FAT RESERVES OF AN OPPORTUNIST AND OF A SPECIALIST IN THE NEGEV DESERT

YEHOSHUA $I-IKEDY AND URIEL N. $AFRIEL Departmentof Zoology,The Hebrew University of Jerusalem,Jerusalem 91904, Israel, andthe Mitrani Centerfor DesertEcology, Jacob Blaustein Institute for DesertResearch, Sede Boqer 84990, Israel

ABSTRACT.--Weextracted fat from samplesof two sympatric speciesresident in the Negev Desertof Israel.The CrestedLark (Galeridacristata), which fed on a wide range of seed sizes,exhibited within-day but not seasonalfluctuations in fat content and carried on average12.9% fat (of lean dry mass).The Desert Lark (Ammomanesdeserti), which fed on a narrowerrange of seedsizes, exhibited smaller within-day fluctuationsin the amountof fat, and carried on average 8.9% fat. We proposethat although the two occupythe same environment and are of the same feeding guild, these differencesare adaptive. When food becomesperiodically scarce, feeding specializationsof the Desert Lark may often sufficeto overcomeshortage, but their absencein the generalistCrested Lark may make it more dependableon fat reserves.Received 9 July 1990,accepted 20 December1990.

SPECIESthat utilize few food types are ex- within the Negev Desert of Israel, the Crested pected to exploit each of them more efficiently Lark consumedseeds that rangefrom 0.3 to 53.5 than speciesthat utilize a wide range of re- mg in mass,whereas most of the seedscon- sources(Pianka 1981a).If high efficiencyis ex- sumedby the DesertLark were lighter than 1.0 pressedin the ability to exploit effectively a mg. The CrestedLark generally was lesstena- resource even when it becomes rare, then food ciousand, dependingon the time of year, moved specialistsshould be less susceptiblethan food between regions and habitats;but the Desert generaliststo a simultaneousreduction in the Lark was much more sedentary($hkedy 1990). level of most resources. Thus, with respectto both food resourcesand Fat storage,however, can be usedas a buffer habitat utilization where both speciesoverlap, againstfluctuations in food availability (King the Desert Lark is a specialistand behavesas if 1972),but the size of the fat reserveis expected its environment is rather stable. The Crested to be determinedby a trade-offbetween benefit Lark is a generalist, frequently changesfood (reducedprobability of starvation)and cost(in- types and habitats, and behaves as if this same creasedrisk of predation)(King and Farner1966; environment is rather variable. Specialization King 1972;Blem 1975, 1976;$tuebe and Ket- vs. generalization are regarded as adaptations terson 1982; Nolan and Ketterson 1983; Pien- to stable vs. unstable environment (Pianka kowskiet al. 1984;Lima 1986,1988; Rogers 1987). 1981b,Futuyma and Moreno 1988).Because both It can be predictedthat specialistsand species species occur in the same environment, their of stable environments should deposit smaller different adaptationsmay reflect differencesin fat reservesthan generalistsand speciesof vari- the "interpretation" of environmental stability. able environments. We explore whether this implied difference in We comparedfat storagein two speciesof "interpreted" environmental stability and in nonmigratorylarks (Alaudidae).The distribu- mode of resourceutilization by co-occurring tion of the (Galerida cristata) ex- speciesis reflected in differencesbetween them tends from southern Sweden in the north to in fat storagetactics. northern Kenya in the South (Hall and Moreau 1970). Most of this range is nondesert,hence METHODS desertmight be marginalhabitat for the Crested The study site was on Halouqim Ridge in the cen- Lark. The distribution of the Desert Lark (Am- tral Negev Desertof Israel near Sede-Boqer(30ø52'N, toomanesdesertl) is confinedto arid regions(Hall 34ø57'E; 475 m above mean sea level). The terrain and Moreau 1970). Both speciesare ground consistsof stonyhillsides, with sparseshrubs (mostly dwelling, and their diets compriseseedlings, Artemisiaherba alba, Hammada scoparia, and Zygophyl- insects,and a large proportionof seeds($hkedy lum dumosum)(Evenari et al. 1982). Mean annual rain- 1990). We found that in a region of sympatry fall is 91 mm. Rainfalloccurs only during winter (Sep-

556 The Auk 108: 556-561. July 1991 July1991] FatReserves in Larks 557

TABLE1. Mean (+SD) wet mass(WM), lean dry mass(LDM), fat content(F), and fat ratio (FR) of all . Level of significance:*** = P < 0.001.

