View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by PDXScholar

Portland State University PDXScholar

Biology Faculty Publications and Presentations Biology

11-1987 The mpI act of Weather on Foraging Behavior

Michael T. Murphy Portland State University, [email protected]

Let us know how access to this document benefits ouy . Follow this and additional works at: http://pdxscholar.library.pdx.edu/bio_fac Part of the Biology Commons, and the Ornithology Commons

Citation Details Murphy, Michael T., "The mpI act of Weather on Kingbird Foraging Behavior" (1987). Biology Faculty Publications and Presentations. Paper 82. http://pdxscholar.library.pdx.edu/bio_fac/82

This Article is brought to you for free and open access. It has been accepted for inclusion in Biology Faculty Publications and Presentations by an authorized administrator of PDXScholar. For more information, please contact [email protected]. The Condor89:721-730 ? The CooperOrnithological Society 1987

THE IMPACT OF WEATHER ON KINGBIRD FORAGING BEHAVIOR'

MICHAELT. MURPHY2 Museum of NaturalHistory, The Universityof Kansas, Lawrence,KS 66045

Abstract. Foragingdata on EasternKingbirds (Tyrannus tyrannus) were collected during the early breedingseason in eastern Kansas to test the hypothesis that foragingrate and other aspects of foragingbehavior vary with weather. Foragingcharacteristics of five ad- ditional kingbirdspecies were also examined to assess Fitzpatrick's(1980) generalization that (Tyrannus spp.) are aerial specialists. In EasternKingbirds, total foragingrate was independentof air temperature,cloud cover, wind speed, and time of day, but the rate of aerialhawking varied directlywith air temperatureand inverselywith cloud cover (both P < 0.05). Effectsof the two variableswere additive.The percentageof foraging movementsthat were aerialhawks also increasedwith temperatureand declinedwith cloud cover,and, hover-gleaningand perch-to-groundsallying were observed mainly during cloudy weather. Sally (i.e., foragingflight) distance correlateddirectly with perch height and air temperature,and largeinsects were captured almost exclusivelyin long upwardor horizontal flights. I interpretthese data to indicate that foragingbehavior and the captureof large, flyinginsects depends on weatherbecause of how it affectsthe activityof insectprey. Foraging data on kingbirdssupport Fitzpatrick's generalization, but the relativeuse of aerialhawking varies considerablyamong species. Resident TropicalKingbirds (T. melancholicus)are the most specialized foragers,whereas the migrant and widely distributedEastern Kingbird appearsto be the most generalized.Certain habitats also appearto favor the use of particular foragingmethods (e.g., outwardstriking in grasslands,and perch-to-groundsallying in drier, open habitats). Keywords: Aerialhawking;foraging behavior; prey capture; Tyrannus tyrannus; weather.

INTRODUCTION mate weather conditions should be a primary Avian ecologists have quantified tyrant flycatch- determinant of flight activity, food avail- er foraging behavior for many purposes ranging ability and therefore foraging behavior of fly- from the study of morphology (Karr and James catchers. However, with only rare exception (Da- 1975, Traylor and Fitzpatrick 1982), resource vies 1977, Foreman 1978), all flycatcher foraging use and community structure (Beaver and Bald- studies to date have explicitly avoided collecting win 1975, Eckhardt 1979, Blancher and Rob- data during poor weather, and as a result, there ertson 1984, Sherry 1984), phylogenetic rela- is a critical shortage of information on how ty- tionships (Fitzpatrick 1980), to analyses of rannid foraging varies with weather. migratory behavior (Verbeek 1975a). Analyses This gap in knowledge is significant for at least of stomach contents have shown that diets are two reasons. From the perspective of under- diverse, but the main flycatcher prey are the Hy- standing the evolution of flycatcher reproductive menoptera, Coleoptera, Orthoptera, Hemiptera, patterns it is essential to determine what factors and Diptera (Bent 1942, Dick and Rising 1965, influence foraging success. In the other large group Hespenheide 1971, Beaver and Baldwin 1975, of aerial foragers, swallows and swifts, foraging Sherry 1984). An unexplored, yet important area and reproductive success vary with weather be- of flycatcher foraging ecology concerns the im- cause of the strong dependence between insect pact of weather on foraging patterns. In view of flight activity and meteorological conditions. Low the dependence between body temperature and air temperatures, precipitation, high cloud cover, capacity for flight in most (Heinrich 1981, and wind all reduce the availability of flying in- Kingsolver 1983a), it seems logical that proxi- sects (Bryant 1973; Hespenheide 1975; Davies 1977; Visscher and Seeley 1982; Kingsolver 1983a, 1983b; Jones 1987). The accepted view IReceived 11 1986. Final 13 September acceptance is that life histories of that the May 1987. species exploit 2 Presentaddress: Department of Life Sciences, In- "aerial plankton" have been shaped largely by diana State University, TerreHaute, IN 47809. high, but short-term temporal and spatial vari-

