Grit-Site Selection of Black Brant: Particle Size or Calcium Content?

2004·6.9 MB·English

by  Derek E Lee

Wilson Bulletin 1 16(4):304-313, 2004
GRIT-SITE SELECTION OE BLACK BRANT: PARTICLE SIZE OR
CALCIUM CONTENT?
DEREK E. LEE? 24 MATTHEW G. HAMMAN? ^ AND JEFFREY M. BLACK'
—
ABSTRACT. We examined selection of grit-ingestion sites by Black Brant {Branta hernicla nigricans) on
South Humboldt Bay, California in relation to particle size and calcium content. We hypothesized that Brant
site selection was dependent primarily upon calcium content and secondarily upon distribution of substrate
particle size. We (1) mapped grit-ingestion sites, (2) ranked their importance by Brant abundance and individual
movement probabilities between sites, (3) characterized Brant gizzard grit and compared it with grit available
at ingestion sites, and (4) compared calcium content and particle-size distribution between ingestion sites and
unused sites, and between primary and secondary ingestion sites. Brant repeatedly congregated at specific,
discrete sites during the 2 years of observation. The distribution of gizzard-grit particle size was right-skewed
toward larger particles (>0.5 mm) relative to the proportional availability of particle sizes in the substrate. We
found no significant differences in calcium content or particle size between sites where grit was inge.sted and
unused sites. Within used sites, the calcium content ofsubstrates at the primary ingestion site was significantly
higher than at the secondary ingestion sites, as ranked by Brant abundance and between-site movement proba-
bilities. Our findings from the field corroborate previous laboratory results, and confirm that calcium is a sig-
nificant ecological factor for this species. Received 14 May 2004, accepted 19 October 2004.
Preferred sites for gizzard-grit ingestion 1954). Calcium is a crucial breeding-season
may be used faithfully by wild bird popula- nutrient for eggshell and skeleton formation,
tions for many decades (Mcllhenny 1932), but but historically it has been relatively neglected
site selection of this resource is understudied. compared with investigations of fat and pro-
It has long been accepted that gizzard grit tein (Alisauskas and Ankney 1992, but see
(hereafter grit) is an essential aid for grinding Ankney 1984). Geese rely to varying degrees
food (Leopold 1931), and captive birds de- upon endogenous reserves for successful
prived of grit experience elevated mortality breeding (Ankney and Macinnes 1978, Rav-
(McCann 1939). In some waterfowl, amount eling 1979, Prop and Black 1998), and al-
of grit in the gizzard and size of grit particles though Black Brant {Branta bernicla nigri-
are related to diet, with more grit and smaller cans) rely on reserves less than other arctic-
particles in the gizzards of herbivores com- breeding waterfowl of similar mass (Ankney
pared to omnivores and carnivores (Thomas 1984), Brant skeletal mass (an index of cal-
et al. 1977, Skead and Mitchell 1983). Leo- cium content) is reduced 27% between pre-
pold (1933) proposed that grit could be a and post-laying (Ankney 1984). Trost (1981)
source of mineral calcium for birds. This is reported that grit consumption by captive fe-
well-documented only for Ring-necked Pheas-
male Mallards {Anas platyrhynchos) peaks in
ant (Phasianus colchicus), a species whose
the spring pre-breeding period; Mallards dem-
grit consumption is driven primarily by the onstrate selectivity in both particle size and
need for calcium and only secondarily—as a calcium content. Additionally, a calcium-de-
grinding material (McCann 1939) and
ficient diet reduces the breeding success of
whose distribution and abundance is influ-
Great Tits {Pants major, Graveland and Drent
enced by the availability of calciferous grit
(Leopold 1931; McCann 1939, 1961; Dale 1997), implying that calcium may be impor-
tant to breeding birds across taxa.
Here, we examine selection of grit-inges-
Wil'dlWiaftee,rfHouwmlboElcdotloSgtyateReUsneiva.r,chArGeartoau,p,CADep9t5,521o,f tion sites by Black Brant on South Humboldt
USA. Bay, California. Ourobjective was to examine
-Current address: PRBO Conservation Science, how calcium and particle size affect site use
4990Shoreline Hwy., Stinson Beach,CA94970, USA. by Brant, hypothesizing that selection of grit-
Current address: Sacramento National Wildlife
Refuge, 752 County Rd. 99W, Willows, CA 95988, ingestion sites was dependent primarily upon
USA. calcium content and secondarily upon the fre-
‘‘Corresponding author; e-mail: [email protected] quency distribution of particle size. We (1)
304
Lee et al. • GRIT-SITE SELECTION OE BLACK BRANT 305
mapped Brant grit-ingestion sites; (2) ranked potential for regional movements (Moore et
site importance by Brant abundance and in- al. 2004).
