in Hemignathus virens on Mauna Kea, Hawaii. I observed ... population on Mauna Kea, did not occur if adequate food supplies remained available during.
THE
INFLUENCE
NESTING
OF NECTAR
SUCCESS
OF THE
AND
RESOURCES
MOVEMENT
COMMON
ON
PATTERNS
AMAKIHI
(HEMIGNATHUS
VIRENS)
CHARLES VAN RIPER III
Departmentof ZoologyandCooperative NationalPark Resources StudiesUnit, Universityof California,Davis,California95616USA
ABSTRACT.--The objectivesof this study were to investigate the influence of nectar resourceson: (1) territory establishment,(2) breeding success, and (3) postbreedingdispersal in Hemignathus virenson Mauna Kea, Hawaii. I observedcolor-bandedindividuals from 1971
to 1981 and manipulatedfood during the 1975 breeding seasonby placingsupplementary nectar resourcesin an area where breeding had not occurredin the previous4 yr. The results of this study demonstratedthat H. virensis intrinsically tied to the nectar resourcesavailable within its breeding territory and that a minimum of 2,000 •l/day (1,400 cat) of nectar must be present before territory establishmentwill occur. This finding was further supportedby the fact that no breedingoccurredin the experimentalareaduring the 6 yr after removal of the supplementarynectar. Breeding successof Hemignathus virenswas influenced by different levels of nectar availability. The mostsuccessful breedingoccurredat nectarlevelsof 20,000-40,000•t/day (14,00028,000 cat). More than 35,000 •t/day (24,500 cat) of nectar within a territory generally inhibited successful breeding,presumablybecausetoo much time was spentin the defenseof the nectar sources.Birds with larger amountsof nectarwithin their territories had concomitantty larger clutch sizes.Postbreedingdispersal,an annual phenomenonin the H. virens populationon Mauna Kea, did not occurif adequatefood suppliesremainedavailableduring the nonbreeding period. Received 30 November1982,accepted 6 September 1983. dition of nectar resources, birds were induced
AVIAN ecologistsoften speakof food-limited populations, but evidence of competition for a
to nest in an area that had not been
used for
specificfood supply during the breeding sea- breeding in the previous 4 yr; and (2) obserson is meager (Stenger 1958, Walsberg 1977). vations of nesting successand postbreeding There seem to be two reasons for this lack of dispersalas a function of the amount of nectar evidence: (1) Mere observations on nesting presentwithin each territory. successand food supply are likely to be equivMETHODS ocal. A multitude of factorsaffect nesting success as well as one another, and this interde-
pendence of factors often obfuscatesresults. Avian nectivores,however, provide an unusual casein which individuals are largely dependent upon only one easily measurable resource.(2) It is difficult to changeone, and only one, factor. Occasionally,it is possibleto conduct an experiment that demonstratesthe relationship between a single ecological parameter and an organism's behavior. During my
studyof the CommonAmakihi (Hemignathus virens) such a situation arose, which enabled me
to examine the effect of nectar resourcesupon territory establishment, breeding success,and postbreedingdispersal. ! shall present results from two kinds of studies:(1) an experiment in which, by the ad-
Studyareaand vegetation.--Thestudy site was a 25ha plot at 2,130m elevationon the southwesternslope of Mauna Kea, Hawaii (Fig. 1). It was locatedin one of the last extensivedry forestsin Hawaii, a high montane savanna dominated by two tree species,So-
phora chrysophylla(Leguminosae) and Myopo•um sandwicense (Myoporaceae).
Sophora is by far the most important nectar source for H. virenson Mauna Kea and will be the only one consideredin this paper. The phenology of Sophora has been studied on Hawaii by Baldwin (1953), Lamoureux et al. (1981), and van Riper (1975b, 1980). Most flowering is in November through June, but a few widely scatteredtrees flower in other months. Flowering on Mauna Kea was greater in 1974 than in either
1973 or 1975.
