The Adaptive Significance of Local Variations in the Bill and Jaw Anatomy of North European Red-Necked Grebes Podiceps Grisegena

Home , Great crested grebe, Podilymbus, Rollandia (bird)

The adaptive significance of local variations in Gr the bill and jaw anatomy of North European Red-necked Grebes Podiceps grisegena be

A to Tb JON FJELDSÅ Re ad FJELDSÅ, J. 1982 : The adaptive significance of local variations in the bill and jaw anatomy of North European Red-necked Grebes Podiceps grise- gr~ gena. - Ornis Fennica 59:84-98. gel Red-necked Grebes from the northernmost breeding populations in Fin- tio land and Russia develop a long and slender bill. This trend, and also SM the large size of East Siberian and Nearctic Red-necked Grebes, might thz be associated with a change in the diet from invertebrates to fish in Fu areas where the piscivorous Great Crested Grebe is missing. A sample of 87 Red-necked Grebes from a Danish wintering area comprised both inl South Scandinavian and slender-billed northern types. These were studi- tu( ed in order to discover to what extent the morphological and anatomical dif variation affects the diet. tio The slender-billed Red-necked Grebes converge with Great Crested Grebes in the ability to open the bill rapidly and keep a strong grip on Fo the prey without sliding of the mandible . Both adaptations may be valu- ha able for catching fish. Study of the stomach contents showed that the ho, Red-necked Grebes of South Scandinavian type fed mainly on annelid, Sul gleaned, from plants and picked up from the bottom, while individuals of the northern type specialized on squids and, fish. th( Although this does not prove anything, it provides support for the Ca assumption that the morphological variation was caused by selection due Tl to the presence or absence of Great Crested Grebes . wl Jon Fjeldsd, Zoological Museum, University of Copenhagen, Universi- no tetsparken 15, DK-2100 Copenhagen 0, Denmark . zol

The Red-necked Grebe Podiceps gri- Finnish birds, was later synonymized. th( segena has two distinct subspecies However, although there is consider- Ar (Storer 1979), the West Palaearctic able overlap with .the other European Cr grisegena and the Nearctic and East birds, my data (Fjeldså 1982) con- cri Siberian holboellii, the latter being firm that there is a clinical change in 19 rather pale and similar in size to the this part of the range . As we move et Great Crested Grebe Podiceps crista- from southwestern Scandinavia to Sio tus. Birds from Balchash ("balcha- near the northern distribution limit of N( schensis") are described as rather the species (inland areas from Kuopio th( long-winged. All the specimens I have to Finnish Lapland, Kola, Archan- ta( seen from this area are, however, gelsk, Dvina) the mean wing length mi within the grisegena range, and hardly increases by 1 .5 0/o and the tarsus by va qualify for taxonomic recognition. 2 .5 0/o, while the bill increases in Gr Similarly, the name schioleri, proposed length by as much as 11 0/o (Fig. 1) ve by Hortling (1929) for the rather big and becomes more slender, its shape Gr J. Fjeldsd : Functional anatomy of North European Red-necked Grebes 85 approaching that of the Great Crested Grebe. The interpretation of this trend will be the objective of this study.

