The Feather Tracts of Swifts and Hummingbirds

Home , Alula

68 CLARK,Pterylosis o/Swt/ls and Hummingbirds. [Jan.[Auk bird. There is no evidencethat griseusinterbreeds with marianet, and I think it shouldbe givefull specificrank. The breeding rangeof griseusextends along the South Carolinacoast as far no•Xhas the •nouthof the SanteeRiver. A glanceat the •nap of South Carolina, will showthat thcre are no salt m•rshcsof any ' extentfrom Georgetown to Southport, N. C., in whichthis wren could breed.

THE FEATHER TRACTS OF SWIFTS AND HUMMINGBIRDS.

BY HUBERT LYMAN CLARK.

Plates II and III.

SOMEyears ago I undertookto obtain•naterial for a studyof the arrange•nentof thefeather tracts in the Swiftsand Hmmning- birds. Through the kindnessof the authoritiesof the United States National Museran, the alcoholic •naterial in that collection wasplaced at •ny disposal,and wascarefully exmnined. Later on, stonebeautiful hunnningbird•naterial frmn Arizona carne into•ny possession through the effortsof Mr. R. D. Lusk,and in 1897,Mr. C. B. Taylor of Kingston,Jmnaica, presented •ne with somevery valuable speci•nens of both swifts and hmmningbirds. In April,1901, a briefstate•nent appear.ed in 'The Auk' concern- ingthe conclusionsto which the study of this•naterial had led •ne, and a •noreextended reference to the•n appearedin 'Science'for January17, 1902. The preparationof theentire report, however, wascontinually postponed in the hopeof obtaining•nore speci- •nens,and in July,1905, through the kindness of Dr. Wit•nerStone, stonealcoholic hmmningbirds frown Brazil wereloaned •ne by the Academyof Natural Sciencesof Philadelphia. As thereis little Vol.1906 XXIII] j CJ,AI•Xr,Pterylosis o[Swi]ts and Hummingbirds. 69 probabilityat presentof securingin this countryfurther material of importance,I havedecided to delayno longerthe publication of this account of the work that has been done and the conclusions reached. It isan interesting fact that t•he pterylography of no group Of birds has received so much attention as has that of the swifts and hummingbirds.•The classicwork of Nitzseh('40) dealsquite fully with the matter,so far as his materialand methodsallowed. Shufeldt('88) has given a veryfull account6f thepteryloses of both swiftsand hummingbirds,making use howbver of Nitzsch'sfig- ures,which are, unfortunately, very inaccurate. Lucas ('92) has describedthept•rylosis of the hummingbirds in Ridgway's paper on that group,giving figures of the dorsaland ventralaspects of Florisuyamellivora. Thompson(:01) has describedin full, and figuredwell, the pterylosisof Patayona9iyas, comparedit with bothCaprimul.qus and Colloealia, and given a figureof thedorsal aspectof the latter. His conclfisionswere criti•ised by me'in 'Science'(January 17, 1902) and Shufeldtpublished a counter criticismin 'Thd Condor'(March-April, 1902), to whicha brief replywas made in the nextissue of the samejournal (May-June, 1902). Discussioncenters around the questionwhether the swifts andhummingbirds have essentially distinct and uniquepteryloses or not,and to thisquestibn we shall return when we have completed an examinationof the evidencein the case. The pterylography of the sw'iftswill be consideredfirst, a brief generaldescription of the cypselinepterylosis being followed by a ghortaccount of each'genusexmnined. The accountof thehummingbirds follows in the sameWay, after whichtim tw6 groupsare' compared, and final conclusions stated. In this connectionI Wishto thank Mr. Lusk, Mr. Taylor, 1Mr. Stone,and especially Mr. F. A. Lucas,for thehelp they have given me in securingthe neededmaterial, without which nothing could havebeen done, andto Mr. Stoneand Dr. J. A. Allerl,I amindebfed for assistance in nomenclature. 70 C•,At•K,Pterylosis o[Sw•[ts and Hummingbirds. [-AukLJan.

CYPSEIA.

The generalpterylosis of the swifts is verydistinctive and rE- xnarkablyuniform, so that the figuresof Choeturapelagica (Plate II) will answerwith slightchanges for any of the otherspecies. The tractsare all clearlydefined and well feathered,while the apteftaare uniformly naked. The foliehead isquite thickly feathered,but with largecreseent-shaped aptefta over the eyes,while the crownand occiputare rathersparsely leathered, in longitu- dinal rows. The throat is fully featheredbut there are a:pteria alongthe raxniof the lowerjaw. The uppercervical tract is very broad,and unitesalong the sidesof the neckwith the lowercervical tract, so that there are no lateral cervicalaptefta. Between the shoulders,it bifurcatesto form the dorsaltract. On the anterior part of the neck,close to the head, is a large and very evident apteriron,ode of the mostcharacteristic features of the pterylosis. It is a verycurious fact that Shufeldt('88 and :02) positivelydenies the existenceof thisapterium in the swifts,while I,ucas ('92) states that "soxneof the swifts"possess it, and Thompson(:01) xnakes no referenceto it in his descriptionof Collocalia,though his figure showsan entireabsence of featherson the anteriordorsal part of the neck. A carefulreSxamination of xny material,figures and notes,leaves no doubt whateverof the existenceof this apterium in everyspecies of swift examined. I can only accountfor Shu- fehlt'sstatements by supposingthat they are basedon his exaxni- nationof Choeturapelagica and A•ronautesmelanoleucus: in Choe- turathis apteriron is notvery large, while in A•ronautesit is smaller than in any other swift. It is not inexplicabletherefore that it wasoverlooked in thesecases, but it is hard to seehow it couldhave beenoverlooked in Cypseloides,which he alsoexamined.--The dorsal tract is divided from the shoulders to the middle of the sacrmn,by anotherlong but narrowapterium, il•to two narrow bandsof nearlyuniform width, but it endsin a singlenarrow band at the baseof the oil gland. The anteriorportion of the spaee betweenthe dorsal and humeral tracts is sometimesseparated from the rest so as to form a s•nallbut quite distinctapterium on eachside of theupper cervical tract. This is wellshown in Cha•- Vol.1000 XXIIF[ / C•mK,Pterylosis o[Swi[ts and Hummingbirds. 71

