Forelimb Joint Mobility and the Evolution of Wing-Propelled Diving in Birds

FORELIMB JOINT MOBILITY AND THE EVOLUTION OF WING-PROPELLED DIVING IN BIRDS

ROBERTJ. RAIKOW? 2 LESLEYBICANOVSKY, • AND ANTHONY H. BLEDSOE •'2 1Departmentof BiologicalSciences, University of Pittsburgh,Pittsburgh, Pennsylvania 15260 USA, and 2TheCarnegie Museum of NaturalHistory, 4400 ForbesAvenue, Pittsburgh, Pennsylvania 15213 USA

ABSTRACT.--Wemeasured wing joint mobility in penguins,alcids, diving-petrels, and nondiving fliers.Great reduction in mobility of the intrinsicwing jointswas found in penguins, but not in alcids or diving-petrels.This reduction is correlatedwith simplificationof the intrinsicwing musculature.In contrast,alcids and diving-petrels,which usetheir wings in both air and water, retain the full functional capacitiesfor flight. Movement through the air probably requiresa capability for subtleand varied motions,forces, and shapechanges that precludestiffening and simplificationof the wing. Hence, the conversionof an aerial wing to a flipper,as in penguins,must be possibleonly after the evolutionaryloss of flight. Received 23 September1987, accepted 2 February1988.

WING-PROPELLEDdiving has evolved at least in penguins. The limb and its skeleton are flat- five times in birds: in the penguins (Sphenis- tened, and the wing is reduced in surface area cidae), auks (Alcidae), diving-petrels (Pelecan- by the lossof differentiated flight feathers,pa- oididae), the extinct Plotopteridae (Olson and tagia, and the alula. The shoulder joint and the Hasegawa 1979), and the dippers (Cinclidae; extrinsicmuscles of the wing are functionally Goodge 1959). The penguins are flightless, as specialized (Bannasch1986a, b, 1987), and the were the plotopterids,the Miocene Lucasauks limb is relatively rigid becausejoint mobility is (Mancallinae), and the recent great auks (Pin- restricted. In all, the wing is converted to a guinus).The diving-petrelsand all living alcids, "flipper" similar in external form to those of however, practice both aerial and underwater other aquatic tetrapods. locomotion. Simpson (1946: 84-92) maintained No detailed analysesof the relationship be- that living alcids and diving-petrels represent tween wing structure and the idea of compro- a stage in the evolution of flightlessnessand mise has been made for forms such as the div- wing-propelled diving through which the ing-petrels and volant alcids. Hudson et al. ancestorof penguins passed.Storer (1960) sug- (1969) describedthe wing musclesof alcidsand gestedfurther that flying auks and diving-pe- suggested(without functional explanation)that trels represent a "compromise" stage between somefeatures might be related to diving. Spring birds well adapted for locomotion in air and in (1971) describedunderwater swimming in two water. speciesof the alcid genusUria. Storer (1960:fig. Becausethe optimum design of the propul- 4) used illustrationsof wing skeletonsof a gull sire organsis differentfor the two media,which (Larus),Razorbill (Alca), Great Auk (Pinguinus), differ substantiallyin densityand buoyancy,an LucasAuk (Mancalla),and penguin (Spheniscus) intermediateadaptive stage probably would in- to represent stagesin the evolution of the ex- volve a lossof efficiencyin eachmedium as the treme aquatic specialization,showing a trend priceof adequacyin the other.Neither Simpson for increasedflattening of the bones and loss nor Storer elaborated on the anatomical corre- of the alula. Pennycuick(1987) suggestedthat latesof a "compromise"wing. If their views are auksand diving-petrelshave reduced wing areas correct, one might expect intermediate stages asadaptations for propulsion through media of to approach some of the anatomical features of greatly different densities. submarine specialists. Storer did not addressa concomitantspecial- The extreme of underwater adaptation occurs ization, the stiffening of the wing through a 446 The Auk 105:446-451. July 1988 July1988] ForelimbJoint Mobility 447 reduction in the ranges of motion of the joints. that deformation through stretchingof the tissuesdid Stiffening may prevent distortion of the wing not occur during measurement. during movement through the dense aquatic medium. That this is a problem is suggestedby RESULTS the observationthat dippersand aukskeep their wings partly folded when swimming (Goodge Jointmobility.--We obtained arcsof motion at 1959, Spring 1971). four joints for 11 volant speciesthat do not use Studies of underwater propulsion indicate their wings underwater, 2 speciesof diving- that alcidsemploy a different method from that petrels,9 of alcids,and 7 of penguins (Table 1). of penguins. In alcids the manus stays in the For all four motions the values in alcids,diving- flexed position during the propulsivestroke, in petrels, and nondiving fliers were similar and which the wing movesdown and backward in overlappedbroadly. In penguinsthe mean val- a rowing action (Spring 1971). In penguins the uesfor flexion of the forearm,manus, and major manus is extended, the wing is rotated so that digit were substantiallylower than those for the leading edge is lower than the trailing edge alcids,diving-petrels, and nondiving fliers, and in the downstroke, which entails little caudal did not overlap the values for these groups. For movement of the wing, and the upstrokeis used protraction of the humerus, the values for pen- to generatethrust as well (Clark and Bemis1979). guins overlappedthe lowest three valuesamong Thus,with regardto methodof underwater pro- the nondiving fliers and the lowestvalue of the pulsion, alcids and penguins can be considered set of alcids. The penguin average values, how- convergent only in the most general sense. ever, were lower than the values for these We attempted to determine whether species groups.Bannasch (1986a) reported a lower es- capableof both flight and wing-propelled div- timate (45ø ) for this motion but, unlike our ing show attributes of wing structure conver- method, limited the fully protractedposition to gent to those of penguins. We measured and a line perpendicular to the body axis. comparedwing-joint mobility in penguins,oth- The mean arc and 95% confidence limits of er wing-propelled divers, and nondiving fliers, the mean for flexion of the forearm, protraction and related thesefindings to differencesin the of the humerus, flexion of the manus, and flex- wing musculature. The results are discussedin ion of the major digit for each speciesare de- relationto the evolutionarytransformation from picted in Figs. 1-4. The reduction in forearm aerial to submarine propulsion. flexion of penguins differed dramatically from the other species(Fig. 1). This pattern was typ-

