PATTERNS
IN THE
NESTLING
PROVISIONING
RHINOCEROS
AND
GROWTH
OF
AUKLETS
D. F. BERTRAM,• G. W. KAISER,2 AND R. C. YDENBERG
Behavioral
EcologyResearch
Group,Departmentof Biological
Sciences,
SimonFraserUniversity,
Burnaby,BritishColumbiaV5A 1S6,Canada
ABSTRACT.--We
studied (1983-1987) the provisioning and growth of nestling Rhinoceros
Auklets (Cerorhinca
monocerata)
in a breedingcolonyon the LucyIslandsand measuredgrowth
(1984-1986) in other coloniesalso on the British Columbia coast(Pine Island and Triangle
Island). There were strong differencesin the pattern of provisioning among the three years.
In 1985 provisioning peaked around the midpoint (30 days) of the nestling period and then
declined rapidly. In 1986, provisioning also declined after the midpoint, but more slowly
than in 1985. In 1987, provisioning remained approximately constant during the entire
breedingseason.We found no evidencethat late-hatchedchickswere consistentlyfed less
than early-hatchedchicks,though the statisticalpower of the testsmade to detect this difference was low. There were significant annual differencesin chick growth. Growth was
fastestin 1985,intermediate in 1984and 1986,and slowestin 1983and 1987.The mean growth
rate on the three colonieschangedin unison from year to year. Hypothesesbasedon seasonal
environmental changesand on systematicchangesin the food requirements of growing
chicksdo not explain the observedpatterns.Our data on chick growth and independentdata
on oceanproductionsuggestthat the two varied directly and in unisonamongyears.However,
provisioning in the latter part of chick development did not appear to reflect directly the
quality of oceanfeeding conditions.We concludethat although large-scalefluctuationsin
oceanproductionare likely to be the dominant influence upon provisioning,the manner in
which parentsrespondto suchvariation is poorly understood.Received
3 January1990,accepted
17 March
1991.
PARENTS
of most seabird speciesprovision
their offspring with food collected at sea and
delivered to the nest. Among the Alcidae this
and provisioning at a number of coloniescollectedover successive
years(Harris 1978,Harris
and Hislop 1978, Wilson and Manuwal 1986,
behavior has been describedfor many species Barrett et al. 1987). Barrett et al. (1987) showed
and has been shown to changewith a variety that the breeding successof Atlantic Puffins
of factors,suchasweather (Birkhead 1976, Hatch
1984), fish stocks (Harris 1978, Harris and Hislop 1978, Hislop and Harris 1983, Springer et
al. 1984, Furness and Barrett 1985, Barrett et al.
1987, Baird 1990), the age of the chick (Harris
and Hislop 1978, Ashcroft 1979, Hudson 1979),
and position of the nest(Nettleship 1972,Harris
1984).The growth of the nestlingsof many specieswas reviewed by Gaston(1985). Many investigators have measured provisioning and
growth in successive
yearsat the samecolony
(e.g. Ashcroft 1979, Vermeer 1979, Gastonand
Nettleship 1981, Wehle 1983), but only a few
studieshave reported measurementsof growth
(FraterculaarcticaL.) was similar over large sec-
tions of the Norwegian coast.Coloniesin the
central area failed completely in three of the
four study years,while thosein the northern
and southern regions had good breeding success.All the coloniesenjoyed good breeding
successin 1983. Barrett et al. (1987) linked these
patternsto the level of herring (Clupeaharengus
Valenciennes) stocks. In Britain it was found
that puffinson St.Kilda were shortof the "best"
prey speciescomparedwith thoseon the Isle
of May and thus had difficulty rearing young
(Harris 1978,Harris and Hislop 1978).We measuredthe growth and provisioningin the Rhinoceros Auklet (CerorhincamonocerataPallas) at
three
colonies
on the British
Columbia
coast.
• Presentaddress:Departmentof Biology,McGill We found that thesemeasureschangedgreatly
University,1205,avenuedocteurPenfield,Montr6al, from year to year, and that the growth of nestlings changed in unison at these colonies.We
Qu6bec H3A IBI, Canada.
2 Present address: Canadian Wildlife Service, P.O.
examine three hypothesesto explain the obBox 340, Delta, British Columbia V4K 3Y3, Canada.
served patterns.
842
The Auk
108: 842-852.
October
1991
October1991]
Provisioning
byRhinoceros
Auklets
843
COASTAL
DOWNWELLING
550
DOMAIN
',,,,,,,,,,
BRITISH
COLUMBIA
Queen •
Islands
TRAI•SITION
ZONE
i•ih
eIsland
Triangle•
Island
500
COASTAL
UPWELLING
DOMAIN
1300
Fig. 1. CoastalBritish Columbia, showing Lucy Islands,Pine Island, and Triangle Island, the locationsof
the three study colonies.Oceanicfeatures(terminologyfrom Ware and McFarlane1989)referred to in the
text are also shown.
METHODS
RhinocerosAukletsoccurthroughoutthe northern
PacificOcean(Vetmeet 1979).Their breedingbiology
and natural history have been documented(Richardson 1961, Leschner 1976, Wilson 1977, Summers and
Drent 1979, Vetmeet 1979, Hatch 1984, Wilson and
Manuwal 1986).The single egg is laid in a burrow in
lateApril or May, and incubatedfrom 39 to 52 days
(average45 days).Parentsfeed their chicksat night
(but see Thorenson 1983) and may remain on the
colonyduring the dark hours,but they departbefore
sunrise.One or both parentsmayfeed the chickeach
night during a single visit to the burrow, although
Richardson(1961)reportsinstancesof three visitsto
a singleburrow in onenight. On somenightsneither
parentmay visit. A parentmay deliver one largefish
(up to 55 g, pers.obs.),or up to 20 smallerfish carried
cross-wise
in its bill. We refer to eachsuchdelivery
as a "bill load." In British Columbia, Pacific sandlance
(Ammodytes
hexapterus
Pallas) is the dominant prey
speciesalthough herring and rockfish(Sebastes
spp.)
are also commonprey (seeVetmeet and Westrheim
1984).
