INTERSPECIFIC
NORTH
ALLOZYME
ATLANTIC
DIFFERENTIATION
WHITE-HEADED
RICHARD
AMONG
LARID
GULLS
R. SNELL i
Department
of Zoology,Universityof Toronto,Toronto,OntarioM5S 1A1 Canada
ABSTR•CT.--Iassessed
patternsof allozymicvariation in Larusargentatus
(Herring Gulls),
L. cachinnans
(Yellow-leggedHerring Gulls), L. fuscus(LesserBlack-backedGulls), L. glaucoides
(Iceland Gull), L. hyperboreus
(GlaucousGulls), and L. marinus(Greater Black-backedGull). I
surveyed34 presumedstructuralgeneloci over479 individuals.All birdswere monomorphic
and fixed for the same allele at each of 26 loci. There were neither
fixed allelic
differences
nor markerallelesfor the differentspeciesacrosseight polymorphicloci. Quantitativegeneflow estimatesare suggestiveof high interspecificgene flow (Nm = 2.1-10.7per generation).
Gene-diversityanalysesprovide evidencethat almostall allozymevariation (>90%) occurs
within colonies,whereasalmostnone (2.3%)is accountedfor by diversityamongspecies.
All thesespecies
aresimilarin bothproteinencodinggenesandmorphology,
whichimplies
overall genetic similarity. My resultsare consistentwith the hypothesisthat these taxa are
of very recent origin, and indicate that speciationneed only involve a minute portion of the
total genome.Additional genetic differentiation such as the accumulationof rare alleles or
fixed allelic differencesat electrophoreticallydetectableloci may simply be an artifactof the
time elapsedsubsequentto speciation.Received
27 March 1990,accepted
16 October1990.
STUDIES
of hybridization are important in attempts to understandthe processesof population differentiationand, ultimately, speciation
(Barton and Hewitt 1985). Among colonially
breeding seabirds,hybridization is thought to
be common (Pierotti 1987 and references there-
in), especiallyamong gulls in the genus Larus.
Pierotti (1987) argued that the specificmaterecognition systems(sensuPaterson 1980, 1981)
of seabirdsare exceedinglysimpleand are cued
to critical morphological charactersof leg/foot
coloration,bill pattern,and perhapssize.Where
recognition systemsof different populations
overlap,rangeexpansionand secondarycontact
may be followed quickly by hybridization. In
general, hybridization is expected wherever
gulls of different taxawith similar leg/foot coloration and bill patternsare sympatric.
Analysisof apparenthybrid zonesin gulls is
complicatedby the incompletedescriptionsof
natural variation of allopatricpopulations.The
earlier conclusions (Smith 1966a, 1969) on the
logisticallyimpossible(Snell 1989).Regardless,
in eastIceland,secondarycontactbetweenLarus
argentatus(Herring Gull) and L. hyperboreus
(GlaucousGull) occurredduring the 1920sfollowing range expansionof argentatus.
Variable
and seemingly intermediate plumage patterns
among Icelandicargentatus
are suggestiveof argentatus x hyperboreus
hybridization (Ingolfs~
son 1970, 1987), though morphometric evidence (Snell 1991) indicatesthat thesepatterns
simply reflectintraspecificvariability.
I usedprotein electrophoresis
to evaluategenetic differentiation in L. argentatus,
L. cachinnans(Yellow-legged Herring Gulls), L. fuscus
(LesserBlack-backedGulls), L. glaucoides
(Iceland Gull), L. hyperboreus,
and L. marinus(Great
Black-backed Gull). (L. cachinnansare consid-
ered argentatusrnichahellis
by some authors.)I
evaluate evidence for (1) patterns of genetic
structuringamongpopulations,(2) argentatusx
hyperboreus
hybridization in Iceland, (3) genetic
differentiation by distance,and (4) gene flow.
extent of hybridization in some charadriiform
birds appear unreliable. Unfortunately, completing the published set of experimental protocols on which Smith's (1966a, b; 1969) con-
clusionswere reportedlybasedwould havebeen
• Present address: Canadian Museum of Nature, P.O.
Box 3443, Station "D," Ottawa, Ontario, KIP 6P4, Canada.
METHODS
I collectedtissuesamplesof argentatus,
cachinnans,
fuscus,
glaucoides,
and hyperboreus
from 14 populations
in Europeand North America (Table 1, Fig. 1). I analyzed an additional compositesample of L. marinus
from eastern Iceland, eastern Canada, and northern
France (n = 5). Tissue samplesof heart, liver, and
pectoralmusclefor each bird (total n = 479 gulls)
