ESTIMATING  EFFECTIVE  NUMBER  OF  BREEEDING  ADULTS,  REPRODUCTIVE   SUCCESS  AND  SPAWNING  DURATION  FOR  WHITE  STURGEON  IN  THE  UPPER   COLUMBIA  RIVER,  CANADA   By   Kathleen  Joan  Jay                                             A  THESIS   Submitted  to   Michigan  State  University   in  partial  fulfillment  of  the  requirements   for  the  degree  of   Fisheries  and  Wildlife  –  Master  of  Science   2014   ABSTRACT   ESTIMATING  EFFECTIVE  NUMBER  OF  BREEEDING  ADULTS,  REPRODUCTIVE   SUCCESS  AND  SPAWNING  DURATION  FOR  WHITE  STURGEON  IN  THE  UPPER   COLUMBIA  RIVER,  CANADA   By   Kathleen  Joan  Jay   The  white  sturgeon  (Acipenser  transmontanus)  population  in  the  Upper   Columbia  River,  Canada,  has  been  undergoing  recruitment  failure  for  several   decades.    Uncertainty  exists  regarding  white  sturgeon  reproductive  ecology   including  the  number  of  adults  annually  contributing  offspring  and  accuracy  of   estimating  the  duration  of  spawning  activity.    The  effects  of  temperature  on  white   sturgeon  larval  development  were  examined  to  improve  estimates  of  fertilization   dates  for  wild  caught  larvae  and  number  of  spawning  days.    Molecular  techniques   allowed  for  the  examination  of  levels  of  recruitment  to  the  egg  or  larval  stage   and/or  reproductive  success  of  adult  fish.    Microsatellite  loci  (N=12)  and  likelihood-­‐ based  pedigree  analysis  were  used  to  quantify  the  number  of  spawning  adults  (N),   effective  breeding  number  (Nb)  and  individual  reproductive  success  from  eggs  and   larvae  collected  in  each  of  two  consecutive  years.    Genetically  derived  estimates  of   numbers  of  spawning  adults  (N)  were  concordant  with  empirical  estimates  of  the   number  of  annual  available  breeders  (Nc)  based  on  sex  ratios  and  maturation  stages   of  adults.    Results  are  fundamental  to  our  understanding  of  white  sturgeon   spawning  strategies  and  can  be  applied  to  revise  ongoing  recovery  strategies  (e.g.   conservation  aquaculture  programs).     This  work  is  dedicated  to  my  late  father,  Bill  Jay,  whose  impact  on  my  life  has  been   immeasurable  and  continues  to  this  day.     iii   ACKNOWLEDGEMENTS       I  would  like  to  thank  first  and  foremost  my  advisor,  Dr.  Kim  Scribner,  for  his   guidance  and  encouragement  throughout  the  course  of  this  project.    While  working   as  a  technician  at  the  Black  Lake  Research  Facility,  he  recognized  my  potential  and   was  patient  as  I  tried  on  the  ‘genetics  hat’.    Thank  you  to  BC  Hydro  for  their   generous  financial  support  towards  research  and  project  costs,  particularly  Dr.   James  Crossman  for  making  this  happen.    My  thesis  work  would  not  have  been   possible  without  the  funding,  equipment,  services,  data  and  advice  James  provided.     I  also  thank  the  Freshwater  Fisheries  Society  of  BC  and  the  Kootenay  Trout   Hatchery  for  providing  me  the  required  space,  equipment  and  eggs  for  my   experiment.    Ron  Ek,  Mike  Keehn,  Chad  Fritz,  Aaron  Wolff,  Ashley  Uittenbogaard,   Sten  Lundgren  and  Gord  Frew  welcomed  me  into  their  workplace  and  homes  and   kept  me  sane  while  microscopically  viewing  thousands  of  larvae  for  hours  on  end.     Field  sampling  would  not  have  been  possible  without  the  help  of  Golder  Associates   Ltd.,  Terraquatic  Resource  Management,  and  BC  Hydro  employees  James  Baxter  and   Dean  Den  Biesen.    Special  thanks  go  to  Marco  Marrello  for  his  invaluable  help  in  the   field.    I  am  forever  indebted  to  him  for  his  efforts,  enthusiasm  and  countless   McDonald’s  teas.    Thank  you  all  for  making  my  field  seasons  so  enjoyable  and  filled   with  laughter.              I  thank  all  my  lab  mates  in  the  Scribner  lab  for  their  insights,  perspectives   and  support.    I  am  sincerely  indebted  to  John  Bauman  for  providing  his  invaluable   expertise,  guidance  and  encouragement.    He  heightened  my  interests  in  sturgeon     iv   research  and  supported  me  every  step  of  the  way.    I  am  deeply  grateful  to  Dr.   Jeannette  Kanefsky  for  her  assistance  and  support  within  the  lab  and  forcing  me  to   step  away  and  take  a  break  (“So  many  bands!”).        I  thank  my  undergraduate   assistants,  Phil  Ganz  and  Kaitlin  Clark,  for  their  tireless  efforts.    Neither  complained   after  countless  hours  of  photographing  and  measuring  what  must  have  seemed  like   the  same  larvae  over  and  over.    I  would  also  like  to  acknowledge  the  contributions   of  my  graduate  committee,  Dr.  Dan  Hayes  and  Dr.  Mike  Wagner,  for  their  advice  and   encouragement.       I  sincerely  thank  my  mom  who  always  supported  me  to  follow  my  interests   although  it  entailed  traveling  to  different  provinces,  countries  and  continents.    I   thank  my  cousin,  Cindy  Gilbert,  who  inspired  me  to  pursue  research  within  this   field.    I  also  thank  my  family  and  friends  who  supported  my  decision  to  go  back  to   school.    I  am  grateful  to  my  friends  within  the  department  and  the  East  Lansing  Beer   Club  who  became  my  support  group  throughout  this  degree,  especially  my  co-­‐ president/founder,  Howie  Singer,  and  good  friend,  Kyle  Herreman.    Special  thanks   go  to  Paige  Howell  for  her  love,  constant  support,  valued  input  and  analytical   assistance.    You  have  been  vital,  in  every  way,  to  the  completion  of  this  thesis.     Thank  you,  Best  Brew,  for  your  faith  in  me,  firm  encouragement  and  much-­‐needed   distractions.   And  finally,  I  thank  the  ‘cup’  experiment.    I  would  gladly  measure  every  rock   in  the  hatchery  again  as  I  would  not  be  where  I  am  today  if  it  was  not  for  this   notorious  experiment.                 v   TABLE  OF  CONTENTS     LIST  OF  TABLES......................................................................................................................vii   LIST  OF  FIGURES .................................................................................................................. viii   THESIS  INTRODUCTION......................................................................................................... 1   CHAPTER  I:  INVESTIGATION  OF  TEMPERATURE  EFFECTS  ON  YOLK-­SAC   LARVAL  DEVELOPMENT  OF  WHITE  STURGEON  (ACIPENSER  TRANSMONTANUS)   AND  ESTIMATION  OF  FERTILIZATION  DATE ................................................................. 5   ABSTRACT............................................................................................................................................5   INTRODUCTION .................................................................................................................................6   METHODS.............................................................................................................................................9   Gamete  Collection,  Fertilization  and  Incubation ................................................................................. 9   Statistical  analysis ........................................................................................................................................ 11   RESULTS ............................................................................................................................................ 12   DISCUSSION ...................................................................................................................................... 14   Management  application:  estimating  fertilization  date  of  wild  caught  yolk-­sac  larvae .. 15   CHAPTER  II:  ESTIMATES  OF  EFFECTIVE  NUMBER  OF  BREEDING  ADULTS  AND   REPRODUCTIVE  SUCCESS  FOR  WHITE  STURGEON  ACIPENSER   TRANSMONTANUS ..................................................................................................................17   ABSTRACT......................................................................................................................................... 17   INTRODUCTION .............................................................................................................................. 19   METHODS.......................................................................................................................................... 23   Study  Area........................................................................................................................................................ 23   Egg  and  larval  collection............................................................................................................................ 26   Developmental  staging  and  estimation  of  fertilization  date........................................................ 29   Genetic  analysis.............................................................................................................................................. 29   Data  Conversion............................................................................................................................................. 31   Pedigree  analysis........................................................................................................................................... 