UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Behavioral transformation from the planktonic larval stage of some marine fishes reared in the laboratory Marliave, Jeffrey Burton 1975

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


831-UBC_1976_A1 M37.pdf [ 19.06MB ]
JSON: 831-1.0093811.json
JSON-LD: 831-1.0093811-ld.json
RDF/XML (Pretty): 831-1.0093811-rdf.xml
RDF/JSON: 831-1.0093811-rdf.json
Turtle: 831-1.0093811-turtle.txt
N-Triples: 831-1.0093811-rdf-ntriples.txt
Original Record: 831-1.0093811-source.json
Full Text

Full Text

THE BEHAVIORAL TRANSFORMATION FROM THE PLANKTOHIC LARVAL STAGE OF SOME MARINE FISHES" REARED ISi THE LABORATORY by JEFFREY BURTON! MARLIAVE BiSc.Flsh., University of Washington, 1970 A THESIS SUBMITTED IN: PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in the Department of Zoology We accept this thesis as conforming to the required standard THE - UNIVERSITY OF BRITISH COLUMBIA September, 1975 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . It i s u n d e r s t o o d that c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Depa rtment The U n i v e r s i t y o f B r i t i s h Co lumbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 i i Abstract This study tested the hypothesis that larvae of benthic species settle only on a substrate which is typical of the adult habitat and that those which f a i l to encounter a suitable substrate will not settle and will die before completion of metamorphosis. A similar hypothesis, that larvae of schooling species will f a i l to metamorphose i f they cannot express adult schooling behavior, was tested by rearing them in isolation through metamorphosis. The range of substrate preferences for settling larvae of intertidal species corresponded to the adult niche breadth. Adults of Artedius lateralis occur in most shallow water habitats on the open coast. Their larvae settled on any substrate offerred. Adult Xiphister atropurpureus live only among pebbles and small rocks on cobble beaches. Their larvae settled only in gravel. Pholis laeta preferred to settle in eelgrass, Bbthragonus swani on gravel and Leptocottus armatus on sand. These species settled directly onto the substrates most characteristic of the adult habitats. Larval Gobiesox maeandricus. however, settled onto smooth seaweeds rather than onto the rocks inhabited by adults. Thus, the preferred substrate for larval settlement was always an element within the adult habitat, but not necessarily the same substrate preferred by the adults. This study demonstrated that metamorphosis does not coincide with the behavioral transformation. Larvae of schooling species adopted adult schooling behavior well before the onset of their gradual metamorphosis. Aulorhynchus flavidus left the plankton i i i and started to school just after hatching. In intertidal species, particular metamorphic changes occurred before the settlement period, which took place more or less abruptly. Immediately after settlement, metamorphic pigmentation occurred rapidly, followed by allometric growth. One benthic species which does not tend to be sedentary as an adult (Blepsias cirrhosus), meta-morphosed completely before settling onto kelp. Considerable variability in the pattern of behavioral transformation and meta-morphosis was witnessed between species. Prevention of schooling (by isolation) did not hinder meta-morphosis of Clupea harengus pallasi or Aulorhynchus flavidus. Only in those benthic species having abrupt settlement and narrow substrate preferences was there a tendency for high mortality rates at transformation. Such mortalities took place only when suitable substrates were unavailable for settlement. The positive phototactic tendencies of larvae changed during settlement only in species which settled into a shaded substrate (Xiphister atropurpureus). Physical characteristics of substrates resulted in the settlement preferences, which are probably based on tactile cues and properties of light transmission. Observations of two species under varying current velocities indicated that species-specific preferences for flow rates also influenced settlement. A subtidal benthic species, Gilbertidia sigalutes. alternated between larval and adult behavior for a long period, during which their vertical migrations became increasingly nocturnal. This repeated settlement and reentry to the planktonic drift would iv maximize the effectiveness of sampling the bottom to locate the rare adult habitat (protected rocky crevices). Species in which the larval and adult habits differ to a great extent usually have an abrupt behavioral transformation which immediately precedes final stages of metamorphosis. Trans-formation of these species occurs only upon encounter with the adult habitat; the probability of such encounter reflects the mortality risk for this stage. Exceptions to these generalizations, such as G. sigalutes or A. flavidus. tend to have either unusually extended or, at the other extreme, unusually brief larval stages. V Table of Contents Page Number Abstract i i Table of Contents . v List of Tables x List of Figures • xi Acknowledgments x i i i I. Introduction * • 1 II. Experiments and Field Observations A. Preface 12 B. Materials and Methods 1. Experimental Animals • IA 2. Rearing • . 16 3. Substrate Preference Studies on Intertidal Fishes . . . . .17 4. The Vertical Migrations and Transformation of Gilbertidia slgalutes . . . . 22 5. Rearing Larvae of Schooling Species in Isolation • • 2b C. Substrate Preference Studies on Intertidal Fishes 1. Artedius lateralis. F. Cottidae 26 2. Blepsias clrrhosus. F. Cottidae, • • • • • • 2 9 3. Leptocottus armatus. F. Cottidae . . . . . 3 3 k, Nautichthys oculofasclatus. F. Cottidae 37 5. Bothragonus swani. F. Agonidae . • 40 6. Xiphister atropurpureus. F. Stichaeidae kz ?• Pholis laeta. F. Pholidae & 8. Gobiesox maeandricus. F. Gobiesocidae . . . . . . . .60 D. The Vertical Migrations and Behavioral Transformation of Gilbertidia sigalutes 1. Adult Habitat 6*t 2. Larval Stage . . . . . . 66 3. Transformation Stage • • • • • • 72 E. Schooling Species 1. Introduction . . 81 2. Clupea harengus pallasi. F. Clupeidae . 82 3. Aulorhynchus flavidus. F. Gasterosteidae 84 III. Discussion A. Intertidal Fishes . . . . . . . . . ... ^ ...,.c. . . 88 B i Gilbertidia sigalutes. F. Cottidae 105 C. Schooling Species 1. Clupea harengus pallasi. F. Clupeidae . 113 2. Aulorhynchus flavidus. F. Gasterosteidae 116 D. Behavioral Transformation in Relation to the Ecology of Marine Fish Larvae . . . . . . . . . . . . 120 E. Suggestions for Future Research . . . . 1 2 6 IV. Summary . . . . . . . . . . . . . 128 Literature Cited 131 Appendix 1. The Rearing of Marine Fish Larvae A. Introduction.. . . . . . 1 3 9 B . Obtaining Material 142 C. Maintaining Egg Masses 148 v i i D. Tanks . . . . . . . . . . . . . . . . . . 153 E. Water Supply 160 F. Lighting . . 164 G. Density Effect 168 H. Feeding 1. General . . . .170 2. Plankton Towing 175 3. Culturing Artemia salina Nauplii 181 Appendix 2. Annotated Illustrations of HE Pacific Fish Larvae of the Neritic Plankton Preface. 184 F. Clupeidae 1. Clupea harengus pallasi ... . . . . . . . . •• . 185 F. Gobiesocidae t , ;\ 2. Gobiesox maeandricus . . . . . . 1 8 7 3. Rimicola musearum • 190 F. Gasterosteidae 4. Aulorhvnchus flavidus 190 F. Clinidae 5. Gibbonsia montereyensis 192 F. Stichaeidae 6. Anoplarchus purpurescens 19^ 7. Xiphister atropurpureus . . . . 197 8. Xiphister mucosus . . . . . . . . 198 F. Pholidae 9. Apodichthys flavidus 200 v i l i 10. Pholis laeta 200 11. Xererpes fucorum 202 F. Anarhiehadidae 12. Anarrhichthys oeellatus . . . 204 F. Gobiidae 13. Coryphopterus nicholsi . . . . . . 204 F. Hexagrammidae 14. Hexagrammos decagrammus 207 15. Ophiodon elongatus 207 F. Cottidae l6.i>'Artedius fenes.tralis 210 17. Ar ted jus lateralis . . 210 18. Blepsias cirrhosus 212 19. Enophrys bison • 214 20. Gilbertidia sigalutes 214 21. Leptocottus arraatus • . . ... . . . 217 22. Nautichthys oculofasciatus . . . •• 217 23. Psychrolutes paradoxus • 219 24. Rhamphocottus richardsoni . . . . . . . . . . . 222 25» Sc orpaenic hthys marmoratus 222 F. Agonidae 26. Bothragonus swani 224 27. Pallasina barbata 224 28. Xeneretmus latifrons 224 F. Cyclopteridae 29. Liparis fucensis • • 228 Appendix 3. Summary of a l l plankton tows made during the winter of 1973/1974 • .- • . . . . . . . . .230 Appendix 4. Catch per unit effort of plankton-tows made during the winter of 1973/1974 (numbers of Gllbertidia  sigalutes per tow, by half month periods). • • . .... 231 Hist of Tables x Table I1.. Species of marine fish larvae which have been reared in the laboratory from hatching through metamorphosis to the juvenile stage. . . . . . . . . . . . . . . . . . . page 5 Table IT. Experimental Organisms:rseason of availability, methods of collection and incubation technique. ...... page 1 5 xi Figure 1. Map of Barkley Sound with detail of Bamfield area 13 Figure 2. Positions of early metamorphic (large pectorals -dotted) and late metamorphic (small pectorals - open) stages of Nautichthvs oculofasciatus exposed to different current speeds. . . . . . . . . . . . . . . . . . . . . 39 Figure 3. Size composition of Xiphister atropurpureus collected in horizontal transect on Helby Island. Tide level .9 m. Dashed lines delimit four arbitrary size groups 48 Figure 4. Distribution of Xiphister atropurpureus size groups in quadrats #1-10 of transect along the beach at Helby Island. Above: ink drawings of inner 60 x 60 cm2 of m2 quadrats (#1-10) in this transect 49 Figure 5» Numbers of Gilbertidia slgalutes caught in surface plankton hauls on Nov. 7» 1973 (open numbers))and Dec. 7. 1973 (circled numbers) 71 Figure 6. Histogram of dusk and night surface catches showing frequency of vertical migrations of Gilbertidia sigalutes. The vertical dotted lines indicate dates on which dusk or night tows yielded no catches. . . . . 76 Figure 7. Time after sunset of ascent to surface of Gilbertidia  sigalutes. Dashed curve: time after sunset of nautical twilight. 79 xii Figure 8. Flow chart of larval l i f e history. Most research emphasizes problems detailed above the dashed line. This paper covers points below that line. . . . . . . . 1 2 1 1 x i i i Acknowledgments Mr. Peter Gardiner, while enrolled in Marine Science 412 at Bamfield Marine Station, agreed to conduct field and labora-tory experiments I1 designed to test the substrate size prefer-ences of adult Xiphister atropurpureus and permitted me to summarize the results he obtained in this thesis. I wish to thank Mr. Gardiner for his efforts. Mr. Rob Knight designed the circuit for the dawn/dusk lighting simulator which was used in my rearing system. I thank Floy KLise Zitten and Deborah A. Roman for their help with the illustrations of larvae. This research was conducted at the Bamfield Marine Station. Staff members Myriam Haylock and John Boom generously devoted much of their private time to helping my project. Similarly, Dan Pace and Bruce Leaman greatly assisted my work. Above a l l , I would like to thank Audrey Marliave for aiding every phase of my research and writing. Drs. N.J. Wilimovsky, N.R.Liley and T.R. Parsons gave many suggestions for developing this project. They and Drs. T.H.. Carefoot, H.D., Fisher and P.W. Hochachka provided useful criticisms of the manuscript. This work was supported by a National Research Council Grant to Dr. N.J. Wilimovsky, a National Science Foundation Graduate Fellowship to the author and a National Research Council Postgraduate Scholarship to the author. 1 I. Introduction Researchers studying the ecology of early stages of marine fishes have been concerned mainly with causes of larval mor-talities. Hjort (1926) contends that larval mortalities, caused either by drift away from suitable adult habitats or failure to initiate feeding, lead to failing year classes. Marr (1956), reviewing the concept of the critical period (a distinct stage in larval development at which certain con-ditions can cause heavy mortalities), points out that most researchers have considered the initiation of feeding at the end of yolk absorption to involve the highest mortality rates of any stage. However, laboratory studies of factors affecting mortality usually concern only the earliest larval stages because of the difficulty of rearing through metamorphosis. Similarly, field studies emphasize early larval stages because of their greater abundance and ease of capture. Thus, Hempelt (I965), in his review of the critical period concept, omits consideration of the transition from the larval to juvenile form since so l i t t l e is known about the ecology of that stage compared with other l i f e stages of marine fishes. Perhaps owing to this lack of emphasis on late larval stages there is confusion over the relation between physical and behavioral changes ln the transition from the planktonlc larval to the adult (juvenile) stage. Contradictory defini-tions of metamorphosis are used in different papers. There is more agreement over the behavioral transformation to adult habits, but discussions of that phenomenon are presumptive and undetailed. 2 The confusion over what constitutes "metamorphosis" in marine fish larvae results primarily from authors using the term without specifying any particular definition. The defini-tion of biological metamorphosis is "a marked and more or less abrupt change in the form or structure of an animal during postembryonic development" (Webster's Third New International Dictionary). This definition covers the common use of the term as the stage at which larvae come to resemble miniature adults. Ahlstrom and Counts (1958) state that "Metamorphosis may be characterized as the period during which marked changes occur in body proportions and body structures without any marked increase in standard length." This definition refers to a stage of physical change which is often accompanied by significant ecological changes (changes in habitat and behavior). Based upon this sort of definition is Hubbs* (19^3) conception of a fish larva as a fish in •'developmental stages well differen-tiated from the juvenile," the juvenile being "young essentially similar to the adult." A l l of these definitions are equivalent. This paper will refer to "metamorphosis" in the sense outlined above (change from larval to juvenile/adult form); Many authors use a different definition of metamorphosis as the endpoint of formation of fin rays. However, adult morphometry usually has not been developed at this stage and there seems to be no ecological significance to the definition. An intermediate viewpoint is taken by Berry and Richards (1973), who distinguish between the larval stage (hatching to develop-ment of adult fin ray complement) and the prejuvenile stage 3 (after fin ray formation, before the end of metamorphosis). This definition emphasizes the fact that metamorphosis into the adult form is not always abrupt. In addition to the confusion over terminology, there are conflicting views on the relation between metamorphosis and the behavioral transformation to the Juvenile/adult habits and habitat. General works on fish biology usually refer to pleuronectiform larvae in discussions of metamorphosis since their eye migration is dramatic, but these discussions tend to be inaccurate and misleading. Marshall (1966) and Wimpenny (1953) incorrectly describe metamorphosis in Pleuronectes platessa as being completed before settlement to the bottom occurs. Shelboume's (1964) photographs of metamorphic stages of this species illustrate that, although eye migration precedes settle-ment, metamorphic pigmentation follows i t . Lagler et a l . (1962) extend the generalization of metamorphosis preceding settlement to a l l pleuronectids while Norman and Greenwood (1963) as well as Rae (1965) make this same statement for a l l pleuronectiform fishes (incorrectly, as shall be seen for Solea solea). Another incorrect generalization about metamorphosis is that physical and behavioral changes are often considered 6ne uniform process. Norman and Greenwood (1963) and Wimpenny (1953) contend that the two phenomena coincide, proceeding at the same rate (that is , the flatfish gradually alters its posture and sinks during metamorphosis). The temporal relation of metamorphosis and settlement of these larvae is not as clearly understood as the more general literature implies. 4 Fish, other than pleuronectiform species, are known which undergo an abrupt metamorphosis immediately after settlement. This pattern of a behavioral transformation that triggers metamorphosis has been witnessed in the field for acanthuroids (Freder, 1949; Randall, 196l) and blenniids (Chapman, in Weber and deBeaufort, 1951). Breder reports a considerable capacity for delay of metamorphosis in Acanthurus hepatus. which may grow several times in size without any structural change until i t encounters a reef and settles, after which i t metamorphoses within 24 hours. Randall also reports delayed metamorphsis in Acanthurus triostegus sandvicensis. which settles at vary-ing sizes (34$ variation in length) during an extended period. Late postlarvae of this species actively swim inshore at night, then settle and metamorphose within 48 hours. Metamorphosis does not occur until just after settlement. Perhaps the low probability of encountering an open ocean reef or island has selected for a high capacity for delay of settlement and meta-morphosis in; these species. Overall, there is very l i t t l e infor-mation on the period during which settlement of benthic species can occur. Relatively few species of marine fish larvae have been successfully reared through metamorphosis in the laboratory. Table I covers those instances of which I am aware. One might expect such studies to include detailed descriptive information on the behavioral transformation, but most authors simply comment that at some point the fish settled to the bottom or that adult habits were adopted. Table I . Species of marine f i s h larvae which have been reared through metamorphosis to the juvenile stage. in the laboratory from hatching-0. Clupeiformes (Isospondyli) P. Clupeidae Clupea harengus harengus Linnaeus Sardinops sagax (Jenyns) • Adult Habitat o o Fish Species (Classified by Order & Family) o- « to qj X X -P o m § § •H C -P O * •H E -p E « o G a <s § «5£ Author yes Rosenthal, 1968 no Bardach, 1968 no Longhurst, 1968 F. Engraulidae Engraulis mordax Girard 0;. Gobiesociformes (Xenopterygii) F. Gbbiesocidae Aspasma minima (Ddderlein) 0.. Gadiformes (Anacanthini) F. Gadidae Gadus morhua Linnaeus-0. Atheriniformes (Beloniformes, Synentognathi) F. Belonidae Belone belone Linnaeus X no Bardach, 1968 no Longhurst, 1968 yes* Shiogaki & Dotsu, 1971 no Rognerud, 188T no Rosenthal &: Fonds, 1973 * f i e l d information CO C+ lo o (B 3* 2 •s? o «• CO • ° to > O n P to co O c+ 4 *1 co o 3* *i « M o ct-|3 co a> O c+ H-a* o « IX 3 to 3 <D a w a. CT o *i to o o ca <+ co I a ca c+ 3* a co f CO I CD CO o CO :3 ca CO CT to > > 3" c+ c+ CO 3* 3" 1 ? 3 !? O H- O a co to ca Mi CO CD CO o fi H CO o o 3 c+ I a CO CO X X X X x x Pelagic x x. X x Benthic 3 o 3 O CO co 2 (0 3 o 3 o 3 3 o o 3 O 3 O 3 3 O O 3 3 Q O Info, on Transf. a o to CO vO vj\ 3 w 3* a H CO vO CT to H co oT 3 TO vO O O 3 CO vO •P-to P-co 4 to to cr 3 to o to vf) * i? ^ 3 o O vO Ic+vO o & CD vO 2 c f H" CO vO 8 s * H" TO W 3" I** M PS-CD 1 00 9 H3 CO vO Ov ON TO 3" fi» M PS-CD »1 9 Table I. (continued)) 0. Perciformes, continued. F. Pholidae ......... Pholis gunnellus=(Linnaeus) F.. Gbbiidae Bathygobius andrei (Sauvage) Lophogobius cyprlnoides (Pallas)) F.. Seorabridae Scomber japonicus Hbuttuyn F. Cottidae Clinocottns•recalvns (Greeley) Oligocottus snyderi Greeley 0.. Pleuronectiformes (Heterosomata) F.-Pleuronectidae Limanda yokohamae (Gttnther) Pleuronectes platessa Linnaeus F. Soleidae Solea solea (Linnaeus)) 0; Tetraodontiformes (Plectognathi) F.. Tetraodontidae Lagocephalus lunaris spadiceus (Rii • » 60 5 e o a. CQ Author X no Qasim, 1955, 1959 £ no Delmonte, et al . , 1968 X no Delmonte, et al . , 1968 no- Bardach, 1968 no Longhurst, 1968 X no Morris, 1951, 1956 X no Morris, 1956 yes Yusa, et al., 1971 yes Yusa, 1974 no Riley, 1966 no Rollefsenj 1939 yes Shelbourne, 1964 no Shelbourne, 1970 no Dannevig, 1948 yes Flflchter, 1965 son)) X no Fujita, 1966 8 Yusa, et a l . (1971) describe a switch from positive to negative phototaxis i n Limanda yokohamae when settlement occurs. Yusa (197^) also describes the sequence of allometric growth preceding settlement, which i t s e l f precedes formation of juvenile pigmentation. Shelbourne (1964) illustrates the same sequence of metamorphic changes i n relation to settlement for Pleuronectes  platessa. Shelbourne mentions that the presence of sand on the tank bottom i s not essential for settlement and completion of metamorphosis i n this species, although he did not perform any controlled comparison of the effect of this natural sub-strate on settlement. Fltfchter (1965) describes larval Solea  solea as repeatedly settling before and during the metamorphic eye migration. Permanent settling of this species also precedes formation of juvenile pigmentation. Fishelson (1963) reports that larval Blennlus pavo rise back to the upper water layers whenever they descend to the tank bottom. However, one even-ing (day 26), the'larvae "began to search for food at the bottom, subsequently hiding in a corner of the aquarium. The following day they already behaved l i k e young B. pavo in nature." That i s , their behavioral transformation took place within 24 hours. Rosenthal (1968) describes the gradual development of school-ing in Clupea harengus harengus. which occurs before metamorphosis. Shaw (1957, 1958) and Williams and Shaw (1971) describe the similar ontogeny of schooling in Menidla menidia. although they did not rear this species past metamorphosis. 9 Only a few studies offer information suggesting that larvae of benthic species display substrate preferences when they are settling. In a paper on the l i f e history of Aspasma minima. Shiogaki and Dotsu (1971) include the field observations that newly settled juveniles cling to the seaweeds Ecklonia cava and Sargassum spp. whereas the adults live on the undersides of stones. Garstang (1900) reared Blennius ocellaris under a variety of conditions and obtained settlement and metamor-phosis only with individuals reared in a prepared tank (with dense algal growth on tank surfaces). Individuals reared in ordinary plunger jars (cylindrical tanks with stirring discs) remained near the surface when they reached the stage of settle-ment, periodically becoming listless and finally dying. Delmonte et al. (1968) reared two tropical gobiid species in prepared tanks. With Lophogoblus cyprinoldes. 173 larvae settled, then developed juvenile pigmentation. However, only 3 Bathygobius  andrei settled, suggesting that the algal growth in the tanks may have been more suitable as a substrate for one species than the other. These last tiro studies suggest that, for benthic species, mortalities may occur at the stage of behavioral trans-formation i f conditions are unsuitable for settlement. The objectives of this study have been to determine which factors stimulate the behavioral transformation in marine fish larvae, at what stage during development the transformation can occur, and whether transformed behavior is crit i c a l to the comple-tion of metamorphosis. A variety of species was reared in the laboratory in order to permit generalizations about conditions 10 for behavioral transformation and the relationship of behavioral changes to metamorphosis. For the pelagic fishes Menidia menidia and Clupea harengus  harengus (mentioned above) conditions for initiation of adult schooling behavior (the behavioral transformation) are well demonstrated. During formation of fin rays, larvae become attracted to conspecifics and begin to approach and follow each other. Schooling is completely developed before metamorphosis begins. My experiments test the dependence of metamorphosis upon expression of transformed behavior in larvae of schooling species. The experiments isolating larvae of schooling species from their conspecifics are included for the sake of a more comparative base in testing the hypothesis that adult behavior is required for successful completion of metamorphosis. For benthic fishes, however, the conditions stimulating the behavioral transformation (settlement) have never before been investigated under controlled conditions. This study emphasizes substrate selection of trans-forming larvae of intertidal fishes, with incidental attention to effects of light intensity and currents. The substrate selection experiments are based on the hypothesis that larvae of benthic species test the substrate, then settle only on a certain substrate(s) which is an essential element of the adult habitat. Some alternatives to this hypothesis are that larvae settle anywhere, then metamorphose and seek the adult habitat as a juvenile, or metamorphose in the plankton, and then settle either selectively or randomly. Also tested in the substrate 11 experiments is the previously mentioned hypothesis that larvae failing to encounter a suitable stimulus (substrate) which results in transformation to adult habits (settlement) will die. Drift dispersal and the manner of accumulation or attrac-tion toward the shore, together with depth or pressure effects, are not considered in this work. 12 I!. Experiments and Field Observations A. Preface Rearing of larvae and experimentation "was conducted in the laboratory at the Bamfield Marine Station,,in Barkley Sound on the west coast of Vancouver Island, Canada. Figure 1 shows the location of the Bamfield Marine Station and other places mentioned in this paper. Throughout this report the tide levels referred to are on the Canadian scale.. The fish nomenclature recommended by the American Fisheries Society is followed (Bailey, et al., 1970).' Plant names are taken from Scagel (1971). Ecological data from fish museum collection notes are referred to by collection numbers coded "BC" for University of British Columbia and "BCPM" for British Columbia Provincial Museum. For the experimental results and field observations, a l l Chi-square values are followed by a symbol indicating significance (S))or insignificance (NS) at the ;005 level. However, for a few values which seem marginally significant, their particular level of significance (.01 or .05) is stated. Figure 1. Map of Barkley Sound with detail of Bamfield area. 14 B. Materials and Methods 1 . Experimental Animals This paper describes the behavioral transformation and metamorphosis of eleven marine fishes, the larvae of which occur in the neritic plankton. The most reliable source of larvae was eggs collected in intertidal and shallow subtidal areas. Larvae of certain species were collected from the plankton. The collection and egg incubation techniques and the spawning period (in the Bamfield region) are listed for each of these species in Table ix. The species are arranged according to general adult habitat and by family. Egg masses of intertidal fishes were incubated in a plexi-glas trough ( 1 2 2 x 1 5 x 15 cm) provided with slots into which pairs of plexiglas frames with No. 351 Nitex screens f i t . The egg masses were wedged inside the pairs of screened frames. / o \ This trough had a flow of fresh seawater (ca. 10 C) at a rate of . 3 liters/cm /rain. Eggs of Olupea harengus pallasi were incubated in aerated, closed systems with modified temperature and salinity (6°C, 15 ppt and 10°C, 20 ppt;,; combinations within the normal T,S range) in order to induce hatches at different times (as described by Alderdice and Velsen, 1 9 7 1 ) . 15 Table II. Experimental Organisms: season of availability, methods of collection and incubation technique. Intertidal Species Cottidae Artedlus lateralis (Girard) Blepsias cirrhosus (Pallas) Leptocottus armatus Girard Nautichthvs oculofasciatus (Girard) Agonidae Bothragonus swani (Steindachner) Stichaeidae Xiphister atropurpureus (Kittlitz) Pholidae Pholis laeta (Cope) Gobiesocidae Gobiesox maeandrlcus (Girard)) Subtidal Benthic Species Spawning Season Occurrence in Plankton Collection of eggs December - Hay ? ? ? February December - Majr February November - Kay Cottidae Gilbertidia sigalutes (Jordan & Starks)) July, August Schooling Species Clupeidae Clupea harengus pallasi Valenciennes Gasterosteidae Aulorhynchus flavidus G i l l March, April-May, June February - April October - March January - May schools, late May ? schools. May, June November - April March - June May - July intertidal, attached under boulders scuba, in kelp holdfasts intertidal, loose under boulders intertidal, loose under boulders intertidal, attached under boulders spawn in lab strip & fertilize from adults or on plants (shallow) plankton net or scuba (on plants)) s Incubation of Eggs Collection of Larvae open flow incubator — — plankton net — dipnet at surface plankton net open flow incubator open flow incubator open flow incubator scuba (dipnet)) parents tend eggs plankton net closed incubator (aerated, modified S & T) open flow incubator or on airstone 16 Materials and Methods 2. Rearing I reared fish using 180 and 1900 l i t e r circular tanks (53 cm diam. x 89 cm high and 144 cm diam. x 120 cm high, resp.) and 930 l i t e r hexagonal tanks (64 cm sides, 90 cm high). The tanks were equipped with through-flowing seawater creating circular currents grading from up to 4 cm/sec peripherally to n i l at tank center. Screened outlets were positioned cen-trally. The turnover rate was once per day in larger tanks and twice per day in the 180 l i t e r tanks. Water temperature o for rearing trials was 10-13 C and salinity around 29 ppt. Artificial lighting consisted of overhead banks of floures-cent lights timed to be l i t from 10:00 to 24:00 (giving about 500 lux at tank surfaces). Single 40 W incandescent bulbs over each tank were timed to increase in brightness from 8:00 to 10:00, then to decrease from 24:00 to 2:00. The shift forward in the light period allowed larvae to be fed plankton caught during dusk and early night. The fish larvae were fed to excess at least twice daily, once with Artemia nauplii and at other times with freshly caught plankton. Plankton was sieved into portions of 0-.25, .25-.5 and .5-1.0 mm or 0-.5 and .5-1.0 mm, depending upon the sizes of larvae being reared. Newly feeding or extremely small larvae (e.g. Artedius lateralis) were fed a plankton fraction of 0-.25 mm, larvae with well established feeding were fed a 0-.5 mm fraction, and the .5-1.0 mm fraction was fed only to very large larvae of certain cottidispecies. Dead food was siphoned from the tank bottoms each day. 17 Materials and Methods 3. Substrate Preference Studies on Intertidal Fishes Growth and behavior were monitored to detect readiness of larvae for behavioral transformation. Growth was monitored by taking samples of different developmental stages of larvae from the rearing tanks and killing them in weak formalin (.5-1$) in seawater. They were preserved in 2-4$ formalin in seawater (concentration determined by size), with Ionol crystals added as a buffer. After the completion of fin ray formation the onset of metamorphosis was indicated by the start of alloraetric growth and pigmentation changes. Behavioral changes such as an increased angle of swimming posture or reduced rheotactic tendencies typically preceded settlement. These changes in swimming behavior, together with the onset of metamorphosis, indicated an impending transformation to the adult habits. When larvae approached the stage of transformation, differ-ent substrates were introduced. Each substrate was placed in a separate plastic tray. Round trays (15 cm diam. x 2 cm deep) were suspended near the surface in the 180 li t e r tanks whereas larger rectangular trays (27 x 33 x 5 cm) were placed on the bottom of 930 or 1900 l i t e r tanks. Early trials in the 1900 li t e r tanks showed no difference in response to substrates suspended near the surface as opposed to placed on the bottom. In certain species the disturbance during removal of a tray would cause some individuals to shift position within the tray or to swim out of the tray. Fleeing tended to occur more often in the smaller trays. They were therefore suspended to permit closer observation of any fleeing while the trays were removed. 