Intraspecific Variation in the Hindlimb Muscles of the Ivory-Billed Woodcreeper and the Blue Jay, with a Review of Other Species

THE CONDOR AUG31 19% A JOURNAL OF AVIAN BIOLOGY UH\~~~~nY Q& lDAH0 Volume 95 Number 3 August 1993

The Condor 95:497-506 0 The Cooper Ornithological Society 1993 ;

INTRASPECIFIC VARIATION IN THE HINDLIMB MUSCLES OF THE IVORY-BILLED WOODCREEPER AND THE BLUE JAY, WITH A REVIEW OF OTHER SPECIES

ROBERT J. RAIKOW,~ ANTHONY H. BLEDSOE,~ TAH~IAT SYED AND ANGELA G. GLASGOW Department of Biological Sciences,University of Pittsburgh,Pittsburgh, PA 1.5260and Section of Birds, The CarnegieMuseum of Natural History, Pittsburgh,PA 15213

Abstract. Systematistshave long used myology in phylogeneticstudies of birds,but have only recentlybegun to assessthe extentof intraspecificvariation in musclesand the poten- tially disruptiveeffect of suchvariation on invsstigationsof phylogeny.We provideinfor- mation on the Ivory-billed Woodcreeper(Xiphorhynchusflavigaster) and the Blue Jay (Cy- arsdtta cristata), and integratethese results with thoseof previousstudies. Bilateraldissection of 19 specimensof X. flarigaster revealedseven major variantsin six different muscles.Ten of the 19 specimens(52.6%) possessedvariant muscles;of these, four were unilateral and seven bilateral. Overall, 1.6% of the individual muscles varied from their species-typicalstates. Six of the 11 variants were found in M. flexor cruris lateralis, suggestingan associationwith that muscle. The 26 specimensof C. cristata exhibited a different pattern. Major variations occurred in only three muscles,but for each of thesea typical condition could not be ascribedbecause variation was extensive and polymorphic. Analysis of similar information for a total of 14 speciesof passerinebirds indicated that variation is common enoughthat many individuals in a sample can be expected to possess at least one variant muscle. However, the probabilities of error in designatingspecies-typical conditions from bilateral dissection of single specimens were low (O.OO-0.06),suggesting that intraspecificvariation does not significantlycompromise phylogeneticanalyses of myological data. Simulation experiments are needed to test this hypothesis. Key words: Cyanocittacristata; Xiphorhynchus flavigaster; myology; variation; phylog eny; passerines.

INTRODUCTION studies of phylogeny (McKitrick 1986; Raikow et al. 1990). In addition, documenting the extent Although systematistshave long used the mus- and nature of species-level myological variation cular system in taxonomic and phylogenetic provides baseline information for discussionsof studies of birds, they have only recently begun the origin and biological meaning of such vari- to assessthe nature and extent of intraspecific ation (e.g., McKitrick 1986; Raikow, in press), (individual) variation in muscles. Such infor- and hence of the origin myological evolution- mation permits the evaluation of proceduresfor of ary novelties. obtaining species-typical myological data from Recent studies (M&trick 1985, 1986; Ber- the dissection of small numbers of specimens. man et al. 1990; Raikow et al. 1990; Rudge and This is important because undetected intraspe- Raikow 1992; Raikow, in press) suggestthat in- cific variation can have disruptive effects on traspecific myological variation in birds is usually infrequent but is occasionally common in I Received 11 January 1993. Accepted23 March specificmuscles (e.g., McKitrick 1986), and that 1993. the muscular system,studied with care, is a source