tufa. The fernoraltracts thoughdiffuse, are very evident,and are usuallylong, so that they extendbackward almost or quite to the tail-coverts,while anteriorly they reachdownward and forward, often to the sternalsand alongthe sidesto the posterior end of the humerals. The latter tracts are narrow or of moderate width, passingover the humeruseither at its head or near the middle,and are clearly connected with the feathering on thepata- gia of thewings and evenwith the uppercervical tract. The large oil glandis almostsurrounded by a smalltract, chiefly made up of the upper tail-coverts. The lowercervical tract is forkedquite far forwardon the neck, and each half passesinto a broad sternaltract which continues without interruption,though becoming narrower, to the belly, whereit endsrather abruptly. In manyswifts just at the point wherethese pteryl•e leave the neckand passon to the breast,an anteriorlydivergent fork of the ventral apteriumnearly severs them; thisis mostevident in Chazturaand Hemiprocne. In some of the large swifts (Hemiprocne,Cypseloides) scattered feathers in the regionof the furculumtend to unite the anterior endsof the 2 sternaltracts. There is nothingelse peculiar or specially noteworthyin the ventralpterylosis, but the sternalsare connected anteriorlyover the shoulderswith the humeraltracts and posteriorly, occasionally,with the fernorals. The characteristicand importantfeatures of this pterylosis maybe summarized as follows: the presenceof supraocularapteria, an uppercervical apterium, and a long and conspicuousspinal apterium,the absenceof lateral cervicalapteria, the large and diffusefernoral tracts, and the continuityand markeddevelop- ment of the lower cervical,sternal and ventral tracts. In addition to thesefeatures, the followingpoints are noteworthy. Aftershafts are presenton the contourfrethers and are oftenvery large. True downfeathers are wantingand filoplumesare not notablyabundant, thoughusually evident on the neckand the posteriorpart of the back. The oil glandnever bears a terminaltuft of feathers. There are always10 rectricesand 10 primaries. The alula consists of 2 or 3 feathers and in some cases there is also a distinct claw. The secondariesare usually8 or 9 in number,and the wing quineubital,but in Cypseloides,Hemiprocne and Mactop- 72 Pteryl•siso/Swi/ts and Hummingbird,•. ,, LFAukJan. teryxthere are 10 or 11sedondaries &nd fhe wing 'is aquihcubita!. Exceptin Macro?teryxand C•llocalia, the skin'on the hand is verydark, and thi• darkeningextends in somecases'not only along the forearmbut nearlyto the shoulder. The aboveaccount of the cypselinepterylosi• is basedupon the examinationof 15 •wifts,representing 10 spediesand 8 g6nera. Nitzsch('40) studiedH•miprocne collaris Tdmm., H. acuta(C•/ps. acutusPr. Max v. Nen Wied.) and Cypselusapus; Shufeldt('88), Chcetura]•elagica, A•ronautes melano'leucus, and Cypseloidesniger; and Thompson(:01), Collocaliaepodiopyga; but sofar aSI know, thepterylosis has not been examined in anyOthers of'the Cypseli. The generawill now be consideredin detail. '

MaCropteryx.

It is especiallynoteworthy that the pterylosisof' this interesting genusdoes 'not differ in any essentialpoint from that of Chetura. The nuchalapterinto is .presentbut is quite small,as the upper cervical tract is comparativelylong and narrow. The fernoral tractsare notablystrong posteriorly. The legsare featheredabout to the tarsal joint. The skin on the hand is not dark. The 10 rectrices,10 primaries,1l secondaries,and3 aiulafeathers are not peculiar,'but the wing is aquincubital.

Specimenexamined.

No. Name. [ Collection. Condition. 1 Macropteryxcoronata(Tickell): U.S. Nat. Museum. iAlcoholic; fair.

0ollocalia. '

The singlespecimen of thisgenus whi'ch I wasable to exmnine wasin poorcondition, but the onlypoints in whichits pt•rylosis appearedtodiffer from that of Cha, tura is in }hdshape of the upp6r cervicaltract, whichis longerand narrower,'and the nudhalapte- Vol.1906 XXlII'] j CLARK,Ptery•osis o/Swifts and Hummingbirds. 73

rium, whichis correspondinglynarrow. The tarsusis feathered part way down in front. The skin on the hand is not specially (lark. There are 10 rectricesi10 .primariesand 9 secondaries and the wingis quincubital. The alula consistsof 2 feathers.

Specimenexamined.

Name. Collection. Condition.

Collocaliasp. ? U.S. Nat. M•seulm Alcoholic; poor.