18 Trays were placed into tanks for periods of 24 hours, then they were removed and the numbers of larvae settled onto different substrates were counted. Results of earlier runs were used as a basis for selecting substrate combinations for later runs, so^results will be presented in chronological order unless otherwise stated. When larvae settled neither on inorganic substrates nor onto the tank walls or bottom, plant materials were introduced. Plant species were selected to provide larvae with the typical variation in plant morphology which could be encountered during drift. Differences in plant structure and surface texture on a size scale relative to the size of larvae were considered most important in selecting plant substrates. In most runs using plants or plastic substitutes for plants, these substrates were suspended in the water without trays, then netted after 24 hours. Replication of substrate preference tests was handled differently according to species because of variation in the extent of the settlement period. Artedius lateralis and Xiphister  atropurpureus were inclined to settle out abruptly. Metamor-phosing individuals tended not to leave a substrate during tray removal. These fish were not replaced into the rearing tank once they had been removed in a tray. Individual Leptocottus  armatus. Bothragonus swani and Pholis laeta tended to settle repeatedly over a more extended period, often swimming in mid-water during their metamorphosis. Those individuals which were settled in substrate trays at the time of tray removal were: 19 placed back into the experimental tank. Gobiesox maeandrlcus tended to settle onto the tank walls and then onto the bottom after their i n i t i a l period of settlement onto plants. Individual Gi maeandrlcus removed with plants were replaced into the experimental tank. All of the species (excluding Nautichthys oculofasciatus and Blepsias cirrhosus) were presented with different size-fractions of inorganic substrate, including sand, gravel, pebbles and rocks. Sand was less than 1 mm diameter, gravel was between 2 and 12 mm diameter, pebbles were from 20 to 40 mm diameter and rocks from 60 to 100 mm diameter. To present the rationale for the substrates offered to each species, a summary of information on the adult habitat will precede a description of the sequence of experiments for each species. When results of runs with natural substrates provided evidence of substrate preferences for a particular species, a r t i f i c i a l substrates were used to determine which parameters of the natural substrates serve as stimuli causing settlement. These specialized experiments are described sepa-rately for each species. Substrates provided different degrees of shade to larvae settling in them. To investigate whether changes in sensitivity to light mediate substrate selection, the phototactic tendencies of these fish before and after the behavioral transformation were tested in a phototaxis apparatus. The apparatus, a modi-fication of that described by Blaxter (1969 )•, is a black plexiglas trough (68 x 22 x 21 cm high, O.D.) with one end 20 of clear plexiglas to permit transmission of light from a point source 50 cm distant. Sheets of heavy paper were used as neutral density f i l t e r s . The tank has a vertical longitudi-nal divider and removable vertical dividers at each end which can close off end chambers 10 x 10 x 21 em high. Untransformed and transformed fish were placed separately into the two central compartments (dividers in place) and dark adapted one half hour. The clear end and top of the phototaxis tank were covered during dark adaptation and the entire arrangement was inside a dark tent. After dark adaptation, the covers were removed, the light source turned on (with filters in place) and the dividers removed. (The tank top would be recovered during the run.) After ten minutes the dividers were replaced and counts made of the numbers of fish in the various compart-ments . The light intensity gradients used in^these phototaxis experiments were probably higher than intensities for response thresholds. Using a vertical light source, Blaxter (1973)) found the response threshold for photopositive reactions in Clupea harengus harengus and Pleuronectes platessa larvae to be from 0.002 to 0.02 lux and nan optimal luminosity or pre-ferendum" (tendency to sink) to occur at 2i'0 lux. However, Gilbertidia sigalutes larvae tested in pilot runs of the photo-taxis tank described above showed no reduction in photopositivity at intensities reaching 10^ lux. Their distribution was random at an intensity of 0.1 lux and below. Three light intensity gradients were used in these experiments. Gobiesox maeandricus was tested at intensity gradients of 1000-300 and 7.5-0.2 lux; Leptoeottus armatus and Xiphister atropurpureus were tested at gradients of 500-100 lux and Pholis laeta at 7.5-0.2 lux. 22 Materials and Methods 4. The Vertical Migrations andTransformation of Gilbertidia sigalutes The periods during which certain intertidal fishes such as Leptocottus armatus. Pholis laeta and Bbthragonus swanl repeatedly settle onto substrates, then reenter the plankton are considerably briefer than their periods of larval develop-ment (up to 1/5 as long). Gilbertidia sigalutes. however, spends about 40 days as a strictly pelagic larva, then alter-nates between benthic adult and pelagic larval behavior patterns for about 120 days (three times as long).. This protracted transformation stage of G. sigalutes involves changes in the nature of their vertical migrations. These changes were moni-tored with plankton tows. Plankton towing was carried on throughout the larval and transformation periods during the winter of 1973/1974. Records were kept of the times and locations of tows, sea and weather conditions, and numbers of G. sigalutes taken. A l l surface tows were five minutes duration at three to four knots speed; subsurface tows (three meters depth))were ten minutes duration at one to 1.5 knots. These data clarify the question of whether vertical migrations occur in the larval stage of this species. These data also clarify the relation of changes in the vertical migrations to the transformation to benthic habits. Records of plankton tows did not indicate at what stage the G. sigalutes larvae start to vertically migrate. This question was studied in the laboratory. Larvae spawned in the laboratory were reared from hatching to the onset of 23 transformation (59 days at the end of the t r i a l ) . Proportions of larvae hovering in contact with the surface and swimming in mid^water were monitored during the first week after hatching to determine whether tendencies to migrate vertically would be expressed in the laboratory under the simulated dawn/dusk light-ing regime. The vertical migrations and repeated settlement of trans-forming G> sigalutes seemed to me to relate to feeding behavior. Gilbertidia sigalutes larvae reared through transformation were observed in order to determine the proportion of individuals swimming versus resting on the tank bottom. Proportions of fish swimming with food present as opposed to swimming with food absent were recorded. The feeding behavior of swimming and settled fish was also observed to determine whether settle-ment affected feeding behavior. Since settlement could involve encounters with adult fish, the effect of adults of G. sigalutes in the vicinity of newly settling G.. sigalutes was tested during January and February, 197^ . Seven adults (one year old, laboratory reared) were introduced into the tank containing newly settling fish. The proportion of newly settled fish resting on the bottom before, during and after the presence of adults was recorded. 24 Materials and Methods 5. Rearing Larvae of Schooling Species in Isolation In order to determine the extent to which transformed behavior is necessary for the completion of metamorphic growth in schooling fishes, larval Clupea harengus pallasi and Aulorhvnchus flavidus were reared in isolation from eonspeeifies. Such isolation prevented the expression of adult schooling behavior. These experiments were conducted using 18 l i t e r circular tubs of black plastic with open flows of seawater which create circular currents. These tubs are fully described in the appendix on rearing techniques (appendix 1). Clupea harengus pallasi larvae were isolated after the completion of fin ray formation and the development of school-ing behavior, but before the onset of allometric growth or guanine pigment deposition (43 days age). Two tubs contained twelve fish each, four tubs contained three fish each and twelve tubs contained single fish (total of 18 isolation tubs, 48 fish). The arbitrary feeding regime was one portion to isolated fish, two portions to groups of three and four portions to groups of twelve, so that isolated fish had lower prey concentrations, but a greater absolute amount of food. Lighting, sanitation and timing of feedings were the same as for the general rearing program. Records were kept of mortalities due to handling (within 48 hours of transfer at start of tria l ) , regular mor-talities and accidental losses (overflows) and a l l losses replaced with fish of the same age from rearing tanks. Forty days after the start of the experiment, a l l fish were killed 25 in 2$ formalin in seawater, measured and ranked according to their stage of metamorphosis. Larval Aulorhynchus flavidus were isolated at hatching. Single larvae were placed into each of sixteen tubs, a group of twelve fish into one tub and over f i f t y larvae into another. These fish were reared for fi f t y days, then killed in 2$ formalin in seawater and measured and ranked for their stage of metamor-phosis. 26 C. Substrate Preferences of Intertidal Fishes 1. Artedius lateralis. F. Cbttidae a. Adult Habitat Clemens and Wilby (1961) describe this species as being generally distributed in shallow water and tidepools, but spawn-ing only among rocks. I have observed adults of this species on solid rock bottom subtidally and on rocky shores of varying exposures. They have been collected from sandy beaches (BC 53-295), rocky tidepools (BC 53-296), boulder beaches (BC 64-46?) and shores with sand, pebble and rock bottom (BC 65-536, BC 64-464). b. General Rearing Results Larvae of A. lateralis (which were reared in a 1900 l i t e r tank) hatched ln early April, 1973. and had a planktonic stage lasting 41-60 days, settlement occurring during this 19-day period. Settlement of Individuals seemed to occur abruptly since metamorphosing larvae were rarely seen swimming in mid-water. Marked increases in pigmentation- immediately followed this rather abrupt settlement. Larvae hatched at 4.0 mm TL* and settled at about 11.0 mm TL. Juveniles are identifiable as adults within two months (at 30 mm TL). c. Substrate Preferences For convenience, coinciding hatches of this species and Xlphister atropurpureus were reared together in a 1900 l i t e r tank. Settlement of the two species coincided. Only the smaller grades of inorganic substrates were presented in two runs testing *TL = total length, tip of snout to end of caudal fin 27 the settlement of A. lateralis* A third run comparing coarse and rounded gravel (gathered from sheltered and exposed shores, respectively) was made at the end of the settlement period, so could not be replicated. Artedius lateralis did not exhibit a significant preference for any of the size grades of inorganic substrates offered. empty tray sand gravel pebbles 1 3 * 3 ( 2 runs)) x| = 1 . 7 , NS In the run comparing coarse and round gravel, a marginally significant preference for coarse gravel occurred (coarse - 1 1 , round - 2 ; x| « 4.92, S. Q 5 ). This cottid showed a marked tendency, however, to settle onto the.tank bottom and walls. Since X. atropurpureus was reared with this species and settlement of the two species coincided, a comparison of the numbers settling into trays as opposed to on the bottom was possible for four runs. In each of the four runs the bottom was siphoned after removal of the substrate trays. Siphoning permitted counts of numbers settling on the bare tank bottom. Artedius lateralis showed a significantly greater preference for the tank bottom than X. atropurpureus. trays bottom A. lateralis 26 125 (4 runs) xf = 107.4, S X; atropurpureus 125 41 This difference should not be construed to mean that A. lateralis prefers a solid substrate to the other inorganic substrates. If expected values are based on the percentage bottom area 28 covered by trays (expected settlement of A. later a l i s t 38 i n trays, 113 on bottom), there i s no significant preference for the trays or the bottom (Xj = 3.1, NS). The difference between these two species i s that X..atropurpurens shows di s t i n c t substrate preferences, but A. la t e r a l i s settles at random on these different substrates. 29 2. Blepsias eirrhosus, F. Cottidae a. Adult Habitat Blepsias eirrhosus inhabits shallow water, usually in eel-grass beds, and spawns on rocks (Clemens and Wilby, 1961). U.B.C. fish museum collection data show this species occupying eelgrass (presumably Zostera) beds (BC 65-567, BC 65-568, BC 65-127), Macrocvstis beds (BC 65-30, BC 65-17) and tidepools fi l l e d with another kelp, Alaria (BC 63-880, BC 63-1073). b. General Rearing Results Seven larvae of this species at different stages of develop-ment were captured from the plankton and reared through metamor-phosis in a 930 l i t e r tank. Unlike most of the other intertidal species studied, metamorphosing B. eirrhosus would never swim out of the current in a tank and hover at the walls. That i s , this species never settled before metamorphosis was complete. After metamorphic growth was complete, Bi eirrhosus remained positively rheotaetic even in the slowest currents (ca. 1 cm/sec). There were no mortalities among these fish, even among ones which did not settle until long after metamorphosis. Blepsias eirrhosus undergoes increases in the relative sizes of its pectoral, soft dorsal and anal fins while the adult pigment pattern develops. Blepsias eirrhosus metamorphoses rapidly in comparison with the other cottids (within two weeks). After this metamorphic stage, fish swimming in mid-water do o so with an increased upward angle (more than 4-5 angle with the surface) and with slower, more undulatory movements. 30 Blepsias elrrhosus c. Substrate Preferences . The small number and varied stages of development of the B. elrrhosus taken from the plankton precluded quantified com-parison of substrate preferences, so different inorganic sub-strates were never intentionally presented to this species. After some of the B. elrrhosus had metamorphosed, substrate' trays; were: placed down to test settlement of Pholis laeta larvae which had been reared in the same tank. Trays with gravel, pebbles and rocks, removed after 48 hours (April 6, 1974), had attracted no settlement of either species. The next run in this tank included trays of rocks, gravel and mud, again with no settlement of B. elrrhosus. Subsequent runs in this tank (during mid-April) included trays with Phyllospadix scouleri. Zostera marina. Macrocvstis integrifolia and Odonthalia flocossa (the rationale for plant choices pertained to P. laeta. and is explained in the section on that species). At least two of the B. elrrhosus had fully metamorphosed, yet no response was observed towards the plants in trays on the bottom (the Macrocvstis blades were placed flat on the tray bottom to imitate the posture of Agarum fimbriatum). During the week following these substrate tray runs, Fucus gardneri. both floating and sunken, was present in the tank. No responses were shown to these plants. On April 19 a strand of Macrocvstis with pneumatocysts intact was introduced (floating). Seven hours after this introduction the three largest B. cirrhosus had settled onto the Macrocvstis. At 24 hours the four largest individuals (over 25 nan TL, a l l 31 fully metamorphosed) were on the kelp, yet s t i l l none showed any response to Fucus. At this point (on April 21) Ulva sp. was introduced (suspended at the surface) and no response was shown to i t in the next 24 hours. During these observations the B. eirrhosus would not flee from the Macrocystis when i t was gently disturbed, but would hover and make efforts to remain resting in place on the blades. On April 24, 1974, two other laminarian algae, Alaria  marginata and Egregia menziesii were added together with some Sargassum muticum (these and the Macrocystis were a l l suspended from the tank rim). Ten counts were made over a period of four days; the plants were shaken to cause the B. eirrhosus to flee at each count. The data suggest no ultimate preference for the Macrocystis. Macrocystis Alaria Egregia Sargassum 11 8 6 3 (10 runs) After this set of observations, however, the B.,eirrhosus started settling onto Fucus as well. When I summarized the observations on this limited number of B. eirrhosus larvae, i t seemed that this species did not settle until after metamorphosis, and then only on :floating plants. These fish never settled onto any bottom substrate, either as larvae or juveniles. The juveniles were held for four months and persisted in hovering on floating plants. Their i n i t i a l settlement took place on a laminarian, Macrocystis  integrifolia. to the exclusion of the fucoid Fucus gardnerl. A week after i n i t i a l settlement the B. eirrhosus started hovering 32 and resting on other laminarian algae as well as on a different fucoid, Sargassum mutlcum. S t i l l later they began to hover on Fucus (previously avoided), indicating a certain amount of general-ization after settling. 33 3.. Leptocottus armatus. F. Cbttidae a. Adult Habitat Adults of this species occur on sand or mud bottom (Hart, 1973; Jones, 1962; BG 57-310, BC 63-1156). I have taken adults from sandy beaches. There is a tendency for settlement to occur in estuaries and for juveniles to move into freshwater (Jones, 1962). Juveniles have been taken i n a brackish water estuary with a rocky bottom (BC 60-413). b. General Rearing Results Leptocottus armatus larvae captured by dipnetting in March, 1973, settled abruptly onto the sides of a 180 l i t e r tank which had a current of over 2 cm/sec. Within a week these larvae had metamorphosed to a cryptically pigmented juvenile form not identi-fiable as adults. In October, 1973, a larger number of these larvae were captured and placed in a 930 l i t e r tank with a much slower flow (under 1 em/sec). In this larger tank the larvae tended to swim in raid-water. These larvae were presented with substrate trays into which they repeatedly settled. During this somewhat protracted settlement period (nine days) larvae were seen to swim In mid-water with increasingdegrees of cryptic juvenile pigmentation and to Increase In size from 12 to 15.5 mm TL. Leptocottus armatus are identifiable to species one month after settlement. They complete allometric growth in about two months (30 mm TL). Leptocottua armatus 3** e. Substrate Preferences Larval L. armatus which were dipnetted from the surface in Bamfield Inlet in October, 1 9 7 3 , were presented with sand and pebble substrates of different shades* The fish preferred to settle onto sand rather than pebbles, regardless of shade. black white dark light sand sand pebbles pebbles 1 5 3 3 3 1 ( 1 run)) X* = 4 9 . 9 , S Trays containing black sand, white sand and a mixture of the two shades were presented twice, a tray containing Fucus gardnerl and Ulva sp. on large rocks being included in the second run. white mixed Fucus & Ulva sand sand on rocks 3 5 1 4 8 2 The sum of the results for the sand trays in these two runs indicates a marginally significant preference for the white and mixed shades of sand over black (x| = 9 . 5 . S Q ^ ). The sum of the results for black and white sand in a l l three of the above runs indicates that these fish significantly preferred 2 the white sand over the black (1^ = 1 2 . 8 , S). The results for the run with Ulva and Fucus on large rocks show a marginally significant preference for the sand (Xj * 8 . 7 9 , S > 0 j ), so that no preference for plants or for rocks would be indicated. It should be noted that the total surface area of the plants and rocks in this tray was several times the surface area of any of the sand trays. black sand 0 5 35 Following these tests, two runs were made with white and mixed sands from the previous tests (commercially obtained coarse sands, up to 1 mm diameter) and with a tray of very fine beach sand (grayish white). No preference existed for the different size grades of sand. coarse coarse fine white mixed white 7 7 5 (2 runs)) x| = 0.33, NS During a l l substrate tray runs after the f i r s t , concurrent counts were made of the numbers of L. armatus resting on smooth and roughened linoleum tiles. These fish showed a slight pre-ference for the rough over the smooth t i l e : rough t i l e smooth t i l e 84 55 (16 counts) - 5.64, S < 0 5 . It should be noted that marked changes occurred in the behavior and growth of these fish during the week-long period of these experiments. During the fi r s t run the majority of the fish were swimming in mid-water, showing no metamorphic pigment changes. Most of the fish that settled in the f i r s t run selected the sand substrate trays. At the time of the last run a l l fish, swimming in mid-water or resting on the bottom, were undergoing metamorphosis. Few fish settled in the substrate trays whereas a large number had settled on the tank sides and bottom. The fish on the tank bottom were captured after removal of the substrate trays in both the f i r s t and last runs of this-entire study. The following table summarizes the results. on bottom ln trays fi r s t mn 18 52 ? X? = 59.8, S last run 69 8 After the i n i t i a l settlement onto a sand substrate, during metamorphosis, fish of this species undergo a generalization of their substrate preferences. After the fir s t substrate tray experiment a group of behaviorally transformed L. armatus was compared to untrans-formed individuals of the same species in the phototaxis appa-ratus. There is no significant difference between the num-bers attracted toward the light and those moving away from i t . l i t end middle dark end untransformed 21 17 0 _ (1 run)) XT * 2.1, NS transformed 17 5 2 Leptocottus armatus juveniles remain photopositive after settling to the bottom. 37 4. Nautichthys oculofasciatus. F. Cottidae a. Adult Habitat Hart (1973) states that this species rests in rock crevices and is taken in shrimp trawls at depths up to 110 m (presumably from soft bottom). The only substrate data from the U.B.C. fish museum show that this species occurs in rocky tidepools (BC 59-109) and on rock and sand bottom with laminarians (BC 63-1007). b. General Rearing A small number of the early larval stages of this species was captured in plankton hauls and reared in a 930 l i t e r tank. Only half of those which had". recently hatched fed successfully in the laboratory. This species hatched at 8.0 - 9.0 mm TL and started metamorphosing after fin ray formation, at 16.0 mm TL. During metamorphosis (16.0 - 30 mm TL), which lasted about one month, the fish began hovering against tank walls and then settled in vertical corners (secondarily on the bottom). Larval N. oculofasclatus developed large, fan-like pectoral fins which spread laterally to permit hovering in mid-water with only a slight sinking rate. These fins were folded against the body during rapid swimming. In currents the larvae usually alternated between hovering with spread pectorals and swimming forward. In advanced stages of metamorphosis (over 20 mm TL) the relative size of the pectorals became reduced by allometric growth. At these stages the fish increasingly tended to sink passively or swim down to the bottom when placed into either s t i l l water or a slow current. Nautiehthys oculofasciatus 38 e. Substrate Preferences As with Blepsias cirrhosus. there were too few individuals of this species to quantify substrate preferences. This species, however, did settle onto tank surfaces. The earlier metamorphic stages would hover near tank walls for extended periods before settling (as did another cottid, Rhamphocottus richardsoni GuVither). Since the periphery of the 930 l i t e r tanks (especially the corners) had virtually no current, seven metamorphosing Nj. oculofasciatus were placed in a 180 l i t e r tank where the current speed could be easily manipulated. In different current speeds the positions of four fish in early metamorphic stages (large pectorals) were compared to those of two late stage fish with small pectorals. Counts were repeated over extended intervals for each current speed, as summarized in Figure 2 (next page). In the f i r s t seven observations the early metamorphic stages behaved differently from the later stage fish, both at high and low current speeds. However, in the last observations the small fish (large pec-torals) may have either habituated to high current speeds or become fatigued so that they settled to the bottom at the end of the observation period. I 40 5. Bothragonus swani. F. Agonldae a. Adult Habitat This species tends to be taken i n moderately exposed areas on boulder/gravel beaches (BC 65-527), in kelp beds (BC 65-582) and in rock/gravel tidepools (BC 65-57*0 • Clemens and Wilby (1961) simply state that B. swani occurs in tidepools. b. General Rearing Results Bothragonus swani were reared from hatching through meta-morphosis with no mortalities. These larvae tended to eat newly hatched f i s h larvae in preference to invertebrate prey items. This species was kept i n a 180 l i t e r tank. Individual B. swani began hovering at the tank walls at 48 days of age. They started to settle periodically in substrate trays at 5^ days of age. This phase of repeated settlement for b rief periods lasted for two weeks (54-68 days )-j after which the juveniles swam to the tank bottom when removed from a tray and placed back into the tank. c. Substrate Preferences The small tank size (180 l i t e r ) dictated use of small substrate trays (15 cm diam., 2 cm deep), which were suspended near the surface to permit observation of any fleeing during tray removal. Only i n these small tanks were substrate trays placed i n the v i c i n i t y of the seawater i n l e t , because of the limited space. Bothragonus swani showed a significant preference for gravel over pebbles or rocks, as the following data show. 41 gravel pebbles rocks 7 0 0 (1 run)) X? = 9.3, S > 0 1 Summed replicates show a slightly significant preference for round or coarse gravel over sand. coarse round sand gravel gravel 2 13 24 31 (10 runs) X 2 • 6.9, However, closer examination of the data shows great variability in the preferences between runs. When the data are categorized as to whether trays were in front of the inlet or away from i t , i t becomes evident that B. swani prefer a substrate nearer to the inlet (in a stronger current). sand round gravel coarse gravel at inlet 6 18 14 away from inlet 6 6 4 (? runs) x | « 1.98, NS Analysing the data as a contingency table yields an insignificant test value, indicating that the effect of the inlet current is independent of the type of substrate. Therefore, substrates may be grouped to isolate the factor of current strength. Since there is a preference for gravels, data for sand substrate are omitted. Comparing lumped types of gravel at the inlet against gravel away from the inlet indicates that these individuals had a significant preference for a substrate to be in a current. at inlet away from inlet 32 10 (7 runs) X^ * 11.5, S 42 6. Xiphister atropurpureus. F. Stichaeidae a. Adult Habitat Adults of X. atropurpureus are limited in their distribu-tion to boulder beaches (Clemens and Wilby, 196l). They spawn under boulders in the intertidal (Wourms and Evans, 1974). More detailed information on adult substrate preferences is presented in the results below, b. General Rearing Results Xiphister atropurpureus larvae were reared through meta-morphosis in five trials. Settlement usually occurred at ages from 41 to 55 days, the settlement period for any hatch lasting an average of 12 days. At a l l stages these larvae swam at a slight upward angle. This swimming angle increased noticeably when the larvae de-veloped to the transformation stage. Melanophores marking traces of the adult facial pigment pattern also appeared at this stage. In the field, larvae reaching transformation were repeatedly observed schooling in the lee of rock outcroppings or in kelp beds (within one meter of large objects). School-ing tendencies were observed in laboratory-reared X. atropurpureus starting at 16-days age. When larvae settled into substrate trays they tended to burrow deeper into a gravel or pebble substrate rather than to flee, when disturbed. After settlement, adult pigmentation rapidly developed (within 48 hours). There was no allometric growth after settlement, although a 20.6$ reduction occurred in the maximum body width (gut) during the f i r s t four days after settlement. 43 Xiphister atropurpureus was the f i r s t species I^reared through the larval period. In the first rearing t r i a l , c r i t i -cal mortalities ensued for larvae which showed the first signs of metamorphic pigment changes as well as a marked increase in their upward angle of swimming. About 90$ of the fish which had been reared to this stage (50-70 total) died within a week before If realized that they were ready to settle. A tray of pebbles suspended in the tank attracted settlement of a l l seven survivors within three days. In another t r i a l with this species (in a 180 l i t e r tank), trays of gravel were suspended, one l i t and one shaded, during the settlement stage. Sanitation was neglected, owing to the obstruction caused by the trays, so that detritus accumulated on the bottom. A high mortality occurred (79$, 96 fish total), likely due to anoxic conditions met by individuals sampling the substrate of the tank bottom, c. Substrate Preferences Since adults occur strictly on cobble beaches where a gradation of inorganic substrate, sizesc.can- usually be found, only the series of these substrates was offered in earlier rearing trials. Reference to the contingency table in the section on A. lateralis (page 27) shows that X. atropurpureus had a significantly greater tendency to settle into the sub-strate trays rather than onto the tank bottom, indicating that this species had a substrate preference. Xiphister  atropurpureus showed a distinct preference for gravel substrate (trays were placed on the tank bottom for 24 or 48 hour intervals). 44 empty tray sand gravel pebbles 0 1 25 8 (1973; 2 runs) = 47.1, S pebbles >gravel 40 82 (1973; 6 runs) X* « 13.4, S The field data revealed that 0-year class X. atropurpureus prefer gravel. In the field I observed that they were aggregated in gravel under rocks. In the laboratory, X. atropurpureus at the transformation stage did not seek rocks or gravel under rocks, but simply gravel. round coarse round gravel gravel gravel under rocks rocks 16 19 1 0 (1974; 4 runs) X3 - 32.6, S A run with transformed and untransformed X.. atropurpureus in the phototaxis apparatus indicated a great reduction in photopositivity after transformation. lighted end dark end transformed 0 6 Cl run) X~ a 13.5, S3 untransformed 5 0 However, settlement into brightly lighted gravel was not signifi-cantly less than into adjacent shaded gravel. lighted gravel shaded^  gravel 10 15 X^  * 1.0, NS Therefore, the reduction in photopositivity may cause juveniles to seek more protected (darker) positions under rocks or deeper in the gravel after they have transformed but probably is not a mechanism inducing them to leave the plankton. Another indi-cation that photoattractions play no major role in the transformation 45 of this species Is that the settlement into substrate trays occurred both during the li g h t and dark periods. Transformed, metamorphosing individuals were placed into a tub containing a peg tray designed to simultaneously test for i n t e r s t i t i a l space size preferences and li g h t intensity preferences. This peg tray was designed to provide a gradation of i n t e r s t i t i a l space sizes. Newly transformed X. atropurpureus average 1.5 mm maximum width (at the gut), whereas after meta-morphosis (in about 24 hours) they average 1.23 mm width. The tray was made with four equal-sized compartments. The compartments have pegs separated by distances of 0.5» 1.3. 1.7 and 3*0 mm (minimum separation). The entire tray has a length-wise divider separating the peg compartments into clear and black halves. The transformed X. atropurpureus had a s i g n i f i -cant preference for pegs with 1.7 mm spaces. Thus they preferred i n t e r s t i t i a l spaces s l i g h t l y larger than their body width (about 1.5 x body width). 0.5 mm 1.3 mm 1.7 mm 3.0 mm 9 30 51 31 (9 runs) X* = 27.6, S Reclassifying the data according to black versus clear sides reveals a significant preference for the dark pegs. dark clear 71 39 (9 runs) X^ = 9.3, S Analysing the peg tray as a contingency table yields an ins i g -2 , nlficant X^ « 0.24, indicating the interstice preference to be independent of the preference for darkness. 46 Since Xi atropurpureus was the one f i s h used i n these sub-strate preference studies which occurs i n great abundance i n the intertidal of the Bamfield area, f i e l d studies were possible which provided supporting evidence for laboratory results. In addition, f i e l d surveys provided information on the distribu-tion of the 0-year age group relative to the older f i s h . Square meter quadrats were staked along vertical transects at a beach on Helby Island grading horizontally from fine gravel to pebbles to small rocks. The pebbles and rocks are overlaid with boulders to an increasing extent toward the more exposed end of the beach. Two of the vertical transects were made i n the middle of the beach where the substrate consists of gravel and pebbles sparsely overlaid with rocks. A l l sizes of f i s h tended to be accumulated under the rocks. The vertical tran-sects were made to determine the t i d a l height range of 0-year f i s h relative the the adult population. t i d a l height (m) # O-vear f i s h # adults July 30, 1973 0 0 2 .3-6 0 3 .9 1 6 1.2 1 9 1.5 0 3 1.8 0 0 2.1 0 1 2.7- 3.0 0 0 August 18, 1973 1.2-1.5' 9 14 1.5-1.8 3 13 1.8- 2.1 0 0 >7 Two quadrats were examined at the more exposed end of the beach where pebbles, crushed shell and small rocks are overlaid with large boulders. tidal height (m) # 0-year fish # adults .6 3 6 .9-1.2 0 5 Although limited, this information suggests that X. atropurpureus larvae settle onto substrates in shallow water, and that they remain more stratified inrtheir vertical distribution than the older year classes. These transects, together with Informa-tion from the many egg mass collections, indicate that a l l year classes of this species occur predominantly in the lower intertidal. A survey was taken of the length frequencies of X. atropurpureus occurring in meter-square quadrats placed at equal intervals on the same beach on Helby Island (see page 15)• This beach grades from fine gravel to boulders, with small rocks and pebbles under the boulders. This survey indicates a shift in size composition of X. atropurpureus corresponding to the shift in substrate size composition along this beach. Figure 3 shows the size composition of X. atropurpureus collected in this transect and the arbitrary size groups of fish which were desig-nated for a laboratory study on substrate size selection. Figure 4 shows the distribution of these size groups in each of the quadrats in this transect, as well as ink drawings made from photographs of the center of each quadrat. 2 20 18 16 14 c r . 1 2 LU | 10 z)- -Z 8 - 6 4 2 0 H - 1 • i 3 1 1 I 4 _ , , 8 9 10 11 12 13 14 15 16 17 18 19 20 LENGTH, cm . . . „ 1 _ . Figure 3« Size composition of Xiphister atropurpureus collected in horizontal transect on :Helby Island. Tide level .9 m. Dashed lines delimit four arbitrary size groups. ~~ co 8 10 40-| a 3 O u. O CD N CO CD CL 30H 20H CD - i 10H 2 3 1 Figure 4. Distribution of Xiphister atropurpureus size groups i n quadrats # 1 - 10 of transect along beach at Helby Island. Above: ink drawings of inner 60 x 60 cm of quadrats (#1-10) i n this transect. 4 5 6 Quadrat Number 10 50 Size group i consisted of 0-year fish* Their settlement occurred in quadrats with considerable gravel substrate.. Size group 4 occurred in quadrats with large boulders covering small rock and pebble substrate. Although this beach showed a generally uniform gradation of substrates i t should be noted that large rocks rested on pebbles in quadrat 2 (with many fish of size group 3) and quadrat 10 had finer substrate and fewer boulders than quadrats 7, 8 and 9, because of protective rock outcroppings at that end of the beach. There were no fish of size group 4 in quadrat 10. A laboratory study was conducted of the substrate preferences of individuals typical of these size groups. Six individuals of each of four size groups (cf. f i g . 3) were presented with substrates averaging 1 (and under), 5, 9, 18 and 30 mm diameter. In each of four runs the fish were introduced into a seawater table with even bands of the five substrates. After varying periods (2, 3# 4, 12 hours) the fish were netted from the different substrates (after draining the tank). The average cross-sectional area of the fish in each group was calculated from the mean depth and width of the fish. Similarly, the average cross-sectional area of the interstitial spaces in each substrate type was calculated from the average diameter of the partioles in each substrate. fish size group average C S . area (cm2) 1 2 3 4 0.2 0.5 1.1 4.5 51 average CS. area of -substrate type average diameter interstitial spaces (cm2) sand under 1 mm n i l fine gravel 5 mm 0.2 large gravel 9 mm 0.6 small pebbles 18 mm 2.5 large pebbles 30 mm 6.9 Each of the four runs (of the laboratory test of substrate preferences by different size groups) gave significant results (for example, Xg * 23.3. S, for run #1). Summarizing the results of these runs as a contingency table shows the differ-ences in substrate preferences of the size groups. 1 Fish Size Group 2 3 v 1 mm sand 0 0 0 0 5 mm gravel 1 0 0 0 9 mm gravel 14 6 0 0 18 mm pebbles 6 17 7 0 30 mm pebbles 0 0 17 24 (4 runs) V » 75.9, S Taking the primary preferences to be for 9 mm gravel for size group 1, 18 ram pebbles for group 2 and 30 mm pebbles for groups 3 and 4, the interstitial space size preference can be calcu-lated as a percentage cross-sectional body size. interstitial interstitial size group fish CS. area space. CS. area apace (# body CS.) 1 0.2 0.6 330# 2 0.5 2.5 580# 3 1.1 6.9 625f b 4.5 6.9 153* 52J the fish never selected a substrate with average interstitial spaces smaller than their cross-sectional body size. The two smallest size groups never selected the largest substrate, which had spaces 15 and J& times their cross-sectional body sizes. Data from egg collections indicate that full-sized X. atropurpureus never spawn under boulders with only sand, gravel or large rocks as substrates. Some component of small rocks, shells or pebbles always occurs in the under-boulder substrate where adult fish spawn (Marliave, ln press). Adults sometimes occur under boulders with gravel substrate, but only where the boulders are not rounded, so that larger spaces are created by the irregular surface of the boulder. Another study was undertaken to investigate an isolated strip of boulders that line a rock promontory in the middle of Brady•s Beach (which is sand). The central portion of these boulders is not covered by sand at any time and has a substrate of pebbles and small rocks. Although such an area would seem suitable for stichaeids and pholids, an extensive search re-vealed none. A transplant was performed to determine whether this area was somehow unsuitable for adults of this species or whether larvae will not settle in this area because of the surrounding sand. If settlement occurs subtidally, then larvae would encounter sand at this site and presumably would not settle. If the transplanted adults were to remain and spawn, then one could conclude the site to be a suitable habitat. The absence of this species would then suggest that larvae have failed to colonize this area because they settle subtidally. 