32   Comparisons  of  genetic  and  empirical  estimates  of  mature  adult  population ..................... 34   RESULTS ............................................................................................................................................ 35   Egg  and  larval  collection  and  estimation  of  fertilization  date.................................................... 35   Pedigree  analysis........................................................................................................................................... 38   Comparison  of  empirical  and  genetic  data ......................................................................................... 43   DISCUSSION ...................................................................................................................................... 45   FINAL  CONCLUSIONS ............................................................................................................52   LITERATURE  CITED ..............................................................................................................58       vi   LIST  OF  TABLES     Table  1.  Mean  total  (hours)  and  cumulative  thermal  units  (CTU)  required  to  reach   developmental  stages  of  white  sturgeon  yolk-­‐sac  larvae  from  time  of   fertilization  at  experimental  temperature  regimes  of  12.5°C,  14.0°C,  15.5°C  and   17.0°C. ................................................................................................................................................ 13   Table  2.    RTi  (relative  time  of  development  stage  i,  as  a  proportion  of  the  total   duration  of  yolk-­‐sac  larval  period)  for  white  sturgeon  larvae  at  four   temperature  treatments  (12.5°C,  14.0°C,  15.5°C  and  17.0°C).    Time  is  quantified   in  hours  and  cumulative  thermal  units  (CTU).................................................................. 13   Table  3.  Relationship  between  developmental  stages  (y)  and  transition  time  to  stage   (hours  and  CTU  post  fertilization;  x)  as  a  function  of  incubation  temperature   following  the  model  y  =  mx  +  b,  where  m  is  the  slope  and  b  is  the  intercept. .... 14   Table  4.  Estimation  of  White  Sturgeon  Nb,  Ns  and  Nk  based  on  pedigree  analyses  of   wild  larvae  of  unknown  parentage  collected  by  means  of  egg  mat  and  drift  net   from  the  Upper  Columbia  River  in  summer  2011  and  2012...................................... 39   Table  5.  Comparison  of  pedigree  analyses  using  reduced  number  of  White  Sturgeon   tissue  samples  to  determine  the  effects  of  subsampling  on  pedigree   reconstruction.    Estimates  include  Nb,  Ns  and  Nk  of  collected  progeny  from  the   Upper  Columbia  River  in  summer  2011  at  the  sites  of  Waneta  and  ALH............. 42   Table  6.  Variation  in  inferred  Upper  Columbia  River  White  Sturgeon  adult  (Table  1)   reproductive  success  between  sampling  years  (summer  2011  and  2012)  and   among  sites  (ALH  and  Waneta)  including  number  of  contributing  inferred   adults  per  estimated  spawning  day  (ESD),  number  of  breeding  partners  per   inferred  adult,  and  number  of  ESD  per  inferred  adult  (mean  ±  SD)....................... 43   Table  7.  Empirical  (Nc)  and  genetic  (Ns;  Table  1)  estimates  of  the  proportion  of  the   total  Canadian  Upper  Columbia  River  White  Sturgeon  population  (1,157  [414,   1889;  95%  CI];  Irvine  et  al.  2007)  in  spawning  condition.    Empirical  estimates   of  number  of  individuals  in  spawning  condition  were  calculated  based  on  sex   ratios  (1:1)  and  maturation  stages  of  adults  captured  via  set  line  during  the   2009  –  2012  broodstock  programs  (June)......................................................................... 44       vii   LIST  OF  FIGURES     Figure  1.  Transboundary  Reach  of  the  Upper  Columbia  River  between  Hugh  L.   Keenleyside  Dam  (HLK),  British  Columbia,  Canada,  and  Gifford,  Washington,   USA.    Egg  mat  and  drift  nets  were  deployed  in  2011  and  2012  within  in  the   Canadian  Portion  of  the  Upper  Columbia  River.    Sampling  sites  include:  (A)   Arrow  Lakes  Generating  Station  (ALH,  river  kilometer  0.1  (rkm),  including   HLK),  (B)  18.2  rkm,  and  (C)  Waneta  (56.0  rkm). ........................................................... 24 Figure  2.  Estimated  spawning  activity  duration  and  sample  size  of  eggs  and  larvae  at   the  spawn  monitoring  sites  of  ALH  and  Waneta  for  the  (A)  2011  and  (B)  2012   spawning  seasons.    Estimated  fertilization  date  was  back  calculated  for  wild   caught  eggs  and  larvae  as  a  function  of  developmental  stage  (Detlaff  et  al.  1993)   and  temperature  (Jay  unpublished;  Parsley,  U.S.  Geological  Survey,   unpublished).    Bars  represent  estimated  fertilization  date  for  samples  collected   at  Arrow  Lakes  Generating  Station  (ALH;  eggs:  dark  grey;  larvae:  light  grey)  and   Waneta  (eggs:  black;  larvae:  white)...................................................................................... 36 Figure  3.  Number  of  inferred  full-­‐sibling  families  per  kin  group  for  each  year  at  the   spawning  sites  of  Arrow  Lakes  Generating  Station  (ALH),  18.2  rkm  and  Waneta.     Numbers  above  box  plots  represent  the  number  of  kin  groups  inferred  at  each   site. ...................................................................................................................................................... 39 Figure  4.  Relationship  between  number  of  spawning  partners  and  number  of   offspring  produced  (as  a  measure  of  reproductive  success)  for  inferred  adults   at  sites  of  Arrow  Lakes  Generating  Station  (ALH;  solid  line)  and  Waneta  in  2011   (dotted  line)  and  2012  (dash/dot  line).  Lines  are  of  best  fit. .................................... 40   viii   THESIS  INTRODUCTION   White  sturgeon  (Acipenser  transmontanus)  are  a  native  fish  of  western  North   America  and  have  undergone  significant  reductions  in  population  abundance  and   distribution  due  to  over-­‐exploitation,  environmental  degradation  and  loss  of  habitat   connectivity  (Billard  and  Leconintre  2001).    Four  of  six  populations  inhabiting   Canada  (Upper  Columbia,  Upper  Fraser,  Kootenay  and  Nechako  rivers)  have  been   listed  as  endangered  under  Canada’s  Species  at  Risk  Act.    Recruitment  failure  has   been  identified  as  the  principle  cause  of  endangerment  (Jager  et  al  2001,  Anders  et   al.  2002,  McAdam  et  al.  2005,  Irvine  et  al.  2007).    Causes  of  recruitment  failure  are   currently  unclear;  however,  regulation  of  river  flow  is  believed  to  be  a  significant   factor  (McAdam  et  al.  2005,  Paragamian  et  al.  2001,  Paragamian  et  al.  2009,  UCWSRI   2002).       River  regulation,  through  hydroelectric  dams,  causes  an  alteration  of  fish   migration  routes,  destruction  of  spawning  habitat,  loss  of  food  resources,  altered   seasonal  flow  and  temperatures  regimes,  changes  in  benthic  substrate  and   effectively  isolates  upper  river  populations  (Brown  et  al.  1992,  McAdam  et  al.  2005).       Isolated  populations  have  increased  vulnerability  of  extinction  through  disease,   catastrophic  events,  genetic  isolation,  and  density-­‐dependent  responses  (Auer   1996).    Studies  suggest  that  these  environmental  changes  affect  all  life  history   phases  of  white  sturgeon  particularly  limiting  egg  and  larval  survival,  which  leads  to   recruitment  failure  (McAdam  et  al.  2005,  Paragamian  et  al.  2001,  Paragamian  et  al.   2009,  UCWSRI  2002).         1   White  sturgeon  exhibit  life  history  characteristics  of  late  age  at  maturity,   infrequent  spawning,  and  low  rates  of  natural  recruitment,  which  further   complicate  restoration  and  recovery  efforts.    White  sturgeon  reach  sexual   maturation  between  11  and  34  years  of  age  (Semakula  and  Larkin  1968).    Once   mature,  white  sturgeon  spawn  intermittently.    Males  spawning  every  two  to  four   years  while  females  spawn  less  frequently,  every  three  to  eleven  years  (Billard  and   Leconintre  2001).    Mortality  rates  during  early  ontogenetic  stages  are  naturally   high.    Eggs  and  larvae  are  particularly  vulnerable  to  extreme  temperatures,  abrupt   temperature  changes,  poor  habitat  availability  and  predation  (Parsley  et  al.  2002,   Parsley  et  al.  1993).    Degradation  to  spawning  and  rearing  grounds  due  to   extraction  of  gravel  or  cobble  stone,  modification  in  water  flow  caused  by   channeling  or  fluctuating  water  levels  below  dams,  and  fluctuation  in  temperature   and  oxygen  levels  can  further  prevent  successful  reproduction  and  recruitment   (Rochard  1990).         The  Columbia  River  Water  Use  Plan  (WUP)  assigns  a  high  priority  to   conservation  and  recovery  of  white  sturgeon  in  the  Upper  Columbia  River  (UCR),   Canada  (UCWSRI  2002);  a  goal  with  the  potential  to  conflict  with  the  river’s   designation  as  a  “working  river”.    Consequently,  management  of  white  sturgeon  is   currently  limited  to  non-­‐operational  habitat  improvements  designed  to  increase   adult  spawning  success  and  juvenile  survival.     