A possible interpretation and how to test it The morphological change in northern Red-necked Grebes may reflect adaptations to climate or to habitat gradients. However, according to general ecogeographical rule, adaptations to a colder climate should give smaller relative change in bill length than in wing length as we move north. Further, the summer temperatures of inland areas change little with lati- tude, and the wintering areas of the different Red-necked Grebe populations (Danish waters and according to Folkestad 1978 northwest Norway) FiG. 1. Bill lengths of Red-necked (dots) have similar temperatures . Finally, and Great Crested Grebes (open circles) from different populations, measured on specimens holboellii shows no variation from the in breeding plumage in the Zoological Mu- subarctic zone, with its bleak lakes, to seums of the Universities of Copenhagen, Oslo the warm climate and rich lakes of the and Helsinki ; the Swedish Museum of Natural Canadian forest/prairie ecotones. History, Stockholm ; the American Museum of Natural History, New York ; the Australian Thus, considering the species as a Museum, Sidney, and the museum of CSIRO's whole, the geographical variation does Wildlife Research, Canberra. Bottom : bill not follow the climate or lake type lengths of the grebes collected in Danish zones. winter quarters and used in this study. It might be that the large size of the Red-necked Grebes in North America is due to the lack of Great ern coastal zone. Population fluctua- Crested Grebes (compare the diet of tions (Thomasson 1953), gene flow cristatus, e.g. in Cramp & Simmons contact through the European popula- 1977, and holboellii, e.g. in Palmer tions and some interspecific contact in et al. 1962). The smaller bill dimen- all European winter quarters might sions in the East Siberian than in the explain the moderate amplitude of the Nearctic holboellii (Fig. 1) agrees with geographical variation of the Europe- the fact that the former has some con- an Red-necked Grebes. Altogether, the tact with Great Crested Grebes on variation in the species coincides well migration and in winter. The size with what we could expect, if the variation of Finnish Red-necked variation was due to interspecific Grebes could also be explained by the competition. More distribution and very sparse breeding of Great Crested biometric data from North Russia Grebes outside the southern and west- would be desirable to reinforce the 86 ORNIs FENNICA Vol. 59, 1982 documentation. Nevertheless, it ap- study of the birds from these important pears to be a reasonable working winter quarters for Red-necked Grebes hypothesis that Red-necked Grebes particularly interesting is the fact that evolve morphological adaptations they have very variable bills (Fig. 1), which permit them to live like Great which suggests that birds of both Crested Grebes in areas where this North Finnish/North Russian and species is missing. Alternatively, we South Scandinavian provenience are may suppose that the two were origin- represented. Altogether 87 Red-neck- ally allopatric ecological counterparts, ed and six Great Crested Grebes were and that the Red-necked Grebe collected after the oil-spill by the managed to invade the range of the Danish Game Research Station, Kale, Great Crested Grebe by means of and sent to the Zoological Museum, "character displacement", evolving a University of Copenhagen. smaller size and smaller bill suited for taking invertebrate prey (for diet in the breeding areas, see Madsen Anatomical studies 1957, Onno 1958, Markuze 1965) . Methods. In methods and terminology, the It is pertinent to test whether the anatomical study follows that of Fjeldså (1981) on Peruvian Grebes . morphological changes affect the birds' The jaws and jaw muscles of one Great ecology. This could be established by Crested and six Red-necked Grebes (including studying whether Red'-necked Grebes two of the most slender-billed) from Hatter change their ecology as we pass north Rev were dissected under a 10 .5-67.5 X zoom Kyowa dissecting microscope . In addi- through Finland. Unfortunately, tion an examination was made on the sculls though, a documented ecotope change of 26 Great Crested and, 19 Red-necked would be insufficient as evidence, as Grebes (including holboellii) . Distances from it could have three or more causes : articulations to insertion points of muscles were measured with caliper rulers, and the It could be due to the change in "power arms" calculated by the appropriate morphology, but it could as well re- trigonometric functions. The potential muscle flect changes in the habitat and food force is determined- from the number of fibres resources, or changes in resource ac- rather than the muscle mass (Bock 1974) . To obtain an index of fibre number, I calculated cessibility due to the lack, in the north, the physiological cross section of the muscles of interference competition from Great as insertion area times the sine of the insertion Crested Grebes (cf. Berndt & Drenck- angle. Since various parts of the muscle may hahn 1974). The data might con- have different insertion angles, the total cross section was calculated, section by section. This sequently be hard to interpret. was done at either end of each muscle, and It may be safer to make the study the mean of the two vallues was usedl. For in a common wintering area, where pennate fibre arrangements, the cross section the interference competition is slight, value. The potential speed, of contraction pinnation angle to give a functionally relevant and where all the birds use the same value. The potential speed of contraction food supply. is judged from the fibre lengths. Some card,- The opportunity to make such a board models, in 4 X scale, were used to determine the effects of coupling ligaments study arose in January 1979, when and changes in action angles of muscles as the 400-500 Red-necked Grebes were bones change position. Lever calculation was killed by an oil-spill on Hatter Rev used to determine muscle torques, while the (55°55'N, 08°50'E) in Danish coastal interaction of muscles was studied by con- structing freebody diagrams (Bock 1974), waters (Larsen 1979) . These birds can elaborated for non-gravity and, static situations, all be assumed to have fed in the and ~disregardling the load against which the same macrohabitat . What makes the muscles act. J. Fjeldsd : Functional anatomy of North European Red-necked Grebes 87

