S02.2: Cranial kinesis in birds: Consequences for the evolution of the jaw apparatus

Van der Klaauw Laboratory, Leiden University, Evolutionary Morphology, PO Box 9516, NL 2300 RA Leiden, The Netherlands, fax 31 71 527 4900, e-mail zweers@rulsfb.leidenuniv.nl

Almost all modern birds have a kinetic skull that makes the upper jaw movable relative to the neurocranium. There are basically two types of cranial kinesis: prokinesis and rhynchokinesis. Whereas two sub-types of rhynchokinesis are distinguished by several authors (e.g., Bock 1963; Zusi 1984), one occurring in paleognaths and the other in neognaths, no classification of sub-types of prokinesis that are connected to specific jaw functions is available yet. Although several functional analyses are available on the construction and mechanism of the skull and jaw apparatus, the evolutionary radiation of the feeding system by varying the type of cranial kinesis has hardly been explained. This paper is a semi-review of selected key analyses and recent studies that use experimental and mechanical techniques connecting the trophic system to the underlying design of the jaw apparatus. A first conclusion is that having a prokinetic jaw apparatus with a streptostylic quadrate has been one of the key factors for the success of the radiation of the avian trophic system, as well as for the passage of neognathous birds through the KT-boundary and beyond. A second conclusion is that both the generality and plasticity of the design of the prokinetic jaw apparatus has strongly facilitated trophic radiation, with or without a locking mechanism for the lower jaw. A third conclusion is that the prokinetic pecking mechanism and the rhynchokinetic browsing mechanism are ancestral for the neognath and paleognath trophic mechanisms, respectively. A fourth conclusion is that it is highly probable that the jaw apparatuses and, hence, the trophic systems have developed closely and in parallel to three different taxa in the Maniraptora. A fifth conclusion is that passage through the KT-boundary is connected to being able to generate increased biting forces through the horizontal force components of the pterygoid muscles which were released due to a mobilised palate and rostrokinetic skull.

The design and mechanism of the kinetic skull and its connection to the jaw apparatus in birds have challenged anatomists for a long time (e.g., Nitzsch 1816; Marinelli 1928; Schumacher 1929; Kripp 1933). However, neither their diversity nor their evolutionary history are fully understood yet. Until recently, the subject has been approached by comparative descriptions and two-dimensional mechanical analyses. However, functional experimental analysis and testing of the initial assumptions were seldom undertaken. Such analyses became recently possible by innovations in both techniques and conceptual approaches. Comparative morphological approaches were brought together by reviews of Bühler (1981) and Zusi (1993) on the avian jaw apparatus and the diversity of the avian skull, respectively, while Zweers et al. (1994) reviewed functional anatomical analyses of the trophic system. The present paper addresses the old question of the relationship between design and mechanism of the muscle-bone apparatus of the feeding system. It is a preliminary version of a more extensive review on avian kinesis. To do this, first the avian cranial kinesis and jaw mechanism are considered from a general point of view. Then, selected, functional anatomical studies are summarised that connect jaw mechanism and trophic system. Finally, some speculations on patterns of trophic evolution and their connection to the underlying apparatus are proposed. The anatomical terminology as proposed in the NAA II (Baumel 1993; Vanden Berge & Zweers 1993) has been used.

In birds, basically two types of cranial kinesis are found. They separate the Neognathae from the Ratitomorphae. The prokinetic skull is characterised by the articulation of the upper mandible as a fronto-nasal hinge that is connected primarily to the antorbital fenestra, whereas the rhynchokinetic skull is characterised by a bending zone in the slenderised, dorsal premaxillo-nasal bar of the upper mandible, which passes above the enlarged nostrils. Prokinesis is found in Neognathae to provide a hinge between frontal and nasal bones. Rhynchokinesis is found as different types in two groups of birds (Bock 1963). One type of rhynchokinesis is present in ratitomorphs. In this type of rhynchokinesis, kinesis in the upper mandible is primarily connected to the bones around the nares shaping the most rostral portion of an irregularly shaped fenestra that interconnects the orbit, the postorbital and antorbital fenestrae, and the nares. I will refer to this type as primary rhynchokinesis. The second type of rhynchokinesis is present in Neognathae. This type is considered to be a condition that has been derived from prokinesis (Zusi 1984). I will refer to this type as secondary rhynchokinesis.

Bock (1964) has analysed the kinetics of the avian skull and proposed a model for a prokinetic skull in connection to mandibular kinetics. In most extant birds, the lower mandible does not move independently from the cranium and upper mandible. This condition is known as coupled kinesis of the avian jaw apparatus and is often considered to be the generalised condition of the avian jaw mechanism. Bock (1964) distinguishes two kinetic mechanisms: (1) uncoupled kinesis; and (2) coupled kinesis. Uncoupled kinesis is called a quadrato-mandibular design in which the postorbital and lacrimo-mandibular ligaments are absent, or in which the quadrato-articular hinge does not show specific adaptations. Coupled kinesis occurs in two conditions: (1) A specific locking mechanism in the facets of the articulation surfaces; and (2) Inextensible postorbital and lacrimo-mandibular ligaments that span the mandibulo-postorbital gap and are loaded by action of the M. depressor mandibulae (cf. Starck 1940; Barnikol 1952; Bock 1964). Von Kripp (1933) and Manger Cats-Kuenen (1961) also described the locking mechanism, but considered the ligaments to be elastic. According to these authors, the unlocking and depression of the lower jaw is hypothesised to occur through the contraction of the protractor muscle of the pterygoid and quadrate by which the quadrate is rotated towards rostro-medio-dorsal, the postorbital ligament is released, and the upper mandible is elevated. Hence, the upper mandible must be elevated prior to the depression of the lower mandible. None of these authors has tested the loading and extensibility of the ligaments, the actual motion of the mandibles, or the muscle action patterns.

