Feeding ecology of the earliest vertebrates

  • Published on

  • View

  • Download


  • ~oological Journal of the Linnean SocieQ (1984), 82: 261-272. With 2 figures

    Feeding ecology of the earliest vertebrates


    Department of

  • 262




    paper is based on the tenet that pre-vertebrates and ancestral vertebrates were microphagous suspension feeders (Jmgensen, 1966). Like ammocoetes and most extant protochordates, they would have employed their pharynx to trap microscopic food particles (phytoplankton and fine detritus. under about 0.2 mm in diameter). This is the traditional view of ancestral vertebrates and recent evidence supports it (Mallatt, 1981): there are similar patterns of cilia tracts in the pharynges of cephalochordates, ascidian 1 unicates, and ammocoetes, and corresponding tracts act to direct cords of food-trapping mucus in the same relative directions (Fig. 1 ) . For exposition of an alternate view, that vertebrates primitively were predaceous, see Jollie (1982) and Gans & Northcutt (1983).

    Assuming a microphagous suspension-feeding ancestry, clues to the poorly documented earliest stage of vertebrate history may lie in a peculiarity of larval lamprey feeding. The ammocoete has been found (Mallatt, 1982) to move water across its feeding apparatus at the lowest rate ever documented for a suspension feeding animal (Table 1). Inability to obtain food-carrying water rapidly implies either that ammocoetes must feed on highly concentrated suspensions to obtain enough nutrient, or else that their metabolic rates are so low that they can survive on a comparatively low rate of food intake. The second possibility is ruled out by the data of Table 2, which indicate ammocoetes expend more metabolic energy per unit of food-carrying water pumped than do other suspension feeders. It follows that ammocoetes require more concentrated f o o d suspensions than do any other known suspension feeder. Available data, although incomplete, are consistent with this view: while most other suspension feeders live and grow in waters containing under 1 mg suspended organic solids per litre (Mallatt, 1982) the lowest concentration at which I was able to obtain ammocoete growth was 4 mg/l (yeast diet: Mallatt, 1983). Suspended food concentrations in the ammocoete habitat have been measured at up to 40 mg/l (J. Moore, in Mallatt, 1982).

    Other fish and amphibians (Table 1 ) pump water at rates (under 200 ml/g/h) nearer those of ammocoetes than of most suspension-feeding invertebrates (over 200 ml/g/h). The mechanical reason for this is not known, (and should be investigated. If it can be assumed, however, that ancestral suspension-feeding vertebrates shared the relatively low pumping rates of later fish, certain conclusions about their ecology necessarily follow. They were not typical suspension feeders: an inability to obtain suspension quickly would have generally excluded them from the open waters of oceans and lakes, which are oligotrophic (< 1 mg/l; Mallatt, 1982). Confined to habitats where suspensions were concentrated, ancestral vertebrates would have been primarily benthic- for organic particles concentrate on and just above substrate surfaces, as they settle from the water column and are then resuspended by bottom currents. This deduction differs from the commonly proposed view that the earliest vertebrates were nektonic, open-water animals (Berrill, 1955; Halstead, 1'369; Jagersten,

    Figure I . Pharyngeal ciliated tracts in larval lampreys and amphioxus, drawn in a composite pharyngeal tube. Only the anterior three pairs of branchial bars are shown. Name.< of the tracts in the lamprey are written above the corresponding names in amphioxus. Directions in which cilia move food-laden mucus are indicated by arrows. The lamprey's gill seam tracts ( i x . 2a) are situated laterally, but initially develop medially as in amphioxus (2b). Given this. the patterns in the two animals correspond closely. Dorsal is above.

  • 264 J. MALLATT

    Table 1. Typical water flow rates across the biological filters of suspension feeders, and

    through the pharynges of fish*

    Animal group

    Suspension feeders Copepods Lamellibranchs Sponges Cladocerans Rotifers Iunicates Frog tadpoles Larval lampreys

    Various fish?

    Average flow rates (ml/g/h)

    1400 800 700 450 300 225 40 30


    *This table summarizes Table 1 in Mallatt (1982). T = 10-20C; all flow rates (F) have been corrected for animal weight (W) according to FzWO 1 5 .