Due to the complexity of the superficial TABLE 1 . Some important distances (mm) and deep muscles and space considerations, and physiological cross section values (0, few aspects can be properly depicted . Fig. 2a mm2) of jaw muscles of one Great Crested shows superficial aspects, Fig . 26 areas of and six Red'-necked Grebes. For the latter fleshy attachments of muscles, Fig . 2c the species, Spearman's rank correlation coefficients narrow (aponeurotic) attachments. Other (rs) are given in respect of the length, large- aspects are shown in numerous illustrations in ness (LBI) and slenderness (SBI) of the in- Bams (1956), Zusi & Storer (1969) and dividual bills. For explanation of the bill Fjeldså (1981) for Podilymbus podiceps, Rol- indices, see the text p . 87 . landia rolland and microptera, Podiceps occi- pitalis, taczanowskii and cristatus . of functionally most P. grisogena P. Table 1 shows some the (6) cris- important parameters and their correlation Range r r r tatus with bill size parameters . The largeness of the s s s (1) bill is expressed by the index LBI : bill length with with with X n X bill depth X bill width. The best culmen LBI SBI way to tell northern Red-necked Grebes in the Hatter Rev data apart from Culmen 35.5- 50.5 southern is to record how long and relatively 45.5 - 0.74 0 .56 thin the bill is . This is expressed as a "slender bill index" (SBI), viz. bill length times the Distal symphysis of 7 .8- ratio of bill length to the sum of bill depth 0.72 0.74 0 .5-2 10 .5 and bill width. mandibular 11 .3 rami of the . the Opening bill Opening of M. depressor 20.9- 0.86 0 .86 0.89 32 bill is due to depression of the lower mand. 0 30.8 mandible, which may be speeded up power arm 5-6 .3 Q:86 0.83 1 .00 7.1 by an increase in the fibre lengths M. add. 84- 0.89 0.81 0.73 120 and power arm of the involved muscle mand. ext. 109 III 0 6.3- 0.83 0.83 0.73 7.6 and by supplementary lifting of the power arm 7 .2 upper mandible . M. pseudo- 42-50 Musculus depressor mandibulae runs temp . 0.72 0.86 0.70 60 from the deltoid fossa of the skull superfic . 0 6-7 0 .77 0.52 0.57 7 .6 with the surrounding crests (with apo . power arm I), its long (rapid) fibres narrowly in- M. pseudo- 8.5-10 0.73 0.76 0.52 11 serted on the posterior fossa of the temp . prof . 0 lower mandible via apo. II and one power arm 7.4- 0:81 8.2 0.70 0.43 9.8 small unnamed aponeurosis . Spanning two joints, it has a complicated effect . M. protract . 17-28 0.70 0.74 0.94 25 quadr. 0 5-.6- If the action line is steeper than the power arm 5.9 0.43 0.46 0.52 6.1 quadrate axis, it swings the quadrate forward, which causes a coupling between mandibular depression and protrusion of the palate bones and upper high SBI (Table 1) . These birds here- mandible . This effect is reinforced by by approach cristatus. The change in the coupling action of the postorbital action angle gives a basis for the pro- ligament . trusion and lifting of the upper There is a slight general increase mandible mentioned above, and ac- with increasing bill size (Table 1). celeration of the gaping . More important, though, is the relative increase in power arm due to a longer, Adduction of the bill. Adduction is more backwards directed, retro- due to several muscles. Some are articular process in specimens with "two-joint" adductors, spanning both 88 ORNis FENNICA VOI . 59, 1982 the cranio-quadratic and quadrato- though with backward sliding of the mandibular joints, which complicates quadrate. The long fibres give high the calculation of their effects. Others speed in the final phase of adduction, are attached directly to the quadrate, when the action angle is optimal. and are "one-joint" muscles, which Some increase, particularly of the may control the movements of this part attached in the temporal fossa, bone. Sliding of the quadrate will be was recorded in long-billed individ- transferred through the palate bones uals. to the upper mandible. Part 3 is the main adductor, a big, Musculus adductor mandibulae ex- pennate muscle whose fleshy origin ternus forms a powerful, multipennate occupies the whole temporal' fossa of complex, which may be divided into the skull with the surrounding crests. three main parts. Many fibres originate from the out- Part 1 is a superficial fan of long side of the complexes muscle, via apo. (rapid) fibres arising from the mem- X from the rostral demarcation of the brane which passes in front of the ear, temporal fossa and from the posterior with some fibres also arising from part of the orbit, above the fleshy ape . III and the top of apo. VIII and origin of M. pseudotemp. superfic. A XII. It is inserted narrowly via the few fibres also arise from apo. VIII big apo . IX on the surangular/dentary and XII and are inserted on apo. junction of the mandible edge. The VIII. The central apo. VII is inserted mean vector at the insertion point as a strong tendon on the coroid pro- parallels the jaw when the bill is cess of the surangular edge. slightly open. Thus, its main effect The muscle gives more efficient ad- must be pulling the mandible back duction than part 2, although with and rotating the quadrate backwards, backward sliding of the quadrate. The which indirectly means that the upper high number of short fibres suggests mandible is depressed. This suggests that force production rather than that the muscle is most efficient in angular acceleration is the main moving big prey items back into the function. throat by quick backward jerks. The variation in size is modest The muscle was strongest in the (Table 1). largest-billed birds, but this involves Musculus pseudotem¢oralis super- no drastic functional variation. ficialis, part l, has a fleshy origin on Part 2 originates via apo,. VIII and the posterior wall of the orbit and is XII on the zygomatic process, with narrowly inserted via the big, super- some fleshy attachments from the ficial apo. XI and a stout tendon in- rostral demarcation of the temporal side the previous muscle on the medial fossa. It is attached with a wide, face of the mandible. The smaller fleshy attachment on the outside of part 2 lies still deeper and is inserted the surangular bone and via apo. VII anterior to part 1 . The high number along its upper edge. It has a slightly of short fibres suggests maintenance more adductive effect than part 1, al- of isotonic grip as the main function,

FIG. 2. Jaw muscles of grebes. - A.. Red-necked Grebe, superficial jaw muscles; M. add. mand. ext. 1, which runs from the dashed line to the aponeurosis IX, is removed. - B. Red-necked Grebe, fleshy attachment areas of the jaw muscles on the skull (B 1), inside of quadrate (B 2), and outside (B 3), upper edge (B 4) and inside (B 5) of lower mandible. - C. Great Crested' Grebe, narrow (aponeurotic) muscle attachments.

J. Fjeldså : Functional anatomy of North European Red-necked Grebes 89 90 ORNIs FENNICA VOI. 59, 198'2

than the quadrate axis (Fig. 3), which means that this two-joint muscle rotates the quadrate forwards, giving slight protrusion of the upper mandible in the early adduction, or counteracting some backwards sliding due to the external adductors. The Great Crested Grebe shows elongation of the skull due to an increase of the area serving as attachment for the pseudotemporalis muscles. The central aponeurosis of the muscle, from a ridge on the skull, also increases, and the power arm as well. This means considerable acceleration of the initial phase of adduction and stabilizing of the final phase, since the effect on the quadrate, according to a free-body diagram, balances that of M. add. mand. ext. 3. The variation in this muscle in the Red-necked Grebes was very slight. Holboellii also showed only slight convergence towards the Great Crested Grebe state.