Thompson(: 02) has given a verydetailed account of theptery- losisof the headand wing Of Collocaliaand a somewhatbriefer descriptionof that of theb.ody. He hasalso figured the dorsal aspectof the pluckedbird. So far as the headand wingare con- cerned,his figureand descriptionsare verygood, but it is neces- sary to take exceptionto his statementsconcerning the cervical andfernoral tracts. He saysthat the dorsaltract "runs morethan half wayup the neck"but doesnot mentionany connection with the pterykeon the head,and his figuredoes not showany; either his specimenwas peculiarand I must believe,abnormal, or else he has overlooked the real connection which exists. He also states that the "pectoraltracts are .... separatedfrom the featherlngof the neck"; if that is correct,I mustconsider the specimenstill •nore unique. The stateme,fftha•'there is "no well-definedfern- oral tract" is not socontradictory to his ownfigure and to what I havedescribed ahoy% as at firstappeam, for on page 324, he refersto a "lumbar"tract in Uolloqalia,which indicates that he thereuses that term for whatis, in thispaper, called "fernoral"; theword"lumbar '• is not used elsewhere inhis paper. It isgreatly to be r•gr•ttedthat tt•es e discrepancies should occur between hispaper and my observations.Further study of the pterylosis of this genusis very desirable. 74 c•A•,, Pterylosiso[Swi[ts and Hummingbirds.

Ch•etura.

(Plate ii, Figs. I-3.)

The gener•dpterylosis is bettershown by the figuresgiven than by any amountof descriptions.The humeraltracts are somewhat narrowerthan in somegenera and lessclearly connected with the fernoral. The dorsalbands are unusuallynarrow, especially posteriorly. The legs are featheredto the tarsal joint. The skin on the hand is very dark, bat not on the forearm. There are 10 rectrices,the first longest,the fifth shortest. Of the 10 primaries, the first is longcst. There are 8 or 9 secondaries,but the first 6 are much longerthan the others,and the ninth, whenpresent is very smallindeed. The wing is quincubital. The alula consists of 2 feathers.

Speci•ae•sexami•ed. No. Name. Locality.Collection. Condition. 1 Cheeturapelagica(Linn.). U.S.Nat.Mus.Alcoholic;fair. 3 Amherst,Mass.[H. lj,.C. " " •t Olivet,Mich. [ " verygood.

Hemiprocne.

The generalpterylosis is essentiallythe same as in Cbo•turo. The ventralcervical apterlure is verywell defined,and is separated fromthe greatventral apteriron by numerousfeathers in the region of the furculum. The fernoraltracts are very broad and extend backwardto the upper tail-coverts,but their anteriorprolonga- tion on the sidesis not speciallynoticeable. The legsare feathered to the tarsaljoint. The skin of the hand is very dark, but that of the forearmis lighter. There are 10 rectrices,of whichthe outerones are longestand the middlepair shortest. Of the I0 primaries,the tenthis the longestand the othersare successively shorterto the first. There are 10 secondariesand the wing is aquincubital.The alulaconsists of 3 feathersand a distinctclaw.

Vol.1900 XXIII] j CLAR•,Pterylosis o/Swi/ts and Hummingbirds. 75

Specimenexamined.

No. Name. I Loctdity.Collection. Condition. 1 Hem•procnezonaris(Linn.)[Jamaica. [U.S.Nat.Mum Alcoholic;very good,

Nitzschrefers to the "lange,kriiftige, dichtfiedrige" fernorals, and to the width of the branches of the dorsal tract, "ziemlieh breit" in collaris,"zweireihig fledrig" or "einreihig"posteriorly in acura. In both, he says,he found "sechszehnSchwingen," whie•his probab]ya mistake.

Cypseloides.

(Plate II, Figs. 4-5.) The generalpterylosis' of this genusis not essentiallydifferent from that of Ch•etura,but the tractsare wider and morethickly feathered. The legsare leatheredto the tarsaljoint. The skin .onthe handand forearm, and even on the upperarm, is verydark. There are 10 rectrices,10 primaries,and 10 secondaries,and the wing is aquineubital. The alula consists of 3 feathers and a little claw.

Specimenexami•ed. No. Name. Locality.Collection. Condition. 1 Cypselo•clesniger(Gmel.).] Jamaica. U.S.Nat.Mus.'Alcoholic; verygood.

Tachornis.

In this genus,the nuchalapterinto is small,but otherwisethe pterylosisis remarkablylike Cb,etura. The legs are sparsely leatheredto the verybase of the toesin front, but the tarsusis bare behind. The skin on the hand is ahnostblack, and the darkened color extendsnearly to the shoulder. There are 8 rectricesand 76 CLARK,Pt'erylosis o]Swi]ts and Hummingbirds. L Jan.

10 primaries; of the latter, the ninth is 1he longest,and eveuthe eighth is longer than the tenth. There are 10 secondarles,the wing is quincubital,and the alula consistsof 2 feathers.

Specimensexamined.

No. Name. [ Collection. Condit•ion. I Tachore,is parva (Licht.). U.S. Nat. Museum.. Alcoholic; fair.

A•ronautes.

This genusis notablefor the very small nuchal apteriumbut in otherrespects is very si•nilarto Chastura.The legsare leath- eredin front to the baseof the toes'. The skin on the hand is xery dark. There are 10 rectrices,and 10 primariesof whichthe niuth is longest. The wing is quincubitaland has 9 secondariesand 3 feathers in the alula.

' Specimensexamined.

No.' Name. [ Collection. Condition. I ••,ti:ronaut•s-- melar•olcucuskBMrd) U.S.Na?. Mll setlln. Alcoholic; g,o•od

Micropus.

