UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The feeding ecology of larval and pelagic juvenile redfish (Sebastes SPP.) on Flemish Cap (47 N, 45 W) Anderson, John T. 1992

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

Item Metadata

Download

Media
831-ubc_1992_spring_anderson_john_t.pdf [ 3.81MB ]
Metadata
JSON: 831-1.0086664.json
JSON-LD: 831-1.0086664-ld.json
RDF/XML (Pretty): 831-1.0086664-rdf.xml
RDF/JSON: 831-1.0086664-rdf.json
Turtle: 831-1.0086664-turtle.txt
N-Triples: 831-1.0086664-rdf-ntriples.txt
Original Record: 831-1.0086664-source.json
Full Text
831-1.0086664-fulltext.txt
Citation
831-1.0086664.ris

Full Text

T H EF E E D I N GE C O L O G Y O FL A R V A L A N DP E L A G I C J U V E N I L ER E D F I S H ( S E B A S T E S S P P . ) O NF L E M I S H C A P ( 4 7 ° N ,4 5 ° W )b yJ o h nT . A n d e r s o nB . S c . , U n i v e r s i t y o f G u e l p h , 1 9 7 3M . S c . , U n i v e r s i t yo fG u e l p h ,1 9 7 8T H E S I S S U B M I T T E DI N P A R T I A L F U L F I L L M E N T O FT H E R E Q U I R E M E N T S F O RT H E D E G R E E O FD O C T O R O FP H I L O S O P H Yi nT H E F A C U L T Y O F G R A D U A T ES T U D I E S( D e p a r t m e n to fZ o o l o g y )W e a c c e p tt h i st h e s i sa s c o n f o r m i n gt o t h e r e q u i r e d s t a n d a r dT H E U N I V I T YO FB R I T I S HC O L U M B I AJ a n u a r y1 9 9 2Ø J o h nT r u m a n A n d e r s o n , 1 9 9 2S i g n a t u r e ( s )  r e m o v e d  t o  p r o t e c t  p r i v a c yNational Libraryof CanadaBibliothêque nationaledu CanadaCanadian Theses Service Service des thèes cariadiennesOttawa. CanadaK1AON4The author has granted an irrevocable nonexclusive licence allowing the Nalional Utxaryof Canada to reproduce, loan, dstiibute or sellcopies of his/her thesis by any means and inany form or format, making this thesis availableto interested persons.The author retains ownership of the copyrightin his/her thesis. Neither the thesis norsubstantial extracts from it may be printed orotherwise reproduced without his/her permission.-L’auteur a accordé une licence irrevocable etnon exclusive permettant a Ia 8ibliothéquenatiönale du Canada de reproduire, prêter,distribuer ou vendre des copies de sa thesede quelque manlére et sous quelque formeque ce soit pour mettre des exemplaires decette these a Ia disposition des personnesintéressées.Lauteur conserve (a propnété du droit d’auteurqui protege sa these. Ni Ia these ni des extraitssubstantiels do celle-ci ne doivent êtreirnprimés ou autrement reproduits sans sonautorisation.CaxiadISBN 0-315-75400-1I n p r e s e n t i n gt h i st h e s i si n p a r t i a lf u l f i l m e n to ft h e r e q u i r e m e n t sf o r a na d v a n c e dd e g r e ea tt h eU n i v e r s i t yo f B r i t i s hC o l u m b i a ,I a g r e et h a tt h eL i b r a r ys h a l l m a k e i tf r e e l ya v a i l a b l e f o rr e f e r e n c ea n ds t u d y . If u r t h e ra g r e et h a t p e r m i s s i o nf o r e x t e n s i v ec o p y i n go ft h i st h e s i sf o rs c h o l a r l y p u r p o s e sm a yb e g r a n t e d b yt h e h e a d o fm yd e p a r t m e n to rb y h i s o rh e r r e p r e s e n t a t i v e s .I t i su n d e r s t o o dt h a t c o p y i n go rp u b l i c a t i o no f t h i st h e s i sf o r f i n a n c i a lg a i ns h a l l n o tb ea l l o w e dw i t h o u t m yw r i t t e np e r m i s s i o n .D e p a r t m e n t o f 0L0 eS 7T h eU n i v e r s i t yo f B r i t i s hC o l u m b i aV a n c o u v e r ,C a n a d aD a t eFY’17D E - 6( 2 / 8 8 )S i g n a t u r e ( s )  r e m o v e d  t o  p r o t e c t  p r i v a c yAbstractThe copepod Calanus finmarchicus dominated the biomass of all invertebratezooplankton species sampled on Flemish Cap during 1977-1983 as sampled by bothcoarse (0.333 mm and 0.505 mm) and fine (0.165 mm) mesh samples. Oithona similisand Oithoma atlamtica were numerically dominant in fine mesh samples. Significantdifferences were observed in the rate of C. fimmarchicus development and absoluteabundance among the years sampled on Flemish Cap.Copepod eggs and nauplii were preferred prey of redfish larvae whereas Oithonaspp. copepodites were not selected by larvae. The preferred prey of juveniles includedOithona copepodites, in addition to eggs and nauplii. This shift at metamorphosis-was a function of changing prey availability associated with the seasonal succession ofzooplankton. Significant differences in diet between 1979 and 1980 compared to 1981resulted from differences in prey availability which appeared to be due to differencesin the seasonal development of C. finmarchicus. When copepod development startedearlier and was faster, the diet of redfish was predominantly Oithona spp. Underthose conditions, the redfish had lower amounts of food per stomach, lower relativecondition and delayed age of metamorphosis at larger sizes.Higher feeding and growth rates occurred in 1980 in association with lower watertemperatures and prey concentrations, compared to 1981. Differences in feedingand growth rates between years were related to a shift in available prey, wheredevelopment of C. finmarchicus occurred earlier in 1981 and the zooplankton weredominated in May by Oithona spp. nauplii and copepodites. It appears that anearlier occurrence and subsequent decline in the abundance of C. finmarchicus eggs11and nauplii was detrimental to the feeding and growth of redfish larvae in 1981.Higher growth and survival rates occurred in 1980 compared to 1981. Similarmean lengths of the population measured in August (Day 215) in 1980 and 1981,compared to different growth rates measured each year, indicates that only largerredfish survived in 1981. It is concluded that variations in physical circulation during this study, 1979—81, would not explain differences in annual variations of abundances. Comparison of relative redfish abundances at different ages demonstratedrecruitment at age 5 was determined during the larval stage for the years 1979-81.111ContentsAbstract iiContents ivList of Tables viiiList of Figures xviAcknowledgements xxii1 General Introduction 11. 1 Theory and Background 11.2 Study Area 42 Invertebrate Zooplankton Community2.1 Introduction2.2 Materials and Methods2.2.1 Field Collections2.2.2 Laboratory Processing2.2.3 Statistical Analyses2.2.4 Estimates of Copepod Development2.3 Results2.3.1 Zooplankton Biomass2.3.2 Zooplankton Community2.3.3 Timing of Annual Copepod Spawning on Flemish Cap . .778810131415151619iv2.3.4 Interannual Differences 232.3.5 Predictions of Peak Spawning of Calanus finmarchicus . . 292.4 Discussion 312.5 Summary 373 Feeding of Redfish3.1 Introduction3.2 Materials and Methods3.2.1 Laboratory Processing3.2.2 Feeding Analyses3.2.3 Selectivity.3.2.4 Redfish Condition3.3 Results3.3.1 Seasonal Patterns in Diet3.3.2 Among Year Differences in Diet .3.3.3 Diurnal Patterns in Feeding3.3.4 Among Year Differences in Feeding3.3.5 Feeding Selectivity3.3.6 Redfish Condition3.3.7 Metamorphosis from Larvae to Pelagic3.4 Discussion3.4.1 Diet and Feeding3.4.2 Metamorphosis3.4.3 Match/Mismatch of Redfish and their Prey.3939404044495253535759616265Juveniles 7172727881V3.5 Summary. 834 Redfish Growth and Feedingand Temperature4.1 Introduction4.2 Methods4.2.1 Growth4.2.2 Prey Concentration4.2.3 Standardized Stomach4.2.4 Temperatures4.2.5 Analyses4.3 Results . .4.3.1 Temporal Comparisons4.3.2 Spatial Comparisons4.4 Discussion4.5 Summary5.3 Physical Dispersal of Redfish Larvae on Flemish Cap 1225.4 Relationship of Larval Survival to Recruitment5.5 Summary6 Literature Cited 131in Relation to Prey ConcentrationContents85858787899293939595991031111141141155 Survival of Redfish on Flemish Cap5.1 Introduction5.2 Larval Survival and Growth Rate127129viA Appendix 153B Appendix 161vuLIST OF TABLESTable 1. Summary of the stations sampled and the number of samples processedfor invertebrate zooplankton analyses from the Flemish Cap P. 169Table 2. Ratios of Calamus fimma’rchicus mean density (rn—3) and weight (mg m3)to Oithona sp. at different times on Flemish Cap from small mesh samples(0.165 mm). Weights were estimated by the sum of total weight in each copepodite stage for each species for each cruise, standardized for the numbers ofstations in each cruise. Copepodite weights from Tremblay (1981) P. 170Table 3. Mean densities of copepod eggs and nauplii (number m3) sampled onFlemish Cap for different depth strata from 0.080 mm mesh samples . . P. 171Table 4. Densities (number m3) of the copepod Calanus finmarchicus copepoditestages on Flemish Cap for the different strata based on the 0.165 mm meshsamples P. 172Table 5. Comparison of Calamus finmarchicus copepodite stage densities (numberm3) in different years for waters 400 rn depth on Flemish Cap, from 0.165mm and 0.333 mm mesh samples P. 173Table 6. Comparison of mean densities (number rn3) of Calanus finmarchicus andother copepods among years on Flemish Cap within different depth strata for0.165 mm and 0.333 mm mesh samples. Other copepods refers to those ‘other’than C. finmarchicus samples P. 174Table 7. Summary of non-parametric analyses of differences among years for cope-pods from 23 April - 9 May 1979 versus 2-9 May 1981 within different depthvu’strata for 0.333 mm and 0.505 mm mesh samples P. 175Table 8. Summary of non-parametric analyses of differences among years for cope-pods from 21-32 May 1980 versus 22-27 May 1981 within different depth strataand for different mesh size samples P. 176Table 9. Non-parametric analyses of differences among years for copepods duringthe July-August period 1979-82 within different depth strata for 0.333 mmmesh samples (one-way ANOVA on rank scores for F-value, P-levels and Dun-cans multiple range test). Other copepods refers to those other than Calanusfinmarchicus. The years refer to the following dates: 10-14 July 1979, 20-29July 1980, 1-4 August 19-81- and 1-3 August 1-982 .. P. 177Table 10. Flemish Cap surface water temperatures (°C) for 10 m depth withinthe central area 200 in water depth bounded by the area 46°30’—47°48’N and44°6’—46°6’W. Mean values were based on temperature data from the MarineEnvironmental Data Service (MEDS), METOC sea surface temperatures, andfrom data collected during directed Canadian research cruises as part of theFlemish Cap Project. Ranges of temperatures represent the 99% C. I. , inparantheses P. 178Table 11. Temperature dependent predictions of days since spawning for copepodite stages of Calanus finmarchicus on Flemish Cap, based on Belehradekequations (see text for explanation) P. 179Table 12. Summary of samples collected on each cruise and the number of redfishanalyzed for diet on Flemish Cap, 1978-82. Redfish Size Class refers to theixmean and range of the number of specimens examined for each mm size P. 180Table 13. Summary of redfish diet expressed as percent occurrence based on bothweight and numbers. Here diet is summarized for each cruise for all sizes ofredfish and cruises are listed chronlogically by season. Diet is ranked fromhighest to lowest in each case. Each prey item represents the best possibletaxonomic identification possible from stomach samples P. 181Table 14. Estimates of prey width for selected prey types as a percent of maximummouth width for different redfish sizes. NI and NVI refer to nauplii stages Iand VI, and CI and CVI refer to copepodite stages I and VI. Width estimatedfor Cthmvs fimmarthirs uauplii was 75% of total length and for copepoditeswas 33%. For Oithona similis naup]ii width was estimated to be 67% of totallength and for copepodites was 34%. Width estimates approximated inclusionof folded antennae P. 184Table 15. Statistical results of comparison of redfish diet (% weight) among yearsduring the period March-May 1979-8 1. Each entry for upper, middle and lowernumbers are the Kolmogorov-Smirnov D test statistic, P-level and the numberof observations n1 and n2. Redfish Sizes refer to the overall size range of fishfor which diet comparisons were made P. 185Table 16. Statistical results summarizing the comparisons of redfish diet (% weight)among years during the July-August period 1980-82 on Flemish Cap for threesize classes: 9 mm, 10-19 mm, 20 mm. Each entry for upper, middle andlower numbers are the Kolinogorov-Smirnov (Dmax) test statistic, the probability level (P-level) and the number of observations n1 and n2. n.s. refers to thextest statistic being less than the critical value where o 0.1, based on tablesL11 and L111 in Siegel and Castellan (1988). NA indicates no comparison waspossible P. 186Table 17. Mean stomach weight (tg) standardized for fish size (mm3) for redfishlarvae sampled on Flemish Cap. Mean standardized stomach weights werecalculated for each size class within each cruise and then an overall mean wascalculated from these mean estimates. (n = number of redfish size classes, s.d.standard deviation, CV = coefficient of variation) P. 187Table 18. Wilcoxon’s test of differences between standardized stomach weights(weight of st&mach divided by the weight of redfish) for each mm length groupsampled May 1980 versus 1981. n1 — refers to 20-26 May 1980 and n2 — to22-27 May 1981 P. 188Table 19. Summary of redfish prey items classified as “Other” in the feeding analyses. The count represents the total number observed in all redfish stomachsanalyzed for each cruise and the percent ranks the relative abundance of theseprey items for each cruise P. 189Table 20. Summary of redfish sizes (mm) and number of stations (n) used in theselectivity analysis using the 0.080 mm mesh zooplankton samples .... P. 190Table 21. Summary of redfish sizes (mm) and number of stations (n) used in theselectivity analysis using the 0.165 mm mesh zooplankton samples .... P. 191Table 22. Summary of dry weight versus length (logio-logio transformed) leastsquares regressions for different cruises. The overall regression includes allxcruises except DAWO79, 10-14 July 1979, in which only small redfish were sampled and dry weights were significantly low. DW — refers to dry weight (tg),SL — refers to standard length (mm) P. 192Table 23. Summary of statistical tests of differences in redfish dry weight (tg) ateach length group (mm) between 20-26 May 1980 versus 22-27 May 1981. n1and n2 refer to the sample size for 1980 and 1981 data, respectively. z — refersto Wilcoxon’s test statistic P. 193Table 24. Summary of dry weight versus fish volume (logio-logio transformed leastsquares regressions for different cruises. The overall regression is for all cruises.Table 25. Summary of regression comparisons in dry weight (/Lg) versus fish volume(tgDW) among all cruises. All paired comparisons between cruises are basedon weighted least squares regressions P. 195Table 26. Summary of fish volume versus length (logio-logio transformed) leastsquares regressions for different cruises. The overall regression is for all cruises.FV — refers to fish volume in units of dry weight (gDW), SL — refers to standardlength (mm) P. 196Table 27. Summary of statistical tests of differences in redfish fish volume (gDW)at each length group (mm) between 20-26 May 1980 versus 22-27 May 1981.n1 and n2 refer to the sample size for 1980 and 1981 data, respectively. z —refers to Wilcoxon’s test statistic P. 197Table 28. Summary of regression comparisons in fish volume (gFV) versus lengthxl’(mm) among various cruises for the July-August period 1980-82. Redfish wereselected to range between 6-23 mm in length P. 198Table 29. Summary of statistical tests comparing the differences in fish volume(igDW) for different length groups (mm) between cruise pairs for the July-August period 1980-82. The sample sizes are indicated in parantheses for meanfish volume values. The overall values are for all length ranges within each cruisecomparison P. 199Table 30. Summary of the number of redfish examined for age and growth ratebased on otoliths for each cruise, and the number of stations in each cruise forwhithageanclgrowtkdatawereavai1ab1...... P. 200Table 31. Copepod egg and nauphi densities (number in3) averaged for each cruisefor waters 400 m depth, from 0.080 mm mesh samples P. 201Table 32. Ranges of variables used in the spatial analyses from 3 cruises, 20-26May 1980, 22-27 May 1981 and 26-30 June 1981. The values represent rangesof means calculated for each variable at each station. Prey concentrations forthe two May cruises are based on 0.080 mm mesh samples and for the June1981 cruise are based on 0.165 mm mesh samples P. 202Table 33. Spearman rank correlations comparing redfish growth and feeding ratesto redfish size and condition, water temperatures and prey biomass for threecruises: 20-26 May 1980 (GADO37); 22-27 May 1981 (GADO51); 26-30 June(HAWPAN). MMGI - mean growth (mm d’), MGI5 = growth rate for the last5 days (mmd1), MGI1O = growth rate for the last 10 days (mmd1), MSTD_Xxlii= diurnally corrected standardized feeding rate (mg mniz9, MEANSIZE =standard length (mm), MDW = dry weight (tg), MDW_SL = relative condition(g mmd), MPOSTX = age (days), SURTEMP = surface water temperatures( 20 in), PREY1BIO = copepod eggs (mg rn3), PREY2BIO = copepodeggs and nauplii (rng rn3), PREY3BIO = copepod eggs, nauplii, Calanusfinmarchicus and Oithona copepodites (rng m3), PREY4BIO = cyclopoid eggsand nauplii, 1981 data only (rng rn3), PREY5BIO = cyclopoid copepodites(mg m3) P. 203Table 34. Spearman rank correlations comparing redfish size, condition and age totemperature and prey biomass for three cruises: 20-26 May 1980 (GADO37);22-27 May 1981 (GADO51); 26-30 June (HAWPAN). MEANSIZE = standardlength (mm), MDW = dry weight (rig), MDW..SL relative condition (tgrnmd1),MPOSTX = age (days), SURTEMP = surface water temperatures (20 m), PREY1BIO = copepod eggs (mg m3), PREY2BIO = copepod eggs andnauphi (mg m3), PREY3BIO = copepod eggs, nauplii, Calamus finmarchicusand Oithona spp. copepodites (mg m3), PREY4BIO = cyclopoid eggs andnauphi (1981 data only) (mg m3), PREY5BIO = cyclopoid copepodites (rngm3) P. 205Table 35. Spearman rank correlations comparing the redfish feeding index to growthrates for three cruises: 20-26 May 1980 (GADO37); 22-27 May 1981 (GADO51);26-30 June (HAWPAN). MMGI- mean growth for all ages (mm d9, MGI5= growth rate for the last 5 days (mm d), MGI1O = growth rate for the last10 days (mm d), MSTD..X = diurnally corrected standardized feeding indexxiv(mg mm’).P. 207Table 36. Regressions analyses of redfish growth, size and relative condition withprey concentrations and temperature, based on spatial differences measuredduring 3 different cruises. Food is standardized food for Prey Type 2 andTemperature (°C) is for surface waters 20 m depth. C = mean growth (mmd—1), G5 = recent 5-day growth (mm d—9, GlO = recent 10-day growth (mmd-1), SL = standard length (mm), DW = dry weight (gig), RC = relativecondition (gDW mm-1) P. 208Table 37. Redfish abundances estimated for each cruise carried out on FlemishCap,- for all lengths- samplei. P. 2.09-Table 38. Redfish population abundances estimated in different years for a) thetime of peak spawning standardized to Day 120 and b) during early Auguststandardized to Day 215. Abundance esitmates for Day 215 are made only forlarvae spawned in April. Projections are made forward and backward in timefrom sampled population estimates using different estimates of mortality (rateof change per day). a) Peak larval abundances on Day 120. b) Pelagic juvenileabundances on Day 120 P. 210Table 39. Abundance estimates of redfish (Sebastes spp.) at different ages. Seetext for data sources. Trawl — number caught per tow, Stomach— numberper cod fish stomach, Larvae and Juveniles — abundance estimated for FlemishCap, CPUE - catch per unit effort P. 211xvLIST OF FIGURESFigure 1. Flemish Cap bank, centered at 47°N, 45°W, is situated west of the GrandBank of Newfoundland. The boxed area represents the maximum area in whichsamples were collected as part of this study P. 212Figure 2. Seasonal changes in plankton volume (ml m3)of invertebrate zooplank..ton on Flemish Cap. The plots represent a composite of data collected indifferent years from different mesh samples: (a) 0.333 mm mesh, (b) 0.165 mmmesh, (c) 0.080 mm mesh. Vertical bars represent one standard deviation aboutthe individual cruise means P. 214Figure 3. Seasonal changes in percent community composition of invertebrate zoo-plankton on Flemish Cap broken down by selected groups. This is a compositepicture of data collected in different years from 0.333 mm mesh samples. Referto the text for details P. 216Figure 4. Seasonal changes in percent community composition of invertebrate zoo-plankton on Flemish Cap broken down by selected groups. This is a compositepicture of data collected in different years from 0.165 mm mesh samples. Referto the text for details P. 218Figure 5. Seasonal changes in the percent composition of Calanus finmarchicuscopepodite stages on Flemish Cap in waters 400 m depth. This is a compositepicture of data collected in different years from 0. 165 mm mesh samples. Referto the text for details P. 220Figure 6. Development times predicted for Calamus finmarchicus in different yearsxvibased on temperature dependent Belehradek equations. CI refers to copepoditestage CI, etc. * refers to date from which projections of stage development weremade P. 222Figure 7. Schematic representation of morphometric measurements made on red-fish indicated on a pelagic juvenile and in box form P.224Figure 8. Comparison of redfish feeding for different cruises/dates on Flemish Cap.Redfish were subdivided into one of 3 length groups: Small: 9 mm, Medium:10-19 mm and Large: 20 mm. In addition, for each length group the meansize and number of fish measured is giycn. The prey item 1ab1s are: A-- pod Eggs; B - Copepod Nauplii; C - Copepod Copepodites; D - Copepods NotSpecified; E - Cyclopoid Eggs; F- Cyclopoid Nauplii; G - Cyclopoid Copepodites; H - Cyclopoid Not Specified; I - Calanus finmarchicus Copepodite; J- Oithona spp. Copepodite; K - Limacina spp. ; L - Euphausiid; M - Phytoplankton; N- Prey Type Not Specified; 0 - Empty stomach P.226Figure 9. Mean size of Limacina sp. (mm) observed in the diet of redfish duringdifferent times of the year for 1979, 1980 and 1981 data P. 231Figure 10. Direct comparison of feeding differences for redfish larvae sampled indifferent years at approximately the same time: 20-24 March 1979 versus 6-13April 1980, 23-27 April 1979 versus 2-9 May 1981, 20-26 May 19890 versus22-27 May 1981. In each case the comparison is for all redfish as there were nosignificant feeding differences among size classes within each cruise. The preyXviiitem labels marked on the figure axis are: A - Copepod Eggs; B - CopepodNauplii; C - Calanoid Copepodites; D - Cyclopoid Nauplii; E - CyclopoidCopepodites; F - Euphausiid Naupili; and G - Limacina spp P. 233Figure 11. Diurnal patterns of redfish stomach weight (tgDW) averaged for eachcruise, ± 2 standard errors. Interpolation between each 2 hour period issmoothed using a spline fit to the means. Sunset (dark shaded) and sunrise(open) are indicated at the bottom of each panel P. 235Figure 12. Redfish stomach weights standardized by redfish dry weight (± 2 se)versus redfish length (mm) for 20-26 May 1980 and 22-27 May 1981 .. . P. 237Figure13. Redfish feeding e1ectivity calculated for eight food trpes for 4 cruisesusing 0.080 mm mesh zooplankton samples. Mean selectivity is plotted withminimum and maximum values calculated for each food type within a cruise.(a: = Chesson’s alpha index for prey species i). The prey item labels are: A -Copepod Eggs; B - Cyclopoid Nauplii; C - Copepod Nauplii; D - Oithona spp.Copepodite; E - Limacina spp. ; F - Euphausiid; G - Calanus finmarchicusCopepodite; H - Copepodites Not Specified; I - Empty stomach P. 239Figure 14. Redfish feeding selectivity calculated for eight food types for 2 cruisesusing 0.165 mm mesh zooplankton samples. (c = Chesson’s alpha index forprey species i). Redfish analyzed 26-30 June 1981 ranged in length from 9-15 mm and 1-4 August 1982 from 16—23 mm. The prey item labels are: A -Copepod Eggs; B - Cyclopoid Nauplii; C - Copepod Nauplii; D - Oithona spp.Copepodite; E - Limacina spp. ; F - Euphausiid; G - Calanus finmarchicusCopepodite; H - Copepodites Not Specified; I - Empty stomach P. 241xviiiFigure 15. Dry weight (pg) versus redfish length (mm) regression residuals plottedagainst redfish length (log10 mm). All cruises with dry weight data exceptDAWO79 P. 243Figure 16. Percent differences in dry weight (zg) and fish volume (mm3) between20-26 May 1980 versus 22-27 May 1981. % difference is calculated relative to20-26 May 1980 P. 245Figure 17. Fish volume (mm3)versus redfish length (mm) regression residuals plotted against redfish length (logio mm). All cruise for which data was availableP. 247Figure 18. lVfetamorphosis of redfish from larvae to pre-juveniles determined bythe degree of flexion at different lengths (mm) for a) 20-26 May 1980 and b)22-27 May 1981. Flexion stages 1 to 3 refer to pre-flexion, in flexion and postflexion, respectively P.249Figure 19. Redfish growth rates (mmd1) versus water temperature (°C) averagedfor the months March through July 1980 and 1981 on Flemish Cap. Solid linesconnect months within each year, dashed lines connect months between yearsP. 251Figure 20. Redfish growth (mm d’) versus feeding rates (tg mm3)averaged foreach cruise on Flemish Cap in 1980 and 1981. The solid lines connect cruiseswithin each year, the dashed line connects cruises conducted at the end of May1980 and 1981 P. 253xixFigure 21. Redfish growth rates (mm d’) versus water temperature (°C) averagedfor each cruise on Flemish Cap in 1980 and 1981. The solid lines connect cruiseswithin each year, the dashed line connects cruises conducted at the end of May1980 and 1981 P. 255Figure 22. Redfish feeding rates (ttg mm—3) versus water temperature (°C) averaged for each cruise on Flemish Cap in 1980 and 1981. The solid lines connectcruises within each year, the dashed line connects cruises conducted at the endof May 1980 and 1981 P. 257Figure 23. Redfish growth rates (mm d—’) versus prey concentrations (mg m3)sampled with 0.080 mm mesh nets for different cruises in 1980 and 1981. Dashedlines connect cruises within each year. a) Prey Type 1— Copepodeggs, b) PreyType 2— Copepod eggs and nauplii, c) Prey Type 3— Copepod eggs and nauplii,Calarius finmarchicus and Oithona spp. copepodites P. 259Figure 24. Redfish feeding rates (mg mm3)versus prey concentrations (mg m3)sampled with 0.080 mm mesh nets for different cruises in 1980 and 1981. Dashedlines connect cruises within each year. a) Prey Type 1 — Copepod eggs, b) PreyType 2— Copepod eggs and nauplii, c) Prey Type 3— Copepod eggs and nauphi,Calamus fiumarchicus and Oithoua spp. copepodites P. 261Figure 25. Redfish growth rates (mm d1) averaged for each station versus theconcentration of copepod eggs and nauplii (log mg m3) for the cruises 20-26May 1980 (GADO37) and 22-27 May 1981 (GADO51). a) mean growth ratesaveraged over all ages, b) growth rates for the last 5 days of life, c) growth ratesfor the last 10 days of life P. 263xxFigure 26. Redfish growth rates (mm d—1) averaged for each station versus theconcentration of copepod eggs and nauplii (log mg rn3) for the cruise 26-30June 1981 (HAWPAN). a) mean growth rates averaged over all ages, b) growthrates for the last 5 days of life, c) growth rates for the last 10 days of life . . P.265Figure 27. Growth rates (mm d—1) calculated over 5-day periods for cohorts ofredfish released on different dates ranging from Days 90 to 140 on FlemishCap in 1980 and 1981. The periods of larval release and of seasonal decline ingrowth rates are indicated at the bottom of each panel P. 267Figure 2& Population abundance-s estimated foi April-released redfish during different cruises, 1978-1983, on Flemish Cap. Values on lines are calculated ratesof change (d—’) between cruises in different years. Hatched area representsestimated peak release period of redfish on Flemish Cap P. 269Figure 29. Mean redfish length (mm) calculated for the April-released redfish sampled during different cruises, 1978-83, on Flemish Cap. The line is the bestexponential fit for all values P. 271Figure 30. Relative estimates of the anticyclonic circulation of surface waters onFlemish Cap based on dynamic height calculations a) during different seasonsfor all available data, and b) for different years during this study, 1979—81 . P.273Figure 31. Relative abundances of redfish calculated at different times for the 3year classes, 1979-81. See text for data sources P. 275xxi.AcknowledgementsI would like to thank my supervisor, Carl Walters, who provided me the opportunity to return to academic life to pursue my PhD degree and, as well, for histhorough assistance in helping to bring my thesis to completion. I am also indeptedto the friends I made at UBC, in particular Jeremy Coffie and Mike Lapointe, whohelped encourage me through some of those long nights while working on my own5,000 km away. In particular, I am indebted to my wife, Margaret, who shares agiant part of this thesis, to my children Jessica and Ryan, and to the rest of myfamily who were always there in support; it would not have been possible withoutthem.xxii1 General Introduction1.1 Theory and BackgroundCauses of recruitment variation in fish have been studied over several decades. During this time many hypotheses have been advanced to explain varying survival duringthe pre-recruit stages of fish (reviewed by Anderson 1988). There is still uncertaintyon which life stage determines year-class strength (Sissenwine 1984). Historically, ithas been considered that year-class strength is largely determined during the firstyear of life and, particularly, during the larval stage (Hjort 1914, Gulland 1965,Templeman 1972, Cushing 1974, Shephard and Cushing 1980). Several recent studies support this observation (Rauck and Zijistra 1978, Leggett et al. 1986, vander Veer 1986, Campana et al. 1989, Hollowed and Bailey 1989, Sundby et al.1989). Among the many factors hypothesized to effect survival, starvation duringfirst feeding (Hjort 1914, Cushing 1975, Lasker 1975, 1981) has received considerable attention. However, starvation alone has not explained variations in the annualsurvival of marine fish larvae (reviewed by Anderson 1988).The theoretical basis of my work is derived from the observation that mortality rates decline with increasing size (Ware 1975, Peterson and Wroblewski 1984,McGurk 1986). The concept that size—specific growth rates interact with size—dependent mortality to determine survivorship in fish populations has long been atenet of fisheries theory (Gulland 1965, Cushing and Harris 1973, Cushing 1974).Simply stated, the theory predicts that annual survival of a cohort is directly relatedto growth rates during the pre-recruit period (Ware 1975, Shepherd and Cushing1980). Factors affecting growth rate in fish presumably will be some function of food1availability (ration) together with the direct and indirect effects of temperature. Theprimary cause of death is assumed to be predation.Fish develop through several life history stages during the first year of life. Transitions from one stage to the next typically result in significant reductions in boththe magnitude and variation of mortality rates. Major reductions in stage-specificmortality from egg through juvenile stages have been demonstrated for plaice (Zijistra et al. 1982, van der Veer 1986), cod (Campana et al. 1989, Sundby et al. 1989)and anchovy (Peterman et al. 1988). Houde (1987) demonstrated for five fish speciesthat growth rate variation and its effect on stage durations during the first year oflife can be a major factor affecting survival. His study was based on the premisethat mortality was stage dependent and transition between life stages occurred atfixed sizes. In particular, he demonstrated prolonged duration of the larval stagedue to slow growth resulted in low survival during the first year of life. In addition,Chambers and Leggett (1987) and Chambers et al. (1988) have demonstrated thatdevelopment rate from larva to juvenile is adaptive, occurring earlier and at smallersizes for fish having higher growth rates. Development rate is defined as the inverseof the time spent in a particular life history stage. Together, these observationsindicate that maximizing development rate, which is directly dependent on growthrate, during the larval stage is a key condition for producing strong year-classes.To summarize, the growth—mortality hypothesis leads to several predictions: cohort survival is directly related to growth rate during one or more life history stagesin the first year of life; fish will maximize growth rate, versus maximizing growthefficiency; growth rate is food limited in nature, at least when survival is lower; development rate is dependent on growth rate, occurring earlier and at smaller sizes,2when growth rate is faster; higher temperatures should result in higher growth rates.Many components of this hypothesis are supported in theory and have been demonstrated in laboratory studies. However, it remains to be demonstrated that growthrate, and the processes affecting growth, are significant determinants of survival innature. Indeed, the question of food limited growth in larval fish is still unresolved.The primary purpose of this study was to examine the relative importance offood availability and temperature on the feeding, condition and growth rate of redfish(Sebastes spp.) larvae on Flemish Cap and, in turn, to relate growth rate to survival.Ultimately, I am interested in the processes which control survival during specificlife history stages during the first year of life in fishes. I do not examine predationmortality in this thesis. I assume the major cause of mortality results from predationwhich is stage-dependent and decreases with each life history stage, as outlinedabove. Therefore, shorter stage duration, as a function of growth and developmentrates, should be related to higher survival where growth and development rates area function of prey availability and temperature.In this study I focus on the feeding ecology of larval and pelagic juvenile redfishand on the relative effects of prey availability and temperature on their growth ratesbetween years and among locations within years. Specific questions asked were:1. Do redfish select prey based on size or type?2. Is there a significant change in diet at metamorphosis from larva to juvenile?3. Do age and size at metamorphosis vary due to differences in growth rate?4. How does growth rate vary dependent on food availability and temperature?5. Is growth rate related to survival?3The remainder of this thesis is divided into 4 sections. The first (Chapter 2)describes the key components of the zooplankton food community in terms of howthey varied among years and what were the possible ramifications on prey availabilityfor redfish. Copepods are known to be the primary prey of redfish in the Irminger Sea(Bainbridge and McKay 1968) and have been reported to dominate the zooplanktonon Flemish Cap (Konstantinov et al. 1985). The next section (Chapter 3) examinesthe feeding ecology of redfish as larvae and pelagic juveniles. Differences in dietand prey selectivity among seasons and years are examined and related to redfishcondition and growth. In Chapter 4, I examine the effects of prey availability andtemperature on growth rate of redfish. Initially I examine the relationship of growthrate to prey concentrations and temperatures for different years, 1980 and 1981.Then I examine spatial differences of food availability and temperature on redfishgrowth rate, size and condition within different cruises. In the final section (Chapter5) differences observed in growth rates between years are related to estimates ofredfish survival. Survival estimated in this study during the first 3 months of lifeis also related to survival at 1, 2 and 5 years of age as an indication of the stagethat determined year-class strength. In addition, the role of physical oceanographiccirculation affecting survival is discussed.1.2 Study AreaFlemish Cap (47°N, 45°W) is an offshore bank situated east of the Grand Bankof Newfoundland and separated from it by the 1000 m deep Flemish Pass (Figure1). The bank is approximately 200 km in diameter at the 400 m isobath and isbounded by two strong ocean currents: the relatively cold, fresh Labrador Current4flows across the northern side of Flemish Cap and down the western and easternsides, while the relatively warm, saline North Atlantic Current flows from west toeast across the southern side. These two currents serve as strong oceanographicboundaries for Flemish Cap and together they contribute to the formation of watersover the Cap (Hayes et al. 1977, Keeley 1982a). An analysis of historic T-S datafor the Flemish Cap area demonstrated the existence of 6 oceanographic areas, ofwhich 3 occurred over Flemish Cap within the 600—800 m isobaths (Keeley 1982b).Distinct was a central area of Flemish Cap bounded by the 200 m isobath. Thecirculation over this central area is described as a weak, anticyclonic gyre that canbe disrupted by storms (Hill et at 1975, Hayes et al. 1977, Ross 1980, 1981, Kudloet aT. 1984). Flemish Cap waters are isothermal in winter, and surface temperaturesincrease from seasonal lows of 3.9 °C (±0.7 sd) in March to peak values of 13.3 °C(±2.9 sd) in September (Drinkwater and Trites 1986).Flemish Cap traditionally supported commercial fisheries for Atlantic cod (Gadusmorhua) and Atlantic redfishes (Sebastes spp.), although at present redfish is dominant. Samples collected during this study demonstrated that redfish accounted for90% or more of all ichthyoplankton sampled. Both cod (Lear et al. 1981) and red-fish (Bainbridge and Cooper 1971, Templeman 1976, Ni 1982, Anderson 1984) areassumed to be discrete Flemish Cap stocks. There is a noteable absence of pelagicfishes on Flemish Cap. Capelin (Mallotus villosus) and sandlance (Ammodytes dubius), while abundant on the adjacent Grand Banks, are seldom observed on FlemishCap (Lilly 1987).Three species of redfish have been reported for Flemish Cap (Templeman 1976,Ni 1982). Sebastes memtella and S. marimus are spawned primarily during April5while peak spawning of S. fasciatus occurs during June (Barsukov and Zakharov1972, Templeman 1976, Penney 1987). Of these three species, S. mentella is themost numerous while S. marimus always is reported in low numbers (Templeman1976, Ni and McKone 1983, Penney et al. 1984). Because S. memtella is numerically dominant and its peak spawning occurs in April, I conclude that it was thepredominant species sampled in this study. In addition, comparative analyses oflarval morphology demonstrated that no single criterion successfully discriminatesamong these three species (Penney 1985, 1987). Also, it appears that S. mentellaand 5. rnarimus larvae are more similar to each other than either of them is to S.fasciatus (Penney 1987). For these reasons, any mixture of species that did occurin my samples should have a negligible effect on my interpretation of results.Several characteristics distinguish Flemish Cap as a good area for research intothe dynamics of larval fish survival: it is a distinct oceanographic area, both biologically and physically; it is a relatively simple marine ecosystem dominated inspring by a single herbivore and one fish species; it is a relatively small area whichfacilitates sampling. These factors led fisheries scientists to propose Flemish Cap asan important area in which to carry out research into factors causing survival variation during the egg and larval stages of demersal fish species, and an internationalprogram was mounted to sample the system during 1979-1983 (reviewed by Lilly1987). Data collected as part of my thesis derived from the Canadian participationin this international program of research on Flemish Cap.62 Invertebrate Zooplankton Community2.1 IntroductionAn underlying theme in recruitment research is that interannual changes in foodavailability during larval fish stages are an important determinant of variation inlarval survival and ultimately recruitment to adult populations. Several hypotheseshave been proposed relating food availability to larval fish survival under varyingconditions (Anderson 1988). Variation in food availability has been attributed tovariable timing of the spring bloom in relation to fish spawning, varying food concentration at the pycnocine in relation to storm mixing, and changes in the magnitudeof spring production. Annual dynamics of the classical marine food chain, consistingof relatively large phytoplankton and herbivores, is still hypothesized to be criticalin determining new fish production in temperate stratified and upwelling systems(Cushing 1989, 1990). Recently, the role and response of copepods in marine temperate systems has been proposed as a more direct link between varying environmentalevents affecting survival of larval fish and eventual recruitment (Runge 1988).Previous studies of invertebrate zooplankton on Flemish Cap have been carriedout in different months spanning the period from February to September and datingfrom as early as 1958 to as recent as 1981 (Vladimirskaya 1967, Konstantinov et al.1985). Calamus finmarchicus was the dominant copepod during the spring in the14 years covered by their work. Samples of nauplii in 1958-61 (Vladimirskaya 1967)and 1970-78 and 1981 (Konstantinov et al. 1985) indicated that spring spawning ofC. finmarchicus was just beginning during late March, early April (Vladimirskaya1967). However, these studies have been limited by the use of a coarse mesh net (70.5 mm), irregular spatial sampling, and no comparison of interannual variability.The purpose of this chapter is to describe the invertebrate zooplankton community on Flemish Cap (47° N, 45° W), including the seasonal timing of copepodspawning and development and interannual differences observed over the 7 years inwhich collections were made, 1977 to 1983. Redfish larval extrusion peaks duringApril, and these larvae comprise 90%, or more, of all ichthyoplankton from Marchto August (Anderson 1984). Redfish feed predominantly on eggs and nauplii stagesof calanoid copepods (Bainbridge and MacKay 1968) and, therefore, are highly dependent on the spring production of copepods. It is in such a system that copepodsare expected to have a direct effect on larval fish survival (Runge 1988).2.2 Materials and Methods2.2.1 Field CollectionsZooplankton were collected during fourteen plankton surveys carried out from 1977-1983 on Flemish Cap. Only the years 1979-81 are adequate to describe the seasonalchanges in zooplankton on Flemish Cap and only in 1980 and 1981 were multiplemesh nets used. Sampling was done using a survey grid of 20 nautical miles (37km) station spacing, ranging from 20 to 56 stations per survey, each survey lasting3-7 days. During 1979-1983 sampling dates (Table 1) ranged from the beginning ofMarch to the first week of August; only one survey per year was conducted in 1977(October), 1978 (July) and 1983 (March).Plankton samples were collected at each station using a 61 cm bongo samplerwith two 0.333 mm mesh nets. In two cruises 0.505 mm mesh nets were used on the61 cm bongo sampler, although this was limited to one side during the cruise 25-308October 1977 and 31 stations during the cruise 2-9 May 1981. During 7 cruises a20 cm bongo sampler was used simultaneously with the 61 cm bongo sampler, using0.165 mm and 0.253 mm mesh nets. Finally, a 50 cm ring net using an 0.080 mmmesh net was used during 4 cruises. A summary of the number of stations sampledand samples processed for species identifications and plankton biomass is given inTable 1.The bongos were towed obliquely from near the bottom or 200 m depth, whicheverwas less, at ?-‘1.25-1.5 m (2.5-3.0 knots) following standard techniques of Smithand Richardson (1977). Payout and retrieval rates were 0.83 and 0.33 m srn’, respectively. Each net was fitted with a GO flow meter and maximum depth wasmonitored using a pressure sensor fitted above the bongo frame.