The growth rate of RhinocerosAuklet nestlingsis
among the slowestin the Alcidae (Ydenberg 1989),
Chicksleave the nestat 50-80% (250-400 g) of adult
weight (Vetmeet and Cullen 1979; pets. obs.)at 45-
60 daysof age,and are independentwhile they complete their development at sea after fledging. RhinocerosAuklets from BritishColumbiaprobablywinter off the California and Oregon coast(Kaiser et al.
1984).
Studyareas.--Theprovisioningstudy was conducted from 1985-1987 on the Lucy Islands (54ø18'N;
130ø37'W),an archipelagoof small,low-lying, heavily
forested islands located in Chatham Sound, 18 km
west of Prince Rupert,BritishColumbia(seeFig. I).
Approximately 21,500 pairs of RhinocerosAuklets
breed on the islands,spreadover mostof the larger
forested
islands.
We studiedgrowth on the Lucy Islandsin 19831987, and at two other Rhinoceros Auklet colonies in
1984-1986.Pine Island (50'58'N; 127ø41'W)is heavily
forested and low-lying, has an estimated 67,000
breeding pairsand lies I0 km northeastof Vancouver
Island. Triangle Island (50'52'N; 129'05'W) is an exposed,treeless,steep-slopedisland 40 km northwest
of VancouverIsland (Fig. I). It is British Columbia's
largest seabird colony and contains about 22,000
breedingpairsof RhinocerosAuklet amongmorethan
half a million pairsof otherseabirdspecies.The three
study coloniesall lie in or immediately adjacent to
the transitionalzone between two major northeast
Pacific production domains--the CoastalDownwelling Domainand the CoastalUpwelling Domain (Fig.
844
BERTRAM,
KAISER,
ANDYDENBERG
Experimentalchickswere hoodedon a rotating basisevery fifth night from the onsetto the termination
of the experiment,or until they left the nest.Chicks
were hooded in the evening between 1700and 2200.
The following morning eachnestlingwas unhooded,
and the uningestedfish collectedby systematically
checking all tunnels of its burrow, through previously formed accessholes. Following washing and
150
14(
120
110
).. 100
(D
90
z
measurement of the fish, chicks were fed their in-
tended meal. We used mass-lengthregressionson
samplescollectedin 1985to calculatethe total weight
of fish in the load. For each year a provisioning
curve
wasestablishedby usinga seconddegreepolynomial
equationto describethe relationshipbetweenburrow
load and chick age.We determined chick age from a
wing length-age regressionestablishedin 1985from
a sampleof chickswhosehatchingdateswere known.
To estimatethe sizeof bill loads,we capturedadults
on the colonyas they returned at night with fish for
their young. Samplingwas never conductedin areas
where experimentswere underway,and repeatsamples from the samearea were separatedby about 1
LU 80
O
70
cc
60
[Auk,Vol. 108
50
40
30
20
10
05
15 2535455565758595105115125
week to reduce disturbance
to the birds.
BURROW LOAD (grams)
Growthstudy.--Themain growth studywascarried
out
on the Lucy Islands in 1985-1987. We used reFig. 2. Histogramof all 790burrow loadscollected
in 1985-1987. The burrow loads are grouped into gressionsof nestlingmasson age during the period
blocksof 10 g each;the numbersalongthe axis(except between 10 and 40 days of age and the slope of the
zero) representthe midpoint of the group. Zero rep- fitted relation as an estimate of the growth rate for
each nestling. We call these growth measurements
resentsburrow loads ranging from 0 to 5 g.
sequential.
In 1987the nestlingsusedin this studyalso
participatedin the hooding experiment (seeabove).
We constructedcompositegrowth curves(Ricklefs
1). Theseare generatedby large-scalewind and ocean
and White 1975),on Lucy, Pine, and Triangle islands
patterns,and shifts in these patternscan causethe
in 1984-1986,and in 1983 and 1987 on the Lucy Istransitionzone to be dominatedby either the Downlands.To establisha compositegrowthcurve,we meawelling or Upwelling domain (Ware and McFarlane
sureda sampleof chicksencompassinga wide range
1989).
of developmentalstagesin early July on eachisland,
Provisioning
study.--Ineachyear28 to 45 chickswere
and again 10 dayslater. We estimatedthe growth rate
locatedin burrowson the LucyIslands,and the amount
for chickson eachislandin eachyear from the slope
of fish provided by their parentswas measured.We
of a linear regressionfitted to the full data set. The
fit the chickswith nylon hoods.Holes cut at the leve!
visitsto the islandswere carriedout by the Canadian
of the narespermitted easybreathing,and a drawWildlife Serviceand are reportedin full by Bertram
string prevented chicks from removing the hoods
and Kaiser (1988).
(Hatch 1984).We usedtwo sizesof hood, depending
on nestlingsize. Parentsvisiting hoodedchicksleft
the fish intended
for the chick
in the burrow.
The
RESULTS
morning following hooding, we located these fish,
measured
them and then fed them to the chicks. We
called fish retrieved in this manner a burrowload(Hatch
1984). A burrow load ranged from no fish to three
bill loads of fish. In 1985 some of the hooded
chicks
apparentlywanderedfrom their burrowsand were
lost. In 1986 and 1987, hooded chicks were tethered
in the nest chamber by 15 cm of nylon masoncord
attachedto a bird band on the leg, and securedto a
peg.This preventedwandering, and appearedto have
no adverseeffectson chicks,or on adult provisioning
behavior. Hooding had no significant negative impact on the growth of auklet chicksrelative to the
growth of unhooded control chicks (Bertram 1988).
Provisioning.--Burrow
loadsranged from 0 to
116 g, and the modal load was 30-35 g (Fig. 2).