MATERIALS AND METHODS ical of flexion of the manusand major digit as well (Figs. 3 and 4). The values for flexion of The range of motion was measuredat the shoulder,. the forearm in alcids and diving-petrels ap- elbow, wrist, and major digit of specimensthat were peared slightly larger on average than for the frozen while freshly dead and subsequentlythawed nondiving fliers (Fig. 1). The penguinsalso had in a refrigerator 2-5 days (depending on size) before a slightly reduced protraction of the humerus measurement.Each specimenwas placed on its back comparedwith the other species(Fig. 2). on a table, and the right wing was extendedflat over a sheet of paper. To measuremotion at eachjoint, the To test for these and other possible differ- proximal element was held securely in place, the po- encesin the arcsof movement, the species(Ta- sition of the joint was marked,and the distal element ble 1) were allocatedto three categories:diving was moved through the range of motion until the nonfliers (penguins), diving fliers (alcids and joint resisted further movement. These points were diving-petrels), and nondiving fliers (remain- marked, and the arc thus circumscribed was measured ing species,including Melanitta nigra, a foot- with a protractor. This was taken as an estimate of propelled diver). For eachmeasurement a group the normal arc of motion. To reduce variation in tech- mean was calculated(Table 2), and a one-way nique, one person (the senior author) moved the ele- analysis of variance was performed. For pro- ment through the range of motion while the other traction of the humerus and flexion of the ma- workersmaintained the specimenin placeand marked the positions. Replicate estimates varied, and each nus, the difference between the mean for pen- measurementwas repeated 10 times and a mean arc guins and that for each other category was of motion was calculated.Although the tissueswere statisticallysignificant (P < 0.05). For flexion of compliant,we observedno pattern of increasein the the forearm and major digit the variancesof the estimatedarc during a seriesof replicates,indicating means were heteroscedastic(Bartlett's box F test 448 R•Kow, B•c^•ovsK¾,^•D BLEDSOE [Auk, Vol. 105

