publications of the INSTITUTE of MARINE SCIENCE Voume II SEPTEMBER Number I 1951 Published by The University of Texas Printing Division Austin Table of Contents _Preface ------·---------------------------------------------------------------------------------------------------------·-------· 5 The Littoral Marine Annelids of the Gulf of Mexico____ __ __ ____ __________ Olga Hartman 7 Brackish-Water and Marine Assemblages of the Texas Coast, with Special Reference to Mollusks______________ _ ________________________ __ _____________________ Harry S. Ladd 125 Foraminifera of the South Texas CoasL__ ________ __________________________ _ _______ Rita J. Post 165 Some Principles of Marine Fishery Biology__ _ ___ ___ _ ________ __Martin D. Burkenroad 177 Preface This volume continues publications in marine science with special reference to the Gulf of Mexico. Papers are edited by the staff of the Institute of Marine Science, and there is no established editor. The undersigned acts largely in an "ex officio" capacity. Occasionally referees from other institutions are utilized when required. Referees take full responsibility for their criticisms and complete signed statements will be furnished all authors. By thus procedure the ofttimes superficial and uncritical re­marks of anonymous referees are avoided. Even at its best the anonymous referee system carries the stigma of "sub rosa" action which seelr\s to be contrary to straight­forward democratic principles. In any case, it indicates a lack of desire to shoulder responsibilities and the system appears to have more harmful potentialities than good ones. This journal is operated on the assumption that authors in general know a great deal more about their papers and how they should be presented than editorial staffs do. Therefore, wishes of the authors are usually followed in public~tion even though they might be questioned by the editorial staff. There is no subscription list to this journal. although individual volumes may be purchased. For that reason no provision is made for free reprints. Manuscripts should be submitted to the under­signed. Gordon Gunter, Acting Director, Port Aransas, Texas. The Littoral Marine Annelids of the Gulf of Mexicol By OLGA HARTMAN Allan Hancock Foundation, University of Southern California The aim of the present study is mainly to publish the presence of a rich and flourishing annelid fauna in the Gulf of Mexico. Keys and illustrations together with the bibliographic citations may permit a fairly complete and accurate identifi­cation of the annelids to be encountered in littoral zones. The collections upon which this study was ·based come largely from intertidal or shallow seas; only a few are from deeper waters. This investigation has disclosed the existence of a vast, diversi­fied, possibly highly endemic fauna in this area. The few deeper water species that have been reported show affinities with those known also from the West Indies and the tropical Atlantic Ocean. One hundred fifty-eight species in 110 genera and 36 families of marine annelids are recorded, most of them for the first time from the Gulf of Mexico. There are 15 new species or subspecies, one new name and one new genus. These are listed by family. LIST OF SPECIES POLYNOIDAE Lepidametria' commensal is Webster Halosydna leucohyba Schmarda Lepidonotus . sub levis Verrill H armothoe aculeata Andrews Lepidonotus variabilis Webster H armothoe trimaculata (Treadwell) POLYODONTIDAE Eupanthalis tubifex (Ehlers) Polyodontes lupina (Stimpson) SIGALIONIDAE Sthenelais articulata Kinberg AMPHINOMIDAE Amphinome rostrata (Pallas) Hermodice carunculata (Pallas) Chloeia viridis Schmarda Hip ponoe multibranchiata (Treadwell) Eurythoe complanata (Pallas) Pareurythoe americana, new species PHYLLODOCIDAE Anaitides erythrophyllus (Schmarda) Eumida sanguine a (Oersted) Eteone hetero poda n. sp. N ereiphylla fragilis (Webster) Eteone alba Webster N ereiphylla, near paretti Blainville Eulalia myriacyclum (Schmarda) HESIONIDAE H esione picta Miiller Podarke, near guanica Hoagland lContribution :no. 75 from the Allan Hancock Foundation of the University of Southernt California. PILARGIIDAE Ancistrosyllis bassi Hartman Loandalia americaTUJ Hartman SYLLIDAE Autolytus brevicirrata Winternitz Odontosyllis enopla Verrill Exogone disp4r (Webster) SynsyUis longigularis Verrill Ehlersia, sp. Tryparwsyllis vittigera Ehlers H aplosyUis spongicola (Grube} Typosyllis corallicoides Augener NEREIDAE Ceratonereis irritabilis (Webster) Nereis oligohaliTUJ (Rioja} Ceratonereis mirabilis Kinberg N ereis pelagica occidentalis Hartman Ceratonereis tridentata (Webster) N ereis riisei Grube Laeonereis culveri (Webster) Perinereis anderssoni Kinherg Lycastopsis tecolutlensis Rioja Perinereis floridana (Ehlers) Neanthes succinea (Frey and Leuckart) Platynereis dumerilii (Audouin and M. Nereis largoensis Treadwell Edwards) NEPHTYIDAE Nephtys bucera Ehlers N e phtys picta Ehlers GLYCERIDAE Glycera americana Leidy Glycera dibranchiata Ehlers ONUPHIDAE Diopatra cuprea (Bose) Onuphis eremita oculata, new subspecies H yalinoecia tubicola (Miiller) Onuphis magna. (Andrews) EUNICIDAE Eunice filamentosa Grube Eunice (Nicidion) kinbergi Webster Eunice floridana (Pourtales) Lysidice ninetta Audouin and M. Edwards Eunice mutilata Webster Marphysa regalis Verrill Eunice rubra Grube Marphysa sanguinea (Montagu) Eunice schemacephala Schmarda Palola siciliensis (Grube) LUMBRINERIDAE Lumbrineris alata, new species Lumbrineris parvapedata (Treadwell) Lumbrineris bassi Hartman Ninoe nigripes gracilis, new subspecies Lumbrineris inflata (Moore) ARABELLIDAE Arabella iricolor (Montagu) Drilonereis cylindrica, new species Drilonereis magna (Webster and Bene­dict) LYSARETIDAE Lysarete brasiliensis Kinberg DORVILLEIDAE Dorvillea rubra (Grube) Dorvillea sociabilis ·('W'ebster) Dorvillea rudolphii (delle Chiaje) ORBINIIDAE flaploscoloplos foliosus, new species N aineris laevigata (Grube) Haploscoloplos fragilis (Verrill) N aineris setosa (Verrill) · H aploscolo plos robustus (Verrill) Phylo ornatus (Verrill) Naineris bicornis, new species Scoloplos (Leodamas} rubra (Webster) SPIONIDAE Dispio uncinata, new genus and species Polydora websteri Hartman Nerine agilis Verrill Prionospio heterobranchia texana, new Polydora armata Langerhans subspecies Polydora caulleryi Mesnil Prionospio treadweUi, new name Polydora hamata Webster Prionospio ? cirri/era Wiren Polydora ligni Webster Spio phanes bombyx ( Claparede) Polydora socialis (Schmarda) MAGELONIDAE Magelona, near californica Hartman CHAETOPTERIDAE Chaetopterus variopedatus (Renier) S piochaeto pterus oculatus Webster CIRRATULIDAE Cirratulus hedgpethi, new species Dodecaceria diceri,a, new species Cirriformia filigera (delle Chiaje) Dodecaceria near concharum Oersted ARENICOLIDAE Arenicola cristata Stimpson OPHELIIDAE Armandia agilis (Andrews) Polyophthalmus pictus (Dujardin) FLABELLIGERIDAE Semiodera roberti,. new species Stylarioides inflata (Treadwell) CAPITELLIDAE Capitella capitata (Fabricius) H eteromastus filiformis ( Claparede) ?Capitellides teres Treadwell N otomastus hemipodus Hartman Dasybranchus lumbricoides Grube N otomastus latericeus Sars Dasybranchus, lunulatus Ehlers M ediomastus californiensis Hartman . MALDANIDAE Axiothella mucosa (Andrews) ?Macroclymene elongata (Webster) Branchioasychis americana Hartman . Maldane sarsi Malmgren Clymenella iorquata calida, new sub­ species OWENIIDAE Owenia fusiformis delle Chiaje PECTINARIIDAE Cistenides gouldii Verrill SABELLARIIDAE Sabellaria floridensis Hartman Sabellaria vulgaris beaufortensis Hartman AMPHARETIDAE Amphicteis gunneri floridus, new subsp. Melinna maculata Webster TEREBELLIDAE Enoplobranchus sanguineus Verrill Pista palmata (Verrill) Loimia medusa \Savigny) T ere bella rubra (Verrill) Loimia viridis Moore T erebellides stroemi Sars Pista cristata (Miiller) T helepus setosus (Quatrefages) SABELLIDAE Branchiomma bairdi (Mcintosh) Megalomma bioculatum (Ehlers) Branchiomma nigromaculata (Baird) Megalomma lobiferum (Ehlers) Chone duneri Malmgren Sabella melanostigma Schmarda F abricia, sp. Sabella micro phthalma Verrill Hypsicomus elegans (Webster) Sabellastarte magnifica (Shaw) SERPULIDAE Eupomatus dianthus (Verrill) Protula tubularia (Mwn from New England southward to Florida and in the Gulf of Mexico west to Texas. It occurs in littoral sands. Haploscoloplos robustus (Verrill), 1873 Plate 21, figs. ~ Scoloplos bustorus Eisig, 1914, pp. 422-423; Hartman, 1942, p. 58, figs. 110-112. Haploscoloplos bustorus Hartman, 1944, p. 340. Haploscoloplos bustoris Hom and Bookhout, 1950, pp. 1-8, pis. 1-4. This comes from southwestern Florida, in sand. The species is clearly separable from H aploscoloplos fragilis (above) with which it may be associated for attaining a greater size, in having a prostomium that is less acutely pointed and for having neu~o­podial flanges that are entire (figs. ~),.not divided. Branchiae are first present from about segment 23. The change from thorax to abdomen is at or near 22 to 24. Inter­cirri (fig. 4) are first present from the first branchial segment and continued back through 50 or more segments but absent thereafter. Setae of both thorax and abdomen are distally pointed and transversely denticulate along the cutti~g edge. Furcate setae have not been identified. Color in life is orange red with bright red branchiae. The specific name, bustorus Eisig (see synonymy, above) originated from the as­sumption that Alcandra robusta Kinberg (1866) and Anthostoma robustum Verrill (1873) were both species of Scoloplos. However, Kinberg's species is not a Scoloplos (Hartman, 1948, p.106) and Verrill's specific name is thus valid. H aploscoloplos robustus is common in more northern parts of eastern United States, from New England south to North Carolina in addition to southwestern Florida; it is not yet reported from the western half of the Gulf of Mexico. Haploscoloplos foliosus, new species Specimens come from Turtle Bayou, Aransas Bay, Texas from decaying vegation or sand beaches, October, 1950 (coli. A. E. Hartman) and from Biloxi, Mississippi, sandy mud flat, shore, December, 1943 (coli. M. W. Williams). Preserved individuals lack color; in life it may be orange red with red branchiae, as in other species of the genus. The body consists of about 200 segments and is .4() to 50 mm. long; the posterior end is closely coiled. The body is thickest in the thorax and slightly depressed; farther back it is cylindrical and tapering to the end. . It terminates in a thick, collar-like pygidium with a pair of long, thread-like filaments inserted at the dorsolateral edge. The prostomium is equitriangular, acutely pointed in front; it lacks eyes or other marks.. The thorax has 17 or 18 segments; these have biramous parapodia from which the setae emerge close to the body wall; there are no lamellar or prolonged structures except for a small, inconspicuous papillar lobe emergent behind the middle neuropodia! fascicle. Thoracic setae are entirely pointed and arranged in fan-shaped, spreading fascicles. ·The proboscis, not everted, is assumed to be voluminous and branched, as typic~l for other species of H aploscoloplos. The transition from thorax to abdomen is at segment 17 to 19, where parapodia come to have . characteristic lobes. In some individuals the body is gradually, in others abruptly, narrower at this place. Branchiae are small at first, present from segment 18, but by the next one they are as large as the notosetallobe and continue in this size proportion through a long region; in segments far back the postsetal lobe comes to be greatly prolonged, so that the branchial tip is considerably surpassed by it. The lateral margins of branchiae have dense rows of long fimbriae; these fringes are continued proximally alon'g the inner margins of postsetal lobes and are most conspicuous in posterior segments. An intercirrus is altogether lacking; the only structure which might be regarded homologous is a slight fleshy boss, visible in early abdominal segments; this is not ciliated nor does it elongate in any part of the body. A ventral cirrus is likewise absent. Instead, the superior edge of the neuropodia! flange, through abdominal seg­ments, is characterized for having a slight prolongation; this comes to be set off at an increasingly greater distance from an inferior auricular lobe by segment 40, and in far posterior segments the 2 parts are abruptly set off by a constriction. Furcate setae have not been identified. Haploscoloplos foliosus differs from other species of the genus for lackin-g intercirri and ventral cirri, · and in having greatly prolonged neurosetal lobes in posterior segments which extend distally beyond the branchial processes. The holotype specimen is from Turtle Bayou . in Aransas Bay, Texas, from muddy sand; others are from Mississippi. Phylo ornatus (Verrill), 1873 Aricia ornata Verrill, 1873, pp. 596-597. Orbinia orn~ta Hartman, 1945, pp. 28, 30. Two individuals come from near Englewood, Florida in sandy shoals. The thorax is divisible into an anterior region of about 14, and a posterior one of about 16 seg­ments; these 2 are abruptly different for having modified hooks in the second region which are lacking from the first one. The transition to abdomen is abrupt at about segment 31: Anterior thoracic neuropodia have 3 to 5 vertical rows of yellow acicular hooks and a single posterior row of pointed, slightly curved setae that flank the hinder row of books. Posterior thoracic neuropodia differ from these in having shorter rows of neurosetae and the uncini of the upper half of the ramus are replaced by a single anterior row of 6 to. 8 heavy, modified spines, characteristic of the genus Phylo Kinberg, 1866. These modified spines emerge from the uppermost part of the neuropodium, near a large parapodial gland provided with .an external pore. As the spines are worn, they are replaced by new_ones and the older pushed successively down to form a vertical series of several in a row with the oldest below. The function of spines with glands remains enigmatical. Branchiae are first present from about the fifth setigerous segment and continued back to the end of the body; they are simple and fimbriated marginally. A conspicuous ventral fringe on segments one to 18 consists of 4 to 14 digitate lobes in a single transverse row present on successive segments; this fringe diminishes in posterior thoracic segments and is absent in the abdomen. A subpodial fringe, proximal to the neuropodium, is present from about segment 15 and continued to the second abdominal segment. Setae include long, pointed, transversely spinous ones in notopodia and neuropodia of thorax and .R-bdomen; uncini are present in most thoracic neuropodia and accomr panied by modified spines in posterior thoracic segments; furcate setae occur in abdominal notopodia. Phylo ornatus is more widely known from New England south to North Carolina in littoral zones; in the Gulf of Mexico it has not been taken beyond Florida. Family SPIONIDAE The spionids are usually small, less than 50 mm. long, pale to reddish or melanis­tically spotted in life. The body is long, vermiform, has a pair of long, extensile palpi projecting forward. The prostomium is clearly visible unless covered by the palpi; it consists of a long dorsal lobe without eyes or with them more or less visible between the palpal bases, or limited to juvenile stages. Parapodial branchiae are usually present as simple filaments from the upper surface of notopodia or they are somewhat fused with the notosetal lobe; rarely they are lacking or accessory (pl. 23, fig. 3). Setae consist of simple pointed ones in both rami,. or als~ simple hooded hooks in neuropodia or also notopodia. About 30 species are recorded from eastern shores of America. At least 18 of these may be expected to occur in the Gulf of Mexico. Except for one which is injurious to oysters, the family has remained largely unknown from the Gulf. Twelve species are recorded below; one represents a new genus and species, another is a new name and most of the others are new records. Key to Species 1. With a modified fifth setigerous segment --------------------------------------Polydora ________ 6 1. Without a modified fifth segment ______________________________________________________________________:_____ 2 2. Median. and posterior segments with accessory branchiae (pl. 23, fi~. ~) ------:·-­ _______ c________________________________________________________________________________________________ n~sp~o unc~nata 2. Without accessory branchiae ------------------------------------------------------------------------------c-----3 3. Branchiae present on most body segments _________________________________ :__________ Nerine agilis 3. Branchiae present on few segments or absent. ------------------------------_________________________ .4 4. Prostomial with frontal horns or processes at anterior margin; without parapodial branchiae --------------------------------------------------------------_________________ Spiophanes bombyx 4. Prostomium without frontal horns --------------------------------------------Prionospio ________ 5 5. Branchiae 12 pairs, each straplike or smooth, without lateral fila~ents --:------.----:-­--------------------------------------------------------------------------------------------------Pnonospw c~rnfera. 5. Branchiae 4 pairs, all pinnately divided --------------------------------Prionospio treadweUi 5. Branchiae 5 pairs, second and third smooth, first, fourth and fifth pinnatety divided ------------------------------------------------·-_________ Prionospio heterobranchia texaRIJ 6. Posterior segments have modified notopodial hooks ---------------------------~------------------7 6. Posterior segments lack special hooks __ ___ ______________________ _ ______ _____ ___________________9 7. Posterior -notopodia have single, sickle-shaped hooks directed dorsally _____________ _ -----------------------------------------------'-------------------------------------------------------Polydora hamata 7. Posterior notopodia have bundles of smooth, acicular hooks --------------------------------8 8. Modified spines of posterior notopodia in close bundles; branchiae number only 5 to 7 pairs ----------------------------------------------------------------------------------Polydora armata 8. Modified spines of posterior notopodia alternate with pointed setae; branchial pairs more numerous ----------------------------------------------------------------Polydora caulleryi 9. Prostomium with a median antenna at anterior end of caruncle ____ Polydora ligni 9. Prostomium without median antenna --------------------------------------------------------------------10 10. Prostomium with many eyespots between palpal bases; hooded hooks with obtuse distal end ---------------------------------------------------------------------------------__ Polydora socialis 10. Prostomium with few (3 or 4) to no eyespots; hooded hooks with acute distal end ----------------------------------------------------------------------------------------------Polydora websteri Nerine agilis Verrill, 1873 Nerine minuta Treadwell, 1939, p. 5, figs. 18-20. Nerine agilis Hartman, 1944, p. 340, pl. 50, fig. 11; Hartman, 1945, p. 31; Fauvel, 1950, pp. 371-372. Collections come from Lemon Bay, Sarasota Co., Florida, Jan. 1938, from Grand Isle, Louisiana (coli. E. H. Behre), from near Rockport, Texas (coli. J. W. Hedgpeth) and frQm Alliga,tor Point, Franklin Co., Florida (coil. L. M. Henry). The prostomium is acutely pointed in front and has 4 eyespots in a nearly transverse line between the palpal bases. ~ranchiae are straplike and present from the second segment, continued back through a long region. Anterior parapodia have only pointed setae; farther back both notopodia and neuropodia have hooded hooks. This forms beds in fine sandy beaches that shift little; the vertical burrows extend down considerably more than .the length of the individual; only the distal aperture is Visible at tidal recession; _in some places the apertures of burrrows are very abundant and closely spaced. Nerine minuta Treadwell (1939) from Port Aransas, Texas is here considered _the same; it was originally .thought distinct because of its small size, 10 mm. instead of 30-40 mm. long. This feature is characteristic of many species from the Gulf of Mexico when compared with those from colder or more northern seas. The species is widely known from New Jersey to North Carolina and French West Africa (Fauvel, 1950) . . Genus Polydora Bose, 1802 This genus may be well represented throughout the Gulf of Mexico. Because of the small size, the species have remained largely unknown or unreported. The history of the generic name and the distribution of fossil records in Texas are given elsewhere (see Stenzel and Turner, 1944, pp. 289-308). Six species are reported below. Polydora websteri Hartman, 1943 Polydora ciliata Kavanagh, 1940, pp. 31-34; Kavanagh, 1941, p. 354. Polydora websteri Hartman, in Loosanoff and Engle, 1943, pp. 70-72, fig. 1; Hartman, 1945, p. 33; Hopkins, 1947, pp. 12-14; Behre, 1950, p. 13. This occurs throughout the Gulf of Mexico, associated with mollusk shells, serpulid tubes, barnacles and calcareous layers of rocks. It forms blisters in nacreous layers of oysters. It is intertidal and aestuarine withstanding brackish conditions. Larvae occur abundantly in the tow in November (based on collections of Dr. S. H. Hopkins). The body is usually small, 20 mm. long or less; the prostomium is distinctly bifid in front, has 3 or 4 weakly developed eyespots or these are nearly lacking. The first segment has neurosetae but the notopodium is represented by a papillar lobe only.