53 On Sept. 13. 1973» 11 adult X. atropurpureus were trans-planted to the 0.9 m tide level at this site. On Sept. 26, 1973, 49 adult X. atropurpureus and 14 adult X. mucosus were transplanted to boulders from the 0.6 to 2.1 m tide levels. On October 12, 1973, a thorough search of this area from the 0.6 to 2.4 m tide levels revealed only 3 X. atropurpureus. a l l at the 0.6 m tide level. Two of these fish had sand adhering to their epidermis, suggesting that the amount of sand suspended in the water by wave action in this type of area may be damaging to these fish. The other fish may have emigrated. 54 7. Pholla laeta. F. Pholidae a. Mult Habitat Pholis laeta has been found among seaweeds and rocks in the intertidal (Clemens and Wilby, 196l), under intertidal boulders (BC 65-3, BC 65-57*0 and in eelgrass beds (BC 65-567, BC 56-638). b. General Rearing Results Pholis laeta larvae were reared through metamorphosis (in a 180 l i t e r tank) with very low mortalities. After the stage of fin ray formation, crowding forced the transfer* of about half the fish to a 930 li t e r tank (where Blepsias eirrhosus and Nautlchthys ooulofaseiatus larvae were being reared). Transfer mortalities were minimal. Individual P. laeta from this hatch settled at various times over a twelve day period (45-57 days age). No interspecific interactions were noted. Pholis laeta at the stage of behavioral transformation were inclined to swim to the sides of tanks (cross-ourrent) and swim back and forth there at the surface (angled upward). They tended to settle repeatedly for very brief periods, swimming as described above during the period when juvenile pigmentation forms. It will be seen that the series of inorganic substrates in i t i a l l y presented attracted only light settlement and fish settled on such substrates usually attempted to flee whenever a tray was disturbed. When a clump of Phyllospadix scouleri was presented the fish which settled would remain within the blades even during vigorous agitation. Pholls laeta 55 c. Substrate Preferences Settlement observations were made simultaneously on the P. laeta in 180 and 930 l i t e r tanks. The series of inorganic substrate grades was offered since this species also occurs on cobble beaches. The results of these runs provided the rationale for succeeding substrate presentations, so the results will be listed roughly in order and annotated accord-ing to sequence and in which tank they were performed. The fi r s t four settlement experiments involved trays with inorganic substrates. A large portion of the larval P. laeta showed slight metamorphic color changes and a ten-dency to swim out of the tank current and along the tank wall at the surface (both with and against the current direction). Despite these appearances of readiness to settle none of the inorganic substrates had a settlement of significant numbers o f f i s h . coarse rounded runs #1,2 sand gravel gravel (180 1 tank)) X§ = 0.34, NS run # 3 coarse (180 1 tank) sand gravel mud xf m 1.0, NS coarse run # 4 rocks gravel mud (930 1 tank) 2 0 1 0 X 2 = 0, NS Those fish which settled onto gravel trays tended to flee as the tray was removed, rather than to wriggle into the substrate. They were never seen burrowing. 56 Since these fish showed no definite tendency to settle either into inorganic substrates or onto the solid tank bottom, i t seemed possible that some sort of plant was the preferred substrate. Equal sized pieces of heavy green polyethylene plastic were suspended in the rearing tank, one piece pleated and the other cut into 10 mm strips (the width of Zostera marina eelgrass in the Banfield region). A comparably sized piece of Fucus gardneri (irregular in shape) was also introduced. After 24 hours these substrates were netted and no significant settle-ment was detected. pleated 10 mm strips run # 5 green plastic green plastic Fucus (180 1. tank) , 4 6 4 « 0.4, NS Although the pleated plastic has the thickness of kelp, the common intertidal sea lettuce, Ulva sp., is very flimsy in comparison. Similarly, the surf eelgrass, Phvllospadix scouleri, is not as wide as Zostera. Phvllospadix in Barkley Sound is usually 3 mm width or less. A run was made comparing pleated plastic, wide plastic strips, irregularly shaped Fucus and Phvllospadix. 1 run # 6 pleated 10 mm strips (180 1. tank) green plastic green plastic Fucus Phvllospadix 1 1 0 64 X2, » 161.9, S Not only did P. laeta show an obvious preference for Phyllospadix. they were tenacious in entwining themselves in the blades, many washings being required to remove them. Another run compared the pleated plastic with thin strips of plastic and a comparably 57 sized piece of Ulva. run # 19 pleated 3 mm strips (180 1. tank) green plastic green plastic Ulva 0 28 4 XJ; = 40.75, S The preference for Phvllospadix corresponds to the preference for plastic strips of the same width. Since thick, broad-bladed plants such as kelp have chemical exudates (Scotten, 1971) which are lacking i n the plastic (which is not preferred unless cut in thin strips), a run was made comparing Macrocvstis integrifolia (with exudates) to Phvllospadix (assumed without exudates). Also, since the structural irregularities of Fucus (not preferred) are on a scale larger than some of the body dimensions of transforming Pi laeta. the extremely fine structured Odonthalia flocossa was introduced. run # 7 Phvllospadix Odonthalia Macrocvstis (930 1 . tank) 9 25 9 0 X| = 25.5 , S The results s t i l l indicated Phvllospadix to be most preferable, although Odonthalia attracted some settlement. Macrocvstis was as ineffective as pleated plastic, so chemotactic attraction to kelps was ruled out. Three runs comparing the two forms of eelgrass indicated the thinner Phvllospadix to be more suitable. run # 8 3 & 10 mm strips Zostera Phvllospadix (180 1 . tank) 31 6 27 X| « 16 .0 , 3 runs # 8, 9, 10 in part Phvllospadix Zostera (180 & 930 1 . tanks) , 51 10 x£ = 27.6 , S Furthermore, this preference was narrowed down to the blades 58 rather than the root clump (including substrate) or some of the sand and gravel which had been amid these roots, run # 10 (180 1. tank) Phvllospadix Phvllospadix sand & gravel 3 & 10 mm blades roots & substrate from roots Zostera plastic strips 15 2 1 4 8 x£ = 21.7, S Summing the results of runs comparing Phvllospadix to plastic , strips (both widths combined) indicated no significant preference. runs # 8. 10 in part Phvllospadix green plastic; (3 & 10 mm) (180 1. tank) , 42 39 X| • O i l l , NS Additional runs with plastic strips established two points: narrower strips are preferable and the degree of light transmission affects preferences. The shaded i n t e r s t i t i a l spaces among black plastic strips are less preferable than lighted spaces (clear or green plastic). 3 mm strips 10 mm strips runs # 11-18 i n part green plastic green plastic (1180 1. tank) . 97 41 XJ = 22.8, S runs # 15-18 (180 1. tank) 10 mm green 3 mm green 3 mm clear 3 mm black 2 32 81 98 18 X^ « 78.8, S A run with the phototaxis apparatus showed no significant change in phototactic tendencies (direction of response) in behaviorally transformed P. laeta when comparing the ends of the tank. However, the transformed f i s h tended to remain unreactive i n the middle compartment to a significantly greater degree than the untransformed f i s h . 59 lighted end middle dark end untransformed 1? 1 0 transformed 5 11 0 (1 run)} xj* « 0, NS x | * 12,92, S 6 0 8. Gobiesox maeandricus. F. Gobiesocidae a. Adult Habitat Hart (1973) states that G. maeandricus clings to rocks in the intertidal and in tidal currents. This species spawns its eggs on the undersides bf intertidal boulders (pers. obs.). b. General Rearing Results At the beginning of June 1973i about 30 G. maeandricus had been reared to 2 5 days, over half of the larval stage. On June 11, 1973» three trays of pebbles (different shades) were placed into their tank, but only X. atropurpureus (12) settled into them. On June 15 two G. maeandricus were removed and preserved (10, 14 mm TL). Both had fully formed pelvic suction discs. Another set of substrate trays (sand, gravel, pebbles) was placed down on June 20, again with only stichaeids settling. Neither were any G. maeandricus ever seen settled onto the sides of the tank, so they presumably died at the stage at which behavioral transformation normally occurs (13-14 mm TL). Late larval stages of G. maeandricus commonly form large schools (numbering in thousands) in shallow subtidal areas. The small numbers reared in the laboratory showed no signs of schooling. Gobiesox maeandricus which had settled would, when disturbed, either flee the substrate and immediately settle elsewhere or remain fast. The tendency to flee may have been greater in fish settled on less preferable substrates. 61 The newly settled f i s h , at 14 mm TL, had suction discs 1.75 om diameter (0.25 mm i n depth). The disc consisted of transparent, unmodified f i n membranes (joined) and a transparent, ridged flap of skin posteriorly. The juveniles, up to a size of about 1? mm TL, remained on plant material. At this stage color metamorphosis remained incomplete and the sucking disc membranes had thickened and developed rudimentary papillae. When the f i s h were approximately 19 mm TL they switched to rock surfaces (or the tank bottom)l At this time the suck-ing disc was f u l l y papillate and 4 x 5 mm in size. Upon settling the f i s h underwent an immediate contraction of melanophores, leaving them a translucent yellow, through which the ground color of the plant showed, whether red, brown or green. The adult pigment pattern did not appear u n t i l the adhesive disc had developed sufficiently to permit adhesion to rock surfaces. Gobiesox maeandrious 62 c. Substrate Preferences Because inorganic substrates failed to induce settlement of larvae reared in 1973* some sort of plant material was presumed to be suitable for settlement. In 1974 wild-caught G. maeandrlcus larvae were offered comparable amounts of various plants, together with equal amounts of pleated and stripped green plastic (suspended in the tank without trays). They showed distinct preferences for broad, smooth surfaces such as Ulva sp., sporophytes of Gigartina sp. or Nereocystis luetkeana and pleated plastic. pleated strips of Fucus Ulva Gigartina* Sargassum Nereocystis* plastic plastic 9 26 2 3 1 23 25 2 2 *young sporophytes (4 runs) X^  = 44.2, S Attempts to determine whether rocks with Ulva attract heavier settlement than bare rocks failed due to a tendency for G. maeandrlcus to settle onto the undersides of suspended trays. There was virtually no settlement onto the rocks themselves. By suspending the peg tray with its base just in the water, i t was found that these transformed fish preferred the black underside to the clear. Although the phototaxis apparatus revealed no significant difference in the photopositivity of transformed and untransformed G. maeandrlcus (comparing lighted and dark ends), a significantly greater proportion of meta-morphosing fish remained unreactive in the middle compartment. ion trans formed transformed clear surface underside black surface underside 24 (4 runs) Xj = 11.2, S l i t end middle dark end 43 12 12 39 3 4 (2 runs) X = 3.2, NS X =32.4, S 64 D. The Vertical Migrations and Behavioral Transformation of Gilbertidia sigalutes 1. Adult Life History Very l i t t l e information on the l i f e history of this species has been published. Clemens and Wilby (1961) report the f i r s t B r i t i s h Columbia specimen having been trawled at 62 m and speci-mens from Puget Sound having been found in sponges from 102 m depth. Hart (1973) states that individuals from 5 to 40 mm in length have been taken i n plankton hauls (maximum recorded length i s 83 mm). Hart reports spawning as occurring during July i n B r i t i s h Columbia waters. In August, 1973 and 1974, adults of this species spawned in the laboratory. The spawning involved a single male defending a territory, courting a group of females and stimulating syn-chronous egg laying i n one spot. The females have very low fecundity for an oviparous marine teleost (15 females produced an average 130 eggs each). Adults normally spawn during August ln Barkley Sound, but an unusual spawn took place in the laboratory during the spring of 1974. Seven adults held in the laboratory were unfed for three months during the winter of 1973-1974. They showed only limited wasting. Although they remained s t i l l during this period, they were reactive whenever disturbed. When these starved adults were placed i n a tank with transforming G. sigalutes at the end of January they ate newly settled f i s h . The adults were removed after two weeks, having eaten about 30 (50$))of 65 the larvae. These feedings were sufficient to stimulate prema-ture sexual maturation and spawning, despite continued starvation subsequent to these feedings. (At the end of transformation, permanently settled juveniles were treated as adults and not eaten by larger Q. sigalutes.) In addition to the laboratory spawnings, experimental populations of G. sigalutes were caged above and below the seasonal thermocline (at 8 and 40 m depth). There was no spawning in the shallow cage whereas a l l females spawned i n the deep water group. The cages differed in depth and in lighting conditions, the deep cage being in relative darkness. Also, the deep cage was probably subject to less water movement than the shallow cage which was in tidal currents. Recent collections by the B.C. Provincial Museum have yielded relatively large numbers of sexually mature adults from different locations with two particular types of subtidal habitat:: boulder slopes (BCPM< 973-129 (5 fish)} BCPM; 974-407 (5), BCPM 974-488 (28), BCPM 974-497 (5) ))and rock c l i f f s with crevices (BCPM 97^ -^ 68 (5), BCPM 974-485 (150)} BCPM 97^ -491 (13), BCPM 974-492 (9), BCPM 974-494 (?) ). Al l of these sites are sheltered from wave action. These fish average 55.1 mm TL (excluding BCPM 974-485). 66 Gilbertidia sigalutes 2. Larval Stage In 1974 Gilbertidia sigalutes were reared from hatching to the early transformation stage, but the t r i a l was stopped long before permanent settlement had occurred. The newly hatched larvae showed a strong tendency to hover at the surface in a neustonic fashion (head up, in contact with the surface film). However, observations were made to detect any tendency to migrate vertically. During the fi r s t week after hatching, repeated counts were made of the numbers swimming in mid-water and the numbers hovering at the surface. Counts made on four occasions during simulated twilight periods averaged 38.5/122 (31.5*) hovering at the surface. At eight different times during the daylight periods counts averaged only 18.4/165 (11.1*) hovering at the surface. In the four daylight counts nearest in time to the four twilight counts, 18.7/121, or 15.^*» were hovering at the surface. These limited observations suggest that yolk sac larvae tend to migrate vertically in the confines of a laboratory tank (0,8m depth) under simulated dawn/dusk lighting conditions. Plankton tows were made throughout the larval period of this species during November and December, 1973. Pairs of consecutive surface and subsurface tbws were made in the same places. A comparison of such pairs of tows made during day-light as opposed to during dusk or night reveals a significant reversal of the distribution of G. sigalutes larvae. 67 surface subsurface day (11 pairs) 7 36 2 dusk/night (24 pairs) 125 11 Xj * 99.0, S In order to unequivocally demonstrate vertical migration, concommitant decreases in subsurface numbers must be demon-strated along with increases in surface numbers in night hauls as compared with day hauls. The above data clearly demonstrate vertical migration in G. sigalutes larvae. Furthermore, a comparison of three pairs of consecutive and adjacent hauls made around nautical twilight (sun 7-12° below horizon; normal human vision difficult) and then after astronomical twilight ('sun 13-18° below horizon; starlight becoming visible) reveals a significant tendency for these larvae to leave the surface after dusk. dusk dark night surface tows (3 pairs) 73 17 X 2 = 17.4, S ; Larvae were caught at the surface regularly during moonlit periods. Appendix 3 (page 230) summarizes a l l tows made during the winter of 1973/1974, excluding weather, lighting, time and position data. Appendix 4 (page 23l) converts the data from appendix 3 into the catch per unit effort. On November 15, 16/and 17, 197^, surface tows were made at various hours. number of tows total catch^ dawn daylight dusk night 11 15 12 9 18 8 11 4 catch per tow 1.64 0.53 0.92 0.45 68 The larvae tended to rise to the surface=both during dawn and dusk, but apparently to a greater extent at dawn* Catch rates differed between Bamfield Inlet and Trevor Channel. In the winter of 1973/1974 daytime surface catches were higher than subsurface catches in Bamfield Inlet (.39 per surface tow versus .3 subsurface). The opposite trend occurred i n Trevor Channel (.75 P®r surface tow versus 1.82 subsurface), despite higher overall catch rates. There i s no downwelling (downward flow of surface waters) within Bamfield Inlet during calm weather. Gilbertidia sigalutes larvae were caught i n nighttime surface tows in Bamfield Inlet only on moonlit nights or during SE windstorms. The vertical mixing during high winds could randomize the distribution of these larvae, but the large numbers caught at the surface would more l i k e l y indicate efforts by the larvae to swim to the sur-face during windstorms which cause intense downwelling. In Trevor Channel, the eastward direction of net surface transport causes continuous downwelling along the southern shore of the Channel (sampling area). Evidence of this i s the regular occurrence of foam and debris patches along this shore. Freshwater runoff, and therefore net transport tend to be very heavy during the late f a l l , so that downwelling i s usually intense during this period. Larval (5. sigalutes were taken along this shore on calm, dark nights. It appears that these larvae remain at the surface at night only i n moonlight or where there i s downwelling. 69 During extended periods of low runoff G. sigalutes larvae were caught at the surface during daytime in Trevor Channel, although in daylight they had less tendency to be near the surface in Trevor Channel than in Bamfield Inlet. The results of consecu-tive tows made along the Channel shore in sunlight and in adja-cent shade from the c l i f f s indicate that G. sigalutes larvae accumulate in the shade of the shore. Eight pairs of tows totalled 6 fish in the 31m and 24 fish in adjacent shade (X 2 » 10.8, S). Another approach to analyzing differences in the horizon-tal distribution of G» sigalutes larvae is to compare catches in pairs of consecutive hauls made within foam lines or patches (at downwellings) and in adjacent open water. Foam occurred primarily at the hydrologic front at the mouth of Bamfield Inlet and along the shore of Trevor Channel. Larval G. sigalutes show a significant tendency to aggregate in downwelling areas. foam adjacent open water 11 pairs 2 (surface tows) 47 18 Xj • 12.06, S Considering the differences in dark night catches along the:shore of Trevor Channel and within Bamfield Inlet (areas with and without downwelling), i t is of interest to compare day and night catches i n a different system, that i s , in foam lines (downwellings) and open water. open water foam ' dajrr dusk night day dusk night # tows 88 46 57 11 7 4 total catch 43 163 41 10 47 22 average catch 0.49 3.54 0.72 0.91 6.71 5.50 70 Aside from the overall increase in numbers in foam areas, there was a marked increase in the average nighttime catch in foam areas as opposed to open waters (almost 8x increase at night, foam vs. open water, compared with under 2x for day and dusk). The results of night tows made along the southern shore of Trevor Channel, as well as at foam lines and during high winds, a l l support the argument that G. sigalutes larvae swim upward against downwellings to a greater extent during darkness than during daylight. That the larvae did not remain at the surface in foam patches during daylight suggests that they sink with downwelling to depths of less intense flow velocity when light-ing is adequate for feeding at such depths. The accumulation of relatively high densities of these vertically migrating larvae in the downwelling area along the southern shore of Trevor Channel was marked by a restricted distribution in this area. Figure 5 shows the numbers caught in the vicinity of Bamfield on two separate occasions. The arrow shows the direction of net transport. The overall d i s t r i -bution evidently underwent l i t t l e change through the larval stage. Other surveys showed that their distribution tended to be stratified against the shoreline in Figure 5. Oh moonlit nights, when they remained at the surface for the longest periods, their distribution was observed to spread as far as halfway across Trevor Channel. However, as mentioned earlier, they occurred predominantly in the shade of the shore on sunny days. Bark ley °«%° Sound P a c i f i c Ocean Vancouver Island direction of net transport Trevor Channel 0 0 © (DV Vancouver Island 01 j Figure 5. Numbers of Gilbertidia ! sigalutes caught in surface plankton ihauls on Nov. 7, 1973 (open numbers) land Dec. 7, 1973 (circled numbers). • 72 Gilbertidia sigalutes 3. Transformation Stage Larval Gilbertidia sigalutes were taken from the plankton during November, 1973, and reared in the laboratory. They started settling onto the tank bottom during late December. Also, transforming fish had been caught from the plankton during January, 1973, and reared similarly. At any one time during January and February, 1973 and 1974, a small propor-tion of the transforming G. sigalutes held in the laboratory were resting on the bottom., Individuals would remain on the bottom for periods of only a few minutes during this earlier part of the transformation stage. These fish settled randomly over the bottom and showed no tendency to seek cover. During March and April the transforming juveniles rested on the bottom for longer periods. As the G. sigalutes settled for longer periods they showed an increasing tendency to seek locations under solid objects or in corners. By the beginning of May the juveniles were behaving as adults, either resting on the bottom (usually under cover) or swimming around the sides of the tank in contact with the walls. The tendency to settle was inhibited by the presence of adults on the bottom. At the start of the transformation period (in the fi r s t half of January, 197*0 over 95* of the larvae were swimming at any one time. From January 15 to January 25 daily averages of numbers swimming were between 80* and 90*. Oh January 25, adult G. sigalutes were introduced into the tank with the transforming larvae, I observed that larvae avoided settling in the vicinity of adults, which would eat them. Percentages of larvae swimming rose to 96$ after introduction of adults, then to 100$ during the fi r s t week of February. The adults were removed on February 8. Within-a week after removal of the adults, percentages swimming dropped to around 60-70$. During February, 1973, repeated counts were made of the positions of transforming G. sigalutes. before and after the introduction of plankton. The results indicate that these fish had a tendency to remain settled on the bottom when plankton was sparse. on bottom swimming (8 runs; 96 counts) unfed 202 405 Xf m 83.9, s; feeding 113 738 It should be pointed out that during early transformation individuals resting on the bottom never fed, despite the fact that copepods accumulated just off the bottom. Feeding took place while the fish were swimming in mid-water. During the protracted behavioral transformation dusk migrations occurred later and later after sunset and dawn migrations occurred much less often. At the same time the cephalic portion of the acoustico-lateralis^system was develop-ing. The trunk portion of this system was enlarged during transformation but did not fully develop until the end of transformation. Laboratory observations indicate that in 74 the adults the cephalic acoustico-lateralis canals are used in feeding, whereas the trunk portion serves in territorial and sexual displays. Gilbertidia sigalutes in the early transformation stage were observed feeding in the laboratory under lighted condi-tions. As in the early larval stages, they visually fixated on prey which hovered immediately in front of their heads in the region of binocular vision. If the prey remained s t i l l they would snap with a forward dart of a few millimeters. Individuals at this stage were never observed feeding while resting on the bottom. Initially they showed no response to introductions of small amphipods. However, later stages started feeding on amphipods during a period of deprivation of plankton. At this later transformation stage and as adults, G. sigalutes would abruptly snap at amphipods swimming near the head, usually under the mandibular pores of the head canals. Ingestion involved buccal pumping and a downward or sideways lunge resulting from backwards thrusts with the pectoral fins and trunk flexure. Such actions were usually directed at prey well outside the visual field. This behavior would permit exploitation of night-time surface aggregations of zooplankton since transforming larvae which have developed this mode of feeding migrate only after dusk. Stomach samples have revealed very heavy feeding by transforming juveniles night-lighted at the surface. These immature fish grow very rapidly while they are s t i l l feeding on plankton (i.e. to 75$ of adult length). 75 After the beginning of February, 1974, most plankton hauls were crudely ranked according to the density of the zooplankton catch (light, net bucket drains almost entirely; moderate, bucket 1/2 f u l l ; heavy, bucket f u l l ; very heavy, bucket well over f u l l ) . During the late transformation period of infrequent migrations, no G. sigalutes were ever taken in tows which yielded light catches. On March 14, four G. sigalutes were taken in night hauls which yielded heavy catches. The next night a l l . catches were extremely heavy and nine fish were caught. No record was made of catch densities on the following night, Maroh 16, when only one fish was taken. On March 17, 18 and 23 a l l night tows yielded light catches (and no G. sigalutes were caught). On March 25 night catches were heavy again and one G. sigalutes was taken. On the succeeding two nights, two and five fish were taken in heavy zooplankton catches. Thus, there was some tendency for G. sigalutes migrations to corres-pond to periods of peak nighttime surface densities of zoo-plankton during this stage of transformation. This could indicate that the fish at this stage are migrating vertically to the surface along with their prey populations. In the field the vertical migrations of G. sigalutes took place less and less frequently during this protracted trans-formation period . Figure 6 shows the catch per tow for dusk and night surface tows made during the winter of 1973/1972*. The vertical dotted lines in the figure indicate dates on which dusk or night tows were made without any catch. The decrease 76 N O V 1973 DEC J A N FEB Figure 6. Histogram of dusk and night surface catches showing freauencv of V - T ^ M I 77 in catch rates after the end of the hatching period (November) probably represents a natural decline in population numbers rather than any increase in net avoidance by larger individuals since this species is extremely, sluggish at a l l growth stages. The important point demonstrated by this graph is that the frequency of migrations decreased during the transformation stage. Since G. sigalutes reared in the laboratory spent increasing periods of time resting on the bottom during this stage, the reduction in migrations in the field presumably resulted from larvae repeatedly settling to the bottom. That transforming larvae settle in the field is indicated by a single collection (BCPM 72-88) of 16 G. sigalutes with "con-spicuous orange pectorals" (field notes of A. Feden, pers. comm.). These fish averaged 41.9 mm TL (76$ of adult size), yet were definitely s t i l l making vertical migrations, judging from the larval body shading (pers. obs.) and the orange pectorals. The largest individuals showed no signs of gonadal development (pers. obs.). This collection was made in shallow water on a flocculent bottom with no rocks, in contrast to the habitats where adults are found. Another change in the pattern of vertical migrations which occurred during this transformation period was that migrating fish reached the surface later and later during dusk. Figure ? graphs points indicating the time after sunset at which a catch was made in a tow immediately preceded by at least three consecutive tows with no catches. Such catches best indicate 78 the time when migrating fish firs t reach the surface. The re-gression- on these points shows a significant (P = .021) trend for migrations to occur later after sunset. The dashed curve shows the time after sunset of nautical twilight (def *n. p.. 6?). The regression equation is Yv= 24.9 + 0.61X. This regression line cannot be extended beyond the period of mid-December to the end of March since before this period migrations occurred at or before sunset (in November and early December, 1973, 53 larvae were taken in 87 daytime surface tows) and after this period there were no migrations. Vertical migra-tions ceased entirely imearly April, 1974. There was a trend for daytime catches to become more rare as vertical migrations became later during dusk. In a total of 166 daytime surface tows taken from mid-December to the end of March, only two transforming fish were taken, both during^January. Again, i t should be noted that these trends ('fewer daytime catches, later migrations) cannot easily be construed to mean that these fish start avoiding the nets during daylight. If these observations were the result of net avoidance the catches would be sporadic and the numbers caught low. Reference to Appendix 3 (page 230) or Figure 6 (page 76)} for example during March, demonstrates that catches are cyclic and that relatively high catches are made on some nights even when migrations have become infrequent. In addition to the shift in the timing of dusk migrations, there is evidence that relatively fewer G. sigalutes migrated 79 180-1 150H 120-f 2 OH NOV ' DEC JAN ' FEB ' MAR 1973 1974 Figure ?. Time after sunset of ascent to surface of Gilbertidia  sigalutes. Dashed curve:: time of nautical twilight. 80 at dawn during the transformation stage compared to the larval stage (cf. page 6?). Dawn tows made on December 30 and 31, 1973, yielded two f i s h i n ten tows. From dusk on January 2 to dawn on January 4, 1974, 20 G. sigalutes were taken in 12 dusk and early night tows (1.67 f i s h per tow), compared with only 2 f i s h i n r l 4 dawn tows (0.14 f i s h per tow). Whereas dawn migrations tended to be more intense than dusk migrations during the larval stage, the opposite trend appeared early during the transformation stage. The tendency was for migra-tions to the surface to become increasingly nocturnal during transformation• E. Schooling Species 81 1. Introduction The behavioral transformation of two schooling species was investigated. One, Clupea harengus pallasi. always occurs in schools as a juvenile or adult whereas Aulorhynchus flavidus forms schools only during the nonbreeding season. Another distinction between these species is that C. harengus pallasi is pelagic and A. flavidus swims near the bottom. The develop-ment of schooling differs between these two species. Inr,these species, both the larval and adult stages are free-swimming. Larvae show no response to conBpecifics (other than avoidance upon contact) before the onset of sohooling. Adults swim as a group with polarized orientation and can swim against currents. Their larvae presumably drift with the currents. The behavioral differences distinguishing the larval and juvenile/ adult stages are primarily those associated with the development of schooling (the behavioral transformation). In the case of C. harengus pallasi another important distinction is the onset of f i l t e r feeding, which appears much later than schooling. Since certain benthic species had suffered mortalities when prevented from settling (transforming) during metamorphosis; I decided to attempt a comparable test on schooling species by preventing schooling during metamorphosis. Individuals were isolated from their conspecifics at that stage. The survival and metamorphic growth of the isolated fish was compared with that of control fish reared under similar conditions In groups. Again, the purpose was to determine the extent to which trans-formed behavior is necessary for successful metamorphosis. 82 2. Clupea harengus pallasi. F. Clupeidae a. Life History Notes Stevenson (1962) surveyed the distribution of C. harengus  pallasi larvae spawned in northern Barkley Sound. He never caught them seaward from the mouth of the Sound and concluded that those swept offshore either die from adverse abiotic or feeding conditions or become so widely dispersed that they f a i l to encounter conspecifics for the onset of schooling. His conclusion that the contingent of larvae swept offshore must die rests on the correspondence between the number of larvae remaining inshore and the number of adults in the returning year class* b. Isolation Experiment This species was isolated only during the actual stage of metamorphosis (allometric growth, pigmentation) to determine whether schooling behavior was a necessary condition for successful completion of metamorphic growth. The larvae reared for this experiment started parallel swimming (in pairs, for brief periods) on day 16. Schooling behavior was fully developed by day 25, at which time two pilot isolations were started (no mortalities occurred). On day 43, after the completion of fin ray formation, the isolation experiment was initiated, as allometric growth was becoming evident in some individuals. Fish were reared in isolation and ln control groups for 40 days, through the normal period for the completion of metamorphic growth. 