[4971 498 ROBERT J. RAIKOW ET AL. of accurate phylogenetic information. However, 038806, 045195, 045196, 067423; Yale: 7061, these generalizations are based on relatively few 8778, 8779, 10226, 12935, 12936). Dissection studies,and the information that is available has was performed under a stereomicroscopeat mag- not been integrated to provide a general assess- nifications of 6X-25X. Specimens were stained ment of intraspecificmyological variation. In this with iodine to enhance contrast (Bock and Shear paper, we provide new information on intraspe- 1972). The muscles dissected and their abbre- cific myological variation for two passerine spe- viations are: CF, M. caudofemoralis; EDL, M. cies, the Ivory-billed Woodcreeper (Xiphorhyn- extensor digitorum longus; EHL, M. extensor thus jkwiguster; Furnariidae: Dendrocolaptinae), hallucis longus; FB, M. fibularis brevis; FCRL, and the Blue Jay (Cyunocitta cristata; Corvidae). M. flexor cruris lateralis; FCRLA, M. flexor crur- In addition, we integrate these results with those is lateralis parsaccessoria; FCRM, M. flexor cruris of earlier studies to provide an overview of the medialis; FDL, M. flexor digitorum longus; FHB, current status of the problem of intraspecific M. flexor hallucis brevis (absent in X. flavigus- myological variation in birds. ter); FHL, M. flexor hallucis longus; FL, M. fibu- Our study is modeled on the work of Raikow laris longus; FPD2, M. flexor perforatus digiti II; et al. (1990) and usesthe following classification FPD3, M. flexor perforatus digiti III, FPD4, M. of intraspecific variation. As a first approxima- flexor perforatusdigiti IV; FPPD2, M. flexor per- tion, a typical condition is defined as one that forans et perforatus digiti II; FPPD3, M. flexor occurs in a majority of the specimens of a given perforanset perforatusdigiti III; FTEM, M. fem- species, and a variant as one that occurs in a orotibialis externus et medius; FTI, M. femo- minority of specimens(Raikow et al. 1990:368). rotibialis intemus; GI, M. gastrocnemius pars Minor variants are slight differences in size or intermedia; GL, M. gastrocnemiuspars lateralis; shape that are either nonbiological in origin or GM, M. gastrocnemius pars medialis; ICR, M. too indistinct to assessobjectively. Major vari- iliotibialis cranialis; IF, M. iliofibularis; IFI, M. ations include the following. Incongruous vari- iliofemoralis intemus; IL, M. iliotibialis lateralis; ants are distinctive, functionally compromised ISF, M. ischiofemoralis; ITCA, M. iliotrochan- aberrancies. Mimicking variants are conditions tericus caudalis; ITCRM, M. iliotrochanteticus atypical in one speciesthat closely resemble con- cranialis et medius (in C. cristata, M. iliotro- ditions typical in other species.Singular variants chantericus cranialis [ITCR] and, as a separate are distinctive and functional, but are unknown muscle, M. iliotrochantericus medius [ITM]); L, in other species.Explosive variation is a special M. lumbricalis; OL, M. obturatorius lateralis; casewith a high frequency of multiple variations OM, M. obturatorius medialis; PIFCA, M. pubo- in a muscle. This classification has its limitations ischio-femoralis pars caudalis; PIFCR, M. pubo- because, in principle, variation akin to poly- ischio-femoralis pars cranialis; PL, M. plantaris; morphism in the population genetical sensede- TCR, M. tibialis cranialis. Minor variants, as fies simple categorization. In addition, in prac- they are described in the introduction and by tice a typical state cannot be assignedwhen only Raikow et al. (1990), were not included in our two variants are observed and are of equal fre- analyses. We dissected 1,140 muscles in X. flu- quency. We treat suchcases separately from those vigaster and 1,6 12 in C. cristata, for a total of for which our classification is appropriate. 2,752 muscles. To test for an association between frequency MATERIALS AND METHODS of variation and muscle or individual specimen, We dissected 19 specimens of X. Jlavigasterbi- we followed the procedure described by Raikow laterally. (Specimen numbers are as follows. Car- et al. (1990), which employs an R x C test of negie Museum of Natural History: 1123, 16 15, independence (G-statistic) and Williams’ (1976) 3780,3781,3908,3912,3915,4081,4082,4189, adjustment for small sample size. For thesetests, 4249, 4335; University of Kansas Museum of the columns (C) of the contingency table were Natural History: 040077,051232,056838; Yale “variant” and “not variant.” A specimen was Peabody Museum: 4812,4813,4814,4858.) We coded as variant if it showed either unilateral or also dissected both hindlimbs of 26 specimens bilateral variation. Association with the fre- ofC. cristata(CarnegieMuseum: 858,2379,2497, quency of suchvariation was testedwith the rows 5123,5146,5203,5338,5971,6038,6075,6076, (R) of the contingency table being muscle names 6127, 6192, 6210, 6302; Kansas: 