Estimates  of  reproductive  success  or  failure  (e.g.,  based  on  egg  and  larval   captures)  at  known  and  suspected  spawning  locations  in  the  UCR  are  critical  to   addressing  management  questions  related  to  recruitment.    In  order  to  meet     2   management  goals,  data  are  required  to  assess  annual  spawner  abundance  and   spawning  success.    My  thesis  research  focuses  on  white  sturgeon  reproductive   ecology  including  estimating  spawning  duration,  number  of  adults  contributing  to   natural  recruitment  and  variation  in  individual  reproductive  success.    In  Chapter  I,  I   investigate  the  effects  of  four  temperature  regimes  (12.5°C,  14.0°C,  15.5°C  and   17.0°C)  on  rate  of  yolk-­‐sac  larval  development  in  a  common  garden  experiment.     With  these  data,  I  was  able  to  estimate  fertilization  and  hatch  dates  of  wild  caught   larvae  from  the  date  of  capture.    Estimating  white  sturgeon  spawning  period  within   the  UCR  using  staged  eggs  has  been  used  as  the  primary  measure  of  spawning   activity  for  many  years  being  the  best  available  metric  describing  start  and  end   dates  of  spawning  activity.    By  including  staged  larvae,  the  estimation  of  spawning   duration  provides  a  better  representation  of  all  spawning  days.      In  Chapter  II,  I  used   microsatellite  loci  and  likelihood-­‐based  pedigree  analysis  to  quantify  the  number  of   spawning  adults,  effective  number  of  breeding  adults,  and  kin  group  number  and   size.    Additionally,  I  examined  individual  reproductive  success  and  spawning   duration,  of  which  were  unknown  for  white  sturgeon  population  prior  to  this  study.     Additionally,  I  compared  genetically  derived  estimates  of  the  number  of  spawning   adults  to  empirical  estimates  based  on  sex  ratios  and  visual  determinations  of   maturation  stages  of  adults  captured  during  the  pre-­‐spawning  season  to  determine   if  one  technique  may  be  substituted  for  estimating  spawning  population  size  if  the   other  technique  is  not  available  to  researchers.    Results  increase  understanding  of   white  sturgeon  reproductive  ecology  and  recruitment,  and  allow  projections  of     3   cohort  levels  of  genetic  diversity.    Similar  data  can  be  applied  to  recovery  planning   and  aquaculture  programs  for  this  and  other  species  of  conservation  concern.             4   CHAPTER  I:  INVESTIGATION  OF  TEMPERATURE  EFFECTS  ON  YOLK-­‐SAC  LARVAL   DEVELOPMENT  OF  WHITE  STURGEON  (ACIPENSER  TRANSMONTANUS)  AND   ESTIMATION  OF  FERTILIZATION  DATE     ABSTRACT   Temperature  is  widely  considered  to  be  the  most  important  factor   influencing  early  ontogenetic  development  in  fishes.    The  effect  of  constant   incubation  temperatures  (12.5°C,  14.0°C,  15.5°C  and  17.0°C)  on  developmental  rates   of  white  sturgeon  yolk-­‐sac  larvae  was  examined  in  a  common  garden  experiment.     Linear  models  revealed  timing  of  development  significantly  increased  as  incubation   temperature  increased  when  quantified  in  hours  (F=240.14,  P=0.0041).     Developmental  rate  was  not  affected  by  temperature  when  rate  of  development  was   quantified  in  cumulative  thermal  units  (CTU)  (F=3.877,  P=0.1878).    Data  provided   an  index  for  the  transition  timing  (hours)  between  yolk-­‐sac  larvae  developmental   stages  as  a  function  of  temperature  to  be  used  for  estimating  fertilization  date  of   wild  caught  yolk-­‐sac  larvae,  adult  spawning  duration  and  number  of  spawning  days   of  white  sturgeon.    Estimation  of  timing  and  duration  of  spawning  activity  is  an   important  tool  for  managing  endangered  populations  in  large  river  systems.       5   INTRODUCTION   Temperature  is  an  important  driver  of  early  development,  affecting  traits   associated  with  life  history  and  survival  (Morbey  and  Ydenberg  2003).    Warmer   temperatures  lead  to  more  rapid  consumption  of  the  yolk-­‐sac  reserves,  earlier  onset   of  first  feeding  (Wang  et  al.  1987,  Hardy  and  Litvak  2004),  and  increases  growth  and   developmental  rates  during  early  ontogenetic  stages  (Atkinson  1994,  Wang  et  al.   1985),  size  at  hatch  (Laurel  and  Blood  2011,  Ojanguren  &  Braña  2003),  and   incidence  of  deformities  (Lahnsteiner  et  al.  2012).    Understanding  the  effects  of   temperature  during  early  ontogenetic  stages  is  important  to  understand  egg  and   larval  behavior,  distribution,  and  adaptability  to  variable  environmental  conditions.     Limited  capabilities  with  respect  to  swimming,  feeding,  and  predator  avoidance,  and   general  capacity  to  respond  adaptively  to  variation  in  the  surrounding  environment   can  result  in  high  rates  of  mortality  in  many  fish  species  during  the  egg  and  yolk-­‐sac   larval  stages  (Miller  et  al.  1988,  Killgore  et  al.  1987,  Rice  et  al.  1987,  Sheperd  et  al.   2000).       Similar  to  other  Acipenseriform  species,  white  sturgeon  are  characterized  by   a  number  of  demographic  and  life  history  characteristics  that  affect  their  ability  to   adapt  to  natural  and  anthropogenically  altered  environmental  conditions.    White   sturgeon  (Acipenser  transmontanus)  are  a  long-­‐lived  species  of  conservation   concern  that  includes  many  endangered  populations  experiencing  recruitment   failure  (Jager  et  al.  2001,  Anders  et  al.  2002,  McAdam  et  al.  2005,  Irvine  et  al.  2007).     Females  spawn  every  3  to  11  years  while  males  spawn  at  intervals  of  2  to  4  years   (Billard  and  Leconintre  2001).    Reproductively  active  adults  congregate  in  late     6   spring  and  summer  months  and  exhibit  a  promiscuous  mating  behavior  (Billard  and   Leconintre  2001).    Gametes  are  broadcasted  over  the  river  substrate  (ideally  gravel   and  cobble)  where  they  remain  until  hatch  with  no  parental  care  provided  (Parsley   et  al.  1993).    Newly  hatched  yolk-­‐sac  larvae  lack  many  structures  necessary  for   movement,  sensory  perception,  and  resource  acquisition  (Beers  1981).    As  a  result,   yolk-­‐sac  larvae  burrow  into  river  substrate  and  generally  remain  there  until   endogenous  yolk  reserves  are  utilized  (McAdam  2011).    Although  females  are  highly   fecund,  white  sturgeon  experience  extremely  high  mortality  during  early  life  stages   due  to  the  inefficient  gamete  fertilization  system,  adult  mating  behavior  and   exposure  of  eggs  and  larvae  to  physical  and  biotic  environmental  conditions  (Billard   and  Leconintre  2001).    Additionally,  female  size,  age  and  health  can  influence  egg   size,  yolk  quality  and  subsequent  larval  size  and  yolk  sac  reserves  (Hardy  and  Litvak   2004,  Parsley  et  al.  unpublished).         In  order  to  determine  the  annual  success  of  white  sturgeon  spawning,   reliable  data  on  white  sturgeon  adult  abundance,  timing  and  locations  of  spawning   activity  is  required  and  important  for  population  management.    However,  in  large   river  systems  these  data  can  be  labor-­‐intensive  and  difficult  to  obtain.    Estimating   white  sturgeon  spawning  period  within  the  UCR  using  staged  eggs  has  been  used  as   the  primary  measure  of  spawning  activity  for  many  years  and  has  been  the  best   available  metric  describing  start  and  end  dates  of  spawning  activity.    Naturally   produced  eggs  can  be  captured  using  methods  including  egg  mats  and  drift  nets   (Crossman  et  al.  2011).    Estimates  of  adult  spawning  duration  and  number  of   spawning  days  is  possible  by  back  calculating  fertilization  dates  based  on  time  and     7   date  of  capture,  developmental  stages  of  eggs  and  mean  incubation  water   temperature  (Golder  Associates  Ltd  2009).    These  methods  can  also  be  applied  to   staged  larvae  to  provide  a  better  representation  of  all  spawning  days,  however,  data   describing  the  effects  of  temperature  on  the  development  of  yolk  sac  larvae  is   required.       Detlaff  et  al.  (1993)  documented  development  during  egg  and  yolk-­‐sac  larval   stages.    Forty-­‐five  distinct  stages  were  classified  from  the  time  of  fertilization  to  the   onset  of  feeding.    These  authors  examined  the  effects  of  temperature  on   development  in  three  sturgeon  species  from  the  Caspian  and  Azov  seas.    Following   this  classification  of  developmental  stages,  Parsley  et  al.  (2011)  quantified  the   effects  of  temperature  on  white  sturgeon  egg  development.    Beer  (1981)  examined   yolk-­‐sac  larval  development  under  a  single,  partially  controlled  temperature  regime   for  white  sturgeon.    Wang  et  al.  (1985)  investigated  the  effects  of  controlled   temperature  conditions  ranging  from  11  to  26°C  on  a  limited  number  of  yolk-­‐sac   larval  developmental  stages.    