319
The Auk 108: 319-328. April 1991
320
RICHARDR. SU•L
[Auk, Vol. 108
TABLE
1. Larusgull samplesites.Samplesizes(malesand femalespooled)indicated.
Latitude
Longitude
Species
n
Year
1 Home Bay, Baffin Island, Canada
2 Svalbard (Longyearbyen)
Locality
69ø00'N
78ø14'N
68ø00'W
15ø39'E
hyperboreus
hyperboreus
hyperboreus
argentatus
argentatus
argentatus
argentatus
argentatus
argentatus
argentatus
47
48
53
61
15
32
31
51
23
30
1985
1987
1986
1986
1986
1986
1986
1987
1988
1987
cachinnans
cachinnans
34
16
1988
1988
fuscus
glaucoides
24
9
1987
1985
5
--
3 Bjarnjarhafnjarfjall,westernIceland
65ø00'N
23ø00'W
4 Skruder, east Iceland
5 Karlsskali, east Iceland
6 Kent Island, New Brunswick, Canada
64ø54'N
65ø01'N
44ø35'N
13ø37'W
13ø40'W
65022'W
7 P. E. I. (Malpeque Bay), Canada
8 Troms•3,Norway
46ø30'N
69ø42'N
63ø50'W
19ø00'E
9 •le deBalanec,
Bretagne,
France
10 lie Ricard,Bretagne,
France
11 Camargue
(•tangduFournelet),
France
48ø25'N
48041'N
43ø21'N
4ø59'W
3ø54'W
4040'E
12
13
14
15
43ø29'N
62ø02'N
69ø00'N
4ø37'E
6047'W
68ø00'W
Camargue (Etang de la Galbre), France
Faroe Islands (Torshavn)
Home Bay, Baffin Island, Canada
Composite from sites4, 6, and 9; see text
were quick-frozen in liquid nitrogen in the field (no
pectoralmusclewas retained from Home Bay birds).
Tissueswere later kept either in liquid nitrogen or
in a -80øC
freezer.
As genetic variability and differentiation among
gulls was known to be slight (Ryttman et al. 1980;
Tegelstromet al. 1980;Johnson1982, 1985;Zink and
Winkler 1983),I sampledpopulationsfrom geographically distant regions to increase the likelihood of
detectinggeneticdifferentiation.Birdswere collected
during the breeding seasonby shooting,trapping, or
druggingwith either tribromoethanolor alpha-chloralose.Where possible,I collectedbirds at breeding
colonies. Otherwise, gulls were obtained at refuse
dumpsnear activebreeding sites.
The nuclear genomeof Laruspopulationswas examined by horizontal starch-gel electrophoresis.A
total of 34 presumptive loci were resolved (Appendix). Techniques followed Shaw and Prasad (1970)
and Harris and Hopkinson (1976). Tissueswere prepared for eachbird by joint homogenizationin a 0.1
M Tris-HC1buffer (pH 7.0) for most enzymes.I used
a 0.01 M dithiothreitol-0.1 M Tris-HCl buffer (pH 6.0)
for peptidases.Tissuehomogenates
were centrifuged
for 2 min in a Beckman microcentrifuge, and supernatants stored at -80øC until used. Running conditions and gel/buffer combinationswere standardand
are available on request. I used 9-10% starch in all
gels.Histochemicalstain recipesgenerally followed
thosein Shaw and Prasad(1970), Clayton and Tretiak
(1972), Turner (1973), Harris and Hopkinson (1976)
or Smith (1976); see St. Louis (1986: appendix 3) for
additionalstainrecipes.Isozymes,when present,were
numbered sequentially from cathode to anode. Al1ozymeswere scoredby lettersbeginning with A, in
decreasingorder of frequency.
Of the 34 loci scored,eight varied within or among
gull populations(see Results).All individuals were
scored for all loci to obtain frequency estimates of
marinus
rare alleles,which provide information on gene flow
(Slatkin 1985, Barton and Slatkin 1986). Sequential
electrophoresisof all individuals (Aquadro and Avise
1982, Hackett 1989) was not performed. However,
initial analysesevaluated resolution of the different
enzyme systemsunder various buffer and gel conditions. Gels for any particular enzyme systemrun
underdifferingelectrophoretic
conditionsdid not differ in either the number of alleles presentor their
relative patterns.
I used BIOSYS-1 (Swofford and Selander 1989) to
calculateallele frequencies,measuresof geneticvariability (mean heterozygosityby direct countand Hardy-Weinberg expectation),the exact significanceof
deviation from Hardy-Weinberg equilibrium within
populationsof individual loci (analogousto Fisher's
exact test), F-statistics, and matrices of Nei's (1978)
unbiasedand Rogers'(1972)D geneticdistance.I used
NTSYS (Rohlf 1989) for permutation tests.
Gene-diversityanalysis(Nei 1977:231) partitioned
total variance in the entire set of samples(HT) as follows: HT = Hc + Dcs+ Ds•, where H c and Dcsare the
gene diversitieswithin and between colonieswithin
subdivisions,and Dstis variation amongsubdivisions
relative to the total. Gene-diversityvalueswere based
on hierarchical F-statistics (Swofford and Selander
1989: 21).
To estimaterates of gene flow among populations
and species,I used Slatkin's (1981) indirect approach
of relating meanp(i) (mean frequencyof allelesfound
in i of d demes)with i/d (number of populationsin
which allele is present/total number of populations).
Also, Nm (the number of migrants between demes
per generation)wasdirectly estimatedusingthe "private" allele approachof Slatkin (1985) and Bartonand
Slatkin (1986)and the Fstmethod of Wright (1978).
Mantel's test (Dietz 1983) was used to evaluate the
significanceof the associationbetween among-locality geographicgreat-circledistance(GCDø) and ge-
April1991]
Allozymes
ofNorthAtlantic
Larids
321
Fig. 1. Geographicmap of sampling localities.
neticdifferentiation(Rogers'D). As Jackson
and Somers (1989) provide evidence that Mantel's test is not
stablewith as few as 500 permutations,I used 10,000
the averagenumber of alleles (r = 0.46, n = 15,
NS), polymorphismlevel (r = 0.53, n = 15, P <
0.05), and, mean Hobs(r = 0.38, n = 15, NS). There
to calculatesignificancelevels.I evaluatedgenetic is a tendencyfor larger samplesto be more varidifferentiationby distanceinterspecifically
amongthe able.
14 populationsof argentatus,
cachinnans,
fuscus,glauStandard errors of mean Hobsestimates were
coides,
and hyperboreus.