Species n WM LDM F FR (%) Crested Lark 118 39.0 _+ 2.7 11.5 + 1.2 1.48 + 0.71 12.9 _+ 7.2 Desert Lark 103 22.0 _+ 2.5 6.6 -+ 0.8 0.59 _+ 0.22 8.9 _+ 3.2 Statistics 59.0'**a 38.9'**a 12.7'*'4 5.47'**b

values, value, Mann-Whitney U test.

tember to May). Mean monthly air temperaturesvary but in CrestedLarks it was lower in September from 9.8øCin January(the coldestmonth) to 25.5øC than in December (Table 2). On average, fat in July (the hottestmonth). During January,air tem- ratio in both specieswas lower in September peraturesmight drop to -IøC (Zangvil and Druian than in other seasons, but this trend was not 1983;data from meteorologicalstation located 2 km significant(Table 2). There were no significant from the study site). seasonal differences in fat ratio. We obtained our data from 118 Crested Larks and 103 DesertLarks collected from February 1987to July Fat ratio calculated over the entire study pe- 1989.Samples of both specieswere securedfour times riod increasedduring the day in CrestedLarks. a year (hereafter seasons):before the breeding season The highest fat ratio was between 1,500 and (February through March, hereafter February),just 1,700,and the ratio later decreased(Fig. 1). The after the breedingseason (July), when CrestedLarks mean increaseof fat ratio over the entire study return to the study site (September),and December. period was 0.5% per hour. In the pooled data Eachshooting expedition lasted approx. 2 weeks.Right of birds secured in December of 1987 and 1988, aftercollection, we weighedeach (wet mass,WM) fat ratio increaseduring the day was higher to the nearest0.5 g (50-gPesola spring balance), mea- than in all other seasonscombined, and higher suredits wing length (WL) to the nearest1 mm (from than that of the Desert Lark (Fig. 1, Table 3). carpaljoint to wing tip; Newton 1969) and sexedit The correlation between fat ratio and time of in the field by dissection.Because the gonadsare small during nonbreedingseasons and there was damage day waslow and not significantin Februaryand causedby the shotgunpellets, we sexedonly 56 Crest- July (Table 3). ed and 49 Desert larks. The birds were deep-frozen The daily increasein fat ratio of Desert Larks 1-4 h after death. We used Moreau and Dolp's (1970) wasmuch slower and not significant(only 0.03% methodsfor determiningdry mass(DM), the amount per hour; Fig. 1, Table 3). of fat (F, extracted with petroleum ether), and lean In February, lean dry mass in both species dry mass(LDM). We then calculatedthe fat ratio by increasedsignificantly during the day (r• = 0.37, calibratingextracted fat with the bird'slean dry mass n = 28, P < 0.05, in the Desert Lark; and rs = (FR = F/LDM) and with wing length (FR1 = F/WL). 0.36, n = 37, P < 0.05, the Crested Lark). Wing The resultsof the statisticalanalysis concerning both measuresof fat ratio (FR and FR1) were similar, thus length also increasedin the Desert Lark (rs = we only present the resultswith fat ratio of fat cali- 0.33, n = 26, P < 0.05) and in the Crested Lark bratedwith lean dry mass. (rs = 0.29, n = 36, P < 0.05). In both species male wet masswas greater than female wet mass, RESULTS and mean lean dry massof Crested Lark males was greater than females(Table 4). There were On average,Crested Larks were heavier, they no significantdifferences between sexesin any stored more fat, and their fat ratio (fat to lean other index. dry mass)was higher than DesertLarks' (Table 1). In July lean dry massof Desert Larks was DISCUSSION 8.7%lower than average;in CrestedLarks, wet masswas 4.4%lower than averageand lean dry There are only a few studiesof fat deposition masswas 9.2%lower (Table 2). This is probably in nonmigratorybirds of nontemperatelati- becausejuveniles (which we did not alwaysdis- tudes (Grant 1965, Ward 1969). The amount of tinguish from adults) were presumably part storedfat (ca. 13%of the lean dry mass)as well of the July samples,and juveniles are lighter as its seasonal fluctuations (with extreme de- than adults (PJtzold 1986). Desert Larks' mean viation of a monthly mean from an annual mean lean dry masswas similar in all other seasons, of <0.3 g fat) in the tropicalYellow-vented Bul- 558 SHKEDYAND SAFRIEL [Auk, Vol. 108