Zusi (1984) reviewed the form, function and evolution of avian rhynchokinesis. He sought to clarify the relationships of cranial kinesis, schizorhiny or holorhiny, the kinematic significance of different kinds of cranial kinesis, and the evolution of rhynchokinesis. He recognised prokinesis, rhynchokinesis, and amphikinesis. He followed Bühler (1981) in subdividing rhynchokinesis into five pragmatic categories: namely proximal, central, distal, double and extensive rhynchokinesis. Zusi (1984) also distinguished holorhinal from schizorhinal skulls (cf. Garrod, 1873). Ratites and tinamous in general have rhynchokinetic and schizorhinal skulls. Ratites have no continuous lateral naso-maxillar bar, and the ends of the discontinuous lateral bar are interconnected by a ligament. Tinamous, however, have a continuous lateral bar, which Zusi (1984) considered to be a primitive condition in Paleognathae (see also Elzanowski 1987). Furthermore, the hinge area in the dorsal bar of the upper jaw lies very far rostral.

There is a long-standing debate about the significance of the special paleognathous palatal complex. The anatomy of the special paleognathous cranium and jaw apparatus has been described by several authors (e.g., Parker 1866; McDowell 1948). All Neognathae lack the strong vomerine connection between the pterygoids and upper mandible of the Paleognathae, in which the orbital and nasal septa are continuous. The latter feature may have necessitated a rostral shift of the hinge towards the nostril area where the premaxilla and nasal bones run parallel to each other. Simonetta (1960), however, suggested that ratite skulls are akinetic, except in the Kiwi Apteryx, and he used unquantified arguments, such as ‘the dorsal bar is too thick’ to support his interpretation. Bock (1963) assumed that the special construction of the palate strongly favours the transmission of large forces onto the upper mandible. However, neither Bock (1963) nor Simonetta (1960) have tested their assumptions. To test these, not only anatomical and two-dimensional mechanics, but also electromyography, three-dimensional kinematics, ligament constraints, and the role of feeding performance must be considered.

A series of functional anatomical analyses use contemporary techniques, such as electromyography, radiography, 3D mathematical modelling, force measurements, as well as performance analysis. These analyses focus on correlating cranial kinesis to jaw motion and analyse the significance of cranial kinesis for specific performances of the trophic system. Unfortunately, quantitative physical models are rarely available (but see Kingsolver & Daniel 1983).

Ancestors of waterfowl (Anseriformes), fowl (Galliformes) and waders (Charadriformes) are generally considered to be close to the ancestors of extant neognathous birds. I selected papers on pecking and gaping in fowl, on filter feeding, pecking and drinking in Waterfowl, and on pecking and probing in Waders. In order to get an overview of the initial models and the major modifications of the cranio-mandibular complex underlying trophic systems in modern extant neognathous birds. I reviewed the literature on pecking and gaping in fowl; on filter feeding, pecking, and drinking in waterfowl; and on pecking and probing in wading birds. I also reviewed the literature on the extremely specialised pecking and drinking behaviour of pigeons (Columbidae) to see to what level of specialisation the pecking behaviour can develop. And I reviewed the literature on the husking and pecking mechanisms of finches (Fringillidae) and waxbills (Estrildinae) to explore the significance of the presence or absence of the postorbital ligament, since that ligament is generally considered to be diagnostic of coupled kinesis of the skull and lower mandible. Finally, some recent, yet unpublished experimental data on paleognathous rhynchokinesis are included to see whether it differs fundamentally from neognathous rhynchokinesis.

Models for the cranio-mandibular design and pecking mechanism in chickens Gallus gallus were described by several authors, such as Zusi (1967) and Van den Heuvel (1992). Van den Heuvel (1992) tested the model on the decoupling of the coupled cranio-mandibular kinesis (cf. Bock 1964) by electromyographic, cinematographic and 2D-mechanical analysis. He concluded that coupled kinesis - in the sense that the postorbital ligament becomes loaded as a result of a contraction of the M. depressor mandibulae - does not play a role during feeding because the contraction of the protractor muscle starts 20 msec before the contraction of the depressor muscle. Many individual chickens bounce their beaks forcefully into the substrate and subsequently gape strongly to remove the top soil from the beak. Meeting the gaping function requires specific adaptations for the forceful depression of the lower mandible, such as a caudally extended retroarticular process for the attachment of an enlarged depressor muscle with an optimised lever arm, as is found in chickens.

Waterfowl filter feed, graze, dabble and grub (root out by digging). Therefore, waterfowl can be considered as an example of both the malleability of the trophic design and the multiple applicability of a single trophic system. The grazing function demands forceful closure of the jaws, whereas dabbling and grubbing functions demand very high gaping forces as well, though they do not require contraction velocities as high as those in filter feeding. Except for that speed, none of these demands is qualitatively very different for the underlying muscle-bone apparatus from what we have seen in fowl, so that a similar design and working mechanism can be expected. Zweers (1974) tested whether a decoupling of the coupled cranio-mandibular kinesis was present in mallards Anas platyrhynchos by using electromyography, cinematography, and static 2D-mechanical analysis. Furthermore, Zweers et al. (1977) performed a radiographic analysis of the filter and pump mechanisms. The anatomical analysis showed that, apart from extended lacrimal and postorbital processes, strong postorbital, lacrimo-mandibular, occipito-mandibular and external jugo-mandibular ligaments are present, whereas the lower mandible bears a large caudal retroarticular process, which allows the attachment of enlarged depressor muscles. In addition, the quadrato-mandibular joint bears a longitudinal ridge along the dorsal edge of the mandible as well as a double condyle at the ventral end of the quadrate. This configuration enables the mandible and quadrate shift rostro-caudally relative to each other. Zweers (1974) concluded that coupled kinesis - in the sense that the postorbital ligament becomes actually loaded as a result of a contraction of the M. depressor mandibulae - does play a role during filter feeding. A doubly coupled, doubly loaded model was proposed to explain the forceful, fast repetition of a cyclic jaw movement pattern. Just before opening the jaw, the lacrimo-mandibular and postorbital ligaments are loaded by the contraction of the depressor muscles. This loaded configuration is released by the contraction of the M. protractor quadrati, which rotates the quadrate towards rostro-medial so that the upper mandible is released and the lower mandible is depressed. The reverse occurs while the jaws are closed.