dividual movement probabilities between South Spit, the sandy peninsula separating
sites; (3) characterized gizzard grit from two South Humboldt Bay from the Pacific Ocean,
time periods, and compared it with grit avail- has one large and many small intertidal sand-
able at ingestion sites; and (4) compared cal- bars along its eastern shoreline that are used
cium content and frequency distribution of by Brant to rest, preen, and ingest grit as the
particle size between ingestion sites and un- tide ebbs (Fig. 1). These sandbars are exposed
used sites, and between primary and second- early in the ebb, long before the water level
ary ingestion sites. is low enough to allow Brant access to the
eelgrass beds. As soon as the substrate is with-
METHODS in —0.3 m of the water surface, Brant flock to
— the sandbar sites to ingest grit.
a sSmtauldly sspeeacigeosoasendtshiatte.breTehdesBalnadckmBorlatnstiins veyOsbsoefrvSaotuiotnhs.B—ayWetocoinddeuntcitfeyda8n8dBrmaanpt stuhre-
: western and northern Alaska, Russia, and the principle grit-ingestion sites. In 2000, we sur-
Northwest Territories (Reed et al. 1998). In veyed the bayside beach of South Spit from a
fall, most Black Brant stage at Izembek La- vehicle on South Spit Road (which runs along
I goon, Alaska, before migrating south, non- the bay shore) 24 times during daylight ebbing
II stop, to coastal lagoons of Washington, tides January-April; the abundance and dis-
: Oregon, California, Baja California, and Mex- tribution of all Brant were mapped on an ae-
li ico (Reed et al. 1998). During winter and rial photo of South Humboldt Bay, and grit-
spring, Brant head north again toward breed- ingestion behavior was noted. To better ob-
jj ailn.g 1g9r9o8u)n.dsNoirnththwearwdesmtiegrrnatNieoanrctisicco(mRpeoedseedt 3s1ervSeouatllhoSfpiStouRtohadBasyurivney2s0,01w,eisnuardvdeiyteidonthteo
ofshorter, stepping-stone flights between stop- entire South Bay 33 times during daylight
overs at bays and estuaries along the west ebbing tides in February and March using a
coast ofNorth America (Reed et al. 1998). We 60X spotting scope from an elevated obser-
conducted our study at South Humboldt Bay, vation site (Bell Hill in Fig. 1); the abundance
California (Fig. 1), an important stopover site and distribution of all Brant were mapped on
for Black Brant during their northward migra- an aerial photo of South Humboldt Bay, and
tion to the breeding grounds (Moore et al. grit-ingestion behavior was noted. We identi-
2004). Brant begin arriving at Humboldt Bay fied 10 main haul-out sites, and observed grit-
in mid-December, peak at —14,000 birds in ingestion behavior at 6 of these (Fig. 1). We
mid-March, and before April the majority based our analyses of grit selection on the
have departed Humboldt Bay (U.S. Fish and three most used grit-ingestion sites (A, B. and
Wildlife Service unpubl. data). C in Fig. 1 ) u—sed by 78% of all Brant.
Humboldt Bay is a 62-km- estuary (Barn- Movement To estimate movement proba-
.
hart et al. 1992); 1,044 ha ofeelgrass (Zostera bilities between grit-ingestion sites, we read
marina) occurs in discrete beds interlaced tarsal bands ofall Brant at the three most used
with a dendritic network of channels. Black grit-ingestion sites (A, B, and C in Fig. 1 ).
Brant feed almost exclusively on eelgrass dur- Brant were banded at major breeding and
ing the non-breeding season, and because molting locations; all ages of indi\idual Black
Brant do not dive, eelgrass can only be ac- Brant have been marked annually with
cessed at low tides (Derksen and Ward 1993). uniquely coded tarsal bands resulting in —8%
South Humboldt Bay contains 709f of the eel- ofthe total population being banded (Sedingcr
grass beds found in Humboldt Bay, and sup- et al. 1993, Ward et al. 1993, Bollinger and
ports 78-949f of the Brant that use Humboldt Derksen 1996). On 24 days between January
Bay each year (Moore et al. 2004). E'rom 350 and April 2()()(). we usetl a 60X spotting scope
km to the south to 600 km to the north, Hum- on the bay-side shore of South Spit to read
boldt Bay is the only large estuary con—taining the leg bands ol Brant ingesting grit at South
substantial beds ofeelgrass (>300 ha) effec- Spit sandbars during ebbing tides.
tively making it an insular study area with low Wc used multi-strata modeling in program
306 THE WILSON BULLETIN • Vol. J16, No. 4, December2004
LIG. I. Study area on South Humboldt Bay, California. Black Brant (Branta beniicla nigricans) using grit-
ingestion sites were observed from Bell Hill. Grit-ingestion sites are marked with solid-line circles (main sites
are labeled A, B, and C). Paired, unused, substrate-sampling sites are indicated by broken-line circles. Roost
sites with very fine sediments that were not sampled are indicated with squares.