GeographicGCDs(in degrees
of arc) were calculated as follows:
GCDø = arccos[(sinA sin B) + (cosA cos B cos P)],
whereGCDø= great-circledistancebetweentwogeographicpoints,A and B;A = latitude of A; B = latitude
large, and in someinstancesthey were as large
as the estimatesthemselves.For each population the mean Hob•+ 1 SE overlapped the mean
Hardy-Weinberg expectedvalue (He•p).Mean
Hobswere highest among hyperboreus
popula-
of B;and P = differencein longitudebetweenA and
tions from Svalbard (0.044 + 0.018) and Home
B. A program (written in QuickBasic4.5) to calculate
a lower triangularGCDømatrix from a seriesof latitude-longitudecoordinatesis availableon request.
Bay (0.041 + 0.021). Mean Hobswas reduced
among conspecificsfrom Bjarnjarhafnjarfjall
(0.013 + 0.009). Among argentatuspopulations,
mean heterozygositywas highest at Troms6
(0.023 + 0.010);it was reducedat Malpeque Bay
RESULTS
(0.018+ 0.010),ile Ricard(0.017+ 0.010),
Generalresultsof allozymeanalyses.--Twentyfour enzyme systemsrepresenting34 presumptive gene loci were resolved satisfactorily.
Twenty-six loci were monomorphic and fixed
for the sameallele in all gulls (Appendix). A1lele frequenciesat the eight variable loci are
shown in Table 2. Average number of alleles
per locus per population varied from 1.0 to 1.3.
Levelsof polymorphismwere low, rangingfrom
2.9% (marinus)to 20.6% (argentatus
from Trom-
s6). Mean observedheterozygosity(Hobs)
of individual populations was 0.044 or less. How-
de Balanec (0.015 + 0.007), and Kent Island
(0.013 + 0.010); and it was lowest at the Icelandic colonies of Skruder (0.012 + 0.006) and
Karlsskali (0.008 + 0.005). Among cachinnans,
meanheterozygosity
waslowat •tangduFournelet(0.013+ 0.008)and ]•tangde la Ga16re
(0.009 + 0.009).The mean heterozygosityestimatesof glaucoides
(0.029 + 0.020),fuscus(0.009
+ 0.005), and the composite marinussample
(0.006 __+0.006) were also low.
In only 6 of 60 locus-by-locuscomparisonsof
individual populationsdid Hobs
and H•xpdiffer
ever,there wasa consistentpositivecorrelation significantly (P -< 0.05, exactprobabilities were
(Spearmanrank test) between samplesize and calculatedfor each polymorphiclocusin each
322
mc}mur•R. SNELL
[Auk, Vol. 108
TABLE2. Frequenciesof alleles at eight polymorphicloci in 15 Larid populations.The sequenceand sample
size of populationsare providedin Table 1. Loci are defined in the Appendix.
Population
hyperboreus
argentatus
Locus
1
2
3
4
5
6
7
EST-1 A
B
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
0.938
0.063
0.823
0.177
EST-2 A
B
C
D
0.830
0.170
0.943
0.057
0.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
0.000
0.000
0.875
0.125
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
GDA A
B
0.915
0.085
0.833
0.167
0.991
0.009
0.926
0.074
0.933
0.067
1.000
0.000
0.952
0.048
GPD A
B
C
0.511
0.489
0.000
0.510
0.490
0.000
0.028
0.972
0.000
0.057
0.943
0.000
0.067
0.933
0.000
0.194
0.806
0.000
0.177
0.823
0.000
IDH
A
B
C
0.979
0.021
0.000
0.958
0.042
0.000
1.000
0.000
0.000
0.934
0.066
0.000
1.000
0.000
0.000
1.000
0.000
0.000
1.000
0.000
0.000
PGI A
B
0.787
0.213
0.865
0.135
0.811
0.189
0.984
0.016
0.933
0.067
0.969
0.031
0.968
0.032
PGD A
B
0.957
0.043
0.854
0.146
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
PGM A
B
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
population). In eachcasethere was a deficiency philopatry, or substructuring of populations
even within
colonies.
of heterozygotes.
Geneticstructuringamongpopulations.--F-sta- Gene-diversity.--Significant Fsvvalues imtistic analysesare typically confined to popu- plied geographicand interspecificvariation in
lations within a single species,even in multi- genomes.Gene-diversityanalysisdetailing patspecies comparisons (e.g. Waples 1987). I terns of allozyme differentiation (e.g. Grudzien
calculated F-statisticsacrossall speciesto reflect et al. 1987) was done by separate hierarchical
the apparent frequencyof interspecifichybrid- analysesof (1) the six species,(2) pink/gray and
ization among gulls (Pierotti 1987) and the pre- yellow-footed gulls, and (3) non-Icelandic hysumed nonindependence of breeding pools.
perboreus,
non-Icelandicargentatus,
and IcelandI used standardizedgenetic variance (FsT)to ic colonies containing apparent or possible ardetermine the position of both the entire set of
populations and also populations among species relative to panmixia (FsT= 0), and to complete fixation for alternative alleles (Fsv= 1).
The Fsvestimatefor eachpolymorphic locuswas
significant (Table 3). The mean Fsv(0.108), averagedacrossall 15 populationsand all six species, indicated relatively little geographicand
interspecific variation in allele frequency.
Mean F•s(measureof inbreeding within populations) equaled 0.081, with all values for individual loci except GPD being positive (Table
3). Statisticallysignificant F•svalues for EST-1
(0.408), EST-2 (0.198), and PGM (0.418) indicate
a deficiencyof heterozygotes.Suchdeficiencies
might be caused by high levels of natal-site
gentatus
x hyperboreus
hybrids.Re•gardless
of
how the set of 15 populations was subdivided,
the within-colony diversity (Hc) accountedfor
almost all allozymic variation (> 90%) (Table 4).