83 During the experiment no mortalities occurred after mortalities caused by handling. However, tubs with groups of three fish overflowed frequently due to outlets clogging with dead food organisms. Losses due to overflows were so numerous that eventually a l l of the fish in these tubs had been introduced as replacements at later developmental stages than the starting group, so that groups of three had to be eliminated from the study. Comparisons are limited to the groups of twelve fish and the Isolated fish. The only effect which isolation seemed to have on individuals of C. harengus pallasi was a slight retardation of development. At the end of the experiment (day 83), fewer of the isolated fish had completed metamorphosis, even though their growth in Size was comparable to that of the grouped fish. Isolated fish averaged 41.2 mm TL (range 36-51 mm) and grouped fish averaged 43.9 mm TL; ( range 38-49 mm). unmetamorphosed metamorphosed isolated 5 5 grouped 1 20' 2 XjL = 8.4, S It should be noted, however, that the isolated fish which did complete metamorphosis averaged 44.4 mm TL. 3. Aulorhynchus flavidus. F. Gasterosteidae a* Life History Notes Because A. flavidus are bottom-associated they are probably bounded by stretches of deep, open water. Limbaugh (1962) found their average depth to be 11 m and their maximum depth 30 m . Limbaugh observed that their egg masses were primarily laid on Macrocystis pyrifera. below the growing tip of the young stipes (encircling the stipe at the base of the pneumatocyst). He also observed nests (blades are interconnected with adhesive threads at the nest site) without any egg masses. He noted that a new batch of eggs would be spawned on a pneumatocyst base i f a previously laid mass were to break free. He further noted that egg masses of this species were sometimes washed ashore at La Jolla, California, during the winter months. Hart (1973) reports hatching in this species as starting two weeks after fertilization and continuing for four to six weeks. Larval A. flavidus form schools near the bottom shortly after hatching-(Limbaugh, I962). b. Field Observations I have observed spawning on Alaria marglnata (early sporo-phytes), young Nereooystls and new growths of Scytosiphon lomentaria. The common feature of these spawning-substrates was that they were new growths (not very sturdy). In every case the mass encircled a constriction just below the growing portion. In 1973, a l l four egg masses which I collected were taken in surface plankton tows in the downwelling area along the 85 shore of Trevor Channel after the end of a two day windstorm. These egg masses had been spawned on Alaria sporophytes and had torn free with the plant material intact. In 1974, two nest sites of this species were located in widely separated areas. Two egg masses were removed from one site shortly after they had been laid and were incubated in the laboratory for use in isolation experiments. Shortly after, several more masses were spawned in their place. Before any of the egg masses at the two field sites started hatching, a moderate windstorm, lasting only one day, caused a l l of the egg masses at both sites (about 12 total)}to be torn loose. Only one of the sites was monitored after this, and further spawning ensued within three weeks, c. General Rearing Observations Aulorhvnchus flavidus normally hatched singly. Egg masses permitted to hatch normally in 1973 hatched over a period of 10 to 14 days. In 197^ a brief exposure to air (15 min.) caused a limited burst of hatching within an hour. Ih 1973 I reared this species in a 1900 l i t e r tank and observed larval schools within a week of the onset of hatching. For the f i r s t two weeks these schools remained within 10 cm of the tank bottom. Both laboratory and scuba diving observations revealed that later stage schools remain fairly closely associated with the bottom. In the 1900 l i t e r tank, home site specificity and roaming during poor food conditions were characteristic of these schools from hatching through metamorphosis. There 86 appeared to be no behavioral changes associated with the period of metamorphic growth. The most interesting aspect of the behavioral transforma-tion of A. flavidus is the rapidity with which i t occurs. Limbaugh (1962) reported larvae forming schools near the bottom shortly after hatching. In the laboratory I observed this species forming schools within one half hour of the onset of hatching (artificially stimulated by air exposure). As soon as the larvae hatched they rose to the surface and struggled to remain there, periodically sinking in quiescence about 20 to 80 mm. Within a half hour they established a horizontal orientation and swam below the surface tension film, sinking slowly toward the bottom. Oh reaching the bottom they immedi-ately started swimming in association with other larvae at the bottom, polarized in orientation and turning into and across the current in unison (thus, schooling by the definition of Breder, 1959). d. Isolation Experiment The same isolation experiment which had been performed on Clupea harengus pallasi was performed on A. flavidus. The difference between the two experiments was that A. flavidus larvae were isolated from hatching through to 50 days age. Although feeding and sanitation conditions were not as favorable as in the preceding experiment, there were virtually no morta-l i t i e s . A l l of the fish metamorphosed completely within the 50 day period, so prevention from adopting schooling behavior was not crit i c a l to physical development in this species either. 87 Isolation had no harmful effects on A. flavidus larvae and juveniles. In fact, more rapid growth was noted in isolated individuals. Isolated fish (16) averaged 4-3.3 mm TL, whereas 12 uncrowded grouped fish averaged 36.0 mm TL and 50 crowded grouped fish averaged 23.7 mm TL. This difference in growth was probably due to feeding conditions. Although food densities were lowest for the isolated fish at the time of feeding, this species regularly went unfed for periods of one or two days. During that length of time food was fi r s t eliminated in the more crowded tub of grouped fish, then in the less crowded one. Some food was always present in tubs with isolated fish. . 8 8 III, Discussion A. Intertidal Fishes The variety of intertidal species studied provides a basis for comparisons, especially within the Family Cottidae. The NE Pacific fish which I studied had a planktonlc larval stage averaging 5 3 days, one hatch settling over about 1 2 days (days 4 7 - 5 8 , average). Usually there were early signs of metamorpho-sis and alterations in swimming behavior which preceded settle-ment. Some species hovered over the bottom for long periods before settling. In the majority of species, definite substrate preferences existed. The brief period prior to final settlement when species such as Pholls laeta. Bbthragonus swani or Leptocottus armatus altered their swimming behavior (swam cross-current) probably constituted an important*part of the behavioral transformation. These species were seen swimming during metamorphic color changes. During this period they must have repeatedly settled in order to sample the different substrates. The differences in the abruptness of settlement cannot be fully explained on the basis of substrate preferences (the range of preferences generally corresponding to the adult niche breadth). For example, Artedlus lateralis. Xiphister atropurpureus and Gobiesox maeandricus showed the most abrupt settlement. The latter two species had very narrow substrate preferences at trans-formation, in marked contrast to the overall lack of preferences in A. lateralis. These three species, however, tend to be more strictly intertidal than the others in this study, as indicated by the fact that their spawning centers above the zero tide level. 89 The only species for which field data on settlement was readily available was X. atropurpureus* Newly settled fish occurred in the center of the adult depth range (i.e. in the lower intertidal), Therefore, sampling of substrates and settlement may occur during a single high tide. The abruptness of the settlement of these three species may have evolved under the influence of the tides. That some fish which had settled into trays, particularly trays with less suitable substrates, tended to flee those trays upon tray removal from the tank probably indicates that they were only temporarily settled. Some aspect of a preferred sub-strate, such as interstitial spaces, color shade or textural qualities, must reduce the tendency to reenter the water column (that is, reduce avoidance responses to the substrate). A fish settled on a preferable substrate might be considered to be in a consummatory situation. The six species used in the substrate tray experiment showed differing substrate preferences. Ih every species (except Artedius  lateralis) one substrate was significantly preferred over a l l others. A generalized summary of these results is charted below. H ID • H CO It) - 1 I I 9 1 5 E « CD h t CQ Artedius lateralis - 0 0 0 - x; mm Leptocottus armatus - X 0 0 0 0 Bothragonus swani - 0 X 0 0 0 -Xiphister atropurpureus - 0 X 0 0 0 m Pholis laeta 0 0 0 - 0 G X Gobiesox maeandrlcus - 0 0 0 0 0 X (key: - not investigated, 6''not preferred, X preferred) 90 Artedius lateralis had no significant preferences, but settled onto any substrate in proportion to its surface area (so that the majority of these fish settled onto the tank walls and bottom). Bothragonus swani settled primarily on gravel, but secondarily on sand. Similarly, X. atropurpureus preferred gravel but also tended to settle onto pebbles. The two species which settled onto plants (P. laeta and G. maeandricus) showed significant preferences for particular types of plants. Preferences for substrates were indicated by the varying behavior of a species on different substrates. For example, P. laeta clung firmly within the bases of Phvllospadix scoulerl blades being washed with seawater, yet fled from gravel as soon as the tray was disturbed. Xiphister atropurpureus burrowed into gravel when disturbed (P. laeta never burrowed into gravel). Bothragonus swani would flee sand while the tray was being removed, yet would repeatedly catch onto a gravel substrate with its hooked body plates while that gravel was being washed around. Gobiesox maeandricus would either flee a plant and take up a new position or remain fast when disturbed. This response seemed to depend on the suitability of that plant as a substrate. Gobiesox  maeandricus which had settled onto unsuitable plants such as Fucus  gardnerl fled after the netting of the plant. Thus, the behavior of a fish on a substrate illustrated its degree of preference for that substrate. The results of runs using a r t i f i c i a l substrates indicate that the structural characteristics of the substrates rather than their chemical composition caused the preferences. For example, Xiphister 91 atropurpureus preferred gravel over larger or smaller grades of inorganic substrate. However, there was no preference for coarse textured gravel over rounded gravel. Settlement into the plexi-glas peg tray demonstrated that distinct preferences existed for particular interstitial space sizes. It should be pointed out that the range of interstice sizes in the peg tray was limited to the range which would occur in gravel and pebbles. The size of interstices in a substrate rather than the type of material seemed to be the most important parameter for thi3 species. The degree of shading was a secondary factor. Pholis laeta preferred to settle into Phvllospadix scouleri or plastic of similar width (no preference existing for the plant over the plastic). Zostera marina, another eelgrass with wider blades, and plastic strips as wide as Zostera were not preferred. Pholis laeta wrapped itself around the blades of eelgrass or plastic. Broader blades were too wide for the fish to securely entwine themselves among the blades. Similarly, although very l i t t l e settlement occurred on Fucus gardneri (none when presented simultaneously with Phvllospadix), considerable settlement occurred on Odonthalla flocossa (about one third the number that settled onto Phvllospadix). Fucus has large dimen-sions relative to the body size of P. laeta. whereas Odonthalla has relatively small dimensions, so that these fish were able to entwine their bodies among the blades of the Odonthalla. Therefore, in this species the attraction to a substrate again seemed to be determined by the dimensions, not the composition (and perhaps not the structure) of the substrate. 92 Pholis laeta settled into eelgrass and X. atropurpureus into gravel, although both species are similar in size and morphology. Also, P. laeta remained photopositive after settling while X. atropurpureus showed an abrupt change to photonegative tendencies at the end of its behavioral transformation. The translucent eelgrass provided light in the protected spaces among the blades whereas the gravel into which X. atropurpureus burrowed was opaque. The phototactic tendencies of the trans-formed fish reflected their habitat preferences and would have reinforced the proper selections. In the field, eelgrass provides shifting interstitial spaces. Gravel, especially under rocks, provides very stable, protected interstices. Although larval X. atropurpureus lacked any preference for gravel situated under rocks (as opposed to open gravel) when they transformed in the laboratory,, they tended to aggregate in such spots in the field. The difference could be due to both the higher light intensities and the wave action in the field, which would cause juveniles to seek darker, more stable interstices under rocks and boulders. In any case, gravel affords greater resistance to water movement than float-ing blades of eelgrass, whieh may reflect differences in the extent to which these species are thigmotactic. Field data suggest Zostera to be a more typical habitat of P. laeta than Phvllospadix. despite the preference for the latter in vlaboratory tests. The smaller dimensions of the 93 Phvllospadix probably make i t more suitable for settlement, but at low tide levels direct wave action could make its inter-s t i t i a l spaces very unstable (although the blade bases are rigid and may be protected from wave shock by the blades).. A more likely explanation for the difference between adult field data and laboratory settlement results is that Zostera grows in sheltered, sedimented areas offering no other interstitial spaces, whereas Phvllospadix is common to rocky shores with a diversity of algal forms and suitable grades of hard substrate. The adult P. laeta which live in diverse microhabitats of rocky shores may have settled predominantly in clumps of Phvllospadix. then later generalized their habitat preferences to include other substrates. The pronounced thigmotaxis of X. atropurpureus was typi-fied by the behavior of these fish in the peg tray. Over seven percent of the settlement into this tray (9 in 121) was into pegs with minimum separations of 0 . 5 mm, less than half the average body width. Of course, this was possible because there was a larger space between four round pegs than the minimum space separating any pair of pegs in the tray. However, these individuals were usually stuok in the pegs, some of them dead, and a few were killed in the process of being extricated. The pegs were in fixed positions whereas gravel can be shifted. Observations of X. atropurpureus during attempts to separate them from gravel substrate indicate that to a limited extent these fish will ' 9 4 shift the gravel around with their wriggling movements. This contrasts to the behavior of P. laeta. Observations of unmetamorphosed P. laeta which had settled onto plastic strips did not reveal any efforts by the fish to force their bodies between blades which were not widely enough separated. Rather, they gently probed until they found a suitable space. In the field, continual shifting of the eelgrass would facilitate the success of such probing. The combination of differences in thigmotactic and phototactic tendencies between P. laeta and X. atropurpureus lead to their different substrate preferences. An interesting point is that although none of the larvae in the „substrate tray experiments showed preferences for pebbles or rocks, the adults of a l l these species except Leptocottus  armatus can be found on boulder beaches which have pebbles and small rocks in the under-boulder substrate. The tendency for some species to settle into smaller-sized substrates (gravel) may be due to the relation of the size composition of the substrate to the body size of the fish. This is certainly the case with X. atropurpureus and may be the case with Bothragonus swani. Adult Pholis laeta also occur on boulder beaches but tend to be associated with plants like Phvllospadix. The same is true of newly transformed juveniles. Adult pholids which live in eelgrass beds are simply more forceful than the newly trans-formed juveniles in moving among the blades, and create 95 close-fitting interstices within the pliable plant material. .No shift in substrate preferences is necessary since the same eelgrass can accommodates diversity of body sizes. With GobleBOx maeandricus. however, the newly settled juvenile had a different substrate preference than the meta-morphosed juvenile (or adult). This species settled onto plants (which occur on rock substrates) at transformation, then clung to rock surfaces when larger. The observations of Shiogaki and Dotsu (1971) suggest a similar shift in substrate preferences for the gobiesocid Aspasma minima. The shift in substrate preferences of G. maeandricus is because of differences in textural composition of the substrates, rather than differences in size composition. Pebbles, rocks and boulders a l l have coarse surfaces compared to plant surfaces. The observed tendency for larvae of fishes having suction discs to settle onto new plant growths or plastic surfaces which have no epiphytes or other growths (bacterial or fungal) probably relates to the texture of these surfaces. The relatively coarse texture of rock surfaces probably does not permit successful suction by the pelvic adhesive disc of the small, newly settled fish. More than the increase in the size of the disc after settlement, the change in morphology probably permits the change in substrates for juvenile G. maeandricus. The thin fin membranes and skin flap of newly settled fish likely serve only as a suction disc in the true sense. The development of papillae probably adds 96 a gripping function to the disc, permitting adhesion to rougher surfaces. Arita (MS) describes papillae of this species as func-tioning to create a watertight seal and to prevent slipping (i.e. to grip) The plants on which G. maeandricus settle sway with the surge, so that the fish create less resistance to water movements than they would on a stationary surface.. This may aid their efforts to remain on the plant. As mentioned, formation of the adult melanin pattern occurs only after the shift to rock substrate, permitting cryptic transmission of the plant base color by newly settled fish. Gobiesox maeandricus was the only species for which the results of the light/dark substrate choice experiments disagreed with the responses to a light intensity gradient (phototaxis tank). These results were obtained with fish not yet fully metamorphosed. The transformed fish, although less responsive than larvae, tended s t i l l to be photopositive in the light gradient. However, they preferred to settle on the dark section of the peg tray underside rather than on the translucent section. This suggests that the response to light intensity in this species is more of a photokinetic than a phototactic response. That i s , i f a transformed fish becomes active in the phototaxis tank i t will tend to remain active while swimming toward the lighted end, whereas i t would not be stimulated i f i t turned toward the dark end. When settling (becoming inactive') the fish would tend to become inactive under a dark surface, whereas under a translucent surface (where light strikes the fish as i t would i f they were on top of a rock) 97 they would tend to remain active rather than settle. This could be one mechanism which is responsible for these fish being found resting on the undersides of boulders at low tides. The light/dark substrate results agree with the phototaxis tank results for Leptocottus armatus. Xiphister atropurpureus and Pholis laeta. Leptocottus armatus and P. laeta remained photoposi-tive after transformation. Leptocottus armatus preferred light over dark sand and P. laeta preferred translucent over black strips of plastic. Xiphister atropurpureus. however, showed a reversal of its phototactic tendencies at moderate intensities (500-100 lux) after transformation. This species preferred black over translucent pegs. This change in photopositivity would reinforce their tendency to worm down into the gravel where i t is darker as well as more stable. As previously mentioned, both Gobiesox maeandricus and Xiphister  atropurpureus settled into substrates different from those typical of the adult. In both species these differences are due to factors relative to body size. Xiphister atropurpureus show preferences for interstitial spaces slightly larger than their cross-sectional body size, so that a gradual shift occurs into larger-sized substrates as the fish mature. Rocky shores often have horizontal exposure gradients resulting in corresponding gradients of substrate size so that this shift in substrate preferences occurs with the fish simply moving along the beach. In the absence of such a gradient the shift must occur from fine to coarse-grained patches on the j shore. In the case of transformed-G. maeandricus the shift in 98 substrate preferences, from plants to rocks, occurs abruptly. However, the new growths of Ulva and kelp onto which the larvae settle grow on rocks, the adult substrate, i n the depth range typical of adults (low intertidal, shallow subtidal). This sh i f t i n habitats therefore involves no migration along the shore. The shift results from metamorphic changes in the adhe-sive disc, together with size increase. Leptocottus armatus larvae showed strong preferences for sand in the laboratory (mud was not tried).. Adults l i v e on soft bottom. There is a tendency for settlement to occur in estuaries and for juveniles to move into freshwater (Jones, 1962). The larvae used i n my laboratory studies were dipnetted as they swam into the freshened surface layers at the mouth of Bamfield Inlet. A primary factor in settlement of the larvae may there-fore be attraction into a gradient of decreasing s a l i n i t i e s before sampling bottom types. Although s i l t y bottom types (typical of estuaries) are preferred, larvae entering an estuary with a less suitable rocky bottom apparently settle with success -(BC 60-413). Artedius l a t e r a l i s showed no preferences, settling on different substrates in proportion to their surface areas. The great diversity of the adult habitat of this species i s reflected in the generalized settling tendencies of transforming larvae. Bothragonus swani showed two types of preferences, one 99 for gravel substrate and one for moderate current (as opposed to s t i l l water). These results agree with the limited informa-tion on this species* adult habitat (exposed rocky shores). That currents can affect transformation in some species is suggested by observations of a limited number of Nautichthys  oculofasciatus. I have observed adult individuals of this species swimming along cracks and angles in rocky faces. This corresponds to the laboratory observation of in i t i a l l y settling larvae hovering in vertical comers of the hexagonal tank. The tendency for fish in this early stage of settlement to reenter the water mass in currents above a minimum velocity rather than to sink and locate a protected spot on the bottom (as do fully transformed juveniles) would serve as a transport mechanism. Larvae starting to settle in areas of strong current along shore areas with few protected spots in the substrate would reenter the water mass for brief periods, resettling periodically until an area sheltered from currents was found. There is some evidence that repeated failure to find a protected habitat would result in habituation* to currents, so that the fish would settle in a less suitable, exposed area. This stage of hovering was also typical of the agonid Bothragonus swani and another cottid, Rhamphocottus richardsoni. Hovering over the substrate seems to be a common mechanism for exploiting drift transport. * habituation: diminishing response to a repeated stimulus 100 Blepsias elrrhosus. unlike most of the other intertidal fishes studied, would never swim out of the current in a tank and hover at the walls. Different metamorphic stages of this f i s h , some f u l l y metamorphosed, remained pelagic u n t i l Macrocvstis was introduced,, resulting i n the rapid trans-formation of a l l the larger specimens. Field data indicate that adults occupy both kelp and eelgrass beds. Macrocvstis beds occur in moderate exposure whereas Zostera beds are characteristic of protected, often estuarine, waters. Both plants form large beds in the shallow subtidal which dampen water movements. These beds also extend to the water surface during lower tides. Since untransformed B. cirrhosus swim near the surface and remain in currents they would tend to encounter these nearshore beds of seaweed where settlement occurs. Individuals swept into i n t e r t i d a l regions at high tide would find the greatest protection from surge and turbulence i n tide-pools with Alarla (which grows to great densities in rocky pools). Blades of Alaria are of a similar size to those of Macrocvstis and, when in a tidepool, Alaria would provide protection from water movements in the same way as Macrocvstis would i n a dense bed. The effect of kelp and eelgrass beds in minimizing water movements i s probably an important factor which induces this species to settle i n such habitats. Aquarium observations on large juveniles indicate that they continue to closely associate with kelp blades at larger body sizes. Certain 101 individuals grew to large sizes (after metamorphosing f u l l y ) before settling. This indicates that there i s a considerable period during which this species may delay transformation (until the proper habitat i s encountered). Museum data on ecological characteristics of collection areas rarely permit determination of the microhabitats of different species. This i s because the usual shallow water collection techniques (rotenone and beach seining) sample over broad areas, preventing much discrimination between small-scale types of habitat (e.g. Phvllospadix clumps. gravel or pebble patches, crevices). Adults of many species often behave abnormally when intro-duced into aquaria, hiding wherever the greatest cover i s available. Therefore, observations of the tendencies of laboratory reared larvae when they transform from pelagic to benthic behavior may be the most accurate way to assess innate habitat preferences of different species, since such f i s h show normal feeding and growth under laboratory conditions. For example. Nautichthys  oculofasciatus. Blepsias eirrhosus and Bothragonus swani a l l tend to be collected from rocky shores, either around kelp beds or in tidepools. The nature of the behavioral transformation of these species under laboratory conditions suggests that in such a rocky intertidal area B. eirrhosus would be associated with the kelp canopy whereas the other two species would occur on the bottom. The N« oculofasciatus would l i k e l y be found 102 in a spot protected from currents and the B. swani on a patch of gravel swept by the current. This entire study has been concerned with the types of habitats which are selected by neritlc fish larvae at the stage of the behavioral transformation from the plankton. The question remains as to how the larvae in the plankton come inshore when they approach the transformation stage. Scuba diving observations are that such larvae accumulate inshore, around islets, rocky points and kelp fringes. Miller (1974) observed marked increases in the abundance of fish larvae in the near shore regions of the Hawaiian Islands. Since these increased densities were not matched by comparable increases in the density of invertebrate zooplankton or passive elements such as eggs. Miller suggests that the larvae might tra-verse current shears into slow inshore waters on the basis of an optomotor response. Miller proposes that such larvae may see the shore or reef and make a directed response.toward i t . Neritic fish larvae may well cross near-shore current shears when they drift into the vicinity of such shears. Harden Jones (1968) explains that a fish within a uniform current cannot orient itself to the current without tactile or visual cues. Larval fish have very limited ranges of visual discrimination (e.g. Hunter, 1972-Engraulls mordax) and i t seems improbable that they would rely on long range visual detections, even i f they had that capacity, since many larvae develop during periods of phytoplankton blooms dense enough to preclude long range vision. 103 More probably the larvae respond to accelerations that would be experienced when entering a current velocity gradient. Current shears created by the boundary drag of kelp beds or prominences would also have turbulence providing additional accelerations. While this thesis was in the exploratory stage, I conducted pilot studies with an apparatus which I designed to create a variable speed current shear in an optically uniform environment. Larval Scorpaenichthys marmoratus. Gilbertidia sigalutes and Hexagrammos decagraremus had marked tendencies to cross the current shear. With the larval G. sigalutes. too sluggish to swim against rapid currents ( 10 cps), the shear was slowed to the extent that a complex circulation with three gyres formed, the central one having an eddy. In this system the G. sigalutes tended to cross the most intense current gradients and then to remain in the eddy. Larvae which orossed current shears upon encounter would be brought close enough to shore to visually contact stationary objects such as kelp plants or rock outcroppings. Once within sight of shore, a larva nearing transformation might tend to maintain its position on the basis of visual fixation on sta-tionary objects. I have repeatedly observed schools of late stage Xiphister atropurpureus larvae hovering in the lee of rock outcroppings or in kelp beds. Similar schools of Gobiesox  maeandricus are very common (but larger in numbers)! Cottids form smaller groups which show some characteristics of schools. 104 In species without swim bladders (larval cottids, agonids, stichaelds and pholids) swimming always has a slight upward angle. This angle Increases visibly in X. atropurpureus. an elongate species, when the larvae develop to the transformation stage. By this stage ossification, elimination of the sub-dermal space and other metamorphic changes could have increased body density to the extent that overcoming the tendency to sink would become fatiguing. This might induce the larvae to swim across the boundary current gradient to the shore where periods of relaxed swimming would cause repeated contacts with the sub-strate. Individuals of a particular species would avoid settling on substrates imparting the wrong tactile cues. Upon contacting the species-typical substrate final settlement would occur. The final settlement is followed by more or less rapid completion of metamorphic development, depending upon the species. The most significant change in behavior at final settlement may be the abrupt reduction in overt activity. (Transformed juveniles are as sedentary as the adults of their species.) This drop-off in energy output for swimming activity may permit1 channeling of metabolic energy into physical develop-ment. That considerable energy input is involved in these final metamorphic changes is suggested by the 21 * reduction in maximum body width of X. atropurpureus during the first four days after their transformation. 1 0 5 B. Gilbertidia sigalutes. F. Cottidae The behavioral transformation of this species involved alternation between pelagic larval and benthic adult behavior together with a gradual change in the vertical migrations typical of the larval stage. There is literature describing different vertical migration patterns in larvae of different marine fishes, as well as literature arguing over the validity of such observations. Considering the existing confusion over vertical migrations in the plankton, the larval vertical migrations of G. sigalutes should be dealt with before discussing the changes in those migrations during the long transformation period. Stevenson (1962) found that the Barkley Sound population of Clupea harengus pallasi larvae tended to migrate during the older stages, which were usually distributed deeper than the younger stages (although the data were not fully significant). Colton, et al.