Introduction Results.......... Discussion...... Summary

4 downloads 0 Views 879KB Size Report
Jan 7, 1986 - 1976, Nolan & Ketterson ..... Nolan & Ketterson 1983) are untenable, and can greatly ... grateful to Dr. C. Lloyd, Dr. M. Pienkowski, M. Kersten.
GEOGRAPHIC VARIATION IN THE LEAN MASS OF DUNLINS WINTERING IN BRITAIN N. C. DAVIDSON, J. D. UTTLEY & P. R. EVANS Department of Zoology, University of Durham, South Road, Durham DHI 3LE, U.K. Received

7

January 1986, revised 14 March 1986

CONTENTS 1.

Introduction

.......... Results.......... 3.1. Mass and body size.. 3.2. Variationsin mass between sites................... Discussion......

2. Materials and methods

)

4.

191. 191

193

193 193 195

4.1. Geographical variations in mass ................... 195 4.2. Implications forstudiesof body condition....... 196 5. Acknowledgements ... ........ 6. 7.

Summary

...

196 196

1. INTRODUCTION

Many studies of the body condition of birds have been based wholly or chiefly on measurements of the total mass of live individuals. This approach has the particular advantage that seasonal changes in the condition of individual birds can be followed if they can be captured several times (Pienkowski e/ al. 1984). Seasonal variation in the total mass of an individual arises in large part from seasonal variation in fat load, as has been shown for several bird species e.g. for shorebirds by Davidson (1981a, b), Dugan el al. (1981). As a result, many authors who have used total mass as an indicator of body condition, have assumed that total mass and lean dry mass do not vary, either seasonally or geographically (e.g. Connell et a\.1960, Rogers & Odum 1964, Elliott et al. 1976, Nolan & Ketterson 1983); they have interpreted all changes in total mass as changes in fat load. However seasonal variation in lean mass has been reported for several species during breeding (Jones & Ward 1976, Ankney 1977, Ankney & Maclnnes 1978, Houston et al. 1983), before migration (Evans 1969, Fry et al. 1972, Baggott 1975, Davidson 1981a, 1983, Davidson & Evans 1986), and during starvation in severe winter weather (Davidson 1981b, Dtgan et al. 1981, Davidson & Evans 1982, Piersma 1984a) so that the assumptions of stability in lean mass do not always hold. Much less attention has been paid to vari-

ation in the lean mass of populations of a species wintering in different areas. Whilst lean mass of some species may remain stable throughout a winter (in the absence of se-

vere weather) at a single wintering site (David-

son 1981a), large geographical differences in the lean mass of House Sparrows Passer domesticus and Starlings Sturnus vulgaris were re-

ported by Blem (1973,1981). Blem found that mass was correlated with latitude and weather conditions, at least over the part of the geographical range he studied. In waders, Davidson (1981a, 1983) showed that within a species, lean mass was generally lower in birds wintering in equatorial and southern hemisphere areas than in northern temperate regions. Davidson e/ a/. (1986) found recently that the pectoral muscles of Dunlins Calidris alpina wintering in Britain were smaller on milder estuaries. Since the pectoral muscles form t5-20Vo of the lean dry mass, it follows that the total lean dry mass could be expected to vary similarly. Here we report on geographical variations in the lean mass and lean dry mass of Dunlins wintering in Britain, in relation to body size and environmental conditions. We consider the implications of such variations in lean mass for the validity of assessing the body condition of live birds from measurements of total mass. 2. MATERIALS AND METHODS

Between 1973 and 1984 a total of 175 adult and 74 juvenile Dunlins of the race alpina were collected in midwinter from the 7 British estuaries shown in Fig. 1. Some birds were killed accidentally during large-scale netting and ringing operations and forwarded to us; most from Teesmouth were collected under licence from the Nature Conservancy Council for studies of body condition and heavy metal loads. Only

birds collected between December and early March, and in the absence of severe weather, Ardea 74 (1986): 191-198

192

GEOGRAPHIC VARIATION LEAN MASS OF DUNLIN

are included in the anlaysis, since the lean mass of Dunlins remains stable at these times, but varies seasonally before 'and after (Davidson

1981a), and can be reduced during severe

weather (Davidson 1981b).

Birds were deep-frozen until analysis, and aged by plumage characteristics (Prater et al. 1977) as immature (in their first winter of life) or adult.

At least one measure of external body size was made on each bird. For most samples wing length (maximum chord) and bill length (exposed culmen) were measured; for samples col-

lected recently we have also measured total head (1.e. head-plus-bill) length (Green 1980) and tarsus plus toe length (Anderson 1975, Piersma 1984b). All measures were taken to the nearest 1 mm using a stopped wing-rule (for

F,r 0Y

o

€,

Fig. 1. Estuaries in Britain from which samples of Dunlins were obtained.

