Winter fattening in the American goldfinch and the possible role …329512/fulltext.pdf · 1982. Michigan and California birds were captured alive using mist nets or traps. American

WINTER FATTENING IN THE AMERICAN GOLDFINCH AND THE POSSIBLE ROLE OF TEMPERATURE IN ITS REGULATION'

WILLIAM R. DAWSON AND RICHARD L. MARSH

Division of Biological Sciences and Museum of Zoology, University of Michigan, Ann Arbor, Michigan 48109; and Department of Biology, Northeastern University, Boston, Massachusetts 02115

(Accepted 11/18/85)

We investigated whether environmental temperature has any causal role in the winter fattening in certain finches of the subfamily Carduelinae. Correlational analyses between fat content of American goldfinches (Carduelis tristis) and various short- and long-term measures of temperature provide no evidence for a proximate role of this environmental variable in determining the degree of fattening of these birds in southeastern Michigan. Their fat content shows the best correlations (r = -.61 to -.63) with the long-term average minimum temperature or record low temperature for the date of capture. Furthermore, inclusion of long-term thermal measures in multivariate analyses excludes from significance temperature conditions surrounding the day of capture. Comparison of American goldfinches wintering in Michigan, California, and Texas, respectively, strengthens the conclusion that environmental temperature does not directly influence their fat content. Taken together, our data on this species favor the hypothesis that temperature is an ultimate, i.e., evolutionarily significant, rather than a proximate factor in winter fattening. Comparisons of American goldfinches, pine siskins (Carduelis pinus), and common redpolls (Carduelisflammea) in Michigan indicate that these similar-sized congeners show different levels of winter fattening under similar winter conditions. The differences in fat content among these species do not correlate in any simple way with their respective overall winter distributions.

INTRODUCTION

Many small northern birds increase de- pot fat during the winter (King 1972; Daw- son, Marsh, and Yacoe 1983b). The adaptive significance of this reserve clearly relates to the increased demands for thermogenesis in the cold and the uncertain supply of food during times of inclement winter weather (see King 1972). However, unlike our knowledge of premigratory fattening (King and Farner 1965; Berthold 1975), that concerning the factors that reg- ulate winter fattening is scanty. The nega-

'We are grateful to the Department of Biology, University of California, Riverside, for space and certain facilities used in work with American goldfinches from the Riverside area. We also wish to thank Dr. Charles D. Fisher, Stephen F. Austin University, Nac- ogdoches, Texas, for obtaining our Texan sample of American goldfinches. Dr. Dennis G. Baker generously provided access to his computer files of temperature data for Ann Arbor, Michigan. Dr. William A. But- temer, Dr. Marshall E. Yacoe, and Mr. James Camil- liere provided us with technical assistance. Our research was supported in part by grants DEB 78-25487, BSR 80-21389, and BSR 84-07952 from the National Sci- ence Foundation to W.R.D.

Physiol. Zool. 59(3):357-368. 1986. C 1986 by The University of Chicago. All

rights reserved. 0031-935X/86/5903-8568$02.00

tive correlation of fat content with ambient temperature over the cooler months of the year has fostered the obvious suggestion that temperature is a proximate-i.e., causal-as well as an ultimate-i.e., evolutionarily significant-factor in the winter fattening cycles of birds (e.g., Baldwin and Kendeigh 1938; Helms and Drury 1960; Newton 1966; O'Connor 1973). However, results of certain investigations of wild populations have not confirmed this (Evans 1969; King and Mewaldt 1981). Alternative hypotheses have been difficult to test ex- perimentally, for captive birds, even in out- door flight cages, often show abnormal winter fattening (King and Farner 1966; West 1972; Pohl and West 1973; White and West 1977; W. R. Dawson, C. Carey, and R. L. Marsh, unpublished observations).

In this study we attempt, through analysis of data on body composition and per- tinent thermal variables for wild populations of American goldfinches (Carduelis tristis), to evaluate hypotheses relating to regulation of winter fattening. We sampled birds of this species near Ann Arbor, Mich- igan, over three winters. Our timing was propitious, for our study coincided with a period of unusually large year-to-year vari-

357

358 W. R. DAWSON AND R. L. MARSH

ation in winter weather (Karl, Livezy, and Epstein 1984). We have supplemented data on Michigan birds with information on American goldfinches wintering in warmer areas and on two other cardueline finches- the common redpoll (Carduelis flammea) and the pine siskin (Carduelis pinus)-that visit southeastern Michigan sporadically in winter.

MATERIAL AND METHODS

American goldfinches were collected at three localities: (1) near Ann Arbor, Wash- tenaw County, Michigan (420 17' N, 83046'W), during 1977-1980; (2) near Riverside, Riverside County, California (33o15'N, 117031'W), in 1978-1979; and (3) near Nacogdoches, Nacogdoches County, Texas (31037'N, 94038'W) in 1982. Michigan and California birds were captured alive using mist nets or traps. American goldfinches were obtained in Texas by shooting them with fine dust loads. We did not identify any of these individuals to subspecies. However, extensive distributional information indicates that the Michigan and Texas birds should be as- signable to Carduelis tristis tristis and the southern California birds to C. t. salica- mans (Howell, Paynter, and Rand 1968). Pine siskins and common redpolls were trapped near Ann Arbor, Michigan, during the winters of 1978-1979 and 1980-1981, respectively. We restricted all our analyses to birds collected after 1100 hours to avoid complications of any diurnal cycle of body constituents. Regression analysis on Amer- ican goldfinches in Michigan during the peak of the winter fattening (Dec.-Feb.) in- dicated no effect of the time of capture on body constituents of birds collected after 1100 hours.