Sophoraflowers occur in terminal clustersof about 14 (œ= 13.78, SE = 1.22). Each flower is approximate-
38
The Auk 101:38-46. January1984
January 1984]
39
Nectar Influenceson CommonAmakihi
i
MAUNA KEA GAME MANAGEMENT AREA
',.iU,AAU CA,,N IKAUA, . . .... . .I N •'
)
•.
•
• I
•'N • N, F///////////////A
I
/
MOLOKAI' / ' LANA[•• •
I /
/ I • •AWAiI / HAWAIIANISLANDS
//
/
STUDY PARKER
,
I
/
)
•
o•5
?
i.,o km
015 miles
RANCH
,
•.,
• I i
KAOHE GAME MANAGEMENT AREA
'--'•/JEEP ROAD (;
/ PUU [AHU MOA,
Fig. 1. Map of the studysite on the southwesternslopeof Mauna Kea, Hawaii.
ly 25 mm long, with four elongated,curved yellow petals.Nectaris producedand storedat the baseof
2.25/•1/cluster, SE = 0.25). The averagecluster, there-
fore, produced33.81/d of nectarover a 24-h period. During 1975, flower abundancethroughout 45 H. asynchronously,and approximatelythree (• = 2.77, virensbreeding territories was measuredat weekly intervals over the nesting cycle (Table 1). Full flow= 0.13, range = 0-8)produce nectar concurrently. Eachflower producesnectarfor 3-5 daysafter open- ering trees averaged 244 flower clusters(SE = 35.3, ing and may remain on the tree for up to 7 daysafter range = 123-460 clusters),while medium flowering treesaveraged54 flower clusters(SE = 194, range = cessationof nectar production. Nectar production was measuredby bagging 19 17-103 clusters). The amount of daily nectar (Nd) clustersconsistingof 244 flowers on 6 trees. Nectar available to each breeding pair of H. virenswas calwasextractedwith micropipettesat first light (0630), culatedusing the formula: midday (1300.),and sunset(1830).This procedurewas Nd = (C,)(F,)(N•)(T), repeatedfor -the entire nectar-producingperiod of the corolla. Individual flowers within a cluster open
•E
each cluster (range 3-10 days). Nectar concentration was measured with an Itago hand-refractometer for 109 samples, and sugar concentrations averaged 19.21% (SE = 0.68, range = 5.8-53.7%). Nectar was
producedin eachclusterat an averagerate of 2.63
where C,= the average number of flower clusters presentat any one time in the territory over the nesting cycle, Fi = the averagenumber of flowers/cluster, Nc= the averageamount of nectar produced by
/d/h (SE = 0.40,range = 1.70-3.36/d) during the day. Very little nectar was produced over the night (•' =
a flower cluster each hour, and T = the time period throughout 24 h when nectar is produced. The av-
40
CHARLES VANRIPER III
LATE APRIL21:50m
[Auk,Vol. 101
1975
NESTS (• 0
IOO
200 50
:500 I00
400
500 feet 150 meters
UNOCCUPIED
DISPUTED Fig. 2. Map of Hemignathus virensbreeding territories during April 1975 at 2,130-m elevation on Mauna Kea, Hawaii. Circled letters are the 3 supplementalfeeding stations;asterisksare the 3 control feeding stations.The cross-hatched areais the experimentallyinitiated territory.