· The fifth modified segment has 5 or 6 pairs of falcate hooks accompanied by pointed, distally flattened setae. Branchiae are present from the seventh segment to near the end of the body or the last 10 to 16 segments lack them. The pygidium is surrounded by a circular collar that is open dorsally. Hooded hooks are present from the eigh.tli neuropodium; they terminate in a bifid tip in which the main fang is about at right angles to the shaft. In life the body is pale or slightly flesh color; pal pi have dusky or purple markings. The gallery is lined by a membranous tissue that may project from the burrow for a short distance. The tube is roughly U-shaped and large enough to permit the animal to tum about. Mud, silt and other inert materials are drawn into it to cause dark stains or mud blisters in calcareous layers. Polydora websteri occurs more widely along eastern shores of the United States to New England. Polydora ligni Webster, 1879 Polydora ligni Mortensen and Galtsofi, 1944, pp. 164-165; Hartman, 1945, p. 32. Many finds are from Point Pefiescal, Laguna Madre, Texas (coli. W. Armstrong Price). They come from , gray mud in slimy tubes embedded in a clump of dead serpulid tubes. Length of body is only 10 to 15 mm.; color in life is pale flesh to white. The prostomium is deeply cleft in front; there are 4 dark eyespots in rectangular arrange­ment. The prostomial caruncle has a median antenna inserted between the palpal bases. The pygidial collar is large and chalky white in life. Branchiae are first present from the seventh setiger and continued through few segments; they are absent from a long posterior !"egion. The processes of tube-building are described by Mortensen and Galtsofi ( 1944) . This occurs more widely along both shores of North America, in littoral zones and may be considered euryhaline. Polydora hamata Webster, 1879 Polydora hamata Webster, 1879, pp. 251-252, figs. 111-118. Six individuals come from Bayou Rigaud, Grand Isle, Louisiana (coli. J. H. Roberts). The prostomium has 4 dark eyes in rectangular arrangement between the palpal bases. Large, yellow dorsal hooks are present in at least 15 terminal segments; they occur singly and stand erect over the dorsum, presenting a conspicuous armature: over the back. The modified hooks of the fifth segment are falcate. The hooded hooks of neuropodia are distinctly obtuse in their distal end. This species is close to P. hoplura Clapar~de, from Europe; in the latter, however, the modified hooks of the fifth segment have an accessory process in the subdistal concavity and the hooded hooks of neuropodia are distally acute. ; Color in life is white in front, brown to flesh in the middle and white or yellowish in back; branchiae are red (froni Webster, 1879). The species is associated with bivalve shells. It is known elsewhere from Virginia. Polydora socialis (Schmarda), 1861 Polydora socialis Hartman, 1941, pp. 310-311, pl. 48, figs. 41--42; Hartman, 1945, pp. 33-34. This comes from Alligator Harbor, Franklin Co., Florida in 6 ft., Aug. 1950 (coli. H. J. Humm). It occupies a long, ~ilt-covered tube and occurs in mud flats. Overall body length is about 10 mm. or less. The prostomium has many small eyespots between the palpal bases; it is cleft in front. . Branchiae are present from the eighth segment; they are small through several segments, larger in the middle and absent from many posterior segments. The pygidium has a broad, flaring disk with dorsal incision. The modified spines of the fifth segment are heavy, acicular, nearly straight and distally blunt. The hooded hooks of neuropodia are distally bifid, with teeth ,forming an obtuse angle with the shaft. Polydora socialis is known from both sides of North America, including Chile north to California, and North Carolina. Polydora armata Langerhans, 1880 Polydora armata Fauvel, 1927, pp. 55-56, fig. 19. Several individuals were taken from a large shell of the Carrier-Shell, Tugurium (Trochotugurium) longleyi Bartsch off southwestern Florida, in 117 fms., by the M/V OREGON, station 243 (see Eunice floridana, above). They occupied drilled burrows, presumably made by the spionid. -The fragments are incomplete, minute, pale or white; some show at the posterior end, the spread stellate arrangement of notosetae, characteristic of the species. This is known from southern California, western Mexico, and various parts of western and southern Europe, in addition to other widely dispersed geographic · areas. Polydora caulleryi Mesnil, 1897 Polydora caulleryi Fauvel, 1927, pp. 54--55, fig. 19. This comes froin mud and silt tubes in crevices of the Carrier-Shell, Tugurium (Trochotugurium) longleyi Bartsch off southwestern Florida, in 117 fms., taken by the M/V OREGON, station 243 (see Eunice floridana, above). Preserved individuals are pale orange and larger than P. armata (above, which comes from the same mollusk). The -modified spines of the fifth segment are coarse, strongly falcate and have the distal fang overhung by a fimbriated cape that appears bristled when seen from the side. The hooded hooks of neuropodia are distally bifid with the teeth forming an obtuse angle with the shaft. This species clearly shows affinities with P. armata Langerhans in the character of the modified spines but they are never bristled in P. armata, in so far as can be observed. The ecologic niche of the two appears to be different. This is the first record for the species from the Western Hemisphere; it is more widely known from Europe. Prionospio ?cirrifera Win~n, 1883 Prionospio cirrifera Fauvel, 1927, p. 62, fig. 21. Specimens come from Lemon Bay-and vicinity, Sarasota Co., Florida, in old snail shell and from fine sandy beaches, Jan. 1938. They measure 20 to 30 mm. long. Branchiae in all cases number 12 pairs, hence the questionable specific identity. They are strap-like, more or less uniformly large and abruptly absent thereafter. The prostomium is rounded in front and has conspicu­ous lateral flanges at its sides. A median caruncle is high and extends back to the front margin of the second segment. The 4 eyes are visible, with the 2 posterior ones much the largest. Hooded hooks are first present from segment 18; they appear gradually in neuro­podia and come to occupy most of the neuropodia! ridge farther back. In notopodia the hooded hooks are first present at segment 40 and continued farther hack. The anus has a large, midventral cirrus and a similar, though much shorter, dorsal process. Prionospio cirri/era Wiren was originally described from the Bering Sea. Various redescriptions of the species, based largely on European finds, record 6 pairs of branchiae (Soderstrom, 1920, p. 237) and 6 to 11 pairs (Fauvel, 1927, p. 62). In the case of the specimens from the eastern Gulf of Mexico, there are always 12 pairs of hranchiae; the latter may thus represent another subspecies. Prionospio treadwelli, new name Prionospio plumosa Treadwell, 1931, p. 3. Not Sars, 1873. Prionospio tenuis Hartman, 1945, p. 32; Behre, 1950, p. 13; Not Verrill, 1880, pp. 176--177. Collections come from Barataria Beach, Grand Isle and vicinity, Louisiana, July, 1942 (coli. E. H. Behre and J. H. Roberts). Others for comparison are from North Carolina. The prostomium has 4 small, inconspicuous eyes in quadrate arrangement. Length attains 15 to 30--40 mm. All hranchiae, numbering 4 pairs, are pinnately divided. There is no transverse dorsal ridge, such as occurs in some species of the genus. The first 8 segments have thick full fascicles of curved, distally pointed setae. The notosetal fascicle resembles the neurosetal one. Abruptly from the ninth segment the neuropodium has different setae. There is a vertical series of about 12 hooded hooks that alternate with an equal number of slender, companion setae, and one inferiormost thick~r, distally pointed seta that projects obliquely downward. This seta resembles the notosetae in thickness hut is distally straight, not curved. Farther hack this arnmgement of hooks and seta is continued but the hooks come to be more conspicuous. Notopodia continue to have pointed setae but ~n segments far back the long pointed ones diminish in number and are accompanied by 2 or 3 hooded hooks that resemble those in neuropodia. Uncini are hooded; they have a main fang that is at right angles to the shaft and are surmounted by a uniserial row of several smaller teeth. This species is readily identified specifically for its abrupt parapodial change be­ tween segments 8 and 9. The nearly related Prionospio tenuis (Verrill) from New England, also has 4 pairs of pinnately branched branchiae. In it, however, the neuropodia} hooks are not present before segment 15. The specific name, Prionospio plumosa Treadwell, is here changed to P. treadweUi~ new nam~, since it is preoccupied by P. plumosa (Sars), 1873. Prionospio treadwelli was first described from Chesapeake Bay, Maryland in 7-48 meters, and later reported from North Carolina as P. tenuis by Hartman (1945). Its known range includes Louisiana. Prionospio heterobranchia texana, new subspecies Prionospio heterobranchia Moore, 1907, pp. 195-197, pl. 15, figs. 1-6 (stein species). This was taken from Turtle Bayou, Aransas Bay, Texas from fine sand mixed with decaying vegetation, October, 1950 (coli. A. E. Hartman) . . Length is about 12 mm. The body is pale or white (preserved) except for the black . eyes on the prostomium. There are 58 segments. The prostomium is prolonged forward like a snout. There is one pair of very large dark eyes near the middle and a pair of much smaller eyespots, slightly in front and to the side of the larger ones. Branchiae number 5 pairs; the second and third pairs are smooth and cirriform; the first, fourth and fifth pairs are plumose with the filaments numerous and laterally ~nserted. Anterior neuropodia have pointed setae above and below. At the twelfth neuro­ podium there is one hooded hook accompanying pointed setae; at the fourteenth there are 2 hooks and farther back there is a gradual increase but the pointed setae continue to be present throughout the length (or 58 segments) . In no to podia the hooded hooks are first present in segment 35 and their number increases gradually going back. The hooks resemble those in neuropodia. The subspecies, texana, differs from the stem species, P. heterobranchia Moore, in that the latter has hooks first present in neuropodium 15, whereas in the subspecies they occur from segment 40. T~e prostomium has a single pair of large eyes and the anterior prolongation is not so pronounced in·the stem species. Prionospio heterobranchia texana has been taken only at Turtle Bayou, Aransas Bay, Texas. The stem species is known only from soft ooze at the deepest part of Eel Pond, Woods Hole, Massachusetts, and has not been recorded since first described (1907). Spiophanes bombyx (Claparede), 1870 Spiophanes bombyx Fauvel, 1927, p. 41, fig. 14; Hartman, 1945, p. 31. This comes from the Gulf of Mexico side of Lemon Bay, Florida and others for compariso!l come from North Carolina. The prostomi urn is a broad, triangular lobe that is widest in front and has attenuate lateral processes or horns. The 4 eyespots are in trapezoidal arrangement, with the anterior pair wider apart. The first ciliary transverse ridge is on the third segment. The first neuropodium has a pair of h~avy spines, distally curved and directed downward. Uncini are first present from segment 16; they are in vertical series of about 8 in a row, accompanied by a long, pointed seta in the lowest position. All setae before the 16th segment are distally pointed. Hooded hooks are distally bifid with the main fang at about right angles to the shaft, not obtuse, as in the European species (see Fauvel, 1927, fig. l4e). The accessory or more distal tooth is much the smaller. Spiophanes verrilli Webster and Benedict (1884, pp. 728-729) from Massachusetts, low water, is a different species. Earlier I (1945, p. 31) had regarded them as possibly the same. They are different in that the hooded hooks are already present from seg· ment 6 in this, and not before segment 16 inS. bombyx; dorsal ciliary ridges are first present on segment 6 in the first and from segmenl 3 in the second. Spiophanes bombyx (Claparede) is more widely recorded from Europe (see Fauvel, 1927, p. 41). Genus Dispio, new Type D. uncinata, new species The prostomium is rounded in front; its caruncle extends hack only to the first segment. There are 2 pairs of small, inconspicuous eyes at the sides, largely concealed by the palpal bases. The palpi are paired, large, thick, longitudinally grooved. The lateral margins of the peristomium at the sides of the prostomial ridge, are enlarged to form flanges that extend back to the first setiger. Branchiae are present from the first setiger and continued back to the posterior end; they are more or less fused with the postsetal, notopodial lobe. All parapodia, including the first are biramous. Notosetae are entirely simple, distally pointed. Neurosetae are similarly pointed in anterior segments; farther back . some of the pointed setae are replaced by hooded hooks that end distally in an entire point and are hooded. The pygidium ends in a simple, collar-like ring with a slight middorsum notch. From about the twentieth segment there are paired accessory branchial tufts that arise from behind the notopodial base; these tufts have 2 to 5 digitately arranged lobes, each with a vascular loop. The notopodial lobes of the first few segments are enlarged and distally serrated. ,; Dispio grossly resembles Spio Fabricius in having branchiae on all segments and hooded hooks only in neuropodia. The prostomium is similarly prolonged in front and lacks frontal horns, and parapodial branchiae are somewhat fused to postsetal lobes. Dispio differs for having accessory branchial lobes and the first few notopodial lobes are serrated. Dispio resembles Nerine Johnston in having hooded hooks that are distally entire, and branchiae fused with no to podia, but in Nerine the notopodia in posterior segments have hooks and branchiae are absent f_~om the first segment. Polybranchia Potts (1928, pp. 693-694) has accessory palmate branchial lobes that resemble those of Dispio, but this genus, known for the single species, P. foxi Potts, differs so sharply from other spionids that it may belong elsewhere; notable are the lack of pointed setae and parapodiallamellae. Dispio uncinata, new species Plate 22, figs. 1-5; plate 23, figs. 1-4 Spio setosa Behre, 195~, p. 13. There ·are 7 individuals; they come from Alligator Point, Franklin Co., Florida, March, l950 (coli. L. M. Henry) and Grand Isle, Louisiana, sandy beach, July, 1942 (coli. E.,H. Behre and J. H. Roberts). Length of a complete specimen (preserved) is 32 mm. for 120 segments; width is 2 to 2.5-mm. at the middle or widest part. A larger specimen that is posteriorly in­complete measures 30 mm. long for 70 segments. There is no color remaining on fixed materials except for that on the palpi. The prostomium is prolonged forward and its anterior, subtriangular part is slightly set off by ·a pair·of lateral emarginations from the larger, lateral-.Iobes that comprise the peristomium. The prostomium extends back between the palpal bases and is only slightly prolonged as a caruncle to the end of the first segment to form a nuchal ridge or enlargement. The raised lateral flanges of the peristomium extend lengthwise along the prostomial ridge (pl. 22, figs. 1, 2). Palpi are thick, longitudinally grooved, marked by widely spaced black transverse bars on the side opposite the longitudinal groove. Four prostomial eyes are present but largely concealed in the grooves between the prostomial caruncle and pal pal bases; they are in trapezoidal arrangement with the anterior pair the wider apart. The first 3 pairs of parapodia differ from those farther back in that their superior notopodial parts are deeply serrated; this serration is most conspicuous in the first segment. The first one (pl. 23, fig. 2) has a large notopodium with a long branchia that is continuous. with a series of 5 or 6 serrations which diminish in size gradually and merge with the frilled postsetal membr.ane; the presetal lobe is much smaller and attached near the base of the large postsetal process. Between these 2 lobes emerge 30 to 50 long, slender setae that are directed upward. The setae are longest at the upper end of the fascicle and diminish gradually going down. The notopodial proc­esses of the first pair thus form a palisaded series behind the palpal bases. The neuropodia of the . first segment are much smaller than notopodia; each consists of a weakly lobed foliaceous postsetal lobe; it carries about 20 slender setae, including 9 shorter ones directed laterally and 6 or more longer ()nes in an inferior tuft. Farther back the second and third notosetal lobes are similarly serrated but de­creasing in the number of lobes so that by the, fourth and fifth segments the divisions have disappeared. · The second notopodium has a superior" tuft of about 6 slender setae similar to that of the first but shorter, and 6 pairs of shorter setae below; each of the pairs includes a broadly limb ate one and. a slenderer companion seta inserted slightly behind the broader one, together forming a double vertical series. In addition, there are 2 or 3 ventralmost, long pointed setae extending out far beyond.the others. The arrangement of setae from the third segment is similar to that of the second except that the larger pointed setae come to be more broadly limbate . . In all 15egments the longest setae are the 2 or 3 uppermost in notopodia, and the 2 or 3 lowermost in neuropodia. In the twentieth segment there are 3 long pointed setae and about 10 pairs of shorter ones in notopodia. Neuropodia have about 11 pairs of shorter ones and 3 slender lowermost ones. 2 1 4. PLATE 22 FIGs. 1 to 5, Dispio uncinata, n. gen., n. sp. FIG. 1, anterior end in· dorsal view, showing prostomium and first five segments, X 22. FIG. 2, anterior end in left lateral view, showing prostomium with first five seg­ments, x 22. FIG. 3, series of hooded hooks with companion setae and two inferiormost setae, X 239. FIG. 4, an accessory branchial tuft showing arrangement of vascular lool>s, x 234. FIG. 5, a hooded hook and companion setae from a posterior neuropodium, x 526.5. PLATE 23 FIGS. I to 4, Dispio uncinata, n. gen., n. sp. FIG. 1, a pair of notopodial setae from a median parapodium, x 368.5. FIG. 2, first parapodium from left side, in posterior view, x 75. FIG. 3, a posterior parapodium in posterior view, x 35. FIG. 4, twenty-fourth parapodium in posterior view, :x: 35. From segment 25 to 27 the neuropodia} paired setae. are replaced by hooded hooks (pl. 23, fig. 3) but the tuft of 2 or 3 slender setae remains as before. The hooks occur in single series of 5 to 8 in vertical rows, and are continued to the posterior end; they are accompanied by slender pointed setae (pl. 22, fig. 3) like those in front (pl. 23, fig. 1) , alternating with the hooks. The hooded hooks are distally entire, slightly curved and the hood extends distally to or slightly beyond the point (pl. 22, fig. 5). Parapodial lamellae are well developed. The presetal ones are short, auriculate; the postsetal lobes are longer but do not extend out as far as setae or hooks. They are more or less completely fused with the branchial processes along their upper sides but ­ the distal ends are free (pl. 23, figs. 3, 4). Postsetallamellae of the first few segments are frilled and larger than those in back. Branchiae are present from the first segment to the end of the body. They are fused with .the notopodial postsetal lobe (pl. 23, fig. 3) ; their inner proximal edges have long fringe that is continued along the base of the parapodium and onto the dersum of the body. In the middorsum of each segment there is a band of similar long fringe, in line with that of the branchial row. The ·anal end tapers gradually to a short, narrow collar with a slight middorsal notch and a posterior aperture; there is no flange or cirrus. Accessory branchiae are first pre'sent from segment 24--28, as simple long lobes that arise from the dorsolateral side of the body behind the notopodial base. The number of lobes increases gradually to 6 or 8, arranged i~ palmate group (pl. 22, fig. 4). Each loop contains a vascular flexure that is continuous with successive ones so that the entire branchia may be considered a prolonged vessel that is thrust into a series of curvatures. This accessory vessel connects, not with the branchial loop in parapodia, but with a lateral vessel in the body. The accessory branchiae appear to be non­ retractile (pl. 23, figs. 3, 4) . Large, elongate oval discs, representing nearly ripe ova, fill the body cavity; each is covered with tt thick sculptured shell. Dispio uncinata is known only from Alligator Point, Florida, sand and from ·Grand Isle, Louisiana (reported as Spio setosa by Behre, 1950). Family MAGELONIDAE These spioniform annelids differ from the SPIONII)AE chiefly for having greatly prolonged paired palpi that are conspicuously papillated. The thorax is set off from the abdomen by a ninth segment in which the setae differ from those in front and back. Both notopodia and neuropodia in the abdomen have hooded hooks. A single species is reported. Magelona, near californica Hartman, 1944 Magelona californica Hartman, 1944, pp. 320-321, pl. 28, figs. 10-14. Six specimens come from Alligator Harbor, Franklin Co., Florida, June, 1950 (coil. L. M. Henry) . The prostomium lacks frontal horns. The ninth or modified segment has full spreading setal fascicles with setae distally tapering and pointed. Abdominal hooks ­are distally bifid with the main tooth at a sharp angle to the shaft and the secondary tooth nearly as large as the m~in one. The formalin-preserved specimens retain no color. except for paired speckled areas on the dor~al side of the thorax; there is no lateral pigment in the abdomen such as occurs in M. californica, from California. Magelona rosea Moore (1907, pp. 201-204, pl. 16, figs. 24--30) from New England is a different species; in it the abdominal hooks are much less sharply beaked distally and the secondary tooth is smaller; the paired pigment patches of the thorax are lacking; the prostomial lateral flanges are obliquely ridged and ventral cirri are present on the tenth parapodium. Magelona californica is elsewhere known from central and southern California. Family CHAETOPTERIDAE Two species are reported. Key to Species 1. Larger, in U-shaped, opaque tubes------------------------------------Chaetopterus variopedatus 1. Smaller, in long, slender, annulated, translucent tubes ____ Spiochaetopterus oculatus Spiochaetopterus oculatus Webster, 1879 Spiochaetopterus oculatus Webster, 1879; pp. 247-249, pl. 8, figs. 98-102; Hartman, 1945, p. 35. Numerous finds come from northern shores of the Gulf of Mexico, from near Biloxi, Mississippi in sand flats with oysters, Dec. 1943 (coli. M. W. Williams), from Barataria Bay and Caminada, near Grand Isle, Louisiana, 1942 and 1948 (coli. J. H. Roberts). The tube is characteristic, cylindrical, transclucent, annulate, resembling a delicate straw; it may be several inches long and only a mm. across; the· distance between successive annulations ranges from 0.5 to 2 mm. On the worm there is a dark spot on the ventrum between segments 6 and 7, followed by a chalky white ventrum on segments 7 to 9. The species is more extensively recorded from New England south to North Carolina. Chaetopterus variopedatus .