Most birds captured alive were immediately killed humanely and then transferred in sealed plastic bags to the laboratory. However, common redpolls were first brought to the laboratory and then killed within 2 h of capture. Fresh body mass of the birds was determined to the nearest milligram. Contents of the crop and proventriculus were removed and weighed with similar precision. Carcasses were stored frozen in plastic bags until analysis of body composition was performed. Carcasses and crop contents were analyzed using Carey et

al.'s (1978) procedures, except that the contour feathers were not removed and the lean dry carcasses were not analyzed for protein. The following body components were determined: crop contents = fresh mass of seed in the crop and proventriculus; dry body mass = mass remaining after the carcass (less crop contents) had been com- pletely dried by lyophilization; lean dry body mass = mass remaining after redry- ing of the lyophilized carcass following extraction with petroleum ether; fat content = difference between dry and lean dry body masses (petroleum ether extracts neutral lipids, predominantly triacylglycerols, and we shall take the change in body mass pro- duced by this extraction as representing fat content); water content = the difference between fresh body mass (less crop contents) and dry body mass.

Temperature variables for Ann Arbor, Michigan, during the study period were obtained from computer files maintained at the University of Michigan. The long-term average temperatures and record temperatures for Ann Arbor have been published in DeWitt and Baker (1980). Thermal data for Riverside, California, and Nacogdoches, Texas, were obtained from National Cli- matic Center summaries (U.S. Environ- mental Data Service 1978, 1979, 1982). Additional information on long-term trends was obtained from a decennial cen- sus of U.S. climate (U.S. Weather Bureau 1965).

Statistical analysis was performed with the aid of MIDAS, a statistical package of the statistical research laboratory at the University of Michigan (Fox and Guire 1976). Multiple comparisons between means were accomplished with one-way analysis of variance (ANOVA), utilizing Scheff6's technique to compare individual means (Sokal and Rohlf 1981). In our multivariate analyses of thermal variables, we used "backward" and "forward" variable- selection schemes (Sokal and Rohlf 1981).

RESULTS

AMERICAN GOLDFINCHES

Body composition of Michigan birds.- Our analysis of body composition of American goldfinches in Michigan covers the period of November through early May

WINTER FATTENING IN AMERICAN GOLDFINCHES 359

to avoid complications involving the breeding season and recruitment ofjuvenal birds into the population in late summer and early fall. As established previously (Carey et al. 1978), significant variation occurs seasonally in all the major body constituents (table 1). The means for lean dry body mass vary from 3.9 g in May to 4.3 g in February, with the difference between these extremes being statistically significant (P = .011). Fat content of Michigan goldfinches increases in late fall, reaching a maximum mean value of 1.7 g or approximately 29% of the dry body mass. This level persists from December through Jan- uary and February. Fat content decreases significantly over the period of March to early May (ANOVA, P = .0001). In early May the mean fat content, 0.6 g, corre- sponds to 14% of the dry body mass.

Body composition ofsouthern California birds.-Mean lean dry body mass of American goldfinches from the vicinity of

Riverside, California, only ranges from 3.9 to 4.0 g in the months in which sampling was conducted (ANOVA, P = .83). Fat content also varies little, remaining between 0.6 and 0.9 g (table 1; ANOVA, P = .08). These values represent only 14%-18% of the corresponding dry body masses. Gold- finches of this more southerly population thus do not exhibit significant winter fattening.

Body composition of Texas birds in early February.-We arranged to have American goldfinches collected in the vicinity ofNac- ogdoches, Texas, to determine whether members of the same taxon to which the Michigan birds are assigned (Carduelis tristis tristis) show fattening while wintering at a latitude similar to that at which we collected our California birds. Lean body mass and fat content for the Texas goldfinches in early February average 4.2 and 1.5 g, respectively (table 1). The latter figure cor- responds to 26% of dry body mass.

TABLE 1

BODY COMPOSITION OF CARDUELINE FINCHES

BODY COMPOSITION PARAMETER

FINCH TYPE, Lean Dry LOCATION, AND m' W Body Mass F

MONTH (N) (g) (g) (g) (g) W/(m' - F) F/(m' - W)

American goldfinches: Michigan:

Nov. (19) ..... 14.33 + .169 8.96 + .139 4.14 + .051 1.23 + .099 .684 .229 Dec. (25) ...... 14.54 +. 133 8.73 + .069 4.11 + .046 1.70 + .091 .680 .293 Jan. (52) ...... 14.61 + .115 8.73 + .073 4.19

_ .039 1.69 + .056 .676 .288

Feb. (22) ...... 14.80 + .188 8.74 + .093 4.34 + .058 1.72 + .118 .668 .284 Mar. (22) ..... 14.01 + .226 8.49 + .207 4.21 + .085 1.31 + .054 .669 .237 Apr. (22) ...... 13.90 + .147 8.72 + .092 4.18 + .041 1.00 + .057 .676 .193 May.(5) ...... 12.48 + .290 7.98 + .122 3.87 + .101 .63 + .123 .673 .140