The humeral tracts in this genus are somewhatpeculiar, for in melbathey see•nto have a small horizontalbranch on the inner sideextending in towardsth• do•,sal,while in asquatorialisand st•'cubeli,they passvery nearlyover the head of the humerusand thus are nearerihe dorsaltract }hah in Chastura.In melba,the nuchalapterlure is'short and wide, but in the'others is long and narrow.The legs are leathered tothe toes in front, in a•'quator•alis andstreUbeii, butonly about half way down the tarsus in melba. The skin on the hand is very dark but not on the forearm. Of Vol.looo XXIII] ] CLXRK,Pterylg•is o[Swifts and Hummingbirds. 77 the 10 rectricesthe middlepair is shortest,the outer oneslongest; in .streubel•,howe. yet, thereis a sexualdifference, for the tail is only a verylittle forkedin- the female,while in the •nalcthe fork is deep and the feathersnarr0wcd: Of the 10 primaries,either the ninth or tenth •nay be longest..The quincubitalwing has 8 or 9 secondariesand the alulais madeup of 2 or 3 feathers.

,Specimensexamined.

No. Name. Collection. Condition.

Alco,h,olic; go,,od. 21 Micropus" •quatorialismelba(Linn.). (Milli.). I]U.S. Nat,• kV[useum. 3 •" cafferstreubkli(Hartl.)(• I " 4 '5 .... Q[, "

Cypselus.

Nitzschdescribes and figtires the pterylos!s of Cypselus apus, but the arrangementof the feathers,and tracts on the head and neckare certainlynot as hc shows-it,unless wc arc to believethat this genusdiffers radically from all otherknown swifts. He says there arc "achtzchnSchwingcn," which is probablycorrect.

TROCHILI.

The hum•ningbirdsare anothergroup in which the pterylosis is rc•narkablyuniform and quite distinctive.so that the figures given of Trochilusalexandri will answerwith slightchanges for any otherspecies. The tractsare well definedand the aptcriaare perfectlybare. The headis not thicklyor uniformlyclothed but on the contrarythe featherstend to form longitudinalrows or narrowtracts with moreor lessdefinite aptcria between. Thmnp- son(: 01) figuresand describesad•nirably no lessthan 10 apteria on the headof Patagona,and otherhummingbirds show the same generalarrangc•ncnt. Of theseapteria, the •nost i•nportantare thoseto whichhe givesthe na•nes/ro•tal and s•praocular,both [Auk 78 C•,•aK, Pterylosiso[ Swi]tsand Humminjbirds. I.Jan.

beingcharacteristic of the Troehili. The upper cervicaltract is completelyunited on eachside of the neckwith the lower cervical for a considerabledistance, so that there is a distinct lateral cervical tract for some little distanceon the side of the neck, which ultimatelydivides to form the real eervieals.The conspicuous and characteristicnuehal apterium occupiesthe upper surface of the greaterpart of the neck. The humeraltracts are well defined,rather short,pass over the upper end of the humerus, and at their posteriorend are usuallyvery slightlyconnected by 2 or 3 featherswith the dorsaltract. The latter is very broad, coveringmost of the back, but containsa long and conspicuous spinal apterium. Posteriorlyit unites more or less completely with the fernoraltracts, though the latter are usuallyquite easily distinguished,asthe feathers composing them are larger and stouter than those which connect them with the dorsal. The lower cervical tract is divided as far forward as the angleof the mouth, into two branches,which for a part of their lengthare unitedwith the uppercervical, as alreadydescribed. Each half passesback- ward, over the sides of the breast, as the sternal tract, which is wide and well developed;they terminateso abruptlyjust posterior to the nmrgin of the sternum that there are practicallyno ventral tracts. Anteriorlythe sternal tractsare connectedover the shoulderswith the featheringon the upperside of the wings and thus with the humerals. The characteristicand important features of this pterylosis may be smnmarizedas follows:the presenceof frontaland supraocularapteria, a verylarge and conspicuousnuehal apterium, and a long and conspicuousspinal apterium, the absenceof lateral cervicalapteria, the s•nallfernorals noticeably connected with the dorsal,the well-markedsternals and practicalabsence of ventrals. In additionto thesefeatures, the followingpoints are worthy of note. Aftershaftsare presenton the contourfeathers though very weak. True downfeathers are wantingand filoplumesare not notablyabundant except on the neckand to a lesserextent on the back. The oil gland never bears a terminal tuft of feathers. There are always10 teetricesand 10 primaries;of the latter, the tenth is the longest,the otherssuccessively shorter. The alula is usuallywanting, though a singlefeather is sometimespresent Vol.1906 XXIII'I j ' CLARK,Pterylosis o[Swi[ts and Hummingbirds. 79 indicatingits position. The secondariesare usually6, but not uncommonly7, in number,and the wing seemsto be generally aquincubital,though in severalcases examined it was clearly quincubital. In oneinstance one wing had 6 secondariesand was clearlyaquineubital as in Patagona,while the otherwing had 7 secondariesand was as certainlyquineubital. Apparentlythis pointin wingstructure is no longerof significancein the hmnmingbirds. The skin on the hand is often very dark, and this color extendsmore or less•narkedly onto the forearm. This accountof the trochilinepterylosis is basedon the careful exa•ninationof 31 hummingbirds,representing 17 speciesand 15 genera,and on Thompson's(: 01) veryvaluable study of Patagona gigas,Nitzsch's ('40) figureand descriptionof Chrysolampismo- schitus,Shufeldt's ('88) accountof Selasphorusplatycercus, and Lucas's('92) figure and brief descriptionof Florisugamellivora. So far as I know,no otherhummingbirds have ever been examined pterylographically.The variousgenera will now be considered in detail.

Campylopterus.

Referenceis madeto this genusonly because I,ucas ('92) states that the skin of the hand is "colored black" as in the swifts and some other hummers.

Florisuga.

The singlespecimen exa•nined showed no peculiarities,except that therewere apparently only 9 primaries,doubtless accidcntal or an abnormal condition. The skin on the hand is black.