(Renier), 1804 Chaetopterus pergamentaceus Leidy, 1888, pp. 73-74. Chaetopterus variopedatus Fauvel, 1927, pp. 77-79, fig. 26; Hartman, 1945, pp. 34--35. · This occurs in low channels in mud flats near Englewood, Florida; the tubes· are large, U-shaped and open to the surface at both ends. It is parchment like and opaque. Commensal crabs of the genera Polyonyx and Pinnixa (Hartman, 1945, p. 35) are apt to be found in the tube. The species has a very wide, cosmopolitan distribution and is known from New England southward; it has been noted at the mouth of the Manatee River, Florida (Leidy; 1888) . Family CIRRATULIDAE Four species in three genera are recorded; two are newly described. Key to Species 1. Penetrating shells; dark green to black in life; lateral tentacles few, limited to anterior end ------------------------------------------------------------------------------------------------------------3 l. Not penetrating shells; lateral tentacles few to many --------· ·--· --------------------------------2 2. Posterior segments provided with only pointed setae (pl. 25, fig. 3) ------------------------------------------------·------------------------------------------------Cirratulus hedgpethi 2. Posterior segments provided with blunt, acicular hooks ___ ____ _ ___ Cirriformia filigera 3. Paired tentacles number less than 5 pairs _________ __ _________ ______ __ ________ Dodecaceria diceria 3. Paired tentacles number more than 10 pairs ______________ __ Dodecaceria, near concharum Dodecaceria diceria, new species Plate 24, figs. 4, 5 Many specimens were removed from drilled runways in the nacreous layers of a large living shell of the Carrier-Shell, Tugurium (Trochotugurium) longleyi Bartsch off southwestern Florida in 117 fms., taken by the M/V OREGON, station 243 (see Eunice floridana, above). Preserved and possibly also in life the color is very dark green to black. The body measures 12 to 15 mm. long, about 0.3 to 0.5 mm. wide at the prostomium and 0.6 to 0.8 mm. wide in the middle third or widest part. The segments number more than 100 and are closely crowded at the posterior fourth to end of the body. Under magni­fication the body rings are seen to be crossed by 6 to 10 closely crowded rings and splashes of dark color. These marks obscure the segmental furrows. The prostomium is a short, triangular lobe (fig. 4) a little wider than long; it has a pair of nuchal slits at the postectal margins; there are no eyes. The prostomium is indistinctly separable from a longer, smoot~ preantennal region; on the ventral side its anterior end forms the lower lip. The first body ring is only about half as long as the peristomial ring. From its anterior margin, at the dorsoectal ends there is a pair of thicker, longer, cylindrical filaments and a similar, slenderer, shorter pair inserted immediately within (fig. 4). A wide space separates these filaments so that the middorsum is exposed. In most specimens only these 2 pairs of tentacular processes are present, hence the specific name. An occasional individual has in addition a very much shorter, smaller pair of filaments inserted behind the first pair but on the following segment. Setae and hooks are nowhere conspicuous and are usually not readily seen because of the deeply pigmented epithelium. The first setiger has about 4 long setae in nolo­podia and 3 in neuropodia. The second has 2 larger and 2 slender setae in notopodia and 3 larger with a slenderer in neuropodia. The third segment has one larger and 2 finer setae in notopodia and one larger and 3 slenderer ones in neuropodia. After that there is less difference in size between the finer and coarser setae. The first 10 seg­ments have only pointed setae. Simple falcate hooks (fig. 5) are present from about the eleventh segment and alternate with longer setae in both rami. In one individual that was minutely,examined, segment 11 has 5 pointed setae above, and 4 setae with 2 hooks below; segment 12 has 4 setae and 2 hooks above and 5 hooks below; segment PLATE 24 FIGS. 1 to 3, Notomastus hemipodus Hartman. FIG. 1, anterior end including thorax and first 2 abdominal segments, seen from right side, x 28. FIG. 2, prostomium and peristomium, seen from the top, x 66. FIG. 3, median abdominal segments seen from above, x 56. FIGs. 4, 5, Dodecaceria diceria, n. sp. FIG. 4, anterior end in dorsal view, x 64. FIG. 5, a parapodial hook seen from the side, x 1485. 13 has both setae and hooks above and 6 hooks below. Farther back the J?resence of the hooks and setae in both rami can usually be distinguished. and there is a gradual _ diminution_of parts going posteriorly. The hooks (fig. 5) are distally falcate and have a series of 7 or more transverse ridges at the geniculate region. The posterior end of the body terminates in a slender tail with a distal aperture. Some individuals show indications of anterior regeneration . of the first 10 or more segments. The genus Dodecaceria Oersted is known for few species. Some, such as D. con­cha rum Oersted, penetrate calcareous structures, others, such as D. fistulicola Ehlers, construct limey matrices. D. dice ria belongs to the group which destroy shells. It differs from all known species in having its tentacular processes limited to only 2 pairs; its acicular hooks are transversely ridged instead of ladle-shaped. Dodecaceria diceria is known only off the western end of the Florida Keys in 117 fms. Dodecaceria, near concharum Oersted, 1843 Dodecaceria concharum Fauvel, 1927, pp. 102-103, fig. 36. One dark olive green specimen comes from Alligator Harbor, Franklin Co., Florida, March 1951 (coli. D. Carpenter). It has 29 segments, is posteriorly incomplete, and measures 8 mm. long and about 1.3 mm. wide. The paired palpi are followed segment­ally by 8 pairs of long filamentous cirri and these by 4 much shorter pairs that come to be gradually shorter going back. Anterior segments have only long, pointed setae in notopodia and neuropodia; these are almost wholly replaced by pale yellow hooks ­in the ·third post-tentacular segment. It differs from typical D. concharum Oersted in having more tentacular cirri and pointed setae through a longer region. Cirriformia filigera (delle Chiaje), 1841 Audouinia filigera Fauvel, 1927, pp. 92-93, fig. 32. Cirriformia filigera Hartman, 1945, p. 35. About 25 specimens were trawled near the pass in Lemon Bay, Sarasota Co., Florida, Jan. 1938, from gravel aggregates and another was taken from a low tide mud flat, near Englewood, Florida at the same time. The prostomium is a conical lobe lacking eyespots. The numerous dorsal tentacles arise across the middorsum of the third or fourth setigerous segment. Lateral tentacles are continued on body segments; they are first present from the first setiger, proximal to the notopodium. In median and posterior segments their origin is gradually up­ward so that they are well removed from the notopodial papilla. Acicular setae. are yellow, present in both notopodia and neuropodia in middle and posterior segments;' they number 2 to 5 in a series and alternate with slender pointed setae. Cirriforniia filigera occurs more w~dely on bo'th sides of the North Atlantic Ocean . and south to Brazil. ­ Cirratulus hedgpethi, new species Plate 25, figs. 1-3 A large, dark brown (preserved) specimen, complete though broken in the middle comes from a bell buoy in 4--5 fms., off Port Aransas, Texas, October 22, 1946 (coli. J. W. Hedgpeth); another posterior fragment i's from Port Aransas, Texas, November, 1948 (coli. J. W. Hedgpeth). The body is thick, much coiled, over 200 mm. long and 8.5 mm. wide in the front or at greatest width; it consists of more than 300 segments. It is somewhat depressed in front, cylindrical farther back or somewhat depressed (fig. 2) and tapers to a slender posterior end with terminal anus. The prostomium is broadly rounded in front (fig. 1) ; eyes are not visible and presumably absent; nuchal organs are visible as crescentic depressions at its postectal margins. The presetigerous region is long, crossed by 6 to 8 deep wrinkles of contraction; complete segmental rings are lacking. The first setigerous segment is biramous; it has on either side, immediately above the notopodial tuft, a reduced bundle of slender branchial filaments that number only about 4 pairs (fig. 1) ; these filaments are unusually slender and short for a cirratulid; leaving bare a broad, median space. Farther back the lateral tentacles are very few, slender and inserted slightly above the notosetal tuft, at least to the middle of the body. Farther back I was unable to make out any lateral tentacles or scars marking their possible surface of attachment. Setae are entirely long, distally pointed. Those in front are in greatest number and farther back they gradually diminish in number and length. There are 2 kinds (fig. 3), including a thicker, somewhat coarser one and an alternating finer one, arranged in a .double series. Both are pale amber, smooth and finely pointed at the tip. There are no blunt acicular or curved spine-like setae in any part of the body. Cirratulus hedgpethi belongs to the group of species in which parapodia have only slender, pointed setae. Cirratulus elongatus Treadwell (1901, p. 204, fig. 56) from Culebra is said to belong here, but it has a prostomium that is acutely pointed in front and the presetigerous region has 3 complete annulations. C. filiformis Keferstein (Fauvel, 1927, pp. 94--95) also belongs to this group 'but agrees with C. elongatus in the characters named above. C. hedgpethi is further characterized for the greatly diminished dorsal and lateral tentacles and a broadly rounded prostomium that lacks eyes. I take pleasure in naming the species for Dr. Joel W. Hedgpeth to whom I am indebted for the finds. C. hedgpethi is known only from Texas littoral zones. Family ARENICOLIDAE A single member of the lug-worms is present. A renicola cristata Stimpson, 1856 Arenicola cristata Ashworth, 1912, pp. 105--111, pis. 5, 8, 10, 13; Fauvel, 1927, pp. 163-164, fig. 57; Warren, 1942, p. 45; Hartman, 1945, p. 37; Behre, 1950, p. 13. I PLATE 25 FIGS. 1 to 3, Cirratulus hedgpethi, n. sp. FIG. 1, anterior end in dorsal view, x 8.5. . FIG. 2, cross section through a posterior body segment, showing right half, x 16.5. FIG. 3, neurosetal fascicle from a far posterior segment, x 177. This occurs in many ~uddy sand beaches throughout the Gulf of Mexico, at least west to Texas. Its surface mark is a large cinder cone exposed at lowest tide line. Tht> following beaches in the Gulf of Mexico have yielded specimens: Godfrey and ~derson's Creeks near Englewood, Florida, Caminada Beach, Barataria Bay, Grand Isle, Louisiana, near Biloxi, Mississippi from sand flats with oysters, and Wakulla Beach, Wakulla Co., Florida. The body is thick, large 50 to 100 mm. long or more. There are 17 setigerous segments, 11 pairs of tufted branchial pairs followed by a caudal region which lacks parapodia. The surface epithelium is strongly areolated and its anterior parts are transversely ringed farther back. Arenicola cristata is otherwise known from limited cosmopolitan areas (see Ashworth, 1912, pp. 105-111, for detailed account). Family OPHELIIDAE Only 2 species are recorded though others are doubtless to be found. Key to Species l. Body smooth, small, seemingly lacking segmental furrows, resembling a minute Amphioxus with dark segmental patches _____ ____ ______,_____ _____ _ _Polyophthalrtzus pictus 1. Body with segmental cirriform branchiae and lateral segmental eyespots ___ ________ _ --------------------------------------------------------------------------------------------------------Armandia agilis Armandia agilis (Andrews), 1891 Ophelina agilis Andrews, 1891, p. 289, pl. 15, figs. 21-26, 28. Armandia agilis Hartman, 1945, p. 37. Individuals come from Alligator Harbor, Franklin Co., Florida, August, 1950 (coli. D. Carpenter), from Lemon Bay, Florida in sands with Chaetopterus and trawled from gravel aggregates. Length attains 15 to 24 mm., width is 1 mm. or less. The body consists of 36 to 39 setigerous segments. The prostomium is long, acutely pointed, its length about twice its width; it has a pair of large, circular nuchal organs at the sides, just in front of the first parapodia. The ~outh on the ventral side, is in line with the first setiger. Lateral eyespots number 20 or 21 pairs; they are black, approximately circular in shape, first present between setiger 6/7 or 7/8 and continued back through setigers 26/27; they are located just in front of the parapodial bases. The anal funnel is long, its length about equal to that of the last 6 setigers; it is transversely barred with light and dark areas, approximately equal to one another and numbering about 20. The posterior end has a very long, midventral cirrus that is about as long as the funnel, and up to 8 shorter processes arranged in a circlet. The alimentary tract contains fine quartz particles. Armandia agilis is more extensively known from North Carolina; in the Gulf of Mexico it is known only from Florida. Polyophthalmus pictus (Dujardin), 1839 Polyophthalmus pictus Fauvel, 1927, pp. 137-138, fig. 48; Rioja, 1946, p. 195. This is recorded from Vercruz, Mexico by Rioja (1946, p. 193). It occurs typically among intertidal algae and rocks. It attains a length of 10 to 20 mm. and is slender, very agile, sometimes highly colored in life, with transverse bars of brown to reddish black. There are 27 or 28 segments. This is known from both sides of North America in littoral zones, in addition to cosmopolitan areas. Family FLABELLIGERIDAE These are the bristle-cage worms, characterized for having tufts of setae directed forward to form a cephalic cage. The surface of the body tends to he roughened due to the presence of epithelial papillae and adherent sand particles. The prostomium and its surrounding parts including the peristomium and its tentacular processes are completely retractile into the oral aperture, and thus visible only by dissection unless these parts are everted by fixing. Two species are here recorded. Key to Species l. Peristomium fomiing a tongue-like process (pl. 26, fig. 1) that extends forward over the prostomium; neuropodia! hooks have a bifid tip (pl. 26, fig. 3) _______ _ ------------------------------------------------------------------------------------------------------Semiodera roberti l. Peristomium without such tongue-like process; neuropodia! hooks with entire tip -------------------------------------------------------------------------------------__________ Stylarioides inflata Stylarioides inflata (Treadwell), 1914 Stylarioides inflata Hartman, 1951 (in press, Pacific Science)_ This was taken from trawled shell aggregates in Lemon Bay, Sarasota Co., Florida. It occupies a U -shaped burrow in friable rocks or shell masses. The anterior end is widest and obliquely abrupt (peristomial cirri withdrawn) to form a plaque on the dorsal side of the first 4 segments. The surface of the body is slightly roughened and covered with a very thin layer of minute sand grains. The surface papillation is limited to anterior margins of segments. The prostomium (seen by dissection) is a small lobe with a pair of large, lenticulated eyes. The oral tentacles, inserted on the peristomial base are in crescentic arrangement, surrounding the dorsal margins of the thick and paired palpi and the prostomium. The oral tentacles are numerous, including about 6 larger and 13 smaller on each side. Setae of the first 2 segments are longer than others and directed forward to en­compass the anterior end. N otosetae are fine, distally pointed and transversely barred. Neurosetae of the first 3 segments are distally pointed and resemble the notosetae~ ­thereafter they are thicker, distally blunt, entire and slightly curved along the shaft. Stylarioides inflata is more widely distributed on the Pacific side of North America, from Oregon to Lower California, Mexico (Hartman, in press). Semioderl!' roberti, new species Plate 26, figs. 1-4 Stylarioides eruca Behre, 1950, p. 14. This comes from Sugar House Bend, Grand Terre, Louisiana, June 24, 1942 (coli. 1. H. Roberts) . A long, posteriorly tapering specimen measures 68 mm. long for 115 segments. It grossly resembles Stylarioides eruca (Claparede) because of its similar body pro­portions and bifid neuropodia} hooks (fig. 3) . However, the surface epitheli urn is much less papillated and the oral tentacles are inserted on a tongue-like process (fig. 1) as characteristic of the genus Semiodera Chamberlin. The first few parapodial parts are prolonged forward and have longer setae than those farther back; the setae form a distinct cephalic Cjlge. The oral structures are everted on this individual. The tongue-like process is longer than wide, ha~ a deep median longitudinal furrow and each half is provided, on jts ventral side, with a closely spaced series of tentacular bases (fig. 2) from which all tentacles have fallen away (assuming that there were tentacles). This tongue over­hangs the comparatively minute prostomial lobe and the palpal bases, from which the palpi have also fallen. The prostomium is a small, triangular lobe with a pair of large eyes at the sides. Each eye has a large dark red circular lens and the embedded parts of the 2 eyes are nearly convergent within the cephalic tissue. The entire prostomial portion is surrounded by a collar-like sheath that closely fits about the eversible structures. Parapodial parts are not conspicuous except in the first few segments where the lobes are triangular, sand-encrusted and projecting forward (fig. l). Farther back the notopodia have a linear series of long, pointed setae, a single, slender papilla on the posterior side that . extends distally nearly as far as the setae, and several smaller papillae."-Neuropodia have linear series of hooks and a few slender papillae in a posterior position. Notosetae are long, slender, distally pointed and transversely barred. Neurohooks are shorter and have a bifid tip (fig. 3) with a few transverse bars along the length. The bifid tip (fig. 4) has a falcate tip and an accessory, flattened process or flap that appears tooth-like only when seen on edge. Semiodera ·roberti differs from other species of the genus for the bifid character of the neuropodia! hooks. A nearly related species is shown in color by Ehlers ( 1887, pl. 42, fig. 6) as Siphonostomum cariboum. I take pleasure in naming the new one for the finder, Mr. fH. Roberts. This is known only from the type locality, Grand Terre, Louisiana. Family CAPITELLIDAE The capitellids are small (a few mm. long) to large (80-100 mm. long) usually cylindrical, purplish red annelids in which the body is divisible into thorax and ab­domen. The prostomium is a small, plain lobe; the proboscis is a soft, unarmed eversible pouch. Parapodia lack lamellar processes or have segmental branchiae that are eversible or stationary. The several species are often associated with mud of ooze and in some instances form beds. Seven species are recorded. PLATE 26 FIGS. 1 to 4, Semiodera roberti, n. sp. FIG. 1, anterior end in ventral view with prostomium and associated parts everted, X 21. . FIG. 2, prostomium and overhanging tentacular process in ventral view, x 53. FIG. 3, a nooropodium from a median segment in anterior view, x 221. FIG. 4, distal end of a neuropodia} hook to show broad accessory process, x 445. The Littoral Marine Annelids of the Gulf of Mexico 101 Key to Species I. Thorax consists of 9 segments ----------------------------------------------------------Capitella capitata I. Thorax consists of 10 segments --------------------------------------------·-·-----?Capitellides teres I. Thorax consists of smooth peristomium and 10 setigerous segments ___ ______________ __ _ -------------------------·--------------------·----------------------------------·-------Mediomatus californiensis I. Thorax consists of 11 setigerous segments (pl. 24, fig. 1) _ _____ _ --·------·····----------------2 l. Thorax consists of 13 setigerous segments ---------·------·--·------··--·-Dasybranchus ___ _____ 4 2. All thoracic segments have pointed setae -----------------------·-··---------Notomastus ____ ___ 5 2. First 5 thoracic segments have pointed setae, last 6 segments have hooded hooks --------------------------·----------···-· ------------------------------------·· Hetemmastus filiformis 3. Prostomium has paired patches of eyespots; first setigerous segment has both notopodia and neuropodia ---------·--·---------··---··-----··--·-·· ·--·-----Notomastus latericeus 3. Prostomium has a single pair of eyes (pl. 24, fig. 2) ; first setigerous segment lacks neuropodia (pl. 24, fig. I) ----------------------·-···--·------·-------·---Notomastus hemipodus 4. Branchiae occur as single or double filaments from some abdominal segments _ ___ _ _ __ _ __ _ _ __ ___ __ _.___ ________ .______ ____ __ _ _._________ __ ..___________ .___.__ ______.___._Dasybranchus lunulatus 4. Branchiae occur as bushy tufts emergent from the upper edge of some abdominal neuropodia ----------------------·······-··-····-···················--···--··-· Dasybranchus lumbricoides ?Capitella capitata (Fabricius), 1780 Capitella capitata Fauvel, 1927, pp. 154-155, fig. 55; Ha-rtman, 1947, pp. 402,404-405. One small specimen comes from Turtle Bayou, Aransas Bay, Texas, October, 1950. (coli. A. E. Hartman) ; it was associated with decaying vegetation in fine sand bottom. It measures about 5 mm. long and is nearly complete. The first 7 setigers have pointed setae only, segments 8 and 9 have hooks only; there are no genital hooks, hence the doubtful identification, indicated above. Capitella capitata is recorded from cosmopolitan regions in littoral zones, especially from lagoons, sloughs or other inlets in mud or mixed bottoms. It has remained unreported in the Gulf of Mexico, but ~ay be recovered from soft muds in vertical burrows. ?Capitellides teres Treadwell, 1939 Capitellides teres Treadwell, 1939, p. 6, figs. 21-24. The original and only known specimen comes from Port Aransas, Texas. Through the courtesy of the American Museum of Natural History, I have re-examined this specimen. It is posteriorly incomplete, consists of the prostomium and 19 setigerous segments. Large ova are seen in segments 11 and 12; they show through the body wall. The first segment is setigerous, with notopodia and neuropod~a. Segments one to 8 have pointed setae above and below; segments 9 and 10 have long handled, hooded hooks. A single large yellow genital spine is visible on the middorsum, between the eighth and ninth segments. There is no sharp separation between thorax and abdomen. There are no visible branchiae on any of the segments. This .specimen resembles a species of Capitella (above) since the first segment is setigerous. It resembles the genotype of Capitellides Mesnil, C. giardi Mesnil, since genital hooks are present .in the female. It departs from species of both these genera in that there are no visible hooded hooks in the first 8 setigerous segments. In this respect there is no known genus in the family (see chart of capitellid genera in Hartman, 1947, p. 402) to define the species. If a new genus is erected for it, its place will be in the group having setae in the first segment and thorax with 10 se~ents. Mediomastus californiensis Hartman, 1944 Mediomastus californiensis Hartman, 1947, p. 4Q8, pl. 46, figs. 3, 4. Five specimens come from Alligator Harbor, Franklin Co., Florida, washed froiD fine sandy tubes on beach, Jan., 1950 (coli. D. Carpenter). The longest measure 20 mm. long and less than half a mm. across. The first segment is a smooth ring; the thorax includes 4 segments with only pointed setae and 6 successive ones with hooded hooks. The abdomen comprises about 80 segments and ends in a long, digitate cauda on the ventral side of the anal pore. The species is otherwise known from Oregon to California, in intertidal muddy sands. H eteromastus filiformis (Claparede), 1864 Heteromastus filijormis Fauvel, 1927, pp. 150-152, fig. 53; Hartman, 1947, pp. 402, 427-428. This is very abundant at the mouth of Godfrey Creek, Lemon Bay, Florida, Jan. 1938; on mud flats at Englewood, Florida; also Bald Point, Ochlockonee Bay, Franklin Co., Florida, March, 1950 (coli. L. M. Henry) and in Turtle Bayou, Aransas Bay, Texas, in fine sand with decaying vegetation, October, 1950 (coli. A. E. HartmanL Others come from Rockport, Texas, in beach at low tide, July, 1946 (coli. J. W. Hedgpeth). Length is 20-40 mm. but the posterior end is apt to be much coiled. The first seg­ment is smooth and lacks parapodia; the next 5 have setae in both notopodia and . neuropodia, the following 6 have long handled hooded hooks above and below; the transition from throax to abdomen is thereafter. Abdominal parapodia are somewhat enlarged and branchial. Color in life is deep maroon; preserved it is much