wnfiwhfly, . yrt.“ . ”f ,. 7 _ r4": v"“~'§‘23’h'w' ‘fix’h‘f figuwrfiwhhfigrw. m: - \r -. r: N3; '? at"- x - ‘ifirfi‘r": "‘ «5-1 ., 9‘, ...__ 31"“) :1"; w: L , .- .. -1 .,. ~17 A - ‘1‘;wa “(27.2?“ x. s; .. .-: ‘5 J. "‘13,? Lynn. _ ‘ >33. 39.3.:- -u,. ”-0 -\ a... M... Wu...“ _ gt - wfillu~"c.~ R ‘ .. » Nl-A~\~‘\ .' v; 1‘. L344 .1. ‘ j J' ‘ an}; 4 R :lU‘ \ .4.ch- ‘ 3 V ,n ’ ‘ my (I): . ‘ q .aq ' AV Mi 1: 4 ‘Z’ml’ " "’KEVS‘J :2 r "‘ 1&7 4 r0“ ? —.-n.: 3-. .- ‘ '1 "cu --A Inn guy-um. unu‘fifu “sy‘ - ‘1." g 4 4-. . .r 1‘ -. £5: ‘ "Ix-.12. .. 3:" 1 «4 Z diir s'U M. s. , w... 9:4 #1,»: «fl .. .. ¢ ‘ v a «44 ‘ .A' ‘L um; ‘I‘IL ‘ 9 ‘ J; } a?“ ' h...“ ka‘y' L ' -. u: g.‘ "37" .. “i‘ . ... 7 r , . ‘ -- ’ . > mxi» l' _ : ”J.- . . \)1 h. .. . -v.. r: - ”(guru ; ...»: ' - ' - " ». _., .. ‘ . ‘4. N v I ' 1. J [v .5 4r ‘ , _ 4. ,Jcflawm (z - ‘- 1“ L . . I, .. 1‘1.’ 'L‘n‘ -c Jarr- . ‘sb'y‘v L, . 3g“... . . w; .‘r - ww’lvwgai, . {aflm'rrzjflad~ "any.” “ I.) p1 w... {I} ‘7- 3,"? ' r Wvflzfi .. aw, , a \i’ “Egan'rz‘w . W1”; n g'b-rtl ff”; 4 m. N v ”‘9'? .._..v v.' 1 ,1. .M , JEN . é ', . t C r I! " '1 .4. L: ’6 .r .J. M: II. 'i 1" f. 2 . n H J.’ K . r t C ‘ ‘_ r 1 '. 1 , . l' y u w Ju—J.‘ ‘ . . ,. 2,5: I . . ; _‘ . ‘ ‘ - . '. ~ . ' -. - "r, r1,“ ‘1" .r' w}. ,? . SEE-1&4. . I ‘ _ o ‘ - - v ' 7’ ’l _. 1‘] ,5; -_ r - {1 m ‘1' _ w .' , ~ . 7 , v. I. q: ,‘_ , . . '1 ‘ f. r, V .Y".,;_"~ , A , . . 3“ . . ..~: ,, . ’7 ,‘ 'V"" n". N.“ | , v , ., y.;';;~ ,. , ’ ~., . ,,. -~ “—4,, .~ ' . ‘ . . -‘. y y . _ - .- r x _ V .1? 9"} r" _ , ,. , 7‘53: - , . , 'T _ , I . ‘ . - . I. ‘. 31-2333". ”3”“ T ‘1' . '1'“ '1' f] ,"'1’|,'V"r‘ .‘ , ‘ ‘ V , ,' - ' ‘ _ > ' ‘ I 11’ L,“ R. 2122! J“, p .. 4'3?"- . ’z’m‘r:v:rfl , v ‘ . , . 1 _. , ‘ . verve-7' a Rf,” 5|, , . ‘ 1.6.0:” 4 u. -Aa‘...’ I." 25'.“ . . , ’ . ‘ V .‘ ‘ » u n, . . . .. C‘r‘éfilut. - a ‘ v. .u . J. ~c r‘" r"-v I .-...o. I. ulnar: "a. . a. ,::,.‘,;a..~.:.~‘:. '4... v‘tf'llw' ”4”» .r. ~mlwur—‘Ol . . V. . . r »v(-a . .w-..-. '- 1.0:,» .rm "bx-4'14. We. ‘ I "1: "3f. 4‘, r I. y" . 3:14 '57 ..g~ .. u v 1’ .- (_ t. ‘ ~ ’- "’J,.r.a..:.':a.._.. gun..." r 4...‘ ' ‘- Inu" 5» .5. liW/llliflifllllIil/IllIl/l This is to certify that the thesis entitled Factors Influencing the Larval Survival and Recruitment of Lake Whitefish in the Upper Great Lakes presented by Russell W. Brown has been accepted towards fulfillment of the requirements for M.S. degree in Fisheries /Z/a /// 94 Major profesy [hue April 12, 1991 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY I MIchIgan State Unlverslty PLACE IN RETURN BOX to remove We checkout from your record. TO AVOID FINES return on or betore date due. [5 an ATE DUE DATE DUE DATE DUE 091914 , *' __J MSU Is An Affirmative ActioNEqual Opportunlty InaItutIon chna-m FACTORS INFLUENCING THE LARVAL SURVIVAL AND RECRUITMENT OF LAKE WHITEFISH IN THE UPPER GREAT LAKES BY Russell W. Brown A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Fisheries and Wildlife 1991 J 5’ 7 l N 631—5 \J‘K ABSTRACT FACTORS INFLUENCING THE LARVAL SURVIVAL AND RECRUITMENT OF LAKE WHITEFISH IN THE UPPER GREAT LAKES BY Russell W. Brown Eggs were collected from two stocks of lake whitefish in Lakes Michigan and Huron to assess the effect of egg composition and prey density on larval growth and survival. Larval fish hatched from six parental females in each stock were fed one of four rations (0, 18, 24, 50 brine shrimp/day) after yolk-sac absorption. Length at hatch, endogenous growth, exogenous growth, and survival were measured during a 42-day laboratory experiment. Length at hatch of larvae was positively related to egg caloric content (R2== 0.780). Endogenous growth for lake whitefish larvae was positively related to percent lipid content (R2 0.896) and total egg lipid content (R?== 0.876) of parental females. These results indicate that egg composition has the potential to significantly influence the growth and survival dynamics of larval lake whitefish. Spawning stock and climate-based recruitment models were developed for lake whitefish in northern Green Bay and the North Shore areas of Lake Michigan. Beverton-Holt and Ricker stock—recruitment functions, ice cover concentrations, wind intensity data, and spring water and air temperature variables were used as model inputs in climate—based recruitment modelling. The climate-based model for northern Green Bay, modelled lake whitefish recruitment as a function of a Beverton-Holt stock- recruitment function, ice concentration, and wind intensity. The climate-based recruitment model (R2 = 0.65) demonstrated improved hindcasting ability when compared to the Beverton- Holt stock-recruitment model (R?== 0.37) for the 1961-1985 cohorts. The climate-based model for the North Shore modelled lake whitefish recruitment as a function of a spring air temperatures, ice concentration, the Beverton- Holt stock-recruitment function, and wind intensity. The climate-based recruitment model (R2 = 0.57) demonstrated improved hindcasting ability when compared with Ricker stock—recruitment model (R2 =- 0.13). Results of this study indicate that climate-based recruitment models have the potential to more accurately forecast lake whitefish cohort strength several years in advance of recruitment into the fishery. ACKNOWLEDGMENTS This study was sponsored by the Michigan Sea Grant College Program (Project Number R/GLF-l9 under Grant Number NA86AA-D-SG043, and Project Number R/GLF-35 under Grant Number NA89AA-D-SG083) from the Office of Sea Grant, National Oceanic and Atmospheric Administration (NCAA), U.S. Department of Commerce, and funds from the State of Michigan. Sue Walker (Grand Traverse Band of Chippewa and Ottawa Indians), Theron King (King's Fisheries), and Forrest Williams (Bayport Fisheries) provided adult lake whitefish for egg samples utilized in the laboratory portion of this study. I also thank student interns Brian Jonckheere, Melissa Treml, Joe Mion, Heather Sysack, and Steve Hero for their assistance during the laboratory studies. Eva Moore of the National Fisheries Research Center, Great Lakes, was especially helpful in obtaining commercial catch and effort data. Raymond Assel of the Great Lakes Environmental Research Laboratory provided invaluable assistance and advice during the climate modelling portion of this thesis. I am indebted to fellow students John Kocik, Dan Hayes, and Bob Sluka for their insightful suggestions and recommendations during several phases of this study. iii I thank the members of my committee, Dr. Richard Hill and Dr. Niles Kevern for their critical reviews of this manuscript. I especially thank Dr. William Taylor for his guidance, patience, and friendship during my graduate career. I would like to express my sincere appreciation for the love and patience of my family (Mom, Dad, and Susan) and my family in Nebraska (Bernie, Sharlene, Sarah, and Nancy) during my education. And finally, I am especially grateful for the love and support of my wife, Amy. iv TABLE OF CONTENTS EASE List Of Tables. OOOOOOOOOOOOOOO O ........... O O O O O O O O O O O O Viii List Of Figures 0 O O OOOOOOOOOOOOOO O ........... O O O O O O O O O O O x List of Appendices..................................... xiii Chapter 1: Effects of Egg Composition and Prey Density on the Larval Growth and Survival of Lake Whitefish Abstract.. ................................ . ...... . 1 Introduction. ......................... ............ 2 Methods........................................... 6 Egg Collection and Incubation................ 6 Laboratory Analysis of Egg Composition....... 8 Laboratory Experimental Design............... 10 Measurement of Survival and Growth........... 13 Data Analysis................................ 14 Results........................................... 17 Characteristics of Sampled Female Lake Whitefish.......................... 17 Egg Composition.............................. 17 Length at Hatch of Larval Lake Whitefish..... 20 Endogenous Growth Rates ............. ......... 22 Exogenous Growth Rates....................... 27 Larval Survival Rates........................ 31 Exogenous Growth Model ....... ................ 34 Discussion......... ....... ........................ 34 Egg Composition.............................. 34 V Tradeoffs Between Egg Size, Egg Number and Egg Composition..................... Point of No Return........................... Relative Importance of Temperature........... Conclusions....................................... Chapter 2: Climate-Based Recruitment Models of Lake Whitefish in Two Areas of Northern Lake Michigan................................... Abstract....................... ......... .......... Introduction...................................... Hypotheses........................................ Methods........................................... Model Construction 0verview.................. Commercial Catch Data and Study Areas........ Commercial Fishery Gear and Effort........... Indices of Spawning Stock and Recruitment.... Stock-Recruitment Relationships.............. Ice Cover Models............................. Wind Speed Data.............................. Spring Water and Air Temperatures............ Climate-Based Recruitment Model Construction. Results........................................... Stock-Recruitment Relationships.............. Northern Green Bay Climate-Based Recruitment Models...................... North Shore Climate-Based Recruitment MOdeISOOOOOOOOOOIOOOOOOOIO_OOOOO0.0..0... vi 37 39 4O 42 43 43 44 48 49 49 51 S3 55 60 61 63 66 67 67 67 72 DiscussiODOCCOOOO0.0...O...OOIOOOOOOOOOIOOOOOOOOOO Relationship between Spawning Stock and Recruitment.......... Climate-Based Recruitment Modelling.......... Other Factors Affecting Recruitment.......... COHClUSionSOo ....... 00...... ....... 0.0.0.0... Literature Cited...... Appendix 1..... ............... ......... Appendix 2........... vii 74 74 78 79 81 83 90 91 Table Table Table Table Table Table LIST OF TABLES Mean length, weight, and age data of all sampled parental females sampled from Bayport (Lake Huron) and Naubinway (Lake Michigan) in November 1989, and parental females of eggs used in the laboratory studies................. ........ . 18 Mean values, standard errors, and ranges of egg composition parameters of female lake whitefish sampled from stocks at Naubinway (n = 14) and Bayport (n = 17) in November, 1989........................... 19 Results of T-Tests and approximate T-Tests to evaluate differences in egg composition parameters between lake whitefish sampled from Bayport and Naubinway.................. 21 Results of regressions between length at hatch of lake whitefish larvae and egg composition parameters of eggs sampled from Bayport and Naubinway. Analysis of covariance was used to determine whether combined or separate regression equations for each stock were used.................... 23 Results of analysis of variance to test for differences between stocks, females, and feeding rations in the endogenous growth, exogenous growth, and survival of larval lake whitefish under laboratory conditions.. 26 Results of regressions between endogenous growth rate of lake whitefish larvae and egg composition parameters of eggs sampled from Bayport and Naubinway. Analysis of covariance was used to determine whether combined or separate regression equations for each stock were used.................... 28 viii Table Table Table Table Table Table 10. 11. 12. Standardized units of effort for the commercial fishery in the Great Lakes as defined by the Great Lakes Fishery Commission (Hile 1962)...................... Percent age composition of lake whitefish sampled in area MM-l during the period 1960-1987. Data sources are shown for each yearOOOOOO0....OOOOCOOOOOOOOOOOOOOOO0.0...O. Percent age composition of lake whitefish sampled in area MM-3 (North Shore) during the period 1960-1987. Data sources are shown for each year......................... Results of ice cover modelling for areas MM-l (northern Green Bay) and MM-3 (North Shore). The averaging periods used to produce each model, R2, root mean square errors (RMSE) and number of observations used in model calibration each winter classification within each area...OOOOO0.0.00IIOOOOIOOOOOOOOOOOOOOCC... Results of Beverton-Holt and Ricker stock- recruitment relationships for lake whitefish in northern Green Bay and the North Shore, Lake Michigan using abundance and year class indices from 1961-1985....... Parameter estimates and squared partial correlation coefficients for climate-based recruitment models of lake whitefish in areas MM-I and MM-3......................... ix 56 58 59 64 68 7O Figure Figure Figure Figure Figure Figure Figure LIST OF FIGURES BASE Map showing sampling locations of spawning lake whitefish in Lakes Michigan and Huron........................ 7 Diagram of the experimental design for the laboratory experiment..................... 12 Mean length of hatching larval lake whitefish as a function of their egg caloric content (cal/egg)................. 24 Mean length of hatching larval lake whitefish as a function of their lipid content (mg/egg)COO...0.000.000.0000.0.000.000.0000 25 Endogenous growth rates (u) of larval lake whitefish during the first 21 days of the laboratory experiment as a function of percent lipid content...................... 29 Endogenous growth rates (u) of larval lake whitefish during the first 21 days of the laboratory experiment as a function of lipid content (mg/egg)..................... 30 Endogenous and exogenous growth rates of larval lake whitefish during the 42 day laboratory experiment. Endogenous growth rates were not significantly different from each other, and exogenous growth rates are expressed as a function of prey rations. Error bars represent 1 one standard error...................... 32 Figure Figure Figure Figure Figure Figure Figure 10. 11. 12. 13. 