Muscles stabilizing the quadrate. Musculus adductor mandibulae post- FIG. 3. The combined effect of two muscles erior has a fleshy attachment on the which may counteract backward sliding of the thirds of the orbital process lower mandible caused by the main adductor, inner two Musculus adductor mandibulae externus part of the quadrate, and its short fibres 3, in three dissected grebes. In the free-body are inserted on the dorsomedial sur- diagrams of the rear end of the lower man- face of the lower mandible, right dible are shown the forces exerted by Muscu- lus pseudotemporalis superficialis parts 1 and ahead of the quadrato-mandibular 2 (M. pst. sp . 1, 2) and their combined joint. A one-joint muscle, it may vector at the quadrato-mandibular joint. This oppose backward sliding of the qua- vector may be separated into a component drate due to the action of the external which follows the quadrate axis (Q axis) and, perpendicular to this, a rotating component adductors, thus both stabilizing and (rotat c.) . In the vector diagrams of the qua- reinforcing an isotonic grip . There drate bone is shown the force exerted by was no obvious variation. Musculus protractor quadrati (m. protr. q.) Musculus pseudotemporalis profundus and the calculated total forward-rotation force XVIII due to this muscle and M. pst. sp. 1, 2. runs with a central apo. from the distal part of ,the orbital process of the quadrate to a wide fleshy although the muscle may, with its attachment on the medial fossa of the short power arm, also give high mandible. Being a long-fibred one- angular acceleration. The vector at the joint muscle, it secures stable and quadrato-mandibular joint is steeper quick adduction about the quadrato-

J. Fjeldsd : Functional anatomy of North European Red-necked Grebes 91

mandibular joint, even in the early pterygoideus pars dorsalis, lateral phase of the adduction, and counter- branch, has a fleshy origin on the acts, die to its action angle, some pro- lateral surface and adjacent ventral trusion of the quadrate caused by M. edge of the pterygoid, and via apo . pseudotemp . superfic. The bulk and XXIX is inserted on the crest just power arm (length of orbital process) ahead of the articulation of the man- did not vary clearly with the bill dible. These two muscles are the most parameters in European Red-necked efficient in retracting the palate, Grebes, but holboellii and above all thereby rotating the upper mandible Great Crested Grebes showed an in- about the frontonasal hinge, so that creased power arm . it is lowered. This supplements the big Musculus protractor quadrati & pte- M. add. mand. ext. 3 in exerting a rygoidei has a wide, fleshy attach- strong grip, and may give a backward ment to the skull and fleshy and apo- push to a big prey item during swal- neurotic attachments (apo.'s XIII, XV, lowing. XVII) to the medial surface of the The power arm (height of upper quadrate and posterior end of the mandible at frontonasal hinge) is long pterygoid. The main function may be in relation to the bill length in most to counteract sliding of the quadrate Red-necked Grebes, but is reduced in due to M. add. mand. ext. 3. In addi- slender-billed birds (including Great tion it may protract the quadrate, Crested Grebes) . This variation ap- which has a kinetic effect, and by an pears to be compensated for, in inward pull on the quadrate it may slender-billed grisegena, by variation inhibit the quadrato-mandibular artic- in muscle size. ulation and permit independent movements of the mandibles. Inframandibular bending. Musculus In Red-necked Grebes, the attach- pterygoideus, pars lateralis, ventro- ment area expands upwards and for- medial branch, originates in the vent- wards as SBI increases (Table 1). In ral cavity of the palatine and via the holboellii it reaches an extreme devel- big apo. XX and XXII. Musculus opment along the entire anterior de- pterygoideus, pars dorsalis, median marcation of M. pseudotemp . superfic. branch, originates from the ventral This is not seen in any Great Crested and ventromedian surface of the pte- Grebes. rygoid. Both are inserted via the complex apo. XXV on the medial crest Kinesis. Kinesis involves the move- of the articular. The muscles supplement of the upper mandible about the ment those mentioned above in de- frontonasal hinge, due both to some pressing the upper mandible. In addi- of the muscles mentioned above and tion they pull the medial crest of the to the palate muscles. articular forwards, thereby rotating Musculus pterygoideus, pars latera- the bone about the quadrato-man- lis, dorsolateral branch, has a wide, dibular articulation, so that the man- fleshy attachment to the dorsolateral dibles bulge in the middle, at the pli- surface of the palatines and further able inframandibular "hinge" . This uses the big apo. XXII along the allows a big prey item to pass between posterolateral edge of the bone. It is the mandibular rami, and pushes down inserted via the medial and postero- its mass by pressure from the upper medial faces of the articular. Musculus mandible. 92 ORNIs FENNICA VOI . 59, 1982 J. F

