338
m•R, Hybridization
andIntergradation
inBirds
SOME
CONCEPTS
INTERGRADATION
OF HYBRIDIZATION
IN WILD
BY ALDEN
POPULATIONS
[Auk
tOct.
AND
OF BIRDS
H. MILLER
APER$ISTENT
conceptionamonga considerablegroupof taxonomists
holdsthat intergradationand hybridization are contrastingphenomena
and that the termsshouldbe usedas antonyms. They regardan intergrading situation as denotinginvolvement of two racesand hybridization as involving two species. Accordingly, all one needs to do is to
find out which phenomenonis displayed where two forms come in
contact, and the troublesome borderline casesbetween race status and
speciesstatus are thereby correctlyindicatedand classified.
This chain of reasoningI regardas fallacious,for it is an oversimplification of the factorsand it doesnot clearly emphasizethe best criteria
for species. This incorrect reasoningseemsto be an outgrowth of a
view now abandonedby nearly all geneticistsand studentsof evolution
that speciespossessone sort of character and racesanother, or in other
words,that speciesdisplay qualitative differermesand racesquantitative differenceswhich blend or intergrade. The truth of the matter
is that both types of charactersmay differentiate racesfrom racesand
speciesfrom speciesand that there is no sharp distinction between
these kinds of charactersanyway.
Intergradation in its typical manifestationin continental areasindicates two things: (1) blendinginheritance of the charactersinvolved,
and (2) frequent,if not free, interbreedingof individuals. Hybridization indicates: (1) non-blendingor alternate type of character manifestation, usually with few genetic determiners involved in any one
character, and (2) free, or partial, or at least a little, interbreeding.
Since the kind of inheritance and the kind of character expression,
whether blending or of discontinuous-alternatetype, are not decisive
criteria of species,what do these situations tell us about the important
matter of freedom of interbreedingand subsequentfertility? One
impliesfreedomor somethingcloseto it, the other may involve a state
of freedombut also a state of rare, sporadicinterbreeding. The two
situations are thus not clearly contrastedin this important matter.
It is well to look into the exact meaningof the words "hybridize"
and "intergrade." To hybridize is to crossor interbreed organisms
that are different, whether of varieties, races, species,or genera. How
differentmust they be? The differencemay be a matter of eye color,
or only of bristle size or number in Drosophila. It may involve black
or white rats of the samespecies,or short and tall corn plants. When,
in wild populations, we can see that hybridization must have taken
VoL 66]
1949
I
MILLI•R,
l-Iybridization
andIntergradation
in Birds
889
place, we use the term. Hybridization is easy to detect when unit or
alternate characters are involved and when parents of very diverse
type interbreed. But, in the wild, the fact of hybridization is not
easily detected from inspectionof the offspringif a bird with a wing
length of 80 millimeterscrosses
with oneof 82 millimeters. The phenotypes do not segregateclearly and we can not, from the partial evidence,seethat hybridization occurredeven though it has. Hybridization takes place in an area of junction (primary or secondary)of
weakly differentiated,blendingracesbetweenslightly differentparents
just as truly as it doesbetweenstrongly contrastingparentsor between
those whose characterssegregatewell.
To intergrademeansto mergeor grade through a seriesof intermediate types. This demandscharactersthat will blend in areasof
interbreeding. The emphasisis placedon the type of character. One
of our terms then referschiefly to the mode of inheritanceor character
expression(intergrade), the other chiefly to the fact of interbreeding
(hybridize)whichmay be of high or low frequency. Hencemy view
that the two are not antonyms, that one processis often, indeed
usually, involved with the other.
Some illustrations may serve to show the interplay of these phenomena. The white-crownedsparrows,Zonotrichialeucophrys,
of the
southernCanadianRockiesrecentlycalledto attention by Rand (Auk,
65: 426, 1948)are differentiatedby meansof the presence(oriantha)or
absence(gambeli)of a black loral area. The changein proportionof
black-loredtypes increasesfrom the JasperPark area south to Waterton Lakes in a freely interbreedingseriesof populations. The characters show no real blending and the population at any one point
appearsmixed or hybrid. Adults of onetype are hybridizing with the
other. Yet, the frequency of occurrenceshifts gradually along a northsouth line; there is gradation in this sense. Because of the evidence
of free interbreeding and evident fertility, these forms are correctly
regarded as races,even though they showevidence of hybridization.
Comparably, the salt marsh song sparrows,Melospiza melodia, of
the San FranciscoBay region possessa character of yellow-colored
underpartswhich is lacking in adjoining upland races. There is no
gradation in this character, accordingto Marshall (Condor, 50: 238239, 1948); it is either present or absent, although when present it
fluctuates in intensity. In a short-range geographicgradient, this
character becomesless and less frequent as upland areas are approached. Free interbreedinghas been fully proved along this gradient; mixed pairs have often been observed. Yellow and white birds
thus hybridize, and hybrid populationsare produced. At the same
340
[Auk
MmL•R,
Hybridization
andIntergradation
in Birds
[Oct.
time, thesebirds display somedifferentiating featuresof dorsalcoloration which are not of the alternate or discontinuous type. These
intergrade in the samepopulationsfrom one extremeto the other and,
as usual in the extremes,there is "normal," but blending, individual
variation. There is no doubt here of racial status even though we
have hybridization.
