II III M M MM M OO M M M I M OM M I M M I MI I I I I O I M I M I I M

-3

2

I

-4 -2

-1

PC1

0

1

2

PC1

Fig. 2. Scatter plots of the first two principal components (PC) for sexually mature, male D. delphis (n = 144) calculated from the subset of a) 15 arcsin transformed fatty acids (FAs), and b) 23 arcsin transformed FAs. The first 2 PC cumulatively account for 83.6% of the total variance in a), and 68.8% of the total variance in b). In both plots, data points corresponding to the inner, middle, and outer layers are marked with the letters I, M and O, respectively.

3.1.3. Classification and regression tree approach CART analysis effectively separated sexually mature, male D. delphis blubber into three layers, selecting 20 : 0, 14 : 1n − 9, Table 4 Jack-knifed classification matrix and mean canonical scores (± SEM) for two significant discriminant functions generated from the arcsin transformed summary variables (∑SFA, ∑MUFA, ∑(n − 3)PUFA, and ∑(n − 6)PUFA) for sexually mature, male D. delphis (ni = 48) Actual

Predicted membership

Membership

Inner

Middle

Outer

% correct

Score 1

Score 2 a

Inner Middle Outer Total

45 6 0 51

3 32 5 40

0 10 43 53

94 67 90 83

2.4 ± 0.1 − 0.3 ± 0.2 − 2.1 ± 0.1

− 0.3 ± 0.2 0.6 ± 0.1 − 0.4 ± 0.2

a

14 : 1n − 7, and 14 : 0 as splitting variables (Fig. 5). The initial split separated the majority (45 / 48) of the outer layer samples from the inner and middle layer samples. Terminal nodes with misclassifications always included at least one middle layer sample: three inner samples were misclassified as middle, one middle sample was misclassified as inner, three outer samples were misclassified as middle, and two middle samples were misclassified as outer. The tree had an overall misclassification rate of 9 / 144. 3.1.4. Blubber stratification in other dolphin groups Visual examination of the most abundant FA revealed noiceable layering differences in the other groups of common dolphins. Fatty acid gradients were apparent across the blubber in all groups. In general, 14 : 1n − 9, 14 : 1n − 7, 14 : 1n − 5, 16 : 1n − 9, and 16 : 1n − 7 increased in mass percent from the inner to outer layer, while 14 : 0, 16 : 0, 18 : 0, 18 : 1n − 7, 20 : 1n − 9, 22 : 5n − 3, and 22 : 6n − 3 decreased in mass percent from the inner to outer layer. In contrast, 18 : 2n − 6 tended to be 5

2

LAYER

FACTOR(2)

significant differences were found for 18 : 1n − 9 or 18 : 2n − 6. A significant increase in mass percent of 14 : 1n − 9, 14 : 1n − 7, 14 : 1n − 5, 16 : 1n − 9, and 16 : 1n − 7 was observed across the blubber when moving from the inner to outer layer. In contrast to this, 14 : 0, 16 : 0, 18 : 0, 18 : 1n − 7, 20 : 1n − 9, 22 : 5n − 3, and 22 : 6n − 3 decreased across the blubber from the inner to outer layer (Fig. 4). Significant layering differences were also found when the summary variables were examined. The three layers differed significantly when ∑SFA, ∑MUFA, ∑(n − 3)PUFA and ∑(n − 6) PUFA were considered together (MANOVA, P b 0.001). ∑SFA and ∑MUFA were present in significantly different quantities in all three layers (ANOVA, all P b 0.001); ∑SFA decreased across the blubber from the inner to outer layer, while ∑MUFA increased across the blubber from the inner to the outer layer. ∑(n − 3)PUFA and ∑(n − 6)PUFA also decreased from the inner layer to the outer, however; only the differences between the inner and outer, and middle and outer layers were significant. ∑(n − 3)PUFA / ∑(n − 6)PUFA was also found to differ between all three layers (ANOVA, P b 0.001), decreasing from the inner to the outer layer.

Inner Middle Outer -1

Mean canonical scores

The second mean canonical score only differed significantly between the inner and middle, and middle and outer layers (ANOVA, both P b 0.001; Scheffé's test with α b 0.05).

-4 -4

-1

2

5

FACTOR(1) Fig. 3. Canonical scores plot for the 2 discriminant functions generated from the summary variables (arcsin transformed) for sexually mature, male D. delphis (n = 48). The centroids of the 3 ellipses are significantly different from one another (MANOVA, P b 0.001).

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

493

Mass % Total FAs and Unknowns

35

30

Inner Middle Outer

25

20

15

10

5

22:6n-3

22:5n-3

20:5n-3

20:1n-9

18:2n-6

18:1n-7

18:1n-9

18:0

16:1n-7

16:1n-9

16:0

14:1n-5

14:1n-7

14:1n-9

14:0

0

Fig. 4. Mass percent values of fatty acids (FAs) commonly present in quantities ≥ 1% in sexually mature, male Delphinus delphis blubber (n = 48). Blubber was divided into inner, middle, and outer layers. Bar values are means; error bars are 1 SEM. With the exception of 18 : 1n − 9, 18 : 2n − 6, and 20 : 5n − 3, FAs were present in significantly different quantities in all 3 layers.