erageH. virensbreeding territory contained4.6 flowering trees (SE = 0.68, range = 1-9 trees). The bird.--Hemignathus virensis a member of the endemic Hawaiian honeycreepers(Drepanidinae). A number of investigatorshave studied its breeding, ecology, and behavior (Baldwin 1953; Berger 1969; Kamil 1978; Kamil and van Riper 1981; van Riper 1975a, b, 1976, 1977, 1978; Kern and van Riper in prep.). It is the most abundant bird in the dry forest on Mauna Kea (van Riper et al. 1978). The diet consistsprincipally of nectarand insects,although it will take various fruits. Territories are mutually exclusive and are reoccupiedeach year by the same individuals. Betweenbreeding periodsthe territorial structurebreaksdown into a home-rangesystem,and most birds undergo postbreedingdispersal,respond-
ing to temporal changesin flowering Sophora. A few pairs may remain on or near their breeding territory but occupy an area that encompasses3-6 times as much space.Thesebirds subsistchiefly upon insects, and conspecificsare permitted to forage within the home range. From 1971through 1981! capturedover 1,000birds by mist net and banded them with aluminum Fish and Wildlife Servicebandsin combination with plastic color bandsfor subsequentindividual identification in the field. I also closelymonitored the population in terms of territory boundaries, movement patterns, and breeding success.
The experiment.--Immediatelyprior to the 1975 breeding season,within an area in which H. virens had not attempted to nest during the previous 4 yr,
NectarInfluences on CommonAmakihi
January 1984]
41
TABLE1. Potential Sophorachrysophylla nectar sup- consistedof three 1,000-ml nectar feeders hung on plies and breedinginformationfrom Hemignathusbright yellow boards.The yellow simulated Sophora virensterritories during 1975, at 2,130 m elevation
flower color and served to attract birds. A control,
on Mauna Kea, Hawaii.
which
Sophora nectar
Territory number
produced daily (/•1)
5-75
32,001
12-75 16-75 17-75 18-75 19-75 21-75 22-75 23-75 24-75
21,938 32,001 21,938 54,924 26,554 32,001 23,755 18,308 20,124
25-75 26-75 27-75 28-75
28,370 18,308 49,478 9,078
29-75 31-75 32-75 34-75 35-75 36-75 37-75
1,816 13,693 28,370 27,386 5,447 21,938 28,370 3,631 20,124 11,878
38-75 39-75 40-75 42-75 b
0
43-75
18,308
44-75 45-75 48-75 50-75 51-75 52-75 53-75
38,412 11,878 5,447 11,878 23,755 3,631 3,631
54-75 55-75 56-75
16,492 10,062 20,124
57-75 58-75
26,554 10,062
59-75 60-75 65-75 66-75
10,062 11,878 13,693 8,246
70-75 71-75
20,124 20,124
75-75
10,062
Clutch size
Number fledged
3
0
(Eggsbroken) 3 2
0 1
3 0
3
2
2
2 3 0 3 2 2 3
0 0
consisted of the same number
of boards and
empty feeders,was establishedin another area (Fig. 2). The nectar feeders were filled with 20% sugar water, approximately the same sugar concentration as Sophoranectar. Feederswere filled on 10 January 1975,well before breeding commenced(the first active nest was found on 6 February),and were maintained daily into the period of postbreedingdispersal (10 June 1975). Observationswere made during 24 sessions, each with a duration of not less than 1 h;
using stopwatch,binoculars,and a tape recorder, I spent a total of 49 h observingbird use at each feeder from
a blind.
RESULTS
3
0 2 2 2 2
2 2
0 1
3
0
2 2 2 3
0a 0 2 1
2
3
1 0
2
2
OP-X, were first observedusing the feederson 21 January; WBk-X concentrated his efforts around Station A, while OP-X frequented the feedersat B. Neither bird had bred in the study area during 1974. Station C was incorporated into the adjacent territory of X-G, an estab-
2 2 2
0 0 0
H. virensfemalesX-PY and X-YY consistently usedthe feedersthroughoutearly February.
2 3
0 0
2
2
(Not followed) 2 (Fate unknown)
Nest predation; not included in calculations. Experimental feeder nest.