    ?Calculated from Sheltons (1970) Table 1; non-hypoxic conditions only: shark, Scylior/pus: 70 ml/g/h; shark, Squulus: 200; eel, Ansuillu: 20; trout, Sulmo: 45, 130; sucker, Cutostomus: 75; bullhead, Ictulurus: 60; carp, Cyprinus: 35.

    fMuch higher rates than this characterize fish that employ ram ventilation instead of muscular pharyngeal pumping. Flow rates for anchovies are about 2000 ml/g/h (Leong & OConnell, 1969).

    1972), but it is consistent with the evidently benthic adaptations (Moy-Thomas & Miles, 197 1) of most ostracoderms.

    A requirement for highly concentrated suspensions would select for the ability to find appropriately rich patches of suspended food, and I reconstruct ancestral fishes as shifting about the surface of the ocean floor exploiting such microhabitats. They are pictured as actively seeking food in a non-homogeneous environment, resembling predators in this respect. This benthonecktonic existence is consistent with the derivation of vertebrate sensory structures (eye, ear, electroreceptive organs) and vertebrate activity metabolism (Ruben & Bennett, 1980), and it can answer the often asked question fJollie, 1977: 79-80) of how benthic suspension feeders could have evolved the features of active existence. T o equate benthic suspension feeding with inactivity ignores the ecology of amphioxus and larval lampreys: both are exceptionally active within the sediments they inhabit and are highly sensitive to changing local conditions (Webb & Hill, 1958; Webb, 1975; and pers. obs.).

    The above line of reasoning depends on an assumption, which should be stated. Some extant fish employ ram ventilation (Roberts, 1975): by swimming constantly with mouth open, they can process water fast enough to live as typical open-water suspension-feeders (e.g. anchovies: Leong & OConnell, 1969). The above model assumes ancestral fish did not primarily utilize ram ventilation (most living fish do not).


    Table 2. For various suspension feeders: ratio of rate at which water is moved (ml/g/h) per unit of oxygen consumed (pl/g/h) *. Note this ratio is

    smallest for larval lampreys


    Copepods 5-60 Lamelli branchs 4-80 Sponges 4-25 Tunicates 13-300 Bryozoa 12-60 Cladocerans (Duphniu) 1.3-2.6 Larval lampreys

  • 266

    Nekton + J. MALLATT

    8. Macraphnpus particulate feeders



    ,? / /?

    6 Macmpha&s 7 Micmphagous 4 Fish suspenvon particulate parasites

    I Microphapus suspension feeders

    2 Depasif &eders 2 b Select big

    mud swallowers

    cans and osteostracans 3 lnmprey ancestors, anaspids, osteostrocans ( 4 Adult lomprey

    7 Many larval ostrocoderms 8 Primitive gnothostome fish

    Figure 2 The initial radiation of vertebrate feeding types, from a proposed microphagous suspension feeding ancestral condition. According to this model, ancestral vertebrates (1 ) were entirely benthic and included no open water species Much of this radiation could have occurred in estuaries Vertebrates would have occupied all feeding zones shown (except 4) by the beginning of the Devonian period Each zone could have been occupied by several different fish groups through convergence, so no rigid phylogenetic sequence is implied Complex life cycles are expected among early fish, with larvae and adults feeding differently In the evolution of gnathostomes ( 8 ) , the sequence 1 +7+8 is considered more likely than the other possibility, 1 +6+8

    Myxinoids feed by foraging through the mud for infaunal organisms such as worms and shrimps ((Strahan, 1963; Shelton, 1978). Gut analyses indicate hagfish ingest small quantities of mud. I t should be emphasized that, although hagfish can also feed epifaunally, teleost fish do not comprise a significant part of their normal diet (Strahan & Honma, 1960).

    Jawless vertebrates possess no hyoid or opercular apparatus so it is reasonable to assume the primitive vertebrate pharynx could produce only weak suction (i.e., via elastic recoil of branchial arches during ventilatory inspiration: Hughes & Ballintijn, 1965; Alexander, 198 1 : 45). Thus mouth, rather than pharyngeal, features are expected to have become specialized for the uptake of heavy particles in early deposit feeders: a variety of oro-bucchal scooping, raking, and suction devices should have existed. While such structures may be echoed in the complex mouth parts of modern lampreys and hagfish (Reynolds, 1931; Dawson, 1963), direct documentation has been limited to the oral plates of some heterostracans (Moy-Thomas & Miles, 1971). It is therefore significant that two recently discovered, well preserved Palaeozoic agnathans exhibit strikingly elaborate mouth parts (Bardack & Richardson, 1977).

