THE
AUK:
A QUARTERLY
JOURNAL
OF
ORNITHOLOGY.
VOL. xxI.
THE
JULY, •9o4.
BIOLOGY
OF
RESPECT
TO
THE
TYRANNIDz•;
THEIR
No. 3'
WITH
SYSTEMATIC
ARRANGEMENT.
BY
I)R.
H.
VON'
IHERING.
TH• systematicarrangementof animalsis usuallybased on
morphologicalcharactersonly, but biological observationsmay
often give us precioushints which may help us to settledifficult
questionsin this respect. Consideringthat the presentsystematic
arrangementof the genera belonging to the great family of
Tyrannidmis far from being a satisfactoryone, I have thought
it usefulto studyin a comparativeway the biologyof the members
of this family.
Of the four subfamiliesaccepted according to the classification
of Mr. Sclater, at least one seems to be unnatural as well as
regardsmorphologicalas biologicalcharacters;that is to say,the
Platyrhynchinm.
In my paper on eggsand nestsof Brazilianbirds (Revista do
Museu Paulista,IV, i899, p. 226)1 describedthe nest and eggs
of .Platyrhynchus
myslaceus
and expressedmy surpriseat their
great differencewhen comparedwith the nests and eggs of the
allied genera. Having obtainedthis year an authenticnestof this
speciesI am able to statethat as regardsthe first describednest
there was a mistake. The nest of the abovementionedspeciesof
.Plalyrhynchus,
which will be fully describedin Vol. V of the
VoNIH•RISG,t7œ010•y
of the Tyrannida•.
3•4
l-Auk
L July
' Revista do Museu Paulista,' is purse-shapedand suspendedat
the extremityof a branch. It has a round openingin the middle
protectedby a shelter above. I have quite similarnestsof
rostrum cinereum,Orchilus auricularis, ]-femitriccusdio•ps,and of
different speciesof.œuscarthmus. The same form of nest is,
therefore,commonto the genera't'latyrhynchus,
Todirostrum,
]ruscarthmusandOrchilus. Moreover,the eggsof the membersof all
thesegeneraare yellowishwhite or brownishwith very fine points
on the larger end.
On the other hand the nestsof Serphaphaffa
are placed among
the divergingboughsof a branch and are cup-shaped,while the
eggsare of a uniform yellowishwhite. Of the sametype are the
nests and eggs of An•eretes and lfapalocercus. The nest of
•'hyllascartes ventralis, however, as Mr. Krone assured me, has
one wall of the nest elongatedabove and recurveal,forming a
somewhatglobular,domedstructure.
A form of nest like that of Serphophaga
is found in the genera
Elainea and •'hyllamyias, and in other Elaineinto,among which,
however, occurs also a second form of nest. This form is illus-
trated by the nest of Ornilhianabsaletum
which I have recently
examined. It is of a pear-shapedform, similar to that of ]ruscarthmus,but not suspendedfrom the top of a twig but fixedat different points on the branches.
Euler hasdescribedthe nestof this speciesdifferently,but the
nest observedby him, whichwasmuchhiddenbetweenmassesof
Tillandsia,was not probablyof a normalform. Besides,O. obsoleturndoesnotoccurin Rio de Janeiroand Bahia,wherethe speciesis representedby O. cinerascens
(Wied), which,in opposition
to Mr. Allen, I do not doubt is identical with O. imberbeScl.
A
similar nest is built by Mionectes
ru)7ventris
(Licht.), as has been
observe,d
by Mr. Krone.
The
nest of Ornithion
forms the transition
between
that
of
Serphapha•aand that of Euscarthmus. We may be justified to
assume that such an artificial and wonderful
construction as the
nest of Euscarthmus is not the work of free invention but is to be
consideredas the resultof developmentfrom a previousform of
nest. We have but to supposethe nest of Ornilhion,insteadof
being fixedon variousbranchessuccessively,
to be placedon one
Vol.
xxIx9o4
-II
Vo>•
IvmR•x•,
B/ology
of theTyrannldcv.
3 15
branch only and we have the suspendednest of ]•uscarthmus. A
very remarkable form of suspendednest among the Tyrannid•e
occurs in the genus2?hynchocyclus,
but I believe it to be nothing
more than an extreme modificationof the Euscarthminenest type.
In this respect it is rexnarkable that the much bristled and flattened bill of Rhynchocydus
is very like that of ]•/atyrhynchus,and
quite different from that of the typical Elainefe.