Specimenexamined. No.[ Name. Collection. Condition. 1 iFlorisugamellivora(Linn.) U.S] Nat. Museum. Alcoholic;very poor.

The figure and d{;scriptionof .this genusgiven by I,ucas ('92) are fairly good,though I find no tcndencyin the sternaltracts to 8O C•,AR•:,...... Pterylosis o[Swi[t•8 and H*emraingbirds. ' L[Auk Jan. connectwith each other near the furcu[um,cid.her in Flo.risugqor anyother hummer, •as'i.ndicated'in Lucas's figu?e. Moreover the headand wing are too closely 'fe•thered, and the humeral 'tracts are not well enoughdefined. Incidentally, ,I wish to dissent stronglyfrom Lucas's statements as to theyeasons for theexist- enceof the variousapter!a: ...

Patagona.

The.pterylosis .of thisgenus, so admirably figured and described by Thompson(: 01), deservesspecial mention because of the presenceof 7 secondariesin the aquincubitalwing, and of a "lateral tract" just outsideof the sternaltract. This small tract occurs in no othergenus of Troehili, so far as I know, but a similartract occursin Todus,according to Nitzsch'sfigure. It would be .in- terestingto know if the tract has any specialsi.gnificanee in Pata- gonaor is eoml•osedof feathersin any way peculiar,or different from thosewhich make up the sternals.

Argyrt•ia. 'Thereis. n9thing exceptional or worthy of notein thisgenus, exceptthat the formulafor the lengthof the rectrices•s 3-2-4-1-5.

Specimenexamined.

l Argyrtriabrevirostris,(Less.). Brazil.Phil. Acad. Nat. Sci. Alcoholic; iair.

Eucephala.

The generalpterylosis is like that of Trochilusalexamtri. There are 6 secondariesin the right wing, with 7 major coverts,and the wing is clearlyaquincubital, but in the left wing there are 7 well developedsecondaries, of nearly equal size, and o.nly7 major coverts,the wingbeing perfectly quincubital. The tail is slightly VoL1900 XXIII-] j C•.ARK,Pterytosis o[Swi[ts and Hummirujbirds. 81 forked, the middle pair of rectricesbeing shortestand the outer pair longest. The legsare featheredjust over the tarsal joint.

SpeOimenexamined.

No.!I Name. l,oeality.Collection. Condition. 1 ?ucephrtlacazrul½o-larvrtta(Gould). I F;razil.iPhil. Acad.Nat.Sci. Alcoholic; poor.

Tha.lura.nia..

Thereis nothingto be saidof thisgenus to distinguishits ptery- lographyfrom otherhummingbirds. The tail is deeplyforked, the oute•feathers being much the longest.

Specimenexamined.

No.i Name. l,()ealitv.i Collection. I Condition.

1 Thalura•ia ,sp.': I Brazil. , Pldl. Acad. Nat. Sci. Alcoholic; good.

La.mpornis.

In this genus,the secondariesare 6 or 7, but the wing is clearly aquincubital. The skin on the hand, and evenon the forearm, is very dark. In the specimensexamined the teetriceswere 2-1- 3-4-5 in one, and 4-5-3-2-1 in the other.

Specimensexamined.

No., Name. i Localitv.i Collection. Condition. 1 I L(tmp,o,rnisviolicauda (Bodd.).' Brazil. iPhil. Acad. NaL Sci.lAleoholic; good. 2 [ I J verypoor. [Auk 82 CLARK,Pterylosis of Swi[tsand Hummingbirds. [Jan.

Chrysolampis.

In his accountof Trochilusmoschitus Nitzsch ('40) refers to the small aftershaft,the absenceof down, the broad, rhombic formof thedorsal tract, the long and rather broad spinal apterium, theslight development of the fernoral tracts, the nuchal apterium, andthe large oil gland. The figuregiven (Plate III, fig. 19) was probablyprepared from a skinand is seriouslydefective in its representationof the tractsand spaces on thehead, neck, and posteriorpart of the back.

• Leucochloris.

There is nothingspecially noteworthy about this genus,except that in both specimensthe right wing had 7 secondariesand the left 6, but all four wingswere clearly aquincubital. In onespeci- men the formula for the rectrlces is 1-2-3-4-5 and in the other it is 3-2-1-4-5.

S?ecimensexamined. Name. Locality.Collection. Condition. (Vieill.).[Phil.Acad. Nat.Sci. Alcoholic;

Eugenes.

The pterylosisof thisgenus is exactlylike Cceligena.There are 6 or 7 secondaries and the formula for the teetrices is 5-4-3--2-1 in the male and 4-3-2-1-5 in the felnale.

Specimensexamined. No.I Na•ne, Sex.Locality. Collection, Condition. l Eugcnes•u•g,e,ns(Swains.). [ • Ariz,,ona. H.L.C.Mcoholic verygood.

•Vol. 10o6 XXIII] ] CLXaX,Pterylos,s o/Swi/ts and Hummingbirds. 83

-CcMigena.

In this genusthe head is more sparselyfeathered than in Pata- gona,with the linesof feathersvery clear,the occipitalapterium is very distinct,the frontal apteriumis very long, and the humeral tract is clearly connectedwith the dorsal. The secondariesare 6 or 7, the wing is aquincubital,and the formula for the rectrices is 4-3-2-5-1.

Specimensexamined. No. Name. Sex.Locality. Collection. Condition. • Cwl{.?(vacle,m?itice(Less.). iz ha. [ H.,!,. C. ,, very,g,ood.

Trochilus.

(PlateIll, Figs.1-3.)