Experimental improvement of a territory.--Territory establishmenthad not been observed in the experimentalarea for a period of 4 yr preceding 1975.Shortly after initiation of the supplemental "nectar" supplies, birds began to
frequentthe area.The majority were H. virens, but the introduced Zosterops japonicus also used feeders throughout the experimental period. Two 22-month-old
H. virens males, WBk-X and
lished resident male. In addition to these birds,
By late FebruaryWBk-X and both femaleshad disappeared;OP-X had taken over sole use of stationsA and B. This male paired with WYYX, a young bird whose complete life history wasnot known. The pair maintaineda territory around these stations throughout the entire breeding seasonand successfullyfledged two young. No birds nestedat the controlsite. Birds were not observednesting in this experimental area during the subsequent6 yr following removal of the supplemental"nectar" supplies. The added
food resources
had little
or no ef-
fect upon other establishedH. virenspairs within the study site. Contiguous territorial boundI establishedthree supplementaryfood stations(Fig. 2). Each station was placed far enough from known aries, as they were delineated in the previous territorial boundaries and from the other two stabreeding season,did not change appreciably tions (average51.8 m) that birds had an adequatearea during 1975. Only the territory of male X-G in which to establisha territory. Eachfeeding station was slightly enlarged to encompassfeeding
42
CHARLES VANRIPER III
Station C. At no time during the breeding season did I record other territory holders from within the studyareaat any of the feeders.The
majorityof birds(otherthan the residentpairs) recorded within the experimental area were yearling floaters(143 of 156 feeding observa-
[Auk, Vol. 101 -2.0
1.8
14-
tions).
Theinfluence of Sophoranectarabundance upon breeding success.--The mean daily nectar available within a territory influenced breeding Success, including the maintenanceof a territory, clutchsize,and the numberof youngproduced. Each H. virensterritory that I recorded in 1975 had Sophoraflowering within its confines exceptthe one with supplementalfood supplies(Table 1). The minimum daily nectar requirement necessaryfor the maintenanceof a breedingterritoryappearsto be 2,000td (1,400 cat).Exceptfor territory 29, however, it appears that in order to raiseyoung successfully, a pair
.2
requires atlea'st 9,000 td(6,300 cal)ofdailynectar. In most of the territories
30,000 td (7,000-21,000 cal) of Sophoranectar produced daily, but production in those territories that were most successfulin fledging
youngrangedfrom 20,000to 40,000td (14,00028,000cat;Fig. 3). It appearsthat birds are unable to defend
more nectar than this and still
successfullyfledge young. Indeed, when nests were dividedinto categories of success, the only
pairswith a disproportionate differencein the amountof daily nectarproducedwithin their territorial
o
there was 10,000-
IO,OOO 2o,ooo
3o,ooo 4o,ooo
5o,ooo
/.•1,OF NECTARPRODUCEDDAILY Fig. 3. The shadedbar graph depictsthe relationship between the number of Hemignathusvirensterritories established and the amount of Sophorachrysophyllanectar containedwithin each territory (n = 45). The solid line is the average number of young producedwithin territoriesin relation to the amount of Sophora chrysophylla nectarwithin thoseterritories (n = 4]).