Developmental  data  of  yolk-­‐sac  larvae  exposed  to   different  temperature  regimes  is  lacking  but  required  to  representing  the  typical   thermal  regimes  experienced  by  early  life  stages  during  different  times  and  in   different  locations  in  natural  river  systems  (Billard  and  Leconintre  2001,  Parsley   and  Beckman  1994,  Wang  et  al.  1985,  Perrin  et  al.  2003,  Parsley  et  al.  1993,   Paragamian  et  al.  2001).       The  main  objective  of  this  study  was  to  quantify  the  effects  of  different   temperatures  on  white  sturgeon  yolk-­‐sac  larval  development  quantified  in  time   (hours)  and  cumulative  thermal  units  (CTU;  Kempinger  1988)  of  exposure.    The     8   development  of  a  temperature  index  for  yolk-­‐sac  larval  staging  can  be  used  as  a   management  tool  to  increase  understanding  of  white  sturgeon  reproductive   ecology.    The  addition  of  staged  larvae  to  past  methods  of  staging  eggs  will  provide  a   better  representation  of  all  spawning  days.     METHODS   Gamete  Collection,  Fertilization  and  Incubation   White  sturgeon  eggs  were  obtained  from  adults  collected  and  held  as  a   captive  broodstock  from  the  Upper  Columbia  River  (UCR)  and  held  at  the  Kootenay   Trout  Hatchery,  Fort  Steele,  British  Columbia.    Spawning  was  induced  with  the  use   of  CCP  (acetone-­‐dried  common  carp  pituitary)  hormone.    Eggs  from  two  females   were  fertilized  with  milt  of  two  males  to  produce  two  full-­‐sibling  families  (F1  and   F2).    Parsley  et  al.  (2011)  documented  maternal  variation  in  development  and   growth  at  the  egg  and  larval  stages.    Therefore  eggs  collected  from  more  than  one   female  were  used.       Fertilized  eggs  from  each  female  were  divided  into  four  groups  and   incubated  in  MacDonald  jars  at  water  temperatures  of  12.5°C,  14°C,  15.5°C  and   17°C.    Hatched  yolk-­‐sac  larvae  were  collected  and  placed  into  family  specific  tanks   (152.4  cm  x  76.2  cm  x  20.3  cm)  that  were  maintained  at  one  of  the  aforementioned   incubation  temperatures  through  the  duration  of  the  experiment.    All  tanks  were   covered  to  eliminate  effects  of  overhead  light  and  lined  with  artificial  substrate  (1”   diameter  sinking  Bio-­‐Spheres;  Dynamic  Aqua-­‐Supply  Ltd.  Surry,  BC)  allowing  yolk-­‐ sac  larvae  to  burrow  into  interstitial  spaces.    Substrate  was  provided  to  prevent  an     9   increase  in  energy  consumption  among  yolk-­‐sac  larvae  searching  for  cover  that  can   reduce  growth  (McAdam  2011,  Boucher  2012)  or  development.     At  hatch,  a  random  subsample  of  10  individuals  from  each  family  and   temperature  treatment  was  euthanized  by  overdose  of  MS-­‐222  (tricaine   methanesulfonate)  and  preserved  in  Prefer  buffer  (solution  of  glyoxcal,  buffer  and   alcohol).    Subsample  collection  and  preservation  (12.5°C:  F1  n=290,  F2  n=280;  14°C:   F1  n=240,  F2  n=240;  15.5°C:  F1  n=180,  F2  n=170;  17°C:  F1  n=170,  F2  n=150)   continued  at  12-­‐hour  intervals  following  hatch  until  complete  absorption  of  the  yolk   sac  (stage  45  with  the  initiation  of  exogenous  feeding).    Preserved  samples  were   randomly  examined  with  respect  to  temperature  treatment  and  family  (to  reduce   observer  bias)  using  a  digital  compound  microscope  (Nikon  SMZ-­‐745t  Stereo   Microscope  with  10X  eyepiece)  and  assigned  a  developmental  stage.    Enumeration   of  stages  corresponded  to  the  classification  by  Dettlaff  et  al.  (1993)  including  10   stages  from  hatch  (stage  36)  to  exogenous  feeding  (stage  45).    Each  developmental   stage  was  associated  with  the  appearance  of  at  least  one  new  morphological  feature;   therefore  stages  were  not  determined  strictly  by  quantitative  changes.    Stage  at   hatch  was  verified  and  not  assumed  to  occur  at  stage  36.    Based  on  previous  studies   (Wang  et  al.  1985,  Beer  1981,  Dettlaff  et  al.  1993),  12  hours  was  assumed  to  be  a   sufficient  time  period  to  allow  for  observations  of  the  10  yolk-­‐sac  larval   developmental  stages.    Since  larvae  were  not  provided  a  food  source,  and  therefore   the  act  of  feeding  could  not  be  verified,  44  was  the  latest  assigned  stage.       Time  at  hatch  was  recorded  and  sampling  intervals  began  when   approximately  50%  of  individuals  hatched  within  a  treatment.    To  determine  the     10   timing  of  transition  between  successive  developmental  stages,  the  post  hatch  time   (hours)  was  assigned  when  50%  of  the  sampled  yolk-­‐sac  larvae  achieved  a  stage.   Timing  of  transition  to  a  given  developmental  stage  was  measured  both  as  time   (hours)  and  cumulative  thermal  units  (CTU;  Kempinger  1988).    The  relative  time  (in   hours  and  CTU)  to  reach  each  developmental  stage  was  estimated  following  the   formula  (Klimogianni  et  al.  2004):   [Eq.  1]  RTi  =  (ti/Tsd)*100%     where  RTi  is  the  relative  time  of  developmental  stage  i,  ti  is  the  time  interval  from  t0   (time  at  hatch;  t0  =  0)  to  developmental  stage  i,  and  Tsd  is  total  duration  of  the  yolk-­‐ sac  larval  period  (here  defined  as  total  time  (hours  or  CTU)  of  development  from   stage  36  to  stage  44  as  described  above).           Statistical  analysis   The  relationship  between  developmental  stage  (y)  and  transition  time  to   stage  (hours  and  CTU  post  fertilization;  x)  as  a  function  of  incubation  temperature   following  the  model  y  =  mx  +  b,  where  m  is  the  slope  and  b  is  the  intercept  was  fitted   by  least  squares  regression  for  each  temperature  treatment.    Analysis  of  variance   (ANOVA)  was  used  to  test  for  differences  in  developmental  rates  and  RTi  among   temperature  treatments  measured  in  both  hours  and  CTU.    Analyses  were   performed  using  the  statistical  software  "R"  (R  Development  Core  Team  2011).   11   RESULTS   All  nine  yolk-­‐sac  larval  stages  were  observed  using  the  12-­‐  hour  interval   subsampling  regime  in  all  four  treatments  (Table  1).    Stage  44  was  observed  at  156-­‐ 168  hours  post  hatch  at  17°C,  192  hours  post-­‐hatch  at  15.5°C,  264  hours  post-­‐hatch   at  14°C  and  324-­‐336  hours  post-­‐hatch  at  12.5°C  (Table  1).    When  quantified  in  CTU,   stage  44  was  observed  at  72.8  -­‐  78.4  at  17°C,  77.6  at  15.5°C,  90.2  at  14°C  and  90.5-­‐ 93.8  at  12.5°C  post  hatch  (Table  1).    Relative  time  (RTi)  to  reach  each  developmental   stage  was  not  affected  by  temperature  when  quantified  in  hours  (F=0.2417,   P=0.8668)  or  CTU  (F=0.9797,  P=0.4042)  (Table  2).    All  regressions  of  the   developmental  stage  and  transition  time  relationship  yielded  significant  F  values  (P   ≤  0.05)  and  R2  values  ranged  from  0.612  to  0.954  (Table  3).    Relative  to  12.5°C   exposure,  the  time  required  to  develop  to  stage  44  from  hatch  decreased,  on   average,  by  20.0%  at  14°C,  41.8%  at  15.5°C  and  50.9%  at  17°C.    The  time  to   transition  significantly  increased  as  temperature  increased  (F=240.14,  P=0.0041).     However,  rate  of  development  was  not  significantly  different  between  temperature   conditions  when  quantified  in  CTU  (F=3.877,  P=0.1878).       12   Table  1.  Mean  total  (hours)  and  cumulative  thermal  units  (CTU)  required  to  reach   developmental  stages  of  white  sturgeon  yolk-­‐sac  larvae  from  time  of  fertilization  at   experimental  temperature  regimes  of  12.5°C,  14.0°C,  15.5°C  and  17.0°C.       Temperature   (°C)   12.5   14.0   15.5   17.0     Hours  Post  Fertilization   Stage           36  (Hatch)   240   196   170   148   37   270   208   188   160   38   288   250   206   184   39   312   268   218   196   40   336   280   230   208   41   396   322   260   232   42   480   382   314   256   43   534   430   338   286   44   570   460   362   310                               12.5   14.0   15.5   17.0   CTU  Post  Fertilization           67.0   65.6   67.9   67.2   75.4   69.7   75.2   72.8   80.4   84.1   82.5   84.0   87.1   90.2   87.3   89.6   93.8   94.3   92.2   95.2   110.6   108.7   104.3   106.4   134.0   129.2   126.1   117.6   149.1   147.6   135.8   131.6   159.1   155.8   145.5   142.8   Table  2.    RTi  (relative  time  of  development  stage  i,  as  a  proportion  of  the  total   duration  of  yolk-­‐sac  larval  period)  for  white  sturgeon  larvae  at  four  temperature   treatments  (12.5°C,  14.0°C,  15.5°C  and  17.0°C).    Time  is  quantified  in  hours  and   cumulative  thermal  units  (CTU).           Temperature  (°C)     Stage   36  (Hatch)   37   38   39   40   41   42   43   44   12.5   14.0   15.5   17.0   RTi  (Hours  Post  Hatch)           0   0   0   0   0.09   0.05   0.09   0.07   0.15   0.20   0.19   0.22   0.22   0.27   0.25   0.30   0.29   0.32   0.31   0.37   0.47   0.48   0.47   0.52   0.73   0.70   0.75   0.67   0.89   0.89   0.88   0.85   1.00   1.00   1.00   1.00       13                               12.5   14.0   15.5   17.0   RTi  (CTU  Post  Hatch)           0   0   0   0   0.09   0.05   0.09   0.07   0.15   0.20   0.19   0.21   0.22   0.27   0.25   0.29   0.29   0.32   0.31   0.39   0.47   0.48   0.47   0.50   0.73   0.70   0.75   0.64   0.91   0.89   0.88   0.82   1.00   1.00   1.00   1.00   Table  3.  