paler. In most specimens the branchial processes are short vesicles; in some from Texas, however, these processes are conspicuously longer, resembling lobes .and they .are continued through many abdominal segments. · Heteromastus filiformis is widely known from both sides of North America, from Europe and elsewhere. Its surface mark is a small cone. Notomastus latericeus Sars, 1851 Notomastus latericeus Fauvel, 1927, p. 143, fig. 49; Hartman, 1947, pp. 402, 411. Finds come from Alligator Harbor and Bald Point, northwestern Florida, March, 1950 (coli: L. M. Henry). The prostomium has paired patches of eyespots. All thoracic segments have pointed setae only. Anterior abdominal segments have the superior edge of neuropodia! ridges prolonged to form branchial lobes. Some other specimens, taken near Beaufort, North Carolina, ~uddy sand, shore, agree fully with these from the Gulf. Color in life is dark red, preserved it is pale. This is known more extensively from Europ~ and other widely scattered geographic areas. It is not yet recorded from the western half of the Gulf of Mexico. Notomastus (Clistomastus} hemipod~s Hartman, 1945 Plate 24, figs. 1-3 Notorriastus (Clistomastus} hemipodus Hartman, 1947, pp. 424-426, pl. 48, figs. 1-5. This comes from Lemon Bay, Sarasota Co., Florida in a sandy shoal, January, 1938. Length is about 30 mm. The prostomium (fig. 2) has a pair of eyespots near the posterior margin. The first segment is smooth or somewhat areolated and lacks setae. The second has notosetae (fig. 1) but no neuropodia. Nephridial apertures, located at the sides in segmental grooves (fig. 1) are limited to posterior thoracic segments. Lateral organs, located between the notopodia and neuropodia, are papillar in the thorax (fig. 1) and eversible pouches (fig. 3) in the abdomen. The proboscis is a large, eversible pouch that is papillated proximally (fig. 1) and smooth, epithelial more distally. This is known more widely from North Carolina; it is not yet recorded from the western end of the Gulf of Mexico. Dasybranchus lunulatus Ehlers, 1887 Dasybranchus lunulatus Ehlers, 1887, pp. 174-177, pl. 45, figs. 5-9; Hartman, 19471 pp. 432--434, pl. 56, figs. 1, 2. Several come from southwestern Florida in sandy mud shoal exposed at low tide, January, 1938. The prostomium has a pair of ocular patches located in front of the nuchal slits. The first segment is smooth; the next 13 have setae in both notopodia and neuropodia. Abdominal segments have long handled hooks only. The epithelium 'of the thorax is nearly smooth except for a few wrinkles. Some abdominal segments have digitate branchial lobes that are emergent from a ~mall pore at the upper edge of parapodia. This is known from the Florida Keys and some West Indian regions in additied in place of the synonymous and more usual "sustainable catch", because the latter becomes liable to misunderstanding in cases where the· potential equilibrium of catch with abundance is being displaced by natural changes (unless the phrase is expanded to the form "catch sustainable under prevailing environmental conditions"). In other words, a given equilibrium yield is not the fixed characteristic of a given stock-level but would vary as systematically as do environmental affects (such as, say currents, which drift eggs and larvae from spawning areas to optimal nursery groUnds). At the same time. social changes may prevent maintenance of a catch equal to the equilibrium yield. Therefore a given bioiogically "sustainable'' catch often does not in reality remain sustainable. This cause& verbal confusion which is avoided by the substitute term, equilibrium yield. However, this inflection point is not the only one with a biological basis. Another, occurring at stock levels above . that for maximum equilibrium yield, is a conse­quenQe of difference in slope of the curve of change in equilibrium catch from that of the curve of change in the corresponding abundance of the stock. The sacrifice of potential catch required to raise abundance above the level for maximum equilibrium yield thus has a point of maximum return. For example, reduction of catch to some 7 per cent below the equilibrium limit which has been calculated fw:. marketable North Sea trawl fish by Baerends (1947, Fig. 1, using the ingenious method of Graham, 1935) would evidently correspond ~o a 25 per cent increase in the equilibrium stock. Greater or lesser hypothetical reduction in the catch would result in proportionately less improvement in hypothetical catch per unit effort. Suppose that both price per ton of fish and costs of a unit of fishing-effort were fixed at a level such that the value of the maximum equilibrium yield would not be a great deal more than the costs of obtaining it. The hypothetical equilibrium . catch. of 93 per cent of the maximum one calculated by Baerends would then be that of maximum net value of the total catch. In reality, of course, changes in tonnage of fish caught, in character of the stock, and in effort expended (all hav­ing interdependence) would entail change both in cost of a u:nit of effort and in the relation of quality and magnitude of catch to demand, hence in price of fish. Theref~re, even assuming all independent factors unchanged, the theoretical point of maximum product of equilibrium yield times corresponding catch per unit effort is by no means necessarily that of maximum sustainable net monetary return from the fishery (let 'alone necessarily optimal in a broad sense). Nevertheless, value judgments are incomplete without consideration of this second biologically defined i:Q.flection point (cf. Graham, 1935, pp. 265, 274); which may be termed that of maximum net product; and which may be defined algebraically as that at which Ce2/fe ==max., when Ce is equilibrium yield, andfe is the fishing-effort required to obtain this catch when the stock has reached equilibrium with it. In addition to the · direct effects of a particular fishery on its particular stock, repercussions on the environment must also he allowed weight in appraising the results of exploitation. The given fishery has in the first place to he considered not alone hut in its relation to others based on stocks comprised ~n the same ecosystem. For example, certain valuable fish (Sciaenops and Cynoscion) of the southern United States annually consume several times their weight of shrimp (Penaeus}, which latter support the most valuable fishery of the region. A third of a century ago, the shrimp catch was about double that of these fish, but sold at only a third of the price per pound, and it might have seemed clear that the fish-stocks ought not to be driven below their level for maximum equilibrium yield. The shrimp catch has subsequently quadrupled, while the commercial catch of the fish remains much the same; and the relative price of shrimp has risen to two-thirds or more. The present yearly consumption of shrimp by these fish is probably on the order of the present annual shrimp catch by man, which latter may he con­siderably more than the average standing crop of shrimp. It .will be seen that the limits. and costs of production of the shrimp would have to be considered in relation to those of the fish; and that a catch of the fish greater than would permit equilibrium might seem advantageous from some points of view. As a more far-fetched example of the necessity of considering indirect fishery. dependent changes, it may be noted that although important North Sea demersal stocks have probably been driven by fishing below the level for maximum equili.. brium yield, the potential North Sea balance of yield and cost of potential food might not necessarily have been much changed (thus, Graham, 1949, remarks the possibility that demersal "trash" has been favored, and herring might conceivably have benefited from reduced predation by cod). A rational trend in human con• .sumption of fish might be toward processed forms, like Indonesian trassi, which ~educe dependence on raw kind and size. Thus, if increasing scarcity of choice North Sea demersal fish were to be more or less quantitatively compensated by increasing abundance of less desired ones, and if this would favor the development of preference for processed substitutes, the ultimate result might conceivably he more advantageous to European ·society than that from restriction of the fishery to an equilibrium catch of demersal species and sizes now valued. In addition, fishery-dependent effects on environment may have more subtle and important consequences than the above. These arise from circular readjuSt­ ments in composition of the biomass which ·may in some cases include significant modification of basic photosynthetic production, as discussed in the next chapter. Beyond any consequences of fishing, the direct and indirect eflects of independent environmental change must also play a part in value judgments. There are numerous indisputable instances in which fishery-independent aflects have overwhelmingly outweighed fishing. There are probably also cases in which it may be doubted if overfishing during a period of abundance would have much influence upon the · time or extent of the next expansion of the stock following an intervening period of unfavorable natural conditions (cf. Burkenroad, 1946a, pp. 52-7). In appre­ priate circumstances, a catch greater than could be indefinitely sustained by the stock might thus salvage what would otherwise be lost (as has been suggested by Gunter, 1951, for Texas bay fishes, which are periodically killed by catastrophic cold). Further, the misidentification of changes in abundance caused by fishery­ independent affects may result in erroneous estimates of equilibrium yield. Thus, use of the term "optimum" to designate an estimate of the maximum equilibrium yield may be injudicious, especially since fishery administrators and legislators, and even biologists themselves, seem sometimes to be confused by the broader connotations of the word. "Rational," as employed by fishery hiologisis, likewise seems to have undesirable implications; since the reasonableness of an exploitation policy depends not only on the relation between its biological effectS .and the end desired, but on the rationality of the end (which, being a social matter, is difficult to judge, especially in advance). The fishery biologist, as such, is trained for objective consideration of the methods and eflects of human predation on the aquatic biota, but not of the more intricate counter-influences upon society, To learn and advise what kind, size and number of fish would be taken over what period by how much of what kind · of effort is a function not to be confused with that of deciding what eflort would be most advantageous to society ( cf. Sette, 1943,, p. 4) . The special function of the fishery biologist with respect to managemetl;t is to determine the requirements for biologically effectual ·exploitation (a phr~ intended to mean, that form of exploitation which will eflectually produce a glveil biological result; and which is therefore noncommittal as to the value of the given purpose to be answered by the biological result) . Attention may now be turned to the chief purpose of the present essay, which is to examine· the degrees and causes of reliability of conclusions concerning bio­ logically effectual exploitation of marine fisheries. It will be shown that, at the present level of development of marine science, the biological efiects of fishing are probably not predictable with requisite range and accuracy by the deductive methods upo11 which (except by accident, as in the North Sea) reliance has bee~ chiefly placed. II. The Effects of Environmental Change Dependent on Fishing. 1. Ultimate limits of potential harvest. >To introduce the problem of marine productivity, it is useful to compare a shallow sea with a fish-pond. Oriental · salt-water ponds utilizing food-chains of minimal length may yield as much as 5,000-7,000 lb. fishjacrejyear (cf. Har­vey, 1950, p. 132), which is more than 300 times the. comparable yield from the North Sea. However, the pond harvest would annually remove some 30-40 pounds of phosphorus per acre. The original upper six inches of pond-bottom might contain several hundred pounds of phosphorus; but even if this store in the bottom were subject to deduction after the first year or so (rather than, as is likely despite periodic draining and ploughing, to accretion), our present con­cern is with the sustainable yield of phosphorus, which could not exceed the income. Assuming a two meter depth of water in the pond, 35 lb./acre of annual phosphorus harvest would amount to some 2 g/m3/yr. Suppose that the natural waters avail­able were to average, say, 20 mg P/m3• Removal of all the phosphorus from 100 fillings of the pond would then be required to balance the year's harvest. Even if it were possible to create a condition in which the producers were all attached and the pond could be completely emptied on every suitable tide without killing the biota, photosynthesis would be slowed as the depletion of the nutrient in a filling ·ot water proceeded, impeding the extraction of phosphorus from intake-water at the rate required to balance the given harvest. A sustained annual yield of 35 lb.jP per acre of pond would thus in practice evidently require a supplementary supply of the nutrient element (as from human wastes) . If, now, the fish-pond is imagined expanded to North Sea dimensions (575,000 km2 X 96m), the anabolic function would have to' be conducted chiefly by phyto­ plankton instead of attached vegetation. To the extent that the producers would now freely drift out of the region considered with the water from which they have removed phosphate, there would be no differential in phosphorus content between in-and out-flowing water, hence the system would have no sustainable yield. However, in addition to simple sinking and turbulence which carry the plankton within range of the bottom fauna, a unit mass of planktonic herbivores is believed to take a quantity of plants per unit time which increases with the standing crop of plants, even when this rate rises above that of assimilation. At densities of phytoplankton too high for the graze to be fully used by the standing crop of planktonic herbivores, an excess ~s therefore' discharged undigested but with its sinking rate enhanced, and 80 remains in the region. At the same time, pla~ktonic herbivores not only in part metamorphose to sessile forms, but are in part consumed by a nekton which can maintain itself within the limits of the system against the outflow; and this process of retention is further assisted by predation of nekton by some of the demersal fishes. A differential between inflow and outflow can thereby be established, which can serve to balance phosphorus losses incurred by the irretrievable burial of nutrient in the bottom, and by terrestrial harvests. Thus, while the long food-chains of the sea dissipate energy, they may in appropriate circumstances increase the local availability of phosphorus to a more-than-compensatory degree. With a depth 50 times that in the model fish-pond, the model North Sea could evidently obtain phosphQrus equivalent to the given fish-pond harvest by complete depletion of two annual fillings of water containing 20 mg P/m3. However, the available North Sea exchange sources evidently average less rich, the rate of water-exchange may not be so rapid, the outgoing water is evidently far from completely depleted, and the unrecoverable losses to the bottom must be consider­able.3 If usable terrestrial contributions ,rom North Europe were less than the Mississippi River discharges to the Gulf of Mexico (therefore probably less than 10,000 tons of phosphorus per annum; cf. Riley, 1938, p. 168), and if losses to the bottom were, say, 50,000 tons P/yr., maintenance of phosphorus balance even under natural conditions would evidently require one annual exchange of the North Sea water mass with the ocean at an average differential of nearly 1 mg P/m3~ The North Sea fish catch removes some 10,000 tons P /yr. If this were a cumu­lative loss, it would probably amount in a century to nearly as much as the equi­librium store of the element on hand in the North Sea.4 It does not, however, seem 3Suppose, for purposes of illustration, that the upper centimeter of North Sea bottom has ·a dry weight of 0.5 g/cc, with an average of 4 per cent nonliving organic matt~r which contains phosphorus to 1 per cent of its own weight; that the rate of deposition is 1 mm per year; and that after 20 years the original upper centimeter has lost· two-thirds of its organic phosphorus by regeneration, the remainder ~ing unrecoverable (c/. Sverdrup, et al., 1946, pp. 230-1, 1015-6). On this basis, there might be some 1,200,000 tons of nonliving organic phosphorus in the upper centimeter of the bottom (on the order of 10 times as much as in the average benthonic standiug crop, to judge from estimates of the latter for the English Channel by Harvey, 1950, p. 129). This accumulation would be in equilibrium with some 150,000 tons P regenerated yearly, and 50,000 tons P permanently buried. If the oceans as a whole are not losing phosphorus, gains .from the land must balance losses to the bottom; hence the above hypothetical North Sea situation would not he typical. An oceanic balance would of course have the whole world and geological periods of time for its maintenance; on the other hand, the present is probably a period of relatively great erosioa and the Mississippi River a relatively great carrier of phosphorus. On this basis, the above hypo­thetical rate of loss to the North Sea bottom sounds rather high, although this argument against it is not telling. A study of the problem would be of importance. 4 Channel winter phosphate maxima (Cooper, 1948, p. 335; Harvey, 1950, p. 135) would he equivalent to a North Sea range between 1.4 g P per m2 of area (equalling the average of the phosphate-poor period in the Channel, 1931-39 and 1949) and 2.1 g P/m2 (equalling the phosphate-rich period in the Channel, 1925-29); or 800,000-1,200,000 tons total. To this should probably be added some 10 per cent for dissolved orranic phosphorus. The winter maximum of total water-borne phosphorus in the Channel is evidently only some 25 per cent above the summer minimum, but unless this difference results wholly from exchange with seasonally depleted ocean water, it might represent a considerable temporary accumulation in the bottom, and the non-planktonic biomass; so that the whole of the average biomass could not be regarded as containing phosphorus additional to the winter phosphate maximum. The average biomass, as derived from the estimate in terms of dry organic matter per in2 in the Channel by Harv.,Y (1950, p. 129), would be on the order of 50 times the annual North Sea fish catch, hut ils phosphorus content per pound would probably average less than half that of fish. This would amount to some 350 mg P/m2 or 200,000 tons (of which the amount allocated by Harvey to the average standing crop of plankton would equal some 65 mg P/m2, or 40,000 tons). To lite corrected sum should be added the amount of phosphorus yearly regenerated from nonlivia& organic matter in the bottom. The average total equilibrium store of potentially usable phoe phorus on hand in the North Sea might thus be some one and a third million tons. likely that this drain by fishing· would have a cumulative effect on the phosphate content of the water {as evidently thought by Baerends, 1947, p. 25), since it would amount to less than 0.2 mg P/m3 /yr., little more than a hundredth of the pre­ sumptive average winter phosphate maximum, some thirtieth of the average differ­ ence b~tween rich and poor years of winter maximum off PI ymouth, and a quarter of the above-assumed natural differential between in-and out-flowing North Sea water. The direct effect of present fishing could thus be compensated by an uncle­. tectable reduction in the average content of total phosphorus in the outflowing water, probably much smaller than what may occur during adjustments of the system's balance to natural fluctuations in phosphorus content-of inflowing water . . At the same time, however, it will be seen. that the ultimate limit for sustainable North Sea yield of organic matter in any form might not exceed the present harvest by tens of times (if so much). A tenfold increase in the harvest would require for balance a complete annual exchange of water at a phosphorus differential averaging some. 2 mg/m3 /vr. more than the excess of irretrievable losses into the bottom over ue-able income irom the land. This might be incompatible with maintenance . of the living balance on which the availability of the phosphorus income depends, since this income cannot be regarded as a fixed quantity independent of the biota. The gross inflow from rivers and ocean is, indeed, independent; but the amount which can be harvested from the system without unbalancing the phosphorus budget depends on how much of what comes in can be kept by the living part of thP-system from flowing out again and from being irretrievably buried. Conse­ qu~ntly, if. phosphorus balance is to be maintained, the North Sea landings have either to be of magnitude inappreciable relative to the natural exchanges, as they evidently are at present; or to be of such nature as to improve the efficiency with which losses to bottom are checked and (or) depletion of outgoing water maintained. Leaving the question of maintenance of North Sea phosphorus balance in the above somewhat indefinite state of resolution, another aspect of the limit of poten­ tial rate of yield of phosphorus may be considered. The phosphorus income pro­ vides a measure of the equilibrium harvest, but not of the basic production of the system. A pond with an income of zero but with a considerable store of phos­ phorus on hand and an organization permitting rapid recirculation of nutrient to the producers could maintain a dense standing crop although without being able to come to equilibrium with any catch at all. In contrast, an appropriately con­ ~tructed model with a considerable income fully utilized could nevertheless have a standing crop too sparse for profitable fishing. In the former model, any rate of continued harvest would result in a corresponding rate of deceleration of the biomass time-derivative, checked only at zero biomass or at abandonment of fishing. In the latter model, the time-derivative would at first decelerate at a decreasing rate in response to a given rate of harvest, and new equilibria could therefore be reached at levels above zero; but since even the initial biomass would (by definition) not be high enough so that the energy-equivalent of the harvest would be com­ parable with what cou\d be produced by some other investment of the energy required for harvesting, fishing would not be profitable. The North Sea is a system in which phosphorus supply evidently limits photo­ synthesis, and in which the natural equilibrium store of this limiting element on hand in one or another form is probably. on the order of thousands of times the maximum daily available income from outside the system. The supply of energy consumed in the living part of the North Sea system must therefore be strongly affected by the rate at which the phosphorus on