14. Endogenous and exogenous survival rates of larval lake whitefish during the 42-day laboratory experiment. Endogenous survival rates were not significantly different from each other, and exogenous survival rates are expressed as a function of prey rations. Daily error bounds represent i one standard deviation........ 33 Specific exogenous growth rate (u) of larval lake whitefish during weeks 4-6 of the laboratory experiment as a function of prey ration. Solid line represents the fit of the Monod threshold-corrected equation to predict specific growth rates as a function of prey ration.................................... 35 Commercial fishery landings of lake whitefish in Lake MiChigan' 1900-1989000000000000000 45 Flow chart of lake whitefish recruitment modelling showing the model inputs, intermediate model parameters, intermediate models and analyses, and final recruitment models produced for each area............. 50 Location map of Lake Michigan showing the two areas for which recruitment models were developed. The recruitment model for northern Green Bay included the entire area within GLFC statistical district MM-l. The recruitment model for the North Shore included the shaded region in the northern portion of GLFC statistical district MM-3. 52 Commercial fisheries landing of lake whitefish in GLFC districts MM-l and MM-3 in Lake Michigan from 1930-1989........... 54 Recruitment model hindcasts and actual cohort abundance index values for the 1961-1985 cohorts of lake whitefish in northern Green Bay, Lake Michigan......... 71 xi Figure 15. Figure 16. Figure 17. Plot of residual values between recruitment model hindcasts and actual cohort index values for the 1961-1985 cohorts of lake whitefish in northern Green Bay, Lake Michigan............................. 73 Recruitment model hindcasts and actual cohort abundance index values for the 1961-1985 cohorts of lake whitefish along the North Shore, Lake Michigan............ 75 Plot of residual values between recruitment model hindcasts and actual cohort index values for the 1961-1985 cohorts of lake whitefish along the North Shore, Lake Michigan............................. 76 xii Appendix 1: Appendix 2: LIST OF APPENDICES Length, weight, and age data of female lake whitefish sampled from Bayport and Naubinway and data for those fish whose eggs were used in laboratory experiments................. Egg composition data for female lake whitefish sampled from Bayport and NaubinwayooOI.IOOIOOOOIOOOOOOOOOOOOOOO. xiii 89 90 CHAPTER 1 Effects of Egg Composition and Prey Density on the Larval Growth and Survival of Lake Whitefish ABSTRACT Eggs were collected from two stocks of lake whitefish in Lakes Michigan and Huron to assess the effect of egg composition and prey density on larval growth and survival. Egg composition parameters including wet weight (mg/egg), dry weight (mg/egg), percent water, caloric content (cal/egg), caloric density (cal/mg), percent lipid content, and lipid content (mg/egg) were measured. Fish hatched from six parental females in each stock were fed one of four rations (0, 18, 24, 50 brine shrimp/larva/day) after yolk- sac absorption. Length at hatch, endogenous growth, exogenous growth, and survival were measured during a 42-day laboratory experiment. Length at hatch of larvae was positively related to egg caloric content (R?== 0.780). Endogenous growth of larvae was positively related to percent lipid content (R?== 0.896) and total egg lipid content (R?== 0.876) of parental females. Exogenous growth and survival of larval lake whitefish was positively related to prey availability. Larval growth was accurately modelled (R2 =- 0.973) as a function of prey ration using a threshold- corrected hyperbolic equation. These results indicate that egg composition has the potential to significantly influence the growth and survival dynamics of larval lake whitefish. 2 INTRODUCTION Mortality rates experienced during the pelagic larval stages are important in determining year class strength in many fish species (Gulland 1965; Chenoweth 1970: May 1973). For many species, larval mortality is concentrated during a relatively short period during early development (Hjort 1914; May 1973). Because mortality experienced during the egg and larval stages is severe, strong year-to-year variations in recruitment may arise from relatively small fluctuations in egg and larval survival rates (Taylor and Freeberg 1984: Freeberg et al. 1990). A number of abiotic factors (i.e. temperature, salinity, and currents) and biotic factors (i.e. starvation, competition, and predation) have been proposed as possible mechanisms controlling recruitment of larval and juvenile fishes. The ability of larval fish to cope with adverse biotic and abiotic factors is dependent upon a number of factors including body size, available energy resources, swimming ability, and reactive distance (Miller et a1. 1988). Body size is an important variable influencing many aspects of the physiology, ecology, and behavior of larval fish (McMahon and Bonner 1983; Peters 1983; Calder 1984; Miller et a1. 1988). The ability of smaller larvae to feed successfully is constrained because they have shorter 3 reaction distances and restricted swimming abilities limiting their searching ability (Blaxter 1986; Webb and Weihs 1986). Body size also influences the susceptibility of larval fish to predation through differential encounter rates and escape ability (Blaxter 1986; Zaret 1980). Swimming ability, largely determined by body size (Webb and Weihs 1986; Miller et al. 1988), also influences the ability of larval fish to maintain their position in optimal areas for growth and survival. During the period following hatching, initial feeding by larval fishes appears to be crucial for survival (Einsele 1963; O'Connell and Raymond 1970) as larvae switch from endogenous to exogenous food resources (Blaxter 1965: O'Connell and Raymond 1970; Blaxter 1971). Survival of larvae through the transitional period from endogenous to exogenous food resources is dependent upon a number of factors including the availability of suitable prey resources. Small increases in body size at hatching or available endogenous energy resources may result in large benefits in terms of the "window of opportunity" when first feeding can occur (Miller et al. 1988). Delayed initial feeding and subsequent mortality can occur if prey density is too low (Lasker et al. 1970; Beers and Stewart 1971), or if an inadequate prey species composition is present (Einsele 1963: Fluchter 1980). The "starvation hypothesis" only focuses on the extreme 4 case when the energy balance remains negative long enough for the larval fish to die (Fritz et al. 1990). Smaller larvae within a species generally have less endogenous reserves and thus are more vulnerable to starvation (Hunter 1981). Delayed initial feeding reduces searching ability and increases larvae susceptibility to predation (Ivlev 1961; Blaxter 1969). Larvae exhibiting slow growth also remain in a size range more vulnerable to predators for a longer period of time. The impact of exogenous prey resources on larval fish growth and survival depends upon the availability of endogenous energy resources. For example, a positive relationship between egg size and larval survival and growth has been demonstrated for arctic char, Salyelings alpingg (Wallace and Aasjord 1984); brown trout, Sglmg tzgtta (Brown 1946; Bagenal 1969); Chinook salmon, gnggxnynghng tshawytscha (Fowler 1972); and rainbow trout, Q; mykigg (Gall 1974). The ability of larvae to survive during starvation was positively related to the yolk content of eggs in trout (Gray 1928; Smith 1958) and herring (Blaxter and Hempel 1963; Schnack 1981). The relative amount of endogenous food reserves may be important in determining larval survival, especially if the production of initial food sources is mismatched in time with the initiation of exogenous feeding by larval fish (Cushing 1982). Once initial feeding has occurred, the ability of 5 larval fish to grow rapidly largely influences their survival. Availability of prey resources (Einsele 1963; Lasker et al. 1970; Beers and Stewart 1971; Fluchter 1980) and water temperature (Miller et al. 1988) are the two ecological resources most important in determining the growth of larval fish once their endogenous food resources have been exhausted. In addition, as larvae grow they become better able to visually detect both food and predators resulting in higher survival rates (Miller et al. 1988). Thus, the quantity and quality of exogenous food resources can significantly effect survival of larval fish and ultimately their year class strength. To evaluate the relative importance of endogenous versus exogenous energy resources on the population dynamics of lake whitefish, a laboratory study was designed to determine the influence of egg composition and prey density on the larval growth and survival of lake whitefish. Specific objectives of this study were to: 1. Evaluate the relationship between egg composition measures and the length of batch of larvae hatching from individual female lake whitefish from two stocks in the upper Great Lakes. 2. Evaluate the relationship between egg composition measures and the endogenous growth of larval lake whitefish. 6 3. Determine the influence of zooplankton prey densities and egg composition on the larval survival and growth of lake whitefish. METHODS qu Collection and Incubation Eggs were collected from two stocks of lake whitefish, one in Lake Michigan and the other in Lake Huron (Figure 1). Eggs were collected from 14 ripe female fish at Naubinway (Lake Michigan) on November 6, 1989. Eggs were collected from 17 ripe female fish at Bayport (Lake Huron) on November 14, 1989. In each case, female fish were manually stripped into a dry bowl. The unfertilized eggs from each female were mixed and a random sample of the unfertilized eggs was collected and frozen for subsequent laboratory analysis of egg composition parameters. The remaining eggs from each female were fertilized with milt from one male fish using dry fertilization methods (Piper et al. 1985). The fertilized eggs were then rinsed in lake water, allowed to water harden, and transported back to the laboratory within six hours of fertilization. Fertilized eggs from each female were incubated separately in a Heath-Tech 8-tray incubator. Eggs were treated with a 1:600 solution of formalin for thirty minutes on a weekly basis to control fungal growth. Fertilized eggs Nauljlnway O oooooo 'QOOQ---------.-------- -- I O Lake Figure 1. Map showing sampling locations of spawning lake whitefish in Lakes Michigan and Huron. 8 were incubated at an average temperature of 3.4.0 C (i 0.3 °)and hatching occurred between 84 and 101 days (Y = 94 days) after fertilization. Length at hatch was measured for a sample of 100 larvae hatching from each of the 31 parental females sampled during egg collection. Laboratory Analysis of qu Composition To obtain measures of egg composition for each female, unfertilized eggs were analyzed in the laboratory to determine egg wet weight (mg/egg), egg dry weight (mg/egg), percent water, egg caloric content (cal/mg and cal/egg), and egg lipid content (percent composition and mg of lipid/egg). To determine mean egg wet weight, three replicate samples of 100 eggs were weighed from each female. Each of these samples was dried in a forced air oven at 60° C for 48 hours and reweighed to obtain dry weight measurements for each female. Percent water measurements were calculated using the following equation: (wet Weight ng)-Lujrwedght (mg)) Percent water- . WEt welght ng) * 100% The remaining egg sample from each female was dried in a forced air oven at 60° C for 48 hours in preparation for measures of energy content. Two major types of measures of energy contained within eggs were utilized in this study. Egg caloric content (cal/egg) is a measure of the total amount of energy 9 contained within the egg, and is primarily a function of egg size and composition. Egg caloric density (cal/mg) measures energy content independently of egg size, and is primarily a function of the proportion of carbohydrate, lipid, and protein contained within the egg (Potts and Wooton 1984). Egg caloric density is most highly correlated with the lipid content, the most energy dense of the three major types of compounds normally found in fish eggs (Potts and Wooton 1984). Caloric content was measured using a Parr adiabatic calorimeter. Three sub-samples of dried eggs weighing 1.0 g each were analyzed for each female. The following equation was used to calculate the caloric content of each sample: H _ AtarW—eI-e2 9 m H9 = gross heat of combustion (cal/mg): At = change in temperature; W = energy equivalent of calorimeter in cal per degree Celsius e1 = nitric acid (HNO3) formation correction: e2 = wire heat of combustion correction; m = sample weight (mg). Egg samples from individual females were pulverized with an electric grinder and the lipid content was 10 determined for three 1.0 g subsamples. Lipids were extracted by the diethyl ether extraction method using a Tecator Soxtec System HT 1043 Lipid Extraction Unit. Lipid content of each sample was calculated using the following formula: (Initial Wt (mg) - Extracted Wt (mg)) P rcent.L' 'd- . . . e um In1t1a1 Welght (mg) a 100% Labo ato Ex e ' a e Eggs from six female fish in each stock were selected to produce larval lake whitefish that were utilized in the laboratory growth experiments. Once hatching began, egg containers were examined and hatching larvae were removed daily. A sample of 100 newly hatched larvae from each female was measured to determine a mean length at hatch. Larvae hatched from each female were divided among four treatment groups of 100 larvae each (Figure 2). Larvae hatching after similar incubation times were selected to avoid the confounding effects of differential incubation times on the length at hatch and endogenous growth rates of larvae. Each treatment group was assigned one of four brine shrimp feeding rations (0, 18, 24, and 50 brine shrimp/fish). Feeding was conducted on a prey/larvae basis 11 to maintain constant feeding levels while the number of remaining larvae changed through the course of the experiment (Taylor and Freeberg 1984). Each experimental unit was fed the corresponding prey ration twice daily, and prey rations were adjusted for changes in the number of larval fish on a daily basis. Dead brine shrimp were siphoned from the bottom of each tank on a daily basis. A large recirculating water bath was used to maintain constant temperatures (6.9° : 0.6) between experimental aquaria. Twenty liter aquaria initially containing 100 lake whitefish larvae were randomly assigned positions within this water bath. Water temperatures within individual aquaria were initially measured daily, and the maximum difference in water temperature between aquaria never exceeded 0.3°. Yolk-sac absorption occurred between 23 and 25 days posthatch for all larval lake whitefish observed in this study. Larvae were not fed for a period of 21 days after hatching to obtain accurate measurements of growth based on endogenous energy resources. Beginning on day 21, larvae were fed 2-day old brine shrimp (Artemia sp.), which were cultured continuously in the laboratory. As such, growth of larval lake whitefish occurring between day 0 and day 21 of life represented the endogenous growth period, while growth occurring between days 21 and 42 of life represented the exogenous growth of the larvae. 