There appears to be some increase respects resemble the unspeciahzed tior in fibre numbers in the muscles in and probably primitive states shown mu long-billed Red-necked Grebes, al- by Golden Grebes Rollandia spp. can though the estimate of ,the physiologic- (Fjeldså 1981) . However, they app- wai al cross section was too inaccurate to roach their congeners Podiceps occipi- of determine whether this is more closely talis and taczanowskii (also described associated with SBI than with the in Fjeldså 1981) by showing some adz general bill size. The power arm (in- adaptations for foliage-gleaning, viz. gri] ward bend of medial articular crest) moderate divergence of the insertion cor. showed no correlation with bill length angles of the adductors and a relative- sup or SBI, while a change in the shrape ly small pseudotemporal muscle with Th of the bill tip and a relative shorten- a short power arm. This means that Re, ing of the symphysis in slender-billed the bird cannot fully contract the shc birds may imply a reduced resistance mandibular adductors without back- cre to bending (Table 1). In Great Crested ward sliding of the mandible. Al- 3 Grebes, on the other hand, the elonga- though Red-necked Grebes are certain- tor tion of the skull involves a relative ly not so specialized as the two above- ten increase -of the part of the jaw which mentioned congeners, the conditional tra lies behind the hinge. This increases phrases are certainly narrower than eqi the potential maximum spread of the in Rollandia. sle mandibular rami. Expansion of the Most of the individual variation in W: foramina on the inframandibular Jaws and muscles may represent direct cot hinge and a shortening of the sym- adjustments of the potential force pro- eqi physis at the bill tip reduce the resist- duction to variation in general size. ma ance to bending. On the other hand, This does not involve true functional dr, the long "arm" from the mandible changes. the joint to the tip of the prearticular Two ;trends were associated with im wing (the hinge) in Great Crested SBI rather than the overall size, and cai Grebes means that torques given by may involve a change in mechanical slil the pterygoid muscles at the medial properties . Slender-billed birds con- Cr articular crest result in reduced torques verge with Great Crested Grebes in Gr at the hinge. Further, the short power showing an increase in the mandibular me arm for kinesis (low height of the depressor and its power arm, which ve: upper mandible at the base) in rela- may speed up the depression of the tion to bill length means that the bill lower mandible and couple this move- in¬ tip cannot exert strong pressure on ment with a lifting of the upper man- tio the prey. Since reliable measurements dible. This specialization in relation fu: of the resistance to bending of the to what we see in most grebes may be Sir bones can be obtained only on fresh important during rapid pursuit of fish, pr birds, there are too many points of as the bird can maintain its stream- gr uncertainty for calculating whether lining, keeping the bill closed, till la) species or populations differ regarding immediately before it can grasp its un the ability to swallow large-sized prey. Secondly, the quadrate pro- m; animals. tractor expands, which means increas- ed ability to control the position of ad Possible functional significance of the the quadrate. In normal' (stout-billed') Re variation. Red-necked Grebes have Red-necked Grebes, the forward rota- Gi strong jaw muscles, which in many tion of the quadrate by the joint ac- an

J. Fjeldsd : Functional anatomy of North European Red-necked Grebes 93

tion of the superficial pseudotemporalis long bill gives a potential increase in muscle and -the quadrate protractor the spreading of the mandibular rami. can do little to counteract the backward rotation due to full contraction of the powerful M. add. mand. ext. 3 . Great Crested Grebes show an Food selection adaptation for maintaining a stable grip even when this muscle is fully Method . The stomach contents of the 87 Red- necked and the six Great Crested Grebes were contracted, since they have a greater examined under the microscope to identify the superficial pseudotemporal muscle. remains of the prey. The number of prey This is not the case in slender-billed items was counted from resistant and char- Red-necked Grebes, but these instead acteristic parts, such as cleithra and otoliths of fish, eye-lenses of squids, and the man- show apparent paradaptatio,n, an in- dibles of most annelids and arthropods . The crease in the quadrate protractor . Fig. state of decomposition of most of the prey 3 shows that the combined forward items was unfortunately too far advanced to torque of the superficial pseudo- allow safe determination of their size, or the mass composition of the diet. temporal muscle and quadrate pro- The numbers of prey items in the stomachs tractor on the quadrate bone is about showed hyperboloid variation, because most equally strong in Great Crested and of the stomachs of the oil-killed birds were slender-billed Red-necked Grebes . strongly decomposed . For this reason, log- arithmic values (log. (N+1)) should be used What is really important is that this in the analysis . combined torque vector approximately equals the component of M. add. mand. ext. 3 which rotates the qua- Stomach contents of Red-necked drate backwards. The net effects is, Grebes. The prey comprised: then, that the short-fibred adductors, - Hydmozoa: 1 Dynamena pumila colony. - Annelida : 381 Nereis spp ., mainly small involved in exerting an isometric grip, N. diversicolor and pelagica, but also 16, big can exert maximum torques without N. virens with estimated total lengths of 25- sliding of the quadrate. Thus Great 40 cm, masses 2560 g; 685 Nephthys sp ., Crested and slender-billed Red-necked 10 Glycera sp . ; 2510 polynoids~, mostly Lepido- notus squamatus ; 139 Pectinaria koreni ; 2 Grebes both manage, by different Hyaloecia sp., 3 unidentified . mechanisms, to maintain a stable and - Arthropoda : 139 Idothea sp. ; 11 Gam- very strong grip. marus sp . ; 2 Mysidacea (undet .) ; 1 Crangon This may be important sp. ; 1 Mallophaga (un&et.) ; 4 Coleoptera when grasp- (Scarabaeidae, Hydrophiliidae and Curculio- ing a slippery fish. That this adapta- nidae undet.) . tion is associated with fish-eating is - Mollusca : 4 Buccinum undatum ; 7 Lit- further indicated by the fact that a torina littorea ; 1 Gibbula sp. ; 1'05 of the similar rapid squid Aloteuthis subulata (mass about functional state has evolved, 5 g) . probably independently, in other - Echinodermata : 1 Holothuria apoda grebes which take fairly big fish (Rol- (Leptosynapta inhaerens?) . landia, Aechmophorus occidentalis; - Pisces : 2 Rutilus rutilus ; 39- Gasterosteus unfortunately I lack aculeatus; 2 Spinachia spinachia ; 1 Merlucius anatomical merlucius ; 3 Pholis gunellus . material on Podiceibs major). Further, all the stomachs had a variable There are no obvious convergent amount of complete and more or less dis- adaptations between slender-billed integrated feathers and sand (probably ingest- Red-necked Grebes and Great ed with annelids), and many stomachs also Crested had some fragments of fresh or dead plants, Grebes for swallowing big prey pieces of plastic and' other indigestible animals, apart from the fact that a roughage . 94 ORNIS FENNICA VOI. 59, 198'2