038805, (30 musclesfor Xflaviguster, with M. gastrocne- MYOLOGICAL VARIATION 499 mitts divided into lateral, medial, and interme- pars accessoriawas a strap-shapedmuscle whose diate bellies for the analysis) or individual spec- insertion onto the femur was adjacent to and imen number. For the variable muscles in C. continuous with the origin of M. gastrocnemius cristata, no typical condition could be designat- pars intermedia (Fig. 1A). However, in eight of ed, therefore we did not perform the R x C tests thirty-eight limbs a variation occurred in which for this species. the muscle was slightly reduced, and its insertion For each specieswe calculated the frequency was separatedby a gap from pars intermedia (Fig. of error in the assessmentof species-typicalcon- 1B). This occurred bilaterally in three specimens ditions under the assumption that one specimen and unilaterally in two (Table 1). It may be con- per speciesis dissectedbilaterally. We used the sidered a mimicking variant, as the size of this procedure described by Raikow et al. (1990) in muscle varies widely among passerine groups. which each cell of a matrix of muscle names by A different variant was observed bilaterally in specimen numbers is scored either “T” (typical specimen YPM 4858 (Table l), where pars ac- condition present on both sides of a specimen; cessoriapossessed two distinct bellies (Fig. lC), “T” entries are designatedby blanks in Table l), a singular variant. “B” (a single variant condition present on both M. pubo-ischio-femoralis pars caudalis (Fig. sides of a specimen), “U” (right and left sides 2). In woodcreepersthis muscle typically arises differed), or “d” (damage to one or both sides by a tendon from the ischium, and becomesfleshy precludeddetermination). From the matrices, we at about the level of the pubis or distad to the calculated the frequency of error as the number latter (Raikow, in press). The narrow, strap- of “B” entries (excluding incongruous variants, shapedbelly inserts on the distal end of the fem- which are easily recognized and would not be oral shaft.In the right leg of specimenKU 040077, used in phylogenetic analysis) divided by the however, the muscle had an anomalous extra number of non-“d” cells. For several species, head that arose from the mid-region of the me- information on the number of damaged or un- dial side of M. flexor cruris medialis (Table 1). assessedmuscles was not available; for thesespe- The belly of this anomalous portion extended cies, frequency of error was calculated under the distad to join the distal one-third of the normal assumption that damaged muscles and undam- belly. In addition, the fasciclesof M. flexor cruris aged muscleshad equal frequencies of variation. medialis were distinctly skewedtoward the fleshy attachments of the anomalous portion. If the RESULTS anomalous fibers contracted they would contrib- ute to the normal action of the muscle. Because IVORY-BILLED WOODCREEPER we know ofno other occurrenceof this condition, hf. iliotrochantericuscranialis et rnedius. Mm. it is best considered a singular variation. In most iliotrochantericus cranialis and iliotrochanteri- passerinesthis muscle is much wider than in the cus medius are small muscles that arise on the woodcreepers(Raikow, in press),and so this ab- ventral margin of the ilium and insert side-by- errant form may represent a deformed atavism. side on the femur in most birds. In 18 specimens M. gastrocnemiuspars lateralis. In specimen only a single muscle occurred, which was termed CM 39 15 this muscle exhibited bilaterally a sec- cranialis et medius because it is thought to be ond, medial head of origin in addition to the the product of fusion between the two muscles usual origin on the lateral side of the crus (Table (Raikow, in press; Rowe 1986). However, in 1). This head arosefleshy from the medial surface specimenCM 3780 medialis occurredin the right of the common origin of M. flexor perforatus leg only (Table l), occupying a space between digiti III and M. flexor perforatus digiti IV, lat- cranialis and M. iliofemoralis internus that usu- eral to M. gastrocnemius pars intermedia. It ally lacks a muscle. Inasmuch as the presenceof passeddistad to join the normal part of the belly a separate M. iliotrochantericus medius is the a short distance beyond the level of the iliofibu- ancestral condition in passerines (Raikow, in laris tendon. This extra mass of muscle tissue press), this variant appears to be an atavistic was structurally integrated with the normal part reappearance, and a mimicking variation. A of the muscle and presumably contributed to its comparable variation in another speciesis illus- function. To our knowledge it does not mimic trated elsewhere (Raikow, in press, Fig. 4B). any condition normal in other species,and thus M. flexor cruris lateralis. In most specimens is a singular variant. 500 ROBERT J. RAIKOW ET AL.