The mode correspondswith the mean (+SD)
bill load measuredon the Lucy Islandsin 19851987 (1985:31.8 + 10.6, n = 74; 1986:34.1 _+
10.6, n = 212; 1987:33.5
+ 8.5, n = 131; overall:
33.5 _+ 9.7, n = 417), which indicates that the
majority of burrow loadsare composedof a single bill load. The largestbill load we collected
was only 55 g. The large burrow loads were
likely composedof two or perhapsthree separate bill loads. The loadswere, by mass,>85%
October
1991]
Provisioning
byRhinoceros
Auklets
Pacific sandlance, and there was no seasonal
845
60
(a) 1985
pattern in the appearanceof other prey species
(unpubl.
data).
There were considerable
differences
50
between
years in the pattern of burrow load deliveries
as chicksaged.In 1985burrow loadsincreased
with chick age and peaked at around 30 days
of age,then they declinedsteadily(Fig. 3a).The
40
estimates
for the two youngest
agegroupsin
30
1985 are based on small samples;more importantly, they are underestimates.Thesewere the
firstburrow
loads
wecollected
and,duetoour
20
inexperience,we probablymissedfinding parts
ofsome
burrow
loads
intheoftenintricate
bur-
10
row systems.(We later found someof them in
side branches.)These errors occurred only in
thesefirst two samplesand the steepnessof the
i
increase
in burrow
loadobserved
in 1985isex-
i
i
i
i
i
i
60
aggerated,though we have no way of estimat-
i
i
i
i
(b) 1986
ing by how much.In 1986the burrowloads •
delivered to young chicks remained roughly
constantuntil chicks were about 45 days old,
afterwhich they declined(Fig.3b).In 1987there
was no decline
in burrow-load
• 50
size as chicks
approachednestdeparture(Fig.3c).The decline
in burrow
loads near the end of the season in
1985 and 1986 is valid accordingto the significant negative quadraticparameterof the polynomial equations(Table 1). The Chi-squarecoefficients differ significantly between years
(1985-1986: t = 51.3, df = 496, P < 0.001; 19861987: t = 30.8, df = 501, P < 0.001). The shape
of the provisioningcurvecouldnot be accounted for by changesin bill-load sizebecausethere
was no consistentseasonalchangein bill-load
O• 30
lO
I
size(Bertram1988),asreportedby otherwork-
60
ers (Leschner1976, Wilson 1977, Hatch 1984).
The declinein provisioningin 1985and 1986
50
might, as many investigatorshave suggested
(e.g.Burger1980),be due to a seasonaldecline
in the food availablearoundthe colony.To test
this, we comparedthe provisioning of early-
I
I
I
I
!
I
I
I
(c) 1987
40
and late-hatchedchicks,reasoningthat if a sea-
sonaldeclinewere responsible,late-hatched
chicks would be fed less than early-hatched
chicks.Chickswere placedinto groupshatched
30
20
10
Fig. 3. Burrow loadsdelivered to hoodedRhinoc-
erosAuklet chicksin relationto their agein 1985(a),
I
1986(b), and 1987(c). For clarity group means(+SD)
are shown, but the regressionline shown is fitted to
all points in eachyear.
10
I
I
20
I
I
30
I
I
40
I
I
50
AGE (days)
I
I
60
846
BERTRAM,
KAISER,
ANDYDENBERG
[Auk,Vol. 108
TAnrE1. Coefficientsof the regressionequationsfitted to the relationshipof burrow load deliveredand
chick age (provisioningcurves)in different years.
Parameters
Intercept
(+SD)
Year
1985
1986
1987
1985-1987
-8.2
34.0
51.5
30.9
Age (x) _+SD
_+ 9.9
_+ 9.6 a
_+ 9.6 a
_+ 5.6 '
3.69
0.87
-0.49
0.97
Age2 (x2) _ SD
_+ 0.64 a
_+ 0.64
+ 0.59
+ 0.36 a
-0.061
-0.019
0.005
-0.018
n
+ 0.009 a
_+ 0.009 a
_+ 0.008
+ 0.005 •
287
209
294
790
• Significantlydifferent from zero with a = 0.05 for one-tailedtests.
within 5 days of each other. To compare the
mean
burrow
load
fed
to the
different
birth-
date groups as they aged, we used Friedman's
methodfor randomizedblocks(outlinedby Sokal and Rohlf 1981: 445). The procedure involved ranking the burrow loads acrossbirthdate groups for each age categoryconsidered,
essentiallycomparingthe burrow loads fed to
early and late chicks while holding age constant.
Growth.--We also comparedthe growth rate
of early-and late-hatchedchicks.Dataare avail-
able from 1986and 1987,and in neither year
waslatechicks'growthsignificantlyslower(Table 3). Thus neither provisioningnor growth
measurements supported the idea that latehatched
chicks were fed less.
We foundsignificantdifferences
amongyears
in growthrate.In theyears1984-1986(for which
we have compositegrowth estimateson all col-
The rankings of burrow loads indicated that
there was no significantdifferencein the weight
of burrow loads fed to chicks of the same age
"early" and "late" in the seasonin any of the
three years(Table 2). Unfortunately the statistical power of these comparisonsis low, and it
is unlikely that a difference of the magnitude
onies), growth was fastest in 1985 on all the
we estimate
by this methodare lower than thosegiven by
the compositemethod,their rankingisthe same.
would
have been
detectable
with
our sample (Bertram 1988).
colonies, slowest in 1986, and intermediate in
1984 (Table 4). For the Lucy Islandswe found
a rise in growth rates from 1983 to 1985, and a
decreasefrom 1985 to 1987. The sequential
growth measurementsmade on the Lucy Islandswere similar.Thoughthe estimatesmade
TABLE2. Burrow loads (g) delivered to chickshatched on different datesin 1985-1987.