Although the predominantform of nest amongthe Tyrannidas
is certainlythe cup-shapedone, we meet also with very different
structures in this family. Covered nests occur in the genera
]'hy/loscartes,•trundinico/a, ]¾tang•us,and •œyiazeleles,
leading us
on to the nest of Ornithionabovedescribed,and to the suspended
nests of ]ruscarthmus and 2•hynchoO,dus. According to Euler
Myiobius barbatus has a suspended,purse-shapednest, while the
nest of •,iobius ngvius is cup-shaped,but is fixed suspended
within the fork o• two diverging branchesin the manner typical
of the nests of Thamnophilus and other Formicariidte. On the
other hand we find nestsof very slight structuremade of a small
numberof slendersticksand rootsin the genera ]'yrannus,
rioriotous,Myiodynastes,Megarhynchus,and others. These nests
are extremelyflat and apparentlynot well suitedto retain the eggs
in safety.
Amongthe Tteniopterime,inhabitantsof the open plains,there
are specieswhichbreed in holesin banks,as is said to be the case
in Argentina with ]'•enioi)teranengetaby Mr. Hudson, while in
Brazil this speciesbuilds its nest on trees. The speciesof
rus and Machetornisbreed in holes of trees, as also does
terairupero. The last-namedspecieslikesto appropriatethelarge
covered mud-nests of Furnarius, and Machetornis prefers the
largethornynestsof •tnumbius. Thus we seeamongthe Tyrannidasthe most different forms of nest structuresrepresented.
In generalthe nestsof speciesthat inhabit the woodsare well
built, and covered carefully with dry plant material in order to
be well hidden.
Some of them, such as that of ]•lainea, are true
masterpiecesof art, being generally ornamentedexternallywith
piecesof lichencarefullyfixedon by spiders'webs. On theother
hand,the Tasniopterin•e
and Tyrannidas,inhabitantsof the pampas and campos,contraryto what would be expected,take little
3t6
Vo:NIHEP.•XC;•
Biolog'y
of l•}eTyran.&le.
['Auk
[.July
care to hide their nests. Everyonewould supposethat these
birds should prefer to build their nestsunder cover of the grass
and form simple structuresof dried grass, as do the speciesof
Sycctlis, .4mmodromus, Embernot,rot, Emberizoides and other
Fringillidze of the campos. Except, however, in the genus
Aleclrurus,I do not knowanyotherexampleamongthe Tyrannidaeof this form of nesting. In general these birds are not very
careful to hide their nests. On the contrarythe largespeciesof
Tyranninae,and the speciesof the allied genera zl•ryiozetetes
and
]'itanffus,seem to prefer to place their nestson isolated trees, as
much exposedas possible. This custom correspondswell with
the bold characters
of these birds.
Taking a generalview of the eggsof Tyrannidze,we find a uni-
formityin colorationwhichis in strongcontrastto the varietyof
formsof their nests. The eggsin this family,as a rule, are white
or cream-whitewith reddishbrownspotsat the largerend. These
spotsare smalland pointed in the Euscarthminze,
whilethey are
obsoletein the buff eggsof the Serphophaginze. The occurrence
of pure white eggsis limited to the genera Copurus,.4rundinicohz,
and a few others. In the genus Jl•ruscivora
the ground color is
somewhatdark brownish. The eggsof the genus•ryiarchusare
remarkable for the elongated form of the numerous red-brown
spots.
If we compare the color of the eggs with the mode of construction of the nests no pronounced correlation is to be found.
Uniformlywhiteor cream-colored
eggsexistin the generaCopurus
and .4rundinicola,which are depositedin coverednests,while the
similar eggs of the Serphophaginae
are laid in open nests. The
eggsof the Euscarthminze,
though depositedin closed nests are
adorned with numerous red spots, while those of •ryivzetetesand
jøitan•us,which are laid in closedand domed-shaped
nests,have
the samelarge,reddishbrownspotsas thoseof the genera25,rannus, zl•rilvulus,and others,the nestsof which are wholly open.
Similar casesprevail in the eggs of other South American birds.
In this respectthe exampleof the A•nericanGallinzeis instructive,
for while as regards the carelessconstructionof the nest no difference is noticeable, the eggs of the Brazilian representativesof
the Gallinzeare white, while those of the Crypturi are distin-
Vol.
XXII
I9o4
-1
VONIHERING,
Biology
oftbeTyrannidce.