MARK (White and Burnham 1999) to esti- size (AICJ (Akaike 1974, Burnham and An-
mate probabilities of banded Brant moving derson 1998). All models were run using the
between grit-ingestion sites. Two strata were logit link function. Goodness-of-fit (GOF)
defined: North (site A), and South (sites B and was assessed in a recaptures-only data struc-
C). Sites B and C were combined based on ture using the bootstrap procedure in program
their proximity and to make the number of MARK with 100 simulations. The general
individuals in each strata more equivalent. model for GOF testing was {5'(site X time)
Model selection was based in Akaike’s infor- /7(site X time)}, indicating that local study
mation criterion corrected for small sample area fidelity {S) and recapture {p) probabilities
Lee et al. • GRIT-SITE SELECTION OF BLACK BRANT 307
i
; varied by site and time. We accepted a general comparable to the best model. To account for
model GOF if its deviance ranked <90 out of model selection uncertainty, model averaging
[
the 100 rank-ordered deviances simulated. was used to create parameter estimates (Burn-
I Due to weather and tides, leg bands could ham and Anderson 1998). Model averaging
not be read every day to estimate daily move- uses AIC^ weights to calculate the weighted
ment probabilities; therefore, we pooled the average ofeach real parameter across all mod-
24 days of band-reading observations into els with AIC^ weights greater than zero.
—
twelve 7-day intervals. Different pooling in- Grit sampling. To compare gizzard grit
tervals were tried with the final selection be- with grit available at grit-ingestion sites, we
ing the shortest time period that met three cri- obtained hunter-donated gizzards to character-
teria: (a) the number of observation periods ize particle-size distributions of gizzard grit.
with no data was 0, (b) the rank of the devi- We sampled substrates at grit-ingestion sites
ance of the general model was <90th out of and compared those to gizzard grit to deter-
i
100 ranked bootstrap simulations, and (c) the mine whetherdifferential selection ofparticles
general multi-strata model {5'(site X time) was occurring. To determine whether calcium
/?(site X time) vl;(site X time)} would con- content or particle size distribution was the
'
verge. Temporal pooling violates the assump- main site-selection factor, we also compared
' tions of instantaneous observations and of a substrate samples from used and unused sites,
closed population during observations, and and between used sites.
[ can lead to biased parameter estimates (Har- During November 2000 and January 2001,
I
grove and Borland 1994); however, pooling hunters donated gizzards, wings, and heads
was necessary to estimate movementrates and from Brant taken on Humboldt Bay. Novem-
i is commonly practiced (Pradel et al. 1997, ber gizzards were more likely to be from birds
' Reed et al. 1998). newly arrived from Izembek Lagoon, Alaska,
I The a priori model set included S, p, and whereas January gizzards were more likely to
ij;(movement probability) as constants (•), be from birds either newly arrived from a
time-dependent (time), and linear trends more southerly stopover site, or from birds
through time (T), with site effects modeled as overwintering on Humboldt Bay. Extraneous
I
(site), (site + time), (site + T), (site X time), portions of the alimentary canal and fat de-
and (site X T), for a total of eight models for posits were removed, gizzards were opened,
j each parameter. We simplified the general and all contents washed into a container. Emp-
1
model systematically, starting with p, then re- ty gizzard wet weight (g), head length (mm),
il ducing S and v|j in turn for a total of24 models age (adult orjuvenile based on plumage char-
tested altogether (Lebreton et al. 1992). The acteristics), and month taken (November or
general model {5(site X time) /?(site X time) January) were recorded for each bird. Color
4^(site X time)} was reduced by hrst ranking of the gizzard grit was classified as either pre-
all eight models for p in the set while holding dominantly white or black in order to assign
S and vj; in their general form. After a parsi- its geographic origin. Organic matter was re-
monious model ofp was found, S was reduced moved from gizzard-grit samples by ignition
by ranking all eight models for S in the set at 500° C for 4 hr. After removal of organic
while holding p in its most parsimonious matter, grit samples were dried at 1()5°C for
form, and i(i in its general form. Finally, i|i was 24 hr and filtered through a stack of five
reduced by ranking all eight models for in sieves (mesh sizes: 0.053, 0.106, 0.25, 0.5,
the set while holding p and .V in their most 1.0, and 2.0 mm) for 5 min in a sediment
parsimonious form (Lebreton et al. 1992). The shaker; portions then were weighed. Propor-
model with the lowest A1C, value was con- tions (by weight) were arcsine transformed hu'
I sidered the best or most parsimonious model. analysis (Zar 1974).