The gray-headedjuneos,Juncocaniceps,of the Kaibab Plateau of
northern Arizona consistof a population of mixed black- and whitebilled birds, the black-billed birds derived from J. c. dorsalis to the
south of the canyon, the white-billed birds from J. c. canicepsin the
mountains to the northward in Utah (Miller, Univ. Calif. Publ. Zool.,
44: 186-189, 1941). Someintermediatebill types occur,but in general
the bill coloration sorts out into a small number of discrete types. In
any event, completefreedom of interbreedingoccursin the Kaibab
population, and the group presentsthe appearanceof a hybrid swarm.
At the sametime the gray of the head blendsbetweentwo extremesin
thesebirds; the group as a wholeshowsa wide range of variability, but
the variationsare gradient, with an intermediateaverage. There are
no availablebreedinggroundsimmediatelynorth or southin which, as
in the precedingexamples,the frequencyof the alternating characters
of bill could gradually shift. Hence we have a singlehybrid swarm,
not a geographiegradient either of blendingor of alternating characters
or both. Yet we seecomplete freedom of interbreeding of the diverse
types (no mate preference)that have taken up residenceon this isolated plateau; moreover, the group of eharactersseparating the extremes is neither great nor especially complicated. Racial status
seemswholly proper.
Other examples may serve to show intergradation where interbreedingis impeded or not free. Theoretically, suchsituationsmight
arise in areas of continuousdistribution owing to differentiating fertility factorsor reproductivedivergenceswhichwouldreducegeneflow
between organismsthat overlap or intergrade in all observablefeatures. Some exampleshave been noted in groupsother than birds,
but it is perhapssignificantthat in birdsexamplesare not evidentthat
do not entail the aiding circumstanceof geographicisolation--a most
important factor of eoursein race and speciesformation.
Loggerheadshrikes,Lanius ludovicianus,
on the mainlandof southern California
close to the shore line
are intermediate
between
the
coastal population generally and the race on the offshore islands
(Miller, Univ. Calif. Publ. Zool., 38: 82, 1931). No passageof shrikes
to and from the islandsis certainly known or is likely. The similarity
of some mainland birds to the island race is probably due to the in-
Vol. 66]
1949
I
MILLER,
Hybridization
andIntergradation
inBirds
341
flueneeof a set of environmentalconditionsalong the shore,which are
similar to conditionson the islands and which operate on island and
mainland birds alike. Here then, we assume,is intergradation caused
by gradingselectiveforcesof the environmentbut certainlynot involving free interbreedingof the extremetypes.
Another instance may be found in the isolated groupsof pygmy
nuthatches,Sitta pygmaea,alongthe detachedmountaintopsof southern California (Grinnell and Miller, Pac. Coast Avif. No. 27: 320-321,
1944). Thesediscretepopulationsintergradevery completelyin characteristics,yet interbreedingmust be very rare between groupsas a
resultof occasionaldispersalacrossthe severalinterveningbarriers.
It is rather widely agreed that speciesusually differ from one
anotherby many morecharactersthan do races. Heritablecharacters
in generalare probablymostoften manifestasblendingfeatures,less
often as distinct alternatefeatures. Therefore,specieswhichdiffer by
many charactersare more likely than racesto show somefeatures
whichwill be of the non-blendingtype and whichwill yield the obvious
pictureof hybridizationwherethey interbreed. This generaltruth is
perhapspartly responsible
for overstressing
the hybrid complexas a
sign of speciesstatus.
In seeking
criteriafor classifying
the borderline
cases
betweenspecies
and race status we obviouslyshouldnot use the mode of inheritance
of the charactersinvolved,but the degreeof freedomof interbreeding
and the degreeof fertility--barriers to free interchangeof geniematerial other than simple geographicisolation. Commonly we are
forcedto guesswhethersuchbarriersexist. Fertility may be lacking
even though hybrids occur. Island forms may be widely separated
yet, couldthey betestednaturally,mightinterbreedfreely. A typical
pictureof contiguous
or continuous
populations
showing
intergradation
rather well assuresthat there is no significantbar to geneflow. If
there is also transgressireintergradation,overlapthroughindividual
variation of every characterin insulargroups,it is a rathersafeassumption that there is potentialfreedomof exchange. But in instancesof
hybrid junctionwe mustlook with carefor the true biologicalcriteria
for species,
namelyfor somedegreeof actualreproductiveisolation.
Suchisolationmay be presentin the form of loweredviability or fertility, even •vken hybridizationis freely undertaken. I suspectit
occurssometimeswherea zone of hybrids betweentwo forms persists
as a narrow belt in the face of full geographiccontinuity and in the
seeming
absence
of ecologic
differentiation
of the rangesof the parental
forms. This possiblyis the situationwhereJuncocanJeeps
and Junco
oreganusmeet in southernWyoming and where Juncooreganusand
3Z•9.
M•r•r•, Hybridization
andIntergradation
inBirds
[Auk
toot.
Juncoaikeni meet in southernMontana; in the last instancethere may
also be a partial ethologicalbarrier contributing to the reproductive
isolation.
In brief, a groupof hybridslinking two formscanneitherbe assumed
to indicatespeciesstatusbecauseof "hybrid appearance"nor can they
safely be supposed,regardlessof their number, to indicate full or even
partial breakdown of reproductive isolation and hence racial status.
The situationmust be completelyexploredto determinethe manner
and degreeof reproductiveisolation,and we shouldnot alwaysexpect
therefrom a simple positive or negative answer.
Museumof VertebrateZoology,Berkeley,California, December15, 1948.