boundaries were those that did not
successfully fledgeany young(ANOVA,df = 39, F = 4.51, P < 0.01; Fig. 4). An example of a "resourceoverload" situation in H. virensis the nesting history of male
this pair built nest#36-75 in the sameterritory, but this year an averageof only 22,000td/day
(15,400 cal) of Sophoranectar was available throughout the entire territory (Table 1). The the 1974 breeding season,this pair built nest pair again had a clutch of two eggs,but this #132-74 in an area of their territory in which time successfullyfledged one young. The yearly mean clutch size of H. virenshas five closelyspacedSophora treessubsequently cameinto heavy flower. At the height of this been shown to be influenced by Sophoranectar flowering, there was over 37,000 td of nectar production(van Riper 1978).The resultsof this being producedeachday just in the immediate experimentsupportmy earlier findings,in that vicinity of the nest.This pair of birds spentan clutch size varied in relation to the nectar inordinate amount of time chasing intruders. available to each pair. Birds in territories with Throughoutegg laying and early incubation more than 30,000 td/day (21,000 cat) of nectar periods,the birds averaged10.2chases/h(SE= producedlarger clutches(Fig. 5). 1.81), as comparedto the mean chasingrate of Theinfluence of Sophoranectaruponpostbreed1.99 chases/h (SE = 0.22) of the Mauna Kea H. ing dispersaL--ByMay 1975, the majority of H. virenspopulation during the breeding season virenshad left the 2,130-m study area in a post(van Riper 1978).Their clutch of two eggswas breeding dispersal.This movement pattern is not incubated,presumablybecauseof the dis- an annual occurrence and coincides with the flowering (van Riper 1978). traction causedby intruders. In 1975,however, cessationof Sophora X-OR
and female
WP-X
in 1974 and 1975. In
January1984]
NectarInfluences onCommon Amakihi
43
60,000
50,000 n• n• W
40,000
W
30,000
20,000
Z
I0,000
ß
NEST ONLY
EGGS NONE SOME ALL ONLY FLEDGEDFLEDGEDFLEDGED
CATEGORIES
OF SUCCESS
Fig.4. Success categories of Hemignathus virens nestsin relationship to theamount of dailySophora chrysophylla nectarwithinthatterritory.Triangles arethemeannectarproduction for eachgroup,andlinesare standard
errors.
For example,during the peak breeding period (April), over 10% of the Sophorabranches throughout the 2,130-m study area bore flowers, while in May lessthan 2% were in flower (van Riper 1980). The birds apparently move around
Mauna
Kea in search of nectar.
The experimental area was the only portion of the 2,130-m study area in which I recorded H. virensin any numbersafter the postbreeding exodus. The pair associatedwith the experimental feeders, along with their young, re-
time birdswould line up on the yellow boards waiting to feed. Although the feeders were an artificial situation that resulted in a more intense use of an
areathan might normally occur,the feeder experiment indicates what happens when food resources become available
to H. virens in the
nonbreeding season.In another area of the 2,130-m study site during 1973-1974, I recorded a banded pair within their territory for 15 consecutivemonths.Within the original breedmained until June, when the feeders were reing territory, two Myoporumsandwicense trees moved. After the fledging of their young (22 were heavily fluxing sap,and this resource,toApril), defense of the feederswaned, until by gether with the insectsit attracted,was apparmid-May there was little territorial aggression ently enough food so that the birds and their observed. During this latter period I recorded young could remain in the areathroughout the for the first time instances of resident H. virens entire nonbreeding period. Paton (1980) has rebreedersfrom other areasof the 2,130-mstudy cently describedthis samephenomenon occurarea, as well as a large number of recently ring in a number of honeyeaters(Meliphagifledged young, using the feeders.During this dae) in Australia. The postbreedingdispersal
44
CHARLES VANRIPER III
of this pair of H. virenswas obviously suppressedby the presenceof a copiousfood supply, as was that of the pair associatedwith the experimental feeders.
[Auk,Vol.101
:5.0 2.9 2.8 2.7
DISCUSSION
•
The resultsof this studydemonstratethat the H. virenspopulation on Mauna Kea is influenced by Sophoranectar resourcesin several respects:(1) territory establishment will not occur
unless
a minimum
amount
of
nectar
is
2.6,
N
•
2.5
-1-
,o
2.4
I-
• 2.:5 2.2
available within an area; (2) breeding effort is most successful
in territories
that
contain
nec-
tar within an optimum range, balanced between the minimum amount necessaryto produce the largest possible clutch size and an upper limit determined by what can be economically defended; and, (3) postbreeding movement,an annual phenomenonin the population, need not occur if adequate food supplies are made available in the nonbreeding season.
Territoryestablishment.--Hemignathus virenson Mauna
Kea defend
strict territorial
in terms of time
and energy to defend a territory. On Mauna Kea this level appearsto be about2,000•1 (1,400 cal) of daily Sophoranectar, becauseonly after nectar
resources
rose above
2.0
I I0,000
I I 20,000:50,000
I 40,000
>40,000
p..1,OF NECTAR PRODUCEDDAILY Fig. 5. The relationship between Hemignathus virensclutch size and the amount of daily nectar available within a territory (n = 26). Triangles are the means; lines are the standard errors.
boundaries.