T^BLE1. Range of motion in avian wing joints.a

Protraction of Flexion of Flexion of Speciesno. humerus forearm Flexion of manus majordigit Spheniscidae I Aptenodytespatagonica 103 (97-106) 23 (17-34) 22 (18-26) 12 (11-15) 2 Aptenodytesforsteri 92 (88-102) 14 (13-18) 13 (10-15) 14 (12-19) 3 Pygoscelispapua 65 (57-72) 26 (20-34) 29 (24-34) 25 (20-31) 4 Pygoscelisadeliae I 100 (91-106) 25 (20-32) 25 (22-29) 17 (13-21) 5 Pygoscelisadeliae 2 27 (20-32) 23 (19-26) 19 (11-23) 6 Pygoscelisantarctica 73 (67-82) 25 (20-30) 21 (19-22) 21 (16-27) 7 Eudyptescrestatus 92 (83-106) 34 (28-37) 30 (25-35) 26 (22-30) 8 Eudypteschrysolophus 62 (57-68) 26 (22-31) 27 (24-31) 16 (14-20) Pelecanoididae 9 Pelecanoidesurinatrix 144 (120-155) 149 (142-155) 112 (95-122) 78 (63-91) 10 Pelecanoidesgeorgicus 157 (148-173) 149 (144-156) 106 (99-119) 83 (65-100) Alcidae 11 Uria aalge 158(144-173) 157 (149-165) 109(102-120) 76 (72-84) 12 Uria lomvia 136 (125-154) 152 (145-163) 110 (98-117) 63 (55-70) 13 Cepphuscolumba 114(87-135) 160(158-162) 101(93-109) 53 (47-60) 14 Synthliboramphusantiquus 155 (130-180) 143 (136-152) 107(95-114) 71 (64-82) 15 Ptychoramphusaleuticus 137 (119-150) 161 (151-170) 114 (110-117) 63 (45-75) 16 Cyclorrhynchuspsittacula 124(102-157) 158(153-164) 119(112-128) 67 (56-84) 17 Aethia cristatella 154 (136-168) 142 (135-148) 117 (105-128) 62 (52-68) 18 Fraterculacirrhata 135 (115-148) 155 (147-160) 113 (100-120) 65 (58-68) 19 Fraterculacorniculata 92 (76-131) 158 (154-164) 97 (90-105) 68 (50-75) Nondiving fliers 20 Larusoccidentalis 137 (113-148) 143 (137-151) 119 (103-126) 60 (49-67) 21 Rissatridactyla 159 (150-169) 159 (147-168) 119 (90-135) 56 (46-75) 22 Melanittanigra 129 (102-139) 132 (120-143) 117 (111-130) 58 (40-72) 23 Colinusvirginianus 151(135-164) 155(142-171) 109(92-121) 64 (58-71) 24 Porzanacarolina 141 (130-155) 146 (130-153) 113 (100-130) 77 (65-85) 25 Columbalivia 1 120 (113-130) 125 (117-132) 114 (107-128) 48 (40-63) 26 Columbalivia 2 106 (95-115) 124 (111-134) 134 (125-144) 45 (39-50) 27 Columbalivia 3 99 (87-115) 113 (102-123) 129 (120-137) 57 (51-75) 28 Coccyzusamericanus 149 (135-160) 152 (140-160) 104 (87-117) 81 (64-98) 29 Ramphastosambiguus 126(109-140) 142(126-152) 113(90-138) 66 (59-75) 30 Colapresauratus 99 (88-107) 125 (103-135) 103(95-122) 34 (29-43) 31 Hylocichlamustelina 85 (71-107) 112 (98-122) 99 (91-107) 43 (36-53) 32 Graculareligiosa 145 (120-165) 134(123-141) 116(96-130) 68 (57-79) Values are in degrees,with mean and range for 10 measurementsat eachjoint in eachspecies. for homogeneityof variances;P < 0.05), and significantlyfrom thoseof nondiving fliers and the one-wayANOVA model was thus inappro- of diving fliers (P < 0.01). The difference be- priate.An approximatetest of equalityof means, tween the means for flexion of the forearm in usingthe Gamesand Howell method(Sokal and diving and nondiving fliers was also statisti- Rohlf 1981: 409-410), indicated that for these cally significant(P < 0.05);the value in diving two motions the means of penguins differed fliers was larger than in nondiving fliers None

T^BLE2. Mean valuesin degreesof arc + I SD for joint movementsin diving nonfliers (penguins), diving fliers(alcids and diving-petrels),and nondivingfliers. n indicatesthe samplesize.

Protraction of Flexion Flexion Flexion of n humerus of forearm of manus major digit Diving nonfliers 8 83.9 + 23.3a 25.0 + 5.5 23.8 + 5.4 18.8+ 5.0 Diving fliers 13 126.6+ 23.3 153.1+ 6.6 109.6+ 6.6 68.1 + 8.5 Nondiving fliers 11 136.9+ 20.6 135.5+ 15.4 114.5+ 9.8 58.2 + 13.4 • Sample size - 7; see Table 1. July1988] ForelimbJoint Mobility 449