TABLE 1. Major variations in the hindlimb muscles of Xiphorhynchusflaviguster. Symbols: B, bilateral; d, muscle damaged;M, mimicking; S, singular;U, unilateral. Abbreviations of muscle names are under Materials and Methods.

Specimen numbers CM CM CM CM YPM CM CM CM YPMYPM KU KU CM CM CM KU CM CM YPM Vari- Muscles 3908 3912 3781 4189 4813 4249 3915 1615 4812 4814 040077 051232 4081 4082 4335 056838 1123 3780 4858 ant Type

ICR d d d 0 IL d IF : ITCA 0 ITCRM d U 1 M IF1 0 FTEM d 0 FTI 0 FCRL B’ B’ B’ U’ U1 B2 6 Ml s* FCRM 0 CF 0 ISF 0 OL 0 OM 0 PIF U 1 s TCR 0 EDL 0 FL FB : GM 0 B 1 s :: PL : FPPD2 B 1 M FPPD3 0 FPD2 0 FPD3 0 FPD4 0 FHL B 1 s FDL 0 EHL L :

Variant with gap at insertion. ?Variant with two bellies.

M. flexor per-ovum et perforatus digiti IZ. In flexor perforatusdigiti IV, just distal to that belly. specimen CM 3781 this muscle was atypically This anomaly occurred bilaterally, and is a sim- wide bilaterally as it arose from the femur and ple example of two adjacent muscles gaining an adjacent lateral collateral ligament (Table 1). This atypical interconnection. It is best considered as is similar to a variation shown elsewhere (Rai- a singular variant (Raikow et al. 1990). kow, in press, Fig. 17D) except that it did not Tendon ossification.Woodcreepers differ from extend far enough to reach the head of the fibula. other passerinesin possessingextensive tendon It is considered to be a mimicking variation be- ossification in the limb muscles. Variation in X. cause it closely resembles the typical condition flavigaster is discussedelsewhere (Bledsoe et al. in treerunners (Rudge and Raikow 1992). 1993) as part of a study of ossification. M.Jlexor hallucislongus. In specimenCM 3780 Analysis of patterns of variation. Seven major an anomalous fusion was observed between M. variations occurredin 6 different muscles,or 20% flexor hallucis longus and M. flexor perforatus of the 30 muscles recognized in this species(Ta- digiti IV (Table 1). A group of fibers about 4.5 bles 1, 5). All of these had a single variant state mm wide passedfrom the distolateral surface of except M. flexor cruris lateralis, which had two. M. flexor hallucis longus to insert on the lateral Three variations were of the mimicking type, and margin of the tendon of the proximal belly of M. 4 were singular. Ten of the 19 specimens(52.6%) MYOLOGICAL VARIATION 501