12 Great Lakes Whitefish e l I I ock Bayport Naublinway Mathias Emmi llllll llllll Feeding Ration ”8 24 0 am a 886868;”; mane: na na 13 5 me t f 'v G wt Survival of larvae was measured by removing and recording dead individuals from each tank on a daily basis. Larval growth of lake whitefish was measured by sampling and measuring the remaining individual larvae on a weekly basis. Larval fish sampled for growth were photographed on the date of hatch and weekly thereafter using slide photography. The fish from each experimental unit were isolated in a grided petri dish and photographed using a standard focal distance on a light table to enhance the resulting image. Slide photographs were then projected onto a Summagraphics digitizing pad and the images of each larval fish were digitized to determine length. A BASIC computer program was written to calibrate the scaling of the digitizing pad with a known scale place on the bottom of each petri dish at the time of photographing. Initial evaluation of this procedure demonstrated that straight objects and curved objects with less than three curves could be accurately measured to within i 0.02 mm. Images of larval fish which either were not horizontal in the photograph or were obscured by the edge of the petri fish were not measured using the digitizing program. This procedure allowed for the accurate measurement of large numbers of living larval fish (approximately 38,000 measurements) with little subsequent mortality (17 fish). 14 W Mean, variance, and standard error estimates were calculated for egg composition parameters measured for eggs from each spawning female. Mean egg composition estimates were tested to detect differences in these estimates between stocks using T-Test analysis. Because non-homogenous variances were detected between stocks for some parameters, an approximate T-statistic was calculated and used for hypothesis testing in these cases (Cochran and Cox 1950: Lee and Gurland 1975). Mean lengths of larval lake whitefish were used to calculate instantaneous daily growth using the following equation (Ricker 1975): (log e1c - log .1 c_1) T u = instantaneous daily growth; lt = mean length (mm) in week t: 1t,1 = mean length (mm) in week t-l; T = time interval in days. Regression analysis was used to test for significant differences between egg composition parameters, length at hatch, and endogenous growth of larval lake whitefish. Analysis of covariance was used to test for significant differences in the regression relationships between stocks. 15 For statistical analysis, survival rates of larval lake whitefish were calculate on a weekly basis. Mean weekly survival rates were calculated using the following equation (Ricker 1975): S = mean weekly survival rate Nt = number of larvae alive at week t N = number of larvae alive at week t-l Three stage analysis of variance was used to test for differences in exogenous growth and survival between feeding rations, progeny of females within stocks, and stocks of lake whitefish. A threshold-corrected hyperbolic growth equation (Freeberg et al. 1990) was calculated as a function of zooplankton feeding rations to model specific growth rates of larval lake whitefish during the exogenous growth period (weeks 3-6). Instantaneous growth rates were plotted as a function of prey ration (zooplankton/larvae) using a Monod threshold-corrected hyperbolic equation for substrate- limited growth (Freeberg et al. 1990): 16 pm, [(z/f) -(z/f) ql " [KB-(Z/f) ,1 + [(z/r) - (z/f) ,1 u = instantaneous daily growth; umax = maximum instantaneous daily growth; (z/f) = prey ration (zooplankton/fish): (z/f)q== maintenance prey ration: and Ks = prey ration producing one-half maximum growth. Maintenance substrate level was estimated from the plot and calculated constants of u" and K‘ using the regression X equation: [(z/rI—(z/fl .1 -( La ).(..1....) [(z/f)-(z/f) ,1 Pm II‘ max 17 RESULTS Characteristics 9f Sampied Femaie Lake Whiiefish Length, weight, and age of sampled female lake whitefish are shown in Appendix 1. Female lake whitefish sampled from the Bayport stock were longer'(I; = 569 mm vs. I." = 537 mm) and heavier (W, = 1713 g vs. "fin = 994 9) than females sampled from the Naubinway stock (Table 1). I found no significant differences in the mean age of female lake whitefish between stocks (Randomization Test, P = 0.1910). To eliminate differences in parental female size as a confounding factor during feeding trials, I selected eggs from a similar range of parental female sizes from each stock. As such, there were no significant differences between stocks in the mean length (Table 1, P = 0.9859) or age (Randomization Test, P = 0.4328) of the parental females of larvae used in the feeding trials. Egg Composiiion Mean values and variance estimates for egg composition parameters including wet weight, dry weight, percent water, cal/mg, cal/egg, percent lipid, and mg lipid/egg were calculated for eggs collected from each female spawner in each stock (Appendix 2). Mean values, standard error estimates, and ranges of egg composition parameters for each stock are shown in Table 2. With the exception of caloric density, the female fish sampled from the Naubinway stock 18 Table 1. Mean length, weight, and age data of all sampled parental females sampled from Bayport (Lake Huron) and Naubinway (Lake Michigan) in November 1989, and parental females of eggs used in the laboratory studies. Field Ba 0 Naubinway Mean Length (1 SE) 569 i 9.3 537 1 13.0 Mean Weight (1 SE) 1713 i 92 1341 i 103 Mean Age 5.9 5.3 Laboratory Barrett. Fascism Mean Length (1 SE) 557 1 19.7 557 i 19.3 Mean Weight (1 SE) 1592 i 184 1514 i 180 Mean Age 5.8 5.8 19 Table 2. Mean values, standard errors, and ranges of egg composition parameters of female lake whitefish sampled from stocks at Naubinway (n = 14) and Bayport (n = 17) in November, 1989. Ba ort Wet Weight (mg/egg) Mean 9.32 SE 0.24 Range 7.49-11.03 Dry Weight (mg/egg) Mean 2.66 SE 0.10 Range 2.19-2.98 Percent Water Mean 71.27 SE 0.44 Range 66.92-75.20 Caloric Content (cal/egg) Mean 17.23 SE 0.39 Range 14.02-19.68 Caloric Density (cal/mg) Mean 6.47 SE 0.03 Range 6.21-6.74 Percent Lipid Mean 19 . 58 SE 0.63 Range 15.44-23.43 Lipid Content (mg/egg) Mean 0.524 SE 0.024 Range 0.372-0.697 Naubinway 8.24 0.34 5.11-10.83 2.61 0.55 1.68-3.19 68.15 0.82 63.67-75.85 17.13 0.69 10.78-21.27 6.56 0.03 6.40-6.73 19.09 0.70 13.80-23.98 0.503 0.033 0.285-0.732 20 had larger ranges and standard errors for measured egg composition parameters. Results of T-Tests and Cochran's T-Tests for non- homogenous variance to test for significant differences in egg composition parameters are shown in Table 3. Eggs sampled from Bayport female lake whitefish had a significantly higher mean wet weight (WB = 9.32 mg) than eggs sampled from Naubinway females (R;== 8.24 mg, Table 3, P = 0.016). However, I found no significant differences in the average dry weight of eggs sampled from female lake whitefish in these two stocks, indicating that the stock differences in wet weight were a function of the significantly higher water content of eggs sampled from Bayport lake whitefish (Table 3, P = 0.006). Eggs sampled from Naubinway lake whitefish had a significantly higher caloric density (§;== 6.56 cal/mg) than eggs sampled from Bayport lake whitefish (i§== 6.47, Table 3). However, there were no significant differences in the caloric content of eggs (cal/egg) or the lipid content (%) of eggs from lake whitefish sampled from the Bayport and Naubinway stocks. Len th a atc of Larval Lake Wh'te '5 Least squares regression was used to test for significant relationships between egg composition parameters -and length at hatch of larval lake whitefish. When significant relationships were indicated, analysis of Table 3. Results of T-Tests and approximate T-Tests to 21 evaluate differences in egg composition parameters between lake whitefish sampled from Bayport and Naubinway. Egg Composition Parameter Wet Weight (mg/egg) Dry Weight (mg/egg) Water Content (%) Caloric Content (cal/egg) Caloric Density (cal/mg) Lipid Content (%) Lipid Content (mg/egg) Test T_yeiee T-Test 2.55 Cochran's T 0.50 Cochran's T 3.23 Cochran's T 0.13 T-Test -2.13 T-Test 0.51 T-Test 0.64 P Value 0.016 0.627 0.006 0.898 0.042 0.612 0.513 22 covariance was used to test for significant differences in the regression equations between stocks. Separate regression equations are reported for parameters when there were significant differences in the regression equations between stocks (Table 4). I found a significant positive relationship between egg caloric content (cal/egg) and the length at hatch of larval lake whitefish (Figure 3). I also found a significant positive relationship between egg lipid content (mg/egg) and the length of hatch of larval lake whitefish (Figure 4). There were also positive relationships between the length at hatch of larval lake whitefish and other egg composition parameters (i.e. lipid content and caloric density) that were highly correlated with egg caloric content (Table 4). No significant relationship between the egg incubation times and the length at hatch of larval lake whitefish was noted (R2 = 0.087, P = 0.759). Endegenoue Growth Ratee The relationship between stock egg source, parental female, and endogenous growth of hatching larvae was tested using two-stage analysis of variance. I found no significant relationship between parental stock and the endogenous growth rate of hatching larvae (Table 5, P = 0.4555). However, there was a significant relationship between parental females and the endogenous growth rate of hatching larvae (Table 5, P < 0.0001). I also found a 23 Table 4. Results of regressions between length at batch of lake whitefish larvae and egg composition parameters of eggs sampled from Bayport and Naubinway. Analysis of covariance was used to determine whether combined or separate regression equations for each stock were used. E q Composition Parameter Stock(s) Caloric Content (cal/egg) Combined Egg Lipid Content (%) Combined Lipid Content (mg/egg) Combined Caloric Density (cal/mg) Naubinway Bayport R2 0.780 0.579 0.228 0.186 0.146 2 < 0.0001 < 0.0001 0.0038 0.0296 0.0468 24 11.4 - 11.3 - .5 .1. N I 11.1 - .a .1 O I 2 ‘11.. F1 ===‘)J7END Mean Length at Hatch (mm) 3 to l 10.8 - Bayport C ‘ Naubinway A 10.7 I l l I l 10 12 14 1s 18 20 22 Egg Caloric Content (cal/egg) ' Figure 3. Mean length of hatching larval lake whitefish as a function of their egg caloric content (cal/egg). 25 11.5 - 11.4 - A E E 11.3 - V .C .9. ‘I" 11.2 - ‘6' .2 «pl 0711.1 - c o .1 5 d) 11.0 - E 10.9 .— Bayport c Naubinway A 10-3 l 1 l 1 l l 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Llpld Content (mg/999) Figure 4. Mean length of hatching larval lake whitefish as a function of their lipid content (mg/egg). 26 Table 5. Results of analysis of variance to test for differences between stocks, females, and feeding rations in the endogenous growth, exogenous growth, and survival of larval lake whitefish under laboratory conditions. Endogenous Growth Rate Souree Degrees of Freedom E-Velee [Egyelge Stock 1 0.56 0.4555 Female 5 9.70 0.0001 Exogenous Growth Rate Source Degrees of Freedom E-Value Ezyelge Stock 1 0.74 0.3945 Female 5 3.48 0.0065 Ration 3 95.01 0.0001 Exogenous Survival §ource Degree§_of Freegem E-Valge Eeyelge Stock 1 1.66 0.2014 Female 5 0.45 0.8102 Ration 3 170.11 0.0001 27 significant relationship between parental females nested within stocks and the endogenous growth rate of their offspring during the first three weeks of the experiment (Nested ANOVA, F = 31.21, P < 0.0001). The relationship between parental females and endogenous growth rates of their larvae was related to egg composition (Table 6). I found a significant relationship between the endogenous growth rate of larvae and egg percent lipid content (Figure 5). Additionally, there was also a significant relationship between endogenous growth rate and egg lipid content (Figure 6). Positive relationships between endogenous growth and other egg composition parameters (i.e. egg caloric content, egg caloric density) that were highly correlated with egg percent lipid content were also observed (Table 6). No significant relationship between egg incubation time and endogenous growth rate was found (R2== 0.052, P = 0.856). Exogenous Growth Ratee The relationships between exogenous growth rates of larval lake whitefish and parental stock, parental female, and feeding ration were tested using three-stage analysis of covariance. There was a significant relationship between parental females and the exogenous growth rate of their offspring during weeks 4-6 post-hatch (Table 5, P = 0.0065). However, there was not a significant relationship between parental stock and the exogenous growth of larval lake 28 Table 6. Results of regressions between endogenous growth rate of lake whitefish larvae and egg composition parameters of eggs sampled from Bayport and Naubinway. Analysis of covariance was used to determine whether combined or separate regression equations for each stock were used. qu Composition Parameter Steek(s) BE 2 Lipid Content (%) Combined 0.896 < 0.0001 Lipid Content (mg/egg) Combined 0.876 < 0.0001 Caloric Content (cal/egg) Naubinway 0.649 0.0019 Bayport 0.562 0.0042 Caloric Density (cal/mg) Naubinway 0.187 0.1153 Bayport 0.115 0.1639 29 0.013 '- o.012 - A :3 V .93 cu .. m 0.011 E 9. a 0.010 *- tn 3 2 g R = 0.896 8, 0.009 - O 13 : ll] 0.008 - Bayport o Naubinway A 0-007 1 J 1 1 1 1 14 16 18 20 22 24 26 Percent Llpld Content Figure 5. Endogenous growth rates (u) of larval lake whitefish during the first 21 days of the laboratory experiment as a function of percent lipid content. 