    Deposit feeding is ecologically related to feeding upon algae that cover rocks (Fig. 2, epilithic algal feeding). This continuum is illustrated by African cichlids where some species facultatively alternate between collecting food particle films from sediment surfaces and removing periphyton from rocks (Fryer & Iles, 1972: 72). Many cichlid species exclusively utilize the latter food source. Epilithic grazing is a reasonable feeding mode to propose for jawless fish, as it is widespread among aquatic animals (Hynes, 1970: ch. 4), biting mouthparts are


    not required (scraping and combing elements can suffice), and the algal food source undoubtedly existed in the Palaeozoic.

    Among extant epilithic algavores, some stream-dwelling teleosts and tadpoles cling to rocks via suctorial lips, not only to scrape off the periphyton but also to resist fast currents (Noble, 1931; Alexander, 1974; Lagler et ,d., 1977). This could have been the condition in anaspid ostracoderms (Parrington, 1958) and ancestral lampreys (Ritchie, 1968), groups which seem to have included fluviatile forms (Denison, 1956; Hardisty, 1979). Extant suctorial teleosts resemble adult lamDrew in mouth structure and both use tidal ventilation for

    I ,

    irrigating the gills when the mouth is engaged (Randall, 1972; Alexander, 1974).

    A behaviour of teleost rock-scraping fish may show how parasitism evolved in lampreys. Neither the loricariid Hypostomus plecostomus nor some carp minnows (Cyprinidae) can be held with fish that produce significant quantities of epidermal mucus, for they scrape off the mucous layer and open welts in the victims sides (Lagler et al., 1977). The nutritional value of fish epidermal mucus is discussed by Gorlick ( 1980).

    After vascular plants became abundant in the Devonian period (Andrews, 1961; Taylor, 1981), periphyton scraped from plants would have been a potential food source for fish.


    For the earliest known vertebrates (Cambrian to Ordovician), fossil evidence points to a fully marine, often near-shore, habitat (Darby, 1982). There was an unexplained preference for sandy sediment of medium grain size.

    For vertebrates that lived somewhat later (Ordovician to Silurian), estuaries and other marine fringing zones have been considered potenl ially important habitats (Northcutt & Gans, 1983), primarily because they could have served as springboards for the subsequent invasion of fresh water (Denison, 1956; Heintz, 1963; Spjeldnaes, 1967). Nutrient considerations also suggest the importance of such environments. If ancestral fish required exceptionally high concentrations of suspended organic particles, this need would have been met by estuarine waters, where land-derived nutrients stimulate phytoplankton growth, and tidal action continually resuspends settled detritus (Barnes, 1974; Teal, 1980). For deposit feeders (Fig. Z), the substrate of estuaries supports rich growths of algae (possibly richer in early Palaeozoic times before large shading plants appeared), bacteria, and small animals (meiofauna). There is fossil evidence that some ostracoderm groups initially inhabited marine-fringing environments (cephalaspidomorphs in the mid-Silurian: Denison, 1956).

    Vertebrates entered fresh waters no later than late Silurian times (Denison, 1956). Initially, eutrophic lakes may have been favourable habitats (based on the cichlid analogy). Palaeozoic rivers and streams, on the other hand, could have provided only limited opportunity: only the largest rivers support plankton (Hynes, 1970), and detritus content in rivers and streams woulcl have been low prior to the widespread occurrence of terrestrial plants, whose fallen leaves support most extant fluviatile food webs (Cummins & Klug, 1980; Pomeroy, 1980). Initially, rock scraping may have been the only mode of feeding widely available to fish in fluviatile environments.

  • 268 J. MALLATT


    An evolution of jawed predaceous fish from microphagous suspension-feeding ancestors necessarily involved two changes: increasing the size of particles taken (microphagy to macrophagy) , and attaining selectivity for individual particles (suspension to particulate feeding). Depending on which change occurred first, two models of gnathostome derivation are possible.

    Alternative 1

    By one scheme (Fig. 2: Stages 1+6+8) a filtrate of originally microscopic food particles came to include macroscopic zooplanktonic organisms of progressively larger sizes. (Zooplankton would have been a valuable energy-rich nutrient supplement if early fish were limited in the rate at which they could obtain suspension, as proposed above.) Later, as discriminatory abilities increased (better vision), capture of the most desirable individual zooplankton organisms would have become possible. Extant anuran tadpoles can be arranged in a continuum of feeding types from microphagous to macrophagous suspension feeders to obligate carnivores (Wassersug & Hoff, 1979; Wassersug & Rosenberg, 1979) and this reflects the pregnathostome radiation proposed here.