Year/hatching
date
Chick
age
(days)
10
15
20
25
30
35
40
45
50
55
4-9 July
29 June-3 July
25.2
0.0
35.7
30.4
68.8
46.3
44.4
51.1
33.7
66.2
40.2
55.4
61.6
28.9
-39.9
-14.5
---
23-28 June
17-22 June
21.3
--
28.9
37.1
37.6
40.6
37.4
49.3
42.3
44.4
33.3
44.7
37.1
26.3
25.6
30.0
11-16 June
--
--
18.1
31.8
54.3
47.2
53.4
53.4
60.3
32.6
19.8
-22.9
14.0
1985 •
1986 b
16-21 July
10-15 July
4-9 July
29 June-3 July
38.2
46.6
40.8
--
32.8
37.8
61.2
42.8
29.7
17.0
61.2
68.6
14.2
27.6
29.5
25.2
18.5
19.5
33.5
17.4
.....
.....
42.9
49.7
34.1
73.9
---
---
---
23-28 June
42.9
40.2
51.0
38.4
48.9
46.9
34.7
--
--
--
---
40.7
48.3
52.4
37.7
35.1
42.4
40.1
54.2
37.7
36.1
43.4
40.8
32.7
39.4
22.3
56.2
---
1987 c
23-28 June
11-16 June
Friedman'sanalysis(seetext) performedon chick ages20-40 days:n = 195; X2= 3.04; df = 4, P > 0.05; NS.
Analysisperformedon chick ages15-30 days:n = 103;X2= 8.6; df = 4, P > 0.05; NS.
All datain analysisn = 276;X2= 0.5;df = 1, P > 0.05;NS.
October
1991]
Provisioning
byRhinoceros
Auklets
TABLE
3. Mean (+SD) growth rates(g/day) basedon
sequential measurementsof chicks hatched early
andlateon the LucyIslandsin 1986and 1987.There
is no significanttendencyfor later-hatchedchicks
to grow more slowly.
847
TABLE
5. Mean (+SD) nestlinggrowthrates(g/day),
basedon sequentialmeasurements,
in 1985-1987
on the LucyIslands.Samplesizesarein parentheses.
Year
Year / hatching
date
Growth
rate
n
F
P
2.18
0.139
,986
23-28 June
5.79 + 1.39
17
29 June-3 July
4-9 July
3.76 + 1.24
4.55 + 2.41
2
4
Growth
1985
1986
1987b
rate a
7.10 + 2.58 (14)
5.40 + 1.66 (23)
5.25 + 1.52 (33)
• F = 5.68; df = 2, 67; P = 0.004. A Student-Newman-Keuls multiple-
rangetestindicates
that1985differsfrom1986and1987,andthat1986
and 1987 do not differ from each other.
bChicksalsoparticipatedin the hoodingexperiment.
1987
11-16 June
5.79 + 1.81
16
23-28 June
5.50 + 1.19
17
0.96
0.346
Following the 1982/1983 E1 Nifio-Southern
Oscillation (ENSO) event, a warm-water mass
Growth rateswere highest in 1985 and lowest
in 1987 (Table 5). We found consistent differ-
ences among the three colonies (Table 4).
Growth rate was alwayshighest on the Lucy
Islands, and lowest on Pine Island. Moreover,
Pine and Triangle islandswere always more
similar to eachother than to Lucy, commonly
67-80%of the growth rate on the Lucy Islands.
DISCUSSION
Provisioning and the growth of Rhinoceros
Aukletchicksweresignificantlydifferentamong
the yearsof our study (Fig. 3, Table 1). Chick
growth was fastestin 1985(when provisioning
was terminated most rapidly) and slowestin
1987 (when provisioning was most prolonged
and did not decline). Rhinoceros Auklet chick
growth at all three coloniesrose and fell synchronouslyduring the study(Table4). We suggestthat an oceanicphenomenoninfluencesthe
entire region and somehowaffectsthe birds. A
mechanismto producesucha large-scaleinfluencewaspostulated(Ware and McFarlane1989).
Severalindependentlines of evidencealsosuggest that ocean production in the region increasedfollowing a low in 1983,peakedin 1985,
and fell again in 1986 and 1987.
moved northward along the western North
American coastline(see Mysak et al. 1982). In
the upwelling and transitional regions(Fig. 1),
this warm-water mass apparently suppressed
upwelling and hence productivity. In upwelling systems,cold saline surfacewatersare most
productive (e.g. Barber and Chavez 1983). We
therefore assumethat sea surfacetemperature
and salinity recordscan, taken together, provide a crude index of oceanproduction.Ocean
temperature(Fig. 4a) and salinity (Fig. 4b) at
northern light stationsalong the BritishColumbia coastindicatethat, after the major 1982/ 1983
ENSO event, production dropped, then rose,
and peakedin 1985;it hassincedeclined again.
In addition, recordsof zooplanktonproduction
(1985-1987; Thomson and Ware 1988), fish fau-
na and marinesalmonidsurvival(K. Hyatt pers.
comm.) and herring recruitment (Haist and
Schweigert 1989)supportthis interpretation of
fluctuationsin ocean production. Although it
is not clear that the ENSO
event
was itself
en-
tirely responsiblefor the observedpattern, the
availableevidenceis consistentwith a 1985peak
in production. The postulated mechanismfits
with the patternin growth measuredat the three
coloniesand alsoexplainswhy they all showed
the samepattern.
The 1982/1983 ENSO (the strongestof such
TABLE
4. Growthrates(g/day)of Rhinoceros
Aukletchickson threeBritishColumbiacolonies(1983-1987),
estimatedfrom the compositegrowth curve.The standarderror for the slopeestimatesrangesfrom 0.04
to 0.1. The samplesize is given in parentheses.