[721] 722 MICHAELT. MURPHY

ation in insect abundance(Lack and Lack 1951; Most observations were of individuals in pas- Bryant1973; O'Connor 1977, 1979;Jones 1987). tures and recentlyabandoned fields where small Likewise,hawkers such as flycatchersdepend on shrubs,fence posts, and fence lines werethe main insect movement for prey detection and capture perch substrates. (e.g., Davies 1977), and Murphy(1983) suggest- Terminologyfollows Fitzpatrick(1980). Sal- ed that weatherhas also been an integralfactor lying refers to the approach flight preceding a in the evolution of temperate-zonebreeding fly- prey capture attempt. All foragingmoves were catcherlife histories.Poor weathernegatively af- initiated by a flight from a perch, and included fects flycatcher reproductive success (Davies aerial hawking(direct flights to capturea single, 1977, O'Connor and Morgan 1982, Murphy flyinginsect), outwardstriking (prey are snatched 1983), yet except for Davies' (1977) study, the from vegetationfollowing a rapid, direct flight), link betweenweather, foraging, and preycapture outward hover-gleaning(prey are located and is not establishedin flycatchers(but see Mahan picked from vegetation following a brief hover- 1964). ing flight),and perch-to-ground sallying(prey are Second, weather's impact on flycatcherfor- picked from the groundafter a direct flight). aging behavioris relevantfor interspecificstud- Foraging were observed from 250 m ies of either niche relationships (Beaver and usingbinoculars under all weatherconditions ex- Baldwin 1975, Verbeek 1975b, Holmes et al. cept rain. Observations were recorded vocally 1978), or assessment of foraging specialization with a hand-held tape recorderand later were. (Fitzpatrick1980). Measuringniche overlap us- transcribed.Upon sighting a foragingbird (in- ing only "fair-weather"studies may be mislead- dicated by previous foragingattempts, and alert ing since diet and foragingbehavior are expected posture),I waited for it to initiate a new foraging to diversify when food becomes scarce (Pyke et sequenceand then recordedtime of day, and all al. 1977), which is most likely to occur when furtherperch heights, foraging movements, perch poor weather depresses insect activity. Since changes without a prey capture attempt being competition is most likely when food is least made, sally distances, and flight trajectories abundant,quantification of communityrelation- (above, below or level with the perch).Horizon- ships and foragingplasticity in flycatchersshould tal flightsoccurred within about 150 of the plane not be restrictedto fair weather. parallelwith the groundsurface. Foraging period In this report I quantify foragingpatterns of duration was also timed to the nearest second EasternKingbirds (Tyrannus tyrannus) in east- with a stop watch. Observationswere continued ern Kansas to addressthe issues raised above. I for as long as a birdwas in sightup to a maximum will test predictions that short-termchanges in of 15 min. I ended all observationsat 15 min to weather are the primary determinantsof vari- avoid overrepresentingforaging under any spe- ability in kingbirdforaging behavior and success cific environmentalcondition (e.g., a local insect (Murphy 1983). I will also assess Fitzpatrick's emergence). (1980) classificationof the genus Tyrannusas Within the study area fences and fence posts aerialhawking specialists by evaluatingtheir de- wereabundant, were of uniformheight and spac- gree of behavioral flexibility through (1) intra- ing, and were frequently used by kingbirds. I and interspecificcomparisons of kingbird for- therefore estimated sally distances and perch aging, and (2) by examining the consistency of heights to the nearestfoot and half-foot, respec- EasternKingbird foraging as weatherchanges. tively, by referringto fence lines. I later con- verted distance to meters. Except in the case of METHODS I was unaole to determinewhether lcanc •lISeCLS I collecteddata on EasternKingbirds in May and capture attempts were successful. However, early June of 1981, 1982, and 1983 in Douglas kingbirdshad to kill largeinsects by beatingthem County,eastern Kansas (site descriptionin Mur- againsta perch,which allowedme to recordtheir phy 1986).Foraging behaviors were recorded op- capture. portunisticallywhile kingbirdnests were located At the end of each foragingperiod, I recorded early in the season. I limited my observationsto air temperatureusing a Schulteis fast reading pre-eggstage birds to controlfor seasonalchanges thermometer(shielded from sun and wind) and in food abundance(Murphy 1986) and for vari- wind speed at breastheight (1.4 m) with a Sims ation in energyneeds due to breedingactivities. hand-held cup anemometer. The stall