    Estuarine waters support rich concentrations of zooplankton as well as phytoplankton (Barnes, 1974). Estuaries thus could have been primary sites for the evolution of zooplanktivory in pregnathostomes. Some modern estuaries support populations of macrophagous suspension-feeding fish (Blaber, 1979; Mann, 1980). Zooplanktonic organisms available as prey in Palaeozoic times included ostracods, protozoans, and the pelagic larvae of many metazoans Uagersten, 1972; Tasch, 1973), but not todays dominant crustacean groups (cladocerans, copepods: also see Miiller, 1983).

    There is some indication, however, that the pre-gnathostome line never passed through a macrophagous suspension-feeding stage. Extant gnathostome fish of this feeding type show rather extreme structural features that contrast markedly with traits of primitive gnathostomes. Gill rakers are elongate and abundant, the pharynx tends to be highly expandable, and the teeth on the upper and lower jaws are degenerate (Leong & OConnell, 1969; Lagler et al., 1977; Moss, 1977; Rosen & Hales, 1981). By contrast, primitive gnathostomes seem to have had well developed jaws and teeth, and only a limited capacity for expanding the pharynx (early osteichthyes: Lauder, 1982; placoderms: Miles, 1967; biting in primitive sharks: Moss, 1977). Acanthodians (Moy-Thomas & Miles, 197 l ) , which initially possessed the typical features of gnathostome fish, diverged markedly in morphology with the apparent adoption of macrophagous suspension feeding in late forms.

    Even if pre-gnathostomes were not macrophagous suspension feeders, other early fish may have been (pteraspid heterostracans: Halstead, 1969).

    Alternative 2 The alternative to the above scheme is that microphagous suspension-feeding

    ancestors abandoned suspension feeding early to take small individual particles (microphagous particulate-feeding) , with an increase in the particle sizes taken


    following later in phylogeny (Fig. 2: Stages 1+7+8). One way this switch to particulate feeding could have occurred is through neoteny. Most larval fish necessarily take individual particles (Harden Jones, 1980), due 10 the tiny sizes of their mouths. This holds true even when their adult stages are suspension feeders (Ruelle & Hudson, 1972; Drenner et al., 1982), and it should have been the case for many Palaeozoic jawless fish*-for even amphiotxus larvae are particulate feeders (Webb, 1969; Gosselck & Kuehner, 1973). I t is proposed that retention of larval proportions and feeding behaviours occurred in pregnathostomes, resulting in a prolongation of particulate feeding throughout the life cycle and the elimination of an originally suspension-feeding stage in the adult; progressively larger particles would have been utilized a!; the fish grew. The proposed neotenic step (Gould, 1977) could explain why comparative anatomists have long considered the adult head of gnathostonne fish to be a better reflector of the vertebrate embryonic plan than are the heads of known aganthans. It should be stressed that the neotenic event advanced here is not equivalent to that sometimes proposed (Berrill, 1955) to have produced vertebrates from tunicates.

    Once particulate feeding had evolved in the lineage leading to gnathostomes, macrophagy would have directly followed. Observations on zooplanktivorous predatory fish indicate preference for the larger prey items present (Hughes, 1980; Gardner, 198 1; Eggers, 1982), and this is expected since 1a.rger prey offer higher yield per capture effort. Assuming no competition, the earliest gnathostomes should have risen rapidly to the top of the food chain. Indeed, it is noteworthy that most of the oldest known gnathostome groups (cladodont sharks, arthrodires, large palaeoniscids, early dipnoans and crossopterygians) seem to have fed on large prey (e.g., other fish, molluscs: Moy-Thomas & Miles, 197 1 ) . The initial gnathostome (or pre-gnathostome) feeding niche of predation on zooplankton could have been retained by small acarithodians and paleoniscids, and by the larvae of other jawed groups. However, today's great range of zooplanktivorous teleosts (Lauder, 1982) seems to have been conspicuously without analogue in the Palaeozoic record.


    ( 1 ) Early vertebrate evolution is discussed in an ecological context (Fig. 2) , using modern animals and environments to evaluate Palaeozoic (counterparts.

    (2) Comparison of the pharynx in protochordates and larval lampreys (Fig. 1 ) supports the idea that ancestral vertebrates were microphagous suspension feeders.