FIGURE 1. Deep muscleson the lateral surfaceof the knee region in Xiphorhynchus juvigustev,’ showing variation in M. flexorcruris lateralispars accessoria.(A) Typical condition,in which parsaccessoria inserts adjacentto M. gastrocnemiuspars intermedia.(B) Variant in which pars accessoriais separatedfrom pars intermediaby a gap. (C) Variant in which pars accessoriais divided into two separatebellies. See text for discussion.Abbreviations: F, femur;FCRLA, M. flexorcruris lateralis pars accessoria; FCRLP, M. flexorcruris lateralispars pelvica; GI, M. gastrocnemiuspars intermedia; T, tibiotarsus. possessedvariant muscles; four of these variants dividual muscle (G,, = 30.00, P > 0.05, df = were unilateral and seven were bilateral. Nine 29) or individual specimen (G,, = 2.5, P > 0.05, specimenshad only one variant muscle, whereas df = 29). However, six of the 11 variant entries one specimen had two. A total of 1,140 muscles in Table 1 are found in M. flexor cruris lateralis, were examined, of which 11 were damaged. Of suggestingan association with that muscle; the the remaining 1,129, eighteen individual mus- lack of statisticalsignificance may reflectthe small cles had variant conditions; thus 1.6% of the in- overall frequency of variation in the matrix. With dividual muscles varied from their species-typ- regard to sex, five of the 10 specimens that pos- ical states,and excluding the unilateral variants, sesseda variant muscle were male, and five were 1.2% of the muscleswere bilateral variants. The female. latter are the most likely to be misinterpreted, even with bilateral dissection. BLUE JAY There was no statistically significant associa- M. gastrocnemiuspars medialis (Table 2). This tion between the occurrence of variants and in- muscle arises from the head of the tibiotarsus 502 ROBERT J. RAIKOW ET AL.

FIGURE 2. Medial surfaceof thigh in Xighorhynchusflaviguster,showing anomalous extra head arising from the surfaceof M. flexor cruris medialis. See text for discussion.Abbreviations: F, femur; P, pubis; PIFCA, M. pubo-ischio-femoralis pars caudalis; FCRLP. M. flexor cruris lateralis pars pelvica; FCRM, M. flexor cruris medialis; X, anomalous head of PIFCA. and adjacent structures. Raikow (1978) de- complete overlap. Seven specimens (27%) were scribed and illustrated variations in the origin. Type I bilaterally, seven (27%) were Type II bi- In all specimensof the Blue Jay that we examined laterally, and five (19%) were Type III bilaterally. the muscle has a deep head and a superficial There was unilateral variation as well. Five spec- head, both arising from the tibiotarsus, plus a imens (19%) were Type I in one leg and Type II patellar band that arises from the patellar liga- in the other, and two specimens (8%) were Type ment. Differences in the size of the patellar band II in one leg and Type III in the other. Nineteen are characterized by its width relative to the pa- specimens(73%) had the same size patellar band tellar ligament. Variations among Blue Jay spec- in both legs, whereasseven (27%) had a different imens require the definition of different types. In size in each leg. Type I the patellar band was narrower than the This is a mimicking variation. The muscle is patellar ligament, ranging from 0.70 to 0.90 of considered polymorphic because its high fre- its width. In Type II the patellar band was the quency of variability did not permit designation full width of the patellar ligament (1 .OO).In Type of a typical condition. III, a new arrangement, the patellar band ex- Al. jibularis longus(Table 3). This muscle orig- tended its area of origin onto the patella itself, inates on the head of the tibiotarsus, where a its relative width varying from 0.10 to 1.OO, a range of variation was observed. In Type I the MYOLOGICAL VARIATION 503

TABLE 2. Variation in M. gastrocnemiuspars me- TABLE 3. Variation in M. fibularislongus in Cya- dialisin Cyanocittacristata. Type I: Patellarband ranged nocittacristata. Type I: Origin completelyfleshy. Type from 70% to 90% of the width of the patellarligament. II: Origin slightlytendinous. Type III: Origin exten- Type II: Patellarband 100%width of patellarligament. sivelytendinous. Type III: Patellarband extendedonto the patella. Specimennumber Right leg Left leg Specimennumber Right leg Left leg CM 6076 I I CM 6076 I I CM 5338 II III CM 5338 II YPM 7061 III III YPM 7061 III III YPM 12935 III III YPM 12935 III III YPM 8778 II II YPM 8778 I CM 5203 III II CM 5203 II II CM 5123 I I CM 5123 II II KU 038805 II I KU 038805 II III CM 6127 I CM 6127 II II YPM 12936 II II YPM 12936 I I CM 6192 II I CM 6192 II CM 6038 II II CM 6038 II II KU 038806 III III KU 038806 I I CM 6302 I I CM 6302 III III CM 2379 II II CM 2379 II I CM 5971 III II CM 5971 II II CM 2497 I I CM 2497 III III CM 6075 III III CM 6075 I I KU 045195 II III KU 045195 III III CM 858 II II CM 858 II II CM 5146 CM 5146 II I KU 45196 I I KU 45196 III II YPM 8779 I I YPM 8779 II II CM 6210 II II CM 6210 I I KU 067423 I I KU 067423 I I YPM 10226 II II YPM 10226 II