I Akaike weights (Burnham and Anderson During low titles of 1 and 15 April 2001.
1998) were computed to denote relative we samplctl substrate at the three grit-inges-
I
strength of evidence supporting each model. tion sites on .South .Spit with highest mean
I Often, several models in the final set oftop- Brafit abimdance. anti at pairetl, unusetl sites
ranked models appear equally plausible, with within 50 m t)f usetl sites. Five samples were
AAIC, values near zero and A1C', weights ct)llectetl at each site except at usetl site A,
308 THE WILSON BULLETIN • Vol. 116, No. 4, December 2004
TABLE I. Table ofmodel selection results for local fidelity (5), recapture (/?), and movement (v}i) probabil-
ities of 322 Black Brant (Branta hernicia nigricans) ingesting grit at two sites on South Humboldt Bay, Cali-
fornia, 2000. While the best model (boldface, AAIC^ = 0) indicated no difference in movement probability
between sites, the two next-best models did include site, and had non-trivial AIC,. weights, thus contributing a
substantial site effect (a north-biased movement probability) to the final model-averaged parameter estimates.
AIC^
Model AIC<. AAIC^ weights k Deviance
{S(T)pisite + time) il>(-)p 687.4 0 0.37 15 241.7
{.S(T) /Hsite + time) vl;(site)} 687.7 0.3 0.31 16 239.9
{5(4) /;>(site + time) il/(site X T)} 689.1 1.7 0.16 18 236.9
(AiT) pisite + time) v|;(T)} 689.6 2.2 0.12 16 241.7
{A(T) /?(site + time) v}i(site + T)} 691.7 4.3 0.04 17 241.6
{5(T)pisite + time) v}i(site + time)} 697.9 10.5 0 26 227.5
“T= lineartemporaltrend,site = differencebetweennorthandsouthsites,time = fulltimedependence,(•) = constant.
where we collected seven samples. Each sam- formed foranalysis (Zar 1974). We used GLM
ple was a volume equivalent to approximately in program NCSS (Hintze 2000) to analyze all
30 g (dry weight) ofsand collected from with- gizzard grit and substrate data. Means are re-
in 1.5 cm of the surface. Samples were taken ported ± SE.
at 5-m intervals along a transect that began 10
m down slope from the high tide line, and ran RESULTS
along the ridge of each sandbar toward the Based on 88 surveys, Brant repeatedly
water line. In spring 2001, at Izembek La- hauled out at 10 specific sandbars along South
goon, Alaska, we also obtained two substrate Spit and at fine-sediment bars within the bay
samples (—30 g each) from the intertidal zone (Eig. 1). These sites were discrete and their
of one of the most heavily used grit-ingestion use by Brant did not change between the 2
sites. All substrate samples were dried, fil- years of observation, nor within an observa-
tered, and weighed using the same method de- tion season. Sites A, B, and C combined were
scribed above for analyzing samples of giz- used by 78% of all Brant. Site A, the north-
zard grit. Our analyses relied upon the as- ernmost grit-ingestion site (Fig. 1), was used
sumption that the gizzard grit we collected by more Brant (site A: mean = 342 birds ±
was ingested from the sites we sampled at Iz- 73, maximum = 1,580) than the two sites with
embek Lagoon and Humboldt Bay. There is next-highest abundance estimates (site B:
no way to confirm the validity of this as- mean = 94 ± 25, maximum - 250; site C:
sumption, but based on the appearance and mean = 105 ± 53, maximum = 200). Fine-
mineral composition of the grit particles, we sediment bars around the perimeters of eel-
are confident they came from the bays in grass beds were also used by Brant hauling
question, if not necessarily from the sampled out during ebb tides, but grit ingestion at these
sites. sites was not documented.
To determine calcium content of substrate Based on 322 individual encounter histo-
samples, we used a sample splitter to split en- ries, the most parsimonious model of ij; (the
tire —30-g substrate samples, one portion of probability of an individual moving between
which was split again to obtain —7-g samples. strata each week) between North (site A) and
Samples were dried at 105° C for 24 hr, cooled South (sites B and C) was constant (Table 1),
in a desiccator, and weighed. Calcium carbon- but model-averaged parameters indicated
ate content was estimated by measuring the North-biased iji (North to South: i|i = 0.123 ±
volume of gas evolved (corrected for sample 0.054; South to North: = 0.287 ± 0.124).
weight, temperature, and pressure conditions) The most parsimonious model of S indicated
when 10 ml of6 A"hydrochloric acid was add- no difference in local study area fidelity be-
ed to the sample and stirred for 5 min (de- tween sites (Table 1). Based on the number of
tailed methods in Machette 1986, modified Brant using grit-ingestion sites, and move-
from Dreimanis 1962). Percent calcium car- ment probabilities between the sites, we des-
bonate of each sample was arcsine trans- ignated the northern site (A) as the primary
Lee et al. • GRIT-SITE SELECTION OF BLACK BRANT 309
TABLE 2. Weight (SE) ofgizzards and grit found in the gizzards ofBlack Brant (Branta herniclanigricans)
shot at Humboldt Bay, California, in November 2000 and January 2001.