hand is returned by the consumers to the producers to be recompounded with energy. The magnitude of the consUJD­ing biomass would be expected to display correspondence to rate of supply of consumable energy, and this relationship might also hold for the fraction of the biomass at high trophic levels. What appears to be such a correspondence between standing crop of marketab~ fish and rate of basic production can, in a crude way, be shown by comparison .of the North Sea with New England trawling grounds. For the North Sea, the average catch of marketed fish per day's absence of English steam-trawlers, 1920­1934, ranged between 1,500 and 3,000 lb. (cf. Baerends, 1947, fig. 3). Assum­ing an average of some 485 acres swept per day's absence (perhaps too high, but compare Parrish, 1949, and Margetts, 1949 a, b), the marketed catch would have run around 3-6 lb./acre (gutted weight). Even just after the first War, the level would evidently have been only some 10--12 lb./acre. Margetts and Holt (1948, pp. 29-31) give data suggesting 3-5 lb./acre in 1938, 12-17 lb./acre immediately after the resumption of fishing following War II, and 7-9 lb./acre a year later. For the catch of haddock alone, the maximum four-year average given by Baerends (1947, Table 6), 1915-18, would amount only to some 5 lb.jacre. Parrish (1948, p. 7) gives gutted weights of haddock per 100 hours' fishing by selected vessels in selected areas which (assuming 27 acres swept per hour) indicate a catch of some 13 lb./acre immediately after resumption of fishing in 1945 (some four tiines the 1938 level). For New England, Herrington (1948, pp. 271-2) gives data on average land• ings per day in 1926-31 and 1941-43 by a selected group of large Georges Bank trawlers. Assuming a sweep of 485 acres/day the resuits would be 30--90 lb. of marketed fish per acre, of which some 65-90 per cent was haddock. In addition, unmarketable fish (some of which might have been saved in the North Sea) amounted to 35-60 per cent of the marketable haddock taken, in years for which data are available. Herrington's sample may have been so selected or manipulated as to be above average, since, for example, whereas he gives 16,500 lb. haddock per vessel per day in 1942, Anderson and Power (1946, p. 87) give days absent and landings of gutted haddock for large trawlers in the southern part of Area XXII in 1922 corresponding only to 9,100 lb. (i.e.,' only 19 lb./acre instead. of 32 lb.). However, the indicated crops would still be far above comparable North Sea levels, unless the average fishing power of the American vessels is much greater than the European. Merriman and Warfel (1948) give data on the catch of small Connecticut drag­gers fishing near the mouth of Long Island Sound in 1943-46. In this area, 90 per cent of the regularly marketed catch was blackback flounder, an exclusively American form (the remainder was mostly cod; while haddock, thovgh sporadic in the area, was absent in the period of study) . These d1Jta are also.not entirely comparable with those for the North Sea and Georges Bank because they refer to round weight of sample catches of all sizes of different species per hour's fishing in different months, and are not weighted for seasonal changes in commercial fishing time. At the authors' figure of t2 acres (or less) trawled per hour, the ·average poundage of regularly and occasionally marketed species in monthly sampling hauls in the . different years amounted to at least 20-60 lb./acre, in addi­tion to forms not marketed which comprised 35-50 per cent of the whole catch. The occasionally marketed species, which would probably have been saved in the North Sea, varied from less than half to equal the amount of regularly marketed ones. It thus appears that the standing crops of marketable demersal fish per unit area in the North Sea, even after several years of rest when thought to be approach­ing the maximum which the region would maintain, may amount to less than those in the poorest seasons for Georges Bank and Block Island Sound after years of uninterrupted fishing of considerable intensity {according to Riley, 1949, Block Island marketed catches run about 25-50 lb. per acre per year, and Georges Bank 7-33 lb./acre/yr. This annual catch per unit area, though considerable, still seems a m~ch smaller proportion of the standing crop than the 5-7 lb./acre/yr. obtained ­when yearly North Sea demersal catch,. excluding war periods, is divided by total North Sea acreage) . The ratio of fishable biomass on the European grounds to that ·on the American grounds might, at equilibrium with equal rates of fishing mortality, evidently be as little as 1 to 4 or 5. Channel phytoplankton production in a phosphate-poor year (which is, of course, probably not representative of the whole North Sea) is estimated by Harvey m2 (1950, p. 107) · to be at least 120-200 g of dry ash-free organic matter per of area per year. From Riley's data for Georges Bank (obtained by a different method; Riley-et al., 1949b, p. 44), Harvey derives an estimate of 270 g/m2/yr.; however, · RiJey's own ,estimate of annual production runs much higher, on the order of 70.0 g in terms of dry organic matter (cf. Riley, 1949, p. 18). For Long Island Sound, Riley found no maximum as high as on Georges Bank, but a much higher production average by months; however, depth of the euphotic zone was probably less; and his various estimates for total annual production in Long Island Sound (cf. Sverdrup et al., 1946, p. 938) suggest a level under 1,000 g/m2 (dry organic matter). Phytoplankton production in the North Sea thus seems likely to be much less than that in New England trawling areas; and it may be that the ratio under comparably representative conditions would approach 1 to 4 or 5. Consequently, it appears possible that the standing crops of marketable demersal fish in these two shallow temperate Atlantic marine ecosystems may tend to be proportional to basic production. Basic production has been shown by Riley et al. (1949 a, b) to be a function of phosphorus supply (in the sense that, given the physical characteristics and the empirically derived physiological constants, the resultant standing crops of plant and animal plankton have been predicted with astonishing accuracy from variation in phosphate alone) . Therefore, quite apart from the matter of phosphorus balance, the effects of fishing on costs of production · ~f animal matter from the North Sea might depend on whether fishery-induced "'""cltange~r in··composition--·of -the--hi~~ ,sig~tl¥....altec.. the . ..r.ate .of re­generation of the store of phosphorus on hand. It is clear, from the effects of the two wars, that the North Sea demersal fishery does . produce large changes in composition of that part of the biomass on which it directly bears. Even if the indicated reduction of the 1938 marketable demersal · fish stock to some one-fourth of the natural level had been balanced by increase in those demersal carnivores of such size and habits as are not much caught, signifi~ant changes in the pressures upon different parts of the stocks of prey of the marketable fish must have ensued. Harvey (1940, p. 129) assigns to demersal fish in the Channel a daily consump. tion of food amounting to half of the daily production of benthos. It seems likely that the marketable demersal fish would be to a considerable extent supported -by; and would be the principal consumers of, the larger and older benthos. Harvey (ibid., p. 131) assumes, as a rough approximation, that about a quarter or a thinl by weight of the Channel invertebrate bottom fauna, comprising the young and small elements, loses by respiration 4 per cent daily on an average; the remainder burns only 1% per cent; and the whole community respires 2 per cent of its . weight per day. If this were taken to represent the state of affairs in 1938, wheJJ, the stock of marketable demersal fish was around one-fourth of that in 1945, it seems conceivable that by 1945 the composition of the stock of benthos might have been greatly changed by the increasing pressure of the larger fishes. If the chailge were great enough, say, to reverse the mass ratio of young and small to large and old benthos, it might evidently raise the average respiratory rate per nnii weight of the whole benthonic invertebrate community nearly to 3 per cent. It is difficult to guess just what would be the corresponding effect on the total mass of benthos; but if, for present purposes, it is assumed that change in total mass would be small relative to that in composition, the benthonic respiratory require-­ment might evidently have increased greatly from its presumptive 1938 level of some 3~0 per cent of average daily phytoplankton production. Phosphate concentration in the North Sea region is evidently driven below the optimum for photosynthesis during all months of favorable radiation and tempera­ ture, even in the bottom water; and there is a decline in total phosphorus in the water-column from winter to summer apparently amounting to some 25 per cent (cf. Harvey, 1948, pp. 357-8), which suggests the possibility of considerable sea­ sonal accumulation in the benthos. Therefore (even if summer accumulation of dissolved organic phosphorus forms a bottleneck in recirculation), an increase in average rate of phytoplankton production of, say, one-fourth might he permitted­ and entailed-by an increase in benthonic respiration of the order assumed above. In other words, the system might operate at an appreciably higher rate of phos­ phorus turnover under natura]· conditions, with a commensurate improvement in its capacity to support a crop of marketable fish. The North Sea demersal fishery might, then, significantly reduce basic production and the biomass time-derivative, even though phosphorus balance was maintained. It seems conceivable that fishing for zooplankton-eaters might also impair the efficiency of the North Sea system (since these nektonic predators may provide a check, vital for maximal production, on the grazing of phytoplankton before the time of the spring and fall blooms; cf. Harvey, 1950, p. 112). Whether a direct fishery for herbivores might in part balance the reduction in rate of turnover of phosphorus brought about by a fishery for predators is not altogether clear. Neither is it obvious whether the effects of fishing would be altogether similar in shallow ecosystems of somewhat different type from the North Sea, as the broad eastern shelves of the continents in temperate latitudes; or Chesapeake Bay with a large income both of terrestrial phosphorus and of sediment; or Georges Bank where there seems great opportunity both for exchange of water with deep sources rich in nutrient, and for loss thereto of the product of photosynthesis. Where grazing rather than nutrient is the limiting factor for utilization of radiant energy by photosynthesis, as presumably in the Antarctic upwelling, serious reduction in the enormous abundance of quick-growing herbivores might require staggering catches. In terms of the models discussed on previous pages, the up­ wellings C}re systems with a phosphorus income high enough to support a large standing crop without regard to the amount and rate of katabolic turnover of nutrient accumulated in the area. This type of productive system might be termed anabolic. If efficient means to catch herbivores as small as Antarctic euphausiids are developed, the ultimate limit of yield of organic matter by the katabolic systems where standing crop is largely dependent on accumulated nutrient (such as the North Sea) will probably remain an academic question, even in case of greatly increased demand for food-stuffs from the sea. Nevertheless, a grasp of the relations of income and metabolic turnover of nutrient to productivity permits the detection of misapprehensions concerning the limits of yield which might influence practical decisions (such as the mutually contradictory ones held by Baerends, 1947, and Taylor, 1951); and at the same time assists apprecia­ tion of the difficulties of deductive identification of the biologically conditioned points of inflection in the curve of diminishing returns from fishing. -2. Particular curves of population-logistics. It has been noted in the preceding subsection that a decline in the stocks of marketable North Sea demersal fish, resulting from fishing, might cause an increase . in the stocks of large and old prey. In this case, sudden cessation of fishing might permit the demersal fish reaching marketable size to draw upon an accumulated food supply larger than would exist at natural equilibrium. Consequently, it seems possible that during the two wars the marketable demersal fish might temporarily have reached levels of abundance higher than naturally sustainable under pre­ vailing physicochemical conditions; and that the rapid postwar declines in abun­ dance of marketable sizes might have reflected not only the resumption of trawling but temporary reduction of the appropriate food-stocks to levels less than natural. If su~h over-expansion of demersal fish-stocks did in fact occur, the limit for maximum equilibrium catch might evidently be placed at a lower st?ck-level than has been estimated from the effects of war. The above suggestion of oscillating adjustment would seem to call for various special sorts of drastic change in relative abundances and growth rates of different size-classes and kinds of North Sea demersal fish during and after the wars, such as have apparently not been observed by workers familiar with the data (although it should be noted that there did occur in the early 1920's a temporary trough in catch per unit effort; which is attributed to the effect of poor brood years by Russell, 1942; but which might conceivably result from overrunning the food supply). Further, a countervailing buffer might exist in the form ·of improvement in katabolic efficiency of the system (see above). However, whether or not evidence of war-cattsed oscillation can be found, the indirect effects of fishing do need con­sideration when deductive methods or extrapolations are used. The general problem may be illustrated by analysis of the curves of population equilibria constructed by Graham (1935) and by Baerends (1947, Fig. l) for marketable North Sea trawl fish. These curves are such that the equilibrium con­ tribution of a unit of weight of stock in a unit time would, when plotted against number of units in the total stock, follow a straight line (the equation for which, indicated by Baerends' Figure l and p. 33, is roughly 2x -170y -4.5 million tons== 0). This straight-line relationship (expressed by Graham, 1935, p. 269, as direct proportionality of rate of natural increase of the stock to the difference between mass of stock at the moment and the maximum which the area could support) means that the contribution by a unit of stock was assumed to be reduced by a constant amount for each unit added to the weight of the total stock, whether this unit increase in the total stock is the pt or the nth. In other words, change in sustainable contribution by a unit of stock was assumed (as "a reasonable first approximation"; Graham, ibid.) to be inversely proportional to change in total stock. However, it seems fairly clear that even under constant fishery-independent con­ ditions, some at least of the factors governing contribution by a unit of stock would not be expected to change in strict inverse proportion to changes in total magnitude of the stock. For example, the change in availability of food which is repre~ted by a unit change in equilibrium stock magnitude of the fishery stocks' prey, would presumably not be constant for a unit change in the amount of prey consumed per unit time by a changing fishery stock. Instead, it would be expected to vary with the changing situation of the prey stock in relation to its own curve of logistics. Hence, changes in energy expended to catch food per unit time by a unit of fishery stock (and in its rate of natural mortality, etc.; all of which might perhaps be subsumed under the heading of rate of utilization of food for net contribution of weight, which is not the same thing for a population as for. an individual) , would have to vary so as to balance the varying change in availability of food. Meanwhile, the change in the stock of prey would presumably have affected the sustainable yield of its own food, in a manner varying with the selectivity of the given :fishery among competing predators and their foods, and with the habits of the food-sto~ (as, predominantly herbivorous, scavenging, or predatory on these or on other predators) , etc. Regular and symmetrical behavior of so intricately balanced a resultant as that indicated may (ceteris paribus) occur, hut it seems possible that, instead, sustain­able catch of a particular stock of a complex ecosystem would often change with magnitude of the stock in an asymmetrical way. In other words, more than one critical stock level might exist for the same :fishery species (for example, one p.eak ·of sustainable catch might occur after :fishing has so far reduced the stock as to permit important development of a· food resource previously driven far below its own level for maximum sustainable yield; a second at stock levels too low to permit epidemic transmission of a disease; and abundance might oscillate back and forth across the trough between, under the impulse of :fishing). In consequence, the interpreter of :fishery data might sometimes even be dealing with a curve of popu­lation growth inherently of compound type; or one so skewed that sustainabl~ catch would tend to be inversely proportional to stock magnitude over consider~le ranges of the latter, resulting in wide plateaus of maximum net product; etc. The difficulty of reliable extrapolation from a curve of population-logistics con­ structed from a limited range of observations on a particular component of the biomass of a complex,ecosystem, may be illustrated in still another way. We have seen that it is co~ceivable that basic production, and with it .the capacity of the North Sea to sustain marketable fish-stocks, might be reduced as these stocks are cut down by fishing. In this .case, the assumed rate of. change of the slope of the upper part of the growth curve constructed by Graham and by Baerends (ibid) for North Sea demersal stocks may be far too rapid (in conformity with which Margetts and Holt, 1948, indicate that stock-levels attained during the recent war may have been above the extreme limit estimated by Graham from the effects of the first war) . It might even, be possible that the point of maximum net product might be consider­ ably farther above· that of maximum equilibrium yield than is indicated in Baerends' curve, with consequent error in proposals for North Sea management (although in this case on the side of modesty) . Another difficulty in extrapolation caused by fishery-dependent changes is tech­nical, although it might sometimes he of such importance as almost to be called fundamental. Curves of logistics of marine fishery stocks are usually defined in terms of relative catch per unit effort (although periodic f:hecks of the relation of catch per unit effort to stock magnitude are generally attempted by tagging; cf. Sette, 1943, pp. 5, 19). Thus, in situations where catch per unit effort of a species may he affected not only by its abundance but by dependent changes in abundance of other species, construction of a reliable population curve presents special difficulties. For example, in hook and line fisheries such as that for Pacific halibut, catch per unit effort must be considerably influenced by changes in relative abundance of bait-stealers or ."trash." "It is not an infrequent occurrence that the halibut men will haul in their gear and find on every hook a large rock cod, much to their disgust" (Chapman, 1942). If the halibut fishery were to favor these competing · species (as it might be expected to do, since the gear and methods are designed for preferential capture of halibut, and spots where the competitors are less annoy­' ing are no doubt favored) , the catch of halibut per skate of hooks would be expected to be reduced by fishing at a faster rate than is the stock of halibut, which would result in systematic distortion of a population curve based on catch per skate. Avail­ able evidence from halibut tagging seems insufficient to resolve this uncertainty.5 IlL The Effects of Environmental Change Independent of Fishing. 