Brown (1969) believed that only where the density of food resourcesexceeded a certain level did it become economical
2.1
this threshold
was
a territory established.The daily nectar available to breeding H. virensis considerably less than has been reported for other nectarivorous species.Gill and Wolf (1975) found that Golden-winged Sunbirds (Nectariniareichenowi)defended a feeding territory only when there was
form an important componentof H. virens'diet, particularly when they are raising young (Baldwin 1953, van Riper 1978). Nevertheless, a certain quantity of nectar appearsto be a prerequisite within an area for successfulbreeding.
Breedingsuccess.--Hemignathus virenspairs in territories with very low and very high quantities
of nectar
were
often
not
successful
in
fledging young. Defense of a greater food sourcewould seemto require a disproportionate amount of time and energy. The experifrom 5,500 to 13,375 •1 (3,850-9,712 cal) of nec- mental nectar feeders provided an unlimited tar available per day. A similar situation exists amountof daily "nectar,"but, as they were sitwith Selasphorus rufusin easternArizona (Ko- uated on the oppositeside of the territory from dric-Brown and Brown 1978). Carpenter and the nest, they were easilydefended. Therefore, MacMillen (1976) reported nectar quantities the amount of nectar being defended appears from six feeding territories of nonbreeding in most casesnot to be the upper limiting facVestiariacoccinea(Drepanidinae), and all were tor of successful breeding,but rather the ability greater than what I recorded for H. virens of the bird to defend the nectar sources. breeding territories. Perhaps, as Hainsworth The range of most successfulbreeding oc(1977) showed in Colibri coruscans,nectar for H. curred in territories in which the daily Sophora virenson Mauna Kea is important as a primary nectar production was between 20,000 and food source becauseof its stationary predict- 40,000•1 (14,000-28,000cal).Thiswasthe range able nature rather than the proximate ability of of maximum fitness, determined on the one the birds to exploit nectarmore efficiently than hand by clutch size (pressureto produce the insects. It must also be remembered that insects largest clutch and therefore most potential
January1984]
NectarInfluences onCommon Amakihi
young) and on the other by the ability of the pair to defend as large an area as possibleand still be able to raiseyoung (pressureto occupy a territory that was easilydefensible).Lessthan half (20 of 44 pairs)of the population defended an optimal nectar supply. The H. virensbreeding on Mauna Kea have contiguousterritories (Fig. 2), and it may be that the large numberof suboptimal territories are a result of differential social dominance forcing certain individuals to contract
their
territories
below
the oth-
45
anna). Kodric-Brown and Brown (1978) were
able to keep migrant Rufous Hummingbirds (Selasphorus rufus)on territories with the addition of an artificial nectarsupply. Pimm (pers. comm.) has demonstrated in southeastern Ari-
zona that with a constantlya-vailablefood supply, three speciesof hummingbirdswill forego seasonalmovements.It thus appearsthat nectar levels control seasonalmovementpatterns in many nectarivorousspecies,and this is particularly true of H. virenson Mauna Kea.
erwise optimal size, as suggested by Hixon (1980).
CONCLUSIONS
There is someevidence to suggestthat elsewhere food may not be as critical to H. virens as it is on Mauna Kea. Eddinger (1970) found that on Kauai, H. virens territories
consisted of
a small circle around the nest, usually extending to a radius of 5.5 m. Baldwin (1953) worked with this specieson Mauna Loa, Hawaii and reported that territories were loosely held and often shifted within limits. Perhaps H. virens usesdifferent territorial strategiesin different habitats, as does Phainopeplanitens(Walsberg 1977).