speed of KINGBIRDFORAGING 723 the anemometerwas about 8 km/hr, hence, I was diversity and equitabilityof use of differentfor- able to measurewinds that exceeded 8 km/ and is definedas where only agingcategories, H'/H'max, hr. The wind measurementsno doubt failed to H' = - pilog pi, and pi is the proportionof each precisely reflect the actual convective microen- foragingbehavior. H'max is the maximum diver- vironment of the , but I believe they accu- sity possible given n foragingcategories, where rately portrayedgeneral environmental condi- pi = 1/n for all i. J varies between 0 and 1, the tions. Cloud cover was estimated by classifying two extreme values representinga specialistus- sky conditions into five categoriesof percentage ing one behavior (0) and a generalist using all cover (i.e., 1 = 0 to 20%, 2 = 21 to 40%, . . . , foragingmethods equally (1). 5 = 81 to 100%). PREDICTIONSAND ANALYSIS RESULTS GENERALFORAGING CHARACTERISTICS Insect flightactivity increaseswith increasingair temperature,clearing sky conditions, and de- BreedingEastern Kingbirds foraged as "sit-and- creasingwind speeds (see referencesabove). As- wait" predatorssince 85%(n = 477) of searches sumingenergy intake is sensitive to insect activ- endingin a foragingflight. Of these, aerialhawk- ity, I expected the following patternsto emerge. ing accounted for 50.0% of all foragingmoves, First, total foragingand hawkingrate would in- followedby outwardstrikes (39.9%), hover-gleans crease with air temperature,and decline with (8.1%),and perch-to-groundsallies (2.0%). Other increasingcloud cover and wind speed. Second, foragingbehaviors observed infrequently at oth- assumingperches were selected to maximize en- er times (e.g., during rain or the emergence of counters with prey, I predicted an increase in aquaticinsects) included gleaningof lepidopter- perchheight with increasingair temperature,de- an larvae from trees by upwardstriking, surface creasingcloud cover, and wind speed.Third, Leck gleaning from water, and multiple captures of (1971) proposedthat long sally distancesin king- aerialprey by hoveringduring a single flight.The birdsreflect good foragingconditions. Long flights latter behavior, describedalso by Blancherand presumablyindicate greateravailability and se- Robertson (1985), involved many consecutive lectivity of prey. I thereforeexpected sally dis- hover-gleanswhile the birdmoved in a slow flight tance to increase with air temperature,and de- over grass covered fields. Frugivoryon mulber- clinewith increasingcloud cover and windspeeds. ries (Morus rubra; Stapanian 1982) was also Optimal foragingtheory (Pyke et al. 1977) pre- common later in the season when fruit was in- dicts also that diet diversity, and in this case cluded in both adult and nestling diets (M. T. foragingdiversity, will increase as the expected Murphy,pers. observ.). rate of energy intake declines. Hence, kingbirds Perch height averaged 1.3 m (SD = 1.53, n = should switch from being foragingspecialists to 49; range= 0.3 to 9.1 m) and mean sally distance generalistsas weatherdeteriorates. was 2.9 m (SD = 1.58 m, n = 49; range = 0.9 I excluded all foraging bouts in which birds to 7.6 m). The median sally distance of 2.1 m were observed for <3 min. For the remainderI was shorterthan the 3.7 m reportedby Via (1979). calculated foraging rate (foraging moves/min), Kingbirds returnedto the same perch used to hawkingrate (number of aerialhawks/min), mean initiate prey captures 53.5% of the time (n = and median perchheight, sally distance,and the 325), similar to Via's reported value (48.8%; percentageof foragingmovements as aerialhawks t-test for percentages,t = 0.7, P >> 0.05). How- (minimum of three foraging movements re- ever, sally distanceinfluenced the probabilityof quired for inclusion). Percentageswere arcsine perchreuse since mediandistance for individuals transformed.Univariate comparisonsof behav- reusing percheswas 1.5 m, compared to 2.5 m ior to weatherwere made using least-squareslin- for new perches(median test, G = 15.6, df = 1, ear regression and correlation analysis. Multi- P < 0.01; Zar 1974). Sally distance also varied variaterelationships were further examined using with flighttrajectory. Downward, horizontal, and step-wise multiple regression(BMDP2R; Dixon upward flights accounted for about 59, 32, and 1981). I also groupedbirds accordingto weather 9% (n = 376) of all attempted prey captures, conditionsand calculatedforaging diversity with respectively. Median (and mean) sally distance respect to weather using a standardmeasure, J for level (2.3 m [2.9 m]) and upwardflights (2.4 (Eckhardt1979). J takes into account both the m [3.8 m]) did not differ significantly(median 724 MICHAELT. MURPHY