    ( 3 ) Assuming that ancestral vertebrates pumped water like extant fish (and did not employ ram ventilation in feeding), they must have obtained food suspensions so slowly as to require eutrophic, mostly benthic habitats. Highly

    * I assume that ancestral marine vertebrates, including pre-gnathostomes, had complex life cycles with planktonic (and planktivorous) larvae and benthic adults. Benthic marine animals usually have pelagic larvae (Mann, 1980), and such a life cycle seems to have been primitive for most marine metazom groups (Jagersten, 1972). Pelagic larvae are not typical for fluviatile or deep-sea animals, however, due to unidirectional water currents instreams, and paucity of food in the deep sea water column (Mann, 1980). By this reasoning, lamprey and hagfish life cycles are not primitive for vertebrates.

  • 270 J. MALLATT

    productive marine fringing environments exhibit the appropriate conditions. Ancestral vertebrates are modelled as relatively active, epifaunal (not infaunal) organisms. (4) From this benthic suspension-feeding ancestry, agnathan groups radiated

    into the related niche of deposit feeding, and then to benthic foraging (hagfish) or feeding on epilithic algae (lamprey ancestors). These feeding modes are all interrelated, as illustrated by the recent radiation of cichlid teleosts in African lakes.

    (5) The evolution of gnathostomes (macrophagous particulate feeders) from microphagous suspension-feeding ancestors could have involved either an initial increase in the particle sizes filtered (intermediate stage = macrophagous suspension feeder), or an initial switch to individual particles (microphagous particulate-feeding intermediates). The latter possibility seems more likely, as extant macrophagous suspension-feeding fish show rather extreme speciali- zations.

    (6) Known fossil palaeozoic gnathostomes seem to have clustered at the top of the food chain (large prey). Although the possibility of preservational bias in the fossil record cannot be ruled out, zooplanktivorous fish seem underrepresented in the Palaeozoic (cf. among modern fish).


    Thanks are extended to Quentin Bone, Carl Gans, Philippe Janvier, Malcolm Jollie, Richard Parker, Barbara Stahl, and Richard Wassersug for critical discussion of the ideas in this paper, and to Barbara Comstock and Sue Rose for typing the manuscript.


    ALEXANDER, R . McN., 1974. Funclional Deszgn in Fishes (3rd edition): Chs 4 & 5. London: Hutchinson. ALEXANDER, R. McN., 1981. The Chordates (2nd edition): Chs 2-4. New York: Cambridge University

    ANDREWS, H.N., 1961. Sludies in Paleobotany. New York: John Wiley & Sons. BARDACK, D. & RICHARDSON, E. S., 1977. New agnathous fishes from the Pennsylvanian of Illinois.

    BARNES, R. S. K., 1974. Estuarine Biology. London: Arnold. BERRILL, N.J., 1955. The Origin of Vertebrates. Oxford: Clarendon Press. BLABER, S. J. M., 1979. The biology of filter feeding teleosts in Lake St. Lucia, Zululand. Journal of Fish

    CUMMINGS, K . W. & KLUG, M. J., 1979. Feeding ecology ofstream invertebrates. Annual Review of Ecology

    DARBY, D. G., 1982. The early vertebratc Astraspis, habitat based on a lithologic association. Journal of

    DAWSON, J. A., 1963. The oral cavity, the jaws, and the horny teeth of Myxine glutinosa. In A. Brodal and

    DENISON, R. H., 1956. A review of the habitat of the earliest vertebrates. Fieldiana: Geology, 11: 359-457. DENISON, R. H., 1961. Feeding mechanisms of agnatha and early gnathostomes. American zoolo& 1:

    DRENNER, R. W., VINYARD, G. L., GOPHEN, M. I(r McCOMAS, S. R., 1982. Feeding behaviour of the cichlid, Sarotherodon galilaeum: selective predation on Lake Kinneret zooplankton. Hydrobiologica, 87: 17-20.

    EGGERS, D. M., 1982. Planktivore preference by prey size. Ecology, 63: 381-390. FENCHEL, T., 1980. Suspension feeding in ciliated protozoa: feeding rates and their ecological significance.

    FRYER, G. & ILES, T . D., 1972. The Cichlid Fishes of the Great Lakes of Africa: Ch. 5 . Edinburgh: Oliver &

    GANS, C. & NORTHCUTT, R. G., 1983. Neural crest and the origin of vertebrates: a new head. Science,


    Fieldiana: Geology, 33: 489-510.