relatively consistent across the blubber layer, while 18 : 1n − 9 and 20 : 5n − 3 tended to be present in the highest concentration in the middle layer. 3.2. Degree of stratification SI values ranged from a low of 30.21 ± 2.67 in short, sexually immature, male D. delphis (n = 5) to a high of 58.58 ± 6.38 in pregnant, non-lactating, female D. delphis (n = 2; Table 5). SI values for male D. delphis differed significantly (ANOVA, P b 0.001), with the short, sexually immature males being significantly less stratified (30.21 ± 2.67; n = 5) than the sexually mature males (50.36 ± 1.52; n = 48). SI values of long, sexually immature males were not significantly less stratified than the sexually mature males. In general, blubber of the older animals was more stratified than the younger animals.

and in the middle and outer layer trees when only sexually mature females were analyzed. D. capensis nPL females separated from the sexually immature and nPnL females in both the middle and outer layer trees when sexually immature and mature animals were analyzed together, and also from the nPnL females in all trees when only sexually mature animals were analyzed. When the simplified data sets were analyzed (i.e., D. delphis females, P vs. nP, and L vs. nL) all trees had lower MRs than the previous analyses of data sets with more detailed reproductive states (Table 6). Of all CART analyses performed, the highest MR (9 / 17) occurred in all three layers in D. delphis females (Table 6). D. delphis sexually mature males often grouped together with “long” sexually immature males, while D. capensis sexually mature males always separated from sexually immature males (Table 6). 3.4. Interspecific differences in FA composition

3.3. Intraspecific differences in FA composition D. capensis were correctly grouped by sex 100% of the time when sexually mature and immature animals were analyzed separately. When sexually mature and immature animals were analyzed together, the middle and outer layers resulted in no misclassifications, while the inner layer had a MR = 1 / 9. In all cases, the trees consisted of only one split, using a SFA or MUFA with ≤17C as the splitting criteria (Table 6). D. delphis also grouped by sex, but typically resulted in higher MRs and more splits per tree. Longer chain FAs such as 22 : 6n − 3, 20 : 5n − 3, 20 : 4n − 6, 20 : 1n − 9, 20 : 0, and ∑(n − 3)PUFA were used as splitting criteria in analyses involving sexually mature animals (Table 6). In general, females that were pregnant and/or lactating tended to separate from animals that were not. D. delphis PL females separated from all other females in all trees when sexually immature and mature animals were analyzed together,

All dolphin groups examined were effectively classified to the species level on the basis of differences in FA composition. All samples were correctly classified in all analyses (i.e. MR = 0) with the exception of the outer layer in sexually mature males, where MR = 2 / 50. Thirteen of the 15 trees produced used only one split to classify the data points into groups. Shorter chain (≤ 16 C) saturated and monounsaturated FAs were selected as splitting criteria in all cases (Table 6). 4. Discussion FA composition of common dolphin blubber is comparable to that of other marine mammal species previously studied (Jangaard et al., 1963; Ackman and Eaton, 1966; West et al., 1979a; Iverson et al., 1997b; Samuel and Worthy, 2004) with 16 : 0 as the predominant saturated FA (SFA), followed by 14 : 0

494

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

LAYER (48,48,48)

20:0<0.055 Inner (48,46,3) 49/97

Outer (0,2,45) 2/47

14:1n-9<0.230 Inner (43,3,0) 3/46

Middle (5,43,3) 8/51

14:1n-7<0.145

20:0<0.145

Inner (36,0,0) 0/36

Inner (7,3,0) 3/10

Inner (4,1,0) 1/5

Middle (1,42,3) 4/46

14:1n-9<0.205 Inner (5,0,0) 0/5

14:0<4.250

Middle (2,3,0) 2/5

Middle (1,35,0) 1/36

Middle (0,7,3) 3/10

Fig. 5. Classification tree for sexually mature, male D. delphis blubber samples (n = 144), grouped by layer. Each square represents a node in the tree and lists the following node information: Mode (most abundant sample type present), number of samples in each group (Inner, Middle, Outer), and Misclassification Rate (MR) = (number of samples from least abundant groups) / (total samples in node). The overall MR for the entire tree was 9 / 144. Fatty acids (FAs) used as splitting criteria for each node are listed, as well as values (mass percent total FAs and unknowns) used to produce indicated splits.

and 18 : 0. Overall, ∑MUFA accounted for the greatest proportion of FA measured, with 18 : 1n − 9 and 16 : 1n − 7 being predominant. PUFAs were predominated by the n − 3 class, primarily 22 : 6n − 3, 22 : 5n − 3, and 20 : 5n − 3. 18 : 2n − 6 and 20 : 4n − 6 were the most common n − 6 FA.

4.1. Blubber stratification In the present study, common dolphin blubber was found to be biochemically stratified, similar to the blubber of numerous other cetacean species. Common dolphin blubber had

Table 5 Mean stratification index (SI) values (±SEM) for all dolphin groups Species

Male Mature

Female a

D. delphis

50.36 ± 1.52 (n = 48)

D. capensis

46.38 ± 19.40 (n = 2)

Immature

44.33 ± 3.15 (long c) (n = 5) 30.21 ± 2.67 (short) (n = 5) 53.76 ± 12.08 (n = 5)

Mature

Immature

PL b

nPL

PnL

nPnL

49.03 (n = 1)

54.16 ± 3.86 (n = 3)

58.58 ± 6.38 (n = 2)

51.70 ± 4.73 (n = 4)

40.25 ± 4.79 (n = 7)

36.23 ± 3.47 (n = 2)

35.03 ± 2.16 (n = 2)

34.95 (n = 1)

SI was calculated as the sum of the absolute values of the differences in mass percent between the inner and outer blubber layers for the 15 FAs commonly present in quantities N1% by mass: 14 : 0, 14 : 1n − 9, 14 : 1n − 7, 14 : 1n − 5, 16 : 0, 16 : 1n − 9, 16 : 1n − 7, 18 : 0, 18 : 1n − 9, 18 : 1n − 7, 18 : 2n − 6, 20 : 1n − 9, 20 : 5n − 3, 22 : 5n − 3, and 22 : 6n − 3. a Mature = sexually mature, Immature = sexually immature. b PL = pregnant and lactating; PnL = pregnant, not lactating; nPL = not pregnant, lactating; nPnL = not pregnant, not lactating. c Long animals had a mean body length of 174 ± 2 cm, short animals had a mean body length of 137 ± 2 cm (mean ± SEM).