The pterylosisof thisgenus is cicaflyshown in thc figuresgiven, and is an excellentexample of the trochilinecondition. The feet are leatheredto the tarsal joint. There are 6 secondariesand the wing is aquincubital.

Specimensexamined.

No. Name. ] Locality.Collection. Condition. 1[Trochil•s alexandri&Bo•rc.I Muls.' I u. S.Nat.Mus. Alcoholic;good. 3 I " Arizona. H.L.C. x•erv good

Calypte.

The pterylosisof this genusshows no peculiaritieswhatever, nor is there anythingnoteworthy in the details,concerning the primaries,secondaries, alula and rectrices. 84 c•,•,a•,Pterylosis o/Swifts and Humminqbtrds. [Jan.[Auk

Specimenexamined. No.I Name. Collection.Condition. I Calypte anna (Less.). U.S. Nat. Museum. ' Alcoholic; fair.

Selasphorus.

In this genus,the feinoraltracts are extendedfurther backwards and downwardsthan in Inost huinmingbirds,so that lhere is a slightconnection between them and the group of under tail coverts. The legsare leatheredto the tarsaljoint. There are 6 secondaries, and the wing is probablynormally aquincubital. Of the 10 teetrices,the first is longestand fifth shortestin the male, while in the femaleof platycercusthe formula is 2-3-1-4-5.

Specimensexamined.

No. Name. Locality. Collection. Condition.

Arizona. U.S.N:•,t.Mus. Alcol?,olic:g,o,od, 4 ] "plctty½•rcus(Swains.). H. L.C. vers•,good.

Mellisuga.

(Plate III, Fig. 4.)

Although no adults were examined,there is no indicationin the nestlingsthat the pteLvlosisdiffers from that of Trochilus. Even in the einbryos,the characteristicfeatures of the pterylosis are evident. In the nestlings,the leathering of the head shows somepeculiarities, which are illustratedin the figuregiven. Vol.1900 XXIII] j C•,.•m<,Pterylos2.s o[Swi[ts and Hummingbirds. 85

Specimensexamined.

No. Name. I Age. Locality.ICollect!on Condition.

1 Mtllisuga minima (Linn.). Embryo. Jamaica. H. L.C. Alcoholic: good. 3 ; " Ne?ing., i', very,!?od.

Orthorhynchus.

There is nothingspecially noteworthy with referenceto this gentis,for evenin detailsit agreeswith Trochilus. Of the 10 reetriees,the Iirst is longest,the fifth shortest.

Specimensexamined.

NO. Name. Collection. Condition. Or•hort•/nchu.sexilis(Gruel.). U. S. Nat,•Museum. •Alco,h,olic;go, o•d.

Stephanoxys. Thelong plulne on the head is located inthe group of feathers betweenthe frontal and occipitalapteria and its "pit" is very conspicuousin the plucked bird. There is nothingspecially peculiar in the general pterylosis,but one wing is apparently quineubital. The tail formulais 2-1-3-4-5. The feet are leathered just over the tarsaljoint, and the skin on the hand alongthe upperanterior edge, and evenon the forearmis very dark.

Specimenexamined.

No.I Name. Locality.Collection. Condition. 1 Stepha•oxysdelalandi•(Vieill):! Brazil. IPhil. Acad. Nat.Sci.lAlcoholic: good. [Auk 86 CL•nx, Pterylosiso/Swi[ts and Hummingbirds. [Jan.

Genus unknown.

An unidentifiedhummingbird (No. 23493 U.S. Nat. Mus.) showsa pterylosisexactly like the others,and is mentionedhere simplyto completethe list of thoseexamined.

COMPARISON OF TIlE TWO GROUPS.

Beforeentering on a detailedcomparison of the two groups whichwe have been examining,it will set someof the factsmore clearlybefore us if we arrangethein in tabularform.

Cypseli. Tinchili.

No. of primaries 10 10 The three longest 10-9 8 10-9 8 No. of feathers in alula 2-3 0-1 Aftershaft Present Present Condition of oil-gland Bare Bare True down Wanting Wanting ,Conditionof wing Either aquincubital or Usually aquincubital. quincubital. No. of secondaries 8-11 6-? No. of rectrices 10 10 Frontalapt eriu,•,n Wanting Present Occipital Indistinct Supraocular " Present Nuchal " Spinal " Lateral-cervical" Wanting Wanting Fernoral tracts Large, separate from Small, connected with dorsal dorsal Skin of hand Usually very dark Often very dark

It will readilybe seenfrom the table andthe foregoing accounts of thetwo groups that the points of differencebetween swifts and hummingbirdsare comparativelyfew, and the import.ant matter to determine is whether these differencesare trivial or not. These pointsare the presence of frontal and occipital apteria in Trochili, their virtual lack of an alula, their small number of secondaries, their verybroad, diamond-shaped dorsal tract and its practical unionwith the femorals,and their almostcomplete lack of ventral tracts. Let us considerthese points under the three following heads,(1) pterylosisof headand neck; (2) pterylosisof wings; (3) pterylosisof trunk and legs. Vol.1906 XXIII] j CLARK,Pterylosis o/Swifts and Hummi.•.gbirds. 87