Type I has a single extra head (Fig. 3A), Type II origin is completely fleshy, Type II has a slightly has two extra heads (Fig. 3B), Type III has scat- tendinous origin, and Type III has an extensive tered fascicles,and Type IV has no distinct extra tendinous origin. Nine specimens (35%) were parts (Fig. 3C). Type I bilaterally, seven (27%) were Type II bi- Six specimens (24%) were Type I bilaterally, laterally, and four (15%) were Type III bilater- one (4%) was Type II bilaterally, three (12%) ally. In addition, two specimens(8%) were Type were Type III bilaterally, and two (8%) were Type I in one leg and Type II in the other, and four IV bilaterally. Unilateral variation also occurred. specimens (15%) were Type II in one leg and Three specimens(12%) were Types I and III, five Type III in the other. (20%) were Types I and II, three (12%) were These are mimicking variations. Because no Types III and IV, and two (8%) were Types I and variant occurredin a majority of specimens,there IV. One leg of specimen YPM 10226 was dam- is no typical state as defined above. Instead, the aged and could not be recorded, the undamaged condition is best described as polymorphic. leg was Type I. A4. extensordigitorum longus(Fig. 3; Table 4). Because no typical condition could be deter- This is the deepestmuscle on the cranial surface mined in C. cristata, this is another caseof poly- of the tibiotarsus. It arises fleshy from the cne- morphism. This condition in Cardinalis cardi- mial and patellar crests and the cranial face of nalis (Raikow et al. 1990) was classified as a the tibia1 shaft. In C. cristata variations were singular variation, but now that the variant is found in the presence, position, and number of known from more than one speciesit is better extra heads arising on the tibiotarsus. Variations considered as mimicking. ranged from two heads, one head, and scattered Analysis of patterns of variation. The pattern fascicles,to the absenceof any heads or fascicles. in C. cristata was distinctly different from that 504 ROBERT J. RAIKOW ET AL.

FIGURE 3. Variation in the originof M. extensordigitorum longus from the tibiotarsusin Cyunocittacristata. (A) Type I, with a singleextra head.(B) Type II, with two extraheads. (C) Type IV, with no extraparts.