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

495

Table 6 Summary of CART results Grouping variable

Species

Individuals

Layer

n

MR

FA used as split criteria

Sex (M, F)

D. delphis

All

I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O I M O

75 75 75 58 58 58 17 17 17 9 9 9 5 5 5 4 4 4 17 17 17 10 10 10 10 10 10 10 10 10 5 5 5 3 3 3 58 58 58 4 4 4 4 4 4 6 6 6 9 9 9 50 50 50 12 12 12

9 / 75 4 / 75 7 / 75 7 / 58 5 / 58 3 / 58 2 / 17 2 / 17 2 / 17 1/9 0 0 0 0 0 0 0 0 9 / 17 9 / 17 9 / 17 4 / 10 5 / 10 3 / 10 1 / 10 2 / 10 1 / 10 1 / 10 2 / 10 0 1/5 2/5 2/5 0 0 0 5 / 58 6 / 58 5 / 58 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 / 50 0 0 0

Iso 15 : 0, 18 : 0, 22 : 6n − 3, iso 16 : 0, 12 : 0, 18 : 1n − 7, 18 : 1n − 9 20 : 4n − 6, 20 : 1n − 9, 12 : 0, 18 : 1n − 9, anti 15 : 0 16 : 1n − 5, 20 : 5n − 3, anti 15 : 0, ∑(n − 3)PUFA, iso 14 : 0 22 : 6n − 3, iso 15 : 0, 18 : 1n − 7, iso 16 : 0 20 : 4n − 6, ∑(n − 3)PUFA, iso 14 : 0, 20 : 0 16 : 2n − 6, iso 14 : 0, 20 : 0, 16 : 1n − 7 16 : 1n − 11, 14 : 0 16 : 1n − 11 16 : 1n − 11, 10 : 0 17 : 0 14 : 1n − 7 14 : 1n − 5 15 : 0 14 : 1n − 9 12 : 0 10 : 0 13 : 0 13 : 0 16 : 1n − 11 Iso 16 : 0 16 : 1n − 11 15 : 1n − 8 13 : 0 Iso 14 : 0, 10 : 0 18 : 1n − 11 14 : 0 14 : 0 18 : 2n − 6 13 : 0 Iso 14 : 0 10 : 0 10 : 0 10 : 0 12 : 0 10 : 0 10 : 0 22 : 1n − 9, 16 : 1n − 7, 14 : 1n − 9 anti 15 : 0 16 : 1n − 9 14 : 0 Iso 14 : 0 Iso 14 : 0 12 : 0 10 : 0 Iso 14 : 0 14 : 0 Iso 14 : 0 Iso 14 : 0 14 : 0 10 : 0 14 : 0 16 : 4n − 1 16 : 4n − 1, 16 : 1n − 11 15 : 0, 14 : 0 13 : 0 13 : 0 14 : 1n − 5

s a. mature

s. immature

Sex (M, F)

D. capensis

All

s. mature

s. immature

Reproductive status (PL b, PnL, nPL, nPnL, I)

D. delphis

All females

Reproductive status (PL, PnL, nPL, nPnL)

s. mature females

Reproductive status (L c, nL)

s. mature females

Reproductive status (P d, nP)

s. mature females

Reproductive status (nPL, nPnL, I)

D. capensis

Reproductive status (nPL, nPnL)

All females

s. mature females

Maturity (M e, IL, IS)

D. delphis

All males

Maturity (M, I)

D. capensis

All males

Species

Both

nPL females

nPnL females

s. immature females

s. mature males

s. immature males

Overall misclassification rate (MR) for each tree is listed. The fatty acid (FA) listed first was used to create the initial split. a s. mature = sexually mature. b PL = pregnant and lactating; PnL = pregnant, not lactating; nPL = not pregnant, lactating; nPnL = not pregnant, not lactating; I = sexually immature. c L = lactating; nL = not lactating. d P = pregnant, nP = not pregnant. e M = sexually mature, IL = immature long = sexually immature, physically mature; IS = immature short = sexually immature, physically immature.