California: Nov.(3) ...... 12.55 + .412 7.91 + .214 3.98 + .011 .66 + .183 .665 .142 Dec. (6) ....... 12.31 + .239 7.78 + .134 3.90 + .099 .63 + .050 .666 .138 Jan. (11) ...... 12.54 + .157 7.79 + .105 4.02 + .121 .73 + .053 .660 .154 Feb. (8) ....... 12.47 + .241 7.82 + .168 3.94 + .090 .71 .079 .665 .153 Mar. (11) ..... 12.76 + .184 8.01 .117 3.87 + .058 .87 + .072 .674 .184

Texas: Feb. (13) ...... 14.05 + .272 8.33 + .128 4.23 + .077 1.49 + .116 .663 .260

Pine siskins; Michigan:

Dec. (22) ...... 15.31 + .193 8.86 + .094 4.03 + .044 2.42 + .127 .687 .375 Jan. (6) ....... 15.72 + .414 8.91 + .227 4.22 + .187 2.60 + .267 .679 .382

Common redpolls: Michigan:

Jan.-Feb. (7) ... 13.70 + .26 7.98 + .139 4.13 + .070 1.59 + .107 .659 .278

NOTE.--m' = body mass less crop contents; W = water content; and F = fat content.


PINE SISKINS AND COMMON REDPOLLS

Several cardueline finches undergo spo- radic eruptive movements apparently linked with reduced availability of food in their normal wintering areas (Bock and Lepthien 1976). Such movements brought considerable numbers of pine siskins and redpolls to the Ann Arbor area during the winters of 1978-1979 and 1980-1981, respectively. Our sampling of these birds established that lean body dry mass averages ~4. 1 g in each species (table 1). However, their mean fat contents differ, being 1.6 g in the redpolls and - 2.5 g in the pine siskins. These values correspond to 28% and 38% of dry body mass, respectively.

'o1' 20 o 10-

AVETMMN I

0~ w A-20

MONTH

FIG. 1.-Seasonal cycles of fat content in American goldfinches during fall, winter, and spring (upper panel) at Ann Arbor, Mich. The thin vertical lines, which are plotted at the respective midpoints of the 2-wk periods to which they pertain, mark the range of values for the birds sampled. The short horizontal lines and the shaded rectangles connected to these lines indicate the means for the various 2-wk samples and the interval represented by the mean + SEM, respectively. The corresponding daily average minimum temperatures (A VE MIN) as well as the record low temperatures for the various dates (EXT MIN) are plotted in the lower panel (based on data for the 99-year interval prior to 1980). The line tracing the record low temperatures is rather erratic owing to pronounced fluctuations in temperature from one date to the next. Its excursions are enclosed in the hatched band to simplify plotting.

20 OCT NOV DEC JAN FEB MAR APR MAY

MONTH

FIG. 2.--Minimum temperatures at Ann Arbor, Mich., for the October-May period of 1977-1978 and 1979-1980. The lines for these years trace the mean temperatures for successive 2-day intervals. (This procedure simplifies plotting but somewhat damps the os- cillations in the curves.) The curve for average daily minimum temperature is included for comparison.

DISCUSSION

WINTER FATTENING IN THE AMERICAN GOLDFINCH

Thermal correlates offattening in southeastern Michigan.-The pattern of fattening evident over the cooler months of the year in free-living American goldfinches in southeastern Michigan suggests an inverse relation with temperature (fig. 1). This relation is also apparent in previous studies of this species, as well as in studies of other passerine birds (see Dawson et al. 1983b). However, the overall inverse correlation between fat content and environmental temperature does not assist appreciably in discerning whether temperature is only an ultimate factor, which has fostered natural selection for increased fattening in winter, or a proximate one, which serves as a stim- ulus for this form of lipid deposition. We have pursued this question by examining correlations of fat content in American goldfinches in Michigan with local thermal variables. Temperature in southeastern Michigan undergoes substantial variation both within and between years. The curve for minimum temperatures for the 99 years preceding 1980 is symmetrical, reaching its lowest value on January 24-26 (fig. 1) (DeWitt and Baker 1980). An essentially parallel curve describes average daily maximum temperatures. In contrast to these smooth curves, the daily temperatures in any given winter fluctuate greatly (fig. 2).


TABLE 2

SIMPLE CORRELATION COEFFICIENTS BETWEEN THE FAT CONTENT OF GOLDFINCHES AND VARIOUS

TEMPERATURE VARIABLES FROM NOVEMBER THROUGH MAY

Variable r Variable r Variable r

MIN................ -.399 MEANMAXLAG4 ..... -.503 MEANMAXLED2 ..... -.491 MAX............... -.421 MEANMAXLAG8..... -.551 MEANMAXLED4 ..... -.526 MINLAGI........... -.355 MEANMAXLAG16 .... -.586 MEANMAXLED8..... -.552 MAXLAG1 .......... -.449 MINLEDI ............ -.328 MEANMAXLED16 .... -.599 MEANMINLAG2 ..... -.396 MAXLED1........... -.406 MEANMINLAG4 ..... -.473 MEANMINLED2 ...... -.429 AVEMIN ............ -.612 MEANMINLAG8 ..... -.520 MEANMINLED4 ...... -.461 AVEMAX ........... -.609 MEANMINLAG16 .... -.554 MEANMINLED8...... -.515 EXTMIN ............ -.621