November2000 January2001
in = 31) 1 in = 14) F-value^
Gizzard wet weight (g) 72.5 (9.9) 102.8 (5.9) <0.001
Grit-sand weight (g) 6.9 (0.3) 10.0 (0.5) 0.001
Gizzard weight/sand weight 10.5 10.3
Grit color Black White
^FromMeststatistic.
grit-ingestion site, and southern sites (B and from Baja California north to Humboldt Bay
C) as secondary grit-ingestion sites. (DEL pers. obs., D. H. Ward pers. comm.).
The activities of Brant observed at the We compared particle-size distributions of
South Spit sandbars were tide-dependent. gizzard grit collected in November with sub-
When waterdepth over the sandbars was >0.5 strate from grit-ingestion sites at Izembek La-
m, no Brant were present. When water depth goon (Fig. 2A), and we compared gizzard grit
was <0.5 m, but sandbars were still sub- collected in January with substrate collected
merged, rafts ofup to hundreds ofbirds would from grit-ingestion sites at Humboldt Bay
jostle over the still-submerged sandbars, rap- (Fig. 2B). In November, gizzard grit was com-
mm
idly upending then coming up with mouths posed mostly of 0.5-1.0 particles. In Jan-
overflowing with sand solution. Once the uary, gizzard grit was made up ofnearly equal
sandbars became exposed as tide waters re- proportions of particles 0.25-0.5 and 0.5-1.0
mm
ceded, the few birds still ingesting grit were in size. Substrate at both Izembek Lagoon
found at the perimeters of the sandbars or at and Humboldt Bay was composed of mostly
mm
puddles of water remaining on sandbar sur- particles 0.25-0.5 in size. In both seasons,
faces. The three main grit-ingestion sites we distributions of gizzard-grit particle size were
focused on were exposed by ebbing tides at right-skewed toward larger particles (>0.5
nearly the same water level (1.8-1.6 m above mm) relative to the proportional availability of
MLLW), so no tidally induced sequence of particle sizes in the substrate (Fig. 2A, B).
accessibility was present. We compared particle-size distributions of
Hunters donated 31 gizzards in November substrate sand from ingestion sites with sand
and 14 in January. All gizzards collected in from unused sites in Humboldt Bay (Fig. 3).
November were significantly shrunken due to We found no differences in particle size dis-
the birds having recently undertaken their tributions between sites where grit was in-
non-stop southward migration. November giz- gested and unused sites. We also compared
zards contained significantly less grit than particle size distributions of sand from pri-
those collected from Brant during their “step- mary (site A, north site) and secondary (sites
ping-stone” northward migration in January, B and C, south sites) ingestion sites (Fig. 4).
although the ratio of gizzard weight to grit Relative to the north site, the south sites' dis-
weight remained constant between seasons tribution was right-skewed, containing larger
(Table 2). After controlling for month of col- proportions of particles 0.25-0.5 and 0.5-1.0
mm
lection, neither gizzard weight, nor grit-sand in size (Fig. 4).
weight differed significantly between adult The mean calcium content (proportion cal-
and juvenile birds; thus, ages were pooled for cium carbonate) of substrate from all grit-in-
subsequent analyses. Grit samples from birds gestion sites used by Brant (mean = 0.100 ±
collected in November contained predomi- 0.042) was not signilicantly higher (/ =
nantly black volcanic sand, indicating an ori- —0.88, df = 31, /^ = 0.19) than at unused sites
gin at Izembek Lagoon, Alaska, the primary (mean 0.088 ± 0.034). However, the cal-
fall staging area for Brant (I). 11. Wart! pers. cium content of substrates at the primary in-
comm.). Grit samples from birds collected in gestion site (site A: mean 0.137 ± 0.036)
Janua—ry contained predominantly white c|uart/ was significantly higher (/ — -5.01, df = 15,
sand as found at known Brant stopover sites P < O.OOl) than at the secondary ingestion
310 THE WILSON BULLETIN • VoL !16, No. 4, December 2004
0.70
I A Novembergizzards BIzembeksand
I
Particlesize(mm)
Particlesize(mm)
FIG. 3. Particle-size distributions (mean ± SE) for
substrate sand at grit-ingestion sites versus sites not
used by Black Brant {Branta bernicia nigricans) on
South Spit, South Humboldt Bay, California, 2000-
2001
.
based upon calcium carbonate content of the
substrate, and secondarily based upon avail-
ability ofparticle size. These results from field
data corroborate the laboratory-based findings
of McCann (1939) for Ring-necked Pheasants
Particlesize(mm) and Trost (1981) for Mallards.