Subdivision of populations by species provided evidence of diversity within taxa (Dcs)
that accounted for 6.9% of the variation, where-
asdiversity amongtaxa(DsT)accountedfor only
2.3% of total genic variability. Subdivision of
populations by leg color (probable mate recognition cues)provided evidencethat diversity
within mate recognition systems(Dcs)accounted for 9.9%of the variance. No genetic variance
was assignedto diversity among mate-recognition systems,and in fact the Dsvestimatewas
slightly negative. Subdivision of populations
April 1991]
TABLE 2.
Allozymes
ofNorthAtlantic
Larids
323
Extended.
Population
argentatus
cachinnans
fuscus
glaucoides
marinus
8
9
10
11
12
13
14
15
1.000
0.000
0.978
0.022
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
0.961
0.029
0.000
0.010
0.913
0.065
0.022
0.000
1.000
0.000
0.000
0.000
0.882
0.118
0.000
0.000
0.938
0.063
0.000
0.000
0.875
0.083
0.042
0.000
0.944
0.056
0.000
0.000
1.000
0.000
0.000
0.000
0.931
0.069
0.978
0.022
0.783
0.217
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
0.196
0.804
0.000
0.130
0.870
0.000
0.150
0.817
0.033
0.191
0.809
0.000
0.156
0.844
0.000
0.021
0.979
0.000
0.333
0.667
0.000
0.000
1.000
0.000
0.912
0.078
0.010
0.978
0.022
0.000
0.950
0.050
0.000
1.000
0.000
0.000
1.000
0.000
0.000
0.979
0.021
0.000
1.000
0.000
0.000
1.000
0.000
0.000
0.931
0.069
0.978
0.022
1.000
0.000
0.971
0.029
1.000
0.000
1.000
0.000
1.000
0.000
0.900
0.100
0.990
0.010
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
0.944
0.056
1.000
0.000
0.980
0.020
1.000
0.000
1.000
0.000
1.000
0.000
1.000
0.000
0.896
0.104
0.944
0.056
1.000
0.000
acrossthe Icelandic argentatusx hyperboreuspatric, and Icelandic subdivisions(Dcs)accountcontactzone (cachinnans,
fuscus,glaucoides,
and ed for only 2.6%of the variance.Diversityamong
marinuswere excludedfrom this analysis)in- these subdivisions (DsT) accounted for 7.1% of
dicatedgene diversitywithin allopatric,sym- total genic variability acrossthe contactzone.
Geneticand geographic
distance.--Nei'sD valT^BLE3. Summaryof F-statistics
at all polymorphic ues are all extremely small (0.000-0.009), are
loci. Significancelevels are associatedwith Chi- comparable to values reported between other
square testsof (1) Ho: F•s= 0, and (2) H0: FsT= 0; *
populationsof gulls (e.g. Karl et al. 1987), and
= P < 0.05, ** = P < 0.01, *** = P < 0.001.Only
populationspolymorphicfor the locusbeingtested may not differ significantlyfrom zero (Table5).
were included in each analysis.
F-statistics
FiT
FsT
b
EST-1
EST-2
GDA
GPD
IDH
PGI
PGD
PGM
Locus
0.408***
0.198'**
0.052
-0.032
0.065
0.113'
0.068
0.418'**
F•sa
0.480
0.244
0.132
0.131
0.098
0.184
0.150
0.458
0.122'**
0.057***
0.085***
0.157'**
0.035***
0.080***
0.088***
0.069**
Mean
0.081
0.181
0.108
' Chi-square= F•s2N(k
- 1), df = [k(k - 1)]/2 (Waples1987),where
N is the total numberof individualssampledfrom populationspolymorphicfor the locusbeing tested,and k is the numberof allelesat
that locus.
bChi-square= 2NFsT(k
- 1), df = (k - 1)(s- 1) (Waples1987),where
N and k are defined above,and s is the number of populationspolymorphicfor the locusbeing tested.
Nei's D values of zero occurredbetween populations within and among species,which indicatesnear geneticidentity between thesegull
taxa. Similarly, Rogers'D distanceswere low
(0.004-0.032).The Mantel's testprovided no evTABLE4. Gene-diversity estimatesof six speciesof
White-headed Gulls. Subdivisionalgroupings of
populationsanalyzedseparately.
Subdivisions
Hc
Dcs
DsT
Speciesa
Leg colorb
0.908
0.909
0.069
0.099
0.023
-0.008
Hybrid zonec
0.903
0.026
0.071
' argentatus,
cachinnans,
fuscus,
glaucoides,
hyperboreus,
and marinus.
bMorphologicalcuesin thesebirds' specificmate-recognition
systems: gray/pink or yellow legs.
ßIcelandicargentatus
or hyperboreus,
allopatricargentatus,
and allopatric hyperboreus.