0.013 '- 0.012 - 0.011 - 0.010 L- P o o o I Endogenous Growth Rate (u) 0.008 Bayport 0 ‘ Naubinway A 0.007 I l l l l l l l 0.35 0.40 0.45 0.50 0.55 0.50 0.85 0.70 0.75 Llpld Content (mg/egg) Figure 6. Endogenous growth rates (u) of larval lake whitefish during the first 21 days of the laboratory experiment as a function of lipid content (mg/egg). 31 whitefish (Table 5, P = 0.3935). I found a significant positive relationship between feeding ration and exogenous growth of larval lake whitefish (Table 5, P < 0.0001), with higher feeding rations resulting in higher exogenous growth rates of larval lake whitefish (Figure 7). Laryal Survivel Rates There were no significant differences in survival rates between any of the treatment groups during the three week period of endogenous growth. Therefore, analysis of survival rates concentrated on the period during weeks 4-6, when larval lake whitefish made the transition from endogenous to exogenous food resources. The relationships between survival rates during weeks 4-6 and parental stock, parental female, and feeding ration were tested using three-stage analysis of variance. There was a significant relationship between feeding ration (zooplankton/larvae) and larval survival during exogenous feeding (Table 5, P < 0.0001). Higher feeding rations resulted in higher larval survival rates throughout the 21- day exogenous feeding period (Figure 8). There were no significant relationships between exogenous survival rate of larval lake whitefish and parental stock (Table 5, P = 0.2067) or parental females (Table 5, P = 0.8102). 32 .00000 0000:000 0:0 H 000000000 0000 00000 .0000000 >009 no c000oc=0 0 00 000000900 000 00000 203000 0aoc0uox0 0:0 .00000 0000 £000 0:0000000 >00000000co00 00: 0003 00000 003000 050:000000 .0003000000 >0o0000000 >00 me 000 000090 500000053 0000 00>000 no 00000 003000 0aoc0uox0 0:0 0:0:000000 .b 005000 0>0D - 0 d m 0 . _ .°.FF msocomoxm msoc0uoucm 1 9N _. 1 o.» _. ..w- u . ...... T H 11% 1 1 1\1.1\.1V :0EEOECEQOON O 1 O c F N 111111 \....\..... £m_u__\c0uxcm_QOON GP 111111 w 1 1 1 1 1 1 1H1 1 .\..\.% :0EEOECMEOON cw ........... 1 9m P (..\ 1 1 ...... \....\. smfiEouxcmEonuN cm I I. I .............. \.. 1W\ 00500.2. >05 1 9.0 w 1 9:. 33 .:00000>00 0000:000 0:0 H 0:0000000 00::00 00000 >000Q .0:00000 >000 00 :000o:00 0 00 000000000 000 00000 00>0>000 000:00000 0:0 .00:00 :000 £000 0:0000000 >00:000u0:000 00: 0003 00000 00>0>000 000:0000:m .0:0E000mx0 >0o000on00 >00 «0 0:0 0:0000 £000000n3 0x00 00>000 no 00000 00>0>000 090:00ox0 0:0 000:00o0cm .0 000000 0500 we mm mm 5 E n o _ _ _ _ _ macsmmoxm 02050005 cwEEBxcEQOON o cwEEoEcmaooN m. 1 :0_n_\c00_:0_000N 0N I fimioycmaofi om D 1 000050 >05 ON 00 cm IV memed 9N 34 ExogenoueiGrowth Mode; A plot of instantaneous growth rate (u) versus zooplankton feeding ration (z/f) is shown in Figure 9. Larval growth of larval lake whitefish was accurately modelled using a threshold-corrected hyperbolic equation: (z/f)-0.277623 -0.104138 “ *[ (z/f)+12.032211-(2*(0-277623)) R2 = 0.973. The instantaneous growth rate associated with each of the four zooplankton feeding rations was within one standard deviation of the curve predicted by the threshold-corrected hyperbolic equation. DISCUSSION Egg Composition I found few significant differences between stocks in the egg composition parameters measured for sampled fish (Table 1). There were significant differences between stocks in the mean wet weight of eggs, which were primarily due to differences in the water content. There were no significant differences in the egg dry weight, egg caloric content, or egg lipid content between stocks. 35 P O m I 0.06 - 2 0.04 _ R = 0.973 Specific Growth Rate (u) '(lJJEI I I l I I I 0 20 40 60 80 1 00 Zooplankton/Fish/Day Figure 9. Specific exogenous growth rate (u) of larval lake whitefish during weeks 4-6 of the laboratory experiment as a function of prey ration. Solid line represents the fit of the Monod threshold- corrected equation to predict specific growth rates as a function of prey ration. 36 One of the principal responses of lake whitefish populations to commercial exploitation is a shift in the age and size structure from older, larger individuals to younger, smaller individuals (Smale 1988). Another population level response to exploitation commonly observed is that somatic growth rates of lake whitefish increase (Ebener and Copes 1985: Smale 1988). Although the age at maturation may decline in response to exploitation, the size at maturation of lake whitefish populations is fairly constant over different levels of exploitation (Smale 1988; Taylor et al. 1991). In response to exploitation, lake whitefish populations would therefore tend to have a higher proportion of younger and smaller female spawners. In this study, the Naubinway stock is a relatively heavily exploited stock comprised of primarily younger, smaller sized fish. The Bayport stock is a relatively lightly exploited stock dominated by older, larger-sized individuals. One of my predictions regarding average egg composition of these two stocks was that the larger and older female spawners in the Naubinway stock would have higher egg quality, as measured by egg composition parameters. Although this study was not designed to assess population responses to exploitation, data were collected from female lake whitefish with sufficient age and size ranges to allow for preliminary conclusions. Based on the analysis of a limited sample size of two stocks, there is no 37 evidence to suggest that exploitation influences the egg quality parameters of lake whitefish measured in this study. Tradeoffs Between Egg Size, Egg Number, and Egg Composition In allocating available energy for reproductive products, spawning female fish face two major tradeoffs. First, available reproductive energy must be allocated between egg size and egg number (Svardson 1949). Second, energy reserves must be allocated between carbohydrate, protein, and lipid components within individual eggs. Life history theory suggests that natural selection will select for females that allocate reproductive energy in a manner to maximize fitness return per unit of ovarian resource invested (Smith and Fretwell 1974: Lloyd 1987; McGinley et al. 1987; Sargent et al. 1987; Winkler et al. 1987; Fleming and Gross 1990). Thus, egg number should vary in response to selection upon egg size, egg composition, and total investment in egg production (Fleming and Gross 1990). Caloric Content vereus Calorie DensiEy In evaluating egg composition of lake whitefish, there appears to be a tradeoff between the caloric content (cal/egg) and the caloric density (cal/mg) in eggs when examined independently of egg size. Parental females having a high egg caloric density (cal/mg) also have a higher percent lipid content. These eggs in turn produce larvae having more endogenous energy reserves. Larvae hatching from eggs with higher lipid contents hatched at a slightly 38 smaller average size, but were able to compensate for this size disadvantage through higher levels of endogenous growth. Larvae with greater endogenous energy reserves may have a number of abilities enhancing survival and growth of these larvae, despite their smaller initial size. Greater endogenous energy reserves would enhance the ability of larvae to delay initial feeding for longer periods of time compared to larvae with less lipid energy reserves. During this time, larvae could also expend larger amounts of energy to actively seek and capture prey resources. Parental females having high egg caloric content (cal/egg) but lower caloric densities had higher levels of protein and carbohydrate. These larvae hatched at a slightly larger size, but had lower levels of endogenous growth. These larvae may have size-related advantages such as enhanced feeding (Hunter 1981; Blaxter 1986; Webb and Weihs 1986) and predator escapement abilities (Zaret 1980; Blaxter 1986), but their limited endogenous energy resources limit the time available to initiate feeding on exogenous prey resources. However, relatively small size differences may translate into substantial differences in survival and subsequent recruitment (Adams and DeAngelis 1987), and thus may be significant in lake whitefish recruitment dynamics. 39 W The presence or absence of a "point of no return" in fish, a point beyond when fish are no longer capable of feeding, has been the point of conflicting reports in the literature (Lasker et al. 1970; Miller et al. 1988: Fritz et al. 1990). The point of no return is defined as the time after which the effects of starvation are irreversible due to destruction of the digestive system (Miller et al. 1988). The point of no return is defined for each individual larva, but can only be measured for a group of larvae. The presence of a point of no return has been reported by Lasker et al. (1970) for anchovy (Engreglie peggex). During the present research, larval lake whitefish not utilized in the experimental design were held at a temperature of 4.0° C for as long as 86 days without feeding. Survival of these larvae ranged from 25 to 86 days after hatching. Repeated attempts to feed larvae after 28 days resulted in no evidence of feeding, despite the fact that larvae were able to survive for weeks beyond this point. This clearly indicates that lake whitefish larvae are capable of surviving beyond the point that they are capable of feeding. If starvation is an important source of mortality under field conditions, these results could have serious implications for the estimation of densities and average size of larval lake whitefish collected during ichthyoplankton sampling. Presence of starving fish could 40 lead to an overestimation of effective larval density or an underestimation of average size of non-starving larval fish. Considering the current emphasis on the analysis of the characteristics of surviving larvae (Fritz et al. 1990), this prolonged survival could be a confounding factor in the identification of characteristics allowing larval fish to survive. Relative Influence of Tempegatgge Endogenous and exogenous growth observed during this study were lower than those reported from previous laboratory experiments using larval lake whitefish (Taylor and Freeberg 1984). There were two causes for the apparent slow growth during this experiment. First, larval lake whitefish in this experiment were not fed for a period of 21 days to allow for the measurement of endogenous growth. Under natural conditions, larval lake whitefish began feeding within 48 hours after hatching (Freeberg et al. 1990). Second, water temperatures in this experiment were ‘maintained at relatively cold temperatures (6.9°1C) to simulate temperatures normally encountered in the Great Lakes, while the earlier study (Taylor and Freeberg 1984) was conducted at 12° C. At this warmer temperature, larval lake whitefish exhausted their endogenous energy reserves between 12 and 15 days post hatch (Taylor and Freeberg 1984) compared to 23 to 25 days post-hatch in the present experiment. 41 Under natural conditions, larval lake whitefish emerge (late March and April) when open water temperatures in the Great Lakes are cold (2° - 4° C) and areas of warmer water (4°-£P C) are located only at the surface and in near- shore areas. Larval lake whitefish located in nearshore areas with warmer water temperatures would be able to achieve growth rates commonly observed under field conditions. At this time of the year, warm water is probably an important ecological resource because it determines the bioenergetic potential for larval growth. In this experiment, starvation was the only source of mortality experienced by larval lake whitefish. Other factors influencing larval survival under field conditions (i.e. predation) were not evaluated. However, if predation is important, I noted that larvae fed lower prey rations during this experiment demonstrated slow growth and impaired swimming ability which may result in higher mortality under natural conditions. 