Kooloos et al. (1989) and Kooloos & Zweers (1991) analysed performance gradients of filter feeding in several anatid species, such as the Mallard Anas platyrhynchos, Shoveler Anas clypeata, and Tufted Duck Aythya fuligula. They studied the kinematic work envelope - which represents the set of all possible motions - of the beaks by positioning heads in a stereotactic holder and manipulate the jaws in any desired position. They found that the gapes and amplitudes of bill rotations are positively correlated with the size of the seeds that are strained during filter feeding. Moreover, they found that jaws elevate and depress in relatively elevated position if small kernels are filtered, but in a relatively depressed position for filtering large kernels. Simulation of these data by stereotactic manipulation of freshly killed specimens shows that the beak kinematics during filter feeding carefully adjust the width of the spaces to the size of the strained seeds, while the water-pumping capacity is kept at a maximal level. Kooloos et al. (1989) and Kooloos & Zweers (1991) also found that predictions from the stereotactic model on the most efficient degrees of maxillary rotation for gathering the staple food were confirmed by the observations from their cinematographic analyses. The message from these studies is that the cranio-mandibular design of waterfowl is apparently so plastic that it can adjust its operation to the size of seeds by modulating the gape and the level of elevation of their upper mandible, while it keeps the pumped water volume at a maximal level.

The characteristic features of wading birds, for example in Sandpipers Calidris, such as substrate probing, a lengthened and slenderised beak, and secondary rhynchokinesis, has been studied extensively. The jaw apparatus has been subjected to 2-D mechanical analysis (e.g., Marinelli 1928; Schumacher 1929; Kripp 1933; Hoerschelmann 1970; Burton 1974), and the sensory apparatus was analysed from an ecosensorial perspective (Bolze 1968; see Piersma et al. 1995 for a review). A complete functional-anatomical analysis of the probing mechanism, however, is not available yet, though parts of the picture have become clear. The anatomy of the trophic system has some striking features: (1) The cranium is distally rhynchokinetic and schizorhinal; and (2) there is a strong, back-slanting postorbital ligament which runs parallel to the quadrate and attaches to a prominent ventro-caudally pointing postorbital process. In addition, an external jugo-mandibular ligament is found, which is similar to that in waterfowl, in that it possesses a meniscus that underlies the lateral condyle of the quadrate. Gussekloo and Van der Meij (personal communication) made high-speed-video analyses of feeding Knots Calidris canutus. Mandibular tip kinesis was observed neither during pecking of small food items nor during drinking. However, the lower mandible is depressed during these small pecking and drinking jaw motions, hence, despite the presence of a strong postorbital ligament, there is no coupled kinesis during small gaping movements. However, if large food items are eaten, rhynchokinesis occurs in both the lower and upper mandibular tips. This type of kinesis is found especially when large food items are swallowed and is correlated with the creation of a wide gape. It was also found that the depression of the lower mandibular tip starts prior to the rhynchokinesis of the upper mandibular tip if a wide gape is generated. Furthermore, the depression of the lower mandibular tip can occur without rhynchokinesis of the upper mandibular tip.

Apparently, the characteristic feeding specialisation of many waders consists of probing substrates. Gerritsen & Meijboom (1986) studied the ecomorphology of sandpipers. Based on statistical analyses of probing bouts of sandpipers in trays containing known numbers of regularly distributed food items, they found that Sanderlings Calidris alba performed significantly fewer probing movements to find a prey item than what would be expected if they had to find prey items by touching them directly. Gerritsen & Meijboom (1986) inferred that these Sanderlings had a ‘remote touch’ capacity for recognising prey items at least 2cm away from the tips of their beak (see also Piersma et al. 1998). They also observed that Sanderlings probe with slightly open and fully straightened beaks. Zweers & Gerritsen (1997) made a deductive analysis of the jaw apparatus under the mechanical demands of probing substrates and sensorial control demands of ‘remote touch’. For example, (1) the characteristic back-slanting of the quadrate was interpreted as a means for shock absorption to minimise the risk of disarticulating the quadrate from the lower mandible, (2) the enlarged ventrocaudally pointing retroarticular process was interpreted as a result of the required increase of the depressor muscle mass and optimisation of its working line, and (3) the occurrence of rhynchokinesis at the very tip of the beak was interpreted as a means to grasp submerged prey items with minimal friction and hold them tightly upon retraction. The question is what role the strong postorbital ligament and the meniscus in the external jugo-mandibular ligament may have. It is proposed that the locking mechanism is effective just beyond the small gape that is necessary for probing to ensure that the mandibles remain straight while they penetrate the substrate, thus decreasing the risk of breaking.