324
R•CmU•D
R. SNELL
[Auk, Vol. 108
TABLE5. Matricesof geneticdistancesamong population samplesof Larusgulls, with Nei's (1978) unbiased
D abovethe diagonaland Rogers'(1972)D below. Populationsare sequencedas in Table 1.
hyperboreus
1
argentatus
2
3
4
5
6
7
1
2
-0.010
0.000
--
0.007
0.008
0.008
0.008
0.007
0.007
0.005
0.005
0.006
0.006
3
4
5
6
7
8
9
10
11
12
13
14
15
0.022
0.027
0.025
0.026
0.028
0.021
0.023
0.027
0.021
0.024
0.030
0.020
0.028
0.028
0.028
0.027
0.028
0.031
0.023
0.026
0.024
0.023
0.027
0.032
0.022
0.030
-0.011
0.001
-0.004
0.010
0.012
0.008
0.008
0.008
0.012
0.009
0.011
0.018
0.008
0.000
0.000
-0.009
0.010
0.008
0.008
0.011
0.010
0.008
0.012
0.017
0.005
0.002
0.001
0.000
-0.005
0.009
0.007
0.012
0.005
0.006
0.015
0.012
0.010
0.002
0.001
0.001
0.000
-0.012
0.010
0.013
0.011
0.010
0.019
0.017
0.014
0.008
0.013
0.017
0.014
0.010
0.019
0.012
0.010
0.011
0.018
0.005
idenceof a significantassociation
amongthe 14 from the six gull species(Table2). The resulting
populationsof the five gull speciesbetween curve was concave(Fig. 2), a pattern consistent
Rogers'D and great-circledistance(Z = 0.14,P with a hypothesisof high gene flow.
Based on Monte-Carlo
= 0.180).
simulations,
Slatkin
Genefiow.--I calculatedan indirect measure (1985)and Bartonand Slatkin (1986)arguedthat
of geneflow (i.e.the observedrelationbetween the mean frequency of "private" alleles (i.e. almeanp[i]andi / d) fromtheallelefrequencydata leles found in only one population) could provide an estimate of Nm, the actual number of
migrantsbetween demes per generation.Barton and Slatkin's (1986: 413) simulations suggest an almost linear relation between loga-
rithm p(1) and logarithm Nm, of the form
log•0[p(1)]= [a log•0(Nm)]+ b,where 0.01 < Nm
< 10. The mean samplesize per population of
these gulls was 31.9, and the mean frequency
of "private" alleles[p(1)]in thesegulls equaled
0.018.The coefficientsa and b vary with mean
North Atlantic Gulls
sample size per population, and were interpolated
between
the values
Barton
and
Slatkin
provided (for n = 25, a = -0.576 and b = -1.11;
for n = 50, a = -0.612 and b = -1.21) to a =
-0.586
and b = -1.14.
Thus, the estimate for
Nm among the populationsof the six speciesis
10.7 birds per generation.
An alternate
0.0
0.2
0.4
0.6
0.8
1.0
estimate
of Nm based on F-sta-
tistics (Wright 1978) was calculatedas Nm =
([FsT
-•) - 1]/4). With the mean FsTaveraged
acrosspopulationsand loci (0.108,Table 3), the
Fig. 2. Inverted-"L" relation of p(i) to i/d, over 15 estimate
populations of argentatus,
cachinnans,
fuscus,hyperboreus,glaucoides,
and marinus.This pattern is consistent
with conflicting
'hypotheses
of (1) high ratesof gene
flow and (2) little or no gene flow among minimally
diverged populations sharing ancestral polymorphisms.
for Nm was 2.1.
DISCUSSION
Geneticstructuring.--If it reflected presumed
genetic cohesivenessand monophyletic origin
TAnLI• 5.
325
Allozymes
of NorthAtlanticLarids
April 1991]
Extended.
argentatus
cachinnans
fuscus
glaucoides
marinus
8
9
10
11
12
13
14
15
0.004
0.004
0.001
0.001
0.000
0.000
0.001
-0.009
0.011
0.009
0.009
0.016
0.014
0.013
0.005
0.006
0.001
0.000
0.000
0.000
0.001
0.000
-0.012
0.005
0.004
0.009
0.013
0.010
0.006
0.005
0.003
0.001
0.001
0.001
0.002
0.001
0.001
-0.013
0.011
0.018
0.018
0.016
0.004
0.004
0.001
0.001
0.001
0.000
0.001
0.000
0.000
0.002
-0.004
0.011
0.010
0.011
0.005
0.005
0.001
0.000
0.000
0.000
0.001
0.000
0.000
0.001
0.000
-0.009
0.009
0.009
0.009
0.009
0.001
0.001
0.001
0.001
0.002
0.001
0.000
0.002
0.001
0.001
-0.015
0.010
0.002
0.002
0.003
0.002
0.002
0.000
0.001
0.000
0.001
0.002
0.000
0.000
0.002
-0.018
0.009
0.009
0.000
0.000
0.000
0.001
0.002
0.001
0.000
0.002
0.001
0.001
0.001
0.003
--
of nominal gull species,intraspecificdiver-
stand the role of selection
and environmental
induction in the productionand maintenance
of phenotypicdifferentiation.Where actuallevels of gene flow are high and populationsare
morphologically differentiated, nonselective
and nongeneticfactorsare implicateddirectly
as causalmechanismsin producingand mainhafnjarfjall conspecifics,and Ile de Balanec taining phenotypicclines.
argentatus
are more similar to fuscusfrom the
Slatkin(1981)arguedthe shapeof the p(i) vs.