We interpreted movement toward the pre-
FIG. 2. Particle-size distributions (mean ± SE) of
gizzard grit from Black Brant (Branta bernicia nigri- ferred site as being calcium driven. That the
cans) collected at Humboldt Bay, California in No- most preferred grit-ingestion site had higher
vember 2000 compared with substrate fromgrit-inges- amounts of calcium carbonate is not surpris-
tion sites at Izembek Lagoon, Alaska (A); and gizzard ing considering thateggs are composed of 10-
grit collected at Humboldt Bay, California in January
2001 compared with substrate from grit-ingestion sites
at Humboldt Bay (B).
sites (sites B and C combined: mean = 0.073
± 0.018).
DISCUSSION
We documented repeated use by Brant,
within and between seasons, of discrete grit-
ingestion sites on South Humboldt Bay. Used
sites were characterized by differential abun-
dance and constant, low levels of asymmetri-
cal movement by individual Brant between
sites. These differences were used to rank the FIG. 4. Particle-sizedistributions (mean ± SE)for
sites as primary and secondary in importance. substrate sand at primary (A) and secondary (B & C)
Only at the level of “within used sites” was grit-ingestion sites used by Black Brant {Branta ber-
nicia nigricans) on South Spit, South Humboldt Bay,
there any evidence of differential selection. California, 2000-2001. Sites were ranked as primary
Our results within used sites indicate that or secondary according to bird abundance and be-
Brant select grit-ingestion sites primarily tween-site movement probabilities.
Lee et al. • GRIT-SITE SELECTION OF BLACK BRANT 311
15% calcium (Gilbert 1971) and Brant egg- lecting paired unused sites at a scale that bi-
shells average 8.2 ± 0.1 g calcium (Ankney ased the results toward no difference). A ran-
I 1984). Calcium deficiency has been linked dom selection of the comparison sites might
with reduced reproductive success in Great have revealed differences. Selection of grit-
Tits (Graveland and Drent 1997). Graveland ingestion sites likely represents a complex in-
and Van Gijzen (1994) found that Great Tits teraction of nutritional requirements, social
could not obtain sufficient calcium from ar- factors, and grit availability.
I thropod and seed food items, but required cal- The right-skewed particle-size distributions
cium-rich supplementary material to meet the ofgrit found in Brant gizzards versus samples
demands of egg laying. Although Brant are of substrate sand could indicate the ability of
I less reliant on endogenous reserves of fat and Brant to ingest larger sand particles; alterna-
protein than other arctic-breeding waterfowl tively, it may reflect differential retention
due to rich food sources being available near times in the gizzard fordifferent particle sizes.
the nest (Ankney 1984), all the calcium re- We believe the difference is due to selective
quirements of the eggs must be available at ingestion because the only study of gizzard
the time of laying. For egg formation, birds particle retention in waterfowl (using captive
mobilize calcium from the skeleton and med- Mallards) found no difference in retention
'' ullary bone (Taylor and Moore 1954, Simkiss rates for particles of different size classes
! 1967), calcium that must be obtained and (Trost 1981). Wear (reduction of grit size
I stored during the nonbreeding season. while resident in the gizzard) could induce
' Movement toward secondary grit-ingestion only a left-skewed distribution relative to
sites could have been due to calcium satiation, available particle sizes. There is some dis-
’ slightly greater levels of the preferred (larger) agreement over the mechanism waterfowl use
' grit-particle sizes at secondary sites, displace- to selectively ingest grit particles (Crome
ment by more dominant competitors, and/or 1985, Kooloos et al. 1989, Nudds 1992,
1 avoidance of disturbance. The dimffmerence in Nudds and Wickett 1994, Mateo et al. 2000),
proportions of particles 0.5—1.0 in size but whatever process the birds use, grit-inges-
between the primary and secondary sites was tion behavior was much more prevalent when
small, but the difference in handling times re- sandbars were still submerged, indicating that
li quired to filter out sufficient quantities of Brant prefer to ingest grit when it is in sus-
these particles during opportune tidal win- pension.
dows may be enough to explain the attraction Calcium as an essential resource should be
I to secondary sites where the preferred particle more closely examined in grit and food sup-
sizes are more abundant. Levels of agonistic plies at stopover and breeding sites for all spe-
behavior were not documented in our study to cies of migratory birds. Similar studies of
i establish whether competition might drive the newly hatched precocial birds, which need
movement toward the secondary sites. Brant, calcium most for skeletal development, would
however, are sensitive to anthropogenic dis- also be instructive. Additionally, calcium re-
turbances (wSchmidt 1999), and disturbance quirements for successful reproduction and
may have influenced some oftheir movements the ability of various species to store and mo-
to secondary sites. The primary site is near bilize skeletal calcium should be determined
the main ship channel between South Hum-
precisely.
boldt Bay docks and both North Humboldt
Bay and the Pacific Ocean, while the second- ACKNOWLHDGMHNTS
ary sites are more remote from human activ-
ities. Wc tliank R. Guadagno (deceased), 1). H. Ward, .1.