Pigeons Columba livia have a cranio-mandibular design that is very similar to that in the Black Crow Corvus corax, which is the model species in many textbooks. However, crows (Corvidae) are not specialised seed eaters. Furthermore, crows perform tip-up drinking in a manner comparable to that of the general avian drinking mechanism as found in chickens (Heidweiller et al. 1992), whereas pigeons have a tip-down mechanism for drinking. Pecking and drinking mechanisms were studied by anatomical, radiographic and cinematographic analyses (Zweers 1982). The skull of the pigeon is prokinetic, and there are strong postorbital and occipito-mandibular ligaments which may cause a coupled kinesis. A thorough re-analysis of the anatomy indicated that several other ligaments were highly constraining the motion possibilities of the jaw apparatus (Van Gennip 1986). Van Gennip & Berkhoudt (1992) developed a mathematical 3-D model of the kinematics of the jaw apparatus. This predicted and confirmed that the locking mechanism of the postorbital ligament is effective, but also that unlocking of the lower mandible can occur either by elevating the upper mandible or by increasing the length of the M. protractor quadrati.

Electromyographic and cinematographic analyses of the pecking mechanism showed that stereotyped, though plastic, motion patterns do occur. Bout & Zeigler (1994a) found that, in the grasping phase, the lower jaw must be ‘unlocked’ prior to its depression through Mm. protractores quadratorum being active a few milliseconds before the onset of a short burst of activity of the Mm. depressores mandibulae. In the subsequent stationing phase, the lower mandible is maintained in an unlocked position when the seed is held between the tips of the beak and, hence, keeps the beak open. Bout & Zeigler (1994a) also found that the M. protractor quadrati and the M. depressor mandibulae were active simultaneously during the stationing phase, so that their actions are additive and cause a sudden and very fast elevation of the upper mandible. Bout & Zeigler (1994b) also analysed the drinking mechanism which consists of a fast opening, a period of sustained opening, and a fast closing of the beak. The myograms show that the contraction of the M. protractor quadrati starts to open the beak. A large burst of contractions of the M. depressor mandibulae starts 20 msec later. During the plateau of sustained opening, only the depressor muscle is active. A comparison of the pecking and drinking mechanisms shows that an unlocking of the lower jaw is a prerequisite for both behaviours.

Finches (Fringillidae: Carduelinae) and waxbills (Passeridae: Estrildinae) are small, seed husking songbirds. They show a wide adaptive radiation of granivory and husking mechanisms for small mono- and dicotyledonous seeds. Even though the ecology and the shape of the bills have been extensively analysed, the functional anatomy of the trophic system has received much less attention. Ziswiler (1965) distinguished two types of seed husking: (1) Crushing, in which a seed shell is crushed by compressing it between the blunt tomial ridge of the lower mandible and a keratinous ridge on the corneous palate; and (2) cutting, in which a seed shell is cut open by back-and-forth motion of the sharp tomium of the lower mandible. In both types, Ziswiler (1965) recognised two phases: opening the shell and removing the shell from the kernel. When Nuijens & Zweers (1997) videotaped the feeding behaviour of Greenfinches Carduelis chloris, they observed a medio-lateral motion of the lower mandible and a unilateral crushing motion, but no rostro-caudal motion of the lower mandible or cutting motion as had been described by Ziswiler (1965). Nuijens & Zweers (1997) suggested that both types of the seed husking mechanism use crushing motions, though in different ways. They described the anatomy of two taxa representing both seed husking types: the Spice Finch Lonchura punctulata as a ‘crusher’, and the Greenfinch as a ‘cutter’ or ‘medio-lateral crusher’. They concluded that a complex of anatomical differences parallels the difference in seed husking mechanisms, which are based on differences in the shape of the articular facets of the quadrato-mandibular joint and in the presence or absence of the postorbital ligament.

Hoese & Westneat (1996) analysed avian cranial kinesis by using quantitative models, including those of four bar linkages and lever systems in the White-throated Sparrow Zonotrichia albicollis. They predicted the type of kinesis under the assumptions of coupled and uncoupled mechanisms and force transfer via either the jugal or the pterygoid. They concluded from their videotapes that the upper and lower jaws are not coupled.

The mechanical role of the postorbital and medial jugo-mandibular ligaments was investigated experimentally by Nuijens & Bout (1998) in the Zebrafinch Taeniopygia guttata, which possesses a postorbital ligament, and the Yellow-fronted Canary Serinus mozambicus, which lacks a postorbital ligament. They measured the resistance of the lower mandible against depression in freshly killed specimens and found unexpected results. The postorbital ligament produces only 20% of the total resistance force on the lower mandible when the upper mandible is fixed in the resting position. Nuijens & Bout (1998, p. 30) concluded that the ‘de-blocking’ of the lower jaw depression ‘by reducing the distance between the insertion of the postorbital ligament and the quadrato-mandibular joint is not the result of tension in that ligament ... In the elevated upper beak position 74% of the drop in the jaw opening resistance is produced by the removing of the jugo-mandibular ligament-meniscus complex...’. They explained this phenomenon by suggesting that the crucial feature that unlocks the lower mandible is not the release of the postorbital ligament, but rather the creation of a space within the quadrato-mandibular joint.

Most Ratitomorphae are cursorial, bulk feeding, herbivorous and omnivorous browsers. Their cranium is considered to be primarily rhynchokinetic (Parker 1866; Mc Dowell 1948). Rheas (Rheidae) do not have a postorbital ligament (Fiedler 1951), but there are no clear references on this feature for other Paleognathae. Yudin (1970; as reported by Elzanowski 1987) considered a ligament that connects the postorbital process to the jugal as a characteristic feature of the Paleognathae and calls it the postorbital ligament. No behavioural and experimental analyses are currently available for review. However, Gussekloo & Zweers (1997) have analysed the feeding mechanism in rheas and have experimentally simulated the movement of the quadrate in freshly killed specimens of ostriches (Struthionidae). Their data deviate strongly from the general picture accepted so far for paleognathous feeding, namely, that rhynchokinesis is an inherent feature of feeding. Gussekloo (personal communication) analysed the kinematics of rheas that were feeding on food items of varying sizes. He found neither rhynchokinesis, nor kinesis of the lower mandibular tip. Also, the size of the gape was not proportional to the size of the food items. Further, Gussekloo (personal communication) developed a 3-D stereoradiographic method using a double X-ray projection in which the movement of the quadrate was simulated by elevating the upper mandible to simulate rhynchokinesis in freshly killed specimens of the Ostrich. He concluded that the quadrate must rotate rostro-medially when rhynchokinesis occurs.