TABLE2. Seasonalvariation (• + SE) in lean dry mass(LDM), fat ratio (FR), and wet mass(WM). •

LDM FR (%) WM Desert Lark February 6.76 + 0.12 9.2 + 0.6 21.60 _+0.36 July 6.06 + 0.11b 9.3 + 0.4 21.44 -+ 0.37 September 6.81 + 0.15 7.2 + 0.6 21.80 + 0.48 December 7.18 + 0.17 9.5 + 0.6 22.54 _+ 0.49

Crested Lark c February 12.01 + 0.12 13.4 + 0.9 39.24 _+0.29 July 10.46 + 0.16b 13.8 + 1.5 37.29 + 0.52 September 11.59 + 0.21a 10.2 + 1.0 38.69 _+0.66 December 12.24 + 0.10 a 13.1 + 1.2 40.26 _+ 0.50

aSeasonal variation analyzedby two-way ANOVA. The year of study and the interactionsbetween the year and the seasondid not affect significantlyany of the independentvariables. Comparisons between seasons were evaluatedby Duncan'smultiple range test only when the F valuesof the seasoneffect were found significantby the two-wayANOVA. bDifferent from all other seasons(P < 0.05). • Wet massin July was significantlylower than in Februaryand in December(P < 0.05). a Lean dry massin Septemberand Decemberare significantlydifferent (P < 0.05). bul (Pycnonotusgoiavier; fig. 2 in Ward 1969)are metabolism,as suggestedfor the tropical Yel- small when compared with temperate birds of low-vented Bulbul (Ward 1969). Indeed, the similar sizes. For example, fat ratios of the daytimebuildup of fat in the CrestedLark may American Goldfinch (Carduelistristis) at Ontario indicate an overnight utilization. The evening in January and February was more than 30% drop in fat ratio shown in Fig. 1 is probablyan (Middleton 1986,see also King and Farner1966, artifact. We shot only foraging birds, and the King 1972,Visser 1978, Dugan et al. 1981,Nolan very few birds shot in the eveningswere prob- and Ketterson 1983, Pienkowski et al. 1984). ably unsuccessfulearlier in the day, and hence The Crested Lark's mean fat ratio (12.9%) was they were relatively lean. The daytimebuild up similar to that of the tropicalbulbuls (13%; Ward of fat in the Crested Lark was evident in De- 1969), but the fat ratio of the Desert Lark was cember,when the nights are longest,and typ- even lower (8.9%). Seasonal fluctuations in fat ified by low averageminimum daily tempera- ratio of both specieswere insignificant,even at tures. No accumulation, however, was detected timeswhen lean dry masseswere relatively low in February, though nights are still long and (July for both speciesand Septemberfor the their temperaturesare nearly as low as in De- CrestedLark). There were no significantdiffer- cember.Our Februarydata may be misleading. ences between sexesin the pattern of fat ac- Although February fat ratio did not increase cumulation in either species, despite differ- during the day, lean dry massand wing length ences in mass between the sexes in the Crested did. This may mean that for some reason we Lark (similar to the SedgeWarbler (Acrocephalus caughtlarger birds late in the day in February. schoenobaenus);Baggott 1986). The only differ- Within-day changes in fat ratio never oc- ence between sexes could have been that of a curred in the Desert Lark. It is therefore pos- greater proportion of water contentsin Desert sible that, instead of utilizing at night the fat Lark malesthan in females,as suggested by the stored during the day, the Desert Lark lowers significant difference between male Desert its body temperatureduring the night, as birds Larks' wet massand the lack of significancein often do during fasting(Baldwin and Kendeigh lean dry mass.As comparedwith birds in tem- 1932, Biebach 1977, Ketterson and King 1977, perate,nondesert regions, the amountof stored Stuebeand Ketterson1982) or during the night fat is low in a climaticallyvery stableand warm (Prinzinger et al. 1989). Only one desert bird, region (the tropics), as well as in the Negev the Common Poorwill (Phalaenoptilusnuttallii), Desert, where seasonaldifferences in temper- engagesin long-term dormancy(Bartholomew atures are high and winter temperaturesmay et al. 1962), and we have no data on desertbirds drop to below 0øC(Zangvil and Druian 1983). that depress their metabolic rate during the It is possiblethat the observedsmall fat stor- night. As an alternative for lower night metab- age servesas an energy sourcefor overnight olism, the Desert Lark may substitutedaytime July1991] FatReserves in Larks 559