Pterylosiso/the Head and Neck.-- In all birds,the distribution of the feathers on the head seems to be more liable to variation than on any other part of the body; amongthe Caprimulgi,for example,we find that the differentgenera exhibit quite a notable diversityin the pterylosisof the head. When we cousiderthere- fore how the whole head of the hummingbirdhas been altered for its peculiarfeeding habits, and how the plumageof the head has been modifiedin connectionwith the developmentof the gorgeousmetallic colors, we shouldexpect a characteristicptery- losis. When on the.other hand, we considerhow nearly in the oppositedirection the modificationof the swift'shead has been, and that there is no developmentof a peculiarhead plumage, it is not strangethat thereare somedifferences in the arrangement of the feathersof the head and neck betweenhummingbirds and swifts. That these differences are due to the modification of the trochilinehead in connectionwith feedinghabits is certainlysug- gested,if not demonstrated,by the pterylosisof the nestlinghum- mingbird. A comparisonof figure4, plateIII, withfigure 3, plate II, certainlysuggests a real resemblancebetween the two. The long,narrow frontal apterium shown by all adult hmnmingbirds, is almostwanting in the nestling,as it is in swifts,while the sparse featberingof the occipitalregion is quite as evident in Chartufa as in Mellisuga. On the other hand, the nestlingshows clearly the supraocularand nuchal apteria, and the absenceof lateral cervical apteria which are such characteristicfeatures of both swifts and hummers. It is a natural conclusiontherefore, that thefundamental plan of the pterylosisof headand neckwas origin- ally similar,if not identical,in the Cypseliand Trochili. Pterylosiso] the Wings.--The general arrangementof the remigesand covertsof the wing is quitesimilar in the two groups under consideration,and such differencesas occur are closely associatedwith the size of the bird. Thus we know that, as a rule, the smaller the bird, the fewer will be the secondaries,lesser covertsand alula feathers,and this is well shownby the swifts and hummingbirds.As regardsthe quincubitalor aquincubital conditionof the wing there is no constantdifference between the two groups,for, as alreadyshown, individual hummingbirds show variabilityon this point, and neither conditionis characteristic 88 CLArtK,Pterylosis o[Swifts and Hummingbirds [AukLJan. of the swifts,as a whole. A comparisonof figures4 and 5 of plate II, with the wings.of Chceturaand Trochilus,will emphasizehow really intermediatebetween the two swifts,in this point,the hummingbird is. There is no essentialdifference in the humeral tracts, thoughthose of the swiftsare more obviouslyconnected with the dorsal and fernoral tracts, than are those of the Trochili. Pterylo,•is'o] the Body.-- In the form of the dorsaltract swifts and hummingbirdsdiffer. In the latter it broadenson the middle of the back,extending downward to the sides,and then narrow- ing againto a point at the oil gland; wereit not for its obvious connectionwith the fernoral tracts on each side, it would thus be an almostperfect diamond in shape. In the swifts,on the con- traw, the tract doesnot extenddownward on the sides nor does it directlyconnect with the fernorals;its shapetherefore is that of an elongatedellipse. In eachcasc, however, there is the well- markedspinal apteriron, a very importantpoint of resemblance. In the swifts,the femoraltracts are well developed,extend forward on the sidesto the posteriorpart of the humerals,and back- wardsto connectwith the uppertail-coverts. In the hummingbirds, however,the fernoralsare weak and only extend towards the dorsalwith whichthey connect. It is evidenttherefore, 'that in the swiftsthe dorsaltract has becomequite sharplydefined, and the femoralsextended to.form a part of the coveringfor the sides,while in the humatingbirdsthe dorsaltract remains more extensive,has not becomeseparated from the fernorals,and pro- videsall the coveringfor the sides. Thus the dorsalpterylosis of the swiftsindicates a greater specializationthan that of the hummingbirds,possibly associated with their extraordinary power of flight. The ventralpterylosis of the Cypscliis essentially identicalwith that of the Trochili,for the apparentabsence of ventraltracts in the latter is a differenceof degree,not of kind, whilethe separationof a lowercervica! apterium from the ventral, in someswifts, is alsoan unimportantcharacter, present in only a few species.

CONCLUSIONS.

In bringingto a closethis account of thepterylography of swifts andhummingbirds, it is worthwhile to considertwo general ques- Vol.1000 XXIII'I j C•.x}m,Pterylosis ofSwifts and Hummingbirds. 89 tions, still open to discussion,which may be stated as follows: Is the pterylosisof a swiftsufficiently similar to that of a hummingbird to givesupport to the viewthat they havea commonancestry ? If so, is that of the Caprimulgisufficiently similar to warrantthe belief that they also have the same ancestry? There are stone other interestingquestions ;vhich arise in connectionwith these, particularlyas to whetherthe swiftsand swallowshave anything in commonand as to what birds are most nearly allied to the groupsconsidered in this paper. But lack of sufficientreliable evidenceforbids any attempt to touch on such questions.Let us turn to thosefirst stated: 1. Is the pterylosisof a swift suf- ficientlysimilar to that of a hummingbirdto give supportto the view that they havea commona•cestry ? In the judgmentof the writer, ¾•s. What may properlybe called the fundamentalplan of the pterylosisis essentiallythe same. That is to say, each group has apparentlysprung from an ancestorin whichthe headwas fairly uniformly,though sparsely leatheredexcept for an apteriumover the eye; the neck had the upper and lower cervicaltracts united anteriorly,so that there xvasno lateral cervicalapterium; there xvasa conspicuousnuchal apterium; the back was well coveredby a broad dorsaltract con- taining an elongatedspinal apterlure; but not fully distinctfrom the femoral tracts or even from the humerals; the lower cervical tract was deeply forked and continuouswith the conspicuous sternaltracts; and the latter were separatefrom eachother, and extendedon to the belly as short,wide, ventral tracts. As regards the dorsaltract, it seemsto me probablethat this ancestralform was a bird in which the humeral, dorsal, and femoral tracts were more or lessconnected, and the two latter at least not sharply defined,somewhat as Nitzsch figuresthe dorsalsurface of Colius; this conditionwas followed by sucha pterylosisas Nitzschshows for Cuculus,in which the spinalapterium has appeared; from this arrangementthe dorsaltract of the hummingbirdwould arise with little change,while the cypselinecondition would follow furthercondensation of the dorsaltract, accompaniedby development of the femorals. In additionto this agreementin fundamentalplan, the swiftsand hummingbirdshave so many details, of more or lessimportance, in common,we may say.further of their ancestralform that it had 10 primarieswith the 3 outermost 90 CLARK,Pterylosis ofSwifts and Hummingbtrds. [Auk].Ja,n.