Colony
1983
1984
1985
1986
1987
7.2(42)
8.9(38)
10.8(30)
8.0(21)
5.4(33)
Pine Island
6.2 (68)
7.2 (53)
5.9 (59)
TriangleIsland
6.7(53)
7.4 (49)
6.4(40)
LucyIslands
848
BERTRAM,
KAISER,
ANDYDENBERG
[Auk,Vol. 108
(a)
---'
14
13
mo.
) 12
Jul
:::D
May
.•it:•
•.,•,,•,•• Apr
•
Mar
1982
1983
1984
1985
1986
1987
YEAR
32.0
(b)
31.8
31.6
31.4
31.2
31.0
Jun
30.8
Apr
Mar
30.6
30.4
30.2
30.0
I
I
I
I
I
1982
1983
1984
1985
1986
I
1987
YEAR
Fig. 4. Mean monthly sea-surface
temperatures(a) and salinities(b) from 1982-1987taken from northern
light stationsin BritishColumbia(Mar. Environ.Data Syst.,CanadaDep. Fish.Oceans,unpubl.).The months
chosenencompassthe developmentalperiod of the larvae and juvenile sandlancethat predominatethe
nestlingdiet. Salinity measurements
were taken from Bonilla,Kains,Langaraand Pine Islands.Temperature
was taken from theseislandsas well as St. Jamesand Egg Island.
events on record (Mysak 1986)) has been considered a factor in seabird mortality and reproductive failure in Peru (Duffy 1983,Hays 1986),
the central Pacific (Schreiber and Schreiber
1984),Oregon (Hodder and Graybill 1985,Bayer 1986) and as far north asAlaska (Hatch 1987)
(seeAinley et al. 1988 for a comprehensivereview). In most cases,changesin prey populations (primarily fish) were consideredbasicto
the problems. There has been some debate as
to whethershiftsin the distributionof fishprey
populations render them unavailable to birds
October
1991]
Provisioning
byRhinoceros
Auklets
(e.g. Duffy 1983)or whether fish abundancedeclines as a result of the decrease in primary
productivity (e.g. Barberand Chavez 1983). In
Peru the evidence suggeststhat the 1982/1983
ENSO disrupted the normal food web, and it
decreasedgrowth and reproductive successfor
most speciesof higher trophic levels (Barber
and Chavez 1983). The consequencesof such
oceanographicchangesdepend on locationand
prey speciesunder consideration.Regardless,
in any attempt to link events on seabird coloniesto oceanographicfluctuations,the question
concerningprey abundanceor availability will
arise(e.g. Briggset al. 1983,Springer et al. 1984,
Baird 1990). One thing that has becomeclear
from such studiesis that different seabirdspecieswill respondto changesin oceanographic/
feeding conditionsin different ways (e.g. Ainley et al. 1988, Hatch and Hatch 1990).
In our study, ocean temperature was lowest
and salinity was highest in 1985 (Fig. 4: a, b)
suggestinga peak in productivity in that year.
Moreover, in 1985, year-class-1sandlance(see
Vermeer and Westrheim 1984) dominated the
bill loads in all three seabird colonies, some-
thing which happened in no other year (Bertram and Kaiser 1988). We suggestthat a high
degree of ocean productivity lead to a strong
year-classfor sandlancealong the entire northern British Columbia coastin 1985. Although a
discussion
of recruitment
variation
in sand-
lance is beyond the scopeof this paper, for a
number
of marine
fishes in the Northeast
Pa-
cific Ocean, recruitment successis strongly influenced by environmental conditions (Hollowed et al. 1987). In British Columbia there is
no commercialfisheryfor sandlance,which may
allow for more direct linkage among oceanproduction, prey populations, and events on seabird coloniesthan in systemswhere seabirdprey
are harvestedcommercially.(For a discussion,
849
ing the later part of the nestling period (Atlantic
Puffin [Harris and Hislop 1978, Ashcroft 1979,
Hudson 1979], Pigeon Guillemont, CepphuscolumbaPallas [Emms 1987], and Black Guillemot,
C. grylleL. [Cairns 1987a]).In RhinocerosAuklets the decline seemsto result from fewer parental visits, rather than from a reduction of the
load delivered
on each visit.
We conclude
this
becausebill loads show no systematicdecline
in massin the latter part of the nestling period.
The decline
in burrow
loads must occur because
parents make fewer visits toward the end.
One hypothesis to explain the decreasein
feeding trips is that there is a seasonaldecline
in abundanceof food in the region surrounding
the colony (e.g. Burger 1980). This "seasonal
deterioration" hypothesis is rejected by our
study on two grounds.We found no evidence
that later-hatched chicks grew significantly
slower than early chicks,as would be expected
if there wasa seasonaldecline in the availability
of food. Nor did the analysisof the size of bur-
row loadsdelivered to chickshatching on different dates (Table 2) indicate any consistent
seasonaldecline.Although the statisticalpower
of the test was low, and we cannot entirely rule
out the possibilityof a seasonaleffect,it seems
unlikely to be large comparedwith the effect
of chick age on parental provisioning. Rhinoceros Auklets in the Gulf of Alaska are provisioned by adults according to the age of the
chickbut not accordingto the time of the season
(Hatch 1984). Evans (1981) reported a similar
result in Dovekies (Alle alle L.) in Greenland.
The inability to detecta seasonaleffecton the
size of burrow loadssuggeststhat the seasonal
deterioration hypothesisis insufficientto account for the observeddecline in feeding rate
to chicks in the late stagesof development.
Moreover, growth in 1985 was "good" at all
three
colonies
studied.
The seasonal
deteriora-
Cooper 1982, MacCall 1984 and Montevecchi et
tion hypothesiswould predict the oppositebecauseprovisioningdeclinedmostrapidly in that
al. 1988.)
year.