Relationship  between  developmental  stages  (y)  and  transition  time  to  stage   (hours  and  CTU  post  fertilization;  x)  as  a  function  of  incubation  temperature   following  the  model  y  =  mx  +  b,  where  m  is  the  slope  and  b  is  the  intercept.       Temperature  (°C)     12.5   14.0   15.5   17.0   R2   m   b   Hours  Post  Hatch   0.020   36.897   0.917   0.027   36.738   0.943   0.038   36.630   0.938   0.043   36.817   0.953                   R2   b   CTU  Post  Hatch   0.0735   36.573   0.912   0.0794   37.314   0.939   0.0763   36.677   0.612   0.0983   36.889   0.953   m     DISCUSSION   Embryonic  development  involves  a  complex  progression  of  cell   differentiation  and  proliferation  that  is  regulated  by  temperature  (Ojangunen  and   Brana  2003).    The  results  presented  here  demonstrate  increased  rearing   temperatures  hasten  larval  development,  consumption  of  yolk-­‐sac,  and  the  onset  of   first  feeding.    The  observed  effects  of  warming  were  similar  to  other  studies   examining  larval  development  among  sturgeon  (A.  fulvescens,  Wang  et  al  1985;  A.   brevirostrum,  A.  oxyrhynchus,  Hardy  and  Litvak  2004;  A.  medirostris,  Van  Eenennaam   et  al.  2005).    Developmental  rate  was  independent  of  temperature  when  measured   in  CTU.    Therefore,  CTU  may  be  a  more  reliable  predictor  than  time  (hours)  of  the   timing  of  transitions  between  life  phases  such  as  hatching,  larval  drift  behavior  and   onset  of  exogenous  feeding.    Additionally,  RTi  was  independent  of  the  temperature   conditions  applied.    The  use  of  RTi  in  yolk-­‐sac  larval  stages  has  been  used  to  scale   ontogenetic  development,  independently  of  temperature  (Klimogianni  et  al.  2004).       14   Results  presented  here  validate  the  consistency  of  RTi  in  white  sturgeon  yolk-­‐sac   larvae  when  exposed  to  varying  temperature  conditions.       Management  application:  estimating  fertilization  date  of  wild  caught  yolk-­sac  larvae     The  ability  of  managers  to  estimate  spawning  days  can  inform  management   of  populations  of  conservational  concern.    Data  can  be  used  to  estimate  adult   spawning  activity  including  duration  of  the  reproductive  season,  and  of  individual   spawning  days  and  responses  to  reproductive  cues  such  as  water  temperatures  and   flow  regimes.    For  species  inhabiting  large  river  systems,  direct  observation  of   mating  can  be  difficult  if  not  impossible.    Estimating  fertilization  dates  of  wild-­‐ caught  yolk-­‐sac  larvae  is  indicative  of  adult  reproductive  ecology  and  can  be   estimated  through  back-­‐calculation  from  the  date  and  time  of  sample  collection   based  on  temperature-­‐dependent  data  developed  from  this  study.    Additionally,  if   larval  collection  is  conducted  to  supply  stocking  programs,  targeting  wild  caught   larvae  representing  a  range  of  ages  for  rearing  increasing  the  probability  of  stocking   offspring  representing  a  large  component  of  the  spawning  population.                 Duration  and  timing  of  adult  spawning  activity  can  provide  insight  into   spawning  behavior.  However,  caution  is  advised  when  making  assumptions   concerning  adult  spawner  abundance  based  on  the  number  of  spawning  days.    Given   the  promiscuous  mating  system  of  white  sturgeon,  number  of  spawning  days  is  not   indicative  of  number  of  individual  spawners  or  adult  sex  ratios  (Chapter  II).     Additionally,  managers  should  not  assume  that  individuals  contributed  to  only  one   spawning  day.    Data  presented  in  Chapter  II  indicated  adults  of  both  sexes  spawn     15   over  multiple  days.    This  behavior  has  also  been  seen  in  lake  sturgeon  (Acipenser   fulvescens)  (Duong  et  al.  2013).         16   CHAPTER  II:  ESTIMATES  OF  EFFECTIVE  NUMBER  OF  BREEDING  ADULTS  AND   REPRODUCTIVE  SUCCESS  FOR  WHITE  STURGEON  ACIPENSER  TRANSMONTANUS       ABSTRACT     Accurate  estimates  of  the  number  of  spawning  adults  (Ns),  effective  breeding   number  (Nb),  and  estimates  of  individual  adult  contributions  to  recruitment  are   required  for  recovery  planning  for  endangered  White  Sturgeon  Acipenser   transmontanus  populations,  many  of  which  are  suffering  from  prolonged  periods  of   recruitment  failure.  We  show  that  genetic  techniques  can  be  used  to  characterize   important  features  of  White  Sturgeon  reproductive  ecology  in  large  rivers  where   census  data  are  extremely  difficult  to  obtain.    We  used  microsatellite  loci  (n=12)  and   likelihood-­‐based  pedigree  analysis  to  estimate  Ns,  Nb,  number  of  kin  groups  (Nk)   and  individual  reproductive  success  of  White  Sturgeon  contributing  to  viable  eggs   and  larvae  collected  in  the  Upper  Columbia  River  in  each  of  two  years.    Estimates  of   mean  annual  Ns,  Nb,  and  Nk  (mean  ±  SD)  were  121.5  ±  34.7,  86.5  ±  10.6,  and  73.5  ±   17.3,  respectively.    Large  variation  in  estimates  of  Ns,  Nb,  and  Nk  was  observed   between  spawning  areas  (n=3),  with  one  spawning  site  representing  61%  of  total   adult  spawning  population.    Variation  in  adult  reproductive  success  was  observed   within  and  among  sites.    Individual  spawning  duration  (1.9  ±  1.1  days)  and  number   of  mates  per  adult  (2.9  ±  2.5)  likewise  varied  spatially  and  temporally.    Based  on  age   of  collected  eggs  and  larvae,  number  of  spawning  days  ranged  from  5  to  19  between     17   years  and  among  sites.    Genetically  derived  estimates  of  annual  adult  spawners   were  lower  but  generally  concordant  with  empirical  estimates  of  available   spawners  (Nc)  based  on  sex  ratios  and  maturation  staging  of  adults  captured   independently  (Ns/Nc  ratio  =  0.683).    Results  increase  our  understanding  of  White   Sturgeon  reproductive  ecology  and  recruitment,  and  allow  projections  of  cohort   levels  of  genetic  diversity.    Similar  data  can  be  applied  to  recovery  planning  and   aquaculture  programs  for  this  and  other  species  of  conservation  concern.               18   INTRODUCTION   Large  inter-­‐annual  variability  in  levels  of  recruitment  and  number  of  adults   contributing  offspring  can  have  considerable  effects  on  genetic  diversity  and   population  abundance.    Reliable  data  on  the  actual  number  of  spawning  adults  (Ns)   in  large  river  systems  is  important  for  population  management  but  can  be  labor-­‐ intensive  and  difficult  to  obtain.    Indices  of  relative  abundance  (e.g.,  catch  per  unit   effort;  CPUE)  or  estimates  of  spawning  population  census  size  derived  using   traditional  means  (e.g.  capture-­‐mark-­‐recapture;  Pledger  et  al.  2013)  are  likely  made   with  high  or  unknown  levels  of  uncertainty,  in  part  because  of  ineffectiveness  of   sampling  gear  and  considerable  uncertainty  over  timing  and  duration  of  occupancy   of  spawning  sites  relative  to  when  surveys  are  conducted.    Therefore,  to  more   effectively  address  conservation  and  restoration  needs,  genetic  data  have  become  a   standard  component  of  many  recovery  programs  (Pemberton  2008;  Anders  et  al.   2011).       Genetic  techniques  allow  biologists  to  examine  aspects  of  recruitment   immediately  after  spawning  (Wirgin  et  al.  1997;  Duong  et  al.  2011),  facilitating   estimates  of  spawner-­‐recruitment  relationships  to  be  made  during  early  life  history   stages  when  data  can  be  interpreted  based  on  environmental  conditions.    Using  a   combination  of  statistical  and  genetic  techniques,  it  is  increasingly  easy  to  use   genetically-­‐based  parentage  or  pedigree  methods  (Blouin  2003;  Wang  2004;  Jones   et  al.  2010)  to  estimate  the  number  of  parents  consistent  with  production  of   offspring  using  genetic  markers  of  disomic  inheritance  (Duong  et  al.  2011).       19   However,  most  population  genetic  analyses  developed  for  diploid  species  are   inapplicable  to  species  like  Sturgeon  that  are  polyploid.    Rodzen  et  al.  (2004)   proposed  a  simple  and  general  solution  to  convert  the  microsatellite  genotypes  of   polyploids  to  diploid  dominant  genotypes.    Wang  and  Scribner  (2013)  have  shown,   using  simulated  and  empirical  data,  that  this  method  can  be  used  to  reconstruct   pedigree  and  parentage  relationships  with  a  modest  number  of  polymorphic  non-­‐ disomic  loci.     Genetic  diversity  is  important  for  long-­‐term  population  viability  by  providing   greater  potential  for  adaptation  to  environmental  change  (Reed  and  Frankham   2003).    Effective  population  size  (Ne)  is  defined  as  the  number  of  individuals  in  an   ideal  population  having  the  same  magnitude  of  random  genetic  drift,  inbreeding  or   loss  of  heterozygosity  as  the  actual  population  (Wright  1931).    Factors  contributing   to  the  reduction  in  Ne  include  variations  in  lifetime  reproductive  success,  skewed   sex  ratios  and  fluctuations  in  population  size  over  time  (Frankham  1995;   Charlesworth  2009;  Waples  2010).    