MEANMAXLAG2.... -.442 MEANMINLEDI6 -.599 MEANEXTMIN ...... -.627

NOTE.-Sample size = 160. In all cases the correlation coefficient (r) is significant, with P < .0001. The temperature variables analyzed and their abbreviations (given in parentheses) are as follows: the minimum and maximum temperatures on the day of capture (MIN and MAX, respectively); the minimum and maximum temperatures on the day before capture (MINLAG I and MAXLAG I, respectively); the mean of the minimum and maximum temperatures for 2, 16 days before capture (MEANMINLAG2, ... MEANMINLAG16 and MEANMAXLAG2, ... MEANMAXLAG16, respectively); the minimum and maximum temperatures on the day after capture (MINLEDI and MAXLEDI), respectively; the mean of the minimum and maximum temperatures for 2, 16 days following capture (MEANMINLED 2, ... MEANMINLEDI6 and MEANMAX- LED 2, MEANMAXLED 16, respectively). The average minimum and maximum temperatures of the capture date for 99 yr previous to 1980 (AVEMIN and AVEMAX, respectively); the record low temperatures for the capture date during the 99 yr previous to 1980 (EXTMIN); and a 7-day-centered moving average of the record low temperatures (MEANEXTMIN).

Minimum temperatures above or below long-term averages for the dates may persist for as much as a month in any particular year, and temperatures differing by 10-15 C from the corresponding average minima may occur routinely for 4-5 days over the months studied.

Were temperature acting as a proximate cue for winter fattening in the American goldfinch, large-scale thermal variation should foster corresponding variation in the fat content of these birds. We might thus expect fat content to be inversely correlated with temperature conditions on or immediately preceding the day of collection. To test this we have undertaken simple and partial correlation analyses involving the following thermal variables (table 2): the minimum and maximum temperatures on the day of collection of the particular bird (MIN and MAX, respectively), the minimum and maximum temperatures on the day before capture (MINLAGI and MAXLAG1, respectively), and the mean minimum and mean maximum temperatures for periods extending back from capture date to 2, 4, 8, or 16 days earlier (MEANMINLAGn and MEANMAX- LAGn, respectively).

We also have considered the hypothesis

that American goldfinches store fat in an- ticipation of temperature changes or inclement conditions associated with the passage of winter weather fronts. To test this, we assessed the correlations of fat content with the minimum and maximum temperatures on the day following capture (MINLED1 and MAXLEDI, respectively), and the mean minimum and mean maximum temperatures for periods of 2, 4, 8, or 16 days beyond the day of capture (MEAN- MINLEDn and MEANMAXLEDn, respectively).

Additionally, we have utilized several models for evaluating temperature as an ultimate factor (table 2). To determine correlations between fat content and average conditions, we used the long-term (99-yr) daily average minimum and maximum temperatures (AVEMIN and AVEMAX, respectively). We also examined correlations between fat content and the 99-year record daily low temperatures (EXTMIN) and between this body component and a moving 7-day average of the record low temperatures, centered on the day of capture (MEANEXTMIN). The moving average was used to smooth the data on record low temperatures, which may vary 5-8 C on successive days (fig. 1).


In a simple correlational analysis of fat content and the variables representing the actual temperatures on the days surrounding the date of capture, the correlation coefficient (r) increases directly with the number of days used in calculating the mean temperatures (table 2). The strongest correlations in this analysis are found with the mean temperatures for the 16 days either preceding or following the day of capture (r = -.55 and -.59 for MEANMINLAG16 and MEANMAXLAG 16, respectively, and -.60 for both MEANMINLEDI6 and MEANMAXLED 16). These means tend to damp day-to-day fluctuations in temperature, yielding values approaching long-term averages for the particular periods. These simple correlational analyses suggest that the American goldfinches neither react to nor anticipate current winter thermal conditions but actually respond in a manner correlated with long-term temperature patterns. Consistent with this suggestion, the strongest correlations of fat content in our study (r = -.61 to -.63) are found with measures based on temperature records for the 99-yr period through 1980. These include AVEMIN, AVEMAX, EXTMIN, and MEANEXTMIN (table 2). All of the correlations in table 2 are driven by the au- tumnal increase and vernal decrease in fat content. If the analysis is restricted to the period from after the first week in Decem- ber to the end of February, fat content is not significantly correlated with any of the thermal variables considered, despite substantial variation in temperature during this period both within and between years (fig. 2). These analyses argue against the hypothesis that fattening is influenced proxi- mately by temperature.

Multivariate techniques also bolster the conclusion that levels of fat stored by American goldfinches are not significantly correlated with the temperatures either on the day of capture or on the immediately preceding or following days. If AVEMAX, AVEMIN, EXTMIN, or MEANEXTMIN is included in the correlational analysis, the partial correlation coefficients for the thermal variables pertaining to actual collection dates for goldfinches are not significant (P > .05). The two variations of stepwise multiple regression (Sokal and Rohlf 1981) that we employed allow determination of which

variables in a multiple-regression model particularly assist in explaining variance in fat content. If MEANEXTMIN is included in the analysis with any combination of the other temperature variables, all of the others are excluded from the model (i.e., if included, they have insignificant partial correlations with fat content). If MEAN- EXTMIN is removed from the analysis, EXTMIN is selected and all other variables are excluded. Likewise, if EXTMIN is removed, AVEMIN is selected and other variables are excluded. The strong correlation of fat content with EXTMIN is due to the shape of the curve describing these temperatures. As with mean fat content, the curve for record low temperatures is relatively flat from late December until late February. At any time during the period, Canadian air masses may enter Michigan and birds may encounter temperatures from -20 to -30 C. This seasonal pattern of record low temperatures may have ex- erted a broadly distributed selective influence on the fattening cycles of American goldfinches in eastern North America, for it appears common to areas in that region that are widely separated geographically (McCalla, Day, and Millward 1978).