- During two surveys, 20-26 March and 10-14 July 1979, oblique tows were done to 125 m depthonly. Ring net samples were towed vertically from 100 m depth. All samples werepreserved in 5% buffered formalin.Tow data from the 14 cruises were extensively analyzed to remove identifiablesources of error. Identification of questionable values of water volume was based onregressions of water volume versus tow time and tow depth versus water volume foreach cruise, compared to the the theoretical calculations of Webster and Anderson(1988). Based on a simulation model, Webster and Anderson (1988) calculated thewater volume versus tow time relationship as: Y = 20.91X, where Y is water volume(m3) and X tow time (mi). Similarly, they estimated the tow depth versus watervolume relationship as: Y = 0.335X, where Y is tow depth and X water volume. Ineach case outliers were corrected, if possible, based on verification of incorrect datarecording or coding. Exceptionally low estimates of water volume generally were9not a problem, indicating clogging of the bongo sampler did not occur as a rule. Ifquestionable, the observations were removed from the data set and water volumewas estimated based on the regression equation for that cruise. In the case of twocruises, 20-29 July 1980 and 3-14 March 1983, the tow data indicated water volumeshigher than expected, in relation to tow times and tow depths. In these cases watervolumes and tow depths were estimated from the above relationships.2.2.2 Laboratory ProcessingPlankton volume was measured using methods similar to those outlined by Smithand Richardson (1977). Large fish and gelatinous zooplankton (> 1 cm3) wereremoved from the sample prior to measurement, usually by use of a coarse meshnet. The remaining sample was then drained using Nitex mesh of the same, orsmaller, size as the original sample until drainage diminished to the occasional drop.Plankton volume, to the nearest ml, was then measured by displacement of waterin a graduated cylinder.Zooplankton were identified to the lowest taxonomic level possible. All large animals (eg. chaetognaths, ctenophores, cnidarians, amphipods, decapods, mysis, larvaceans, euphausiids, fish, ostracods, pteropods, polychaetes) were initially removedand idenitified to species, if possible, and counted. After removal and processingof the large organisms, the remaining animals were subsampled, identified, stagedand counted. Large mesh samples (0.505 mm and 0.333 mm mesh, 61 cm bongo)were split using either Folsom or Motoda plankton splitters, or the beaker techniqueof Van Guelpen et al. (1982). Samples were split into successively smaller aliquotsuntil the smallest aliquot (no smaller than 1/512) contained .—‘lOO individuals of10the dominant taxon. This and other abundant taxa were completely enumerated inthe smallest aliquot, whereas less numerous taxa were counted from progressivelylarger subsamples. In the case of samples from 0.165 mm and 0.080 mm mesh nets,the total sample volume was made up to a known amount and 1.0 ml subsampleswere withdrawn using a Stempel pipette. For all samples at least 300 individualswere identified and counted, where i100 individuals of the most common taxonwere identified. Where one species completely dominated the sample, counting wascontinued to ensure that 30-50 individuals of the next most abundant taxon wereidentified and counted. Taxa were identified as completely as possible using availableliterature, and abundant copepod species were staged where possible.- Afl zooplankfthi analyses wee carried out unde contràt Through the Department of Fisheries and Oceans, St. John’s, Nfld. with different personnel performingthe work in some years. However, in each case the methods were the same. Qualitycontrol measures were built into the subsampJing, counting and identification procedures. For each component of work, which conformed to each mesh size for eachcruise, 10% of all samples were randomly selected for re-analysis. In addition, theselection of samples was stratified across the total time spent sorting each year suchthat samples were re-analyzed from the beginning, middle and end of the total sorting period. For all but rare species, subsampling and counting error was expectedto be less than 10%. When error exceeded this amount, samples processed to thatpoint were completely re-analyzed. However, throughout the entire project this situation occurred rarely and was usually attributed to one sorter at the beginning ofthe sorting period. Overall, re-analysis of samples indicated error associated withthe laboratory processing of samples was < 5% for all but rare species. Taxonomic11verifications were done throughout the different sorting and identification contractsby known experts from the National Oceanographic Identification Center, NationalMuseum, Ottawa, the Atlantic Reference Centre, Huntsman Marine Science Centre,St. Andrew’s, N. B. and by Dr. G. Gardner, Memorial University of Newfoundland,St. John’s, NF.All data were entered on coding forms, transformed into machine readable codeand extensively edited. Species and taxonomic classification were coded according tothe method of Foy and Anderson (1986). All data base management and statisticalanalyses were done using the Statistical Analysis System (SAS 1985). Abundance(individuals m2) was calculated asCDNllr2Land density (individuals m3) was calculated asCN llr2Lwhere C is the number of plankton collected, D is the maximum sampled depth (m),L is the length of the tow path (m), and r is the radius of the net opening (m).During most cruises zooplankton were collected using more than one gear type.Ideally, these data would be combined into a single data set to quantitatively estimate invertebrate zooplankton across all size classes 0.080 mm to 0.333 mm.However, this was not practical and data from each sampler has been analyzed independently. Biomass and plankton densities from the vertically hauled cone net(50 cm, 0.080 mm mesh) were considerably greater than from the bongos. This wasprobably due to the fact the vertical tows were done to 100 m, compared to 200m, and the occurrence of zooplantkon predominantly in the upper part of the water12column. Comparison between the small (20 cm, 0.165 mm mesh) and large (61cm, 0.333 mm and 0.505 mm meshes) bongo samplers is much more direct as theywere towed simultaneously. For Calanus fimmarchicus copepodite stages CI—CVI, Ihave compared catch rates to evaluate which stages were representatively sampledby each sampler. These results are summarized in Appendix A.2.2.3 Statistical AnalysesStatistical analyses for abundance differences were done using parametric techniqueson square root transformed data whenever possible. The assumption of homogeneityof variances was tested for t-test comparisons using the F-ratio test and for multiplemean comparisons using either Bartlett’s chi-square test or Levene’s test (Brown andForsythe 1974). When the assumption of homogeneity was rejected, non-parametrictechniques were used. For comparison between two means Wilcoxon’s test wasused (Sokal and Rohif 1969). For comparison of more than two means, tests fordifferences were done using a standard unbalanced ANOVA on ranked scores ofthe original data (SAS 1985). The ANOVA on ranked scores is considered to bea more robust analysis than the Kruskal-Wallis test with the advantage that theranked data can be analyzed using a posteriori techniques routinely available forparametric analyses. In all cases Duncan’s multiple range test was used testing fordifferences at P < 0.05. For comparisons of relative stage frequency differences theKolmogorov-Smirnov test was used (Siegel 1956) by adapting this test within a SASprogram that made use of standard SAS procedures.132.2.4 Estimates of Copepod DevelopmentPredictions of stage duration and peak spawning dates of Calanus finmarchicus werebased on the temperature dependent Bèlehr.dek equations of Corkett et al. (1986)which are of the general formD =a(T+awhere D is development time (days), T is temperature (°C), and a, a and b arefitted parameters. Following Corkett et at (1986), I assumed that the a parameteris stage dependent where CI = 6419, CII = 8014, CIII = 9816, CIV = 11601, CV= 13526 and CVI = 17477. For all equations a = 10.60 and b = —2.05.Backcalculations of stage development times to peak spawning dates of C. fin-marchicus using the relation of Corkett et al. (1986) for 1980 and 1981 were based onobservations during the last week of May each year. In 1980 copepodite stage CIIIdominated while in 1981 stage CIV dominated. Development times for stages CIIIand CIV were based on average April/May temperatures. For stage CII copepoditespredictions were based on a weighted average of April and May temperatures, whereApril temperatures were weighted by a factor of 2. Finally, estimates for stage CIwere based on average April temperatures.It is assumed that early stages of marine copepods, and specifically C. finmarchicus, occur in the upper water column (Krause and Trahms 1982, Williams and Conway 1988, Williams et al. 1987). As they develop copepods are known to undergodiel vertical migrations, in which case they would be subject to changing diurnaltemperature regimes. Temperature was estimated as the average temperature at 10m depth for central Flemish Cap waters lying within the 200 in contour. This esti14mate will minimize cold or warm water influences due to meanders of the Labradoror North Atlantic currents that may occur in the broader areas of Flemish Cap.To estimate temperature at 10 m depth all Nansen bottle and bathythermographtemperature data available from the Marine Environmental Data Service, Ottawa(MEDS) were selected and averaged monthly. The area chosen to be 200 m depthwas bounded by 46°40’-47°20’ N and 44°30’-45°20’ W.2.3 Results2.3.1 Zooplankton BiomassPlankton volume is a measure of total biomass of the plankton community on Flemish Cap. A composite plot of data collected during 12 cruises suggests that peakbiomass of the mesoplankton (0.333 mm mesh nets) occurred around the end of Mayto the beginning of June (Figure 2a). Overall, values ranged from 1.92 ml 10m3 inlate March, prior to spring spawning of copepods, to a peak value of 6.22 ml 10rn3during late May 1981. Values generally decreased by late July and early August.Plankton biomass measured from 0.165 mm mesh samples from 6 cruises alsoindicated a peak in biomass during late May (Figure 2b). The only July observationwas in 1982 and plankton biomass was relatively high compared to observationsearlier in the year for 1980 and 1981. This was similar to the high 1982 Julyplankton biomass as measured by the larger 0.333 mm mesh samples.Plankton biomass data for 0.080 mm mesh samples were only available for 2cruises in 1980 and 2 cruises in 1981 (Figure 2c). Plankton biomass decreased fromearly to late May in 1981, in contrast to plankton volume measured by coarsermeshed nets. This decrease probably represented the declining biomass of copepod15eggs and nauplii stages following peak spawning, as sampled by the 0.080 mm meshnets.In summary, peak plankton biomass occurred during late May for the zooplanktonsampled by 0.165 mm and 0.333 mm mesh nets. For fine mesh samples peak planktonbiomass probably occurred earlier, around the end of April to beginning of May.2.3.2 Zooplankton CommunityThe invertebrate zooplankton community on Flemish Cap, based on 0.333 mm mesh61 cm bongo samples, was dominated by calanoid copepods of which Calanus finma’rchicus was the dominant species (Figure 3). Data collected at different timesduring 1977-1982 are plotted here as a composite figure, representing general conditions observed during different seasons on Flemish Cap. In constructing this plot Ihave used data from all cruises carried out in different seasons April to November,assuming that interannual variations in zooplankton community development areless than that observed across these eight months. In two cases cruises occurred onthe same dates: end of May 1980—1981, and beginning of August 1981—82. In bothcases I have used the 1981 data as this was the year in which the most sampling wasdone (4 cruises) and should represent the chronology of succession most accurately.Interannual comparisons are discussed in a later section.During late March total copepods accounted for 53% of the invertebrate zooplankton on Flemish Cap. The percent occurrence increased rapidly to 91% byearly August and declined to 72% by the end of October. The dominant species wasCalanus fimmarchicus which ranged from 37% to 69% of all zooplankton sampled,with this percent steadily increasing from the March to August period after which it16declined to 62% in October. There was some variation in relative abundance amongyears (Table 2). For example, C. finmarchicus represented 67% of the total zoo-plankton 1-4 August 1981 and 83% 1-3 August 1982. Other medium sized copepods(2.0 to 5.5 mm TL) that occurred abundantly were Metridia lucens, M. louga andC. glacialis. Of the large calanoid copepods (5.0 to 10 mm TL) only C. hyperboreusand Euchaeta norvegica were observed.Non-copepod species groups represented in the collections were summarized into12 species groups and their relative contributions to the plankton are plotted inFigure 3. These species groups included: ostracods, gastropods, echinoderms, larvacea, euphausiids, amphipods, chaetognaths, ctenophores, cnidaria, polychaetes,adiolarias,and “others”. In each ease one or two species dominated within each -group. The ostracods were dominated by Conchoecia elegans and C. obtusata,gastropods by Limacina sp. larvaceans by Oikopleura sp. with both 0. vanhoeffemi and 0. labradoriensis being observed, and euphausiids by Meganictyphanesno’rvegica. Other species groups occurring in low densities included echinoderms,hyperiid amphipods, chaetognaths (Sagitta sp. ), ctenophores (Pleurobrachia sp. ),jellyfish (Cnidaria sp. ) and polychaetes. The species groups summarized here as‘other’ ranged from 9% to 30% abundance of all species observed (Figure 3).Samples of smaller zooplankton from finer mesh samples (0.165 mm mesh, 20cm bongos) again indicated that copepods numerically dominated the zooplanktoncommunity, ranging from 55% to 82% during the period April through to July(Figure 4). This figure was constructed similar to Figure 3, except that the onlydates where more than one cruise was available were at the end of May 1980—81. Asabove, I have used the 1981 cruise data in this composite plot. The non-copepod17species groups used were the same as for the 0.333 mm mesh samples. The dominantspecies from these small mesh samples was the cyclopoid copepod Oithoma sp.dominated by 0. similis and 0. atlantica. Together they comprised from 20% to42% of all species sampled by the finer mesh nets, and 37% to 51% of all copepods.By contrast, C. fimmarchicus numerically represented only 15% to 43% of all speciessampled. Similar to samples from the coarser mesh nets (0.333 mm) there was anincrease in the percent composition of copepods during the period sampled fromearly April to early August.Comparison of densities for C. fimmarchicus and Oithona sp. collected using 0.165mm mesh nets demonstrated that Oithona sp. was more numerous in 5 of 7 cruises(Table 2). Weights were estimated by totalling weights for each copepodite stage foreach species for each cruise, standardized for the number of stations in each cruise.Copepodite weights were based on values published by Tremblay (1981). In earlyApril 1980, densities were equal for the Flemish Cap area and during 1-3 August1982 densities of Oithoma sp. were lower than C. finmarchicus. At other times, representing the period from May-July, Oithona sp. was 1.1—2.5 times more abundantthan C. finmarchicus. When comparing biomass, however, C. finmarchicus dominated in all 7 cruises ranging from 8—137 times more in total weight (Table 2). Thelargest differences in total biomass occurred during early April and early Augustwhen abundances of C. fimrnarchicus were equal to or greater than Oithoma sp. andthe community was dominated by copepodite stages CVI and CV, respectively.Of the many species of small copepods (0.5 to 1.5 mm TL) sampled from the finermeshed nets, 0. similis dominated in all but one cruise (n=7). The second mostabundant species was 0. atlantica, although Microcalanus pygmaeus was very similar18in abundance. These species were followed by Pseudocalanus sp. , Scolecithricellaminor and Oncaea borealis, in descending order of abundance.A second notable difference in the finer mesh samples was the abundance offoraniiniferans and radiolarians, of which Globigerina sp. was the dominant speciesidentified. Together this group (Sarcodina) accounted for 10% to 43% of zooplanktonenumerated.In summary, the invertebrate zooplankton of Flemish Cap was dominated bycalanoid copepods. Numerically C. finma’rchicus dominated in coarser mesh sampleswhereas Oithona sp. was of equal or greater abundance in the finer mesh samples. Interms of biomass C. finmarchicus was the dominant zooplankton species on Flemish- Cap. Finer mesh samples indicated that the foram Globigerina sp. was the mostnumerically dominant non-copepod observed on the Cap.2.3.3 Timing of Annual Copepod Spawning on Flemish CapCopepod spawning on Flemish Cap is examined first based on egg and naupliidata available from 0.080 mm mesh samples collected during 1980 and 1981. Forcomparison of spatial differences these data were divided into one of three depthstrata: 200 m, 201-400 m and > 400 m depth. Comparison of densities among thestrata demonstrated that during 7-14 April 1980 and 2-9 May 1981 egg densities weregreatest in waters 200 m depth (Table 3). These data indicate spawning occurredthroughout the Flemish Cap area but was most concentrated in the shallowest watersoverlying Flemish Cap. This is supported by the nauplii data from the 0.080 mmmesh samples where density again was highest in waters 200 m depth during thetwo earliest cruises (Table 3). During 7-14 April 1980 and 2-9 May 1981 the nauplii19densities, however, were lowest in the 201-400 m depth strata and higher again indeeper waters. These observations suggest spawning may have occurred both overcentral Flemish Cap and in deep oceanic waters surrounding the Cap. Comparisonof mean egg and nauplii densities among strata for these cruises demonstrated theobserved density differences among strata were not statistically different.By the end of May in 1980 egg densities were highest in the 20 1-400 m depthstratum and lowest in deep oceanic waters surrounding Flemish Cap (Table 3). Inlate May 1981 egg densities were still highest in waters 200 m depth. Naupliidensities were similar in all depth strata during late May 1980 and lowest in 201-400 m depth waters in late May 1981. Comparison of mean densities among stratademonstrated no significant differences during 21-31 May 1980. However, during22-27 May 1981 eggs were significantly different among strata, being highest overcentral Flemish Cap 200 m depth (P=0.0650); also nauplii were significantly moreabundant in waters > 400 m depth (P=0.0439) for this period.The decrease in egg densities for waters within the 200 m isobath from early Aprilto late May demonstrates that peak calanoid spawning occurs sometime during midto late April (Table 3). Nauplii data support this observation. Densities were lowduring the first week of April 1980, averaging only 183 nauplii m3 in waters 200m depth. By the first week of May 1981 densities increased to 1226 nauplii m3.Together these data support the observation that copepod spawning had recentlystarted during early April and peaked before the beginning of May.These observations are for Crustacean eggs collectively and represent a compositeof seasonal changes from data collected in two different years. Nevertheless, thegeneral pattern of copepod spring spawning, and in particular, for the dominant20copepod C. fimmarchicus, is well represented. The data suggest that spawning beganfirst and most intensively in shallow waters 200 m depth overlying central FlemishCap during the early part of April. By early May egg densities had decreasedand nauplii densities increased indicating peak spawning had passed. By late Maycopepod spawning continued to remain high in shallow waters but may have spreadto deeper waters in the 20 1-400 m depth strata. In deep oceanic waters surroundingthe Cap spawning also occurred and it was often higher than that in the 201-400 mdepth strata. While egg densities were high in deep waters (> 400 m) they neverexceeded those over central Flemish Cap ( 200 m depth).The progression of copepod spawning and spring production is examined herefui’ther based on copepodite stages of C. fimmarchicus as sampled by 0.165 mm meshnets and occurring 400 m depth on Flemish Cap. During the cruise 7-14 April1980 only stage CVI C. finmarchicus copepodites were observed in any significantabundance, supporting the observation that spawning was probably underway atthat time (Day 100, Figure 5). Chronologically twenty-five days later, during 2-9May 1981, C. fimmarchicus copepodites were dominated by stages CII and CIII, withpeak densities of 128 m3 and 161 m3, respectively. By late May 1981 (Day 150)stage CIV was clearly dominant at densities of 172 m3 although stage CV were nowappearing in greater numbers with densities averaging 83 m3. By late June 1981stage CV was now dominant with density averaging 96 m3. Stage CV continued todominate in subsequent cruises through July, averaging 270 m3 during 10-14 July1979 and 362 m3 1-3 August 1982.Comparison of C. finmarchicus copepodite densities among the three depth stratasupport the observations from the egg and nauplii data that spawning was most21concentrated in the shallow waters overlying Flemish Cap. During 7-14 April 1980stage CVI dominated in all three depth strata and was most abundant in waters200 m depth, averaging 99 m3 compared to 54 m3 in the 201-400 m stratum and 42m3 in waters > 400 m depth (Table 4). Twenty-five days later during 2-9 May 1981stages CIII and CIV dominated in waters 200 m depth, while stages CII and CIIIdominated within the 20 1-400 m depth stratum and stage CIII dominated over deepwaters > 400 m depth. This indicates copepodite development was more advancedover the shallow, central waters and lagged by approximately one development stagewithin the other strata. The more rapid development observed for 200 m depthcould be the result of both an earlier start of spawning and warmer temperaturesfound in these waters.Surveys later in the year during late May 1980 and 1981, late June 1981 andJuly 1979 and August 1982 (using 0.165 mm mesh samples) demonstrated thatdominance of one particular stage occurred throughout all 3 depth strata. Therefore,differences in relative stage abundances among depth strata observed earlier in theyear had disappeared by the end of May indicating no differences in the progressionof development now existed among strata. However, in all cases copepodite densitiesof the dominant stage were highest in waters 400 m depth indicating higherconcentrations occurred in waters over Flemish Cap.The progression of spring secondary production for C. fimmarchicus can be summarized as spawning being underway by early April and peaking sometime duringmid to late April in shallow waters 200 m depth overlying Flemish Cap. Overalldensities of copepod eggs and nauphi and C. finmarchicus copepodites were higherover Flemish Cap than in deeper oceanic waters surrounding the Cap. Copepod22development proceeded through April and May with stage CV C. finmarchicus appearing towards the end of May and dominating from June through to August,indicating diapause had set in. Finally, copepodite density was always greatest inwaters 400 m depth. Based on the single stage dominance observed throughoutthe period late May to August, it appears there is a single spring spawning peak ofC. fimmarchicus on Flemish Cap.2.3.4 Interannual DifferencesComparison of differences in copepod seasonal development and abundances amongyears must be done with caution due to possible differences among years in thetiming of spawning, the temperature dependent rate of copepod development andstage dependent mortalities.The data indicate there were significant differences in plankton biomass amongyears at all times for which direct comparisons could be made. Plankton volumemeasured from 0.333 mm mesh samples in 1980 was always significantly lower thanin 1981 (P < 0.05). This was true in late May and July and, by extrapolation, forthe June period as well (Figure 2a). Plankton biomass for both 0.080 mm and 0.165mm mesh samples was also significantly different during late 1\’Iay 1980 and 1981(Figures 2b and 2c; P < 0.05). In addition, observations 23 April - 9 May 1979were significantly lower than 2-9 May 1981 for waters 400 m depth demonstratingplankton biomass in 1981 also was relatively high compared to 1979 and 1980 (P <0.00001).Mean plankton biomass data from 0.333 mm mesh samples for observations in4 years from mid-July to the first week of August ranged widely, from 0.116 ml23m3 in 1978 to 0.532 ml m3 in 1982 (Figure 2a). Comparison of means amongyears indicated an increase in biomass across years from 1978 to 1981. It shouldbe noted that only 4 observations were available for July 1978 and there are noplankton volume data from 1979. Comparison of differences among years duringthe July-August period for waters 400 m depth demonstrated 1-3 August 1982was significantly higher (P < 0.0001) than 1-4 August 1981 and 20-29 July 1980,which were not significantly different from each other. For waters > 400 m depth1982 was still the highest year and there was no difference between 1981 and 1980(P=0.0275).The only direct comparison of copepod eggs and nauplii from 0.080 mm meshsamples was late May 1980 and 1981. Comparison of densities indicated eggs weremore abundant in late May 1980 than in 1981, whereas nauplii were more abundantin late May 1981 compared to 1980. Egg densities in waters 400 m depth averaged945 eggs m3 in 1980 compared to 287 eggs m3 in 1981, higher by a factor of 3.3.Conversely, nauplii densities in waters 400 m depth were 2.2 times higher in 1981compared to 1980. For waters over Flemish Cap 400 m depth egg:nauplii ratioswere 1.81:1 in 1980 compared to 0.25:1 in 1981. For comparison between years,divided by depth strata, eggs were significantly more abundant in waters 400 mdepth in 1980 (P=0.0154) whereas there was no significant difference in waters > 400m depth. Nauplii were more abundant during 1981 in all depth strata (P <0.00001).These observations suggest that copepod spawning and seasonal development wasmore advanced by the end of May in 1981 compared to 1980.Comparison of differences in C. fimmarchicus copepodites among years indicatesthere were significant differences in both the relative rate of copepodite development24and absolute abundances. Direct observations can be made among years for 0.165mm and 0.333 mm mesh samples at various times (Table 5).Comparing relative stage densities of C. finmarchicus copepodites from 23 April- 9 May 1979 to 2-9 May 1981 for 0.333 mm mesh samples indicates developmentwas more advanced in 1981. This observation is based on the absence of stages CIand CII in 1981, and the dominance of stage CIV in 1981 compared to dominanceof CVI in 1979 (Table 4). Comparison of stage frequencies demonstrated differenceswere significant (Kolmogorov-Smirnov test, P < 0.01). However, this comparisonmust be qualified by the fact that stages CT-CuT are undersampled by 0.505 mmmesh nets (Appendix A) and that although the mean cruise dates only differed by5 days the cruise in 1979 started on 23 April, 10 days earlier than in 1981. - -Copepodite stage densities from 0.165 mm mesh samples, in which copepoditestages CI—CV are quantitatively sampled (Appendix A), demonstrated that development in late May 1980 significantly lagged that of 1981 by one stage (KolmogorovSmirnov test, P < 0.01). This was supported by the 0.333 mm mesh data. Inboth cases stage CIII was dominant in 1980 and stage CIV in 1981 at the time ofsampling in late May. Results comparing late May 1980 and 1981 were supportedby other calanoid species for which stage abundance data are available. Calanushyperboreus stage abundance was the same as C. fimmarchicus, being dominated bystage CIII in 1980 and CIV in 1981. Stage abundance data for Microcalanus sp.also indicated development was more advanced in 1981. In 1981 stage CVI clearlydominated whereas in 1980 stages CIII to CV were relatively much more abundant,compared to CVI.By the mid-July to early August period stage CV dominated in all years 1979-251982. Analysis of differences in percent composition of copepodite stages demonstrated significant differences occurred. Kolmogorov-Smirnov tests of significancebetween years demonstrated. fimmarchicus stage frequencies were different in0.165 mm mesh samples for 1979 compared to 1982 (P < 0.01). Stage densitiesindicated stage CV dominated both years but that in 1979 stage CIV was relativelyless abundant and stage CVI was more abundant than was the case in 1982. Thisresult suggests copepodite stage development was more advanced in 1979 comparedto 1982. This could occur due to an earlier spawning time and/or a more rapid rateof development but was not due to time of collection as samples were collected 22days earlier in 1979.Further comparisons during summer based on 0.333 mm mesh data demonstratedno significant differences occurred between 1980 and 1982, but both years weresignificantly different from 1981 (P < 0.01). Stage densities indicated, again, thatstage CV dominated in all years but that stage CIV was relatively less abundantin 1981 compared to 1980 and 1982. Therefore, development of C. fimmarchicusappeared to be more advanced in 1981 than in the other two years. Given thatthe date of sampling was identical for 1981 and 1982 this difference in copepoditestages would result from an earlier spawning time and/or more rapid developmentrate. Samples in 1980 were collected approximately 7 days earlier than in 1981and, therefore, the earlier sampling date may have contributed to the significantdifference between these years.A direct comparison could not be made between 1979 and 1980 or 1981 data asmesh sizes were not the same. However, by inference to results comparing 1979 to1982 for 0.165 mm mesh data, I conclude that 1979 was significantly more advanced26than 1980. I base this on the fact 1979 was more advanced than 1982 (comparing0.165 mm mesh samples) and that there was no difference in relative stage densitiesin 1980 versus 1982 (comparing 0.333 mm mesh samples). In addition, comparing1979 to 1981, CIV:CV catch ratios were 0.113 in 1979 (both 0.165 and 0.253 mmmesh data) and 0.026 in 1981 (0.333 mm mesh data). Therefore, copepodite stageCIV was relatively more abundant in 1979 indicating that development lagged behind that of 1981. Given these comparisons it appears that copepodite developmentwas most advanced in 1981 followed by 1979 followed in turn by 1980 and 1982,which were not different from each other.These observations of relative stage differences in summer (July-August) are supported by comparisons earlier In the seasoit For 1980 and 1981 during late Maycopepodite development was more advanced in 1981 which was still the case in July-August. Comparing 1979 and 1981 at the beginning of May development was againmore advanced in 1981. Therefore, 1981 copepod development was more advancedthan 1979 and 1980 early in the season and this was still the case for summer observations. The observations are distributed unevenly in time but overall the datasupport the inference that copepodite development was different among years andthat these differences occurred first early in the season.Comparison of 0.333 mm mesh samples from 23 April— 9 May 1979 to 0.505 mmmesh samples from 2-9 May 1981 demonstrated abundance of all copepod specieswas greater in 1981, especially so in waters 400 m depth (Table 6). Significantdifferences among years occurred for C. finmarchicus and other copepods for alldepth strata (Table 7). During the same time in late May 1980 and 1981 meancopepodite density for C. finmarchicus and other copepods was significantly higher27throughout the Flemish Cap area in 1981 (Table 8). For example, in waters overFlemish Cap 400 m depth C. fimmarchicus was higher in 1981 by a factor of 4.3for 0.165 mm mesh data (Table 6). Statistical comparisons among years for differentdepth strata demonstrated C. finmarchicus was not significantly different in waters200 m depth, but was significantly higher in the 201-400 m, > 400 m and 400m depth strata (Table 8). lVIean densities of copepods other than C. finmarchicuswere 2.3-6.6 times greater in 1981 compared to 1980 for 0.165 mm mesh data andwere significanity different for all meshes, all depth strata (Tables 6 and 8).Assuming copepodite development was more advanced in 1981 it is possible tocorrect for stage specific mortality to allow a more direct comparison of abundancedifferences. As stage CIII was dominant in 1980, correcting for mortality associatedwith development from stage CIII to CIV allows a direct comparison to stage CIVdensity in 1981, the dominant stage at that time. Average copepodite mortality istaken to be 0.01 d1 (Davis 1984) and average development time for CIII copepoditesis approximately 6.6 days at 4.8 °C (see below). Observed density of stage CIIIcopepodites in May 1980 was 34.8 m3 which would decrease to 32.5 m3 aftercorrection for mortality. This compares to an observed density of 200.3 m3 forMay 1981 stage CIV copepodites, higher by a factor of 6.2. Therefore, this directcomparison indicates Calanus abundance was significantly higher in 1981. However,the delayed seasonal development in 1980 may mean that copepodites had not fullyrecruited to stage CIII; the bulk of the spring production were still nauplii. Whilecomparison of nauplii densities from 0.080 mm mesh samples indicted nauphi weremore abundant in 1981, > 94% of these nauplii were Cyclopoida. Without a detailedknowledge of the spring spawning curve of Calamus finmarchicus and the rate of28development each year it is difficult to compare densities.Comparison of densities later in the season, following the spring bloom, may bemore valid. Mean copepod densities among 4 years during the period from 10-14July to 1-4 August again demonstrated significant differences occurred. Comparisonof 0.165 mm mesh samples for the two highest years, 10-14 July 1979 and 1-3 August1982, demonstrated there were no differences among years for either C. finmarchicusor other copepods. However, comparison of 0.333 mm mesh data from 1-3 August1982, 20-29 July 1980 and 1-4 August 1981 demonstrated significant differences didoccur among years in all strata for C. finmarchicus and for waters in the 20 1-400m and < 400 m depth strata for copepods other than C. fimma’rchicus (Table 9).For C. finmarchicus 1980 and 1981 aensities were nOt different and were lower than1982 for all strata except <200 m, based on Duncan’s multiple range test on rankedscores (Table 9). For copepods other than C. finmarchicus density was lower in 1980than both 1981 and 1982, which were not significantly different from each other, for20 1-400 m and 400 m strata. Therefore, density of C. finmarchicus in 1980 and1981 were the lowest during the July to early August period while 1979 and 1982were highest. For copepods other than C. fimmarchicus 1980 was significantly lowerthan the other three years. These comparisons were made without correction formortality due to different sampling dates.2.3.5 Predictions of Peak Spawning of Calanus finmarchicusSurface water temperatures over Flemish Cap historically increase from winter lowsin February to maximum values in August and September (Anderson 1984, Drinkwater and Trites 1986). The period of maximum increase in temperature occurs during29June each year. During the three years considered here, 1979-1981, mean surfacewater temperatures were significantly different during the spring heating period.The coldest year was 1980 and the warmest years were 1979 and 1981 (Table 10).Temperatures in 1980 were not significantly different than the longterm mean temperatures while during April and June 1979, and May and June 1981 temperatureswere significantly warmer than the longterm mean (Anderson 1984).Predictions of the date of peak spawning for C. finmarchicus in 1980 and 1981were backcaiculated, using temperature dependent Bêlehridek equations, from observations at the same time during late May each year. By late May au regions ofFlemish Cap were dominated by a single copepodite stage. In addition, this period is well past the date of1eàk spawning. Finally, stage CIII dominated in 180,whereas stage CIV dominated in 1981. Backcalculations predict that peak spawningoccurred 36.5 days earlier in 1980 and 38.2 days earlier in 1981. These times correspond to peak spawning on 16 April 1980 and 15 April 1981 (Figure 6). Therefore,in spite of a difference in copepodite stage dominance in late May 1980 and 1981the dates of peak spawning are predicted to differ by only one day. By comparison,the estimated time to develop from CIII to CIV was 6.7 days in 1980 and 5.9 daysin 1981 (Table 11). Therefore, while conditions indicated seasonal development ofC. fimmarchicus was more advanced in 1981 by late May most of the observed difference can be explained due to differences in temperature dependent developmentas opposed to an earlier spawning time.Comparison of C. finmarchicus development to that predicted from longtermmean temperatures for Flemish Cap will give some insight into conditions on theCap during the three years surveyed (1979-81). Development of C. fimmarchicus30based on longterm temperatures for surface waters overlying Flemish Cap 200m depth were scaled relative to an estimated date of peak spawning on Day 107(15 April). Development patterns during 1979-81 were compared to that predictedfrom longterm temperatures (Figure 6). In this case 1979 predictions were alsoscaled to Day 107 as there were no post-peak spawning observations for that year.It is apparent from these comparisons that 1980 development of C. fimmarchicus isestimated to have been slower than that predicted from the longterm mean, whereasboth 1979 and 1981 were faster. In 1980 the slower development occurred duringnauplii development in April, and by stage CI development lagged the longtermprediction of copepod development by 2.3 days. For stages CII-CV this lag thenremained constant. In 1981 development through all nauphi stages to CI equalledthat predicted for the longterm mean. However, development from CI to CV duringMay began to increase compared to the longterm mean and by stage CV preceededit by 3.8 days. In 1979 it is estimated that development was faster than the longtermmean for both nauplii and copepodite stages. By stage CI it was 4 days faster andby stage CV it was 5.4 days faster.2.4 DiscussionThe invertebrate zooplankton community on Flemish Cap is typical of the northwest Atlantic ocean in terms of the dominance of Calanus finmarchicus. This speciesnumerically dominated the larger mesh samples (0.333 mm), was the second mostabundant species in small mesh samples (0.165 mm) and was always dominant interms of biomass. The numerical dominance of C. finmarchicus has been reportedfor Davis Strait where it made up 35.5% of the total zooplankton for the period31May-October, ranging from 18-64% (Huntley et al. 1983). On the Scotian ShelfC. firimarchicus represented 60% of the total biomass on Emerald Bank during theperiod February-September (Tremblay and Roff 1983). Similarly, on Georges BankC. finmarchicus dominated the invertebrate zooplankton biomass, although Pseudocalanus sp. was the most abundant copepod, followed by C. finmarchicus (Davis1987). Investigations carried out over several years on Flemish Cap demonstratedthat copepods dominated the plankton from March to July constituting 88-98%of the total zooplankton by number (Konstantinov et al. 1985). C. fin.ma’rchicusnumerically ranged from 39-66% while Oithona similis ranked second ranging from28-57% of the total zooplankton. These results were for a 30-cm Juday net usingNo. 