penguins diving-petrels alcids non-divingfliers Fig. 3. Arcs of flexion of the manus in penguins, diving-petrels,alcids, and volant speciesthat do not penguins diving-petrels alcids non-divingfliers engagein wing-propelleddiving. The format is as in Fig. 1. Arcs of flexion of the forearm in penguins, Fig. 1. diving-petrels, alcids, and volant speciesthat do not engagein wing-propelled diving. The horizontalbars represent mean arcs based on 10 measurements,and Wing musculature.--Penguinspossess virtual- the vertical lines are the 95% confidence intervals of ly a full complementof extrinsicmuscles, those the means. The numbers along the abs½issacorre- that arise on the pectoral girdle or other struc- spond to speciesnumbers listed in Table 1. tures and insert on the humerus (Table 3). These musclescontrol the operation of the wing as a whole, and their retention as a functional sys- of the other differencesamong the means for tem is consistentwith the wide range of move- each motion (Table 2) was statisticallysignifi- ments at the shoulder. cant (P > 0.05). The reduction in intramembral mobility of We were unable to measure accurately rota- penguins is closely correlated with modifica- tion at the shoulder or elevation and depression tion of the correspondingintrinsic muscles(Ta- of the wing. Bannasch(1986a) measuredrota- ble 3). Most of these muscles either are con- tion in penguinsas part of a functional analysis verted to tendons that lack contractile function but did not measureother groups of birds. Ma- or are lost entirely (Schreiweis 1982). These nipulation of our specimenssuggested that none changes, however, occur in a pattern that re- of the specieshas substantial(if any) reduction tainsthe capabilityof providing limited activity of the arc of elevation and depression. at each joint. No joint is immobilized, but each is reduced to a restricted arc of movement (Ta- ble 1).

penguins diving-petrels alcids non-divingfliers Fig. 2. Arcs of protractionof the humerusin pen- penguins diving-petrels alc,ds non-divingfliers guins, diving-petrels, alcids, and volant speciesthat Fig. 4. Arcs of flexion of the major digit in pen- do not engagein wing-propelled diving. The format guins, diving-petrels, alcids, and volant speciesthat is asin Fig. 1. The secondspecimen of Pygoscelisadeliae do not engagein wing-propelleddiving. The format (5) was not suitable for this measurement. is as in Fig. 1. T^BLE3. Musclesof the wing in penguins.

Muscles a Condition b

Humerus Scapulohumeraliscranialis Absent Scapulohumeraliscaudalis Normal Subscapularis Normal Subcoracoideus Normal Coracobrachialiscranialis Weak or vestigial Coracobrachialis caudalis Normal Pectoralis Normal Supracoracoideus Normal Latissimusdorsi pars cranialis Normal Latissimusdorsi pars caudalis Normal Deltoideusmajor Weak Deltoideus minor Normal

Forearm Extension Tricepsbrachii Normal Flexion Bicepsbrachii Absent Brachialis Normal

Rotation Pronatorsuperficialis Absent Pronatorprofundus Absent Supinator Weak Ectepicondylo-ulnaris Tendinous Entepicondylo-ulnaris Absent Manus Extension Extensormetacarpi radialis Weak Extensorlongus alulae Normal Extensorlongus digiti majoris Tendinous Flexion Flexorcarpi ulnaris Tendinous Flexordigitorum superficialis Tendinous Flexordigitorum profundus Tendinous Extensordigitorum communis Tendinous Extensormetacarpi ulnaris Tendinous Ulnometacarpalisdorsalis Normal Ulnometacarpalisventralis Absent Alula Extensor brevis alulae Absent Abductor alulae Absent Flexor alulae Absent Adductor alulae Absent Extensordigitorum communis Tendinous Major digit Extension Interosseus dorsalis Tendinous Abductordigiti majoris Weak Extensorlongus digiti majoris Tendinous Extensordigitorum communis Tendinous Flexordigitorum superficialis Tendinous Flexordigitorum profundus Tendinous Flexion Interosseus ventralis Normal Flexordigiti minoris Normal Functionalgroupings follow Raikow(1985). Data from Schreiweis (1982). July1988] ForelimbJoint Mobility 451