As stated above, no attempt was ment by bill length gave rs 0.03 ; by made to calculate the mass composi- bill depth r,. 0 .20; by LBI r, 0.18 ; by tion of the diet, but squids probably SBI rs 0 .28 ; all P>0.10). This does predominated, followed by Lepidono- tus and big Nereis virens. not indicate that morphologically This diet differs greatly from the different birds select different parts winter diets described by Madsen of the winter quarters. (1957) . This is possibly because he examined birds accidentally collected Variation with grebe morphology. near coasts, and not birds from the Rank correlation tests on the average typical Danish wintering habitats, log (N+1) values for each six individ- which may be on distant offshore uals in a series were done for all the shallows, generally habitats with regularly occurring prey species. Some Scoters Melanitta sp. and Eider Ducks grebes could not be used in the an- Somateria mollissima. alysis because the head or bill was damaged. Habitat variation. Insofar as it is not Lightly armoured animals which due to differences in the stage of de- mainly crawl on algae, rocks, etc. composition, great individual varia- (Idothea, Lepidonotus) were negative- tion in stomach contents certainly re- ly correlated with the bill length (rs flects differences in habitat and -0.62, P<0.02 ; rs -0 .69, P<0.01) individual habits. Before we analyse and LBI (rs -0.70 ; P<0.01 ; the significance of bill size, it is necessary to check whether morpho- rs -0.68, P=0 .01) rather than with logically different grebes systematical- SBI (rs -0.36, P>0.05; rs -0.45, ly select different microhabitats. P=0.05) . Still stronger negative Individuals of Nereis pelagica and correlations, with SBI as well were Nephthys measuring 5-15 cm must found for Pectinaria, whose sand seem very similar to a grebe in tubes project only slightly above the appearance, mode of swimming and bottom ooze (rs -0 .70, P<0.01 ; r s the way in which they can be mandib- -0.84, P<0.01 ; rs -0.77, P<0.01) . ulated. However, they select markedly Annelids which often swim freely different habitats. Nereis pelagica show slighter correlations with the occurs in offshore areas with rocks, morphology. However, Nereis with boulders and crevices, particularly mandibles measuring less than 5 mm places with Fucus or Laminaria. showed a slight negative correlation Nephthys inhabits offshore mud- with the bill length (r s -0.66, P< bottoms. Some grebes had taken only Nereis, other numerous 0.02) . Maybe there is a tendency for Nephthys, large-billed grebes to eat a greater which suggests different feeding habitats. However, the ratio between the proportion of big Nereis (LBI r s two prey types was not closely cor- -0.52, P<0.05, but SBI rs -0 .40, related with the grebe morphology P>0.05) . (Spearman's rank correlation tests; Rapid nekton was mainly taken by ratio calculated for sub-samples of long-billed birds. In -the case of Alo- each six specimens in a series). If the teuthis, there was a stronger correla- birds were arranged according to the tion with the bill length (rs -0.80, wing lengths, rs was 0.02. Arrange- P<0.01) and LBI (rs -0.84, P<0.01)

J. Fjeldsd : Functional anatomy of North European Red-necked Grebes 95

Fcc. 4. Number of prey items (semilog. scale) in the grebe stomachs in relation to the "slender bill index" (SBI) . The top diagram shows the numbers of the squid Aloteuthis subulata and fish, the bottom diagram the numbers of the worms Pectinaria and Lepidonotus and the crustacean Idothea. Large symbols mean that the stomach did not contain any of these prey types. Round symbols represent stomachs of Red-necked Grebes, square symbols stomachs of Great Crested Grebes .

than with SBI (rs -0.68, P=0.01) . eating is associated with the long and Fishes were too poorly represented for slender bill type characteristic of a rank correlation test. One typical North Finnish/North Russian birds. South Scandinavian bird had taken No prey type showed significant 11 Gasterosteus . If this bird is omit- correlations with the wing lengths or ted, the 12 birds which had taken fish bill depths . were among the 30 longest-billed birds In conclusion, Red-necked Grebes or the 13 birds with the highest SBI of South Scandinavian types may (Fig. 4) . According to one-sample runs mainly feed by foliage-gleaning and tests, this gives P=0.34 and PG0.02, picking from the bottom, supplement- respectively, which suggests that fish- ed by capture of some squids and

96 ORNIs FENNICA VOI. 59, 1982

TABLE 2 . Similarity between diets of Red-necked and, Great Crested Grebes from the Hatter Rev winter quarters (calculated from the percentage occurrence of Idothea, Lepidonotus, Ne- reis, Nephthys, Glycera, Pectinaria, Gastropoda, Aloteuthis and fish in series of six stomachs) . The Red-necked Grebes are arranged in order of increasing wing-length, bill-length, LBI and SBI. The two samples with the highest SBI correspond to the North Finnish/Russian breeding population of Red-necked Grebes.