The diversity of cranio-mandibular design among birds is actually much greater than what may even have become clear from the examples discussed so far. However, the summaries and references in the previous sections lead to some preliminary conclusions for the Neognathae. (1) The pecking system and the prokinetic cranio-mandibular muscle-bone design - including the relative location of all bony and muscular elements, such as the postorbital ligament and the jugo-mandibular ligament with its meniscus - are generalisations of the ancestral system of the Neognathae. This system can be modified into a secondary rhynchokinesis to fulfil highly specific functional demands. (2) This design is characterised not only by being highly modifiable, but also by a general applicability. The tendency towards being highly modifiable and the generalisation of the ancestral design may now be understood as a prerequisite for the enormous trophic diversity of modern birds. Much less can be concluded for the Paleognathae as yet. (1) The browsing system and the primary rhynchokinetic design - including the specific palato-pterygoid-vomero-parasphenoid construction and the specific postorbito-jugal and jugo-mandibular ligaments - are generalisations considered of the ancestral system that are characteristic for Paleognathae. (2) So far, however, primary rhynchokinesis has not been found to serve any specific functional demands (but see below). Hypotheses for the evolutionary pathways towards the two proposed ancestral conditions need a robust sauropsid phylogeny.

To understand the evolution of the avian kinetic skull, we need to know how avian prokinesis and rhynchokinesis are related to conditions of cranial kinesis in tetrapods, in particular to those in the dinosaurian Coelurosauria, because this taxon is currently considered to be the ancestral taxon of birds. Several types of cranial kinesis are distinguished in early reptilian diversification. These diverging evolutionary lines are characterised by an increase of biting performance via different lines of development of both temporal fenestration and cranial kinesis. At the beginning, there is the ancestral anapsid reptilian condition in which the skull is a closed dermatocranial box with only small bilateral orbits and nostrils. Three evolutionary lines emerged in early reptiles. First, an anapsid line became modified by an emargination of the originally closed, caudally facing post-temporal fossa, as is found in extant turtles. This feature allowed the expansion of the jaw adductor muscles and, hence, an increase in biting forces, but no cranial kinesis was developed. Second, there is a synapsid-mammalian line in which a temporal fenestra develops initially and later enlarges and merges with the orbit, but in which the cranium remains akinetic. Increased biting forces are developed by the creation of extra space for the enlarged jaw adductor muscles caudal to the orbit.

Third, there is a diapsid condition of tetrapod skulls. These skulls possess an upper and a lower temporal fenestra caudal to the orbit. This evolutionary line splits into two branches, namely one leading to the radiations of the Ichthyosauria, Plesiosauria, and Lepidosauria, such as the lizards and snakes in which the ventral temporal fenestra may disappear, another leading to the Crocodilomorpha, Pterosauria, and Dinosauria. Theropod dinosaurs possess an additional antorbital fenestra rostral to the orbit. The theropod-coelurosaurian lineage is now generally considered to comprise birds (Padian 1998; Quang et al. 1998). Fenestration has strongly increased in Maniraptera. The general phylogeny of the Coelurosauria adopted here is based on a number of contemporary systematic analyses ( see Zweers & Vanden Berge 1998). Despite some controversy (see Chiappe 1995; Fedduccia 1995), it is accepted that bird ancestry is located in the dinosaurs within the Maniraptora very close to the Oviraptoridae (Padian 1998; Qiang et al. 1998; see Zweers & Vanden Berge 1998 for a review), though it is not accepted that birds take root exclusively in dromaeosaurid Maniraptorans.

The modified diapsid conditions comprise various types of cranial kinesis. First, there is metakinesis which is characterised by a motion possibility at the level of the temporal fenestrae in or between the parietal and some occipital bone. Second, there is mesokinesis which is characterised by a motion possibility in the roof of the orbit in or between the parietal and frontal bones. Third and fourth, the avian modifications of prokinesis and rhynchokinesis which were characterised by motion possibilities in the roof of the antorbital fenestra and the nostril, respectively. The ancestral avian cranium is considered to have been highly fenestrated and to have been at the transition of becoming kinetic by merging the dorsal and ventral temporal fenestrae and by losing the squamoso-quadrato-quadratojugal, postorbito-jugal, and lacrimo-jugal bars. Zweers et al. (1997) and Zweers & Vanden Berge (1998) have proposed a triple diversification and, hence, a triple ancestry, in the Maniraptora for the avian trophic system and its underlying muscle-bone jaw apparatus in the Late Jurassic and Early Cretaceous. This proposal is based on the appearance of three different types of cranial kinesis - mesokinesis, prokinesis, and rhynchokinesis, respectively - all of them allowing the generation of extra biting forces by releasing the horizontal component of the force generated by the pterygoid muscles to depress the upper mandible. These cranial types were parallelled by three different types of decoupling the palate from the cranium, and also by three different types of trophic specialisation. These three types were closely resembling the design of the cranium, jaw apparatus and the (assumed) trophic system of the Dromaeosauridae, Ornithomimidae and Troodontidae, respectively. The extinct group comprising the archeopterimorphs, Enantiornithes and Alvarezsauridae was linked to the pre-mesokinetic Dromaeosauridae. The extinct Hesperornithidae and the Ratitomorphae were linked to the pre-primary rhynchokinetic Ornithomimidae. The Neognathae were connected to the pre-prokinetic Troodontidae. If the cranium, jaw apparatus and trophic system were taken as the criterium for ancestry and descendence, then it was assumed that each of these three complexes has an ancestor, and these ancestors on their turn have a common ancestor within the Maniraptora. Finally, in extant birds, the temporal fenestrae, the antorbital fenestra, and the orbit were merged into a single large, irregularly shaped fenestra resulting in either a primary rhynchokinetic skull as found in Ratitomorphae, or a prokinetic skull as in Neognathae. The generalised trophic system of extant Ratites - the browsing system - is considered as their ancestral system, and the pecking system as that of the extant Neognathae.