-
3 17
guishedby the mostbrilliant colorsof brown,red,blue,and green.
It is true that the Pici, Psittaci,and other birds that lay their eggs
in hollowtrees,all have white eggs,but eggs of the same color
are found also in the open nests of the Trochilid•e and of the
Columb•e. Returningto the Tyrannid•ewe find the eggof g(achetornisrixosawhollydifferentfrom thoseof the true T•eniopterin•e,
and resemblingin its numerous,
large, somewhatelongatedbrownish spots the eggs of the genera EmzPidonomus
and g(yiarchus,
especiallythose of the latter. These facts throw doubt on the
correctness
of the generallyacceptedsystematicpositionof g/irachelornis. So far as regardsthe egg of 2"yrannus
auranlioalronolatus
Lafr. & D'Orb., it belongs to the genus 2"yrannusand not to
2•mpidonotnus.
These differences,therefore, cannot be explained by the so-
calledlaw of ' natural selection,'but bear relationsto the genetic
affinities and the inner movementswhich, independentlyof the
supposed'natural selection,'determinedthe individualvariation
as well as the phylogenetic
development
of the organisms.
After what I have stated it is evident that the systematic
arrangementof the Tyrannid•e in its present form can only be
consideredas provisional,and it may be well altered when a general anatomical study of the whole group has been made. At
present the systematicsectionsare only based on a restricted
number of external characters,principally on the form of the
tarsi, feet and bills.
These characters are in intimate connection
with the manner of life. In this way we are exposedto the
danger of confoundingessentialtypical characterswith adaptive
ones. I think that such a mistake took place on the occasionof
the formation of the subfamily T•eniopterin•e. This section
embracesformswith strong feet, strong and elongatedtarsi, and
slenderelongatedbills, characterswhich seem to resultfrom the
life on the groundon the pampasand campos,which thesebirds
inhabit.
In generalthis subfa•nilymay be considereda very natural
one. The predominantcolorsof the speciesare gray, white, and
black. These colorsare not common in the family Tyrannidaeas
a whole,and theyare evidentlyto be consideredas being acquired
charactersand not of a phylogeneticvalue. This is provedby
3•8
VONIHERING,tit'ology
Ofthe 2"yrann,'dce.
ß
f.July
the fact that in the speciesof Cn•aleg•us,Zichenaps,
and othersin
whichthe malesare wholly black, the femalesand youngare of a
brownishcoloror have a spotted plumage. Amongthe more or
lesssimilarmembersusually placed in this subfamilytwo monotypic genera are completelydifferent in their coloration,namely,
SiwiOygz•
and Machetornz•,
which in my opiniondo not belongto
this subfamily,but to the Elainein•e. 3[achetornisseemsto me to
be allied to Pitangus,and Sz?apygis
to A/fianectes,
Capsiempis,
and
similar genera. While 3[achetornis,
at least in its mode of life,
resemblesthe T•eniopterina•,Sz?apygt?
inhabitsthe woodslike the
Elainein•e.
That the Platyrhynchin•ereally consistof two differentsubfamilies, Euscarthmin•eand Serphophagin•e,
we have shownabove.
With the biologicaldifferences
correspond
suchimportantmorphological ones, principally those of the form of the bill, that the
separation here proposedwill probably be accepted as being
naturally founded.
In order to obtain a natural classificationof the Tyrannida• it
is necessaryto get an idea of the phylogeneticdevelopmentof
the family. In this respectthe Tyrannin•e,judging from their
largedimensionsand their large, somewhatdepressedbills, donot
representthe original form, but, as I think, an extremebranch of
the family. Otherspecialized
branchesare foundin the Euscarthminceand T•eniopterina•. The latter offer not only a coloring
somewhatuncommonin this family, but also casesof decided
sexual dimorphism,which evidently representsa specialization
acquiredwithin the subfamily.
Excludingfrom the Elaineinaethe Pitanginae--large birds with
strongbills that biologicallymuch approximateto true Tyrannin•e
-- the Elainein•eevidentlyrepresentthe groupmostnearly allied
to the ancestorsof the Tyrannida•. These forms are also those
which have the nearest relations with the Pipridge. Strongly
developedsyndactylism,
whichis one of the charactersdistinguish-
ingthe latter,is alsoveryremarkable
in manygenera
of the
Elainein,e,as for examplein the genus 2'yranniscus.