.013 r2= r2 .106 2.8 1.6 PP<.05 w 0 2.1 4 0 o S0.8 0 z 1.4 - *

0.7 * *

0.0 0.0 I 2 3 4 5 CLOUD COVER ABSENT TOTAL ,- FIGURE2. Hawkingrate (number ofhawks/min) vs. cloudcover for Eastern Kingbirds from Douglas Coun- 0.8 ? ty, Kansas. 0-

0.4

0.0 regressionindicated that effects of temperature 12 15 1s 21 242i and cloud cover were additive, and that despite AIR TEMPERATURE(0C) uncontrolledvariation in absolute insect abun- dance among years (Murphy 1986), recent past FIGURE1. Foragingrate (number of moves/min)as and differences birds in a functionof ambientair temperaturefor (a) all for- weather, among hunger agingtactics, and, (b) just hawking behaviors (number level and body condition, the two-variablemod- of aerialhawks/min) for Eastern Kingbirds from Doug- el of air temperatureand cloud cover explained las County,Kansas. a quarterof the variation in hawking rate (r = 0.501, df = 2, 37, P < 0.01). The partialcorre- lation of both and cloud cover were test, G = 0.47, df = 1), but the median of their temperature < when effects of the other combined distributions was longer than for significant(P 0.05) variablewere controlled time downward flights (1.9 m [2.3 m]; G = 5.80, statistically.Adding of to the model little additional df= 1, P < 0.025). Largeinsects were also more day explained variation = df = P < likely to be capturedin upward(6.0% of all for- (r 0.537, 3, 36, 0.01). and level aging moves) (3.4%)flights compared PERCHHEIGHT AND SALLYDISTANCE to downwardflights (0.4%)(x2 = 6.96, df = 2, P < 0.025). Perch height did not vary with air temperature (r = 0.238), wind speed(r = -0.134), cloud cover FORAGINGRATE (r = -0.050), or time of day (r = -0.092). Sally Foragingrates varied widely (X = 1.2 moves/ distancealso variedindependently of wind speed min, SD = 0.76, n = 40), but essentiallynone of (r = -0.049), cloud cover (r = -0.042), and time the variabilitywas relatedto air temperature(r = of day (r = -0.171), but did tend to vary directly 0.110; Fig. la), wind speed (r = -0.033), cloud with temperature(r = 0.250, P = 0.12, n = 40). cover (r = 0.027) or time of day (r = -0.048). As is common in flycatchingbirds (Pinkowski Furtherrestriction of the analysis to timed ob- 1977, Greig-Smith 1983, Moreno 1984), perch = servations >5 min, or min did not affectthe height and sally distance were correlated (r results. _7 0.441, n = 40, P < 0.01), and even higher so if I examined hawking rate separatelybecause birds that perched in trees were excluded (r = kingbirds(Tyrannus spp.) preferentiallycapture 0.623, n = 36, P < 0.001). I thus reexamined insects by aerial hawking (Fitzpatrick 1980). sallydistance in relationto temperatureand perch Hawkingrate increased with air temperature(r = height using multiple regression analysis. The 0.395, P < 0.02; Fig. lb) and decreased with two-variablemodel was significantfor birds that cloud cover (r = -0.325, P < 0.05; Fig. 2). did not use trees as perches (r = 0.683, df = 2, Hawkingrate did not vary with wind speed (r = 33, P < 0.01), but the partialcorrelation of sally 0.072), but tended to decline as the day pro- distanceand temperaturewas not (r = 0.250). A gressed(r = -0.235, P = 0.15). Stepwisemultiple plot of residual sally distance (effects of perch KINGBIRD FORAGING 725

2 ; I*r ?

0 0 0 3

* 0

AIR TEMPERATURE (o0)(?C) FIGURE3. Residualsally distance (perch height effects removed) vs. ambientair temperaturefor Eastern Kingbirdsfrom Douglas County, Kansas. Each point is basedon themean of all salliesfrom individual foraging bouts.Open circles are for bouts lasting >3 min,but <5 min, whereasclosed circles are for bouts lasting >-5 min.The solid line is theleast-squares linear regression of residualdistance vs. temperaturefor all pointsexcept the two low valuesnear 21?C (P = 0.03;see text). height removed) vs. temperature(Fig. 3) sug- ,I found that the percentageof hawksvar- gested that the nonsignificantpartial correlation ied directlywith air temperature(r = 0.378, P = of sally distance with temperaturewas due to 0.02), inversely with cloud cover (r = -0.408, two low values near 2 1C. I observed both birds P = 0.01), but was independent of wind (r = for just over 3 min, and it is possible that the 0.053). Temperatureand sky conditionstogether short observation periods resulted in spurious accounted for about 27% of the variation in measurements.Residual sally distance and tem- hawkingpercentage (r = 0.523, df = 2, 34, P < peraturecorrelated significantly when either the 0.01; % hawks = 27.0 + 1.83[TEMP] - lattertwo points were excluded (r = 0.380, df = 4.53[SKY]).Temperature and cloud cover were 33, P = 0.03; Fig. 3), or if the analysiswas limited not related (r = -0.129, P > 0.40), hence both to foragingepisodes lasting > 5 min (r = 0.391, variablescontributed significantly (P < 0.05) to df = 24, P = 0.054; Fig. 3). the two-variableregression model. Foragingdi- varied over the of air FORAGINGDIVERSITY versity substantially range temperaturesand sky conditions, mainly due to The fact that hawkingrate, but not total foraging the high diversitiesobserved at intermediateval- rate,varied significantly with air temperatureand ues of both environmental variables. Foraging cloud cover suggestedthat the frequencyof use diversitywas greatestbetween 15 and 180C,and of differentbehaviors varied with weather. In- at about 50%sky cover. deed, this was the case (Table 1). At low air tem- INTRA-AND INTERSPECIFICCOMPARISONS peratureskingbirds hawked infrequently and re- lied mainly on outwardstriking from vegetation. I have attempted to summarize each study of At the highestair temperatures they hawked about kingbirdforaging (Table 2) within Fitzpatrick's 70% of the time. Cloudy skies also produced a frameworkso as to facilitatecomparisons. With sharp drop in the frequencyof hawking. Inter- the exception of Tatschl's data (which reported estingly, perch-to-groundsallying and hover- substratesfrom which prey were taken instead gleaningwere almost never observed except un- of foragingmovements), aerial foraging (hawking der cloudy conditions. Using the percentageof plus sallying)was reportedas the usual foraging each bird's foragingmovements that were aerial method for all species, accounting for between 726 MICHAELT. MURPHY

TABLE1. Variationin foragingbehavior and foraging diversity (J) by EasternKingbirds as airtemperature (0C)and cloud cover varied during foraging.