Bock (1963, p. 49) mentioned that ‘the problem as to why rhynchokinesis evolved in ratites remains - whether it was a primary functional change associated with a new feeding method, or whether it was secondary, the result of some still-unknown primary demand’. Of course, the same question must be answered for prokinesis. Zweers et al. (1997) proposed that a demand for increased biting force, which was assumed to have been generated by the evolution of rhynchokinetic or prokinetic conditions, may have been the selective force for survival at the Cretaceous-Tertiary boundary. They also hypothesized that at least three types of early neognathous ecomorphs with prokinetic skulls and one type of early paleognathous ecomorphs with a rhynchokinetic skull - all having evolved extra biting forces at the upper mandible - were capable of passing the KT-boundary because they would have been able to rely on food resources that were much less likely to have been destroyed by the KT-disasters. The hypothesised ecomorphs were (1) grebe-like catchers eating insect larvae and fish as aquatic pursuit hunters; (2) plover-like shore-probers eating burrowed crabs, worms, and molluscs; (3) rail-like nothornid slicers eating roots and tubers; and (4) tinamou-like browsers eating tough dry food as terrestrial bulk feeders.

Unfortunately, the above considerations consider only the phenomenon of cranial kinesis and its effect on increased biting force, which cannot be related yet more specifically to cranio-mandibular design. Bock (1964) discussed the significance of mandibular movements that are coupled to cranial kinesis. He mentioned six roles for coupled cranial kinesis: (1) Maintaining the mandible in a closed position without muscular effort; (2) gaping mechanisms; (3) maintenance of the primary axis of orientation; (4) increased speed of the closure of the jaws; (5) shock absorbing mechanism; (6) attachment of the jaw closing muscles. These suggestions have not been tested yet, and they are interpreted here as secondary adaptations compared to the primary demand for increased biting forces. The evolution of particular neognathous and paleognathous postorbital and quadrato-jugal ligaments may also be studied by analysing developmental stages, by hybridisation experiments, and by transplantation of neural crest cell populations. The following hypotheses need further testing (1) These ligaments may evolve late in the development from the perimysium of jaw adductor muscles and periosteum if an increase in force resistance to open the jaws were advantageous; (2) these ligaments, being isolated features, can easily be influenced by mutational change and can experience rapid adaptation.

As mentioned earlier, pecking and the design of the underlying muscle-bone system of the jaw apparatus are believed to have evolved as part of the neognathous ancestral trophic system (Zweers et al. 1994, 1997). Most neognathous trophic adaptations can be described as modifications of that behavioural pattern and anatomical design of the jaw apparatus and mouth (Zweers et al. 1994). The pecking behaviour may show some adaptations or omissions if the anatomy of the feeding apparatus has undergone specialisation. But the basic mechanisms of catching, ingestion, and swallowing are maintained. It is the same for the drinking, water intake, and swallowing mechanisms, as they represent basically a tip-up mechanism with three different patterns (see Zweers 1992 for a review). To understand the evolution of the cranio-mandibular design in relation to the radiation of the trophic system, we need to understand how modifiable the cranio-mandibular system is. And in order to do this we need to define the design space - describing the accumulation of all possible modifications of the system - by investigating the conditions that constrain the physical and biological modifiability of the design of the trophic system. Though several qualitative analyses have addressed the evolution of trophic systems at lower taxonomic levels, such analyses are very rare above the family level. Zweers & Vanden Berge (1997), in one such attempt, proposed a hypothesis that infers the evolutionary pathways in the cranio-mandibular design space of the fowl-wader-waterfowl complex, which is based on the flexible adaptation of the cranio-mandibular design under selection from specific trophic demands. They hypothesise that the pecking system is ancestral in neognaths and developed the following scenario (Zweers et al. 1994). At Late Cretaceous or KT-boundary shores, a small wader evolved with a somewhat slenderised and lengthened beak to peck and superficially probe the substrate to discover submerged food items. From this ancestor, a radiation gives rise to pecking-probing plover-like descendants. Additional adaptive radiations give rise to taxa that probe the substrate more deeply to oystercatchers (Haematopidae) that forage by direct touch, and to sandpiper-like and watersnipe-like birds (Scolopacidae) that forage by remote touch. Moreover, as a feature that occurs in conjunction with the behaviour of probing in wet mud, a drop of water may run up and down the beak as a result of adhesion and of opening and closing the beak. If a food item were trapped in that drop of water, and, therefore, were transported along the lengthened and slenderised beak to the rictus to be swallowed, this action would be of selective advantage due to its rapid food transport. This phenomenon is observed in Phalaropes Phalaropus (Rubega 1994) and some Sandpipers Calidris (Gerritsen 1989) and may, indeed, have been selected for as a means of rapid food transport. The food transport mechanism can be strongly improved in its performance if suction is applied instead of adhesion as a transport mechanism. The elevated tongue can easily serve as a piston, albeit a leaking one, in the mouth cavity serving as a cylinder to generate suction when it is retreated. The tongue can force water out of the mouth cavity during protraction. Two separate evolutionary lines towards suction feeding may have developed: (1) Protraction of the tongue in an elevated position and a subsequent forcing of the water out of the front and sides of the bill, as in flamingoes (Phoenicopteridae) (Zweers et al. 1995); and (2) protraction of the tongue in a depressed position and a forcing of the water out of the mouth cavity in the next cycle when the elevated tongue is retracted, as in waterfowl (Kooloos et al. 1989). This hypothesis, however, needs further quantitative experimental testing.