The strongest correlation that we obtained (for fat content and MEANEXT- MIN; table 2) explains only 38% of the variance in fat content. This seems much lower than the correlations reported in previous studies of winter fattening in small passerines (King and Farner 1966; Evans 1969). However, the respective mean fat contents for the birds collected on particular dates (e.g., 10-20 individuals for each date in Evans 1969) were employed in these earlier correlation analyses, rather than our procedure of using individual values. If we examine the relation of the mean fat contents of goldfinches for 2-wk intervals (fig. l) to either AVEMIN or MEANEXTMIN temperatures, our correlation coefficients reach ~ -.93, thus accounting for nearly 87% of the variance.

Geographic variation in body composition of American goldfinches in winter.- Fattening is a prominent component of winter acclimatization of the American goldfinch in Michigan (Carey et al. 1978). This accumulation of additional energy substrate entails costs relating to extra feed-


ing beyond that required for immediate maintenance, synthesis of triacylglycerols, and transport of added weight in flight. These considerations led us to hypothesize that winter fattening would be either less prominent or absent in populations of goldfinches wintering in southern areas, where temperatures are milder, growing seasons longer, and risks of coverage of food by snow or ice lower than in the northern United States and southern Canada. Our results for American goldfinches from southern California are consistent with this hypothesis, for these birds show insignificant winter fattening (table 1; ANOVA, P = .08). As noted previously, American goldfinches in that area are assigned to a different subspecies than are those in the eastern United States and eastern Canada (Howell et al. 1968). This led us to extend our comparison to southern birds assigned to the same taxon as our Michigan birds, by utilizing individuals wintering in east Texas. In early February, the fat content of these latter birds resembles that of Michigan birds at the same time of year (table 3; ANOVA with Scheff6's multiple comparisons, P = .77). The subsample of Michigan birds in table 3 was chosen to bracket the date of capture of the Texas birds and to be similar in size, but no statistically significant difference in fat content exists between the Texas birds and the entire sample of Michigan birds monitored in January and February (table 1; ANOVA with Scheffe's multiple comparisons, P = .11). On the other hand, the fat content of the American goldfinches in Texas during early February averages approximately twice that of their counterparts in California (table 3; ANOVA with Scheff6's multiple comparisons, P < .0001), despite the similarity (1) in temperature conditions both on and immediately before the capture dates and (2) in latitudes for the localities where the two samples were obtained.

INTERSPECIFIC COMPARISONS OF BODY COMPOSITION

IN CARDUELINE FINCHES

Comparison of the extent of winter fattening in the American goldfinch, pine siskin, and redpoll in southeastern Michigan is facilitated by the similarity in their lean dry body masses (table 1). Surprisingly, the fat content of the most boreal of these birds,

the common redpoll, does not differ significantly from comparable values for the American goldfinch (ANOVA, P = .59). Moreover, the mean value that we obtained for the redpoll, 1.6 g, is bracketed by those of 1.7 and 1.2 g calculated from lipid indices presented for redpolls obtained near Fair- banks, Alaska, during January and Febru- ary, respectively (White and West 1977). This comparison is complicated by the fact that the Fairbanks samples include common redpolls, hoary redpolls (Carduelis hornemanni), and intergrades (White and West 1977), but the fact that the January- February fat content of the Alaskan birds is not appreciably greater than those of common redpolls or American goldfinches in southeastern Michigan is striking. Per- haps this uniformity is somehow linked with the fact that some exchange of common redpolls occurs between Alaska and the Great Lakes area (Troy 1983). It may also indicate that the requirement of Alas- kan redpolls for greater fat storage in ad- vance of long, cold winter nights is alle- viated by their capacity for seed storage in the buccal pouches and crop (White and West 1977). Equally striking is the fact that the mean fat content of the pine siskins that we obtained in the vicinity of Ann Arbor during winter is significantly higher than the comparable values for common redpolls and American goldfinches captured at the same locality (table 1; ANOVA with Scheff6's multiple comparisons, P < .0001). Were our comparison confined to the pine siskin and American goldfinch, it would have been tempting to conclude that the greater fat content of the siskins was somehow associated with the fact that their dis- tribution extends into more boreal regions than does that of the American goldfinches. However, the data on redpolls suggest a more complex situation. Information on the annual cycle of body composition and on the degree of variation of fat content among populations of pine siskins is needed before we can adequately interpret the information thus far obtained.

REGULATION OF WINTER FATTENING IN

CARDUELINES AND OTHER PASSERINES

Three general patterns of geographic variation in winter fattening seem to exist within species of small birds. (However,

TABLE

3

BODY

COMPOSITION

OF AMERICAN

GOLDFINCHES

IN WINTER

TEMPERATURE

BODY

COMPOSITION

PARAMETER

(C)

Lean Dry Body

m

Cr

W

Mass

F

Jan.