We found no difterence between used and ,S. .Sedinger, anil especially, R. Poetter, who supplied
unused sites with respect to particle size or Alaskan sand. We thank G. .Susieh lor help in orga-
calcium content. No other factors were inves- nizing hunter contributions anti R. “Bud” Burke ot
Humboldt .State University lor allowing us to use his
tigated, thus the factor responsible for specific lab, including the Ghittiek apparatus. We acknowledge
site use by Black Brant at Humboldt Bay re- the support ot the Mono Bay Black Brant Group. C'al-
mains unclear. It could have been a function itbrnia W'aterlowI Association, and the ('alifornia.State
of scale in our experimental design (i.e., se- Department ot F ish and (iame (Duck Stamp Account).
I
312 THE WILSON BULLETIN • Vol. 116, No. 4, December2004
We also acknowledge the support of PRBO Conser- Hintze, j. L. 2000. Number cruncher statistical sys-
vation Science. tem: NCSS 2000. Kaysville, Utah.
LITERATURE CITED Kooloos, j. G. M., a. R. Kraaheveld, G. E. J. Lan-
GENBACH, AND G. A. ZwEERS. 1989. Comparative
Akaike, H. 1974. A new look at statistical model iden- mechanics offilterfeeding inAnasplatyrhynchos.
tification. IEEE Transactions on Automatic Con- Anas clypeata, and Aythya fuligula (Aves, Anser-
trol AC19:716-723. iformes). Zoomorphology 108:269-290.
Alisauskas, R. T. and C. D. Ankney. 1992. The cost Lebreton, J.-D., K. P. Burnham, J. Clobert, and D.
of egg laying and its relationship to nutrient re- R. Anderson. 1992. Modeling survival and test-
serves in waterfowl. Pages 30-61 in Ecology and ing biological hypotheses using marked animals:
management ofbreeding waterfowl (B. D. J. Batt, a unified approach with case studies. Ecological
A. D. Afton, M. G. Anderson, C. D. Ankney, D. Monographs 62:67-118.
H. Johnson, J. A. Kadlec, and G. L. Krapu, Eds.). Leopold, A. 1931. Report on a game survey of the
University of Minnesota Press, Minneapolis. north central states. Democrat Printing Company,
Ankney, C. D. 1984. Nutrient reserve dynamics of Madison, Wi.sconsin.
breeding and molting Brant. Auk 101:361-370. Leopold, A. 1933. Game management. Charles Scrib-
Ankney, C. D. and C. D. MacInnes. 1978. Nutrient ner’s Sons, New York.
reserves and reproductive performance of female Machette, M. N. 1986. Calcium and magnesium car-
Lesser Snow Geese. Auk 95:459-471. bonates. Pages 30-33 in Field and laboratory pro-
Barnhart, R. A., M. J. Boyd, and J. E. Pequegnat. cedures used in a soil chronosequence study (M.
1992. The ecology of Humboldt Bay California: J. Singer and P. Janitzky, Eds.). U.S. Geological
an estuarine profile. Biological Report, no. 1. U.S. Survey Bulletin, no. 1648.
Fish and Wildlife Service, Washington, D.C. Mateo, R., R. Guitart, and A. J. Green. 2000. De-
Bollinger, K. S. and D. V. Derksen. 1996. Demo- terminants oflead shot, rice, and grit ingestion in
graphic characteristics of molting Black Brant ducks and coots. Journal ofWildlife Management
nearTeshekpukLake, Alaska. JournalofFieldOr- 64:939-947.
nithology 67:141-158. McCann, L. J. 1939. A study ofthe grit requirements
Burnham, K. P. and D. R. Anderson. 1998. Model of certain gallinaceous game birds. Journal of
selection and inference: a practical information- Wildlife Management 3:31-41.
theoretic approach. Springer-Verlag, New York. McCann, L. J. 1961. Grit as an ecological factor.
Crome, E H. j. 1985. An experimental investigation American Midland Naturalist 65:187-192.
of filter feeding on zooplankton by some special- McIlhenny, E. a. 1932. The Blue Goose in its winter
ized waterfowl. AustralianJournal ofZoology 33: home. Auk 49:279-306.
849-862. Moore, J. E., M. A. Colwell, R. L. Mathis, and J.