Our measuresof provisioning on the Lucy
Islands revealed patterns that changed in step
with indices of ocean productivity on the British Columbia coast. Parental provisioning
reachedapproximately the samepeak level in
each of the three years, but it declined most
rapidly in the year of higher productivity and
not at all in the year of lowest productivity. In
many semiprecocialalcid speciesthe rate of parental provisioningprogressivelydeclinesdur-
A secondhypothesisto explain the decline
in provisioning holds that the decline reflects
the decreasingenergy demands of older nestlings. In the Pigeon Guillemot (Koelink 1972)
and the Dark-rumped Petrel (Pterodroma
phaeopygiasandwichensis
Ridgway;Simonsand Whittow 1984),the energybudgetpeaksin the middle of the development period and then falls
toward fledging. Provisioningcurvesshouldbe
similar from year to year if parent Rhinoceros
see Crawford
and Shelton
1978, Furness and
850
BERTRAM,
KAISER,
ANDYDENBERG
Auklets provision chicks based on the chick's
energetic needs. The fact that the provisioning
curves are so different from year to year suggeststhat this hypothesisis insufficient to account for the variable decline in feeding rates
as chicksapproachfledging.
A corollary of the chick-energeticshypothesis is that chick fledging weights should be
similar between years. Although we have no
supportingdatafrom our study,meanfledging
weights on Triangle Island ranged from 51 to
69% of adult body weight in 1975-1978 (Ver-
[Auk,Vol. 108
influenceupon provisioningpatternsof British
Columbia Rhinoceros Auklets, the manner in
which parentsshould respondto suchenviron-
mental variation remainsopen for investigation.
ACKNOWLEDGMENTS
This work was made possible by the Canadian
Wildlife Service,which provided financial and much
logisticalsupport. Additional logistical support was
received from the Canadian CoastGuard, especially
Don and Kathy Richards.M. Healey, D. Blackbourn,
meet and Cullen 1979). The fact that fledging K. Hyatt, and D. Ware of the PacificBiologicalStation
weightswere more variablebetweenyearsthan were extremelyhelpful in spottingand tracing for us
the correlation between auklet behavior and largewithin years (Vermeer and Cullen 1979) also
scaleoceanicchanges.A. J.Gaston,W. C. Leggert,and
arguesagainstthe chick-energeticshypothesis. an anonymousreviewer providedvaluablecomments
We suggest that parental provisioning was on the manuscript. Financial support was also re-
affectedby interannual changesin oceanpro-
ceivedfrom the Anne Vall6e EcologicalFund,NSERC
duction, but the manner in which these large- grant U0451 to R. C. Ydenbergßand an NSERCscholscale changeswere manifested in the provi- arship to D. F. Bertram. Finally, we thank the many
sioning of nestlings is not obvious.A possible volunteersand assistantswho helped in the field.
explanation is that parents vary provisioning
LITERATURE CITED
from year to year accordingto the postfledging
D. W. ANDERSON,K. T.
survival prospectsof their offspring, which in AINLEY, D. G.ß H. R. CARTF•Rß
BRIGGS,
M. C. COULTER,F. CRUZ,J. B. CRUZ,C. A.
turn are affectedby annual changesin ocean
VALLE, S. I. FEFER,S. A. HATCH, E. A. SCHREIBER,
conditions.Although we presentlyhave insufR. W. SCHREIBERß
•: N. G. SMITH. 1988. Effects of
ficient datato test this idea, it appearsto warrant
the 1982-1983 E1 Nifio-Southern
Oscillation on
further attention. For parental RhinocerosAukPacificOceanBird Populations.Pp. 1747-1758fn
lets the colony is believed to be a dangerous
Proceedingsof the 19th International Ornithoenvironment. For example, Bald Eagles (Hallogical Congress(H. OuelletßEd.). OttawaßUniv.
iaeetusleucocephalus
L.) are known to congregate
of Ottawa Press.
on coloniesand prey on auklets as they arrive ASHCROFT,
R. E. 1979. Survival rates and breeding
or depart (DeGrange and Nelson 1982, Kaiser
biology of puffins on SkomerIslandßWales.Ornis
Scandinavica
10: 100-110.
1989).Consequently,
we proposethatwhen food
availability is high and the survival of chicks BAIRDßP. T. 1990. Influence of abiotic factors and
prey distributionon diet and reproductivesucat seais likely to be high (asin 1985),selection
cess
of three seabird speciesin Alaska. Ornis
favors parentsthat fledge young early, which
Scandinavica 21: 224-235.
increasestheir own survival prospects.Clark
BARBER,
R. T., & F. P. CHAVEZ. 1983. Biological conand Ydenberg (1990; see also Ydenberg 1989)
sequencesof El Nifio. Science 222: 1203-1210.
showedtheoretically that predation risk at sea- BARRETT,R. T., T. ANKER-NILSSEN,F. RIKARDSEN,K.
bird coloniesselectsfor "reluctant"parents,who
VALDE, N. gov, •: W. VADER. 1987. The food,
ceaseprovisioning when possible(seeNortongrowthand fledgingsuccess
of NorwegianPuffin
Griffiths 1969, Hauser and Fairbanks 1988).
chicks Fratercula artica in 1980-1983.
Ornis Scan-
dinavica 18: 73-83.
It hasbeen advocatedrepeatedlythat seabird
R.D. 1986. Breedingsuccess
of seabirdsalong
chick diets and growth rates may act as indi- BAYER,
the mid-Oregon coastconcurrentwith the 1983
catorsof marine resources(e.g. Cairns 1987b).
El Nifio. Murrelet
67: 23-26.
However, despite the fact that parent seabirds
BERTP,•M,
D.
F.
1988.
The
provisioning of nestlings
are capable of working harder than normally
by parent RhinocerosAuklets (Cerorhinca
mono(reviewed by Ydenbergand Bertram1989),we
cerata).M.Sc. thesisßBurnaby, British Columbia,
know very little about how parental provisionSimon Fraser Univ.
ing effort should vary with chick age in reß& G. W. KAISER.1988. Monitoring growth
sponseto a range of oceanfeeding conditions.
and diet of nestling RhinocerosAuklets to gauge
Although our studyimplicateslarge-scaleflucprey availability. Can. Wildl. Serv. Tech. Rep.