1. Results of failure to discriminate fishery-independent changes: Needler (1948) points out that even if Pacific halibut fluctuations were pre­ dominantly natural in cause, regulation of fishing might raise the levels of abun­ 5Data on bait-loss or preoccupation of hooks therefore seem important for assured evaluation of changes in halibut catch per skate. Indications of rate of bait-loss in the past might still he obtainable, if records of bait supply have been included in fishermen's logs, and if the quantity of bait damaged in hauling, discarded when gear is taken up for the day or otherwise replaced . without having been eaten, the quantity of unrecorded bait caught and used during the trip, etc., are not so great as to obscure the ratio of recorded bait-supply to catch. A less direct approach might be made by comparison of relative catches of those sizes of halibut and "trash" which are fully vulnerable to both hook and Ott(;T·trawl, when both types of gear are used in the same area during the same period. This method might simultaneously afford some indication &f the minimum density of fishable halibut stock corresponding to a unit catch per skate; which would aid in resolving the contradiction retween results of calculation of absolute magnitudes of halibut stock from rate of fishing mortality and catch by Burkenroad (1948), and those from catch per unit effort and catch, by Thompson (1950; a method questioned by Burkenroad, 1950). dance about which the fluctuations occur; and a similar point of view is indicated by Kesteven (1950). However, although restriction would be expected to raise the catch per unit effort at the expense of the catch, it would not be expected to improve both of these values unless the stock had been driven by fishing below. the level for maximum equilibrium ·catch. But if the stock were fluctuating because of natural changes, to determine whether it was above or below the critical level would be difficult.6 For example, suppose that a stock were being fished at a rate considerably leM than the maximum sustainable under prevailing conditions, and was in consequence declining toward an equilibrium with this rate of catch. An acceleration of tiWt decline, by a natural change which was nevertheless insufficient to prevent ultimate attainment of equilibrium with the catch, would produce an appearance which might for a time be mistaken for overfishing. Restriction of the fishery as a result of the mistake would prevent full use of the resource, and might at the same time itself impede realization that overfishing had not occurred. As a concrete illustration, the Area 3 stock of Pacific halibut wa_s apparently capable during the early 1940's of achieving equilibrium with an annual catch much greater than 27 million pounds (since catch per unit effort rose from 121 pounds in 1940 to 151 pounds in 1944, at a fairly steady level of catch averaging over 27 million pounds; according to data of the International Fisheries Commission). But the annual catch was never as much as 27 million pounds from the beginning of the fishery until 1927; indeed, the average catch, 1915-1926, was only some 18 million pounds; and even in the period of highest pre-regulation catch, 1924-1930, the maximum was only 30 and the average 27 million pounds. The hi~tor.y of the Area 3 halibut stock is nevertheless believed to indicate rapid decline from a virginal abun­dance in 1912-1915 considerably greater than that in 1944, to a pre-regulation 6A natural change in, say, phosphate or temperature, affecting absolute quantity of food supply available to a stock or the ratt at which it can he consumed by the individual, would obviously entail change in maximum equilibrium yield of the consuming stock. A change affecting only percentage of hatch or larval survival might also, less directly, modify the maximum equilibrium catch, since if the magnitude of the brood stock require-d to maintain the rate of recruitment were changed, so also would he the relative food supply per fish at the stock level in equilibrium with a given rate of catch. Fluctuations in phosphate supply (such as the notable one off Plymouth) might cause correlation between larval survival and later food supply of the year classes, not only for a particular species, hut for a number of kinds simultaneously. Increase in phosphate causing improved survival 'of larvae would also he expected to result in an increase in the phosphoms store on hand in the system; which might not he drained away for many years after return to a lower level of income (Cooper, 1948). Hence, large year classes resulting from the fortunate concurrence of the critical larval period with a surge of planktonic food induced by improved phosphate supply, might find exceptional forage in later years as well. Thus, the fact that dominant year-classes, rtcognizahle as such at their earliest appearance, may yet show growth and survival as good as that of much smaller broods of the same species, does not necessarily mean that the food supply is not limiting in later life. Correlation of larval survival and exceptional later food supply might also he perniitted by identity of conditions favorable to survival of both a predator and its particular prey; but. unle$S the predator were narrowly limited in choice of food, buffering might he expected to level out the effects of fluctuations not causing change in the total biomass supported by the system. Th~ disappearance of the eel-grass does not seem to have had the drastic effect on coastal fisheries which might have been anticipated from pioneer quantitative studies on marine food-relationships.; Composition of the fauna has changed since the eel-grass die-off (thus, scallops became rare m regions not providing alte.rnative supports for the attached young), hut available information does not seem to suggest the expected massive diminution in intracoastal fishable standing crop!$ as a whole. This might result simply from utilization by phytoplankton of nutrient no looger taken by eel-grass, with resultant maintenance of the level of hasio production. nnmmum much less than in 1944 (average catch per unit effort over 200 pounds in l915-l6, 94 pounds in 1926 and 65 pounds at minimum in 1930). If these suppositions ·concerning relative stock-magnitudes are correct, and if maximum equili­ brium yield of Area 3 halibut in the 1940's was considerably more than 27 million pounds, then it is hard to see why the stock should have been driven far below present magnitudes by annual catches which were so much less than 27 million pounds; unless natural conditions were formerly less favorable. But if natural conditions in Area 3 have changed to this extent, it cannot be thought certain that the pre-regulation stock was ever driven below the level capable of yielding the maximum sustainable under the natural conditions then existent. It is true that, if the stock at the beginning of regulation had in fact been below the critical level, the legal restriction (which prevented the. catch from increasing as fast as demand recovered) should have improved the rate of response of the stock to improvement of natural conditions. On the other hand, if the stock was still ahov·e its level for maximum sustainable yield when regulation of the fishery was begun, the wastage of potential catch from 1931 to date would· comprise not only the increase in the stock during this period (which could he recovered), hut also the cumulative unrecove~able loss resulting from the relative depression of the rate of addition which would be expected to develop as population density increased. If improvement in natural conditions has been large, this unrecoverable loss of catch would have been compensated only in part by that fraction of the rise in catch per unit effort resulting from the regulation. Loss from the restric­ tion of the fishery might thus by now he large enough to .outweigh any risks which would have been taken in 1931 if the regulations had been designed in the form of a controlled experiment able to yield conclusive evidence of the biological situation. It can hardly he questioned that the Area 3 stock would not have increased so much as it did, without the regulations. However, whether or not the Area 3 stock was ever below the level for maximum equilibrium yield, it is difficult to believe that it was not far above any critical level in 1944. . Therefore, it is not at all . obvious what is the exact purpose which is intended to be served by Area 3 quota-levels; since gains from reduction of costs of production and of competition among members of the halibut fishermen's and dealers' guilds, etc., may evidently be counterbalanced by injuries to consumers, independent fishermen and dealers, ports adjacent to Area 3, etc. The lack of clarity in official thinking about the public interest which is thus ·mooted, gives reason for doubt of the desirability of restrictions which· are not designed to be self-correcting, especially in fisheries which may he subject to significant influence from natural change in environment. 7 7Tbe possibility has been suggested that changes in abundance of Area 3 halibut might follow a regular pattern of natural fluctuation, and that in this case, at some time not many years after 1944, "abundance might be expected to begin to drop precipitously despite whatever restriction of the ·fishery" (Burkenroad, 1948; a suggestion made public on January 10, 1947, before data . later than those for 1944 were at hand). The Report of the International Fisheries Commission for 1949 (1951), received after the present paper was in galley-proof, reveals that events have more or less conformed to the above prediction. Catch per skate on Area 3 declined from .151 lb. in 1944 to llO lb. in 1949, a loss of about half the gain registered between the beginning of regulation in 1931 and the 1944 peak. This loss might amount to more than a hundred million pounds of halibut, and is thus not accounted for by the catch unless there was no recruitment or gain from excess . of growth over natural mortality at all. The Commission states (ibid.) that changes in the fishery during recent years have been such that _actual abundance of the stock in 1949 was more than 200 per cent of that in 1930, even though 2. Difficulty of discriminating fishery-independent changes by the deductive method: In the case of Pacific halibut fishing-area 3, it is fairly obvious that (I) the apparent magnitude of the fishable stock (as calculated on the assumption that catch per unit effort is proportional to stock magnitude and that absolute magnitude of fishable stock in a reference year is given by the quotient of the catch divided by the fishing mortality rate estimated from a tagging experiment) shows changes much greater than would be expected by subtraction of the catches from· the stock (Burkenroad, ·1948); (2) the characteristics of the catches do not permit absolute magnitude of stock to he estimated as a quantity small enough to allow the changes: in stock-magnitude to be accounted for by the fishery, even when allowance is made for effects on rate of addition (Burkenroad, 1950); (3) the recent equilibrium yield is considerably greater than the catches which were accompanied by a decline in abundance from a virgin level presumably much greater than at present to a minimum apparently much less than at present (above) . In the case of · Area 2, from which a 'larger harvest than in Area 3 has been taken for a longer time, the same simple approaches as the above-mentioned do not reveal an impressive discrepancy between the catch and the changes in ahundance.8 However, compatibility between a cause and its supposed effect is not the same catch per skate was only 170 percent. If this is true, it would seem that abundance in 1944 may have been around 275 per cent of 1930, which would perhaps be comparable with that in the early days of the fishery. Such a return to an abundance near that of _1915-16, despite a . catch which had for many years been much greater than any prior to 1924, would provide additiQnal indication that changes in magnitude of the Area 3 stock have been chiefly natural and little affected by the fishery. If abundance at the peak prior to the decline which began in 1945 did in fact actually reach near the 1915-16 level, the possibility that there is a natural cycle regular as regards both period and amplitude seems strengthened, although the period of this cycle would apparently be under 30 years instead of "around thirty-four years" as originally suggested. It seems worth note tlu!t the marine cycle off Peru has a period of about seven years and that of New England starfish about fourteen years, so that a period of quadruple seven years for a halibut cycle might fit a widespread pattern. A renewed upturn of Area 3 abundance would on this basis be expected around 1959. The Commission (ibid., p. 23) reports that analysis of the commercial catch from one of the regions of Area 3 reveals that its age-composition in 1949 was similar to that when sampling was first undertaken in 1927, 13 and 14 year olds being predominant in a range mainly from 10 to 16 years. The Commission states that this "suggests very strongly" that such part of the recent decline in abundance on Area 3 as is "not attributable to changes in the fishery may have been a normal result of the reduced number of spal'o"Ders ... during the late 1920's and early 1930's . . ." Such a suggestion hardly seems warranted by the facts. Fish of age 13 years in 1945--49 would have been the product of the presumably rising brood-stock of 1933-37. A decline resulting from reduced brood-stock should therefore apparently have been ending, not beginning, in 1945 (and it should be noted that Thompson, 1950, p. 36, in a comparable but inverse attempt to explain an . inconvenient change in the stock, states that the timing of the "secondary increase following 1936 ... is a very strong argument in favor of the explanation that the spawning stocks were restored and are beginning to provide new young to the catch") . Actually, no close, simple, short-term correlation of brood-stock with recruitment would seem· expected in a case where there is a long sea-drift of planktonic larvae to nursery-grounds, followed by a ten-year interval before recruitment to the commercial stock begins; and no such relation seems to be exhibited in the rest of the available data, from either Area. Past parallelism of catch per skate on Areas 2 and 3, discussed in the next subsection, suggest!{ that a sharp decline to match that on Area 3 is due on Area 2; and the Commission (ibid.) reports a recent reduction in recruitment which may anticipate this event. Since Area 2 recruitment is complete at the age of eight or nine years, attribution of such a forthcoming decline to the low level of the brood-stock around 1930 will hardly be considered, and a situation which should throw light on the Commission's attitude;; might develop. 8The history 9f this fishery is as follows: When serious commercial fishing began on Area 2 in 1888, availability of large halibut was evidently enormous. Indeed, Thompson and Freeman (1930, p. 20) suggest that exploratory reports of scattered fish only in Area 3 at this time w~ merely relative (''Doubtless they were scattered in comparison to the wealth of fish found to thing as satisfactory proof of causal relation.ship. The data of fishery biology are in a majority of cases so incomplete, inexact .and unrepeatable that to draw definite conclusions from them demands an element of bias. It is therefore not only a legiti­ mate but a desirable and useful procedure in this field to compare any conclusions reached with those obtainable from the operation of an opposed bias on the same data. For best results, the contradictory working hypothesis should be maintained with sufficient genuine conviction to empower a penetrating pursuit through the mazes of inference, but not so strongly as to result in the unconscious concealment of ~~sults unfavorable to the hypothesis pursued. In the case of the Area 2 halibut fishery, the hypothesis that there has been an important change independent of fishing would lead to particular scrutiny of the reported rate of fishing mortality, which plays a strategic part in the above-described appearance of congruity between rate of catch and rate of decline of catch per unit effort. The reliability of this mortality-estimate is open to question because of the peculiar discrepancy between Area 2 and Area 3 in ratio of estimated stock magnitude to catch per unit effort (Burkenroad, 1948, pp. 97-8). The distribution of tags among halibut in the experiments of the 1920's is dbscribed as having been deliberately related to the distribution of fishing-effort (Thompson and Herrington, 1930, pp. 28, 54). In Area 3, returns indicated that the mature fish which preponderate in the catch move freely and rapidly over the entire coast above the boundary between Areas 2 and 3; so that the tagged fish would probably tend to become more or less equally available for recapture no matter where in the Area they were tagged. In contrast, the Area 2 catch has, since early in the 1920's at least, consisted chiefly of immature fish, which are shown by the southward"). They state (ibid., p. 27) that before 1900, and even up to 1909, a catch of 1,000 pounds per skate of the gear then used on the inshore banks (probably equivalent to more than 500 lb. per standard unit effort) "was not considered fully satisfactory fishing, causing a move to new grounds." Up until 1905, the maximum reported annual catch from Area 2 was only 28 inillion pounds, but in the period 1907-1913 it ranged from 50 to 60 million pounds. ·By 1910, even on the newly opened offshore grounds to which the fishery was extending in conse­ • quence of declining inshore abundance, catch per standard unit effort was only some 270 lb., though this still consisted chiefly of mature fish (Thompson and Freeman, 1930, p. 34; Thompson, Dunlop and Bell, 1931, p. 51). By 1915, average catch pe-r unit effort was less than 120 pounds, and the annual catch after this date only once exceeded 33 million pounds. The winter fishery for spawning adults died out not long after 1915 (Dunlop, 1937). During the last six years of the 1920's the annual catch averaged only 26 million lb. (maximum, 27 million lb.); yet catch per unit effort continued to decline, from 56 lb. in 1924 to 35 lb. in 1930. Thereafter, at catches never below 22 million lb., catch per unit effort began to rise. It was above 90 lb. by the late 1940's, at catches above 27 million lb. It will be seen that, if the Area 2 catch per unit effort corresponding to maximum equilibrium yield had been more than 120 lb., and the maximum equilibrium yield had been less than 35 million pounds, the history of the fishery could be explained on the basis that at some time during the decade (1906-15) when Area 2 annual catch was more than sustainable, the stock was driven below the level capable of yielding the maximum sustiiinable. Such an explanation would be consistent with the estimate from tagging that fishing-mortality in Area 2 around 1926 was about 40 'per cent per year (Thompson and Herrington, 1931), which implies an average fishable stock in 1926 of some 65 million pounds. Assuming for present purposes that catch per unit effort was proportional to stock-magnitude, the 1915 stock would have been more than twice that of 1926 or, say, 145 million pounds; whilst in 1905 the stock was presumably at least six times as great as in 1926 or, say, 400 million ' p~ds. The decline in the fishable stock, 1905-15, would then have amounted to some 250 million pounds. The sum of catches during the decade was evidently around 550 million pounds; while additions to the stock might well have been no more than 300 million pounds (assuming that the stock level for maximum rate of addition was passed through during the decade, and that the maximum addition amounted to no more than 35 million pounds per annum). Thus, the estimated decline in stock agrees with the estimated excess of catch over additions. tag-returns to remain (until grown) on the particular banks where first found. To obtain a representative rate. of tag-return in Area 2, therefore, the distribution of tags would have to be made proportional to that of the population rather than of the fishing-effort. The Area 2 fishing grounds extend along some 800 miles of coast, subdivided for statistical purposes into sectors 60 miles long. Thirteen sectors (No, 5-17) are wholly included in the Area. In both the 1925 and the 1926 tagging experi­ments, some 96 per cent of the Area 2 fish effectively marked were liberated within only four of these sectors (No. 10, 11, 13, 15; see ibid., Fig. 2 and pp. 10, 109­111). The fishery in these four .sectors where tagging was concentrated employed 60 per cent of the total effort in, Area 2 and yielded 63 per cent of the total catch (in 1928-29; ibid., Table 5 and pp. 52-54). Data indicating what proportion of the fishable halibut population of Area 2 occupied the four sectors of concentrated tagging are not available. For present purposes and with appropriate reservations, it might be assumed that the area populated by halibut per linear sector, and the rate of.addition to the stock _per unit area populated, averaged about the same in these four sectors as in the nit¥ where little tagging was done; and that the average density of fishable stock on the populated area in a sector was about proportional to the average catch per unit effort there. 9 On these admittedly questionable assumptions, the four sectors with 96 per cent of the tags would have contained only 34 per cent of the fishable Area 2 population. This presumptive 34 per cent of the population yielded 63 per cent of the catch. Consequently, the annual rate of fishing mortality among the tagged stocks may have been more than three times that on the rest of Area 2; and nearly double that for Area 2 as a whole. 9Against these assumptions, it may be argued that those sectors of Area 2 with a more extensive area of banks or a more rapidly adding stock would (if the catch is the most important affect of catch per unit effort) be those where the fishery is concentrated. Fishing effort is presumably distributed among the sectors of an Area according to relative profitableness, which is presumably governed by the balance of numerous factors including, on the one hand, relative _value of the catch per unit effort, and on the other, distance from po-rt, hydrographic features, etc. Each sector would thus be expected to have a different relative attractiveness, even if value of catch per unit effort were equal on all. This locally characteristic non-biological attractiveness would change with development of new ports, a differently sized or powered fleet, better means of navigation, etc. But provided relative non-biological attractiveness changed less than did relative value of the sector's catch per unit effort, and provided non-biological factors were not over­whelmingly predominant in relative profitableness, those sectors on which abundance could ~ least rapidly reduced by a given level of sustained catch might be expected to become those of concentrated fishing effort. However, although there is indicated to be some positive correlation between catch or effort and catch per unit effort on the secto-rs of Area 2 (according to data for the period 1928-29 given by Thompson and Herrington, 1930, Table 5 and pp. 52-62), this correlation is not great enough to be significant (rcf =+ .31, with a I in 3 expectation of a better result by chance alone; rcc = +.39, with a 1 in 5 expectation by chance). Since the reported range of catch per unit effort on the different secto·rs of Area 2 was fairly considerable, from 34 pounds . to. 53 . pounds, the poor correlation with effort suggests that other factors than catch per unit effort may have had a predominant influence on the distribution of effort among the sectors (a con~ elusion in agreement with the indication from the same Table 5 that, on the sectors of Area 3, there was no positive correlation at all between catch or effort and catch per unit effort, despite a range in the latter from 62 pounds to ll4 pounds). Thus, there is no reason to think that the most heavily fished sectors of Area 2 in 1926 were necessarily those with the largest populations or fastest rates of addition. A direct investigation of populated area and average population density in different sectort of Area 2, will probably be required to settle conclusively the question here raised. Thus, if the fishing rate among the immature Area 2 stocks tagged in the 1920's was 40 per cent/0 the overall rate on Area 2 at that time might have Been only some 20 per cent. This would greatly reduce the mentioned discrepancy between ratio of estimated stock magnitude to catch pet unit effort on Areas 2 and 3. Possible con­firmation that overall Area-2 rate of fishing mortality in 1926 was lower than has been thought is offered by the unexpected! y sparse returns . said to have resulted from the still incompletely reported tagging experiments performed on Area 2 adults in 1939-l940.U 1 If the annual fishing-mortality rate on Area 2 around 1926 was only 20 per cent or less, the decline in Area 2 population before 1930 might evidently have been too great to be accounted for by the difference between the catch removed and the additions to the stock which would have been expected had natural conditions been as favorable as recently (compare footnote 8). This discrepancy is, however, at best much less striking than the comparable one which has been demonstrated for Area 3; and so merely provides encouragement, rather than serious support, for the working hypothesis that fishery-independent changes have strongly affected abundance on Area 2 as well as Area 3. Casting about for further promising leads, a purposively biased mind might be struck by the remarkable parallelism in changes in catch per unit effort on Areas 2 •and 3, during the whole 34-year period for' which records are available for both stocks (c/. Thompson, 1950, Fig. 1, 2). The similarity holds over a great range of change, not only for general trend but for the year-to-year steps (differences in 10Ricker ( 1945, p. 88) · points out that this 40 per cent rate is based on the instantaneous rather than the annual corrected rate of tag-return, which latter would be 1-e-...0=0.33 (although, if the average stock had not been declining, the ratio of annual catch to average stock would evidently approximate the instantaneous rate of tag return). He also notes that the particular extrapolation performed by Thompson and Herrington results in a rate theoretically somewhat too high. Ricker (1948, pp. 86-91) points out that the vulnerability of the fish tagged averaged too low; and also that tag-loss might not have bf:en negligible as thought by Thompson and Herrington (1930), in which case upward revision of the postulated fishing mortality rate would be required (however, the evidence, ibid., that strap-tags were rarely lost seems fairly satis­ factory). Finally, thf:· decline with time in return of tags from the 1926 experiment seems to have been less than that in the 1925 experiments from which the 40 per cent rate was calculated (Thompson and Herrington, 1930, Table 9) ; though fishing-effort was increasing. Taken altogether, it is difficult to judge just what the ratio of annual catch to average &lock-magnitude during the year may have been among the Area 2 fishable stocks on which. the tagging experiments were performed, but somewhere between 30 per cent and 50 per cent seems likely. 