Dale, E H. 1954. Influence of calcium on the distri- M. Black. 2004. Staging of Pacific flyway Brant
bution of the pheasant in North America. Trans- in relation to eelgrass abundance and site isola-
actions of the North American Wildlife Confer- tion, with special consideration ofHumboldt Bay,
ence 19:316-323.
California. Biological Conservation 115:475-486.
Derksen, D. V. and D. H. Ward. 1993. Life history Nudds, T. D. 1992. Patterns in breeding waterfowl
and habitat needs of the Black Brant. Fish and communities. Pages 540-567 in Ecology and
Wildlife Leaflet 13.1.15. U.S. Fish and Wildlife
management ofbreeding waterfowl (B. D. J. Batt,
Service, Washington, D.C.
A. D. Afton, M. G. Anderson, C. D. Ankney, D.
Dreimanis, a. 1962. Quantitative gasometric deter-
mination ofcalcite anddolomiteby usingChittick H. Johnson, J. A. Kadlec, and G. L. Krapu, Eds.).
University of Minnesota Press, Minneapolis.
apparatus. Journal of Sedimentary Petrology 32:
520-529. Nudds, T. D. and R. G. Wickett. 1994. Body sizeand
seasonal coexistence ofNorth American dabbling
Gilbert, A. B. 1971. The egg: its physical and chem-
ical aspects. Pages 1379-1399 in The physiology ducks. Canadian Journal ofZoology 72:779-782.
aanndd bBi.ocMh.emFirsetermyano,ftEhdes.d)o.mAesctaidcemfiocwlPr(eDs.s,J.NBeelwl Pradetoln,.R.1,99N7..RIinoduixv,idaua.lTatmuirsnioeverr, aanmdonJg.-Dw.inLteebrrien-g
York. teal in Camargue: a mark-recapture study. Journal
Graveland, j. and R. H. Drent. 1997. Calciumavail- ofWildlife Management 61:816-821.
ability limits breeding success of passerines on Prop, J. and J. M. Black. 1998. Food intake, body
poorsoil. JournalofAnimal Ecology66:279-288. reserves and reproductive success of Barnacle
Graveland, J. and T. Van Gijzen. 1994. Arthropods Geese Branta leucopsis staging in different habi-
and seeds are not sufficient as calcium sourcesfor tats. Norsk Polarinstitutt Skrifter 200:175-193.
shell formation and skeletal growth in passerines. Raveling, D. G. 1979. The annual cycle ofbody com-
Ardea 82:299-314. position in Canada Geese with special reference
Hargrove, J. W. and C. H. Borland. 1994. Pooled to the control ofreproduction. Auk 96:234-252.
population parameter estimates from mark-recap- Reed, E. T, E. G. Cooch, F. Cooke, andR. A. Goudie.
ture data. Biometrics 50:1129-1141. 1998. Migration patterns of Black Brant in
Lee et al. • GRIT-SITE SELECTION OE BLACK BRANT 313
Boundary Bay,BritishColumbia.JournalofWild- Taylor, T. G. and J. H. Moore. 1954. Skeletal deple-
life Management 62:1522-1532. tion in hens laying on a low calcium diet. British
Schmidt, R E. 1999. Population counts, time budgets, Journal of Nutrition 8:112-124.
and disturbance factors of Black Brant (Branta Thomas, G. J., M. Owen, and P. Richards. 1977. Grit
bernicla nigricans) at Humboldt Bay, California. in waterfowl at the Ouse Washes, England. Wild-
M.Sc. thesis, Humboldt State University, Areata, fowl 28:136-138.
California. Trost, R. E. 1981. Dynamics ofgrit selection and re-
Sedinger, J. S., C. J. Lensink, D. H. Ward, R. M. tention in captive Mallards. Journal of Wildlife
Anthony, M. L. Wege, and G. V. Byrd. 1993. Management 45:64-73.
Current status and recent dynamics of the Black Ward, D. H., D. V. Derksen, S. P. Kharitonov, and
Brant Branta bernicla breeding population. Wild- V. V. Baranyuk. 1993. Status of Pacific Black
fowl 44:49-59. Brant Branta bernicla nigricans on Wrangel Is-
SiMKiss, K. 1967. Calcium in reproduction and phys- land, Russian Eederation. Wildfowl 44:39-48.
iology. Chapman and Hall, London, United King- White, G. C. and K. P. Burnham. 1999. Program
dom. MARK: survival estimation from populations of
Skead, D. M. and R. j. H. Mitchell. 1983. Grit in- marked animals. Bird Study 46:S120-S139.
gested by waterfowl in relation to diet. South Af- Zar, j. 1974. Biostatistical analysis. Prentice-Hall,En-
rican Journal of Wildlife Research 13:32-34. glewood Cliffs, New Jersey.