42 Conclusions There were no apparent differences in the egg composition, larval growth, and larval survival dynamics of lake whitefish sampled from the Naubinway stock (Lake Michigan) and the Bayport stock (Lake Huron). Length at hatch of larval lake whitefish was primarily a function of egg size, as measured by the overall egg caloric content. Endogenous growth of larval lake whitefish was highly dependent on the egg lipid content, as measured by the percent lipid content and the mg lipid per egg. Exogenous growth and survival of larval lake whitefish were positively related to prey resource availability and both were sensitive to small changes in the availability of zooplankton prey resources. Variation in egg composition has the potential to be important in determining larval growth and survival and eventual recruitment of lake whitefish. Chapter 2: Climate-Based Recruitment Models of Lake Whitefish in Two Areas of Northern Lake Michigan Abstract Spawning stock and climate-based recruitment models were developed for lake whitefish in northern Green Bay and the North Shore areas of Lake Michigan. Abundance and cohort strength indices were calculated for the 1961-1985 year classes based on catch and effort data from the commercial fishery for the 1958-1989 year classes of lake whitefish in each area. Beverton-Holt and Ricker stock- recruitment functions, ice cover concentrations, wind intensity data, and spring water and air temperature variables were used as model inputs in climate-based recruitment modelling. The climate-based model for northern Green Bay modelled lake whitefish recruitment as a function of a Beverton-Holt stock-recruitment function, ice concentration, and wind intensity. The climate-based recruitment model (R2 = 0.65) demonstrated improved hindcasting ability when compared to the Beverton-Holt stock-recruitment model (R2== 0.37) for the 1961-1985 cohorts. The climate-based model for the North Shore modelled lake whitefish recruitment as a function of a spring air temperatures, ice concentration, the Beverton- Holt stock-recruitment function, and wind intensity. The 44 climate-based recruitment model (R2 = 0.57) demonstrated improved hindcasting ability when compared with Ricker stock-recruitment model (R? = 0.13). Results of this study indicate that climate-based recruitment models have the potential to more accurately forecast lake whitefish cohort strength several years in advance of recruitment into the fishery. Introduction The lake whitefish is an important component of the commercial fishery of the upper Great Lakes, and currently comprises over 50% of the commercial landings in Michigan waters (Kinnunen 1990). Over the past century, there have been large fluctuations in the commercial landings and catch-per unit effort (CPUE) of lake whitefish (Figure 10) that have been primarily caused by variable levels of population abundance (Christie 1963; Taylor et al. 1987). Several potential causes for the observed variation in abundance of lake whitefish include overfishing (Patriarche 1977), habitat degradation (Christie 1963), sea lamprey (Petromvzon marinus) depredation (Smith and Tibbles 1980), competition with exotic species of fish (Christie 1974), and variable recruitment (Christie 1963; Lawler 1965; Patriarche 1977; Smale 1988). Correlative studies of lake whitefish recruitment 45 N so 5» C” CD tn 7 I i j 0’ C) I Yield (kg x 106) E A CD I 53 tn i (’ I I I I I I I I 1900191019201930194019501960197019801990 Year Figure 10. Commercial fishery landings of lake whitefish in Lake Michigan, 1900-1989. 46 suggest that much of the variability in the cohort strength and recruitment may be caused by variation in climatic influences, such as water temperatures and wind. Christie (1963) found a relationship between lake whitefish recruitment and winter and spring temperatures in Lake Ontario, with cold Novembers followed by warm Aprils providing conditions associated with the production of strong year classes. Strong year classes in Lake Erie resulted only when an optimal set of water temperatures for spawning, incubation, and development was observed (Lawler 1965). Strong year classes in Lake Erie occurred when fall temperatures fell below 6.1 C prior to spawning; the fall and winter temperature decline was steady with few fluctuations; and spring temperatures increased slowly and late, providing a long incubation period at optimal temperatures of 2.0 to 4.0 C. Miller (1952) found an inverse relationship between fall wind intensity and lake whitefish year class strengths in seven Alberta Lakes. These correlations between fall and winter weather conditions and year class strength of lake whitefish suggest that weather related variables significantly influence egg and larval survival. In Grand Traverse Bay, Lake Michigan, ice cover over lake whitefish spawning grounds appears to reduce the intensity of wind generated waves and currents, directly influencing egg survival (Taylor et al. 1987; Freeberg et 47 al. 1990). Egg survival in Grand Traverse Bay during an ice covered winter (1983—84) was approximately 3.4 times greater than during an ice free year (1982-83). Additionally, the recruit per stock ratios for a lake whitefish stock along the North Shore of Lake Michigan were positively related to the maximum extent of ice cover on a lake wide basis (R2== 0.56), and historical trends in winter severity were positively associated with lake whitefish catch by the North Shore commercial fishery (Smale 1988). In addition to late fall and winter climatic conditions influencing spawning and egg survival, spring climatic conditions appear to be important in determining survival of emerging lake whitefish larvae. Freeberg et al. (1990) found that the abundance of copepod zooplankton between 0.7 and 1.1 mm in total length significantly influenced the growth and survival of larval lake whitefish in Grand Traverse Bay, Lake Michigan. Spring climatic conditions, including the timing of ice dissipation and the warming of shallow water areas, appear to influence the timing of the initiation of production of copepod zooplankton, which in turn impacts larval survival of lake whitefish. Identification and quantification of the factors influencing cohort strength of lake whitefish would enhance the management of the fishery by enabling forecasts of lake whitefish abundance several years in advance of their recruitment to the fishery. The goal of this study was to 48 identify key biotic and abiotic variables determining cohort strength and to incorporate these variables into recruitment models for lake whitefish in two areas of northern Lake Michigan. The specific objectives of this study were to: 1. Identify the biotic and abiotic variables important in determining the egg and larval survival of lake whitefish. 2. Incorporate these variables into a quantitative recruitment model to hindcast cohort strength of lake whitefish in two areas of northern Lake Michigan. Hypotheses Four hypotheses about the effects of biotic and abiotic variables were used as a basis for developing climate-based recruitment models during this study. 1. In the absence of climatic variables, the cohort strength and recruitment of lake whitefish is primarily a function of the size of the spawning stock, and that the relationship between stock size and recruitment can be approximated by a Ricker or a Beverton-Holt stock- recruitment relationship (Ricker 1975). 49 2. Wind generated currents and waves during and after spawning are important in determining egg survival of lake whitefish. 3. Ice cover over lake whitefish spawning areas protects eggs from wind generated currents and waves, thereby enhancing the survival of lake whitefish eggs. 4. Spring water temperatures are important in determining the production of zooplankton food resources, which in turn influence the growth and survival rates of hatching lake whitefish larvae. Methods Model gonstructiog Overview Analysis of commercial fishery and climatic data produced three models for each area modelled, two traditional stock-recruitment models and a climate-based recruitment model. A flow chart of the modelling process is shown in Figure 11. Commercial catch and effort data were used to calculate abundance and cohort strength indices, which were in turn used to calculate Ricker and Beverton— Holt stock-recruitment models for each area. Freezing degree day data and ice cover observations were used to construct ice cover models for each area. Outputs from ice Model Inputs lnterrnedlete Model Parameters 100m000000001 and Analyses FmflFhmummuu 0000: Fumflng heCMWN Decree-Dav: Obeervetlone Figure 11. Flow chart of lake whitefish recruitment modelling showing the model inputs, intermediate model parameters, intermediate models and analyses, and final recruitment models produced for each area. 51 cover modelling including the timing and duration of ice cover, wind intensity data, spring temperature data, and lake whitefish stock abundance were analyzed using stepwise regression to produce a climate-based recruitment model for each area. In summary, a Ricker stock-recruitment model, a Beverton-Holt stock-recruitment model, and a climate-based recruitment model were constructed for each area. Commercial Catch Data and Studv Areae Records of commercial landings and standardized effort of the commercial fishery have been recorded since 1931 by state agencies and are currently maintained by the United States Fish and Wildlife Service at the National Fisheries Research Center - Great Lakes. Because of the difficulty of partitioning the effects of lamprey mortality from those of variable recruitment prior to 1958, commercial catch data from 1958 to 1989 were used to develop abundance and cohort strength indices for the 1961-1985 cohorts in this study. Two areas in northern Lake Michigan were selected for recruitment modelling based on the consistent availability of commercial catch and age structure data. These areas were Great Lakes Fisheries Commission (GLFC) statistical district MM-l including the Michigan waters of northern Green Bay and GLFC statistical district MM-3 including the north-central and northeastern areas of Lake Michigan (Figure 12). Because important spawning areas for lake whitefish are located in the northern portion of area MM-3, 52 nnnnnnnnnnnnnnn ..J .-.-.-Q-‘-- . ............. ................. . ................. ................... 1: ..... . E Lake 2 I Michigan - ............. .................. .................. .................... ----4°3-~~ .............. Figure 12. Location map of Lake Michigan showing the two areas for which recruitment models were developed. The recruitment model for northern Green Bay included the entire area within GLFC statistical district MM-l. The recruitment model for the North Shore included the shaded region in the northern portion of GLFC statistical district MM-3. 53 a recruitment model was developed for the northern half of this area. The commercial catch of lake whitefish over the past 30 years has fluctuated, but catch data suggests that similar recruitment mechanisms may be operating in each area (Figure 13). Each of these areas has multiple stocks of lake whitefish which contribute to the commercial landings within each district (Ebener and Copes 1985). Because of the nature of the data set and the limited data available from previous studies, modelling of individual stocks was not attempted. Commercial Fishery Gear and Effort Six types of gears have been used to fish for lake whitefish in area MM-l since 1958 including 2-inch gill nets, 4-inch gill nets, shallow water trap nets, deep water trap nets, pound nets, and otter trawls. Seven types of gears have been used to fish for lake whitefish in area MM-3 including 2-inch gill nets, 4-inch gill nets, shallow water trap nets, deep water trap nets, pound nets, otter trawls, and seines. Catch and effort data from 2-inch gill nets, deep water trap nets, otter trawls, and seines were not used to index abundance or cohort strength because of the limited catch or time frame that these gears were fished. Catch and effort data from three principal gears, 4-inch gill nets, shallow water trap nets, and pound nets, were used to index abundance and cohort strength of lake whitefish in each 54 1.6 .. . —— Green Bay (MM-1) 1,4 _ ----- North Shore (M M-3) «=2 ,. >< C” 5.. J: 4) fl (0 0 I J 1 930 1940 1950 1960 1970 1980 1990 Year Figure 13. Commercial fisheries landings of lake whitefish in GLFC districts MM-l and MM-3 in Lake Michigan from 1930-1989. 55 catch of lake whitefish in area MM-l and 94.6 % of the commercial catch in area MM-3 during the period 1958-1989. Catch per unit effort was calculated for each gear during each calendar year for each area. Commercial fishing effort was standardized using measures of effort defined by the GLFC (Hile 1962; Table 7). Indices of Spawning Stock and Beereitmehe Annual indices of spawning stock abundance were calculated based on methods outline by Hile (1962). Mean annual catch per unit effort was calculated for each gear in each year using data for the entire period of 1958-1989. An abundance index was then calculated using the following formula (Hile 1962): all gear: Abd.