[Ardea 74

wing length and tarsus plus toe length) or calipers (for bill length and total head length). Most samples were dried to constant mass in vacuum ovens at 50'C, then fat was extracted in a Soxhlet apparatus with 60-80 "C bp petroleum ether, and samples were redried to constant mass (the lean dry mass). Lean mass was calculated as the total mass minus the mass of fat. Samples from Conway Bay and Portsmouth

Harbour were dried in convection ovens at 80'C and fat was extracted using chloroform (Pienkowski et al. 1979). There was no differ-

ence in the percentage water content of samples

analysed

in

these two ways, so data from all

samples were included in the analysis. For most

samples, the pectoral muscles (pectoralis major and supracoracoideus) of the right side of the body were excised before drying, and processed in the same way as the residual carcass. The mass of the pectoral muscles was added to that of the residual carcass to give total lean dry mass and total lean mass. Variations in the pectoral muscle mass of these samples is reported in Davidson et al. (1986). We treat all birds in our samples as winter residents on the estuary from which they were collected, since Dunlins move little between estuaries in Britain between December and early

March (Pienkowski & Evans 1984). Also the body condition of known (ringed) residents was the same as that of birds of unknown origin (Davidson et al.1986). Winter weather conditions at each site were assessed by use of the 30-year mean midwinter (December/January) air temperature for the nearest coastal weather station to each sampling site, published by the British Meteorological Office (1916). This measure reflects not only the average severity of winter temperatures at a site, but is correlated also with the relative harshness of conditions during particularly severe winters (Dugan 1981). Statistical analyses were made using SPSS (Nie e/ al. 1975). Step-wisg multiple regressions were used to examine the within-site variations of mass with body size measures. Between-site variations in mass were assessed by analysis of variance and by partial correlation. Since between-site differences in mass could arise merely from variations in the proportions of differ-

1e861

GEOGRAPHIC VARIATION LEAN MASS OF DUNLIN

193

3.2. VARIATIONS IN MASS BETWEEN SITES

ent-sized birds present at each site, wing length and bill length (the correlates of body size available for each site) were entered as covariates in ANOVA before testing between-site variations. Masses were transformed to their cube roots be-

3.1. MASSANDBODYSIZE

Since body size accounted for much of the variation in mass at each site, the effects of body size were excluded in analyses of between-site variation by the prior inclusion of wing length and bill length as covariates in ANOVA. There remained significant variation between sites in lean dry mass in both adults (1-way ANQVA, Fo.r+s : 2.32, p : 0.036) and immatures (1-way ANOVA, Fo,ss : 2.69, p : 0.023). Lean mass

For each age class in which sample sizes were )10, larger individuals had greater lean dry mass and lean mass. Table 1 gives the correlation coefficients between mass and body size de-

' ' ':::'

fore correlations were made with linear measures of body size.

3. RESULTS

also varied between sites in adults (Fo.r+s : 11.44, p < 0.001) and immatures (Fu.r, : 11.05, Further anarvses under-

rived from stepwise multiple regressions, using wing length and bill length as measures of body size, since they.were available for most samples. The inclusion of both measures of body size improved the correlations in all but one case. Variability in wing length and bill length within samples from each site accounted for a substantial part of the variability in both lean dry mass and lean mass at most of the sites: coefficients of determination (r2) for lean dry mass were 31-90% in adults and l&-64Vo in immatures, and for lean mass were 3114% in adults and 53-:77% in immatures. When additional body size measurements were available, their inclusion increased coefficients of determination by up to 14% (Table 1). There was little consistency between sites or age-classes as to which measure of body size accounted for most variation in mass, although in each sample all the body size measures were intercorrelated.

o

l,::..':"'

20

e18 E

=16 co

)o

1L

12

n

20

16 60

3010

7

113 33 3?220

32

r_--T----l-1

|----r---l-1 3456

3/+56oC

Fig2. Lean dry mass (LD) of Dunlins in relation to

mean

midwinter air temperature (Ta) of wintering site. The mean (horizontal bar) + 1 standard deviation (thick vertical bar), range (thin vertical line) and sample size are shown. For adults, LD0.33 -- 2.5778 -0.0133Ta, r : -0.16, n = 175, p : 0.030; for immatures, LD033 : 2.5922 -0.0300Ta, r : -0.27, n:74, p: 0.021. Dotted horizontal lines show means adjusted to the mean bill length and wing length of the whole sample.

Table1. Correlationcoefficientsof leandrymass(LDM0.33),andleanmass(LMo3l)withmeasuresof thebodysizeof Dunlins wintering in Britain. Body size measures are wing length (WL), bill length (BL), total head length (THL) and tarsus plus toe length (TT)

WL and BL onlv

r(LDM) adults Menai Straits Firth of Forth Wash Severn

Teesmouth Portsmouth Harbour immatures Menai Straits Firth of Forth Teesmouth

I

16 23 48

.66 .82 .84 .56 .56

l0

l9

32

20

11

26

best estimate

r(LM)

.86 .56

16

.57 .70 .92 .56 .87

23 10

.85

.95

.79 .60 .80

.74 .88 .82

19

.90

.84

1

.73

.93

b

20

r(LM)

.75 .83

BL;

32

.65

r(LDM) .67 .84 .88 .59

Size measures available for each sample were: a WL and

.56

size measurel

b c

b b c A

WL. BL and THL;

b c a

c

WL, BL, THL and TT

194

GEOGRAPHIC VARIATION LEAN MASS OF DUNLIN

taken to examine whether these between-site differences were related to latitude and temperature.

Lean dry mass (uncorrected for body size variations) proved to be higher on more northern estuaries in immatures (r : 0.31, n : 74, p : 0.008) but not in adults (r : 0.04, n : 175, p : 0.057). Lean dry mass was also higher on colder estuaries, in both adults and immatures (Fig. 2). When body size (as measured by bill length) and midwinter air temperature were held constant, the correlation between lean dry mass and latitude disappeared (adults r : -0.05, P:0.269; immatures r:0.17, p: 0.083). However when body size (bill length) and latitude were held constant, the correlations between lean dry mass and temperature improved (adults r : -0.45, p < 0.001; immatures

r:-0.41,p