Amongthe Piprida•the samefact is observableas in the Tyrannida•,namely,tb•tt sexual dimorphismin colorationexistsonly in
the more highly organizedforms. In the subfamilyof Piprinse
¾ol.
XXl']
19o4
.•
Vo• I•w•,
Bt'olo•y
oftheTyrannld,
e.
319
the very striking and beautifulcolorationis found only in the
adult males,while the femalesand young malesretain uniform
olive colors, and it is alsoonly among the adult males that we
meet with suchabnormal charactersas enlargedstemsof the primariesand secondaries,
erectfrontal feathers,and elongatedtailfeathers,while the Ptilochlorin•eresemblethe Elainein•enot only
in coloration,but also in the rather small and bristled bill.
These
factsinduceus to concludethat the Pipridgeand Tyrannidmhave
descended from a common ancestral form, the nearest relatives of
whichare the Elaineintoamongthe Tyrannidmwith the Ptilochlorinzeamongthe Pipridze. The commonancestors
musthave been
birds of small size, with pronouncedsyndactylismof the outer
toes, with rather small, somewhat compressedand bristled bill,
andof uniform
olivecolor. Thefrequent
occurrence
of •tyellow
coronalpatchamongthe Pipri&e as well as the Tyrannidmleads
us to supposethat this ornamentmay havebeentransferredfrom
the common ancestors, which were inhabitants of the woods.
From the Elaineine branchof the Tyrannidmoriginated,besides
the Euscarthmin•e
and Serphophaginm,
whosebiologicalconditions
are nearly the same, two great sectionsof inhabitants of the
campos,mostlylarge-sizedbirds, the Tzeniopterinm
and the Pitangine-Tyrannin•e.
With thesegeneralresultsthegeographical
distributionaccords.
As is generallythe casewith the wood-inhabitingbirds,the distribution
of the Elainein•e
of Brazil is a somewhat restricted
one.
While a numberof spediesare distributedthroughthe forestregion
o[ Brasil,only a few rangethroughGuiana and Central America
to Mexico. The Euscarthmin•ein this respectalso do not diverge
much from the Elainein,e, but the Serphophaginae,preferring
open plains and river banks, do occurnot only in the campos
but the maj6rity of them is restrictedto the Andine Region.
These two groups of camposinhabitingTyrannidm are wholly
differentnot only in their way of life but alsoin their geographical
distribution.
The habits of the T•eniopterin•eare terrestrial. They run on
the ground and have in relation therewith elongatedtarsi and
strongambulatorialfeet, seekingtheir insectfood on the ground.
They inhabitthe pampasand the camposof centralBrazil,being
320
VoNIH•,RnqG,
)9œ01o•ry
of theTyrannt'clce.
[-Auk
L July
representedin the littoral zone by but few species. On the other
hand, manyspeciesand genera are adaptedto live in the Andes,
wherethey occur from Patagoniato Colombia,but no speciesof
these Andine forms passesinto Mexico and Texas. For this reason I think it to be right to separatethe genusSayarnisfrom the
T•eniopterin•e,and to unite it to the Tyranninm,in the societyof
which it is found in North
America
and from which it does not
differ regardingits biology.
The Pitangin•eand Tyrannin•e,on the contrary,are of verywide
geographicaldistribution. Thougt• preferringthe campos,they
avoidthe treelessplains. They are not ground-walkers,
but capture insects as they fly like Flycatchers. They are very active,
courageous
birds of large size and goodflight, and their geographical distribution therefore,as a rule, is very wide, some of them
occurringfrom Argentinato North America. Amongthe seventyeight speciesof Tyrannid•e living in the State of S. Paulo fortythree belong to the Elaineintoand the allied groups of arboreal
life, and of theseten, or 23 per cent, have a relativelywide geographical distribution. Among the six Pitangin•eonly the two
speciesof Canopiasand Sirystesare restrictedto Brazil, while the
species of Zegatus, 3[yiazetetes,fitangus, and Myiodynas/esare
representedeven in the southernparts of North America by the
samespeciesor by little different local races. Amongthe sixteen
Tyrannina• of S. Paulo all have a very extensivegeographical
distribution except Blacicus cinereus(Spix) and 2ryrannusalbagularis Burm., so that more than 80 per cent of the Pitanginae
and Tyranninm of S. Paulo have very wide geographical
distribution.
These facts of geographicaldistributionshowus that the only
systemof nomemclaturewell applicable to the discussionof zo6geographicalproblems is the trinomial.