Foragingbehaviors" Strike Hover P-to-G nbJ Temperature 13-14 0.358 0.566 0.056 0.019 53 0.668 15-16 0.361 0.556 0.000 0.083 36 0.820 17-18 0.488 0.512 0.000 0.000 43 1.000 19-20 0.493 0.400 0.093 0.013 75 0.716 21-22 0.553 0.404 0.042 0.000 47 0.753 23-24 0.671 0.205 0.096 0.027 73 0.659 25-26 0.714 0.190 0.095 0.000 21 0.710 Skycover 1 (clear) 0.676 0.309 0.015 0.000 68 0.629 2 0.644 0.311 0.044 0.000 45 0.713 3 0.633 0.366 0.000 0.000 30 0.947 4 0.506 0.376 0.082 0.035 85 0.746 5 (total) 0.342 0.526 0.096 0.035 114 0.756 aValues are the proportionof each foragingbehavior. Hawk = aerialhawk; Strike = outwardstrike; Hover = hover-gleaning; P-to-G = perch-to- groundsally. bSample size refersto numberof preycapture attempts.

70 and 90% of a species' foragingmovements. DISCUSSION The was of exception my study Eastern King- My purposein measuringthe impact of weather birds in Kansaswhere aerial accounted hawking on kingbirdforaging was twofold. I first sought for 50%of This dif- only prey captureattempts. to test predictionsarising from Murphy's(1983) fers fromthe of about greatly hawkingpercentage suggestionthat EasternKingbird life historyevo- 90%in the two studiesof Eastern from Kingbirds lution has been driven mainly by the influence West and Ontario. im- Virginia Although it is of short-term,weather-induced changes in food to realize that Tatschl's classification portant availabilityon reproductivesuccess. Support for method differedfrom the other note that studies, this hypothesis was obtained by showing that fartherwest in Kansas EasternKingbirds foraged kingbirdforaging patterns, and in particularat- from even more than individuals in vegetation tempted prey capturerates, varied with weather my study. in the same fashion as insect availability (see Considerable variabilityalso exists both with- references above). Second, and from a more in and in the use of other among species foraging methodological standpoint, I attempted to de- methods. for Perch-to-groundsallying, example, termineif avoidingpoor weatherwhile collecting was uncommon in all relatively species except data might lead to underestimatesof foraging Western(T. verticalis)and Cassin'skingbirds (T. breadth. At least for open-country flycatchers, vociferans).Sallying, the successive capture of "fair-weather"studies do appear to underesti- several in flying insects one foragingflight, was mate behavioralflexibility. observed in at least four species, but was only moderatelycommon in one populationof West- THEIMPACT OF WEATHER ern Kingbirds.Gray Kingbirds (T. dominicensis) foragedfrom vegetation about a quarterof the Although total foragingrate was unaffectedby time, whereas Tropical Kingbirds (T. melan- the measuredmeteorological variables, weather cholicus)appear to be the most behaviorallyspe- did influence foragingbehavior and very prob- cialized forager.They relied on aerial hawking ably had a markedeffect on total energyintake. nearly 95% of the time. Although conclusions For example, the frequency and rate of aerial about Scissor-tailedFlycatchers (T. forficatus) are hawkingrose with increasingair temperatureand limited because the data are not strictly com- declining cloud cover. Perch-to-groundsallying parable, scissortailsappear to capture most in- and hover-gleaningincreased mainly when skies sects directly from vegetation. were clouded over (Table 1), indicating a shift KINGBIRDFORAGING 727

TABLE2. Foragingmode profiles of six speciesof kingbirds.Data are presented as percentagesof the total number(n) of foragingmoves. T indicatestrace use (< 1%).Numbers 1 through9 underspecies headings refer to literaturesources. Brackets enclose methods that were not distinguishedin the originalsources.

Species" Foraging EKB WKB CKB STF TKB GKB methodb 1 2 3 4 1 5 5 6 1 7 8 9 Hawking 17 92 86 50 26 58 70 84 8 94 71 O-Strike 2 81 8 1 559 - 8 86 88 26 Hover f8f8 12 4088 7 iff1I P-G-Sally 2 1 2 15 17 22 9 6 1 3 Sally - - - T - 18 7 - - 2 - - Gr-Hop ------10 - - Fruit - - - T - - - - - T 3 T n 247 68 2,443 406 352 162 124 183 67 2,718 425 34 aEKB = EasternKingbird; WKB = WesternKingbird; CKB = Cassin'sKingbird; STF = Scissor-tailedFlycatcher; TKB = TropicalKingbird; GKB = GrayKingbird. bHawking = aerialhawking; O-Strike = outwardstriking; Hover = hover-gleaning;P-G-Sally = perch-to-groundsallying; Sally = > 1 preycapture/ foragingflight; Gr-Hop = > 1 preycapture/flight to the ground;Fruit = frugivory. Sources:1-Tatschl 1973; 2-Via 1979; 3-Blancher and Robertson1985; 4-this study;5-Goldberg 1979; 6-Landres and MacMahan1980; 7- Foreman1978; 8 -Fitzpatrick 1980;9- Ricklefsand Cox 1977.