Red-necked Grebes Indices of overlap with Great Crested Grebes arranged by

wing- 0.101 0.089 0.165 0.164 0.241 01.189, 0.202 0.281 0.270 0.192 0.090, 0.1,81 0.419 length bill- 0.0,92 0.127 0.096 0.109 0.135 0.408 0.102 0.100 0.184 0.096, 0.190 0,227 0.49,2 length LBI 0.091 0.097 0.122 0.120 0.197 0.085 0.297 0.166 0.231 0.119 0.291 0.102 0.203 SBI 0.005 0.098 0.101 (1`105 0.096 0.126 0.184 0.1617 0.094 0 .135 0.165 0.59-6 0.630

maybe some large annelids by the Table 2 shows how the diet similar- largest-billed birds. The slender-billed ity between the grebe species increases northern birds specialize in eating with the wing length, bill length, LBI squids and fish. and SBI of the Red-necked Grebes . Since all the size parameters are part- Stomach contents of Great Crested ly connected, all the series show an Grebes . The stomachs of the six Great increasing trend. However, the devel- Crested Grebes contained, besides the opment of an extremely slender bill, usual feather mush and some plant evident in North Finnish/Russian Red- debris, remains of 1 Idothea, 9 Lepi- necked Grebes and 12-15 individuals donotus, 10 small and one large Ne- in the present sample, is definitely the reis, 1 Nephthys, 2 Glycera, 29 Alo- factor most clearly causing a shift in teuthis, 93 Gasterosteus, 1 Spinachi-a, the diet of Red-necked Grebes to- 1 Micromesistius poutassou. Squids wards that of the Great Crested and fish clearly predominated by Grebe. mass.

Diet similarity . The similarity of the Discussion diets of the Red-necked and Great Crested Grebes can be calculated as The anatomical approach suggests that the North Finnish/Russian Red-necked s Grebe type converges with allopatric 2 Y~ xi yl Great Crested i=1 Grebes in some C2 - functional properties -of the feeding apparatus x2 -1- E y2 . The study of the feeding i=1 i=1 ecology shows that the morphological/ anatomical characters characteristic of (Morisita 1959), where the two pre- these northern Red-necked Grebes is dators utilize each of the i . . . s re- just the kind of development which source states to the extents of x and y, best secures a convergence in diet. The respectively. Identical diets give the concordant results of the two app- value 1 .00. roaches show that functional anatomy, J. Fjeldsd: Functional anatomy of North European Red-necked Grebes 97 largely neglected in contemporary these Red-necked Grebes will resemble European ornithology, may give a Great Crested Grebes in their summer basis for ecological predictions, or at diet as well. least be smplortant for explaining eco- We cannot prove here that the geo- logical differences between species or graphical variation in Red-necked populations. Grebes was caused by selection due to Similarly, :studies of the relation be- the presence or absence of Great tween the bill anatomy and the diet Crested Grebes (character displace- of Horned Grebes Podiceps auritus ment/character release) . However, (Fjeldså 1973, reanalysed in Fjeldså previous writers (e.g., On-no 1958, 1982) provide evidence that what Markuze 1965, Berndt & Drenkhahn birds eat is, at least in part, determin- 1974) have reported that the two spec- ed by their anatomy/morphology . The ies are well separated from each other presence of inherent "search images" ecologically, when breeding sympatric- for prey, or learning of the diet dur- ally. Our present results give reason ing feeding by the parents cannot be to believe that they occur as ecological excluded as factors affecting the counterparts when allopatric, and fur- choice of food. The diet described ther that the anatomical/morphological here, in winter quarters with annelids change shown in allopatry is and squids, can, however, hardly be precisely the change which best secures due to learning during feeding by the a switch to the diet of the Great parents. The individual variation in Crested Grebe. Although this does not prey rather suggests that a substantial prove anything it certainly makes dis- part of the development of food selec- placement a likely explanation of the tion is steered by the anatomy, which variation. excluded some prey animals, makes it There has been considerable sceptic- inconvenient or uneconomic to feed ism recently regarding the validity of on some and easy to feed on others. the theory of character displacement This might, by ":trial and error", lead (see e.g. Grant 1972, Connell 1980). to specialization on those prey animals The evidence in favour of -it may be which give the highest yield/effort discounted by arguing that if suf- ratio (cf. theory of optimal feeding, ficiently many species are studied, see e.g. Krebs & Cowie 1976). some of the observations will happen Morphology-related variations in to accord with the theory, even if the diet could not be documented in greb- theory is wrong. This argument can- es with an unpredictable food supply not, however, be used to reject the or a supply so sparse that !the grebes evidence found in grebes,, since need to eat every potential food type throughout the world, in every case of available (Fjeldså 1981, 1982 and local contact between two closely alli- unpublished results on Australian ed grebe species, the variation is as grebes) . This is predicted by the predicted by the theory (Fjeldså 1982) . theory of optimal feeding (Schoener 1971) : Specialization. on an optimal fraction of the food supply requires Acknowledgements . My best thanks are due that it is easy to find enough food. If to the Game Biology Station, Kalo, for provid- this is the case in the breeding habit- ing the grebes on which this study was made . Mrs. Ursula Friis did the final typing of the ats of the North Finnish/Russian Red- manuscript, and Dr. Marc Ballan made the necked Grebes, we may postulate that language corrections.