496

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

decreasing levels of ∑SFA and increasing levels of ∑MUFA, from the inner to the outer layer, as has also been observed in bottlenose dolphins (Samuel and Worthy, 2004), killer whales (Worthy et al., 2003; Krahn et al., 2004), and fin whales (Balaenoptera physalus) (Ackman et al., 1965; Lockyer et al., 1984). These gradients have been hypothesized to correspond to the thermal gradient that exists across the blubber layer of animals that inhabit cold water. As the outer blubber layer is exposed to lower temperatures than the inner layer, a reduced level of SFA, which pack close together at low temperatures, and an increased level of MUFA, which do not pack together as readily as SFA, will better allow this outer blubber to remain fluid (Pond, 1998). The decrease in ∑SFA across common dolphin blubber observed in the present study is in contrast to the work of Koopman et al. (1996), who found an increase in ∑SFA from 23.18 ± 3.90 to 31.16 ± 5.34 mass percent from the inner to the outer layer in harbor porpoise blubber (data listed are for the “III Dorsal” body site, i.e., dorsal location, approximately at the trailing edge of the dorsal fin). It is likely that this discrepancy may be attributed to the different methods used to prepare esters from the lipid extract, rather than to species differences. Koopman et al. (1996) were interested in quantifying short chain FA such as isovaleric acid (iso 5 : 0), and therefore prepared butyl esters, as they are much less volatile than the methyl esters prepared in this study. As a result, ∑SFA as calculated by Koopman et al. (1996) included values for iso 4 : 0 and iso 5 : 0, both of which increase from the inner to the outer layer (0.56 ± 0.77 to 1.71 ± 0.90, and 3.11 ± 2.09 to 8.61 ± 4.05 mass percent, respectively). These short chain FAs were not measured in this study, and were therefore not included in our ∑SFA calculation. Additionally, preparation and analysis of different types of esters can result in different mass percent values for FA in the same sample. Because FAs are expressed as a percentage of all FAs identified in a given sample, when butyl esters are prepared, a greater number of FA are identified (i.e., short chain FA), and mass percent values of FA are proportionately lower than if methyl esters had been prepared. This produces lower ∑SFA values overall, and may contribute to the reduced magnitude of between layer differences. Examination of FA commonly present in quantities ≥ 1% by mass in mature male common dolphins revealed that levels of most individual FAs differed significantly between all three layers. All other FAs measured appear to follow the same pattern as the more abundant FAs of the same class (e.g., SFA, MUFA). The middle layer values were typically intermediate between the inner and outer layer values (e.g., Table 2), supporting the suggestion of Koopman et al. (1996) that a continuous gradient in FA composition runs across the entire blubber layer. Koopman et al. (1996) tested for significant differences in levels of individual FAs between the inner and outer blubber layers in mature, male, harbor porpoise. Similar to the results of Koopman et al. (1996), 18 : 1n − 9 was not found to differ significantly across the complete depth of blubber of mature, male, short-beaked common dolphins. Animals are capable of synthesizing 18 : 1n − 9 (Cook, 1985; Nelson, 1992) in addition to obtaining it from the diet. Consistently high levels

(approximately 30% by mass, this study) of 18 : 1n − 9 across the complete depth of blubber suggests that this FA may contribute to outer blubber fluidity, as well as being involved in metabolic activities. Animals feeding on diets deficient in essential FA (i.e., 18 : 2n − 6 and 18 : 3n − 3) have been shown to desaturate and elongate 18 : 1n − 9 to produce 20 : 3n − 9. This new FA is able to act as a partial substitute for 20 : 4n − 6, and when incorporated into phospholipids, adequately performs some membrane functions, but does not act as a precursor for prostaglandins (Cook, 1985). As 18 : 1n − 9 may be involved in such metabolic activities, it is reasonable that such high quantities would be maintained in the inner blubber layer, where they would be readily accessible in a time of need, in addition to the high quantities contributing to membrane fluidity in the outer layer. The magnitude of 18 : 2n − 6 levels (approximately 1.2% by mass) measured in mature, male, short-beaked common dolphins in this study was similar to the levels measured in mature, male, harbor porpoise by Koopman et al. (1996). However, unlike the situation found in harbor porpoise, no gradient across the blubber layer was found in common dolphins. Because 18 : 2n − 6 is an essential FA, it is available only through the diet (Cook, 1985). It is therefore logical to expect to find the highest levels of 18 : 2n − 6 in the more metabolically active inner blubber layer. It is possible that the lower than expected levels of 18 : 2n − 6 in the inner layer may be due to the physiological state of the animals collected, with the 18 : 2n − 6 demand exceeding current intake, thereby resulting in no gradient. CART analysis effectively separated the inner blubber layer samples from the outer blubber layer samples, but was less successful at separating out the middle blubber layer samples. The number of misclassifications involving inner and middle samples was roughly equal to the number of misclassifications involving outer and middle samples, and therefore does not indicate that the middle layer is more similar to either the inner or the outer layer. However, the initial split in the tree separated the majority (45 / 48) of the outer layer samples from the inner and middle layer samples, which suggests that the middle layer may be most similar to the inner layer. This suggestion is further supported by the ANOVA results, where 20 : 5n − 3 was not found to differ significantly between the inner and middle layers in sexually mature, male D. delphis. Samuel and Worthy (2004) found that the inner and middle layers tended to group together in male and non-lactating female bottlenose dolphins, while the outer and middle layers tended to group in lactating females. The tendency for the middle layer to reflect either the inner or outer layer was attributed to the reproductive status, and therefore the differing physiological and/ or dietary requirements of lactating animals. The suggestion of variable FA composition in the middle layer with respect to state of lactation is supported by the work of Aguilar and Borrell (1990), who found that reproductive status affected the lipid content of the inner and middle blubber layers in fin whales. Iverson et al. (1995) found that lactating animals selectively mobilize FA such as 20 : 5n − 3 from maternal blubber into the milk. The ability of female mammals to forage while lactating would maintain levels of such PUFA in the blubber during