Mean

STATE

(N)a

(g)

(g)

(g)

(g)

(g)

W/(m

- Cr -

F)

F/(m

- Cr -

W)

Minimumb

Minimum

MI (13)

15.08

.25

.328

+ .046

8.87

.15

4.35

.07

1.53

+ .12

.671

.260

-11.7

-8.6

TX

(13)

14.07

+ .27

.030

+ .010

8.33

+ .13

4.23

+ .08

1.49

+ .12

.663

.260

4.7

3.2

CA

(10)

12.51

+ .24

0

7.84

+ .14

3.97

+ .15

.71 + .07

.663

.152

3.7

3.2

NOTE.-rm

= total

body

mass;

Cr = crop

contents;

MI

= Michigan;

TX

= Texas;

and

CA

= California.

Other

abbreviations

are

given

in table

1.

a Dates

of collection

for the three

localities

were:

MI,

Jan.

30-Feb.

15; TX,

Feb.

2; and

CA,

Jan.

19-Feb.

13.

b Mean

of the minimum

temperatures

for 4 days

before

the capture

of each

bird.

SObtained

from

records

of the U.S.

Weather

Bureau

for the collection

locality.


before summarizing these, we should note the possibility that a portion of the apparent diversity may be a product of our attempting in some cases to infer trends in winter fattening by use of information on body mass.) In the case of house sparrows (Passer domesticus) in North America, a direct correlation exists between fat content of most winter individuals and latitude (Blem 1973), latitude providing a rough index of severity of winter climate. Only house sparrows from Churchill, Manitoba, the most northern population sampled, depart significantly from this relationship. Extent of winter fattening in house finches (Car- podacus mexicanus) also may correlate with winter climate, although free-living individuals from only two localities have been sampled thus far. House finches from Riv- erside, California, do not show winter fattening, whereas those from Boulder, Col- orado, deposit small but significant amounts of additional fat at this season (Dawson et al. 1983a). Gambel's white- crowned sparrows (Zonotrichia leucophrys gambelii) manifest quite a different pattern. On the basis of measurements of body mass of winter birds dispersed over a wide north- south span, extent of winter (and vernal premigratory) fattening appears to be in- dependent of latitude (King and Mewaldt 1981), in contrast to the prediction of King and Farner (1966). Data on fat content of redpolls (White and West 1977; table 1 of this study) suggest a similar pattern. Ob- servations on American goldfinches estab- lish a complex picture of geographic variation in winter fattening. These birds deposit substantial fat during the colder months of the year in the north central United States, New Jersey, and southern Canada (Wiseman 1975; Carey et al. 1978; Prescott 1983; A. L. A. Middleton, unpublished data; table 1 of this study), but those from southern localities may or may not undertake this form of energy storage (ta- bles 1, 3 of this study). Clearly, substantial differences in geographic patterns of energy storage during winter do exist among birds. Perhaps these relate to the vagility of the various species; those we have discussed range from relatively sedentary (house sparrow, house finch) to irregularly migratory or nomadic (redpoll, American goldfinch), to regularly migratory (Gambel's

white-crowned sparrow). Genetic differences among conspecific populations also may be involved in some cases. Such a conclusion is bolstered by the tenacity with which Alaskan common redpolls hold their normal annual schedule of fattening when transferred to an environment in which the native European representatives of this species follow quite a different schedule (Pohl and West 1976). Perhaps the absence of winter fattening in American goldfinches in southern California (table 1) is similarly a manifestation of genetic differences from the eastern representatives of the species. Banding returns indicate that the Riverside population is largely sedentary (W. A. But- temer and W. R. Dawson, unpublished observations). Perhaps natural selection has favored the birds' avoiding winter fattening owing to the advantages in an equable environment of avoiding the costs of deposition and transport of the added mass.

Our study was undertaken in hopes of gaining a general understanding of the role of environmental temperature in regulation of winter fattening in passerine birds. This is a complex task, for although the temperatures incorporated into meteorological records constitute a useful index of the severity of winter climates, other variables are involved. Moreover, interspecific differences in foraging and roosting habits may lead to variation at the same locality in the extent of the problems posed by convective and radiational heat loss. Nevertheless, there still is utility in attempting to define the role of temperature.

Correlation of winter fattening with environmental temperature has fostered the suggestion that temperature acts as a proximate cue for this form of lipid deposition. On the basis of data on captive house sparrows, Kendeigh et al. (1969) suggested that winter fattening resulted from increased overnight loss of body mass in the cold, followed by increased food intake and fat synthesis the next day. Such a response might be explained by an increase in the activity of fat-synthesizing enzymes, like that ob- served after the termination of fasting in chicks of the domestic fowl (Goodridge 1968; Silpanata and Goodridge 1971; Fischer and Goodridge 1978). However, prolonged exposure to reduced environmental temperatures does not cause in-

366 w. R. DAWSON AND R. L. MARSH

creased activity of these enzymes in such chicks (Belnave 1972b). All of the data demonstrating a response to short-term thermal conditions involve measurements of body mass rather than direct determination of fat content (Baldwin and Ken- deigh 1938; Odum 1949; Helms and Drury 1960; Newton 1966; Kendeigh et al. 1969; O'Connor 1973). Interpretation of this information is hindered by the possibility of significant variation in other body constituents (King and Farner 1966; Carey et al. 1978; Dawson et al. 1983b; table 3 of this study) and in the amount of food present in the digestive tract (Evans 1969; table 3 of this study).