(i960) demonstrated both net avoidance and migration in larval C. harengus harengus. noting that their nightiday catch ratio decreased with depth. Ryland (1964) found Pleuronectes platessa larvae to be stratified at eight meters depth during daylight and dispersed at night, which really cannot be interpreted as indicating vertical migration. In a laboratory investigation, Blaxter (1973) used a vertical light chamber with an automated observation system to monitor diurnal movements of various growth stages of larval C. harengus  harengus and P. platessa. Both species tended to be neustonic 106 at hatching* The older stages of both species, however, would rise to the surface at night and remain near the surface until dawn. These results for P. platessa conflict with Ryland's data* Other patterns of migration occur in other species• Hartmann (1970) reports Scomberesox saurus larvae remaining neustonic and then migrating during the "juvenile" stage. Bartlett and Haedrich (1968) concluded from a neuston towing survey that larval Makaira nigricans migrate to the surface at dawn and remain there until dusk, since these larvae were almost absent from night neuston hauls. Contradictory statements have been made about vertical migration in larval Sebastes marinus. Magnusson, et al*(1965) conclude that these larvae migrate to a more marked extent than the invertebrate zooplankton. Kelly and Barker (196l), however, state that this species does not vertically migrate. Raitt (1964), correlating echo sounding data to the depth distribution of S. marinus larvae also concludes that no migrations occur with these larvae. Russell (1926) was one of the first to propose that some marine fish larvae migrate vertically. Russell considers that Sardlna pllchardus and Callionymus lyra larvae migrate to the surface at night since he caught increased numbers, in night-time surface hauls. However, his data show no concommitant decreases in numbers caught at depth during the night. Furthermore, this upward increase in the nighttime distribution of C. lyra larvae did not occur on moonlit nights. Bridger 107' (1956) suggests that the sort of data presented by Russell results from visually mediated net avoidance. Bridger observes that the night:day catch ratio depends upon the size (swimming capability) of larvae. The significant increases in dusk and nighttime surface catches of Gilbertidia sigalutes larvae could not have been due ton daytime net avoidance since subsurface catches at dusk and night showed comparable decreases. Also, surface catches were much more predictable on moonlit than on dark nights. The vertical migra-tions of larval G. sigalutes were usually most pronounced during dawn and dusk. These larvae also showed definite responses to downwellings, so that the character of their migrations varied with weather conditions and in different areas. Together with the marked vertical migrations and aggrega-tions at downwellings, two other aspects of the larval d i s t r i -bution of G, sigalutes merit discussion. Although this species has a very low fecundity for an egg laying marine teleost the larvae were caught in relatively great abundance. The only larvae which I have consistently caught in greater numbers during any season have been Clupea harengus pallasi and Thaleichthys  pacificus. both extremely prolific species. The other interesting aspect of the distribution of G. sigalutes larvae is the confined area in which they were taken. The high densities must result from the constricted areal distribution. The tendency for larvae to swim against downwellings in dark-ness could be to prevent dispersal in subsurface waters. Similarly, 108 vortical migration to the surface for brief periods at dawn and dusk would minimize drift due to surface transport. Migrating ten the surface during twilight periods would permit feeding to continue when lighting at depth would be inadequate for the visual feeding of these larvae. Also, surface concentrations of plankton very often peak during twilight. In order for vertical migrations to evolve as a mechanism for limiting horizontal drift,.there must be strong selection favoring limited dispersal. Spawning in the laboratory and in cages in the field indicated that group spawning occurs. One male synchronized maturation in several females which laid their eggs simultaneously. The females tended the egg mass while the male patrolled the nest site area. Reproductive success probably depends to some extent upon maintaining a -minimum population density. This may be achieved through limitation of larval drift dispersal. In addition to group spawning, another factor which would select for limited dispersal is the rarity of the adult habitat. Since this species occurs in sheltered sites inside crevices or among boulders, suitable habitats would only be common in protected regions with subtidal c l i f f s and rubble piles. Such areas tend to be separated by extensive stretches of gradually sloping soft bottom. Since such subtidal rocky areas would be stable environments compared to intertidal terrain, long range dispersal might be less important than maintenance of critical population density. During the transformation period G. sigalutes migrated to the surface less and less frequently until they had settled permanently. The intermittent migrations probably indicate repeated settlement. Considering the low probability of an individual settling in the vicinity of a hole or crevice not occupied by an.adult, the reentries into the plankton probably function as a dispersal mechanism at this stage. The seeming paradox of having evolved ways to limit dispersal during the pelagic larval stage and having evolved a tendency to reenter the pelagic zone partly for the purpose of dispersal during transformation can be explained in two ways. The limit to lar-val dispersal is macrodistributional. If this limitation of dispersal is successful, then transforming Juveniles need only search on a micredistributional basis for suitable homesites. If they were to settle in an area of soft bottom and attempt to traverse the area in search of a suitable, protected spot they would be highly susceptible to predation by benthic or epibenthic organisms. If they remained s t i l l during the day, however, their dull coloration would serve as camouflage. Reentry to the plankton for brief periods would bring about limited drift transport. The bottom collections by the B.C.. Provincial Museum demonstrate that these transforming individ-uals do settle onto unsuitable soft bottom and the adults only occun on rocky slopes and c l i f f s in protected waters. Another difference between the larval and transformation stages is the nature of their vertical migrations. Larval 110 stages seemed to remain at the surface at night only i n down-welling areas or in moonlight. Although transforming juveniles always migrated during darkness they were never taken at night at foam lines, which suggests that f i s h in this stage are proba-bly not responsive to downwelling. Juveniles are probably large enough not to be entrapped i n downwelling. Not aggrega-ting within downwellings could permit the transforming juveniles to d r i f t or move randomly in the horizontal so that the sam-pling effectiveness (with regard to resettlement) would be greater. The other aspect of migrations during the transforma-tion stage i s their correspondence to high surface plankton densities, underscoring the importance that the f i s h exploit the plankton during this stage of rapid growth. Nocturnal feeding on zooplankton in surface waters i s evident for trans-forming G. sigalutes from the f i e l d . The-shift-to migrations in increasing darkness probably relates to a s h i f t from visual to "distant touch" feeding. One important difference between the laboratory and f i e l d situations was that the migrations of transforming f i s h i n the f i e l d occurred with relative uniformity. That i s , a considera-ble portion of the population would migrate on a particular evening. Furthermore, migrations often took place on several consecutive evenings, then ceased entirely for a longer time. In the laboratory a f a i r l y constant proportion was resting on the bottom. No observations were made to discern whether diurnal activity patterns took place i n the laboratory during this stage. Of course, the laboratory conditions were constant I l l relative to the natural environment. Changing weather and feeding conditions may have induced the irregularity of the migrations in the field. In the laboratory the proportion of transforming fish resting on the bottom significantly related to feeding conditions, more settling when unfed than when feeding. These laboratory comparisons were made over <•• brief periods. In the field more prolonged periods of low or high plankton abundance or longer term changes in conditions which affect zooplankton migrations could have affected the migrations of G. sigalutes. The effect of repeated settlement on the growth and metabo-lism was not investigated. Transforming G. sigalutes caught in the plankton have usually had very f u l l gut contents. This large energy intake must be accompanied by high energy-costs for swimming since this species is negatively bouyant. The periods of sedentary habits could permit the rapid growth evident in this stage. A l l of the other species I have encoun-tered metamorphose (and transform) at a small fraction of their adult size. A prolonged transformation in G. sigalutes may have evolved as a mechanism of exploiting the rich larval food source (plankton) for growth to a size adequate for sexual maturation (larval behavior supporting juvenile growth).' Laboratory observations of transforming G. sigalutes avoiding settlement in the presence . of cannibalistic adults illustrate another aspect of the problem of finding a suita-ble habitat. If a G, sigalutes settles in a suitable area already occupied by adults i t risks being eaten whenever 112 i t enters a protected spot. I f such individuals can detect the presence of adults in time to escape, then they may search for nearby cracks or reenter the plankton. The a b i l i t y to repeatedly/settle affords an effective population density regulating mechanism. Aside from regulating densities, the protracted period of repeated settlement could allow the adult spawning popula-tion to exploit the spring plankton blooms through the new year class. The productivity of subtidal crevices during different seasons would be d i f f i c u l t to assess. However, adult G. sigalutes seem to have evolved the capacity to withstand starvation and then to mature and spawn after eating just a few settling juveniles. In areas densely populated with adults, the repeatedj efforts of immature f i s h to settle could permit these adults to gain tremendous energetic input by preying on the newly settling f i s h . This could provide an energy base for the i n i t i a t i o n of sexual maturation in the population. 113 C. Schooling Species 1. Clupea harengus pallasi. F. Clupeidae The significantly lower degree of metamorphic development in-the isolated fish could be solely due to slightly slower growth rates since the unmetamorphosed fish averaged shorter lengths. Although prevention of schooling with conspecifics might somehow affect the behavior (e.g. feeding efficiency) or the physiological state in a detrimental fashion* the slightly reduced growth of isolated fish may simply have resulted from the lower food densities they were offered. A third possibility is that the growth differences resulted only from the natural variation in growth rates together with the small numbers involved. The number of larvae to be isolated was chosen on the basis of the number whichiwould be adequate to demonstrate significant mortalities resulting from prevention of schooling during the stage of metamorphosis. No such mortalities occurred. This experiment casts doubt upon Stevenson's (1962)'suggestion that metamorphosing larvae too widely dispersed to form schools would die. Schooling is not crit i c a l to metamorphosis into the juvenile form, but isolated juveniles might suffer without the advantages of schooling behavior. Another possibility would be that larvae f a i l to survive i f isolated before fin ray formation, the stage when schooling behavior develops. Schooling develops gradually in both C. harengus pallasi and in the atherinid Menidia menidia. 114 Williams and Shaw (1971) reared M. menidia in isolation well beyond the end of the stage of development of schooling, evi-dently without significant mortalities. They noted that paral-l e l swimming starts at 15 days age in this species and school-ing is fully developed at 25 days, exactly the spanrfor school-ing development in C. harengus pallasi. Isolation during the period of schooling development probably would also not be critical to the survival of G. harengus pallasi larvae. Isolationsof early larval stages of C. harengus pallasi which are undergoing the onset of schooling probably occurs only to a slight extent ln Barkley Sound. Stevenson (1962)• in following the drift dispersal of larvae from the Toquart Bay spawn, found that larvae are f i r s t carried to the mouth of Barkley Sound 14 days after hatching. The numbers drifting out to sea increase from that point on. Since schooling ten-dencies start developing by day 16, the majority of larvae swept offshore have already developed some schooling tenden-cies and may have had schooling experiences withhconspeeifics. Assuming that larvae isolated at sea do not return-as adults (Stevenson, 1962), rather than the behavioral transformation being critical to successful metamorphic development, i t is likely a case of isolated herring larvae developing to the juvenile stage but never surviving to maturity. If advanced herring larvae were to survive in isolation better than juveniles, l t would probably be due to the difference in their feeding behavior. Until the completion of metamorphosis, larvae feed strictly by biting (visual fixation on single prey). Rosenthal (1968) describes the breakdown of larval schools whenever feeding occurs (larvae do not f i l t e r feed). I fi r s t noted f i l t e r feeding in juvenile herring well advanced beyond the end of metamorphic growth. It was during f i l t e r feeding that the best defined schooling was observed in the rearing tank. Schooling disappeared completely whenever a large size fraction of plankton was introduced, stimulating visual (biting) feeding inrthe juveniles. Berry and Richards (1973) note that g i l l raker formation continues into the juvenile stage in fish species in"general, which explains the late development of fi l t e r feeding. 0?Connell (1972) calculates that adults of the clupeoid Engraulis mordax cannot sustain their metabolic demands without a combination of f i l t e r feeding and "biting 0 feeding. Leong and 0*Cdnnell (1969) associate the abandonment of schooling in adult E. mordax only with intense "biting" feed-ing. One possible explanation of why juveniles dispersed at sea would f a i l to survive to maturity would be that f i l t e r feeding may only be expressed by elupeoids during schooling., I know of no literature, however, on the feeding of adult elupeoids in isolation. 116 Schooling Species 2. Aulorhynchua flavidus, F. Gasterosteidae All of the species of neritic fish larvae I have observed swam vertically to the surface Immediately after hatching, then established horizontal swimming posture under the surface tension film. The surface is probably the best baseline against which larvae can gauge their postural orientation. Perhaps more impor-tant, however, is that rising to the surface permits entry into surface current drift. The onset of schooling in A. flavidus occurred abruptly, before any metamorphic growth had occurred. However, the most significant ecological difference between the abrupt behavioral transformation of this species and the gradual transformation of Clupea harengus pallasi is the difference in the temporal extent of their periods of drift dispersal. Aulorhynchus  flavidus had the briefest planktonlc stage of any fish I studied. The brevity of the period bf pelagic drift in A. flavidus relates to the unusually protracted nature of its hatching. Morris (1956) describes the hatching of egg masses of inter-tidal fishes as "explosive." I have found this synchronized type of hatching is typical of the intertidal and shallow subtidal species with which I have worked. Protracted hatch-ing may have evolved as a means of maximizing planktonic drift dispersal' by different phases of tidal currents and shifting patterns of wind-driven surface currents. However, this highly randomized type4 of dispersal is only effected over a limited distance owing to the very rapid transformation to a home site-specific, schooling behavior. 1 1 7 As long as an egg mass of A. flavidus remains fixed in the spawning site maximum dispersal is ensured within a suita-ble environment (where parents have succeeded in spawning)• However, this strategy severely limits long range dispersal and could presumably limit genetic flow between isolated popu-lations. Larval drift in A, flavidus is undoubtedly much more limited than the swimming capabilities of adults. Under changing environmental conditions schools of adult A. flavidus might emigrate to a new home site in their foraging efforts, but as mentioned, adults do not range beyond shallow-bottom waters. Larval drift in A. flavidus may be too brief to permit passage across channels or other open bodies of deep water, but this species has evolved a long range dispersal mechanism with certain advantages over long distance emigration or larval drift. This species lays eggs which are adhesive to eaoh other but not to the plant blade stem around which they are laid. They^also tend to select delicate plants. The consequence of choosing this sort of spawning site is that the egg masses are frequently torn off during stormy weather and carried in the wind-driven surface currents. The firm egg membranes pro-tect embryos from wave shock which would occur irr surface waters during windstorms. Larvae would probably be much more suscepti-ble to mechanical damage during dri f t transport in windstorms. Furthermore, an egg mass would not be adversely affected by drifting within a water mass of low plankton abundance for an extended period. 118 Storm dispersal of egg masses seems to occur commonly in this species, and parental behavior appears adapted to periodic losses during storms. New egg'masses are spawned to replace those loosened by storms. The nest-guarding male might be stimulated to resume courtship behavior (disinhibited)iupon removal of the egg mass in the same manner as a reproductive male of Gasterosteus aculeatus (Bbl, 1959; Sevenster-Bol, 1962)} which is related to A. flavidus on the family level. This tendency to replace egg masses carried away by storms would increase the probability that hatching would occur in the home site area where larvae would have a high probability of surviving. Egg masses drifting with surface currents would settle out at some point since they are slightly negative in bouyancy. This would likely happen in calm spots such as inshore around kelp beds or in protected bays, suitable areas for larval development. Although many egg masses would be lost due to predation, sinking in unsuitable areas or washing ashore, very extensive drift could occur during windstorms. Since wind-driven surface currents average 3-5$ of the wind*s speed (Fairbridge, 1966), a storm averaging 30 mph winds for 48 hours could carry an egg mass 4-3-72 miles (69-115 km). The tendency to spawn egg masses on fragile plant material may have evolved not only so that egg masses would break free easily, but also in order that some plant material would remain with the egg mass (as observed). Plant material would greatly increase surface area without much effect on density, increasing flotation in moderately turbulent waters. Since hatching occurs 119 during 25 to 75* of the time that an egg mass exists, there i s a high probability that an egg mass, i f broken free i n a storm,, w i l l have already started hatching. Furthermore, after a certain amount of hatching the security of the mass* attachment to the plant would decrease, increasing the chance of storm dispersal during the l a t e r hatching of the egg mass. Limited hatching would occur en route. Whenever the egg mass settled the pro-tracted hatching would allow for short range larval d r i f t d i s -persal in-changing t i d a l currents. Aulorhynchus flavidus has a period of larval d r i f t lasting less than one hour, the larvae forming bottom-associated schools as soon as they establish horizontal orientation. Other species have larval d r i f t periods over 1000 times as long. Aulorhynchus  flavidus has evolved a protracted hatching period (perhaps a polymorphism for varied rates of embryonic development or for hatching at varied states of maturity) which serves to randomize short range dispersal by this brief planktonlc d r i f t . Long range dispersal i s probably mediated to the greatest extent by storms loosing egg masses from their precarious places of attachment and transporting them i n the wind-driven surface currents. The adults replace egg masses removed by storms, ensuring larval settlement within"the parental habitat. 120 D. Behavioral Transformation in Relation to the Ecology of Marine Fish Larvae Most field and laboratory studies of marine fish larvae are concerned with factors affecting mortality rates at differ-ent stages of larval development. Baranenkova (1965) correlates high mortalities in the early larval stages of Gadus morhua with stormy weather. Critical periods associated with the i n i -tiation of feeding and the end of yolk absorption are definitely linked to drift into water masses of low plankton abundance (Shelbourne, 195?! Laurence, 197*0. Certainly, drift into improper abiotic conditions could hinder development so that proper feeding could not begin before yolk absorption, as i l l u s -trated by the low temperature limits to development of larval Sardinops sagax (Lasker, 1964). My study indicates that drift into areas with improper substrates can prevent the behavioral transformation of some species. Thus, the random nature of planktonlc larval dr i f t results in critical periods in certain species at stages when major changes in behavior and physiology must occur for the continuation of development. That i s , hatching (or the onset of active swimming), yolk absorption (the initiation of external feeding) and the behavioral trans-format ionn(to adult habits) may a l l become critical periods under certain conditions. Figure 8 is a flow chart of different possibilities of what may happen during larval drift, according to the present work and the results of other researchers (such as those above). The significance of the critical period concept is that the larval stage includes different adaptive strategies and the greatest vulnerability occurs during the transitions. EATEN DURING'HATCHING OR' ASCENT TO SURFACE Fig. 8. Flow chart of l a r v a l l i f e history, detailed above the dashed l i n e . This paper Most research emphasizes "problems covers points below that l i n e . -122 Planktonlc larval stages permit predation on large numbers of suitably sized organisms while the larvae are carried more or less randomly by the currents. As exemplified by the extreme types of behavioral transformation in Aulorhynchus flavidus and Gilbertidia sigalutes. the timing and mechanisms of the trans-formation of a particular species are most important in determin-ing- the range of drift dispersal and in insuring that the suita-ble adult habitat can be located. Assuming that a species can only succeed in-the habitat to which i t is evolutionarily adapted, the behavioral transformation must then be crit i c a l at least to the genetic survival (reproductive success)'of an individual. The transformation would only constitute a critical period in the strict sense i f there were a relatively high probability of failure under some normal circumstances. The extent to which the behavioral transformation is a critical period depends upon its abruptness, its temporal re-lation to metamorphosis and the precision of substrate require-ments. Invgeneral, i f larvae of species strictly inhabiting the rocky intertidal (notably Xiphister atropurpureus and Gobiesox maeandricus) were to drift into a vast area of unin-terrupted sand or mud bottom, then the stage of behavioral transformation would become a critical period. For example, the transformation of a hatch of X.-atropurpureus occurred during about one week in the laboratory. If larvae spawned within the Broken Group (cf. Fig, 1, pg. l6))were retained within the current gyre in that area (Doe, 1952), then there would be a high probability of the larvae encountering a suitable 123 gravel substrate during the stage when settlement from the plankton should occur since the islands tend to have very patchy intertidal habitats. If, however, the X. atropurpureus were 1 to drift out of that gyre into Loudon Channel just before the stage of behavioral transformation, then they would be swept along a uniform mud shoreline for up to two weeks (Stevenson, 1962), longer than their settlement period, and total mortality of that drift contingent could follow. Selection favors mechanisms reducing the chance of a critical period and various adaptations occur which ensure that different species will find a suitable substrate in time for settlement. As discussed previously, intertidal species tend to exhibit decreased rheotactic tendencies together with signs of increased density (in species without swim bladders:: cottids, agonids, pholids and stichaeids);when the stage of behavioral transforma-tion approaches. These changes probably bring the larvae closer to shore and result in tactile sampling of different substrates, settlement occurring when a suitable substrate is encountered. Larval Leptocottus armatus are attracted into estuaries before the stage of settlement, increasing the proba-bi l i t y of their finding sediment substrates. After fin ray formation, X» atropurpureus and G. maeandricus form schools inshore. I have seen such schools only in eddies around rocky prominences and kelp beds; i f ever they were over sand i t was within one or two meters of rocky substrate. The period of accumulating in nearshore schools lasts much longer than the 124 settling period, thus increasing the chance of finding suitable substrate patches for settling. The schooling itself is probably not involved in the behavioral transformation, but simply advan-tageous to the pelagic larvae trying to remain near shore. More advanced larvae of these species probably settle as indi-viduals, perhaps after dropping out of the school on the basis of differential swimming effort (increasing with increasing body density) in a manner similar to the phenomenon of size segregation in schools. Schooling species undergo no dramatic change of habitat or activity level at the stage 6*f behavioral transformation. In these species metamorphic growth occurs relatively slowly and shows no correspondence to the onset of schooling, probably because there is never a sudden decrease in the energy required for routine activity which could be exploited for a rapid growth phase. No critical period has been associated with the onset of schooling in Clupea harengus pallasi or Aulorhynchus flavidus. Similarly, Blepsias eirrhosus showed l i t t l e correspondence between its settlement and its earlier, abrupt metamorphosis. Upon metamorphosis this species started swimming more slowly, at a marked upward angle.. The reason for the greater capacity for delayed settlement in this species is that this slow pelagic swimming of juveniles differs l i t t l e from their hovering at an upward angle amongst kelp blades. Settlement involves no abrupt change to sedentary habits in this species. An even more gradual shift in activity patterns is evident in the transformation of Gilbertidia sigalutes. which alternates to an increasing extent from being an active pelagic "larva" to a sedentary benthic "juvenile." 126 E. Suggestions for Future Research 1. The mechanisms of inshore accumulation of larval neritic fishes nearing transformation need to be determined and explained. 2. The critical nature of substrate type for transformation should be quantified. Just before transformation, larvae from one hatch could be placed into separate tanks, the bottom of each tank covered with a different substrate (e.g. mud, gravel and solid bottom). Differential survival through metamorphosis (settlement rate) would provide coefficients of substrate suitability (cf. suggestion #4). 3. The contingency of food supply could be tested or respirometry used to determine the energetic aspects of transformation. For example, a pilot study with newly transformed Gobiesox  maeandricus larvae placed into four tubs (#1 - fed, with sub-strate; #2 - starved, with substrate; #3 — fed, no substrate; #4 - starved, no substrate) showed that the transformed larvae which were fed but prevented from remaining settled died three to six times faster than starved larvae permitted to remain settled. This suggests that a relation between settlement and metamorphosis rests in the metabolic energy budget. 4. Known current patterns, general bottom types and larval development rates could be incorporated into a simulation model to predict the ranges of mortality rates which could result from drift patterns within a particular area. 5. Just as the cottid species Nautichthvs oculofasclatus and Blepsias elrrhosus are not readily differentiated with respect 127/ to adult habitat except by examination of substrate preferences at settlement, sibling species of a complex genus like Sebastes might be similarly differentiated i f their larvae were reared in the laboratory through the behavioral transformation. 6. The mierohabitats where settlement of commercial groundfish species occurs should be studied through laboratory rearing. Adult habitats of such species are relatively well known. If settlement occurred in more restricted habitats than those in which adults were found, such findings could guide restric-tion of commercial fishing areas for protection of nursery grounds. 7. The floating top of the kelp canopy should be compared to the submerged portion with respect to settlement of Sebastes species to clarify the effects of kelp harvesting on recruitment. *28 IV. Summary A. In this thesis the relationship between the behavioral transformation and metamorphosis from the larval stage is described. In intertidal species the two phenomena occurred abruptly and a l l but simultaneously. For the intertidal and shallow water species the sequence of metamorphic changes and the behavioral transformation were ordered to permit success-ful adoption of juvenile habits without placing the untrans-formed planktonlc larvae at any extreme disadvantage. In the subtidal cottid Gilbertidia sigalutes the two phenomena occurred simultaneously, but extremely slowly (ambivalence expressed between larval and adult behavioral modes). In the schooling species the behavioral transformation (onset of schooling) preceded a gradual metamorphosis. B. The most striking feature of the behavioral transformation of intertidal species was their settlement to the bottom. The range of substrate preferences of settling larvae corresponded to the adult niche breadth. Most species (those with relatively narrow adult niches) showed significant preference for a parti-cular substrate. That substrate was an essential element of the adult habitat, although often not the actual adult substrate. Physical characteristics of substrates resulted in settlement preferences, which were probably based on tactile cues and properties of light transmission. 129 C. The protracted behavioral transformation of Gilbertidia  sigalutes consisted of an alternation between settling to the bottom and reentering the plankton. During the larval stage crepuscular vertical migrations occurred. During the trans-formation stage G. sigalutes. when in the plankton, migrated progressively later after sunset. These migrations became rarer as they became more nocturnal. There is some indication that in the field, individuals simultaneously reentered the plankton, perhaps during peaks of nighttime surface zooplankton abundance. Laboratory observations and field collection data indicate lack of discrimination in settlement sites during the briefer early settlings. Towards the end of the transformation period individuals settled for increasingly extended periods and showed preferences for crevices and other protected places (adult habitat). D. In both Clupea harengus pallasi and Aulorhynchus flavidus the behavioral transormation to schooling habits occurred prior to metamorphosis, but in neither species did prevention of schooling (by isolation) seriously hinder metamorphosis. In C. harengus pallasi schooling developed gradually, whereas A. flavidus expressed fully developed schooling as early as a half hour after hatching, as soon as conspecifics were located near the bottom. E. . The special adaptations of A. flavidus for dispersal are discussed in light of its exceedingly rapid behavioral trans-formation from the planktonlc stage. Long range dispersal 130 probably involves dri f t of egg masses loosened during windstroms. F. The behavioral transformation did not seem to constitute a critical period except in benthic species having an abrupt settlement and narrow substrate preferences. In such species, critical mortalities probably occur at transformation only when the planktonlc larvae dri f t into areas of uniform sub-strate of an unsuitable type for settlement. G5- It is hypothesized that mortality of larvae which failed to find a suitable substrate for settlement resulted from a metabolic deficit. Such a deficit could have resulted from the energy needed for metamorphic growth having been used to sustain continued swimming in mid-water. Literature Cited: 131 Ahlstrom, E.H. and R.C..Counts. 1958. Development and d i s t r i -bution of Vincinguerria lucetia and related species i n the eastern Pacific. U.S.F.W.S. Fish. B u l l . 58: 363-416. Alderdice, D.F. and F.P.J. Velsen. 1971. Some effects of s a l i n i t y and temperature on early development of Pacific herring (Clupea p a l l a s i ) . J. Fish. Res. Board Can. 28:.1545-1562. Arita, G.S. MS. A comparative study of the structure and function of the .adhesive apparatus of the Cyclopteridae and Gobiesocidae. M.Sc. Thesis. 1967. Univ. of Br i t i s h Columbia. 90 pp. Aronovitch, T.M. and L.V. Spectorova. 1973* Some data on the feeding of larvae of the turbot Scophthalmus maeoticus under laboratory conditions. In Problems of Rational Marine Fisheries and Reproduction of Marine Fish and Shellfish, (in Russian, with English summary). Trudy V.N'.I.R.O. 94: 149-156. Bailey, R.M., J.E. Fitch, E.S.. Herald, E.A. Lachner, C C . Lindsey, CR. Robins and W.B. Scott. 1970. A l i s t of common and s c i e n t i f i c names of fishes from the United States and Canada (third edition). Amer. Fish. Soc. Spec. Pub. 6: 1-149. Bainbridge, V. I965. A preliminary study of Sebastes larvae in relation to the pj-anktonic environment of the Irminger Sea. I.C.N.A.F.. Spec. Pub. 6 : 303-308. Bainbridge, V. and B.J. McKay. 1968. The feeding of cod and redfish larvae. I.C.N.A.F. Spec. Pub. 7:- 187-217. Baranenkova, A.S. 1965. Notes on the condition of formation of the Arcto-Ndrweglan tribe of cod of the 1959-1961 year-classes dur-ing the f i r s t year of l i f e . I.C.N.A.F.. Spec. Pub. 6: 397-410. Bardach, J.E. 1968. The Status and Potential of Aquaculture, Particu-l a r l y Fish Culture. Vol. 2(1). Part III. Fish Culture, (pp. 22-34 report work of G.0. Schumann on marine f i s h larvae) Barnabe, G. 19?4. Mass rearing of the bass Dicentrarchus labrax L. pp. 749-753. In J.H.S.. Blaxter (ed.). The Early Life History of Fishes. Springer-Verlag, Berlin. 765 PP« Bartlett, M.R.. and R.L. Haedrich. 1968. Neuston nets and South Atlantic larval blue marlin (Makaira nigricans). Copeia 1968: 469-474. Berry, F.H. and W.J., Richards. 1973* Characters useful to the study of larval fishes. N.M.F.S. Middle A t l . Coastal Fish. Center Tech. Pub. It 48-65. Bhattacharyya, R.Nv 1957. The food and feeding habits of l a r v a l and post-larval herring in the northern North Sea. Mar. Res. Scot. 1957 (3). 16 pp. 132 Bishai, H.M.. 1959. The effect of water currents on the survival and distribution of fish larvae. J. Cons. 25s 134-146. Blaxter, J.H.S. 1962. Herring rearing - IV. Rearing beyond the yolk-sac stage. Mar. Res. Scot. 1962 ( l ) . 18 pp. Blaxter, J.H.S.. 1969. Visual thresholds and spectral sensitivity of flatfish larvae. J. Exp. Biol. 51: 221-230. Blaxter, J.H.S. 1973. Monitoring the vertical movements and light responses of herring and plaice larvae. J. Mar. Biol. Ass. U.K.. 53:;635-647. 2 pi. Bbgorov, V.G. 1941. Biological seasons in the plankton of various seas. Doklady Akad. Nauk S.S.S.R. 31(4). (in Russian) (not available - not read) Bol, A.C.A. 1959. A consummatory situation: the effect of eggs on the sexual behavior of the male three-spined stickleback (Gasterosteus aculeatus L.). Experientia. 15:; 115-116. Breder, CM. Jr. 1939. On the l i f e history and development of the sponge blenny, Paraclinus marmoratus (Steindachner). Zoologica. 24: 487-496. 4 pi. Breder, CM. Jr. 1941. On the reproductive behavior of the sponge blenny, Paraclinus marmoratus (Steindachner). Zoologica. 26: 233-235. 3 pi. Breder, CM. Jr. 1949. On the taxonomy and the postlarval:stages of the surgeonfish, Acanthurus hepatus. Cbpeia 1949: 296. Breder, CM. Jr. 1959. Studies on social groupings in fishes. Bull. Amer. Mus. Nat. Hist. 117: 393-482. Bridger, J.Pli 1956. On day and night variations in catches of fish larvae. J. Cons. 22::42-57. Budd, P.L. 1940. Development of the eggs and early larvae of six California fishes. Calif. Fish & Game Fish Bull. 56. 50 pp. 13 pi. Clemens, W.A. and G.V. Wilby. 1961. Fishes of the Pacific coast of Canada. Bull. Fish. Res. Board Can. 68. 443 pp. Colton, J.B7;, K.A.. Honey and R.F. Temple. I960. The effectiveness of sampling methods used to study the distribution of larval herring in Maine. J. Cons. 26: 180-190. Dannevig, A. 1948. Rearing experiments at the Flodevigen seafish hatchery 1943-1946. J. Cons. 15: 277-283. 133 Delmonte, P.J., I. Rubinoff and R.W, Rubinoff. 1968. Laboratory-rearing through metamorphosis of some Panamanian gobies. Cbpeia. 1968::411-412. Doe, L.A.E. 1952. Currents and net transport in Loudon Channel, April 1950. J. Fish. Res. Board Can. 9:: 42-64. Fairbridge, R.W. (ed.). 1966. The Encyclopedia of Oceanography. Reinhold Pub. Corp., New York. 1021 pp. Fishelson, L. 1963. Observations on littoral fishes of Isreal. II. Larval development and metamorphosis of Blennius pavo Risso (Teleostei, Blenniidae). Isreal J. Zool. 12: 81-91. Fitch, J.E. and R.L. Lavenberg. 1968. Deep-Water Teleostean Fishes of California, p. 119-120. Univ. of Calif. Press, Berkeley. 155 PP# FWchter, J. 1965. Versuche zur brutaufzucht der seezunge Solea solea in kleinen aquarien. Helg. Wiss. Meeresunt. 12:: 395-A03. (German with English abstract) Fujita, S. 1966. Egg development, larval stages and rearing of the puffer, Lagocephalus .lunaris spadiceus (Richardson). Jap. J. Ich. 13s 162-168. (Japanese with English summary) Fujita, S. and K. Uchida. 1959. Breeding habits and rearing of larvae of a blennioid fish, Ernogrammus hexagrammus (Temminck et Schlegel). Sci. Bull. Fac. Agr. Kyushu Univ. 17:: 283-289. (Japanese with English summary) Garstang, W. 1900. Preliminary experiments on the rearing of sea-fish larvae. J. Mar. Biol. Ass. U.K. 6: 70-93. Gorbunova, N.N..1962. Spawning and development of greenlings (family Hexagrammidae). pp. 121-183. In Rass, T.S. (ed.). Greenlings:;Taxonomy, Biology, Interoceanic Transplantation, (translation I.P.S.T. 1970). Harden Jones, F.R.1968. Fish Migration. Edw. Arnold (Pub.) Ltd., London. 325 PP. Hart, J.L. 1973. Pacific fishes of Canada. Bull. Fish. Res. Board Can. 180. x i i + 740 pp. Hartmann, J. 1970. Juvenile saury pike (Scomberesox saurus Walb.), an example of ichthyoneuston. J. Cons. 33*: 245-255. Hempel, G. 1965. On the importance of larval survival for the population dynamics of marine food fish. Calif. Coop. Fish. Invest. Rep. 10:: 13-23. 