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

lactation and will influence the degree to which the layers resemble one another. If stores of mobilized FA are not replenished, the FA composition of the mobilized layer will change over the course of lactation. Assuming that stores of PUFA are depleted as the animal lactates, mobilization of FA from the inner layer will result in the inner layer becoming more similar to the middle layer. In contrast, if PUFA are mobilized from the middle layer, the middle layer will become more similar to the outer layer. Variation in FA composition of the different blubber layers may also be caused by changes in diet. Bernard and Hohn (1989) noted differences in feeding habits between pregnant and lactating spotted dolphins, and Young and Cockcroft (1994) noted significant differences in the species of prey consumed by lactating versus non-lactating common dolphins. Depending on the FA composition of consumed prey items, and assuming that excess FAs are deposited in the inner blubber layer, the FA composition of the inner layer may become even more different than that of the middle layer, causing the middle and outer layers to appear more similar. 4.2. Degree of stratification The blubber of sexually mature, male common dolphins was more highly stratified than that of sexually immature males, in agreement with the results of Koopman et al. (1996), who suggested that this difference is due to changes in lipid metabolism, and therefore, changes in FA composition of the outer blubber layer over time. PUFA, and n − 3 PUFA in particular, were present in only very small quantities in the outer blubber layer of older porpoises (Koopman et al., 1996). It is also possible that differences in the degree of stratification may be attributed to differences in diet, with prey items rich in certain FA resulting in steep FA gradients across the blubber. Sexually immature males in the present study were further classified according to body length. “Short” animals were significantly shorter (137 ± 2 cm) than “long” animals (174 ± 2 cm) (ANOVA, P b 0.001; Scheffé's test with α = 0.05), which were more similar in length to sexually mature animals (186 ± 1 cm). While blubber of both groups of sexually immature animals were less stratified than the sexually mature animals, the “short” immature group was also less stratified than the “long” immature group. Dietary differences were the most likely causative factor since the longer immature dolphins were physically mature and were thus capable of catching prey items similar to those caught by the sexually mature animals. This ability would likely be due to a combination of larger body size and more foraging experience over the shorter, younger, and probably less experienced physically immature dolphins. An increase in the degree of stratification with sexual maturity was also observed in female D. delphis, similar to harbor porpoises (Koopman, 2001). Stratification index values for mature females were quite variable with respect to reproductive status, and by extension, dietary differences (e.g., Bernard and Hohn, 1989). Unfortunately, small sample sizes precluded statistical testing, and visual examination did not reveal any correlations.

497

Stratification index values for D. delphis from the present study (∼ 40.2 to 58.6) were higher than the SI values (27.5– 53.6, n = 13) calculated by Koopman (2001) for the same species. This could potentially be a function of the method by which blubber was divided into layers. Koopman (2001) divided blubber into inner and outer layers, discarding a layer of blubber between the two. In the present study, blubber was divided into five layers, and blubber between the inner and middle, and middle and outer layers was discarded (Fig. 1.). As such, the inner and outer layers examined in the present study were potentially thinner than the inner and outer layers examined by Koopman, which may have incorporated some of the blubber considered “middle” in the present study. As concentrations of FA in the middle layer were intermediate between inner and outer layer values (present study) then including portions of middle blubber in outer or inner blubber samples could potentially result in less pronounced differences between the inner and outer layers. Until standardized methods of analyzing fatty acid composition are adopted, it will remain difficult to determine whether such differences exist solely due to method, or are the result of phylogenetic, environmental or other factors. 4.3. Intraspecific differences in FA composition Foraging differences between male and female common dolphins have been observed (Young and Cockcroft, 1994; Chou et al., 1995), and this was expected to result in gender differences in FA composition. As dietary FAs are initially deposited in the inner blubber layer (Lockyer et al., 1984; Aguilar and Borrell, 1990; Koopman et al., 1996), it was expected that this would be where the FA composition of males and females would be most easily distinguished. Contrary to this expectation, the highest misclassification rates occurred in the inner layer analyses for both species of common dolphin. FA selected as splitting criteria in the CART analyses with sex as the grouping variable, and involving sexually mature D. delphis often included long chain PUFA of dietary origin (e.g., 22 : 6n − 3, 20 : 4n − 6, 20 : 5n − 3). In contrast to this, trees produced in all other CART analyses grouped by sex used FA with carbon chains ≤ 17C as splitting criteria. However, the majority of these shorter chain FAs were oddchained (e.g., 13 : 0, 15 : 0, 17 : 0) and hence also of dietary origin. Vertebrates are typically capable of biosynthesising only even-chained FA (Cook, 1985). Regardless of chain length, the selection of dietary FA as splitting criteria suggests that differences in diet between the two sexes do exist. Of the nine CART trees produced for D. capensis using sex as the grouping variable, only the tree for the inner layer, with sexually mature and immature animals combined, resulted in a misclassification (1 / 9). FA composition was expected to vary with reproductive status, as differences in lipid metabolism and diet selection have been observed in animals of different reproductive condition. Based on evidence that female mammals selectively mobilize certain FA from the blubber layer during lactation (Iverson et al., 1995), differences in the blubber FA composition