longest,and 10 rectrices;probably 9 secondaries,an alula of 3 feathers,an aftershafton the contourfeathers, no down feathers, filoplumesrather few and mainlyconfined to the neck and back, the feet leatheredto the tarsal joint, the oil glandwithout a terminal tuft of feathers, and the skin on the hand dark colored. Whether the wing was originallyaquincubital or not is opento questionbut judging from the wingsof the hummingbirdsex- amined,the evidencewould seem to favor the viewthat the quincubital conditionof most of the swiftsis a specializedcondition, and the ancestralform was probablyaquincubital. In the subsequentdevelopment of the Cypseliand Trochili, the pterylosisof the head and neckbecame more specialized in the latter group,while that of the back has beenmore modified in the swifts; with the wings,specialization has occurred in both groups,perhaps to a greaterextent in Trochili,which have lost the alula, and 2 or 3 secondaries,and are apparentlyapproaching the quincubitalcondition. Shufeldt ('88) in attemptingto show the absenceof relationshipbetween swifts and hummingbirds lays muchstress on the differencein the numberof secondaries, the formof the dorsaltract, the absenceof supraocularand nuchal apteriain swiftsand the absenceof fernoraltracts and of black skin on the hand in hummingbirds.As has been repeatedly shownin the previouspages, his positionis clearlybasedon insuf- ficient or unreliable evidence; the number of secondariesis almost whollya matterof size,the form of the dorsaltract is a real dif- ferencebut not inexplicable,the supraocularand nuchalapteria are not absent in swifts, and the femoral tracts and black skinon the hand are not wanting in hummingbirds. 2. Is the pter.ylosiso] theCaprimulgi sufficiently similar t• that o] swi]tsor hummingbirdstogive support to theview that they have a commonancestry with either? In thejudgment of the writer,No. Althoughthe linear arrangement of the featherson the head of the Caprimulgi (seeClark, '94) seemsat first somewhatsimilar to that in Patagona(see Thomp- son, : 01) the morethe pterylosesof the twogroups are compared, the more evident it seems that there is no indication of a common fundamentalplan in the two. The upper cervicaltract in the Caprimulgiis quitenarrow with no traceof a nuchalapterium, Vol.1906 XXIII] j CLARK,Pterylos•s o/Swt/ts and Hummivjlbirds. 91 and is moreor lessseparate from the dorsaltract; th• lowercervical is alsonarrow and not dividedso far up on the throat as even in the swifts. Lateral cervicalaptefta are very evident. The dorsaltract of the Capri•nulgiis clearlyderived from an ancestral form in whichthe anteriorand posteriorportions were not equally well developed;•noreover the spinal apterironis not long and narrow,but shortand wide, with sharplateral angles. Probably the eaprimulgineform is derivedfrom such a conditionas is shown in Nitzsch'sfigure of Falco brachypterus,rather than frmn a form like that of Cucul•ts. The hmneraltracts in Cypseliand Trochili are near the dorsaland there are somefeathers connecting them therewith,but in Caprimulgithe humeralsare narrow and distinct, at somedistance from the dorsal. The form and position of the femoraltracts is clearlydifferent in the Caprimulgifrom thosein the Cypseliand Trochili, and the marked contractionof the sternal tracts to form the ventrals is a minor characteristic of the Caprimulgi. Someof the detailsin which the Cypseliand Trochili agreeare not the samein the Caprimulgi,as for example, the black skin on the hand and the last threeprimaries being the longest; the developmentof the alula and the presenceof 12 or more secondariesin Caprimulgimay alsobe mentioned,though no weightshould bq attached to suchdifferences taken by them- selves.

BIBLIOGRAPHY.

'40. NITZSCH,C. L.--System der Pterylographie. '88. Sa•Jr•L•)T, R. W. -- Studies of the Macrochires,etc.--Jour. Linn. Soc., Zool., Vol. XX., pp. 299-394. '92. Lvc•s, F. A.- Pterylosis (of Trochilidm) in R. Ridgway's The Humming Birds.-- Rep. U. S. Nat. Mus. for 1890, p. 290. '94. CLARK,H. L.--The Pterylography of Certain American Goatsuck- ers and Owls.-- Proc. U.S. Nat. Mus., Vol. XVII, pp. 551-572. :01. . The Pterylosisof Podargus:with noteson the Pterylography of the Caprimulgi.-- Auk, Vol. XVIII, pp. 167-171. 01. THO•rSON,D'ARcY, W.-- On the Pterylosisof the Giant Humming- Bird (Patagonagigas).-- Proc. Zo61.Soc. Lo•)don, 1901, pp. 311-324. 02. CLaRet,H. L.-- Are Humming Birds Cypseloidor Caprimulgoid?-- Science,N. S., Vol. XV, No. 368, p. 108. 02. SH•Jr•L•)T,R. W.--Pterylosis of Humming Birdsand Swifts.- Condor, Vol. IV, No. 2, p. 47. 02. CLaRet,H. L.-- Communication.--Condor, Vol. IV, No. 3, p. 75.