38 gauze (0.500 mm) and are similar to results reported here using the coarsermesh nets.The timing of spring Calanus spawning reported here was mid-April and peakbiomass of 0.165 mm and 0.333 mm mesh samples occurred during sampling in lateMay. These observations are supported by previous investigations on Flemish Cap(Vladimirskaya 1967, Konstantinov et al. 1985). Both the time of spring spawningand subsequent peak plankton biomass occurred approximately one month later onFlemish Cap than in shelf waters to the west and south. On Georges Bank spawningof Calanus finmarchicus appears to occur in March (Fig. 24.3 Davis 1987). Thisalso appears to be true of Browns Bank on the southern Scotian Shelf as McLarenand Corkett (1986) report spawning of C. finmarchicus had not occurred by lateFebruary whereas nauplii were abundant by mid-April. Similarly, Calanus spawningis reported to begin in late February or March on the southern Grand Bank andby late March nauplii constituted 80-90% of the population (Vladimirskaya 1967),32indicating mass spawning occurred during March. Samples taken 17-31 March 1960and 6 April - 2 May 1958 indicated that Calanus spawning had just started onFlemish Cap, was intensively underway in the Labrador Current (47° N) and wasmostly finished on the northern Grand Bank (47 ° N) with the plankton beingdominated by nauphi at that time (op. cit.).Peak biomass on Emerald Bank, Scotian Shelf occurred in May and declinedthrough June to minimum values during July-September (Tremblay and Roff 1983).This cycle was largely due to C. finmarchicus which made up 60% of the annual cope-pod biomass. On Georges Bank Davis (1987) reports that peak plankton biomassoccurred during July, coincident with peak abundance of C. finmarchicus, althoughhe states this peak may have occurred in May owing to the variability in the dataanalyzed. Due to the high abundance of C. fin.marchicus, its relatively large bodysize and the fact it has a large increase in body mass between successive stagescompared to smaller species such as 0. similis (McLaren and Corkett 1986), thisspecies of copepod totally dominates the annual biomass cycle on Flemish Cap andin other shelf areas of the northwest Atlantic.Therefore, Flemish Cap is characterized as a marine ecosystem dominated by C.fimmarchicus which spawns about one month later than on the Grand Banks, ScotianShelf and Georges Bank and peak plankton biomass probably occurs one month lateras well. The timing of C. finmarchicus spring spawning is known to be closely linkedto the initiation of the spring bloom (Krause and Trahms 1983) and the magnitudeof production dependent on the concentration of phytoplankton (Runge 1988). Thiswould indicate the spring bloom does not start until April on Flemish Cap, which issupported by chlorophyll observations made during March-May 1979-81 on Flemish33Cap (J. Anderson, unpubl. data).Comparison among years in this study reflect significant differences in the rateof development, but not necessarily in the time of peak spawning. These differencesin development of C. finmarchicus can be explained by differences in surface watertemperatures in different years. When food is not limiting then development ratewill vary directly due to water temperatures (Landry 1983, Corkett et al. 1986).Differences in water temperatures mostly will be a function of the seasonal heatingcycle due to increased solar radiation, but also may vary due to advective mixingof cold (Labrador Current) or warm (North Atlantic Current) waters onto FlemishCap. Analysis of formation of the mixed layer depth during 1979-1981 on FlemishCap indicated measureable differences among these 3 years on Flemish Cap wherethe MLD was, on average, much deeper early in the year in 1979 (March-April).However, this did not adversely affect the surface warming of water in 1979 astemperatures were always high (J. Anderson, J. Booth, and R. Keeley, unpub. data).Differences in C. finmarchicus abundance observed among years in this study,and of total zooplankton biomass, could result from differences in phytoplanktonconcentration during the time of spawning (Runge 1988). Density of C. finmarchicusin 1981 was 2.4x higher in early May than in 1979, and in late May was 4.3xhigher than in 1980 for waters 400 m depth. During the period from mid-Julyto the first week of August density varied by a factor of 4.5 in the four years 1979-82. Plankton volume measured during this same period in 1978, 1980-82 rangedfrom 0.205—0.533 ml rn3, a factor of 2.6x. Lower food (phytoplankton) would alsoreduce growth rate and extend the development rate of copepods (Vidal 1980, Davis1984), as was observed in 1980 compared to 1981. Chlorophyll and nutrient data34collected as part of our work on Flemish Cap is insufficient to address the questionof interannual differences in phytoplankton concentrations. The question of whetherherbivorous zooplankton are ever food limited during spring production in temperatewaters is unresolved. Davis (1984) demonstrated that development and survival ofPseudocalanus sp. and Paracalanus parvus were not food limited on Georges Bank.Development was found to be temperature dependent. Vidal (1980) demonstratedthat growth and development rate of Calanus pacificus and Pseudocalamus sp. werereduced at low food concentrations in laboratory experiments. He did not relatethese results to natural conditions. However, the effects of food concentrations werestage dependent, having the least effect on copepodite stage CIII (the youngestmeasured)and most on stage CVI (op. cit.). On the other hand, low temperaturesretarded the development of early stages proportionally more than that of the laterstages. These results suggest that the effect of food concentrations on growth anddevelopment of copepods in the earliest stages may be insignificant compared totemperature effects.As spawning of C. fimmarchicus is dependent on spring phytoplankton concentrations, an earlier start of the spring bloom should result in an earlier spawningof Calanus. Recently, Ellertson et al. (1989) reported the annual maximum occurrence of C. finmarchicus stage CI was inversely related to water temperatures,sampled at a single station 1960-1984 (n=16). Maximum occurrence of stage CIvaried by 48 days (1 April to 19 May) for mean surface (0-30 m) water temperaturesthat ranged from 0.7-3.5°C in March and from 1.9-4.4°C for April. They concludedthat these differences resulted from earlier spawning times and were not due to differences in development rates. Differences in temperature dependent development35times for their data are estimated to be 16.2 days for March, 11.3 days for April and13.7 days for average March-April temperatures; based on development to stage CIfrom Corkett et al. (1986). Therefore, temperature dependent development couldaccount for approximately 30% of the observed difference in timing. These resultssuggest that the onset of spawning may have a more significant effect on interannualdifferences in seasonal development of C. finmarchicus.The degree to which Calamus spawning was earlier on Flemish Cap cannot easily be determined. Simulation modeling of the onset of the spring bloom based onSverdrup’s critical depth model indicated spring production occurred 7 days earlier in 1980 than in 1981 (J. Anderson, J. Booth, and R. Keeley, unpub. data).By comparison, the red1ction for 1979 was 32 days later compared to 1980, and1979 was a year of significantly warmer temperatures throughout the spring season.Therefore, it is not obvious that the onset of the spring bloom is directly related totemperature. Analysis of chlorophyll and nutrient data collected in 1980 and 1981demonstrated that the seasonal phytoplankton cycle was more advanced by the endof May 1981, than in 1980 (J. Anderson, J. Booth, and R. Keeley, unpub. data).This conclusion was based on comparisons of the chlorophyll to nitrate ratio fordifferent cruises. The result is in agreement with the more advanced development ofcopepods but it cannot be demonstrated clearly whether the differences result fromfaster development versus an earlier start. Obviously, the relative importance of theonset of Calamus spawning versus temperature dependent development can only bedetermined by direct observation. However, I conclude that both mechanisms maycontribute significantly to interannual differences in seasonal development, especiallyif warmer temperatures can be directly related to an earlier onset of spawning.36Interannual differences observed in the timing and rate of copepod developmentand also the absolute magnitude of spring spawning could have a significant effecton larval fish growth and survival. It has been hypothesized that the timing of thespring bloom in temperate oceans will effect differences in larval fish survival andultimately fish recruitment (Cushing 1975, 1990). However, it has been difficultto demonstrate such a direct relationship. More recently, it has been hypothesizedthat copepods will act as a direct link between phytoplankton and fisheries variability in classic marine ecosystems dominated by larger copepods (Runge 1988).This hypothesis places an emphasis on differences in the timing and magnitude ofspring primary production acting through copepod seasonal dynamics which willhave a direct effect - on differences in larval fish survival among years. - Ellertsenet al. (1989) have demonstrated mismatch of C. finmarchicus nauplii (prey) withthe annual occurrence of cod larvae for a number of years. However, it remains tobe demonstrated that such significant differences in copepod development rate andproduction do have a significant effect on larval fish feeding, growth and survival.2.5 Summary1. Flemish Cap was dominated by the Calanoid copepod Calanus finmarchicus.It constituted 37% to 69% of all zooplankton sampled with coarse mesh nets(0.333 and 0.505 mm) and 15% to 33% in fine mesh nets (0.165 mm) duringthe March-August period. Owing to its relatively large size C. finmarchicusdominated the biomass of all species sampled in both coarse and fine meshsamples.2. The dominant copepod species in fine mesh samples (0.165 mm) were Oithona37similis and Oithona atlantica which together constituted from 20% to 42% ofall species sampled.3. Calanus finmarchicus formed a discrete population on Flemish Cap with peakspawning occurring in mid-April beginning first in waters 200 m depth. Peakbiomass occurred in late May and probably early June coincident with peakabundance of late stage copepodites.4. Significant differences were observed in the rate of C. fimmarchicus developmentand absolute abundance among years sampled on Flemish Cap. Differences indevelopment were assumed to be related to differences in the seasonal heatingof surface watersonFlemish Capin differei years. Higher temperatures mayact to initiate an earlier onset of spring production as well as directly increasethe rate of copepod development.383 Feeding of Redfish3.1 IntroductionSince Hjort (1914), fisheries research into the causes of recruitment fluctuation hascentered on the role of successful feeding by fish larvae. Based on laboratory estimates of prey densities necessary to maintain growth, compared to field measurements, it is hypothesized that prey concentrations in nature are limiting and thatstarvation mortality is high. May (1974) reviewed the available literature regardingthe “critical” phase of first feeding larvae and concluded there was no evidence thatsurvival at this stage affects year-class formation. Peterman and Bradford (1987)tested Lasker’s (1975, 1981) first feeding hypotheses for Northern anchovy, based on13 years of data, and found no relationship to year-class strength. However, Fortierand Leggett (1985) demonstrated increased mortality in capelin at the time of firstfeeding, consistent with Hjort’s (1914) hypothesis. Theilacker (1986) reported starvation mortality in jack mackerel in offshore populations of 70% compared to only12% for inshore fish. While starvation mortality has been reported during the larvalperiod, there is no evidence that it significantly affects recruitment (O’Connell 1980,McGurk 1985, 1989, Hewitt et al. 1985, Theilacker 1986, Fortier and Leggett 1985,Buckley and Lough 1987). Some studies have examined larval diet in relation togrowth and survival, but comprehensive estimates of prey availability were lacking(Cohen and Lough 1983, Ware and Lambert 1985). Recently estimates of prey selectivity by larval fish have been reported (Kane 1984, Peterson and Ausubel 1984,Monteleone and Peterson 1986), but these studies have been restricted in time andnot related to cohort growth or survival. More than a decade after the review of39May (1974) the role of feeding by larval fish, and how feeding variations among yearsaffects differences in growth and survival, is still unclear (Leggett 1986).The purpose of this chapter is to examine the feeding ecology of redfish (Se bastesspp.) on Flemish Cap during the larval and pelagic juvenile stages, using fielddata from research cruises carried out on Flemish Cap from 1979 to 1982. Thediet of redfish is described and estimates of selectivity determined for each stage.Differences in feeding between years are related to relative condition and size atmetamorphosis.3.2 Materials and Methods3.2.1 Laboratory ProcessingIchthyoplankton used for the feeding analyses were collected as part of a field program using bongo nets that were towed obliquely from 200 m, or from near thebottom when water depth was < 200 m. These samples were collected on a 24 hbasis such that the set of samples for each cruise were collected at all times of theday. A detailed summary of collection methods is contained in Anderson (1984).Redfish larvae were selected for feeding and morphometric analyses from preserved samples (3-5% Formalin) collected during 12 research cruises on FlemishCap (Table 12). When possible, redfish larvae were examined from each stationsampled within a cruise. In each case a maximum of 3 fish were selected for eachmm size class, choosing fish that were in good condition to ensure the integrity ofall measurements made. In 1979 often more than 3 redfish per mm size group wereexamined, averaging 2.3-3.7 fish per mm size group, compared to 1.0-1.7 for otheryears (Table 12). A comparison of mean lengths (±95%CI) between larvae subsam40pled for morphometric and gut analyses to the original sample was done to verifythat the subsample was representative of the population of larval redfish originallysampled. In all but one cruise the differences in means ranged from 0.0 - 0.6 mmand were not different based on a comparison of the 95%CI. For the cruise labelledGADO2O, the 95%CI’s did not overlap. The mean length of larvae subsampled was7.2 ±0.46 mm compared to 6.3 ±0.11 mm for the original sample. While comparisons based on the 95%CI are not statistically robust the measure is an indicationof possible differences among means. There was no systematic bias in the size ofredfish subsampled, with the possible exception of GADO2O where slightly largerfish were examined, on average.Prior to examination of stomach contents, each fish was re-measured for standardlength and morphometric measurements were made of body height, head width, headheight, and maxillary length (Figure 7). Standard length was measured from thetip of the snout to the end of the notochord for larval fish and to the posterioredge of the hypural plates for metamorphosed fish (juveniles). Head width wasmeasured dorsally or ventrally immediately posterior to the eyes and head height wasmeasured sagitally at the same point as head width. Body height was the maximumbody depth measured sagitafly at, or immediately posterior to, the insertion ofthe pectoral fin. Maxillary length was measured from the tip of the snout to theposterior end of the maxilla. Measurements of standard length were made to thenearest mm while all other measurements were made to the nearest 0.1 mm. Forselected cruises a measure of metamorphosis from the larval to the pelagic juvenilestage was made based on the degree of notochord flexion. Flexion of the notochordoccurs due to the formation of the hypural plates during ossification, and complete41formation of these plates is a good indicator of the completion of ossification andmethamorphosis to pelagic juveniles. It was relatively straightforward to distinguishbetween pre-fiexion and flexion larvae as the notochord transformed from straightto slightly bent. It was more difficult to distinguish the transition from flexion topost-flexion based on the angle of the notochord. Determination of the post-flexionstage was based on interpretation of the complete formation of the 4 distinguishablehypural plates, although this proved to be somewhat subjective.Stomach contents were examined from each fish by carefully slicing open thestomach using a dissecting needle. All stomach contents were identified to the mostaccurate taxonomic level possible. Parts of copepods and other food items werealso identified and recorded as fragments. Taxonomic classification followed Foy andAnderson (1986), while life history stage was assigned based on one of five categories:eggs, nauphi, copepodites, larvae and juvenile. Under each of these headings effortwas made to distinguish between cyclopoids and calanoids, NI-NVI, CI-CVI, andlarval forms such as gastropod veligers. In the case of copepod nauplii, classificationto one of six stages was often not possible. Therefore, nauplii were categorized intoone of 3 size classes: 0.2 mm, 0.21 - 0.40 mm, 0.41 mm, although this was onlydone beginning in 1981. Similarly, for small cyclopoid copepodites when specificstage identification was not possible they were classified into one of 3 size classes:0.4 mm, 0.41- 0.80 mm, 0.81 mm. In the case of the shelled gastropod (Limacinasp. ) the shell was measured along its longest axis.Estimates of stomach weights were based on published, measured or derivedweights (mg wet weight) for each species/stage level of classification. In many casesweight was determined from a calculation of size based on the assumption that 1 cc42= 1 g wet weight. Procedures used for determining these weights are summarizedin Appendix B.Finally, redfish dry weight was determined by drying each specimen at 55 °C for24 h on a tared pan and then storing the specimen in a desiccator for at least 24 hprior to weighing. Weights were measured to the nearest tg using a Perkin-Elmerelectrobalance (Model AD—2) making 2-3 replicate weighings to ensure a consistantvalue was obtained. Note that these measures of dry weight were done on specimensfrom which the stomach contents had been removed. Every effort was made toensure that pieces of the stomach wail that may have become detached during theremoval of the stomach contents were included in the weight determination.As a inaependent measure of redfish weight, fish volume was calculated for eachspecimen based on the morphometric measurements and assuming an idealized fishshape approximated by 2 triangles and a rectangle (Figure 7). This measurementis included here as an alternate measure of redfish size, both as a check on length(SL) and dry weight measurements and because dry weights were not determinedfor all cruises. The head was assumed to be triangular in shape from the tip ofthe snout to immediately posterior of the eyes. From immediately posterior of theeyes to the pectoral fin the fish was assumed to be a box. And from the pectoralfin to the tip of the notochord, or hypural plates in metamorphosed fish, the shapewas assumed to be a triangle. In this way fish volume was calculated based on thefollowing formula:FV = (O.33(HH . HW)(O.14SL) + (BH. HW)(O.11SL)) + O.33(BH. HW)(O.75SL)where FV = fish volume (mm3), 11W = head width (mm), RH = head height (mm),43BR = body height (mm) and SL = standard length (mm).3.2.2 Feeding AnalysesStomach contents were summarized into 15 prey categories based on their relativeimportance (% weight) to the diet of redfish during the different times sampled.These categories served to reduce the many prey types into a more manageable subset for subsequent analyses. The rationale for choosing these categories included consideration of prey identification limitations, variations in prey identifications amongtechnicians and with increased experience, uniqueness of the prey category and relative importance to the diet during one or more cruises. A detailed summary ofeach prey category follows:1. Copepod Eggs: These were predominantly Cala-nus finmarchicus, at least in1979 and 1980. Eggs were not staged so identification must be inferred fromthe predominant species present in the water.2. Copepod Nauplii: In 1979 and 1980 these were predominantly C. finmarchicusand, therefore, the category is equivalent to Calanoid Nauplii (Appendix B).The type (Calanoida, Cyclopoida) is inferred from the in. situ data. In 1981 fooditems were mostly classified as either Calanoida or Cyclopoida and, therefore,the distinction is explicit. Information on nauplii for 3 size classes is availablefor 1981 and 1982.3. Copepod Copepodites: Food items in this category are mostly calanoid cope-pods, dominated by C. finmarchicus. This is inferred as copepodite stages canbe readily distinguished as Calanoida or Cyclopoida and, as with eggs andnauplii, can be inferred from in situ data.444. Copepods Not Specified: This is a miscellaneous category that was never adominant food class but served to summarize all other Copepoda that couldnot be identified. It would include both Calanoida and Cyclopoida species for1979 and 1980 but probably only Calanoida in 1981.5. Cyclopoid Eggs: These are eggs, and possibly egg masses, that were identifiedmostly on the basis of size.6. Cyclopoid Nauplii: Here identification was based on both size and shape.Most items in this category are for 1981 and 1982 data.7. Cyclopoid Copepodites: Here identifications were based on size and shape.Items in this category apply to 1980, 1981 and 1982 data and at least July1979 data as well.8. Cylcopoid Not Specified : This was never an abundant food category.9. C. finmarchicus Copepodites: These were easily identifiable, stages CI to CVI.10. Oithona Copepodites: These were easily identifiable, stages CI to CVI. Mostly0. similis.11, Limacina sp. This small shelled gastropod (sea butterfly) was an importantfood item for small redfish (6-10 mm) only in two cruises: GADO19 and ATC331.12. Euphausiacea: Euphausids observed in the diet included eggs and nauphi(GADO2O) and furcilia larvae in both HAWOO2 and ATC331. They were notabundant but, when present, relatively large sizes contributed significantly tostomach contents.4513. Phytoplankton: These were mostly dinoflagellates and never occurred as anabundant item by weight (i. e. < 1% by weight). Numerically they only showedup in GADO37 (1% by number) and GADO51 (4.2% by number).14. Other: This category served to sum all other food items. The dominant fooditems that fell in this category are summarized in the results. Earlier in theseason (March - May) other food items were exclusively Protozoa of whichtintinnids and Globigerina were the most common. Later in the year a fewdifferent food items occurred but these always represented a small componentof the diet by weight.------15. Empty:--Empty stomachs would -have-neither yolk- sac-remains nor -food in -the stomach. The earliest time period sampled (GADO19) had an average of18.5% redflsh with empty stomachs. In all other cruises empty stomachs rangedbetween 2.8 - 8.3 % empty (based on counts). There was no apparent seasonalor yearly trend other than the high incidence of empty stomachs in GADO19.For comparison of redflsh diet among years, and analysis of prey preference,further partitioning of the data was done. Redfish in this study ranged in lengthfrom 4—26 mm in length and represented the period from release as first feedinglarvae at 5—7 mm through metamorphosis at 8—14mm to relatively large pelagicjuveniles in the 15—30 mm size range. Therefore, redflsh were divided into one of3 size classes. These size classes were chosen to represent larval redflsh ( 9 mm),redflsh undergoing metamorphosis (10-19 mm) and pelagic juvenile redflsh sampledmostly during the summer period ( 20 mm). Note that the 10-19 mm size classwill include not only metamorphosing redfish but pelagic juveniles as well. However,46redfish in the upper end of this size range were seldom abundant in our samples.In addition, redfish diet was summarized into a smaller number of prey categoriesfor these comparisons. These categories emphasized dominant feeding patterns observed during each cruise and served to clarify important feeding characteristicswhile eliminating categories that contributed little or nothing to the diet. The dominant prey categories were: copepod eggs (both calanoid and cyclopoid), calanoidnauplii, calanoid copepodites, cyclopoid nauplii and cyclopoid copepodites. Twoother prey items that occurred abundantly during some cruises were euphausiid fur-cilia larvae and the shelled gastropod Limacina sp. For completeness all other fooditems were included as “Other”.Feeding comparisons among years were possible for four periods ranging fromlate March to early August. The periods compared were:• 20-24 March 1979 — 6-13 April 1980.• 23-27 April 1979 — 2-9 May 1981.• 20-26 May 1980 — 22-27 May 1981.• 10-14 July 1979 — 22-28 July 1980 — 1-4 August 1981 — 1-3 August 1982.In the first comparison the cruise dates differed by approximately 2 weeks. However,both cruises occurred before the seasonal increase in redfish larval extrusion and themean and size range of larvae were similar. The second comparison occurred justafter the time of peak redfish release (Anderson 1984, Penney and Evans 1985), fortwo cruises that differed in time by only 1.5 weeks. However, the mean size of redfishwere 6.3 mm in 1979 compared to 8.3 mm in 1981 although the range of lengthswere similar: 4-12 mm and 6-13 mm, respectively. For this comparison redfish were47compared for two size classes: 9 mm and 10-19 mm. The third comparison wasfor nearly the same dates during the end of May 1980 and 1981, and redfish sizewas not different. The final comparison was among 4 years during the period frommid-July to the first week of August. These comparisons were made for all threesize classes of redfish. Statistical analyses for feeding differences among cruises werebased on the Kolmogorov-Smirnov test (Siegel 1956).As a measure of the amount of food redfish ate per unit body size, stomachweights of redfish were standardized by summing the contents of each fish, dividingby the length of the fish cubed and then computing standard statistics for eachcruise. I used length to standardize stomach content weight, since redfish dry weightmeasurements were not available for all cruises. This is a relative index of stomachweight to fish size measured as standard length, intended as a comparison among allcruises. For cruises which occurred later in the season mean standardized stomachweights were determined for both small (larvae 11 mm) and large (generallyjuveniles 12 mm) length groups.A direct comparison of standardized stomach weights between years was made for20-26 May 1980 versus 22-27 May 1981. In this case summed stomach weights weredivided by dry weight. This is a more appropriate method of standardizing stomachweights for statistical comparison because it results in a dimensionless value thatremoves known differences in dry weight which occurred between years. Statisticaldifferences in stomach content weights were determined for each mm length groupbased on Wilcoxon’s test (SAS 1985).Diurnal feeding patterns were determined within each cruise by summing thestomach weights of each redfish sampled and then averaging these weights for each482 hour period over the entire cruise. Sunset and sunrise were based on a standardtable used by the St. John’s weather station, adjusted to GMT.3.2.3 SelectivityEstimates of food preference by redfish were based on Chesson’s (1978) cr index.Of the many indices available Chesson’s cj has the advantage that selectivity isstandardized for the relative abundance of all prey types, allowing direct comparisonamong samples when abundance varies (Lechowicz 1982, Pearre 1982). This indexcan be thought of as representing the predator’s perception of the value of a preytype in relation to both its abundance and the abundance of all other prey typesavailable (Lechowicz 1982). When the primary interest about the feeding ecology is --to calculate the rank order of feeding preference- then the choice of selectivity indexis not considered to be critical (opt. cit. ). Selectivity (cj) for each prey type wascalculated asTi/P1—>=i rk/pkwhere r is the proportion of prey type i in the diet and p, the proportion in thewater.Redfish diet for calculation of r was calculated from stomach contents for allfish examined within each mm size class at each station sampled within a cruise.Calculation of pj for zooplankton prey in the environment was based on 0.080 mmmesh samples for 4 cruises (GADO35, GADO37, GADO5O, GADO51) and 0.165 mmmesh samples for 2 cruises (HAWPAN, ATC331). Sampling and processing methodsfor the zooplankton samples are outlined in Chapter 2. In both cases, diet andenvironment, samples were based on integrated tows from 0-200 m for bongo samples49and 0-100 m for the 0.080 mm ring net samples. However, larval redfish are knownto occur in the upper 50 m of the water column which typically will be above thepycnodine (K. Frank, Bedford Institute of Oceanography, Dartmouth, Nova Scotia,unpubi. data). In addition, it is known that early stages of copepods (Krauseand Trahms 1982, Williams and Conway 1988, Williams et al. 1987) and Limacinasp. (Perry and Neilson 1988) are concentrated in the upper mixed layer as well.Therefore it is assumed that the larval fish co-occurred with their prey, at leastduring the early part of the season from March through June. By July-August thedegree of co-occurrence is less well known. However, it is typical of fish in theirfirst year of life to begin diurnal migrations as they grow, which would put them inassociation with a greater variety of food items.Older copepod stages are known to undergo diurnal migrations, occurring in surface waters at night and migrating to subsurface waters by day. Therefore, estimatesof selectivity for copepodite stages, and possibly later nauplii stages as well, maybe influenced by the fact that these prey items do not co-occur with the redfish,or only do so for part of each 24 h period. Therefore, selectivity against such preyitems may be due, in part, to their unavailability due to vertical separation. Thedegree to which this is true remains speculative due to uncertainty about diurnalvertical migrations of copepods and to uncertainty about diurnal feeding patternsof redfish; these patterns may be timed to prey migrations. Pelagic juvenile redfishmost probably undergo diurnal migrations as well.Redfish larvae and plankton nets are expected to sample the environment differently in some cases. Plankton may not be eaten by redfish because: they are toobig for larvae to ingest but are sampled by plankton nets (set A), or they are not50at the same depths as the redfish and, therefore, are not available as prey (set B).Conversely, plantkon may not be sampled by nets but are eaten by redfish because:plankton are too small to be retained by the plankton nets (set C) or, particularlyfor larger redfish, plankton are too big and can avoid the plankton sampler (set 0).In the calculation of selectivity, to avoid the case where zooplankton were sampledbut not eaten (sets A & B) and where prey items were eaten but not sampled (setsC & 0) I only calculated feeding selectivity for the intersection set (set E) thatincluded plankton that were both eaten by redfish and sampled by 0.080 mm meshnets.RediSIc 1-I calculated r as the relative proportion of each prey type in the diet summedfor all redfish examined within each redfish size class (mm) at each station sampledwithin a cruise. Prey types were the 15 prey categories summarized previously. Icalculated p. based on the standardized zooplankton data (numbers m2) as sampledby the 0.080 mm and 0.165 mm mesh nets after summarizing all zooplankton intothe 15 prey categories. Finally, an overall estimate of prey preference was calculatedfor each cruise by averaging the as’s for each mm size class. Neutral selection was51calculated as 1/n, where n was the number of prey types which occurred in set Efor all redfish within each redfish size class for each cruise.3.2.4 Redfish ConditionThe relative condition of fish can be measured in a variety of ways. I based mymeasure of condition on the size of redfish relative to its length in each year (i. e.cruise). Size was measured as both dry weight (pg) and fish volume (mm3 convertedto equivalent units of gDW). Fulton’s Condition Factor ( K = W/L3x 10” ) assumesthat growth is isometric (i. e. b = 3). However, this was not true in all cases wherea comparison of K among years was possible. The Relative Condition Factor (ER = W/Lb x 10”, where b is estimated in each case from W = aLb) assumes thatthe slope- is equal among all samples being compared. This assumption also was notvalid for my comparisons among years. Therefore, neither K or KR were suitablemeasures of relative condition.Ricker (1973, 1975) proposed that Model II regression analysis be used to estimateweight-length relationships due to the covariance of both variables; and specificallythat the geometric mean functional regression (GMFR) be used. Cone (1989) argued that predictive (least-squares) regressions should be used as a measure of fishcondition, as opposed to the GMFR. Reasons for this include the ease of intrepretation due to familiar and readily testable outputs from regression analysis. Thechoice between using predictive regression analysis (Model I) or GMFR (Model II)depends on the question being asked. If the objective is to obtain the best functional (theoretical) relation between length and weight then the appropriate analysisis GMFR. However, if the objective is to predict weight from length then use the52predictive least-squares regression (Jensen 1986). My objective is to compare therelative size of redfish larvae among years, and at different times of the year, asopposed to deriving the best theoretical relationship of weight to length for redfish.For this reason I analyzed differences in relative condition using predictive regression(Model I, least-squares).All comparisons were based on log-log transformed data for redfish of similar sizeranges. In each case the residual variances were tested for homogeneity (Sokal andRohif 1969). When they were unequal the data were weighted by the reciprocal of theMean Square Error (MSE) prior to analysis of covariance. In this way the mean andslope remain unchanged and the weighted least-squares estimates are B.L.U.E. (bestlinear unbiased estimates) (SAS 1985). Regression slopes were tested for differences - -(H0: = b2) and if H0 was accepted then the intercepts were tested for differences(H0: a1 = a2) using analysis of covariance (SAS 1985).When analysis of covariance demonstrated that regression slopes were different(i. e. reject H0 : = b2) then differences in redfish size (dry weight, fish volume)were compared for each length (mm) between cruises in different years. Wilcoxon’snon-parametric test was used to test for differences in size for each length classwhen size-length comparisons rejected the null hypothesis of homogeneous residualvariances. All comparisons were based on normal data.3.3 Results3.3.1 Seasonal Patterns in DietObservations of redfish diet for the period April to August during 3 years, 1979-81,demonstrated there were significant differences in diet both among years and with53the changing biological season. The seasonal change in diet was characterized by anincreasing number of prey types during all 3 years. For example, during 20-24 March1979 and 6-13 April 1980 three food types made up 98% and 99.8% of the diet byweight, respectively (Table 13). By the end of May this had increased to 12 foodtypes making up 98.7% of the diet in 1980 (GADO37) and 9 food types making up98.7% of the diet in 1981 (GADO51). By July-August during 1979-1982 there were12-27 food types which made up 98.9-99% of the diet by weight. In addition, foodtypes tended to increase in size as the fish grew, with a shift from eggs and nauplii toCyclopoid and later Calanoid copepodites. However, the strong differences in dietobserved in 1981 compared to 1979 and 1980 make it difficult to generalize aboutthe diet of Flemish Cap redfish in a simple way.The main food items early in the season were copepod eggs and nauplii, for theperiod March to the end of May 1979 and 1980. Copepod nauplii were the dominantfood item by weight in all cruises during this period, except 6-13 April 1980 wheneggs were most important accounting for 85.8% of the diet, compared to naupliiwhich only accounted for 14% (Table 13). In 1981 comparable feeding observationsdemonstrated the dominant food item was cyclopoid copepodites followed by nauplii.A detailed examination of feeding differences among years is presented in the nextsection.The best comparison of seasonal feeding differences is made by comparing thediet of small ( 9 mm) redfish larvae, as these sizes were most always larvae (premetamorphosis) allowing a standard comparison for different seasons and differentyears. In 1979 and 1980 copepod eggs and nauplii (calanoid) made up 78.5% to99.8% of the diet by weight during the March-May period (Figure 8). In 198154copepod eggs and nauplii, which were primarily Cyclopoida, made up 56.1% to63.3% during this same period, whereas cyclopoid copepodites accounted for 34.6- 42.6% of the diet. However, by July and early August eggs and nauplii of bothcalanoid and cyclopoid copepods made up as little as 4.5% in 1979 and 30.7-43.1%in 1980-82 (Figure 8). In contrast, Oithona spp. copepodites made up 51.8% to81.6% of the diet during the July-August period. This strong seasonal change indiet from eggs and nauplii to cyclopoid copepoclites for redfish larvae 9 mm inlength reflects seasonal shifts in prey availability.Limacina sp. was a component of the diet of small redflsh in most cruises (Table13, Figure 8). However, only in two instances did it constitute a notable proportionof the diet. During 20-24 March 1979 and 1-3 August 1982 it made up 13.8% and17.5% of the diet by weight, respectively. It is noteworthy that these two cruisesrepresented the earliest and latest observations available indicating Limacina sp.may be relatively more important in the diet both prior to the onset of springcopepod production and possibly late in the summer period. In all other cruises itranged from 0-2% of the diet by weight. There was a notable seasonal trend in thesize of Limacina in the diet of redfish larvae, being initially large on average during20-24 March 1979 followed by a sharp decrease for the 6-13 April 1980 sample, afterwhich size more or less steadily increased to a maximum during 1-3 August 1982(Figure 9). While these data support a strong seasonal trend in the size of Limacina,they represent a composite of observations collected over 4 years, and the degree towhich interannual differences were present cannot readily be discerned.Two estimates of mouth width are available for Flemish Cap redfish. Head widthwas measured immediately posterior to the eye which was just behind the extension55of the maxilla. Therefore, head width can be regarded as a direct measure of maximum mouth width. Similarly, mouth width can be estimated from maxillary lengthbased on the empirical relationship of Shirota (1970). Comparison of the two measures of mouth width indicated they were not different: Y — —0.01 + 1.08X, (F =16882, P < 0.0001, R2 = 0.905, n = 1769), where Y is Mouth Gape estimated fromShirota (1970) and X is Head Width. The slope was not different from one andthe intercept was not different from zero. Therefore,I consider either measure ofmaximum mouth width is appropriate. For estimates of mouth width used hereI regressed head width versus standard length for all available data (8 cruises) toderive the linear relationship:HW = —0.724 ± 0.226SLwhere 11W is head width (mm) and SL is standard length (mm) (F=21,897, P<0.0001,R2=0.913, n=2098).I have compared maximum mouth width for redfish 6, 10, 15, 20 and 25 mmlength to C. fimmarchicus eggs, and C. finmarchicus and 0. similis nauplii andcopepodites (Table 14). Based on knowledge of the diet and preferred prey typesfor larvae of these different lengths at different times of the year the range of preferredprey sizes relative to maximum mouth width ranged from 2% to 73%. These valuesrepresent the extremes of 6 mm larvae feeding on Calamus finmarchicus NVI and 25mm juveniles feeding on Oithoma similis NI (Table 14). However, I conclude thatthe optimum size of prey was more likely in the range 5% to 30% for redfish larvae asthey grew through the first 3 months of life. I base this range on my interpretationof the results in Table 14, considering that small larvae would probably tend towards56smaller prey items not above 30% and larger redfish would tend towards larger preyitems not below 5%.Finally, comparison of head width to standard length indicates that mouth widthincreases at a greater rate. For example, at 6 mm length mouth width is 10% ofbody length while at 25 mm it is 20%. Therefore, it appears redfish have evolvedto increase mouth size relatively quickly during ontogeny, which is important tosuccessful feeding in a changing environment with an ever increasing range preyfield size and type.3.3.2 Among Year Differences in DietFeeding differences among years were apparent during 1979-1981 for the April-Mayperiod. By July most differences in diet had disappeared. The most notable difference was a shift in diet to cyclopoid nauplii and copepodlites during May 1981,compared to a predominance of calanoid eggs and nauplii in 1979 and 1980.Comparison of redfish diet for the earliest time period indicated significant differences (Table 15). During 20-24 March 1979 redfish ate mostly copepod nauplii(66.7% by weight) followed by copepod eggs (17.5%) and Limacina sp. (13.8%). Incontrast, during 6-13 April 1980 redfish ate predominantly copepod eggs (85.8%)followed by nauplii (14%) (Figure 10). These results suggest that Calanus finmarchicus spawning was much more advanced in 1979 by the end of March than it was inthe second week of April 1980.Comparison of redfish diets near the time of peak spring redfish extrusion in 1979and 1981 also showed significant differences (Table 15). Diet during 23-27 April 1979was dominated by copepod nauplii (47.8%), copepod eggs (30%) and euphausiid57nauplii (13.5%). In contrast, larval diet during 2-9 May 1981 was dominated byOithoma sp. copepodites (35.8%) and nauplii (32.3%). Calanoid nauplii only madeup 11.8% of the diet during this period in 1981. Comparison of diet between yearsfor redfish size classes 9 mm and 10-13 mm demonstrated diet was significantlydifferent in both cases (Table 15). In 1979 the diet of small (< 9 mm) redfish was notsubstantially different from medium (10-19 mm) redfish (Figure 8). In 1981 mediumsized redfish ate relatively more C. finmarchicus copepodites and less cyclopoidnauplii and copepodites, compared to small redfish which ate no C. finmarchicuscopepodites (Figure 8). However, both size classes of redfish in 1981 predominantlyate cylcopoid nauplii and copepodites. Therefore, the main difference in diet wassimilar for both size classes of redfish in 1979 versus 1981.During the end of May 1980 compared to 1981 redfish diet was again significantlydifferent, and the difference mirrored that observed at the beginning of May in 1979versus 1981 (Table 15). During 20-26 May 1980, redfish ate mostly copepod nailphi (calanoid) (72.1%) followed by eggs (12.2%) and some Oithona sp. copepodites(7.1%). In contrast, redfish during 22-27 May 1981 ate mostly Oithona copepodites(53%) followed by cyclopoid nauplii (39.4%) (Figure 10). Therefore, 1981 can becharacterized by redfish eating relatively older stages of cyclopoid copepods compared to 1979 and 1980.Comparison of feeding among years during July-August 1979-82 for three sizeclasses of redfish ( 9 mm, 10-19 mm, 20 mm) did not reveal an overall patternof differences among years for the 3 size classes (Table 16). In all 3 years smallredfish in July-August ate mainly Oithoma sp. copepodites, being 51.8—53.8% of thediet by weight. Similarly, medium sized redfish ate mainly Oithoma sp. copepodites:581980-72.3%, 1981—53.8%, 1982—89.3%. Only for redfish in the largest size class wasthere a notable difference among years. Diet of these large redfish was dominatedin all 3 years by copepodites of both Oithona sp. and C. fimmarchicus. However, in1980 Oithoua dominated by weight whereas in 1981 and 1982 C. finma’rchicus wasdominant (Figure 8). These differences were not significant (Table 16).The only significant difference observed in feeding was for small redfish sampled10-14 July 1979 compared to 22-28 July 1980 and 1-4 August 1981 (P , 0.05, Table16). In both cases the difference was due to small redfish in 1979 eating few copepodeggs and nauplii (< 5% by weight) compared to 1980 and 1981 when eggs and naupliiaccounted for > 40% of the diet. During 1-3 August 1982 eggs and nauplii accountedfor 30.7% of the diet. Statistical comparison of the diet of small redfish between10-14 July 1979 and 1-3 August 1982 showed no significant differences.3.3.3 Diurnal Patterns in FeedingDiurnal feeding of redfish is characterized by daylight feeding with peak stomachweights consistently observed before sunset (Figure 11). After sunset, stomachweights generally decreased until the following morning, suggesting that feeding hadlargely ceased and gut contents were being digested. Minima in stomach weightsoccurred around sunrise. However, there was some variation in this pattern amongcruises. For example, redfish stomach weights sampled during 20-26 May 1980reached a minimum 0.5 hours after sunrise and began to increase after this time(Figure 11). However, stomachs sampled during 26-30 June 1981 reached a minimum 0.5 hours after sunrise but did not begin to increase in weight until 4 hourslater (Figure 11). Finally, for the combined results of the four July-August cruises59the minimum occurred 2 hours after sunrise and stomach weights did not begin toincrease until 6 hours after sunrise (Figure 11). The remaining cruises show thedelay in the onset of feeding less clearly. Once feeding began there was a steadyaccumulation of stomach content weight through the remaining daylight hours.A delay in increasing gut contents following sunrise would partly result from alag in the accumulation of gut contents following the initiation of feeding. However,in some cases the delays were too long. In addition, there was a seasonal trend inthe delay which suggests a different explanation. Redfish larvae are known to occurthroughout the upper mixed layer depth but to be concentrated near the pycnocine(Kenchington 1991, K. Frank, Bedford Institute of Oceanography, Dartmouth, N.S.,unpublished data). The mixed layer depth has formed each year by the end of Mayand typically is 20-30 m deep (Drinkwater and Trites 1986). Therefore, any lagin the onset of feeding following sunrise may result from sufficient light for activefeeding reaching the 20-30 m depths where redfish larvae occur. The seasonal delayin the time of the onset of feeding from the end of May to the end of June and finallyfor July-August may be due to older redfish (pelagic juveniles) occurring deeper inthe water column; however, I do not have data to verify the depth distribution ofredfish at these older ages.The diurnal feeding pattern was less pronounced during four cruises early in theyear in 1979 and 1981 which sampled redfish larvae generally 11 mm, and in whichfeeding was generally poorer. This occurred 20-24 March and 23-27 April 1979 and2-9 May and 22-27 May 1981 (Figure 11). The weak definition of a diurnal feedingpattern for these 4 cruises seems to suggest that feeding might be suboptimal, suchthat normal diurnal feeding patterns were not established. In other words, it was60difficult for these redfish to fill their guts even with mid-day light conditions.3.3.4 Among Year Differences in FeedingStandardized stomach weights, as a measure of the amount of food redfish ate,varied both among years and for different periods within years. In general, theindex increased later in the season. This was most apparent for data collected in1981 where the index increased from a low of 0.085 to values > 0.3 during theperiod 2-9 May to 1-4 August 1981 (Table 17). A similar seasonal increase was alsoobserved in 1979 and 1980. Therefore, redfish consistently ate more food per unitbody length as the season progressed. One reason this would occur is the longerfeeding periods due to daylight as the season progressed. With increasing hours ofdaylight conditions improved for redfish feeding due to the increased time availablefor foraging. These results are in contradiction with the observed delay in onset ofdaily feeding as the season progressed (Section 3.3.4).The seasonal increase in the amount of food redfish ate per unit size would occurif pelagic juvenile redfish are better feeders than larvae. There are three lines ofevidence in support of this observation. First is the increase in the index between22-27 May and 26-30 June 1981, which can be directly related to the differenceobserved for larvae and pelagic juveniles, respectively. The second is the differenceobserved