134 Hettler, W.F. Jr. 1973. Rearing menhaden larvae (part 2). N.M.F.S..Middle Atl. Coastal Fish. Center Tech. Pub. 1: 149-157. Hirano, R. 1969. Rearing of black sea bream larvae. Symposium on culture and propagation of sea breams. Bull. Jap. Soc; Sci. Fish. 35: 567-569. (Japanese with English summary on pp. 603-604)} Hjort, Ji. 1926. Fluctuations in the year classes of important food fishes. J. Cons. l : : 5 - 3 8 . Hubbs, Carl L. 1943. Terminology of early stages of fishes. Copeia 1943: 260. Hunter, J.R. 1972. Swimming and feeding behavior of larval anchovy Engraulis mordax. N.M.F.S. Fish. Bull. 70: 821-838. Hunter, J.R. and G.L.. Thomas. 1974. Effect of prey distribution and density on the searching and feeding behavior of larval anchovy Engraulis mordax. pp. 559-5?4. In J.H.S., Blaxter (ed.). The Early Life History of Fishes. Springer-Verlag, Berlin. 765 pp. Jones, A.C. 1962. The biology of the euryhaline fish Leptocottus armatus Girard (Cottidae). Univ. Calif. Pub. Zool. 67: 321-368. Jones, A.J. 1972. An inexpensive apparatus for the large-scale hatching of Artemia salina L. J. Cons. 34: 351-356. Kasahara, S., R. Hirano and Y. Oshima. I960. A study on the growth and rearing methods of the fry of black porgy, Myllo macrocephalus (Basilewsky). Bull. Jap. Soc. Sci., Fish. 26: 239-244. Kelly, G.F. and A.M. Barker. 1961. Vertical distribution of young redfish in the Gulf of Maine. Rapp. Proc. Verb. C.P.I.E.M. 150:: 220-233. Lagler, K.F., J.E. Bardach and R.R. Miller. 1962. Ichthyology. John Wiley and Sons, Inc., New York. 545 pp. Lasker, R. 1964. An experimental study of the effect of temperature on the incubation time, development, and growth of Pacific sardine embryos and larvae. Copeia 1964:: 399-405. Laurence, G.C. 1974. Growth and survival of haddock (Melanogrammus  aegleflnus) larvae in relation to planktonlc prey concentra-tion. J. Fish. Res. Board Can. 31::1415-1419. 1'35 Leong, R.J.H. and CP. 0*Connell. 1969. A laboratory study of particulate and f i l t e r feeding of the northern anchovy (Engraulis mordax). J. Fish. Res. Board Can. 26: 557-582. Limbaugh, C. 1962. Life history and ecological notes on the tubenose, Aulorhynchus flavidus. a hemibranch fish of western North America. Copeia 1962: 549-555. Longhurst, A.R> 1968. (summary of larval rearing techniques of G.O. Schumann) U.S.F.W.S. Bur. Comm. Fish. Circ. 303:: 23-24. Magnusson, J., J . Magnusson and I. Hallgrimsson. 1965. The "Aegir" redfish larvae expedition to the Irminger Sea in May 1961. Rit. Fisk. 4(2): 1-79. Marliave, J.B:.. (in press). Seasonal shifts in the spawning period of a northeast Pacific intertidal fish. J. Fish. Res. Board Can. Marr, J.C. 1955. The "critical period" in the early l i f e history of marine fishes. J. Cons. 21:160-170. Marshall, N.Bw 1966. The Life of Fishes. World Pub. Co., Cleveland. 402 pp. Marshall, S.M. and A.P. Orr. 1952. On the biology of Calanus  finmarchicus. VTI. Factors, affecting egg production. J. Mar. Biol. Ass. U.K. 30: 527-548. 1 pi. McHugh, J.L. and B.W. Walker. 1948. Rearing marine fishes in the laboratory. Calif. Fish & Game. 34: 37-38. Miller, J.M. 1974. Nearshore distribution of Hawaiian marine fish larvae: effects of water quality, turbidity and currents, pp. 217-231. In J.H.S. Blaxter (ed.). The Early Life History of Fishes. Springer-Verlag, Berlin. 765 pp. Morris, R.W. 1951. Early development of the cottid fish, Clinocottus  recalvus (Greeley). Calif. Fish & Game. 3?: 281-300. Morris, R.W. 1956. Some aspects of the problem of rearing marine fishes. Bull. Inst. Oceanog. Monaco. 1082. 61 pp. Norman, J.R. and P.H. Greenwood. 1963. A History of Fishes. Hill and Wang, New York. 398 pp. O^onnell, CP. 1953. The l i f e history of the cabezon Scorpaenlchthys  marmoratus (Ayres). Calif. Fish & Game Fish Bull. 93. 76 pp. O'Cbnnell, CP.. 1972. The interrelation of biting and filtering in the feeding activity of the northern anchovy (Engraulis  mordax). J. Fish. Res. Board Can. 29:: 285-293. 136 Okamoto, R. 1969* Rearing of Red Sea bream larva. Symposium on culture and propagation of sea breams. Bull. Jap. Soc. Sci. Fish. 35J 563-566. (Japanese with English summary on p. 603) Peppar, J.L. MS. Some features of the l i f e history of the cockscomb prickleback, Anoplarchus purpurescens G i l l . M.Sc. Thesis. I 9 6 5 . Univ. British Columbia. 159 pp. Qasim, S.Z. 1955. Rearing experiments on marine teleost larvae and evidence for their need for sleep. Nature (London). 175: 217-218. Qasim, S.Z. 1959. Laboratory experiments on some factors affect-ing the survival of marine teleost larvae. J. Mar. Biol. Ass. India. 1:: 13-25. Rae, B.B. 1 9 6 5 . The Lemon Sole. Fishing News (Books) Ltd., London. 1 0 6 pp. Raitt, D.F.S. 1964. Scottish redfish larval investigations in 1962 with some observations on mid-oceanic echo traces. J; Cons. 29* 65-72. Randall, J.E. 1961. A contribution to the biology of the convict surgeonfish of the Hawaiian Islands, Acanthurus triostegus  sandvicensis. Pac. Sci. 15: 215-272. Rene, F. 1974. Rearing of gilt-head Sparus aurata. p. 747. In J.H.S. Blaxter (ed.). The Early Life History of Fishes. Springer-Verlag, Berlin. 765 pp. Ricketts, E.F., J. Calvin and J.W. Hedgpeth. 1968. Between Pacific Tides. 4th Ed. Stanford Univ. Press, Stanford, Calif. 614 pp. Riley, J.D. 1966. Marine fish culture in Britain. VII. Plaice (Pleuronectes platessa L.) post-larval feeding on Artemia  salina L. nauplii and the effects of varying feeding levels. J. Cons. 30: 204-221. Rognerud, C. I887. Hatching cod in Norway. (Report of the codfish hatchery at Flodevigen, Norway, for the year 1886.) Bull. U.S. Fish Comm. 7: 113-119. (translation) Rollefsen, G. 1939. .Artificial rearing of fry of sea water fish. Rapp. Proc. Verb. C.P.I.E.M. 1 0 9 : : 1 3 3 . Rosenthal, H. 1968. Beobachtungen vfber die Entwicklung des Schwarmverhaltens bei den Larven des Herings Clupea harengus. Mar. Biol. 2 :73-?6. (German with English summary) 137/ Rosenthal, H. and M. Fonds. 1973. Biological observations during rearing experiments with the garfish Belone belone. Mar. Biol. 21i:203-218. Rosenthal, H. and G. Hempel. 1970. Experimental studies in feeding and food requirements of herring larvae (Clupea  harengus L.). In J.H. Steele (ed.). Marine Food Chains. Univ. Calif. Press, Berkeley, pp. 344-364. Russell, F.S. 1926. The vertical distribution of marine macro-plankton. IH. Diurnal observations on the pelagic young of teleostean fishes in the Plymouth area. J. Mar. Biol. Ass. U.K.. l4(N.S.): 387-414. Ryland, J.S. I963. The swimming speeds of plaice larvae. J. Exp. Biol. 40: 285-299. Ryland, J.S. 1964. The feeding of plaice and sand-eel larvae in the southern North Sea. J. Mar. Biol. Ass. U.K. 44: 343-364. Ryland, J.S. 1966. Observations on the development of larvae of the plaice, Pleuronecte3 platessa L., in aquaria. J. Cons. 30: 177-195. Scagel, R.F. 1971. Guide to common seaweeds of British Columbia. British Columbia Prov. Mus. Handbook 27. 330 pp. Schultz, L.P.- and A.C. DeLacy. 1932. The eggs and nesting habits of the crested blenny, Anoplarchus. Copeia 1932: 143-147. Scotten, H.L. 1971. Microbiological aspects of the kelp bed environ-ment, pp. 315-318. In W.J. North (ed.). The Biology of Giant Kelp Beds (Macrocvstis) ln California. J. Cramer Pub., Germany. 600 pp. Sevenster-Bol, A.C.A. 1962. On the causation of drive reduction after a consummatory act (in Gasterosteus aculeatus L.). Arch. Neerl. Zool. 15: 175-236^ Shaw, E. 1957. Preliminary studies on the ontogeny of schooling behavior in the silversides Menidia menidia. BioL>Bull. 113: 354. (seminar abstract) Shaw, E. 1958. The development of schooling behavior in the genus Menidia. Biol. Bull. 115: 324. (seminar abstract) Shelbourne, J.E.. 1957. The feeding and condition of plaice larvae in good and bad plankton patches. J. Mar. Biol. Ass. U.K. 36: 539-552. Shelbourne, J.E. 1964. The art i f i c i a l propagation of marine fish. Adv. Mar. Biol. 2: 1-83. 138 Shelbourne, J.E.. 1970. Marine f i s h cultivation: priorities and progress in Britain. In Marine Aquiculture. Oregon State Univ. Press. Corvallis. 172 pp. Shiogaki, M. and Y<. Dotsu. 1 9 7 1 . The l i f e history of the cli n g -f i s h , Aspasma minima. Jap. J. Ich. 18: 76-84. (Japanese with English summary) Stevenson, J.C. 1962. Distribution and survival of herring larvae (Clupea pallasi Valenciennes) in British Columbia waters. J . Fish. Res. Board Can. 19: 735-810. Taylor, F.H.C. 1964. Life history, and-present status of B r i t i s h Columbia herring stocks. Fish. Res. Board Can. B u l l . 143. 81 pp. Weber, M. and L.F. deBeaufort (eds.). 1951. The Fishes of the Indo-Australian Archipelago, vol. IX. E.J.. B r i l l , Leiden. 484 pp. Wilby, GW. 1937. The l i n g cod, Ophiodon elongatus Girard. B u l l . Fish. Res. Board Can. 54. 24 pp. Wilkie, D.W. MS. Colour pigments in the penpoint gunnel Apodichthys  flavidus and their ecological significance. Unpub. M.Sc. Thesis. 1966. Univ. of B.C. 143 pp. Williams, M.M. and E. Shaw. 1971. Modiflability of schooling behavior i n fishes: the role of early experience. Amer. Mus. Nov. 2448: 1-19. Wimpenny, R.S. 1953. The Plaice. Edward Arnold & Co., London. 145 pp. Wourms, J.P. and D. Evans. 1974a. The annual reproductive cycle of the black prickleback, Xiphister atropurpureus. a Pacific Coast blennioid f i s h . Can. J. Zool. 52:: 795-802. Wourms, J.P. and D. Evans. 1974b,, The embryonic development of the black prickleback, Xiphister atropurpureus. a Pacific Coast blennioid f i s h . Can. J. Zool. 52::879-887. 3 p i . Yusa, T. 19?4. Early l i f e history of Limanda vokohamae (Gunther). pp. 675-676. In J.H.S. Blaxter (ed.). The Early-Life History of Fishes. Springer-Verlag, Berlin. 765 pp. Yusa, T., CR. Forrester and C. Iioka. 1971. Eggs and larvae of Limanda vokohamae (Gunther). Fish. Res. Board Can. Tech. Rep. 236. 21 pp. 139 Appendix 1. The Rearing of Marine Fish Larvae A. Introduction . Research on marine fish larvae is limited by their seasonal occurrence. Most fish spawning has evolved to coincide with the spawning and various developmental stages of prey organisms. This has been described in the. ••biological spring" (Bogorov, 1941; Baranenkova, 1965) and "serial effect" (Bainbridge & McKay, 1968) concepts. The majority of the neritic fishes of the NE Pacific spawn during spring, although the temporal extent of spawning varies considerably, from the week-long spawning of localized Clupea harengus pallasi populations to the four-month spawning period of Xiphister atropurpureus populations on specific beaches (Marliave, in press). Literature on the temporal and spatial extent of spawning has, for many species, proven inaccurate (see Appendix 2). Various species are rare enough that their egg masses cannot be found predictably. Thus, the first constraint on these investigations is the limited availa-b i l i t y of the larvae. In this study, rearing of larvae was started in two manners. Larvae of fifteen species were hatched from eggs and reared, although some species never reached metamorphosis. Nine species were taken from the plankton by various methods and reared. These two basic approaches had different complications, but there was usually no choice in approach for any one species. The availability of egg masses or the capacity of larvae to withstand capture with a net determined which approach was taken. 140 The laboratory f a c i l i t y for rearing larvae was varied within limits. These limits varied between species. In par-ticular, current speeds, lighting conditions and feeding regimes had to be suited to the individual species. Conditions had to be changed for different growth stages as well. In a l l cases, the continuous functioning of every aspect of the setup was v i t a l to successful rearing. Failures in rearing were usually linked to problems with the feeding regime or the seawater system. Feeding the larvae was by far the most labor. "Wild" plank-ton was relied upon in these studies. The rapid turnover of plankton populations forced use of techniques designed to opti-mize the catch of ongoing blooms as well as to provide informa-tion on changes occurring i n localized plankton communities. Even during the height of the spring season, brief periods occurred when sufficient amounts of zooplankton of a suitable size and type were not available. Cultured organisms (Artemia  sallna nauplii) were substituted at these times. Using plankton as food posed supply problems, whereas feeding Artemia nauplii created nutrition problems. Since the emphasis on larval rearing in this investigation was to provide maximum numbers of prejuvenile or metamorphosing fi s h , there was never a proper analysis of any of the problems which arose. Rather, on the basis of information from the literature, past experiences and intuitive judgment, pragmatic solutions were sought-whenever there were mortalities or growth deficiencies. Therefore, no attempt w i l l be made to discuss the merits of this technique relative to some fundamentally different approaches used elsewhere. Rather, the incidents which led to the adoption of more successful approaches will be explained so that the logic involved in developing a rearing system will become evident. 142 B. Obtaining Material The simplest approach to a study of the behavioral trans-formation of fish larvae would be to capture the larvae just before they enter that stage. The difficulty of obtaining viable larvae in sufficient numbers from plankton hauls forced me to attempt rearing newly hatched larvae of various species. Later in these studies there was some success in capturing prejuveniles with a scuba diving technique. A variety of approaches was developed both for obtaining egg masses and late-stage larvae. The most obvious way to obtain eggs was to collect them from the field. Overall, low tide collections proved the most reliable and efficient means for obtaining material. For example, 58 egg masses of Xiphister atropurpureus were found under the boulders on four particular beaches during the winter/ spring spawning of 19?4. Eggs of X. atropurpureus and other species whose spawning is centered above the O-tide level were sought during at least several days of each tide series. Scuba diving collections yielded eggs between periods of low tides, but with greatly reduced efficiency and with the disad-vantage that nest-tending parents fled immediately upon being disturbed. Nest-tending pholids and stichaeids remain wrapped around their eggs at low tide and can readily be taken for positive identification of the eggs. However, in other cases the proximity of an adult to an egg mass may lead to an incorrect identification. 143 Despite the inconvenience and relative ineffectiveness of scuba collecting, the spawning of many species centers in the shallow subtidal, often within beds of laminarian seaweeds. In some cases the guarding behavior of adults will indicate the position of well hidden or camouflaged egg masses, as with Hexagrammos decagrammus. Aulorhynchus flavidus or Coryphopterus  nicholsi. In the case of Clupea harengus pallasi. two collection techniques proved useful. During moderate low tides, herring eggs laid on plants such as Zostera marina or Sargassum mutlcum were collected from a boat by hand or with a gaff. Eggs of this species can easily be stripped and fertilized. With this technique, eggs can be placed on whatever substrate lends itself tp use in the laboratory. This method was used with & J mm mesh plastic screen as substrate. The primary advantages of stripping eggs were knowing exactly when development had started and having an inert substrate which would not affect the eggs. When refrigera-tion was used to retard development of herring eggs naturally spawned on several types of seaweed, the darkness of the refrigera-tor led to the death and spoilage of the plants, which caused egg mortalities• The final technique used for obtaining eggs was to allow adults to spawn in laboratory tanks. This proved particularly useful with Gilbertidia sigalutes. since the adults of this species are inaccessible. Larval G. sigalutes reared through metamorphosis in the laboratory spawned during their f i r s t 144 summer (in August). Gilbertidia sigalutes larvae were the one species of fish larva sufficiently abundant and hardy to permit collection of reasonable numbers of viable larvae from plankton hauls. However, i t was very difficult to estimate the age of the smallest individuals from the plankton before larvae had been reared from eggs spawned in the laboratory. From November 4 to December 9, 1973, 326 G. sigalutes larvae were taken in 213 5-minute surface plankton hauls with a .25 m net of #351 Nitex. Although not a l l of these larvae were kept, around 70* of those retained survived to show visible growth. As is mentioned in the main text, these larvae showed a tendency ta migrate vertically. Thus an average of four larvae per 5-minute haul were captured in dusk surface tox*s. Sufficient numbers of these larvae for most experimental require-ments were obtained from plankton hauls. The other larvae taken from the plankton and reared through metamorphosis in sufficient numbers for transformation to be observed were Psychrolutes paradoxus. Blepsias eirrhosus. Nautichthys  oculofasciatus and Rhamphocottus rlchardsoni, a l l in the family Cottidae (as is G. sigalutes). One reason that these cottid larvae survived capture is their large size, although large clupeoid, hexagrammid or osmerid larvae always died after being caught. Cottid, agonid and liparld larvae seem to have evolved the capacity to withstand severe physical disturbances, perhaps because of their Inshore distributions. Prejuvenile pleuronectids also seem to be very hardy. 145 Patchiness seemed typical of the surface distribution of many of the abundant larvae* Any attempt to follow a collec-tion schedule based on guesses of catch sizes had to take into account the spatial and temporal discontinuities in larval distribution due to patchiness and diurnal behavioral patterns. Different growth stages showed marked differences in their d i s t r i -bution. Newly hatched larvae of a l l species tended to be neus-tonie, but there was no typical distribution pattern for older stages. Experimental planning which counted on live capture of larvae in plankton tows took into consideration variability in the age of specimens. Where low numbers of certain cottid larvae were experimentally manipulated during their transforma-tion they were observed as individuals or as members of distinct growth stages. One technique for the capture of planktonlc fish larvae was to set out night lights and dipnet those which were attracted*. This technique yielded prejuveniles of Blepsias cirrhosus in 1973. but not in sufficient numbers for their transformation to be accurately observed. The most successful dipnet collec-tions were of Leptocottus armatus prejuveniles on March 11, 12, 23 and October 15 of.1973* Chance observations of surface aggregations of larvae of this species led to collections of sufficient numbers of larvae of uniform development (30 in March, 150 in Oct.) to allow laboratory observation of their trans-formation. 146 The only technique for live capture of larvae other than plankton towing or dipnetting was scuba diving. Attempts to capture larvae by diving, both with dipnet and slurp gun, failed during 1973 until a long handled, counterweighted net was used. Counterweighting permitted more rapid swinging of the net when underwater. The net had a 1.5 kg lead weight behind the handle grip, a 20 cm diameter mouth and an 80 cm bag of 1 mm mesh. However, Goblesox maeandricus and Aulorhynchus flavidus larvae caught in this net always died until a plastic holding bag was included. When the larvae were netted the diver grasped the cod end of the net with the larvae inside and put the entire net into the plastic bag. With the other hand the plas-tic bag was closed around the forearm and the net handle while the net was everted to free the larvae from the net inside the plastic bag. The larvae were transported to the boat inside the bag, which was slipped into a bucket at the water surface. This technique proved effective. On June 7, 1974, about 200 G. maeandricus were taken in a dive lasting less than ten minutes. Both G. maeandricus and A. flavidus school as larvae, making them susceptible to net capture by a scuba diver. This tech-nique was not fully exploited since i t was developed late in the study. Various species showed different behavioral tendencies which facilitated their capture with specialized techniques. Swimming speeds particularly affected susceptibility to capture. As Bishai (1959))found for Clupea harengus harengus. larvae 147 have roughly the same range of swimming speeds with respect to body size as adult fishes. This range seems to apply especially to the more elongate larvae. That is, i f a larva can swim at a burst speed of about ten lengths per second, then the larger larvae should be too fast to capture. Later stage Aulorhynchus  flavidus (over about 30 mm TL) and any clupeid larvae at the schooling stage proved too fast to capture using a handnet. With some species, notably Clupea harengus pallasi. capture was impracticable because of the speed and fragility of the larvae. Eggs of C_. harengus pallasi were predictably available in? unlimited quantity so that rearing was the logical approach with this species. In the case of Gilbertidia sigalutes the larva is exceptionally sluggish (as is the adult) and tends to migrate to the surfaoe at dusk. These behavioral traits, to-gether with the typical cottld ability to withstand handling, permitted ready capture by plankton towing. G. sigalutes eggs were only available by inducing adults to spawn in the labora-tory or in cages. In the case of A. flavidus. prejuvenile fish were easily obtained either by collecting egg masses and rearing the larvae or by scuba capture of schooling larvae. 148 C.. Maintaining Egg Masses The discussion of methods for obtaining material omits mention of the developmental stage of egg masses at the time of collection. Collecting advanced stages of development can save considerable laboratory time which would be devoted to car-ing for eggs as they develop. For example. Clupea harengus  pallasi may most predictably and conveniently be obtained by stripping commercially caught adults, but they must then be tended through their embryonic development. If a natural spawn is ac-cessible, then with minimal monitoring the eggs can be collected several days before hatching. However, egg masses with embryos showing guanine eye pigment may be near hatching, in which case special care is required to prevent hatching and subsequent damage during transport. Poorly developed egg masses can be transported in a bucket of seawater i f care is taken to avoid temperature change or oxygen depletion. Portable air pumps can be used or the eggs transported in damp towels or seaweed. If the eggs normally hang in a drop form (e.g. Coryphopterus nicholsi) or in the case of a fish such as Porichthys notatus with hatched embryos hanging by the yolk sac, the eggs must be taken with the rock. The rock is propped in'the normal position with the eggs or embryos free of contact with the container. The physical shock of being scraped from a stone, exposed to light and, for prolonged periods, to air, together with the shocks of a boat ride, usually induced hatching in well-developed 149 egg masses. The degree of maturity of the hatch greatly affected the viability of the larvae, both premature and delayed hatches being less viable. Either an overextended air exposure (longer than would occur during a low low tide) or a physical shock caused premature hatches. If extremely premature, only a partial hatch would occur, in which case the viability of the remaining eggs was usually re-duced by the same disturbance. However, i f the egg mass developed in poor aeration, the innermost embryos would have developed at a retarded rate, so that protracted hatching would occur for the entire mass. Morris (1956) 'noted reduced viability inrlarvae from protracted hatches. In the laboratory a partial hatch due to poor aeration served as an indicator that the remaining embryos would show reduced viability. Protracted hatching normally happens with large egg masses of hexagrammids such as Ophlodon elongatus and Hexagrammos  decagrammus and the cottid Scorpaenichthys marmoratus. With Aulorhynchus flavidus protracted hatching seems to have evolved to an extreme degree, the eggs hatching singly over a one to two week period. For most species, however, development occurs at a uniform rate and the entire mass hatches within hours unless unfavorable conditions have been met. The technique described by Morris (1956) 'for propping an egg mass over an airstone to provide aeration works comparably to propping an egg mass in front of an inlet of fresh, well aerated seawater. Neither technique works well unless a rapid flow occurs around the egg mass. 150 The most successful method found for maintaining egg masses was to fix them between fine-meshed screens in a relatively high velocity open flow of seawater ( . 3 l/cm2/min). The incubator was a plexiglas trough, 122 x 15 x 15 cm, with sets of plexiglas bars affixed at intervals inside to create slots. Egg masses were wedged between screens (#351 Nitex) on plexiglas frames. These pairs of framed screens were tightly placed into the slots in the trough. The screen covered a 10 x 10 cm hole in a frame. At one end of the trough an inlet pipe was directed toward the screens. At the opposite end a 2.0 cm I.D.. Tygon tube siphoned the water out over a drop of 1 . 3 m (at a rate of 3 0 1/min.). An additional consideration in the collection and mainten-ance of egg masses was the provision for replicate rearing of any one species. For many species, spawning continued for a longer time than the normal period of larval development, so the field collections provided ample material for replicates. Often, how-ever, the rarity of the species or the briefness of the spawn-ing period prevented repeated collection of egg masses. In these cases the eggs were divided into lots and maintained at different temperatures to provide hatches at different times. As long as these temperatures remained within the extremes normally occurring in the field during that species' spawning period, the viability of the eggs was not affected. However, care was taken not to subject the eggs to rapid temperature changes. Since varied temperatures could readily be obtained only by using non-circulating refrigerated~or heated bath systems, intense 151 aeration was required. In such a noncirculating system putre-faction became a problem. If the eggs were in a firm mass they were washed daily under a spray of cold water. Seawater was changed daily in closed systems at elevated temperatures and ' weekly in closed refrigerated systems. Seawater used in closed systems was pasteurized by heating to 70°C, then cooling to ambient temperature two or three times in succession within 24 hours. Antibiotics such as streptomycin/penicillin (Ryland, 1966; Shelbourne, 1970))were not used in this study. Eggs of Ophiodon elongatus. Gilbertidia sigalutes and Clupea harengus were incubated at varied temperatures. Clupea harengus pallasi eggs are laid on plant substrates, so that putrefaction of the plants may affect egg viability. o This occurred with eggs held at 6 C in an unlit refrigerator. Stripping herring eggs onto an inert substrate proved highly effective and would have eliminated this difficulty with extended incubation at lowered temperature. Alderdice and Velsen (1971) determined the viability of C. harengus pallasi eggs incubated under different salinity and temperature combinations. The ex-tremes for development periods of optimum (90$) viability are 8 days at 6°C, 15 ppt and 14 days at 10°C, 20 ppt. These com-binations, as well as inlet surface waters of about 9°C and 15 ppt were used. However, the plant substrate rotted in the 6°C lot. The primary purpose of replicating rearing attempts was to guarantee a supply of experimental material in the event of a 1 5 2 rearing failure, as well as to provide material for experimental replication. The contingencies which caused rearing failures were so numerous that replications were necessary with as much separation of supplies and facilities as possible. As will be discussed later, feeding conditions tended to vary most. The purpose of attempts to stagger hatching times in the C. harengus  pallasi eggs was to avoid the coincidence of the critical feed-ing initiation with a drop in plankton abundance. 153 D. Tanks Although difficulties obtaining eggs or larvae and then providing food for the larvae may have occupied the greatest portion of time involved in rearing, the greatest technical ob-stacle to laboratory rearing of marine fish larvae was the impossi-b i l i t y of duplicating conditions found in the pelagic habitat. With the exception of certain larval cottids and agonids such as Gilbertidia sigalutes and Bothragonus swani, and the earliest larval stages of Hexagrammos decagrammus. most larvae in the laboratory could not withstand contact with solid objects or handling with dipnets. Normally most species do not contact solid obstacles, except perhaps at hatching i f the eggs are laid in the intertidal or shallow subtidal. These larvae evolved under conditions where physical contact is only with zooplank-ters of comparable size. Thus, most larvae injured themselves by their responses to physical contact with large solid objects. The typical larval escape response to a physical contact was rapid and undirected swimming. In species with a greater tendency to turn, once the larva collided with a tank wall, its undirected spiralling movements caused repeated collisions with the wall, resulting in disorientation and cessation bf normal behavioral patterns such as feeding. Rearing systems which tended to have a high incidence of such physical contacts never sustained low enough mortality rates for any larvae to reach the stage of meta-morphosis. Therefore, the principal objective of tank design was to minimize contacts of the larvae with the sides, bottom and piping in the tank. 154 The least suitable tank design f o r rearing larvae is a standard rectangular glass aquarium. Transmission o f light through the glass sides would attract the photopositive larvae, hence the larvae would tend to accumulate in the corners where their random movements would be most likely to cause collision w i t h the sides or bottom. The f i r s t tanks tried for rearing were round and had opaque sides. However, these 180 1 tubs (89 cm high x 53 cm diam.))were of reflective green plastic. Early attempts to rear newly hatched Artedius lateralis and Xiphister  atropurpureus in these plastic tubs failed. None of the larvae of these species initiated feeding. The larvae accumulated around the side of the tank at the surface and repeatedly had avoidance reactions to C o n t a c t with the side. High mortality rates developed before the endpoint of yolk absorption, indica-ting physical disturbances rather than the failure to feed as the primary cause of mortality. Within a few days of complete yolk absorption mortalities were 1 0 0 $ . In March of 1973 these tanks were painted black on the inside, after which Xiphister  atropurpureus. Pholis laeta, Bothragonus swani and Gilbertidia  sigalutes were reared in these tanks from hatching through meta-morphosis. Experiences with other tanks indicated that flat black is t h e best color for a rearing receptacle because of its lack of reflectivity. The attempts to rear X. atropurpureus in these 180 1 black tanks resulted in only one success in eight attempts, whereas ' attempts to rear this species in 1900 1 wooden stave tank3 155 resulted i n four successes i n four attempts. The r a t i o of the 2 3 tank surface area (cm ) to tank volume (cm ) was .08 i n the 180 1 tanks and .03 i n the 1900 1 tanks. One probable reason for the lower rearing success i n the small tanks i s that the larvae had a greater pr o b a b i l i t y of contacting a s o l i d surface i n the small tanks because of t h e i r higher surface area/volume r a t i o . The only constraint on the upper l i m i t to the size of a rearing tank was the a v a i l a b i l i t y of food. Except during the months of the spring plankton blooms, great d i f f i c u l t y was met i n supplying an optimal density of wild plankton to the 1900 1 tanks. An hexagonal tank design, discussed below, proved to be large enough (930 1) that larvae tended not to contact the sides, yet small enough to supply plankton to with reasonable e f f o r t . Although a l l the rearing tanks were about one meter i n depth, experience with shallow tanks used for p i l o t studies on l a r v a l behavior indicated that depths of .3 m or less were unsuitable for rearing, especially i f the bottom were s u f f i c i e n t l y r e f l e c t i v e to cause phototactic attractions toward i t . The reason f o r using moderate tank depth was that larvae i n the yolk sac stage usually showed p e r i o d i c i t y i n a c t i v i t y . That i s , a newly hatched larva usually alternated between swimming up to and hovering at the surface and passively sinking. Even the most robust larvae would have passive periods l a s t i n g as long as a few seconds, long enough to sink 10 to 20 cm. The l e a s t robust larvae re-mained passive, l y i n g on the bottom, for periods of a few minutes 156 or longer. Most larvae would make abrupt attempts to flee the bottom upon contacting i t , often skittering in a spiral along the bottom before swimming upward. These efforts may have caused mechanical damage leading to mortalities. Larvae which remained passive on the bottom rarely survived, perhaps owing to infection from the bacterial film which invariably formed on the bottom. Larvae past yolk sac absorption were successfully reared in shallow isolation tubs having a flat black bottom. However, newly hatched larvae should have depths over .5 m. Failures to rear certain species through metamorphosis may well have been due to develop-mental changes in depth preferences. The rearing in this study was conducted with an open-flow system. This involved positioning inlets and outlets in the tanks, and created some degree of current. The rheotactic ten-dencies of larvae were made to benefit rearing success by greatly reducing contacts with the tank sides. Observations of Scorpaenichthys  marmoratus and Gilbertidia sigalutes larvae in a plexiglas tank designed to create a current shear indicated that the larvae would seek a current velocity into which they could orient at a normal swimming speed. If burst speeds were required to swim into a current, the larvae would usually swim cross-current until they reached a current velocity more easily stemmed. In a rearing tank with the inlet away from its sides this meant the larvae would head toward the sides where boundary drag slowed the current. Therefore, too rapid a flow resulted in many larvae contacting the tank walls. Bishai (1959) has studied swimming speeds of 157 various larvae, including Clupea harengus harengus. and Ryland (1963))has reported speeds of Pleuronectes platessa larvae. Both authors have found sustained swimming speeds to be in the range of 1 cm/sec, for newly hatched larvae. Current speeds in the rearing tanks generally graded from about 4 cm/sec. at the inlet to n i l at the tank center. The most effective inlet position was at the side of the tank, with the flow distributed by a vertical pipe with small holes directing the flow parallel to the tank side. This created a current gradient with the great-est flow rate near the sides, decreasing toward the central axis of rotation. By adjusting the flow rate, a current gradient was established in which the larvae took up positions midway between the sides and the center. In the smaller (180 1) tanks the holes in the inlet pipes were 1 mm diameter, compared with 2 mm diameter holes in the inlets of the larger tanks. The inlet hole size determined the amount of seawater turnover which could be achieved for a given current speed. The rearing tanks in this study had turnover rates of once or twice per day, which is comparable to that (2x/day) used by Shelbourne ( 1 9 7 0 ) . The higher turnover rates helped mediate certain aspects of seawater quality which will be dis-cussed in the following section. Perhaps the greatest problem with high turnover rates was the outflow. Outlets were always in the center of the tank. Newly hatched larvae would swim into the outlets and become trapped against or injured by contact with the screen. To avoid this, 158 outlets were positioned centrally at a depth 20 cm above the bottom. Larvae tended to be distributed near the surface, but in the event of their sinking they were more likely to swim within a few centimeters of the bottom than 20 cm above i t . A large area outlet screen had a lesser velocity flowing through its meshes and was more easily avoided. The most effective outlets in this study were sections of PVC pipe about 15 cm long and 10 cm in diameter. Stainless steel screens of 1 mm mesh were fixed onto both ends of the pipe by heat welding and a 2.5 cm I.D.. outlet tube was tightly wedged into a hole in the pipe midway between the screened ends. The outlet tube would then extend out over the top of the tank to form a self-actuating siphon. That i s , the tube formed a loop outside the tank with a hole at the top of the loop. The top of the loop was fixed at a height corresponding to the desired water level in the tank. Once the tube had been'filled, whenever the tank level increased beyond the desired level, the tube would start siphon-ing but would stop when the surface level reached the level of the airhole. A different system was designed for the purpose of rearing Clupea harengus pallasi and Aulorhynchus flavidus larvae in iso-lation. A seawater table, 212 x 105 x 29 cm, was built with an overflow at a height of 22 cm. An inlet pipe (2.6 cm) ran the length of the table, overflowing at a height 27 cm above the table overflow. Eighteen circular tubs, 30 cm high and 34 cm diameter, were constructed of mat finish black ABS plastic with 159 6 cm diameter outlet holes cut centrally in the bottom and covered with black-dyed #351 Nitex mesh.. These tubs were placed in the seawater table, raised off the table bottom by three 1.5 om nubs under each itub. A length of 4 mm I.D.. Tygon tubing ran from the inlet pipe to each tub, fixed in a position 20 cm up the side and directed around the circumference so that a circular current gradient was set up in each tub. The height of the inlet pipe overflow above the seawater table overflow determined the head of pressure in the lengths of tubing to the isolation tubs, re-sulting in equal current speeds and turnover rates in a l l tubs. The water levels in the tubs would equalize with that of the seawater table through the bottom outlet screens, the overall level being determined by the table overflow height. 