498

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499

of lactating and non-lactating females were anticipated. In the present study, lactating females tended to, but did not always separate from non-lactating females. Small cetaceans, including common dolphins, forage throughout an extended (∼6 months to 2 years) lactation period (Boyd et al., 1999). It is possible, that while common dolphins may selectively mobilize particular FA in milk production, the continued foraging by these animals reduces the need to mobilize FA stores to the same extent as has been observed in phocid seals (e.g., Iverson et al., 1995). In the present study, sexually mature and immature long D. delphis grouped together, while sexually immature short animals formed their own group. This provides evidence that differences in FA composition, as attributed to differences in feeding habits, are more likely correlated with body size (physical maturity) than sexual maturity, as discussed above. 4.4. Interspecific differences in FA composition There was only one misclassification when species was used as the grouping variable, supporting the idea that these two species of dolphins may be feeding on different prey items in order to co-exist in the same area (niche partititioning) (Heyning and Perrin, 1994). Even female animals, with the same reproductive status and therefore facing similar nutritional demands, were consistently separated from one another. In conclusion, this study demonstrates that common dolphin blubber is biochemically stratified, with inner, middle and outer blubber layers distinguishable from one another on the basis of FA composition. The majority of FAs were distributed along gradients across the blubber layer. Levels of MUFA were highest in the outer layer, while SFA and PUFA were highest in the inner layer. Blubber stratification was most pronounced in mature dolphins. FA composition was found to differ between dolphins grouped by sex, reproductive status and species. Selection of dietary FA as splitting criteria in CART analyses suggests that differences in FA composition observed in the present study may be attributed to dietary differences between these groups of animals. These results highlight the need to collect full-depth blubber cores in future studies where the intention is to use fatty acid data to make dietary inferences. It is important that the innermost layer of blubber be analyzed for this purpose. It is equally important that blubber be divided into layers in a consistent manner such that between-study comparisons of data may be possible. Acknowledgements We would like to thank S. Chivers, K. Robertson, K. Danil (SWFSC, NMFS), J. Heyning, and D. Janiger (LACM) for providing archived blubber samples. We thank S. Iverson, S. Lang, and D. Cowan for their technical expertise. We appreciate the constructive comments of B. Fadely, M. Lander and R. Watson on earlier versions of this manuscript. This manuscript was also improved by comments provided by two anonymous reviewers. Funding for this work was provided by: The Mooney Graduate Student Travel Grant, the Office of

Graduate Studies (TAMU), the Research Management Office (TAMUG), and a research award to GAJW from the Texas Higher Education Coordinating Board Advanced Research Program.

References Ackman, R.G., 1972. The analysis of fatty acids and related materials by gas– liquid chromatography. In: Holman, R.T. (Ed.), Progress in the Chemistry of Fats and Other Lipids, vol. 12. Pergamon Press, Oxford, pp. 165–284. Ackman, R.G., 1991. Application of gas–liquid chromatography to lipid separation and analysis: qualitative and quantitative analysis. In: Perkins, E.G. (Ed.), Analyses of Fats, Oils and Lipoproteins. American Oil Chemists' Society, Champaign, pp. 270–300. Ackman, R.G., Eaton, C.A., 1966. Lipids of the fin whale (Balaenoptera physalus) from north Atlantic waters. III. Occurrence of eicosenoic and docosenoic fatty acids in the zoolplankter Meganyctiphanes norvegica (M. Sars) and their effect on whale oil composition. Can. J. Biochem. 44, 1561–1566. Ackman, R.G., Eaton, C.A., Jangaard, P.M., 1965. Lipids of the fin whale (Balaenoptera physalus) from North Atlantic waters. Can. J. Biochem. 43, 1513–1520. Ackman, R.G., Hingley, J.H., Eaton, C.A., Logan, V.H., Odense, P.H., 1975a. Layering and tissue composition in the blubber of the northwest Atlantic sei whale (Balaenoptera borealis). Can. J. Zool. 53, 1340–1344. Ackman, R.G., Hingley, J.H., Eaton, C.A., Sipos, J.C., Mitchell, E.D., 1975b. Blubber fat deposition in mysticeti whales. Can. J. Zool. 53, 1332–1339. Aguilar, A., Borrell, A., 1990. Patterns of lipid content and stratification in the blubber of fin whales (Balaenoptera physalus). J. Mammal. 71, 544–554. Akin, P.A., Peltier [Robertson], K.M., Miller, R.B., 1993. Techniques for the preparation and examination of reproductive samples collected from dolphins in the eastern tropical Pacific. NOAA Tech. Mem., NMFS-SWFSC-192. Bernard, H.J., Hohn, A.A., 1989. Differences in feeding habits between pregnant and lactating spotted dolphins (Stenella attenuata). J. Mammal. 70, 211–215. Best, N.J., Bradshaw, C.J.A., Hindell, M.A., Nichols, P.D., 2003. Vertical stratification of fatty acids in the blubber of southern elephant seals (Mirounga leonina): implications for diet analysis. Comp. Biochem. Physiol. B 134, 253–263. Borobia, M., Gearing, P.J., Simard, Y., Gearing, J.N., Béland, P., 1995. Blubber fatty acids of finback and humpback whales from the Gulf of St. Lawrence. Mar. Biol. 122, 341–353. Boyd, I.L., Lockyer, C., Marsh, H.M., 1999. Reproduction in marine mammals. In: ReynoldsIII III, J.E., Rommel, S.A. (Eds.), Biology of Marine Mammals. Smithsonian Institution Press, Washington, pp. 218–286. Brown, D.J., Boyd, I.L., Cripps, G.C., Butler, P.J., 1999. Fatty acid signature analysis from the milk of Antarctic fur seals and Southern elephant seals from South Georgia: implications for diet determination. Mar. Ecol. Prog. Ser. 187, 251–263. Chou, L., Bright, A.M., Yeh, S., 1995. Stomach contents of dolphins (Delphinus delphis and Lissodelphis borealis) from North Pacific Ocean. Zool. Stud. 34, 206–210. Christie, W.W., 1972. The preparation of alkyl esters from fatty acids and lipids. In: Gunstone, F.D. (Ed.), Topics in Lipid Chemistry, vol. 3. Logos Press Ltd, London, pp. 171–197. Cook, H.W., 1985. Fatty acid desaturation and chain elongation in eukaryotes. In: Vance, D.E., Vance, J.E. (Eds.), Biochemistry of Lipids and Membranes. Benjamin-Cummings Publishing Co, Inc., Menlo Park, pp. 181–211. Cowan, D.F., Worthy, G.A.J., 1991. Body site variation in the structure of the blubber in Tursiops truncatas. Abstracts of the Ninth Bienniel Conference on the Biology of Marine Mammals. Chicago, IL, 5– 9 December 1991, p. 14. Evans, W.E., 1975. Distribution, differentiation of populations, and other aspects of the natural history of Delphinus delphis Linnaeus in the Northeastern Pacific. Ph.D. dissertation, University of California, Los Angeles. Evans, W.E., 1994. Common dolphin, white-bellied porpoise, Delphinus delphis Linnaeus, 1758. In: Ridgeway, S.H., Harrison, R.J. (Eds.), Handbook of Marine Mammals, vol. 5. Academic Press Ltd., London, pp. 191–224.