among four cruises which sampled both small (larvae) and large (juveniles)redfish. In each case the index of standardized stomach weights was considerablyless for larvae (range 0.055-0.242) than for pelagic juveniles (range 0.311-0.642)(Table 17). The final reason is based on the weight—standardized comparison ofstomach contents made between 20-26 May 1980 and 22-27 May 1981 (Figure 12).61All length groups from 7-11 mm had significantly greater standardized stomachweights in 1980 (Table 18). The mean difference tended to decrease with increasingfish length and by 12 mm the means were no longer statistically different (Table 18).At 12 mm in length redfish had predominantly metamorphosed to pelagic juveniles.Therefore, the difference in feeding which occurred between 1980 and 1981 wasconfined to the larval stage. Once redfish metamorphosed, in 1981, the difference instandardized stomach weights disappeared. Overall, standardized stomach weightsof juveniles were consistently high, whereas those for larvae were high only for 20-26 May 1980, when redfish larvae were growing relatively fast and were in bettercondition. All other cruises that sampled predominantly redfish larvae indicatedfeeding was relatively poor. These results support the conclusion that redfish becomebetter feeders after metamorphosis.3.3.5 Feeding SelectivityFeeding selectivity was calculated for 4 cruises where 0.080 mm mesh zooplanktondata were collected. The times examined were 6-13 April 1980, 2-9 May 1981, 20-26May 1980 and 22-27 May 1981. The overall result clearly demonstrated that larvalredfish, which ranged in length from 6-12 mm for these cruises, consistently selectedcopepod eggs and nauplii and did not select calanoid and cyclopoid copepodites.Calanoid copepodites always had a lower selectivity than cyclopoid copepodites.During 6-13 April 1980 copepod eggs were selected more strongly than nauphi. Inthe other 3 cruises copepod nauplii were selected more strongly than eggs. Forthe cruise 2-9 May 1981 selectivity results indicated that cyclopoid eggs were notselected for, however, this was not true for observations made 3 weeks later, 22-2762May 1981 nor the two cruises in 1980 (Figure 13).In each cruise Limacima sp. was neutrally selected for, on average. All other preytypes were classified as “Other”, and in each case they were collectively negativelyselected for by redfish (i. e. less than neutral). The Other prey types that occurred inthe diet for these 4 cruises were exclusively Protozoa (Table 19). The most commonfood types were tintinnids, Globige’rima sp. and radiolarians, although radiolarianswere only abundant as an Other food type in the 20-26 May 1980 cruise (73% ofOther prey types). Globige’riua spp. typically range in size from 0.270-0.800 mm(Newell and Newell 1977), and would be retained by both the 0.080 mm and 0.165mm nets. These observations for the dominant Other food types in the diet werealso true for all other cruises in which stomach data were examined for the periodMarch-June 1979-1981 (Table 19). Later in the season Other food types which alsooccurred in the diet included Oikopleura (ZAGOO4), Parathemisto (HAWPAN), fisheggs (ZAGOO4) and Isopoda (HAWOO2). Prey types classified here as Other werealways a small component of the overall redfish diet by weight (Table 13).There was considerable variation of selectivity values for all prey types (Figure13). This variation represents variable selectivity within each prey type as well asdifferences observed among stations within each cruise. Variability within each preytype could result from aggregating prey types into general categories. This wouldresult in large size ranges for nauplii (NI-NVI) and copepodites (CI-CVI), as well asinterspecies differences. If selectivity varied within each of these prey categories dueto either size or species differences, then this would result in variation in selectivityvalues. For these 4 cruises selectivity of each prey type was remarkably consistentfor each size class of fish, within the numerically predominant size range. Selectivity63was calculated for each cruise for all length classes but, in practical terms, relates tothe predominant length classes due to the small number of samples examined in thetails of each size distribution. The redfish size information is summarized in Table20.It was also possible to calculate prey selectivity for two cruises later in the season:26-30 June 1981 and 1-3 August 1982 (Table 21). For these two cruises zooplankton were collected using 0.165 mm mesh nets. Selectivities calculated from thesesamples are biased by the fact that 0.165 mm mesh nets will not quantitativelysample copepod eggs and nauphi (Davis 1980). However, these nets will retain atleast Pseudocalanus sp. stage CI (0.4 mmTL), and possibly smaller (Davis 1980).A statistical approximation gave a mean retention of 0.26 +0.08 mm, which corresponds closely with the size range of Oithona sp. CI which has a size range of 0.23± 0.07 mm (Murphy and Cohen 1978). Therefore, prey selectivity would be overestimated for copepod eggs and nauplii but not for copepodite stages which wouldbe quantitatively sampled by 0.165 mm mesh nets.Results from these two cruises demonstrated that redfish positively selected cyclopoid copepodites later in the season, while the fish were increasing in size (Figure14). These copepodites were primarily 0. similis and 0. atlantica. However, the fishstill did not select for calanoid copepodites, which were primarily C. finmarchicusstage CV at these times of the year (Chapter 2). During both cruises euphausiid furcilia larvae were neutrally selected (Figure 14). If all prey types eaten wereincluded in the calculation of neutral selection then euphausiid larvae would havebeen slightly selected for as a preferred prey type. However, furcilia larvae werean unimportant component of the diet accounting for < 0.7% by both weight and64number (Table 13).For the cruise 26-30 June 1981 the dominant prey types were copepodites, although copepod eggs and nauplii did account for 20.1% of the diet by weight. Laterin the season during 1-3 August 1982 only 2.7% of the diet was made up of copepodnauplii with the rest being cyclopoid and calanoid copepodites. Therefore, it appears there is a shift in prey selectivity to cyclopoid copepodites later in the seasonafter the redfish metamorphose to the pelagic juvenile stage.Redfish sizes during the June and August cruises fell into different size modes. Inboth cruises the majority of redfish occurred in the largest size range (Table 21). For26-30 June 1981 the redfish chosen for selectivity analysis were 9-15 mm in lengthand for 1-3 August 1982 they were 16-23 mm in length. Selectivity results for thesmaller size fish in each cruise differed from those for the larger fish. For 6-8 mm fishin June 1981 and 6-7 and 10-11 mm fish in August 1982 the redfish did not selectfor Oithoma copepodites. This is similar to the results for redfish analyzed for theApril—May period in 1980 and 1981 where the total size range was 6-12 mm (Figure13). Therefore, selectivity results for all redfish 12 mm for the period spanningearly April to early August were similar.3.3.6 Redflsh ConditionDry Weight versus Length:Regression of dry weight on length from 6 cruises in 1980 and 1981 gave theoverall relationship DW = 0.715 . SL3’59 (Table 22). This regression excludesthe July 1979 data which all fell substantially below the line. This exponent isgreater than 3.0 indicating growth of redfish was not isometric. In 1981 there was a65pattern of increasing exponent and R2 with time (Table 22). A plot of the residualsrevealed a definite shift to positive values for redfish lengths greater than —13 mm(Figure 15). Fish of this length were post-metamorphosis, indicating that once fishmetamorphosed to juveniles they increased in weight per unit length more rapidlythan larval fish. This explains the increasing exponents observed for 1981 data asmore pelagic juveniles occurred in successively later samples.On the basis of weight-length differences, redfish were in better condition in1980 than in 1981. This comparison was made between GADO37 and GADO51at the end of May each year. Least-squares regression demonstrated the rate ofincrease in weight as a function of length was much lower in 1980 than in 1981(H0 : = b2; F = 65.40, P < 0.0001). However, comparison of these regressionsover all length groups demonstrated that redfish < 14 mm weighed more in 1980than in 1981. Comparison of mean weight for each mm length group demonstratedredfish 6-11 mm in length weighed significantly more in 1980 than in 1981. Only at12 mm length was there no difference (Table 23). A plot of the percent differencein weight revealed that 1980 redfish 6-10 mm in length weighed 34.4—50.1% more,after which the difference in weight rapidly began to disappear (Figure 16). Redfish> 10 mm (9-10 mm) are undergoing, or have undergone, metamorphosis indicatingthat the difference in weight which occurred between these years only occurred forthe larval stage. The rapid recovery in weight for larger redfish (metamorphosed)in 1981 also explains the much larger regression coefficient of DW-SL, compared to1980 redfish.The only other direct comparison that can be made between years was 10-14 July1979 to 1-3 August 1981. However, only small redfish < 11 mm occurred in 197966which prevents comparison of DW-SL among years for redfish from the April-Mayrelease period during the summer period. Comparison of the intercept of DW-SLregressions for redfish 11 mm in length for these two cruises demonstrated redfishweighed significantly less in 1979 than in 1981 (P < 0.00001). The predictiveregression for 1981 redfish 11 mm was DW = 0.769 SL2999 (P < 0.0001, R286.5, n = 14). The residual variances were homogeneous and there was no significantdifference between slopes (F = 0.08, P < 0.7769).Due to the low number of observations in 1981 for redfish 11 mm in length Ialso compared redfish 11 mm from all other cruises to July 1979 data. The resultwas the same, demonstrating that July 1979 redfish also weighed significantly lessthanall other redfish sampled in this length range from earlier times in the season(P < 0.0000 1). The predictive regression for redfish 11 mm from all cruises exceptDAWO79 (6 cruises) was DW 1.15 . SL2929 (P < 0.0001,R2 = 72.5,n = 1631).The residual variances were homogeneous and there was no significant differencebetween slopes of the 6-cruise regression and DAWO79 (F = 0.31, P < 0.5758).Dry Weight versus Fish VolumeRedfish weight was approximated by converting Fish Volume (mm3) to units ofdry weight (/Lg). Fish weight estimated this way is only approximate as weight isderived from the relation 1 cm3 = 1 g wet weight and then converted to dry weightby the factor 0.1. However, it does represent a measure of allometric fish growththat is independent of dry weight measurements and is not strictly dependent onstandard length. Here I wanted to test that Fish Volume (FV, tg dry weight) wasequivalent to dry weight among years and among cruises within years. Specifically,I wanted to determine if there were any significant differences in either slopes (func67tional relationship) or intercepts (magnitude). Except for GADO35 the regressioncoefficients all ranged from 0.907 - 1.047, close to the expected value of 1.0 (Table24). Comparison of residual variances among these cruises indicated in each casethat the variances were heteroscedastic. Therefore, weighted least-squares regression was used to compare the DW-FV relationship for selected pairs of cruises. Thetests of slopes indicated there were no significant differences among slopes for thecomparisons when the size ranges compared were similar (Table 25). There werefour comparisons where intercepts were significantly different (P <0.05, Table 25).In the first case the GADO37 intercept was considerably larger than GADO51, indicating redfish weighed more (DW) per unit size (FV) in 1980 than in 1981. In theremaining comparisons, all for 1981, there was a seasonal trend of increasing redlishweight per fish volume: GADO5 1—HAWPAN-HAWOO2.The regression slope of DW-FV was much smaller for GADO35 than for all othercruises (Table 24). However, redfish in GADO35 only ranged from 5-8 mm in length,with most fish being confined to 6 and 7 mm. To test if the functional relationshipwas in fact different for GADO35, I compared slopes between GADO35 and GADO37,GADO5O and GADO51 for redfish 8 mm only. In each case the slopes were notstatistically different (F = 1.13,P < 0.2886;F = 1.89,P < 0.1706;F = 1.60,P <0.2071, respectively). However, all comparisons demonstrated the intercepts diddiffer significantly (F = 102.1,P < O.0001;F = 244.8,P < 0.00001;F = 48.6,P <0.0001, respectively). These comparisons demonstrated GA0035 redfish weighedless per unit volume than GADO37 (1980) but more than CADO5O and GADO51(1981) redfish.Therefore, the functional relationship between DW and FV was the same among68years and cruises when similar size ranges were compared, but there were significantdifferences in magnitude (weight per unit volume) in 1980 compared to 1981 andthere were some differences among cruises within a year.Fish Volume versus LengthFish Volume data were available from all cruises in 1980 and 1981 as well as theAugust cruise in 1982. The relationship between Fish Volume and redfish length(mmSL) was very similar to the dry weight versus length comparisons. Fish weight(FV) is expressed here as tg dry weight, based on the assumption that 1 cc = 1g wet weight and dry weight is 0.1 of wet weight. This conversion is simply donefor convenience so that comparisons of FV to dry weight (DW) will be in the sameunits. Overall, the best relationship was obtained by including all 8 cruises for whichthere were data (Table 26). A plot of the residuals revealed a strong positive skewat redfish lengths greater than —‘13 mm (Figure 17), similar to that for DW-SL.The regression coefficient was greater for the FV-SL regression than for the DWSL regression indicating that redfish increased at a greater rate in body size, asestimated by Fish Volume, than they do in biomass, as estimated by dry weight(F 320.7, P <0.00001).Statistical comparison of the regression slopes between GADO37 and GADO51demonstrated that redfish in May 1981 increased in size (FV) at a greater ratethan the 1980 redfish (F = 72.97,P < 0.0001). This was similar to the results forthe regression of DW on SL. However, comparison of mean FV for each mm lengthgroup demonstrated that redfish 8-12 mm in length were bigger in 1981 than in 1980.Only 7 mm redfish were significantly bigger in 1980 (Table 27). Percent differencein weight, calculated the same as for DW, demonstrated an increasing difference69in fish volume for redfish 7-12 mm length (Figure 16). There was a large increasein this difference at 11 and 12 mm length. Therefore, in 1981 redfish sampled atthe end of May were increasing in FV at a greater rate than in 1980. The FV-SLrelationship resulted from redfish that were bigger throughout the larval stage (6-10mm) in 1981 and also the metamorphosing fish.It was possible to compare Fish Volume versus length among the years 1980-82 for the July - August sampling period. Comparison of the regression slopesamong these 3 years for redfish 6-23 mm length indicated no statistical differences(Table 28). In all cases the residual variances were equal. Tests of the interceptsgave confusing results as the maximum (HAWOO2, 1981) and minimum (ATC331,i982) intercepts were not significantly different from each other whereas the middleintercept value (ZAGOO4, 1980) was significantly different from the other two (Table28). This result appears to arise from the much larger sample size in 1980.To further compare possible size differences among these 3 cruises I tested fordifferences in mean size (FV) between cruise pairs for different length ranges. Ineach case I combined 2 or more mm length groups in order to increase the numberof observations for comparison, especially as specimens did not occur in all mmsize classes in each cruise. These tests for differences indicated the only significantdifference occurred for redfish 20-21 mm in length between 1980 and 1982 (Table 29).A test of differences across all size groups among these 3 cruises indicated there wasno significant difference overall (Kruskal-Wallis test, x2 = 3.09, P < 0.2131, 2df).Therefore, it appears reasonable to conclude that the sizes of redfish during theJuly - August period 1980-82 were not different. The observation that differencesin larval size at length observed earlier in the season, as larvae, had disappeared by70August is consistent with the hypothesis that only the largest redfish survive eachyear.3.3.7 Metamorphosis from Larvae to Pelagic JuvenilesMetamorphosis from larvae to pelagic juveniles was based on the onset of notochordfiexion. This occurs due to the formation of the calcareous hypural plates whichform in the tail section, and is concurrent with the process of ossification in redfish.Measurement of the onset of fiexion is more straitforward as the notochord changesfrom strait to slightly bent. It is more difficult to determine the completion offlexion and, therefore, there tends to be more variance in these observations. Comparison in the size of metamorphosis was only made between 20-26 May 1980 and22-27 May 1981, as these cruises sampled redfish approximately during the periodof metamorphosis.The results clearly showed that redfish began flexion at a smaller size in 1980than in 1981. In 1980 redfish 8 mm length were 47.5% in flexion and at 9 mmthey were 77.3% in flexion (Figure 18). By comparison, in 1981 only 23.4% of the9 mm larvae had entered flexion, and 55.4% of the 11 mm larvae. (Figure 18).Therefore, not only did redfish begin flexion at a larger size in 1981 but it occurredover a larger size range compared to 1980. Comparison of the percent of larvaethat were either in flexion or had completed flexion as a function of length wasstatistically different (P<0.0001, Kolmogorov-Smirnov test,n1=277,n2=69240 for1980 and 1981, respectively). The large difference in sample sizes was due to thefact that all larvae were measured for flexion for the 1981 samples as part of thesample processing procedure, whereas a representative subsample was measured for71flexion for the 1980 data.Considering growth and metamorphosis of redfish size as a continuous processover all lengths, then the size at which the proportion of redfish in flexion exceeded50% differed by 2.72 mm between 1980 and 1981. This calculation was derivedfrom the data in Figure 18 by comparing the exact length at which 50% of theredfish occurred in the “flexion” stage each year. Based on growth rates for thesefish (Penney and Evans 1985), a difference of 2.72 mm equates to 17 days at 1980growth rates and 25 days at 1981 growth rates.3.4 Discussion3.4.1 Diet and FeedingPrevious studies on the feeding of redfish larvae and juveniles were limited to onlyone year of observation and were primarily descriptive in nature (Einarrson 1960,Bainbridge 1965, Bainbridge and McKay 1968, Marak 1974). Redfish larval diet,reported here for Flemish Cap, was similar to previous studies in the predominanceof copepod eggs and nauphi. Marak (1974) reported differences in feeding due toareas, time of day and season (July versus September). In each case he concludedthe feeding differences resulted from differences in food availability, which were notmeasured. There were no apparent differences in feeding due to depth of the water• column in which the fish were caught (op. cit. ).My observations on redlish, for sizes that ranged from 6-28 mm, demonstratedthat both the maximum prey size and the variety of prey types increased in the diet,as redfish increased in size, from March to August. These traits are characteristicof all fish larvae. Examples include: redfish (Bainbridge and McKay 1968, Marak721974), cod and haddock (Kane 1984), flounder, sole and dab (Last 1978), Atlanticherring (Cohen and Lough 1983), mackerel (Peterson and Ausubel 1984, Ware andLambert 1985), sandlance (Ryland 1964, Monteleone and Peterson 1986), northernanchovy, Pacific sardine and jack mackerel (Arthur 1976). Only in highly selectivefeeders such as plaice (Shelbourne 1962, Ryland 1964, Last 1978) and English sole(Gadomski and Boehiert 1984) that feed exclusively on Oikopleura sp. does thevariety of prey types not increase with increasing fish size, although the size ofOikopleura does.It is also characteristic of fish larvae that they consume prey much smaller thanthe maximum prey size they are capable of ingesting. I estimated optimum preysizes for redfish ranged - approximately from 5% to 30% of mouth size, and thatredfish mouth size increased at a faster rate than leng.th. Theilacker and Dorsey(1980) reported that fish larvae generally eat prey 13-38% of their maximum mouthwidth. Cohen and Lough (1983) reported that herring larvae fed on prey sizesapproximately 20% of the maximum (range 10-28%); and that this did not changeover a large size range of 5-35 mm through the period September to the followingFebruary. Frank and Leggett (1986) reported that capelin larvae grew faster andhad a higher survival when feeding on optimum prey sizes 0.040-0.051 mm equivalentspherical diameter. These sizes equate to 18-23% of the maximum mouth width ofcapelin larvae.A change in diet will result from changes in prey availability, when prey selectivityhas not changed. Seasonal differences in prey availability were best demonstratedfor redfish larvae ( 9 mm). The diet of these larvae differed seasonally while preyselectivities did not. Larval diet early in the season was dominated by copepod eggs73and nauphi whereas by July-August the diet was dominated by Oithoma spp. copepodites. Regardless of the season, larvae always preferred copepod eggs and naupliiwhile not selecting Oithoma spp. copepodites. The size ranges of these different preyare very similar. C. finmarchicus eggs are approximately 0.135 mm in diameter(Bainbridge and McKay 1968) while the nauplii range from 0.22-0.61 mm in length(Ogilvie 1956). By comparison, Oithoria similis nauplii range from 0.115-0.215 (Gibons and Ogilvie 1933) and copepodites from 0.230-0.515 mm total length (Conwayand Minton 1975).This seasonal shift in diet can be interpreted with respect to the seasonal production cycles of Calanus finmarchicus and Oithona similis, the two dominant species onFlemish Cap (Chapter 2). The seasonal dynamics of these two species are very different. C. fimma’rchicus is characterized as a large, herbivorous calanoid that spawnsin spring in direct proportion to the spring bloom of diatoms (Krause and Trahms1983, Runge 1988). While it may produce more than one generation each year, it isessentially univoltine in terms of its annual production (McLaren and Corkett 1986).It develops rapidly at rates directly dependent on water temperatures (Corkett etal. 1986), and stage development within the nauplii and copepodite stages resultsin relatively large increases in size and biomass at each stage (McLaren and Corkett1986). On Flemish Cap spawning began in March and peaked sometime in April.By the end of June the spring production of eggs and nauphi was complete, withcopepodites (CV) predominating (Chapter 2).In contrast, Oithona similis is a small, omnivorous cyclopoid. It spawns independently of the spring bloom and is multivoltine, producing a number of generationswith more or less continuous development for at least half the year (McLaren and74Corkett 1986). Its relatively small sizes for all stages result in small biomass increasesas each stage develops within each generation (op. cit.). While its development rateis temperature dependent, the increase in development rate is slow compared toCalanus finmarchicus, based on development rates given by McLaren (1978) for 0.similis and Corkett et al. (1986) for C. fimma’rchicus. For example, at 7°C it takesC. finma’rchicus 18 days to develop from an egg to stage CI, whereas it takes 0.similis 37 days.Interpreting the seasonal diet change of redfish larvae with respect to the zoo-plankton indicates that available prey in the preferred size range of these larvae shiftsfrom eggs and nauplii of C. finmarchicus to nauplii and copepodites of 0. similis.Because redfish larvae did not actively select Oithona copepodites, the seasonalreplacement of C. finmarchicus nauplii with Oithona copepodites may be disadvantageous to feeding, growth and, possibly, survival of redfish. This observation issupported, in part, by standardized stomach weights for small redfish (Table 16).Values during April and May were always high compared to values in June—Augustwhich ranged from high to low in different years.Differences in larval diets and, therefore prey availability, also occurred for 1981compared to 1980 and 1979. Redfish sampled in May 1981 predominantly ate cyclopoid nauplii and copepodites, compared to redfish in 1979 and 1980 that atecalanoid eggs and nauplii. As noted previously, the size ranges of these prey typesare approximately equal. The selection against Oithona spp. copepodites by redfish larvae, together with the significantly lower feeding rates and relative conditionof larvae in 1981, suggests that this food type was an inadequate replacement forcalanoid nauplii.75Differences observed between years in larval feeding rates and condition, do notdirectly relate to prey concentrations. Comparing observations at the end of May1980 and 1981, densities and total biomass of copepods were greater in 1981 (Chapter 2). For the surveyed area, Oithoria spp. and Calanus finmarchicus were approxi-.mately 3.7 and 2.4 times more abundant in 1981, respectively. While total copepodconcentrations were higher in 1981 compared to 1980, Oithona was relatively moreabundant than C. fimmarchicus. In addition, copepod nauplii were 4.2 times moreabundant in 1981, although 94% of these nauplii were classified as cyclopoids. Theseresults indicate that total prey concentrations were greater in 1981 and this was especially true for Oithona spp. This difference in prey availability was reflected inthe diet of redfish larvae that was dominated numerically by small cyclopoid nauphiand in biomass by Oithona copepodites in 1981. However, eventhough prey concentrations were higher in 1981, the amount of food eaten, as weight per unit fish,was lower; as was relative condition of the redfish larvae. Simply put, in 1981 highprey concentrations of a small type (number m3) resulted in many small prey beingeaten (prey items per unit fish) but, expressing this food as weight, the amountswere small (rig per unit fish) compated to 1980. Feeding on many small prey items(cyclopoid nauplii) versus a few large items (calanoid nauplii) is generally consideredto be disadvantageous to growth and survival (Pyke 1984). These results suggestthat feeding conditions in 1981 were poor, not because of lower prey concentrationsbut because these prey were, on average, one-tenth the size of the predominant preyeaten in 1980.Feeding on non-preferred prey types, in this case, Oithona copepodites, must bedisadvantageous to growth and survival as well; eventhough such prey fall within the76preferred size range. Other studies have demonstrated different selectivities basedon prey types within the same size ranges. Checkley (1982) demonstrated thatherring larvae raised in the laboratory on natural zooplankton food did not selectAcartia sp. The preferred prey of herring larvae 12 mm was Pseudocalanus sp.and it switched to Oithona sp. for larvae 13 mm length. Similar observations weremade for sandlance larvae sampled in the field where Acartia hudsonica nauplii andcopepodites were not selected by fish that ranged in size from 13.5—25.5 mm, onaverage (Monteleone and Peterson 1986). The preferred foods of these larvae werenauplii and copepodites of Temora iongicornis. Field studies of herring carried outover 3 years on Georges Bank demonstrated that the larvae did not eat Oithona spp.copepodites in spite of their local abundance. Instead they favoured Centropages sp.and Pseuclocalanus sp. copepodites, and grew faster in years in which Pseudocalanussp. predominated in their diet (Cohen and Lough 1983). Similarly, both cod andhaddock larvae selected against Oithona sp. for larvae 4-10 mm length (Kane 1984).The preferred prey of small larvae (< 6 mm) were copepod eggs and nauplii 0.25-0.35 mm in size. For larger larvae (6-10 mm) the preferred prey were Pseudocalanusminutus copepodites and copepod eggs. In each of these studies of prey selection, thecalanoid copepods were all small species. Females range in total length from 0.8-1.5mm, with cephalothorax lengths from 0.70-1.2 mm (Wilson 1932, Lawson and Grice1970, Conway and Minton 1975). These sizes are comparable to Oithona simiiis and0. atiantica which have female total lengths of 0.5-0.8 mm and cephalothorax lengths0.4-0.7 mm (Conway and Minton 1975, Gardner and Szabo 1982). Therefore, it isapparent that fish larvae select prey based on type, as well as size. This is also clearlythe case for plaice (Shelbourne 1962, Ryland 1964) and English sole (Gadomski and77Boehlert 1984) that only feed on Oikoplettra sp.Reasons why fish larvae have higher or lower selection for a particular prey typeare not known. In some cases it is hypothesized that fish larvae prefer smaller,relatively sessile prey that would be easily captured (Drenner et ai. 1978). Alternatively, it has been hypothesized that larvae prefer moving, relatively active preybecause they are easier to see (Peterson and Ausubel 1984). In studies reportingthe non-selection of Oithona spp. as a prey item it was felt that the copepodites ofthis cyclopoid species hold their antennae at right angles to their body, effectivelyincreasing the prey width thereby making them difficult to ingest (Cohen and Lough1983, Kane 1984). It has also been hypothesized that nutritional status of potential- prey may determine prey preference (Checkley 1982). -3.4.2 MetamorphosisThe diet of pelagic juvenile redfish was increasingly dominated by copepoditesof Oithona spp. and Calarius fimmarchicus as the season progressed and the fishgrew. Selectivity analysis demonstrated that juveniles actively selected Oithomacopepodites, in addition to copepod eggs and larvae. Chronologically, the switchin diet from Calanus eggs and nauphi to Oithona copepodites appears to beginsometime in June. This coincides with the metamorphosis of redfish from larvaeto pelagic juveniles, the end of the spring development cycle when C. finmarchicusreaches stage CV (i. e. cliapause), and the increased biomass of Oithoma relative toC. finmarchicus in the zooplankton community (Cli apter 2). Bainbridge and McKay(1968) reported a similar shift in the diet of redfish from Calanus eggs and naupliito copepodites at 14-16 mm, which corresponds to redfish that have recently meta78morphosed. The transition from larva to juvenile may be an important phase inthe life history of redfish. The switch to Oithona copepodites demonstrates thatcapturing this prey type is accomplished best during the juvenile stage and not during the larval stage. When they metamorphose redfish larvae acquire fin rays andvertebrae which are expected to increase their swimming performance. In addition,they also develop gill rakers at this time which should aid in the retention of smallfood particles (Einarrson 1960, Bainbridge and McKay 1968), such as the naupliiand copepodites of Oithona spp.My observations on standardized stomach weights of juvenile redfish suggest thattheir ability to actively select Oithona copepodites resulted in increased feedingrates. This was true when comparing values for larvae in early and late May 1981(0.085 and 0.132 g mm3) versus juveniles sampled in late June and early August(0.35 1 and 0.311 ug mm3) of the same year (Table 16). During each cruise in 1981,the diet was dominated by copepodites (Table 13). Similarly, the relative stomachweight of 12 mm larvae (recently metamorphosed) sampled during late May 1981was not significantly different from that measured for 1980 larvae, whereas relativestomach weights for all larval sizes were significantly lower in 1981.It appears that there is a clear advantage to metamorphosing from the larval tothe juvenile stage based on an improved ability to feed. A similar feeding responsewas reported for the freshwater roach, where juveniles increased the variety of foodtypes eaten with a concomittant increase in the quantity of food eaten (Wieser etal. 1988). More specifically, I have shown that there is a direct feeding advantagefor pelagic juvenile redfish in their ability to eat Oithona copepodites. Feeding ona wider prey spectrum would enable redfish to optimize their feeding which should79lead to increased growth, and ultimately survival, during the pelagic period of highgrowth rates. An ability to successfully feed on Oithona copepodites is necessarydue to the successional nature of the copepod community on Flemish Cap. As theseason progresses availability of Calamus eggs and nauplii declines, to be replacedby Oithona copepodites. C. finmarchicus copepodites were not selected by juvenileredfish sampled in this study up 23-28 mm in length. Therefore, this prey type maynot be important to redfish feeding at this developmental stage.Chambers and Leggett (1987) and Chambers et al. (1988) reported that winterflounder larvae which grew faster went through metamorphosis earlier, at a smallersize. In addition, fish which grew faster and metamorphosed earlier at a smaller size-- were larger than slower growing fish that metamorphosed later, when compared a acommon age. However, it was not apparent to them why an earlier metamorphosisat small sizes is advantageous. Chambers et al. (1988) state, “At this point in ourunderstanding of the ecology of marine fishes we lack estimates of the benefits andcosts of being large-at-age and metamorphosing early as opposed to being large atmetamorphosis but experiencing a longer larval period.”Results from my study agree with those reported by Chambers and Leggett (1987)and Chambers et al. (1988). In 1980 redfish entered, and completed, metamorphosisat smaller sizes than in 1981. Redfish larvae in 1980 had greater quantities of food intheir stomachs, per unit size, indicative of higher feeding rates. They also had higherrelative condition. Previously, Penney and Evans (1985) reported that Flemish Capredfish in 1980 grew significantly faster than in 1981 and that the time of peakspawning was not different; indicating their ages would be similar each year whensampled at the end of May. Therefore, metamorphosis of redfish in 1980 appears to80have occurred at smaller sizes for younger fish that were growing more quickly. Thehigher incidence of feeding observed in 1980, based on standardized stomachs, mayhave resulted in the significantly greater condition and growth which was, in turn,responsible for the smaller size at metamorphosis that year. My results demonstratean improved feeding ability of juvenile redfish, which was shown in higher feedingrates and condition. Given the seasonal succession of zooplankton on Flemish Capit may be reasonable to hypothesize that an earlier metamorphosis is advantageousto the feeding, growth and survival of redfish.3.4.3 Match/Mismatch of Redfish and their PreyThe increased selectivity for Oithonacopepodites by redfish juveniles appears to bean adaptation to the seasonal succession of the zooplankton community. Redfishspawn in association with Calamus fimmarchicus spawning when the preferred foodsof redfish larvae (eggs and nauphi) are plentiful. As larvae grow and metamorphoseto juveniles the change in prey selectivity to include Oithoma copepodites coincideswith the end of the spring production of C. finmarchicus eggs and nauplii. Theassociation indicates there is a temporal match of redfish larvae with C. fimmarchicuseggs and nauplii.Ellertsen et al. (1989) have reported such a match for cod larvae, and demonstrated that it varied dependent on the onset of spring spawning of C. finmarchicus.They reported that the time of peak cod spawning only varied by one week over 11years, 1976-86, in contrast to peak spawning of Calamus finmarchicus which variedby almost 7 weeks as measured in 16 years for the period 1960-84. The variablespawning times of C. finmarchicus compared with the relatively stable spawning81time of cod resulted in good and poor matches of cod larvae with their food (op.cit.). These observations are consistent with the original match/mismatch hypothesis of Cushing (1975) if we assume the onset of the spring bloom is directly relatedto the onset of peak C. fimmarchicus spawning. Poor matches of cod larvae withtheir food occurred during warmest and coldest years; in particular, early spawningof C. fina’rchicus occurred during warmer years. In addition, Ellertsen et al. (1989)reported the food of cod larvae is predominantly C. fimmarchicus eggs and naupliibut during the warmest year recorded, 1960, that cod larvae fed predominantly onOithona spp. (Sysoeva nad Degtyareva 1965, ref. by Ellertsen et al. 1989) and thatthis year was one of very low recruitment.-These results are very similar to those reported here for redfish. Analysis of thezooplankton demonstrated that seasonal development was earlier in 1981, a year ofsignificantly warmer temperatures (Chapter 2). The difference between years couldbe explained by temperature dependent development but, also, may have been dueto an earlier spawning. There is some evidence from redfish diet to indicate thatthe spring production of Calamus did begin earlier in 1981. If peak spawning of C.fimmarchicus occurred on Day 107, as predicted earlier, and nauplii developmenttime in 1981 averaged 25 days (Table 11, Chapter 2) then any nauplii spawned onDay 107 would be expected to metamorphose to copepodites by Day 132. During thecruise 2-9 May 1981 (Day 126), calanoids represented 28.8% of all nauplii sampledby 0.080 mm mesh nets, compared to 71.2% which were cyclopoid. If peak spawninghad occurred on Day 107 then one would have expected a greater number of Calanusnauplii at this time; and that they would have constituted a greater proportion ofredfish diet. Together, these observations suggest that the peak spawning of C.82fimmarchicus may have occurred earlier than Day 107 in 1981. Penney and Evans(1985) reported that the time of peak redfish extrusion was not different in 1980 and1981. Therefore, if spawning of C. fimmarchicus was earlier and development fastercompared to spawning of redfish, then a mismatch may have occurred in 1981. Sucha mismatch is supported by observations on larval diet and feeding rates.3.5 Summary1. The diet of redfish larvae was dominated by copepod eggs and nauplii ofCalanus finmarchicus. As redfish grew both the size and number of prey typesincreased in their diet. Preferred prey size typically ranged from 5% to 30% oftheir mouth width.2. The preferred prey of redfish larvae were Calanus finmarchicus eggs and naupliiand Oithona spp. nauplii. Neither calanoid nor cyclopoid copepodites wereselected as prey by redfish larvae.3. There was a shift in diet associated with metamorphosis from larvae to pelagicjuveniles. The preferred prey of juveniles were the same as larvae but also included Oithona copepodites. Copepodites of C. finmarchicus were not selectedby pelagic juvenile redfish ( 23 mm) sampled in this study.4. Significant differences were observed in the diet of larval redfish between 1979and 1980 compared to 1981. In 1979 and 1980 the diet, by weight, consistedpredominantly of calanoid eggs and nauplii, whereas in 1981 the diet consistedprimarily of cyclopoid nauplii and copepodites. As selectivity was the samethis diet shift reflects differences in prey availability.835. The predominantly cyclopoid diet of redfish in 1981 was associated with loweramounts of food in the stomachs, lower relative condition, delayed size and ageof metamorphosis.6. Differences between years in larval redfish diet, feeding rates, and condition werenot apparent in the pelagic juveniles sampled in July-August. This pattern isconsistent with the theory of size-dependent survival, where only larger redfishsurvived each year.7. Size at metamorphosis was smaller in 1980 when redfish larvae were eatingmore, had higher relative condition and were growing (Penney and Evans 1985)---fast.