160 E. Water Supply The use of an open-flow system obviated the need for aeration, cooling or filtering of the seawater. Similarly, through-flowing seawater greatly reduced problems of evaporative salinity ohanges, gas supersaturation, toxicity from long term curing or oxidation and some aspects of organic waste buidup. The one advantage of a closed, recirculating system would be that only a small quantity of seawater would be required. Also, with proper ultraviolet equipment, sterilization of the seawater would be achieved with a recirculating system. However, since there was an unlimited supply of high quality seawater available for this study, the simpler option of using an open-flow was taken. The high quality of the seawater supply at the Bamfield Marine Station is assured by the location on the west coast of Vancouver Island. The seawater intakes are at a depth of 20 m, so that the water is essentially oceanic. The salinity remained constant around 29 ppt, and the temperature varied only season-ally, from 9°C in mid-winter to 12°C during the summer. This temperature range was less extreme than that of surface waters in the Bamfield region. The temperature of the top few centi-meters of water in the tanks usually increased by 1-2°C over the temperature of the remainder of the tank. No species of larva ever demonstrated any response to this temperature discontinuity. Pilot observations were conducted in a thermocline apparatus which maintained a 4°C discontinuity (usually around 9-13°C) over a depth of 20 cm in a tank 72 cm high. These studies 161 indicated that most species of larvae exhibit no response to a temperature discontinuity which is great enough to cause mortali-ties. No yolk-sac larvae survived in the thermocline tank, a l -though survival could be achieved in the same tank by eliminating the temperature discontinuity. During February of 1973, which was the fi r s t month of opera-tion 1 of the Bamfield Marine Station seawater system, gas super-saturation in the seawater caused various types of "gas disease" in larvae. Gas bubbles formed under the epidermis of larval Gilbertidia sigalutes. in the pores of the developing acoustico-lateralis system. This condition improved when a splash column was installed to desaturate the water. Artedius lateralis larvae ingested air bubbles. If the bubbles coalesced inside the gut, the blockage resulted in starvation. The two serious difficulties which arose repeatedly with the seawater quality were temperature changes during shutoffs and the occasional occurrence of detritus in the unfiltered water. Mortalities which occurred during seawater system failures were almost certainly due to anoxia caused by rapidly increasing meta-bolic demands at warmer temperatures, coupled with depletion of the available oxygen in stagnant tanks. Such mortalities only occurred during the months of the seasonal thermocline, from May to October. Marked increases in turbidity followed particular winter storms and decline phases of heavy phytoplankton blooms. Clogging of outlet screens sometimes caused overflows. Flows in the incubator decreased because of clogging of the screens holding the egg masses. The solution to clogging problems was 162 vigilance. During the heaviest blooms, screen cleanings were necessary at least once every six hours. Tank outlet screens were brushed and incubator screens removed and washed. Larvae never seemed affected by increased turbidity although associated overflows caused moderate losses. Water quality tended to be affected by detritus on the tank bottom, especially i f fungal growths were allowed to develop. Tank cleaning was avoided during 1973, except when odors became noticeable or there was visible detritus on the bottom. The clarity of the water was never affected, so cleanings on a weekly basis seemed sufficient. Avoiding disturbances of the larvae was considered more important than sanitation. However, an epidemic in a crowded 1900 1 tank followed a period during which I was absent and the tanks had remained uncleaned despite develop-ment of considerable microbial growth on the bottom. Due to this incident daily bottom siphoning was begun. The tank sides were scraped and sponged on a monthly basis to remove fouling organisms, primarily bryozoans, barnacles, sponges and tunicates. Bryozoans were observed to k i l l small numbers of newly hatched larvae, but the other organisms appeared harmless. The standard hatchery practice of using separate sets of equipment on each tank for the purpose of avoiding spread of infections was not instituted until May of 19?4. At that time a brief epidemic occurred in an overcrowded, underfed tank of Clupea harengus pallasi. In order to prevent contamination of other tanks holding the same and other species, separate sets 163 of siphons and brushes were used. This precaution may have been responsible for the other tanks of herring larvae not suffer-ing the infection. In any case, such precautions should have been taken as a matter of course. 16k F. Lighting The problems discussed in the previous sections were more easily solved than the problem of determining the effect of lighting upon the survival and growth of the larvae. The effect of the light's direction on the distribution of larvae in a tank tended to be marked and these photoattractions were used to advantage. Lighting intensity affected many aspects of larval behavior. Abrupt changes seemed harmless to some species whereas the same changes caused fatalities among certain stages of other species. Assessment of the f u l l importance of lighting became so complicated that the most pragmatic approach was to supply lighting as closely duplicating natural conditions as possible. Observations in a phototaxis apparatus on the behavior of 1-day Scorpaenichthys marmoratus larvae and Gilbertidia sigalutes larvae of varying ages revealed that these larvae are photoposi-tive at light levels of 1 lux and greater (up to at least 10^ lux). Their distribution became random with respect to light at an intensity of 0.1 lux. Therefore, in tanks which did not have lighting directly overhead larvae tended to aggregate on the side nearest the light source. In tanks of this sort strips of black polyethylene were fixed aroung the edges so that the perimeter of the water was shaded. Shading the tank sides greatly reduced early mortalities due to physical injuries. Yolk-sac stages of many larvae have neustonic distributions, the larvae hovering at the surface in a nearly vertical posture. At this stage rheotactic responses tend to be feeble, whereas photopositive 1 6 5 . responses are more extreme than i n l a t e r stages. Hence the larvae were attracted toward the middle of the shaded tank. Most species of larvae showed changes i n depth d i s t r i b u t i o n as they developed. The end of the neustonic phase generally corresponded to a decrease i n photopositive tendencies. In the case of the hexagrammid Ophiodon elongatus. the larvae tended to be dis t r i b u t e d near the bottom a f t e r yolk absorption (6-days age). The l i g h t i n g was not reduced and heavy mo r t a l i t i e s started at 20 days of age. Whether a reduction of l i g h t i n g would have caused these larvae to have fewer contacts with the bottom and thereby permitted successful rearing of them was not determined. Another hexagrammid species, Hexagrammos decagrammus. showed high survival rates during the f i r s t two weeks of rearing. When they had reached 20 days of age these larvae l o s t the tolerance to handling c h a r a c t e r i s t i c of t h e i r e a r l i e r stages. At the same time i t was noticed that the larvae reacted v i o l e n t l y to l i g h t s being turned on or o f f . Turning on l i g h t s i n the middle of the night caused many of these larvae to c u r l up and quiver. This response seemed to precede f a t a l i t y . Not one of t h i s species was reared through metamorphosis. Throughout these studies the overhead row of flourescent l i g h t s was controlled with an Intermatlc timer switch, with the lk hour l i g h t period shifted forward so that l i g h t s turned o f f at midnight. This permitted larvae to feed on plankton caught at dusk. 166 In April of 1974, during the rearing of Clupea harengus pallasi. a dawn/dusk simulator was installed which gave two hour periods of increasing and decreasing incandescent lighting before and after the 14 hour period of flourescent light. This setup consisted of a double-switched Intermatic timer which controlled a reversi-ble polarity motor driving a household rheostat. The motor revolved at one third revolution per hour. One timer switch connected a circuit driving the motor in one direction, turning the rheostat up over a period of two hours. This would brighten incandescent bulbs over the two hour period from 8:00 to 10:00 am, then leave them on for the 14 hour day period while the motor was switched off. At midnight the double-switched timer would switch on another circuit driving the motor in the reverse direction for two hours, gradually turning the bulbs off. In practice, the drive periods were slightly extended, as the rheostat would turn fully three quarters of a turn. A piece of Tygon tubing connec-ting the motor to the rheostat acted as a slip clutch. When the dawn/dusk simulator was installed the laboratory-reared herring larvae immediately displayed diurnal activity patterns never before expressed. During the dusk period they would rise to within 10 cm of the surface and swim about rapidly, feeding actively. Shortly after the installation of the simu-lator the herring larvae started to descend to the bottom half of the tank during the early hours of the light period. Yolk-sac Gilbertidia sigalutes larvae tended to rise to the surface in increased numbers during the dawn and dusk periods.. Whether 167 these more natural behavior patterns affected the viability of laboratory-reared larvae would be difficult to assess. One test would be to use the dawn/dusk simulator in an attempt to rear a sensitive species such as Hexagrammos decagrammus which has not been successfully reared under a simpler lighting setup. 168 G. Density Effects Numbers of larvae available affected the strategy for their rearing. If the spavm of a species was available in abundance, then a large starting number was used to cushion accidental mortalities. Brief mortalities from overflows, seawater shutoffs or other contingencies were anticipated by starting with far more larvae than would be required for experiments. Starting numbers above approximately 10,000 (in 1900 1 tanks) created feeding problems and increased the possibility of epi-demics. (The only two epidemics occurred in the two rearing attempts using extremely high starting numbers, upward of 500,000.) : Under severe crowding (over about 10 larvae per liter) there seemed to be a factor of physical interference between larvae. This was due to the typically stratified distribution of larvae in a tank. In a 200 1 tank, 2,000 larvae would often be d i s t r i -buted primarily within 10 cm of the surface, resulting in densi-ties of over 100 per l i t e r . Localized densities on this order resulted in larvae colliding with each other and displaying avoidance responses which disrupted their normal feeding behavior. Even in successful rearing attempts, daily mortality rates of 10% or over occurred during certain stages or when conditions became temporarily unfavorable. However, whenever such mortality rates were sustained continuously, the population would be deci-mated long before any fish reached transformation. Only when conditions would be attained where mortalities a l l but ceased during at least half the rearing period could reasonable numbers of larvae be raised to transformation. 169 When a small s t a r t i n g number was used i n a larger tank, excess food had to be provided to obtain adequate densities for feeding. In order to optimize use of tank space and feeding e f f o r t s , mixed rearing was practised. Different species of l a r -vae tended to take up d i f f e r e n t positions i n the tank and often fed on d i f f e r e n t size-fractions of plankton. Unrelated species displayed varying s e n s i t i v i t i e s to pathogens and to inclement conditions. However, some combinations may have resulted i n . competition leading to the lowered s u r v i v a l of one species. When a r e l a t i v e l y small number of clupeids was reared with stichaeids the overall s u r v i val was high, whereas an attempt with equal proportions resulted i n lowered s u r v i v a l of the stichaeids, perhaps due to t h e i r slower rates of feeding and growth. Combi-nations of c o t t i d and stichaeid, pholid and stichaeid, and gobiesocid and stichaeid larvae succeeded w e l l . However, an attempt to rear newly hatched stichaeids with agonid larvae led to a predation problem. 170 H. Feeding I. General The problem of obtaining food for fish larvae is complicated by the varying suitability (morphology, size, behavior) of zoo-plankton species in different blooms. Larvae never reacted to dead food organisms. No attempt was made to use ar t i f i c i a l or prepared foods. The above discussion of mixed rearing mentions that species-specific food preferences occur. Larvae of Ophiodon  elongatus. with their large gape, were presumed to prey on other fish larvae, as did Bothragonus swani. However, this was not the case. They preferred copepods and brachyuran zoea. Larval 0. elongatus have poorly developed dentition in comparison with :larval B. swani. suggesting that large teeth rather than a large gape indicate preference for fish larvae. Other larvae usually ate copepod or balanoid nauplii when younger and calanoid copepods when older. A hierarchy of pre-ferences generally existed. For example, Gilbertidia sigalutes larvae preferred newly hatched fish larvae over copepodites, but always consumed copepodites before Artemia nauplii. The fact that individual preferences also occurred was verified whenever Artemia salina nauplii and wild plankton were introduced simultaneously into a tank of transluscent larvae such as Xiphister atropurpureus or Clupea.harengus pallasi. The guts pf some individuals would f i l l with orange Artemia nauplii while others would remain clear (that is, f i l l with transluscent zooplankton). Rosenthal and Hempel (1970)) have documented 171 individual prey preferences in Clupea harengus harengus. as well as developmental changes in preferences. Although mixed gut contents were less common, some Xiphister atropurpureus larvae did feed f i r s t on Artemia and then on copepodites, presumably when the Artemia density had been reduced. Although cultured Artemia nauplli were usually acceptable as prey items, there was a tendency.to nutritional deficiencies (reduced growth rate, listlessness) whenever Artemia was exclu-sively fed. Bardach (1968) and Blaxter (1962) describe Artemia as a nutritionally incomplete food source. Shelbourne (1970)) reports failures rearing Pleuronectes platessa larvae using a certain type of Artemia nauplii as food, whereas eggs from other sources resulted in successful rearing. Although there has been DDT contamination of Artemia eggs harvested in certain regions of the U.S.A. (K.K. Chew, pers..comm., Hettler, 1973). i t is more probable that Artemia lacks certain nutrients present in the organisms normally fed upon by larvae i n the ner i t i c plankton. Rollefsen (1939), who introduced Artemia to mariculturists by successfully rearing P. platessa on Artemia nauplii, obtained his eggs from Rumania, Italy and Spain. Mixed sources of Artemia eggs may provide a complete nutritional base. Rosenthal and Hempel (1970) report Artemia nauplii to be more slowly and less completely digested than copepod nauplii (by £. harengus harengus). In February of 1974 a mixed culture of Xiphister atropurpureus and Pholis laeta which had been primarily fed Artemia nauplii showed increases in mortalities as well as reduced growth rates. 172 Feeding these larvae large amounts of zooplankton resulted in increased activity, visible growth and a cessation of mortali-ties within one week. Scorpaenichthys marmoratus was the only larva with which rearing was attempted using exclusively Artemia nauplii as food. This was due to a lack of zooplankton during August and September. These larvae failed to survive half the larval stage. The oldest surviving _S. marmoratus were obviously starving. These larvae rapidly consumed huge quantities of Artemia. so that their starva-tion was possibly due to inadequate numbers of feedings as well as to poor nutritional quality of the food. An additional advantage of feeding wild plankton rather than Artemia was that a greater proportion of a zooplankton feed-ing would remain viable over a Zk hour period.. Although a large fraction of the plankton was dead at the time of introduction, the surviving fraction could live for an indefinite period. Artemia nauplii, on the other hand, generally survived only a few hours, although a small proportion often lived one or two days. The advantage of feeding Artemia rather than wild plankton was availability, barring culturing failures. During the spring plankton blooms Artemia was cultured strictly as a supplemental food, f u l l reliance being placed upon i t only on days when con-tingencies resulted in minimal catches of plankton. The food ration was limited only during periods of low plankton availability when numerous tanks of larvae were being reared simultaneously. Larvae were deliberately fed to excess 173 whenever possible. Observations on the feeding of various larvae in the laboratory, together with comparisons of larval abundance around and away from foam patches, indicated tendencies for many species of larvae to accumulate in plankton patches. Hunter and Thomas (197*0 demonstrated that larval Engraulis mordax which enter a surface patch of Gymnodinium w i l l tend to remain within the patch on the basis of modified locomotor behavior. Pleuronectes platessa larvae in good and bad plankton patches d i f f e r in. their condition factors, indicating dependence on heavy plankton abundance for survival (Shelbourne, 1957). Rosenthal and Hempel (1970); determined the required food density for week-old Clupea harengus harengus larvae to be upwards of 21-24 nauplii per l i t e r , a density which would only occur in patches. Similarly, Aronovitch and Spectorova (1973) determined the optimal prey concentration for larval Scophthalmus maeoticus to be very high (1-2 per ml). The only problem with feeding to excess in laboratory rearing was the necessity of daily tank cleanings. Rather than attempt estimates of food densities, the time to dissipation for small surface aggregations of certain zoo-plankters or Artemia nauplii was monitored. The formation of such patches is in part a function of overall densities. Therefore, the time required for the larvae to significantly reduce the food density would be reflected in vthe time during which patches per-sisted. Gross overfeeding resulted in such patches remaining from one feeding-to the next. The elimination of such patches within an hour or two of feeding was taken to indicate underfeeding. 174 More i m p o r t a n t t h a n t h e amount f e d was t h e t i m i n g o f f o o d i n t r o d u c t i o n . As m e n t i o n e d i n t h e s e c t i o n on l i g h t i n g ( p g . 156), t h e l i g h t p e r i o d was s h i f t e d f o r w a r d i n o r d e r t o p e r m i t f e e d i n g i m m e d i a t e l y a f t e r i n t r o d u c t i o n o f p l a n k t o n c a u g h t d u r i n g t h e e v e n -i n g . L a r v a l f e e d i n g has b e e n f o u n d t o be v i s u a l i n many s p e c i e s ( B l a x t e r , 1969b - P l e u r o n e c t e s p l a t e s s a ; B a i n b r i d g e , 1965 - S e b a s t e s  m a r i n u s ; B a i n b r i d g e and M c K a y , 1968 - Gadus morhua ; B h a t t a c h a r y y a , 1957 - C l u p e a h a r e n g u s h a r e n g u s ) . A r t e m i a w o u l d t h e n be i n t r o d u c e d i n l a t e m o r n i n g when t h e z o o p l a n k t o n had been l a r g e l y e l i m i n a t e d . However , s i n c e t h e l a r v a e r e q u i r e d more t i m e t o d i g e s t A r t e m i a . t h i s f e e d i n g o r d e r was r e v e r s e d d u r i n g m i d - s p r i n g when ample p l a n k t o n c o u l d be c a u g h t i n d a y l i g h t . W i t h t h i s f e e d i n g s c h e d u l e l a r v a e had g u t c o n t e n t s a t a l l t i m e s . Q u i t e o f t e n , h o w e v e r , p l a n k t o n tows w o u l d be made d u r i n g v a r i o u s p a r t s o f t h e d a y t o p r o v i d e more f r e q u e n t f e e d i n g s . 175 H. 2. Plankton Towing Aside from plankton population cycles, which generally changed on an order of days, there were hourly distribution changes resulting from various factors. Vertical migratory patterns usually followed a diurnal rhythm, with peak surface abundance occurring during dusk. However, tidal flow patterns also affected plankton abundance. Salinity gradients moving in and out of inlets usually marked discontinuities in plankton density. Tidal flows and freshwater runoff affected zooplankton d i s t r i -butions, particularly at these hydrologic fronts. Downwellings, marked by foam lines,, typically occurred at hydrologic fronts. Often tremendous amounts of plankton accumulated at or to one side of these foam lines. Weather also affected plankton distribution. Parallel foam lines (windrows) running lengthwise down a channel or inlet usually marked increased plankton densities. Normal winds caused such small-scale effects on plankton abundance whereas gale force winds had dramatic effects on the overall abundance. Storms resulted in highly unpredictable plankton cycles. Surface plank-ton sometimes increased during a windstorm, yet at other times entirely disappeared. However, vertical migrations were usually less pronounced during storms. At the end of a storm lasting several days or longer there was often a marked diurnal vertical migration. These abrupt fluctuations in plankton availability were more typical of winter when plankton was less abundant and storms more frequent. Rather than to attempt prediction of storm-induced changes in abundance, such changes were directly monitored. 176 During the months of January and February extended stormy periods followed by calm, sunny periods (lasting a week or more) typically brought about brief phytoplankton blooms in the inle t s . Such localized blooms were invariably followed by brief blooms of calanoid copepods. Marshall and Orr (1952) related copepod egg extrusion to high abundance of diatoms. These occurrences were important since copepod nauplii and copepodites are ideal food for larval f i s h and could be used as an alternate to the less nutritious Artemia (which were relied upon to a great extent during the winter). The period for which a particular bloom could be exploited as food for larvae depended upon the organism involved. A bloom of calanoid dopepods provided food during the naupliar and copepo-dite stages. Some species of copepods were small enough to be fed upon as adults. Thus, copepod blooms could provide food for periods of up to a month. Balanoid nauplii remained small enough to be fed upon for almost a week. Blooms of brachyuran zoea, however, provided useful food for only a few days or less since the zoea became too large and spinous after their f i r s t molt. The only phytoplankton bloom of value as larval food during my experiments was one brief bloom of Noctiluca. Certain blooms caused d i f f i c u l t i e s to plankton towing efforts. Diatom blooms often resulted in such rapid net clogging that l i t t l e useful plankton could be collected. Rainstorms creating heavy runoff often caused diatoms within the inlets to form dense mats which severely clogged nets. Blooms of medusae and ctenophpres 17? not' only caused net clogging, but also reduced the v i a b i l i t y of the remaining catch due to the stinging of t h e i r neumatocysts. Plankton tows were made from & J m pontoon boat (20 or 25 HP 2 outboard engine) with a three meter long net of .25 m mouth area (351 micron mesh). Plankton tows were made to provide food as we l l as information on s p a t i a l and temporal patterns of plankton abundance (for planning future towing). During the i n i t i a l phases of a localized bloom eff o r t s were concentrated on maximizing catches from that bloom and tracking changes i n i t s d i s t r i b u t i o n . Blooms tended to be l o c a l i z e d to particular water masses i n p a r t i c u l a r i n l e t s , so that peak densities would occur at s p e c i f i c t i d a l phases when the surface area of that water mass • o was constricted, presuming that the organisms were tending to maintain a position near the surface. Diurnal v e r t i c a l migra-tions also had to be taken into consideration. For example, i f peak densities were occurring at the mouth of an i n l e t at high tide during l a t e afternoon, those peak densities could be expected to i n t e n s i f y when the high tide shifted to dusk, but then perhaps to dissipate when the high tides came i n during darkness. Of course, quite d i f f e r e n t d iurnal migratory patterns occur with d i f f e r e n t zooplankters. In addition to s h i f t i n g density patterns (due to water movements or v e r t i c a l migrations), increases i n average size and decreases i n numbers had to be anticipated. Toward the expected end of a s p e c i f i c bloom, plankton tow-ing efforts had to be shifted from tracking that bloom to samp-l i n g other areas i n search of new blooms. Although exploratory 178 towing of this sort usually emphasized tows in different areas at different times of the day, more marked density differences were sometimes found by comparing surface and subsurface tows. The subsurface tows were made by attaching a depressor to the net (so that the. net towed at about three meters depth). During-times of year other than spring (when abundant surface plankton-could usually be found) diurnal movements of calanoid copepod populations tended to be predictable enough that density shifts according to depth could be followed with the towing depth, down to four meters.. During the period of the seasonal thermocline (May through October), density increases occurred over shallow depths in the region of marked temperature changes. Temperature discontinuities seemed to be shallower and more abrupt in inlets. At this time of year vertical migrations appeared to be most brief (confined to dusk), whereas when the seasonal thermocline was completely dissipated (January and February) peak surface densities tended to continue through the dark hours. The length of tows was determined by plankton abundance in order to obtain catches most nearly f i l l i n g the bucket yet not clogging the meshes of the cod end of the net. Tows were timed to be five, ten or fifteen minutes i n duration. Longer tows tended to yield higher proportions of moribund zooplankton due to the extended period during which the earlier caught por-tion had been subjected to high flow rates, turbulence and con-tacts with the bucket screen. 179 A p l a s t i c k i t c h e n s i e v e (one m i l l i m e t e r d i a m e t e r mesh o p e n -i n g s ) was u s e d t o s t r a i n p l a n k t o n as i t was p o u r e d i n t o a h o l d i n g b u c k e t on t h e b o a t . T h i s e l i m i n a t e d medusae and c t e n o p h o r e s •  '. • ( w h i c h w o u l d s t i n g o r g a n i s m s i n t h e c a t c h ) a s w e l l a s m y s i d s , a m p h i p o d s , l a r g e copepods and l a r g e z o e a ( w h i c h w o u l d c a u s e m e c h a n i c a l damage t o o t h e r z o o p l a n k t o n ) . . A l s o , r e m o v a l o f l a r g e r e l e m e n t s l o w e r e d oxygen c o n s u m p t i o n i n t h e c a t c h b u c k e t b y l o w e r -i n g t h e o v e r a l l d e n s i t y o f o r g a n i s m s . The p l a n k t o n i n t h e h o l d i n g b u c k e t was e i t h e r a e r a t e d w i t h a p o r t a b l e b a t t e r y pump o r f r e q u e n t l y s t i r r e d . S t i r r i n g p r e v e n t e d t h e p l a n k t o n f r o m s e t t l i n g o n t o t h e b o t t o m o f t h e b u c k e t where a n o x i c c o n d i t i o n s and m e c h a n i c a l d i s t u r b a n c e s f r o m s p i n e s and appendages w o u l d l o w e r t h e v i a b i l i t y o f t h e c a t c h ; I f u n a e r a t e d , p e r i o d i c a l l y s p l a s h i n g f r e s h s e a w a t e r i n t o t h e p l a n k t o n b u c k e t w o u l d i n c r e a s e t h e o x y g e n t e n s i o n and v i s i b l y r e t a r d t h e s i n k i n g r a t e o f t h e p l a n k t o n . A s e r i e s o f tows was r a r e l y l o n g e r t h a n one and one h a l f h o u r s , i n o r d e r t o p r o t e c t t h e v i a b i l i t y o f t h e c a t c h , w h i c h w o u l d b e s o r t e d i m m e d i a t e l y f o r f e e d i n g . P l a n k t o n was s i z e - s i e v e d b e f o r e f e e d i n g , w i t h b r a s s s i e v e s o f .25 and .5 mm d i a m e t e r mesh o p e n i n g s . W i t h n o r m a l c l o g g i n g , t h e #351 N i t e x mesh o f t h e p l a n k t o n n e t t e n d e d t o c a t c h l i t t l e w h i c h w o u l d p a s s t h r o u g h a .25 mm s i e v e . T h u s , s i z e f r a c t i o n s o f 0-250 m i c r o n s , 250-500 m i c r o n s and 500-1000 m i c r o n s were u s e d t o f e e d d i f f e r e n t l a r v a l s t a g e s . O n l y t h e s m a l l e s t f r a c t i o n was used when f e e d i n g y o l k s a c l a r v a e o f s p e c i e s w h i c h h a t c h e d w i t h a r e l a t i v e l y s m a l l m o u t h , i n c l u d i n g A r t e d i u s l a t e r a l i s , X i p h i s t e r  a t r o p u r p u r e u s . P h o l i s l a e t a and C l u p e a h a r e n g u s p a l l a s i . N e w l y 180 hatched larvae of other species and older larvae of a l l species were fed both the medium and smaller fractions, when\ the small fraction was not required for other rearing projects. The larger fraction was used only on the late larval stages of Gilbertidia  sigalutes. Blepsias eirrhosus. Nautichthys oculofasciatus and Clupea harengus p a l l a s i . Size sieving provided a means for eliminating unsuitable food organisms. The relative s u i t a b i l i t y of different zooplank-ters as food for f i s h larvae was not investigated beyond noting accumulations of particular species which were selected against as prey. Such zooplankters were either large or more spinous than preferred prey items, so that sieving tended to sort out these organisms. S t i l l , the changing species composition of the plankton sometimes resulted in most of the catch being unsuitable. 181 H. 3» Culturing Artemia salina Nauplii Regardless of plankton ava i l a b i l i t y , Artemia salina nauplii were hatched on a daily basis to provide supplemental food which could be solely relied upon i f the plankton failed. Two Artemia incubators were in continuous operation, being started on a staggered basis so that one was harvested and set up again each day. At 27°C Artemia eggs hatch in 18 to 36 hours. Harvesting at 48 hours assured complete hatches. However, early.hatched nauplii had used their yolk material and grown by that time. Therefore, when small-sized fractions were needed for feeding newly hatched larvae, the tubs would be harvested twice, f i r s t at 24 hours and then at 48 hours. The 48 hour harvest could then be size-sieved with a •25 mm brass sieve to separate larger nauplii. Artemia were harvested by shutting off the aeration and wrapping the entire tub in black polyethylene sheeting, leaving a strip along one side open to l i g h t . A lamp was placed at the tub wall, shining about 10 cm off the bottom. Egg shells would sink or float while the nauplii' aggregated in front of the li g h t . In 20 minutes the nauplii would be aggregated over the bottom toward the lighted side. Settling the hatch longer than 30 minutes would result in anoxic conditions causing the nauplii to sink to the bottom. This anoxia would lessen the v i a b i l i t y of the hatch and i t would be d i f f i c u l t to siphon the sunken nauplii without also taking eggs. Therefore, after 20 to 25 minutes an opening was created in the layer of eggs at the surface by gently blowing directly downward. A siphon was then inserted nearly 182 to the tub bottom and held with the forefinger extending about 2 cm beyond the end of the tube to avoid bringing i t close enough to the bottom to pick up eggs. Care was taken not to s t i r up eggs on the bottom with the forefinger nor to disturb the surface layer of eggs. The majority of nauplii were removed by siphon-ing- 1/4 to 1/3 of the volume of the culture. When a hatch of Artemia had been harvested and fed to the fis h larvae, the incubation tub (plastic, 50 cm high, 24 1 capacity) was repeatedly scrubbed and rinsed and the aeration and heating implements thoroughly wiped. Seawater (in plastic buckets);which had been immersed in a hot bath for the duration of the harvest-ing procedure would then be diluted with cold seawater to obtain 2?