DISCUSSION D. ScottWood, and J. C. Vanden Bergeprovided help- ful criticismsof the manuscript.This materialis based We found that the diving-petrels and extant on work supported by the National Science Foun- speciesof alcids are not intermediate in intrin- dation under grantsBSR-8314729 and BSR-8617896to sicwing mobility between nondiving fliers and R. J.Raikow, and by The CarnegieMuseum of Natural penguins.Furthermore, alcids retain a full com- History under a Rea Fellowship Grant to A. H. Bled- plement of wing muscles, as diving-petrels soe. probablydo, in contrastto penguins,which have a highly reduced set of musclesand but a ves- LITERATURE CITED tige of contractile force at each intrinsic joint. BANNASCH,R. 1986a. Morphologisch-funktionelle Thus, the living alcids and diving-petrels rep- Untersuchungam Lokomotionsapparatder Pin- resent adaptive intermediatesonly in the lim- guine als Grundlagefi•r ein allgemeinesBewe- ited sense that they have balanced a reduced gungsmodelldes "Unterwasserfluges," tell I. Ge- wing area(Pennycuick 1987) with the structural genbaursMorphol. Jahrb. 132: 645-679. specializationsnecessary for flight and the be- --. 1986b. Morphologisch-funktionelleUnter- havioral modificationsrequired for wing-pro- suchungam Lokomotionsapparatder Pinguine pelled diving. In addition, studies of under- als Grundlage ftir ein allgemeines Bewegungs- water propulsion indicate that alcids employ a modell des "Unterwasserfluges,"tell II. Gegen- different method from penguins (Spring 1971, baurs Morphol. Jahrb. 132: 757-817. 1987. Morphologisch-funktionelle Unter- Clark and Bemis1979). Thus, the interpretation suchung am Lokomotionsapparatder Pinguine of alcids as an adaptive compromiseis overly als Grundlage ftir ein allgemeines Bewegungs- simplistic (see also Raikow 1985: 81). modell des"Unterwasserfluges," tell III. Gegen- The lack of wing rigidity and muscularsim- baurs Morphol. Jahrb. 133: 39-59. plification in volant alcids implies constraints CLARK,B. D., •; W. BEMIS. 1979. Kinematics of swim- on the evolution of a flipper-like wing that ming of penguins at the Detroit Zoo. J. Zool. probably relate directly to the requirementsof London 188: 411-428. flight. Maintenance of the pressure gradients GOODGE,W.R. 1959. Locomotionandotherbehavior entailed in flight requireschanges in the shape of the dipper. Condor 61: 4-17. and orientation of the wing surfacesduring the HUDSON,G. E., K. M. HOFF,J. VANDENBERGE, & E. C. TRIVETTE.1969. A numerical study of the wing flapping cycle.These involve rapid and variable and leg musclesof Lari and Alcae. Ibis 111:459- movementsprovided by a combinationof joint 524. mobility and varied muscularactions. Thus, the OLSON, $. L., & Y. HASEGAWA. 1979. Fossil counter- capacityfor flight probablyprecludes structural parts of giant penguinsfrom the North Pacific. modificationslike those of the wings of pen- Science 206: 688-689. guins. PENN¾CUICK,C.J. 1987. Flight of auks (Alcidae) and We submit that the reduction and loss of in- other northern seabirdscompared with southern trinsic muscles,and the stiffening of the wing, Procellariiformes: ornithodolite observations. J. cannot evolve in flying birds. Penguins prob- Exp. Biol. 128: 335-347. ably passedthrough an auklike stage in which RAIKOW,R.J. 1985. Locomotor system. Pp. 57-147 in Form and function in birds, vol. 3 (A. S. King they used their wings for both aerial and un- and J. McLellan& Eds.). London, Academic Press. derwater propulsion, but they must have aban- $CHREI•WEIS,D.O. 1982. A comparative study of the doned aerial flight before they developed their appendicular musculature of penguins (Aves: morphologicalspecializations of reduced mus- Spenisciformes).Smithsonian Contrib. Zool. 341: culature and restrictedjoint mobility. !-46. SIMPSON,G.G. 1946. Fossilpenguins. Bull. Am. Mus. ACKNOWLEDGMENTS Nat. Hist. 87:1-100. $OKAL,R. R., & F. J. ROHLF. 1981. Biometry, 2rid ed. For making specimensavailable we are grateful to San Francisco, W. H. Freeman. Kenneth C. Parkesand D. ScottWood (The Carnegie SPRING,L. 1971. A comparison of functional and Museum of Natural History, Pittsburgh, Pennsylva- morphologicaladaptations in the CommonMurre nia), JosephR. Jehl Jr. and Frank S. Todd (Hubbs-Sea (Uria aalge)and Thick-billed Murre (Uria lomvia). World ResearchInstitute, San Diego, California), and Condor 73: 1-27. Sievert A. Rohwer and Christopher S. Wood (Uni- STORER,R. W. 1960. Evolution in the diving birds. versity of Washington,Seattle, Washington). Parkes, Proc. 12th Int. Ornithol. Congr.: 694-707.