Faroe Islands than to Ile Ricard conspecifics). i/d relation could be used to estimate levels of
White-headedgulls have differentiatedso little gene flow among populations.Slatkin suggestthat theseelectrophoreticdata cannotbe used ed that,whereasan "r"-shapedp(i) vs. i/d patto either evaluatealternativetaxonoraichypoth- tern would provide evidence of low levels of
esesor reconstructhistorical patternsof diver- gene flow (reflectingmany "private" allelesocgence.Similarly, thesedataprovided no insight curringat very high frequency,includingat or
into genetic structuring among mate-recogni- nearfixation),a reversed-"L"shapewould protion systems.
vide evidence of high rates of gene flow (abGeneticevidence
for Icelandic
argentatusx hy- senceof fixed allelic differences,coupledwith
perboreus hybridization.--Hybridpopulations few private allelesat very low frequency).On
arepredictedto be morevariablethanallopatric the basisof a reversed-"L"p(i) vs. i/d pattern
onesboth morphologically(Schuelerand Ris- and evidenceof morphologicaldifferentiation
ing 1976)andelectrophoretically
(Corbin1981). amongpopulations,Zink (1986: 107) proposed
There is no evidenceIcelandicpopulationsof "that local environmentalconditionsacting
gence among populationswould be predicted
to be lessthan interspecificdivergence.I found
no evidence of this. Interspecific genetic distanceswere in many instanceslessthan intraspecificones(e.g.Home Bayhyperboreus
aremore
similar to sympatricglaucoides
than to Bjarnjar-
eitherargentatus
$rhyperboreus
aremorevariable duringthe nestlingperiodshapeinherentphethan allopatricpopulationsof either species
morphologically(Snell 1991) or genetically
(thesedata). However, Corbin's(1981) prediction assumes
the parentaltaxahavethemselves
differentiatedto a notabledegree.As thesespeciesarenot differentiatedallozymically,the genetic data are not useful for testing Corbin's
hypothesis.
Genefiow.--Reliable estimatesof gene flow
would be of considerableimportanceto under-
notypic plasticity, effecting spatial patterns."
Though this is possible,there is no reasonto
think morphologicalpatternsin these gulls
simply reflect proximate environmental con-
ditions. Speciesare morphologicallydistinct
even where they breed sympatrically(e.g. argentatus,fuscus,and marinuson and near Ile Bal-
anec,orglaucoides
andhyperboreus
at HomeBay).
The reversed-"L"p(i) vs. i/d patterns observedwith these13 Laruspopulations(Fig. 2),
326
RICHARt>
R. S•qEI•I•
Fox Sparrows(Passerella
iliaca;Zink 1986:fig. 6),
Northern Flickers (Colaptesauratus;Grudzien et
al. 1987: fig. 2), and Cory's Shearwater (Calonectrisdiomedea;Randi et al. 1989: fig. 3) are
consistent with a hypothesis of high rates of
gene flow. However, the reversed-"L" patterns
are equally consistent with a hypothesis of
shared ancestral polymorphisms among recently divergedpopulations,and low or no gene
flow. The datasets(gulls, sparrows,flickers,and
shearwaters)allow no estimationof gene flow
by Slatkin's (1981) graphical approach.
Both direct estimates of Nm (2.1 and 10.7)
suggesta fairly high gene flow among taxa, in
[Auk,Vol. 108
Bay, or argentatusand marinuson Kent Island).
Thus, there is no reasonto assumethat patterns
of allele frequency and divergence in the 34
surveyedloci reflectpatternsof variation in the
presumablyfew genesthat controlthe heritable
aspectsof thesemorphological cuesof the materecognition system.
Changesat loci that control the SMRSsmay
be critical to the genetic differentiation underlying the processof speciation.Patterns of genetic divergence developed over time in isolated populations (such as the accumulation of
rare alleles, "private" alleles, or fixed allelic differencesat electrophoreticallydetectableloci)
keeping with the apparentfrequencyof inter- may be irrelevant to the essentialprocessof
specifichybridizationin gulls.The discrepancy speciation.
in actualestimatesmay be explainedby the fact
that the higher value, resulting from use of the
ACKNOWLEDGMENTS
"private" allele approach,lies outsidethe linear
portion of the log Nm vs. log p(1) relation. The
Collection of material for this study has required
robustnessof this estimate is also questionable considerableinternational cooperation.The followgiven that p(1) is the averagefrequency of the ing individuals helped in collecting: JosepheePani"private" alleles, and in this case,only three pak, Limakee Illuak, and Hugh Graham-Marr (Home
alleles were found,
far fewer
than Slatkin's
(1985) suggestedminimum of 20.
Recentdivergence.--Theconclusionthat these
six sp..•cies
have only recently divergedand that
there is high overall genetic similarity is sup-
ported by observationsof phenetic similarity
(Snell 1991), near genetic identity among species, shared fixed alleles at 26 of 34 loci, absence
of fixed allelic differences, shared common al-
leles at polymorphic loci and few rare or "private" alleles.Alternatively, the 34 loci surveyed
may be nonrepresentativeof overall patternsof
genetic variation, though there is no evidence
for or against this idea. Also, it is possiblethat
speciationwas not recent, but that evolution in
the nuclear genome of gulls has been constrained
for unknown
reasons.
Pierotti (1987) argued cogently that the specific mate-recognitionsystems(SMRSs)of gulls
and other seabirdsare exceedingly simple and
probablyinvolve the morphologicalcuesof leg/
foot coloration,bill patterning, and size. These
sixspeciesof gulls are very similar phenetically,
with significantdifferencesamong populations
being essentially restricted to morphological
size, plumage pattern, and leg/foot coloration
(Snell 1991). The SMRS characters of size and
leg/foot colorationevidently have a geneticbasis, at least in part, as they are distinct, even in
sympatry (e.g.glaucoides
and hyperboreus
at Home
Bay);BrianSnell (Kent Island);NelsonHurry (Prince
Edward Island); Agnar Ingolfssonand J6n Halld6r
(Bjarnjarhafnjarfjall);
J. Halld6r, Skarph•thinn Th6rissonand Thorhallur Borgarson(Skruder);Doret(•
Bloch,Genevi&veDesportes,Eydfin Stefansson,and
Monrad Mouretsen (Faroe Islands);Erling Norday
(Svalbard);E. Norday, Karl-Birger Straum,and Robert
Barrett (Troms6); David Ashworth and John Walms-
ley (Camargue);and D. Ashworth,Marie-H&l•ne Cadiou, Ewinn Kalgariou, Jean-PaulLinard, and Michel
Quern• (Bretagne).