Ser. No. 48.
tuationsin oceanproductivity as the dominant
October1991]
BIRKHEAD, T. R.
Provisioning
byRhinoceros
Auklets
1976.
Effects of sea conditions on
ratesat which guillemotsfeed chicks.Br. Birds
69: 490-492.
BPaGGS,
K. T., D. B. LEWIS,& W. B. TYLER. 1983. Cal-
ifornia seabirdpopulationchangesearly in the
1982-83E1Nifioevent.PacificSeabirdGroupBull.
10: 46.
BURGER,
J. 1980. The transitionto independenceand
postfledgingparental care in seabirds.Pp. 367447 in Behaviorof marine animals(J. BurgerßB.
Olla, and H. Winn, Eds.).New York, Plenum Press.
CAImqS,
D.K. 1987a. The ecologyand energeticsof
chick provisioningby BlackGuillemots.Condor
89: 627-635.
1987b.
Seabirds as indicators
of marine food
supplies.Biol. Oceanogr.5: 261-271.
CLARK, C. W., & R. C. YDENBERG. 1990. The risks of
parenthood.I. Theoryandapplications.
Evol.Ecol.
4: 21-34.
851
H.qRms,
M.P. 1978. Supplementary
feedingof young
puffins. J. Anim. Ecol.47: 15-23ß
1984. The puffinßCarltonßT. and A.D. Poyser.
., & J. R. G. HISrOE. 1978. The food of young
puffinsßJ. Zool. 185:213-236ß
HATCHßS. A. 1984. Nesting diet and feeding rates
of RhinocerosAuklets in AlaskaßPp. 106-115 in
Marine birds: their feeding ecologyand commercialfisheriesrelationships(D. N. Nettleshipß
G. A. Sangerßand P. F. SpringerßEds.).Proceedings of the PacificSeabird Group Symposium,
1982. OttawaßCan. Wildl. Serv. Spec. Publ.
ß 1987.
Oscillation
Did the 1982-1983 E1 Nifio-Southern
affect seabirds in Alaska? Wilson Bull.
99: 468-474ß
, & M. A. HATCHß1990. Componentsof breeding productivity in a marine bird community:
key factors and concordanceßCan. J. Zool. 68:
1680-1690.
CP,•W•ORD,
R. J. M., & P. A. SHErION. 1978. Pelagic HAUSER, M.D., & L. A. FAIRBANKS. 1988. Motherfish and seabirdinterrelationshipsoff the coasts
offspringconflictin vervet monkeys:variation in
of South West and South Africa. Biol. Conserv.
responseto ecologicalconditionsß
Anim. Behav.
14: 85-109.
36: 802-813ß
DEGRANGE,
A. R., & J. W. NEtSON. 1982ßBald Eagle
predation on nocturnal seabirds. J. Field Ornithol.
53: 407-409.
Biol. Conserv.
26: 227-238.
EMMS,S. K. 1987. The adaptivesignificanceof the
patternof nestingdispersionin the PigeonGuillemot. M.Sc. thesis,Burnaby,British Columbia,
Simon
Fraser Univ.
EVANS,P. G.H.
1986.
Effects of the 1982-83 El Nino on
Humbolt Penguin coloniesin Peru. Biol. Conserv. 36: 169-180.
DUFf, D.C. 1983. Environmentaluncertaintyand
commercialfishing: effectson Peruvian guano
birds.
HAYS, C.
234-239ß
HODDER,J., & M. R. GRAYBILL.1985ß Reproduction
and survival of seabirdsin Oregon during the
1982-1983
1981. Ecologyand behaviour of the
Little Auk Alle alle in west Greenland.
HISLOP,
J.R. G., & M.P. Hm•ms. 1983. Recentchanges
in the food of young PuffinsFraterculaarcticaon
the Isle of May in relationto fish stocks.Ibis 127:
Ibis 123:
1-18.
FURhESS,R. W., & J. COOPER. 1982. Interactions be-
tween breedingseabirdand pelagicfish populationsin the SouthernBenguelaRegion.Mar.
Ecol.Prog.Ser. 8: 243-250.
ß & R. T. BARRETT.1985. The food requirements and ecologicalrelationshipsof a seabird
community in North Norway. Ornis Scandinavica 16: 305-313.
GASTON,
A. J. 1985. Developmentof the young in
the Atlantic alcidae.Pp. 319-354 in The Atlantic
Alcidae:the evolution,distributionandbreeding
biologyof the auksinhabitingthe Atlantic ocean
and adjacent waters (D. N. Nettleship and T. R.
Birkhead, Eds.). LondonßAcademic Press.
ß & D. N. NETTLESHIP. 1981. The Thick-billed
Murres of PrinceLeopoldIsland--a studyof the
breeding ecologyof a colonial high Arctic seabird. Can. Wild. Serv. Monogr. Ser. No. 6.
HAIST, V., & J. F. SCHWEIGERT.
1989. Stock assessmentsfor British Columbia Herring in 1988and
forecastsof the potentialcatchin 1989.Can. Manuscr.Rep. Fisheriesand Aquatic Sci. No. 2028.
E1 Nifio.
Condor
87: 535-541.
HOLLOWED,A. B., K. M. BAILEYß& W. S. WOOSTER.
1987.
Patterns in the recruitment
of marine fish-
esin the northeastPacificOceanßBiol.Oceanogr.
5: 99-131.
HUDSON,P.J. 1979. The parent-chickfledging relationshipof the puffinßFraterculaarctica.J.Anim.
Ecol. 48: 889-898ß
KAISER,
G. W. 1989. Nightly concentrationof Bald
Eaglesat an auklet colony.NorthwestßNatur. 70:
12-13ß
ßD. F. BERTRAM,& D. POWELL. 1984. A band
recovery for the RhinocerosAuklet. Murrelet 65:
57.