°C (+ 1°C). Incubation had to be started at the proper tem-perature for complete hatches to occur. The 27°C temperature was maintained using a 150 W submersible heater controlled by a separate thermostat unit (minimizing problems of electric shock). The key factors for successful mass hatches proved to be maintenance of constant temperature (27°C), adequate aeration and circulation preventing settlement of eggs. Pockets of eggs settling into tank corners failed to hatch. Aeration served the dual purpose of oxygenation and circulation when a i r stones were positioned (with weights);in the center of the tank bottom and air provided at an intensity which raised water two tb three centimeters above the water level (41-45 cm deep). Less intense aeration was inadequate for keeping-all eggs in* suspension. 183 For each hatching run, 100 ml of Artemia eggs were used. An optimal hatch from 100 ml of eggs would y i e l d about four' m i l l i o n n a u p l i i (Jones, 1972);. The overall procedure for harvest-ing and setting up an Artemia incubation tub i s outlined as follows:: 1. s t a r t f i l l i n g sink(s) with hot water 2. stop aeration, cover tub and place lamp for settlement of hatched n a u p l i i (20-30 min.) 3. immerse buckets of seawater i n hot bath 4. siphon out settled n a u p l i i 5. feed n a u p l i i to f i s h larvae 6. clean tub, heater and aeration equipment 7. f i l l tub with heated and cold seawater mixed to obtain 27°C 8. place heater i n , s t a r t aeration and s t i r i n eggs. Introduction of Artemia eggs into rearing tanks should be avoided. Larger larvae tended not to ingest these eggs, but smaller individuals which regularly ate eggs showed marked reduc-tions i n growth. Newly feeding Clupea harengus p a l l a s i larvae were able to defecate the eggs so that blockage of the gut may have been less of a problem than malnutrition. However, w e l l over half of the smallest larvae reared (Artedius l a t e r a l i s ) showed a tendency to eat Artemia eggs. Gut blockage occurred i n A. l a t e r a l i s . This may l a r g e l y account for the low overall survival of hatches of A. l a t e r a l i s . Another problem with Artemia eggs i n rearing tanks was the d i f f i c u l t y of removing f l o a t i n g eggs from the surface. Careless harvesting resulted i n rapid buildups of Artemia eggs on tank surfaces, often associated with organic films caused by b a c t e r i a l growth. 184 Appendix 2. Annotated Illustrations of NE Pacific Fish Larvae of the Neritic Plankton Preface This appendix includes illustrations of larvae with notes on spawning season, spawning site, egg mass characteristics and season of occurrence of larvae. Detailed description of morpho-metries and meristics is omitted, largely because the continuous changes in these characters during larval development make their application in identification difficult,. A brief statement' is included for each species on the characters which distinguish i t from similar larvae which have- also been described'.. In every case except Clupea harengus pallasi. Anoplarchus purpurescens. Hexagrammos decagrammus and Artedius lateralis. the illustrations are the fi r s t I know of for these stages. The illustrations are numbered to correspond to the written descriptions and their captions l i s t the age and size of larvae from top to bottom. C. Osteichthys 0. Clupeiformes F. Clupeidae 1. Clupea harengus p a l l a s i Valenciennes Spawning of t h i s species i n the Bamfield area was observed i n 1 9 7 3 and 1974.. Spawning i n Bamfield I n l e t occurred during early March, midway i n the length of the I n l e t at the narrows, at i n t e r t i d a l l e v e l s , primarily on Fucus gardneri. Spawning took place i n Grappler I n l e t i n a s i m i l a r l o c a t i o n , but egg deposition was on Zostera marina. Eggs are attached singly. They are colorless (translucent)) a f t e r f e r t i l i z a t i o n , becoming silver-gray toward hatching. Larvae occur i n the plankton from l a t e March to mid-May, when- schools begin to form. The early l a r v a l stages (2-day and 14-day larvae) are i l l u s -trated i n a paper by Taylor (1964). These larvae are recognized by t h e i r elongate body, as we l l as by the absence of an adipose f i n and by a snout-anus length of about 3 / ^ body length. They are primarily confused with osmerid larvae (with an adipose f i n ) which occur abundantly during- l a t e summer and f a l l , not spring. The i l l u s t r a t i o n s are of the stages at which the i s o l a t i o n experiment was started and f i n i s h e d . 1 . Clupea harengus pal l a s i . (5 mm scale) 43-day, 29 mm TL; 83-day, 46 mm TL (artist, J. Marliave) 0..Gbbiesociformes ' F. Gobiesocidae 2. Gbbiesox maeandricus (Girard) Gobiesox maeandricus spawns from November to April in Barkley Sound and during April and May in Burrard Inlet (Stanley Park). Embryonic development probably takes much longer before the buildup of the seasonal thermocline (around April). Up to five or six fish may be present during mating, but only a male remains to tend the egg mass. The male eats inviable eggs. The egg mass is a monolayer of hemispherical eggs (flat at the point of attachment). The^eggs are attached to the underside of a boulder on a rocky substrate. The eggs are 2 mm diameter and bright yellow, becoming golden brown toward hatching. Concentric rings of eggs are laid, the most recently developing mode being outermost (200-500 eggs per mode). The embryos and newly hatched larvae (6 mm TL) have a con-spicuous bilobed yolk with a single o i l droplet. Late-stage larvae accumulate in large schools around kelp beds from the last week in May until late June in Barkley Sound. Larval schools occur in Burrard Inlet during July. Larval G. maeandricus develop the pelvic sucking disc during the stage of caudal fin ray development. The disc is complete in rudimentary form (small, no papillae) when anal and dorsal ray development starts. This short-bodied larva is distinguished by the presence of a swimbladder (absent in cottid larvae), melanic countershading over the entire gut, nasal melanophores and a translucent yellow 188 Gobiesox reaeandricus. (2 mm scale) 0-day, 6 mm TL (side & ventral views); 2-day, 6 mm TL; 6-day, 8 mm TL, (artist, J. Marliave), 189 Gobiesox maeandricus. (5 mm scale) 54-day, 14 mm TL (side- & ventral views); 65-day, 18 mm TL (side & ventral views). Settled stages, (artist, J. Marliave). 190 body color. In this region i t can be confused with Rimicola  muscarum larvae. which are smaller and have a pelvic sucking disc at hatching. 3. Rimicola muscarum (Meek and Pierson) - Not Illustrated In Barkley Sound, egg masses of R. muscarum are found during late May and early June on kelp blades (Chris Lobban and Dan Pace, personal communication). The hemispherical eggs are 1.3 mm dia-meter, yellow, and laid in a monolayer of concentric rings. The embryos have'-a bilobed yolk with a single o i l droplet. Embryos develop the pelvic sucking disc. Larvae hatch in about one month at a total length of 4 mm. They grow to 4.3 mm TL iri five days. They appear identical to larvae of Gobiesox maeandricus but are markedly smaller; without nasal pigment and with the pelvic sucking disc at hatching. 0 . Gasterosteiformes F. Gasterosteidae 4. Aulorhynchus flavidus G i l l This species spawns in Barkley Sound during May and June. Its spawning behavior is described in detail by Limbaugh (1962). Territorial males repeatedly mate with different females, accumu-lating up to about six egg masses (150-600 eggs per mass) attached to the blades of newly grown:^  seaweed. The eggs are 2 mm diameter and are amber to brown in color. Larvae hatch during June and July and are almost always found schooling around kelp beds. The larvae are recognized by their unaligned lateral melanin bands (anterior band on dorsal gut surface, posterior band mediolateral on musculature). Larvae develop the elongate snout during fin ray formation. 191 4. Aulorhynchus flavidus. (5 mm scale) 0-day, 8.5 mm TL; 18-day, 16 mm TL; 32-day, 23 mm TL; 45-day, 30 mm TL (artist, F. Zitten). 192 0. Perciformes F. Clinidae 5. Gibbonsia montereyensis Hubbs Two egg masses, the embryos of which had unpigmented eyes, were taken on April 25, 1974 (by Myriam Haylock) from the top of a boulder on a moderately exposed shore. The masses (about 100 eggs each) were together in coraline algae, with Polyneura  latlssima blades matted over them and stuck together with adhe-sive threads,' The eggs were identified as those of G. montereyensis. the only clinid recorded from Barkley Sound,.on the basis of their complete resemblance to egg masses of another clinid, Paraclinus marmoratus. described and illustrated with photographs by Breder (1939). That i s , the eggs were 1 mm diameter, held together with viscous threads. The embryos had large stellate xanthophores and melanophores over the entire body and yolk. The xanthophores and melanophores contracted prior to hatching, . as happens with Paraclinus. Gibbonsia montereyensis hatch at 6 mm TL. The larval swim bladder f i l l s within 12 hours. The larvae are highly photopositive at hatching, less so at 10 hours. They hover at the surface during their f i r s t day, then start swimming in mid-water on day 2. Oh July 29, 1974 a 22 mm TL juvenile G. montereyensis was taken in a scuba diving rotenone collection on SW Helby Island. It was unpigmented, indicating very recent settlement. The larval stage therefore lasts about two months. I I 5. Gibbons ia inon tereyens i s . (2 mm scale) 0-day, 6 mm TL; (5 mm scale) from f i e l d , settled, 22 mm TL (a r t i s t , J. Karliave). 194 Breder (1939) stated that juvenile Paraclinus marmoratus settle at about 25 mm TL. However, in his subsequent publication, Breder (l94l) proposed that larval settlement occurs within 24 hours since he detected a reversal of phototactic tendencies in 0-day larvae held i n a petri dish. This change in behavior probably indicates a s h i f t from neustonic to planktonlc behavior as noted for G. montereyensis. Larval G"., mon tereyens is are recognized by the melanophores on the yolk, over the swim bladder and on the gut over the anus, together with a ventral row of melanophores on the t a i l portion of the musculature. Xanthophores are conspicuous in two pairs over the nape, one pair on the gut over the anus and two pair on the dorsal t a i l musculature just anterior to the caudal peduncle. F. Stichaeidae 6. Anoplarchus purpurescens G i l l This species is much more abundant in inside waters such as Burrard Inlet than on the outside of Vancouver Island in Barkley Sound. The spawning season, nest site and egg mass have been described by Schultz and DeLacy (1932) and Peppar (,MSJ). IJfound that A. purpurescens spawns during late winter (January, February) on protected shores under boulders in the lower and mid-intertidal. They prefer an under-bbulder substrate of crushed s h e l l . The eggs are tended by the female. Peppar ( MSJ) reports t e r r i t o r i a l guarding of eggs by females, stating that he never found a male under a nest rock. I found 21 egg masses with the attendant female (2 without). 195 Six. of these masses (28*)')were being tended in the presence of a male which was under the same rock without interposed obstacles. In three instances two females were tending egg masses under one rock. In a fourth instance there were three females tend-ing nests under the same rock, a l l within one to two body lengths of each other. It would appear that under conditions of high population density and limitation of preferred habitats, terri-toriality breaks down. It might be expected that under such conditions the population would spread into marginal habitats, but this was not observed at Lumberman's Arch (Stanley Park, Vancouver) in 1975. The eggs are small (1.4 mm diameter), and colored lime green to white during early development, becoming silvery gray. Peppar (1965))found egg masses having between 2,000 and 3,000 eggs. Hatching occurs from late February through March at Lumberman's Arch. The larvae hatch at 7.5 mm TLl They are described and illustrated by Peppar ('MS';). They resemble Xiphister sp. larvae except that they have fewer dorsal gut melanophores (1-4, usually 3, per side). It should be noted that the dorsal gut melanophores of stichaeids alternate in position from side to side, so that a lateral view, as in my illustration, may give the appearance of twice as many melanophores on one side. Ih stichaeid and pholid larvae the^snout-anus:total length ratio approximates that of adults of the species. The relative gut length increases slightly during metamorphosis. Thus, Anoplarchus purpurescens. (2 ram scale) 0-day, 7 .5 ram TL (artist, J. Marliave). 197 stichaeid larvae have a snout-anus length about 2/5 of total length (versus 1/2 to 2/3 in pholids). Neither pholid nor stichaeid larvae have swim bladders. Peppar Cjp§5))reports larvae of Anoplarchus purpurescens as becoming negatively phototaetie and leaving the plankton after a few days. I maintained larval A. purpurescens ln the laboratory for six days, one group starved and one group fed.. (Larvae normally feed within one hour of hatching in this species.) The starved larvae started to accumulate in the shaded perimeter of the tank at one to two days age. The fed larvae, however, remained positively phototaetie after three days age. Starvation also caused different swimming behavior. ?• Xiphister atropurpureus (Kittlitz) Spawning occurs during late winter and spring under boulders on moderately exposed shores. Pebble.^ and shell substrates are preferred. Spawning occurs later on more exposed beaches (Marliave, in press). The spawning season (April, May) on the open coast of California (Wourms and Evans, 19?4a) coincides with that on the open coast of Vancouver Island at Jordan River. The egg masses are spherical, 20 to 50 mm diameter, and usually have about 500 to 2,000 eggs (large fish will produce up to about 8,000 eggs). This a greater fecundity than the average 500 eggs reported by Wourms and Evans (19?4a)i The eggs are white in color and about 2 mm in diameter. The embryo-logy of this species is described by Wourms and Evans (19?4b), 198. Larvae hatch at 7 to 10 mm TL (depending on size of parent, or egg, and maturity of hatch). They are elongate, with a snout-anus length about 2/5 total length. The yolk has a single o i l droplet. There are about seven melanophores on each side of the dorsal gut surface, 12 small melanophores on the dorsal t a i l musculature and about 25 melanophores on the ventral t a i l musculature. Larval X. atropurpureus with fully developed fin rays start schooling inshore around kelp beds or rocky prominences (late May, early June in Barkley Sound). 8. Xiphister mucosus (Girard) - Not Illustrated Xiphister mucosus occurs syrapatrically with Xiphister  atropurpureus.•the adults spawning under boulders in the lower intertidal (at the same tide levels as X. atropurpureus). This is contrary to the statements by Ricketts, et a l . (1968) and Wourms and Evans (1974a) that only juveniles occur inter-tidally. The male wraps around the egg mass under a boulder, as does the male of Xj, atropurpureus. However, individuals of X. mucosus are generally larger and produce egg masses of pro-portionally larger size. The eggs are white, approximately 2.5 mm diameter. They^ number about 2,000 to 15,000 per mass. Ih any one locality, the spawning season of the two Xiphister species coincides. The larvae are identical to those of X. atropurpureus. except that they are usually larger (11 mm TL at hatching)i 199 i _ i . Xiphister atropurpureus. (5 mm scale) 0-day, 7 mm TL, 6-day, 9 mm TL; 32-day, 17 mm TL; 60-day, 20 mm TL; 60-day, settled, 21 mm TL. (artist, F. Zitten). 200 F. Pholidae 9. Apodichthys flavidus Girard One egg mass of Apodichthys flavidus. at the hatching stage, was found under a boulder in the lower intertidal at Jordan River in late April, 1974. The eggs were 3 mm diameter. The mass was tended by a lone male. Clemens and Wilby (1961) report spawning in Burrard Inlet (Second Narrows) during January and that pairs of adults coil around the white egg masses. Wilkie ('MS-) reports that both parents tend the eggs and that the period of incubation lasts 2f months. Wilkie' found newly spawned eggs in Burrard Inlet in January and February, so hatching may also occur during April in this area. Larvae hatch at 13 mm TL. Wilkie (1MSS) reared this species and observed settlement after 50 days at 25 mm TL. He did not illustrate the larvae. Larvae of this species have a snout-anus length 3/5 of total length. The gut length is relatively greater than in larvae of Pholis laeta. The larvae have prominent dorsal gut melano-phores (about 10 on each side) which are especially large over the anus. They have a similar number of smaller melanophores on the ventral t a i l musculature. At the caudal peduncle there are also several melanophores on the dorsal musculature. The ventral gut surface has a row of many fine melanophores. 10. Pholis laeta (Cope)) A pair of adult Pholis laeta was found coiled around a mature egg mass under a rock in the lower intertidal in Bamfield Inlet in early February, 1974. The eggs hatched In late February. I. J 9. Apodichthys flavidus. (5 mm scale) 0-day, 13 mm TL ( a r t i s t , J. Marliave), 202 The newly hatched larvae are about 9 mm TL. The early stages do not have dorsal gut melanophores as do Apodichthys  flavidus larvae, but there are three large melanophores on the dorsal gut over the anus. The ventral' gut and ventral t a i l musculature have rows of numerous fine melanophores. At the onset of f i n ray formation, a fine row of melanophores becomes visible on the dorsal gut surface. The snout-anus length i s 1/2 total length. The snout-anus lengths are obviously less than half of total length in the stichaeid >larvae described here and over 1/2 total length in the other pholid species. 11. Xererpes fucorum (Jordan & Gilbert) - Not Illustrated An adult male Xererpes fucorum was found coiled around an eyed egg mass under a rook in the lower intertidal in Bamfield Inlet in early December, 1973, indicating a f a l l spawning i n contrast to the late winter or spring spawning of the other pholids and stichaeids described ln this appendix. The eggs are probably white W h e n l a i d . They are s i l v e r -gray when eyed. They measure 2 mm diameter. The larvae hatch at 11 mm TL, intermediate i n size between Pholis laeta and Apodichthys flavidus. The larvae have melano-phore patterns and a snout-anus:total length ratio similar to those of Apodichthys flavidus. However, Xererpes fucorum larvae are smaller and the gut melanophores over the anus appear larger relative to other melanophores than in A..flavidus. Only these melanophores over the anus are; plainly v i s i b l e i n my specimens of 0-day X. fucorum larvae, although wild specimens may d i f f e r . 203 i 1 10. Pholis l a e t a . (5 mm scale) 0-day, 9 mm TL; 10-day, 14 mm TL; 27-day, 16 mm TL; 39-day, 25.5 mm TL; 52-day, s e t t l e d , 29 mm TL. ( a r t i s t , F. Zit t e n ) . 204 F. Anarhichadidae 12. Anarrhichthys ocellatus Ayres A single larval Anarrhichthys ocellatus was caught in a daytime surface plankton tow in the mouth of Bamfield Inlet. This specimen was 65 mm TL fresh (56 mm TL fixed in 3* formalin in seawater), with a snout-anus length of only 13 mm (11.5 mm fixed)\ The body proportions of the late stage larva are much like those of settled juveniles. There are approximately 223 myotomes visible in the larval specimen. F.. Gobiidae 13. Cbryphopterus nicholsi (Bean) The droplike orange eggs of this species are laid in a monolayer, hanging from the underside of a rock on a sandy bottom in protected waters-. The male digs a depression under the rock prior to spawning and tends the nest site during incubation. The eggs change in color from orange to silver as the embryos mature.. Hatching takes place in July. These larvae are recognized by their small size, moderate body proportions, wellldeveloped swim bladder with dense melanin-and large melanophores on the ventral trunk musculature and on the dorsal musculature at the caudal peduncle. There are no melanophores on the dorsal gut at the anus. There is a row of small ventral gut melanophores. 2 0 5 | . i 1 2 . Anarrhichthvs ocellatus. from plankton. ( 5 mm scale) 6 5 mm TL (artist, F. Zitten). 13. Coryphopterus nicholsi. (1 mm scale) 4-day, 3.5 mm TL (artist, J. Marliave). 207 F. Hexagrammidae 14. Hexagrammos decagrammus (Pallas) Hexagrammos decagrammus spawns brown egg masses on rocks ~ ih the kelp zone during November in the area of Bamfield Inlet. The larvae hatch in late December at 7 to 8 mm TL. They are iridescent green-blue with many melanophores. They are illustrated by Gbrbunova (1962); but the drawings are poorly reproduced in the translated edition. The larvae have a moderate body length and a snort gut (snout-anus';length 1/3 total length). Larvae have several lateral and dorsal rows of melanophores, a dense occipital patch of melanophores, nasal melanophores and melanophores over the dorsal surface of the gut. The fin fold develops an intense iridescent blue color in the region where the dorsal and anal fins develop. These colors are bleached in preservation, i 5. Ophiodon elbngatus Girard This species lays large (2.5-3.0 mm diameter))eggs, pinkish white at fertilization becoming gray with maturation. Eggs are laid in shallow subtidal areas among rocks and the egg mass is tended by the male (Wilby, 1937). Larvae hatch at about 10 mm TL (9.5 to 11 mm TL, depending on maturity at hatching). Larvae are a very light grayish-silver color with no hint of iridescent blue .(cf. Hexagrammos  decagrammus). The jaw of Ophiodon larvae is more prominent than in Hexagrammos. h^e melanophore patterns are similar. Larval Ophiodon reared in the laboratory formed schools near the tank bottom within three weeks of hatching. 208 14. Hexagrammos decagrammus. (2 mm scale) 0-day, 7 mm TL; 1-day, 9 mm TL; 5-day, 11.5 mm TL; 24-day, 13 mm TL (artist, J. Marliave) 209 i i 15. Ophiodon elongatus. (5 mm scale) 0-day, 9.5 mm TL (premature); 0-day, 10.5 mm TL; 13.5 mm TL, from plankton; 56 mm TL, from plankton; 92-day, 55 mm TL, settled (artist, J. Marliave). 210 F. Cottidae 16. Artedius fenestralls Jordan & Gilbert - Not Illustrated Egg masses of this species are laid in clusters under boulders in the lower intertidal from February to May (in Barkley Sound). Males presumably return to the nest site with different females, since several masses of one developmental stage may be present. Males usually are not present when the tide is out. Each egg mass is uniformly colored, but different masses vary in color from slate gray to light blue to maroon to deep purple. The masses are globular, slightly compressed and laid over-lapping each other (adhering to the underside of the boulder). Larvae hatch at 3,7 mm TL. They have a short gut and body. They have ventral melanophores on the t a i l musculature and a red pigment patch (with melanophores) over the gut. There is no nape pigment at hatching, in contrast to larvae of the genus Clinocottus (Budd, 19^0); Morris, 1951). Clinocottus sp. also have a yellow, rather than red, pigment patch over the gut. Yolk sac Artedius fenestralls are identical to the same stage of A. lateralis except that they are perhaps a bit smaller. By the onset of fin ray formation A. fenestralls larvae have a patch of melanophores on the nape, but none on the head, in contrast to A. lateralis. 17. Artedius lateralis (Girard) Egg /masses of Artedius lateralis are laid i n clusters and positions similar to those of A. fenestralls. They are also laid during the same season. Artedius lateralis is more common on the outer coast than A. fenestralls. but is less common 211 17. Artedius l a t e r a l i s , (2 mm scale) 0-day, 4 mm TL; 30-day, 8 mm TL; 48-day, 11 mm TL, settled; 58-day, 14 mm TL, settled ( a r t i s t s , 'J. Marliave & F. Zi t t e n ) . 212 in inland waters. Spawning of A. l a t e r a l i s takes place from December u n t i l June in Barkley Sound. The egg masses of A. l a t e r a l i s are more compressed than those of A. fenestralis. tapering to a monolayer at the edges of the mass. They are also different i n color, varying from yellow to orange to red. The hatched larvae of A. l a t e r a l i s are described by Budd (1940), This species resembles A. fenestralis at hatching, but later develops a patch of melanophores in the occipital region (hone on the nape). This species hatches at 4 mm TL. Larval A. l a t e r a l i s develop late r a l extensions of the red pig-ment patch over the gut. The extensions distend the body wall over the dorsal insertion of the pectorals. 18. Blepsias eirrhosus (Pallas) I have never loeated eggs of this species. Clemens and Wilby (1961) state that B. eirrhosus spawns in the summer, attaching "clear lig h t brown" eggs in clusters to rocks in shallow water. However, Hart (1973) reports females ripening during early February in Puget Sound. In Barkley Sound I have taken-B. eirrhosus larvae from the plankton from late February to mid-April. Spawning probably occurs during late winter. The larvae are always brown colored. Melanophores are large and evenly spaced over the body back to the caudal peduncle, which i s free of pigment u n t i l metamorphosis. Blepsias eirrhosus larvae may occur in small schools after f i n ray formation. The metamorphosed juveniles swim in the plankton prior to settlement. Blepsias eirrhosus, from plankton. (5 mm scale) 10 mm 14 mm TL, 19 mm TL, 25.5 mm TL (artist, F. Zitten). 2i# 19* Enophrys bison (Girard)) Enophrys bison lays eggs on the exposed surface of rocks in the lower intertidal in late winter and the male guards the eggs (E. De Martini, personal communication), .-v. - Larvae are in 'the plankton during early spring. The four large preopercular spines project at right angles to the head and can be seen plainly from a dorsal or a frontal view. They are distinguished by the spiny head and the Artedius- Clinoeottus body type. 20. Gilbertidia sigalutes (Jordan & Starks) Gilbertidia sigalutes lives in sheltered subtidal crevices. Spawning occurs during August in Barkley Sound. A large male courts several females simultaneously, fending off ripe males which may approach by biting their heads. The large (2.3 mm diameter))pink eggs are laid on solid substrate in a monolayer. The female parents tend the eggs until hatching (late October, early November). The females prevent fungal fouling by brushing their bodies over the eggs. They also fan the eggs during the last few weeks of Incubation. The male patrols the nest from a distance of a few body lengths, signalling to prevent the approach of conspecifics. At hatching the females remove the larvae from the nest site and spit them toward the surface. The planktonlc larval and juvenile stages are described in the main text. The larvae are readily recognized by their rotund body, purple-black in color, with bright orange pectoral fins. In the juveniles the orange color of the pectorals recedes to the I I 19. Enophrys bison, from plankton. (2 mm scale) 10 mm TL. (artist, J. Marliave). ro VJ1 216 20. Gilbertidia sigalutes. from plankton. (5 mm scale) 13 mm TL, 15 mm TL, 25 mm TL, 34 mm TL (artist, J. i 7 mm TL, Marliave). 217 f i n tips, being replaced by yellow. At this stage the head canal pores enlarge to a noticeable size and the body color becomes mottled brown and blaok. 21. Leptocottus armatus Girard I have never located the spawn of Leptocottus armatus. Hart (1973) reports spawning i n February. Jones (1962) reports spawningrfrom October to March in California. Larvae were taken from the plankton in Barkley Sound from October to March. The spawning site i s unknown. Eggs are 1.4 mm diameter and egg masses vary i n color from cream to orange (Jones, 1962). Jones illustrates newly hatched larvae of 4 and 5 mm TL (probably premature). They lack nasal or interorbital pigment and have black pigment patches over the gut. However, head pigment develops later. Older larvae are recognized by the lobed melanin patch over the gut and the extension of the ventral row of melanophores on the caudal musculature up onto the hypural plates. Their body length i s relatively greater than in the Artedius-Clinocottus type. The snout-anus length i s 1/3 total length for Leptocottus versus almost 1/2 (0.44) total length for Artedius sp. 22. Nautichthys oculofasciatus (Girard) Eggs of this species are red colored and 2 mm diameter. They are l a i d on the undersides of rocks i n the mid-intertidal (in Burrard Inlet). The larvae havd been taken from the plankton from January to early May i n Barkley Sound. Eggs were taken in Burrard Inlet during Ap r i l , 1975. 218 21. Leptocottus armatus. from plankton. (2 mm scale) 8 mm TL; 12 mm TL; 13 mm TL, settled (artists, J. Marliave & F. Zitten). 219 Larvae hatch at 9 mm TL. They are distinguished by their large pectoral fins and their orange-brown color, patches of which extend onto the median fin fold. The pectoral fins are fringed with this color. Later stages develop larger pectorals and then an elongate f i r s t dorsal fin. After formation of fin rays, lateral rows of body spines develop. Most prominent are the spines along the lateral line. It should be noted from the illustrations that this and some other species of cottids, as well as?some agonids, develop rays of the pectoral fins before hatching-. These species a l l use their pectoral fins to hover and to maneuver. Other fish larvae typically .pass through development of the median fins before pectoral rays form. 23. Psvchrolutes paradoxus Qflnther I know, nothings of the spawning of this species. The larvae were caught in the plankton during February, March and May in Barkley Sound. Larval P. paradoxus are easily mistaken for larval Gilbertidia  sigalutes. However, P. paradoxus larvae are in the plankton later during the year, are relatively smaller and have relatively more melanin. The melanin obscures the orange of the pectoral fins. Also, this species develops a dorsal notch in the caudal fin fold during formation of fin rays. This caudal notch is absent in G. sigalutes larvae. Finally, both species hatch with fully developed pectoral fins and can be distinguished by the ray counts (15-1? in G. sigalutes versus 20-23 in P. paradoxus. 220 22. Nautichthys oculofasciatus. from plankton. (5 mm scale) 9.5 mm TL; 13 mm TL; 1? mm TL; 26 mm TL, settled (artist, F. Zitten) I I 23. Psychrolutes paradoxus, from plankton. (2 mm scale) 10.5 mm TL 13 mm TL; 14 mm TL; 13 mm TL, settled ( a r t i s t , J . Marliave). 222 . Psychrolutes paradoxus settles at a small size (13-14 mm TL), with the head length 25* of total length. After settlement the head grows to a relatively larger size (head 30* of TL at 18 mm TL). No tendency for repeated settlement was noted. 24. Rhamphocottus richardsoni Gunther The spawn of this species was never located. Hart (1973) reports that "Yellow to orange eggs are produced during winter." Larvae were taken from the plankton in Barkley Sound from February to early May. Some specimens taken in early February had fin rays. This species has an unusual head profile, even as a larva. This profile, together with the short body, preopercular spines and uniform orange-brown color (not extending onto the caudal peduncle), distinguish larvae of this species. After fin ray formation, spines form over the entire body. This species settles in a manner similar to Nautichthys  oculofasciatus. the metamorphosing fish hovering against walls. 25• Scorpaenichthys marmoratus (Ayres) - Not Illustrated One cluster of egg masses of this species was found tended by a parent on moderately exposed bedrock at four meters depth. The egg masses hatched during the entire month of August. Larvae hatch at about 6 mm TL, with a large yolk including a single o i l droplet. The larvae have melanophores over the gut and musculature. The larvae become black with iridescent blue overtones (especially over the gut). Pigment is absent on the caudal peduncle. Except for color, the early-stage larvae resemble those of Blepsias eirrhosus. The later larvae resemble Rhamphocottus  richardsoni, except that they have a more blunt snout. Larvae of Scorpaenichthys marmoratus are well illustrated by O'Connell (1953). 223 24. Rhamphocottus richardsoni. from plankton. (2 mm scale) 10 mm TL; 11.5 mm TL; 15 mm TL, settled (artist, D. Roman). ," 224 F. Agonidae 26. Bothragonus swani (Steindachner) An egg mass of B. swani was found within the rhizomes of a Macrocystis integrifolia holdfast* The eggs are 2 mm diameter, brown in color. The larvae hatch at 7.5 mm TL with fully developed peetoral fins (similar to those of Nautichthys oculofasciatus). This species has large, thin teeth at hatching and prefersito eat fish larvae. Like other agonid larvae, the body plates are in the form of spines during the larval period. 27. Pallasina barbata (Steindachner) I know nothing of the spawning of this species, except that the larvae were eaught during April in Barkley Sound. Pigmentation of these larvae is entirely melanic, contrasting with the black and white pigment patches on a l l stages of Xeneretmus  latifrons larvae. Fin position, fin ray counts, body spines (plates) and head morphometry distinguish larvae and planktonlc juveniles of this species. 28. Xeneretmus latifrons (Gilbert) Recently hatched larvae of this species occur commonly in the plankton of Barkley Sound during March and April. Fitch and Lavenberg (1968) report spring spawning of this species. The dorsal fin fold of this elongate larva has four prominent white patches on the margins (not evident in the illustrations). There is also a white patch over the anus. The dorsal white patches accentuate black patches under them. The jaws are very melanic. The body develops a hexagonal lattice pattern of melanin. 225 i - _ i 26. Bothragonus swani. (2 mm scale) 1-day, 7*5 mm TL; 11-day, 10 mm TL; 26-day, 12 mm TL; 48-day, 16.5 mm TL (artists, J. Marliave & F. Zitten). i 1 ro ON 27. Pallasina barbata. from plankton. (5 mm scale) 10 mm TL, 17 mm TL, 28 mm TL (artist; J. Marliave). 28. Xeneretmus latifrons. from plankton. (5 mm scale) 7 mm TL, 10 mm TL, 21 (artist, J. Marliave). F. Cyclopteridae 22B 29. Liparis fucensis Gilbert Ih late May, 1974, Dan Pace collected a newly laid egg mass of this species from among the tubes of the tubeworm Eudlstylia  polymorpha. The egg mass was attended by a ripe male. The eggs were attended by a ripe male. The eggs are small (almost 1 mm diameter) and light orange in color. At ll°Cr, the eggs developed to hatching in two weeks (15 to 16 days), turning to a silver-gray color. They reached the eyed stage in 10 days. Larvae hatch at 3 mm TL. The pectoral fins have a honey-comb (hexagonal) melanin pattern. The fin fold appears granular in transmitted light. The jaw is subterminal and the mandible develops melanic color. There is a ventral row of melanophores on the t a i l musculature. An unusual feature is a second ventral row of melanophores which develops on the fin fold. The early larval stages are laterally compressed (as are pleuronectid larvae) and the large subdermal spaee extends around the cranial region. Late larval stages of liparids resemble the adults. Ih the field, liparid larvae were observed to prefer clean polyethylene sheets to similar sheets covered with diatom growth for settlement. 230 io -4 5:H 0 io H B 5-1 p O G O P O ^ © ^ Q ft 2 T pp p o p JU3 o op -POZpl JOO-OD 1 o • S P o o R ISA a i o - i 5-4 o o o a © pi C L © o 1' -©o ©o 1 o o o o PPOP DO. 2P poo O ^ oo o o PPO O ~ p . O o O o oo o oo o OO Q oo o OO o oooo_o oooo 0 1 Q O O O I O ^ nooo_o" ooo_o_, ooo_o_i poptpr o o o 4s m o s ° M oo ppp o o o o 5 oo o o 5 oo o OOO oo o OQOOQOO O 1 UQOOOO OOO M P Q O O O Q O O g O Q O O O O O _ oppppoooi £283 La o p o o Uo P Q o oo oo. oo op -QQ or o o °9 oo o NOVESER D E C E M B E R JANUARY FEBRUARY MARCH APRIL M A Y Appendix J. Summary of a l l plankton tows made during the winter of 1973/197-+. C i r c l e s represent subsurface tows, squares - surface tows, and diamonds - surface tows taken at dawn. Each dot represents one G i l b e r t i d i a sigalutes taken during a tow. 231 Appendix Catch per unit e f f o r t of plankton tows made during the winter of 1973/197^ (nunbers of G i l b e r t i d i a sigalutes per tow, by half month periods). Circles indicate zero values. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items