The following organizationsallowed generoususe
of facilities:
Northwest
Territories
Renewable
Re-
sourcesat Clyde River;BowdoinCollegeBiologyStation on Kent Island;the BiologyDepartmentof the
University of Iceland; the Museum of Natural History, Faroe Islands; the Institute for Arctic Research
at the University of Tromsi5(in Troms6and on Sval-
bard);the Sysselmann,Svalbard;the BiologyDepartment of the University of Brest, France;Centre de
Recherchessur la BiologiedesPopulationsd'Oiseaux,
Brest;and the StationBiologiquede la Tour du Valat,
Camargue. Logistichelp was provided by David Par-
kin (University of Nottingham) and Joe Tigullaraq
(Clyde River).
Electrophoresiswas done at the Laboratoryof Analytical Systematics,Royal Ontario Museum. Allan J.
Baker, Carol M. Edwards, Mark Peck, and Vince L. St.
Louis provided advice.A. J.Baker,JoriC. Barlow, Fred
Cooke,Mark Engstrom,Gillian F. Gailey,Thaddeus
A. Grudzien, RogerHansell, Tom Parsons,Jim Rising,
and Kermit Ritland provided numerouscriticismsof
this manuscript.
April 1991]
Allozymes
of NorthAtlantic
Larids
This researchwas supported by the Natural Sciencesand EngineeringResearchCouncil of Canada
through a grant (A5999) to J. Rising. Additional sup-
327
ulations of Larusargentatus
and Larusfuscusin
northwestern EuropeßBiol. J. Linn. Soc. 24: 349363.
port was received from the Canadian Wildlife Service,
KARL,S. A., R. M. ZINK, & J. A. MARTEN. 1987. Allo-
the CanadianFederal Ministry of Indian and Northern Affairsthroughthe Universityof Toronto'sArctic
Working Group, the ChapmanFund of the American
Museum of Natural History, and a Sigma Xi grant-
zyme analysisof the California Gull (Laruscalifornicus).Auk 104:767-769ß
NEI, M. 1977. F-statisticsand analysisof gene di-
in-aid.
I received
an Ontario
Graduate
and Natural
Sciencesand Engineering ResearchCouncil postgraduate scholarship.
versity in subdivided populationsßAnn. Hum.
Genetics
41: 225-233ß
ß 1978. Estimationof averageheterozygosity
and genetic distancefrom a small number of individuals.
Genetics
PATERSON, H. E. H.
LITERATURE CITED
30: 1003-1019.
BARTON,
N.H., & G. M. HEWITT. 1985. Analysisof
hybrid zones.Annu. Rev. Ecol.Syst.16: 113-148.
ß & M. SLATKIN. 1986. A quasi-equilibrium
theory of the distribution of rare alleles in a sub-
divided population.Heredity 56: 409-415.
CLAYTON,J. W., & D. T. TRETIAK. 1972. Amine-citrate
buffersfor pH control in starchgel electrophoresis. J. Fish. Res. Board Can. 29: 1169-1172.
CORmN,K. W. 1981. Genic heterozygosityin the
White-crowned Sparrow: a potential index to
boundariesbetweensubspecies.
Auk 98:669-680.
DIETZ, E. J. 1983. Permutation tests for association
betweentwo distancematrices.Syst.Zool. 32:2126.
GRUDZIEN,
T. A., W. S. MooRE,J. R. CooK, & D. TAGLE.
1987. Genicpopulationstructureand geneflow
in the Northern Flicker (Colaptes
auratus)hybrid
zone. Auk 104: 654-664.
HACKETT,
S.J. 1989. Effectsof varied electrophoretic
conditionson detectionof evolutionarypatterns
in the Laridae.
Condor
91: 73-90.
HARRIS, H., & D. A. HOPKINSON. 1976. Handbook of
enzyme electrophoresisin human genetics.Amsterdam, North Holland
A comment
on mate rec-
ognition systems.Evolution 34: 330-331.
AQUADRO,
C. F., & J. C. AVISE. 1982. Evolutionary
geneticsof birds.VI. A reexaminationof protein
divergenceusing varied electrophoreticconditions. Evolution
89: 583-590ß
1980.
Publ. Co.
1981. The continuing search for the unknown and unknowable: a critique of contemporary ideas on speciation.South African J. Sci.
77: 113-119.
PIEROTTI,
R. 1987. Isolating mechanismsin seabirds.
Evolution
41: 559-570.
RANDI, E., F. SPINA, & B. MASSA. 1989. Genetic vari-
ability in Cory's Shearwater (Calonectris
diomedea). Auk 106: 411-417ß
ROGERS,
J. S. 1972. Measuresof genetic similarity
and genetic distanceßPp. 145-153 in Studies in
geneticsßVII. Univ. Texas Publ. 7213.
ROHLF,F.J. 1989. NTSYS-PC.Numerical taxonomy
and multivariate analysissystem.Setauket,Exeter Publishing Ltd.
RYrTMAN, H., H. TEGELSTROM,
& H. JANSSON. 1980.
Isozymedifferencesin three relatedLarusspecies
(Aves). Hereditas 92: 117-122.