H.R. Smith, G.A.J. Worthy / Comparative Biochemistry and Physiology, Part B 143 (2006) 486–499 Ferrero, R.C., Walker, W.A., 1995. Growth and reproduction of the common dolphin, Delphinus delphis Linnaeus, in the offshore waters of the North Pacific Ocean. Fish. Bull. 93, 483–494. Fitch, J.E., Brownell Jr., R.L., 1968. Fish otoliths in cetacean stomachs and their importance in interpreting feeding habits. J. Fish. Res. Board Can. 25, 2561–2574. Folch, J., Lees, M., Sloan-Stanley, G.H., 1957. A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226, 497–509. Fredheim, B., Holen, S., Ugland, K.I., Grahl-Nielson, O., 1995. Fatty acid composition in blubber, heart and brain from phocid seals. In: Blix, A.S., Walløe, L., Ulltang, Ø. (Eds.), Whales, Seals, Fish and Man. Elsevier Science, Amsterdam, pp. 153–168. Grahl-Nielsen, O., Andersen, M., Derocher, A.E., Lydersen, C., Wiig, Ø., Kovacs, K.M., 2003. Fatty acid composition of the adipose tissue of polar bears and their prey: ringed seals, bearded seals and harp seals. Mar. Ecol. Prog. Ser. 265, 275–282. Heyning, J.E., Perrin, W.F., 1994. Evidence for two species of common dolphins (genus Delphinus) form the Eastern North Pacific. LACM Contr. Sci., vol. 442, pp. 1–35. Hooker, S.K., Iverson, S.J., Ostrom, P., Smith, S.C., 2001. Diet of northern bottlenose whales inferred from fatty-acid and stable-isotope analyses of biopsy samples. Can. J. Zool. 79, 1442–1454. Iverson, S.J., Sampugna, J., Oftedal, O.T., 1992. Positional specificity of gastric hydrolysis of long-chain n − 3 polyunsaturated fatty acids of seal milk triglycerides. Lipids 27, 870–878. Iverson, S.J., Oftedal, O.T., Bowen, W.D., Boness, D.J., Sampugna, J., 1995. Prenatal and postnatal transfer of fatty acids from mother to pup in the hooded seal. J. Comp. Physiol., B 165, 1–12. Iverson, S.J., Arnould, J.P.Y., Boyd, I.L., 1997a. Milk fatty acid signatures indicate both major and minor shifts in the diet of lactating Antarctic fur seals. Can. J. Zool. 75, 188–197. Iverson, S.J., Frost, K.J., Lowry, L.F., 1997b. Fatty acid signatures reveal fine scale structure of foraging distribution of harbor seals and their prey in Prince William Sound, Alaska. Mar. Ecol. Prog. Ser. 151, 255–271. Iverson, S.J., Lang, S., Cooper, M., 2001a. Comparison of the bligh and dyer and folch methods for total lipid determination in a broad range of marine tissue. Lipids 36, 1283–1287. Iverson, S.J., MacDonald, J., Smith, L.K., 2001b. Changes in diet of freeranging black bears in years of contrasting food availability revealed through milk fatty acids. Can. J. Zool. 79, 2268–2279. Iverson, S.J., Field, C., Bowen, W.D., Blanchard, W., 2004. Quantitative fatty acid signature analysis: a new method of estimating predator diets. Ecol. Monogr. 74, 211–235. Jangaard, P.M., Ke, P.J., 1968. Principal fatty acids of depot fat and milk lipids from harp seal (Phoca groenlandica) and hooded seal (Cystophora cristata). J. Fish. Res. Board Can. 25, 2419–2426. Jangaard, P.M., Ackman, R.G., Burgher, R.D., 1963. Component fatty acids of the blubber of fat from the common or harbour seal Phoca vitulina concolor De Kay. Can. J. Biochem. Physiol. 41, 2543–2546. Jefferson, T.A., Myrick Jr., A.C., Chivers, S.J., 1994. Small cetacean dissection and sampling: a field guide. NOAA Tech. Mem., NMFS-SWFSC-198. Johnson, R.A., Wichern, D.W., 1998. Applied Multivariate Statistical Analysis, 4th ed. Prentice Hall, Inc., Upper Saddle River, NJ. Jones, R.E., 1981. Food habits of smaller marine mammals from Northern California. Proc. Calif. Acad. Sci. 42, 409–433. Käkelä, R., Hyvärinen, H., 1993. Fatty acid composition of fats around the mystacial and superciliary vibrissae differs from that of blubber in the Saimaa ringed seal (Phoca hispida saimensis). Comp. Biochem. Physiol. B 105, 547–552. Koopman, H.N., 2001. The structure and function of odontocete blubber. Ph.D. dissertation, Duke University, Durham, NC. Koopman, H.N., Iverson, S.J., Gaskin, D.E., 1996. Stratification and age-related differences in blubber fatty acids of the male harbour porpoise (Phocoena phocoena). J. Comp. Physiol. B 165, 628–639. Koopman, H.N., Iverson, S.J., Read, A.J., 2003. High concentrations of isovaleric acid in the fats of odontocetes: variation and patterns of