11These mature fish were recovered from all parts of Area 2, in a fashion like that on Area 3 in the 1920's (Int. Fish. Comm., 1949). The low rate of return has been interpreted (ibid.) as meaning that "the rate of return from tagging on spawning congregations is not representative of the population as a whole," and, as pointed out to me by Dr. W. E. Ricker (in litt.), the fishing season in Area 2 had indeed bt-come so limited that in 1940 the adults mav not have returned from the spawning grounds in time to become fully vulnerable to fishing. . However, although such reduced vulnerability would invalidate the present interpretation of the 1940 experiments, it is a special feature of the recent situation which does not seem to bear upon interpretation of the results of the tagging experiments in the 1920's. Objection to the present view that rate of tag-return from adults on Area 3 was representative might of cour~e be made on other grounds; for example, adults dispersing from spawning grounds may home to their banks of origin. Heavily fished juvenile stocks might be expected to contribute relatively fewer adults to the spawning populations. Hence, if the spawners afterwards home, and if they spend much of their time on the home banks rather than in transit to and from the spawning grounds, tagging on the Area 3 spawning grounds in the 1920's might have yielded dispropor­tionately low returns. However, since evidence of homing would presumably be conspicuous among records of halibut tagged while young and recovered after maturity, the lack of report that homing occurs suggests that it does not. Since the foregoing was set in print, Report 15 of the International Fisheries Commission, for 1949 (1951) has been received. On p. 16, it states that "Much biological evidence was and is at hand to demonstrate the relative lack of interchange of stocks between subsections of Area 2· direction of which result from only five occasions of out-of-line change) ; and it is as about as marked during the 16 years before regulation began in 1931, a8 in the 18 years afterward. Coefficient of correlation of catch per unit effort on the two Areas is (by rank difference, r5 ) +.88 for both the pre-and the post-regul~tion years. The closeness of the relationship is strikingly brought out by plotting catch per unit effort in one Area against that in the other, for each year (Figure I). It will be seen, from the remarkably compact distribution of the points around the regression line drawn in Figure I, Relation of each year's average catch per unit effort in Area 3 of the Pacific halibut fishery to that in Area 2, during the period 1915-1946. Data from Thompson (1950) and International Fisheries Commission (1948). Regression Y=1.85 X -1.5; s7,x=19 lb. Years prior to regulation of the fishery (1915-1930) denoted by o; years after regulation began (1931-1946) denoted by +· that relative apparent abundance in one Area can he predicted quite closely (per­haps even within the limits of variation in ratio of catch per unit effort to actual abundance) from that in the same year in the other Area. Expressed in another way, although the ratio of annual average catch per unit effort on Area 3 to that on Area 2 has a fairly considerable range of variation (34-year mean, 1.8±0.3), this ratio shows no very significant trend (Table I, Figure II). D~ring the years before regulation, 1915--1930, the ratios ranged from 1.4 to 2.6, mean 1.9; during the ... Tagging experiments on spawning grounds in Area 2 showed that the mature fish migrated within relatively restricted sections of the area adjacent to the place of marking. . . ." It is impossible to interpret this flat contradiction of Report 14 for 1948, which states (p. 21) that "The Cape St. James experiment in the winter of 1939-1940 corroborated the belief that Cape St. James was a spawning congregation point for the Area 2 stock as a whole. •.• As in the case of the Cape St. James experiment, recoveries [from similar winter tagging on grounds between Hippa Island and Cape Spencer] were made throughout Area 2." This inexplicable conflict, con­cerning a matter which is not only of theoretical interest but which affects a proposal for change of the regulations, seems to point up the Commission's neglect of its obligation to submit details of its evidence for public inspection. 18 years after regulation, from 1.3 to 2.1, mean 1.7. For 1915-1922, the mean is 1.8; 1923-1930, 1.9; 1931-1939, 1.7; 1939-1948, 1.7. ,Such a relationship suggests as a possibility that catch per unit effort is governed by the same factor on both Areas. Against this, it must be granted that even if abundance on Area 3 had been heavily influenced by factors independent of fishing (as indicated by preceding analyses), Area 2 might simply have been kept fished down to the same level of profitableness as Area 3. But if this had been the case, the ratio of the fishing-efforts required to maintain the parallelism of catch per unit effort on the two ,grounds would have depended on whether or not the two stocks were equally over-or under-fished. If the Area 3 stock were above the level required for maximum equilibrium yield, the catch required to prevent a unit increase in Area 3 stock-magnitude would be expected to become less the greater the stock. If the Area 2 stock were below the level for maximum equilibrium yield, then the greater this stock became the greater the catch which would have been · required to prevent further increase in stock magnitude. Thus, since the changes in catch per unit effort were parallel, the ratios of catch or effort ·On the two Areas might provide a clue to the situations of the stocks in relation to their respective inflection points. The catch and the effort on Area 3 tended through 1922 to decline somewhat. re~ative to those on Area· 2 (Table I; Figure II). In 1923-1925, however, these ratios showed a steep rise after which a relatively high, quite constant level has been maintained. The correlation between these parallel ratios of catch and effort -< ~ ,,30 " 1'115 :Z.50,. 0.50 o.rs 1.00 ~1.50 ~.00 'RATio, AREA~ FIGURE II Ratios of effort, catch and catch per unit effort on Pacific halibut Area 3 to those on Area 2 in the same year, by years during the period 191~1946 (see Table I). Ratios of effort denoted by a dot; of catch by a circle; of catch per unit effort by an encircled dot. on the two Areas is high (r• == + .82), but neither ratio alone shows any clear simple relationship to the ratio of catch per unit effort on the two Areas.12 TABLE I Ratios ~tween catch, effort and catch per unit effort on Pacific halibut Area 3 and those on Area 2 (from data presented by Thompson, 1950, and International Fisheries Commission, 1948, 1949). Ratio, Area 3/Area 2 Ratio, Area 3/Area 2 Year .C~ttch Effort Catch/effort Year -Catch Effort Catch/efl'ort 1915 0.53 0.23 2.25 1932 0.90. 0.55 1.64 6 0.62 0.35 1.76 3 0.95 0.59 1.62 7 0.54 0.28 1.92 4 1.05 0.68 1.56 8 9 0.39 0.46 0.28 0.29 1.42 1.58 35 6 1.08 1.10 0.75 0.68 1.44 1.62 20 0.40 0.23 1.76 7 1.12 0.61 1.84 1 0.39 0.21 1.83 8 1.12 0.66 1.68 2 0.34 0.16 2.14 9 1.04 0.55 1.88 3 4 0.74 0.94, 0.28 0.48 2.63 1.96 40 1 1.06 1.17 0.56 0.59 1.89 2;()() 25 1.09 0.59 1.82 2 1.16 0.57 2.05 6 1.01 0.56 1.81 3 1.15 0.63 1.84 7 1.22 0.69 1.78 4 1.05 0.59 1.78 8 9 0.99 ].15 0.65 0.64 1.52 1.83 45 613 1.17 1.14 0.72 0.82 1.63 1.40 30 1.18 0.64 1.86 713 1.02 0.71 1.42 1 0.93 0.53 1.76 813 1.01 0.77 1.31 This presumably means that the short-term fluctuations in ratio of catch per unit effort were. of no significance for the present line of reasoning. Therefore, on 12There is some slight correlation between ·the catch or effort ratio and the ratio of catch per unit effGrt four years earlier; and r. for the effort ratio may be brought up to +.47 (with rejec· tion of the null hypothesis around the 1 per cent level) if the effort ratios of the early period :tre raised in relation to those after 1927 by adding the quantity 0.35 to each. Such weighting of the early ratios might be justified on the ground that the sharp change in effort ratio from the level of 1915-1922 to that of 1928-1948 suggests an economic change which greatly increased the relative attractiveness of Area 3 (presumably the rapid changeover to diesel power after 1922; see Thompson and Freeman, 1930, pp. 46-7). However, the. interpretation of such a delayed relationship would not be clear even if it were real. Four years seems rather long for lag in fishermen's rtaction to changes in relative attractive­ness resulting from changes in catch per unit effort. 13There is considerable confusion in the statistics of the halibut fishery, some of which seems unnecessary or even deliberately fostered. The values for 1946--48 used in obtaining the ratios given in Table I above are especially uncertain. . Without going into details of minor conflict, it seems desirable to state that the following caches per skate have been arbitrarily chosen here among the alternatives which Thompson ( 1950) and the Commission's Reports for 1947 and 1948 permit: Area 2, 1946, 86 lb.; 1947, 86 lb.; 1948, 93 lb.; Area 3, 1946, I20 lb.; 1947, 122 lb.; 1948, 122 lb. Concerning the values for Area 3 in 1947 and 1948, some further remarks seem justified. The Report for 1947 gives "size of the stock" for Area 3 as 89 per cent greater than in 1930, which would evidently represent 1.89 X 64.7=122 lb. per standard skate. The Report for 1948 sayS that Area 3 abundance "was the same as in 1947"; but adds that "More detailed studies ... indicate that the catch per unit effort in Area 3 has increased 137 per cent from 1930 to 1948". This would naturally be taken to mean that the Report for 1947 had been found in error and that the catch per skate in 1947 and 1948 was actually 2.37 X 64.7=153 lb., slightly more than the 1944 level. This value was consequently used in preceding drafts of the present paper, and it was not until the Commission's Report for 1949 ( 1951) was received, after the present paper was in galley-proof, · that the true state of affairs was grasped. The Report for 1949 (p. 21) remarks that on Area 3 subsequent to 1944 "a gradual recession occurred.... By 1944 . . catch per skate had declined to about llO pounds...." The Report goes on to say in a footnote that "on account of the marked. changes in the fishery and the complex character of the stocks" the Area 3 index of comparison with 1930 requires modifications which would lift the value for 1949 from an uncorrected 70 pet cent to "more than 100 per cent above the 1930 level." It thus appears that the "139 per cent" the hypothesis that fishing was the principal affect of catch per unit effort on at least one of the Areas, the lack of strong trend1in the ratios both of catch per unit effort and of catch or effort in the period 1925-1948 seems to imply that the stocks on the two Areas were in a comparable situation as regards their respective levels for maximum equilibrium yield. This argument is not a conclusive one (since, for example, the slopes of the population-growth curves of the two stocks might have chanced to differ in just t4e manner required to offset the difference in relation of the stocks to their levels for maximum equilibrium, during the period of observation). It does however afford a simple explanation of the peculiar fact that not only the changes in catch hut the effects of these changes on the rate of natural increase of the stocks have been similar (or insignificant) on both Areas. If it is assumed, in accordance with the foregoing, that the stocks of Areas 2 and 3 must (unless governed primarily by a common natural factor) have maintained a comparable relationship to their respective levels for maximum equilibrium yield, the question still remains open whether they were both simultaneously above or both simultaneously below this level at any given time. The statistics of the Area 3 fishery indicate that natural changes affected abundpnce so much more than did fishing as to leave no secure basis for belief that this stock has ever been over­ fished. Therefore, it seems possible that the Area 2 stock may not have been driven below the level capable of maximum sustainable yield either. This is not to assert that Area 2 could have sustained a catch of 50-60 million pounds per year even under the most favorable conditions, but merely to question whether the period of high catch was sufficiently long-sustained to reduce the stock below its level for maximum equilibrium yield.H given in the Report for 1948 must have referred to the same 122 lb. per skate as did the "89 per cent" in the Report for 1947. In other words, the higher and more favorable-sounding value given for 1948 was obtained by a change in method of calculation introduced without explicit notice; and means to compare it with other years than 1930 are still lacking .two years later. Since nothing is said in the Report for 1949 concerning the presumably similar need for correc­ tion of the raw catch per skate on Area 2 (where there has apparently been no marked decline as yet, since the 1949 catch per skate is given as "over 90 pounds"), it is difficult to believe that the Commission has been concerned to clarify the Area 3 situation. Details of the method of correc­ tion which has been adopted will be of interest, since such operations are notoriously difficult. The Report for 1949 {p. 19) also introduces an "abbreviated" table of catches in Areas 2 and 3, which "differ from the reported catches, published elsewhere, as they include corrections for poundage reported from the wrong area and for commercial poundage estimated to have been landed but not reported." This Table selects so few years as not to be usable for present purposes; and since those corrections which are given are not large enough to make a significant difference in present results, they have been ignored. However, it seems worthwhile to emphasize here that when serial data are corrected or amended, it is customary to make every effort to facilitate full comparison with values previously offered. . 14Thompson and Herrington (1930, p. 7l) have calculated from the results of the tagging experiments on Arta 2 in the 1920's that the rate of natural mortality for the preponderantly · immature and relatively heavily fished stocks tagged was considerably greater than the reported rate of individual -growth given by Thompson and Bell (1934). This estimate of natural mor­tality rate can be questioned, and so (I am informed by Dr. Ricker) may be the estimated rate of growth; but if anything approaching this situation existed, it would be unexpected in an overfished stock of a species such as halibut. This would encourage the view that additions to the Area 2 stock during the 1920's might, if natural conditions had not been worsening at a rate sufficient to outweigh the favoring effects of declining population density, have risen to a level more than sufficient to balance the demand permitted by the rising costs of production. It is of interest, in the same connection, to remember the opinion of Thompson and Freeman (1930, p. 20) that virgin abundance on Area 2 was far greater than that on Area 3. The ratio of magnitudes of different stocks of a given species at their natural equilibrium levels might be expected (ceteris paribus) to be similar to the ratio of the maximum equilibrium yields of these different stocks. Therefore, if the Area 3 stock was able to increase by an amount probably a-round a hundred million pounds between 1940 and 1944, despite a yield of more than 27 million lb. per 'Y~r, it might be suspected that the decline of Area 2 abundance was considerably more rapid than The results of scrutiny of the Area 2 data from a viewpoint opposed in bias to that taken by the International Fisheries Commission thus justify the procedure, since they provide some reason to think that fishery-independent changes in Area 2 stock­magnitude might have been of importance. On the other hand, application of this procedure to the ?\orth Sea demersal fishery has not resulted in discoveries such as might require much modification of prevailing beliefs. Unlike the regulation of the Pacific halibut fishery, for which there is no control situation, the two great wars have provided what is, in effect, a controlled experi­ment in restriction of \"orth Sea trawling. The results of this repetition (which was independent of abundance and price) of a sudden large cut and a subsequent restoration of fishing effort prove that catch here has an immense influence on subsequent catch per unit effort, and one sufficient to drive the stocks below their critical levels. Suppose that the pre-1915 demersal stocks had been at or above the level for maximum equilibrium yield; and that therefore, as population grew denser .in consequence of war-restricted fishing, they would (ceteris paribus} have been unable to add to themselves yearly as much as the amount of the prewar catch. In this case, so great an absolute increase in the marketable stocks as is suggested on the basis of rate of prewar tag returns by the rise in catch per unit effort found in 1919 would not have been expected from the war-caused reduction in catch. But the increase during the second war, quadrupling the 1938 availability, \vas again to or beyond the 1919 level {and is thought to have resulted from improve­ment in survival rather than in growth-rate or broods; cf. Baerends, 1947; Graham, 1948) ; and resumption of fishing has again been followed by a sharp decline in abundance. Thus, an explanation of \"orth Sea data seemingly indicative of over­fishing in terms of fishery-independent natural change would require a highly improbable number of coincidences. Improbable occurrences are by no means to be treated as if impossible, and there are some few features of the· North Sea record which might favor suspicion that fishery-independent changes may have been more important than is usually grantedY the fishery would call for (unless the virgin ratio of stock-density to stock-magnitude on Area 2 was much greater than that on Area 3; which is not suggested by available data on relative extent of populated grounds in the two Areas). L>For example, there seem to have been rather high correlations between catch per unit effort in the New England trawl-fishery for adult haddock, and that in various parts of the North Sea demersal fishery, at least in certain periods (compare Herrington, 1948, with Thompson, 1930, and Baerends, 1947). In particular, for North Sea sail-trawlers, 1919-1935, correlation with Georges Bank is r. =+ .81. The meaning of these parallelisms has not been very carefully considered; hut they do suggest the possibility of influence on the North Sea stocks by a factor common to the whole North Atlantic. Whether this presumed common factor is economic or natural, and if natural how important relative to fishing, is not clear. SimultanEOus occurrence of dominant year classes on both sides of the Atlantic is known, but apparently not for haddock. Haddock abundance may, however, he subject to massive long-term natural change in New England waters (Burkenroad, 1948, pp. 114-5), and it is conceivable that a parallel trend might affect North Sea trawl catches. World-wide change in general economic conditions might cause parallel changes in North Sea and New England abundances through changes in fishing intensity; but the relationship of abundance of New England haddock to fishing is by no means clear (c/. Herrington, 1948, especially the discussion, pp. 279-81; and also Schuck, 1949. It will be observed that Schuck is mistaken in thinking that the high correlation between catch and decline in year-class abundance means that changes in the catch account for 66 per cent of the changes in decline. Instead comparison of catch and decline with initial abundance indicates that change in the latter probably to a considerable extent controls both the former. Likewise, Schuck's observation, that when the regression line of catch against decline is extrapolated it passes through the origin, does not mean that natural mortality is nil, but merely that if there were no haddock to start with, both catch and decline would be zero). A common economic factor might influence efficiency of the fishing However, these features are not very imposing ones, and apart from them, it seems necessary to believe that, either the 1914 and 1938 North Sea demersal stocks were below the level for maximum equilibrium catch and began to add themselves at a more rapid rate as population density and average age increased during the wars, or the stock-levels of 1919 and 1945 were greater than could be sustained under existent natural conditions, even if there had been no increase in fishing. There is some theoretical basis for entertaining consideration of the latter possi­bility, since a sudden great change in fishing mortality rate might result in an oscillating adjustment of the relations of predator to prey (see p. 193 above). However, even with benefit of bias against claims of North Sea overfishing, it is difficult to doubt that the present Nort~ Sea demersal fishery has so great an effect on abundance that it would be economically feasible to fish the stocks below the critical level in the near future, even if this has not already been done in the past. Thus, the practical question to which the possibility of North Sea fishery-dependent oscillation relates is the exceedingly intricate one of when or whether which com­ponents of the demersal fishery stock have already been reduced below their levels for maximum equilibrium yield. This question is not of the same gravity as the one about Pacific halibut raised in the present section (which latter is, essentially, whether fishery-independent changes have been so great as to make relatively insignificant the effects of the fishery) ; and it need not here be pursued. The foregoing indicates that predictions concerning effects of change in exploita­tion wn abundance of North Sea demersal stocks have some degree of reliability, whereas those for Pacific halibut are questionable. This difference does not seem to arise entirely from differences in the duration, extent and quality of the research devoted to these fisheries. Instead, it seems to result from the fact that regulation of the halibut fishery is an uncontrolled experiment, so that reasoning about its results is necessarily deductive. In contrast, the twice-repeated restriction and subsequent resumption of North Sea fishing caused by the wars permits use of i.J)ductive reasoning. Except for the wars, it would be difficult to be sure that the North Sea· had been overfished.16 eHort rather than abundance; but corrections for introduction of Vigneron-Dahl rig, radio, fath­ometer, etc., have evidently been made in the New England index (Herrington, 1948, pp. 250-254), and would presumably not be necessary for North Sea sail-trawlers. General economic conditions may of course also influence efficiency of fishing effort in ways fot" which correction is difficult. A fall in ratio of prices to charges on capital might affect both the real efficiency of the effort (which can be regarded as a function of the proportion of the fishable stock which is removed from a unit area by a unit effort) and the apparent efficiency (which could be treated as a function of the ratio between the average quantity of fish on the unit areas· actually fished and the average quantity per unit area over the whole range of the stock), by forcing the accomplishment of more (and therefore less selective) fishing upon the SUrvivors of the pinch. Considerable effects of general economic conditions on efficiency of effort have been recognized in some fisheries (e.g., Alaska razor clam; Thompson and Weymouth, 1935). However, there is no close correlation between catch per unit effort in the New England haddock fishery and some others in which efficiency of effort might be expected to have been similarly ~ected by economic changes (as for example, Pacific halibut, which although it too displayed its. lowest catch per unit effort at the beginning of the great depression, otherwise has a quite different record). Therefore, it seems possible that such affects were relatively unimportant. 