from passive to active foraging. Insects com- tances also tended to be short in comparisonto monly use the sun to bask and raise body tem- other kingbirds(see referencesabove). peratureto the levels necessaryfor flight (King- As indicated by the diversity index, J, king- solver 1983a, 1983b), and I suggest that when birdsexhibited high behavioral flexibility in their skies were cloudy kingbirdswere forced to ac- mode of preyattack (compare to Eckhardt1979). tively search for prey because few insects were However, foragingdiversity did not changewith flying. Unlike the other variables wind had no weatheras expected.Optimality models (Pykeet measurable influence on hawking rate, or any al. 1977) yielded the predictionthat foragingdi- otherforaging trait. Foreman (1978) showed that versity would increase linearly as the expected foragingrate and success of Scissor-tailedFly- rate of energyintake fell, i.e., as air temperatures catchersvaried directly with temperatureand in- droppedand cloud cover increased.Instead, for- versely with cloud cover, but he found also that aging diversity peaked at intermediateweather strongwinds depressedforaging success. The un- conditions. I interpretthese, and the previously importanceof wind in my study probablystems discussed patterns, as indicating that limits to from a lack of strong winds during measure- insect flight duringpoor weatherwere so severe ments. that kingbirds had to almost abandon aerial Leck (1971) suggested and Davies (1977) hawking under such conditions. They switched showed that long flights by aerial hawkersindi- to strikingand hover-gleaninginsects from vege- cated attempts to capturelarge prey. I therefore tation using short, downwardflights. As weather predictedand verified that sally distance would improved insects began to fly and kingbirdsin- vary directlywith temperature.Moreover, most creasedthe frequencyofhawking. However, flying capturesof largeinsects were by long horizontal, insects were probablynot yet abundantenough but especiallyupward flights. I suspectthat higher to permit abandoningforaging from vegetation. air temperatureincreased insect flight activity As a result,at intermediateconditions kingbirds and permitted kingbirds to 'pursue larger, and foragedas extremegeneralists. With furtherim- more abundantflying prey. Unlike the other for- provements in weather, kingbirds switched to aging variablesperch height was independentof mainlyaerial hawking and the use of longerflights weather.This lack of significance,and the gen- to capturelarge insects. Foragingdiversity there- erallylow perchheights that I recorded(compare fore tends to be minimal at extreme environ- to Foreman 1978, Goldberg 1979, Via 1979, mental conditions, but for differentreasons. Blancherand Robertson 1984), probably stem The switch from outward striking to aerial from the low perch diversity in the study area. hawkingprobably has a major impact on paren- As a result of the use of low perches, sally dis- tal feedingcapacity. Flycatchers feed their young 728 MICHAELT. MURPHY with individual, whole prey (Bent 1942; pers. tricolor,"uniquely combines aerial hawking with observ.),and hawkingresults usually in the cap- outward striking."But, A. tricolorand Eastern ture of largeprey in a single flight. On the other Kingbirdsin my study aerialhawk and outward- hand, outward striking from vegetation yields strikeat virtuallyidentical frequencies.Presum- few largeprey items, and involves high levels of ably this is relatedto similarity of habitat. activity.Hence, during poor weatheradults might maintain a positive energy balance by making KINGBIRDS:SPECIALIST OR GENERALISTFORAGERS? numerous,short feeding flights on a concentrated supply of small prey. But, it is doubtfulwhether Fitzpatrick (1980) arbitrarily designated that parentsfeeding a brood of youngat a distantnest specialistforagers use one preycapture technique can use this tactic to keep young fed adequately. for greaterthan 50%of all attackson prey. Thus, Given that largeinsect flightactivity declines at despite differencesamong kingbirdsin whether all levels of the air space when temperaturefalls or not they foragefrom the groundor vegetation (Jones 1987), the poor reproductivesuccess of (Table2), all kingbirdsstudied are aerial hawking kingbirdsduring inclement weather seems best specialists.The only uncertaintyis over the Scis- explainedas an inabilityby the parentsto capture sor-tailed Flycatcher, which may concentrate the largerinsect food of the young(Murphy 1983). much of its foragingfrom vegetation.This would Mahan (1964) showed that Eastern Phoebes not be surprisinggiven that Eastern Kingbirds (Sayornisphoebe) decrease food deliveryrates to appearto increasethe use of strikingfrom vege- nests when wind speeds increaseand light inten- tation in similar habitats(see also Blancherand sity falls. But, he also found that delivery rate Robertson 1985). The importanceof habitat in increasedas air temperaturefell. I suggest this establishingdifferences among species in "mi- was becausephoebes were unable to capturelarge nor" foragingspecializations is furthersuggested insects, and like kingbirds,switched to foraging by the frequentuse of perch-to-groundsallying on small prey when air temperatureswere low. by Westernand Cassin'skingbirds (Table 2). Be- A question that now arises is how much of the cause of the availabilityof bare ground in drier differencein the use of aerial hawkingbetween portions of western North America, perch-to- kingbirdsin Kansas and easternNorth America groundsallying is a possible foragingtechnique. (Table 2) is due to weather. By assuming "fair- It is therefore not surprisingthat Bent (1942) weather"conditions (air temperature of 250Cand reportedthat ground-foragingbeetles (Carabidae clear sky), and using the multiple regression and Cicindelidae)were more common in the diets equationrelating hawking percentage to temper- of WesternKingbirds than any other tyrant. atureand cloud cover (see above), I estimatethat Of the species listed in Table 2, only the Trop- EasternKingbirds hawk for insects about 70% ical Kingbirdis a resident, and it appearsto be of the time during good foragingconditions in the most specializedforager. Sherry (1984) pre- Kansas.I believe the remainingdifference is due viously noted greaterspecialization among res- to habitat. Most of my observationswere made idents comparedto migrantsin other tyrants.He in pasturesand grasslands,which may favor the suggestedthat temporalvariation