- t is hypothezed that the smaller size at mamorphosi results inimproved feeding which positively affects condition, growth and survival ratesof redfish.8. Feeding rates appeared to be more dependent on sufficient concentrations of thecorrect prey types, as opposed to total prey concentrations, within preferredsize ranges.9. Differences in feeding observed between 1980 and 1981 appeared to result froma better and poorer match, respectively, of redfish larvae with their preferredprey types.844 Redfish Growth and Feeding in Relation to Prey Concentration and Temperature4.1 IntroductionThe importance of food limitation on the annual feeding, growth and, ultimately,survival of larval fishes has not been resolved. Previously starvation has been hypothesized as a primary source of larval mortality. While starvation has been reported to occur in nature, it has not been linked significantly to annual variationsin abundance of larval fish populations (O’Connell 1980, Buckley 1984, Hewitt etal. 1985, Titeilacker 1986, Buckley and Lough 1987, McGurk 1989). The results offield and modelling studies designed to determine the importance of prey concentrations on feeding, growth and survival rates of larval fish are contradictory. Insome instances prey concentrations are considered adequate for larval fish feedingand growth, based on the observation that larval fish have no significant impacton their prey populations (Cushing 1983, Dagg et al. 1984, Peterson and Ausubel1984, Monteleone and Peterson 1986, Kendall et al. 1987, Fortier and Harris 1989,Thompson and Harrop 1991). However, each of these studies was carried out atsmall time and space scales that could not resolve annual differences in larval feeding or growth of the population being studied. In contrast, other field and modellingstudies have demonstrated that prey concentrations may be limiting to larval fishes(Lasker 1975, 1981, Cohen and Lough 1983, Govoni et al. 1985, Buckley and Lougli1987, Bollens 1988, Kiorboe et al. 1988, Ellertsen et al. 1989, Fortier and Gagne1990, Freeberg et al. 1990, Sundby and Fossum 1990). Some of these studies werecarried out at sufficient temporal and spatial scales to resolve annual differences infeeding and growth of the population (Cohen and Lough 1983, Freeberg et al. 1990,85Sundby and Fossum 1990). Effects of prey concentrations on larval mortality havereceived less attention. Taggart and Leggett (1987) reported local prey concentrations had no effect on larval capelin mortality during 3 years of study. However, thisstudy occurred in a small area relative to the entire population and only measuredmortality of recently released larvae over short periods of 24 hours. Peterman andBradford (1987) demonstrated significant mortality of 19-day old larvae for a population of northern anchovy studied over 13 years, consistent with Lasker’s (1975,1981) food-limitation hypothesis. However, mortalities associated with these first-feeding larvae were not correlated with recruitment at age 1 (Peterman et al. 1988).While reports of prey limitations on feeding, growth and survival of larval fishes atlocal scales are contradictory, there is some evidence that prey concentrations maysignificantly affect annual growth and annual variations in abundance of larval fishpopulations.Growth rate of fishes is known to be a function of ration, temperature and fishsize (Brett 1979). Growth is expected to increase with increasing ration, independent of fish size and within a preferred temperature range. Ration is a functionof prey availability, usually measured as prey concentration, although what constitutes prey availability in nature is a complex question. When ration is not limiting,growth will increase linearly with temperature, through a fish’s preferred temperature range. Through the entire temperature range of a fish species growth rate is adome-shaped function. However, optimum temperature for growth is dependent onfood availability, or ration (Brett 1979); as ration is reduced so is the optimum temperature for growth. Also, when ration is limiting an increase in temperature willresult in lower growth rates, due to the physiological increase in maintenance costs86when food energy is already limiting. This interaction of ration and temperature ongrowth rates precludes a simple interpretation, and emphasizes that examination ofthe effects of food availability on vital rates of larval fishes must include the effectsof temperature.The purpose of this chapter is to examine the effects of prey availability andtemperature on the feeding, growth and relative condition of larval and pelagicjuvenile redflsh on Flemish Cap. Comparisons are examined between years, 1980 and1981, as well as spatially within different cruises each year. The spatial interactionof temperature with prey concentrations is also evaluated.4.2 Methods4.2.1 GrowthGrowth rate of redflsh during the larval and pelagic juvenile stages was estimatedseveral ways. The simplest was based on differences in mean lengths between cruisesfor 1980 and 1981 and an overall estimate of growth rate from all available cruises.These results were previously published by Anderson (1984). Relative condition,considered here as a comparative index of growth, was estimated as weight perlength ([Lg mm’).A comprehensive estimate of growth rate was obtained from otoliths. This workwas carried out by R. Penney as part of his MSc. thesis and previously published(Penney and Evans 1985). A summary of the otolith preparation and reading techniques is contained in Penney and Evans (1985). I initiated the otolith work in 1980with my technician Randy Penney. During that year we attempted to raise and feednewly extruded redflsh larvae aboard our research vessel to determine both the age87of onset of increment formation and whether or not increment formation was daily.Newly extruded redfish larvae were obtained from adult ripe females obtained bybottom trawl during our cruise 6-13 April 1980. These larvae were kept in 4 litercontainers within a continuosly running sea water bath of surface water. Naturalzooplantkon were obtained approximately every 4-6 hours and injected into the larval containers. While these larvae were kept alive for many days, daily subsamplingrevealed that they were not feeding or growing. The experiment was terminatedat the end of the cruise. It was clear that a laboratory should be set up onshorewhere conditions could be much more carefully controlled and monitored. I assignedthis work to R. Penney, who was unsuccessfully in laboratory culture attempts thefollowing year, 1981 (Penney 1984). However, we obtained otoliths from larvae collected in the field during 1980 and 1981 and R. Penney did the otolith analyses aspart of his MSc. thesis (Penney 1984). Preliminary results were published in Penneyand Anderson (1981).Back-calculations of growth rate were estimated from the measurement of otolithincrement width during each 5-day period following the broad check ring associatedwith release from the adult female to the water. As the diameter of the otolith waslinearly related to redfish length, these 5-day measurements can directly be convertedto redfish length and estimates of growth rate determined, following estimationof length at extrusion. I followed the methods of Penney and Evans (1985) toestimate growth rate for each redfish during each 5-day period of growth. Here Ihave summarized growth rate different ways. For overall estimates of growth ratefor any particular period, such as each month or the period of time during whicheach of our research cruises took place, I have estimated the mean growth rate for88these periods from all fish, all 5-day periods. In this way growth rate for theseperiods is estimated from the survivors and will be an overestimate of the meanpopulation growth rate which actually occurred during these periods of time if therewas differential survival dependent on growth rates.For estimates of growth rate averaged for each station within a particular cruiseI estimated mean growth rate 3 ways: 1) as the mean growth rate for each redfishfor all 5-day periods; 2) as the mean growth rate for the last 5-day period oniy foreach redfish; 3) as the mean growth rate for the last two 5-day periods (last 10 daysof growth) only for each redfish. Age (days) is available for each redfish where theotolith was examined, being simply the sum of rings counted following the extrusioncheck ring. Otolith estimates of growth rate and age were not available from allstations (Table 30).4.2.2 Prey ConcentrationQuantitative estimates of prey concentration using 0.080 mm mesh nets are availablefrom 4 cruises: 6-13 April 1980, 20-26 May 1980, 2-9 May 1981 and 22-27 May1981. Catches from these cruises were dominated by redfish larvae as opposed topelagic juveniles. By the end of June 1981 pelagic juveniles predominated and, whilestill eating copepod eggs and nauplii, they were also selecting and eating cyclopoidcopepodites as well (Chapter 3). For 26-30 June 1981, prey were sampled with0.165 mm mesh nets. Based on the redfish feeding results and considerations ofsample collection and prey identifications, I summarized prey concentration intofour different categories of prey type:89Prey Type 1: Copepod eggs only, which were predominantly Calanoida. This preytype represents both a preferred prey item and should reflect the seasonal abundance of prey in relation to the spring production of Calanus finmarchicus.Prey Type 2: Copepod eggs and nauplii for both Calanoida and Cyclopoida. Thisprey type will represent all the preferred prey items for larval redflsh, with nodiscrimination based on the large size ranges represented for both eggs andnauplii.Prey Type 3: All copepod eggs and nauplii (Type 2) plus all copepodite stagesof Calanus finmarchicus and Oithona spp., the two predominant copepods onFlemish Cap The-0M80nim rneshsarnples most certainly underestimatedCalanus copepodites but they are included here as an indication of total prey,knowing that the larger copepodites both were not eaten and would be poorlysampled.Prey Type 4: Calanoid eggs and nauplii only. This type only applies to the 1981data as no distinction was made between Calanoida and Cyclopoida during thesampling procedure for 1980 samples. It serves as a comparison of the relativeabundance of these two species groups as food for redfish larvae in 1981, andas a general comparison to the 1980 data.To estimate prey concentrations I expressed these prey types as both numbers(rn—3) and biomass (rngDW m3). Of the two estimates, biomass may be moremeaningful as it will normalize prey concentrations for prey size differences whichoccurred within these 4 prey types. Prey concentrations were summed across allspecies for each station and averaged for each cruise. Biomass was estimated by90assigning a dry weight value to each category, as follows:Species Group Prey Item Biomass(mgDW)Calanoida Eggs 0.00 14Nauplii 0.01Copepodites NS 0.085Calanus fimmarchicus 0.125Cyclopoida Eggs 0.00014Nauplii 0.0045Copepodites NS 0.0136Oithoma spp. 0.0136The sources for these biomass estimates are included in Appendix B. Biomass forother prey types was not included for the following reasons: they were a smallcomponent of the diet, such as Limacima sp.; they were too large to be sampledby 00&I mm. mesh nets such as euphausiids they were iioLspecffied to specier orspecies group, such as Copepoda Not Specified, making it impossible to determinean appropriate biomass value.I also estimated prey concentrations for the cruise 26-30 June 1981 (HAWPAN)based on zooplankton sampled by the 20-cm bongo sampler using 0.165 mm meshnets. These concentrations are not directly comparable to the 0.080 mm meshsamples collected using 50-cm diameter ring net samples due to different samplingprocedures (Chapter 2). In addition, the 0.165 mm mesh samples would underestimate the abundance of copepod eggs and smaller nauphi. However, redfish at thistime were mostly pelagic juveniles feeding actively on larger zooplankton prey andI have included estimates of prey concentration for comparative purposes. In thesesamples I included one additional prey type category:Prey Type 5: Oithona spp. copepodites and all other cyclopoid copepodites. Thisprey type was a preferred prey type of pelagic juvenile redfish, based on selec91tivity results, and would have been quantitatively sampled by 0.165 mm meshnets.These prey types accounted for 84.3% to 95.8% as maximum values, based onaverages measured during each cruise (Table 13). In the case of copepod eggs (PreyType 1) they only accounted for 3.4-12.2% of prey eaten.4.2.3 Standardized Stomach ContentsAs an estimate of feeding rate I calculated the amount of food per size of each redfish,expressed as redfish stomach content weight divided by redfish weight. For cruisemeans of standardized stomach contents, I averaged the total set of observationsbecause sampling occurred continuously over each24-h period for the duration ofeach cruise, 5-7 days. However, to compare estimates of feeding among stationswithin each cruise it was necessary to standardize each station for diurnal feedingdifferences (Chapter 3, Figure 11). For each cruise I determined the 2-h period ofmaximum feeding, then compared the mean of maximum feeding to the mean of allother 2-h periods for that cruise. Based on these 2-h corrections I then adjustedthe observed contents weight of each stomach by the correction factor for that 2-hperiod. For example, if feeding during one 2-h period was 0.75 that of the maximum2-h period, I adjusted all redfish stomach weights for that 2-h period and that cruiseby a factor of 1.33. Because larval fish digestion rates are typically less then 24-h,supported in this study by marked diurnal differences in feeding, and because I havestandardized all stomach contents for diurnal differences my values can be regardedas weight of food eaten per weight of redfish per day (d’). Therefore, I regard thesevalues as representative of feeding rate for the day of capture, noting that this is92not a true measure of feeding rate but an index of relative feeding success.4.2.4 TemperaturesMonthly estimates of temperatures were determined from the historical data setprovided by MEDS, as described in Chapter 3. For temperatures at each stationwithin each cruise, temperature of surface waters was estimated for the top 20 metersat each station. This depth was chosen as it was always less than the pycnocinedepth. Temperature is available at 0, 10 and 20 m intervals based on samplingcarried out during these cruises.4.2.5 Analyses-Initially c6rnparisons of differences in growth rates, feeding rates, temperatures andprey concentrations were made among cruises for 1980 and 1981. These comparisonswere based on mean values calculated for each cruise.To further explore the relationships of redfish growth and feeding to temperatureand prey concentrations I examined differences among stations within each cruise forwhich data were available. These comparisons were based on the premise that theremay have been spatial differences in temperature and prey concentrations withineach cruise which would have effected differences in redfish growth and feeding rates.My procedure was to first run simple rank correlations on pairs of variables and toexamine scatter plots for data outliers and non-linear (functional feeding) trends.For the correlation analysis I used all measures of redfish growth rates (mean over allages, most recent 5-day and 10-day growth rates), size (length, dry weight), relativecondition (weight/length), age and feeding rates (standardized for redfish size andtime-of-day). I included estimates of prey concentration for all types summarized93(Types 1-5) and examined both abundance (numbers rn3) and biomass (mgDWm3). Interpretation of these simple correlations would demonstrate the relationshipbetween variable pairs and, taken together, allow for a broader interpretation ofrelationships among variables.To examine the combined effects of temperature and prey availability on redfishgrowth rates I used multiple regression analysis. This analysis was carried out for3 cruises containing a comprehensive set of growth, temperature and prey concentration data: 20-26 May 1980, 22-27 May and 26-30 June 1981. The cruises at theend of May each year also serve as a direct comparison between 1980 and 1981. Theinitial correlation analysis demonstrated a significant relationship between growthrate and both redfish length and age. As growth rate in fish is typically size/agedependent, I standardized growth rates by removing the effects of size. As a measure of prey, I used Prey Type 2 for the two May cruises: all copepod eggs andnauplii as measured by 0.080 mm mesh samples. Based on feeding and selectivityanalyses these prey should best represent a measure of food expected to positivelyaffect growth rate. For the June cruise I used Prey Type 5, the predominant preytype (Table 13). Food availability is taken as the concentration of these prey typesat each station, logo transformed to stabilize variances. Temperatures were averagevalues for 0-20 m depth. The main effects model tested was,Growth = a0 +a1Food +a2Ternperature + E.Because there may be a significant relationship between food and temperaturein these spatial data sets, I first tested for significant multicollinearity (interaction)between the independent variables. Mulitcollinearity was concluded if the Condition94Number > 30, the Variance Inflation Factor > 10 and when any eigenvalue was <0.01, together with an examination of the Type II Sums of Squares (SAS 1985). Inaddition, regression analyses were carried out for redfish length (mm), dry weight(ag) and relative condition (igDW mm—’) dependent on food and temperature, asfor growth rates. All statistical analyses were carried out using regression proceduresin SAS (SAS 1985).4.3 Results4.3.1 Temporal ComparisonsIt was possible to compare mean monthly growth rates to temperature for the periodMarch to July, 1980 and 1981 (Figure 19). Each year growth rate initially decreasedfrom March to April, then increased from April through June in 1980 and July in1981. Growth rate was higher in 1980 compared to 1981 at similar temperaturesacross the entire range of temperatures. For monthly means, growth rate was higherin 1980 at lower temperatures than in 1981, except for July 1980 when temperatures were higher. The difference in temperature was small for March and April butby IVIay the difference averaged 1.4 O and by June it was 2.9 °C. These observations do not support the concept that higher temperatures result in higher growthindependent of year , and indicate that food (ration) may have been the limitingfactor. Unfortunately there are no comparable measures of monthly estimates ofprey concentrations. A direct interpretation of feeding and growth rate responses totemperatures in these data is compromised by: the seasonal increases in redfish size;their metamorphosis from larvae to pelagic juveniles; the seasonal succession of theprey community; and increased daylight and, therefore, feeding period. All of these95factors are expected to affect feeding and growth rates independent of temperatures.The following results are based on average values estimated during each samplingperiod (cruise) in 1980 and 1981. Comparing growth to feeding indicates that,overall, growth rate increased with feeding rate (Figure 20). However, there was anobvious difference between years, with growth rate being much higher in 1980 thanin 1981 at comparable feeding rates. This implies that the higher growth rates in1980 were more than a simple function of feeding rate, expressed here as total foodbiomass per fish weight per day (d’). In this comparison it must be rememberedthat growth rate is expressed as an average over all ages for each fish, whereas feedingis only estimated as an average level of feeding for each fish on the day of capture.In 1980 growth rate initially increased with feeding rate from early April to the endof May then declined slightly with a further increase in feeding as measured duringJuly. This result is consistent with the known decline in growth rate for these older,pelagic juveniles (Penney and Evans 1985). In 1981 growth rate increased withfeeding rate from early May (GADO5O) to the end of June (HAWPAN). However,there was a large increase in growth rate as measured during the first week ofAugust 1981 while feeding rate actually declined. Such a result would occur if therewas differential survival of redfish from June to August, where the faster growingredfish had a higher survival rate. Finally, it is noteworthy that growth rate wassubstantially higher in late May 1980 than in late May 1981 (dashed line, Figure 20).As these samples were collected at the same time, for similar size redfish larvae, itemphasizes the magnitude of difference in feeding and growth rates which occurredbetween these two years, and that lower feeding rates correlated with lower growthrates for larvae in 1981 compared to 1980. As a further comparison, redfish sampled96during June and July of 1981 were feeding at comparable rates to those measuredduring late May 1980, although these rates were for much older larvae, which mayaccount for the lower growth rates.Results comparing growth and feeding rates to temperature measured during eachcruise demonstrate both rates increased at higher temperatures (Figures 21 and 22).Again, the difference between years was apparent. Both growth and feeding rateswere higher in 1980 at lower temperatures when comparing 20-26 May 1980 to 22-27 May 1981 (Figures 21 and 22). It is noteworthy that growth rates appear tobe converging at the highest temperatures, suggesting differences which occurredearlier in the season were disappearing for the survivors by July—August (Figure21). This would occur if the faster growing fish in 1981 had a higher survival rate.As with the data compared for monthly means (Figure 19), the simple relationshipbetween these rate measurements and temperature is complicated by other factors.Comparison of growth and feeding rates across four different measures of preyconcentration demonstrate similar responses for each prey type (Figures 23 a-d, and24 a-d). The only consistent difference was the higher estimate of growth rate versusfeeding rate for 6-13 April 1980 (GADO35). This would occur if faster growing redflshfrom this sampling period had a higher survival rate, as growth rate estimates arebased on backcalculations of redflsh sampled later in the year.Results for copepod eggs (Prey Type 1) demonstrate that higher growth andfeeding occurred in 1980 at higher egg concentrations. However, within each yearhigher growth and feeding occurred at lower concentrations, for cruises which occurred later in the season (Figures 23a, 24a). Obviously, copepod egg concentrationsalone do not adequately represent prey concentrations responsible for redfish larval97feeding and growth within each year.The best representation of prey concentration should be the combined biomassof copepod eggs and nauplii (Prey Type 2), based on larval prey selectivities (Chapter 3). The results demonstrate that higher growth and feeding rates occurred in1980 at lower concentrations of this prey index (Figures 23b, 24b). Within years,it appears that food concentration had no effect in 1980 whereas increasing preyconcentration resulted in higher growth and feeding rates in 1981. Comparing preyconcentrations of copepod eggs and nauplii at the end of May 1980 and 1981, itis apparent that overall prey concentrations were much higher in 1981. This wastrue for both biomass (Figures 23b, 24b) and densities where total nauplii were 4.2times more abundant (Table 31). It was also true at this time that densities ofcalanoid copepods were higher, as was total plankton biomass (Chapter 2). Therelationships of total copepod egg, nauplii and copepodite densities to growth andfeeding rates was similar to that of just eggs and naupii (Figures 23c, 24c). Thedensity of nauplii in late May 1981 was very high, due to a large number of cyclopoidnauplii (94% of total). The number of calanoid nauplii was low, being on averageonly 137 m3 (Table 31). The only prey type that had lower concentrations at thistime was copepod eggs which, in 1981, were about one-half of 1980 densities (Table31). Cyclopoid copepodite densities at the end of May 1980 and 1981 were similar,averaging 407 m3 in 1980 and 456 m3 in 1981. These were primarily Oithonaspp., as sampled by 0.080 mm mesh nets. Nauplii in the 1980 samples were notdesignated as cyclopoid versus calanoid, and it is not known what the proportionswere. However, it is clear that the number of cyclopoid nauplii were very much lessthan in 1981 as total nauplii in 1980 were only 25% of cyclopoid nauplii in 198198(Table 31). It is also known that the seasonal development of Calanus finmarchicusand other calanoids had not progressed so far by this time in 1980 (Chapter 2).Therefore, it is possible that the majority of nauplii sampled in 1980 may have beencalanoid nauplii, although this cannot be stated with certainty. Overall, the resultsindicate that the much higher prey concentrations in 1981 resulted from a significantincrease in small cyclopoid nauphi.Prey concentration differences, considering only calanoid eggs and nauplii, demonstrate that higher growth and feeding rates occurred in 1980 at higher prey concentrations (Figures 23d, 24d). In 1980 an increase in growth rate and a large increasein feeding rate occurred at the end of May, compared to the beginning of April,and this was related to an increase in nauplii over eggs. In 1981 higher growth andfeeding occurred at the end of May at much lower concentrations of calanoid eggsand nauphi compared to the beginning of May.During 2-9 May 1981 (GADO5O) prey concentrations were high by all measures.Copepod eggs were equivalent to those measured 20-26 May 1980 (GADO37) whilenumbers of calanoid and cyclopoid nauplii exceeded those of both 1980 cruises (Table31). During this cruise cyclopoid nauplii accounted for 67% of all copepod nauphiand total prey concentrations of copepod eggs and nauplii were the highest measuredof any cruise.4.3.2 Spatial ComparisonsSpatial differences among stations within cruises in redfish growth, size, age, feedingand condition were compared to differences in surface water temperatures and preyconcentrations within cruises. These comparisons effectively remove season and year99differences in order to further explore the effects of food availability and temperatureon redfish. Within each cruise there were large ranges in values of all variables (Table32). The values represent means for each variable at each station, which covereda grid on Flemish Cap at 37 km (20 nm) station spacing. It is assumed that eachstation is an independent estimate, and representative of conditions at that location.Prey types are as described previously and represent prey biomass concentrations(mgDW m3).Consider first the spatial results for the cruises 20-26 May 1980 (GADO37) and22-27 May 1981 (GADO51) which sampled predominantly redfish larvae of the samesize at the same time each year (Chapter 3). There were significant correlationsbetween redfish growth rates with both length and age (Table 33). These comparisons demonstrate that larger, older redfish were growing faster. The exception wasmean growth rate (over all ages for each fish) during 20-26 May 1980, indicatingthat average growth calculated over the life of each redfish was not dependent onage or length at this time. In each cruise age was correlated with length (r=0.40,P=0.0067 and r=0.40, P=0.0347, respectively) and dry weight (r=0.69, P=0.0001and r=0.39, P=0.0468, respectively). The correlations of the most recent 5-day and10-day growth rate estimates with dry weight were lower but also indicated thatlarger redfish were growing faster (Table 33). In both cruises, mean growth rateswere negatively correlated with dry weight, although the associated probabilities indicate there were no significant relationships. For 26-30 June 1981 (HAWPAN) all 3measures of redfish growth rates were correlated with length, and mean growth ratewith dry weight, but none were correlated with age (Table 33). These correlationsindicate that juvenile redfish growth rate at this time was still dependent on size100but that bigger redfish were not necessarily older. Relative condition was negativelycorrelated with mean growth rate during 20-26 May 1980 and positively correlatedduring 26-30 June 1981. All other correlations of condition with growth rates werelow (Table 33).There were fairly high correlations between redfish lengths, dry weight, relativecondition and age with temperatures for all 3 cruises (Table 34). There were nostrong correlations of redfish length with any of the 5 measures of prey concentration.However, dry weight and relative condition were positively correlated with copepodegg and nauplii biomass during 20-26 May 1980. The only consistent correlations ofgrowth rates with temperature and prey concentrations were for the most recent 5-day and 10-day estimates for 20-26 May 1980 (Table 33). However, the dependenceof growth rates on redfish size, and age, somewhat confounds a direct interpretationof growth rate dependence on food and temperature. These correlations do indicatethat large, older redfish were found at stations with higher temperatures, and theywere growing faster at this time. Prey biomass appeared to have a significant effecton redfish size, condition and growth rate only for 20-26 May 1980. In late May1981 there was no clear dependence of redfish size, growth rate and condition onprey biomass; in fact, the highest correlations were negative (Tables 33 and 34).Diurnally corrected estimates of feeding rates did not significantly correlate withprey concentrations for these 3 cruises, and only marginally correlated with temperature for 26-30 June 1981 (Table 33). Therefore, there is no indication that feedingrate, as estimated here, was dependent on either food or temperature as measured ateach station within these 3 cruises. There were no significant correlations of feedingand growth rates among stations for 20-26 May 1980 (Table 35, Figure 25). For10122-27 May 1981 and 26-30 June 1981 the correlations were higher than during 20-26May 1980, and especially for the most recent 5-day and 10-day growth estimates.Noting the differences in feeding rates between years, these correlations indicatethat feeding rate in late May 1980 may have been at, or above, optimum levels (i. e.excess ration) with the result that there was no correlation with growth rate amongstations.For the regression analyses of food and temperature effects on redfish, the initialtests for all three cruises indicated there were no significant interactions betweentemperature and prey concentration based on Condition Factors (CF) values < 30.Therefore, in each case the main effects model testing the independent effects of foodand temperature on redfish growth rates, length, dry weight and relative conditionwere carried out. However, for 20-26 May 1980 and 26-30 June 1981 CF values were26.3 and 23.5, respectively, and Eigenvalues were < 0.01 in both cases, indicating thepotential for an interaction between the independent variables for these two cruises.There were no significant regressions of growth rates with food and temperatureat P < 0.10 (Table 36). In contrast, there were highly significant regressions fordry weight and relative condition for May 1980 and 1981 and redfish length forMay 1981. In each case only the temperature coefficient was significantly differentthan zero with positive slopes (Table 36). These regressions explained from 26.9 to57.1% of the variance in fish size (length, weight) and relative condition. The highlysignificant values for May 1981 were largely driven by observations at one stationwith high values of both redfish size and temperatures. When this station wasremoved only the regression for dry weight remained significant at P <0.01 and theexplained variation dropped to 19.1%. While this one station sampled only 4 redfish102there is no specific reason to exclude the observation from the data set. Simplescatter plots did not indicate any non-linear relationships for prey concentrationwith any of the independent variables, indicative of a functional feeding relationship.Therefore, these results indicate that during May each year higher temperatureswere related to larger redfish with higher relative condition. While we know fromthe simple correlations that bigger redfish were growing faster, there were no directeffects of either temperature or prey concentration on growth rates after removingthe effects of size.4.4 DiscussionAnalyses of the effects of food availability and temperature on the feeding, growth,and condition of redfish in different years of this study demonstrated there were nosimple, linear relationships of increased growth rates with higher values of eitherprey concentrations or temperatures. Prey concentrations and temperatures wereboth higher in 1981, a year of lower feeding, growth and relative condition. Considering that increases in vital rates are not linear over a broad range of values foreither prey concentrations or temperatures, then my results suggest that feedingand growth conditions in 1981 were not optimal. The obvious difference in comparing prey each year was the dominance of Oithona spp. nauplii and copepoditesin May 1981. Comparing years, the prey field in 1980 was dominated by eggs andnauplii of Calanus finmarchicus compared to nauplii and copepodites of Oithona in1981. In 1981 redfish ate predominantly cylopoid nauplii by number and cyclopoidcopepodites by weight (Chapter 3). C. finmarchicus nauplii are approximately 10times larger than Oithona nauplii and are approximately the same size as Oithona103copepodites. While nauplii of both species were preferred prey of redfish larvae,copepodites were not (Chapter 3). These results indicate that feeding on manysmall nauphi in 1981 was disadvantageous to the feeding and growth of redfish larvae, compared to feeding on relatively fewer, larger nauplii. In addition, feedingon non-preferred prey types, Oithoma copepodites, must also be disadvantageous.When these conditions occur, measures of total prey concentrations, without regardfor prey preference and prey size differences, will be misleading in interpreting thedynamics of feeding and growth of larval fish.For the spatial analyses, simple correlations among variables at the end of May1980 indicated that larger, older redfish larvae were growing faster at stations thathad higher temperatures and more copepod eggs and nauplii, but that this wasnot the case at the end of May 1981. While there were no significant correlationsfor feeding rate among stations, the mean values of feeding and growth rates wereconsiderably higher in 1980 than in 1981 (Figure 20). These results indicate thatgrowth was food limited in 1981. Temperatures, which were higher in 1981, did notresult in higher growth rates, nor was there a significant correlation of growth andtemperature among stations in 1981. These results occurred for prey concentrationsthat were significantly higher in 1981 (Chapter 2) and for redfish larvae that wereapproximately the same size and density each year (Chapter 5).The high correlations of redfish length, weight and relative condition with preybiomass and temperature indicates that larger fish in better condition occurredwhere waters were warmer and prey more plentiful in 1980, but not in 1981. Thesespatial results agree with those based on cruise means. The spatial results indicateredfish on Flemish Cap during late May 1980 were responding to better feeding104and temperature conditions in different areas. In 1981, either the redfish were notcapable of responding to improved conditions or, more likely, such conditions didnot exist. Because temperatures were higher in 1981, it is likely the major differencewas in the prey.The regression analyses of spatial variations among stations for 3 cruises demonstrated that prey concentrations had no effect on redfish while temperatures weresignificantly related to redfish size (length, weight) and condition, but not to growthrates, corrected for fish size. Bigger redfish grew faster, as expected, and they tendedto be older. The association of bigger, older, faster growing redfish with warmer water temperatures may be due to some spatially dependent association where redfishlarvae were released relative to the temperature field on Flemish Cap. On the otherhand, my observations on relative changes in redlish growth rates measured duringdifferent cruises (Section 4.3.1) were consistent with size-dependent survival. Thesespatial results could be interpretted as stations with bigger, older, faster growingredfish at higher temperatures had higher survival rates.Together, the temporal and spatial results of this study indicate that the dominant effect on redfish feeding, growth and condition occurred between years and wasprimarily a function of prey size (cyclopoid versus calanoid nauplii) and prey type(Oithoma copepodites versus Calamus nauplii). Within each year, there was no clearassociation of food and temperature with growth rates.The non-significant associations of redfish feeding and growth with prey availability among stations each year indicates that redfish larvae may not graze downlocal concentrations of prey. In this way, prey concentrations, per se, were not limiting to growth among locations, as reported by other workers (Lasker 1975, 1981,105Cohen and Lough 1983, Govoni et al. 1985, Buckley and Lough 1987, Bollens 1988,Kiorboe et al. 1988, Ellertsen et al. 1989, Fortier and Gagne 1990, Freeberg et al.1990, Sundby and Fossum 1990). However, significantly lower growth occurred in1981 compared to 1980 and my interpretation is that this resulted from food limitation, not temperature. Again, an important observation of this study is thatfood limitation can occur, limiting feeding and growth of fish larvae, but this foodlimitation was not measured as summed prey concentration, or food availability forprey categories that ignored species differences.The lack of a significant association of prey concentrations and availability withredfish feeding, growth and condition may arise from inadequate sampling of redfishand their prey. My data were based on integrated tows from 0-200 m for redfishand 0-100 m for their prey. While I have assumed that redfish and their preyvertically co-occurred on Flemish Cap, the integrated nature of my tows may haveobscured possible vertical differences in prey concentrations. Other studies reportthat redfish larvae occur predominantly in surface waters (Kenchington 1990, K.Frank, per. comm., Bedford Institute of Oceanography) while copepod eggs andnauplii are widely reported to occur in surface waters (Krause and Trahrns 1982,Williams and Conway 1988, Williams et al. 1987). Based on these other studies theassumption that redfish larvae co-occurred with their prey is probably reasonable.Fortier and Gagner (1990) measured the effects of food availability and temperature on growth rates of spring and fall spawned herring during one year. Theyreported that growth rates were food limited initially and temperature limited during the late postlarval stage. However, temperature appeared to have little directeffect as growth rates of spring and fall herring were not significantly different, av106eraging 3.02 and 3.26 % d—’, respectively, for temperatures that were substantiallydifferent, ranging from 8.8—9.7 and 4.1—5.8 OC, respectively. For spring herring,growth rate increased linearly with food availability based on data collected overtime. However, temperature also increased with food availability which obviouslywas a function of time (season). Therefore, there must be an interaction of foodavailability with temperature—they are not independent. This, of course, would betrue of any field study and makes direct interpretation of effects impossible, withoutsampling at sufficiently fine scales and appropriately partitioning the measured effects. Only by comparing differences among years will it be possible to evaluate therelative importance of food versus temperature in effecting differences in feeding,growth and survival.Campana and Hurley (1989) developed general age and temperature dependentgrowth models for cod and haddock. They ignored food limitation in their primaryanalysis, but assumed that food limitation may account for some of the unexplainedvariance. However, closer examination of their work indicates that seasonal changesin prey availability may, in fact, be more important in describing the seasonal growthdynamics of cod and haddock larvae. I base this conclusion on several observations.First, in their model “optimum” temperatures for growth were estimated to be 5.925°C for cod and 6.701 O for haddock. However, these temperatures appear to betoo low based on laboratory work of Laurence (1978) where growth of both cod andhaddock was higher at 9—10 °C than at either 4 or 7 0C, in the presence of excessration. This indicates that optimum temperatures for growth are higher than predicted by Campana and Hurely, when food is not limiting. Due to the interaction ofration with temperature, optimum temperature for growth will be lower when food107is limiting (Brett 1979). This indicates food limitation may have been responsiblefor the seasonal decrease in growth rates, not temperature. Secondly, the optimumtemperatures predicted for growth (5.925 and 6.701 °C) occurred on 13 May 1985(Day 133) for cod and 23 May 1985 (Day 143) for haddock, based on their temperature equations. Compared to peak temperatures predicted by their model tobe 13.8 °C during 3-11 September 1985 (Days 246—254), then temperature wouldappear to have a prolonged and significant limiting effect on seasonal growth of codand haddock. However, consider that a seasonal decline in growth rates for codand haddock began between 13—23 May 1985, based on the dates of temperatureoptima. Data for Browns Bank (McLaren and Corkett 1986) and Georges Bank(Davis 1987) both indicate Calanus nauplii peak in April and have declined in Mayas part of the seasonal production cycle. As Calanus nauplii are the primary foodof these larvae (Bainbridge and Cooper 1968, Kane 1984), these data indicate thatfood limitation in May would result in seasonal declines in growth rates, rather thantemperatures which were still increasing. In addition, we should consider that therewould be a seasonal decrease in optimum temperatures for growth resulting fromfood limitation. Therefore, I conclude that temperature in the model of Campanaand Hurley (1989) is, in fact, a proxy measure of the seasonal decline in copepodnauplii and, therefore, available food.The pattern of seasonal growth rate of redfish on Flemish Cap, based on otolithanalyses, was similar in 1980 and 1981 (Figure 7, Penney and Evans 1985). Followingrelease, and during the period of first feeding, there was an initial lag in growth rateswhich was followed by an increase to maximum values that remained relativelyconstant for a period of time. Finally, a seasonal decline in growth rates occurred108during the latter period in which growth rates were determined. This general patterncharacterizes growth rates for both years and for each 5-day cohort examined.To explore this relationship further I replotted these data using date (Day of theyear) as the abscissa (Figure 27). Three observations immediately become apparent,in addition to the interannual difference in growth rates. First, the seasonal declinein growth rates occurred somewhere between the end of May to the beginning ofJuly in 1980 and 1981. This is significant because surface water temperatures werestill increasing at this time; maximum values typically occur in August—September(Anderson 1984, Drinkwater and Trites 1986). Therefore, growth rate underwent aseasonal decline each year in spite of increasing water temperatures.Second, the seasonal decline in growth rates occurred in a relatively short periodof time relative to the range of ages. In 1980 this decline occurred in a 15 day periodbetween Day 170 to 185, and in 1981 it occurred in a 20 day period between Day145 and 165 (Figure 27). While there was a trend for growth rates of older larvae todecrease earlier, this difference was small compared to their ages. For example, in1980 growth rate of the cohort of redfish released on Day 90 began to decrease onDay 175 at 85 days of age, whereas growth rate of redfish released on Day 130 beganto decrease on Day 185 at 55 days of age. Therefore, the seasonal decline in growthrate occurred over a 10 day period for redfish that were 30 days different in age.Using the growth rate and size of release estimates calculated by Penney and Evans(1985) these larvae were 21.3 mm and 16.5 mm in length, respectively. Therefore,both groups of redfish were juveniles and the decline in growth rate does not appearto be due to ontogenetic changes. Similarly in 1981, growth rate of redfish releasedon Day 90 began to decrease on Day 145 at 55 days of age and larvae released on109Day 140 began to decrease on Day 165 at 25 days of age. These ages correspondedto sizes of 14.3 mm and 11.0 mm, respectively. Together, these observations indicatethe seasonal decline in growth rates is primarily due to food limitation, and is not afunction of temperature, fish size, age or development. Finally, the seasonal declinein growth rates occurred earlier in 1981 than 1980 (Figure 27).The difference in seasonal decline of growth rates occurred in phase with thedifferent rates of Calanus finma’rchicus development observed each year. This seasonal decline in redfish growth rates also coincided approximately with the periodin which C. finmarchicus are recruiting to copepodite stage CV, entering diapauseand thereby signalling the end of the spring production period. I hypothesize theobserved decline in redfish growth rates results from the end of the spring productionof calanoid copepods. At this time C. finmarchicus are steadily recruiting from nauph to the older, larger copepodite stages which are not available as food (Chapter3). However, redfish which have metamorphosed are capable of positively selectingsmall copepodites of Oithona (Chapter 3). This is an obvious, and necessary, feedingadvantage given the sucession of copepods on Flemish Cap. However, this increasedforaging ability does not appear to offset the seasonal decline in growth rate, evenin 1980, a year of higher growth rate. Redfish that are still larvae will find themselves in ever increasing difficulty: the abundance of Calanus eggs and nauplii aredeclining and they are not capable of successfully feeding on Oithona copepodites,while Oithona nauplii are of insufficient size to maintain high growth. Therefore,the poorer feeding conditions in 1981, which lead to slower growth and delayed ageof metamorphosis, can be related to an earlier end to the spring production cycle ofC. finmarchicus, as described below.110Temperatures in 1981 were warmer than the longterm mean (Table 11, Chapter2). Previously, warmer temperatures have been related to an earlier start of Calamusfimmarchicus spawning as well as faster development (Ellertsen et al. 1989, Chapters2 and 3), contrasted with a constant spawning time of redfish each year (Penney andEvans 1985). Here, and in Chapter 3, I have related an earlier spring development ofthe preferred prey of redfish larvae to lower feeding and growth rates, lower relativecondition, delayed size and age of metamorphosis and an earlier decline in seasonalgrowth rates. These observations demonstrate that the timing of spring productionmight have a significant effect on redfish feeding and growth dynamics, in supportof the match/mismatch hypothesis of Cushing (1975).The dynamics of larval fish feeding, growth and survival occur during a periodof exponentially increasing temperature, increased daylength and time available tofeed, seasonal succession of the zooplankton community, and eventually a seasonaldecline in both abundance and biomass of prey at the end of the spring productioncycle. This entire process occurs at the individual level on the order of 1.5 to 3 weeksand at the community level within two months. There is a natural compounding ofthese factors that defies simple explanation. However, it may be there is a directlink between the seasonal dynamics of Calarius finmarchicus and larval survival, ashypothesized by Runge (1988).4.5 Summary1. Comparison of feeding and growth rates of redfish larvae for different cruisesdemonstrated feeding and growth were lower in 1981, and that this occurredduring a year when temperatures and total prey concentrations were higher.1112. Lower larval redfish feeding and growth rates in 1981 