in X jlavigaster (Table 5). Major variations occurred in three of 31 muscles (9.7%). For each DISCUSSION of the three variable muscles, no typical condi- The levels of variation that we observed in X tion could be designated because variation was flavigaster and C. cristata are generally similar extensive and polymorphic. This indicates an as- to those reported for other speciesof passerine sociation between individual muscle and the oc- birds (Table 5). The levels are similar in the over- currence of variation, although it precludes use all frequency of atypical muscles among each se- of our statistical tests;in any event, the tests are ries, expressedas the percentageof atypical mus- unnecessaryto document such an association in cles among all individual muscles, in the this clear-cut instance. There is no evidence of percentageof different muscles that have at least association between individual specimen or sex one atypical example, and in the percentage of and the occurrence of variation, because each specimens having at least one atypical muscle. individual is equivalent in possessingone of the For the latter measure, the value of 0% for C. several possible statesfor each of the three vari- cristata does not reflect the absenceof variation. able muscles. Rather, each of the three variable muscles had MYOLOGICAL VARIATION 505 several conditions, none being present in over TABLE 4. Variation in M. extensordigitorum longus 50% of the sample. This situation is similar to in Cyanocittacristata. Type I: One anomaloushead in origin. Type II: Two anomalous heads in origin. Type the explosive variation reported by McKitrick III: Anomalous fasciclesin origin. Type IV: No anom- (1986) for M. flexor cruris lateralis in tyrannid alous heads or fasciclesin origin. flycatchers. Comparative studies of myology often rely Specimen number Right leg Left leg upon dissection of one specimen per species,un- CM 6076 I der the assumption that the musculature of the CM 5338 I IV dissectedspecimen is typical for the speciesas a YPM 706 1 I I whole. This procedure usually cannot be ex- YPM 12935 I panded to include extensive series because dis- YPM 8778 IV III CM 5203 IV I section is so labor-intensive and the number of CM 5123 II I available specimens is often limited. The use of KU 038805 I III single specimensnonetheless introduces the po- CM 6127 II II tential for error in assessmentsof species-typical YPM 12936 III III IV IV conditions. CM 6192 CM 6038 I II The frequencies of errors in assessingspecies- KU 038806 I III typical conditions are estimated in the last col- CM 6302 I II umn of Table 5 (see Materials and Methods for CM 2379 III IV explanation) and range from 0% to 6.3%, with CM 5971 III III CM 2497 I an average of 1.9%. These values are low, sug- CM 6075 gestingthat the bilateral dissectionof single spec- KU 045195 imens provides accurate myological informa- CM 858 IV III tion. An estimate of the improvement in accuracy CM 5146 I I III III provided by bilateral (as opposed to unilateral) KU 45196 YPM 8779 II I dissection can be obtained from Table 5. The CM 6210 IV IV percentage of atypical muscles among all indi- KU 067423 III vidual muscles is equivalent to the frequency of YPM 10226 - error in assessingspecies-typical conditions under unilateral dissection of a single specimen per species.This averages2.6%, an increase in error for use in developing reference descriptions. Ad- of 0.7% over bilateral dissection. ditional dissectionsmay be performed when uni- Although intraspecific variation in passerine lateral variation is detected, or when between- hindlimb musclesis comparatively rare when the speciesvariation is observed in muscles known full set of muscles is considered, a fairly sub- to be highly variable in other species. Our un- stantial proportion of the individuals of most of derstanding of the frequency of error in assessing the speciesin Table 5 possessesat least one atyp- species-typical conditions is, nonetheless, based ical muscle. The percentage of individual spec- upon comparatively few species.In addition, al- imens having atypical muscles (Table 5) is as though the levels of error in Table 5 are low, they high as 65% for large series (n > 10) and 80% may have the potential to induce errors in phy- for certain small series (n = 5) averaging 47.8% logenetic reconstructions. Hence, dissections of for the 14 speciesas a whole. The overall level additional speciesand simulation studies of the of variation is actually somewhat higher, because effectsthat the levels of error in Table 5 might explosive variation is excluded from the figures have on phylogenetic reconstructionsare needed. in Table 5. The number of different musclesthat include atypical examples is smaller, however, ACKNOWLEDGMENTS and the number of atypical individual muscles For the loan of specimens we are grateful to K. C. among the total dissectedis smaller still. Parkesand J. Loughlin, The CarnegieMuseum of Nat- These considerations support the contention ural History; R. F. Johnstonand T. J. Davis, The Uni- versity of Kansas Museum of Natural History; and F. of Raikow et al. (1990) that the best approach is C. Sibley, Peabody Museum of Natural History, Yale bilateral dissection of single specimens, aug- University. We also thank S. L. Berman, M. C. mented by a series of specimens of one species McKitrick, and D. W. Rudge for providing informa- 506 ROBERT J. RAIKOW ET AL.