Additionally,  fragmentation  and  isolation  of   habitat  and  species’  ecological  characteristics  including  attributes  of  the  mating   system,  generation  length,  and  inbreeding  interval  influences  Ne  (Waples  1990).     Accordingly,  Ne  reflects  the  effects  of  evolutionary  processes  on  population  levels  of   genetic  diversity.    Therefore  estimates  of  Ne  and  annual  recruitment  are  important   variables  to  understand  for  endangered  species  (Charlesworth  2009;  Waples  2010),     20   as  small  populations  are  at  risk  of  extinction  through  demographic  stochasicity,   genetic  drift  and  environmental  variation  (Braude  &  Low  2010).       The  effective  breeding  number  (Nb)  represents  a  measure  of  effective  size  for   a  single  reproductive  season.      This  metric  is  important  for  understanding  the   ecological  dynamics  within  a  spawning  season  and  is  similarly  influenced  by  factors   affecting  general  estimates  of  Ne  (Waples  2002).    Thus,  Nb  can  vary  among  years,   especially  when  inter-­‐annual  fluctuations  in  environmental  conditions  affect   recruitment  (Myers  1998).    For  example,  if  few  adults  produce  a  large  proportion  of   annual  progeny,  the  ratio  between  Nb  and  actual  spawning  population  size  (Ns)  can   be  low.    For  semelparous  species,  the  relationship  between  Nb  and  Ne  is  well   established  (Nunney  2002;  Waples  2002).    However,  for  long-­‐lived,  iteroparous   species,  such  as  Sturgeon,  Ne  is  difficult  to  estimate  because  Ne  depends  on  variance   in  lifetime  reproductive  success  among  adults  in  the  population  (Hill  1972).     Accordingly,  there  has  been  considerable  interest  in  estimating  Nb.  (e.g.,  Duong  et  al.   2013).       White  Sturgeon  are  a  native  fish  of  western  North  America  and  have   undergone  significant  reductions  in  population  abundance  and  distribution  due  to   over-­‐exploitation,  environmental  degradation  and  loss  of  habitat  connectivity   (Billard  and  Leconintre  2001).    Despite  documentation  of  annual  spawning  activity   over  the  past  20  years,  the  Upper  Columbia  River  (UCR)  population  is  listed  as   endangered  and  has  been  experiencing  recruitment  failure  over  the  past  several     21   decades  (Hildebrand  and  Parsley  2013).    To  date,  estimating  the  fertilization  date  of   collected  eggs  has  provided  the  sole  measure  of  spawning  activity  and  is  currently   the  best  available  metric  of  spawning  duration  within  the  UCR.    However,  despite   annual  spawning  activity  (e.g.,  based  on  egg  and  larval  captures)  being  detected  for   this  population,  little  information  is  available  regarding  the  number  of  annual   breeders,  including  estimates  of  individual  reproductive  success,  at  known  or   suspected  spawning  sites  within  the  UCR.  This  information  is  critical  in  the   development  of  recovery  strategies  (Fisheries  and  Oceans  2014)  and  when   measuring  progress  towards  recovery.       The  main  objective  of  this  study  was  to  estimate  the  annual  UCR  White   Sturgeon  Nb  and  Ns.    Given  the  large  size  of  the  UCR,  the  highly  dispersed  nature  of   known  or  suspected  spawning  areas,  and  low  population  abundance,  we   hypothesized  that  the  number  of  adults  spawning  at  each  spawning  site  would  be  a   small  proportion  to  the  entire  spawning  population.    An  additional  objective  was  to   characterize  aspects  of  the  species’  reproductive  ecology  including  individual  adult   reproductive  success,  spawning  duration,  mate  number,  and  spawning  group   composition.  We  also  examined  the  efficacy  of  using  empirical  data  regarding  the   sexual  maturity  stage  of  adult  White  Sturgeon  to  predict  the  number  of  contributing   adults  to  an  annual  spawning  season.      We  hypothesized  that  independent  estimates   of  spawning  population  size  derived  with  the  use  of  both  genetic  (Ns)  and  empirical   (Nc)  data  would  be  concordant  and  therefore  would  highlight  the  importance  of   employing  multiple  methods  to  determine  annual  spawning  population  size  when     22   possible.    Additionally,  results  could  determine  if  one  technique  may  be  substituted   for  estimating  spawning  population  size  if  the  other  is  not  available  to  researchers.       METHODS   Study  Area    We  monitored  spawning  of  White  Sturgeon  during  2  consecutive  years   (2011-­‐2012)  within  a  57  km  reach  of  the  most  downstream  main-­‐stem   impoundment  of  the  Canadian  section  of  the  UCR  between  Hugh  L.  Keenleyside  Dam   (HLK)  at  river  kilometer  (rkm)  0.0  to  the  USA-­‐Canada  border  (rkm  57)  in  British   Columbia,  Canada  (Figure  1).    The  population  of  White  Sturgeon  in  this  section  of   the  Columbia  River  is  estimated  at  1,157  (95%  C.I.  414-­‐1900;  Irvine  et  al.  2007)   with  an  additional  2,003  sturgeon  estimated  (95%  C.I.  1093-­‐3223)  to  reside  south   of  the  USA  border  (Howell  and  McLellan  2007).    Within  the  UCR,  fidelity  to  specific   locations  is  high  (>65%;  BC  Hydro  2013)  though  demographic  patterns  (McAdam   2012)  and  movement  data  (Howell  and  McLellan  2007)  suggest  movement  between   habitats  in  the  Canadian  and  USA  areas  occurs  (Figure  1).    This  section  of  the  UCR  is   highly  regulated  by  hydro-­‐electric  generation  and  storage  dams  controlling  flows   from  three  major  rivers  including  the  Columbia  River,  the  Kootenay  River,  and  the   Pend  D’Orielle  River.    This  altered  hydrograph  is  one  of  several  potential  reasons  for   recruitment  failure  (Gregory  and  Long  2008).         23     Figure  1.  Transboundary  Reach  of  the  Upper  Columbia  River  between  Hugh  L.   Keenleyside  Dam  (HLK),  British  Columbia,  Canada,  and  Gifford,  Washington,  USA.     Egg  mat  and  drift  nets  were  deployed  in  2011  and  2012  within  in  the  Canadian   Portion  of  the  Upper  Columbia  River.    Sampling  sites  include:  (A)  Arrow  Lakes   Generating  Station  (ALH,  river  kilometer  0.1  (rkm),  including  HLK),  (B)  18.2  rkm,   and  (C)  Waneta  (56.0  rkm).       24   Within  the  Canadian  portion  of  the  Columbia  River,  White  Sturgeon   reproduction  occurs  from  mid-­‐June  through  August  at  two  known  spawning  sites,   downstream  of  Arrow  Lakes  Generating  Station  (ALH)  and  Waneta  Dam  (Waneta)   (Figure  1).    ALH  (rkm  0.1)  is  located  beside  HLK  and  regulates  flow  from  the  Arrow   Lakes  Reservoir  to  meet  requirements  of  the  Columbia  River  Treaty.    Waneta  (rkm   56.0)  regulates  flow  of  the  Pend  d’Orielle  River  into  the  UCR.    At  ALH,  sampling  was   conducted  downstream  of  both  the  ALH  tailraces  and  immediately  downstream  of   the  HLK  spillways.    Sampling  at  Waneta  was  conducted  downstream  of  the  tailrace.     The  known  geographical  boundaries  of  the  ALH  and  Waneta  spawning  areas  based   on  egg  and  larval  captures  are  small,  covering  approximately  0.1  and  0.16  km2,   respectively.    Further  details  regarding  the  spawning  areas  and  their  geographical   boundaries  are  available  in  Terraquatic  (2011)  and  Golder  (2009)  for  ALH  and   Waneta,  respectively.    An  additional  site  located  at  18.2  rkm  (18.2  rkm)  was  also   sampled  (Figure  1)  based  on  i)  suitable  spawning  habitat  (e.g.  substrate  and  water   velocities),  ii)  egg  and  larval  collection  at  this  site  in  previous  years,  and  iii)   identified  movements  of  mature  adult  white  sturgeon  to  this  area  during  the   spawning  period  (BC  Hydro  2013).    The  exact  geographical  boundaries  of  where   sturgeon  are  spawning  near  site  18.2  rkm  are  unknown  but  the  sampling  location   represents  the  downstream  extent  of  where  spawning  may  occur.    White  Sturgeon   from  the  UCR  are  also  known  to  spawn  annually  at  two  locations  south  of  the  USA   border  (Howell  and  McLellan  2007),  though  these  locations  were  not  monitored  in   this  study.    Though  inter-­‐annual  exchange  between  different  spawning  areas  is   unknown,  adults  making  spawning  related  movements  within  the  Canadian  section     25   of  the  UCR  tend  to  remain  within  the  specific  river  section  they  were  residing  within   (e.g.  adults  residing  within  10  km  of  18.2  tend  to  spawn  in  this  area)  and  this  has   been  repeatable  across  multiple  years  (details  in  Hildebrand  and  Parsley  2013).     Though  some  work  has  suggested  that  White  Sturgeon  in  the  Canadian  portion  of   the  Columbia  River  may  have  historically  had  population  substructuring  (Nelson   and  McAdam  2012),  other  work  looking  at  current  levels  of  genetic  diversity,  found   that  white  sturgeon  in  the  transboundary  reach  were  not  genetically  different  from   downstream  populations  (Drauch  Schreier  et  al.  2013)  and  additional  investigations   of  historic  genetic  population  structure  are  ongoing.   Spawning  in  this  population  typically  occurs  when  water  temperatures   exceed  14oC  and  UCR  flows  are  on  a  descending  pattern  (Hildebrand  et  al.  1999;  BC   Hydro  2013).  