were related to high abundances of both small prey (cyclopoid nauplii) and non-preferred prey (Oithonaspp. copepodites). These observations emphasize that measures of prey availability must be based on concentrations of prey of both preferred type andsize, which will vary with the species composition among years and will not bemeasured simply by size or type alone.3. Spatial comparisons among stations within 3 cruises of food availability andtemperature effects on redfish feeding, growth, length, weight and relative condition did not clearly demonstrate linear dependencies. Correlation analysesindicated redfish larvae in late May 1980 were larger, growing faster and inbetter condition at stations with higher prey concentrations and temperature,but that there were no significant relationships in late May 1981. In contrast,redfish length, weight and relative condition for larvae and juveniles were allpositively related to temperature and had no association with prey availability.4. Based on differences in cruise means between years and correlation analysesamong stations within cruises, it appears prey concentrations were limiting tolarval growth in late May 1981 for all locations within Flemish Cap and forsome locations in late May 1980 within Flemish Cap.5. Measures of feeding rate demonstrated significant differences between yearsbut were not significantly related to any other parameter measured amongstations, indicating either there was no local effect or that feeding rate was apoor measure of overall feeding success.6. Differences in feeding, growth and condition between years were related to a112shift in the prey field, where development of Calanus finmarchicus occurredearlier in 1981 and the zooplankton were dominated in May by Oithoma spp.nauplii and copepodites. It appears that an earlier occurrence and subsequentdecline in the abundance of C. fimmarchicus eggs and nauphi was detrimentalto the feeding and growth of redfish larvae in different years.1135 Survival of Redfish on Flemish Cap5.1 IntroductionIt has been hypothesized that survival rate of fish will be directly related to growthrate during one or more life history stages in their first year of life (Ware 1975,Shephard and Cushing 1980, Anderson 1988). In particular, fish that grow fasterduring the larval stage will have higher survival than slow growers during any givenyear, and a year of faster growth for the annual cohort will be a year of highersurvival as well. Results from Flemish Cap demonstrated that redfish grew faster in1980 compared to 1981 (Penney and Evans 1985) and that they had higher relativecondition, were feeding at higher rates, and began metamorphosis at smaller sizesand younger ages (Chapters 3 and 4). These results indicate that redfish passedthrough the larval period faster as a function of both higher growth rates as wellas higher development rates (reciprocal of the larval period), and that this resultedfrom improved feeding conditions in 1980.The purpose of this chapter is to examine evidence that survival was related togrowth rate during the two years of this study. Abundance data are available for1979 and these data are examined as well, in relation to survival in 1980 and 1981.The effect of physical dispersal of redfish larvae from Flemish Cap is examined as apossible abiotic factor which may have differentially affected survival during these3 years. Finally, relative abundances of juveniles for these 3 years are examined atdifferent points in the life history up to 5 years of age as an inidication of the relativeimportance of larval survival to eventual year-class formation.1145.2 Larval Survival and Growth RateIchthyoplankton sampling was carried out on Flemish Cap at different times during1978 to 1983, ranging in date from early March to the first week of August. Onecruise was carried out in October 1977 but no fish larvae were caught. In mostcases a 42 station grid was sampled over Flemish Cap at 37 km (20 nm) stationspacing. However, in two cases a larger 56 station grid was sampled (16-23 July1978, 22-28 July 1980) and in 3 cruises only 20 stations were sampled over centralFlemish Cap (10-14 July 1979, 1-4 August 1981, 1-3 August 1982). At each stationan oblique 61-cm bongo tow was done from the surface to 200 m depth or within5 m of the bottom, whichever was less, following standard procedures (Smith andRichardson 1977). Payout and retrieval rates were approximately 0.83 and 0.33 ms1, respectively. The sampling and laboratory procedures, together with stationlocations and redfish distributions (numbers m2) at each station are described ina paper published before I began my PhD thesis (Anderson 1984) and I feel thereis no need to detail these procedures again here.As part of my thesis work, I undertook a careful compilation of the original datainto a computer data base and carried out extensive editing of the original data.These procedures are described in detail in the Materials and Methods section ofChapter 2. As a result of the editing procedures, many of the cruise abundanceestimates I reported originally (Anderson 1984) have changed. This occurred primarily because of corrections for volumes of water ifitered from which standardizedestimates of abundance at each station were made.To estimate redfish population abundance for each cruise, I calculated an areal115expansion for each station (37x37 km2) and summed all stations for an estimate ofthe redfish ichthyoplankton population on Flemish Cap at the time of sampling. In 5cruises samples were available for all 42 stations in a 6x7 grid (Table 37). In 2 cruisessamples were available from 41 stations and in 1 cruise from 38 stations (Table 37).Two cruises sampled 56 and 54 stations on a larger 7x8 station grid which addedlines on the south and east sides of the 42-station grid (Table 37). For these cruisesno corrections to total abundances were made. In the first case, a difference of 1station represented a small percent of the total number of stations sampled andthe missing station was located on the outside of the grid. As well, there was nodirect way of estimating abundance at the unsampled locations. For the 2 cruisesin which 38 and 56 stations were sampled, both occurred during July when redfishabundances were low and confined to central waters over the bank, mostly withinthe 400 m isobath. In the case where 38 stations were sampled, all four missingstations were at the outside of the 42-station grid where catches were generally low.It is possible, however, that this estimate of population abundance (for July 1980)may be slightly low. The extra stations for the 56-station grid were added along thesouthern and eastern sides. Examination of these stations demonstrated 9 stationswith zero catch and 5 stations with catches of 0.1—1.4 redfish m2. These compare topeak station abundances of> 30 m2. These distributions are all plotted in Figure3 of Anderson (1984) and I feel there is no need to reproduce these distributionalmaps here. Finally, our last Flemish Cap cruise was done in early March 1983 andwas not reported in my original publication (Anderson 1984). Of the 54 stationssampled, redfish were only caught at one station in the extreme SE corner of the56-station grid. Overall, I felt there was negligible bias introduced in any of these116cruises to estimates of abundance for a ‘42-station’ grid area which covered FlemishCap bank and conclude that these abundance estimates will adequately representthe redfish population at the different times sampled.In 3 cruises only 20 stations were sampled. However, these cruises occurred ineither July or August (Table 37) and in each case the stations were centered over thecentral waters of Flemish Cap where the majority of redfish larvae were observedin 1978 and 1980 when the whole area was surveyed (Figure 3, Anderson 1984). Toestimate total abundance for the ‘42-station’ Flemish Cap area for these cruises Icompared the abundances of the 20-station areas, as sampled in 1981 and 1982, tototal abundances for the two years in which the entire area was surveyed. This wasonly possible for 2 cruises as the cruise in July 1979 did not capture any juvenileredfish from the April-release period. These comparisons indicated that the 20-station area sampled would have caught from 67.5% to 93.5% of total redfish caught,with an overall mean of 83.0%. Therefore, for the 20-station cruises I estimatedabundances for a ‘42-station’ grid area assuming that we sampled approximately83% of the redfish population . These results are presented in Table 37.For cruises during June through August in different years there were obviousmodes in the length frequencies. To estimate the abundance of April-released redfish(i. e. from the peak release which occurred late April to early May, separate fromredfish released later in the season as part of the second release period) I estimatedpopulation abundance for the 26-30 June 1981 cruise for all redfish 8 mm length,and for all July-August cruises for redfish 12 mm length. These results arepresented in Table 37.Redfish population estimates during the first 3 months of life for redfish released117in the spring, referred to here as April-released redfish, are plotted in Figure 28. Ihave calculated the instantaneous rates of change between different cruises in eachyear, 1979-81. The rate of increase in spring was rapid, with little release underwayduring March and peak values occurring one month later. There is an abrupt endto spawning by the beginning of May with very small abundances of 5-7 mm larvaeoccurring by late May in both 1980 and 1981. The seasonal trend in abundancesagrees with the time of peak release estimated by Penney and Evans (1985) (Figure28). Mortality rate estimated from late May (Day 145) to July (Day 207) in 1980averaged —0.059 d’ (Figure 28). In 1981 mortality rate increased from a low valueduring May of —0.035 d’, to —0.062 d’ from the end of May to the end of June,to —0.103 d1 from the end of June to the first of August (Figure 28). To comparewith 1980, mortality over the period from the end of May (Day 145) to August (Day215) averaged —0.083 d1. It is likely the low mortality rate estimated for May 1981(—0.035 d’) was due, in part, to ongoing release of small redfish which would biasthis rate down. However, by the end of May in both 1980 and 1981 there were few6-7 mm redfish, indicating the spring spawning was over and mortality estimatedfor the main spring spawning of redfish from these dates would not be biased bycontinued spawning. Therefore, these data indicate that mortality rate was greaterin 1981 than in 1980, with the highest rate of mortality occurring between the endof June through to the first week of August in 1981.I standardized population abundance estimates for two times: a) for the peakrelease period at the end of April (Day 120) and, b) for pelagic juveniles in August(Day 215). The time of peak release was estimated to have occurred between Day110—119 (April 20—29) in 1980 and 1981 (Penney and Evans 1985 and Figure 28).118Therefore, I standardized larval redfish abundance to Day 120 using N0 = N etfor different estimates of z (d’) (Table 38). Similarly, abundance estimates weremade for redfish at approximately 3 months age, standardized to Day 215 for 1978—1982 (Table 38). I estimated population abundances in some years based on differentinstantaneous rates of change. In 1979 the increase between Days 115 to 120 was high(0.118 d’) and it is doubtful this rate of increase continued to Day 120. Therefore,my estimate for Day 120 is probably too high, and I choose the estimate of 7.1 x 1012as peak abundance in 1979. In 1980 and 1981, my backcalculations were based onmortalities estimated from Day 126—145 in 1981 and from Day 145-207 in 1980. Foreach year I conclude the estimate based on —0.035 d1 as the most probable mortalityrate, as this was calculated for the month of May in 1981 and we would expect thevalue to be lower due to ongoing release during the first part of May. Therefore, Ipredict my best estimates of peak population abundance in 1980 and 1981 are 1.1x iO’ and 6.9 x 1012, respectively (Table 38). For juvenile abundances estimatedfor Day 215 each year I estimated both high and low mortality rates in 1978 witheither value being as likely. In 1979 I have included population estimates calculatedover the entire period, Day 120 to 215, at high mortality rates for comparison only.We know mortality in 1979 was very high. The estimate for 1980 was based onmortality calculated from Day 145-207 that year while for 1981 and 1982 there wasno time adjustment to the original survey estimate.As a direct comparison of mortality over the first 3 months of life, I also calculatedmortality based on my estimates of peak abundances on Day 120 to abundancesestimated for Day 215 each year (Table 38). In 1980 the average rate of mortality was—0.053 d compared to —0.070 d’ in 1981. Finally, it is apparent that mortality rate119in 1979 was extremely high. Abundance measured at the end of April 1979 occurredduring the time of peak spawning estimated for 1980 and 1981 (Penney and Evans1985 and Figure 28), was the highest measured during any survey and comparesclosely with peak values estimated for 1980 and 1981 (Table 38). However, duringJuly 1979 only small, recently released redfish larvae were caught during our survey,indicating complete mortality of the April-released cohort. This approximates amortality rate of —0.3 d’.It is inevitable that sampler avoidance occurs during ichthyoplankton samplingand that this would increase with fish size. However, my treatment of the dataemphasizes a comparison between years where I assume that any sampler avoidancewould be the same each year. Unless there were size or time dependent differencesaffecting avoidance among years my comparisons of relative survival should be accurate. However, there is some evidence that significant sampler avoidance did notoccur in my data collections based on mean size. Increases in fish length are expected to be exponential during the first few months of life (Ricker 1979). I fitan exponential rate of increase in redfish length for all data, all years (Figure 29).Comparison of mean sizes during July—August of four years indicates mean size wasat, or greater than, that expected based on an exponential increase in length. Thisindicates there was no significant evidence of fish avoidance evident in smaller red-fish being caught than expected based on an exponential rate of increase in length.I interpret this result to be that there was no significant avoidance for redfish larvaeup to approximately 20 mm length.Comparing these mortality estimates to growth rates previously calculated byPenney and Evans (1985) and also presented in Chapter 4, it is clear that higher120growth rate in 1980 was related to lower mortality (higher survival). Average growthrate during the first 3 months of life in 1980 was 0.16 mm d’ and mortality was—0.053 d’. In 1981 average growth rate was 0.11 mm d’ and mortality was —0.070d’. Therefore, one might conclude that higher growth rate was related to highersurvival in these two years.It is interesting to compare data plotted for mortality rates and mean redfish size.In 1981 mortality rate increased with time and was highest from June to August(Figure 28). Mean lengths of redfish were the same at the end of May 1980 and 1981(Figure 29). By July 1980 mean size of April-released redfish was 17.6 mm (Day 207)while in 1981 it was 20.1 mm (Day 215). Based on an average growth rate of 0.16 mmd’ in 1980, projection of mean size to Day 215 in 1980 would be 18.9 mm, whichis similar to, but less than, 1981 on the same date. Knowing that growth rate wasslower in 1981 this result would occur if differential size-dependent survival occurredbetween Day 145 and Day 215 in 1981, with only the larger redfish surviving. Meansize estimated for the end of June 1981 (Day 179) indicates redflsh were muchsmaller than expected, compared to the exponential increase in size caluclated forall data, all years (Figure 29). This indicates that smaller, slower growing redfish stillcomprised a major part of the population in late June 1981 and suggests that sizedependent mortality occurred during July, which compares with the high estimatesof mortality calculated for this period (Figure 28). This part of the season coincideswith the end of spring Calanus production, the seasonal metamorphosis of redfishfrom larvae to pelagic juveniles and continued exponential heating of surface waters(Chapters 2 and 3). The lower growth and survival measured in 1981 occurredduring a year of higher than average water temperatures and during a year in which121the seasonal production of Calamus fimmarchicus was completed earlier (Chapter 2).Redfish growth rates and zooplankton development data are not available for1979. However, it is apparent that mortality was extremely high during the first3 months of life (Table 37, Figure 28) and this occurred during a year of aboveaverage water temperatures, in fact the warmest of the 3 years (Table 10, Chapter2). These observations are consistent with those of 1980 and 1981, and supportthe proposal that high water temperatures during spring cause earlier spawning andfaster seasonal development rate of Calamus fimmarchicus which is detrimental tothe feeding, condition, growth, and survival of redfish larvae.5.3 Physical Dispersal of Redfish Larvae on Flemish CapFor many years it has been hypothesized that the dispersal (drift/retention) of fisheggs and larvae during their planktonic phase is an important determinant of year-class strength. Hjort (1914) was the first to clearly state this. More recently theeffect of physical factors on transport and variable retention within spawning areashas been re-emphasized as an important regulator of marine fish survival (Parrish etal. 1981, Iles and Sinclair 1982). Retention within spawning areas appears to be thestrategy of some species and stocks: several pelagic species spawning in upweffingareas (Parrish et al. 1981), several herring stocks in both the east and west NorthAtlantic (Ties and Sinclair 1982, Fortier and Gagner 1990), most demersal species onthe Scotian Shelf (O’Boyle et aT. 1984), and capelin on the southern Grand Bank(Frank and Carscadden 1989). Conversely, drift away from spawning areas duringthe egg and larval phase is an alternative strategy of some species and stocks: manypelagic species spawning in upwelling areas (Nelson et al. 1977, Parrish et al. 1981,122Bailey 1981, Kruse and Tyler 1989), certain herring stocks in the North Sea (Cushing1986, Bartsch et al. 1989), capelin in the St. Lawrence estuary (Fortier and Leggett1983) and North Sea plaice (Bergman et al. 1988), to mention a few. Therefore,both retention in and drift from spawning areas are successful life history strategiesduring the planktonic stages of fish. In reality, every variant and combination ofthese two “strategies” exist. Fish stocks evolve to live in the regime and are adaptedto the physical environment that is present to enusre stable recruitment and stocksize over time.It is convenient to consider the detrimental effects of physical transport as occurring due to one of two mechanisms; either advective losses as the result of variationsin mean flows or Ekman transport of surface waters off shelf areas. Studying theeffects of physical transport of surface waters on the survival of fish eggs and larvaeis not trivial. It involves studying dynamics relating to the seasonal formation ofthe upper mixed layer depth (MLD), the interaction of barodinic and barotropicforces, and biological adaptations such as ovoviviparous release of larvae, benthiceggs, spawning in seasons or locations of favourable transport, stage dependent eggbuoyancies which alter depth distribution, and diurnal vertical migrations of fish larvae and swimming ability in fish as they grow and develop. Few studies have beencarried out at sufficient scales to examine the effect of these interacting processeson the physical dispersal of fish eggs and larvae. Indeed, there is little evidencethat dispersal resulting from variations in physical circulation during the planktonicphase directly affects survival, and eventual year-class strength. It is relatively easyto show that it takes a major change in transport or retention in a marine system tohave a major effect on recruitment, campared to minor (i.e. ±0.02 d1) variability123in mortality or growth rates to cause a similar change in recrulitment level. Only inupwelling areas may this be important (Bailey 1981, Parrish et al. 1981, Hollowedand Bailey 1988, Kruse and Tyler 1989). In the northwest Atlantic, Myers andDrinkwater (1989a) demonstrated 14 of 17 demersal fish stocks showed a negativerelationship between recruitment and the number of warm core rings. They concluded the overall relationship supported the hypothesis that increased warm corering activity reduced the recruitment of these stocks due to entrainment of surfacewater off the shelf. However, they acknowledged the correlative nature of their analysis and that other mechanisms may in part account for the observed relationship.In a companion study Myers and Drinkwater (1989b) failed to show any relationshipof recruitment for these same stocks to the effects of cross-shelf Ekman transportduring the egg and larval stages. Recently, a comprehensive study on the dispersalof eggs and larvae of cod and haddock on Browns Bank, Scotian Shelf was carriedout over 3 years (Campana et al. 1989). This study failed to demonstrate anysignificant mortality associated with advection off the bank. Presently, there is noconclusive evidence that physical transport of eggs and larvae away from offshorespawning banks is a prime determinant of survival during these stages. On the otherhand, it is likely that marine fish have evolved adaptations to reduce the effects oftransport on survival (Parrish et at 1981, Myers and Drinkwater 1989b).Previous studies on Flemish Cap indicate that significant changes in water masscharacteristics do not occur on a year-to-year basis. Instead, changes occur cyclically over 3-5 year intervals (Hayes et al. 1977, Keeley 1982a). These cycles appearto result from low frequency atmospheric forcing of the subarctic gyre and its subsequent interaction with the Labrador and North Atlantic Currents (Hayes et at1241977). Similarly, over the years 1946-86 for the northwest Atlantic, Myers andDrinkwater (1989b) found the dominant variability in Ekman transport occurredon decadal time scales. The three years of my study, 1979-81, occurred during aperiod of higher temperatures and salinities relative to long-term mean conditions(Keeley 1982a). Therefore, there was no apparent shift in oceanographic conditionsthat might occur coincident with a switch from warm-saline to cold-fresh years.Flemish Cap waters represent a mixture of Labrador Current and North AtlanticCurrent water types (Hayes et al. 1977, Keeley 1982a). The circulation on FlemishCap is characterized by a weak anticyclonic circulation (‘ 3-5 cm/s residual circulation) that can be disrupted by storms (Hayes et al. 1977, Hill et al. 1975, Ross1980, 1981, Kudlo et al 1984). Loder et al. (1988) reported a mean annual recirculation time of 67-78 days for near-surface waters along the 400 m isobath and amean residence time of 32-40 days for near-surface waters within the 400 m isobath.They concluded there is sufficient residency time “to be favourable to biological processes.” This would be true for the larval redlish stage which varied from 10-33 daysduration, and for Calanus finmarchicus nauplii which typically developed to copepodites in less than 21 days. In addition, the gyre tends to become more stable withthe transition from winter to summer conditions (Kudlo et al. 1984). Therefore,residency time would be expected to increase with the seasonal formation of thepycnodine, which is noticeably developed by May (Anderson 1984, Drinkwater andTrites 1986). The existing oceanographic information supports the inference thatwaters over Flemish Cap are retained for sufficient periods of time for the growthand development of redfish larvae and Calanus nauplii.On Flemish Cap it has been hypothesized that interannual variation in the125strength of the anticyclonic gyre may affect the retention of plankton (reviewedby Lilly 1987). For many years the U.S.S.R. has carried out a program of researchon Flemish Cap calculating dynamic height charts (Kudlo et al. 1984). As a simpleevaluation of this hypothesis I calculated a relative index of gyre strength from thesedata. Drifter studies have demonstrated that surface water tends to leave FlemishCap in the southeast quadrant, independent of wind forcing (Ross 1981). Therefore,I calculated a simple index of gyre strength by calculating the difference in dynamicheight between waters over Flemish Cap and waters to the southeast or east of thecentral bank (< 200 m depth). These data originate from dynamic height chartsproduced by the U.S.S.R., as published by ICNAF/NAFO. They spanned the years-1971—81 but only in 1974 and 1978—81 were there 4 or more charts per year. Nodata were available in 1972, and only one observation was available in 1971, 1973,1975 and 1976.Relative gyre strength for the entire data series ranged between 2—8 (Figure 30a).Dynamic height was always greater over Flemish Cap and there was considerablevariation among years in the index for any given date. There was a trend to largervalues in summer and fall (June to October) with values increasing from about 4—8.This seasonal trend was previously reported by Kudlo et at. (1984). Comparison ofrelative gyre strength for the 3 years of this study, 1979—81, indicates similar estimates of gyre strength among years (Figure 30b). The approximate critical periodfor larval redfish would fall roughly between Day 120—150 (May). The low observation on Day 126 in 1979 is the only indication of weak gyre strength among theseyears. It is not apparent from these data that variation in the mean anticycloniccirculation on Flemish Cap would explain among year differences observed in redfish126survival.5.4 Relationship of Larval Survival to RecruitmentResults of this study demonstrate substantial differences in larval survival amongthe three years, 19 79-81. An important question in fisheries science is whether or notsurvival during the larval period relates to eventual recruitment. Redfish on FlemishCap generally recruit to the fishery at 5 years of age, while they do not mature untilapproximately 10 years of age (B. Atkinson, Northwest Atlantic Fisheries Center,St. John’s, Newfoundland, Personal Communication). During this relatively longperiod predation mortality on juveniles is expected to be important in determiningeventual recruitment into the fishery (Sissenwine 1984).-To assess the age at which year-class size was determined for these 3 years Ihave estimated relative abundance at different ages: the time of release (age 0);abundance of pelagic juveniles at 3 months of age immediately following the larvalperiod; juvenile abundance at 1 and 2 years of age, and abundance at age 5 whenthese redfish recruited to the fishery (Table 39). Estimates of peak larval abundancesat age 0 (Day 120) and at approximately 3 months of age (Day 215) were takenfrom Table 38 (Table 39). Relative abundance data is also available for redfishat approximately 1 and 2 years of age for these years (Table 2, Lilly and Gavaris1982). They estimated abundance based on an index of cod stomach contents andabundance of juvenile redfish caught in the small mesh liner of an Engels highlifttrawl. Both data sets derive from fisheries trawl surveys during January—Februaryeach year. Abundance estimates from the trawl must be viewed with caution due tothe highly skewed nature of the data in which a few very large catches occur (Lilly127and Gavaris 1982) and the poor design of this net for capturing juvenile redfish. Inparticular, the trawl catches in 1981 were affected by one very large catch (op. cit.).There is good evidence that cod predation is primarily directed at juvenile redfishin these size ranges (Lilly 1980, Lilly and Gavaris 1982). Therefore, I believe thecod stomach index is best for estimating the relative abundance of 1 and 2 year oldredfish. Finally, I have taken estimates of stock size and recruitment at 5 years of agefrom Power and Atkinson (1989). Catch and effort data were compiled from ICNAFand NAFO Statistical Bulletins for 1959-1985 and were used in a multiplicativemodel to derive a standardized catch rate index for each year (op. cit. ). Theestimates of catch and CPUE for 1979-81 are presented in Table 39. To estimaterecruitment I have used the catches at 5 years of age based on the Soviet commercialfishery (Vascov et aT. 1987) and then standardized these catches based on the effortdata presented by Power and Atkinson (1989) (Table 39). Note that age 5 was usedalso by Templeman (1976) as his estimate of year-class strength for Flemish Capredfish.To compare relative abundance at each age for the years 1979-81, I set the abundance for the highest year to 1 and scaled the other 2 years relative to it (Figure 31).For stock size and age 5 years recruitment I have plotted the catch rate indices. Forlarvae (age 0) and juveniles (3 months age) I used the estimates standardized to Days120 and 215. For juveniles at approximately ages 1 and 2 years I have used the codstomach index. These data show that stock sizes in 1980 and 1981 were very similarwhile it was much lower in 1979. Note that the fishing effort was 1.7 to 2 timeshigher in 1979 than in either 1980 and 1981, and was the fifth highest reported forthe entire data series 1959-1987. Relative larval abundance estimated at the time of128peak release indicated 1980 was higher than either 1979 or 1981 which were similarto each other. By 3 months of age, 1979 was zero while 1981 was only 12% of 1980abundance. This scaling of relative abundances remained the same from 3 monthsof age through to 5 years of age. Note there was no estimate available for age 2(1.7 years) in 1981. These data indicate that year-class strength for these 3 yearswas established during the first 3 months of life. These results do not preclude thepossibility that juvenile predation mortality may significantly determine eventualyear-class strength in some years. However, these results demonstrate that survival,particularly during the larval stage, was important in determining recruitment ofFlemish Cap redfish. Similar results have recently been reported for a number ofspecies (Rauck and Zijistra 1978, Leggett et al. 1986, van der Veer 1986, Campanaet al. 1989, Hollowed and Bailey 1989, Sundby et al. 1989). Together, these studiesdemonstrate that high survival during the larval stage is a necessary, although notnecessarily sufficient, condition for recruitment of marine fish.5.5 Summary1. Abundance of pelagic juvenile redfish at approximately 3 months of age variedby more than 2 orders of magnitude during 5 years of observation. Averagemortality estimated from the date of peak release (Day 120) to approximately3 months of age (Day 215) was —0.053 d’ in 1980, —0.070 d’ in 1981 and >—0.2 d’ in 1979.2. Sequential measures of population abundance in 1981 indicated that mortalityrate increased with time and was greatest (—0.103 d’) between measurementsmade 26-30 June to 1-3 August 1981.1293. Higher growth and survival rates occurred in 1980 compared to 1981.4. Similar mean sizes of the population measured in August (Day 215) in 1980 and1981, compared to different growth rates measured each year, indicates therewas differential size-dependent survival, with primarily larger redfish survivingfrom the 1981 cohort.5. The role of physical dispersal effecting larval redfish survival was reviewed andit is concluded that variations in physical circulation during this study, 1979—81,would not explain differences in survival rates of larvae.6. Comparison of relative redfish abundances at different ages demonstrated thatrelative reèruitmerit levels were determined during the first 3 months of life forthe years 1979-81.1306 Literature CitedAnderson, J. T. (1984). Early life history of redfish (Sebastes spp.) on FlemishCap. Can. J. Fish. Aquat. Sci. 41: 1106-1116Anderson, J. T. (1988). A review of size dependent survival during pre-recruitstages of fishes in relation to recruitment. J. Northw. Ati. Fish. Sci. 8: 55-66Arthur, D. K. (1976). Food and feeding of larvae of three fishes occurring in theCalifornia current, Sadinops saga; Engraulis mordax, and Trachurus symmet‘ricus. Fish Bull. 74: 517-530Bailey, K. M. (1981). Larval transport and recruitment of Pacific hake Merlucciusproductus. Mar. Ecol. Frog. Ser. 6: 1-9Bainbridge, V. (1965). A preliminary study of Sebastes larvae in relation to theplanktonic environment of the Irminger Sea. ICNAF Spec. Pubi. No. 6:303-308Bainbridge, V. , Cooper, C. A. (1971). Populations of Sebastes larvae in the northAtlantic. ICNAF Res. Bull. 8: 27-35Bainbridge, V. , Corlett J. (1968). The zooplankton of the NORWESTLANTsurveys. ICNAF Spec. Publ. No. 7: 101-122Bainbridge, V. , Mckay, B. J. (1968). The feeding of cod and redfish larvae.ICNAF Spec. Publ. 7: 187-217Barsukov, V. V. , Zakharov, C. P. (1972). Morphological and biological characteristics of the American redfish. Tr. Ployarn. Nauchno-Issled. Proektn. Inst.131Morsk. Rybn. Khoz. Okeanogr. 28: 143-173 (Fish. Res. Board Can. Trans.Ser. No. 2488, 1973)Bartell, S. M. , Breck, J. E., Gardner, R. H., Brenkert, A. L. (1986). Individualparameter perturbation and error analysis of fish bioenergetic models. Can. J.Fish. Aquat. Sci. 43: 160-168Bartsch, J. , Brander, K., Heath, lvi., Munk, P., Richardson, K., Svendson, E.(1989). Modelling the advection of herring larvae in the North Sea. Nature340: 632-636Bergman, M. J. N. , Veer, H. W. van der, Zijistra, J. J. (1988). Plaice nurseries:effects on recruitment. J. Fish. Biol. 33, Suppl. B: 201-218Bollens, S. M. (1988). A model of the predatory impact of larval marine fish onthe population dynamics of their zooplankton prey. J. Plank. Res. 10: 887-906Brattelid, T. E. , Matthews, J. B. L. (1978). Studies on the deep-water pelagiccommunity of Korsfjorden, western Norway. The dry weight and caloric content of Euchaeta movegica (Copepoda), Boreomysis arctica (Mysidacea) andMeganyctiphanes norvegica (Euphausiacea). Sarsia 63: 203-211Brett, J. R. (1979). Environmental factors and growth. Fish Physiol., 8: 599-675Brown, M. B. , Forsythe, A. B. (1974). Robust tests for the equality of variances.J. Amer. Stat. Assoc. 69: 364-367Buckley, L. J. (1984). RNA-DNA ratio: an index of larval fish growth in the sea.Mar. Biol. 80: 291-298132Buckley, L. J. , Lough, R. G. (1987). Recent growth, biochemical compositionand prey field of larval haddock (Melanogrammus aeglefimus) and Atlantic cod(Gadus mo’rhua) on Georges Bank. Can. J. Fish. Aquat. Sci. 44: 14-25Campana, S. E. , Frank, K. T., Hurley, P. C. F., Koeller, P. A., Page, F. H.,Smith, P. C. (1989). Survival and abundance of young Atlantic cod (Gadusmorhua) and haddock (Melanogrammus aeglefinus) as indicators of year-classstrength. Can. J. Fish. Aquat. Sci. 46 (Suppl. 1): 171-182Campana, S. E. , Hurely, P. C. F. (1989). An age- and temperature-mediatedgrowth model for cod (Gadus morhua) and haddock (Melamogrammus aeglefin.us) larvaein the Gulf of Maine. Can. J. Fish. Aquat. Sci, 46: &)3-61aChambers, R. C. , Leggett, W. C. (1987). Size and age at metamorphosis inmarine fishes: an analysis of laboratory-reared winter flounder (Pseudopleuromectes americanus) with a review of variation in other species. Can. J. Fish.Aquat. Sci. 44: 1936-1947Chambers, R. C. , Leggett, W. C., Brown, J. A. (1988). Variation in and amongearly life history traits of laboratory-reared winter flounder Pseudopleuronectesamericanus. Mar. Ecol. Prog. Ser. 47: 1-15Checkley, D. M. (1982). Selective feeding by Atlantic herring (Clupea harengus)larvae on zooplankton in natureal assemblages. Mar. Ecol. Prog. Ser., 9:245-253Checkley, D. M. (1984). Relation of growth to ingestion for larvae of Atlanticherring Clupea harengus and other fish. Mar. Ecol. Prog. Ser. 18: 215-224133Chesson, J. (1978). Measuring preference in selective predation. Ecology 59: 211-215Cohen, R. E. , Lough, R. G. (1983). Prey field of larval herring Clupea harenguson a continental shelf spawning area. Mar. Ecol. Prog. Ser., 10: 211-222Cone, R. S. (1989). The need to reconsider the use of conditon indices in fisheryscience. Trans. Am. Fish. Soc. 118: 510-514Conway, D. V. P. , Minton, R. C. (1975). Identification of the copepodid stagesof some common calanoid copepods. Dept. Agr. Fish. Scot. Mar. Lab. NewSer. No. 7Corkett, C. J. , McLaren, I. A., Sevigny, J. M. (1986). The rearing of the Calanoidcopepods Calamus fimmarchicus (Gunnevus), C. glacialis Jaschnov and C. hyperboreus Kroyer with cmment on the equiportional rule. In: G. Schriever, H.K. Schminke and C.t. Shih (eds.) Proc. 2’” Intl. Conf. on Copepoda. Syllogeus58, p. 539-546Cushing, D. H. (1974). The possible density-dependence of larval mortality andadult mortality in fishes. In J. H. S. Blaxter (ed.) The early life history of fish.Springer-Verlag, Berlin, p. 103-111Cushing, D. H. (1975). Marine ecology and fisheries. Cambridge Univ. Press.Cambridge and New YorkCushing, D. H. (1983). Are fish larvae too dilute to affect the density of theirfood organisms? J. Plankton Res. 5: 847-854134Cushing, D. H. (1986). The migration of larval and juvenile fish from spawningground to nursery ground. J. Cons. in Explor. Mer 43: 43-49.Cushing, D. H. (1989). A difference in structure between ecosystems in stronglystratified waters and in those that are only weakly stratified. J. Plankton Res.11: 1-13Cushing, D. H. (1990). Plankton production and year-class strength in fish populations: an update of the match/mismatch hypothesis. Adv. Mar. Biol. 26:249-293Cushing, D. H. , Harris, J. G. K. (1973). Stock and recruitment and the problemof density-dependence. Rapp. P.-v. Re un. Cons. mt. Explor. Mer 164:142-155Dagg, M. J. , Clarke, M. E., Nishiyama, T., Smith, S. L. (1984). Production andstanding stock of copepod nauplii, food items for larvae of the walleye pollockTheragra chaleogramma in the southeastern Bering Sea. Mar. Ecol. Prog. Ser.19: 7-16Davis, C. S. (1980). Preliminary estimates of copepod extrusion from 0.333 mmand 0.165 mm mesh plankton nets. NAFO 5CR Doc. 80/IX/132 Ser. No.N206, 17 p.Davis, C. S. (1983). Lavoratory rearing of marine calanoid copepods. J. Exp.Mar. Biol. Ecol. 71: 119-133Davis, C. S. (1984). Interaction of a copepod population with the mean circulation on Georges Bank. J. Mar. Res. 42: 573-590135Davis, C. S. (1987). Zooplankton life cycles. In: R. H. Backus and D. W. Bourne(eds.) Georges Bank, MIT Press, Cambridge, pp. 256-267Drenner, R. W. , Strickler, J. R., Obrien, W. J. (1978). Capture probability: therole of zooplankter escape in the selective feeding of planktivorous fish. J. Fish.Res. Board Can. 35: 1370-1373Drinkwater, K. F., Trites, R. W. (1986). Monthly mean temperature and salinityin the Grand Banks region. Can. Tech. Rep. Fish. Auat. Sci. 1450: 111 p.Einarrson, H. (1960). The fry of Sebastes in Icelandic waters and adjacent seas.Hit Fisk. II Bindi. Vol. II, Nr. 7, 67 p.Ellertson, B. , Fossum, P., Solemdal, P., Sundby, S. (1989). Relation betweentemperature and survival of eggs and first-feeding larvae of northeast Artic cod(Gadus morhua L.). Rapp. P.-v. Reun. Con. Tnt. Explor. Mer 191: 209-219Fortier, L. , Gagne, J. A. (1990). Larval herring (Clupea ha’rengus) dispersion,growth and survival in the St. Lawrence estuary: match/mismatch or membership/vagrancy? Can. J. Fish. Aquat. Sci. 47: 1898-1912Fortier, L. , Harris, R. P. (91989). Optimal foraging and density-dependent competition in marine fish larvae. Mar. Ecol. Prog. Ser. 51: 19-33Fortier, L. , Leggett, W. C. (1983). Vertical migrations and transport of larvalfish in a partially mixed estuary. Can. J. Fish. Aquat. Sci. 40: 1543-1555Fortier, L. , Leggett, W. C. (1985). A drift study of larval fish survival. Mar.Ecol. Prog. Ser. 25: 245-257136Fortier, L. , Leggett, W. C. (1984). Small-scale covariability in the abundance offish larvae and their prey. Can. J. Fish. Aquat. Sci. 41: 502-512Foy, M , Anderson, J. T. (1986). Marine plankton computer codes for the Northwest Atlantic Fisheries Centre. Can. Data Rep. Fish. Aquat. Sci. No. 595Frank, K. T. , Carscadden, J. E. (1989). Facators affecting recruitment variabilityof capelin (Mallotus villosus) in the northwest Atlantic. J. Cons. in.t Explor.Mer 45: 146-164Frank, K. T. , Leggett, W. C. (1986). Effect of prey abundnce and size on thegrowth and survival of larval fish: an experimental study employing large volume enclosures. Mar. Ecol. Prog. Ser. 34: 11-22Freeb erg, M. H. , Taylor, W. W., Brown, R. W. (1990). Effect of egg and larvalsurvival on year-class strength of Lake Whitefish in Grand Traverse Bay, LakeMichigan. Trans. Am. Fish. Soc. 119: 92-100Gadomski, D. M. , Boehiert, G. W. (1984). Feeding ecology of pelagic larvaeof English sole (Parophrys vetulus) and butter sole (Isopsetta isolepis) off theOregon coast. iViar. Ecol. Prog. Ser. 20: 1-12Gardner, G. A. , Szabo, I. (1982). British Columbia pelagic marine Copepoda:an indentification namual and annotated bibliography. Can. Soec. Pubi.Fish.Aquat. Sci. 62, 536 p.Gibons, S. G. , Ogilvie, H. S. (1933). The developmental stages of Oithoma helgolandica and Oithona spimirostris, with a note on the occurrence of body spinesin cyclopoid nauplii. J. Mar. Biol. Assoc. U.K. 18: 529-550137Giguere, L. A. , Cote, B., St-Pierre, J.-F. (1988). Metabolic rates scale isometrically in larval fishes. Mar. Ecol. Prog. Ser. 50: 13-19Govoni, J. J. , Chester, A. A., Ross, D. E., Ortner, P. B. (1985). An observationof episodic feeding and growth of larval Leiostomus xamthu’rus in the northernGulf of Mexico. J. Plank. Res. 7: 137-146Guland, J. A. (1965). Survival of the younges stages of fish and its relation toyear-class strength. ICNAF Spec. Pubi. 6: 363-371Hallowed, A. B. , Bailey, K. M. (1988). New perspectives on the relationshipbetween recruitment of Pacific hake (Merluccius productus) and the ocean environment. Can. Spec. Publ. Fish. Aquat. Sd. 108: 207-220 -Hayes, R. M., D. G. Mountain, and T. C. Wolford. (1977). Physical oceanography and the abiotic influence on cod recruitment on Flemish Cap. ICNAF Res.Doc. 77/VI/54, Ser. No. 5107, 33 p.Hewitt, R. P. , Theilacker, G. H., Lo, N. C. H. (1985). Causes of mortality inyoung jack mackerel. Mar. Ecol. Frog. Ser. 26: 1-10Hill, H. W., P. G. W. Jones, J. W. Ramster, and A. R. Folkard. (1975). Thecurrent system east of Newfoundland Grand Bank. ICNAF Spec. Pubi. 10:41-55Hjort, J. (1914). Fluctuations in the great fisheries of northern Europe reviewd inthe light of biological research. Rapp. P.-v. Re Un. Cons. i. Explor. Mer20: 1-228138Houde, E. D. (1987). Fish early life dynamics and recruitment variability. Am.Fish. Soc. Symp. No. 2: 17-29Huntley, M. , Strong, K. W., Dengler, A. T. (1983). Dynamics and communitystructure of zooplankton in the Davis Strait and Northern Labrador Sea. Arctic36: 143-161Ties, T. D. , Sinclair, M. (1982). Atlantic herrng: stock discreteness and abundance. Science 215: 627-633Jensen, A. L. (1986). Functional regression and correlation analysis. Can. J.Fish. Aquat. Sci. 43: 1742-1745Kane, J. (1984). The feeding habits of co-occurring cod and haddock larvae fromGeorges Bank. Mar. Ecol. Prog. Ser. 16: 9-20Keeley, J. R. (1982a). Temperature and salinity anomalies along the Flemish Capsection in the 1970’s. NAFO Sci. Coun. Studies 5: 87-94Keeley, J. R. (1982b). Regional differences in water types on the Flemish Cap.NAFO SCR Doc. No. 42, Serial No. N531, 8 p.Kenchington, T. J. (1991). Vertical distribution and movements of larval redfishes (Sebastes spp.) in the southern Gulf of St. Lawrence. J. Northw. Atl.Fish. Sci. 11: 43-49Kendall, A. W., Jr. , Clarke, M. E., Yokiavich, M. M., Boehlert, G. W. (1987).Distribution, feeding, and growth of larval walleye pollock, Therag’ra chalcogramma,from Shelikof Strait, Gulf of Alaska. Fish. Bull. 85: 499-521139Klumpp, D. W. , Westernhagen, H. von (1986). Nitrogen balance in marine fishlarvae: influence of developmental stage and prey density. Mar. Biol. 93:189-199Kiorboe, T. , Munk, P., Richardson, K., Christenson, V., Paulsen, H. (1988).Plankton dynamics and larval herring growth, drift and survival in a frontalarea. Mar. Ecol. Prog. Ser. 44: 205-219Konstantinov, K. G. , Turuk, T. N., Plekhanova, N. V. (1985). Food links ofsome fishes and invertebrates on Flemish Cap. NAFO Sci. Coun. Studies 8:39-48Krause, M. , Trahms, J. (1982). Vertical distribution of copepods (all developmental stages) and other zooplankton during spring bloom in the Fladen Groundarea of the North Sea. Neth. J. Sea Res. 16: 217-230Krause, M. , Trahms, J. (1983). Zooplankton dynamics during FLEX’76. In: C.Sinderman and J. Lenz (eds.) North Sea Dynamics, Springer-Verlag, Berlin, p.632-661Kruse, G. H. , Tyler, A. V. (1989). Exploratory simulation of English sole recruitment mechanisms. Trans. Am. Fish. Soc. 118: 101-118Kudlo, B. P., , V. A. Borovkov, and N. G. Sapronetsdaya. (1984). Water circulation patterns on Flemish Cap from observations in 1977-82. NAFO Sci Coun.Studies 7: 27-37Landry, M. R. (1983). The development of marine calanoid copepods with comment on the isochronal rule. Limnol. Oceanogr. 