    Biology, 15: 37-59.

    and Systematics, 10: 147-172.

    Paleontology, 56: 1187-1 196.

    R. Fange (Eds), The Biology of Myxine: 231-255. Oslo: Universitetsforlaget.


    Microbial Ecology, 6: 13-25.


    220: 268-274.


    GARDNER, M . B., 1981. Mechanisms of size selectivity by planktivorous fish: a test of hypotheses. Ecology, 62: 571-578.

    GORLICK, D. L., 1980. Ingestion of host fish surface mucus by the Hawaiian cleanimg wrasse, Labroides

    GOSSELCK, F. & KUEHNER, E., 1973. Investigations on the biology of Branchiostoma senegalense larvae of f

    GOULD, S. J., 1977. Ontogeny and Phylogeny. Cambridge, Ma.: Belknap, Harvard. HALSTEAD, L. B., 1969. The Patterns of Vertebrate Evolution. Edinburgh: Oliver & Boyd. HARDEN JONES, F. R., 1980. The nekton: production and migration patterns. In R . S. K. Barnes & K.

    Mann (Eds), Fundamenlals of Aquatic Ecosystems: 1 19-142. Boston: Blackwell. HARDISTY, M. W., 1979. Biology of the Cyclostomes. London: Chapman and Hall. HEIN'I'Z, A,, 1963. Phylogenetic aspects of myxinoids. In A. Brodal & R. Fange (Eds), The Biology of Myxine:

    HUGHES, G. M. & BALLANTIJN, C. M., 1965. The muscular basis of the respiratory ]pumps in the dogfish

    HUGHES, R. N., 1980. Strategies for survival of aquatic organisms. In R. S. K. Barnes & K. Mann (Eds),

    HYNES, H. B. N., 1970. The Ecolou ofRunning Waters. Bungay, Suffolk: Richard Clay (Chaucer Press) Ltd. JAGERSTEN, G., 1972. Euolution of the Metazoan LiJe Cycle. New York: Academic Press. JOLLIE, M., 1977. The origin of the vertebrate brain. Annals o f the N e w York Academy ofSciences, 299: 74-86. JOLLIE, M., 1982. What are the 'Calcichordata'? and the larger question of the origin of the chordates.

    JORGENSEN, C. B., 1966. Biology of Suspension Feeding. (International Series of Monographs in Pure and

    LAGLER, K. F., BARDACH, J. E., MILLER, R. R. & PASSINO, D. R. M., 1977. Ichthyology (2nd edition):

    LAUDER, G. V., 1982. Patterns of evolution in the feeding mechanism of actinopterygian fishes. American

    LEONG, R. J. H. & O'CONNELL, C. P., 1969. A laboratory study of particulate and filter feeding of the

    MALLATT, J , , 1979. Surface features of structures within the pharynx of the larval lamprey Petromyzon

    MALLATT, J,, 1981. The feeding mechanism of the larval lamprey Petromyzon marinuc. Journal of ,Soology,

    MALLAI'T, J., 1982. Pumping rates and particle retention efficiencies of the larval lamprey, an unusual

    MALLKIT, J., 1983. Laboratory growth of larval lampreys (Lampetra (Entosphenus) tridtntata Richardson) a t

    MA", K., 1980. Benthic secondary production. In R. S. K. Barnes & K. Mann (Eds), Fundamentals of

    MILES, R. S., 1967. The cervical joint and some aspects of the origin of the F'lacodermi. Cofloques

    MOSS, S. A,, 1977. Feeding mechanisms in sharks. American zoologist, 17: 355-364. MOY-THOMAS, J. A,, & MILES, R. S., 1971. Palueozoic Fishes (2nd edition). Philadelphia: Saunders. MULLER, K. J., 1983. Crustacea with preserved soft parts from the upper Cambrian of Sweden. Lethuia, 16:

    93-109. NEWELL, R. C., JOHNSON, L. G. & KOFOED, L. H., 1977. Adjustment of the components of energy

    balance in response to temperature rhange in Ostrea edulis. Oecologia (Berl in) , 30: 97-1 10. NOBLE, G. K., 1931. Biology of the Amphibia (Paperback edition, 1959). New York: Dover Press. NORTHCUTT, R. G. & CANS, C., 1983. The genesis of neural crest and epidermal placodes: a