TABLE3. Spearmanrank correlationbetween fat ra- tio (FR)and time of day (h). Levelsof significance: * = P < 0.05, ** = P < 0.001.

Crested Desert CRESTED LRRK Lark n Lark n February 0.15NS 37 -0.001 NS 28 July 0.28NS 30 -0.21 NS 27 September 0.41' 22 0.16NS 25 December 0.75** 29 0.12 NS 23 •LRRKTotal 0.40** 118 0.04 NS 103

more fat than the smaller Desert Lark. The over- riding factor is presumablythe difference in "interpretation"by the two speciesof the same ßb CRESTED LRRK environment, viewed by the opportunist spe- cies as unstable,and by the specialistas rela- tively stable.We proposethat the food and hab- itat opportunism of the Crested Lark are safeguardedby reservestorage, at the cost of excessivemass, which may impair foraging and predator avoidance.The food and habitat spe- cializations of the Desert Lark enable this spe- ESERT LRRK cies to be lessdependent on fat storage.It may be more risky, at times of prolonged food scar- city, but the risk may be balancedby improved 7 9 11 I5 19 mobility. TIME (h) It can alsobe argued that it is easierfor a diet Fig. 1. Daily changesin meanfat ratio of Desert generalist than for a specialist to collect less Larksand CrestedLarks (a) for the entire studyperiod preferred food types when the preferred ones and(b) only for birdscollected in December1987 and sufficefor maintenanceonly. When food be- 1988. Vertical lines are 1 SE. comes scare, the specialist is still sufficiently efficient to obtain its maintenance require- fat storageby liver glycogenstorage for over- ments, but it has difficulties in shifting to ad- night utilization. ditional food typesfor fat storage.This may not The stored fat in the two desert specieswe apply to the Desert Lark becauseits main food studied may also buffer between-daysfluctua- type (small seeds)is more abundant by mass tions in foraging success.This might be more than large seeds,both in time and space(Shke- significant for the Desert Lark than for the dy and Safriel in prep.). Furthermore, if the CrestedLark. The relatively stablebody con- DesertLark barely managedto satisfyits main- dition of the Desert Lark suggestsan "inter- tenancerequirement and the CrestedLark nev- pretation" of the desert environment as a sea- er experiencedshortages, this shouldhave been sonally stableenvironment. The CrestedLark, reflectednot only by an averagefat storagein on the other hand, had a low lean dry massin the Desert Lark smaller than that in the Crested September,after late-summerutilization of large Lark, but also by a between-individualsvari- desertvalleys and patchesof desertagriculture. ance around mean fat ratio in the Desert Lark Thus, even though it left its breeding habitats larger than that in the Crested Lark. But, the and utilized other desert habitats, it still did not between-individuals coefficient of variation (%) maintain body massthe year round. for Desert Lark's fat ratio (35.9 + 2.81) was low- Calder(1974) suggested that large speciesare er (data from Table 1, z = 3.66, P < 0.001) than better bufferedagainst environmental variabil- that of the CrestedLark's (55.8 + 4.63), although ity and should therefore, store lessfat. But the when analyzing each seasonseparately (data larger CrestedLark stored on the average 45% from Table 2) the coefficients of variation of the 560 SHADYAND SAFRIEL [Auk, Vol. 108