28: 614-624140Lasker, R. (1975). Field criteria for survival of anchovy larvae: the relation between inshore chlorophyll maximum layers and successful first-feeding. Fish.Bull. 73: 453-462Lasker, R. (1981). Factors contributing to variable recruitment of the NorthernAnchovy (Engraulis mordax) in the California Current: contrasting years, 1975through 1978. Rapp. P.-v. Reun. Cons. Jut. Explor. Mer 178: 375-388Last, J. M. (1978). The food of four species of pleuronectiform larvae in the Eastern English Channel and Southern North Sea. Mar. Biol. 45: 359-368Laurence, G. C. (1977). A bioenergetic model for the analysis of feeding andsurvivfa potential of winter flounder, Pseudopleu’ronectes americamus, larvaeduring the period from hatching to metamorphosis. Fish. Bull. 75: 529-546Laurence, G. C. (1978). Comparative growth, respiration and delayed feedingabilities of larval cod (Gadus morhua) and haddock (Melamog’rammus aeglefinus) as influenced by temperature during laboratory studies. Mar. Biol. 50:1-7Lawson, T. J. , Grice, G. D. (1970). The developmental stages of Centropagestypicus Kroyer (Copepoda, Calanoida). Crustaceana 18: 187-209Lear, W. H. , Wells, R., Templeman, W. (1981.) Variation in vertebral averagesfor year-classes of Atlantic cod, Gadus morhua, on Flemish Cap. J. Northw.Atl. Fish. Sci. 2: 57-60Lechowicz, M. J. (1982). The sampling characteristics of electivity indices. Oecologia 52: 22-30141Leggett, W. C. (1986). The dependence of fish larval survival on food and predator densities InS. Skreslet (ed.) The role of freshwater outflow in coastal marineecosystems Springer-Verlag, Berlin, p. 117-137Leggett, W. C. , Frank, K. T., Carscadden, J. E. (1986). Meteorological andhydrographic regulation of year-class strength in capelin (Mallotus villosus).Can. J. Fish. Aquat. Sci. 41: 1193-1201Lilly, C. R. (1987). Synopsis of research related to recruitment of Atlantic cod(Gadus morhua) and Atlantic Redfishes (Sebastes sp.) on Flemish Cap. NAFOSci. Coun. Studies 11: 109-122Lilly, C. R. (1980). Year-class strength of redfish and growth of cod on FlemishCap. ICNAF Sel. Papers 6: 35-39Lilly, G. R. , Gavaris, C. A. (1982). Distribution and year-class strength of juvenile redfish, Sebastes sp., on Flemish Cap in the winters of 1978-82. J. Northw.Atl. Fish. Sci. 3: 115-222Loder, J. W., C. K. Ross, and P. C. Smith. (1988). A space- and time-scale characterization of circulation and mixing over submarine banks, with applicationto the northwestern Atlantic continental shelf. Can. J. Fish. Aquat. Sci. 45:1860-1885Marak, R. R. (1974). Food and feeding of larval redfish in the Gulf of Maine. InJ. H. S. Blaxter (ed.) The early life history of fish, Springer-Verlag, Berlin, p.267-275142May, R. C. (1974). Larval mortality in marine fishes and the critical period concept. In J. H. S. Blaxter (ed.) The early life history of fish, Springer-Verlag,Berlin, p. 3-19McGurk, M. D. (1985). The role of starvation in the population dynamics oflarval Pacific herring, Clupea harengus palasi. PhD Thesis, Univ. BritishColumbiaMcGurk, M. D. (1986). Natural mortality of marine pelagic fish eggs and larvae:role of spatial patchiness. Mar. Ecol. Prog. Ser. 34: 227-242McGurk, M. D. (1989). Advection, diffusin and mortality of Pacific herring larvaeClupea harengus pallasi in Bamfield Inlet, British Columbia. Mar. Ecol. Prog.Ser. 51: 1-18McLaren, I. A. (1978). Generation lengths of some temperate marine copepods:estimation, prediction, and implications. J. Fish. Res. Board Can. 35: 1330-1342McLaren, I. A. , Corkett, C. J. (1986). Life cycles and production of two copepodson the Scotian Shelf, eastern Canada. In: G. Schriever, H.K. Schminke andC.-.t Shih (eds.) Proc. of 2nd Intl. Conf. on Copepoda, Syllogeus 58. pp.362-368M6ller, H. (1979). Significance of coelenterates in relation to other plankton organisms. Meeresforsch 27: 1-18Monteleone, D. M. , Peterson, W. T. (1986). Feeding ecology of American sandlance Ammodytes americanus larvae from Long Island Sound. Mar. Ecol. Prog.143Ser. 30: 133-143Murphy, J. A. , Cohen, R. E. (1978). A guide to the developmental stages ofcommon coastal, Georges Bank and Gulf of Maine copepods. NMFS Lab. Ref.No. 78-53, 56 p.Myers, R. A. , Drinkwater, K. (1989a). The influence of Gulf Stream warm corerings on recruitment of fish n the northwest Atlantic. J. Mar. Res. 47: 635-656Myers, R. A. , Drinkwater, K. (1989b). Offshelf Ekman transport and larval fishsurvival in the northwest Atlantic. Biol. Oceanog. 6: 45-64Nelson, W. R. , Ingham, M. C., Schaaf, W. E. (1977). Larval transport and year-class strength of Atlantic menhaden, Brevoortia tyrammus. Fish. Bull. 75:23-41Newell, G. E. , Newell, R. C. (1977). Marine plankton: a practical guide. Hutchinson, London, 244 p.Ni, 1.-H. (1982). Meristic variation in beaked redfishes, Sebastes marimus, compared to beaked redfishes of the Northwest Atlantic. Can. J. Fish. Aquat. Sci.39: 1664-1685Ni, 1.-H. , iVicKone, W.D. (1983). Distribution and concentration of redfishes inNewfoundland and Labrador waters. NAFO Sci. Coun. Stud. 6: 7-14O’Boyle, R. N. , Sinclair, M., Conover, R. J., Mann, K. H., Kohler, A. C. (1984).Temporal and spatial distribution of ichthyoplankton communities of the Scotian Shelf in relation to biological, hydrological, and physiographic features.Rapp. P.-v. Re un. Cons. i. Explor. Mer 183: 27-40144O’Connell, C. P. (1980). Percentage of starving northern anchovy, Engraulis moardx,larvae in the sea as estimated by histologici methods. Fish. Bull. 78: 475-489Ogilvie, H. S. (1956). Copepod nauphi (I). Cons. Tnt. Explor. Mer, Zoopl. Sh.50, ‘Ip.Omori, M. , Ikeda, T. (1984). Methods in marine zooplankton ecology. JohnWiley & Sons, New YorkParrish, R. H. , Nelson, C. S., Bakun, A. (1981). Transport mechanism and reproductive success of fishes in the California Current. Biol. Oceanogr. 1:175-203Pearre, S. Jr. (1980). The copepod width—weight relation and its utility in foodchain research. Can. J. Zool. 58: 1884-1891Pearre, S. Jr. (1982). Estimating prey preference by predators: uses of variousindices, and a proposal of another based on x2 Can. J. Fish. Aquat. Sci. 39:914-923Penney, R. (1984). Age, growth, and morphology of larval redfish, Se bastes sp.(Pisces: Scorpaenidae) on Flemish Cap, 1980-81. Msc. Thesis, Memorial University of Newfoundlan, 179 p.Penney, R. (1985). Development of deep-water redfish(Sebastes mentella) larvae,with comparative notes on newly extruded Acadian redfish (Sebastes fasciatus)from the Flemish Cap. Can. J. Zool. 65: 1167-1180Penney, R. W. (1987). Comparative morphology of preextrusion larvaeof thenorth Atlantic sharp-beaked redfishes, Sebastes mentella and Se bastes fasciatus145(Pisces: Scorpaenidae). Can. J. Zool. 63: 1181-1188Penney, R. W. , Anderson, J. T. (1981). Otolith analysis of age and growth oflarval redfish (Sebastes sp.) on Flemish Cap, 1979 and 1980. NAFO SCR Doc.81/IX/117, Ser. No. N423, 15 p.Penney, R. W. , Power, D. J., Atkinson, D. B. (1984). Species proportion andsome reproduction aspects of three redfish species (Se bastes marinus, S. fasciatus and S. mentella) on the Flemish Cap, March 1983. NAFO SCR Doc.84/VI/22, Ser. No. N797Penney, R. W. , Evans, G. T. (1985). Growth histories of larval redfish (Sebastesspp.) on an offshore Atlantic fishing bank determined by otolith increnièntanalysis. Can. J. Fish. Aquat. Sci. 42: 1452-1464Perry, R.I. , Neilson, J. D. (1988). Vertical distributions and trophic interactionsof age-0 Atlantic cod and haddock in mixed and stratified waters of GeorgesBank. Mar. Ecol. Prog. Ser. 49: 199-214Peterman, R. M. , Bradford, M. J. (1987). Wind speed and mortality rate of amarine fish, thenorthern anchovy (Engraulis mordax). Science 235: 354-356Peterman, R. M. , Bradford, M. J., Lo, N. C. H., Methot, R. D. (1988). Contribution of early life stages to interannual variability in recruitment of northernanchovy (Engraulis mordax). Can. J. Fish. Aquat. Sci. 45: 8-16Peterson, W. T. , and Ausubel, S. J. (1984). Diets and selective feeding by larvaeof Atlantic mackerel Scomber scombrus on zooplankton. Mar. Ecol. Prog. Ser.17: 65-75146Peterson, I. , Wroblewski, J. S. (1984). Mortality rate of fishes in the pelagicecosystem. Can. J. Fish. Aquat. Sci. 41: 1117-1120Power, D. , Atkinson, D. B. (1989). Status of redfish in NAFO Division 3M.NAFO SCR Doc. 89/56, Ser. No. N1636, 12 p.Pyke, G. H. (1984). Optimal foraging theory: a critical review. Ann. Rev. Ecol.Syst. 15: 523-575Rauck, G. , Zijistra, J. J. (1978). On the nursery aspects of the Waden Sea forsome commercial fish species and possible long-term changes. Rapp. P.-v. Reun. Cons. in Explor. Mer 172: 266-275Ricker, W. E. (1973). Linear regression in fishery research. J. Fish. Res. BoardCan. 30: 409-434Ricker, W. E. (1975). A note concerning Professor Jolicoeur’s comments. J. Fish.Res. Board Can. 32: 1494-1498Ricker, W. E. (1979). Growth rates and models. In W. S. Hoar, D. J. Randall, J.R. Brett (eds.) Fish Physiology Vol. VIII, Academic Press, New York, London,p. 279-352Ross, C. K. (1980). Moored current meter data from Flemish Cap, January-July1979. NAFO SCR Doc., No. 128, Serial No. N200, 13 p.Ross, C. K. (1981). Drift of satellite-tracked buoys on Flemish Cap, 1979-80.NAFO Sci. Coun. Studies 1: 47-50Runge, J. A. (1988). Should we expect a relationship between primary production147and fisheries? The role of copepod dynamics as a filter of trophic variability.Hydrobiol. 167/168: 61-71Ryland, J. S. (1964). The feeding of plaice and sandeel larvae in the southernNorth Sea. J. mar. biol. Ass. U.K. 44: 343-364SAS. (1985). SAS User’s Guide: Statistics Version 5 Edition. Statistical AnalysisInstitute Inc., Cary, North CarolinaShelbourne, J. E. (1962). A predator-prey size relationship for plaice larvae feeding on Oikopleura. J. mar. biol. Ass. U.K. 42: 243-252Shephard, J. G. , Cushing, D. H. (1980). A mechanism for density dependentsurvival of larval fish as the basis of a stock-recruitment relationship. J. Cons.39: 160-167Shirota, A. (1970). Studies on the mouth size of fish in the larval and fry stages.Bull Jap. Soc. Sci. Fish. 36: 353-368. (Fish. Res. Board Can. Transi. Ser.No. 1978)Siegel, S. (1956). Non-parametric statistics for the behavioural sciences, McGrawHill, TorontoSissenwine, M. P. (1984). Why do fish populations vary? In R. M. May (ed.)Exploitation of marine communities, Springer-Verlag, Berlin, p. 59-94Smith P. E. , Richardson, S. L. (1977). Standard techniques for pelagic fish eggand larval surveys. FAO Fish. Tech. Pap. No. 175Sokal R. R. , Rohif, F. J. (1969). Biometry, W. H. Freeman and Co. San Francisco148Sundby, S. , Fossurn, P. (1990). Feeding conditions of Arcto-norwegian cod larvaecompared with the Rothschild-Osborn theory on small-scale turbubence andplankton contact rates. J. Plank. Res. 12: 1153-1162Sundby, S. , Bjorke, H., Soldal, A. V., Olsen, S. (1989). Mortality rates during theearly life stages and year-class strength of northeast Arctic cod (Gadus morhusL.). Rapp. P.-v. Re un. Cons. in Explor. Mer 191: 351-358Sverdrup, H. U. (1953). On the condition for the vernal blooming of phytoplankton. J. Cons. mt. Explor. Mer. 18: 287-259Sysoeva, T. K. , Degtyareva, A. A. (1965). The relation between the feeding ofcod larvae and pelagic fry and the distribution and abundance of their principalfood organisms. ICNAF, Spec. Pubi. 6: 411-416Taggart, C. T. , Leggett, W. C. (1987). Short-term mortality in post-emergentlarval capelin Mallotus villosus. II. Importance of food and predator density,and density-dependence. Mar. Ecol. Prog. Ser. 41: 219-229Templeman, W. T. (1972). Year-class success in some North Atlantic stocks ofcod and haddock. ICNAF Spec. Pubi. 8: 223-238Templeman, W. T. (1976). Biological and oceanographic background of FlemishCap as an area for research on the ressons for year-class success and failure incod and redfish. ICNAF Res. Bull. 12: 91-117Theilaker, G. H. (1986). Starvation-induced mortality of young sea-caught jackmackerel, Trachu’rus symmetricus, determined with histological and morphological methods. Fish. Bull. 84: 1-17149Theilaker, G. , Dorsey, K. (1980). Larval fish diversity, a summary of laboratoryand field research. TOC Workshop Rep. No. 28, P. 105-142Thompson, A. B. , Harrop, R. T. (1991). Feeding dynamics of fish larvae onCopepoda in the western Irish Sea, with particular reference to cod Gadusmo’rhua. Mar. Ecol. Prog. Ser. 68: 213-223Thompson, K. R. , Loucks, R. H., Trites, R. W. (1988). Sea surface temperaturevariability in the shelf-slope region of the northwest Atlantic. Atmos.-Ocean26: 282-299Tremblay, M. J. (1981). Zooplankton community structure, biomass and production of the Scotian Shelf, with special reference to Emerald Bank. M.Sc. Thesis,University of GuelphTremblay, M. J. , Roff, J. C. (1983). Production estimates for Scotian Shelf cope-pods based on mass specific P/B ratios. Can. J. Fish. Aquat. Sci. 40: 749-753Van Guelpen, L. , Markie, D. F., Duggan, D. J. (1982). An evaluation of accuracy, precision, and speed of several zooplankton subsampling techniques. J.Cons. mt. Explor Mer. 40: 226-236Vascov, A. , Rudneva, G., Galuzo, A. (1987). Estimation of stock size and totalallowable catch of beaked redfish in Division 3LN and 3M for 1988. NAFOSCR Doc. 87/20, Ser. No. N1300, 10 p.Veer, H. W. van der (1985). Impact of coelenterate predation on larval plaicePleu’ronectes platessa and flounder Platichthys flesus stock in the western Wadden Sea. Mar. Ecol. Frog. Ser. 25: 229-238150Veer, H. W. van der (1986). Immigration, settlement, and density-dependentmortality of a larval and early postlarval 0-group plaice (Pleuronectes platessa)papulation in the western Wadden Sea. Mar. Ecol. Prog. Ser. 29: 223-236Vidal, J. (1980). Physioecology of zooplankton. II. Effects of phytoplankton concentration, temperature, and body size on the development and molting ratesof Calanus pacificus and Pseudocalanus sp. Mar. Biol. 56: 135-146Vladimirskaya, E. V. (1967). Seasonal population dynamics of Calanus finma’rchi(Gunner) in the northwestern Atlantic, 1958-61. ICNAF Res. Bull. 4:41-52Ware, D. M. (1975). Relation between egg size, growth and natural mortality oflarval fish. J. Fish. Res. Board Can. 32: 2503-2512Ware, D. M. , Lambert, T. C. (1985). Early life history of Atlantic mackerel(Scomber scombrus) in the southern Gulf of St. Lawrence. Can. J. Fish.Aquat. Sci. 42: 577-592Webster, I. , Anderson, J. T. (1988). A model for simulating the towing performance of the bongo sampler. Deep-Sea Res. 34: 1277-1291Werner, E. E. , Gilliam, J. F. (1984). The ontogenetic niche and species interactions in size-structured populations Annu. Rev. Ecol. Syst. 15: 393-425Wieser, W. , Forstner, H., Schiemer, F., Mark, W. (1988). Growth rates andgrowth efficiencies in larvae and juveniles of Rutilus rutilus and other cyprinidspecies: effects of temperature and food in the laboratory and in the field. Can.J. Fish. Aquat. Sci. 45: 943-950151Williams, R. , Conway, C. V. P., Coffins, N. R. (1987). Vertical distributions ofeggs, nauplii and copepodites of Calanus helgolandicus (Copepoda: Crustacea)in the Celtic Sea. Mar. Biol. 96: 247-252Williams, R. , Conway, D. V. P. (1988). Vertical distribution and seasonal numerical abundance of the Calanidae in oceanic waters to the south-west of theBritish Isles. Hydro. 167/168: 259-266Williams, R. , Lindley, J. A. (1980). Plankton of the Fladen Ground during FLEX76 III. Vertical distribution, population dynamics and production of Calanusfinmarchicus (Crustacea: Copepoda). Mar. Biol. 60: 47-56- Wilson, C. B. (193). The copepods of theWoods Hole region. SmithsonianInstit. U.S. Nati. Museum Bull. 158. U. S. Govern. Printing Office (1972 ed.),635 p.Ziljstra, J. , Dapper, R., Witte, J. (1982). Settlement, growth and mortality ofpost-larval plaice (Fleuronectes platessa) in the western Wadden Sea. Neth. J.Sea Res. 15: 250-272152A AppendixComparison of Catch Rate Differences Among Small and LargeSamplers for Calanus finmarchicus Copepodite StagesThe purpose of this Appendix is to present results of an analysis of catch ratedifferences for C. fimmarchicus copepodites sampled by small (20 cm, 0.165 mmmesh nets) and large (61 cm, 0.333 mm and 0.505 mm mesh nets) bongo samplersused in this study. The null hypothesis tested was that there were no differences incatch rate between samplers for each stage.Catchrate data for 0165ffim and 0.333inmnmesh nètswere compared for sixcruises on Flemish Cap (47°N, 45°W) in which the two mesh sizes were sampledsimultaneously on 20 cm and 61 cm bongo samplers, respectively (Table Al). Inone cruise 31 tows were also done using 0.505 mm mesh nets in place of the 0.333 mmmesh nets. The mean catch rate (numbers m3) was calculated for each copepoditestage (CI-CVI) in each mesh size for all six cruises.The numbers of zooplankton caught by each tow within a cruise often exhibitstrong positive skewness and the assumptions needed to validate the conventionalpaired-t test for the comparison of the catches by mesh size, for each copepoditestage, appeared to be invalid. Therefore, I used the Wilcoxon matched-pairs signedranks test (Siegel 1956), which is not dependent on any distributional assumptions.It was suspected, a priori, that the small sampler would retain more of the smallercopepodite stages than the large sampler, but that the reverse would apply. Therefore, one-sided tests were used. The difficulty that arises with the intermediate153copepodite stages was resolved by testing against the reverse alternative if the nullhypothesis was initially accepted.I considered that each cruises was an independent test of the same null hypothesis.I followed the method of Fisher (1932), which utilizes the geometric mean of the individual p-values Under the assumption of no difference, the individual p-values (pi)should be uniformally distributed on (0,1) whence —2 log(p) would be distributedas a x2 on 2n degrees of freedom, where here n is the number of cruises (n=6). (W.Warren, Northwest Atlantic Fisheries Center, St. John’s, Newfoundland, personalcommunication).Analyses of catch rate differences between small and large (0.333 mm mesh nets)clearly demonstrated that Calanus finmarchicus copepodite stages CI and CII wereundersampled by the large sampler, while stages CV and CVI were undersampled bythe small sampler. These results are apparent from the individual tests of differenceswithin each cruise (Table A2) and from Fisher’s test for all cruises (Table A3). Mysamples were processed by a different laboratory in 1980 than in 1981 and 1982.It appears that staging of CI and CII copepodites for 0.333 mm mesh nets maynot have been carried out in all cases. This is suggested by zero catches in fourcruises for stage CI and three cruises for stage CII (Table A2). Therefore, catchrate differences for stage CI and CII for these cruises are not valid. Nevertheless,the differences observed for two cruises for stage CI and three cruises for stage CIIare convincing (Table A2).Catch rate differences for stage CIII demonstrated no significant difference amongthe small and large samplers. This was true for six cruises where probabilitiesassociated with rejecting the null hypothesis ranged from 0.13 14 - 0.4974 (Table A2).154There was some variability in results for different cruises as GADO5O data indicatedthat the small sampler caught almost twice as many stage CIII copepodites as thelarge sampler. The mean catch rate differences were not a result of an extremevalue. Indeed, measures of skewness for these samples were lower compared to thatsampled for other cruises (Table A2). The small sample size for this cruise mayaccount for the observed difference which would tend to cancel out as more sampleswere included in the calculation of mean density.Results for stage CIV are less clear. Overall I conclude there is a measureabledifference in catch rates—the large sampler caught more stage CIV copepodites thanthe small sampler. This conclusion is based on the results of the Fisher’s overall testwhich demonstrated that we would reject the null hypothesis and on the observationthat in all cases the mean catches of 0.333 mm mesh nets were greater. For eachcruise the differences between the means were not large, suggesting that the smallermesh nets representatively caught stage CIV copepodites.There is some variability in the analyses of catch rate differences among cruisesfor most stages. For example, stage CII had one cruise in which the probabilityassociated with rejecting the null hypothesis was 0.1906, compared to probabilitiesof 0.0065 and 0.0081 for the other two cruises. Similar results were obtained forother stages (Table A2). In each case the Fisher’s test demonstrated that the overallresult was in favour of rejecting the null hypothesis. The variability among cruisesis probably due to the skewed nature of the data where a few large catches alwaysoccurred (Table A3).Comparison of catch rates between 0.165 and 0.505 mm mesh nets (small andlarge samplers, respectively) demonstrated that stages CI - CIII were undersampled155by the 0.165 mm mesh nets while stages CV and CVI were undersampled by the0.505 mesh nets (Table A4). As with the 0.165:0.333 mm mesh comparison it appearsstages CI and CII were not staged for the 0.505 mm mesh samples. Therefore,there is no direct comparison available for these data. However, given the highlysignificant differences for the 0.165:0.333 mm mesh comparisons we conclude thesame result applies here. There was no significant difference in catch rates forstage CIV at P=0.1588. It might be argued that more samples may demonstrate adifference in catch rates for stage CIV based on comparisons presented in Table A2where individual probability values higher than this were nevertheless considered tocontribute to an overall conclusion of significant differences in catch rate between0.165 and 0.333 mm mesh nets.156Table Al. Summary of cruises and samples processed for analyses of catchrate differences in Calanus finmarchicus copepodite stages on Flemish Cap.Number of samples61 cm Bongo 20 cm BongoCruise no. Date 0.505 mm 0.333 mm 0.165 mmATC331 1—3 Aug. 1982— 21 21GADO35 7—14 Apr. 1980— 40 39GADO37 21—31 May 1980— 56 41CADO5O 2—9 May 1981 31 10 41GADO51 22—27 May 1981—- 42-- 41HAWPAN 26—30 June 1981- 42 44157158Table A2. nalysis of catch rate differences between 0.165 mm and 0.333 nun mesh netsfor different copepodite stages of Calanus finmarchicus sampled during six cruises.The conventional maximum value of Wilcoxon’t T is given in brackets. For CI—Cill thealternative test is catch 0.333 > catch 0.165 while for CIV—CVI the alternative iscatch 0.165 > catch 0.333.0.165 mm Mesh 0.333 nun MeshProb.Stage N Mean Skewness Mean Skewness Wilcoxon T (one sided)Stage CIATC331 10 3.6143 1.075 0— 0 (27) 1/1024 0.00098GD035 12 0.9168 2.466 0.0162 3.464 0 (33) 6/4096 = 0.00146G?D037 35 19.3426 5.091 0.8146 1.969 29 (315) 0.0000014GADO5O 8 93.7625 2.797 0— 0 (18) 1/256 = 0.0039G1D05l 15 10.9947 2.030 0— 0 (60) 1/32768 0.00003I4AWPAN 7 16.1184 1.362 0— 0 (14) 1/128 = 0.0078Stage CIIATC331 8 2.5910 2.558 0— 0 (18) 1/256 = 0.0039GADO35 17 0.3914 2.929 0.1694 0.818 58 (76) 0.1906GADO37 41 28.2278 2.419 12.1437 2.946 245 (430) 0.0081GADO5O 9 106.8197 1.508 0— 0 (22) 1/512 0.00195GADO51 28 8.0821 2.593 4.1343 2.148 94 (203) 0.0065HWPJ3N 6 8.2673 1.983 0— 0 (10) 1/64 0.0156Stage CIII-ATC331 20 5.4699 3.123 6.4256 2.337 75* (105) 0.1314GA0035 29 0.2500 4.293 0.2357 3.312 171* (217) 0.1573G?D037 41 42.6613 1.227 49.8164 4.168 422* (430) 0.4561GZsDO5O 9 142.0682 0.347 75.5482 1.489 9 (22) 33/512 = 0.06445G)D05l 41 44.2226 2.384 46.6323 2.315 430 (430) 0.4974HAWPN 30 3.1658 1.686 2.6500 2.108 211* (232) 0.3292Stage CIVATC331 21 73.7041 1.472 85.5424 0.983 93 (115) 0.2171GADO35 35 1.1327 2.222 1.3856 1.715 189 (315) 0.0196G?D037 41 26.0747 1.950 31.1729 2.492 395 (430) 0.3228GAD050 9 47.3430 1.852 104.8872 0.726 8 (22) 25/512 = 0.0488GAIDO51 41 167.3330 1.495 199.4791 1.210 254 (430) 0.0111HAWPAN 39 14.7989 2.003 15.3721 2.329 351 (390) 0.2931Stage CVATC331 21 326.4441 3.074 258.4392 1.739 86**(115) 0.1526G1D035 36 1.0756 2.373 1.0616 1.616 250 (333) 0.0961GADO37 41 3.6157 2.757 4.4884 1.633 292 (430) 0.0363G?D050 9 3.9643 1.672 9.5499 0.808 7 (22) 19/512 = 0.0371GADO51 41 83.4466 1.397 103.6961 1.514 203 (430) 0.0016T1AWPAN 40 89.5095 1.570 98.1962 1.674 347 (410) 0.1986Stage CVIATC331 22 5.1645 0.790 6.9870 1.354 76 (126) 0.0849GADO35 37 40.6106 1.565 46.2756 1.536 211 (351) 0.0170G?D037 41 18.2187 0.978 24.9437 1.064 283 (430) 0.0300GADO5O 9 11.7379 2.378 18.1867 0.813 11 (22) 52/512 = 0.1016G?D05l 41 20.5782 2.302 23.1940 0.657 212 (430) 0.0023H?WPJN 40 11.3607 1.944 11.8662 1.281 351 (410) 0.2139* The Wjlcoxon statistic is conventionally taken as the smaller of T and N (N+1)/2—T. For aone—side test one should, however, be concerned with probability of obtaining a value lessthan or equal to T where the range of T is (0, N(N-i-l)/2). Thus, for example, for CIII,ATC331 the T of 75 should be interpreted as 210—75 = 135 with a p—value against thealternative of catch 333> catch 165 of 1.0—0.1314 = 0.8686.** Conversely, for CV, ATC331, the T of 86 should be interpreted as 231—86 = 145 with ap—value against the alternative of catch 165 > catch 333 of 1.0—0.526 = 0.8474.Table A3. Summary of Fisher’s combined probablilities test for all cruises combined for each copepodite stage of Calamus finma’rchicus. CI - refers to copepoditestage I, and so on. x2 - refers to the test statistic value. P - refers to the probabilitylevel.Statistic CI CII CIII CIV CV CVIx2 99.04 58.89 13.06 30.69 34.34 39.90P 8.6x10’ 1.9x107 0.3647 0.0022 0.0006 0.00011590TableA4.Analysisofcatchratedifferencesbetween0.165mmand0.505mmmeshnetsfordifferentcopepoditestagesofCalanusfinmarchicussampledduringonecruise(GAD050).TheconventionalmaxitflumvalueofWilcoxon’sTIsgiveninbrackets.ForCI—Clilthealternatetestiscatch0.505>catch0.165whileforCIV—CVIthealternateiscatch0.165>catch0.333.0.165mmMesh0.333mmMeshProb.StageNMeanSkewnessMeanSkewnessWilcoxonT(onesided)CI2572.20443.5280—0(162)1/33554432=2.98x10CII30134.52812.9610—0(232)1/1.037Kb=9.31x10CIII31161.19032.24765.47322.74987(248).0000CIV31104.14081.410104.91540.999197(248).1588CV3112.93131.48815.66750.760157(248).0373CVI3111.86761.10621.60211.79662(248).00013B AppendixSummary of Zooplankton Weights Estimated for Redfish StomachContents CalculationsTo estimate the weight of prey items eaten by redfish it was necessary to calculatethe expected weight for each prey item, multiplied by its numerical occurrence inthe diet. In most cases this was based on published estimates from previous studies.However, the weight of many zooplankton is calculated based the equality that 1 cm3is equivalent to 1 g wet weight, with conversion to dry weight based on the factor 0.1.Weight estimated in thi way is approxiniate but shoukfrêfiect the relative contribu-tion of these items to the diet, based on weight. In some cases weight was estimatedin the laboratory. These estimates were largely done in 1980 by technicians employedby Marine Research Associates, St. Andrew’s, N. B. (MRA). For copepods, weight isoften estimated based on empirically derived power equations (Pearre 1980). However, when I compared predicted to measured weight I discovered that there wassignificant disagreement among different estimates and that all published equationsunderestimated the older copepodite stages, for Calamus fimmarchicus and Oithomasimilis. Based on available stage-specific weight data I fit power curves for Calanusfinmarchicus and Oithona similis, the two predominant species on Flemish Cap.These equations are: DW = 0.0151TL2975 and DW = 0.013TL2’74,respectively.My estimates for C. fimmarchicus only predict up to stage CIV with any degree ofaccuracy. Therefore, I only used estimates from power equations to estimate theearlier copepodite stages. Weights estimated for the later stages were based on the161published studies where dry weight for each stage was determined separatley. For0. similis the power equation did not accurately estimate the increase in weightwith stage. However, there are few data available for Oithona, and where data werelacking for specific stages I used my predicted estimate of weight.A complete summary of weights estimated for prey species and stages is listed inTable Bi. Tables B2 and B3 list the taxonomic names associated with the speciesand stage codes used in Table Bi.162Table Bi: Summary of zooplankton prey biomass estimates.PREY STAGE BIOMASS REFERENCE(mg wet)BACCLA 700 0.00002 MRABACORD 700 0.00005 Same as BIDORD 700BIDORD 700 0.00005 NRA coding sheetsBIVCLA 500 0.01 MBACALASP 400 0.00357 Same as CALSBO 420CALASP 405 0.0395 Mean of CALFIN 411 & 412CALASP 411 0.02550 Same as CALFIN 411CALASP 412 0.05355 Same as CALFIN 412CALASP 413 0.1326 Tremblay (1981) for C. fin.CALASP 414 0.41225 “ “ “CALASP 415 1.17810 “ “ “CALASP 700 0.02550 Assumed wgt. for C. fin. 411CALFIN 200 0.0014 Bainbridge and McKay (1968)CALFIN 400 0.13260 Same as CALFIN 413CALFIN 404 0.79518 Mean of CALFIN 414 & 415CALFIN 411 0.02550 Tremblay (1981)CALFIN 4J,2. 0.05355CALFIN 413 0.13260CALFIN 414 0.41225CALFIN 415 1.17810CALFIN 416 0.075 Bainbridge & McKay (1968)CALFIN 417 1.80200 Tremblay (1981)CALFIN 700 0.1326 Same as CALFIN 413CALGLA 411 0.05417 JTA, estimate for C. fin.CALGLA 412 0.16820CALGLA 413 0.5432 Tremblay (1981)CALGLA 414 1.6686CALGLA 415 4.2781CALGLA 416 9.350 “CALGLA 417 9.350 Tremblay (1981), for CVICALHYP 413 1.9559 Tremblay (1981)CALHYP 414 1.9652CALHYP 415 34.0255CALHYP 416 81.2422CALSBO 200 0.0014 Bairibridge and McKay (1968)CALSBO 351 0.000453 JTA, estimated for C. fin.CALSBO 352 0.00357 IICALSBO 353 0.0163CALSBO 405 0.0395 Mean of CALFIN 411—412 (Tremblay 1981)CALSBO 411 0.005 B&M (1968), small copepods (*)CALSBO 412 0.005 “ “ “CALSBO 413 0.050 “ e I?CALSBO 414 0.050 “ “163Table Bi: Continued...PREYSP PREYSTG BIOMASS REFERENCE(mg wet)CALSBO 415 0.075 B&M (1968), small copepodsCALSBO 417 0.075 “ “ ItCALSBO 420 0.00357 JTA, estimated for C. fin.CALSBO 421 0.04435 “CALSBO 422 0.12818 “CALSBO 500 0.045 B&M (C. fin. Cl—Cu)CALSBO 700 0.045 It Il ItCHRDIV 700 0.00002 Same as BACCLACONCSP 700 2.0945 1/2 of Volume where diameter=0.2 mmCOPSBC 200 0.0014 B&MCOPSBC 300 0.01 B&MCOPSBC 351 0.000453 Same as CALSBO 351COPSBC 400 0.1326 Same as CALFIN 400COPSBC 700 0.045 Same as CALSBO 700COSCIN 700 0.00005 MRA - used their value for BIDORDCRUCLA 500 0.01 Same as CRUCLA -CRUCLA 700 0.01 Same as COPSBC 300CYCSBO 200 0.00006 Sazhina (1980) for Oithona similisCYCSBO 351 0.00085 Moller (1979)CYCSBO 352 0.00807 JTA, estimated for 0. sun.CYCSBO 420 0.00807 “ “ “CYCSBO 421 0.01360 Tremblay (1981), 0. sim. CI—CVCYCSBO 700 0.01360 Same as CYCSBO 421DINCLA 700 0.00002 MRAEUCHSP 413 0.7655 Brattelid & Matthews (1978)EIJPORD 560 0.80 or 0.50 LGL Coding Sheets — both are correctFRITIL 700 0.0218 JTA, estimated volume as a cylinderGLOBIG 700 0.005 Same as GLOBSP 700GLOBSP 700 0.005 Same as for GRASBCGRASBC 700 0.005 MRAISOORD 700 0.5 Same as Furcilias 2—3 mm (B&M 1968)LIMASP 500 0.001 MRA Coding SheetsLIMASP 590 (not listed) Estimated volume as a sphereMESTEN 413 0.0187 Same as PSEMIN 413MICPYG 417 0.0466 Same as PSEMIN 417MICPYG 700 0.0155 Assumed 1/3 of Stage 417OIKOSP 700 0.0218 Same as FRITIL 700OITATL 400 0.01360 Same as OITSIM 400OITATL 417 0.01785 Same as OITSIM 417OITATL 418 0.01785 Same as OITSIM 418OITATL 421 0.01360 Same as OITSIM 421OITHSP 200 0.00006 Sazhina (1980)164Table Bi: Cont’d...PREYSP PREYSTG BIOMASS REFERENCE(mg wet)OITHSP 351 0.00085 Moller (1979)OITHSP 352 0.00807 JTA, estimated for 0. sim.OITHSP 400 0.007 B&MOITHSP 413 0.01360 Same as OITSIM 413OITHSP 414 0.01360 Same as OITSIM 414OITHSP 420 0.00807 JTA, estimated for 0. sim.OITHSP 421 0.01360 Tremblay (1981), 0. sim. CI-CVOITHSP 415 0.01360 Tremblay (1981), 0. sim. CI—CVOITHSP 416 0.01785 Same as OITSIM 416OITHSP 417 0.01785 Same as OITSIM 417OITHSP 418 0.01785 “ “OITHSP 700 0.01360 Same as OITSIM 700OITSIM 400 0.01360 Same as OITSIM 413 (Trernblay 1981)OITSIM 411—415 0.01360 Tremblay (1981)OITSIM 416 0.01785OITSIM 417 0.01385OITSIM 418 0.01785 Same as OTISIM 417OITSIM 420 0.00807 JTA, esitmated for 0. sim.OITSIM 421 0.01360 Tretnblay (1981), 0. sim. CI—CVOITSIM 700 0.01360 “ “ONCASP 200 0.00006 Same as OITSIMONCASP 411 0.01360 “ONCASP 412 0.01360ONCASP 415 0.01360 “ONCASP 416 0.01785 Same as OITSIM 417ONCASP 420 0.00807 Same as OITSIM 420ONCASP 700 0.01360 Same as OITSIM 700ONCBOR 417 0.01785 Same as OITSIM 417ONCBOR 418 0.01785 Same as OITSIM 417ONCBOR 420 0.00807 Same as CYCSBO 420ONCBOR 421 0.01360 Same as CYCSBO 421OSTCLA 200 0.180 0.2 miii egg diameterPARASP 700 0.5 Same as Furcilias 2-3 mm, B&MPARPAR 411 0.0060 LGLPARPAR 416 0.05355 Treunblay (1981)PERIOR 700 0.00006 MRAPSEUSP 413 0.040 LGLPSEMIN 411—413 0.0187 Tremblay (981)PSEMIN 414—415 0.0646PSEMIN 417 0.0816165Table Bi: Cont’d...PREYSP PREYSTG BIOMASS REFERENCE(mg wet)PSEMIN 700 0.0187 Same as PSEMIN 411—413PROPHY 200 0.0014 Same as CALFIN egg weightPROPHY 500 0.0218 Same as FRITILPROPHY 700 0.005 Salues from: .20 .003 .005.090 both MRA & LGL Coding SheetsRADSBC 700 0.001 B&MSCOMIN 417 0.0816 Same as PSEMIN 417THYLON 560 0.8 B&M for Furcilias 3—4 mmTHYSSP 560 1.2 B&M for Furcilias 4—5 mmTHYSSP 700 0.5 Assumed furcilias 2—3 mm or calyptopeslarvae (B&M)TINORD 700 0.00009 Averaged value from B&M(*)— “small copepods”, refers to Pseudocalanus, Temora and CentropagesMRA — Marine Research Associates Ltd., St. Andrew’s, N.B.LGL - LGL Ltd., St. John’s, Newfoundland.166Table B2. Taxonomic names, listed alphabetically, associated with the apiha codesof Table Bi, after Foy and Anderson (1986).Species Code Taxonomic Name Species Code Taxonomic NameBACCLA Bacillariophyceae ISOORD IsopodaBACORD Bacillariales LIMASP Limacina sp.BIDORD Bidduiphiales MESTEN Mesocalanus tenuicornisBIVCLA Bivalvia MICPYG Microcalanus pygmaeusCALASP Calanus sp. OIKOSP Oikopleura sp.CALFIN C. finmarchicus OITATL 0. atlan,ticaCALGLA C. glacialis OITHSP Oithona sp.CALHYP C. hyperboreus OITSIM 0. similisCALSBO Calanoida ONCASP Omcaea sp.CHRDIV Chromophyta ONCBOR 0. borealisCONCSP Conchoecia sp. OSTCLA OsteichthyesCOPSBC Copepoda PARASP Parathemisto sp.COSCIN Coscinodiscus sp. PARPAR Paracalanus parvusCRUCLA Crustacea PERIOR PeridinialesCYCSBO Cyclopoida PROPHY ProtozoaDINCLA Dinophyceae PSEUSP Pseudocalanus sp.EUCHSP Euchaeta sp. PSEMIN P. minutusEUPORD Euphausiacea RADSBC RadiolariaFRITIL Fritillaria sp. SCOMIN Scolecithricella minorGLOBIG Globigerina THYLON Thysanoessa longicaudataGLOBSP Globigerina sp. THYSSP ThysanoessaTINORD Tintinnida167Table B3. LIfe history stages associated with the numeric codes of Table Bi, afterFoy and Anderson (1986).Stage Code Stage Stage Code Stage200 Eggs (ns) 415 Copepodite Stage V300 Nauphi (ns) 416 Copepodite Stage VI351 Nauplii (<0.20 mm) 417 Female Copepodite352 Nauphi (0.21-0.40 mm) 418 Male Copepodite353 Nauphi (>0.41 mm) 420 Copepodite (0.4 mm)400 Copepodite (ns) 421 Copepodite (0.4-0.8 mm)404 Copepodite Stage IV-V 422 Copepodite (>0.8 mm)405 Copepodite Stage I-Il 500 Larvae (ns)411 Copepodite Stage I 560 Furcilia412 Copepodite Stage II 590 Veliger413 Copepodite Stage III 700 Life Stage Not Specified414 Copepodite Stage IV168Table1.SummaryofstationssampledandsamplesprocessedforinvertebratezooplanktonanalysesfromtheFlemishCap.61cmbongo20cmbongo50cmringnetDisplacementvolumeVesselandDayCruiseNo.DateofyearStations0.505mm0.333mm0.165mmQ.253mni0.080mm0.080mm0.165mm0.333mm0.SO5nunDawson77Oct.25—30,197730140441Gadus11July14—22,197820056**44Gadus19Mar.20—26,197982425548Gadus20Apr.23—May9,1979121426566Dawson79July10—14,197919320131.3Gadus35Apr.7—14,1980100414039*4242384238Gadus37May21—31,1980147425641*424241130Zagreb4July20—29,1980207564179Gadus50May2—9,198112641311041*4040391446Gadus51May22—27,1981145424241*42424162Hawke/PandoraJune26—30,1981179424244*44178Hawke2Aug.1—4,1981215202029Cameron331Aug.1—3,1982214202121*2137Gadus75Mar.3—14,19836859*1411meshnotspecified,both505pmand333pm*samplescollectedbutnotprocessed**samplesprocessedbutdatanotedited_3Table 2. Ratios of Calanus finmarchicus mean density (m ) and weight (mg mto Oithona sp. at different times on Flemish Cap from small mesh samples(0.165 mm). Weights were estimated by the sum of total weight in eachcopepodite stage for each species for each cruise, standardized for the numbersof stations in each cruise. Copepodite weights from Tremblay (1981).C. fin: Oith9na C. fin: 9ithonaDate (numbers m ) (mg mpril 7—14, 1980 1.01:1 137:1May 2—9, 1981 0.52:1 8:1May 22—27, 1981 0.53:1 23:1May 21—31, 1980 0.88:1 39:1June 26—30, 1981 0.40:1 29:1J-u-lyI0-14-,-l9-7-- 0-47:l 34:1August 1—3, 1982 1.23:1 78:1170Table 3. Copepod egg and nauplii mean densities (number m) sampled onFlemish Cap within different strata from 0.080 mm mesh samples.Bottom depthsDate Stage (00 m 201—400 m 400 m >400 mApril 7—14, 1980 Eggs 2315 1985 2059 1012Nauplii 183 74 99 106May 2—9, 1981 Eggs 1335 286 519 321Nauplii- 1226 8I 912 1Q69Nay 20—26, 1980 Eggs 736 1000 945 552Nauplii 490 531 522 448May 22—27, 1981 Eggs 633 191 287 308Nauplii 1398 1049 1130 1429171Table 4. Density (number m) of Calanus finxnarchicus copepodite stages onFlemish Cap within different strata from 0.165 mm mesh samples.. Copepodjte densityDate Day of year stage <200 m 200—400m >400 m?pril 7—14, 1980 100 NS 63.643 8.425 8.228100 CI 0.477 1.844 0.206100 CII 1.559 0.572 0.225100 CIII 1.664 0.250 0.365100 CIV 3.419 1.402 1.349100 cv 5.953 1.359 1.274100 CVI 98.557 54.091 41.802May 2—9, 1981 126 MS 5.488 0.105126 CI 60.423 88.590 69.673126 CII 51.905 164.752 115.916126 CIII 114.790 166.585 165.887126 CIV 135.698 103.639 77.854126 cv 15.161 12.241 12.205126 CVI 15.259 16.112 8.902May 22—27, 1981 145 CI 24.859 3.135 11.407145 CII 7.268 7.274 13.961145 CIII 29.905 65.362 40.181145 CIV 143.356 215.495 145.46714-5 cv 61.145 9-9.381 7-.63145 CVI 12.437 18.890 23.210May 21—31, 1980 147 MS 17.091 35.043147 CI 11.144 14.562 25.625147 CII 24.922 25.893 37.438147 CIII 54.350 31.290 56.358147 CIV 7.742 16.156 37.744147 CV 1.856 4.131 4.795147 CVI 8.351 19.000 22.300June 26—30, 1981 179 CI 6.977 15.300179 CII 4.113 9.025179 CIII 3.810 3.713 8.080179 CIV 16.860 20.688 16.430179 Cv 198.937 112.498 53.915179 CVI 9.500 11.255 15.563July 10—14, 1979 193 MS 7.177 8.267 5.611193 CI 10.180 5.224 3.818193 CII 9.410 9.202 3.818193 CIII 3.788 12.768 0.000193 CIV 26.842 33.783 1.793193 CV 323.340 251.706 223.921193 CVI 38.386 64.167 36.068August 1—3, 1982 214 CI 5.500 3.256 2.353214 CII 2.188 1.164 5.850214 CIII 6.767 5.226 22.876214 CIV 122.679 69.464 28.192214 CV 212.710 395.829 154.292214 CVI 2.476 5.651 8.364172zn)indifferentyearsforwaters1400mdepthon•Table5.ComparisonofCalanusfinmarchicuscopepod.itestagedensities(nuaberFlemishCap,from0.165nunand0.333mmmeshsamples.HMeshsize=0.333mmMeshsize=0.165sunCopepoditedensityCopepoditedensityDateCICIICXXICXVCVCVICICIICIIICIVCVCVIApril23—May9,19798.613.910.87.96.849.5vsMay2—9,1981——36.663.18.27.5May21—31,19801.28.535.820.26.125.113.725.734.814.43.616.7vsNay22—27,1981—3.250.3267.2119.522.210.47.357.0200.391.317.5July10—14,19796.99.39.430.7277.854.8vsAugust1—3,19823.81.55.881.3355.14.9July20—29,1980—3.410.514.455.28.7vsAugust1—4,1981——2.32.174.817.8vsAugust1—3,1982—-—5.193.4274.96.3Table 6. Comparison of mean densities (number m) of Calanus finmarchicus and othercopepod among years on Flemish Cap within different depth strata for 0.165 mm and 0.333 mmmesh samples.StrataSpecies group Date <200ni 201—400m <400m >400m Total?%esh size = 0.165Calanus May 21—31, 1980 72.0 90.3 86.2 179.1 138.3firimarchicus vsMay 22—27, 1981 262.9 395.1 367.3 288.7 325.1July 10—14, 1979 413.3 367.1 383.9 271.2 366.6vsAugust 1—3, 1982 348.5 475.1 447.0 210.8 413.2Other May 21—31, 1980 100.2 219.6 193.0 265.4 233.7Copepods vsMay 22—27, 1981 1367.2 741.1 872.9 660.7 759.1July 10—14, 1979 1166.2 832.5 953.9 312.9 855.3vsAugust 1—3, 1982 504.0 376.0 404.4 179.9 372.31sh size = 0.333Calanus April 23—May 9, 1979 91.1 98.0 95.8 135.2 115.8finmarchicus vsNay 2—9, 1981* 191.7 237.3 225.9 188.2 207.7Nay 21—31, 1980 245.1 75.3 95.9 129.5 109.7vsNay 22—27, 1981 324.3 505.8 467.6 337.1 396.1July 10—14, 1979** 343.5 305.1 319.1 225.4 304.7vsJuly 20—29, 1980 157.0 59.8 85.1 57.6 73.1vsAugust 1—4, 1981 168.3 76.0 96.5 37.1 90.5vsAugust 1—3, 1982 267.8 411.2 379.3 223.6 357.1Other April 23—May 9, 1979 35.4 61.9 53.6 73.3 63.6Copepods vsNay 2—9, 1981* 96.2 162.3 145.8 105.8 126.4May 21—31, 1980 21.4 54.2 50.3 50.7 50.4vsMay 22—27, 1981 76.7 139.3 126.1 110.8 117.7July 10—14, 1979** 231.3 165.1 189.2 62.1 169.6vsJuly 20—29, 1980 20.1 16.0 17.1 23.2 19.8vsAugust 1—4, 1981 31.3 29.0 29.5 39.9 30.6vsAugust 1—3, 1982 37.9 48.4 46.1 33.9 44.3* 505 pm mesh samples** estimated by predicted catch ratios for 0333:0165 mm mesh nets forCalanus finmarchjcus = ,cO.83l1; Other Copepods xO.l983174Table 7. Summary of non—parametric analyses of differences among years forcopepods from ?ipril 23—May 9, 1979 versus May 2—9, 1981 within differentdepth strata for 0.333 mm and 0.505 mm mesh samples.StrataSpecies group <200m 201—400m 0Om >400mCalanus 2.4749’ 3.6348 4.3737 1.7266finmarchicus 0.01332 0.0003 0.0000 0.0842Other 2.1920 3.4520 4.2569 1.9671Copepods 0.0284 0.0004 0.0000 0.0492Total 2.3335 3.9132 4.7244 1.7782Copepods 0.0196 0.0001 0.0000 0.0754Wilcoxon Z—values2 Probability175Table 8. Summary of non—parametric analyses of differences among years forcopepods from May 21—31, 1980 versus Nay 22—27, 1981 within different depthstrata and for different mesh size samples.Sample StrataSpecies meshgroup (mm) <200m 201—400m <400m >400mCalanus 0.080 n.s. —2.4530 n.s.finmarchicus 0.05632 0.01420.165 n.s. —3.6442 —4.1174 1.64620.0003 0.0000 0.09970.333 n.s. 4.9519 5.2448 —3.11960.0000 0.0000 0.0018Other 0.080 —2.1651 —3.7745 —4.4960 —5.2726copepods 00304 0.0002 0.0000 0.00000.165 —2.1651 —3.6879 —4.2997 3.66700.0304 0.0002 0.0000 0.0002- 0.333 2.165l - 4.05 4.4277 —.3170.0304 0.0000 0.0000 0.0009Total 0.080 —2.1651 —3.4012 —4.1748 —4.3938Copepods 0.0304 0.0007 0.0000 0.00000.165 —2.1651 —3.6879 —4.3909 3.59890.0304 0.0002 0.0000 0.00030.333 n.s. 5.0014 5.2258 —4.39380.0000 0.0000 0.0000Wilcoxon Z—values2 Probability176-JTable9.Non—parametricanalysesofdifferencesamongyearsforcopepodsduringtheJuly—Augustperiod1979—82withindifferentdepthstratafor0.333mmmeshsamples(one—wayANOVAonrankscoresforF—value,P—levelsandDuncansmultiplerangetest).OtherCopepodsreferstothoseotherthanCalanusfininarchicus.Theyearsrefertothefollowingdates:July20—29,1980,August1—4,1981,andAugust1—3,1982.Strata<200in201—400in<00m>400mSpeciesF—valueOMR’TestF—valueIMR’restF—valueDMR’TestF—valueDMR1TestgroupP—level(P<0.05)P—level(P0.b5)P—level(P<0.05)P—level(P<0.05)Calanusn.e.—29.991982117.05198214.241982.1finmarchicus0.000119810.000119810.029219801980I19801981Othern.e.—9.13198219.121982n.s.Copepods0.000519810.0004198119801980Totaln.e.