9+8 ORNIS FENNICA VOI. 59, 1982

Selostus : Pohjois-Euroopan härkälintu- FJELDSÅ, J. 19,73 : Feeding and habitat selection of the horned jen nokan ja leuan anatomian alueelli- grebe, Podiceps auritus (Aves), in the breeding season. - Viden sesta muuntelusta ja sen adaptiivisesta sk . Medd . Dansk Naturhist. Foren. 136 : merkityksestä 57-95. FJELDSÅ, J. 19 81 : Comparative ecology of Härkälinnun pohjoisimpien pesimäkantojen Peruvian grebes . A study on the mechan- linnuilla Suomessa ja Venäjällä on pitkä ja isms of evolution of ecological isolation. kapea nokka. Tämä rakennepärre, samoin kuin - Vidensk. Medd . Dansk Naturhist. Itä-Siperian ja Etelä-Amerikan härkälintujen Foren. 144:125-246 . suuri koko, johtunee siitä, että laji on siirtynyt FJELDSÅ, J. 1982 : Ecological character dis- ruokavaliossaan selkärangattomista kaloihin placement and character release in grebes . niissä osissa asuinaluettaan, mistä silkkiliikku - Ibis : in press. puuttuu. Kuvassa 1 on esitetty museonäytteistä FOLKESTAD, A. O. 1978, : Winter censuses and härkälintujen (mustat ympyrät) ja silkkiuik- studies on the Red-necked Grebe Podiceps kujen (avoimet ympyrät) nokan pituus eri grisegena along the Norwegian coast. - osissa maapalfloa . Tanskan talve~htimisalueilla Anser Suppl. No . 3:84-89 . tammikuussa 197'9 öljytuhon uhreista kerätyt GRANT, P. R. 1972 : Convergent and, divergent 87 härkälintua edustivat sekä Etelä-Skandina- character displacement . - Biol . J. Linn. vian lyhytnokkaista että pohjoisempaa pitkä- Soc. 4:3!9--68 . nokkais!ta kantaa . Näistä yksilöistä tutkittiin, HORTLING, I. 1929': En ny geografisk ras av missä määrin rakenne-erot (ks . kuvia 2 ja 3) gråhakedopping : Podiceps griseigena vaikuttavat ruokavalioon (kuva 4) . schioleri? - Ornis Fennica 6 :11.-14. Anatomisten ominaisuuksiensa ansiosta pi,t- KREBS, J. R. & R. J. COwIE 19 76 : Foraging känokkaiset härkälinnut pystyvät silkkiuikun strategies in birds. - Ardea 64 :9 8-116. tavoin aukaisemaan nokkansa äkkiä ja pitä- LARSEN, A. H. 1919 : Oliedoden. - Vågen mään saalista lujassa puristuksessa, mikä on 1979'(2) :23. epäilemättä eduksi kalanpyynnissä . Pitkänok- MADSEN, F . J. 1957 : On the food habits of kaiset yksilöt olivatkin mahanäytteiden perus- some fish-eating birds in Denmark. - teellia erikoistuneet kalojen ja mustekalojen Dan. Rev. Game Biol . 3 :19-83 . saalistukseen. Sen sijaan lyhytnokkais,et härkä- MARKUZE, K. 1965 : K ekologii vidov pogonok linnut olivat syöneet etupäässä nivelmatoja v delte Volgi so svjazom s rybovodstvom. pohjasta tai kasvillisuudesta . Tulokset osoitta- -- Ornitologia 7 :244257 . vat, että anatomiset ominaisuudet määräävät MORISITA, M. 1959 : Measuring the inter- paljolti sen, mitä lintu käyttää ravinnokseen . specific associations and similarity be- Ne viittaavat myös siihen, että härkälintujen tween communities. - Mem. Fac. Sci., morfologian maantieteellinen muunte~lu joh- Kyushu Univ. Ser. E31 :65-80 . tuisi luonnonvalinnan kautta sil'kkiuikun läs- ONNO, S. 1958 : Sravniteljnaia ekologija vidov näol~osta tai puuttumisesta. pogonok Estonii. - Akad . Nauk . Estn . SSR. Inst . Zool . Bot. Tartu. Disserta- tion . References PALMER, R. S. (ed.) 1962 : Handbook of North BAMS, B. A. 1956 : On the relation between American birds. Vol. 1. - New Haven- the attachment of the jaw muscles and London . STEGEL, the surface of the skull in Podiceps cris- S . 1956 : Non-parametric statistics for tatus L., with some notes on the properties the behavioural sciences . - Kogakusha. of the head . - Proc . K. ned. Akad . Wet. STORER, R. R. 197'9 : Order Podicipediformes . Ser. C 59 :82-101, 248-262. - In E. MAYR & G. W. COTTRELL BERNDT, R. K. & D. DRENCKHAHN 1974 : Vo- (eds .) : Check-list of birds of the world', gelwelt Schleswig-Holsteins. Vol . 1 . - Vol. 1 . 2nd ed .: 140-155. Cambridge Kiel . Mass. Bocx, W. J. 1'9'74 : The avian skeletomuscular THOMASSON, K. 1953 : Die Verbreitung des system. - In D. S . FARNER & W. B. KING Rothalstauchers, Podiceps g. griseigena (eds.) : Avian biology, Vol. IV : 119- Bodd . in N.W .-Europa. - Zool . Bidrag 257 . New York . Uppsala 30 :157-166,. CONNELL, J. H. 1980 : Diversity and the co- Zusf, R. L. & R. W. STORER 1969 : Osteology evolution of competitors, or the ghost of and myology of the head and neck of the competition past. - Oikos 35 :134-138 . pied-billed grebes (Podilymbus) . -Misc . CRAMP, S. & K. E . L . SIMMONS 197'7 : Birds of Publ . Miss . Zool . Univ. Mich. 139 :1-49 . the Western Palearctic . - Oxford . Received March 1982