I acknowledge the discussions on an earlier version of this manuscript with Dr. J.C. Vanden Berge. I thank Dr. S.W.S. Gussekloo and Mrs. M. Van der Meij of Leiden University for allowing me to use some recent results of their kinematical analyses on rheas and knots for this review.

Barnikol, A. 1952. Korrelationen in der Ausgestaltung der Schädelform bei Vögeln. Morphologisches Jahrbuch 92: 373-414.

Baumel, J.J. 1993. Handbook of avian anatomy: Nomina Anatomica Avium, 2nd ed. Cambridge; Publications of the Nuttall Ornithological Club 23: 779pp.

Bock, W.J. 1963. Cranial evidence for Ratite affinities. Proceedings XIII International Ornithological Congress 1: 39-54.

Bock, W.J. 1964. Kinetics of the avian skull. Journal of Morphology 114: 1-42.

Bolze, G. 1968. Anordnung und Bau der Herbst’schen Körperchen in Limicolenschnäbeln im Zusammenhang mit der Nahrungsfindung. Zoologischer Anzeiger 181: 321-355.

Bout, R. & Zeigler H.P. 1994a. Drinking behavior and jaw muscle (EMG) activity in the pigeon (Columba livia). Journal of Comparative Physiology 174: 443-450.

Bout, R. & Zeigler H.P. 1994b. Jaw muscle (EMG) activity and amplitude scaling of jaw movements during eating in the pigeon (Columba livia). Journal of Comparative Physiology 174: 433-442.

Bühler, P. 1981. Functional anatomy of the avian jaw apparatus. In: King, A.S. & McLelland, J. (eds) Form and function in birds 2. Oxford; Academic Press: 439-468.

Burton, P.J.K. 1974. Feeding and the feeding apparatus in waders: A study of anatomy and adaptations in the Charadrii. London; British Museum (Natural History): 150pp.

Chiappe, L.M. 1995. The first 85 MY of avian evolution. Nature 378: 349-355.

Elzanowski, A. 1987. Cranial and eyelid muscles and ligaments of the tinamous (Aves: Tinamiformes). Zoologische Jahrbücher, Anatomie 116: 63-118.

Feduccia, A. 1995. Explosive evolution in Tertiary birds and mammals. Science 267: 637-638.

Fiedler, W. 1951. Beiträge zur Morphologie der Kiefermuskulatur der Oscines. Zoologische Jahrbücher, Anatomie 71: 235-288.

Garrod, A.H. 1873. On the value in classification of a peculiarity in the anterior margin of the nasal bones of certain birds. Proceedings of the Zoological Society, London 1873: 33-38.

Gerritsen, A.F.C. 1989. Feeding techniques and the anatomy of the bill in Sandpipers (Calidris). PhD Thesis, University of Leiden, The Netherlands.

Gerritsen, A.F.C. & Meijboom, A. 1986. The role of touch in prey density estimation by Calidris alba. Netherlands Journal of Zoology 36: 530-562.

Gussekloo, S.W.S. & Zweers, G.A. 1997. Feeding adaptations in the Greater Rhea (Rhea americana; Ratitae). Journal of Morphology 232: 262.

Heidweiller, J.; Van Loon, J.A. & Zweers, G.A. 1992. Flexibility of the drinking mechanism in the adult chicken (Gallus gallus) (Aves). Zoomorphology 111: 141-159.

Hoerschelmann, H. 1970. Schnabelform und Nahrungserwerb bei Schnepfenvögeln (Charadriidae und Scolopacidae). Zoologische Anzeiger 184: 302-327.

Hoese, W.J. & Westneat, M.W. 1996. Biomechanics of cranial kinesis in birds: testing linkage models in the White-throated Sparrow (Zonotrichia albicollis). Journal of Morphology 227: 305-320.

Kingsolver, J.G. & Daniel, T.L. 1983. Mechanical determinants of nectar feeding strategy in hummingbirds: energetics, tongue morphology, and licking behavior. Oecologia 60: 214-226.

Kooloos, J.G.M.; Kraaijeveld, A.R.; Langenbroek, G.E.J. & Zweers, G.A. 1989. Comparative mechanics of filter feeding in Anas platyrhynchos, Anas clypeata and Aythya fuligula (Aves, Anseriformes). Zoomorphology 108: 269-290.

Kooloos, J.G.M. & Zweers, G.A. 1991. Integration of pecking, filter feeding and drinking mechanisms in Waterfowl. Acta Biotheoretica 39: 107-140.

Kripp, D. von, 1933. Beiträge zur mechanischen Analyse des Schnabel-Mechanismus. Morphologisches Jahrbuch 72: 541-566.

Manger Cats-Kuenen, C.W.S. 1961. Casque and bill of Rhinoplax vigil (Forst.) in connection with the architecture of the skull. Archief Koninklijke Nederlandse Akademie van Wetenschappen 2, 53: 1-51.

Marinelli, W. 1928. Ueber den Schädel der Schnepfe. Palaeobiologie 1: 135-160.