KOELINK,
A. F. 1972. Bioenergetics
of growth in the
PigeonGuillemotßM.Sc.thesis,VancouverßUniv.
British
Columbia.
LESCH•qER,
L. 1976. The breeding biology of Rhinoceros Auklet
on Destruction
Islandß M.Sc. the-
sis,SeattleßWashington,Univ. Washingtonß
MACCAI,
I•, A.D. 1984. Seabird-fisherytrophic interactionsin eastermPacificboundary currents:
California and Peru. Pp. 136-148 in Marine birds:
their feeding ecologyand commercialfisheries
relationships(D. N. NettleshipßG. A. Sangerß
and
P. F. SpringerßEds.).Proceedingsof the Pacific
852
BERTRAM,
KAISER,
ANDYDENBERG
SeabirdGroupSymposium,
1982.CanadianWildl.
Serv. Spec.Publ.
THOMSON, R. E., & D. M. WARE. 1988.
fisheriesfailures. Biol. Oceanogr.5: 153-161.
MYSAK,L.A. 1986. E1Nifio, interannual variability
and fisheries in the Northeast Pacific Ocean. Can.
J. Fish. Aquat. Sci. 43: 464-497.
Fish. Oceans, Canada.
THORENSON,
A.C. 1983. Diurnal activity and social
displaysof RhinocerosAuklets on Teuri Island,
Japan.Condor 85: 373-375.
VERMEER,
K. 1979. Nesting requirements,food and
breeding distribution of RhinocerosAuklets, Cerorhinca monocerata,and Tufted Puffins, Lunda cir-
, W. W. HSIEH, & T. R. PARSONS.1982. On the
rhata. Ardea
relationship between interannual baroclinic
wavesand fish populationsin the Northeast Pacific. Biol. Oceanogr.2: 63-103.
NETTLESHIP,
D. N. 1972. Breeding successof the
67: 101-110.
, & L. CULLEN. 1979. Growth of Rhinoceros
Auklet and Tufted Puffins on Triangle Island,
British
Common Puffin [Fraterculaarcfica(L.)] on different
habitats at Great Island, Newfoundland.
La Perouse
project:third annual progressreport, 1987.Dep.
MONTEVECCHI,
W. A., V. L. BIRT,& D. K. CAIRNS. 1988.
Dietarychangesof seabirdsassociated
with local
[Auk, Vol. 108
Columbia.
Ardea
67: 22-27.
, & S. J. WESTRHEIM.1984. Fish changes in
diets of nestling RhinocerosAuklets and their
implications. Pp. 55-105 in Marine birds: their
feeding ecology and commercial fisheries relationships(D. N. Nettleship, G. A. Sanger,and P.
F. Springer,Eds.).Proceedingsof the PacificSea-
Ecol.
Monogr. 42: 239-268.
NORTON-GRIFFITHS,
M. 1969. The organization, con-
trol and developmentof parental feeding in the
Oystercatcher
(Haematopus
ostralegus).
Behaviour
bird GroupSymposium,1982.Ottawa,Can.Wildl.
Serv. Spec.Publ.
34: 55-114.
RICHARDSON,
F. 1961. Breedingbiology of the Rhinoceros Auklet on Protection Island, Washing-
WARE, D. M., & G. A. MCFARLANE. 1989.
ton. Condor 63: 456-473.
production domains in the northeast Pacific
Ocean.Pp. 359-379 in Effectsof oceanvariability
on recruitment and an evaluation of parameters
used in stock assessmentmodels (R. J. Beamish
and G. A. McFarlane, Eds.).Can. Spec.Publ. Fish.
Aquat. Sci. 108.
WEHLE,D. H. S. 1983. The food, feeding and de-
RICKt•FS,R. E., & S.C. WHITE. 1975. A method for
constructingnestlinggrowth curvesfrom brief
visits to seabirdcolonies.Bird-Banding46: 135140.
SCHREIBER,
R. W., & E. A. SCHREIBER.1984. Central
Pacific seabirds and the E1 Nifio
Southern
Oscil-
velopmentof young Tufted and Horned puffins
lation:1982-1983perspectives.
Science225:713-
in Alaska.
716.
SIMONS,T. R., & G. C. WHITTOW.1984. Energeticsof
WILSON,U.W.
Edition. San Francisco,W. H. Freeman and Co.
SPRINGER,
A.M., D. G. ROSENEAU,
E. C. MURPHY,AND
M. I. SPRINGER.1984. Environmental controls of
Condor
85: 427-442.
1977. A study of the biology of the
Rhinoceros Auklet on Protection Island, Wash-
breedingDark-rumpedPetrels.Pp. 159-181in
Seabirdenergetics
(G. C. Whittowand H. Rhan,
Eds.).New York, Plenum Press.
SOKAL,R. R., & F. J. ROHLF. 1981. Biometry, 2nd
Fisheries
--,
ington. M.Sc. thesis,Seattle,Univ. Washington.
& D. A. MANUWAL.1986. Breedingbiology
of the RhinocerosAuklet in Washington.Condor
88: 143-155.
marine food webs: food habitats of seabirds in
YDENBERG,
R. C. 1989. Growth-mortalitytrade offs
and the evolutionof juvenile life historiesin the
Alcidae. Ecology70: 1496-1508.
the easternChukchi Sea.Can. J. Fish. Aquat. Sci.
--,
41: 1202-1215.
SUMMERS,
K. R., & R. H. DRENT. 1979. Breeding bi-
ologyand twinningexperiments
of Rhinoceros
Auklets on Cleland Island, British Columbia.
Murrelet
60: 16-27.
& D. F. BERTRAM. 1989.
Lack's clutch size
hypothesisand brood enlargement studieson colonial
seabirds.
Colon.
Waterbirds
12: 134-137.