TABLE 5. Intraspecific variation in the hind limb musclesof passerinebirds. ’

%of different muscles No. of % of having % of all muscles specimens atypical individual No. of typi- having examples ;nh&; FFfq;tz;;y speci- tally or pnly- SpXkS mens Reference present %zz mbrphism atypical ’ (%) Xiphorhynchusflavigaster 19 This study 30 52.6 20.0 1.6 1.2 Cyanocitta cristata 26 This study 02 9.7 9.1 6.3 Dendrocolaptes certhia 8 Raikow, in press :: 62.5 10.0 2.3 2.1 Margarornis squamiger 5 Rudge and 31 80.0 12.9 1.9 0.6 Raikow 1992 Premnoplex brunnescens I Rudge and 31 57.12 9.7 1.8 1.4 Raikow 1992 Premnornis guttiligera 5 Rudge and 31 0 0 0 0 Raikow 1992 Cardinalis cardinalis 23 Raikow et al. 1990 32 61.0 9.8 2.0 1.5 Hylocichla mustelina 25 Raikow et al. 1990 31 36.0 12.9 1.4 1.0 Passer domesticus 40 Berman et al. 1990 32 65.0 12.5 2.4 2.4 Myiozetetes similis 12 McKitrick 1985, 1986 32 O2 3.1 3.6 2.3 Tyrannus verticalis 14 McKitrick 1985, 1986 32 12.52 6.2 4.0 2.2 Tyrannus melancholicus 7 McKitrick 1985, 1986 32 42.9 6.2 1.3 1.3 Pitangus sulphuratus 5 McKitrick 1985, 1986 32 40.0 12.5 2.2 1.9 Myiodynastes maculatus 5 McKitrick 1985, 1986 32 80.0 9.4 2.8 1.9 ’ ’ Excludingminor variations.See text for discussion. i Excludingcases in which a muscleis so variablethat no typical conditioncan be defined(see p. 504). Theseinclude the following:Cyanocitta crisfata:M. gastrocnemiuspars medialis.M. tibularis longus,and M. extensordigltorum longus,;Premnoplex brunnescens: M. gastrocnemluspan medialis;Myiozetes similis: M. flexorcruris lateralis; Tyrannusverticali& M. flexorcruris laterals 1Including both unilateraland bilateral variants.

tion about their published studies. The illustrations (Aves: Passeriformes)and its possiblesignificance. were drawn by W. R. Filer. This material is basedupon J. Zool. Lond. 209125l-270. work supported by the National Science Foundation RAIKOW, R. J. 1978. Appendicular myology and re- under Grant No. BSR-8617896 to R. J. Raikow. lationships of the New World nine-primaried os- tines (Aves: Passeriformes).Bull. Carnegie Mus. 7: l-43. LITERATURE CITED RAIKOW, R. J. In press. Structure and variation in the hindlimb musculature of the woodcreepers BERMAN, S. L., M. CIBISCHINO, P. DELLARIPA, AND L. (Aves: Passeriformes:Dendrocolaptinae). Zool. J. MONTREN. 1990. Intraspecific variation in the Linn. Sot., Lond. hindlimb musculatureofthe House Sparrow.Con- RAIKOW, R. J., A. H. BLEDSOE, B. A. MYERS, AND C. dor 92: 199-204. J. WELSH. 1990. Individual variation in avian BLEDSOE, A. H., R. J. RAIKOW, AND A. G. GLASGOW. musclesand its significancefor the reconstruction 1993. The evolution and functional significance of phylogeny. Syst. Zool. 39:362-370. of tendon ossificationin woodcreepers(Aves: Pas- ROWE, T. 1986. Homology and evolution of the deep seriformes: Dendrocolaptinae). J. Morphol. 2 15: dorsal thigh musculaturein birds and other Rep- 289-300. tilia. J. Morphol. 189:327-346. BOCK, W. J., AND C. R. SHEAR. 1972. A staining RUDGE, D. W., AND R. J. RAIKOW. 1992. Structure, method for grossdissection of vertebrate muscles. function, and variation in the hindlimb muscles Anat. Anz. 130:222-227. of the Margarornis assemblage (Aves: Passeri- MCKITRICK, M. C. 1985. Pelvic myology of the king- formes: Furnariidae). Ann. CarnegieMus. 6 1:207- birds and their allies (Aves: Tyrannidae). Ann. 237. Carnegie Mus. 54:275-3 17. WILLIAMS, D. A. 1976. Improved likelihood ratio MCKITRICK, M. C. 1986. Individual variation in the tests for complete contingencytables. Biometrika flexor cruris lateralis muscle of. the Tyrannidae 63~33-37.