in the diversity use of nonhawkingforaging methods (especially and abundance of available prey for migrants, outwardstriking from vegetation).The main in- especiallyin physicallystressful grassland habi- sect prey of kingbirdsin Kansas are Hymenop- tats, was muchgreater than for tropicaltyrannids tera, Orthoptera,and Coleoptera(Dick and Ris- (see Wiens and Rotenberry1979 for insect data). ing 1965), which tend to be concentratedat the Greater variability in insect availability forces air-vegetationinterface. Insect density is prob- migrants to be opportunistic foragers, as the ably very high in this region, making foraging EasternKingbird data demonstrate. I suggestfur- profitableat or just above the grass surface. ther that the level of foraging specializations Tatschl (1973) reportedan even greaterreli- within migrantsis a function of the habitat and ance on foraging from vegetation (81%). Al- climatic diversity encountered over their geo- though his figure appears somewhat high, our graphicrange. Present information (Table 2) can data concurand suggestthat living in grasslands not be used to test this idea, but the data are favors using vegetation as a foragingsubstrate. suggestive. Eastern Kingbirds have the widest Furthermore,Fitzpatrick (1980) stated that a distributionin temperatelatitudes of any king- South Americangrassland specialist, Alectrurus bird, and they appear to be the most variable KINGBIRD FORAGING 729 foragers,especially if frugivoryis included (Sta- DIXON, W. J. 1981. BMDP-81. Univ. of California panian 1982; E. S. Morton, pers. comm.). Press, Berkeley. is a relative and as ECKHARDT, R. C. 1979. The adaptive syndromes of "Specialist" term, applied two guilds of insectivorousbirds in the Colorado by Fitzpatrick(1980), does not separatetyrants Rocky Mountains.Ecol. Monogr.49:129-149. using similar foraging behaviors adequately. FITZPATRICK,J. W. 1980. Foraging behavior of Neo- Nonetheless, it is a useful concept for examining tropicaltyrant flycatchers.Condor 82:43-57. between behaviorand traits FOREMAN, L. D. 1978. Activity patternsand foraging relationships foraging behavior of the Scissor-tailed such as or mi- Flycatcher(Musci- reproduction(Murphy, unpubl.) voraforficata). Ph.D.diss., A&M Univer- gration(Verbeek 1975a). To examine such ques- sity, CollegeStation. tions it is necessary to determine how species GOLDBERG, N. H. 1979. Behavioral flexibility and respond behaviorally to variation in the envi- foragingstrategies in Cassin's and Westernking- ronmental abundance. birds(Tyrannus vociferans and T. verticalis)breed- parametersaffecting prey ing sympatricallyin riparianhabitats in central For temperate zone kingbirds I believe this is Arizona. Ph.D.diss., Univ. of Illinois at Urbana, weather, and secondarilyhabitat. Expectations Champaign. are that specialists will show little variation in GREIG-SMITH, P. W. 1983. Use of perches as vantage behavior as weather but points duringforaging by male and female Stone- foraging changes, gen- chats Saxicola Behaviour eralistslike the will torquata. 86:215-236. EasternKingbird show high HEINRICH,B. 1981. Insectthermoregulation. John H. foragingflexibility. Wiley and Sons, New York. HESPENHEIDE,H. A. 1971. Food preferenceand the extentof overlapin some insectivorousbirds, with ACKNOWLEDGMENTS special referenceto the Tyrannidae.Ibis 113:59- 72. Fieldwork was supported by fundsfrom the Alexander HESPENHEIDE,H. A. 1975. Selectivepredation by two E.Bergstrom Research Award of theNortheastern Bird swiftsand a swallowin CentralAmerica. Ibis 117: BandingAssociation, the FrankM. ChapmanFund of 82-99. the AmericanMuseum of Natural and a History, by HOLMES, R. T., T. W. SHERRY, AND S. E. BENNETT. SummerFellowship from the Universityof Kansas 1978. Diurnal and individual in GraduateSchool. NSF grant BSR 830065 to Dr.George variability the S. Bakkenof IndianaState University provided me foragingbehavior of American Redstarts (Seto- with funds I phaga ruticilla).Oecologia (Berl.) 36:141-149. while preparedthe finalversion of this G. 1987. Time and constraints manuscript.Comments by PeterStacey, an anony- JONES, energy during mous and Tom im- incubation in free-livingswallows (Hirundorus- reviewer, especially Sherrygreatly tica): an experimentalstudy using precision elec- provedthe manuscript. tronic balances.J. Anim. Ecol. 56:229-245. KARR, J. R., AND F. C. JAMES. 1975. Eco-morpho- LITERATURE CITED logicalconfigurations and convergentevolution in species and communities, p. 258-291. In M. L. D. AND BEAVER, L., P. H. BALDWIN.1975. Ecological Cody and J. M. Diamond [eds.],Ecology and evo- and the of and overlap problem competition sym- lution of communities.Belknap Press, Cambridge, patryin the Westernand Hammond'sflycatchers. MA. Condor 77:1-13. G. 1983a. A. C. 1942. Life histories of North American KINGSOLVER,J. Thermoregulationand flight BENT, in Colias butterflies: elevational patterns and flycatchers,larks, swallows, and their allies. U.S. Natl. Mus. Bull. 179. mechanisticlimitations. Ecology64:534-545. BLANCHER,P. J., AND R. J. ROBERTSON.1984. Re- KINGSOLvER, J. G. 1983b. Ecological significance of source use by sympatric kingbirds. Condor 86: flightactivity in Coliasbutterflies: implications for 305-313. reproductive strategy and population structure. BLANCHER,P. J., AND R. J. ROBERTSON.1985. A com- Ecology64:546-551. parison of EasternKingbird breeding biology in LACK, D., AND E. LACK. 1951. The breeding biology lakeshoreand upland habitats. Can. J. Zool. 63: of the Swift, Apus apus. Ibis 93: 501-548. 2305-2312. LANDRES, P. B., ANDJ. A. MACMAHON. 1980. Guilds BRYANT, D. M. 1973. The factors influencingthe and community organization:analysis of an oak selection of food by the House Martin (Delichon woodland avifauna in Sonora, Mexico. Auk 97: urbica[L.]). J. Anim. Ecol. 42:539-564. 351-365. DAVIES,N. B. 1977. Prey selection and the search LECK, C. F. 1971. Some spatial and temporal di- strategy of the Spotted Flycatcher (Muscicapa mensionsof kingbirdforaging-flights. Wilson Bull. striata):a field study on optimal foraging.Anim. 83:310-311. Behav. 25:1016-1033. MAHAN,H. D. 1964. The effect of the environment DICK,J. A., ANDJ. D. RISING. 1965. A comparison on growth, development, and temperature regu- of foods eaten by EasternKingbirds and Western lation in nestling Eastern Phoebes. Ph.D.diss., Kingbirdsin Kansas. Bull. Kansas Ornithol.Soc. Michigan State Univ., East Lansing. 16:23-24. MORENO,J. 1984. Search strategies of Wheatears 730 MICHAELT. MURPHY