- ‘1 1 all gear: CPUEZ -1 £91 ABDi = abundance index in year i; Cji = lake whitefish catch in pounds by gear j in year i: CPUEj mean catch per unit effort of gear j for the years 1960-1987; Eji = total effort using gear j in year i. 56 Table 7. Standardized units of effort for the commercial fishery in the Great Lakes as defined by the Great Lakes Fishery Commission (Hile 1962). Gear Type Standardized Unit of Effort Shallow Water Trap Net one lift of one net Pound Net one lift of one net 4-Inch Gill Net one lift of 1000 linear feet of gill net 57 Several factors confounded efforts to partition abundance indices into cohorts and to estimate indices of cohort abundance. Recruitment age was not constant in either stock over the time period considered in this study. As the northern Green Bay and North Shore stocks gradually recovered from extremely low levels of abundance in the early 1960's, somatic growth rates declined and the mean age at recruitment increased by nearly two years (Smale 1988). In addition, the age at first recruitment was delayed in larger cohorts because of reduced somatic growth rates resulting from intraspecific competition within the cohort (Ebener and Copes 1985; Smale 1988). In order to account for variable age of recruitment, cohort analysis procedures were used to partition abundance indices into indices of cohort abundance. Percent age composition data were available for northern Green Bay (area MM-l) including Piehler (1967), NcComb (1989), and unpublished data from the Michigan Department of Natural Resources (Table 8). Percent age composition data for the North Shore (area MM-3) were available from Piehler (1967), Brown (1968), Patriarche (1977), Rybicki (1980), and Scheerer and Taylor (1985) and unpublished data from the Michigan Department of Natural Resources and the tribal management sources (Table 9). The abundance index value for each year was partitioned into the cohorts contributing to the abundance index in each 58 Table 8. Percent age composition of lake whitefish sampled in area MM-l during the period 1960-1987. Data sources are shown for each year. ear I; _11 IX 2 y; El; Source 1958 58 33 9 0 0 0 MDNR, Unpublished 1959 84 14 2 0 0 0 MDNR, Unpublished 1960 74 16 10 0 0 0 MDNR, Unpublished 1963 62 31 7 0 0 0 MDNR, Unpublished 1964 88 8 4 0 0 0 MDNR, Unpublished 1966 94 6 0 0 0 0 Piehler 1967 1968 38 38 20 1 2 0 MDNR, Unpublished 1969 28 31 25 4 10 0 MDNR, Unpublished 1970 12 48 36 3 0 0 MDNR, Unpublished 1971 0 91 8 1 0 0 MDNR, Unpublished 1972 2 73 22 2 0 0 MDNR, Unpublished 1973 3 58 34 1 0 0 MDNR, Unpublished 1979 0 53 39 6 1 0 NcComb 1989 1980 0 29 61 8 1 0 NcComb 1989 1981 1 41 53 5 0 0 NcComb 1989 1982 1 33 52 10 3 l NcComb 1989 1983 0 54 31 12 2 0 NcComb 1989 1984 0 44 53 3 0 0 NcComb 1989 1985 0 3 81 16 0 0 NcComb 1989 59 Table 9. Percent age composition of lake whitefish sampled in area MM-3 (North Shore) during the period 1960-1987. Data sources are shown for each year. Year II _II IV 2 VI N Source 1960 84 14 1 l 0 151 Brown 1968 1961 82 13 2 0 0 153 Brown 1968 1962 55 40 4 0 0 109 Brown 1968 1965 100 0 0 0 0 72 Piehler 1967 1966 25 68 4 0 O 681 Piehler 1967 1967 29 31 35 2 0 1232 Brown 1968 1968 9 74 8 8 0 611 Patriarche 1977 1969 25 58 16 1 O 96 Patriarche 1977 1970 44 54 2 1 0 169 Patriarche 1977 1971 8 81 10 1 0 296 Patriarche 1977 1972 28 68 2 2 0 141 Patriarche 1977 1973 1 65 30 1 0 141 Patriarche 1977 1974 0 15 69 8 0 113 Patriarche 1977 1975 2 92 4 1 O 133 Rybicki 1980 1976 1 60 38 0 0 339 Rybicki 1980 1977 4 36 36 1 0 133 Rybicki 1980 1978 O 44 36 7 0 586 Rybicki 1980 1979 2 63 31 0 0 399 Rybicki 1980 1980 0 78 21 l 0 513 Scheerer & Taylor 1985 1981 0 12 83 5 0 738 Scheerer & Taylor 1985 1982 0 4 37 55 4 452 Scheerer & Taylor 1985 1983 0 14 37 33 16 606 Smale 1988 1984 0 8 52 26 14 361 Smale 1988 60 year based on the age composition data available in each area. The partitioned abundances were summed for all years that a cohort contributed to the commercial catch to calculate a cohort abundance index value for each year class. Commercial catch and age composition data were used to calculate cohort abundance index values for the 1961-1985 cohorts within each area. Stock-Recruitment Relationships Abundance and cohort strength indices were used to calculate a Beverton-Holt (Beverton and Holt 1957) and a Ricker (Ricker 1954; Ricker 1985) stock-recruitment relationship for each area. The Beverton-Holt stock- recruitment relationship for each area was based on the following form: 1 R1--—————— a+£ P R = Recruitment Index Value P = Abundance Index Value 0, fl = dimensionless fitting parameters Parameters for the Beverton-Holt stock-recruitment relationship were estimated by regressing l/R on 1/P to estimate 0, and a was estimated by substitution. The Ricker stock-recruitment relationship for each area 61 was based on the following formula: I?- a.Pe?BP The Ricker stock-recruitment relationship was determined by regressing (logeR - logeP) against P to estimate [3, and a was estimated through substitution. Ice Cover Models Ice cover models based on mean daily temperatures were developed by Raymond Assel of the Great Lakes Environmental Research Laboratory (National Oceanic and Atmospheric Administration). Air temperatures collected at Escanaba, Michigan were used to calculate mean daily temperatures and freezing degree days. Weighted average daily temperatures for various averaging periods were used to estimate spatially averaged ice concentrations for each study area. The following model was a representative of the general form of the ice cover models. I) C C. It- Co+ 2 [(3:2- thdt] 1‘1 Del 62 It = ice concentration on day t Co, Ci, i = 1,2,..j..n,t empirical coefficients Del = (t - Li), 1 = l,2,..j..n, t > Li Lj = a given date before day y, for i = j n = number of terms in the summation Tt = average temperature (° C) on day t The model given in the above illustration contains "n" terms and some models also have a significant intercept term "Co". The C's are temperature weighting factors (coefficients of regression) for the integrated daily temperature for averaging periods of 5, 10, 30, 60, 90, 120, 150, and 180 days. The ice cover data used to calibrate the models were divided into three parts, based on the winter severity classes of above normal, normal, and below normal. Seasonal maximal freezing degree-day accumulations were used for the winter severity classification analysis. Averaging periods for each winter classification within each lake area were determined from a R2 analysis using the SAS statistical package RSQUARE. The "Del" term in the models defines the averaging periods. The model was calibrated for each lake area using the observed spatially averaged ice cover data in the Great Lakes Environmental Research Laboratory's 20-year (1960-1979) computerized digital ice concentration data base. The averaging periods, 63 R2, root mean square errors (RMSE) and number of observations used in model calibration each winter classification within each area are given in Table 10. Each set of models was used to produce mean daily ice cover concentrations for each area for the winters of 1960- 1987. Because the critical level of ice cover concentration necessary to protect lake whitefish eggs was not known, the number of days with ice cover exceeding several levels were used as recruitment model inputs. Ice concentrations of O, 10, 20, 30, 40, 50, 60, and 70 % were chosen to produce seven ice cover variables for each area. The number of days of ice concentration exceeding each of these ice concentration levels was calculated for each area. Wind Speed Data Wind speed data collected at Green Bay, Wisconsin and Sault St. Marie, Michigan were used as model inputs in areas MM-l and MM-3, respectively. Two-day average wind speeds (knots) were calculated on a daily basis during November, December, and January for the years 1960-1987. Average wind speed levels above 8, 10, and 12 knots were selected to define days of intense wind activity. The number of days above these wind speed levels was tabulated for each ice cover level within each area. Ice cover data were used to define the period when intense wind events could impact survival of lake whitefish eggs. I hypothesized that intense wind events would only 64 Table 10. Results of ice cover modelling for areas MM-l (northern Green Bay) and MM-3 (North Shore). Asterisks indicate regression models containing an intercept term. The averaging periods used to produce each model, R?, root mean square errors (RMSE) and number of observations used in model calibration each winter classification within each area. Northern Green Bay, Area MM-l Winter Severity Averaging Pegieds (Days) ifi mg; m Above Normal * 10, 60, 120, 150 0.62 13.9 55 Normal * 5, 30, 60, 120, 180 0.73 12.2 117 Below Normal 10, 90, 120 0.94 16.7 26 North Shore, Area MM-3 Wigter Severity Averaging Periegs (Qeys) BE ELISE m Above Normal 10, 50, 90, 180 0.97 13.6 60 Normal 5, 60, 90, 180 0.96 13.9 163 Below Normal * 5, 10, 30, 120, 150 0.70 18.8 33 65 influence egg survival of lake whitefish prior to the formation of some critical concentration of ice cover over spawning areas. To define biologically meaningful wind variables, I only considered wind events occurring between the earliest possible spawning date and the formation of various levels of ice cover. The starting date for tabulating intense wind events was November 1 in all cases, because this date corresponds with the earliest date of lake whitefish spawning in each area. Because no data were available to define a critical level of ice concentration necessary to protect major spawning areas, several ice concentration levels were selected to define the period of vulnerability of lake whitefish eggs to wind events. Ice concentration levels above 0, 20, 40, and 60 % ice cover were selected as ice concentration levels to determine the ending date for wind speed tabulation. The date when ice cover concentration first exceeded a given critical ice concentration was used as the end date for tabulated intense wind events. For example, the number of days between November 1 and the date when ice cover first exceeded 20 % in a given year was tabulated to create a single wind speed variable. In this way, the three wind speed levels (8, 10, and 12 knots) and the four ice cover concentration levels (0, 20, 40, 60 %) were used to calculate 12 wind speed variables for each lake area . 66 Spring Wate; and Air Iemperafiupee Mean monthly spring water temperatures for March, April, and May were available from a municipal water intake at Menominee, Michigan for the period 1960-1987. These temperature data were used to produce three spring temperature variables used as inputs in recruitment modelling in area MM-l. Water temperature data could not be located for area MM-3. Mean monthly spring air temperatures taken at Sault St. Marie, Michigan were used to produce three monthly mean spring temperature variables used as inputs in recruitment modelling in area MM-3. Climate-Based Recruitmenp Model gonstructiop All variables including stock-recruitment values, ice cover data, wind intensity data, and spring temperature data were used to predict year class strength index values using a least squares multiple regression approach. Stepwise multiple regression analysis was conducted using the SAS statistical software package and the variable selection model STEPWISE. Additional variables were added to each model until none of the variables outside had a significant F statistic (a = 0.15), and every variable in the model has a significant F statistic. Each resulting model was tested for autocorrelation using a Durbin-Watson d statistic (Durbin and Watson 1951) and heteroscedasticity using analysis of the covariance matrix (White 1980). 67 RESULTS Stock-Recruitment Relationships Stock-recruitment parameter estimates for both areas are shown in Table 11. Both stock-recruitment relationships indicated a significant relationship between the abundance and cohort strength indices for the northern Green Bay area. The Beverton-Holt stock-recruitment relationship (R2== 0.373, P = 0.0007) provided a slightly better fit than the Ricker stock-recruitment relationship (R?== 0.330, P = 0.0016) for the northern Green Bay area. In the North Shore area, the Ricker stock-recruitment relationship indicated a significant relationship between abundance and cohort strength indices (Table 11). Although these stock-recruitment models each explained less than 15% of the variation in recruitment, the Ricker stock- recruitment function (R?== 0.126, P = 0.0458) provided a slightly better fit than the Beverton-Holt stock—recruitment relationship (R2== 0.091, P = 0.0786) for the North Shore area. Northern Green Bay Ciimate-Beeeg Recppitmept Medele Lake whitefish recruitment in northern Green Bay was best modelled as a function of the Beverton-Holt stock- recruitment function, the number of days of ice cover exceeding 40%, and the number of days of average wind speeds above 10 knots prior to the onset of 40% ice cover: 68 Table 11. Results of Beverton-Holt and Ricker stock-recruitment relationships for lake whitefish in northern Green Bay and the North Shore, Lake Michigan using abundance and year class indices from 1961—1985. Northern Gpeep Bay Mpgei Aiphg egg; Adjusted R2 P~yelpe Beverton-Holt 0.0035091 0.50842096 0.373 0.0007 Ricker 2.4132768 -0.00504502 0.330 0.0015 North Shore 0005.1 8120.4 Beta 8__1__3_d 'usted Z 2:18.111; Beverton-Holt 0.0055245 0.2153229 0.091 0.0785 Ricker 3.0296718 -0.0091653 0.126 0.0458 69 1 R1 "' Bo + B]. j— + 52 1500 + 63 W10,40§ a + A1 Ri = recruitment index value of lake whitefish year class i; Ai = abundance index value for year i: a, fi = Beverton-Holt stock-recruitment constants [30, B1, 32, 03 = recruitment model regression constants; 1““ = number of days of ice cover exceeding 40%; w1o,4ox = number of days of 2-day average wind speed above 10 knots prior to the formation of 40% ice cover. The partial R?‘s for each parameter in this model are listed in Table 12. This recruitment model was used to hindcast recruitment index values for the years 1961-1985, and these values were plotted against actual recruitment index values calculated from catch, effort, and age composition data (Figure 14). Model hindcasts for the period 1971-1985 were more accurate than hindcasts for the period 1961-1970 based on mean absolute deviations between the model hindcasts and actual recruitment index values (Median Test, Xfi = 4.44, P = 0.035). Recruitment during this earlier period was highly variable, which may have been caused by extremely low levels of spawning stock during these years. For this area, there 70 Table 12. Parameter estimates and squared partial correlation coefficients for climate-based recruitment models of lake whitefish in areas MM-l and MM-3. Squared Model Parameter Partial Area Parameter Estimate gorrglation Beverton-Holt SR 0.768 0.399 MM-l Ice (40%) 1.135 0.216 Wind (40%,8 knots) 0.985 0.033 Beverton-Holt SR 1.949 0.139 MM-3 Ice (70%) 0.702 0.155 Wind (20%,8 knots) 0.957 0.031 Air Temp (May) 1.509 0.252 71 250 - 200 X (D 'O .E 150 ..