The use of binomialsas employedin the excellentHand-list of
Dr. Bowdler Sharpe may be more advantageousfor collection
purposes,but it combinesin a very inconvenientmannerwelldefinedspecieswith local races. Such facts as the vast distribu-
tionof _tailangus
sulphuralus(L.) and Myiazetetes
sireills(Spix)
are completelyhiddenby the useof binomialnomenclature.
It is alsoamongthesebirdsthat wemeettruemigratoryforms,
Vol.
XXi-]
x9o4
-1
YONIHERING,
Bœolo•y
oftheTyrannœdte. 32I
so far as suchexist among the Tyrannid•e. This fact is in intimate relationwith the special biologicalconditionsof the campos.
No migratorybirds at all exist amongthe wood-inhabitingTyran~
nidae,nor amongthe Pipridae,Formicariid•eand other families of
the forests.
True migratorybirds are scarcelyrepresentedin SouthAmerica
and are essentiallyrestrictedto two familiesof insectivorous
birds,
the Hirundinidaeand the Tyrannidae. In SouthBrazil, from Rio
Grande to S. Paulo, I have observedmigratoryhabits in the following species:-Myiodynasles
solilarius(Vieill.).
]>yracephalus
rubineus(Bodd.).
2"yrannus
melancholicus
Vieill.
Muscivoratyrannus(L.).
As I am preparing&paperonthissubjectto be published
in
' Aquila' I will not discussit in the present paper. It is evident,
however,from the precedingdeductions,that in biological respects
the family of Tyrannidaeis one of the most interestingof the
NeotropicalAvifauna, strongly contrastingwith the uniformity
which in this regard prevailsin mostof the other characteristic
families.
Although the object of this essaywas only to refer to some
general biological features and habits common to certain sub-
families I neverthelessthink it useful to give briefly the results
of my observations,as they may be of service to a subsequent
worker who will undertakethe necessarysystematicrevisionof
the family Tyrannidae. They are as follows:
(•) The Ta•niopterina•represent a very natural systematic
group but as usually arranged include some strange elements,
such as Sayornis, Sisopygis,and probably Jkrachetornis,
which
should be removed to other subfamilies.
(2) The Platyrhynchina•
of the systematicarrangementof Mr.
Sclater contain two quite different sections, the Euscarthmina•
and the Serphophagina•.
(3) The Elaineina•contain someaberrant formswhich should
be removedto other subfamilies. For example,the genusRhynchacyclus
shouldgo to the Euscarthmina•,
and the generaZe•atus,
•yiozetetes, Conopias,]>itan•us,Sirystesand •yiadynastes should
form a subfamily,Pitangina•,a sectionwhich biologicallyis inti-
322
T•xv•gN•g,
TheOrt'gqn
ofMt•ralœon.
I-Auk
L July
mately related to the Tyrannin•e, while morphologicallyit is
intermediate
between
the latter
and the Elainein•e.
(4) The Tyrannin•e form a natural sectionwith which perhaps
the Pitangin•eshould be united.
S. Paulo, Brazœ1,9 Arov.,•9o3ß
A
DISCUSSION
OF
THE
BY
ORIGIN
OF
MIGRATION.
P. A. TAVERNER.
ONE of the first, if not the very first, phenomenaof animate
nature to be noticed by primeval man, must have been that of
migration; and from that day to this it has been,to a greater or
lessextent,a subjectof great interestto students. In the present
day it has been approachedfrom many different sides, and
thoughmany pointshave been pretty well cleared up, othersare
still envelopedin a haze throughwhich the fundamentalprinciples are but barely visible,while others still remain shroudedin
a dense,impenetrablecloudof mystery.
The methodsby whichbirds find their wayto far distantpoints,
the mannerof their migrations,etc., lie without the scopeof this
paper, and will not be referred to here. Upo.n these points we
all await the publicationof the resultsof the investigations
now in
progress,whenprobablymany obscurepointswill be clearedup.
Migration consistsof two movements,
one in the spring,away
from the winter station; and the other in the fall, towardsit again.
The reason of the latter
is self-evident.
There
is a lack of food.
If they did not returnin the fall they would perish of hunger,if
not of cold. From general observations,it seemsas if the former
had a larger influence than the latter, and it is the northward
movementthat needs explanation. Why should a bird leave a
warm land of plentyto journey to a countrybut half recovered
from
the
frozen
embraces
of
an
arctic
climate?
It
seems