Postbreeding tnovement.--Movementpatterns have been linked to changesin food availability for a numberof nectarivorousspecies(Skead 1967;Gill and Wolf 1975;Carpenter 1976, 1978; Carpenter and MacMillen 1976). Stiles (1973) felt that as a group, nectar feedersare the most
nomadic of small land birds. Movement patterns in the nectarivorousguild of the Drepanidinae have been linked with changes in flowering on Hawaii (Baldwin 1953, MacMillen and Carpenter 1980). I have also found that the end of seasonalSophoraflowering on Mauna Kea occursconcurrently with H. virens postbreedingdispersal,at which time the birds
move around the mountain seekingpatchesof flowering trees(van Riper 1978).In fact, a reasonablegeneralizationbasedupon thesevaried accountsis probably that avian specieswhose life habitsinclude relianceupon a nectarsource demonstratemovementpatternsin one form or another in at least some period of their annual cycle. As has been demonstrated in this study,
The resultsof this study demonstratethat the Hemignathus virenspopulation on Mauna Kea is intrinsically tied to its food supplies.It is also evident that a lack of food, in this caseSophora nectar,was probablythe limiting factor that, in the 4 yr previousto and the 6 yr following the experiment, had prevented this speciesfrom breedingwithin the experimentalarea. Further credenceis lent to this argument by the facts
that the artifically stimulatedterritory was the only one of 45 that totally lacked Sophoraflowering throughout the 1975breeding seasonand that the pair associatedwith the supplemental food was able to fledge two young successfully in a year of lower than normal fledgling productivity. Nectar resources also influenced breeding success in that the birds respondedto differentlevelsof this food supply.The lowest level (approximately 2,000 /•l/day; 1,400 cal) appearsto be essentialfor territory establishment but in most casesis not adequatefor successfulbreeding. The next threshold is reached at approximately 10,000 /•l/day (7,000 cal), which
seems to be the minimum
amount
nec-
essaryfor feeding young. The maximum nectar level is reached abruptly at 35,000 /•l/day (14,500 cal), after which it appearsthat successfulbreeding cannotusually take place,presumablybecausea disproportionateamount of time has to be spent in defense.Finally, postbreeding dispersalwas found to be influenced by the amount of food resourcesavailable within a region.
however, these movements can be overridden ACKNOWLEDGMENTS
if food suppliesremain available at a fixed location. Stiles (1973) found that in California
many of the recently introduced nectar producing plants have curtailed movementpatterns in the Anna's Hummingbird (Calypte
I thank F. L. Carpenter,A. Kamil, A. Kodric-Brown, R. MacMillen, S. Pimm, C. J. Ralph, S. G. van Riper, and L. Wolf
for their
comments
on various
drafts of
this paper. Alan Kamil helped to collectnectar data.
46
CHARLES VANRIPER III
[Auk,Vol. 101
C. Barnesand M. Murphy typed the manuscript.Fi- LAMOUREUX, C. H., D. MUELLER-DOMBOIS, & K. W. BRIDGES.1981. Temporal variation of organism nancial supportfor the field phaseof this research was provided by World Wildlife Fund Grant US-35, groupsstudied:treesßPp. 391-407 in Island ecosystems: biological organization in selected the Frank M. ChapmanMemorial Fund, and Earthwatch.
Contract
CX 8000 7 0009 from
the National
Park Service in conjunction with a National GeographicSocietyGrant provided partial supportduring preparationof the manuscript.
Hawaiian communities (D. Muller-Dombois, K.
W. Bridges,and H. L. Carson,Eds.).Stroudsburg, Pennsylvania, Hutchinson RossPubl. Co. MACMILLEN, R. E., & F. L. CARPENTER. 1980ßEvening
roostingflightsof the honeycreepers Hirnatione LITERATURE CITED
sanguineaand Vestiariacoccineaon Hawaii. Auk 97: 28-37.