Further,  UCR  adults  cannot  be  observed  congregating  to  spawn  due  to   the  water  depth  (6.0  ±  3.0  m;  mean  ±  SD)  and  relatively  high  flow  volume  (daily   average  510.0  to  1754.4  m3.s-­‐1  during  the  spawning  season).    Therefore  spawning   was  documented  through  the  collection  of  progeny.   Egg  and  larval  collection   White  Sturgeon  are  broadcast  spawners  exhibiting  a  promiscuous  and   aggregate  mating  system  where  gametes  from  multiple  females,  fertilized  by   multiple  males,  are  dispersed  over  large  sections  of  rivers  (Billard  and  Leconintre   2001).    Therefore,  passive  techniques  including  egg  mats  and  drift  nets  have  been   successfully  used  in  the  past  to  collect  the  demersal  eggs  and  drifting  yolk-­‐sac  larvae   (McCabe  and  Beckman  1990;  Parsley  et  al.  1993).    Egg  mats  consisted  of  a  0.76  m  by   0.91  m  steel  frame  enclosing  latex  coated  animal  hair  filter  material.    Drift  nets     26   consisted  of  a  1.3  m  diameter  stainless  steel  frame  (D-­‐shaped)  with  a  0.6  m  by  0.8  m   opening  and  4  m  tapered  plankton  net  (0.16  cm  delta  mesh  size)  with  a  removable   collection  cup  attached  to  the  frame.    Sampling  gear  was  deployed  on  top  of   substrate  (cobble  to  boulders,  all  sites)  on  the  river  bottom  using  a  fixed  anchor   system  that  remained  constant  throughout  the  sampling  period.    Egg  mats  (ALH   2011,  n=5;  Waneta  2011,  n=7;  Waneta  2012,  n=12)  were  placed  at  the  site  of   spawning  activity  and  drift  nets  (ALH  2011,  n=8;  18.2rkm  2011,  n=4;  Waneta  2011,   n=1;  ALH  2012,  n=8;  18.2rkm  2012,  n=2;  Waneta  2012,  n=2)  were  set  immediately   downstream  from  the  spawning  site.       Sampling  was  conducted  from  mid-­‐June  to  mid-­‐August.    Due  to  river   hydrology,  sampling  gear  and  effort  varied  among  monitoring  sites.    Egg  mats  were   deployed  only  at  the  known  spawning  sites  of  Waneta  and  ALH  for  24-­‐hour  sets.     Drift  nets  were  deployed  at  all  sites  for  24  hours  excluding  Waneta,  where  nets  were   set  for  only  3  hours  due  to  time  constraints  and  river  hydrology.    Sampling  location   within  a  site  remained  consistent  across  years  (Figure  1).    The  period  of  sampling   coincided  with  typical  thermal  regimes  (12-­‐  18°C)  documented  during  White   Sturgeon  spawning  activity  (Hildebrand  et  al.  1999;  Parsley  and  Beckman  1994;   Paragamian  et  al.  2001;  Perrin  et  al.  2003).    Due  to  warmer  temperature  influences   of  the  Pend  D’Oreille  River,  sampling  at  Waneta  was  initiated  on  June  13  and  June  11   during  the  2011  and  2012  sampling  seasons,  respectively.    Sampling  at  ALH  and   18.2  rkm  commenced  July  11  in  2011  and  was  delayed  to  July  26  in  2012  due  to  high   water  flows  that  prevented  gear  deployment.    Sampling  was  terminated  at  all  sites   on  August  17  and  16  in  2011  and  2012,  respectively.         27   Live  eggs  collected  from  the  river  were  placed  in  incubation  trays  until  hatch   to  obtain  a  sufficient  genetic  tissue  sample.    Incubation  trays  were  suspended  in  the   UCR  3  m  below  the  water  surface  in  a  stacked  configuration  on  a  30  kg  anchor   system  downstream  from  each  spawning  site  in  areas  of  lower  flow  (<  1  m/s).     Incubation  trays  consisted  of  a  middle  plate  of  plexiglass  (180  mm  x  200  mm  x  6   mm)  with  100  perforations  (6  mm  in  diameter)  distributed  in  a  rectangular  grid   (10x10).    Two  similarly  perforated  plexiglass  plates  (180  mm  x  180  mm  x  3  mm),   with  1  mm  plastic  screen  secured  to  one  side,  were  placed  on  either  side  of  the   middle  plate  to  enclose  the  eggs  within  the  incubator.    Incubation  trays  remained  in   the  river,  in  situ,  until  all  yolk-­‐sac  larvae  successfully  hatched  within  a  tray.       Upon  hatch,  yolk-­‐sac  larvae  were  euthanized  with  an  overdose  of  tricaine   methanesulfonate  (MS-­‐222)  and  preserved  in  95%  ethanol.    Captured  drifting  larvae   were  also  euthanized  and  preserved  in  95%  ethanol.    Due  to  the  large  number  of   eggs  and  larvae  collected  at  Waneta,  a  random  subsample  of  eggs  (~20%)  was   preserved  in  Prefer  (solution  of  glyoxcal,  buffer  and  alcohol)  for  assignment  of   developmental  stage  (see  below),  a  random  subsample  of  eggs  (~20%)  was   incubated  to  obtain  genetic  tissue,  and  a  random  subsample  of  captured  drifting   larvae  (~20%)  was  euthanized  and  preserved  in  95%  ethanol.   Water  temperatures  were  recorded  hourly  throughout  the  period  of   sampling  using  VEMCO  Minilog-­‐II-­‐T  data  loggers  placed  at  all  sampling  sites.    Data   loggers  were  also  paired  with  each  incubation  station.       28   Developmental  staging  and  estimation  of  fertilization  date   Preserved  eggs  and  larvae  were  randomly  examined  with  respect  to  date,   stage,  and  site  (to  reduce  observer  bias)  using  a  digital  compound  microscope   (Nikon  SMZ-­‐745t  Stereo  Microscope  with  10X  eyepiece)  and  assigned  a   developmental  stage.    Enumeration  of  stages  corresponded  to  the  classification  by   Dettlaff  et  al.  (1993),  including  embryonic  stages  (1  through  35;  fertilization  to  pre-­‐ hatch)  and  yolk-­‐sac  larval  stages  (36  through  45;  hatch  to  exogenous  feeding).    Each   developmental  stage  was  associated  with  the  appearance  of  at  least  one  new  feature   therefore  stages  were  not  determined  strictly  by  quantitative  changes.    No   preserved  samples  had  developed  beyond  stage  45.       Fertilization  date  for  collected  eggs  and  larvae  was  estimated  by  back-­‐ calculation  from  the  recorded  date  and  time  of  preservation  based  on   developmental  stage  (eggs,  Parsley,  U.S.  Geological  Survey,  unpublished;  yolk-­‐sac   larvae,  K.  Jay  unpublished),  and  mean  incubation  water  temperature.    The  estimated   age  (hours)  was  subtracted  from  the  preservation  date  and  time  to  determine  the   estimated  date  and  time  of  fertilization  (i.e.  spawning  date).    Calculated  fertilization   dates  provided  an  estimation  of  spawning  duration  for  each  spawning  site.    The   poor  condition  of  collected  samples  at  18.2  rkm  prohibited  the  estimation  of   fertilization  date;  therefore  estimated  spawning  duration  was  only  calculated  for   ALH  and  Waneta  sampling  sites.   Genetic  analysis   The  total  number  of  2011  larvae  collected  was  subsampled  for  genotyping   due  to  disparity  in  total  sample  sizes  between  spawning  sites.    In  order  of  capture     29   date,  every  other  Waneta  larval  sample  (~50%)  and  every  fourth  ALH  larvae   (~25%)  were  genotyped.    Due  to  very  low  numbers  (n  =  33),  all  larvae  collected  at   18.2  rkm  (100%)  were  selected  for  genotyping.    All  larvae  collected  in  2012  were   genotyped.    DNA  was  extracted  from  larval  tissue  samples  using  QIAGEN  DNeasy®   kits  (QIAGEN  Inc.)  according  to  manufacturers’  protocols.    DNA  was  quantified  using   a  Nanodrop  spectrophotometer  and  all  samples  were  diluted  to  a  constant   concentration  (20  ng/µl)  for  use  in  Polymerase  Chain  Reactions  (PCR).    Individuals   were  genotyped  using  12  microsatellite  loci  including  AciG-­35,  AciG-­2,  AciG-­53,  AciG-­ 140  (Bork  et  al.  2008),  Atr-­105  (Drauch  and  May  2007),  Atr-­107,  Atr-­109,  Atr-­117,   Atr-­1101,  Atr-­1173,  Atr-­100  and  Atr-­113  (Rodzen  and  May  2002;  Drauch  and  May   2007).    PCR  reactions  were  conducted  to  amplify  100  ng  DNA  in  25  μl  reaction   mixtures  containing  2.5  μl  of  10X  PCR  Buffer  (0.1  M  Tris-­‐HCl,  15  mM  MgCl2,  0.5  M   KCl,  0.1%  gelatin,  0.1%  NP-­‐40,  0.1%  Triton-­‐X);  additions  of  1  μl  MgCl2  (25  mM)  (0.5   μl  MgCl2  for  Atr-­109;  1.5  μl  MgCl2  for  Atr-­107;  2  μl  MgCl2  for  Atr-­1101)  for  all   reactions  excluding  Atr-­105  and  Atr-­117;  2.5  μl  deoxynucleotide  triphosphates   (dNTPs;  0.8  mmol!L-­‐1);  1  μl  of  fluorescently  labeled  forward  and  unlabeled  reverse   primers  (10  pmol/μl)  and  one  unit  of  Taq  DNA  polymerase  (5U/μl).       All  PCR  reactions  were  conducted  using  a  Robocycler  96  thermal  cycler   (Stratogene).    The  PCR  conditions  were  94°C  for  2  min,  followed  by  35  cycles  (33   cycles  for  Atr-­109;  37  cycles  for  Atr-­107  and  AciG-­2)  of  1  min  at  94°C,  1  min  for   primer-­‐specific  annealing  temperatures  (55°C  for  AciG-­35,  AciG-­53;  56°C  for  Atr-­100,   Atr-­105,  Atr-­107,  Atr-­109,  Atr-­113,  Atr-­117,  Atr-­1101,  Atr-­1173;  57°C  for  AciG-­2;  58°C   for  AciG-­140),  72°  for  2  min,  and  a  final  extension  for  5  min  at  72°C  (excluding  Atr-­   30   105,  Atr-­107,  Atr-­109,  Atr-­117,  Atr-­1173  and  AciG-­35).    PCR  products  were  run  on  6%   denaturing  polyacrylamide  gels  and  genotypes  were  visualized  using  a  Hitachi   FMBIO  II  scanner.    Allele  sizes  were  determined  using  commercially  available  size   standards  (MapMarkerTM,  BioVentures  Inc.)  and  based  on  several  standard   samples  of  known  genotype.    