A far different conclusion concerning the relation of winter fattening to short-term or year-to-year fluctuations in temperature emerges if the analysis is confined to deter- minations of body composition rather than of body mass. Analysis of fat reserves of yellow buntings (Emberiza citrinella) and American goldfinches suggests that the extent of these reserves at the end of winter days is insensitive to short-term and, in the case of the goldfinches, year-to-year variations in environmental temperature (Evans 1969; table 2 of this study). This is not to deny that a significant negative correlation between fat content and environmental temperature exists over the cooler months of the year. However, this correlation be- comes substantially stronger when long- term statistics on temperature rather than actual temperatures on or near the capture date are used (Evans 1969; table 2 of this study). Perhaps EXTMIN has been an im- portant selective factor affecting the regulation of evening fat reserves during winter, for natural selection should favor survival of birds carrying sufficient fat reserves to sustain adequate thermogenesis through the most severe night possible on a given date. In this view temperature would have to be classed as an ultimate rather than a proximate factor in winter fattening, the direct cue being some other variable such as pho- toperiod (with an appropriate lag in response by the bird). However, the possibility exists that temperature acts as a proximate cue for this form of energy storage but that the response is slow enough so that adjustments to short-term temperature fluctuations are damped, producing refractoriness

to daily temperature fluctuations. In the case of the American goldfinch, at least, we are inclined to discount this explanation on the basis of the similarity of midwinter fat contents of eastern American goldfinches from Texas and Michigan (table 3). This expression of the fattening response in the goldfinches wintering in a relatively mild climate merits some further comment. Banding returns from the U.S. Fish and Wildlife Service suggest that birds that winter in the southern United States do not represent a discrete subpopulation of eastern American goldfinches. Such birds are not infrequently found in northern areas in other winters (W. A. Buttemer and W. R. Dawson, unpublished observations). Con- versely, banding returns also indicate that some of the American goldfinches captured at the latitude of southern lower Michigan (~42O N) in winter have been found in preceding or subsequent winters at latitudes of 33o-47o N. With this variability in both direction and extent of movements, it is conceivable that natural selection acts to maintain winter fattening in individuals that must contend with cold climates in at least some years, any disadvantage associated with carrying this fat in more southerly wintering locales being relatively slight. An alternative interpretation of the expression of winter fattening in the Texas birds arises from the hypothesis that selection acts most intensively under extreme conditions. The same northern air masses that contrib- ute to the variability of weather conditions in Michigan also sporadically penetrate into the southern United States. The impact on places such as Nacogdoches, where our Texas birds were obtained, is illustrated by comparing the long-term minimum temperatures (MEANMIN) and EXTMIN for that locality and for Riverside, where our California birds were captured (data from U.S. Weather Service summaries of data collected before 1960). Riverside, Califor- nia, the site where we found no winter fattening, has January MEANMIN and EXTMIN of 3.2 and -6.0 C, respectively. The January MEANMIN for Nacogdoches, Texas, is also 3.2 C. However, the EXTMIN for this month is -20 C (U.S. Weather Bu- reau 1965)! Natural selection for winter fattening in eastern American goldfinches


thus would be advantageous even to those individuals wintering in the south.

The question remains of how to reconcile the conflicting interpretations of the thermal dependence of winter fattening that arise from analyses of data on body mass and body composition. One possible explanation involves birds' adjustments in the rate of food intake and processing under different thermal regimes. To maintain a constant level of fattening in the face of varying energy demands, birds must mod- ulate food intake appropriately. If feeding

rate increases with cold stress, diurnal body mass will increase owing to an increased amount of food in the digestive tract. Chronically elevated food intake may also increase the size of the liver (Balnave 1972a). Finally, some birds, including cardueline finches, may buffer nighttime use of fat reserves by adjusting the extent to which they fill their crops before going to roost. Our data on crop contents of Amer- ican goldfinches obtained in Michigan and Texas in midwinter (table 3) are consistent with such a suggestion.

LITERATURE CITED

BALDWIN, S. P., and S. C. KENDEIGH. 1938. Variations in the weight of birds. Auk 55:416-467.

BALNAVE, D. 1972a. The influence of environmental temperature on the metabolism and liver composition of the immature male chick. Comp. Biochem. Physiol. 42A:425-435.

1972b. The effect of temperature and length of exposure on liver composition and hepatic li- pogenic enzyme activity in the immature male chick (Gallus domesticus). Comp. Biochem. Phy- siol. 43B:999-1007.

BERTHOLD, P. 1975. Migration: control and metabolic physiology. Pages 77-128 in D. S. FARNER and J. R. KING, eds. Avian biology. Vol. 4. Academic Press, New York.

BLEM, C. R. 1973. Geographic variation in the bio- energetics of the house sparrow. Ornithological Monogr. 14:96-121.

BOCK, C. E., and L. W. LEPTHIEN. 1976. Synchronous eruptions of boreal seed-eating birds. Am. Nat. 110: 559-571.

CAREY, C., W. R. DAWSON, L. C. MAXWELL, and J. A. FAULKNER. 1978. Seasonal acclimatization to temperature in cardueline finches. II. Changes in body composition and mass in relation to season and acute cold stress. J. Comp. Physiol. [B] 125: 101-113.

DAWSON, W. R., R. L. MARSH, W. A. BUTTEMER, and C. CAREY. 1983a. Seasonal and geographic variation of cold resistance in house finches Car- podacus mexicanus. Physiol. Zool. 56:353-369.