    PARRINGTON, F. R., 1958. O n the nature of the anaspida. In T. S. Westoll (Ed.), Studies on Fossil

    POMEROY, L. R., 1980. Detritus and its role as a food source. In R. S. K. Barnes & K. Mann (Eds),

    POTTER, I . C. & ROGERS, M. J., 1972. Oxygen consumption in burrowed and unborrowed ammocoetes

    RANDALL, D. J., 1972. Respiration. In M. W. Hardisty & I. C. Potter (Eds). The Biology ofLamprtys , vol. 2 :

    REPETSKI, J. E., 1978. A fish from the upper Cambrian of North America. Science, 200: 529-531. REYNOLDS, T . L., 1931. Hydrostatics of thc suctorial mouth of the lamprey. Universily of California

    RITCHIE, A., 1968. New evidence on Jamoytius kerwoodi White, an important ostracoderm from the Silurian

    ROBERTS, J. L., 1975. Active branchial and ram gill ventilation in fishes. Biological Bulletin, 148: 85-105.

    phthirophagus (Labridae), and its effect on host species preference. Copeia, 1980 ( 4 ) : 86:1-868.

    the northwest African coast,. Marine Biology, 22: 67-73.

    9-2 I , Oslo: Universitetsforlaget.

    (Sqyl iortpus canicula). Journal of Experimental Biology, 43 : 363-383.

    Fundamentals of Aquatic Ecosystems: 162-184.

    zoological Journal of the Linnean Society, 75: 167- 188.

    Applied Biology/Zoology Division, 27.) Oxford: Pergammon Press.

    Ch. 5. New York: John Wiley & Sons.

  • 272 J. MALLATT

    ROSEN, R. A. & HALES, D. C., 1981. Feeding of paddlefish, Polydon spathula. Copeia, 1981 ( 2 ) : 441-455. RUBEN, J. A. & BENNETT, A. F., 1980. Antiquity of the vertebrate pattern of activity metabolism and its

    possible relations to vertebrate origins. Nature (London), 286: 886-888. RUELLE, R. & HUDSON, P. L., 1977. Paddlefish (Polyodon spathula): growth and food of young of the year

    and a suggested technique for measuring length. Transactions of 6he American Fisheries Society, 106: 609-613. SHELTON, G., 1970. The regulation of breathing. In W. S. Hoar & D. J. Randall (Eds), Fish Physiology, vol.

    IV: 293-359. New York: Academic Press. SHELTON, R. G. J., 1978. O n the feeding of the hagfish Myxine glutinosa in the North Sea. Journal o f t h e

    Marine Biological Association of the Unitcd Kingdom. 58: 81-86. SPJELDNAES, N., 1967. The Paleoecology of the Ordovician vertebrates of the Harding Formation

    (Colorado, U.S.A.). Colloques Intcmationaux du Centre National de la Recherche ScientiJique, 163: 11-20. STRAHAN, R., 1963. The behaviour of Myxine and other myxinoids. I n A. Brodal & R. Fange (Eds), Thc

    Biology of Myxinc: 22-32. Oslo: Universitetsforlaget. STRAHAN, R. & HONMA, V., 1960. Notes on Paramyxine atami Dean and its fishery in the Sado Strait, Sea

    of Japan. Hong Kong University Fisheries Journal, 3: 27-35. TAGHON, G. L., SELF, R. F. L. & JUMARS, P. A., 1978. Predicting particle selection by deposit feeders: a

    model and its implications. Limnology and Oceanography, 23: 752-759. TASCH, P., 173. Paleobiology of the Invertebrates. New York: John Wiley. TAYLOR, T. N., 1981. Paleobotany: An Introduction to Fossil Plant Biology. San Francisco: McGraw-Hill. TEAL, J. M., 1980. Primary production of benthic and fringing plant communities. In R. S. K. Barnes & K.

    THOMSON, K. S., 1971. The adaptation and evolution of early fishes. The Quarterb Review of Biology, 46:

    WASSERSUG, R. & HOFF, K., 1979. A comparative study of the buccal pumping mechanism of tadpoles.

    WASSERSUG, R. & ROSENBERG, K., 1979. Surface anatomy of the secretory epithelium in the branchial

    WEBB, J. E., 1969. On the feeding and behavior of the larva of Branchiostoma lanceolafum. Marine Biology, 3:

    WEBB, J. E., 1975. The distribution of amphioxus. 5jmposium of the