BALDWIN,P. H. 1953ß Annual cycle, environment PATON,D.C. 1980. The importanceof manna,honand evolution in the Hawaiian honeycreepers eydewand lerp in the dietsof honeyeaters. Emu 80: 213-226. (Aves:Drepaniidae). UnivßCalifornia Publ. Zool. 52: 285-398.
BERGER, A. J. 1969. The breeding season of the Hawaii Amakihi. Occ. Pap. Bernice P. Bishop Mus. 24: 1-8.
BROWN, J. L. 1969ß Territorial behavior and population regulation in birds. Wilson Bull. 81: 293329.
SKEAD,C.J.
1967ß The sunbirds of Southern Africa.
Capetown, South Africa, Balkema. STENGER, J. 1958. Food habits and available food of
Ovenbirdsin relation to territory sizeßAuk 75: 335-346.
STILES, F.G. 1973. Foodsupplyand the annual cycle of the Anna Hummingbird. Univ. California
Publ. Zool. No. 97. CARPENTER, F.L. 1976. Plant-pollinatorinteractions in Hawaii: pollination energeticsof Metrosideros VAN RIPER, C., III. 1975a. Parasites of the Hawaii collina(Myrtaceae).Ecology57:1125-1144. Amakihi (Loxops virens virens). US/IBP Island 1978ß A spectrum of nectar-eatercommuEcosystems IRP Tech. Rept. No. 62. nitiesß Amer. Zool. 18: 809-819. 1975b. The phenology of mamane (Sophora , & R. E. MACMILLEN. 1976. Threshold model chrysophylla) and naio (Myoporurnsandwicense) on of feeding territoriality and testwith a Hawaiian Mauna Kea,Hawaii, asrelatedto the annual cycle of the Amakihi (Loxopsvirens)and Pallia (Psittihoneycreeper.Science194:639-642. EDDINGER, C.R. 1970ßA study of the breeding birostra bailleui).US/IBP Island EcosystemsIRP TechßRept. No. 51. ology of four speciesof Hawaiian honeycreepers (Drepanididae).Unpublished PhßD.disser1976. Superimpositionof an Amakihi nest tation. Honolulu, Hawaii, Univß Hawaii. on one of an Elepaio.Condor 78: 267-269ß GILL,F. B., & L. L. WOLFß1975. Economicsof feeding 1977. The use of sheepwool in nest conterritoriality in the Golden-winged Sunbirdß struction by Hawaiian birds. Auk 94: 646-651. Ecology56: 333-345. ß 1978ßThe breedingecologyof the Amakihi (Loxopsvirens)and Pallia (Psittirostrabailleui)on HAINSWORTH, F. R. 1977. Foraging efficiencyand parentalcarein Coribricoruscans. Condor79: 69Mauna Kea, HawaiißUnpublished PhßD.disser75.
tation. Honolulu, Hawaii, Univ. Hawaii.
HIXON,M.A. 1980. Food productionand competi1980. The phenologyof the dryland forest tor density as the determinants of feeding terof Mauna Kea,Hawaii, and the impactof recent environmentalperterbations.Biotropica12:282ritory size.Amer. Natur. 115:510-530. 291. KAMIL,A.C. 1978. Systematicforagingby a nectar, J. M. SCOTT,& D. M. WOODSIDE. 1978. Patfeedingbird, the Amakihi (Loxops virens).J.Comp. terns in the distribution and abundance of the Physiol. Psychol.92: 388-396. Palila on Mauna Kea, Hawaii. Auk 95: 518-527. , & C. VAN RIPER,III. 1981. Within-territory division of foraging spaceby male and female WALSBERG, G. E. 1977. Ecology and energeticsof Amakihi (Loxops virens).Condor 84: 117-119. contrastingsocialsystemsin Phainopeplanitens KODRIC-BROWN, A., & J. H. BROWN. 1978ß Influence (Aves: Ptilogonatidae). Univß California Publ. of economics,interspecificcompetition, and sexual dimorphism on territoriality of migrant Rufous Hummingbirds. Ecology59: 285-296.
Zool.
No.
108.