TABLE4. Mean (+SD) wet mass(WM), lean dry mass(LDM), fat ratio (FR in %) and fat (F) of male and femaleCrested Lark and DesertLark. Sample sizes are in parentheses.Levels of significance:** = P < 0.01, *** = P < 0.001,NS = not significant.

Crested Lark Desert Lark

Males Females z Males Females z WM 39.9 + 2.0 (44) 37.7 + 2.6 (29) 3.8*** 22.7 + 2.3 (38) 21.0 + 1.7 (32) 3.5*** LDM 11.9 + 1.0 (35) 11.0 + 1.1 (21) 2.9** 6.8 + 0.8 (25) 6.6 + 0.6 (24) 0.9 NS FR 14.4 + 8.8 (35) 15.1 + 9.7 (21) 0.5 NS 8.7 + 3.4 (25) 8.9 + 3.7 (24) 0.3 NS F 1.7 + 0.8 (35) 1.6 + 0.9 (21) 0.1 NS 0.6 + 0.2 (25) 0.6 + 0.2 (24) NS

Desert Lark were significantly lower only in Negev: the challenge of a desert. Cambridge, July (t = 2.61, df = 76, P < 0.05). Harvard Univ. Press. FUTUYMA, J. D., & G. MORENO. 1988. The evolution of ecologicalspecialization. Ann. Rev. Ecol.Syst. ACKNOWLEDGMENTS 19: 207-233. GRANT, P.R. 1965. The fat condition of some island We thank N. Bernstein, B. Pinshow, and an anon- birds. Ibis 107: 350-356. ymous reviewer for useful comments.N. Rivlin, O. HALL, B. P., & R. E. MOREAU. 1970. An atlas of spe- Eitan, A. Verechzon, and A. Amir helped with labo- ciationin African passerinebirds. London,Trust- ratory work, Y. Heller provided laboratoryspace, and ees Br. Mus. the Mitrani Center for Desert Ecology provided lo- KETTERSON,E. D., & J. R. KING. 1977. Metabolic and gistic support. This research was supported by the behavioral responsesto fasting in the White- Fund for BasicResearch of the Israel Academy of crownedSparrow (Zonotrichia leucophrys gambelii). Sciences and Humanities, and a contribution from the Physiol. Zool. 50: 115-129. Sebafamily. This is publicationNo. 114of the Mitrani KING, J. R. 1972. Adaptive periodic fat storageby Center for DesertEcology. birds. Pp. 200-217 in Proc. 15th Int. Ornithol. Congr. (K. H. Voous, Ed.). Leiden, Brill. LITERATURE CITED , & D. S. FARNER.1966. The adaptive role of winter fattening in the White-crowned Sparrow BAGGOTT,G. K. 1986. The fat contentsand flight with commentson its regulation.Am. Nat. 100: rangesof four warbler specieson migration in 403-418. north Wales. Ringing & Migr. 7: 25-36. LIMA, S. L. 1986. Predation risk and unpredictable BALDWIN,S. D., & S. N. C. KENDEIGH.1932. Physi- feeding conditions:determinants of body mass ology of the temperature of birds. Sci. Publ. in birds. Ecology 67: 377-385. Cleveland Mus. Nat. Hist. 3: 1-196. 1988. Initiation and termination of daily BARTHOLOMEW,G. A., J. W. HUDSON, & t. R. HOWELL. feeding in Dark-eyed Juncos:influences of pre- 1962. Body temperature, oxygen consumption, dation risk and energy reserves.Oikos 53: 3-11. evaporative water loss,and heart rate in the Poor- MIDDLETON,A. L. A. 1986. Seasonal changes in will. Condor 64: 117-125. plumage structureand body compositionof the BIEBACH,V. H. 1977. Reduktion des energiestoff- American Goldfinch, Carduelis tristis. Can. Field- wechselsund der Korpertemperaturhungernder Naturalist 100: 545-549. Amseln (Turdusmerula). J. Ornithol. 118:294-300. MOP,EAU, R. E., & R. M. DOLP. 1970. Fat, water, weights BLEM,C.R. 1975. Geographicvariation in wing-load- and wing-lengths of autumn migrants in transit ing of the House Sparrow. Wilson Bull. 87: 543- on the northwestcoast of Egypt. Ibis 112: 209- 549. 228. 1976. Patternsof lipid storageand utiliza- NEWTON,I. 1969. Winter fattening in the bullfinch. tion in birds. Am. Zool. 16: 671-684. Physiol. ZooI. 42: 96-107. CALDER,W. A. 1974. Consequencesof body size for NOt,N, V., JR.,& E. D. KETTERSON.1983. An analysis avian energetics.Pp. 86-151 in Avian energetics of body mass,wing length, and visible fat de- (R. A. Paynter, Ed.). Cambridge, Publ. Nuttall positsof Dark-eyed Juncoswintering at different Ornithol. Club, No. 15. latitudes. Wilson Bull. 95: 603-620. DUGAN, P. J., P. R. EVANS, L. R. GOODYER,& N. C. P.•TZOLD, R. 1986. Heidelerche und Haubenlerche DAVIDSON. 1981. Winter fat reserves in shore- Lullula arboreaund Galeridacristata. Wittenberg, birds: disturbanceof regulatedlevels by severe Lutherstadt, A. Ziesmen Verlag. weather conditions. Ibis 123: 359-363. PIANKA,E.R. 1981a. Competition and niche theory. EVENARI, M., L. SHANAN, & N. TADMOR. 1982. The Pp. 167-196 in Theoreticalecology. Principles and July1991] FatReserves inLarks 561