DAWSON, W. R., R. L. MARSH, and M. E. YACOE. 1983b. Metabolic adjustments of small passerine birds for migration and cold. Am. J. Physiol. 245: R755-R767.

DEWITT, B. H., and D. G. BAKER. 1980. Ann Arbor weather calendar. DeWitt & Associates, Ann Arbor, Mich.

EVANS, P. R. 1969. Winter fat deposition and overnight survival of yellow buntings (Emberiza citrinella). J. Anim. Ecology 38:415-423.

FISCHER, P. W. F., and A. G. GOODRIDGE. 1978. Co- ordinate regulation of acetyl coenzyme A carbox- ylase and fatty acid synthetase in liver cells of the developing chick in vivo and in culture. Arch. Biochem. Biophys. 190:332-344.

Fox, F. J., and K. E. GUIRE. 1976. Documentation for MIDAS. 3d ed. Statistical Research Laboratory, University of Michigan, Ann Arbor.

GOODRIDGE, A. G. 1968. The effect of starvation and starvation followed by feeding on enzyme activity and the metabolism of [U-14C]glucose in liver from growing chicks. Biochem. J. 108:667-673.

HELMS, C. W., and W. H. DRURY, JR. 1960. Winter and migratory weight and fat: field studies on some North American buntings. Bird-Banding 31:1-40.

HOWELL, T. R., R. A. PAYNTER, JR., and A. L. RAND. 1968. Subfamily Carduelinae. Pages 207-306 in R. A. PAYNTER, JR., ed. Check-list of birds of the world. Vol. 14. Museum of Comparative Zoology, Cambridge, Mass.

KARL, T. R., R. E. LIVEZY, and E. S. EPSTEIN. 1984. Recent unusual mean winter temperatures across the contiguous United States. Bull. Am. Meteo- rological Soc. 65:1302-1309.

KENDEIGH, S. C., J. E. KONTOGIANNIS, A. MAZAC, and R. R. ROTH. 1969. Environmental regulation of food intake by birds. Comp. Biochem. Physiol. 31:941-957.

KING, J. R. 1972. Adaptive periodic fat storage by birds. Pages 200-217 in K. H. Voous, ed. Proceedings of the 15th International Ornithological Congress. Brill, Leiden.

KING, J. R., and D. S. FARNER. 1965. Studies of fat deposition in migratory birds. Ann. N.Y. Acad. Sci. 131:322-440.

1966. The adaptive role of winter fattening in the white-crowned sparrow with comments on its regulation. Am. Nat. 1010:403-418.

KING, J. R., and L. R. MEWALDT. 1981. Variation of body weight in Gambel's white-crowned sparrows in winter and spring: latitudinal and photoperiodic correlates. Auk 98:752-764.

MCCALLA, R. J., E. E. D. DAY, and H. A. MILLWARD. 1978. The relative concept of warm and cold spells of temperature: methodology and application. Arch. Meteorologie Geophysik Bioclimatologie [B] 25:323-336.

NEWTON, I. 1966. Fluctuations in the weights of bull- finches. Br. Birds 59:89-100.

O'CONNOR, R. J. 1973. Patterns of weight change in the house sparrow Passer domesticus (L.). Pages 111-125 in S. C. KENDEIGH and J. PINOWSKI, eds. Productivity, population dynamics, and systematics ofgranivorous birds. Polish Scientific, Warsaw.

ODUM, E. P. 1949. Weight variations in wintering white-throated sparrows in relation to temperature and migration. Wilson Bull. 61:3-14.


POHL, H., and G. C. WEST. 1973. Daily and seasonal variation in metabolic response to cold during rest and forced exercise in the common redpoll. Comp. Biochem. Physiol. 45A:851-867.

1976. Latitudinal and population specific differences in time of daily and seasonal functions in redpolls (Acanthis flammea). Oecologia 25:211- 227.

PRESCOTT, K. W. 1983. Weight, fat, and wing mea- surement variations of adult American goldfinches in New Jersey. North Am. Bird Bander 8:149-152.

SILPANATA, P., and A. G. GOODRIDGE. 1971. Synthesis and degradation of malic enzyme in chick liver. J. Biol. Chem. 246:5754-5761.

SOKAL, R. P., and F. J. ROHLF. 1981. Biometry: the principles and practice of statistics in biological research. 2d ed. W. H. Freeman, San Francisco.

TROY, D. M. 1983. Recaptures ofredpolls: movements

of an irruptive species. J. Field Ornithology 54:140- 151.

U.S. ENVIRONMENTAL DATA SERVICE. 1978. Clima- tological data: California. Vol. 82.

1979. Climatological data: California. Vol. 83. 1982. Climatological data: Texas. Vol. 87.

U.S. WEATHER BUREAU. 1965. Climatic summary of the United States-supplement for 1951-60. Gov- ernment Printing Office, Washington, D.C.

WEST, G. C. 1972. The effect of acclimation and acclimatization on the resting metabolic rate of the common redpoll. Comp. Biochem. Physiol. 43A: 293-310.

WHITE, C. M., and G. C. WEST. 1977. The annual lipid cycle and feeding behavior of Alaskan redpolls. Oecologia 27:227-238.

WISEMAN, A. J. 1975. Changes in body weight of American goldfinches. Wilson Bull. 87:390-411.

Documents

Winter fattening in the American goldfinch and the possible role …329512/fulltext.pdf · 1982. Michigan and California birds were captured alive using mist nets or traps. American