(Oenanthe oenanthe) and Stonechats (Saxicola STAPANIAN, M. A. 1982. Evolution of fruitingstrat- torquata):adaptive variation in perchheight, search egies among fleshy-fruited plant species of eastern time, sally distanceand inter-perchmove length. Kansas. Ecology 63:1422-1431. J. Anim. Ecol. 53:147-159. TATSCHL,J. L. 1973. Niche exploitationamong three MURPHY,M. T. 1983. Clutch size in the Eastern closely related flycatchers in Kansas. Ph.D.diss., Kingbird:factors affecting nestling survival. Auk Kansas State Univ., Manhattan. 100:326-334. TRAYLOR, M. A., JR., AND J. W. FITZPATRICK. 1982. MURPHY,M. T. 1986. Temporalcomponents of re- A surveyof the tyrantflycatchers. Living Bird 19: productivevariability in EasternKingbirds (Ty- 7-50. rannustyrannus). Ecology 67:1483-1492. VERBEEK, N.A.M. 1975a. Northernwintering of fly- O'CONNOR,R. J. 1977. Growthstrategies of nestling catchers and residency of Black Phoebes in Cali- .Living Bird 16:209-238. fornia.Auk 92:737-749. O'CONNOR,R. J. 1979. Egg weights and brood re- VERBEEK, N.A.M. 1975b. Comparativefeeding be- ductionin the EuropeanSwift (Apus apus). Condor havior of three coexisting tyrannid flycatchers. 81:133-145. Wilson Bull. 87:231-240. O'CONNOR,R. J., ANDR. A. MORGAN.1982. Some VIA, J. W. 1979. Foragingtactics of flycatchersin effects of weatherconditions on the breedingof southwesternVirginia, p. 191-202. In J. G. Dick- the Spotted FlycatcherMuscicapa striata in Brit- son, R. N. Connor,R. R. Fleet, J. A. Jacksonand ain. Bird Study 29:41-48. J. C. Kroll [eds.], The role of insectivorousbirds PINKOWSKI,B. C. 1977. Foragingbehavior of the in forestecosystems. Academic Press, New York. EasternBluebird. Wilson Bull. 89:404-414. VISSCHER, P. K., AND T. D. SEELEY.1982. Foraging PYKE,G. H., H. R. PULLIAM,AND E. L. CHARNOv. strategyof honeybee colonies in a temperatede- 1977. Optimalforaging theory: a selective review ciduous forest. Ecology 63:1790-1801. of theory and tests. Q. Rev. Biol. 52:137-154. WIENS, J. A., ANDJ. T. ROTENBERRY. 1979. Diet niche RICKLEFS,R. E., ANDG. W. Cox. 1977. Morpholog- relationshipsamong North American grassland and ical similarityand ecologicaloverlap among pas- shrub-steppe birds. Oecologia (Berl.) 42:253-293. serinebirds on St. Kitts,British West Indies. Oikos ZAR, J. H. 1974. Biostatistical analysis. Prentice-Hall, 29:60-66. Englewood Cliffs, NJ. SHERRY,T. W. 1984. Comparativedietary ecology of sympatricinsectivorous Neotropical flycatchers (Tyrannidae).Ecol. Monogr.54:313-338.