applications(R. M. May, Ed.). London, Blackwell SFIKEDY,Y. 1990. Niche breadth and geographical Sci. distribution of the Desert Lark (Ammomanesdes- 1981b. Evolutionary ecology. New York, ertl) and the Crested Lark (Galeridacristata). Ph.D. Harper & Row. dissertation,Jerusalem, Hebrew Univ. Jerusalem. PIENKOWSKI, M. W., P. N. FERNS, N. C. DAVIDSON, & STUEBE,M. M., & E. D. KETTERSON.1982. A study of D. H. WORRALL. 1984. Balancing the budget: fasting in Tree Sparrows(Spizella arborea) and measuringthe energy intake and requirements Dark-eyed Juncos(Junco hyemalis): ecological im- of shorebirdsin the field. Pp. 29-56 in Coastal plication. Auk 99: 299-308. waders and wildfowl in winter (P. R. Evans, J. VISSER,J. E. 1978. Fat and protein metabolism and D. Goss-Custard, and W. G. Hales, Eds.). New mortality in the Coot Fulicaatra. Ardea 66: 173- York, Cambridge Univ. Press. 183. PPaNZINGER,R., I. L•JEEEN,& K. -L. SC•JC•MANN. 1989. WARD,P. 1969. Seasonaland diurnal changesin the Energy metabolismand body temperaturein 13 fat content of an equatorialbird. Physiol. Zool. sunbird species(Nectariniidae). Comp. Blochem. 42: 85-95. Physiol. 92A: 393-402. ZANGVIL,A., & P. DRUIAN. 1983. Meteorologicaldata ROGERS,C. M. 1987. Predation risk and fasting ca- for Sede Boqer.Desert MeteorologyPapers, Se- pacity:Do winteringbirds maintain optimal body ries A, No. 8. mass?Ecology 68: 1051-1061.