499

accumulation in blubber vs. stability in the melon. J. Comp. Physiol. B 173, 247–261. Krahn, M.M., Herman, D.P., Ylitalo, G.M., Sloan, C.A., Burrows, D.G., Hobbs, R. C., Mahoney, B.A., Yanagida, G.K., Calambokidis, J., Moore, S.E., 2004. Stratification of lipids, fatty acids and organochlorine contaminants in blubber of white whales and killer whales. J. Cetacean Res. Manag. 6, 175–189. Lockyer, C.H., McConnell, L.C., Waters, T.D., 1984. The biochemical composition of fin whale blubber. Can. J. Zool. 62, 2253–2562. Nelson, G.J., 1992. Dietary FAs and lipid metabolism. In: Chow, C.K. (Ed.), Fatty Acids in Foods and Their Health Implications. Marcel Dekker, Inc., New York, NY, pp. 437–471. Olsen, E., Grahl-Nielsen, O., 2003. Blubber fatty acids of minke whales: stratification, population identification and relation to diet. Mar. Biol. 142, 13–24. Pabst, D.A., Rommel, S.A., McLellan, W.A., 1999. The functional morphology of marine mammals. In: ReynoldsIII III, J.E., Rommel, S.A. (Eds.), Biology of Marine Mammals. Smithsonian Institution Press, Washington, DC, pp. 15–72. Parry, D.A., 1949. The structure of whale blubber, and a discussion of its thermal properties. Q. J. Microsc. Sci. 90, 13–25. Perrin, W.F., Coe, J.M., Zweifel, J.R., 1976. Growth and reproduction of the spotted porpoise, Stenella attenuata, in the offshore eastern tropical Pacific. Fish. Bull. 74, 29–269. Pond, C.M., 1998. The Fats of Life. Cambridge University Press, Cambridge,UK. Rosel, P.E., Dizon, A.E., Heyning, J.E., 1994. Genetic analysis of sympatric morphotypes of common dolphins (genus Delphinus). Mar. Biol. 119, 159–167. Samuel, A.M., Worthy, G.A.J., 2004. Variability in fatty acid composition of bottlenose dolphin (Tursiops truncatus) blubber as a function of body site, season, and reproductive status. Can. J. Zool. 82, 1933–1942. Shoda, L.K.M., Worthy, G.A.J., Wade, T.L., 1993. Lipid stratification and contaminant distribution in the blubber of bottlenose dolphins. Abstracts of the Tenth Biennial Conference on the Biology of Marine Mammals. Galveston, TX, 11– 15 November 1993, p. 99. Silva, M.A., 1999. Diet of common dolphins, Delphinus delphis, off the Portuguese continental coast. J. Mar. Biol. Assoc. UK 79, 531–540. Smith, R.J., Hobson, K.A., Koopman, H.N., Lavigne, D.M., 1996. Distinguishing between populations of fresh- and salt-water harbour seals (Phoca vitulina) using stable-isotope ratios and fatty acid profiles. Can. J. Fish Aquat. Sci. 53, 272–279. Smith, S.J., Iverson, S.J., Bowen, W.D., 1997. Fatty acid signatures and classification trees: new tools for investigating the foraging ecology of seals. Can. J. Fish Aquat. Sci. 54, 1377–1386. Steele, R.G.D., Torrie, J.H., 1980. Principles and Procedures of Statistics: A Biometrical Approach, 2nd ed. McGraw-Hill, Inc., New York. Young, D.D., Cockcroft, V.G., 1994. Diet of common dolphins (Delphinus delphis) off the southeast coast of Southern Africa: opportunism or specialization? J. Zool. London 234, 41–53. Walton, M.J., Henderson, R.J., Pomeroy, P.P., 2000. Use of blubber fatty acid profiles to distinguish dietary differences between grey seals Halichoerus grypus from two UK breeding colonies. Mar. Ecol. Prog. Ser. 193, 201–208. West, G.C., Burns, J.J., Modafferi, M., 1979a. Fatty acid composition of blubber from the four species of Bering Sea phocid seals. Can. J. Zool. 57, 189–195. West, G.C., Burns, J.J., Modafferi, M., 1979b. Fatty acid composition of Pacific walrus skin and blubber fats. Can. J. Zool. 57, 1249–1255. Worthy, G.A.J., Edwards, E.F., 1990. Morphometric and biochemical factors affecting heat-loss in a small temperate cetacean (Phocoena phocoena) and a small tropical cetacean (Stenella attenuata). Physiol. Zool. 63, 432–442. Worthy, G.A.J., Samuel, A.M., Worthy, T.A.M., 2003. Composition of killer whale (Orcinus orca) skin and blubber: implications for fatty acid signature analysis. Fifteenth Biennial Conference on the Biology of Marine Mammals. Greensboro, North Carolina, December 14–20, 2003.