McDowell, S. 1948. The bony palate of birds. Part I. The Paleaognathae. The Auk 65: 520-549.

Nitzsch, C.L. 1816. Ueber die Bewegung des Oberkiefers der Voegel. Deutsches Archive für Physiologie 2: 361-380.

Nuijens, F.W. & Bout, R.G. 1998. The role of two jaw ligaments in the evolution of Passerines. Zoology 101: 24-33.

Nuijens, F.W. & Zweers, G.A. 1997. Characters discriminating two seed husking mechanisms in Finches (Fringillidae: Carduelinae) and Estrildids (Passeridae; Estrildinae). Journal of Morphology 232: 1-33.

Padian, K. 1998. When is a bird a bird? Nature 393: 729-730.

Parker, W.K. 1866. On the structure and development of the skull in the Ostrich tribe. Philosophical Transactions 1866: 113-183.

Piersma, T.J.; Van Gils, J.; De Goei, P. & Van der Meer, J. 1995. Holling’s functional response model as a tool to link the food-finding mechanism of a probing shorebird with its spatial distribution. Journal of Animal Ecology 64: 493-504.

Piersma, T.J,; Van Aelst, R.; Kurk, K.; Berkhoudt, H. & Maas, L. 1998. A new pressure sensory mechanism for prey detection in shorebirds. Proceedings of the Royal Society, London B 265: 1-7.

Qiang, Ji.; Currie, P.J.; Norell, M.A. & Shu-An, Ji. 1998. Two feathered dinosaurs from northeastern China. Nature 393: 753-761.

Rubega, M.A. & Obst, B.S. 1994. Prey transport by Phalaropes using surface tension: a novel solution to the problems of planktivory. The Auk 110: 169-178.

Schumacher, S. 1929. Zur Mechanik und Verwendungsart des Schnepfenschnabels. Zeitschrift für Morphologie und Oekologie der Tiere 15: 90-108.

Simonetta, A.M. 1960. On the mechanical implications of the avian skull and their bearing on the evolution and classification of birds. Quarterly Review of Biology 35: 206-220.

Starck, D. 1940. Beobachtungen an der Trigeminus-Muskulatur der Nashornvögel nebst Bemerkungen über einige Besonderheiten des Vogelschädels. Morphologisches Jahrbuch 84: 585-623.

Vanden Berge, J.C. & Zweers, G.A. 1993. Myologia. In: Baumel J.J. (ed.) Handbook of avian anatomy: Nomina Anatomica Avium, 2nd ed. Cambridge; Publications of the Nuttall Ornithological Club 23: 189-247.

Van den Heuvel, W.F. 1992. Kinetics of the skull in the chicken (Gallus gallus domesticus). Netherlands Journal of Zoology 42: 561-582.

Van Gennip, E.M.S.J. 1986. The osteology, arthrology and myology of the jaw apparatus of the pigeon (Columba livia). Netherlands Journal of Zoology 36: 1-46.

Van Gennip, E.M.S.J. & Berkhoudt, H. 1992. Skull mechanics in the pigeon, Columba livia, A three-dimensional kinematic model. Journal of Morphology 213: 197-224.

Yudin, K.A. 1970. The biological significance and evolution of cranial kinesis in birds. Trudy Zoology Institute 47: 32-66.

Ziswiler, V. 1965. Zur Kenntniss des Samenöffnens und der Struktur des hörnernen Gaumens bei körnerfressenden Oscines. Journal für Ornithologie 106: 1-48.

Zusi, R.L. 1967. The role of the depressor mandibulae muscle in kinesis of the avian skull. Proceedings of the United States National Museum 123: 1-28.

Zusi, R.L. 1984. A functional and evolutionary analysis of rhynchokinesis in birds. Smithsonian Contributions of Zoology: 395: 1-40.

Zusi, R.L. 1993. Patterns of diversity in the avian skull. In: Hanken, J. & Hall, B.K. (eds) The Skull 2. Chigaco; University of Chicago Press: 391-438.

Zweers, G.A. 1974. Structure, movement and myography of the feeding apparatus of the Mallard (Anas platyrhynchos L.). Netherlands Journal of Zoology 24: 323-467.

Zweers, G.A. 1982 The feeding system of the pigeon (Columba livia L.). Advances in Anatomy, Embryology and Cell Biology 73: 1-108.

Zweers, G.A. 1992. Behavioral mechanisms of avian drinking. Netherlands Journal of Zoology 42: 60-84.

Zweers, G.A. & Vanden Berge, J.C. 1997. Evolutionary transitions in the trophic system of the wader-waterfowl complex. Netherlands Journal of Zoology 47: 255-287.

Zweers, G.A. & Vanden Berge, J.C. 1998. Birds at geological boundaries. Zoology 100: 183-202.

Zweers, G.A. & Gerritsen, A.F.C. 1997. Transitions from pecking to probing mechanisms in waders. Netherlands Journal of Zoology 47: 161-208.

Zweers, G.A.; Berkhoudt, H. & Vanden Berge, J.C. 1994. Behavioral mechanisms of avian feeding. Advances in Comparative and Environmental Physiology 18: 241-279.

Zweers, G.G.; De Jong, F.; Berkhoudt, H, & Vanden Berge, J.C. 1995. Filter feeding in Flamingoes (Phoenicopterus ruber). The Condor 97: 297-324.

Zweers, G.A.; Gerritsen, A.F.C. & Van Kraneburg, P.J. 1977. Mechanics of feeding of the Mallard (Anas platyrhynchos L.). Contributions in Vertebrate Evolution 3: 1-109.

Zweers, G.A.; Vanden Berge, J.C. & Berkhoudt, H. 1997. Evolutionary patterns of avian trophic diversification. Zoology 100: 25-57.