- C “5’ 9:: E 100 0 0 m 50 - —I— Actual Recruitment Index "0- Model Predlctlon l l 1 l l L l l l L l I l 1961 1965 1969 1973 1977 1981 1985 Year Figure 14. Recruitment model hindcasts and actual cohort abundance index values for the 1961-1985 cohorts of lake whitefish in northern Green Bay, Lake Michigan. 72 is a trend toward increasing recruitment index and model prediction index values through the time period modelled reflecting the gradual recovery of lake whitefish stocks. Residual values between the predicted and actual recruitment index values were plotted as a function of time in Figure 15. There was no apparent time related trend in the pattern of the residual values from this regression model. North Shore Climate-Based Recruitment Models North Shore recruitment was best modelled as a function of the Beverton-Holt stock-recruitment function, the number of days of ice cover exceeding 70%, the number of days of average wind speeds above 10 knots prior to the onset of ice cover, and the average spring air temperatures in May: 1 R1 " Bo + B1 + pz I703 + B3 W10,os + ‘34 TMay a._fi_ Ai Imx = number of days of ice cover exceeding 70%; wwcm = number of days of 2-day average ' wind speed above 10 knots prior to the formation of ice cover: TMay = mean average air temperature in May . The partial R?‘s for each parameter in this model are listed in Table 12. This model was used to hindcast recruitment index 73 60 40 - r .«e 2° ' N “3 , 3 o A I- i (D g . Y '42‘) y .40 — I '45" I I I I I I I I I I I I I 1961 1965 1969 1973 1977 1981 1985 Year Figure 15. Plot of residual values between recruitment model hindcasts and actual cohort index values for the 1961-1985 cohorts of lake whitefish in northern Green Bay, Lake Michigan. 74 values for the years 1961-1985, and these values were plotted against actual recruitment index values calculated from catch, effort, and age composition data (Figure 16). Model hindcasts for the period 1973—1985 were no more accurate than hindcasts for the period 1961—1972, based on differences in mean absolute deviations (Median Test, X% = 0.17, P = 0.6820). The model accurately hindcasted the largest year class during this period (1977), but was less successful in hindcasting the large 1971 year class. Residual values between the predicted and actual recruitment index values showed no apparent time-related trends (Figure 17). DISCUSSION Relationship between Spawning Stock and Recruitment Based on the spawning stock and climate-based recruitment modelling results, it appears that lake whitefish recruitment in northern Green Bay is primarily dependent upon the abundance of spawning stock, and is less dependent upon climatic influences such as ice cover, wind intensity, and spring temperatures over the time period studied. The stock-recruitment relationships for the northern Green Bay area explained a greater proportion of the variation in recruitment when compared to the relationships calculated for the North Shore area. Also, 75 250 r 1 a 3 200 - H X 0 U E 150 E .’ o .‘ E .' :L'." .‘ E 100 0 0 n: 50 —I— Actual Recruitment Index p "0- Model Prediction 0 I l I I I I L I I I I I I 1961 1965 1969 1973 1977 1981 1985 Year Figure 16. Recruitment model hindcasts and actual cohort abundance index values for the 1961-1985 cohorts of lake whitefish along the North Shore, Lake Michigan. 76 60 40 '- 2! 20 cu b = 1 I I A E O (D g ,. -20 '- -40 - '6“) I I I I I I. I I II. I I I I 1961 1965 1969 1973 1977 1981 1985 Year Figure 17. Plot of residual values between recruitment model hindcasts and actual cohort index values for the 1961-1985 cohorts of lake whitefish along the North Shore, Lake Michigan. 77 the partial Ra's from climate-based recruitment modelling indicate that the stock—recruitment term made a greater contribution to the northern Green Bay model than it made to the North Shore model. One possible cause for differences in the importance of climatic factors between the two areas is related to variation in ice cover. Because northern Green Bay is relatively protected from wind and lake currents, there is less variation in the timing and extent of ice cover formation. During the years in which ice cover was modelled, ice cover normally formed during or just after lake whitefish spawning. Because there was little variation in the timing of ice cover formation, there was also less »variation in the wind events affecting egg survival prior to ice cover formation. In contrast, ice cover and wind events were more variable in the North Shore area, and this climatic variability appears to be more important in determining egg survival, cohort strength, and eventual recruitment of lake whitefish in this area. A second possible reason for the stronger relationship between stock size and recruitment is related to the range [of stock abundances present during the modelling time period. While abundance and cohort strength indices are site-specific and not directly comparable between sites, the range of stock abundances appears to be larger for the northern Green Bay area as reflected by catch data (Figure 78 13). Low levels of abundance prior to 1970 may have amplified the apparent relationship between stock abundance and recruitment. Climate—Based Recruitment Modelling Incorporation of climate variables increased the recruitment hindcasting ability of recruitment models in both areas. Ice cover was an important variable in both models, although the concentration of ice cover that was most significant in predicting recruitment was different in each area. Several levels of ice concentration were used as inputs in recruitment modelling because of uncertainty regarding the critical level of ice cover concentration necessary to protect lake whitefish eggs. In the northern Green Bay area, 40% ice concentration was identified as being most significant, while in the North Shore area, 70% was most significant. These differences in critical ice cover concentrations may be due to differences in the lake basin morphology and prevailing storm patterns in each area. The importance of abiotic variables such as ice cover and wind speed to the recruitment process of lake whitefish in different areas of the Great Lakes is dependent upon a number of factors including the quantity and quality of spawning substrate and lake basin morphology. Ice cover may play a more critical role in protecting lake whitefish eggs in areas with marginal spawning substrate quality by protecting eggs from wind and wave action. In these areas, 79 lake whitefish eggs are more vulnerable to displacement and destruction by abrasion than in areas with quality spawning substrate. Freeberg et al. (1990) found that ice cover was critical for protecting lake whitefish eggs in the marginal spawning areas of Grand Traverse Bay, Lake Michigan. The two areas modelled in this study, northern Green Bay and the North Shore occur along a geological band of limestone, which produces substrate that is optimal for incubation of lake whitefish eggs (J. Reckhan, Ontario Ministry of Natural Resources, personal communication). The relatively large availability of quality spawning areas may be one reason why these two areas have consistently had the highest production of lake whitefish in Lake Michigan (Baldwin et al. 1979). Other Factors Affecting Recruitment Several factors not included as inputs in recruitment modelling may have a significant influence on the recruitment of lake whitefish in these two areas. Variation in the production of spring zooplankton resources has been identified as a mechanism which influences the larval survival of lake whitefish (Reckhan 1970; Freeberg 1985; Freeberg et al. 1990). Only one set of surrogate variables, spring water and air temperatures, is directly related to the production of zooplankton prey resources. Site-specific studies of fine-scale nature are needed to identify factors important in controlling the production of these prey 80 resources. Other biotic factors such as competition and predation may be important factors influencing recruitment mechanisms. Competition between larvae of several species including lake herring, alewife, and rainbow smelt may be a locally significant factor influencing lake whitefish recruitment. Likewise, predation by adults of several species on lake whitefish including rainbow smelt (Loftus and Hulsman 1987) and alewife (Hoagman 1974) also may be significant in some areas. Results of this study provide a basis for identifying and incorporating climatic variables into recruitment modelling. These recruitment models provide a basic model which predicts the potential recruitment of lake whitefish based on the egg and larval survival of lake whitefish. While variation in climatic variables appears to be important in determining recruitment of lake whitefish, recruitment models incorporating climatic variables and only one biotic variable (stock abundance) can not be expected to explain all of the variability in lake whitefish recruitment. Other biotic factors influencing prey productivity and availability, intra- and interspecific competition, and predation need to be incorporated into further efforts to model lake whitefish recruitment in order to produce finer-scale models for specific areas of Lake Michigan. 81 Conclusions Lake whitefish recruitment in northern Green Bay was modelled as a function of the Beverton-Holt stock- recruitment function, the number of days of ice cover exceeding 40%, and the number of days of average wind speeds above 10 knots prior to the onset of 40% ice cover. The climate-based recruitment model (R2== 0.65) for northern Green Bay demonstrated improved hindcasting ability when compared to the Beverton-Holt stock—recruitment model (R2== 0.37) for the 1961-1985 cohorts. North shore recruitment was modelled as a function of the average spring air temperatures in May, the number of days of ice cover exceeding 70%, the Beverton-Holt stock-recruitment function, and the number of days of average wind speeds above 10 knots prior to the onset of ice cover. The climate-based recruitment model (R2== 0.57) for the North Shore demonstrated improved hindcasting ability when compared with Ricker stock-recruitment model (R2== 0.13). Based on the stock—recruitment and climate-based recruitment modelling results, the abundance of spawning stock is more important in determining lake whitefish recruitment in northern Green Bay than in the North Shore area. Climatic factors such as ice cover, wind intensity, and spring temperatures appear to be more important in determining lake whitefish recruitment in the North Shore area than in northern Green Bay. 82 Recruitment models incorporating both abiotic and biotic variables have the potential to produce forecasts of lake whitefish cohort strength several years in advance to recruitment into the fishery. 83 Literature Cited Adams, S. M. and D. L. DeAngelis. 1987. Indirect effects of early bass-shad interactions on predator population structure and food web dynamics. Pages 103-117, In W. C. Kerfoot and A. Sih (eds.), Predation in Aquatic Ecosystems. University of New England Press, Hanover, NH. Bagenal, T. B. 1969. The relationship between food supply and fecundity in brown trout, Salmo trutta. Journal of Fish Biology 1:167-182. Baldwin, N. 8.; R. W. Saalfield; M. A. Ross; and H. J. Buettner. 1979. Commercial fish production in the Great Lakes 1867-1977. Great Lakes Fishery Commission Technical Report No. 3. Beers, J. R. and G. L. Stewart. 1971. Microzooplanktors in the plankton communities of the upper waters of the eastern tropical Pacific. Deep-Sea Research 18:861- 883. Beverton, R. J. H. and S. J. Holt. 1957. On the dynamics of exploited fish populations. United Kingdom Ministry of Agriculture and Fisheries, Fisheries Investigations Series 2. 19:533p. Blaxter, J. H. S. 1965. The feeding of herring larvae and their ecology in relation to feeding. California Cooperative Oceanic Fisheries Investigations Report 10:79-88. Blaxter, J. H. S. 1969. Development: eggs and larvae. Pages 177-252, In W. S. Hoar and D. J. Randall (eds.) Fish Physiology, Volume 3. Academic Press, New York and London. Blaxter J. H. S. 1971. Feeding and condition of Clyde herring larvae. Rapports et Proces-Verbaux des Reunions, Conseil pour L'Exploration de la Mer 160:128-136. Blaxter, J. H. S. 1986. Developement of sense organs and behavior of teleost larvae with special reference to feeding and predator avoidance. Transactions of the American Fisheries Society 115:98-114. 84 Blaxter, J. H. S. and G. Hempel. 1963. The influence of egg size on herring larvae (Clupea harenqus L.). Journal du Consiel, Consiel International pour L'Exploration de la Mer 28:211-240. Brown, M. E. 1946. The growth of brown trout (Salmo trutta L.) I. Factors influencing the growth of trout fry. Journal of Experimental Biology 22:118-129. Brown, R. J. 1968. Population structure and growth characteristics of the whitefish in northern Lake Michigan, 1929-1967. M. S. Thesis, University of Michigan, Ann Arbor, Michigan 61p. Calder, W. A. III 1984. Size, function and life history. Harvard University Press, Cambridge, MA. 431p. Chenoweth, S. B. 1970. Seasonal variation in condition of larval herring in the Boothbay area of the Maine Coast. Journal of the Fisheries Research Board of Canada 27:1875-1879. Christie, W. J. 1963. Effects of artificial propagation and weather on recruitment in the Lake Ontario whitefish fishery. Journal of the Fisheries Research Board of Canada 20:597-603. Christie, W. J. 1974. Changes in the fish species composition of the Great Lakes. Journal of the Fisheries Research Board of Canada 31:827-854. Cochran, W. G. and G. M. Cox. 1950. Experimental Designs. New York: John Wiley & Sons, Inc. Cushing, D. A. 1982. Climate and fisheries. Academic Press, London, England. 373p. Durbin, J. and G. S. Watson. 1951. Testing for serial correlation in least squares regression. Biometrika 37:409-428. Ebener, M. A. and F. A. Copes. 