16This statement rests upon the results of analysis of Graham's attempt to demonstrate over­fishing of North Sea cod through calculations from growth and age-census data (1935, pp. 206--269). It is the validity of the reasoning, not of the conclusion, which is here examined. Graham deduces that if the fishing rate for North Sea cod were sufficiently reduced to raise average age of the fish caught from 2.5 years to 3.5 years, the yield then sustainable would at least equal that obtained at the higher expenditure of fishing effort. The argument is based on calculations from the observed relative numbers of fish of different ages in the heavily fished In other words, use of inductive methods seems as yet to be required for assurance of biologically effectual exploitation. IV. The Inductive Method in Marine Fishery Biology. The penetrating study by· Sette (1943) of requirements for deductive approach to the problem of biologically effectual exploitation of California pilchard, and the relatively enormous expenditures which have been made for this investigation, suggest that the costs of adequate research along these lines will not soon be forthcoming for a majority of fisheries. At the same time, it remains to be seen whether the pilchard investigation will in fact succeed in identifying the inflection points, by whateyer elaboration of deductive methods rcf., Tester, 1949, on the necessity of experimental investigation of the effects of fishing on British Columbia stuck: the observed aYerage weights and awrage rates of individual growth of these fish at different ages; and ··some probable values'' for natural mortality at different ages. Graham assumes in his table of calculations that natural mortality and growth at tht different ages, and the number of young annually reaching fishable size, would all remain the same even if the densitv of the stock were increased. On .this basis, the ''rate of natural increa...~" would change only as did the age-composition of the stock. Graham's calculations show that, for a moderate change in average age of the stock, reduced mortality would thus more than compensate for slower growth. In order _to reduce the new equilibrium yield below that in the previous younger population, "either the change in natural mortality rate or the new natural mortality rate has to he quite ridiculous, from the information we now have ... we are only proposing to raise the average age from ZY2 to 3%, in a fish which is first mature at about 5 years of age, and it is really very unlikely that in these young fish the [average] natural mortality rate [for the whole fishable stock] would thereby be raised. In fact, we would expect it to fall. Yet if we make the change only zero, we find that the new mortality rate has to he as high as 41 per cent" (which is double the rate assumed even for the youngest fishable age-group in the heavily fished stock). However, the cun·es of rate of natural mortality and of growth against age would in reality he expected to he transposed bodily to new levels with increase of stock density (in conse­quence of increase in rates of cannibalism, disease, competition fer food, etc.), and these changes might also he expected to reduce the percentage of starting young which livt to reach fishable size. The question is, whether the curve of these various rate-changes against stock-density would rise steeply enough to offset the net gains from increasing age calculated by Graham. In the particular case, the observed effects of the wars do indeed make it likely enough that the net result of so modest an increase in density in a stock so young would be advantageous. But, as far as Graham's cited data go, there is nothing to oppose other assumptions. Suppose recruitment and weight, growth, and mortality of the different age-classes at Time 2 in Graham's Table were slightly modified from those at Time 1, enough to reduce G, from 0.65 to, say, 0.60. Then the ~It and Mz required if Y2 is to drop below Y1 become concordant enough with the new assumption that average natural mortality of the fishable stock would be somewhat increased at the new balance of a\·erage age and stock-density. In fact, Graham's assumptions are such as would make almost any heavily fished stock appear to he below the level for maximum equilibrium yield. In this sense, he has assumed the point at issue, and his argument is circular. What he has really done is not to prove overfishing by age and growth-census data, hut to construct one among a variety of models permitted by the data. Now suppose that, as a result of Graham's argument, regulation of trawling had been under­taken; with results more or less corresponding to those which he assumed would ensue. Such demonstration of insignificant change in other parameters than fishing mortality would have afforded a strong presumption that the demersal stocks really had been overfished. But still, the question would remain whether there might perchance have been a coincident natural change in, say, temperature, currents or phosphate f which, it should be noted, might cause parallel changes in a number of species-masses), to which a critical part of the improvement in "rate of natural increase" should have been credited. To exclude the possibility of coincident fishery-independent change, it would thus have been necessary, say, to relax the regulation&, then tighten them again, to see if the same results would be repeated. It is such conclusive evidence against independent changes which has been provided for the North Sea by the wars, and which in the present state of development of marine science can hardly be obtained by deduction, even with better data than Graham's, from observations on recruitment, growth and death (or catch and catch per unit effort either) in a narrow or uncontrolled range of circumstances (c/. Sette, 1943, p. 26). her~ing; also the remarks of Needler, 1948, pp. 170-1). As has been indicated on preceding pages· by analysis of demersal fishery situations, the fact that a stock is pelagic and is subject to relatively great natural mortality and variation in availability merely· enhances the basic difficulties of deductive approach, which seem to exist for almost all marine fishery problems. Nevertheless,· with or without an adequate basis, marine fisheries are going to be r~gulated. In many cases, a good part. of what is phrased as "conservation in the public interest" really concerns immediate advantages to limited special groups; in many other cases it is simply not realized that knowledge of the effects of the fishery is inadequate; in still others, the immediate risks from failure to regulate are considered graver than the possible long-term loss from misguided regulation. In all these cases, there are risks of indefinitely continued injury to society, because th~ form of the restrictions is not such as to provide assured evidence of their efiectua1ity and they may therefore be maintained indefinitely even though actually disadvantageous. Among politically active interests concerned with fishing, there is usually opposi­tion as well as favor toward restrictions, such that a compromise concerning changes in the status quo ante would often be politically feasible. Such a compro­mise could permit design of regulations in a form equivalent to controlled ex­periment.H Although the prop·osal by Tester in 1946 (unpublished) for an adequately controlled sea-fishery experiment (by means of an alternation between two years without restriction in the British Columbia herring fishery and one of complete closure, over a period of at least 12 years) was rejected in favor of a program which cannot be expected to yield as conclusive results (partial removal of restrictions on one local population while they are rigidly maintained for another), this pioneering proposal appears to have been a feasible one, economically and otherwise. It may he expected that as the principles involved become more familiar, good experimental design will be sought rather than resisted. To the extent that the more significant effects of exploitation are definitely known, the indirect costs to society of the use of fishery regulations for large-scale con­trolled experiment could not be justified. However, it is the present suggestion that (except where use of inductive reasoning has been made possible by accidents) available proofs of past or prospective benefits of regulation of marine fisheries 1 7For example, sportsmen, commercial fish-catchers, and many deep-water shrimp trawlers urge the closure of Texas bays to market-shrimping. Other trawlers, especially those with smaller boats, object. Knowledge of the effects of population-density of young shrimp on their. rates of growth and mortality is insufficient for assurance concerning the effects of nursery­ground trawling on subsequent abundance of adults. Although shrimp are (when available) an important item in tht: food of prized bay fish, and although the young of these fish are occasionally caught in bay trawling, it is uncertain just what would be the sum effect on fish of closure of the bays to trawling (at any rate, those who think closure will improve shrimp catches must to some extent contradict the opinion that it will improve fish catches). Mere closure of the bays (as has been done for some years past in South Carolina) will evidently not in itself yield decisive evidence of its effects. For example, recent high salinities in the Aransas area caused by a drought of several years' duration, are pro-bably involved in the sharp decline in abundance of young white shrimp; and closure of the bays to trawling will presumably at some point be followed by the end of the drought, with inextricable confusion of causes and effects. If opposition to Texas closure is sufficiently balanced against approbation, it should be possible to obtain support for a trial-and-error compromise consisting of regularly alternating years of opening and closure; which, if the results are properly observed, should in time provide statistically admissible evidence on which to base final action. often have an unfortunate degree of resemblance to the Butcher's proof of iden.tifi­ cation of the Jubjub (Dodgson, 1876). If this be true, it should be a leading principle with marine fishery biologists to explain and emphasize the need for use of sound inductive methods to test the effects of fishery regulations. The limitations of the experimental situation seem to require that control;; be provided by alternation of periods of relative freedom with those of restriction, to an extent and frequency (and on a schedule) permitting calculation of the probability that observed gains or losses correlated with restriction were caused by something else. ACKNOWLEDGMENTS Tolerance of my fumbling progress toward grasp of the principles of fishery biology and management has been actively displayed by a number of colleagues (as well as my students)~ to whom I wish to express my appreciation. Particular mention should be made of criticism, information, advice or encouragement received from Michael Graham, Gordon Gunter, A. G. Huntsman, G. W. Jeffers, Daniel Merriman, W. E. Ricker, G. A. Riley, A. L. Tester, E. F. Thompson and L.A. Walford (who are, however, in no way responsible for or necessarily in sympathy with views here presented) . The thinking here brought to focus was in part supported by grants made some years ago by the General Education Board of the Rockefeller Foundation to the Chesapeake Bay Fisheries Commission and to the Survey of Marine Fisheries of North Carolina. It is hoped that the present formulation may aid in attainment of the objectives for which these grants were made. It may not be necessary to emphasize that the concepts and data here used (and perhaps, though unintentionally, misused) have all had a long history of elaboration at the hands of a host of able men, and that the limited number of references cited has no relation to the extent of the sources to be acknowledged. SUMMARY l. a. The major present function of marine fishery biology is, to predict the effects of exploitation upon fishery stocks, in order to supply premises for the social and economic predictions upon which are based the value judgments repre­sented by management policy. b. The tangible net gains to society which are at present obtainable from pre­scriptive restriction of sea-fishing would apparently at best be relatively small; but the accompanying improvement in knowledge of interactions between man and . environment may have great importance. 2. a. The limits of potential marine harvest are in some regions probably strongly influenced by the rate of metabolic turnover of the store 9f limiting nutrient on hand in the system. In such katabolic ecosystems (as the North Sea) it is probable that fishing may reduce not only the standing crops of marketable fish, but also the ability of the system to support them, as a result of its indirect effect on respiratory rate of the biomass. b. A second type of indirect fishery-dependent effect which may yet prove to be of importance is, oscillatory adjustment of predator-prey relationships to sudden changes in fishing-intensity. It is suggested that the accuracy of extrapola­tions from popuiation-curves for North Sea demersal stock which have been based on observations of war-caused changes, may have been significantly affected by such oscillation. 3. a. It is impossible to question seriously the conclusion that North Sea demersal stocks have been or can easily be driven by commercial fishing below the levels for one or both biological inflection points in the curve of diminishing returns from increased fishing. / In contrast, however, serious doubt of current beliefs concerning effects of exploitation on Pacific halibut is legitimate, since data for the latter do not exclude the possibility of major effects from fishery-independent changes in the environment." b. It therefore seems possible that effort saved through management of the Pacific halibut fishery has been more than matched by 'countervailing sacrifice of catch. The difference in reliability of predictions concerning North Sea demersal and Pacific halibut fishery stocks appears to rest upon use of inductive methods in study of the former fishery, an approach there permitted by the accidents of war. 4. a. It is concluded that the degree of accuracy of biological predictions attain­able by use of the usual purely deductive methods is unlikely to he sufficient to permit reliable socioeconomic predictions. b. Regulations which are to be applied to marine fisheries should therefore as a matter of principle be designed in form equivalent to controlled experiment, hy proVision for arbitrarily scheduled periodic relaxations of restriction sufficient in extent and number to permit statistical analysis of the results. REFERENCES Anderson, A. W., and E. A. Power. 1946. Fishery Statistics of the United States, 1942. Stat. _ Dig. U.S. Fish Wildlife Serv., 11:1-248. Baereiids, G. P. 1947. (Translated, 1950, from Pap. Dept. Fish. Netherlands Min. Agr. Fish Food Supply, 36, as: . The rational exploitation of the sea fisheries with particular reference to the fish stock of the North Sea. Spec. Sci. RepL: Fish., U.S. Fish Wildlife Serv., 13:1-103.) Burkenroad, M. D. 1946a. General• discussion of problems involved in starfish utilization. Bull. Bingham Oc. Coli., 9(3) :44-58. ----.. 1946b. The development of marine resources in Indonesia. Far Eastern Quart., Feb., 1946:189-99. ----.. 1948. Fluctuations in abundance of Pacific halibut. Bull. Bingham Oc. Coli., 11 (4) :81-129. ----.. 1950. Population dynamics in a regulated marine fishery [Review of Thompson, 1950]. Texas J. Sci., II, 3:438---41. Chapman, W. L. 1942. The latent fisheries of Washington and Alaska. California Fish Game, 28(4) :182-198. Cooper, L. H. N. 1948. Phosphate and fisheries. J. Mar. Bioi. Assn. U. K., 27:326-36. Dnnlop, H. A. 1937. Why are there separate areas? Circ. Int. Fish. Comm., 5:1-5. Graham, M. 1935. Modern theory of exploiting a fishery, and application to North Sea trawling. J. Cons. Perm. Int. Expl. Mer., X, 3:264-74. ----. 1948. Scientific meeting on the effect of war on the stocks of commercial food fishes-Proceedings. Rapp. Proc-Verb. Cons. Perm. Int. Expl. Mer., 122:5-6. ----.. 1949. The Fish Gate. (2d ed.). Latimer Trend & Co., Plymouth, 199 pp. Gunter, G. 1951. The import of catastrophic mortalities for marine fisheries along the Texas coasL J. Wildlife Man. (in press). ijarvey, H. W. 1948. The estimation of phosphate· and of total phosphorus in sea waters. J. Mar. Bioi. Assn. U.K., 27:337-59. -:-----.. 1950;. On the production of living matter in the sea off Plymouth. Ibid., 29:97-137. Herrington, W. C. 1948. Limiting factors for fish populations, some theories and an example . . Bull. Bingham Oc. Coli., XI, 4:221-79. International Fisheries Commission. 1948. Regulation and investigation of the Pacific halibut fishery in 1947. Rept. Int. Fish. Comm., .13:1-35. · . 1949. Regulation and investigation of the Pacific halibut fishery in 1948. Ibid., 14:1-30. Kask, J. L. 1937. Halibut tagging experiments. Circ. Int. Fish. Comm., 6:1-5. Kesteven, G. L. 1950. Essay heview. j. Cons. Perm. Int. Expl. Mer., XVI, 2. Margetts, A. R. 1949a. Experimental comparison of fishing capacities of different trawlers and trawls. Rapp. Proc.-Verb. Cons. Perm. Expl. Mer., 125:72-81. ----. 1949b. Experimental comparison of fishing capacities of Danish seiners and trawlers. I bid., 125 :82-90. Margetts, A. R., and S. J. Holt. 1948. The effect of the 1939-1945 war on the English North. Sea trawl fisheries. I bid., 122:27-46. Merriman, D., and H. E. Warfel. 1948. Studies on the marine resource!'> of southern New England. VII. Analysis of a fish population. Bull. Bingham Oc. Coli., ll ( 4) :131-64. Needler, A. W. H. 1948. Estimating fishing intensities. Bull. Bingham Oc. Coli., XI, 4:16~73. Parrish, B. B. 1948. The haddock stocks in the North Sea during the second half of 1945. Rapp. Proc.-Vcrb. Cons. Perm. Int. Expl. Mer., 122:47-54. ----. 1949. Fishing capacities of gears used by E;nglish and Scottish research vessels. Ibid., 125:91-6. Ricker, W. E. 1945. A method of estimating minimum size limits fer obtaining maximum yield. Copeia, 1945, 2 :84-94. ----. 1948. Methods of estimating vital statistics of fish populations. Indiana Univ. PubL Sci. Ser., 15:1-101. Riley, G. A. 1938. The significance of the Mississippi River drainage for biological conditions in the Northern Gulf of Mexico. J. Mar. Res. Sears Found., I (l) :60-74. ---·-. 1949. Food from the sea. Sci. Amer., Oct., 1949:16-19. Riley, G. A., and R. von Arx. 1949. Theoretical analysis of seasonal changes in the phyto plankton of Husan Harbor, Korea. J. Mar. Res. Sears Found., VIII, 1:60-72. Riley, G. A., H. Stommel, and D. F. Bumpus. 1949. Quantitative ecology of the plankton of the Western North Atlantic. Bull. Bingham Oc. Coli., XII, 3:1-169. Schuck, H. A., 1949. Relationship of catch to changes in population size of New E:1gland haddock. Biometrics, 5 ( 3) :213-31. Sette, 0. E. 1943. Studies on the Pacific pilchard or sardine (Sardinops caerulea). I. Structure oi a research program to determine how fishing affects this resource. Spec. Sci. Rept. U.S. Fish Wildlife Serv., 19:1-27 (reprinted, 1950, Spec. Sci. Rept. U.S. Fish Wildlife. Serv.: Fisheries, 15: 1-30) . Sverdrup, H. U., M. W. Johnson, and R. H. Fleming. 1942. The oceans, their physics, chemistry and general biology. Prentice-Hall, Inc., New York. 1087 pp. Taylor, H. F. (and associates). 1950. Survey of marine fisheries of North Carolina, w_ith a comprehensive view of the economic:; of national and world fisheries. Univ. -North Carolina Press, Chapel Hill. 555 pp. Tester, A. L.. 1949. '1 he efu<:acy of catch limitations in regulating the British Columbia herring, fishery. Trans. Roy. Soc. Canada, XLII Oil, 5 I :135-63. Thompson, D'A. W. 1930. On fluctuations in the abundance of cod. Rapp. Proc.-Verb. Cons. Perm. Int. Expl. Mer., 65. Thompson, S. H., and F. W. Weymouth. 1935. Conditions of razor clam fishery in vicinity of Cordova, Alaska. Invest. Rept. U.S. Bur. Fish., I, 29:1-14. Thompson, W. F. 1950. The effect of fishing on stocks of halibut in the Pacifi::-. PuU. Fish. Res. lnst. Univ. Washington; Univ. Press, Seattle. 60 pp. Thompson, W. F., and F. H. Bell. 1934. BioJo_:ical statistics of the Pacific halibut fishery. (2). Effect of changes in intensity upon to:.Ul yield and yield per unit of rear. Rept. l!lt. Fish. Comm., 8:1-49. Thompson, W. F., H. A. Dunlop, and F. H. Bell. 193f. Biological statistics of the Pacific halibut fishery. (l). Changes in yield of a standardized unit of gear. Ibid., 6:1-108. Thompsc.n, W. F., and N. L. Freeman. 1930. History of the Pacific halibut fishery". Ibid., 5:1-61. Thompson, W. F., and W. C. Herrington. 1930. Life history of the Pacific halibut. (l) Marki~ experiments. Ibid., 2:1-137.