SCHUELER,
F. W., & J. D. RISING. 1976. Phenetic ev-
idence of natural hybridization. Syst. Zool. 25:
283-289.
SI-mW,C. R., & R. PP,•S^D. 1970. Starchgel electrophoresisof enzymes: a compilation of recipes.
Biochem.
Genet.
4: 297-320.
SL^TKIN,M. 1981. Estimatinglevels of gene flow in
natural populations. Genetics99: 323-335.
ß 1985. Rare allelesasindicatorsofgene flow.
Evolution
39: 53-65.
INGOLFSSON,
A. 1970. Hybridization of Glaucous SMITH, I. 1976. Chromatographic and electrophoretic techniques. Vol. 2, Zone electrophoresis,
Gulls Larushyperboreus
and Herring Gulls Larus
argentatusin Iceland. Ibis 112: 340-362.
1987. Hybridization of Glaucousand Herring gulls in Iceland. Stud. Avian Biol. 10: 131140.
INTERNATIONAL UNION OF BIOCHEMISTRY. 1984.
En-
zyme nomenclature, 1984. New York, Academic
Press.
JACKSON,
D. A., & K. M. SOMERS.
1989. Are probability estimatesfrom the permutation model of
Mantel's test stable?Can. J. Zool. 67: 766-769.
JOHNSON,
C. 1982. The taxonomyand systematics
of
Larusargentatus
and Larusfuscus
in northwestEurope.Ph.D. dissertation,Univ. Nottingham,U.K.
1985. Biochemicalgeneticvariationin pop-
4th ed. London, W. Heinemann
Medical Books.
SMITH,N. G. 1966a. Evolution of some arctic gulls
(Larus):an experimentalstudy of isolatingmechanisms.Ornithol. Monogr. 4.
ß 1966b. Adaptationsto cliff-nesting in some
arctic gulls (Larus).Ibis 108: 68-83.
ß 1969. Polymorphismin Ringed Plovers.Ibis
111: 177-188.
SNELL,R.R. 1989. Statusof Larusgulls at Home Bay,
Baffin Island.
Colon. Waterbirds
12: 12-23.
ß 1991. Variably plumagedIcelandicHerring
Gulls reflectfoundersnot hybrids.Auk 108:329341.
ST. Louis, V. L. 1986. Genetic and morphometric
328
•IC•-L•RD
R. SlYELL
microevolution among introduced and ancestral
populationsof the EurasianTree Sparrow.M.Sc.
thesis, Univ. Toronto.
SWOFFORr),D. L., & R. B. SELANr)ER. 1989. BIOSYS-
1. A computerprogram for the analysisof allelic
variation in population geneticsand biochemical
systematics,
release1.7.Champaign,Illinois Natural History Survey.
TEGELSTROM,H., H. JANSSON,& H. RYTTMAN. 1980.
Isozyme differences among three related larid
species(Aves). Hereditas 92:117-122.
TURNER, B. J. 1973. Genetic variation of mitochon-
drial aspartateaminotransferasein the teleostCyprinodon
nevadensis.
Comp. Biochem.Physiol.44:
89-92.
WAPLES,
R. S. 1987. A multispeciesapproachto the
analysisof geneflow in marine shorefishes.Evolution
41: 385-400.
WRIGHT,S. 1978. Evolution and the geneticsof populations. Vol. 4, Variability within and among
natural populations. Chicago, Univ. Chicago
Press.
ZINK,R.M. 1986. Patternsand evolutionarysignificance of geographic variation in the Schistacea
group of the fox sparrow (Passerella
iliaca).Ornithol. Monogr. 40.
ß & D. W. WINKLER.
1983.
Genetic and mor-
phologicalsimilarity of two CaliforniaGull populations with different life history traits. Biochem.Syst.Ecol. 11: 397-403.
[Auk, Vol. 108
APPENDIX. Enzymesare listed by Enzyme Commission number (International Union of Biochemistry
1984), where possible.
Polymorphic
loci:1.1.1.8(Glycerol-3-phosphate
dehydrogenase,
GPD);
1.1.1.42(Isocitratedehydrogenase,
IDH); 1.1.1.44(6-Phosphogluconate
dehydrogenase,PGD); 2.7.5.1(Phosphoglucomutase,
PGM); 3.1.1.1(Esterase,EST-1, EST-2);3.5.4.3 (Guanine deaminase;GDA); 5.3.1.9 (Phos-
phoglucose
isomerase,
PG1).
Monomorphic
loci: 1.1.1.1 (Alcohol dehydrogenase,ADH); 1.1.1.14
(Sorbitoldehydrogenase,
SDH); 1.1.1.37(Malatedehydrogenase;
MDH1, MDH-2); 1.1.1.49(Glucose-6-phosphate
dehydrogenase,
G6PDH);
1.4.1.2(Glutamatedehydrogenase,
GLUD); 1.11.1.7(Peroxidase,
PER);
1.15.1.1(Superoxidedismutase;
SOD-l, SOD-2);2.6.1.1(Glutamic-oxalacetictransaminase,GOT); 2.7.3.2 (Creatine kinase; CK-1, CK-2, CK-
3); 2.7.4.3(Adenylatekinase;AK-1, AK-2);3.1.1.1(Fluorescent
esterase,
FI-EST);3.1.3.2(Acid phosphotase,
ACP); (Erythrocyteacid phosphotase;EAP-1, EAP-2); 3.4.11 (Peptidases;PEP-A, PEP-E,PEP-C);3.5.4.4
(Adenosidedeaminase,ADA); 4.2.1.3 (Aconitase;ACON-1, ACON-2);
5.3.1.8(Mannosephosphateisomerase,MP1).