1985. Population statistics, yield estimates, and management for two lake whitefish stocks in Lake Michigan. North American Journal of Fisheries Management 5:435-448. Einsele, W. G. 1963. Problems of fish—larvae survival in nature and in the rearing of economically important middle European freshwater fishes. California Cooperative Oceanic Fisheries Investigations Report 10:24-30. 85 Fleming, I. A. and M. R. Gross. 1990. Latitudinal clines: a trade-off between egg number and size in Pacific salmon. Ecology 71:1-11. Fluchter, J. 1980. Review of the present knowledge of rearing whitefish (Coreqonidae) larvae. Aquaculture. 19:191-208. Fowler, L. G. 1972. Growth and mortality of fingerling Chinook salmon as affected by egg size. Progressive Fish Culturist 34:66-69. Freeberg, M. H. 1985. Early life history factors influencing lake whitefish (Coreqonus clupeaformis) year-class strength in Grand Traverse Bay, Lake Michigan. M. S. Thesis, Michigan State University, 89p. Freeberg, M. H., W. W. Taylor, and R. W. Brown. 1990. Effect of egg and larval survival on the year-class strength of lake whitefish in Grand Traverse Bay, Lake Michigan. Transactions of the American Fisheries Society 119:92-100. Fritz, E. S., L. B. Crowder, and R. C. Francis. 1990. The National Oceanic and Atmospheric Administration Plan for recruitment fisheries oceanography research. Fisheries 15:25-31. Gall, G. A. E. 1974. Influence of size of eggs and age of female on hatchability and growth of rainbow trout. California Fish and Game Journal 6(1):26-35. Gray, J. 1928. The growth of fish: II. The growth rate of the embryo of Salmo fario. British Journal of Experimental Biology 6:110-121. Gulland, J. A. 1965. Survival of the youngest stages of fish and its relation to year class strength. Special Publication ICNAF 6:363-371. Hile, R. 1962. Collection and analysis of commercial fishery statistics in the Great Lakes. Great Lakes Fishery Commission Technical Report 5:24p. Hjort, J. 1914. Fluctuations in the great fisheries of northern Europe viewed in light of biological research. Rapports et Proces-Verbaux des Reunions, Conseil pour L'Exploration de la Mer 20:1-228. 86 Hoagman, W. J. 1974. Feeding by alewives (Algsg pgggggharggggs) on larval lake whitefish (Qo regonus glupgafgrmis) in the laboratory. Journal of the Fisheries Research Board of Canada 31:229-230. Hunter, J. R. 1981. Feeding ecology and predation on marine fish larvae. Pages 33- 79, In R. Lasker (ed. ), Marine Fish Larvae. Washington Sea Grant, University of Washington Press, Seattle, WA. Ivlev, V. S. 1961. Experimental Ecology of the Feeding of Fishes. Translation by Yale Univ. Press, New Haven, Conn. 302 p. Kinnunen, R. E. 1990. Michigan 1989 commercial fish harvest. Commercial Fisheries Newsline 9(2):7-10. Lasker, R., H. M. Feder, G. H. Theilacker, and R. C. May. 1970. Feeding, growth, and survival of E_gragli§ mgrdax larvae reared in the laboratory. Marine Biology 5:345-353. Lawler, G. H. 1965. Fluctuations in the success of year classes of whitefish populations with special reference to Lake Erie. Journal of the Fisheries Research Board of Canada 22:1197-1227. Lee, A. F. S. and J. Gurland. 1975. Size and power of tests for equality of means of two normal populations with unequal variances. Journal of the American Statistical Association 70:933-941. Lloyd, D. G. 1987. Selection of offspring size at independence and other size-versus-number strategies. American Naturalist 129:800-817. Loftus, J. E. and P. F. Hulsman. 1986. Predation on larval lake whitefish (Coregonus glupeafgrmis) and lake herring (Coregonus artedii) by adult rainbow smelt (Qsmerus mordax). Canadian Journal of Fisheries and Aquatic Sciences 43:812-818. May, R. C. 1973. Larval mortality in marine fishes and the critical period concept. Pages 17-23, In J. H. S. Blaxter (ed.), The Early Life History of Fish. Proceedings of a symposium held at Oban, Scotland, May 17-23, 1973. McGinley, M. A., D. H. Temme, and M. A. Geber. 1987. Parental investment in offspring in variable environments: theoretical and empirical considerations. American Naturalist 130:370-398. 87 McMahon, T. A. and J. T. Bonner. 1983. On size and life. Scientific American Inc., New York, NY. 255p. Miller, R. B. 1952. The relative sizes of whitefish year classes as affected by egg planting and the weather. Journal of Wildlife Management 16:39-50. Miller, T. J., L. B. Crowder, J. A. Rice, and E. A. Marschall. 1988. Larval size and recruitment mechanisms in fishes: toward a conceptual framework. Canadian Journal of Fisheries and Aquatic Sciences 45(9):1557-1570. NcComb, T. N. 1989. Trapnet mortality and vital statistics of lake whitefish in Green Bay and northeastern Lake Michigan. Masters Thesis. University of Wisconsin, Stevens Point, Wisconsin. 96p. O'Connell, C. P. and L. P. Raymond. 1970. The effect of food density on survival and growth of early post yolk- sac larvae of northern anchovy (Enqraulis mordax Girard) in the laboratory. Journal of the Exploration of Marine Biology and Ecology 5:187-197. Patriarche, M. 1977. Biological basis for the management of lake whitefish in the Michigan waters of northern Lake Michigan. Transactions of the American Fisheries Society 106:295-308. Peters, R. H. 1983. The ecological implications of body size. Cambridge University Press, Cambridge, UK. 329p. Piehler, G. R. 1967. Age and growth of common whitefish, Coregonus clupeaformis (Mitchell), of northern Lake Michigan and Bay de Noc areas. M. S. Thesis, Michigan State University, East Lansing, Michigan 94p. Piper, R. G.; I. B. McElwain, L. E. Orme; J. P. McCraren; L. G. Fowler: and J. R. Leonard. 1982. Fish Hatchery Management. U. S. Department of the Interior, U. S. Fish and Wildlife Service. Washington, D.C. 517p. Potts, G. W. and R. J. Wooton. 1984. Fish Reproduction. Academic Press, Inc., London, UK. 410p. Reckahn, J. 1970. Ecology of young lake whitefish (Coregonus clupeaformis) in South Bay, Manitoulin Island, Lake Huron. Pages 437-460, In C. C. Lindsey and C. S. Woods (eds). Biology of Coregonid Fishes. University of Manioba Press, Winnipeg, Canada. 88 Ricker, W. E. 1954. Stock and recruitment. Journal of the Fisheries Research Board of Canada 11:559-623. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bulletin of the Fisheries Reseach Board of Canada 191:382p. Rybicki, R. W. 1980. Assessment of lake whitefish populations in northern Lake Michigan. Michigan Department of Natural Resources, Fisheries Division unpublished report (PL 88-309) Lansing, Michigan, 27p. Sargent, R. C., P. D. Taylor, and M. R. Gross. 1987. Parental care and the evolution of egg size in fishes. American Naturalist 129:32-46. Scheerer, P. D. and W. W. Taylor. 1985. Population dynamics and stock differentiation of lake whitefish in northeastern Lake Michigan with implications for their management. North American Journal of Fisheries Management 5:526-536. Schnack, D. 1981. Studies on the mortality of Pacific herring larvae during their early development, using artificial in situ containments. Rapports et Proces- Verbaux des Reunions, Conseil International pour L'Exploration de la Mer 178:135-142. Smale, M. A. 1988. A comparative analysis of lake whitefish (Coregonus clupeaformis) population dynamics in northeastern Lake Michigan. M. S. Thesis, Michigan State University, East Lansing, Michigan. 264p. Smith, B. R. and J. J. Tibbles. 1980. Sea lamprey (Petromvzon marinus) in Lakes Huron, Michigan, and Superior: history of invasion and control, 1936-78. Canadian Journal of Fisheries and Aquatic Sciences 37:1780-1801. Smith, C. C. and S. D. Fretwell. 1974. The optimal balance between size and number of offspring. American Naturalist 108:499-506. Smith, S. 1958. Yolk utilization in fishes. Pages 33-53, In D. Rudnick (ed.), Embryonic Nutrition. University of Chicago Press, Chicago, Illinois. Svardson, G. 1949. Natural selection and egg number in fish. Institute of Freshwater Research, Drottningholm 29:115-122. 89 Taylor, W. W. and M. H. Freeberg. 1984. Effect of food abundance on larval lake whitefish, gorgggnug Mitchell, growth and survival. Journal of Fish Biology 25(6):733-74l. Taylor, W. W., M. A. Smale, and M. H. Freeberg. 1987. Biotic and abiotic determinants of lake whitefish (Coregonus ngpeafgrmis) recruitment in northeastern Lake Michigan. Canadian Journal of Fisheries and Aquatic Sciences 44(Supp. II):313-323. Taylor, W. W., M. A. Smale, and R. W. Brown. 1991. An evaluation of size versus age dependent life history characteristics of lake whitefish stocks in the Upper Great Lakes. Proceedings of the International Symposium on the Biology and Management of Coregonid Fishes (In Press). Wallace, J. C. and D. Aasjord. 1984. An investigation of the consequences of egg size for the culture of Arctic charr, Salyelinus alpings (L.). Journal of Fish Biology 24:427-435. Webb, P. W. and D. Weihs. 1986. Functional locomotor morphology of early life stages of fishes. Transactions of the American Fisheries Society 115:115-127. White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrics 48:817-838. Winkler, D. W. and K. Wallin. 1987. Offspring size and number: a life history model linking effort per offspring and total effort. American Naturalist 129:708-720. Zaret, T. M. 1980. Predation and freshwater communities. Yale University Press, New Haven, CT. 187p. 90 Appendix 1. Appendix 1. Length, weight, and age data of female lake whitefish sampled from Bayport and Naubinway and data for those fish whose eggs were used in laboratory experiments. Field Bayport Naubinway IDI Length Weight Age ID# Length Weight Age 1 605 2000 7 1 572 1000 6 3 522 1300 5 4 499 550 5 5 587 1900 6 7 507 650 4 7 557 1750 5 10 567 1250 6 9 582 1800 6 13 607 1475 7 11 551 1550 5 16 542 960 5 13 587 1850 6 19 509 900 4 15 529 1250 5 22 613 1810 7 17 506 1150 4 25 546 1150 5 19 595 1750 7 28 585 1490 6 21 632 2350 8 31 468 550 4 23 598 2000 7 34 508 800 5 25 541 1525 5 37 537 950 6 27 560 1350 6 40 455 375 4 29 641 2550 8 31 535 1550 5 33 543 1500 5 Mean 569 1713 5.9 Mean 537 994 5.3 SE 9.1 89.3 SE 12.5 105 Laboratory Bayport Naubinway IQi Length Weight Age 121 Length Weigh; Age 3 522 1300 5 1 572 1000 6 7 557 1750 6 7 507 650 4 15 529 1250 5 13 607 1475 7 17 506 1150 4 22 613 1810 7 19 595 1750 7 34 508 800 5 21 632 2350 8 37 537 950 6 Mean 557 1591 5.8 Mean 557 1114 5.8 SE 18 168 SE 18 164 91 Appendix 2. Appendix 2. Egg composition data for female lake whitefish sampled from Bayport and Naubinway. Bayport Wet Weight(mg)/egg Dry Wet(mg)/egg Percent Water Fish # Mean SE Mean SE Mean SE 1 10.16 0.247 2.92 0.000 71.21 0.710 3 8.01 0.134 2.41 0.014 69.83 0.585 5 10.24 0.129 2.98 0.019 70.93 0.367 7 9.46 0.199 2.55 0.005 73.05 0.571 9 9.91 0.233 2.87 0.011 70.97 0.637 11 8.37 0.564 2.41 0.022 70.90 1.653 13 8.88 0.176 2.65 0.003 70.16 0.575 15 7.49 0.064 2.19 0.031 70.69 0.535 17 10.16 0.272 2.96 0.081 70.84 0.127 19 8.65 0.326 2.57 0.048 70.24 0.894 21 11.00 0.200 2.94 0.085 73.32 0.346 23 9.94 0.064 2.66 0.005 73.26 0.213 25 8.15 0.133 2.48 0.018 69.50 0.340 27 8.98 0.078 2.44 0.003 72.79 0.223 29 10.89 0.490 2.69 0.054 75.20 0.852 31 9.51 0.039 2.68 0.007 71.78 0.099 33 8.66 0.072 2.86 0.010 66.92 0.289 Total 9.32 0.242 2.66 0.055 71.27 0.443 Naubinway Wet Weight(mg)/egg Dry Wet(mg)/egg Percent Water Fish # Mean SE Mean SE Mean SE 1 8.34 0.091 2.96 0.010 64.46 0.413 4 7.88 0.052 2.52 0.017 68.01 0.379 7 6.98 0.215 2.30 0.010 66.93 0.948 10 10.83 0.308 3.05 0.011 71.72 0.884 13 8.78 0.068 3.19 0.016 63.67 0.454 16 7.37 0.050 2.52 0.024 65.76 0.179 19 9.47 0.078 2.29 0.015 75.85 0.230 22 8.71 0.134 2.79 0.012 67.90 0.406 25 8.94 0.178 2.84 0.017 68.16 0.452 28 7.60 0.071 2.57 0.022 66.18 0.280 31 8.42 0.166 2.42 0.017 71.22 0.681 34 8.97 0.199 2.68 0.022 70.07 0.502 37 8.01 0.089 2.64 0.003 67.08 0.358 40 5.11 0.052 1.68 0.022 67.07 0.436 Total 8.24 0.339 2.60 0.097 68.15 0.822 Appendix 2. 92 Appendix 2 (Cont.) Egg composition data for female lake whitefish sampled from Bayport and Naubinway. Bayport Cal/Egg Cal/g Fish Number Mean SE Mean SE 1 19.7 0.027 6.74 0.0091 3 15.2 0.091 6.30 0.0050 5 19.0 0.124 6.38 0.0089 7 16.8 0.037 6.61 0.0038 9 19.0 0.078 6.63 0.0104 11 15.4 0.145 6.41 0.0104 13 17.0 0.030 6.43 0.0094 15 14.0 0.201 6.39 0.0039 17 18.8 0.514 6.34 0.0191 19 16.8 0.317 6.55 0.0117 21 19.1 0.552 6.50 0.0030 23 16.8 0.039 6.34 0.0070 25 15.9 0.116 6.41 0.0077 27 16.0 0.029 6.55 0.0092 29 16.7 0.340 6.21 0.0150 31 17.7 0.060 6.60 0.0139 33 18.9 0.066 6.59 0.0050 17.2 0.386 6.47 0.0033 Naubinway Cal/Egg Cal/g Fish Number Mean SE Mean SE 1 19.7 0.069 6.66 0.0079 4 16.4 0.116 6.49 0.0139 7 15.0 0.068 6.52 0.0104 10 20.6 0.079 6.73 0.0093 13 21.3 0.116 6.67 0.0125 16 16.4 0.160 6.52 0.0107 19 15.0 0.101 6.58 0.0058 22 18.1 0.079 6.47 0.0066 25 18.9 0.119 6.63 0.0160 28 17.0 0.144 6.63 0.0067 31 15.5 0.112 6.40 0.0113 34 17.7 0.151 6.59 0.0136 37 17.4 0.046 6.61 0.0161 40 10.8 0.143 6.41 0.0049 17.1 0.688 6.56 0.0256 93 Appendix 2 (Cont.) Appendix 2. Egg composition data for female lake whitefish sampled from Bayport and Naubinway. Bayport Lipid Content Percent Lipid Content (mg/egg) Fish Number Mean SE Mean SE 1 22.81 0.475 0.666 0.011 3 15.45 0.440 0.373 0.025 5 23.43 0.466 0.697 0.010 7 18.85 0.244 0.480 0.014 9 19.39 0.482 0.557 0.008 11 17.65 0.520 0.425 0.007 13 21.52 0.440 0.570 0.010 15 17.60 0.188 0.386 0.020 17 15.44 0.466 0.458 0.007 19 16.18 0.195 0.415 0.010 21 22.22 0.134 0.652 0.022 23 20.55 0.538 0.546 0.011 25 19.64 0.091 0.488 0.007 27 21.69 0.132 0.530 0.011 29 16.73 0.372 0.450 0.010 31 22.08 0.357 0.593 0.006 33 21.72 0.454 0.622 0.011 Total 19.6 0.629 0.524 0.024 Naubinway Lipid Content Percent Lipid Content (mg/egg) Fish Num Mean SE Mean SE 1 23.23 0.423 0.688 0.024 4 20.09 0.581 0.506 0.014 7 17.71 0.462 0.408 0.007 10 23.98 0.382 0.732 0.012 13 20.67 0.399 0.660 0.020 16 16.62 0.584 0.419 0.021 19 18.26 0.128 0.417 0.006 22 18.96 0.544 0.530 0.010 25 16.78 0.356 0.477 0.014 28 21.16 0.393 0.544 0.012 31 13.80 0.186 0.334 0.009 34 20.24 0.428 0.543 0.006 37 18.75 0.442 0.494 0.024 40 16.95 0.556 0.285 0.019 Total 19.09 0.703 0.502 0.033