—31.601982I17.321982I4.261982ICopepods0.000119810.00011981j0.02881981198019801980DuncansNultipleRengeTestTable 10. Flemish Cap surface water temperatures (°C) for 10 m depth withinthe central area < 200 m water depth bounded by the area 46°30’ — 47°48’ N and44°6’ — 46°6’ W. Mean values were based on temperature data from the MarineEnviromental Data Service (MEDS) (Bottle and BT), METOC sea surface temperatures, and from data collected during directed Canadian research cruises as partof the Flemish Cap Project. Range of temperatures represent the 99% C. I. , inparantheses.Month Long-Term YearMean 1979 1980 1981January 4.55 3.95(4.05 — 5.05) (3.04 — 4.86)February 3.44(3.09 — 3.79)March 3.83 4.39 3.34 3.81(151 — 4.15) (&05— 4.13) (282 — 3.8G) (170— 192) -April 4.36 5.67 3.84 4.32(4.11 — 4.61) (5.32 — 6.02) (3.58 — 4.10) (4.19 — 4.45)May 5.74 6.30 5.58 6.98(5.47 — 6.01) (5.86 — 6.74) (5.36 — 5.80) (6.83 — 7.13)June 7.12 10.20 7.57 10.47(6.68 — 7.56) (9.61 — 10.79) (7.17 — 7.97) (9.93 — 11.01)July 11.71 12.18 12.32 11.55(11.34 — 12.08) (11.99 — 12.37) (11.29 — 13.35) (11.42 — 11.68)August 12.80 12.29 12.33(12.31 — 13.29) (11.95 — 12.63) (12.13 — 12.53)September 12.99 14.08(12.12 — 13.86) (13.25 — 14.91)October 9.83(9.15 — 10.51)November 8.18(7.77 — 8.59)December 5.96(5.22— 6.70)178Table 11. Temperature dependent predictions of days since spawning forcopepodite stages of Calanus finmarchicus on Flemish Cap, based on Belehradekequations (see text for explanation).Copepodite stageYear NI CI CII CIII CIV CV CVI1979 2.4 21.1 25.7 31.0 36.6 42.7 55.21980 2.9 26.9 31.0 36.5 43.2 50.4 65.11981 2.7 25.2 27.9 32.3 38.2 44.6 57.6179Table 12. Summary of samples collected on each cruise and number of redfishanalyzed for diet on Flemish Cap, 1978-82. Redfish Size Class refers to the meanand range of the number of specimens examined for each mm size.Cruise Sampling Date Samples Total Number Redfish Size Class Empty(Stations) Examined Mean Range (%)GADO11 16-23 July 1978 1 17 1.7 1-4 23.5GADO19 20-24 March 1979 25 135 2.3 1-6 18.5GADO20 23-27 April 1979 62 748 2.8 1-6 6.6DAWO79 10-14 July 1979 19 242 3.7 1-20 5.0GADO35 6-13 April 1980 37 218 1.2 1-3 6.0GADO37 20-26 May 1980 56 861 1.2 1-3 5.3ZAGOO4 22-28 July 1980 30 204 1.1 1-3 8.3GADO5O 2-9 May 1981 39 279 1.1 1-3 7.9GADO51 22-27 May 1981 34 248 1.1 1-2 4.4HAWPAN 26-30 June i981 38 213 1.0 1-2 2.8HAWOO2 1-4 August 1981 10 33 1.1 1-2 6.1ATC331 1-3 August 1982 8 27 1.0 1 3.7180Table 13. Summary of redfish diet expressed as percent occurrence based on bothweight and numbers. Here diet is summarized for each cruise for all sizes of redfish,and cruises are listed chroulogically by season. Diet is ranked from highest to lowestin each case. Each prey item represents the best possible taxonomic identificationfrom stomach samples.Cruise Prey Species Frequency Prey Species Frequency(% Weight) (% Number)GADO19 Copepod Nauplii 66.6 Limacina sp. 41.1(1979) Copepod Eggs 17.5 Copepod Eggs 37.2Limacina sp. 13.8 Copepod Nauplii 19.8Globigerina sp. 2.0GADO35 Copepod Eggs 85.8 Copepod Eggs 97.5(1980) Copepod Nauplii 14.0 Copepod Nauplii 2.2Limacina sp. 0.1 Limacina sp. 0.2GADO2O Copepod Nauplii 47.8 Copepod Eggs 79.7(1979) Cop o&Eggs 3{U) .CopepadNaupliiEuphausiid Nauplii 13.5 Euphausiid Nauplii 1.0Copepods NS 5.6 Limacina sp. 0.5Euphausiid Eggs 1.5GADO5O Cyclopoid Nauplii 32.2 Cyclopoid Nauplii 77.3(1981) Oithona Cop. (420) 20.3 Calanoid Nauplii 6.5Oithona Cop. (421) 15.5 Oithona Cop. (420) 5.1Calanoid Nauplii (352) 7.6 Copepod Eggs 4.9Calanoid Nauplii (353) 4.2 Oithona Cop. (421) 2.3Limacina sp. 1.2GADO37 Copepod Nauplii 72.1 Copepod Eggs 50.0(1980) Copepod Eggs 12.2 Copepod Nauplii 41.2Oil,hona Cop. NS 3.6 Oithona Cop. NS 3.0Calanoida NS 3.2 Limacina sp. 1.3Oithona NS 2.1 Dinophyta 1.00. similis CVI 1.4 Oithona NS 0.9GADO51 Cyclopoid Nauplii 39.4 Cyclopoid Nauplii 79.7(1981) Oithona Cop. (420) 32.9 Oithona Cop. NS 9.5Oithona Cop. (421) 20.1 Copepod Eggs 4.2Copepod Eggs 3.4 Dinophyta 4.2Limacina sp. 1.4181Table 13. (Cont’d...)Cruise Prey Species Frequency Prey Species Frequency(% Weight) (% Number)HAWPAN Oithona Cop. (421) 34.0 Cyclopoid Nauplii (351) 47.0(1981) Oithoma Cop. (420) 32.7 Calanoid Eggs 25.6Cyclopoid Nauplii 10.6 Oithoma Cop. (420) 15.3Calanoid Eggs 9.5 Oithona Cop. (421) 9.40. similis CVI Fern 3.0 Limacina sp. 0.7C. hyperboreus CIII 2.0 Calanoid Nauplii (352) 0.7C. finmarchicus CIV 1.7C. finmarchicus CV 1.2Calanus sp. 0.7DAWO79 0. similis Cop. NS 36.1 Copepod Nauplii 45.3(1979) 0. similis CV 28.9 0. similis Cop. NS 20.20. similis 8.5 0. similis CV 16.2Copepod Copepodites 8.3 Limacina sp. 4.80. sirnilis CIV 6.9 Copepod Eggs 4.2C. finmarchicus Cop. NS 3.2 0. similis CIV 3.8Copepod Nauplil 2.7 0. similis CVI 3.6C. finmarchicus CIV 2.0ZAGOO4 0. similis Cop. (421) 47.9 0. similis Cop. 50.1(1980) 0. similis Cop. (420) 6.0 Calanoid Eggs 16.40. similis CVI M 5.8 Cyclopoid Eggs 11.7C. glacialis CV 5.1 Calanoid Nauplii 5.90. similis CVI F 4.2 0. similis CVI M 3.8C. finmarchicus CVI F 3.9 Calanoid Nauplii 3.6Fish Eggs 3.5 0. similis CVI F 2.8C. finmarchicus CV 2.8 Cyclopoid Nauplii 2.3C. finmarchicus CIV 2.7 C. finmarchicus CI 0.7C. glacialis CVI F 2.2 C. finmarchicus CII 0.3C. glacialis CIII 2.1Calanoid Eggs 2.0Calanoid Nauplii 1.8C. finmarchicus CI 1.5C. finmarchicus CII 1.5C. finmarchicus CIII 1.4182Table 13. (Cont’d...)Cruise Prey Species Frequency Prey Species Frequency(% Weight) (% Number)HAWOO2 C. finmarchicus CV(1981) Oi1iona Cop. (421)C. finmarchicus CVI FOithona Cop. (420)Thyssanoesa NS0. similis CVI FCalanoid NSM. pygmaeus CVI FEuphausiid FurciliaC. finmarchicus CIIIM. pygmaeus NSC. finmarchicus CII0. similis Cop. (421)C. finmarchicus CV0 irnilis Cop 42O)C. finmarchicus CIVThyssanoesa Furcilia0. similis CVI FEuphausiid FurciliaCyclopoid NaupliiOithona Cop. (421)Oiihona Cop. (420)Cyclopoid NanpliiCalanoid EggsCalanoid Nauplii0. szmths CVI FM. pygmaeus NS0. similis Cop. (421)0. similis Cop. (420)Cyclopoid Np1iiCalanoid Eggs0. simths CVI F35.131.42394.22.235.131.011.68.24.32.71.331.323.214.112.25.52.41.81.51.21.21.00.933.222.0-. 17614.53.52.72.51.4ATC331(1982)183Table 14. Estimates of prey width for selected prey types as a percent of maximum mouth width for different redfish sizes. NI and NVI refer to nauphi stages Iand VI, and CI and CVI refer to copepodite stages I and VI. Width estimated forCalarius fimmarchicus nauplii was 75% of total length and for copepodites was 33%.For Oithoma similis nauplii width was estimated to be 67% of total length and forcopepodites was 34%. Width estimates approximated inclusion of folded antennae.a) Calanus fimmarchicusb) Oithona similisRedfish Length Mouth Width Calanus finmarchicus(mm) (mm)6 0.63101520251.542.673.804.93Eggs21.48.85.13.52.7NI26.210.76.24.33.4NVI72.729.717.212.19.3CI32.213.27.65.34.1CVI112.746.126.618.714.4Redfish Length Mouth Width Oithoma similis(mm) (mm) NI NVI CI CVI6 0.63 12.2 22.9 12.4 27.810 1.54 5.0 9.4 5.2 11.415 2.67 2.6 5.4 2.9 6.620 3.80 2.0 3.8 2.1 4.625 4.93 1.6 2.8 1.6 3.6184Table 15. Statistical results of comparisons of redfish diet (% weight) amongyears during the March-May period 1979-81. Each entry for upper, middle and lowernumbers are the Komolgorov-Smirnov test statistic Dmax, the probability level (Flevel) and the number of observations n1 and n2. Redfish Sizes refers to the overallsize range of fish for which diet comparisons were made.Cruise Redfish Sizes DmaxComparison (mm) P-level(ni, n2)20-24 March 1979 vs. 4-8 0.6836-13 April 1980 0.001(135, 218)23-27 April 1979 vs. 4-9 0.6572-9-May 1981- 0.001(692, 237)10-13 0.3130.05(56,42)20-26 May 1980 vs. 6-14 0.80422-27 May 1981 0.001(861, 248)185Table 16. Statistical results of comparison of redfish diet (% weight) among yearsduring the July-August period 1979-82 on Flemish Cap for three size classes: 9mm, 10-19 mm, > 20 mm. Each entry for upper, middle and lower numbers arethe Kolmogorov-Smirnov test statistic (Dmax), the probability level (P-level) andnumber of observations n1 and n2. n.s. refers to the test statistic being less than thecritical value where a 0.1, based on Tables L11 and L111 in Siegel and Castellan(1988). NA indicates no comparison was possible.Cruise Size ClassComparison 9 mm 10-19 mm 20 mm10-14 July 1979 vs. 0.386 0.048 NA22-28 July 1980 P < 0.01 n.s.(232, 24) (9, 143)10-14 July 1979 vs. L354 0i561-4 August 1981 P < 0.1 n.s.(232, 13) (9, 8)10-14 July 1979 vs. 0.262 0.155 NA1-3 August 1982 n.s. n.s.(232, 6) (9, 14)22-28 July 1980 vs. 0.151 0.155 0.1161-4 August 1981 n.s. n.s. n.s.(24, 13) (143, 8) (37, 7)22-28 July 1980 vs. 0.174 0.196 0.2671-3 August 1982 n.s. n.s. n.s.(24, 6) (143, 14) (37, 7)1-4 August 1981 vs. 0.278 0.237 0.1511-3 August 1982 n.s. n.s. n.s.(13, 6) (8, 14) (12, 7)186Table 17. Mean stomach contents weight (tg) standardized for fish size (mm3)for redfish larvae sampled on Flemish Cap. Mean standardized stomach weightswere calculated for each size class within each cruise and then an overall mean wascalculated from these mean estimates. (n = number of redfish size classes, s.d. =standard deviation, CV = coefficient of variation).Year Cruise Redfish Sizes n Mean s.d. CV(mm) (tg/mm3) (%)1979 GADO19 4-7 4 0.124 0.0441 35.6GADO2O 5-11 7 0.281 0.0836 29.80AW079 4-11 8 0.344 0.0749 21.81980 GAD035 5-7 3 0.109 0.0040 3.7GADO37 6-14 9 0.381 0.Q7S3 19.8ZAGOO4 5-11 7 0.055 0.0437 79.512-24 13 0.449 0.1581 35.21981 GADO5O 6-11 6 0.085 0.0369 43.4GADO51 6-11 6 0.109 0.0378 34.7HAWPAN 6-7 2 0.070 0.0170 24.38-16 9 0.351 0.0378 10.8HAWOO2 7-11 5 0.088 0.0504 57.313-24 10 0.311 0.1921 61.81982 ATC331 6-11 4 0.242 0.1839 76.016-23 8 0.642 0.2422 37.7187Table 18. Wilcoxon’s test of difference between standardized stomach contentsweights (weight of stomach/weight of redfish) for each mm length group sampledMay 1980 versus 1981. n1 - refers to 20-26 May 1980 and n2 to 22-27 May 1981.Length Group n1,n2 z P-level(mm)6 44, 2 -0.9965 0.31907 123, 12 -4.0012 0.00018 146, 55 -4.2499 0.000019 153, 78 -5.3535 0.0000110 145, 26 -2.7553 0.005911 105, 48 -4.8148 0.0000112 52, 10 -0.2010 0.8407188Table 19. Summary of redfish prey items classified as “Other” in the feedinganalyses. The count represents the total number observed in all redfish stomachsanalyzed for each cruise and the percent ranks the relative abundance of these preyitems for each cruise.Cruise Date Prey Item Count PercentGADO19 20-24 March 1979 Globigerima sp. 10 90.9Tintinnida 1 9.10AD035 6-13 April 1980 Granuloreticulosia 1 50Radiolaria 1 50GADO2O 23-27 April 1979 Tintinnida 10 83.3Radiolaria 1 8.3Crustacea 1 8.3GADO5O 2-9 May 1981 Tintinnida 3 75Porifera - 1- 25GADO37 20-26 May 1980 Radiolaria 284 72.8Porifera 83 21.3Tintinnida 18 4.6Granuloreticulosia 5 1.3GADO51 22-27 May 1981 Tintinnida 10 90.9Porifera 1 9.1RAWPAN 26-30 June 1981 Globigeriria sp. 3 42.9Tintinnida 2 28.6Porifera 1 14.3Parathemisto larva 1 14.3ZAGOO4 22-28 July 1980 Fish Eggs 80 95.2Oikopleura sp. 3 3.6Fritillaria sp. 1 1.2HAWOO2 1-4 August 1981 Isopoda 1 100ATC331 1-3 August 1982 Tintinnida 1 100189Table 20. Summary of redfish sizes (mm) and number of stations (n) used in theselectivity analysis with 0.080 mm mesh zooplankton samples.Cruise Date Size n Mean SizeGADO35 6-13 April 1980 5 3 6.56 337 3381GADO37 20-26 May 1980 5 3 8.76 187 378 409 4010 3711 2712 613 1GADO5O 2-9 May 1981 6 10 8.57 258 329 2910 1511 1112 313 1GADO51 22-27 May 1981 6 2 9.4778 259 3110 1911 1912 913 1190Table 21. Summary of redfish sizes (mm) and number of stations (n) used in theselectivity analysis with 0.165 mm mesh zooplankton samples.Cruise Date Size n Mean SizeHAWPAN 26-30 June 1981 6 1 10.471819910 2311 2412 1913 1814 815 816 1ATC331 1-3 August 1981 6 1 6.57110 1 10.511 116 1 19.417 318 219 220 321 222 223 1191Table 22. Summary of dry weight versus length (log1o-lo transformed) leastsquares regressions for different cruises. The overall regression includes all cruisesexcept DAWO79, 10-14 July 1979, in which only small redfish were sampled anddry weights were significantly low. DW — refers to dry weight (pg), SL — refers tostandard length (mm).Year Cruise Regression n R2 F-level Size (mm)1979 DAWO79 DW=-1.7083+2.8622SL 240 73.8 0.0001 3-111980 GADO35 DW=1.2481+1.4363SL 211 19.8 0.0001 5-8GADO37 DW=0.5249+2.529SL 855 76.2 0.0001 5-141981 GADO5O DW=-0.2697+3.1O5OSL 261 81.4 0.0001 6-13GADO51 IDW=-0.5188+3.44158L 244 85.3 00001 6-13HAWPAN DW=-0.7819+3.7960SL 211 94.1 0.0001 5-16IIAWOO2 DW=-0.7688+3.7833SL 31 98.4 0.0001 5-26OVERALL DW=-0.1459+3.1590SL 1818 82.5 0.0001 5-26192Table 23. Summary of statistical tests of differences in redfish dry weight (tg) foreach length group (mm) between 20-26 May 1980 versus 22-27 May 1981. n1 and n2refer to the sample size for 1980 and 1981 data, respecitvely. z refers to Wilcoxon’stest statistic.Length Group n1,n2 z P-level Percent(mm) Difference6 39, 3 -2.3935 0.0167 34.37 136, 15 -5.8011 0.00001 50.18 155, 58 -8.6939 0.00001 36.69 160, 82 -9.5851 0.00001 36.110 151, 27 -6.3198 0.00001 34.411 110, 49 -4.8624 0.00001 16.312 37, 9 -1.4122 0.1579 11.0% Difference = difference between years relative to 1980 data.193Table 24. Summary of dry weight versus fish volume (loglo-loglo) transformedleast squares regressions for different cruises. The overall regression is for all cruises.DW- refers to dry weight (ag), FV - refers to fish volume (jtgDW).Cruise Regression n R2 P-level Size (mm)GADO35 DW=0.6842+0.723FV 183 29.7 0.0001 5-8GADO37 DW=0.3249+0.907FV 802 89.6 0.0001 5-13GADO5O DW=-0.1798+1.008FV 259 93.8 0.0001 6-13GADO51 DW=-0.1081+0.979FV 234 93.3 0.0001 6-12HAWPAN DW=-0.2941+1.047FV 201 93.6 0.0001 5-16HAW0O2 DW=-0.4167+1.047FV 27 99.2 0.0001 5-26Overall DW=0.1681+0.929FV 1711 87.2 0.0001 5-26194Table 25. Summary of regression comparisons in dry weight versus fish volumeregressions among various cruises. All paired comparisons between cruises are basedon weighted least squares regressions.Cruise Length Range Slope Comparison df InterceptComparison (mm) F-value P-level (total) P-levelGADO37-GADO5O-GADO51- All (5-35) 26.52 0.0001 1527 NAHAWOO2-HAWPANGADO5O-GADO51-HAWPAN- All (5-35) 8.11 0.0001 724 NAHAWOO2GADO37-GADO51 6-12 (all) 2.26 0.1331 973 0.00001GADO37-HAWPAN 5-16 (all) 37.7 0.0001 1004 NAGADO37-HAWPAN 6-12 2.63 0.1051 880 0.0001GADO5O-GADO51 6-13 (all) - 1.47 - 0.2256 494 0.1227 -GADO5O-HAWPAN 5-16 (all) 2.47 0.1170 461 0.4760GADO51-HAWPAN 5-16 (all) 6.92 0.0088 436 NAGADO51-HAWPAN 6-12 0.32 0.5708 376 0.0383HAWPAN-HAWOO2 5-26 (all) 3.43 0.0652 229 NAHAWPAN-HAWOO2 6-26 2.36 0.1257 224 0.0125HAWPAN-HAWOO2 6-16 0.34 0.5626 210 0.4950NA — refers to Not Applicable, as the slopes were statistically different.All — refers to all available lengths for each cruise were included in the initial comparison. Sizeranges for each cruise are listed in Table 23.195Table 26. Summary of fish volume versus length (loglo-loglo transformed) leastsquares regressions for different cruises. The overall regression is for all cruises. FV- refers to fish volume in units of dry weight (pg), SL - refers to standard length(mm).Cruise Regression n R2 P-level Size (mm)GADO35 FV=0.7012+2.086SL 184 71.7 0.0001 5-8GADO37 FV=0.1760-I-2.835SL 806 87.2 0.0001 5-14ZAGOO4 FV=-0.3367+3.5195L 201 97.4 0.0001 5-28GADO5O FV=-0.0990+3.0915L 276 88.8 0.0001 6-13GAD051 FV=-O.4225-I-3.516SL 237 90.6 0.0001 6-12HAWPAN FV=-0.3009+3.466SL 201 91.5 0.0001 5-16HAWOO2 FV=-0.2262-j-3.360SL 28 98.5 0.0001 5-37ATC331 FV=-0.3922+3.522SL 26 95.1 0.0001 6-23OVERALL FV=-0.5030+3.582SL 1966 95.1 0.0001 5-37196Table 27. Summary of statistical tests of differences in redfish fish volume(tgDW) for each length group (mm) between 20-26 May 1980 versus 22-27 May1981. n1 and n2 refer to the sample size for 1980 and 1981 data, respecitvely. zrefers to Wilcoxon’s test statistic.Length Group n1,n2 z P-level Percent(mm) Difference6 37, 2 1.3053 0.1918 -19.37 118, 15 -2.7421 0.0060 18.98 145, 57 1.7250 0.0845 -6.79 151, 81 2.5342 0.0113 -8.210 150, 28 3.0184 0.0025 -13.711 107, 47 8.3180 0.00001 -34.612 55, 8 3.5816 0.0003 -31.2t % Difference = difference between years relative to 1980 data.197Table 28. Summary of regression comparisons in fish volume (gDW) versuslength (mm) regressions among various cruises for the July-August period 1980-82.Redfish were selected to range between 6-23 mm in length.Cruise Regression n Parameter ComparisonsComparison Slopes Interceptsa)ZAGOO4 FV=2.780+3.534SL 196 F=0.62 F=7.33HAWOO2 FV=2.985-+-3.4595L 25 P=0.4326 P=0.0073b)HAWOO2 FV=2.985+3.4595L 25 F=0.32 F=0.001ATC331 FV=2.541+3.522SL 26 P=0.5770 P=0.9556c)ZAGOO4 FV=278O+3.534SL 19G F=D.02 F=8.35ATC331 FV=2.541+3.5225L 26 P=0.9005 P=0.0042198Table 29. Summary of statistical tests of differences in fish volume (pgDW) fordifferent length groups (mm) between cruise pairs for the July-August period 1980-82. The sample sizes are indicated in parantheses for mean fish volume values. Theoverall values are for all length ranges within each cruise comparison.1981 1980 1982 — 1981 1980 — 1982Size Mean FV z Size Mean FV z Size Mean FV z(mm) (pg DW) P-level (mm) (pg DW) P-level (mm) (pg DW) P-level5-7 271 (6) 0.6617 6-7 308 (6) 0.6390 6-7 281 (16) 0.0369232 (15) 0.51 292 (5) 0.52 308 (6) 0.978-10 957 (5) 1.0410 10-11 1886 (3) -0.2887 10-11 2434 (15) -1.5401858 (10) 0.30 1574 (2) 0.77 1886 (3) 0.1211-13 1463 (1) NA 16-18 10411 (6) -0.6390 16-17 9368 (48) 0.15453185 (27) 9727 (5) 0.52 9616 (4) 0.8814-16 6185 (3) -0.6347 20-23 16579 (10) 0.7319 18-19 12867 (37) 0.50536905 (28) 0.53 18848 (7) 0.46 12933 (4) 0.6120-21 17246 (33) -I.886615491 (7) 0.0622-23 22665 (8) -0.918619721 (3) 0.36Overall Mean (n) z Overall Mean (n) z Overall Mean (n) zP-level P-level P-level1980 9617 (197) -1.7545 1981 7617 (26) -1.1298 1980 9617 (197) 0.31671981 7617 (26) 0.08 1982 9825 (27) 0.26 1982 9825 (27) 0.75199Table 30. Summary of the number of redfish examined for age and growth ratebased on otoliths for each cruise, and the number of stations in each cruise for whichage and growth data were available.Year Cruise Dates Redfish StationsExamined1980 GADO35 6-13 April 147 31GADO37 20-26 May 443 45ZAGOO4 22-28 July 166 291981 GADO5O 2-9 May 21 5GADO51 22-27 May 234 28HAWPAN 26-30 June 277 39HAWOO2 1-4 August 7 5200Table 31. Copepod egg and nauplii densities (number m3) averaged for eachcruise for waters 400 m depth, from 0.080 mm mesh samples.Prey Type GADO35 GADO37 GADO5O GADO51Calanoid Eggs N/A N/A 907 409Cyclopoid Eggs N/A N/A 239 179Total Eggs 1461 909 1146 588Calanoid Nauplii 63 504 629 137Cyclopoid Nauplii N/A N/A 1273 1997Total Nauplii 63 504 1902 2134Total Eggs & Nauplii 1524 1413 3048 2722201Table 32. Ranges of variables used in the spatial analyses from 3 cruises, 20-26May 1980, 22-27 May 1981 and 26-30 June 1981. The values represent ranges ofmeans calculated for each variable at each station. Prey concentrations for the twoMay cruises are based on 0.080 mm mesh samples and for the June 1981 cruise arebased on 0.165 mm mesh samples.Variable 20-26 May 1980 22-27 May 1981 26-30 June 1981(GADO37) (GADO51) (HAWPAN)Length 7.0—10.1 7.0—12.3 9.0—16.5Mean Growth (mm d’) 0.138—0.174 0.095—0.140 0.087—0.1495-Day Growth (mm d’) 0.130—0.192 0.091—0.147 0.062—0.14510-Day Growth (mm d’) 0.125—0.190 0.104—0.141 0.085—0.158Age (d) 7.7—26.9 7.3—62.0 21.0—114.0Dry Weight (pg) 360.0—1373.0 417.7—1903.0 447.0—6006.9Condition (pg mm’) 48.0—138.1 47.1—158.6 51.4—400.5Feeding Rate (pg mm’) 0.121—1.639 0.096—0.957 0.064—1.081Temperature (°C) 3.5—6.6 4.3—10.3 5.2—11.7Prey Type 1 (mg m3) 0.0001—0.610 0.003—0.364 0.0001_0.003*Prey Type 2 (mg m3} 0.002-4.762 0.102—2-4Th 0-.0O03--O.O37Prey Type 3 (mg m3) 0.214—12.380 2.166—16.182 0.332—2.454Prey Type 4 (mg m3) N/A 0.029—0.868 0.0001_0.037*Prey Type 5 (mg m3) 0.10—8.07 1.544—12.963 0.068—1.476Prey Type 6 (mg m3) N/A 0.073—1.929 0.0002_0.0070**— Prey Types underestimated by 0.165 mm mesh nets.202Table 33. Spearman rank correlations comparing redfish growth and feeding ratesto redfish size and condition, water temperatures and prey biomass for three cruises:20-26 May 1980 (GADO37); 22-27 May 1981 (GADO51); 26-30 June (HAWPAN).MMGI - mean growth (mm d1), MGI5 = growth rate for the last 5 days (mm d1),MGI1O = growth rate for the last 10 days (mm d’), MSTDX = diurnally correctedstandardized feeding rate (mg mm’), MEANSIZE = standard length (mm), MDW= dry weight (tg), MDW.SL = relative condition (/Lg mmd’), MPOSTX = age(days), SURTEMP = surface water temperatures ( 20 m), PREY1BIO = copepodeggs (mg m3), PREY2BIO = copepod eggs and nauplii (mg m3), PREY3BIO =copepod eggs, nauphi, Calamus finmarchicus and Oithona copepodites (mg m3),PREY4BIO = cyclopoid eggs and nauplii, 1981 data only (mg m3), PREY5BIO =cyclopoid copepodites (mg m3).203C1UISZ-fiAV?AISPE?JJIA CORRELATION COEF?TCXESTS / PROS UDZ1 RO:2.O0 / I1SZZ OFMEANSIZE PlOW IWWSL MpOsrX SURTEIIP PIET1SIC PRZT2BIO PRET3SEOOSSER0NSPRET4BIO PRETS8XOKMGIPlotSCRUXSEGAD037SPEAPJVN CORRELATION COEFFICIENTS / PROS ) R( VNOER EO:RNO0 / PSUPISER 07MEANSIZE MOW MDWSL MPOSTX SURTE1P PRZY13IO PRET2SIO PREt38100.24941. —0.17923 —0.29398 —0.15299 —0.05072 —0.06204—0.05311 —0.051450.1366 0.2885 0.0774 0.3157 0.7407 0.7153 0.7549 0.762437 37 37 45 45 37 37 370.55145 0.26861.0.0004 0.107937_3714G110 0.526-79 0.238030.0008 0.156037 370.182310.2801370.151970.3692370.45839 0.32991 0.31564 0.33310 —0.087720.0015 0.0269 0.0570 0.0440 0.605745 45 37 37 370.42630 0.28986 0.34650 0.39403 0.027260.0035 0.0534 0.0345 0.0158 0.872745 45 37 37 37OSSERVATIONSPREY4SIO PRETSOXO—0.05311 0.000000.7549 1.000037 370.33310 —0.043150.0440 0.799837 370.39403 0.069230.0158 0.613937 37MSTDX 0.08208—0.19601 —0.19974 —0.09675 —0.07884 —0.0940P —0.16765 0.09408 —0.16765 0.152420.6053 0.2135 0.2047 0.589 0.6197 0.55iS 0.2886 0.5534 0.2886 0.335242 42 42 37 42 42 42 42 42 42CRUtSE.GAD051SPEARMAN CORRELAXON COEFFICIENTS / PROS > RI UNDER B0:REO—0 / NUMBER OF OSSERV?..TIONSMEANSIZE PtDW ?IDWSL K?OSTX SURTEMP PRET1BIO ?RET2SIO PRET3BIO PRET4SIO PRETSBIOP1MG 1 0.43186 —0.13519 —0.18062-0.0117 0-.4928 0.357728 28 2814015 0.671590 .000128KGI1O 0.653530.0002280.343730.0733280.270390.1640280.517310.048280.481730.0094280.559460.0020280.284070.1429280.207440.2895280.2101.8—0.06291—0.04598 0.135740.44554 0.23S6,0.2830 0:7505 0.8163 0.4910 0.0175 0.189228 28 28 21 28 280.006020.9757280.065130.741928—0.17225 —0.301590.3808 0.118828 28—0.22995 —0.278050.2391 0.152028 280.151610.4412210.246850.20542114510 X 0.04325 0.34258 0.39104 0.35829 0.13247—0.06069—0.11.199—0.071350.8081 0.0510 0.0244 0.0612 0.4552 0.7332 0.5283 0.611434 33 33 21 34 34 34 340.025730.8966280.128080.5160280.153800.4346280.291740.132028—0.14530 —0.070130.4123 0.693534 34?4MGI 0.56502 0.32184 0.30104 0.11643 —0.19337 0.22527 0.15811 —0.32567 0.19955 —0.013740.0014 0.0949 0.1195 0.4803 0.3149 0.4593 0.4504 0.0901 0.3733 0.671829 25 28 39 29 13 25 25 22 2114015 0.30148 —0.02956—0.03996 0.04485 —0.11873 —0.27473 0.17196 —0.22167 0.21015 —0.059660.1120 0.8813 0.8400 0.7863 0.5396 0.3637 0.4111 0.2569 0.3462 0.763029 28 28 39 29 13 25 28 22 28140110 0.32660 0.03339 0.01697 —0.02733 —0.09188 —0.03297 0.26697—0.13738 0.27191 0.039110.0838 0.8661 0.9317 0.8681 0.6355 0.9149 0.1970 0.4857 0.2209 0.765129 25 28 39 29 13 25 28 22 2814510 X 0.17509 0.00728 0.00924 —0.25349 0.28657—0.06765—0.1.1997 0.05350 —0.20464 0.048180.3144 0.9669 0.9580 0.1931 0.0951 0.1034 0.5203 0.7602 0.3059 0.783435 35 35 28 35 16 31 IS 27 35204Table 34. Spearman rank correlations comparing redfish size, condition and ageto temperature and prey biomass for three cruises: 20-26 May 1980 (GADO37); 22-27 May 1981 (GADO51); 26-30 June (HAWPAN). MEANSIZE = standard length(mm), MDW = dry weight (zg), MDWSL = relative condition (tg mmd’),MPOSTX = age (days), SURTEMP = surface water temperatures (<20 m), PREY1BIO= copepod eggs (mg m3), PREY2BJO = copepod eggs and nauplii (mg m3),PREY3BIO = copepod eggs, nauphi, Calamus fimma’rchicus and Oithona copepodites (mg m3), PREY4BIO = cyclopoid eggs and nauplii, 1981 data only (mgm3), PREY5BIO = cyclopoid copepodites (mg m3).205SURTENP PREY1BE0 PRET2SIO PRET3BXO PRE’r4810 PREYSBIOCRUISEHAWPANSPJAN CORRELATION COEFrICIEZtrS / PROS ) IRIStYRTEMP PRET1SIO pR5y2B10UNDER !O:tRO=0 / NTKEEZ OFPRET3BIO PRZT4BIO PRETS8IOOBSERVATIONSCZUXISE.GAD037SPEARNAZ9 CORRELATION COEFrICIEN’rS / PROS> RI UNDER EO:RRO0 / 1IUI1BER orSURTEMP PREY1BIO PRE2BIO PRET3SIO PREY4BIO PREYSBXONE?.IfSIZE 0.28191. 0.19108 0.20946 —0.06113 0.20946 —0.028770.0705 0.2254 0.1831 0.7003 0.1831 0.856542 4: 42 42 42 42MOW 0.70408 0.46796 0.46358 —0.03136 0.46358 0.040760.0001 0.0018 0.0020 0.8437 0.0020 0.797742 42 42 42 42 42I4DWSL 0.66648 0.46131. 0.47346 0.02666 0.47346 0.093100.0001 0.0021 0.0015 0.8669 0.0015 0.557642 42 42 42 42 42MPOSTX 0.72564 0.47734 0.38947—0.10706 0.38947 —0.086670.0001 0.0026 0.0172 0.5282 0.0172 0.610045 37 37 37 37 37OBSERVATIONSOBSERVATIONSCRUISE.GAD051SPEARMAN CORRELATION COEFFICIENTS / PROS > R( UNDER MO:RBO—0 / NuMBER orMEABStZE 0.30525—0.04130 0.12694 0.08291. 0.20612 0.119090.0702 0.3110 0.4607 0.6307 0.2278 0.419136 36 36 36 36 36MOW 0.44953 0.17604 —0.17447 —0.23830 0.09626 —0.250330.0087 0.3271 0.3315 0.1817 0.5941 0.160033 33 33 33 33 33MOW St. 0.41544 0.12826 —0.18750 —0.30849 0.03175 —0.332220.0162 0.4769 0.2961 0.0807 0.8608 0.051933 33 33 33 33 33KPOSI’X 0.61515 0.08298 —0.35637 —0.18804 0.15941 —0.114410.0005 0.6746 0.0627 0.3379 0.3344 0.562128 28 28 28 23 28MEASIZE 0.32308—0.12745 —0.24143 —0.15909 —0.11365 —0.141150.0511 0.6259 0.1831 0.3540 0.5647 0.38*537 17 32 36 21 36NOW 0.44572 —0.16912 —0.23243 —0.07284—0.14732 —0.191250.0064 0.5164 0.2005 0.6729 0.4544 0.263836 17 32 36 23 36MDWSL 0.44044—0.24510—0.241.96—0.05302—0.15729 —0.210300.0072 0.3430 0.1121 0.7587 0.3399 0.218336 17 32 36 28 36MPOSTX 0.11285 0.06593—0.41546 —0.27375—0.19389—0.161780.5600 0.1305 0.0389 0.1587 0.3873 0.410129 13 28 28 22 23206Table 35. Spearman rank correlations comparing redfish feeding to growth ratesfor three cruises: 20-26 May 1980 (GADO37); 22-27 May 1981 (GADO51); 26-30June (HAWPAN). MMGI- mean growth (mm d1), MGI5 = growth rate for thelast 5 days (mm d’), MGI1O = growth rate for the last 10 days (mm d—’), MSTD_X= diurnally corrected standardized feeding rate (mg mm1).Feeding Cruise GrowthMean 5-Day 10-DayMSTDX GADO37 0.1017 -0.0059 -0.10120.5492 0.09722 0.551137 37 37GADO51 0.0493 0.2107 0.22330.8034 0.2818 0.253328 28HAWPAN 0.0487 0.3749 0.38480.8056 0.0493 0.043228 28 28207Table 36. Regression analyses of redfish growth, size and relative condition with prey concentrations and temperature, based on spatial differences measured during 3 different cruises. ForGADO37 and GADO51 prey concentration was Prey Type 2 (logio mg m3) and for HAWPANit was Prey Type 5 (logio mg m3). Temperature (°C) is for surface waters < 20 m depth. G= mean growth (mm a—i), G5 = recent 5-day growth (mm d1), GlO = recent 10-day growth(mm d1), SL = standard length (mm), DW = dry weight (pg), RC = relative condition (pgDWCruise Dependent Prey Temperature df F-value P-value R2Variable Coefficient Coefficient (%)GADO37 G -0.001 -0.001 36 0.26 0.7698 2.5G5 0.040 0.003 36 1.77 0.1851 9.5GlO 0.017 0.002 36 0.61 0.5484 3.5SL i.65i 0.182 36 1.61 0.3252 6.4DW -228 184*** 36 8.99 0.0007 34.6RC -8.i 14.i*** 36 6.25 0.0049 26.9GADO51 G 0.050 0.001 27 0.59 0.5636 4.5G5 -0.087 -0.003 27 1.73 0.1978 12.2GiO -0.018 -0.002 27 1.16 0.3309 8.5SL 1.100 O.496*** 27 9.31 0.0010 42.7DW -616 204*** 27 16.66 0.0001 57.1RC -40.4 15.0*** 27 13.72 0.0001 52.3HAWPAN G -0.017 _0.004** 27 2.36 0.1155 15.9G5 -0.016 _0.005* 27 1.95 0.1629 13.5GiO 0.005 -0.005 27 1.33 0.2826 9.6SL -0.160 0.174 27 0.54 0.5916 4.1DW -247 338 27 1.52 0.2388 10.8RC -3.1 21.5* 27 1.62 0.2187 11.5*- significantly different than zero at P <0.10**- significantly different than zero at P < 0.05- significantly different than zero at P < 0.01208Table 37. Redfish abundances estimated for each cruise carried out on Flemish Cap, for alllengths sampled.Cruise Year Date Mid-Date Stations Total April-releasedAbundance AbundanceGADO11 1978 16-23 July 201 56 2.3 x 1011 1.5 x 1011GADO19 1979 20-24 March 80 42 1.1 x 1011GAD02O 23-27 April 115 42 7.1 x 1012DAWO79 10-14 July 193 20 nilGADO35 1980 6-13 April 101 41 3.1 x 1012GADO37 20-26 May 145 42 4.4 x 1012ZAG0O4 22-28 July 207 38 1.5 x 1011 1.1 x 1011GADO5O 1981 2-9 May 126 41 5.6 x 1012GADO51 22-27 May 145 42 2.9 x 1012HAWPAN 26-30 June 179 42 3.8 x 1011 3.5 x 1011HAWOO2* 1-4 August 215 20 1.4 x 1010 8.8 x ioATC331* 1982 1-3 August 214 20 1.0 x 1011 9.3 x 1010GADO75 1983 3-14 March 68 54 6.2 x io*— Estimate made assuming 20 stations sampled 83% of the total redfish population.209Table 38. Redfish population abundances estimated in different years for a)the time of peak spawning standardized to Day 120 and b) during early Auguststandardized to Day 215. Abundance esitmates for Day 215 are made only forlarvae spawned in April. Projections are made forward and backward in time fromsampled population estimates using different estimates of mortality (rate of changeper day).a) Peak larval abundances on Day 120.Year Day Interval Mortality Abundance(d’)1979 115 —> 120 0.118 1.3 x 1013115 na 7.1 x 10121980 120 <— 145 0.035 1.1 x iO’0.059 1.9 x 10131981 120 <— 126 0.035 6.9 x 1012120 <— 145 0.035 6.8 x 10120.059 1.3 x 10-b) Pelagic juvenile abundances on Day 215.Year Day Interval Mortality Abundance(d1)1978 201 —> 215 -0.05 7.3 x 1010-0.07 5.5 x 10101979 120 —> 215 -0.07 1.7 x 1010120—> 215 -0.10 9.7 x iO1980 207 —> 215 -0.059 7.1 x 10101981 215 na 8.8 x io1982 214 na 9.3 x 1010210Table 39. Abundance estimates of redfish (Sebastes spp.) at different ages. Seetext for data sources. Trawl — number caught per tow, Stomach— number per codstomach, Larvae and Juveniles— abundance estimated for Flemish Cap, CPUE -catch per unit effort.Year Parent Stock Larvae JuvenilesCatch CPUE (Age 0) (3 Months)(t) (t/h)1979 20074 1.58 7.2x10’ 01980 15957 2.07 1.1 x1013 7.lxlO’°1981 13891 2.16 6.9 xlO’2 8.8 x109Year 1—Year 1—Year 2—Year 2—Year 5—Year 5—YearTrawl Stomach Trawl Stomach Catch OPUE(number) (number h’)1979 0.30 0.04 0.37 0.02 1027 90.51980 88.5 3.51 433.8 1.08 .8200 762.41981 167.3 1.35 NA NA 2571 241.5211Figure 1. Flemish Cap bank, centered at 47°N, 45°W, is situated west of theGrand Bank of Newfoundland. The boxed area represents the maximum area inwhich samples were collected as part of this study.21249° 47° 45° 43° 41°50°49°48°47°46°45.213Figure 2. Seasonal changes in plankton volume (ml m3) of invertebrate zooplankton on Flemish Cap. The plots represent a composite of data collected in different years from different mesh samples: (a) 0.333 mm mesh, (b) 0.165 mm mesh,(c) 0.080 mm mesh. Bars represent one standard deviation about cruise means.2141.2(a) 0.333mm1.00.80.6 T0.4 T + ITI. 1 1I I(b) 0.165mu-iC-- 00.6A—0.4C-) TA-JC’,0.0(c) 0.080mm3.0* 1979A 198019812.0 0 19820 19831.0••1AI0.0March April May June July August215Figure 3. Seasonal changes in percent community composition of invertebratezooplankton on Flemish Cap broken down by selected groups. This is a compositepicture of data collected in different years from 0.333 mm mesh samples. Refer tothe text for details.216Other100-—•Zooplankton90EuphausiidaeGastropoda__80Copepoda>-Oithonasp.DCalanustinmarchicus0M(H“ z 0 I.40z -J300 0 N20 10- 0-IIIIiiIIAprilMayJuneJulyAug.Sept.Oct.Nov.Figure 4. Seasonal changes in percent community composition of invertebratezooplankton on Flemish Cap broken down by selected groups. This is a compositepicture of data collected in different years from 0.165 mm mesh samples. Refer tothe text for details.218Other100•ZooplanktonSarcodina90::::c1:.zCalamus(<,////OtinmarchicusAprilMayJuneJulyAugustFigure 5. Seasonal changes in the percent composition of Calamus fimmarchicuscopepodite stages on Flemish Cap in waters 400 m depth. This is a compositepicture of data collected in different years from 0. 165 mm mesh samples. Refer tothe text for details.220FREQUENCYPERCENT(%)-1).0)0).)0)C).0)0)C)o0000000000000C)C)00IIIIIIIIIIIII:1I”.)/-‘I’)______‘I..-J00I I-cm____________.ti.-.-......-.-•..•.•.•.•.•.•.•.•.a...L11‘.—,t—II:ii—i——&(.0—_____0)0).•::.:-:-:::•:•:•:-:•o1-‘Cl)-1 G) rn•.•::..•,•..•,.-‘—&F..).10)CD0F..)..0F..).0)0000000000()00000000IIIIIIIIII1III:1F.)10)0:0o::C.m—C0:-:1-(I_--I____________Figure 6. Development times predicted for Calanusfinmarchicusin different yearsbased on temperature dependent Belehradek equations. CI refers to copepoditestage CI, etc. * refers to date from which projections of stage development weremade.2221981a CuCI1980a CIII*a clva cv1t’.)a CI1979a CII*a CIII*a clva cvI CIIa CIa cliiLONGT’MEANa Civa cv*PEAKSPAWNINGNAUPLIICICIICIIICIVCVCOPEPODITESDAY90100110120MONTHAPRIL130140150MAY160170JUNE180Figure 7. Schematic representation of morphometric measurements made onredfish.224M U-’0 (0) 0 -‘ 0 I 0 0 I-a,I0 3 3 m -n Cl)(I) I-Figure 8. Comparison of redfish feeding for different cruises/dates on FlemishCap. Redfish were subdivided into one of 3 length groups: Small: 9 mm, Medium:1049 mm and Large: 20 mm. In addition, for each length group the mean sizeand number of fish measured is given. The prey item labels are: A - Copepod Eggs;B - Copepod Nauplii; C - Copepod Copepodites; D - Copepods Not Specified; E -Cyclopoid Eggs; F - Cyclopoid Nauplii; G - Cyclopoid Copepodites; H- CyclopoidNot Specified; I - Calamus finmarchicus Copepodite; J - Oithona spp. Copepodite; K- Limacina spp. ; L - Euphausiid; M - Phytoplankton; N - Prey Type Not Specified;o - Empty stomach.226rndIU4I(U(UF’1IHa.4I.’(U(Urn1’1>.I310sl0zI-I-3IIaa401IIx‘II.!IIa0C40zII.3H1.114II0El0zII.3IiaC)C401.31LIIIa0I40zII.314I.IIa0I4EE(Urn0I‘IC0zIJ.3H13034LILI3(LIzI-ItHa313.133I401I1.3H1.314.1IIa0I4JnE0I‘IC(UEm-IUC“III-COimc£0Nc’4—00‘I111’1’oo000•CVC00H(UEl’)ECU(Umc00•00‘I00VC00000CCVCFr eqUencyweight)byppp00000000p 0—•1• III&ma0p&pp00000000000I•.I.It\) [‘3 ODI r .1•0 D NI.1113113I H LUr I 1 011)3IIt’11)3Ia U NI’rn 0 HIIC.Ix Ir I’ Il.1.0II r1)I11303I 0 V NI’>1r ruCIx Ir IX Iz Tp a 0 NXl0 H1 CHIC ‘A IC IX Il joro 0a100 N1 >)I.<I NH C.Ix Ir IX lzI C I 0 N310(“4 ‘13 3‘I) I) II) C C r ID 0I) p 1 I I H C A r I z 0IID w H r ID 0Frequency(byweight)0aaa0000000 I11111117DHI.u I) p 1IaaI0aa•000000o00000II_IIII““u”iwIlli///1I/Iflh1i!//Ih/i/?/Il///I//1///I/,:1’D ah3‘13 aF I •1:igiI. W3II ‘13A r I z 0 >.a 01 0 I 17Ja I H’aI•0o00000—I•II1.1:1’4I.riiiiI3•1iciI a30 H71111113C I 13i:i12 IC)“In>Ia1C9aLI-lCaIraIz 10I >1 C.IiIx Iz 10 H a]0 0 aIi0 H CuIC. AIr II Ix,-) p 0 0 a 1LiVilDllhiflhhfflluhIlIIIglluhIIIA2 IilllwllnI. a 13 (j3 ‘1 I? 103 3II‘JO (I33I a 0 H C IMp 0 a 1p I >1 C. AIr Ix Iz lo ,-a 0 0 aHIIC.Ir Ix Iz toI) 100‘1a 1•<C.10AIr Ix Ix,lIu,,uhIuhI,i1,IIiflIi) a 0 CrC z•PTaa aI) a -C a aFrequencybyweight) a•oo00000I•I•T1 In I0 N_________I.•,,mammmimi0a03 3Ir Ix Iz TIw 010 I0 pI0xiiNatj 3Ix IriIx Iz 10 T>IInIi)IDJ)III ‘10N1111111111111111111111111111111c_______Ixprj0I Iz i.eFigure 9. Mean size of Limacima sp. (mm) observed in the diet of redfish duringdifferent times of the year for 1979, 1980 and 1981 data.2310.140.13 SO.12--0.1_i.III—zLU_j 0.09I0.080.070.06’____________________________________________________60 90 120 150 180 210 240DAY OF YEAR2 32Figure 10. Direct comparison of feeding differences for redfish larvae sampledin different years at approximately the same time: 20-24 March 1979 versus 6-13April 1980, 23-27 April 1979 versus 2-9 May 1981, 20-26 May 19890 versus 22-27May 1981. In each case the comparison is for all redfish as there were no significantfeeding differences among size classes within each cruise. The prey item labels are: A- Copepod Eggs; B - Copepod Nauplii; C - Copepod Copepodites; D - Copepods NotSpecified; E- Cyclopoid Eggs; F - Cyclopoid Nauplii; G - Cyclopoid Copepodites;H - Cyclopoid Not Specified; I - Calamus finmarchicus Copepodite; J - Oithoma spp.Copepodite; K - Limacina spp. ; L - Euphausiid; M - Phytoplankton; N - Prey TypeNot Specified; 0 - Empty stomach.233I—IC,LU>-co>-C-)zUIDaUILLI.ILII>->-C-)zUIDaUILLIIC,UI>->-C-)zLiiDaUI234A B C D E F G90807060504030201005040302010080706050403020100A B C 0 E F GA B C D E F GF000TYPEFigure 11. Diurnai patterns of redfish stomach weight (gDW) averaged for eachcruise, + 2 standard errors. Interpolation between each 2 hour period is smoothedusing a spline fit to the means. Sunset (dark shaded) and sunrise (open) are indicatedat the bottom of each panel.235STOMACHWEIGHT(ugDW)—10(..R.00 o00000 00000000_..: 0•0o—g.a/ o/-<I gSa10/10.g1’ •..—STOMACHWEIGHT(ugOW)1000010 00e0000010IIIIIIeB—o—,,) ,/s____0/—._-•__0o—/-.%‘\g0/,%..IN•a-II•1’ 0/faa’•‘—/STOMACHWEIGHT(ugDW)00tO0000oIa’ 0/1il0/0ba0a10tOg-‘—1010oO0P.)0e0/____—I•_—co0—.0—‘——•.S\•_0I.I‘:-----—-—I,/____•.\ 0—\a’•,B—P.)000STOMACHWEIGHT(ugOW)-‘1010 eQP.)0010I1oIIIIIa‘N 010-‘§-‘C8‘-Si)%510I.I=P30ô000000000010IIIIaI00/800-I\%•S..‘a\i•_.)8,(a’10‘a)/— ••-—__at0a00000I0000100a—I /Ia,C.) >a•FI.,C Ii0 b%‘S...-—S%•)• --/‘a) I /—‘a,‘S•\tO0ô’O000000IIIIIIIIII4-Imoc0800IMI.,1%)I!10-.4a’‘\-AI•,5Sa,—,——a.I—..— I—1/ i_/9EFigure 12. Redfish stomach weights standardized by redfish dry weight (+ 2 se)versus redfish length (mm) for 20-26 May 1980 and 22-27 May 1981.2370.6•20-26MAY,1980•22-27MAY,19810.5WLU IO.40WI—Ui N z ci)0.1-0.0W56789101’l1’21LENGTH(mm)Figure 13. Redfish feeding selectivity calculated for eight food types for 4 cruisesusing 0.080 mm mesh zooplankton samples. Mean selectivity is plotted with minimum and maximum values calculated for each food type within a cruise. (o =Chesson’s alpha index for prey species i). The dashed line is neutral selection forprey types (, where n is the number of prey items. The prey item labels are: A -Copepod Eggs; B - Cyclopoid Nauplii; C - Copepod Nauplii; D - Oithoma spp. Copepodite; E - Limacina spp. ; F - Euphausiid; G - Calanus finmarchicus Copepodite;H - Copepodites Not Specified; I - Empty stomach.239E 2-9MAY,1981-0;-.I-—-—?-2I C.) w-3-J Ui ci)-4.C!, E-66-13APRIL,1980LI...1AEF-:a.LtT1020-26MAY,19800—1-2-4,-5.-6-7. 0.-1-2-’-3 4.-7-M 0IIIBC0EFGHI..22-27MAY,1981FOODITEMAEFGHIBCDEFGHIFOODITEMFigure 14. Redfish feeding selectivity calculated for eight food types for 2 cruisesusing 0.165 mm mesh zooplankton samples. (a, = Chesson’s alpha index for preyspecies i). Redfish analyzed 26-30 June 1981 ranged in length from 9-15 mm and1-4 August 1982 from 16—23 mm. The prey item labels are: A - Copepod Eggs; B -Cyclopoid Nauplii; C - Copepod Nauplii; D - Oithoma spp. Copepodite; E - Limacinaspp. ; F - Euphausiid; G - Calamus fimmarchicus Copepodite; H - Copepodites NotSpecified; I - Empty stomach.241>-II0LU-JLUCl)0zLULUU-0--1 --:-2 ->- -3-I0LU-5--JLUCl)0zLULU-8 -U--926-30 JUNE, 1981G HI I I IA B C D E F0—1-2-3-4-5-6-7-8-9FOOD ITEM242Figure 15. Dry weight (pcg) versus redlish length (mm) regression residuals plotted against redflsh length (logjo mm). All cruises with dry weight data exceptDAWO79.243I I0.8 0.9p i p1.0 1.1 1.2I I1.3 1.4LENGTH (log 10mm)DCCC- o ocRiroRBCCCC0.80.70.60.5-J0.4Cl)w--——---—— 0.2I 0.1!oCCC0-QJCC0.6 0.7C C244Figure 16. Percent differences in dry weight (ag) and fish volume (mm3)between20-26 May 1980 versus 22-27 May 1981. % difference is calculated relative to 20-26May 1980.24560 50 40 30 20 10 0-10-20-30-40Figure 17. Fish volume (mm3) versus redfish length (mm) regression residualsplotted against redfish length (log10 mm). All cruise for which data was available.247LUD-J0>IC’)LL.I I I I0.8 0.7 0.8 0.9LENGTH (log10 mm)Cl,-JDC,,uJ0CCa0C0.50.40.30.20.0-0.2-0.3-0.40rDQC000 CI I I I I1.0 1.1 1.2 1.3 1.4 1.5 1.6248Figure 18. Metamorphosis of redfish from larvae to pre-juveniles determined bythe degree of fiexion at different lengths (mm) for a) 20-26 May 1980 and b) 22-27May 1981. Flexion stages 1 to 3 refer to pre-flexion, in flexion and post-flexion,respectively.249100908070w30201001009080700xLu 60-JLL50zL 4003020100138 7 8 9 10 11 12REDFISH LENGTH (mm)250Figure 19. Redfish growth rates (mm d’) versus water temperature (°C) averaged for the months March through July 1980 and 1981 on Flemish Cap. Solid linesconnect months within each year, dashed lines connect months between years.2510.18-JUNE.1980JULY0.17-/\/\\0.16—\\/\\0.15-/MAy”0.14-/‘/ /r0.13-/\MARCH::\PRIL’\113TEMPERATURE(°Q)Figure 20. Redfish growth (mm d1) versus feeding rates (jtg mm3) averagedfor each cruise on Flemish Cap in 1980 and 1981. The solid lines connect cruiseswithin each year, the dashed line connects cruises conducted at the end of May 1980and 1981.2530.16-015-0.14-G35,“H2!//\‘0.13-/LU/V/ /0.12-I./H-P00.11-/II G51G500.09—IIII0.080.130.180.230.280.330.380.430.48FEEDINGRATE(pgmrrr3)Figure 21. Redfish growth rates (mm d’) versus water temperature (°C) averaged for each cruise on Flemish Cap in 1980 and 1981. The solid lines connectcruises within each year, the dashed line connects cruises conducted at the end ofMay 1980 and 1981.2550.16 ::G70.13— :: 0.10-G50/0.09—IIIIII345678910111213TEMPERATURE(°C)Figure 22. Redfish feeding rates (tg mm3) versus water temperature (°C) averaged for each cruise on Flemish Cap in 1980 and 1981. The solid lines connectcruises within each year, the dashed line connects cruises conducted at the end ofMay 1980 and 1981.2570.46-AZ40.42-0.38-\c)20.34-2OD0.30G37H2w0.26-0 z0.22-w LU10LL.vo0.14-G510.10-G50IIII345678910111213TEMPERATURE(°C)Figure 23. Redfish growth rates (mm d’) versus prey concentrations (mg m3)sampled with 0.080 mm mesh nets for different cruises in 1980 and 1981. Dashedlines connect cruises within each year. a) Prey Type 1 — Copepod eggs, b) PreyType 2 — Copepod eggs and nauplii, c) Prey Type 3 — Copepod eggs and nauplii,Calanus fimmarchicus and Oithona spp. copepodites.2590.16G370.15G370.14-•G35G350.13-0.12--0.11-G-51G-51Si.-2IG-50A)COPEPODEGGSB)COPEPODEGGS&NAUPLII0.09-_____________________________________G-501.1.1IIIIIIIIIIQ0.050.070.090.110.130.150.170.190.210.00.10.20.30.40.50.60.70.80.91.01.10.16•G37C)COPEPODEGGS&NAUPLIID)1980-COPEPODEGGS&G37PLUSCALANUSNAUPLII0.15-FINMARCHICUS&OITHONA1981-CALANOIDEGGS&/SP.COPEPODITESNAUPLII0.14-G35•G350.13-0.12-0.11-G-51G-510.10-‘198‘—---—-J98iG-500.09-IIIIIIIIII123456789100.150.170.190.210.230.250.270.290.31PREYCONCENTRATION(mgm3)Figure 24. Redfish feeding rates (mg mm3)versus prey concentrations (mg m3)sampled with 0.080 mm mesh nets for different cruises in 1980 and 1981. Dashedlines connect cruises within each year. a) Prey Type 1 — Copepod eggs, b) PreyType 2 — Copepod eggs and nauplii, c) Prey Type 3 — Copepod eggs and nauplii,Calamus finmarchicus and Oithona spp. copepodites.2610.30-0.30—G37A)COPEPODEGGSG37B)COPEPODEGGS&NAUPLII0.25-0.25—I\ICI0.20-0.20-0) 10.15-0.15-IG51IG51-IC,,iA2.20.10-G350.10,-G35,G5OG50.IIIIIIIIIIIIIIIIItw0.050.070.090.110.130.150.170.190.210.00.10.20.30.40.50.60.70.80.91.01.10.30-0.30-•G37C)COPEPODEGGS&NAUPLHD)1980-COPEPODEGGS&G374PLUSCALANUSNAUPLIIWFINMARCHICUS&QITHONA1981-CALANOIDEGGS&0.25-ISP.COPEPODITES0.25-NAUPLIIII0.20-0’0.20-cr3!0)g‘—I II0.15-/015G51G51.I—-.-1°1I0.10-•G350.10--—_G50G501________________________________IIIIIIIIIIIIIII123456789100.150.170.190.210.230.250.270.290.31PREYCONCENTRATION(mgm3)Figure 25. Redfish growth rates (mm d1) averaged for each station versus theconcentration of copepod eggs and nauphi (log mg m3) for the cruises 20-26 May1980 (s- GADO37) and 22-27 May 1981 (. - GADO51). a) mean growth ratesaveraged over all ages, b) growth rates for the last 5 days of life, c) growth rates forthe last 10 days of life.2632640.04 -MEAN GROWTH0.03-0.02 -0.01 -a0.00 •a1 •q• •..-0.01--0.02 --0.03 -I I I I0.03 -• a 5-DAY GROWTH0.02 -F..o.oo- Ti.I ••a• •%I— I-0.01-••IIt•.. ... . .-0.02-..•-0.03-___________________________________________________I I I I I0.040.03 10-DAY GROWTH002-010. ..— •• •0.00 -___ I•.••• I •. • • •-0.01 - —a •-0.02--0.03-I I I I I-1 0 1 2 3 4COPEPOD EGGS & NAUPLII (1og10mg)m3)Figure 26. Redfish growth rates (mm d’) averaged for each station versus theconcentration of copepod eggs and nauplii (log mg m3) for the cruise 26-30 June1981 (HAWPAN). a) mean growth rates averaged over all ages, b) growth rates forthe last 5 days of life, c) growth rates for the last 10 days of life.2652660.04 - V0.03-VMEAN GROWTH0.02-0.01- . ...•. . .0.00-.•• •-0.01-•V-0.02--0.03 --0.04 -0.04 -5-DAY GROWTH0.03 -.. 0.02-••0.01 • -• •0.00- • •. • • •I • •I— • •.-0.01-O • •-0.02-•-0.03—-0.04—0.04 -10-DAY GROWTH0.03—••0.02—0.01- .• •.••••0.00- • • ••••••-0.01—• •-0.02—-0.03—-0.04--0 1 2 3VCOPEPOD EGGS & NAUPLII (1og10mg)m3)Figure 27. Growth rates (mm d’) calculated over 5-day periods for cohorts ofredfish released on different dates ranging from Days 90 to 140 on Flemish Cap in1980 and 1981. The periods of larval release and of seasonal decline in growth ratesare indicated at the bottom of each panel.267EELUI0026890 100 110 120 130 140 150 160 170 180 190 200 210 220-RELEASE DECIJNE1981ELUII00.20 -.18.16-.14-.12-.10-.08I I I I F F I 1 I I90 100 110 120 130 140 150 160 170 180 190 200 210 220I— I I Ir4-----------’iRELEASE DECLINEDAYFigure 28. Population abundances estimated for April-released redfish duringdifferent cruises, 1978-1983, on Flemish Cap. Values on lines are calculated rates ofchange (d1) between cruises in different years. Hatched area represents estimatedpeak release period of redfish on Flemish Cap.26930-,‘79/025-RELEASE82‘81/.‘7920IIIIIIIIII60100150200DAYFigure 29. IVIean redfish length (mm) calculated for the April-released redfishsampled during different cruises, 1978-83, on Flemish Cap. The line is the bestexponential fit for all values.27125-20182818081806090120150180210240DAYFigure 30. Relative estimates of the anticyclonic circulation of surface waters onFlemish Cap based on dynamic height calculations a) during different seasons forall available data, and b) for different years during this study, 1979—81.273910_._ . . •iI II I II I I I I I I I I I I30 60 - 90 120 130 180 210 240 —270 300 330 360AI I8762I I. — IB I II II9B87.62109- 1979*- 1980•- 1981I .1 I I I I I I I0 30 60 90 120 130 180 210 240 270 300 330 360IyoY274Figure 31. Relative abundances of redfish calculated at different times for the 3year classes, 1979-81. See text for data sources.2759LRELATIVEABUNDANCEopo01IIIIIIIIIo0-<mCl,..(D(D(D-OCD

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086664/manifest

Comment

Related Items