To  minimize  error,  all  genotypes  were  independently   scored  by  two  experienced  lab  personnel  and  verified  again  after  data  were  entered   into  electronic  databases.    Errors  in  genotyping  were  empirically  checked  by  blindly   re-­‐genotyping  a  random  10%  of  all  samples  within  a  year.    Reported  genotyping   error  was  the  ratio  between  observed  number  of  allelic-­‐errors  and  total  number  of   alleles  compared  (Bonin  et  al.  2004).   Data  Conversion   Due  to  the  polyploid  nature  of  the  White  Sturgeon  genome  (Blacklidge  and   Bidwell  1993),  microsatellite  alleles  were  treated  as  dominant  data.    Following   Rodzen  et  al.  (2004),  each  individual  phenotype  was  converted  into  a  1  x  n  vector,   where  n  is  the  number  of  bands  at  the  locus.    Each  band  at  a  given  microsatellite   locus  was  indexed  as  1  if  the  band  was  present  (dominant)  or  indexed  as  0  if  the   band  was  absent  (recessive)  within  an  individuals  phenotype.    Therefore,  an   individual  phenotype  showing  bands  1,  3  and  7  at  an  8-­‐band  microsatellite  locus   was  converted  to  a  1  x  n  vector  of  [1,  0,  1,  0,  0,  0,  1,  0],  yielding  eight  dominant   markers.    This  process  was  repeated  for  each  microsatellite  locus,  and  data  were   combined  to  produce  a  1  x  nT  vector,  where  nT  is  the  total  number  of  bands  across   all  microsatellite  loci  for  each  individual.           31   Pedigree  analysis   Pedigree  analyses  were  conducted  using  COLONY  v2.0.4.0  (Jones  and  Wang   2010).    This  software  uses  a  maximum  likelihood  method  to  estimate  pedigree   relationships  among  offspring  by  identifying  networks  of  full-­‐sibling  and  half-­‐sibling   families  using  their  multi-­‐locus  genotypes  of  dominate  data  while  incorporating   genotyping  errors.    Within  a  random  sample  of  individuals  (with  respect  to  kin),  the   frequencies  of  full  and  half  sib  dyads  were  used  to  estimate  Nb  of  the  population   (see  Wang  2009).    For  each  analysis,  male  and  female  polygamy  was  assumed.    The   full-­‐likelihood  method  was  used  with  most  of  the  default  parameter  settings  (e.g.   dioecious,  diploid,  no  inbreeding,  a  single  run  of  medium  length,  medium  likelihood   precision,  no  sibship  prior,  no  update  of  allele  frequencies).    The  full-­‐likelidhood   method  assigns  all  sampled  individuals  to  various  inferred  relationships  (full-­‐ siblings,  half-­‐siblings,  unrelated)  jointly  and  is  the  most  accurate  COLONY  method  as   verified  by  simulated  and  empirical  data  analyses  (Wang  2004;  Wang  and  Santure   2009).    Due  to  sampling  methods  (collection  of  offspring  of  unknown  parentage)   COLONY  parameters  of  sibship  size,  number  of  parent  candidates,  paternal  sibships,   maternal  sibships  and  population  allele  frequency  were  unknown  or  zero.    The   value  for  allelic  dropout  rate  was  set  at  0  for  all  loci  across  both  years.    The  values   for  rate  of  other  kinds  of  genotyping  errors  (including  mistyping  and  mutations)   were  0.008  and  0.003  for  the  2011  and  2012  data,  respectively.    These  error  rates   were  determined  through  the  reanalysis  of  a  ~10%  random  subset  of  individuals   per  year.  The  transformation  of  polyploidy  codominant  phenotypes  to  diploid   dominant  phenotpyes  could  cause  an  apparent  distortion  of  Mendeliam  segregation,     32   and  thus  may  lead  to  the  split  of  large  full  sib  families  in  the  likelihood   reconstruction  (Wang  and  Scribner  2013).  However,  when  genotyping  error  rates   are  permitted  at  each  locus,  rare  phenotypes  will  be  considered  to  be  due  to   genotyping  errors,  and  the  large  sibship  will  not  be  incorrectly  split  during  pedigree   reconstruction  (Wang  and  Scribner  2013).    To  test  whether  higher  genotyping  error   rates,  relative  to  empirically  calculated  values,  resulted  in  differences  in  sibship   reconstruction,  additional  genotyping  error  rates  (0.02  and  0.04;  as  suggested  by   Wang  and  Scribner  (2013))  were  used  for  the  2011  ALH  data.    To  test  whether   subsampling  biased  results  by  underestimating  Ns,  Nb,  and  Nk,  additional  analyses   were  conducted  for  the  2011  ALH  and  Waneta  data.    Of  the  total  samples  collected,  a   subsample  of  the  Waneta  (25%  and  12%)  and  ALH  (12%  and  6%)  samples  were   systematically  selected  (i.e.,  every  fourth  sample  collected  in  order  of  capture  data   to  represent  25%  of  total  capture)  and  analyzed  in  COLONY.               Pedigree  analyses  were  conducted  separately  for  each  sampling  site  and  year   to  estimate  Nb,  Ns  and  number  of  kin  groups  (Nk)  assuming  random  mating.    Sex  of   inferred  adults  was  unknown.    For  each  analysis,  a  replicate  run  was  conducted   using  the  same  data  and  parameter  values  but  different  random  number  seeds   allowing  for  the  comparison  of  maximum  likelihood  estimates  and  best  pedigree   configurations  as  well  as  to  evaluate  program  convergence  for  each  data  set.    An   analysis  of  variance  (ANOVA)  was  used  to  test  for  differences  in  Nk  between  sites   and  years.     33   With  the  combination  of  the  estimated  fertilization  dates  and  genetic   analyses  data,  the  proportion  of  full-­‐sibling  individuals  estimated  to  be  fertilized   within  24  and  48  hours  of  each  other  was  calculated.    Additionally,  the  duration  of   spawning  per  inferred  adult,  the  mean  and  standard  deviation  of  number  of   contributing  adults  per  estimated  spawning  date,  number  of  spawning  partners  per   inferred  adult,  and  number  of  estimated  spawning  days  an  inferred  adult   contributed  progeny  to  were  calculated.   Comparisons  of  genetic  and  empirical  estimates  of  mature  adult  population   Estimates  of  spawning  population  size  derived  with  the  use  of  genetic  and   empirical  data  were  compared.    Spawning  population  size  based  on  genetic  data  was   estimated  as  described  above.    Empirical  estimates  of  the  Canadian  UCR  population   in  spawning  condition  each  year  was  calculated  based  on  visual  determination  of   maturation  stages  of  adults  captured  during  the  pre-­‐spawning  season  (early  to  mid-­‐ June)  and  corresponding  sex  ratios.      In  2009  through  2012,  adult  White  Sturgeon  were  captured  within  the   Canadian  portion  of  the  UCR  using  baited  (frozen  kokanee)  setlines  in  a  medium  line   configuration  (ten  20/0  circle  hooks)  (for  capture  and  processing  see  BC  Hydro   2013).    Setline  sampling  sites  were  spatially  balanced  and  randomly  selected   throughout  the  entire  spatial  extent  of  the  UCR  from  0.0  rkm  to  56.0  rkm  (Figure  1).     Each  adult  (>150cm)  was  surgically  examined  to  determine  sex  and  maturity  stage.       Sexually  mature  males  were  indicated  with  having  large,  cream  to  whitish  testes   that  were  deeply  lobed  and  filling  most  of  the  abdominal  cavity  (UCWSRI  2006).     Sexually  mature  females  were  identified  with  having  large  dark  ovaries  filling  much     34   of  the  abdominal  cavity.    Black  eggs  contained  tight  in  the  ovary  exhibited  a  distinct   “bulls-­‐eye”  with  a  diameter  greater  than  3  mm  (UCWSRI  2006).    The  number  of   available  annual  spawners  (Nc)  and  proportion  of  the  total  Canadian  UCR   population  (1,157  [414,  1889;  95%  CI];  Irvine  et  al.  2007;  1:1  sex  ratio)  in  spawning   condition  were  estimated  based  on  the  proportion  of  total  sexually  mature  adults   captured.    This  method  of  estimation  was  then  compared  to  inferred  Ns  determined   by  pedigree  analysis  using  the  2011  genetic  data.    Comparisons  were  conducted   only  using  the  2011  genetic  data,  since  not  all  sampling  sites  were  represented  by   the  2012  genetic  data.   RESULTS   Egg  and  larval  collection  and  estimation  of  fertilization  date   Egg  and  larval  collection  in  2011  extended  from  July  4  to  August  6  at  Waneta   (n=466),  August  2  to  August  12  at  ALH  (n=417)  and  July  26  to  August  9  at  18.2  rkm   (n=33).    Water  temperatures  increased  throughout  the  sampling  period  ranging   from  10.8°C  to  19.8°C  (Figure  2a).    Based  on  back  calculating  fertilization  dates  from   developmental  stages  of  collected  eggs  and  larvae,  spawning  at  Waneta  was   estimated  to  have  occurred  from  June  30  through  August  3  with  19  spawning  days   at  water  temperatures  ranging  from  11.8°C  to  18.1°C  (Figure  2a).    Spawning  was   estimated  to  have  occurred  at  ALH  over  a  duration  of  5  days  from  August  1  to   August  5  when  water  temperatures  were  14.8°C  to  16.1°C  (Figure  2a).    Spawning   activity  was  multimodal  at  Waneta  with  three  distinct  peaks  while  one  estimated   spawning  peak  was  found  at  ALH.       35     A.  180   Total  samples  collected   160   140   120   100   ALH:  Larvae   ALH:  Eggs   Waneta:  Larvae   Waneta:  Eggs   80   60   40   20   0   30-­‐Jun   07-­‐Jul   14-­‐Jul   21-­‐Jul   28-­‐Jul   04-­‐Aug   Es